Epigenetics - PDF Free Download (2024)

EPIGENETICS -- -ncRNA

EDITED

C. David Allis The Rockefeller University, New York

Thomas Jenuwein Research Institute of Molecular Pathology (IMP), Vienna

Danny Reinberg HHMIIRobert Wood Johnson Medical School University of Medicine and Dentistry ofNew Jersey

Marie-Laure Caparros Associate Editor, London

COLD SPRING HARBOR LABORATORY PRESS Cold Spring Harbor, New York

•

http://www.cshlpress.com

Epigenetic Mechanisms That Operate in Different Model Organisms

s. cerevisiae s. pombe

N. crassa

C. elegans

Drosophila

Mammals

A. thaliana

14 Mb

40 Mb

100 Mb

180 Mb

3,400 Mb

150 Mb

GENOMIC FEATURES

Genome size

12 Mb 6,000

5,000

10,000

20,000

14,000

-25,000

25,000

1.45 kb

1.7 kb

2 kb

5 kb

35-46 kb

2 kb

Average number of introns/gene

,,;1

6-8

4-5

% Genome as protein coding

1-1.5 (Hs)

Number of genes Average size of genes

EPIGENETIC FEATURES

Histone acetylation

H3K4 methylation

H3K36 methylation

H3K79 methylation

H3.3 histone variant

ON/OFF

SWI/SNF ATPase complexes

+ (+)'

CHD1 ATPase family

+ (+)'

(+)'

SWR1 ATPase

(+)'

+ (+)'

ON/OFF

ISWI ATPase

ON/OFF

IN080 ATPase

MI-2 ATPase

+ (+)'

OFF

+ (+)'

OFF

CENP-A centromeric histone variant

OFF

H3K9 methylation b

OFF

HP1-like proteins

OFF

RNA interference

OFF

H4K20 methylation'

OFF

H3K27 methylation

+ (+)" +9

OFF

Polycomb repressive complexes

OFF

DNA methylation

OFF

DNA methylation binding proteins

OFF

Imprinting

+ +'

Abbreviation: (Hs) hom*o sopiens. , Epigenetic feature considered to be present based on sequence hom*ology but no functional data. b There is evidence that H3K9 methylation is found at active chromatin regions; however, the functional significance of this is unknown. , H4K20 tri-methylation is not present in S. cerevisiae, whereas all three H4K20 methylation states are present in multicellular organisms. d Drosophila possess very low levels of DNA methylation. , Mutated Dnmt2. Dnmt2 (Pp) and MBD-domain proteins (Ce, Cb, Pp). Dnmt2 and MBD-domain proteins (Dm). hChromosome- or genome-wide rather than gene-specific.

EPIGENETICS All rights reserved. © 2007 by Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York Printed in the United States of America Publisher Acquisition Editors Development Director Project Coordinator Permissions Coordinator Production Editor Desktop Editor Production Manager Cover Designer

John Inglis Alexander Gann and David Crotty Jan Argentine Inez Sialiano Carol Brown Pat Barker Lauren Heller Denise Weiss Lewis Agrell

Front cover artwork: Depicted is a schematic representation of the chromatin template. Epigenetic regulation affects and modulates this template through noncoding RNAs (ncRNA) that associate with it, covalent modification of histone tails (mod), methylation of DNA (Me), remodeling factors (blue oval), and nucleosomes that contain standard as well as variant histone proteins (the yellow nucleosome). In the background is a representation of several model organisms in which epigenetic control has been studied. From top left: Pair of mouse chromosomes that may differ in their genomic imprint; a S. cerevisiae colony, showing epigenetically inherited variegation of gene expression; anatomy of C. elegans; illustration of T. thermophila, showing the large "active" macronucleus and the smaller "silent" micronucleus; D. melanogaster; maize section with kernel color variegation; Arabidopsis flower. Library of Congress Cataloging-in-Publication Data Epigenetics / edited by C. David Allis, Thomas Jenuwein, Danny Reinberg ; Marie-Laure Caparros, associate editor. p.cm. Includes bibliographical references and index. ISBN-13: 978-0-87969-724-2 (hardcover: alk. paper) 1. Genetic regulation. 1. Allis, C. David. II. Jenuwein, Thomas. III. Reinberg, Danny. [DNLM: 1. Epigenesis, Genetic. 2. Gene Expression Regulation. QU 475 E64 2006] 1. Title. QH450.E655 2006 572.8'65--dc22 2006028894

10 9 8 7 6 5 4 3 2 Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Cold Spring Harbor Laboratory Press, provided that the appropriate fee is paid directly to the Copyright Clearance Center (CCC). Write or call CCC at 222 Rosewood Drive, Danvers, MA 01923 (978-750-8400) for information about fees and regulations. Prior to photocopying items for educational classroom use, contact CCC at the above address. Additional information on CCC can be obtained at CCC Online at http://www.copyright.com/. All Cold Spring Harbor Laboratory Press publications may be ordered directly from Cold Spring Harbor Laboratory Press, 500 Sunnyside Blvd., Woodbury, New York 11797-2924. Phone: 1-800-843-4388 in Continental U.S. and Canada. All other locations: (516) 422-4100. FAX: (516) 422-4097. E-mail: [emailprotected]. For a complete catalog of all Cold Spring Harbor Laboratory Press publications, visit our World Wide Web Site http://www.cshlpress.com/.

Long before epigenetics changed from little more than a diverse collection of bizarre phenomena to a well-respected field covered by its own textbook, a talented group of foresighted molecular biologists laid a rich foundation upon which the modern era of chromatin biology and epigenetics is based. This group includes Vince Allfrey, Wolfram Harz, Hal Weintraub, Alan Wolffe, and Abe Worcel. This book is dedicated to their collective memory. Their passion and commitment to the study of chromatin biology inspired all of us who followed their work, and we now profit from their many insights.

Contents

Preface, ix

Epigenetics: From Phenomenon to Field, 1

Epigenetic Regulation of Chromosome Inheritance, 265 Gary H. Karpen and R. Scott Hawley

Daniel E. Gottschling

A Brief History of Epigenetics, 15

Epigenetic Regulation of the X Chromosomes in C. elegam, 291

Gary Felsenfeld

Susan Strome and William G. Kelly

Overview and Concepts, 23

C. David Allis, Thomas Jenuwein, and Danny Reinberg 4

Epigenetics in Saccharomyces cerevisiae, 63

Michael Grunstein and Susan M. Gasser

Position-Effect Variegation, Heterochromatin Formation, and Gene Silencing in Drosophila, 81 Sarah c.R. Elgin and Gunter Reuter

DNA Methylation in Mammals, 341 En Li and Adrian Bird

Fungal Models for Epigenetic Research: Schizosaccharomyces pombe and Neurospora crassa, 101 Robin C. Allshire and Eric U. Selker

Dosage Compensation in Mammals, 321 Neil Brockdorff and Bryan M. Turner

19 6

Dosage Compensation in Drosophila, 307 John C. Lucchesi and Mitzi I. Kuroda

Genomic Imprinting in Mammals, 357 Denise P. Barlow and Marisa S. Bartolomei

Germ Line and Pluripotent Stem Cells, 377 M. Azim Surani and Wolf Reik

Epigenetics of Ciliates, 127 Eric Meyer and Douglas 1. Chalker

Epigenetic Control of Lymphopoiesis, 397 Meinrad Busslinger and Alexander Tarakhovsky

RNAi and Heterochromatin Assembly, 151 Robert Martienssen and Danesh Moazed

Epigenetic Regulation in Plants, 167

Nuclear Transplantation and the Reprogramming of the Genome, 415 RudolfJaenisch and John Gurdon

Marjori Matzke and Ortrun Mittelsten Scheid

23 10

Chromatin Modifications and Their Mechanism of Action, 191 Tony Kouzarides and Shelley 1. Berger

Epigenetics and Human Disease, 435 Huda Y. Zoghbi and Arthur 1. Beaudet

Epigenetic Determinants of Cancer, 457 Stephen B. Baylin and Peter A. Jones

Transcriptional Silencing by Polycomb Group Proteins, 211 Ueli Grossniklaus and Renato Paro

Transcriptional Regulation by Trithorax Group Proteins, 231 Robert E. Kingston and John ltv. Tamkun

Appendices WWW Resources, 477

Histone Modifications and References, 479

Histone Variants and Epigenetics, 249 Steven Henikoff and M. Mitchell Smith

Index, 491 vii

Preface

his advanced textbook on "Epigenetics" is truly a reflection of many talented colleagues and individuals, all of whom made this book possible and a rewarding experience. However, without hesitation, the editors want to thank Marie-Laure Caparros (London), without whom this project would have never materialized. Early in the process, it became evident that the editorial team needed help in coordinating such a large project, particularly for keeping the dialogue and editorial feedback with the >40 colleagues who agreed to provide outstanding chapter contributions, only to realize that we wanted more than their expert reviews and attention to detail. Marie-Laure has been instrumental in keeping the momentum moving forward, has bravely exchanged critical comments when needed, has informed all of us on the many deadlines, and has provided necessary coherence to make embryonic chapters come to life. Without her, this book would not have been possible. We are also grateful to our individual assistants, who forever kept us on our toes: Elizabeth Conley (David Allis), Christopher Robinson (Thomas Jenuwein), and Shelli Altman (Danny Reinberg). All of them are the unsung heroes of this book. We thank all of them for their innumerable contributions, large and small, and their unending patience with each of us and our quirky styles and shortcomings as editors. Discussions for such a book took initial form on the coattails of the outstanding 69th Cold Spring Harbor Symposium on Epigenetics in the summer of 2004, but were seeded in early 2003 and formally commissioned by CSHL Press through Alex Gann and other colleagues. This was followed by formulating an editorial team between David Allis, Thomas Jenuwein, and Danny Reinberg. The first concrete outline for this project, including the brainstorming of various chapters and potential contributing authors, was done on the picnic bench at the FASEB meeting on Chromatin and Transcription in Snow-

mass, Colorado, July 2003. We were then very fortunate to confirm the lineup of contributing colleagues who are the leaders in their field. In the early planning stages, a vision crystallized for a different concept. Ideally, we sought to ask not for a compilation of expert reviews which might soon be outdated. Rather, we wanted to compile a set of conceptual chapters, from pairs of experts, that highlight important discoveries for students in chromatin biology and for colleagues outside the epigenetics field. In keeping to a conceptual outline, we hoped to have a more long-lasting impact. Also, by including many diagrams and illuminating figures, and appendices, we hoped to list most of the systems and epigenetic marks currently known. The General Summaries were aimed as a stand-alone precis of the topics covered in each chapter, preceded by "teaser" images to entice the reader to investigate. The figures have been another important hallmark for this book; particularly, the examples provided in the Overview and Concepts chapter. Here, Stefan Kubicek, a Ph.D. student from the Jenuwein lab at the IMP (Vienna), and Marie-Laure Caparros have been the masters of the diagrams. They honed draft upon draft of figures (sometimes only from sketches) for the chapters, such that we could gain a more coherent presentation. Several postdocs and Ph.D. students (Gabriella Farkas, Fatima Santos, Heike Wohrmann, and others) in the labs of several authors also kindly contributed to the excellent illustrations in this book. However, we were unable to convert all of the contributions, and some figures have remained as submitted. We are also particularly grateful to Monika Lachner, Mario Richter, Roopsha Sengupta, Patrick Trojer, and other Ph.D. students and Postdocs in the Allis, Jenuwein, and Reinberg laboratories for amending, proofreading, and finalizing the tables and summaries that are displayed in the appendices. Here, Dr. Steven Gray (St. James Hospital, Dublin) has been

PREFACE

particularly instrumental in validating and providing additional information for the table that lists all the currently known histone modifications. Where appropriate, submitted chapters were sent out for comments from other colleagues who provided important input for streamlining and clarifying some of the complex concepts. Not all of this input could be converted into the revised and final versions, but the comments and suggestions helped to shape many of the chapters and the overall framework of the book. Here, we are indebted to G. Almouzni, P. Becker, H. Cedar, V. Chandler, W. Dean, R. Feil, A. Ferguson-Smith, M. Gartenberg, S. Grewal, M. Hampsey, E. Heard, R. Metzenberg, V. Pirrotta, F. Santos, T. Schedl, D. Solter, R. Sternglanz, S. Tilghman, and others. Finally, we acknowledge the intellectual and, in some cases, emotional contributions made by all of our colleagues in the field who provided the chapters to make this book what it is. Their contributions, by way of writ-

ten chapters and drawings, stand by themselves. But what may not be obvious is the feedback and cross-fertilization that all of them had with the editorial team to help shape and guide the book as it took form. The Overview and Concepts chapter itself reflects their feedback, as in early drafts, we put too much of our own colors and bias into the sentences. For their wisdom and for bringing us a deeper perspective and balance, we thank them, and we admit that any deficiencies and mistakes there are ours. Financial support for this book has come from CSHL Press (New York), the Epigenome FP6 NoE (European Union), IMP (Vienna), the Rockefeller University (New York), and the Howard Hughes Medical SchoolRobert Wood Johnson Medical School (Piscataway, N~w Jersey). Critical contributions were also made by Upstate Serologicals (Lake Placid, New York) and AbCam (Cambridge, UK), leading suppliers of epigenetic-based reagents and tools. CDA, TJ,DR

APT

Epigenetics: From Phenomenon to Field Daniel E. Gottschling Fred Hutchinson Cancer Research Center, Seattle, Washington 98109

CONTENTS 1. Introduction, 2

3.4

Prions, 9

2. A History Of:1P1 netics at Cold Spring Harbor Symposia, 2

3.5

New Phenomenon, 70

4. Closing Thoughts, 10

3. The 69th Sympo . m, 8 3.7 .The Histone Code Hypothesis, 8

Acknowledgments, 11

3.2

Dynamic Silent Chromatin, 8

References, 11

3.3

Nuclear Organization, 9

2 •

CHAPTER

1 Introduction

In the summer of 2004, the 69th Cold Spring Harbor Symposium on Quantitative Biology covered the topic of "Epigenetics;' and many of the authors of this book were in attendance. As an observer at this Symposium, I knew this was going to be an interesting meeting. It started simply enough by trying to define epigenetics. After a week of querying participants about this, it became clear that such a request was akin to asking someone to define "family values"-everyone knew what it meant, but it had a different meaning for each person. Part of the reason for the range of opinions may be understood from the etymology of "epigenetics" as explained by David Haig: The word had two distinct origins in the biological literature in the past century, and the meaning has continued to evolve. Waddington first coined the term for the study of "causal mechanisms" by which "the genes of the genotype bring about phenotypic effects" (see Haig 2004). Later, Nanney used it to explain his realization that cells with the same genotype could have different phenotypes that persisted for many generations. I define an epigenetic phenomenon as a change in phenotype that is heritable but does not involve DNA mutation. Furthermore, the change in phenotype must be switch-like, "ON" or "OFF;' rather than a graded response, and it must be heritable even if the initial conditions that caused the switch disappear. Thus, I consider epigenetic phenomena to include the lambda bacteriophage switch between lysis and lysogeny (Ptashne 2004), pili switching in uropathogenic Escherichia coli (Hernday et al. 2003), position-effect variegation in Drosophila (Henikoff 1990), heritable changes in cortical patterning of Tetrahymena (Frankel 1990), prion diseases (Wickner et al. 2004a), and X-chromosome inactivation (Lyon 1993). The 69th Symposium came on the 100th anniversary of genetics as a field of study at Cold Spring Harbor Laboratory, making it very timely to consider epigenetics. Given this historical context, I thought it appropriate to provide an examination of epigenetics through the portal of previous Cold Spring Harbor Symposia. Although the 69th Symposium was the first dedicated to the topic, epigenetic phenomena and their study have been presented throughout the history of this distinguished series. The history I present is narrowed further by my limitations and likings. For a more complete and scholarly portrayal, I can recommend the more than 1000 reviews on epigenetics that have been written in the past five years. In presenting this chronological account, I hope to convey a sense of how a collection of apparently disparate

phenomena coalesced into a field of study that affects all areas of biology, and that th~tudy of epigenetics is founded upon trying to explain the unexpected-perhaps more than any other field of biological research. 2 A History of Epigenetics at Cold Spring Harbor Symposia

In 1941 during the 9th Symposium, the great Drosophila geneticist H.I. Muller described developments on his original "eversporting displacement," in which gross chromosomal rearrangements resulted in the mutant mosaic expression of genes near the breakpoint (Muller 1941). By the time of this meeting, he referred to it as "position effect variegation." It was well established that the affected genes had been transferred "into the neighborhood of a heterochromatic region;' that the transferred euchromatic regions had been "partly, but variably, transformed into a heterochromatic condition-'heterochromatized'," and that addition of extra copies of heterochromatic chromosomes "allowed the affected gene to become more normal in its functioning." This latter observation was an unexpected quandary at the time, which we now know to be the result of a titration of limiting heterochromatin components. At the 16th Symposium (1951), a detailed understanding of the gene was of high priority. This may explain why little progress had been made on understanding position-effect variegation (PEV), although more examples were being discovered. However, the opening speaker noted that PEV would be an exciting area for future research (Goldschmidt 1951). Barbara McClintock noted that chromosomal position effects were the basis of differencesfn'~ableloci" of maize, and she speculated that the variatiof of mutability she observed likely had its roots in the same mechanisms underlying PEV in Drosophila (McClintock 1951). By the time of the 21st Symposium, McClintock's ideas about "controlling elements" had developed (McClintock 1956). Two were particularly poignant with regard to epigenetics. In the Spm controlling element system, she had uncovered variants that allowed her to distinguish between trans-acting factors that could "suppress" a gene (reduce or eliminate its phenotypic expression) rather than mutate it. She also noted that some controlling elements could suppress gene action not only at the locus where it had inserted, but also at loci that were located some distance on either side of it. Others were discovering this "spreading effect" as well. J. Schultz presented a biochemical and physical characterization of whole Drosophila that contained

EPIGENETICS:

different amounts of heterochromatin (Schultz 1956). Although the work was quite primitive and the conclusions drawn were limited, the work represented early attempts to dissect the structure of heterochromatin and demonstrated just how difficult the problem would be. Two talks at the 23rd Symposium were landmarks with respect to our present-day Symposium. First, R.A. Brink described his stunning observations of "paramutation" at the R locus in maize. If two alleles (R sl and R') with distinct phenotypes as hom*ozygotes are combined to form a heterozygote, and this RsI/R' plant is in turn crossed again, the resulting progeny that contain the Rr allele will always have an Rsl phenotype, even though the Rsl is no longer present (Brink 1958). However, this phenotype is metastable-in subsequent crosses the phenotype reverts to the normal R' phenotype. He meant for the word paramutation "to be applied in this context in its literal sense, as referring to a phenomenon distinct from, but not wholly unlike, mutation." Second, D.L. Nanney went to great lengths to articulate "conceptual and operational distinctions between genetic and epigenetic systems" (Nanney 1958). In essence, he defined epigenetics differently from how it had been originally intended by Waddington (for details, see Haig 2004). He found it necessary to do so in order to describe phenomena he observed in Tetrahymena. He found evidence that the cytoplasmic history of conjugating parental cells influenced the mating-type determination of resulting progeny. His definition encompassed observations made by others as well, including Brink's work on the R locus and McClintock's work noted in the 21st Symposium. Mary Lyon's recently proposed hypothesis of X~chro mosome inactivation in female mammals (Lyon 1961) was of considerable interest at the 29th Symposium. S. Gartler, E. Beutler, and W.E. Nance presented further experimental evidence in support of it (Beutler 1964; Gartler and Linder 1964; Nance 1964). Beutler reviewed multiple examples of mosaic expression of X-linked genes in women, supporting the random nature of X inactivation. From careful quantitative analysis of an X-linked gene product, Nance deduced that X inactivation occurred before the 32-cell stage of the embryo. The 38th Symposium on "Chromosome Structure and Function" represented a return to examining eukaryotic chromosomes-significant progress had been made studying prokaryotic and phage systems, and consequently, bacterial gene expression had dominated much of the thinking in the burgeoning field of molecular biology. However, an appreciation for chromatin (DNA with histones and nonhistone proteins) in eukaryotes was building, but it was unclear whether it played a role in chromosome structure

FROM

PHENOMENON

FIELD.

or function, or both (Swift 1974). Nevertheless, several groups began to speculate that posttranslational modification of chromatin proteins, including histones, was associated with gene transcription or overall chromosome structure (Allfrey et al. 1974; Louie et al. 1974; Weintraub 1974). There was only a hint of epigenetic phenomena in the air. It had been hypothesized that repetitive DNA regulated most genes in eukaryotes, partly based on the fact that McClintock's controlling elements were repeated in the genome. However, it was reported that most repeated DNA sequences were unlinked to genes (Peaco*ck et al. 1974; Rudkin and Tartof 1974). From these observations, the idea that repeated elements regulated gene expression lost significant support from those in attendance. More importantly, however, these same studies discovered that most of the repetitive DNA was located in heterochromatin. The 42nd Symposium demonstrated that in four years, an amazing number of technical and intellectual advances had transformed the study of eukaryotic chromosomes (Chambon 1978). This included the use of DNA restriction enzymes, development of recombinant DNA technology, routine separation of proteins and nucleic acids, the ability to perform Southern and northern analysis, rapid DNA and RNA sequencing, and immunofluorescence on chromosomes. The nucleosome hypothesis had been introduced, and mRNA splicing had been discovered. Biochemical and cytological differences in chromatin structure, especially between actively transcribed and inactive genes, comprised the primary interest at this meeting. However, most relevant to epigenetics, Hal Weintraub and colleagues presented ideas about how chromatin could impart variegated gene expression in an organism (Weintraub et al. 1978). The 45th Symposium was a celebration of Barbara McClintock's discoveries-~~le genetic elements (Yarmolinsky 1981). Mechanistic) studies of bacterial transposition had made enormous progress and justifiably represented about half the presentations, whereas others presented evidence that transposition and regulated genomic reorganization occurred not only in maize, but also in other eukaryotes-including flies, snapdragons, Trypanosomes, Ascobolus, and budding yeast. In the context of this meeting, all observed variegated expression events were ascribed to transposition. Moreover, there was a reticence to seriously consider that controlling elements were responsible for most gene regulation (Campbell 1981), which led some to suggest that "the sole function of these elements is to promote genetic variability." In essence, the idea that heterochromatin was responsible for the regulated expression in position-effect

4 • CHAPTER

variegation was called into question. With respect to future epigenetic studies, perhaps the most noteworthy discussion was the firm establishment of "silent mating cassettes" in Saccharomyces cerevisiae (Haber et al. 1981; Klar et al. 1981; Nasmyth et al. 1981; Rine et al. 1981). Leading up to the 47th Symposium, a general correlation had been established in vertebrate systems that the overall level of cytosine methylation in CpG DNA sequences was lower for genes that were transcribed than for those that were not. However, there were exceptions to this generalization, and more detailed analysis was presented that methylation of a specific area of a gene's promoter was most important (Cedar et al. 1983; Doerfler et al. 1983; La Volpe et al. 1983). On the basis of restriction/modification systems of bacteria, it was thought that DNA methylation prevented binding of key regulatory proteins. Furthermore, it had been shown that DNA methylation patterns could be mitotically inherited in vertebrates, which led to the hypothesis that DNA methylation could serve as a means of transcriptional "memory" as cells divided through development (Shapiro and Mohandas 1983). Another major epigenetic-related finding was the identification of DNA sequences on either side of the "silent mating cassettes" in budding yeast that were responsible for transcriptional repression of genes within the cassettes-these defined the first DNA sequences required for chromosomal position effects (Abraham et al. 1983). "The Molecular Biology of Development" was the topic for the 50th Symposium, and it too encompassed a number of important advances. Perhaps one of the most exciting developments was the overall awareness that fundamental molecular properties were conserved throughout evolution-e.g., human RAS functioned in budding yeast, homeo box proteins were conserved between flies and humans (Rubin 1985). New efforts to understand chromosome imprinting began with the development of nuclear transfer in mice (Solter et al. 1985). These studies revealed that parent-of-origin information was stored within the paternal and maternal genomes of a new zygote; it was not just the DNA that was important, but the chromosomes contained additional information about which parent they had passed through, and the information was required for successful development of an embryo. Part of the answer was thought to lie in the fact that differential gene expression was dependent' on the parental origin of a chromosome (Cattanach and Kirk 1985). There were a number of studies aimed at understanding the complex regulation of the bithorax complex, but

notably, E.B. Lewis made special mention of the curious nature of known trans regulators of the locus; nearly all were repressors of the locus (Lewis 1985). Thus, maintaining a gene in a silenced state for many cell doublings was imperative for normal development. This contrasted with much of the thinking at the time-that gene activation/induction was where the critical regulatory decisions of development would be. DNA transformation and insertional mutagenesis techniques had recently been achieved for a number of organisms. One particularly creative and epigeneticrelated use of this technology came in Drosophila. A P-element transposon with the white eye-color gene on it was created and "hopped" throughout the genome (Rubin et al. 1985). This provided a means to map sites throughout the Drosophila genome where PEV could occur. This meeting also highlighted the first genetic approaches to dissecting sex determination and sex chromosome dosage compensation in Drosophila (Belote et al. 1985; Maine et al. 1985) and Caenorhabditis elegans (Hodgkin et al. 1985; Wood et al. 1985). The 58th Symposium highlighted the celebration of the 40th anniversary of Watson and Crick's discovery. Part of the celebration was a coming-out party for epigenetic phenomena: There was identification of new phenomena, beginnings of molecular analysis of other phenomena, and sufficient progress had been made in a number of systems to propose hypotheses and to test them. In trypanosomes, the family of Variable Surface antigen Genes (VSG) located near telomeres are largely silenced, with only one VSG expressed at a time. Although this organism does not appear to contain methylated DNA, it was re~orted that the silenced VSG genes contained a novel minor base: ~-D-glucosylhydroxymethyl uracil (Borst et al.)1993). This base appeared to be in place of thymidine-irrthe DNA. Parallels between this base and cytosine methylation in other organisms were easy to draw-the modifications were important for maintaining a silenced gene. But how the base was introduced into the DNA, or how it imparted such a function, was unclear. Progress had also been made in vertebrate epigenetic phenomena, including chromosomal imprinting and X inactivation (Ariel et al. 1993; Li et al. 1993; Tilghman et al. 1993; Willard et al. 1993). It had become clear by this time that numerous loci were subject to imprinting in mammals; only one allele was expressed in diploid cells, and expression was dependent on parental origin. The Igf2-H19 locus was of particular interest, primarily because it contained two nearby genes that were regulated in opposing fashion. Igf2 is expressed from the paternal

EPIGENETICS:

chromosome while the maternal copy is repressed, whereas the paternal allele of H19 is repressed and its maternal allele is expressed. Interestingly, methylated CpG was observed just upstream of both genes on the paternal chromosome. It was proposed that the differential methylation regulated access of the two genes to a nearby enhancer element-the enhancer was closer to, and just downstream of, H19 (Tilghman et al. 1993). A mutually exclusive competition between the two genes for the enhancer was envisioned; when the H19 gene was methylated, the enhancer was free to activate the more distant Igf2 gene. Support for the idea that DNA methylation played a regulatory role in this process came from mouse studies. Mutation of the first vertebrate gene encoding a S-methyl-cytosine DNA methyltransferase in ES cells showed that as embryos developed, the paternal copy of H19 became hypomethylated and the gene became transcriptionally active (Li et al. 1993). An important step in the way in which sMeCpG mediated its effects came from the purification of the first sMeCpG DNA-binding complex (MeCP1) (Bird 1993). Not only did it bind DNA, but when tethered upstream of a reporter gene, MeCP1 caused the gene to be repressed. Although this did not explain regulation at the Igf2-H19 locus, it did provide a potential mechanism to explain the general correlation between DNA methylation and gene repression. Genetic mapping over a number of years had identified a portion of the human X chromosome as being critical for imparting X inactivation. Molecular cloning studies of this X-inactivation center led to the discovery of the Xist gene (Willard et al. 1993), an -17-kb noncoding RNA that was expressed only on the inactive X chromosome. The mouse version of Xist was surprisingly hom*ologous in st~~re and sequence and held the promise of being an excelllnt model system to dissect the way in which this RtlAJunctioned to repress most of the X chromosome. Two notable findings were described in Neurospora (Selker et al. 1993). First, it was shown that cytosine DNA methylation was not limited to epG dinucleotides but could occur in seemingly any DNA context. Second was the amazing description of the phenomenon of repeatinduced point mutation (RIP). Sequences become "RIP'd" when there isa sequence duplication (linked or unlinked) in a haploid genome and the genome is put through the sexual cycle via conjugation. Two events occur: Both copies of the duplicated DNA pick up G:C ----7 A:T mutations, and DNA within a few hundred base pairs of the RIP'd sequences becomes methylated. This double attack on the genome is quite efficient-SO% of unlinked

FROM

PHENOMENON

FIELD

loci succumb to RIP, whereas tightly linked loci approach lOO%-and readily abolishes gene function. The brown gene in Drosophila, when translocated near heterochromatin, displays dominant PEV; the translocated copy can cause repression of the wild-type copy. In searching for enhancers and suppressors of this transinactivation phenomenon, Henikoff discovered that duplication of the gene located near heterochromatin increased the level of repression on the normal copy (Martin-Morris et al. 1993). Although the mechanism underlying this event remained mysterious, it was postulated that the phenomenon might be similar to RIP in Neurospora, although it had to occur in the absence of DNA methylation, which does not occur in Drosophila. Paul Schedl elucidated the concept of chromosomal "boundary elements" (Vazquez et al. 1993). The first were located on either side of the "puff" region at a heat shock locus in Drosophila and were defined by their unusual chromatin structure-an -300-bp nuclease-resistant core bordered by nuclease hypersensitive sites. It was postulated that such elements separated chromatin domains along the chromosome. Two in vivo assays supported this hypothesis: (1) When bordering either side of a reported gene, boundary elements effectively eliminated chromosomal position effects when the construct was inserted randomly throughout the genome. (2) The boundary element was also defined by its ability to block enhancer function. When inserted between a gene promoter and its enhancer, the boundary element blocked the gene's expression. Although not as well defined, the concept of boundary elements was also developing in other organisms, especially at the globin locus in mammals (Clark et al. 1993). Budding yeast shone the light on a mechanistic inroad to chromatin-related epigenetic phenomena. It had already been established that the silencers at the silent mating-type loci were sites for several DNA-binding proteins. Their binding appeared to be context-dependent, as exemplified by the Rap1 protein, which not only was important in silencing, but also bound upstream of a number of genes to activate transcription (for review, see Laurenson and Rine 1992). Over the years, numerous links had been made between DNA replication and transcriptionally quiescent regions of the genome. The inactive X chromosome, heterochromatin, and silenced imprinted loci had all been reported to replicate late in S phase relative to transcriptionally active regions of the genome. In addition, it had been shown that the establishment of silencing at the silent mating-type loci required passage through S phase, suggesting that silent.chromatin had to be built on newly repli-

6 •

CHAPTER

cated DNA. Thus, there was great interest when one of the Furthermore, overexpression for Sir3 caused it to silencers was found to be an origin of DNA replication, and "spread" inward along the chromatin fiber from the its origin activity could not be separated from silencing telomere, suggesting that it was a limiting component of function (Fox et al. 1993). Furthermore, mutants in the silent chromatin and could "polymerize" along the chrorecently identified origin recognition complex (ORC) were matin (Renauld et al. 1993). Taken together, there found to cripple silencing (Bell et al. 1993; Fox et al. 1993). appeared to be a large interaction network important for The discovery that telomeres in Saccharomyces ceresilencing-the Sir proteins initiated assembly at telomeric visiae exerted PEV, just like that seen in Drosophila, DNA, due to their interaction with Rap1, and then polybrought another entree into dissecting heterochromatic merized from the telomere along the chromatin fiber, structure and its influence on gene expression. Reporter presumably by binding to the tails of histones H3 and H4. genes inserted near telomeres give variegated expression Switching between transcriptional states in variegated in a colony. The repressed state is dependent on many of telomeric expression appeared to be the result of a compethe same genes (SIR2, SIR3, SIR4) as those required for tition between silent and active gene expression (Aparicio silencing at the silent mating-type loci. Several key aspects and Gottschling 1994; described in Weintraub 1993). If the about the silent chromatin structure and the regulation of transcriptional activator for a telomeric gene was deleted, the variegated expression were described. It is worth notthe gene's basal transcriptional machinery was insufficient ing that heterochromatin is defined cytologically as confor expression and the gene was constitutively silenced. Conversely, overexpression of the activator caused the densed chromatin, but silent chromatin in S. cerevisiae has never been visualized in this way. Nevertheless, telomeric gene to be expressed continuously-the gene was because of similarities to PEV in Drosophila, there was never silenced. In the absence of SIR3 (or SIR2 or SIR4), enthusiasm to consider silent chromatin in yeast to be a basal gene expression was sufficient, whereas increased functional equivalent of heterochromatin (described in dosage of SIR3 increased the fraction of cells that were Weintraub 1993). silenced. Although a transcriptional activator could overFrom the yeast studies, a number of fundamental concome silencing throughout the cell cycle, it was most effeccepts began to come to light. First, the importance of histive when cells were arrested in S phase, presumably when tone H3 and H4 became evident. In particular, the chromatin was being replicated and, hence, most susceptiamino-terminal tail of histones H3 and H4 appeared to ble to competition. Somewhat surprisingly, cells arrested in be directly involved in the formation of silent heterochroG/M also could be easily switched, suggesting that silent matin (Thompson et al. 1993). Specific mutants in the chromatin had not yet been fully assembled by this time. tails of these histones alleviated or crippled silencing and Silent chromatin in yeast was shown to be recalciled to the notion that both the net charge of the residues trant to nucleases and DNA modification enzymes, sugon the tails and specific residues within the tails congesting that the underlying DNA was much less tributed to silencing. In addition, these early days of chroaccessible relative to most of the genome (described in matin immunoprecipitation (ChIP) demonstrated that Thompson et al. 1993). the lysines in the amino-terminal tail of histone H4 were It also appeared that there was a hierarchy of silencing hypoacetylated in regions of silent chromatin relative to within the yeast genome: The telomeres were the most the rest of the genome. Moreover, one of the histone sensitive to perturbation, HML was next, and HMR was mutants identified histone H4 K16, which could be acety-~e least sensitive. In fact, when the SIR] gene was lated, as critical for forming silent chromatin. mutated, the normally completely silenced HM loci disTelomeres appeared to provide the simplest system in played variegated expression (Pillus and Rine 1989). which to develop an understanding of how Sir proteins Finally, Sir3 and Sir4 were localized to the nuclear mediated silencing. The concept of recruiting silencing periphery, as were the telomeres. It was proposed that proteins was being developed. Briefly, the telomeric DNAthe nucleus was organized such that the nuclear envebinding protein, Rap1, was found to interact with Sir3 lope provided a special environment for silencing (Paland Sir4 by two-hybrid methods (described in Palladino ladino et al. 1993). et al. 1993). Thus, Rap1 could "recruit" these Sir proteins Schizosaccharomyces pombe also has silent mating casto the telomeric region of the genome. There was evisettes that were suspected to behave similarly to those in dence that Sir3 and Sir4 could bind to one another, and S. cerevisiae. However, in S. pombe, there was an added most importantly, Sir3 and perhaps Sir4 interacted with twist to the story of mating-type switching. In an elegant the tails of histones H3 and H4 (Thompson et al. 1993). set of experiments, Amar Klar proposed how a "mark" is

EPIGENETICS:

imprinted on one strand of DNA in a cell (Klar and Bonaduce 1993). The mark is manifested, after two cell divisions in one of the four granddaughter cells, as a double-stranded break that facilitates mating-type switching. This yeast does not have any known DNA modifications (methylation, etc.), hence, a different type of mark was postulated to be left on the DNA strand. The topic of the 59th Symposium was "The Molecular Genetics of Cancer." The concept of epigenetic regulation in oncogenesis had begun to develop after the idea of tumor suppressor genes became established. There had been a couple of studies supporting such a notion, but an interesting twist to the story came in studies of Beckwith-Wiedemann syndrome and Wilms' tumor patients. Mutations in both types of patients had been mapped to a locus that included the imprinted H19IGF2 genes. Feinberg et al. (1994) discovered "loss of imprinting" (LOI) for these genes in affected patientsthe maternal locus lost its imprint, H19 was repressed, and IGF2 was expressed. Thus LOI, which in principle could occur elsewhere in the genome, could cause either biallelic expression and/or extinction of genes critical in oncogenesis. In the couple of years leading up to the 63rd Symposium on "Mechanisms of Transcription," several important developments occurred that would affect the molecular understanding of several epigenetic phenomena. Histone-modifying enzymes were identifiedspecifically, histone acetylases and deacetylases. Some of these enzymes played critical roles in regulating gene expression and provided an entry into gene products that directly affected PEV and silencing. The tip of this iceberg was presented at the Symposium (see Losick 1998). Molecular dissection of the Sir3 and Sir4 silencing proteins in yeast revealed the polyvalent nature of their interactions and revealed how the network of interactions between all the Sir proteins, the histones, and various DNA-binding factors set up silent chromatin. In addition, the molecu ar details of how various loci (telomeres, the rDNA, HM loci, and double-stranded breaks) could compete for the limited supply of Sir proteins were shown. By crippling the ability of a specific locus to recruit silencing factors, Sir protein levels were increased at the other loci (co*ckell et al. 1998). This provided direct evidence that principles of mass action were at work and that silencing at one locus could affect the epigenetic silencing at other locian idea originally put forth in studies on PEV in Drosophila, but not yet tested (Locke et al. 1988). Another finding explained how DNA methylation could regulate gene expression through chromatin. This

FROM

PHENOMENON

FIELD.

came with the identification of protein complexes composed of MeCP2, which bind both methylated DNA and histone deacetylases (Wade et al. 1998). Methylated DNA could serve as a point of recruiting deacetylases to a locus and thus facilitate silencing of nearby genes. The concept of boundary elements was extended from Drosophila to mammals, with clear evidence provided at the ~-globin locus, thus indicating that chromatin' boundaries were indeed likely conserved in metazoans and perhaps all eukaryotes (Bell et al. 1998). The 64th Symposium on "Signaling and Gene Expression in the Immune System" provided evidence about how monoallelic expression arose, and that it might be more widespread than previously thought. Monoallelic expression at the immunoglobulin loci had been obvious in lymphocytes for some time-it guaranteed the production of a single receptor type per lymphoid cell (Mostoslavsky et al. 1999). The allele to be expressed was chosen early in development, apparently at random: Both alleles began in a repressed state, but over time one became demethylated. It was unclear how a single allele was chosen, but the phenomenon appeared at other loci, too, where the necessity of monoallelism was not obvious. For instance, only one allele of genes encoding the cytokines IL- 2 and IL-4 was expressed (Pannetier et al. 1999). The most significant epigenetics-related talk at the 65th Symposium concerned the discovery that the Sir2 protein was a histone deacetylase (Imai et al. 2000). This was the only Sir protein that had clear hom*ologs in all other eukaryotes and that regulated PEY. It seemed to be the enzyme primarily responsible for removing acetyl moieties from histones in silent chromatin. Furthermore, because it was an NAD-dependent enzyme, it linked the regulation of silencing (heterochromatin) to cellular physiology. The 68th Symposium on "The Genome of hom*o sapiens" was an important landmark in genetics, and although there is still much genetic work to be done, the complete sequencing of this and other genomes signified that it was time to move "above genetics"-a literal meaning of epigenetics. This historical account highlights several themes shared with many other areas of research. First, it demonstrates the episodic nature of advances in epigenetics. Second, as molecular mechanisms underlying epigenetic phenomena began to be understood, it made it easier to connect epigenetics to biological regulation in general. Third, it showed that people whom we now consider to be scientific luminaries had made these connections early on-it just took a while for most others to "see" the obvious.

8 • CHAPTER

3 The 69th Symposium

A few general principles have been identified over the years that are common to all epigenetic phenomena, and they serve to guide experimental approaches in the search for a detailed understanding. First, the differences between the two phenotypic states ("OFF" and "ON") always have a corresponding difference in structure at a key regulatory point-form translates into function. Hence, identifying the two distinct structures, the components that compose them, and the compositional differences between them have been the primary tasks. Second, the distinct structures must have the ability to be maintained and perpetuated in a milieu of competing factors and entropic forces. Thus, each structure requires selfreinforcement or positive feedback loops which ensure that it is maintained and propagated over many cellular divisions; in some cases, such as X-chromosome inactivation, this appears to be on the order of a lifetime. Many of the mechanistic principles defined in the earlier symposia continued to be refined in the 69th Symposium, but there were also new developments. To put these new developments in context, it is important to note that two other discoveries had a major impact on epigenetics. One was the discovery of RNA interference and related RNA-based mechanisms of regulation. The other was the discovery of mechanisms underlying the prion hypothesis. Both of these fields have advanced rapidly in the past decade, with some of the studies contributing to knowledge about chromatin-based epigenetics and others providing new perspectives about heritable transmission of phenotypes. Many of the accomplishments reported at the Symposium are detailed in the chapters of this book, so I eschew discussing these topics here. However, I will touch upon a few advances that caught my fancy and are not covered within these pages. At the end, I will try to distill the most important concepts I took away from the meeting. 3.1 The Histone Code Hypothesis

In considering histone modifications and their potential information content, there were many discussions about the "histone code hypothesis" (Jenuwein and Allis 2001). Most of those I participated in, or overheard, were informal and rather lively. The proponents of the "code" cite examples such as tri-methylation of histone H3 at K9 and its greater affinity for the HP1 class of heterochromatin proteins (Jenuwein and Allis 2001). Those on the other side cite biochemical and genetic evidence that the net charge on the amino-terminal tail of histone H4, irre-

spective of which position the charge is at, has dramatic effects on DNA binding or phenotype (Megee et al. 1995; Zheng and Hayes 2003). Grunstein presented data that included genome-wide analysis of histone acetylation modifications and chromatin-associated proteins using specific antibodies and ChIP-Chip in S. cerevisiae (Millar et al. 2004). His focus was on the epigenetic switch associated with H4K16 acetylation for binding, or not binding, particular chromatin proteins-thus supporting the histone code hypothesis. Although not discussed, some of his data appeared to support reports from others that for much of the genome, there is no correlation between specific histone modifications and gene expression (i.e., all active genes have the same marks, and these marks are not present on inactive genes) (Schubeler et al. 2004; Dion et al. 2005). Taking all the results together, I suspect that both specific modifications and general net charge effects will be used as mechanisms for regulating chromatin structure and gene expression. 3.2 Dynamic Silent Chromatin

I must confess that, on the basis of static images of heterochromatin and the refractory nature of silent chromatin, I was convinced that once established, a heterochromatic state was as solid as granite. Only when it was time for DNA replication would the impervious structure become relaxed. In thinking this way, I foolishly ignored principles of equilibrium dynamics I had learned in undergraduate chemistry. However, these lessons were brought home again by studies of silent chromatin and heterochromatin, where it was shown that silencing proteins of yeast (Sir3), and heterochromatin proteins in mammalian cells (HP1), were in a dynamic equilibrium-proteins were rapidly exchanged between heterochromatin and the soluble compartment-even when the chromatin was in its most impervious state (Cheng a~artenberg 2000; Cheutin et al. 2003). The realization of its dynamic qualities forced a different view of no.wAn epigenetic chromatin state is maintained and propagated. It suggests that in some systems the epigenetic state can be reversed at any time, not just during DNA replication. Hence, we can infer that mechanisms of reinforcement and propagation for silenced chromatin must function constantly. Methylation of histones was widely held to be the modification that would indeed impart a "permanent" mark on the chromatin (for review, see Kubicek and Jenuwein 2004). In contrast to all other histone modification (e.g.,

EPIGENETfCS:

phosphorylation, acetylation, ubiquitination), there were no enzymes known that could reversibly remove a methyl group from the amine of lysine or arginine. Furthermore, removing the methyl group under physiological conditions by simple hydrolysis was considered thermodynamically disfavored and thus unlikely to occur spontaneously. Those thinking that methylation marks were permanent had their belief system shaken a bit by several reports. First, it was shown that a nuclear peptidylarginine deiminase (PAD4) could eliminate monomethylation from histone H3 at arginine (R) residues (Cuthbert et al. 2004; Wang et al. 2004). Although this methyl removal process results in the arginine residue being converted into citrulline, and hence is not a true reversal of the modification, it nevertheless provided a mechanism for eliminating a permanent methyl mark. Robin Allshire provided a tantalizing genetic argument that the tis2 gene from S. pombe reversed dimethylation on histone H3 at K9 (R. Allshire, pers. comm.). He may have been on the right track, because a few months after the meeting, the unrelated LSD 1 enzyme from mammals was shown to specifically demethylate di- and monomethyl on histone H3 at K4 (Shi et al. 2004), reversing an "active" chromatin mark. Quite interestingly, LSD1 did not work on trimethylated H3K4-thus, methylation could be reversed during the marking process, but reversal was not possible once the mark was fully matured. However, Steve Henikoff presented a way by which a permanent trimethyllysine mark could be eliminated. He showed that the variant histone H3.3 could replace canonical histone H3 in a replication-independent transcriptioncoupled manner (Henikoff et al. 2004). In essence, a histone that contained methyl marks for silencing could be removed and replaced with one that was more conducive to transcription. When total chromatin was isolated, histone H3.3 had many more active chromatin methylation marks (e.g., K79me) on it than canonical histone H3 did. In con~idering all these result~, it s~ems t.h~ the:e may . not be a sImple molecular modIficatIOn Wl~hl hlstones that serves as a memory mark for propagatin the silent chromatin state through cell division. Rath , there must be a more tenuous set of interactions that increase the probability that a silent state will be maintained, although they do not guarantee it. 3.3 Nuclear Organization

Correlations between nuclear location and gene expression have been made for many years (Mirkovitch et al. 1987). These observations began to drive the notion that

FROM

PHENOMENON

FIELD.

there were special compartments within the cell where gene expression or silencing was restricted. It was argued that this organization was necessary to keep the complexity of the genome and its regulation in a workable order. This idea was supported by studies in S. cerevisiae, where telomeres are preferentially located at the nuclear periphery, as are key components of the silencing complex, such as Sir4 (Palladino et al. 1993). Mutations that released the telomeres, or Sir4, from the nuclear periphery resulted in a loss of telomeric silencing (Laroche et al. 1998; Andrulis et al. 2002). Furthermore, artificially tethering a partially silenced gene to the periphery caused it to become fully silenced (Andrulis et al. 1998). In an insightful experiment, Gasser showed that if the teloineres and the silencing complex were both released from the periphery, and free to move throughout the nucleus, telomeric silencing was readily established (Gasser et al. 2004). Thus, there does not appear to be a special need for localizing loci to a compartment. This is more consistent with the findings that rapid movement of chromatin proteins on and off chromosomes can still mediate effective regulation such as silencing. Perhaps some of the localization is necessary to keep high local concentrations of relevant factors under special (stressful?) conditions. Alternatively, this may represent a combination of domains put together through evolution that worked long ago, but had no ultimate purpose. 3.4 Prions

Wickner provided an overview and criteria for defining prions, and from his description it is clear that they are part of the epigenetic landscape (Wickner et al. 2004a,b). In the simplest molecular sense, prions are proteins that can cause heritable phenotypic changes, by acting upon and altering their cognate gene product. No DNA sequence changes occur; rather, the prion typically confers a structural change in its substrate. The beststudied and understood class of prions causes soluble forms of a protein to change into amyloid fibers. In many cases, the amyloid form reduces or abolishes normal activity of the protein, thus producing a change in phenotype. Wickner defined another class of prions that do not form amyloid filaments. These are enzymes that require activation by their own enzymatic activity. If a cell should have only inactive forms of the enzyme, then an external source of the active enzyme is required to start what would then become a self-propagating trait, as long as at least one active molecule was passed on to each cell. He provided two examples and the expectation

10 •

CHAPTER

that this class of proteins will define a new set of epigenetic mechanisms to pursue. Si presented preliminary evidence that a prion model may explain learned memory in Aplysia (Si et al. 2004). Protein translation of a number of stored mRNAs in neuronal cells is important for the maintenance of shortterm memory in this snail. He found that a regulator of protein translation, CPEB, can exist in two forms, and that the activated form of CPEB acts dominantly to perpetuate itself. Testing of this idea is still in its early days, but it offers an exciting new way of considering the issue of how we remember. 3.5 New Phenomenon

The description of a new and unexpected phenomenon always holds our imagination. One presentation in particular held my thoughts for weeks after the Symposium. Standard genetic analysis of mutant alleles of the HOTHEAD gene, which regulates organ fusion in Arabidopsis, revealed that normal rules of Mendelian genetics were not being followed (Lolle et al. 2005). It was discovered that if heterozygous HOTHEAD/hothead plants self-fertilized and produced a hom*ozygous hothead/hothead plant, and then this hom*ozygous hothead/hothead plant was allowed to self-fertilize, the progeny from this hom*ozygous parent reverted to a HOTHEAD/hothead genotype at a frequency of up to 15%. This stunning level of wild-type reversion produced an exact duplicate, at the nucleotide level, of the wild-type gene seen in the earlier generations. This reversion was not limited to the HOTHEAD locus-several other loci had similar frequencies of reversion to wildtype alleles. However, all the reversions required that the parent be hom*ozygous hothead/hothead. The gene product of HOTHEAD did not offer an obvious explanation as to how this could occur, but discussions certainly suggested that an archival copy of the wild-type gene was transmitted, perhaps via RNA, through successive generations. Although it could be argued that this phenomenon is outside the purview of "epigenetics"-due to the change in DNA sequence-the heritable transmission of the putative archived copy does not follow normal genetic rules. Nevertheless, this phenomenon has enormous implications for genetics, especially in evolutionary thinking.

4 Closing Thoughts

So, what more needs to be done to understand epigenetic mechanisms? For the most part, we are still collecting (discovering) the components. Just as the full sequence of

a genome has greatly facilitated progress in genetics, a clearer understanding for epigenetics will likely come when all the parts are known. It is encouraging to see the great strides that have been made in the last decade. I confess that I cannot discern whether we are close to, or far away from, having an accurate mechanistic understanding about how epigenetic states are maintained and propagated. The prion-based phenomenon may be the first to be understood, but those that are chromatin-based seem the farthest off. The polyvalent nature of interactions that seem to be required to establish a silenced state on a chromosome increases the complexity of the problem. This is further compounded by the dynamic nature of silent chromatin. The ability to know more about movement of components in and out of chromatin structures requires application of enhanced or new methods for an eventual understanding. Whereas chromatin immunoprecipitation has been important in establishing which components reside in a structure, it has temporarily blinded us to the dynamics. I suspect that, given the complexity, simply measuring binding and equilibrium constants between all the components and trying to derive a set of differential equations to simulate epigenetic switches may not be an effective use of resources, nor will it necessarily result in better comprehension. Rather, I speculate that a n'ew type of mathematical approach will need to be developed and combined with new experimental measuring methods, in order to eventually understand epigenetic events. Part of this may require development of in vitro systems, that faithfully recapitulate an epigenetic switch between states. The idea of competition between two states in most epigenetic phenomena likely reflects an "arms race" that is happening at many levels in the cell, followed by attempts to rectify "collateral damage." For instance, silencing proteins may have evolved to protect the genome from transposons. However, because silencing proteins work through the ubiquitous nucleosomes, some critical genes become repressed. To overcome this, histone modifications (e.g., methylation of H3K4 and H3K79) and variant replacement histones (H2A.Z) evolved to prevent silencing proteins from binding to critical genes. Depending on subsequent events, these changes may be co-opted for other processes-e.g., repression of some of the genes by the silencing proteins may have become useful (silent mating loci). The silencing mechanisms may have been co-opted for other functions as well, such as promoting chromosome segregation. And so it goes... I look forward to having the genomes of more organisms sequenced, because this might lead us to understand

EPIGENETICS:

an order of events through evolution that set up the epigenetic processes we see today. For instance, S. cerevisiae does not have RNAi machinery, but many other fungi do. By filling in some of the phylogenetic gap~ between species, we may discover what events led to S. cerevisiae no longer "needing" this system. Perhaps more than any other field of biological research, the study of epigenetics is founded on trying to understand unexpected observations, ranging from H.}. Muller's position-effect variegation, to polar overdominance in the callipyge phenotype (Georges et al. 2004). The hope of understanding something unusual serves as the bait to draw us in, but we soon become entranced by the cleverness of the mechanisms employed. This may explain why this field has drawn more than its share of light-hearted and clever minds. I suspect it will continue to do so, as we develop a deeper understanding of the cleverness, and as new and unexpected epigenetic phenomena are discovered.

Acknowledgments

I thank my colleagues at the University of Chicago and the Fred Hutchinson Cancer Research Center for making my own studies on epigenetics so enjoyable, and I thank the National Institutes of Health for financial support.

References Abraham J., Feldman J., Nasmyth K.A., Strathern J.N., Klar A.J., Broach J.R., and Hicks J.B. 1983. Sites required for position-effect regulation of mating-type information in yeast. Cold Spring Harbor Symp. Quant. BioI. 47: 989-998. AlIfrey v.G., Inoue A., Karn J., Johnson E.M., and Vidali G. 1974. Phosphorylation of DNA-binding nuclear acidic proteins and gene activation in the HeLa cell cycle. Cold Spring Harbor Symp. Quant. BioI. 38: 785-801. Andrulis E.D., Neiman A.M., Zappulla D.C., and Sternglanz R. 1998. Perinuclear localization of chromatin facilitates transcriptional silencing. Nature 394: 592-595. Andrulis E.D., Zappulla D.C., Ansari A., Perrod S., Laiosa c.v., Gartenberg M.R., and Sternglanz R. 2002. Escl, a nuclear periphery protein required for Sir4-based plasmid anchoring and partitioning. Mol. Cell. BioI. 22: 8292-8301. Aparicio O.M. and Gottschling D.E. 1994. Overcoming telomeric silencing: A trans-activator competes to establish gene expression in a cell cycle-dependent way. Genes Dev. 8: 1133-1146. Ariel M., Selig S., Brandeis M'-, Kitsberg D., Kafri T, Weiss A., Keshet 1., Razin A., and Cedar H. 1993. Allele-specific structures in the mouse Igf2-H19 domain. Cold Spring Harbor Symp. Quant. BioI. 58: 307-313. Bell A., Boyes J., Chung J., Pikaart M., Prioleau M.N., Recillas E, Saitoh N., and Felsenfeld G. 1998. The establishment of active chromatin domains. Cold Spring Harbor Symp. Quant. BioI. 63: 509-514. Bell S.P., Marahrens Y, Rao H., and Stillman B. 1993. The replicon

FROM

PHENOMENON

FIELD.

model and eukaryotic chromosomes. Cold Spring Harbor Symp. Quant. BioI. 58: 435-442. Belote J.M., McKeown M.B., Andrew D.J., Scott TN., Wolfner M.E, and Baker B.S. 1985. Control of sexual differentiation in Drosophila melanogaster. Cold Spring Harbor Symp. Quant. BioI. 50: 605-614. Beutler E. 1964. Gene inactivation: The distribution of gene products among populations of cells in heterozygous humans. Cold Spring Harbor Symp. Quant. BioI. 29: 261-271. Bird A.P. 1993. Functions for DNA methylation in vertebrates. Cold Spring Harbor Symp. Quant. Bioi. 58: 281-285. Borst P., Gommers-Ampt J.H., Ligtenberg M.J., Rudenko G., Kieft R., Taylor M.C., Blundell P.A., and van Leeuwen E 1993. Control of antigenic variation in African trypanosomes. Cold Spring Harbor Symp. Quant. BioI. 58: 105-114. Brink R.A. 1958. Paramutation at the R locus in maize. Cold Spring Harbor Symp. Quant. BioI. 23: 379-391. Campbell A. 1981. Some general questions about movable elements and their implications. Cold Spring Harbor Symp. Quant. BioI. 45: 1-9. Cattanach B.M. and Kirk M. 1985. Differential activity of maternally and paternally derived chromosome regions in mice. Nature 315: 496-498. Cedar H., Stein R., Gruenbaum Y., Naveh-Many T, Sciaky-Gallili N., and Razin A. 1983. Effect of DNA methylation on gene expression. Cold Spring Harbor Symp. Quant. BioI. 47: 605-609. Cham bon P. 1978. Summary: The molecular biology of the eukaryotic genome is coming of age. Cold Spring Harbor Symp. Quant. BioI. 42: 1209-1234. Cheng TH. and Gartenberg M.R. 2000. Yeast heterochromatin is a dynamic structure that requires silencers continuously. Genes Dev. 14: 452-463. Cheutin T, McNairn A.J., Jenuwein T, Gilbert D.M., Singh P.B., and Misteli T 2003. Maintenance of stable heterochromatin domains by dynamic HPI binding. Science 299: 721-725. Clark D., Reitman M., Studitsky V., Chung J., Westphal H., Lee E., and Felsenfeld G. 1993. Chromatin structure of transcriptionally active genes. Cold Spring Harbor Symp. Quant. BioI. 58: 1-6. co*ckell M., Gotta M., Palladino E, Martin S.G., and Gasser S.M. 1998. Targeting Sir proteins to sites of action: A general mechanism for regulated repression. Cold Spring Harbor Symp. Quant. BioI. 63: 401-412. Cuthbert G.L., Daujat S., Snowden A.W., Erdjument-Bromage H., Hagiwara T, Yamada M., Schneider R., Gregory P.D., Tempst E, Bannister A.J., and Kouzarides T 2004. Histone deimination antagonizes arginine methylation. Cell 118: 545-553. Dion M.E, Altschuler S.J., Wu L.E, and Rando O.J. 2005. Genomic characterization reveals a simple histone H4 acetylation code. Froc. Natl. Acad. Sci. 102: 5308-5309. Doerfler W., Kruczek 1., Eick D., Vardimon L., and Kron B. 1983. DNA methylation and gene activity: The adenovirus system as a model. Cold Spring Harbor Symp. Quant. BioI. 47: 593-603. Feinberg A.E, Kalikin L.M., Johnson L.A., and Thompson;fS.1994. Loss of imprinting in human cancer. Cold Spring Harbor symp) Quant. Bioi. 59: 357-364. Fox C.A., Loo S., Rivier D.H., Foss M.A., and Rine J. 1993. A transcriptional silencer as a specialized origin of replication that establishes functional domains of chromatin. Cold Spring Harbor Symp. Quant. BioI. 58: 443-455. Frankel J. 1990. Positional order and cellular handedness.]. Cell Sci. 97:205-211. Gartler S.M., and Linder D. 1964. Selection In mammalian mosaic cell

12 •

CHAPTER

populations. Cold Spring Harbor Symp. Quant. BioI. 29: 253-260. Gasser S.M., Hediger P., Taddei A., Neumann P.R., and Gartenberg M.R. 2004. The function of telomere clustering in yeast: The circe effect. Cold Spring Harbor Symp. Quant. BioI. 69: 327-337. Georges M., Charlier c., Smit M., Davis E., Shay T., Tordoir X., Takeda H., Caiment P., and co*ckett N. 2004. Toward molecular understanding of polar overdominance at the ovine callipyge locus. Cold Spring Harbor Symp. Quant. BioI. 69: 477-483. Goldschmidt R.B. 1951. The theory of the gene: Chromosomes and genes. Cold Spring Harbor Symp. Quant. Bioi. 16: 1-11. Haber J.E., Weiffenbach B., Rogers D.T., McCusker J., and Rowe L.B. 1981. Chromosomal rearrangements accompanying yeast matingtype switching: Evidence for a gene-conversion model. Cold Spring Harbor Symp. Quant. BioI. 45: 991-1002. Haig D. 2004. The (dual) origin of epigenetics. Cold Spring Harbor Symp. Quant. BioI. 69: 67. Henikoff S. 1990. Position-effect variegation after 60 years. Trends Genet. 6: 422-426. Henikoff S., McKittrick E., and Ahmad K. 2004. Epigenetics, histone H3 variants, and the inheritance of chromatin states. Cold Spring Harbor Symp. Quant. Bioi. 69: 235-243. Hernday A.D., Braaten B.A., and Low D.A. 2003. The mechanism by which DNA adenine methylase and PapI activate the pap epigenetic switch. Mol. Cel/12: 947-957. Hodgkin J., Doniach T., and Shen M. 1985. The sex determination pathway in the nematode Caenorhabditis elegans: Variations on a theme. Cold Spring Harbor Symp. Quant. BioI. 50: 585-593. Imai S., Johnson P.B., Marciniak R.A., McVey M., Park P.D., and Guarente L. 2000. Sir2: An NAD-dependent histone deacetylase that connects chromatin silencing, metabolism, and aging. Cold Spring Harbor Symp. Quant. Bioi. 65: 297-302. Jenuwein T. and Allis C.D. 2001. Translating the histone code. Science 293: 1074-1080. Klar A.J., and Bonaduce M.J. 1993. The mechanism of fission yeast mating-type interconversion: Evidence for two types of epigenetically inherited chromosomal imprinted events. Cold Spring Harbor Symp. Quant. BioI. 58: 457-465. Klar A.J., Hicks J.B., and Strathern J.N. 1981. Irregular transpositions of mating-type genes in yeast. Cold Spring Harbor Symp. Quant. BioI. 45: 983-990. Kubicek S. and Jenuwein T. 2004. A crack in histone lysine methylation. Cel/1l9: 903-906. La Volpe A., Taggart M., Macleod D., and Bird A. 1983. Coupled demethylation of sites in a conserved sequence of Xenopus ribosomal DNA. Cold Spring Harbor Symp. Quant. Bioi. 47: 585-592. Laroche T., Martin S.G., Gotta M., Gorham H.C., Pryde P.E., Louis E.J., and Gasser S.M. 1998. Mutation of yeast Ku genes disrupts the subnuclear organization of telomeres. Curro BioI. 8: 653-656. Laurenson P. and Rine J. 1992. Silencers, silencing, and heritable transcriptional states. Microbiol. Rev. 56: 543-560. Lewis E.B. 1985. Regulation of the genes of the bithorax complex in Drosophila. Cold Spring Harbor Symp. Quant. BioI. 50: 155-164. Li E., Beard c., Forster A.C., Bestor T.H., and Jaenisch R. 1993. DNA methylation, genomic imprinting, and mammalian development. Cold Spring Harbor Symp. Quant. BioI. 58: 297-305. Locke J., Kotarski M.A., and Tartof K.D. 1988. Dosage-dependent modifiers of position effect variegation in Drosophila and a mass action model that explains their effect. Genetics 120: 181-198. Lolle S.J., Victor J.L., Young J.M., and Pruitt R.E. 2005. Genome-wide non-mendelian inheritance of extra-genomic information in Arabidopsis. Nature 434: 505-509. Losick R. 1998. Summary: Three decades after sigma. Cold Spring Har-

bor Symp. Quant. BioI. 63: 653-666. Louie A.J., Candido E.P., and Dixon G.H. 1974. Enzymatic modifications and their possible roles in regulating the binding of basic proteins to DNA and in controlling chromosomal structure. Cold Spring Harbor Symp. Quant. BioI. 38: 803-819. Lyon M.P. 1961. Gene action in the X-chromosome of the mouse (Mus musculus L.). Nature 190: 372-373. - - - . 1993. Epigenetic inheritance in mammals. Trends Genet. 9: 123-128. Maine E.M., Salz H.K., Schedl P., and Cline T.W. 1985. Sex-lethal, a link between sex determination and sexual differentiation in Drosophila melanogaster. Cold Spring Harbor Symp. Quant. BioI. 50: 595-604. Martin-Morris L.E., Loughney K., Kershisnik E.O., Poortinga G., and Henikoff S. 1993. Characterization of sequences responsible for trans-inactivation of the Drosophila brown gene. Cold Spring Harbor Symp. Quant. BioI. 58: 577-584. McClintock B. 1951. Chromosome organization and genic expression. Cold Spring Harbor Symp. Quant. BioI. 16: 13-47. - - - . 1956. Controlling elements and the gene. Cold Spring Harbor Symp. Quant. BioI. 21: 197-216. Megee P.c., Morgan B.A., and Smith M.M. 1995. Histone H4 and the maintenance of genome integrity. Genes Dev. 9: 1716-1727. Millar C.B., Kurdistani S.K., and Grunstein M. 2004. Acetylation of yeast histone H4 lysine 16: A switch for protein interactions in heterochromatin and euchromatin. Cold Spring Harbor Symp. Quant. BioI. 69: 193-200. Mirkovitch J., Gasser S.M., and Laemmli U.K. 1987. Relation of chromosome structure and gene expression. Philos. Trans. R. Soc. Lond. B BioI. Sci. 317: 563-574. Mostoslavsky R., Kirillov A., Ji Y.H., Goldmit M., Holzmann M., Wirth T., Cedar H., and Bergman Y. 1999. Demethylation and the establishment of K allelic exclusion. Cold Spring Harbor Symp. Quant. BioI. 64: 197-206. Muller H.J. 1941. Induced mutations in Drosophila. Cold Spring Harbor Symp. Quant. BioI. 9: 151-167. Nance W.E. 1964. Genetic tests with a sex-linked marker: Glucose-6phosphate dehydrogenase. Cold Spring Harbor Symp. Quant. Bioi. 29: 415-425. Nanney D.L. 1958. Epigenetic factors affecting mating type expression in certain ciliates. Cold Spring Harbor Symp. Quant. BioI. 23: 327-335. Nasmyth K.A., Tatchell K., Hall B.D., Astell c., and Smith M. 1981. Physical analysis of mating-type loci in Saccharomyces cerevisiae. Cold Spring Harbor Symp. Quant. BioI. 45: 961-981. Palladino P., Laroche T., Gilson E., Pillus L., and Gasser S.M. 1993. The positioning of yeast telorneres depends on SIR3, SIR4, and the integrity of the nuclear membrane. Cold Spring Harbor Symp. Quant. BioI. 58: 733-746. Pannetier c., Hu-Li J., and Paul W.E. 1999. Bias in the expression of IL-4 alleles: The use of T cells from a GFP knock-in mouse. Cold Spring Harbor Symp. Quant. BioI. 64: 599-602. Peaco*ck w.J., Brutlag D., Goldring E., Appels R., Hinton c.w., and Lindsley D.L. 1974. The organization of highly repeated DNA sequences in Drosophila rpe/aiWgaster chromosomes. Cold Spring Harbor Symp. Quant. Bioi. 38: 405'+416. Pillus L. and Rine J. 1989. EPigenet;Jinheritance of transcriptional states in S. cerevisiae. Cel/59: 637-647. Ptashne M. 2004. A genetic switch: Phage lambda revisited, 3rd edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Renauld H., Aparicio O.M., Zierath ED., Billington B.L., Chhablani

EP/GENETICS:

S.K., and Gottschling D.E. 1993. Silent domains are assembled continuously from the telomere and are defined by promoter distance and strength, and by SIR3 dosage. Genes Dev. 7: 1133-1145. Rine J., Jensen R., Hagen D., Blair 1., and Herskowitz 1. 1981. Pattern of switching and fate of the replaced cassette in yeast mating-type interconversion. Cold Spring Harbor Symp. Quant. BioI. 45: 951-960. Rubin G.M. 1985. Summary. Cold Spring Harbor Symp. Quant. BioI. 50: 905-908. Rubin G.M., Hazelrigg T., Karess R.E., Laski EA., Laverty T., Levis R., Rio D.C., Spencer EA., and Zuker C.S. 1985. Germ line specificity of P-element transposition and some novel patterns of expression of transduced copies of the white gene .. Cold Spring Harbor Symp. Quant. BioI. 50: 329-335. Rudkin G.T. and Tartof K.D. 1974. Repetitive DNA in polytene chromosomes of Drosophila melanogaster. Cold Spring Harbor Symp. Quant. Bioi. 38: 397-403. Schubeler D., MacAlpine D.M., Scalzo D., Wirbelauer c., Kooperberg c., van Leeuwen E, Gottschling D.E., O'Neill L.P., Turner B.M., Delrow J., et al. 2004. The histone modification pattern of active genes revealed through genome-wide chromatin analysis of a higher eukaryote. Genes Dev. 18: 1263-1271. Schultz J. 1956. The relation of the heterochromatic chromosome regions to the nucleic acids of the cell. Cold Spring Harbor Symp. Quant. BioI. 21: 307-328. Selker E.U., Richardson G.A., Garrett-Engele P.W., Singer M.J., and Miao V. 1993. Dissection of the signal for DNA methylation in the 1;-11 region of Neurospora. Cold Spring Harbor Symp. Quant. Bioi. 58: 323-329. Shapiro L.J. and Mohandas T. 1983. DNA methylation and the control of gene expression on the human X chromosome. Cold Spring HarborSymp. Quant. BioI. 47: 631-637. Shi Y, Lan E, Matson C., Mulligan P., Whetstine J.R., Cole P.A., and Casero R.A. 2004. Histone demethylation mediated by the nuclear amine oxidase hom*olog LSD 1. Cell 119: 941-953. Si K., Lindquist S., and Kandel E. 2004. A possible epigenetic mechanism for the persistence of memory. Cold Spring Harbor Symp. Quant. BioI. 69: 497-498. Solter D., Aronson J., Gilbert S.E, and McGrath J. 1985. Nuclear transfer in mouse embryos: Activation of the embryonic genome. Cold Spring Harbor Symp. Quant. BioI. 50: 45-50. Swift H. 1974. The organization of genetic material in eukaryotes: Progress and prospects. Cold Spring Harbor Symp. Quant. BioI. 38: 963-979. Thompson J.S., Hecht A., and Grunstein M. 1993. Histones and the

FROM

PHENOMENON

FIELD.

regulation of heterochromatin in yeast. Cold Spring Harbor Symp. Quant. BioI. 58: 247-256. Tilghman S.M., Bartolomei M.S., Webber A.L., Brunkow M.E., Saam J., Leighton P.A., Pfeifer K., and Zemel S. 1993. Parental imprinting of the H19 and Igf2 genes in the mouse. Cold Spring Harbor Symp. Quant. BioI. 58: 287-295. Vazquez J., Farkas G., Gaszner M., Ddvardy A., Muller M., Hagstrom K., Gyurkovics H., Sipos L., Gausz J., Galloni M., et al. 1993. Genetic and molecular analysis of chromatin domains. Cold Spring Harbor Symp. Quant. BioI. 58: 45-54. Wade P.A., Jones P.L., Vermaak D., Veenstra G.J., Imhof A., Sera T., Tse c., Ge H., Shi Y.B., Hansen J.c., and Wolffe A.P. 1998. Histone deacetylase directs the dominant silencing of transcription in chromatin: Association with MeCP2 and the Mi-2 chromodomain SWI/SNF ATPase. Cold Spring Harbor Symp. Quant. Bioi. 63: 435-445. Wang Y., Wysocka J., Perlin J.R., Leonelli L., Allis CD., and Coonrod S.A. 2004. Linking covalent histone modifications to epigenetics: The rigidity and plasticity of the marks. Cold Spring Harbor Symp. Quant. Bioi. 69: 161-169. Weintraub H. 1974. The assembly of newly replicated DNA into chromatin. Cold Spring Harbor Symp. Quant. BioI. 38: 247-256. - - - . 1993. Summary: Genetic tinkering local problems, local solutions. Cold Spring Harbor Symp. Quant. Bioi. 58: 819-836. Weintraub H., Flint S.J., Leffak 1.M., Groudine M., and Grainger R.M. 1978. The generation and propagation of variegated chromosome structures. Cold Spring Harbor Symp. Quant. Bioi. 42: 401-407. Wickner R.B., Edskes H.K., Ross E.D., Pierce M.M., Baxa D., Brachmann A., and Shewmaker E 2004a. Prion genetics: New rules for a new kind of gene. Annu. Rev. Genet. 38: 681-707. Wickner R.B., Edskes H.K., Ross E.D., Pierce M.M., Shewmaker E, Baxa D., and Brachmann A. 2004b. Prions of yeast are genes made of protein: Amyloids and enzymes. Cold Spring Harbor Symp. Quant. BioI. 69: 489-496. Willard H.E, Brown CJ., Carrel L., Hendrich B., and Miller A.P. 1993. Epigenetic and chromosomal control of gene expression: Molecular and genetic analysis of X chromosome inactivation. Cold Spring Harbor Symp. Quant. BioI. 58: 315-322. Wood W.B., Meneely P., Schedin P., and Donahue L. 1985. Aspects of dosage compensation and sex determination in Caenorhabditis elegans. Cold Spring Harbor Symp. Quant. BioI. 50: 575-583. Yarmolinsky M.B. 1981. Summary. Cold Spring Harbor Symp. Quant. BioI. 45: 1009-1015. Zheng C, and Hayes p. 2003. Structures and interactions of the core histone tail domains. Biopolymers 68: 539-546.

HAP

A Brief History of Epigenetics Gary Felsenfeld National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20892-0540

CONTENTS 1. Introduction, 16

5. The Role of Chromatin, 18

2. Clues from Genetics and Development, 16

6. All Mechanisms Are Interrelated, 19

3. DNA Is the Same in All Somatic Cells of an Organism, 17

References, 21

4. The Role of DNA Methylation, 17

16 •

CHAPTER

1 Introduction

The history of epigenetics is linked with the study of evolution and development. But during the past 50 years, the meaning of the term "epigenetics" has itself undergone an evolution that parallels our dramatically increased understanding of the molecular mechanisms underlying regulation of gene expression in eukaryotes. Our present working definition is "the study of mitotically and/or meiotically heritable changes in gene function that cannot be explained by changes in DNA sequence" (Riggs et al. 1996). Until the 1950s, however, the word epigenetics was used in an entirely different way to categorize all of the developmental events leading from the fertilized zygote to the mature organism-that is, all of the regulated processes that, beginning with the genetic material, shape the final product (Waddington 1953). This concept had its origins in the much earlier studies in cell biology and embryology, beginning in the late 19th century, that laid the groundwork for our present understanding of the relationship between genes and development. There was a long debate among embryologists about the nature and location of the components responsible for carrying out the developmental plan of the organism. In trying to make sense of a large number of ingenious but ultimately confusing experiments involving the manipulation of cells and embryos, embryologists divided into two schools: those who thought that each cell contained preformed elements that enlarged during development, and those who thought the process involved chemical reactions among soluble components that executed a complex developmental plan. These views focused on the relative importance of the nucleus and cytoplasm in the developmental process. Following Flemming's discovery of the existence of chromosomes in 1879, experiments by many . investigators, including Wilson and Boveri, provided strong evidence that the developmental program'feSicled in the chromosomes. Thomas Hunt Morgan (1911) ultimately provided the most persuasive proof of this idea through his demonstration of the genetic linkage of several Drosophila genes to the X chromosome. From that point onward, rapid progress was made in creating linear chromosome maps in which individual genes were assigned to specific sites on the Drosophila chromosomes (Sturtevant 1913). Of course, the questions of classic "epigenesis" remained: What molecules within the chromosomes carried the genetic information, how did they direct the developmental program, and how was the information transmitted during cell division? It was understood that both nucleic acid and proteins were pres-

ent in chromosomes, but their relative contributions were not obvious; certainly, no one believed that the nucleic acid alone could carryall of the developmental information. Furthermore, earlier questions persisted about the possible contribution of the cytoplasm to developmental events. Evidence from Drosophila genetics (see below) suggested that heritable changes in phenotype could occur without corresponding changes in the "genes." This debate was dramatically altered by the identification of DNA as the primary carrier of genetic information. Ultimately, it became useful to redefine epigenetics so as to distinguish heritable changes that arise from sequence changes in DNA from those that do not. 2 Clues from Genetics and Development

Whatever the vagaries of the definition, the ideas and scientific data that underlie the present concept of epigenetics had been accumulating steadily since the early part of the 20th century. In 1930, H.]. Muller (Muller 1930) described a class of Drosophila mutations he called "eversporting displacements" ("eversporting" denoting the high rate of phenotypic change). These mutants involved chromosome translocations (displacements), but "even when all parts of the chromatin appeared to be represented in the right dosage-though abnormally arranged-the phenotypic result was not always normal." In some of these cases, Muller observed flies that had mottled eyes. He thought that this was probably due to a "genetic diversity of the different eye-forming cells;' but further genetic analysis led him to connect the unusual properties with chromosomal rearrangement, and to conclude that "chromosome regions, affecting various characters at once, are somehow concerned, rather than individual genes or suppositious 'gene elements.''' Over the next 10 to 20 years, strong evidence provided by many laboratories (see Hannah 1951) confirmed that this variegation arose when rearrangements juxtaposed the white gene with heterochromatic regions. During that period, chromosomal rearrangements of all kinds were the object of a great deal of attention. It was apparent that genes were not completely independent entities; their function could be affected by their location within the genome-as amply demonstrated by the many Drosophila mutants that led to variegation, as well as by other mutants involving translocation to euchromatic regions, in which more general (non-variegating) position effects could be observed. The role of transposable elements in plant genetics also became clear, largely through the work of McClintock (1965).

A second line of reasoning came from the study of developmental processes. It was evident that during development there was a divergence of phenotypes among differentiating cells and tissues, and it appeared that such distinguishing features, once established, could be clonally inherited by the dividing cells. Although it was understood at this point that cell-specific programming existed, and that it could be transmitted to daughter cells, how this was done was less clear. A number of mechanisms could be imagined, and were considered. Particularly for those with a biochemical point of view, a cell was defined by the multiple interdependent biochemical reactions that maintained its identity. For example, it was suggested in 1949 by Delbruck (quoted in Jablonka and Lamb 1995) that a simple pair of biochemical pathways, each of which produced as an intermediate an inhibitor of the other pathway, could establish a system that could switch between one of two stable states. Actual examples of such systems were found somewhat later in the lac operon of Escherichia coli (Novick and Weiner 1957) and in the phage switch between lysogenic and lytic states (Ptashne 1992). Functionally equivalent models could be envisioned in eukaryotes. The extent to which nucleus and cytoplasm each contributed to the transmission of a differentiated state in the developing embryo was of course a matter of intense interest and debate; a self-stabilizing biochemical pathway would presumably have to be maintained through cell division. A second kind of epigenetic transmission was clearly demonstrated in Paramecia and other ciliates, in which the ciliary patterns may vary among individuals and are inherited clonally (Beisson and Sonneborn 1965). Altering the cortical pattern by microsurgery results in transmission of a new pattern to succeeding generations. It has been argued that related mechanisms are at work in metazoans, in which the organization of cellular components is influenced by localized cytoplasmic determinants in a way that can be transmitted during cell division (Grimes and Aufderheide 1991). 3 DNA Is the Same in All Somatic Cells of an Organism

Although chromosome morphology indicated that all somatic cells possessed all of the chromosomes, it could not have been obvious that all somatic cells retained the full complement of DNA present in the fertilized egg. Nor until the work of Avery, MacLeod, and McCarty in 1944, and that of Hershey and Chase (1952), was it even clear that a protein-free DNA molecule could carry

BRIEF

HISTORY

EPIGENETICS

•

genetic information, a conclusion strongly reinforced by Watson and Crick's solution of the structure of DNA in 1953. Work by Briggs and King (1952) in Rana pipiens and by Laskey and Gurdon (1970) in Xenopus had demonstrated that introduction of a nucleus from early embryonic cells into enucleated oocytes could result in development of an embryo. But as late as 1970, Laskey and Gurdon could state that "It has yet to be proved that somatic cells of an adult animal possess genes other than those necessary for their own growth and differentiation." In the paper containing this statement, they went on to show that to a first approximation, the DNA of a somatic cell nucleus was competent to direct embryogenesis when introduced into an enucleated egg. It was now clear that the program of development, and the specialization of the repertoire of expression seen in somatic cells, must involve signals that are not the result of some deletion or mutation in the germ-line DNA sequence when it is transmitted to somatic cells. Of course, there are ways in which the DNA of somatic cells can come to differ from that of the germ line, with consequences for the cellular phenotype: For example, transposable elements can alter the pattern of expression in somatic cells, as demonstrated by the work of Barbara McClintock and other plant geneticists. Similarly, the generation of antibody diversity involves DNA rearrangement in a somatic cell lineage. This rearrangement (or more precisely its consequences) can be considered a kind of epigenetic event, consistent with the early observations of position-effect variegation described by Muller. -However, much of the work on epigenetics in recent years has focused on systems in which no DNA rearrangements have occurred, and the emphasis has therefore been on modifications to the bases, and to the proteins that are complexed with DNA within the nucleus. 4 The Role of DNA Methylation

X-chromosome inactivation provided an early model of this kind of epigenetic mechanism (Ohno et al. 1959; Lyon 1961); the silenced X chromosome was clearly chosen at random in somatic cells, and there was no evidence of changes in the DNA sequence itself. In part to account for this kind of inactivation, Riggs (1975) and Holliday and Pugh (1975) proposed that DNA methylation could act as an epigenetic mark. The key elements in this model were the ideas that sites of methylation were palindromic, and that distinct enzymes were responsible for methylation of unmodified DNA and DNA already methylated on one strand. It was postulated that the first methylation

18 •

CHAPTER

event would be much more difficult than the second; once the first strand was modified, however, the complementary strand would quickly be modified at the same palindromic site. A methylation mark present on a parental strand would be copied on the daughter strand following replication, resulting in faithful transmission of the methylated state to the next generation. Shortly thereafter, Bird took advantage of the fact that the principal target of methylation in animals is the sequence CpG (Doskocil and Sorm 1962) to introduce the use of methylation-sensitive restriction enzymes as a way of detecting the methylation state. Subsequent studies (Bird 1978; Bird and Southern 1978) then showed that endogenous CpG sites were either completely unmethylated or completely methylated. The predictions of the model were thus confirmed, establishing a mechanism for epigenetic transmission of the methylation mark through semiconservative propagation of the methylation pattern. In the years following these discoveries, a great deal of attention has been focused on endogenous patterns of DNA methylation, on the possible transmission of these patterns through the germ line, on the role of DNA methylation in silencing gene expression, on possible mechanisms for initiation or inhibition of methylation at a fully unmethylated site, and on the identification of the enzymes responsible for de novo methylation and for maintenance of methylation on already methylated sites. Although much of the DNA methylation seen in vertebrates is associated with repetitive and retroviral sequences and may serve to maintain these sequences in a permanently silent state, there can be no question that in many cases this modification provides the basis for epigenetic transmission of the state of gene activity. This is most clearly demonstrated at imprinted loci (Cattanach and Kirk 1985) such as the mouse or human Igf2/H19 locus, where one allele is marked by DNA methylation, which in turn contro~pressionfrom both genes (Bell and Felsenfeld 2006; Hark et al. 2000). At the same time, it was clear that this could riot be the only mechanism for epigenetic transmission of information. For example, as noted above, position-effect variegation had been observed many years earlier in Drosophila, an organism that has extremely low levels of DNA methylation. Furthermore, in subsequent years, Drosophila geneticists had identified the Polycomb and Trithorax groups of genes, which appeared to be involved in permanently "locking in" the state of activity, either off or on, respectively, of clusters of genes during development. The fact that these states were stably transmitted during cell division suggested an underlying epigenetic mechanism.

5 The Role of Chromatin It had been recognized for many years that the proteins

bound to DNA in the eukaryotic nucleus, especially the histones, might be involved in modifying the properties of DNA. Well before most of the work on DNA methylation began, Stedman and Stedman (1950) proposed that the histones could act as general repressors of gene expression. They argued that since all somatic cells of an organism had the same number of chromosomes, they had the same genetic complement (although this was not demonstrated until some years later, as noted above). Understanding the subtlety of histone modifications was far in the future, so the Stedmans operated on the assumption that different kinds of cells in an organism must have different kinds of histones in order to generate the observed differences in phenotype. Histones can indeed reduce levels of transcript far below those commonly observed for inactive genes in prokaryotes. Subsequent work addressed the capacity of chromatin to serve as a template for transcription, and asked whether that capacity was restricted in a cell-type-specific manner. In a 1963 paper, Bonner (Bonner et al. 1963) prepared chromatin from a globulin-producing tissue of the pea plant, and showed that when E. coli RNA polymerase was added, and the resulting transcript translated in an in vitro system, globulin could be detected. The result was specific to this tissue. With the advent of hybridization methods, the transcript populations from such in vitro experiments could be examined (Paul and Gilmour 1968) and shown to be specific for the particular tissue from which the chromatin was derived. Other results suggested that this specificity reflected a restriction in access to transcription initiation sites (Cedar and Felsenfeld 1973). Nonetheless, there was a period in which it was commonly believed that the histones were suppressor proteins that passively silenced gene expression. In this view, activating a gene simply meant stripping off the histones; once that was done, it was thought, transcription would proceed pretty much as it did in prokaryotes. There was, however, some evidence that extended regions of open DNA did not exist in eukaryotic cells (Clark and Felsenfeld 1971). Furthermore, even if the naked DNA model was correct, it was not clear how the decision would be made as to which histone-covered regions should be cleared. The resolution of this problem began as early as 1964, when Allfrey (Allfrey et al. 1964) had speculated that histone acetylation might be correlated with gene activation, and that "active" chromatin might not necessarily be stripped of histones. In the ensuing decade, there was

great interest in examining the relationship between histone modifications and gene expression. Modifications other than acetylation (methylation and phosphorylation) were identified, but their functional significance was unclear. It became much easier to address this problem after the discovery by Kornberg and Thomas (1974) of the structure of the nucleosome, the fundamental chromatin subunit. The determination of the crystal structure of the nucleosome, first at 7 A and then at 2.8 A resolution, also provided important structural information, particularly evidence for the extension of the histone amino-terminal tails beyond the DNA-protein octamer core, making evident their accessibility to modification (Richmond et al. 1984; Luger et al. 1997). Beginning in 1980 and extending over some years, Grunstein and his collaborators (Wallis et al. 1980; Durrin et al. 1991), applying yeast genetic analysis, were able to show that the histone amino-terminal tails were essential for regulation of gene expression, and for the establishment of silent chromatin domains. The ultimate connection to detailed mechanisms began with the critical demonstration by Allis (Brownell et al. 1996) that a histone acetyltransferase from Tetrahymena was hom*ologous to yeast transcriptional regulatory protein GcnS, providing direct evidence that histone acetylation was connected to control of gene expression. Since then, of course, there has been an explosion of discovery of histone modifications, as well as a reevaluation of the roles of those that were known previously. This still did not answer the question of how the sites for modification were chosen in vivo. It had been shown, for example (Pazin et al. 1994), that Ga14-VP16 could activate transcription from a reconstituted chromatin template in an ATP-dependent manner. Activation was accompanied by repositioning of nucleosomes, and it was suggested that this was the critical event in making the promoter accessible. A fuller understanding of the significance of these findings required the identification of ATP-dependent nucleosome remodeling complexes such as SWI/SNF and NURF (Peterson and Herskowitz 1992; Tsukiyama and Wu 1995), and the realization that both histone modification and nucleosome remodeling were involved in preparing the chromatin template for transcription. It was not clear how information about the state of activity could, employing these mechanisms, be transmitted through cell division; their role in epigenetic transmission of information was thus unclear. The next important step came from the realization that modified histones recruited, in a modification-specific way, proteins that could affect the local structural and functional

BRIEF

HISTORY

EPIGENETICS

states of chromatin. It was found, for example, that methylation of histone H3 lysine 9 resulted in the recruitment of the heterochromatin protein HP1 (Bannister et al. 2001; Lachner et al. 2001; Nakayama et al. 2001). Furthermore, HP1 could recruit the enzyme (Suv39 h 1) that is responsible for that methylation. This led to a model for propagation of the silenced chromatin state along the region through a processive mechanism (Fig. 1a). Equally important, it provided a reasonable explanation of how that state could be transmitted and survive through the replication cycle (Fig. 1b). Analogous mechanisms for propagation of an active state have been proposed that involve methylation of histone H3 lysine 4 and the recruitment of Trithorax group proteins (Wysocka et al. 2005). Different kinds of propagation mechanisms have been suggested that depend on variant histones rather than modified histones (Ahmad and Henikoff 2002; McKittrick et al. 2004). Histone H3 is incorporated into chromatin only during DNA replication. In contrast, the histone variant H3.3, which differs from H3 by four amino acids, is incorporated into nucleosomes in a replication-independent manner, and it tends to accumulate in active chromatin, where it is enriched in the "active" histone modifications (McKittrick et al. 2004). It has been proposed that the presence of H3.3 is sufficient to maintain the active state, and that after replication, although it would be diluted twofold, enough H3.3 would remain to maintain the active state. The consequent transcription would result in replacement of H3 containing nucleosomes with H3.3, thus perpetuating the active state in the next generation.

6 All Mechanisms Are Interrelated These models finally begin to complete the connection between modified or variant histones, specific gene activation, and epigenetics, although of course there is much more to be done. Whereas these mechanisms give us some ideas about how the heterochromatic state may be maintained, they do not explain how silencing chromatin structures are first established. It has only recently become clear that this involves the production of RNA transcripts, particularly from repeated sequences, which are processed into small RNAs through the action of proteins such as Dicer, Argonaute, and RNA-dependent RNA polymerase. These RNAs are subsequently recruited to the hom*ologous DNA sites as part of complexes that include components of the Polycomb group of proteins, thus initiating the formation of heterochromatin. There is now also evi-

20 • C HAP

T ER 2

/)(\~Jn/)(\9n ,

strand A strand B '!

"''6'''V

tI strandA~

"''6

"'V

DNA

replication

~strandB

maintenance DNA methylation

Figure 1. Mechanisms for Maintaining a Pattern of DNA Methylation and a Histone Modification during DNA Replication (0) A mechanism for maintaining a pattern of DNA methylation during DNA replication. During replication, the individual DNA strands, with a specific methylation pattern at CpG or CpXpG residues, become paired with a strand of newly synthesized, un methylated DNA. CpG on one strand has a corresponding CpG on the other. The maintenance DNA methyltransferase recognizes a hemimethylated site, and methylates the cytosine on the new strand, so that the pattern of methylation is undisturbed. (b) A general mechanism for maintaining a histone modification during replication. The modified histone tail (m) interacts with a protein binder (pb) that has a binding site specific for that modification. pb, in turn, has a specific site for the enzyme (e) which carries out that histone modification. e, in turn, can then modify an adjacent nucleosome. During replication, the newly deposited histones which are interspersed with parental histones can thus acquire the parental modification. A similar mechanism would allow propagation of histone modifications from a modified region into an unmodified one at any stage of the cell cycle.

dence that the same mechanisms are required for maintenance of at least some heterochromatic regions. In a way, these stable cyclic reaction pathways are reminiscent of Delbruck's 50-year-old model, of a stable biochemical cycle that maintain~ate of the organism. We now knm/ of countless examples of epigenetic mechanisms at work in the organism. In addition to imprinting at many loci, and the allele-specific and random X-chromosome inactivation described above, there are epigenetic phenomena involved in antibody expression, where the rearrangement of the immunoglobulin genes on one chromosome is selectively inhibited, and in the selection for expression of single odorant receptor genes in olfactory neurons (Chess et al. 1994; Shykind et al. 2004). In Drosophila, the Polycomb group genes are responsible for establishing a silenced chromatin domain that is maintained through all subsequent cell divisions.

Epigenetic changes are also responsible for paramutation in plants, in which one allele can cause a heritable change in expression of the hom*ologous allele (Stam et al. 2002). This is an example of an epigenetic state that is inherited meiotically as well as mitotically, a phenomenon documented in plants but only rarely in animals (Jorgensen 1993). Much of the evidence for the mechanisms described above has come from work on the silencing of mating-type locus and centromeric sequences in Schizosaccharomyces pombe (Hall et al. 2002). In addition, the condensed chromatin structure characteristic of centromeres in organisms as diverse as flies and humans has been shown to be transmissible through centromereassociated proteins rather than DNA sequence. In all of these cases, the DNA sequence remains intact, but its capacity for expression is suppressed. This is likely in all cases to be mediated by DNA methylation, histone mod-

ification, or both; in some cases, we already know that to be true. Finally, the epigenetic transmission of "patterns;' described above for Paramecia, now extends to the prion proteins, which maintain and propagate their alternatively folded state to daughter cells. Although this has been presented as a sequential story, it should more properly be viewed as a series of parallel and overlapping attempts to define and explain epigenetic phenomena. The definition of the term epigenetics has changed, but the questions about mechanisms of development raised by earlier generations of scientists have not. Contemporary epigenetics still addresses those central questions. Seventy years have passed since Muller described what is now called position-effect variegation. It is gratifying to trace the slow progress from observation of phenotypes, through elegant genetic studies, to the recent analysis and resolution at the molecular level. With this knowledge has come the understanding that epigenetic mechanisms may in fact be responsible for a considerable part of the phenotype of complex organisms. As is often the case, an observation that at first seemed interesting but perhaps marginal to the main issues turns out to be central, although it may take a long time to come to that realization. References Ahmad K and Henikoff S. 2002. The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9: 1191-1200. Allfrey V.G., Faulkner R., and Mirsky A.E. 1964. Acetylation and methylation of histones and their possible role in the regulation of RNA synthesis. Proc. Natl. Acad. Sci. 51: 786-794. Avery O.T, MacLeod CM., and McCarty M. 1944. Studies on the chemical nature of the substance inducing transformation of pneumococcal types. f. Exp. Med. 79: 137-158. Bannister A., Zegerman P., Partridge J., Miska E., Thomas J., Allshire R., and Kouzarides T 2001. Selective recognition of methylated lysine 9 on histone H3 by the HPI chromo domain. Nature 410: 120-124. Beisson J. and Sonneborn TM. 1965. Cytoplasmic inheritance of the organization of the cell cortex in Paramecium aurelia. Proc. Natl. Acad. Sci. 53: 275-282. Bell A.C. and Felsenfeld G. 2000. Methylation of a CTCF-dependent boundary controls imprinted expression of the Igf2 gene. Nature 405: 482-485. Bird A.P. 1978. Use of restriction enzymes to study eukaryotic DNA methylation. II. The symmetry of methylated sites supports semiconservative copying of the methylation pattern. f. Mol. Bioi. 118: 49-60. Bird A.P. and Southern E.M. 1978. Use of restriction enzymes to study eukaryotic DNA methylation. I. The methylation pattern in ribosomal DNA from Xenopus laevis. f. Mol. BioI. 118: 27-47. Bonner J., Huang R.C, and Gilden R.Y. 1963. Chromosomally directed protein synthesis. Proc. Natl. Acad. Sci. 50: 893-900. Briggs R. and King n. 1952. Transplantation of living nuclei from blas-

B R f E F H f 5 TOR Y

E P f G ENE TIC 5

•

tula cells into enucleated frogs' eggs. Proc. Natl. Acad. Sci. 38: 455-463. Brownell J.E., Zhou J., Ranalli T., Kobayashi R., Edmondson D.G., Roth S.Y., and Allis CD. 1996. Tetrahymena histone acetyltransferase A: A hom*olog to yeast Gcn5p linking histone acetylation to gene activation. Cell 84: 843-851. Cattanach B.M. and Kirk M. 1985. Differential activity of maternally and paternally derived chromosome regions in mice. Nature 315: 496-498. Cedar H. and Felsenfeld G. 1973. Transcription of chromatin in vitro. f. Mol. BioI. 77: 237-254. Chess A., Simon I., Cedar H., and Axel R. 1994. Allelic inactivation regulates olfactory receptor gene expression. Cell 78: 823-834. Clark R,J. and Felsenfeld G. 1971. Structure of chromatin. Nat. New BioI. 229: 101-106. Doskocil J. and Sorm P. 1962. Distribution of 5-methylcytosine in pyrimidine sequences of deoxyribonucleic acids. Biochim. Biophys. Acta 55: 953-959. Durrin L.K, Mann R.K., Kayne P.S., and Grunstein M. 1991. Yeast histone H4 N-terminal sequence is required for promoter activation in vivo. Cell 65: 1023-1031. Grimes G.W. and Aufderheide K.J. 1991. Cellular aspects of pattern formation: The problem of assembly. Monogr. Dev. BioI. 22: 1-94. Hall I.M., Shankaranarayana G.D., Noma K., Ayoub N., Cohen A., and qrewal S.I. 2002. Establishment and maintenance of a heterochromatin domain. Science 297: 2215-2218. Hannah A. 1951. Localization and function of heterochromatin in Drosophila melanogaster. Adv. Genet. 4: 87-125. Hark A.T., Schoenherr CJ., Katz D.J., Ingram R.S., Levorse J.M., and Tilghman S.M. 2000. CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/lgf2 locus. Nature 405: 486-489. Hershey A.D. and Chase M. 1952. Independent functions of viral protein and nucleic acid in growth of bacteriophage. f. Gen. Physiol. 36: 39-56. Holliday R. and Pugh J.E. 1975. DNA modification mechanisms and gene activity during development. Science 187: 226-232.. Jablonka E. and Lamb M.J. 1995. Epigenetic inheritance and evolution: The Lamarckian dimension. Oxford University Press, New York, p.82. Jorgensen R. 1993. The germinal inheritance of epigenetic information in plants. Philos. Trans. R. Soc. Land. B BioI. Sci. 339: 173-181. Kornberg R.D. and Thomas J.O. 1974. Chromatin structure; oligomers of the histones. Science 184: 865-868. Lachner M., O'Carroll D., Rea S., Mechtler K., and Jenuwein T. 2001. Methylation of histone H3 lysine 9 creates a binding site for HPI proteins. Nature 410: 116-120. Laskey R.A. and Gurdon J.B. 1970. Genetic content of adult somatic cells tested by nuclear transplantation from cultured cells. Nature 228: 1332-1334. Luger K., Mader A.W., Richmond R.K, Sargent D.P., and Richmond T.J. 1997. Crystal structure ofthe nucleosome core particle at 2.8 A resolution. Nature 389: 251-260. Lyon M.P. 1961. Gene action in the X-chromosome of the mouse. Nature 190: 372-373. McClintock B. 1965. The control of gene action in maize. Brookhaven Symp. BioI. 18: 162-184. McKittrick E., Gafken P.R., Ahmad K, and Henikoff S. 2004. Histone H3.3 is enriched in covalent modifications associated with active chromatin. Proc. Natl. Acad. Sci. 101: 1525-1530. Morgan T. 1911. An attempt to analyze the constitution of the chromo-

22 • C HAP

T ER 2

somes on the basis of sex-linked inheritance in Drosophila. f. Exp. Zool. 11: 365-414. Muller H.J. 1930. Types of visible variations induced by X-rays in Drosophila. f. Genet. 22: 299-334. Nakayama J., Rice J.c., Strahl B.D., Allis C.D., and Grewal S.l. 2001. Role of histone H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science 292: 110-113. Novick A. and Weiner M. 1957. Enzyme induction as an all-or-none phenomenon. Proc. Natl. Acad. Sci. 43: 553-566. Ohno S., Kaplan W.D., and Kinosita R. 1959. Formation of the sex chromatin by a single X-chromosome in liver cells of Rattus norvegicus. Exp. Cell Res. 18: 415-418. Paul J. and Gilmour R.S. 1968. Organ-specific restriction of transcription in mammalian chromatin. f. Mol. BioI. 34: 305-316. Pazin M,J., Kamakaka R.T., and Kadonaga J.T. 1994. ATP-dependent nucleosome reconfiguration and transcriptional activation from preassembled chromatin templates. Science 266: 2007-2011. Peterson c.L. and Herskowitz 1. 1992. Characterization of the yeast SWIl, SWI2, and SWI3 genes, which encode a global activator of transcription. Cell 68: 573-583. Ptashne M. 1992. A genetic switch: Phage A and higher organisms, 2nd edition. Blackwell Science, Malden, Massachusetts and Cell Press, Cambridge, Massachusetts. Richmond T.J., Finch J.T., Rushton B., Rhodes D., and Klug A. 1984. Structure of the nucleosome core particle at 7 A resolution. Nature 311: 532-537. Riggs A.D. 1975. X inactivation, differentiation, and DNA methylation. Cytogenet. Cell Genet. 14: 9-25. Riggs A.D. and Porter T.N. 1996. Overview of epigenetic mechanisms.

In Epigenetic mechanisms ofgene regulation (ed. Y.E.A. Russo et al.), pp. 29-45. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Riggs A.D., Martienssen R.A., and Russo Y.E.A. 1996. Introduction. In Epigenetic mechanisms of gene regulation (ed. Y.E.A. Russo et al.), pp. 1-4. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Shykind B.M., Rohani S.c., O'Donnell S., Nemes A., Mendelsohn M., Sun Y., Axel R., and Barnea G. 2004. Gene switching and the stability of odorant receptor gene choice. Cell 117: 801-815. Starn M., Belele c., Dorweiler J., and Chandler Y. 2002. Differential chromatin structure with a tandem array 100 kb upstream of the maize bl locus is associated with paramutation. Genes Dev. 16: 1906-1918. Stedman E. and Stedman E. 1950. Cell specificity of histones. Nature 166: 780-781. Sturtevant A. 1913. The linear arrangement of six sex-linked factors in Drosophila, as shown by their mode of association. f. Exp. Zool. 14: 43-59. Tsukiyama T. and Wu C. 1995. Purification and properties of an ATPdependent nucleosome remodeling factor. Cell 83: 1011-1020. Waddington C.H. 1953. Epigenetics and evolution. Symp. Soc. Exp. BioI. 7: 186-199. Wallis J.W., Hereford L., and Grunstein M. 1980. Histone H2B genes of yeast encode two different proteins. Cell 22: 799-805. Wysocka J., Swigut T., Milne T. Dou Y., Zhang X., Burlingame A., Roeder R., Brivanlou A., and Allis C.D. 2005. WDR5 associates with histone H3 methylated at K4 and is essential for H3 K4 methylation and vertebrate development. Cell 121: 859-872.

HAP

Overview and Concepts C. David Allis,l Thomas Jenuwein, 2 and Danny Reinberg 3 lThe Rockefeller University, New York, New York; 2Research Institute of Molecular Pathology, Vienna, Austria; 3UMDNJ-Robert Wood Johnson Medical School, Piscataway, New Jersey

CONTENTS 1. Genetics Versus Epigenetics, 25

10. RNAi and RNA-directed Gene Silencing, 42

2. Model Systems for the Study of Epigenetics, 26

11. From Unicellular to Multicellular Systems, 44

3. Defining Epigenetics, 28

12. Polycomb and Trithorax, 45

4. The Chromatin Template, 29

13. X Inactivation and Facultative Heterochromatin, 47

5. Higher-Order Chromatin Organization, 31

14. Reprogramming of Cell Fates, 49

6. The Distinction between Euchromatin and Heterochromatin, 34

15. Cancer, 50

7. Histone Modifications and the Histone Code, 36 8. Chromatin-remodeling Complexes and Histone Variants, 39

16. What Does Epigenetic Control Actually D07,52 17. Big Questions in Epigenetic Research, 55 References, 56

9. DNA Methylation, 41

GENERAL SUMMARY The DNA sequencing of the human genome and the genomes of many model organisms has generated considerable excitement within the biomedical community and the general public over the past several years. These genetic "blueprints" that exhibit the well-accepted rules of Mendelian inheritance are now readily available for close inspection, opening the door to improved understanding of human biology and disease. This knowledge is also generating renewed hope for novel therapeutic strategies and treatments. Many fundamental questions nonetheless remain. For example, how does normal development proceed, given that every cell has the same genetic information, yet follows a different developmental pathway, realized with exact temporal and spatial precision? How does a cell decide when to divide and differentiate, or when to retain an unchanged cellular identity, responding and expressing according to its normal developmental program? Mistakes made in the above processes can lead to the generation of disease states such as cancer. Are these mistakes encoded in faulty genetic blueprints that we inherited from one or both of our parents, or are there other layers of regulatory information that are not being properly read and decoded? In humans, the genetic information (DNA) is organized into 23 chromosome pairs consisting of approximately 25,000 genes. These chromosomes can be compared to libraries with different sets of books that together instruct the development of a complete human being. The DNA sequence of our genome is composed of about 3 x 109 bases, abbreviated by the four letters (or bases) A, C, G, and T within its sequence, giving rise to well-defined words (genes), sentences, chapters, and books. However, what dictates when the different books are read, and in what order, remains far from clear. Meeting this extraordinary challenge is likely to reveal insights into how cellular events are coordinated during normal and abnormal development. When summed across all chromosomes, the DNA molecule in higher eukaryotes is about 2 meters long and therefore needs to be maximally condensed about 1O,OOO-foid to fit into a cell's nucleus, the compartment of a cell that stores our genetic material. The wrapping of DNA around "spools" of proteins, so-called histone proteins, provides an elegant solution to this packaging problem, giving rise to a repeating protein:DNA polymer known as chromatin. However, in packaging DNA to better fit into a confined space, a problem develops, much as

when one packs too many books onto library shelves: It becomes harder to find and read the book of choice, and thus, an indexing system is needed. Chromatin, as a genome-organizing platform, provides this indexing. Chromatin is not uniform in structure; it comes in different packaging designs from a highly condensed chromatin fiber (known as heterochromatin) to a less compacted type where genes are typically expressed (known as euchromatin). Variation can enter into the basic chromatin polymer through the introduction of unusual histone proteins (known as histone variants), altered chromatin structures (known as chromatin remodeling), and the addition of chemical flags to the histone proteins themselves (known as covalent modifications). Moreover, addition of a methyl group directly to a cytosine (C) base in the DNA template (known as DNA methylation) can provide docking sites for proteins to alter the chromatin state or affect the covalent modification of resident histones. Recent evidence suggests that noncoding RNAs can "guide" specialized regions of the genome into more compacted chromatin states. Thus, chromatin should be viewed as a dynamic polymer that can index the genome and potentiate signals from the environment, ultimately determining which genes are expressed and which are not. Together, these regulatory options provide chromatin with an organizing principle for genomes known as "epigenetics," the subject of this book. In some cases, epigenetic indexing patterns appear to be inherited through cell divisions, providing cellular "memory" that may extend the heritable information potential of the genetic (DNA) code. Epigenetics can thus be narrowly defined as changes in gene transcription through modulation of chromatin, which is not brought about by changes in the DNA sequence. In this overview, we explain the basic concepts of chromatin and epigenetics, and we discuss how epigenetic control may give us the clues to solve some long-standing mysteries, such as cellular identity, tumorigenesis, stem cell plasticity, regeneration, and aging. As readers comb through the chapters that follow, we encourage them to note the wide range of biological phenomena uncovered in a diverse range of experimental models that seem to have an epigenetic (non-DNA) basis. Understanding how epigenetics operates in mechanistic terms will likely have important and far-reaching implications for human biology and human disease in this "post-genomic" era.

a v E R V lEW 1 Genetics Versus Epigenetics

Determining the structural details of the DNA double helix stands as one of the landmark discoveries in all of biology. DNA is the prime macromolecule that stores genetic information (Avery et a1. 1944), and it propagates this stored information to the next generation through the germ line. From this and other findings, the "central dogma" of modern biology emerged. This dogma encapsulates the processes involved in maintaining and translating the genetic template required for life. The essential stages are (1) the self-propagation of DNA by semiconservative replication; (2) transcription in a unidirectional 5' to 3' direction, templated by the genetic code (DNA), generation of an intermediary messenger RNA (mRNA); (3) translation of mRNA to produce polypeptides consisting of linear amino to carboxyl strings of amino acids that are colinear with the 5' to 3' order of DNA. In simple terms: DNA H RNA ~ protein. The central dogma accommodates feedback from RNA to DNA by the process of reverse transcription, followed by integration into existing DNA (as demonstrated by retroviruses and retrotransposons). However, this dogma disavows feedback from protein to DNA, although a new twist to the genetic dogma is that rare proteins, known as prions, can be inherited in the absence of a DNA or RNA template. Thus, these specialized self-aggregating proteins have properties that resemble some properties of DNA itself, including a mechanism for replication and information storage (Cohen and Prusiner 1998; Shorter and Lindquist 2005). Additionally, emerging evidence suggests that a remarkably large fraction of our genome is transcribed into "noncoding" RNAs. The function of these noncoding RNAs (i.e., non-protein-encoding except tRNAs, rRNAs, snoRNAs) is under active investigation and is only beginning to become clear in a limited number of cases. The origin of epigenetics stems from long-standing studies of seemingly anomalous (i.e., non-Mendelian) and disparate patterns of inheritance in many organisms (see Chapters 1 and 2 for a historical overview). Classic Mendelian inheritance of phenotypic traits (e.g., pea color, number of digits, or hemoglobin insufficiency) results from allelic differences caused by mutations of the DNA sequence. Collectively, mutations underlie the definition of phenotypic traits, which contributes to the determination of species boundaries. These boundaries are then shaped by the pressures of natural selection, as explained by Darwin's theory of evolution. Such concepts place mutations at the heart of classic genetics. In contrast, non-Mendelian inheritance (e.g., variation of embryonic growth, mosaic skin coloring, random X inac-

AND

CON C E P T 5

tivation, plant paramutation) (Fig. 1) can manifest, to take one example, from the expression of only one (of two) alleles within the same nuclear environment. Importantly, in these circ*mstances, the DNA sequence is not altered. This is distinct from another commonly referred to non-Mendelian inheritance pattern that arises from the maternal inheritance of mitochondria (Birky 2001). The challenge for epigenetic research is captured by the selective regulation of one allele within a nucleus. What distinguishes two identical alleles, and how is this distinction mechanistically established and maintained through successive cell generations? What underlies differences observed in monozygotic ("identical") twins that make them not totally identical? Epigenetics is sometimes cited as one explanation for the differences in outward traits, by translating the influence of the environment, diet, and potentially other external sources to the expression of the genome (Klar 2004; see Chapters 23 and 24). Determining what components are affected at a molecular level, and how alterations in these components affect human biology and human disease, is a major challenge for future studies. Another key question in the field is, How important is the contribution of epigenetic information for normal development? How do normal pathways become dysfunctional, leading to abnormal development and neoplastic transformation (i.e., cancer)? As mentioned above, "identical" twins share the same DNA sequence, and as such, their phenotypic identity is often used to underscore the defining power of genetics. However, even twins such as these can exhibit outward phenotypic differences, likely imparted by epigenetic modifications that occur over the lifetime of the individuals (Fraga et a1. 2005). Thus, the extent to which epigenetics is important in defining cell fate, identity, and phenotype remains to be fully understood. In the case of tissue regeneration and aging, it remains unclear whether these processes are dictated by alterations in the genetic program of cells or by epigenetic modifications. The intensity of research on a global scale testifies to the recognition that the field of epigenetics is a critical new frontier in this post-genomic era. In the words of others, "We are more than the sum of our genes" (Klar 1998), or "You can inherit something beyond the DNA sequence. That's where the real excitement in genetics is now" (Watson 2003). The overriding motivation for deciding to edit this book was the general belief that we and all the contributors to this volume could transmit this excitement to future generations of students, scientists, and physicians, most of whom were taught genetic, but not epigenetic, principles governing inheritance and chromosome segregation.

26 •

C HAP T E R 3

Figure 1. Biological Examples of Epigenetic Phenotypes

Barr body polytene chromosomes

twins

epigenetic biology yeast mating types

cloned cat

mutant plant

~ lEU]

blood smear

tumor tissue

2 Model Systems for the Study of Epigenetics

The study of epigenetics necessarily requires good experimental models, and as often is the case, these models seem at first sight far removed from studies using human (or mammalian) cells. Collectively, however, results from many systems have yielded a wealth of knowledge. The historical overviews (Chapters 1 and 2) make reference to several important landmark discoveries that have emerged from early cytology, the growth of genetics, the birth of molecular biology, and relatively new advances in chromatinmediated gene regulation. Different model organisms (Fig. 2) have been pivotal in addressing and solving the various questions raised by epigenetic research. Indeed, seemingly disparate epigenetic discoveries made in various model organisms have served to unite the research community. The purpose of this section is to highlight some of these major findings, which are discussed in more detail in the following chapters of this book. As readers note these discoveries, they should focus on the fundamental principles that investigations using these model systems have exposed; their collective contributions point more often to common concepts than to diverging details. Unicellular and "lower" eukaryotic organisms-Saccharomyces cerevisiae, Schizosaccharomyces pombe, and

Epigenetic phenotypes in a range of organisms and cell types, all attributable to non-genetic differences. Twins: Slight variations partially attributable to epigenetics (© Randy Harris, New York). Barr body: The epigenetically silenced X chromosome in female mammalian cells, visible cytologically as condensed heterochromatin. Polytene chromosomes: Giant chromosomes in Drosophila salivary glands, ideally suited for correlating genes with epigenetic marks (reprinted from Schotta et al. 2003 [©SpringerD. Yeast mating type: Sex is determined by the active MAT locus, while copies of both mating-type genes are epigenetically silenced (©Alan Wheals, University of Bath). Blood smear: Heterogeneous cells of the same genotype, but epigenetically determined to serve different functions (courtesy Prof. Christian Sillaber). Tumor tissue: Metastatic cells (left) showing elevated levels of epigenetic marks in the tissue section (reprinted, with permission, from Seligson et al. 2005 [©Macmillan]). Mutant plant: Arabidopsis flower epiphenotypes, genetically identical, with epigenetically caused mutations (reprinted, with permission, from Jackson et al. 2002 [©MacmillanD. Cloned cat: Genetically identical, but with varying coat-color phenotype (reprinted, with permission, from Shin et al. 2002 [©Macmillan]).

Neurospora crassa-permit powerful genetic analyses, in part facilitated by a short life cycle. Mating-type (MAT) switching that occurs in S. cerevisiae (Chapter 3) and S. pombe (Chapter 6) has provided remarkably instructive examples, demonstrating the importance of chromatinmediated gene control. In the budding yeast S. cerevisiae, the unique silent information regulator (SIR) proteins were shown to engage specific modified histones. This was preceded by elegant experiments using genetics to document the active participation of histone proteins in gene regulation (Clark-Adams et al. 1988; Kayne et al. 1988). In the fission yeast S. pombe, the patterns of histone modification operating as activating and repressing signals are remarkably similar to those in metazoan organisms. This has opened the door for powerful genetic screens being employed to look for gene products that suppress or enhance the silencing of genes. Most recently, a wealth of mechanistic insights linking the RNA interference (RNAi) machinery to the induction of histone modifications acting to repress gene expression was discovered in fission yeast (Hall et al. 2002; Volpe et al. 2002). Shortly afterward, the RNAi machinery was also implicated in transcriptional gene silencing in the plant Arabidopsis thaliana, underscoring the potential importance of this regulation in a wide range of organisms (see Section 10).

a v E R V lEW

AND

CON C E P T S

S.cerevisiae

ospomb' epigenetic model organisms

c. elegans

Tetrahymena

maize

Arabidopsis

Other "off-beat" organisms have also made disproportionate contributions toward unraveling epigenetic pathways that at first seemed peculiar. The fungal species, N. crassa, revealed the unusual non-Mendelian phenomenon of repeat-induced point mutation (RIP) as a model for studying epigenetic control (Chapter 6). Later, this organism was used to demonstrate the first functional connection between histone modifications and DNA methylation (Tamaru and Selker 2001), a finding later extended to "higher" organisms (Jackson et al. 2002). Ciliated protozoa, such as Tetrahymena and Paramecium, commonly used in biology laboratories as convenient microscopy specimens, facilitated important epigenetic discoveries because of their unique nuclear dimorphism. Each cell carries two nuclei: a somatic macronucleus that is transcriptionally active, and a germ-line micronucleus that is transcriptionally inactive. Using macronuclei as an enriched starting source of "active" chromatin, the biochemical purification of the first nuclear histone-modifying enzyme-a histone acetyltransferase or HAT-was made (Brownell et al. 1996). Ciliates are also well known for their peculiar phenomenon of programmed DNA elimination during their sexual life cycle, triggered by small noncoding RNAs and histone modifications (Chapter 7). In multicellular organisms, genome size and organismal complexity generally increase from invertebrate (Caenorhabditis elegans, Drosophila melanogaster) or

Figure 2. Model Organisms Used in Epigenetic Research Schematic representation of model organisms used in epigenetic research. S. cerevisiae: Mating-type switching to study epigenetic chromatin control. S. pombe: Variegated gene silencing manifests as colony sectoring. Neurospora crassa: Epigenetic genome defense systems include repeat-induced point mutation, quelling, and meiotic silencing of unpaired DNA, revealing an interplay between RNAi pathways, DNA and histone methylation. Tetrahymena: Chromatin in somatic and germ-line nuclei are distinguished byepigenetically regulated mechanisms. Arabidopsis: Model for repression by DNA, histone, and RNA-guided silencing mechanisms. Maize: Model for imprinting, para mutation, and transposon-induced gene silencing. C. elegans: Epigenetic regulation in the germ line. Drosophila: Position-effect variegation (PEV) manifest by clonal patches of expression and silencing of the white gene in the eye. Mammals: X-chromosome inactivation.

plant (A. thaliana) species to "higher," and to some, "more relevant," vertebrate organisms (mammals). Plants, however, have been pivotal to the field of epigenetics, providing a particularly rich source of epigenetic discoveries (Chapter 9) ranging from transposable elements and paramutation (McClintock 1951) to the first description of noncoding RNAs involved in transcriptional silencing (Ratcliff et al. 1997). Crucial links between DNA methylation, histone modification, and components of the RNAi machinery came through plant studies. The discovery of plant epialleles, with comic names such as SUPERMAN and KRYPTONITE (e.g., Jackson et al. 2002), and several vernalizing genes (Bastow et al. 2004; Sung and Amasino 2004) have further provided the research field with insights into understanding the developmental role of epigenetics and cellular memory. Plant meristem cells have also offered the opportunity to study crucial questions such as somatic regeneration and stem cell plasticity (see Chapters 9 and 11). For understanding animal development, Drosophila has been an early and continuous genetic powerhouse. Based on the pioneering work of Muller (1930), many developmental mutations were generated, including the homeotic transformations and position-effect variegation (PEV) mutants explained below (also see Chapter 5). The homeotic transformation mutants led to the idea that there could be regulatory mechanisms for establishing and

C HAP T E R 3

maintammg cellular identity/memory which was later shown to be regulated by the Polycomb and trithorax systems (see Chapters 11 and 12). For PEV, gene activity is dictated by the surrounding chromatin structure and not by primary DNA sequence. This system has been a particularly informative source for dissecting factors involved in epigenetic control (Chapter 5). Over 100 suppressors of variegation [Su(var)] genes are believed to encode components of heterochromatin. Without the foundation established by these landmark studies, the discovery of the first histone lysine methyltransferases (HKMTs) (Rea et al. 2000) and the resultant advances in histone lysine methylation would not have been possible. As is often the case in biology, comparable screens have been carried out in fission yeast and in plants, identifying silencing mutants with functional conservation with the Drosophila Su(var) genes. The use of reverse genetics via RNAi libraries in the nematode worm C. elegans has contributed to our understanding of epigenetic regulation in metazoan development. There, comprehensive cell-fate tracking studies, detailing all the developmental pathways of each cell, have highlighted the fact that Polycomb and trithorax systems probably arose with the emergence of multicellularity (see Sections 12 and 13). In particular, these mechanisms of epigenetic control are essential for gene regulation in the germ line (see Chapter 15). The role of epigenetics in mammalian development has mostly been elucidated in the mouse, although a number of studies have been translated to diverse human cell lines and primary cell cultures. The advent of gene "knock-out" and "knock-in" technologies has been instrumental for the functional dissection of key epigenetic regulators. For instance, the Dnmtl DNA methyltransferase mutant mouse provided functional insight for the role of DNA methylation in mammals (Li et al. 1992). It is embryonic-lethal and shows impaired imprinting (see Chapter 18). Disruption of DNA methylation has also been shown to cause genomic instability and reanimation of transposon activity, particularly in germ cells (Walsh et al. 1998; Bourc'his and Bestor 2004). There are approximately 100 characterized chromatin-regulating factors (i.e., histone and DNA-modifying enzymes, components of nucleosome remodeling complexes and of the RNAi machinery) that have been disrupted in the mouse. The mutant phenotypes affect cell proliferation, lineage commitment, stem cell plasticity, genomic stability, DNA repair, and chromosome segregation processes, in both somatic and germ cell lineages. Not surprisingly, most of these mutants are also involved in disease development and cancer. Thus, many of the key advances in epigenetic

control took advantage of unique biological features exhibited by many, if not all, of the above-mentioned model organisms. Without these biological processes and the functional analyses (genetic and biochemical) that delved into them, many of the recent advances in epigenetic control would have remained elusive. 3 Defining Epigenetics

The above discussion begs the question, What is the common thread that allows diverse eukaryotic organisms to be connected with respect to fundamental epigenetic principles? Different epigenetic phenomena are linked largely by the fact that DNA is not "naked" in all organisms that maintain a true nucleus (eukaryotes). Instead, the DNA exists as an intimate complex with specialized proteins, which together comprise chromatin. In its simplest form, chromatin-i.e., DNA spooled around nucleosomal units consisting of small histone proteins (Kornberg 1974)-was initially regarded as a passive packaging molecule to wrap and organize the DNA. Distinctive forms of chromatin arise, however, through an array of covalent and non-covalent mechanisms that are being uncovered at a rapid pace (see Section 6). This includes a plethora of posttranslational histone modifications, energy-dependent chromatin-remodeling steps that mobilize or alter nucleosome structures, the dynamic shuffling of new histones (variants) in and out of nucleosomes, and the targeting role of small noncoding RNAs. DNA itself can also be modified covalently in many higher eukaryotes, by methylation at the cytosine residue, usually but not always, of CpG dinucleotides. Together, these mechanisms provide a set of interrelated pathways that all create variation in the chromatin polymer (Fig. 3). Many, but not all, of these modifications and chromatin changes are reversible and, therefore, are unlikely to be propagated through the germ line. Transitory marks are attractive because they impose changes to the chromatin template in response to intrinsic and external stimuli (Jaenisch and Bird 2003), and in so doing, regulate the access and/or processivity of the transcriptional machinery, needed to "read" the underlying DNA template (Sims et al. 2004; Chapter 10). Some histone modifications (like lysine methylation), methylated DNA regions, and altered nucleosome structures can, however, be stable through several cell divisions. This establishes "epigenetic states" or means of achieving cellular memory, which remain poorly appreciated or understood. From this perspective, chromatin "signatures" can be viewed as a higWy organized system of information storage that can index distinct regions

GENETICS

EPIGENETICS

muta';o", mod

remodeler

ncRNAs

inherited

stable?

germ line

soma

species

variability

Figure 3. Genetics Versus Epigenetics GENETICS: Mutations (red stars) of the DNA template (green helix) are heritable somatically and through the germ line. EPIGENETlCS: Variations in chromatin structure modulate the use of the genome by (1) histone modifications (mod), (2) chromatin remodeling (remodeler), (3) histone variant composition (yellow nuc/eosome), (4) DNA methylation (Me), and (5) noncoding RNAs. Marks on the chromatin template may be heritable through cell division and collectively contribute to determining cellular phenotype.

of the genome and accommodate a response to environmental signals that dictate gene expression programs. The significance of having a chromatin template that can potentiate the genetic information is that it provides multidimensional layers to the readout of DNA. This is perhaps a necessity, given the vast size and complexity of the eukaryotic genome, particularly for multicellular organisms (see Section 11 for further details). In such organisms, a fertilized egg progresses through development, starting with a single genome that becomes epigenetically programmed to generate a multitude of distinct "epigenomes" in more than 200 different types of cells (Fig. 4). This programmed variation has been proposed to constitute an "epigenetic code" that significantly extends the information potential of the genetic code (StraW and Allis 2000; Turner 2000; Jenuwein and Allis 2001). Although this is an attractive hypothesis, we stress that more work is needed to test this and related provocative theories. Other alternative viewpoints are being advanced which argue that clear combinatorial "codes;' lilee the triplet genetic code, are not lil<.ely in histones or are far from established (Schreiber and Bernstein 2002; Henil<.off 2005). Despite these uncertainties, we favor the general view that a combination of covalent and non-covalent mechanisms will act to create chromatin states that can be

V E R V lEW

AND

CON C E P T 5

templated through cell division and development by mechanisms that are just beginning to be defined. Exactly how these altered chromatin states are faithfully propagated during DNA replication and mitosis remains one of the fundamental challenges of future studies. The phenotypic alterations that occur from cell to cell during the course of development in a multicellular organism were described by Waddington as the "epigenetic landscape" (Waddington 1957). Yet the spectrum of cells, from stem cells to fully differentiated cells, all share identical DNA sequences but differ remarkably in the profile of genes that they actually express. With this knowledge, epigenetics later came to be defined as the "Nuclear inheritance which is not based on differences in DNA sequence" (Holliday 1994). Since the discovery of the DNA double helix and the early explanations of epigenetics, our understanding of epigenetic control and its underlying mechanisms has greatly increased, causing some to describe it in more lofty terms as a "field" rather than just "phenomena" (see Wolffe and Matzke 1999; Roloff and Nuber 2005; Chapter 1). In the past decade, considerable progress has been gained regarding the many enzyme families that actively modify chromatin (see below). Thus, in today's modern terms, epigenetics can be molecularly (mechanistically) defined as "The sum of the alterations to the chromatin template that collectively establish and propagate different patterns of gene expression (transcription) and silencing from the same genome." 4 The Chromatin Template

The nucleosome is the fundamental repeating unit of chromatin (Kornberg 1974). On the one hand, the basic chromatin unit consists of a protein octamer containing two molecules of each canonical (or core) histone (H2A, H2B, H3, and H4), around which is wrapped 147 bp of DNA. Detailed intermolecular interactions between the core histones and the DNA were determined from landmark studies leading to an atomic (2.8 A) resolution X-ray picture of the nucleosome assembled from recombinant parts (Fig. 5) (Luger et al. 1997). Higher-resolution images of mononucleosomes, as well as emerging higherorder structures (tetranucleosomes) (Schalch et al. 2005), continue to capture our attention, promising to better explain the physiologically relevant substrate upon which most, if not all, of the chromatin remodeling and transcriptional machinery operates. The core histone proteins that make up the nucleosome are small and higWy basic. They are composed of a

30 • C HAP

T ER

1 genome DNA

stored information

chromatin

organized information Figure 4. DNA Versus Chromatin

epigenomes

~~\ >25,000 genes identical DNA sequence

@....,-.--.

...

\???

Jt<

globular domain and flexible (relatively unstructured) "histone tails," which protrude from the surface of the nucleosome (Fig. 5). Based on amino acid sequence, histone proteins are highly conserved from yeast to humans. Such a high degree of conservation lends support to the general view that these proteins, even the unstructured tail domains, are likely to serve critical functions. The tails, particularly of histones H3 and H4, in fact hold important clues to nucleosomal variability (and hence chromatin), as many of the residues are subject to extensive posttranslational modifications (see back end paper for standard nomenclature used in this textbook and Appendix 2 for a listing of known histone modifications). Acetylation and methylation of core histones, notably H3 and H4, were among the first covalent modifications to be described, and were long proposed to correlate with positive and negative changes in transcriptional activity. Since the pioneering studies of Allfrey and coworkers (Allfrey et al. 1964), many types of covalent histone modifications have been identified and characterized; these include histone phosphorylation, ubiquitination, sumoylation, ADP-ribosylation, biotinylation, proline isomerization, and likely others that await description (Vaquero et al. 2003). These modifications occur at specific sites and residues, some of which are illustrated in Figure 6 and listed in Appendix 2. Specific enzymes and enzymatic complexes, some of which are highlighted in the follow-

>200 different cell types

???

The genome: Invariant DNA sequence (green double helix) of an individual. The epigenome: The overall chromatin composition, which indexes the entire genome in any given cell. It varies according to cell type, and response to internal and external signals it receives. (Lower pane{) Epigenome diversification occurs during development in multicellular organisms as differentiation proceeds from a single stem cell (the fertilized embryo) to more committed cells. Reversal of differentiation or transdifferentiation (blue lines) requires the reprogramming of the cell's epigenome.

ing overview and individual chapters, catalyze these covalent markings. Because these lists will continue to grow in years to come, our intent was to mention only individual marks and enzymes that can illustrate what we feel are important general concepts and principles. In certain chromatin regions, nucleosomes may contain histone variant proteins in place of a core (canonical) histone. Ongoing research is showing that this compositional difference contributes to marking regions of the

Figure 5. Nucleosome Structure (Left) A 2.8 A model of a nucleosome. (Right) A schematic representation of histone organization within the octamer core around which the DNA (black line) is wrapped. Nucleosome formation occurs first through the deposition of an H3/H4 tetramer on the DNA, followed by two sets of H2A/H2B dimers. Unstructured aminoterminal histone tails extrude from the nucleosome core, which consists of structured globular domains of the eight histone proteins.

OVERVIEW

chromosomes for specialized functions. Variant proteins for core histones H2A and H3 are currently known, but none exists for histones H2B and H4. We suspect that histone variants, although often minor in terms of amount and accordingly more difficult to study, are bountiful in the information they contain and essential to contributing to epigenetic regulation (for more detail, see Section 8 and Chapter 13). 5 Higher-Order Chromatin Organization

Chromatin, the DNA-nucleosome polymer, is a dynamic molecule existing in many configurations. Historically, chromatin has been classified as either euchromatic or heterochromatic, stemming from the nuclear staining patterns of dyes used by cytologists to visualize DNA. Euchromatin is decondensed chromatin, although it may be transcriptionally active or inactive. Heterochromatin can broadly be defined as highly compacted and silenced chromatin. It may exist as permanently silent chromatin (constitutive heterochromatin), where genes will rarely be expressed in any cell type of the organism, or repressed (facultative heterochromatin) in some cells during a specific cell cycle or developmental stage. Thus, there is a spectrum of chromatin states and a longstanding literature suggesting that chromatin is a highly dynamic macromolecular structure, prone to remodel-

CONCEPTS.

phosphorylation

KPH

histone H3

135 aa

G GKGGKGLGKGGAKRHR VLRDNIQGITKPAIRRLAR

histone H4

102 aa

~A~~~:lIAPATGGV

~-;;9;-------

acetylation

methylation (arginine)

methylation (active lysine)

.....- - - - - - - '

GRG5QGG~ARAKAKSRSSRAGLQFPVGRVHRLLRKGNY

129 aa

methylation (repressive lysine) Ub

ubiquitylation

PEPAKSAPAPK~G

ing and restructuring as it receives physiologically relevant input from upstream signaling pathways. Only recently, however, has excellent progress been made unraveling molecular mechanisms that govern these remodeling steps. The textbook, ll-nm "beads on a string" template represents an active and largely "unfolded" interphase configuration wherein DNA is periodically wrapped around repeating units of nucleosomes (Fig. 7). The chromatin fiber, however, is not always made up of regularly spaced nucleosomal arrays. Nucleosomes may be irregularly packed and fold into higher-order structures that are only beginning to be observed at atomic resolution (Khorasanizadeh 2004). Differential and higher-order chromatin conformations occur in diverse regions of the genome during cell-fate specification or in distinct stages of the cell cycle (interphase versus mitotic chromatin). The arrangement of nucleosomes on the ll-nm template can be altered by cis-effects and trans-effects of covalently modified histone tails (Fig. 8). cis-Effects are brought about by changes in the physical properties of modified histone tails, such as a modulation in the electrostatic charge or tail structure that, in turn, alters internucleosomal contacts. A well-known example, histone acetylation, has long been suspected to neutralize positive charges of highly basic histone tails, generating a localized expansion of the chromatin fiber, thereby enabling better access of

Me M

AND

KKAVTKAQKKDSKKRKRSRKESYSV

125 aa

Figure 6. Sites of Histone Tail Modifications The amino-terminal tails of histones account for a quarter of the nucleosome mass. They host the vast majority of known covalent modification sites as illustrated. Modifications do also occur in the globular domain (boxed), some of which are indicated. In general, active marks include acetylation (turquoise Ac flag), arginine methylation (yellow Me hexagon), and some lysine methylation such as H3K4 and H3K36 (green Me hexagon). H3K79 in the globular domain has anti-silencing function. Repressive marks include H3K9, H3K27, and H4K20 (red Me hexagon), Green = active mark, red = repressive mark.

32 • C HAP

T ER

nucleosomes length: 2 m

_____

~ DNA]

11 nm

h·Istone mold'f" I Icatlons histone H1

fffJ!JB

]

30nm

]

300-700 nm

]

1.5~m

domain organization

f"'~ I

mitotic condensation

j length: 10

~m BIi~ chromosome

Figure 7. Higher-Order Structuring of Chromatin The ll-nm fiber represents DNA wrapped around nucleosomes. The 30-nm fiber is further compacted into an as-yet-unconfirmed structure (illustrated as solenoid conformation here), involving linker histone Hl. The 300-700-nm fiber represents dynamic higher-order looping that occurs in both interphase and metaphase chromatin. The 1.5-!.Lm condensed chromosome represents the most compacted form of chromatin that occurs only during nuclear division (mitosis or meiosis). It is not yet clear how mitotic chromosome-banding patterns (i.e., G- or R-banding) correlate with particular chromatin structures.

transcription machinery to the DNA double helix. Phosphorylation, through the addition of net negative charge, can generate "charge patches" (Dou and Gorovsky 2000) that are believed to alter nucleosome packaging or to expose histone amino termini by altering the higher-order folded state of the chromatin polymer (Wei et al. 1999; Nowak and Corces 2004). In much the same way, linker histones (HI) are believed to promote the packaging of higher-order fibers by shielding the negative charge of linker DNA between adjacent nucleosomes (Thomas 1999; Khochbin 2001; Harvey and Downs 2004; Kimmins and Sassone-Corsi 2005). The addition of bulky adducts, such as ubiquitin and ADP-ribose, may also induce different arrangements of the histone tails and open up nucleosome arrays. The extent to which histone tails can induce chromatin compaction through modification-dependent and -independent mechanisms is not clear. Histone modifications may also elicit what we refer to as trans-effects by the recruitment of modification-bind-

ing partners to the chromatin. This can be viewed as "reading" a particular covalent histone mark in a context-dependent fashion. Certain binding partners have a particular affinity and hence are known to "dock" onto specific histone tails and often do so by serving as the chromatin "Velcro" for one polypeptide within a much larger enzymatic complex that needs to engage the chromatin polymer. For instance, the bromodomain-a motif that recognizes acetylated histone residues-is often, but not always, part of a histone acetyltransferase (HAT) enzyme that exists to acetylate target histones (see Fig. 10 in Section 7) as part of a larger chromatin-remodeling complex (Dhalluin et al. 1999; Jacobson et al. 2000). Similarly, methylated lysine residues embedded in histone tails can be read by chromodomains (Bannister et al. 2001; Lachner et al. 2001; Nakayama et al. 2001) or similar domains (e.g., MBT, tudor) (Maurer-Stroh et al. 2003; Kim et al. 2006) to facilitate downstream chromatin-modulating events. In some cases, for instance, the association of chromodomain proteins precipitates the spreading of heterochromatin by the histone methyltransferase (HKMT)-catalyzed methylation of adjacent histones which can then be read by chromodomain proteins (Chapter 5). Histone modifications of both the tail regions and the globular core region (Cosgrove et al. 2004) can also target ATP-dependent remodeling complexes to the II-nm fiber required for the transition from poised euchromatin to a transcriptionally active state. This mobilization of nucleosomes may occur by octamer sliding, alteration of nucleosome structure by DNA looping (for more detail, see Chapter 12) or replacement of specific core histones with histone variants (Chapter 13). ATP-dependent chromatin remodelers (such as SWI/SNF, an historically important example) hydrolyze energy to bring about significant changes in histone:DNA contacts, resulting in looping, twisting, and sliding of nucleosomes. These non-covalent mechanisms have been shown to be critically important for gene regulatory events (Narlikar et al. 2002) as much as those involving covalent histone modifications (see Chapter 10). The finding that specific ATP-dependent remodelers can shuffle histone variants into and out of chromatin provides a means to link cis, trans, and remodeling mechanisms. Understanding, in turn, how these interconnected mechanisms act in a concerted fashion to vary epigenetic states in chromatin is far from complete. More compact and repressive higher-order chromatin structures (30-nm) can also be achieved through the recruitment of linker histone HI and/or modificationdependent or "architectural" chromatin-associated factors

cis-effects

V E R V lEW

AND

CON C E P T S

•

mmodification binder

mod

Figure 8. Transitions in the Chromatin Template

binder

( cis/trans) cis-effects: A covalent modification of a histone tail

trans-effects

histone variant mod

histone replacement

regular histone

such as heterochromatin protein 1 (HPl) or Polycomb (PC). Although it is commonly held that compaction of nucleosomal chromatin (ll-nm) into a 30-nm transcriptionally incompetent conformation is accomplished by the incorporation of linker histone HI during interphase, the functional and structural dissection of this histone has, until recently, been difficult (Fan et al. 2005). One likely problem underlying these studies is the fact that histone HI occurs as different isoforms (~8 in mammals), making it difficult to do detailed genetic analyses. Thus, there is redundancy between some HI isoforms whereas others may hold tissue-specific functions (Kimmins and SassoneCorsi 2005). Interestingly, HI itself can be covalently modified (phosphorylated, methylated, poly(ADP) ribosylated, etc.), raising the possibility that cis and trans mechanisms currently being dissected on core histones may well extend to this important class of linker histone, and also to nonhistone proteins (Sterner and Berger 2000). Considerable debate has taken place over the details of the way in which the 30-nm chromatin fiber is organized. In general, either "solenoid" (one-start helix) models, wherein the nucleosomes are gradually coiled around a central axis (6-8 nucleosomes/turn), or more open "zigzag" models, which adopt higher-order self-assemblies (two-start helix), have been described. New evidence, including that collected from X-ray structure using a model system containing four nucleosomes, suggests a fiber arrangement more consistent with a two-start, zigzag arrangement of linker DNA connecting two stacks of nucleosome particles (Khorasanizadeh 2004; Schalch et al. 2005). Despite this progress, we note that linker his-

residue results in an altered structure or charge that manifests as a change in chromatin organization. trans-effects: The enzymatic modification of a histone tail residue (e.g., H3K9 methylation) results in an affinity for chromatin-associated protein (mod binder, e.g., HP1). The association of a mod binder (or associated protein complexes) causes downstream alterations in chromatin structure. Histone replacement: A covalent histone modification (or other stimulus) can signal the replacement of a core histone with a histone variant through a nucleosome-remodeling exchanger complex.

tone is not present in the current structures, and even if it were present, the 30-nm chromatin fiber compacts the DNA only approximately 50-fold. Thus, considerably more levels of higher-order chromatin organization exist that have yet to be resolved outside of light- and electronmicroscopic examination, whether leading to interphase or mitotic chromatin states. Despite structural uncertainties, recent results in living cells have now established the existence of multiple levels of chromatin folding above the 30-nm fiber within interphase chromosomes. A noteworthy advance was the development of new approaches to label specific DNA sequences in live cells, making it possible to study the dynamics of chromatin opening and closing in vivo in real time. Interestingly, these results reveal a dynamic interplay of positive and negative chromatin-remodeling factors in setting higher-order chromatin structures for states more or less compatible with gene expression (Fisher and Merkenschlager 2002; Felsenfeld and Groudine 2003; Misteli 2004). Organization into larger looped chromatin domains (300-700 nm) occurs, perhaps through anchoring the chromatin fiber to the nuclear periphery or other nuclear scaffolds via chromatin-associated proteins such as nuclear lamins. The extent to which these associations give rise to meaningful functional "chromosome territories" remains unclear, but numerous reports are showing that this concept deserves serious attention. For instance, clustering of multiple active chromatin sites to RNA polymerase II (RNA pol II) transcription factors has been observed, and similar concepts seem to apply to the clustering around replicating DNA and DNA polymerase. In

34 •

C HAP T E R 3

contrast, clustering of "silent" heterochromatin (particularly pericentromeric foci) and genes localized in trans has also been documented (see Chapters 4 and 21). How these associations are controlled and the extent to which nuclear localization of chromatin domains affects genome regulation are not yet clear. There is, nonetheless, an increasing body of evidence showing correlations of an active or silent chromatin configuration with a particular nuclear territory (Cremer and Cremer 2001; Gilbert et al. 2004; Janicki et al. 2004; Chakalova et al. 2005). The most condensed DNA structure is observed during the metaphase stage of mitosis or meiosis. This permits the faithful segregation of exact copies of our genome (one or two copies of each chromosome, depending on the division at hand), via chromosomes, to each daughter cell. This condensation involves a dramatic restructuring of the DNA from a 2-m molecule when fully extended, into discrete chromosomes measuring on average 1.5 )..lm in diameter (Fig. 7). This is no less than a 1O,000-fold compaction and is achieved by the hyperphosphorylation oflinker (HI) and core histone H3, and the ATP-dependent action of the condensin and cohesin complexes, and topoisomerase II. Exactly how non-histone complexes engage mitotic chromatin (or M-phase chromatin modifications), and what rules dictate their association and release from chromatin in a cell-cycle-regulated fashion, remain to be determined (Bernard et al. 2001; Watanabe et al. 2001). Here, the well-known mitotic phosphorylation of histone H3 (i.e., serines 10 and 28) and members of the HI family may provide important clues, but genetic and biochemical experiments have yet to yield full insights into what the function of these mitotic marks is. Interestingly, a formal theory has been proposed that specific methylation marks, when paired with more dynamic and reversible phosphorylation marks, may act as a "binary switch" in histone proteins, governing the binding and release of downstream effectors that engage the chromatin template (Fischle et al. 2003a). Using HPI binding to methylated histone H3 on lysine 9 (H3K9me) and mitotic serine 10 phosphorylation (H3SlOph) as a paradigm, evidence in support of a mitotic "methyl/phos switch" has recently been provided (Daujat et al. 2005; Fischle et al. 2005; Hirota et al. 2005). Specialized chromosomal domains, such as telorneres and centromeres, serve distinct functions dedicated to proper chromosome dynamics. Telomeres act as chromosomal ends, providing protection and unique solutions to how the very ends of DNA molecules are replicated. Centromeres provide an attachment anchor for spindle microtubules during nuclear division. Both of these specialized

domains have a fundamental role in the events that lead to faithful chromosome segregation. Interestingly, both telomeric and centromeric heterochromatin is distinguishable from euchromatin, and even other heterochromatic regions (see below), by the presence of unique chromatin structures that are largely repressive for gene activity and recombination. Moving expressed genes from their normal positions in euchromatin to new positions at or near centromeric and telomeric heterochromatin (see Chapters 4-6) can silence these genes, giving rise to powerful screens described earlier that sought to identify suppressors or enhancers of position-effect variegation (PEV) or telomere-position effects (TPE; Gottschling et al. 1990; Aparicio et al. 1991). Centromeres and telomeres have molecular signatures that include, for example, hypoacetylated histones. Interestingly, centromeres are also "marked" by the presence of the histone variant CENP-A, which plays an active role in chromosome segregation (Chapter 14). Thus, the proper assembly and maintenance of distinct centromeric and pericentromeric heterochromatin is critical for the completion of mitosis or meiosis, and hence, cellular viability. In addition to the well-studied centromeric and pericentromeric forms of constitutive heterochromatin, progress is being made into mechanisms of epigenetic control for centromeric (and telomeric) "identity." Clever experiments have shown that "neocentromeres" can function in place of normal centromeres, demonstrating that DNA sequences do not dictate the identity of centromeres (Chapters 13 and 14). Instead, epigenetic hallmarks, including centromere-specific modification patterns and histone variants, mark this specialized chromosomal domain. Considerable progress is being made into how other coding, noncoding, and repetitive regions of chromatin contribute to these epigenetic signatures. How any of these mechanisms relate, if at all, to chromosomal banding patterns is not known, but remains an intriguing possibility. Achieving an understanding of the epigenetic regulation of these portions of unique chromosomal regions is needed, highlighted by the fact that numerous human cancers are characterized by genomic instability, which is a hallmark of certain disease progression and neoplasia. 6 The Distinction between Euchromatin and Heterochromatin

This overview has been divided into discussions of euchromatin and heterochromatin, although we acknowledge that multiple forms of both classes of chromatins exist. Euchromatin, or "active" chromatin, consists

a v E R V lEW largely of coding sequences, which only account for a small fraction (less than 4%) of the genome in mammals. What molecular signals then mark coding sequences with the potential for productive transcription, and how does chromatin structure contribute to the process? An extensive literature has suggested that euchromatin exists in an "open" (decompacted), more nuclease-sensitive configuration, making it "poised" for gene expression, although not necessarily transcriptionally active. Some of the genes are ubiquitously expressed (housekeeping genes); others are developmentally regulated or stressinduced in response to environmental cues. The cooperation of selected cis-acting DNA sequences (promoters, enhancers, and locus control regions), bound by combinations of trans-acting factors, triggers gene transcription in concert with RNA polymerase and associated factors (Sims et a1. 2004). Together these factors have been highly selected during evolution to orchestrate an elaborate series of biochemical reactions that must occur in the appropriate spatial and temporal setting. Does chromatin provide an "indexing system" which better ensures that the above machinery can access its target sequences in the appropriate cell type? At the DNA level, the AT-rich vicinity of promoters is often devoid of nucleosomes and may exist in a rigid noncanonical B-form DNA configuration, promoting transcription factor (TF) occupancy (Mito et a1. 2005; Sekinger et a1. 2005). However, TF occupancy is not enough to ensure transcription. The recruitment of nucleosome-remodeling machines, through the induction of activating histone modifications (e.g., acetylation and H3K4 methylation), facilitates the engagement of the transcription machinery by pathways that are currently being defined (Fig. 9 and Chapter 10). Exchange of displaced histones with histone variants after the transcription machinery has unraveled and transcribed the chromatin fiber ensures integrity of the chromatin template (Ahmad and Henikoff 2002). Achieving fully mature mRNAs, however, also requires posttranscriptional processes involving splicing, polyadenylation, and nuclear export. Thus, the collective term "euchromatin" likely represents a complex chromatin state(s) that encompasses a dynamic and elaborate mixture of dedicated machines that interact together and closely with the chromatin fiber to bring about the transcription of functional RNAs. Learning the "rules" as to how, in the most general sense, the "activating machinery" interacts with the transcription apparatus as well as the chromatin template is an exciting area of current research, although due to its dynamic nature, it may not strictly classify as

AND

CON C E P T S

epigenetics, but more as transcription and chromatin dynamics studies. What then defines "heterochromatin?" Although it is historically less well studied than euchromatin, new insights suggest that heterochromatin plays a critically important role in the organization and proper functioning of genomes from yeast to humans (although S. cerevisiae has a distinct form of heterochromatin). Underscoring its potential importance is the fact that 96% of the mammalian genome consists of noncoding and repetitive sequences. New mechanistic insights, underlying the formation of heterochromatin, have revealed unexpected findings. For example, non-sequence-specific transcription, which produces double-stranded RNA (dsRNA), is subject to silencing by an RNA interference (RNAi)-like mechanism (see Section 10 below). The production of such dsRNAs acts as an "alarm signal" reflecting the fact that the underlying DNA sequence cannot generate a functional product, or has been invaded by RNA transposons or viruses. The dsRNA is then processed by Dicer and targeted to chromatin by complexes dedicated to initiating a cascade of events leading to the formation of heterochromatin. Using a variety of model systems, remarkable progress has been made dissecting what appears to be a highly conserved pathway leading to a heterochromatin "locked-down" state. Although the exact order and details may vary, this general pathway involves histone tail deacetylation, methylation of specific lysine residues (e.g., H3K9), recruitment of heterochromatin-associated proteins (e.g., HPl), and establishment of DNA methylation (Fig. 9). It is likely that sequestering of selective genomic regions to repressive nuclear domains or territories may enhance heterochromatin formation. Interestingly, increasing evidence suggests that heterochromatin may be the "default state;' at least in higher organisms, and that the presence of a strong promoter or enhancer, producing a productive transcript, can override heterochromatin. Even in lower eukaryotes, the general concepts of heterochromatin assembly seem to apply. Hallmark features include hypoacetylated histone tails, followed by the binding of acetylation-sensitive heterochromatin proteins (e.g., SIR proteins; for details, see Chapter 4). Depending on the fungal species (e.g., budding vs. fission yeast), varying amounts of histone methylation and HPI-like proteins exist. Even though these genomes are more set to a general default state of being poised for transcription, some heterochromatin-like genomic regions are present (mating loci, telomeres, centromeres, etc.) that are able to suppress gene transcription and genetic recombination when test genes are placed in these new neighborhoods.

36 • C HAP

T ER

transcription unit (gene)

DNA repeats (noncoding)

messenger RNA _ _ _ _~ gene

-----------.~~

.............

transcription factor binding

noncoding dsRNAs

remodeling complex recruitment

RITS complex recruitment

activating histone modifications

repressive histone methylation

histone variants

DNA methylation

Figure 9. Distinction between Euchromatic and Heterochromatic Domains

accessible information

restricted information

euchromatin

heterochromatin

What useful functions might heterochromatin serve? The defmition of centromeres, a region of constitutive heterochromatin, correlates well with a heritable epigenetic state and is thought to be evolutionarily driven by the largest clustering of repeats and repetitive elements on a chromosome. This partitioning ensures large and relatively stable heterochromatic domains marked by repressive «epigenetic signatures:' facilitating chromosome segregation during mitosis and meiosis (Chapter 14). Here, it is noteworthy that centromeric repeats and the corresponding epigenetic marks that associate with them have been duplicated and moved onto other chromosome arms to create «silencing domains" in organisms such as fission yeast. Constitutive heterochromatin at telomeres (the protective ends of chromosomes) similarly ensures stability of the genome by serving as chromosomal «caps:' Last, heterochromatin formation is known to be a defense mechanism against invading DNA. Collectively, these findings underscore a general view that heterochromatin serves important genome maintenance functions which may rival even that of euchromatin itself.

Summary of common differences between euchromatin and constitutive heterochromatin. This includes differences in the type of transcripts produced, recruitment of DNA-binding proteins (i.e., transcription factor [TF]), chromatin-associated proteins and complexes, covalent histone modifications, and histone variant composition.

In summary, the broad functional distinction between euchromatin and heterochromatin can thus far be attributed to three known characteristics of chromatin. First is the nature of the DNA sequence-e.g., whether it contains AT-rich «rigid" DNA around promoters, repetitive sequences and/or repressor-binding sequences that signal factor association. Second, the quality of the RNA produced during transcription determines whether it is fully processed into an mRNA that can be translated, or whether the RNA is degraded or earmarked for use by the RNAi machinery to target heterochromatinization. Third, spatial organization within the nucleus can playa significant sequestering role for the maintenance of localized chromatin configurations. 7 Histone Modifications and the Histone Code

We have explored how histone modifications may change the chromatin template by cis-effects that alter internucleosomal contacts and spacing, or the trans-effects caused by histone and non-histone protein associations

with the template. What is the contribution and biological output of histone modifications? Patterns of chromatin structure that correlate with histone tail modifications have emerged from studies using bulk histones, suggesting that epigenetic marks may provide "ON" (i.e., active) or "OFF" (inactive) signatures. This has come through a long history of mostly correlative studies showing that certain histone modifications, notably histone acetylation, are associated with active chromatin domains or regions that are generally permissive for transcription. In contrast, other marks, such as certain phosphorylated histone residues, have long been associated with condensed chromatin that, in general, fails to support transcriptional activity. The histone modifications shown in Appendix 2 summarize the sites of modification that are known at this time. Here, we stress that these reflect modifications and sites that may well not be exhibited by every organism. How are histone modifications established or removed in the first place? A wealth of work in the chromatin field has suggested that histone tail modifications are established ("written") or removed ("erased") by the catalytic action of chromatin-associated enzymatic systems. However, the identity of these enzymes eluded researchers for years. Over the last decade, a remarkably large number of chromatin-modifying enzymes have been identified from many sources, most of which are compiled in Appendix 2. This has been achieved through numerous biochemical and genetic studies. The enzymes often reside in large multi-subunit complexes that can catalyze the incorporation or removal of covalent modifications from both histone and non-histone targets. Moreover, many of these enzymes catalyze their reactions with remarkable specificity to target residue and cellular context (i.e.) dependent on external or intrinsic signals). For darity, and by way of example, we discuss briefly the four major enzymatic systems that catalyze histone modifications, together with their counterpart enzymatic systems that reverse the modifications (Fig. 10) (Vaquero et al. 2003; Holbert and Marmorstein 2005). Together, these antagonistic activities govern the steady-state balance of each modification in question. Histone acetylases (HATs) acetylate specific lysine residues in histone substrates (Roth et al. 2001) and are reversed by the action of histone deacetylases (HDACs) (Grozinger and Schreiber 2002). The histone kinase. family of enzymes phosphorylate specific serine or threonine residues, and the phosphatases (PPTases) remove phosphorylation marks. Particularly well known are the mitotic kinases, such as cyclin-dependent kinase or

V E R V lEW

AND

CON C E P T 5

aurora kinase, which catalyze the phosphorylation of core (H3) and linker (H 1) histones. Less clear in each case are the opposing PPTases that act to reverse these phosphorylations as cells exit mitosis. Two general classes of methylating enzymes have been described: the PRMTs (protein arginine methyltransferases) whose substrate is arginine (Lee et al. 2005), and the HKMTs (histone lysine methyltransferases) that act on lysine residues (Lachner et al. 2003). Arginine methylation is indirectly reversed by the action of deiminases, which convert methyl-arginine (or arginine) to a citrulline residue (Bannister and Kouzarides 2005). Methylated lysine residues appear to be chemically more stable. Lysine methylation has been shown to be present in mono-, di-, or tri-methylated states. Several tri-methylated residues in the H3 and H4 animo termini appear to have the potential to be stably propagated during cell divisions (Lachner et al. 2004), as well as the H4K20mel mark in Drosophila imaginal discs (Reinberg et al. 2004). Recen dy, a lysine-specific "demethylase" (LSD 1) was described as an amine oxidase that is able to remove H3K4 methylation (Shi et al. 2004). The enzyme acts by FAD-dependent oxidative destabilization of the aminomethyl bond, resulting in the formation of unmodified lysine and formaldehyde. LSD 1 was shown to be selective for the activating H3K4 methylation mark and can only destabilize mono- and di-, but not tri-methylation. This demethylase is part of a large repressive protein complex that also contains HDACs and other enzymes. Other evidence suggests that LSDI can associate in a complex together with the androgen receptor at target loci and demethylate the H3K9me2 repressive histone mark to contribute to transcriptional activation (Metzger et al. 2005). A different class of histone demethylases has been characterized to work via a more potent mechanismradical attack-known as hydroxylases or dioxygenases (Tsukada et al. 2006). One of these only destabilizes H3K36me2 (an active mark), but not in the tri-methyl state. This novel jumonji histone demethylase OHDMl) contains the conserved jumonji domain, of which there are around 30 known in the mammalian genome, suggesting that some of these enzymes may also be able to attack other residues as well as a tri-methyl state (Fodor et al 2006; Whetstine et al 2006). Considerable progress has been made in dissecting the enzyme systems that govern the steady-state balance of these modifications, and we suspect that much more progress will be made in this exciting area. It remains a challenge to understand how these enzyme complexes are regulated and how their physiologically relevant substrates

38 • C HAP

T ER 3

(lysine)

(serine, threonine)

(arginine)

(lysine)

HAT

kinase

PRMT

HKMT

nTase

~inase (citrulline)

QeOXidase hydroxylase

Figure 10. Histone-modifying Enzymes Covalent histone modifications are transduced by histone-modifying enzymes ("writers") and removed by antagonizing activities. They are classified into families according to the type of enzymatic action (e.g., acetylation or phosphorylation). Protein domains with specific affinity for a histone tail modification are termed "readers." (HAT) Histone acetyl transferase; (PRMT) protein arginine methyltransferase; (HKMT) histone lysine methyltransferase; (HDAC) histone deacetylase; (PPTase) protein phosphatases; (Ac) acetylation; (P) phosphorylation; (Me) methylation.

and sites are targeted. In addition, it remains unclear how covalent mechanisms affect epigenetic phenomena. Histone modifications do not occur in isolation, but rather in a combinatorial manner as proposed for modification cassettes (i.e., covalent modifications in adjacent residues of a particular histone tail, e.g., H3K9me and H3SlOph or H4S1ph, H4R3me, and H4K4ac) and trans histone pathways (covalent modifications between different histone tails or nucleosomes; see Fig. 11). Intriguingly, almost all of the known histone modifications correlate with activating or repressive function, dependent on which amino acid residue(s) in the histone amino termini is modified. Both synergistic and antagonistic pathways have been described (Zhang and Reinberg 2001; Berger 2002; Fischle et al. 2003b) that can progressively induce combinations of active marks, while simultaneously counteracting repressive modifications. It is, however, not known how many distinct combinations of modifications across the various amino-terminal histone positions exist for any given nucleosome, because most of the studies have been carried out on bulk histone preparations. In addition to the amino termini, modifications in the globular histone fold domains have recently been shown to affect chromatin structure and assembly (Cosgrove et al. 2004), thereby influencing gene expression and DNA damage repair (van Attikum and Gasser 2005; Vidanes et al. 2005). It is also worth noting that several of the histone-modifying enzymes also target non-histone substrates (Sterner and Berger 2000; Chuikov et al. 2004). Figure 11 illustrates two examples of established hierarchies of histone modifications that seem to index tran-

scription of active chromatin or, in contrast, pattern heterochromatic domains. These studies provoke the question of whether there is a "histone code" or even an "epigenetic code." Although this theoretical concept has been highly stimulating, and has been shown to be correct in some of its predictions, the issue as to whether a code actually exists has remained largely open. As a comparison, the genetic code has proven extremely useful, because of its predictability and near universality. It uses for the most part a four-base "alphabet" in the DNA (i.e., nucleotides), forming what is generally an invariant and nearly universal language. In contrast, current evidence suggests that histone-modification patterns are likely to vary considerably from one organism to the next, especially between lower and higher eukaryotes, such as yeast and humans. Thus, even if a histone code exists, it is not likely to be universal. This situation is made considerably more complicated when one considers the dynamic nature of histone modifications, varying in space and in time. Furthermore, the chromatin template engages a staggering array of remodeling factors (Vignali et al. 2000; Narlikar et al. 2002; Langst and Becker 2004; Smith and Peterson 2005). However, chromatin immunoprecipitation assays (ChIP), when examined on genome-wide levels (ChIP on chip), have begun to decipher nonrandom and somewhat predictable patterns in several genomes (e.g., S. pombe, A. thaliana, mammalian cells), such as strong correlations of H3K4me3 with activated promoter regions (Strahl et al. 1999; Santos-Rosa et al. 2002; Bernstein et al. 2005) and of H3K9 (Hall et al. 2002; Lippman et al. 2004; Martens et

o al. 2005) and H3K27 (Litt et al. 2001; Ringrose et al. 2004) methylation with silenced heterochromatin. Perhaps the limitation of the histone code is that one modification does not invariantly translate to one biological output. However, modifications combinatorially or cumulatively do appear to define and contribute to biological functions (Henikoff 2005).

V E R V lEW

AND

CON C E P T S

•

mobilize nucleosomes and/or alter nucleosomal structure. Chromatin-remodeling activities often work in concert with activating chromatin-modifying enzymes and can generally be categorized into two families: the SNF2H or ISWI, and the Brahma or SWI/SNF family. The SNF2H/ISWI family mobilizes nucleosomes along the DNA (Tsukiyama et al. 1995; Varga-Weisz et al. 1997), whereas Brahma/SWl/SNF transiently alter the structure of the nucleosome, thereby exposing DNA:histone contacts in ways that are currently being unraveled (see Chapter 12). Additionally, some of the ATP-hydrolyzing activities resemble "exchanger complexes" that are themselves dedicated to the replacement of conventional core histones with specialized histone "variant" proteins. This ATPcosting shuffle may actually be a means by which existing modified histone tails are replaced with a clean slate of variant histones (Schwartz and Ahmad 2005). Alternatively, recruitment of chromatin-remodeling complexes, such as SAGA (Spt-Ada-GcnS-acetyltransferase) can also be enhanced by preexisting histone modifications to ensure transcriptional competence of targeted promoters (Grant et al. 1997; Hassan et al. 2002).

8 Chromatin-remodeling Complexes and Histone Variants

Another major mechanism by which transitions in the chromatin template are induced is by signaling the recruitment of chromatin "remodeling" complexes that use energy (ATP-hydrolysis) to change chromatin and nudeosome composition in a non-covalent manner. Nudeosomes, particularly when bound by repressive chromatin-associated factors, often impose an intrinsic inhibition to the transcription machinery. Hence, only some sequence-specific transcription factors and regulators (although not the basal transcription machinery) are able to gain access to their binding site(s). This accessibility problem is solved, in part, by protein complexes that

SIGNALS

!!!!!

H2BK123ub

H3K9ac

I \ H3K4me I H3K9ac \

--i

H4K16ac

--i

0--

H3K79me

HP1

active

H3K9me HP1

H3K9me

--i

I H4K16ac

0--

H4K20me

\ DNA methylation Figure 11. Coordinated Modification of Chromatin

H4K20me

repressed

The transition of a na"ive chromatin template to active euchromatin (left) or the establishment of repressive heterochromatin (right), involving a series of coordinated chromatin modifications. In the case of transcriptional activation, this is accompanied by the action of nucleosome-remodeling complexes and the replacement of core histones with histone variants (yellow, namely H3.3).

40 •

C HAP T E R

In addition to transcriptional initiation and establishing the primary contact with a promoter region, the passage of RNA pol II (or of RNA pol I) during transcriptional elongation is further obstructed by the presence of nucleosomes. Mechanisms are therefore required to ensure the completion of nascent transcripts (particularly of long genes). In particular, a series of histone modifications and docking effectors act in concert with chromatin-remodeling complexes such as SAGA and FACT (for facilitate chromatin transcription) (Orphanides et al. 1998) to allow RNA pol II passage through nucleosomal arrays. These concerted activities will, for example, induce increased nucleosomal mobility, displace H2A/H2B dimers, and promote the exchange of core histones with histone variants. As such, they provide an excellent example of the close interplay between histone modifications, chromatin remodeling, and histone variant exchange to facilitate transcriptional initiation and elongation (Sims et al. 2004). Other remodeling complexes have also been characterized, such as Mi-2 (Zhang et al. 1998; Wade et al. 1999) and INO80 (Shen et al. 2000), which are involved in stabilizing repressed rather than active chromatin. Compositional differences of the chromatin fiber that occur through the presence of histone variants contribute to the indexing of chromosome regions for specialized functions. Each histone variant represents a substitute for a particular core histone (Fig. 12), although histone variants are often a minor proportion of the bulk histone content, and thus more difficult to study than regular histones. An increasing body of literature (for review, see Henikoff and Ahmad 2005; Sarma and Reinberg 2005) documents that histone variants have their own pattern of susceptibility to modifications, likely specified by the small number of amino acid changes that distinguish them from their family members. On the other hand, some histone variants have distinct amino- and carboxy-terminal domains with unique chromatin-regulating activity and different affinities to binding factors. By way of example, transcriptionally active genes have general histone H3 exchanged by the H3.3 variant, in a transcription-coupled mechanism that does not require DNA replication (Ahmad and Henikoff 2002). The replacement of core histone H2A with the H2A.Z variant correlates with transcriptional activity and can index the 5' end of nucleosome-free promoters. However, H2A.Z has also been associated with repressed chromatin. CENP-A, the centromere-specific H3 variant, is essential for centromeric function and hence chromosome segregation. H2A.X, together with other histone marks, is associated with

histone H3 H3

HFD

1135

H3.3

+~FD

1135

CenpA

HFD

1133

histone H4 H4 histone H2A H2A

macroH2A

HFD HFD HFD HFD

histone H2B H2B

HFD

H2A.Z H2A.X

histone H1 (7 isoforms)

1129

H127

]142 I I 7_Met'j{._1355

1125

c:::;:JI.m• •:::;:::;:::;:::;=:J1179-223

Figure 12. Histone Variants Protein domain structure for the core histones (H3, H4, H2A, H2B), linker histone Hl, and variants of histones H3 and H2A. The histone fold domain (HFD) where histone dimerization occurs, and regions of the protein that differ in histone variants (shown in red) are indicated.

sensing DNA damage and appears to index a DNA lesion for recruitment of DNA repair complexes. MacroH2A is a histone variant that specifically associates with the inactive X chromosome (Xi) in mammals (for more details on histone variants, see Chapter 13). Importantly, and in contrast to the commonly held textbook notion that histones are synthesized and deposited only during S phase, synthesis and substitution of many of these histone variants occurs independently of DNA replication. Hence, the replacement of core histones by histone variants is not restricted to cell cycle stages (i.e., S phase), but can take immediate effect in response to ongoing mechanisms (e.g., transcriptional activity or kinetochore tension during cell division) or stress signals (e.g., DNA damage or nutrient starvation). Elegant biochemical studies have documented chromatin remodeling or exchanger complexes that are specific for replacement of distinct histone variants, such as H3.3, H2A.Z, or H2A.X (Cairns 2005; Henikoff and Ahmad 2005; Sarma and Reinberg 2005). For instance, replacement ofH3 with the H3.3 variant occurs via the action of the HIRA (histone regulator A) exchanger complex (Tagami et al. 2004), and H2A is replaced by H2A.Z through the activity of the SWRI (Swi2/Snf2-related ATPase 1) exchanger complex

OVERVIEW

(Mizuguchi et aI. 2004). Together, these substitutions allow variant nucleosomes to build particularly active chromatin. For some of the other histone variants, the mechanism of targeting and exchange remains to be determined, whether via an ATP-dependent histone exchanger complex or a histone chaperone protein. It has now even been postulated that exchanger complexes may exist to substitute modified histones with their unmodified counterparts, as a mechanism for erasing more robust epigenetic marks that reside in the histone amino termini.

9 DNA Methylation DNA methylation is the oldest epigenetic mechanism known to correlate with gene repression (Razin and Riggs 1980). It is present to varying degrees in all eukaryotes except yeast. This modification consists of the addition of a methyl group at cytosine residues of the DNA template. It occurs at CpG dinucleotides in mammals, whereas other symmetric, asymmetric, and non-CpG methylation patterns are known in N. crassa and plants. The distribution of methylated DNA along the genome shows enrichment at noncoding regions (e.g., centromeric heterochromatin) and interspersed repetitive elements (transposons) but not in the CpG islands of active genes (Bird 1986). In fact, the increasing levels of DNA methylation correlate with a relative increase in noncoding and repetitive DNA content in the genomes of higher eukaryotes (see Fig. 15 in Section 11). Experimental evidence indicates that this is because DNA methylation serves mainly as a host defense mechanism to silence much of the genome of foreign origin (i.e., replicated transposable elements, viral sequences, and other repeated sequences). DNA methyltransferases (DNMTs) are the "effectors" of DNA methylation, and catalyze either de novo (i.e., at novel sites) or maintenance methylation of hemimethylated DNA following DNA replication (see Chapter 18). Loss of the ability to maintain DNA methylation can result in several diseases, such as ICF (Immunodeficiency, Centromeric instability, and Facial abnormalities) (see Chapters 18 and 23 ). Deregulation in the levels of DNA methylation is also a contributing factor to cancer progression (Chapter 24). What are the signals that direct DNMTs to methylate certain regions of DNA? Currently it is known that highly repetitive tandem repeat sequences of the genome (e.g., pericentromeric chromatin) rely on the repressive H3K9 methylation marks to direct DNA methylation de novo, as evidenced in N. crassa and plants (see Chapters 6 and 9). Interspersed repeats can also signal de novo DNA

AND

CONCEPTS

methylation, described in the context of RIP in N. crassa (Tamaru and Selker 200l) and retrotransposon silencing in the male germ line of mammals. A protein responsible for the latter has been identified-Dnmt3L-and may function by scanning the genome to identify high levels of hom*ology-heterology junctions that signal the requirement for DNA methylation (Bourc'his and Bestor 2004). In 'plants, RNAs provide the signal for de novo DNA methylation, through a unique mechanism termed RNAdependent DNA methylation (RdDM; see Chapter 9). There is evidence that chromatin remodelers of the SWI/SNF family are in some way necessary for the global patterns of DNA methylation, as demonstrated in plants (Jeddeloh et aI. 1999) for the DDMI protein and mammals by the Lshl hom*olog (Yan et aI. 2003). Last, the HKMT-PcG protein Ezh2 may also be involved in directing DNA methylation at certain promoters in mammals (Vire et aI. 2005). Once established, the way in which DNA methylation may function to silence chromatin is not entirely clear, although evidence points to trans-regulation. Binders of methylated cytosines, called methyl-CpG-binding domain proteins (MBD), can be considered the DNA methylation equivalent to binders (or readers) of modified histone motifs (Fig. 13). For example, the methylcytosine-binding protein (MeCP2) binds methylated CpGs and recruits HDACs to mediate repressive histone marks (see Chapter 18). DNA methylation is also known to disturb the recognition sites of transcriptional regulators (e.g., CTCF) that are involved in genomic imprinting (see Chapter 19). The existence of methylated DNA at imprinted loci that silence either the maternal or paternal allele in plants and placental mammals suggests that in the course of evolution they uniquely harnessed this epigenetic mechanism to stabilize gene repression. Interestingly, in marsupials, there is a lack of DNA methylation at imprinted loci, indicating that its involvement in mammalian imprinting is a relatively recent evolutionary event (see Chapters 17 and 19). Conversely, in dipteran insects such as Drosophila, DNA methylation has largely been lost as a functional epigenetic mechanism (Lyko 2001). Highly repetitive regions of the mammalian genome that are typically methylated become increasingly mutagenic when unmethylated, to the extent of causing global genomic instability (Chen et al. 1998). Chromosomal abnormalities ensue, which are a major cause of many diseases and cancer progression (see Section 15). This underlines the crucial role that DNA methylation plays in genome integrity. Conversely, individual methylated cyto-

42 •

C HAP T E R 3

regulation

:r NH 2

mutation

H20 \,

NH 3 /

deamination

10 RNAi and RNA-directed Gene Silencing

5-methylCytosine

information, although there is no direct evidence for this. However, RNA methylation appears to be "sensed" by certain Toll-like receptors (transmembrane receptors that recognize common pathogen molecular motifs) to mediate innate immunity (Ishii and Akira 2005), corroborating such a hypothesis. This raises the interesting possibility that RNA methylation may yet prove to be a third form of methylation-based epigenetic modulation.

Thymine

Figure 13. DNA Methylation and Deamination Cytosine nucleotides that are methylated (Me) may be bound by methyl DNA-binding proteins (MBD). Unbound 5-methylcytosine is prone to spontaneous mutation through deamination (reaction shown in the lower panel), resulting in as-methyl CpG to TpA transition in the DNA sequence.

sine bases have a high propensity to spontaneously mutate. Thus, over time, C-T transitions occur through a deamination reaction (Fig. 13), but this characteristic is also thought to be beneficial for protecting the host genome because it permanently deactivates parasitic DNA sequences such as transposons. In a different context, this same chemical reaction is actively catalyzed by the activation-induced deaminase, or AID. Expression of the enzyme in Band T cells causes "somatic hypermutation" at the immunoglobulin (Ig) locus. This is an important mechanism for expanding the repertoire of antigen receptors and hence strengthening the immunity of mammals (Petersen-Mahrt 2005; Chapter 21). AID expression observed in early mammalian development has led to the suggestion that it may provide an alternative route to demethylating DNA, although this would happen at the risk of increased point mutation rates. DNA methylation and histone methylation are prominent mechanisms for epigenetic regulation of the genome. Noncoding RNAs, as described in the next section, are important primary triggers for inducing silent chromatin. It is also known that RNA molecules can be heavily methylated at the sugar or nucleoside backbone. Moreover, methylation at the 3' end of small noncoding RNAs has been shown to stabilize these molecules (Yu et al. 2005). Intriguingly, Dnmt2 was recently identified as a tRNA methyltransferase (Goll et al 2006). It is therefore plausible that, similar to DNMTs and HKMTs, RNA methyltransferases may exist as "writers" of epigenetic

The knowledge that constitutive heterochromatin at centromeres and telomeres plays an instructive role in genome integrity has contributed to a paradigm shift in the way that repetitive noncoding "junk" DNA is viewed. Is it possible that these repetitive sequences serve a nonwasteful purpose that is only beginning to be elucidated? Is it even possible that such DNA sequences are not completely"silent?" This possibility has stemmed from a fundamental series of discoveries that linked RNAi to the formation of silent chromatin (heterochromatin). RNAi is a host defense mechanism that breaks down dsRNA species into small RNA molecules (known as short interfering RNA or siRNA). This process ultimately leads to RNA degradation or the use of the small RNAs to inhibit translation, known as posttranscriptional silencing (PTGS). The more recently discovered transcriptional gene silencing (TGS) mechanism, leading to heterochromatin formation, was discovered through the convergence of independent lines of investigation into chromatin and the RNAi machinery. On the one hand, much was known about repressive DNA methylation (in fungi, plants, and mammals), chromatin modifications (e.g., H3K9me3), and chromatin-associated factors (HP 1) that are characteristic of heterochromatin domains. On the other hand, researchers were making headway in identifying factors of the RNAi machinery (e.g., Dicer, Argonaute, RNA-dependent RNA polymerase or RdRP). The most convincing progress that tied together these two seemingly divergent fields came from elegant studies in S. pombe, where mutations of any component of the RNAi machinery resulted in defects in chromosome segregation (Hall et al. 2002; Reinhart and Bartel 2002; Volpe et al. 2002). This was brought about by the inability to stabilize centromeric heterochromatin and underscored the likely Widespread role of RNAimediated mechanisms in producing silent heterochromatin domains. It also highlighted the importance of heterochromatin beyond transcriptionally silencing

genes, to a role in maintammg genome integrity and hence viability, as shown by the requirement of centromeric heterochromatin for the process of chromosome segregation. Emerging evidence also suggests that siRNAs are required in defining other specialized regions of functional heterochromatin, such as telomeres. Transcription from both DNA strands of S. pombe pericentromeric repeats and the detection of processed siRNA derivatives provided strong evidence that the dsRNA derivative was the critical substrate to target the RlTS complex to the centromeres for silencing (Fig. 14) (Verdel et al. 2004). Furthermore, clr4 mutants (the S. pombe ortholog of mammalian Suv39h HKMT) failed to process dsRNA into siRNAs, strengthening the case for the interplay between the RNAi machinery and heterochromatin assembly (Motamedi et al. 2004). Exactly how siRNAs, generated by the RNAi machinery (i.e., Dicer, Argonaute, RdRP), initiate heterochromatin assembly or guide it to appropriate genomic loci is still unknown. A model has emerged in which a complex interaction between the RNAi machinery complex RITS and centromeric repeats leads to a self-reinforcing cycle of heterochromatin formation involving Clr4, HDACs, Swi6 (ortholog to mammalian HP1), and cohesin, probably via Ago-directed annealing of RNA:RNA hybrids to the nascent transcript (Fig. 14) (see Chapters 6 and 8). In Tetrahymena, a similar RNA-mediated targeting mechanism has been recognized to direct the unique case of DNA elimination that occurs in the somatic nucleus. In this case, transcription occurs from both strands of the internal eliminated segment (IES) sequences in the "silent," germ-line (micronuclear) genome at the appropriate stage of the sexual pathway (Chalker and Yao 2001; Mochizuki et al. 2002). Along the same lines as the RNAi-dependent TGS model, a scan RNA (scnRNA) model was proposed to explain how DNA sequences in the parental macronucleus can epigenetically control genomic alterations in the new macronucleus, involving small RNAs (for more detail, see Chapter 7). These exciting results provide the first demonstration of an RNAi-like process directly altering a somatic genome. This raises the intriguing possibility that intergenic RNAs produced at the V-DJ locus (Bolland et al. 2004) may potentially direct DNA sequence elimination during V-DJ recombination of the immunoglobulin heavy chain (IgH) locus in B cells and the T-cell receptor (TCR) loci in T cells. In plants, there are a number of orthologs for many of the RNAi components, resulting in a variety of RNA silencing pathways that can act with greater specificity for particular DNA sequences, although there is some redun-

dsRNA

V E R V lEW

AND

CON C E P T 5

siRNAs

amplification Figure 14. RNA-directed Heterochromatin Formation

Complementary dsRNA transcripts, produced by transcription of both strands or the folding back of inverted repeat transcripts, result in the generation of siRNAs through the action of Dicer (top). Incorporation of the siRNAs into the RITS complex through binding with the Argonaute protein (Ago) activates the complex for targeting to complementary DNA or nascent RNA. The complex attracts Clr4, which transduces histone H3K9me2. The modification-specific binder, Swi6, binds to these modified histones, facilitating the spread of a repressive chromatin domain. The action of RdRP amplifies the levels of siRNAs by using existing siRNAs as primers, reinforcing the targeting capacity of the RITS complex to specific regions of DNA.

dancy between factors. Studies of RNAi-mediated TGS in plants have revealed a novel class of RNA polymerasesRNA polymerase IV (or RNA pol IV)-that may transcribe DNA solely at heterochromatic regions (Herr et al. 2005; Pontier et al. 2005). Also unique to plants is the demonstration that RNAi pathways directly affect DNA methylation (Chan et al. 2004) (explained in detail in Chapter 9). RNAi-like chromatin effects have also been uncovered in Drosophila and mammals. For instance, RNase A treatment of permeabilized mammalian cells rapidly removes heterochromatic H3K9me3 marks, suggesting that an RNA moiety may be a structural component of pericentromeric heterochromatin (Maison et al. 2002). Ablation of siRNA processing factors in vertebrates impairs H3K9 methylation and HP1 binding at pericentromeric heterochromatin (f*ckagawa et al. 2004). Intriguingly, embryonic stem (ES) cells still proliferate, but fail to differentiate, in dicer-null mutants (Kanellopoulou et al. 2005), suggesting a currently not-understood connection between the RNAi machinery, non coding RNAs, and mammalian development. In Drosophila, silencing of the tandem arrays of the miniwhite gene, subject to PEV also appears to be dependent on the RNAi machinery (Pal-Bhadra et al. 2004).

44 •

C HAP T E R 3

Collectively, these studies indicate a crucial, and probably primary, role for noncoding RNAs in triggering epigenetic transitions and heritably maintaining specific chromatin states of the chromatin template. In fact, these noncoding RNAs have provided the answer for how diverse repetitive sequences in different organisms achieve heterochromatinization through an RNA-targeted mechanism. In an effort to identify more targets of RNAi, the sequencing of small RNAs has revealed that they are largely transcribed from endogenous transposons and other repetitive sequences in plants, Drosophila, and mammals, among other organisms (Almeida and Allshire 2005; Bernstein and Allis 2005). Together, these results indicate that RNAi has evolved, in part, to maintain genomic stability by silencing mobile DNA elements and viruses, and is a conserved mechanism across most eukaryotic species. It now appears, however, not only that RNA silencing represses invading sequences, but also that this basic mechanism has been harnessed by the cell for the heterochromatinization of centromeres, thereby ensuring correct chromosome segregation and genome integrity. Together, the above examples indicate a striking variation to the central dogma of gene control that is beginning to emerge as follows: DNA -'?noncoding RNA-'?chromatin-'?gene function. The idea that noncoding RNAs would actively participate in RNAi-like mechanisms which also target locus-specific domains for chromatin remodeling and gene silencing was never anticipated. 11 From Unicellular to Multicellular Systems

The 5,000-6,000 genes contained in the genomes of budding and fission yeasts are sufficient to regulate basic metabolic and cell division processes. There is, however, no requirement for cell differentiation, because these unicellular organisms are essentially clonal and, as such, repetitive "immortal" entities. In contrast, mammals code for ~ 25,000 genes required in ~ 200 different cell types. Understanding how multicellular complexity is generated and coordinated from the same genetic template is a key question in epigenetic research. A comparison of the genome sizes between yeasts, flies, plants, and mammals indicates that genome size significantly expands with the complexity of the respective organism. There is a more than 300-fold difference between the genome sizes of yeast and mammals, but only a modest 4-5-fold increase in overall gene number (Fig. 15). However, the ratio of coding to noncoding and

repetitive sequences is indicative of the complexity of the genome: The largely "open" genomes of unicellular fungi have relatively little noncoding DNA compared with the highly heterochromatic genomes of multicellular organisms. In particular, mammals have accumulated considerable repetitive elements and noncoding regions, which account for the majority of its DNA sequence (52% noncoding and 44% repetitive DNA). Only 4% of the mammalian genome thus encodes for protein function (including intronic sequences). This massive expansion of repetitive and noncoding sequences in multicellular organisms is most likely due to the incorporation of invasive elements, such as DNA transposons, retrotransposons, and other repetitive elements. Although these represent a burden for coordinated gene expression programs, they also allow genome evolution and plasticity, and a certain degree of stochastic gene regulation. The expansion of repetitive elements has even infiltrated the transcriptional units of the mammalian genome. This results in transcription units that are frequently much larger (30-200 kb), commonly containing multiple promoters and DNA repeats within untranslated introns. In contrast, plants, with similarly large genomes, generally possess smaller transcription units with smaller introns, because they have evolved defense mechanisms to ensure that transposon insertion within transcription units is not tolerated. There are important organismal differences that manifest in the types of epigenetic pathways utilized despite high degrees of functional conservation for many mechanisms across species. Differences are, in part, believed to be related to genome size. The vast expansion of the genome with noncoding and repetitive DNA in higher eukaryotes requires more extensive epigenetic silencing mechanisms. This correlates with the fact that mammals and plants employ a full range of repressive histone lysine methylation, DNA methylation, and RNAi silencing mechanisms. Another challenge accompanying multicellularity is how to coordinate and maintain multiple cell types (cellular identity). This is a delicate balance, involving the Polycomb (PcG) and Trithorax (trxG) groups of protein complexes for genome regulation. The PcG proteins, in particular, correlate with the emergence of multicellularity (see Section 12). Cells within multicellular organisms can be functionally divided into two major compartments: germ cells (totipotent and required for transmission of genetic information to the next generation) and somatic cells (the differentiated "powerhouse" of an organism). There are important questions of how the germ-cell compartment

o V E R V fEW

14 Mb

125 Mb

13.8 Mb

AND

CON C E P T S

180 Mb

2.4 Gb

3.2 Gb

DROSOPHILA

MOUSE

HUMAN

t;) S.CEREVISIAE

S.POMBE

ARABIDOPSIS

- 6,000 GENES

- 5,000 GENES

> 25,000 GENES

> 14,000 GENES

> 25,000 GENES

16 chromosomes

3 chromosomes

5 chromosomes

4 chromosomes

20 chromosomes

23 chromosomes

gene (2 kb)

.•

gene (5 kb)....

gene (2 kb)···

gene (50 kb) ...

p repeats

simple vs. complex gene organization

Figure 15. Pie Charts of Organismal Genome Organization Genome sizes are indicated for the major model organisms used in epigenetic research at the top of each pie chart. The increase in genome size correlates with the vast expansion of noncoding (i.e., intronic, intergenic, and interspersed repeat sequences) and repeat DNA (e.g., satellite, LINE, SINE DNA) sequences in more complex multicellular organisms. This expansion is accompanied by an increase in the number of epigenetic mechanisms (particularly repressive) that regulate the genome. Expansion of the genome also correlates with an increase in size and complexity of transcription units, with the exception of plants; they have evolved mechanisms that are intolerant to insertions or duplications within the transcription unit. P = Promoter DNA element.

maintains totipotency of its epigenome and what mechanisms are involved in erasing, establishing, and maintaining cell fate (cell memory). Because one germ cell can give rise to another germ cell, it essentially has an infinite proliferative potential, as do unicellular "immortal" organisms. However, to fulfill this role, germ cells are for the most part "resting" and unresponsive to external stimuli, so that integrity of their epigenome can be protected. Indeed, mammalian oocytes can be retained in a resting state for more than 40 years. Similarly, adult stem cells (multipotent) are largely a dormant cell population, proliferating (and self-renewing) only when activated by mitogenic stimuli to enter a restricted number of cell divisions. Thus, the makeup of the epigenome is challenged by many intrinsic (e.g., transcription, DNA replication, chromosome segregation) and external (e.g., cytokines, hormones, DNA damage, or general stress responses) signals, particularly if somatic differentiation has forced cells to leave the protective germ-cell and stem-cell environment. 12 Polycomb and Trithorax

Among some of the main effectors that can transduce signals to the chromatin template and participate in maintaining cellular identity (i.e., provide cellular memory) are members of the PcG and trxG groups of genes (Ringrose

and Paro 2004). These genes were discovered in Drosophila by virtue of their role in the developmental regulation of the Hox gene cluster and homeotic gene regulation. PcG and trxG have since been shown to be key regulators for cell proliferation and cellular identity in multicellular eukaryotes. In addition, these groups of genes are involved in several signaling cascades that respond to mitogens and morphogens; regulate stem cell identity and proliferation, vernalization in plants, homeotic transformations and transdetermination, lineage commitment during B- and T-cell differentiation, and many other aspects of metazoan development (see Chapters 11 and 12). We now briefly address what is known about how the PcG and trxG families of genes convert developmental cues into an "epigenetic memory" through chromatin structure. The PcG and trxG groups of proteins function for the most part antagonistically: The PcG family of proteins establish a silenced chromatin state and the trxG family of proteins in general propagate gene activity. The molecular identification of the Pc gene known to stabilize patterns of gene repression over several cell generations provided the first evidence for a molecular mechanism for cellular or epigenetic memory. As well, PC provided an example of a chromodomain-containing protein with a high degree of similarity to the chromodomain of the heterochromatin-associated protein HP1 (Paro and Hog-

46 •

C HAP T E R

ness 1991). As mentioned above, chromodomains are well documented to be specific histone methyl-lysine binding modules (illustrated in Fig. 10). Approximately 20 PcG genes and at least 15 distinct trxG genes have been identified in Drosophila. Functional analyses have shown that these groups of genes constitute a spectrum of diverse proteins yet are higWy conserved between eukaryotes. PcG genes encode products that include DNA-binding proteins (e.g., TIl), histone-modifying enzymes (e.g., Ezh2), and other repressive chromatin-associated factors that contain a chromodomain with affinity for H3K27me3 (e.g., PC). trxG genes encode transcription factors (e.g., GAGA or Zeste), ATP-dependent chromatin-remodeling enzymes (e.g., Brahma), and HKMTs such as Ash1 and Trx (or its mammalian hom*ologs MLL, Setl, and the MLL family). In most instances, the trxG and PcG families of proteins function as components of diverse complexes to establish stable chromatin structures that facilitate the expression or silencing of developmentally regulated genes (see Chapters 11 and 12). Despite recent advances, the mechanism by which PcG- or trxG-containing complexes are targeted to developmentally regulated chromatin regions is not well understood. In Drosophila, heritable gene repression requires the recruitment of PcG protein complexes to DNA elements called polycomb response elements (PREs). Equivalent sequences in mammals have remained elusive. It is unclear how PcG protein complexes cause long-range silencing in a PRE-dependent manner, because PREs are usually located kilobases from the transcription start site of target genes. It can be postulated that repulsion or recruitment of PcG complexes may be discriminated by changes in transcriptional activity, or differences in productive versus nonproductive mRNA processing (Pirrotta 1998; Dellino et a1. 2004; Schmitt et a1. 2005). Current models support PcG binding through interaction with DNA-binding proteins and the affinity of the chromodomain within the PC protein for H3K27me3-modified histones (Cao et a1. 2002). However, PcG complexes can also associate in vitro with nucleosomes that lack histone tails (Francis et al. 2004), and furthermore, PRE elements have reduced nucleosome density (Schwartz et al. 2005). The most logical explanation for some of these disparate observations is that PcG binding in vivo would initially require interaction with DNA-bound factors that is then stabilized by association with nucleosomes and modified H3K27me3 in the adjacent chromatin region. Clearly, more research is needed to link existing evidence of how PcG complexes are targeted to regions of chromatin and how they medi-

ate repression. This is likely to be organism-dependent, because there is great heterogeneity in the PcG complexes (see Chapter 11). Trithorax group proteins maintain in general an active state of gene expression at target genes and overcome (or prevent) PcG-mediated silencing. This transition is even less well understood, but recent evidence suggests that an RNA-based mechanism could provide the trigger for the recruitment of Ash 1 to target promoters (Sanchez-Elsner et a1. 2006). A number of transient and stable changes in chromatin structure are thought to ensue, perhaps facilitated by intergenic transcription that can establish an open chromatin domain and mediate active histone replacement. Documented chromatin changes include the incorporation of "active" histonelysine methylation marks by trxG HKMTs such as Trx and Ash1, and the reading of these marks (e.g., the WDR5 recognition of H3K4me; Wysocka et a1. 2005). The action of trxG ATP-dependent chromatin-remodeling factors such as Brahma is also required, although how these mechanisms interrelate has yet to be fully determined (for more detail, see Chapter 12). Many PcG and trxG proteins cooperate to maintain a tightly controlled level of repressed heterochromatin versus active euchromatin in a normal cell. In mammalian somatic interphase nuclei, the nuclear morphology reveals that constitutive domains of pericentromeric heterochromatin are grouped into 15-20 foci (see Fig. 16). Deregulation of cell fate and proliferation control, which leads to developmental abnormalities and cancer, frequently displays abnormal nuclear morphologies. For example, the nuclear organization in PML-Ieukemia (related to mixed lymphocyte leukemia [MLL]) cells shows an absence of pericentromeric foci (Di Croce 2005). In contrast, senescent (nonproliferating) cells display a nuclear morphology with large ectopic heterochromatin clusters (Narita et a1. 2003; Scaffidi et al. 2005). Thus, nuclear morphology appears to be a good marker for distinguishing between normal and aberrant cell states, indicating that nuclear architecture may yet playa regulatory role in maintaining specialized domains of chromatin. The study of histone modification levels is another indicator of cell normality or abnormality. Many of these changes are attributed to the deregulation of PcG (e.g., Ezh2) or trxG (e.g., MLL) HKMTs, contributing to the progression and even metastatic potential of a tumor (see Section 15). Indeed, the increase in overall levels of either of the above-mentioned proteins is associated with increased risk of prostate cancer, breast cancer, multiple myeloma, or leukemia (Lund and van Lohuizen 2004;

a v E R V lEW Valk-Lingbeek et al. 2004). In other cases of neoplastic transformation, there is a manifest decrease in repressive histone marks and increase in overall acetylation states (Seligson et al. 2005) causing elevated levels of gene transcription and genomic instability. Clearly, changes in the global control of chromatin, possibly through perturbation of histone-modifying enzymes, affects the functionality of the genome and disrupts the proper gene expression profile of a normal cell. In the case of cellular senescence, an increase in repressive histone marks is also an indicator of cellular dysfunction. This, concomitant with reduced definition of histone acetylation, can reinforce and even increase the levels of silent chromatin, blocking cellular plasticity and driving cells into an antiproliferative state (Scaffidi et al. 2005). This is largely an age-related effect, although the disease state, progeria, can prematurely advance aging. Conversely, when repressive pericentromeric methyl marks are decreased in mutants lacking the transducing enzyme (Suv39h), cells display increased rates of immortalization, no longer senesce, and show greater rates of genomic instability (Braig et al. 2005). These examples illustrate that chromatin deregulation, demonstrated by the levels of characteristic histone marks, often trans-

AND

CON C E P T S

oocyte

r·········· "- germ cells

13 X Inactivation and Facultative Heterochromatin

PeG-mediated gene silencing and X-chromosome inactivation are prime examples for developmentally regulated transitions between active and inactive chromatin states (see Fig. 17), often referred to as facultative heterochromatin. This is in contrast to constitutive heterochromatin (at, e.g., pericentromeric domains), which may by default be induced at noncoding and highly repetitive regions. Facultative heterochromatin occurs at coding regions of the genome, where gene silencing is dependent on, and sometimes reversible by, developmental decisions specifying distinct cell fates. One of the best-studied examples for facultative heterochromatin formation is the inactivation of one of the two X chromosomes in female mammals to equalize the dosage of X-linked gene expression with males that possess only one X (and a heteromorphic Y) chromosome (Chapter 17). Here, chromosome-wide gene silencing of

-~-

somatic cells ............ llstress ll

CELLULAR IDENTITY

Polycomb and Trithorax

duced by PeG and trxG enzymes, and nuclear morphology, is proving to be an important indicator of disease progression.

"protected"

~fertilized.

•

......... ·stress"

···········stress ll

Figure 16. Cellular Identity by PcG and trxG Proteins Two cell compartments are established during embryogenesis, distinguished by their differentiation potency: They are germ cells (totipotent) and somatic cells (including stem cells) with restricted differentiation potentials. The plasticity of a germ or stem cell's genome expression potential is reflected in reduced levels of repressive histone marks which are no longer visible at pericentromeric foci. Normal proliferating cells typically have a nuclear morphology showing 15-20 heterochromatic foci. Polycomb- and Trithorax-containing complexes operate in specifying the epigenetic and, hence, cellular identity of different lineages. They also function in response to external "stress" stimuli, promoting cellular proliferation and appropriate gene expression. Loss of genome plasticity and proliferation potential occurs in senescent (aging) cells, reflected by abnormally large heterochromatic foci and an overall increased level of repressive histone marks. Highly proliferating tumor cells, however, exhibit changes in the balance of repressive and activating histone marks through the deregulation of PcG and trxG histone-modifying enzymes. This is accompanied by perturbed nuclear morphology.

C HAP T E R 3

the inactive X chromosome (Xi) induces a high degree of Xi compaction that is visible as the Barr body, localized in the nuclear periphery of female mammalian cells. How the two alleles of the X chromosomes are counted and how one particular X chromosome is chosen for inactivation are challenging questions in today's epigenetic research. X inactivation involves a large (~17 kb) noncoding RNA, Xist, which appears to act as the primary trigger for chromatin remodeling at the Xi. Although there is the potential to form dsRNA between Xist and the antisense transcript Tsix (expressed only before the onset of X inactivation), no compelling evidence exists for RNAidependent mechanisms being involved in the initiation of X inactivation. The X-inactivation center (XIC) and likely DNA "entry" or "docking" sites (postulated to be specialized repetitive DNA elements that are enriched on the X chromosomes) playa role for Xist RNA to associate and function as a scaffolding molecule, decorating the Xi in cis. Xist promotes the recruitment and action of both PRCl (polycomb respressive complex) and PRC2 complexes, involved in establishing a stable inactive X chromosome. PRC2 components include, for example, the HKMT chromatin-modifying enzyme, EZH2, which catalyzes H3K27me3. PRC1 complex binding may be promoted by both H3K27me3 and histone-modification-independent means, whereas other components of

euchromatic gene repression (e.g. Polycomb)

facultative heterochromatin

the complex, such as the Ring1 proteins, ubiquitinate H2A. Such is the heterogeneity of PcG complexes that different components can act independently of other complex components. The chromatin modifications, PcG complex binding, the subsequent incorporation of the histone variant macroH2A along the Xi, and extensive DNA methylation all contribute to generating a facultative heterochromatin structure along the entire Xi chromosome. Once a stable heterochromatic structure is established, Xist RNA is no longer required for its maintenance (Avner and Heard 2001; Heard 2005). A similar form of monoallelic silencing is genomic imprinting, which also uses a noncoding or antisense RNA to silence one allelic copy in a parent-of-origin-specific manner (Chapter 19). It is currently not clear whether and how Dicer-mutant mouse ES cells would affect the processes of X inactivation or genomic imprinting. The general paradigm of dosage compensation, a classic epigenetically controlled mechanism, has also been addressed in other model organisms, notably C. elegans (Meyer et al. 2004; Chapter 15) and Drosophila (Gilfillan et al. 2004; Chapter 16). It is not yet clear whether dosage compensation occurs in birds, despite the fact that they are heterogametic organisms. In Drosophila, dosage compensation between the sexes occurs not by X inactivation in the female, but by a twofold up-regulation from the single X chromosome in the male. Intriguingly, two non-

constitutive heterochromatin

- - -..~.........t ---aberrant RNA 111

dsRNAs=

Dicer

mis-processed RNA ???

dispersed nuclear distribution

Barr body female cells

Figure 17. RNA Directed Induction of Repressed Chromatin States

heterochromatic foci

Different forms of silent chromatin have different primary signals, but many are likely to be RNA transcript-related (from aberrant transcripts, to Xist RNA, to dsRNAs), depending on the nature of the underlying DNA sequence. This triggers the establishment of a collection of chromatin changes, including a combination of histone modifications (H3K9, H3K27, and H4K20 methylation), the binding of repressive proteins or complexes (e.g., PC or HP1) to the chromatin, DNA methylation, and the presence of histone variants (e.g., macroH2A on the inactive X chromosome). Facultative or constitutive heterochromatin shows visible clustering in the nucleus. Euchromatic repression cannot be determined by nuclear morphology patterns.

a v E R V lEW coding RNAs, roXl and roX2, are known to be essential components, and their expression is male-specific. Although similar mechanistic details probably exist between flies and mammals, it is clear that activating chromatin remodeling and histone modifications, notably MOF-dependent H4K16 acetylation on the male X chromosome, plays a key role in Drosophila dosage compensation. Exactly how histone-modifying activities, such as the MOF histone acetyltransferase, are targeted to the male X chromosome remains a challenge for future studies. Furthermore, ATP-dependent chromatin-remodeling activities, such as nucleosome-remodeling factor (NURF), are thought to antagonize the activities of the dosage compensation complex (DCC). Together, this section and Sections 10 and 11 have described mechanisms for RNA-directed chromatin modifications, as they occur for constitutive heterochromatin, the Xi chromosome, and, possibly, also PcG-mediated gene silencing. On the basis of the intriguing parallels, one might postulate that an RNA moiety(s) or unpaired DNA would provide an attractive primary trigger for stabilizing PcG complexes at PREs or compromised promoter function, where they may "sense" the quality of transcriptional processing. Aberrant or stalled elongation and/or splicing errors could spur the interaction between PRE-bound PcG and a promoter, resulting in transcriptional shutdown. Thus, initiation of PcG silencing would be induced by the transition from productive to nonproductive transcription. The extent to which trxG complexes may utilize RNA quality control and/or processing of primary RNA transcripts as part of maintaining transcriptional "ON" states is beginning to be unraveled (Sanchez-Elsner et al. 2006). 14 Reprogramming of Cell Fates

The question of how cell fate can be altered or reversed has long intrigued scientists. The germ cell and early embryonic cells distinguish themselves from other cell compartments as the "ultimate" stem cell by their innate totipotency. Although cell-fate specification in mammals allows for around 200 different cell types, there are, in principle, two major differentiation transitions: from a stem (pluripotent) cell to a fully differentiated cell, and between a resting (quiescent or Go) and a proliferating cell. These represent the extreme endpoints among many intermediates, consistent with a multitude of different makeups of the epigenome in mammalian development. During embryogenesis, a dynamic increase of epigenetic modifications is detected in the transition from the fertil-

AND

CON C E P T S

•

ized oocyte to the blastocyst stage, and then at implantation, gastrulation, organ development, and fetal growth. Most of these modifications or imprints may be erased via transfer of a differentiated cell nucleus to the cytoplasm of an enucleated oocyte. However, some marks may persist, thereby restricting normal development of cloned embryos, and a few could even be inherited as germ-line modifications (g-mod) (see Fig. 18), which, in mammals, are likely to include DNA methylation. Liver regeneration and muscle cell repair are exceptions of mammalian tissues that can regenerate in response to damage or injury, although most other tissues are unable to be reprogrammed. In other organisms, such as plants and Axolotl, certain somatic cells can actually reprogram their epigenome and reenter the cell cycle to regenerate lost or damaged tissue (Tanaka 2003). In general, however, reprogramming of somatic cells is not possible unless they are engineered to recapitulate early development upon nuclear transfer (NT) into an enucleated oocyte. This was first demonstrated in cloned frogs (Xenopus), and more recently by the generation of Dolly, the first cloned mammal (Campbell et al. 1996; see Chapter 22). Three major obstacles to efficient somatic reprogramming in mammals have been identified. First, certain somatic epigenetic marks (e.g., repressive H3K9me3) are stably transmitted through somatic cell divisions and resist reprogramming in the oocyte. Second, a somatic cell nucleus is unable to recapitulate the asymmetry of reprogramming that occurs in the fertilized embryo as a consequence of the differential epigenetic marks inherited by the male and female haploid genomes (see Mayer et al. 2000; van der Heijden et al. 2005; Chapter 20). Third, transmission of imprinted loci that are particularly important in fetal and placental development is not faithfully maintained upon NT (Morgan et al. 2005). Most cloned embryos abort, suggesting that perturbed epigenetic imprints represent a major bottleneck for normal development and could be the cause for the poor efficiencies of assisted reproductive technologies (ART) and the reduced vigor of cloned animals. The use of embryonic stem cells versus somatic cells shows greatly enhanced reprogramming potential. The demonstration that quiescent cells (a frequent characteristic of stem cells) have a reduction in global H3K9me3 and H4K20me3 states could be a factor indicating enhanced plasticity of the epigenome (Baxter et al. 2004). This is also consistent with the fact that "immortal" unicellular organisms (e.g., yeast) with a largely open and active genome lack several repressive epigenetic mechanisms.

50 • C HAP

T ER

embryonic g-modmodmod mod mod-

differentiated

normal

mod

NT? mod modtumor

Figure 18. Reprogramming by Nuclear Transfer During the lifetime of an individual, epigenetic modifications (mod) are acquired in different cell lineages (left). Nuclear transfer (NT) of a somatic cell reverses the process of terminal differentiation, eradicating the majority of epigenetic marks (mod); however, some modification that would also be present in the germ line (g-mod) cannot be removed. During neoplastic transformation (from a normal to tumor cell), caused by a series of genetic mutations (red stars), epigenetic lesions accumulate. The epigenetic lesions (mod), but not the mutations, can be erased through reprogramming upon NT. This approach evaluates the interplay between genetic and epigenetic contributions to tumorigenesis. (Figure adapted from R. jaenisch.)

Another feature of normal epigenetic reprogramming in mammals, postfertilization, is its distinct asymmetry. This can first be attributed to different programs of epigenetic specification in the male and female germ cells (Chapters 19 and 20). The sperm genome is largely made up of protamines, although there is a residual but significant level of CENP-A (an H3 histone variant) and other putative epigenetic imprints (Kimmins and Sassone-Corsi 2005), whereas the oocyte is made up of regular nucleosome-containing chromatin. Once fertilized, the sperm and oocyte haploid genomes have another cycle of reprogramming involving DNA demethylation and exchange of histone variants. The modifications can either enhance or balance epigenetic differences of the two parental genomes before nuclear fusion, in the first cell cycle. During differentiation of embryonic (i.e., inner cell mass [ICM]) and extraembryonic (i.e., trophectoderm [TE] and placenta) tissues, different DNAmethylation and histone-modification profiles are established between lineages (Morgan et al. 2005). Somatic cloning cannot faithfully recapitulate these patterns of reprogramming, showing rapid but less extensive demethylation in the first cell cycle, and perturbed DNA methylation and histone lysine methylation between ICM and TE cells. A closely related concern in somatic cell reprogramming is the fate of imprinted gene loci. For normal embryonic development to proceed, correct allelic expression at imprinted loci is required (Chapter 19).

This was demonstrated by the seminal experiments that generated uniparental embryos (Barton et al. 1984; McGrath and Solter 1984; Surani et al. 1984). Androgenetic embryos (both genomes are of male origin) exhibited retarded embryonic development but hyperproliferation of extraembryonic tissues (e.g., placenta). In gyno- or parthenogenetic embryos (both genomes are of female origin), the placenta is underdeveloped. A parent-specific imprint must therefore be established in the germ cell following erasure of preexisting marks (Chapter 20). It is believed that this occurs for approximately 100 or more imprinted genes, largely involved in systems of resource provision for embryonic and placental development (e.g., Igf2 growth factor). Intriguingly, there is evidence that imprinting may be perturbed during in vitro culture of embryos produced by ART or nuclear transfer (Maher 2005). 15 Cancer

There is a delicate balance between self-renewal and differentiation. Neoplastic transformation (also similarly referred to as tumorigenesis) is regarded as the process whereby cells undergo a change involving uncontrolled cell proliferation, a loss of checkpoint control tolerating the accumulation of chromosomal aberrations and genomic aneuploidies, and mis-regulated differentiation (Lengauer et al. 1998). It is commonly thought to be caused by at least one genetic lesion, such as a point mutation, a deletion, or a translocation, disrupting either a tumor suppressor gene or an oncogene (Hanahan and Weinberg 2000). Tumor suppressor genes become silenced in tumor cells. Oncogenes are activated through dominant mutations or overexpression of a normal gene (proto-oncogene). Importantly, an accumulation of aberrant epigenetic modifications is also associated with tumor cells (see Chapter 24). The epigenetic changes involve altered DNA methylation patterns, histone modifications, and chromatin structure (see Fig. 19). Thus, neoplastic transformation is a complex multistep process involving the random activation of oncogenes and/or the silencing of tumor suppressor genes, through genetic or epigenetic events, and is referred to as the "Knudson twohit" theory (Feinberg 2004; Feinberg and Tyko 2004). To illustrate, silencing of the retinoblastoma (Rb) gene, a tumor suppressor, causes loss of checkpoint control, which not only provides a proliferative advantage, but also promotes a "second hit" by affecting downstream functions related to chromatin structure which maintain genome integrity (Gonzalo and Blasco 2005). Inappro-

OVERVIEW

priate activation of an oncogenic product such as the myc gene can have a similar effect (Knoepfler et al 2006). One question raised by current research is, To what extent do aberrant epigenetic changes contribute to the incidence and overall behavior of a tumor? This was addressed by NT experiments using a melanoma cell nucleus as the donor (Hochedlinger et a1. 2004). Any genetic lesions of the donor cell remain; however, NT erases the epigenetic makeup. The tumor incidence of cloned mouse fetuses was then studied, indicating that the spectrum of tumors that arose de novo varied greatly, consistent with different contributions of epigenetic modifications in different tissues that trigger neoplastic progression. DNA hypomethylation (as opposed to hypermethylation) can occur at discrete loci or over Widespread chromosomal regions. DNA hypomethylation was, in fact, the first type of epigenetic transition to be associated with cancer (Feinberg and Vogelstein 1983). This has turned out to be a widespread phenotype of cancer cells. At the individual gene level, DNA hypomethylation can be neoplastic due to the activation of proto-oncogenes, the derepression of genes that cause aberrant cell function, or the biallelic expression of imprinted genes (also termed loss of imprinting or LO!) (see Chapters 23 and 24). On a more global genomic scale, broad DNA hypomethylation, particularly at regions of constitutive heterochromatin, predisposes cells to chromosomal translocations and aneuploidies that contribute to cancer progression. This effect is recapitulated in Dnmtl mutants (Chen et a1. 1998). The genomic instability that ensues when there is DNA hypomethylation is due likely

AND

CONCEPTS

to the mutagenic effect of transposon reactivation. With attention turning to the essential role that repressive histone modifications play in maintaining heterochromatin at centromeres and telomeres, evidence has emerged that if these marks are lost, genome instability also results, contributing to cancer progression (Gonzalo and Blasco 2005). Conversely, DNA hypermethylation is concentrated at the promoter regions of CpG islands in many cancers. Silencing of tumor suppressor genes through such aberrant DNA hypermethylation is particularly critical in cancer progression. Recent studies have revealed that there is considerable cross talk between chromatin modifications and DNA methylation, demonstrating that more than one epigenetic mechanism can be involved in the silencing of a tumor suppressor gene. As an illustration, it is known that the tumor suppressor genes, p 16 and hMLH1, are silenced by both DNA methylation and repressive histone lysine methylation in cancer (McGarvey et a1. 2006). The deregulation of chromatin modifiers is implicated in many forms of cancer. Certain histone-modifying enzymes become oncogenic, such as the PcG protein EZH2 and the trxG protein MLL, and exert their effect through perturbing a cell's epigenetic identity, which consequently either transcriptionally silences or activates inappropriate genes (Schneider et a1. 2002; Valk-Lingbeek et a1. 2004). It is clear that the epigenetic identity is crucial to cellular function. In fact, the pattern of global acetyl and methyl histone marks is proving to be a hallmark for the progression of certain cancers, as demonstrated by a study in prostate tumor progression (Seligson et a1. 2005).

Figure 19. Epigenetic Modifications in Cancer

normal

tumor (0) Aberrant epigenetic marks at cancer-

OFF

oncogene

tumor suppressor ON

oncogene

tumor suppressor

OFF

...

b 5-aza zebularine

...

SAHA ~

Dnmt inhibitors tumor suppressor

OFF

HDAC inhibitors tumor suppressor ON

causing loci typically involve the derepression of oncogenes or silencing of tumor suppressor genes. Epigenetic marks known to alter a normal cell include DNA methylation, repressive histone methylation, and histone deacetylation. (b) The use of epigenetic therapeutic agents for the treatment of cancer has consequences on the chromatin template, illustrated for a tumor suppressor locus. Exposure to Dnmt inhibitors results in a loss of DNA methylation, and exposure to HDAC inhibitors results in the acquisition of histone acetyl marks and subsequent downstream modifications, including active histone methyl marks and the incorporation of histone variants. The cumulative chromatin changes lead to gene re-expression.

52 •

C HAP T E R 3

The development of drug targets inhibiting the function of the chromatin-modifying effector enzymes has opened up a new horizon for cancer therapeutics (see Fig. 19). The use of DNMT and HDAC inhibitors is in the most advanced stages of clinical trials in this new generation of cancer therapeutics. Zebularine and SAHA are, respectively, two such inhibitors. They are particularly beneficial for cancer cells that have repressed tumor suppressor genes (J.e. Cheng et al. 2004; Garcia-Manero and Issa 2005; Marks and Jiang 2005), because treatment leads to transcriptional stimulation. A major proportion of repressive histone lysine methylation is lost during treatment, most probably due to transcription-coupled histone exchange and nucleosome replacement; however, these inhibitors do not significantly alter H3K9me3 at target promoter regions (McGarvey et al. 2006). It remains to be resolved whether repressive marks that persist could induce subsequent re-silencing of tumor suppressor genes when treatment is paused, thereby counteracting the benefit of "epigenetic therapy." It is possible that a dual epigenetic therapy strategy, using DNMT and HDAC inhibitors, may promise a better prognosis in clinical trials. Identification of inhibitors to other classes of histone modifiers, namely HKMTs and PRMTs, is currently in the development phase. There are approximately 50 SET domain HKMTs alone in the mammalian genome. Most of the well-characterized enzymes, such as SUV39H, EZH2, MLL, and RIZ, have already been implicated in tumor development (Schneider et al. 2002). Thus, highthroughput screens (HTS) are being employed in efforts to identify small-molecule inhibitors that could be used in exploratory research and, eventually, cancer therapy. All the classes of histone-modifying enzymes are suited for such an approach, as their specific substrate-binding sites (i.e., to histone peptides), in contrast to generic cofactor (e.g., acetyl-CoA and SAM) binding sites, would allow more selective drug development. HTS have been successful for HDACs (Su et al. 2000), PRMTs (D. Cheng et al. 2004), and HKMTs (Greiner et al. 2005). For the transfer of knowledge to occur from basic to applied research, both hypothesis-driven and empirical approaches are required to ultimately define the efficacy and usefulness of any histone-modifying enzyme inhibitor. For instance, selective HKMT inhibitors against MLL or EZH2 may be valuable therapeutic agents for leukemia or prostate cancer. Alternatively, the use of a SUV39H HKMT inhibitor, which would seem counterintuitive because of the necessity of this enzyme in maintaining constitutive heterochromatin and genome

stability, may still preferentially sensitize tumor cells. In addition, analysis of the HDAC inhibitor SAHA has revealed that it may operate through additional pathways that are distinct from transcriptional reactivation (Marks and Jiang 2005). For example, HDAC inhibitors can also sensitize chromatin lesions, inhibiting efficient DNA repair and permitting genomic instabilities that can trigger apoptosis in tumor cells. These observations will have to be monitored when assessing the efficacy of dual combination therapies. Judging from the results to date, however, it is conceivable that combination therapy using HDAC and HKMT inhibitors may be more selective in killing pro-neoplastic cells by driving them into information overflow and chromatin catastrophe. It is hoped that continued research will identify the viable candidates for efficient epigenetic cancer therapy. 16 What Does Epigenetic Control Actually Do?

Approximately 10% of the protein pool encoded by the mammalian genome plays a role in transcription or chromatin regulation (Swiss-Prot database). Given that the mammalian genome consists of 3 X 10 9 bp, it must accommodate ~ 1 X 10 7 nucleosomes. This gives rise to an overwhelming array of possible regulatory messages, including DNA-binding interactions, histone modifications, histone variants, nucleosome remodeling, DNA methylation, and noncoding RNAs. Yet, the process of transcriptional regulation alone is quite intricate, often requiring the assembly of large multiprotein complexes (> 100 proteins) to ensure initiation, elongation, and correct processing of messenger RNA from a single selected promoter. If DNA sequence-specific regulation is so elaborate, one would expect the lower-affinity associations along the dynamic DNA-histone polymer to be even more so. On the basis of these considerations, rarely will there will be one modification that correlates with one epigenetic state. More likely, and as experimental evidence suggests, it is the combination or cumulative effect of several (probably many) signals over an extended chromatin region that stabilizes and propagates epigenetic states (Fischle et al. 2003b; Lachner et al. 2003; Henikoff 2005). For the most part, transcription factor binding is transient and lost in successive cell divisions. For persistent gene expression patterns, transcription factors are required at each subsequent cell division. As such, epigenetic control can potentiate a primary signal (e.g., promoter stimulation, gene silencing, centromere definition) to successive (but not indefinite) cell generations by the

a v E R V lEW

heritable transmission of information through the chromatin template (Fig. 20). Interestingly, in S. pombe, Swi6-dependent epigenetic variegation can be suppressed for many cell divisions during both mitosis and meiosis (Grewal and Klar 1996) by histone modifications (most probably H3K9me2). Analogous studies were performed in Drosophila using a pulse of an activating transcription factor to transmit cellular memory for Hox gene expression during the female germ line (Cavalli and Paro 1999). In both of these examples, epigenetic memory is mediated by chromatin alterations that comprise distinct histone modifications and, most likely, also the incorporation of histone variants. If histone modifications function together, an imprint may be left on the chromatin template that will help to mark nucleosomes, particularly if a signal is reestablished after DNA replication (Fig. 20). For even more stable inheritance, collaboration between histone modifications, histone variant incorporation, and chromatin remodeling will convert an extended chromatin region into persistent structural alterations that can then be propagated over many cell divisions. Although explained for the inheritance of transcriptional "ON" states, a similar synergy between repressive epigenetic mechanisms will more stably lock silenced chromatin regions, which is further reinforced by additional DNA methylation. The DNA double helix can be viewed then as a selforganizing polymer which, through its ordering into chromatin, can respond to epigenetic control and amplify a primary signal into a more long-term "memory." In addition, many histone modifications probably evolved in response to intrinsic and external stimuli. In keeping with this, chromatin-modifying enzymes require cofactors, such as ATP (kinases), acetyl-CoA (HATs), and SAM (HKMTs), whose levels are dictated by environmental changes (e.g., diet). Thus, the altered conditions can be translated into a more dynamic or stable DNA-histone polymer. An excellent example is the NAD-dependent HDAC, Sir2, which acts as "sensor" for nutrients and life span/aged cells (Guarente and Picard 2005; Rine 2005). Understanding how these environmental cues are cast into biologically relevant epigenetic signatures, and how they are read, translated, and inherited, lies at the heart of current epigenetic research. It is, however, important to stress that epigenetic control requires an intricate balance between many factors and that functional interaction is not always faithfully reestablished after each cell division. This is a functional contrast with genetics, which involves alteration of the DNA sequence, which is always stably propagated

AND

CON C E P T 5

through mitosis and meiosis, if the mutation occurs in the germ line. An important question arising from the above considerations is how the information contained in the chromatin is maintained from mother to daughter cells. If a cell loses its identity, through disease, misregulation, or reprogramming, is this identity loss accompanied by changes in chromatin structure? Bulk synthesis of most core histones is highly regulated during the cell cycle. Transcription of the core histone genes generally occurs during the S phase, the stage when DNA is replicated (replication coupled). This "coordination" assures that as the amount of DNA is doubled in the cell, there are sufficient core histones to be deposited onto the newly replicated DNA, and thus, the packaging of the DNA occurs simultaneously with DNA replication. As presented above, various regions of chromatin may have distinct differences in histone modifications that program the region to be either transcribed or not. How do domains of the newly synthesized daughter chromatin retain this crucial information for appropriate gene expression? How is the program faithfully templated from one cell generation to the next, or through meiosis and germ-cell formation (sperm and egg)? These central questions await future investigation. Although initial studies indicated a semiconservative process, wherein a new H3/H4 tetramer is deposited, followed by the incorporation of two new H2A/H2B dimers, recent data have challenged this hypothesis. In this recent model, the "new" H3 and H4 polypeptides, which may already carry several posttranslational modifications, are incorporated as newly synthesized H3/H4 histone dimers together with the "old" H3/H4 dimers segregating between the mother and daughter DNA. If this is the case, then the modified, parental H3/H4 dimers would now also be present with the newly synthesized dimers on the same DNA. Their co-presence may then dictate that appropriate modifications are placed on the newly added dimers (Tagami et al. 2004). This model is attractive and might help explain the inheritance of histone modifications, and thus, the propagation of epigenetic information through DNA replication and cell division. However, more evidence is needed to support the validity of this or other intriguing models to explain the transmission of chromatin marks through cell division. In closing this chapter, we ask, Does epigenetic control differ in a fundamental way from basic genetic principles? Although we may wish to view Waddington's epigenetic landscape as being demarcated patches of activating versus repressive histone modifications along

54 • C HAP

T ER

OFF

transient signal primary signal

OFF

marked

OFF

recurring signal

epigenetic pro g ram

primary signal epigenetic control

\.-.. •

J -r-.r..v-rT "

))/j~~))~!,.

\.~

partial loss

Figure 20. Epigenetic Potentiation of a Primary Signal (Memory/Inheritance) Classic genetics predicts that gene expression is dependent on the availability and binding of the appropriate panel of transcription factors (TF). Removal of such factors (i.e., a primary signal) results in the loss of gene expression, and thus constitutes a transient activating signal (top). Chromatin structure contributes to gene expression, where some conformations are repressive and others active. The activation of a locus may therefore occur through a primary signal and result in the downstream change in chromatin structure, involving active covalent histone marks (mod) and the replacement of core histones with variants (e.g., H3.3). Through cell division, this chromatin structure may only be reestablished in the presence of an activating signal (denoted "recurring signal"). Epigenetic memory results in the maintenance of a chromatin state through cell division, even in the absence of the primary activating signal. Such a memory system is not absolute, but involves multiple levels of epigenetic regulation for remodeling chromatin structure. The dynamic nature of chromatin means that although a chromatin state may be mitotically stable, it is nonetheless prone to change, hence affecting the longevity of epigenetic memory.

the continuum of the chromatin polymer, this notion could easily be overinterpreted. It is only in recent years that we have learned about the major enzymatic systems through which histone modifications might be propagated. This has shaped our current thinking about the stability, and hence the inheritance, of certain histone marks. In addition, it is underscored by the recent discoveries showing that mutations in chromatin-modifying activities, such as nucleosome remodelers (Cho et al. 2004; Mohrmann and Verrijzer 2005), DNMTs (Robertson 2005), HDACs or HMKTs (Schneider et al. 2002), as they are frequently found in abnormal development and neoplasia, are telling examples of the ultimate power of genetic control. As such, tumor incidence in these mutant mice is generally regarded as a genetic disease. In contrast, alterations in nudeosome structure, DNA methylation, and histone modification profiles-that are not caused by a mutated gene-would classify as "true"

epigenetic aberrations. Excellent examples of these more plastic systems are stochastic decisions in early embryonic development, reprogramming by nuclear transfer, transcriptional memory, genomic imprinting, mosaic X inactivation, centromere identity, and tumor progression. Genetics and epigenetics are thus closely related phenomena, and inherent to both is their propagation through cell division, which, for genetic control, also comprises the germ line, if mutations occur in germ cells. In the case of other-often too easily categorized-epigenetic modifications, we do not know whether they only reflect a minor and transient response to changes in the external environment or significantly contribute to phenotypic differences that can then be maintained over many, but not indefinite, somatic cell divisions, and occasionally affect the germ line. Even with our greatly improved knowledge of epigenetic mechanisms today, there is little, or no, novel support for Lamarckism.

a v E R V lEW 17 Big Questions in Epigenetic Research

This book discusses the fundamental concepts and general principles that explain how epigenetic phenomena occur, as puzzling as they may seem. Our ultimate goal is to expose the reader to the current understanding of mechanisms that guide and shape these concepts, drawing upon the rich biology from which they emerge. In just a few years, epigenetic research has prompted exciting and remarkable insights and breakthrough discoveries, yet many long-standing questions remain unanswered (see Fig. 21). Although it is tempting to draw broad-brush conclusions and to propound general rules from this progress, we caution against this tendency, suspecting that there will be many exceptions that break the rules. For example, it is clear that striking organismal differences occur. Notably, from unicellular to multicellular organisms, the extent and type of histone modifications, histone variants, DNA methylation, and use of the RNAi machinery does vary. There are, however, plenty of reasons for renewed energy in research programs designed to gain molecular insights into epigenetic phenomena. Elegant biochemical and genetic studies have already successfully dissected many of the functional aspects of these pathways, in an unprecedented manner. It could therefore be predicted that careful analysis of epigenetic transitions in different cell types (e.g., stem versus differentiated; resting versus proliferating) will uncover hallmarks of pluripotency (Bernstein et aI. 2006; Boyer et al. 2006; Lee et al. 2006). This will most likely be valuable in diagnosing which chromatin alter-

AND

CON C E P T 5

ations are significant during normal differentiation as compared with disease states and tumorigenesis. For example, using large-scale mapping approaches with normal, tumor, or ES cells-"epigenetic landscaping" along entire chromosomes (Brachen et al. 2006b; Squazzo et al. 2006; Epigenomics AG, ENCODE, GEN-AU, EPIGENOME NoE)-it is anticipated that the knowledge generated could be harnessed for novel therapeutic intervention approaches and work toward promoting a worldwide consortium to map the entire human epigenome (Jones and Martienssen 2005). It is conceivable that differences in the relative abundance between distinct histone modifications, such as the apparent underrepresentation of repressive histone lysine tri-methylation in S. pombe and A. thaliana, may reflect the greater proliferative and regenerative potential in these organisms as compared to the more restricted developmental programs of metazoan systems. In addition, the functional links between the RNAi machinery, histone lysine methylation, and DNA methylation will continue to provide exciting surprises into the complex mechanisms required for cell-fate determination during development. Similarly, an enhanced understanding of the dynamics and specificity of nucleosome-remodeling machines will contribute to this end. We predict that more "exotic" enzymatic activities will be uncovered, catalyzing epigenetic transitions through modifications of histone and non-histone substrates. It would appear that chromatin alterations, as induced by the above mechanisms, act largely as a response filter to the environment. Thus, it is hoped that this knowledge can ultimately be applied to enhanced therapeutic strategies for resetting

ENVIRONMENT

! ! ! epigenetic code?

C E L

F A r I:

regeneration?

epigenetic inheritance? mod

remodeler

nature of cellular memory? germ line imprint?

stem cells?

cell type identity?

ncRNAs

non-coding RNAs?

aging? epigenetic dysfunction?

1 1 1 ENVIRONMENT

Figure 21. Big Questions in Epigenetic Research The many experimental systems used in epigenetic research have unveiled numerous pathways and novel insights into the mechanisms of epigenetic control. Many questions, as shown in the figure, still remain and require further elucidation or substantiation in new and existing model systems and methods.

C HAP T E R 3

some of an individual's epigenetic response that contribute to aging, disease, and cancer. This includes tissue regeneration, therapeutic cloning (using ES cells and their derivatives), and adult stem cell therapy strategies. It is believed such strategies will extend cellular life span, modulate stress responses to external stimuli, reverse disease progression, and improve assisted reproductive technologies. We predict that understanding the "chromatin basis" of pluripotency and totipotency will lie at the heart of understanding stem cell biology and its potential for therapeutic intervention. Many fundamental epigenetic questions remain. For example, What distinguishes one chromatin strand from the other allele when both contain the same DNA sequence in the same nuclear environment? What defines the mechanisms conferring inheritance and propagation of epigenetic information? What is the molecular nature of cellular memory? Are there epigenetic imprints in the germ line that serve to keep this genome in a totipotent state? If so, how are these marks erased during development? Alternatively, or in addition, are new imprints added during development that serve to "lock in" differentiated states? We look forward to the next generation of studies (and students) bold enough to tackle these questions with the heart and passion of previous generations of genetic and epigenetic researchers. In summary, the genetic principles described by Mendel likely govern the vast majority of our development and our outward phenotypes. However, exceptions to the rule can sometimes reveal new principles and new mechanisms leading to inheritance that have been underestimated, and in some cases, poorly understood previously. This book hopes to expose its readers to the newly appreciated basis of phenotypic variation-one that lies outside of DNA alteration. It is our hope that the systems and concepts described in this book will provide a useful foundation for future generations of students and researchers alike who become intrigued by the curiosities of epigenetic phenomena.

References Ahmad K. and Henikoff S. 2002. The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9: 1191-1200. Allfrey Y.G., Faulkner R., and Mirsky A.E. 1964. Acetylation and methylation of histones and their possible role in the regulation of RNA synthesis. Proc. Natl. Acad. Sci. 51: 786-794. Almeida R. and Allshire R.C 2005. RNA silencing and genome regulation. Trends Cell BioI. 15: 251-258. Aparicio O.M., Billington B.L., and Gottschling D.E. 1991. Modifiers of position effect are shared between telomeric and silent mating-

type loci in S. cerevisiae. Cell 66: 1279-1287. Avery O.T., Macleod CM., and McCarty M. 1944. Studies on the chemical nature of the substance inducing transformation of pneumococcal types. Induction of transformation by a desoxyribonucleic acid fraction isolated from pneumococcus Type III.]. Exp. Med. 79: 137-158. Avner P. and Heard E. 2001. X-chromosome inactivation: Counting, choice and initiation. Nat. Rev. Genet. 2: 59-67. Bannister A.J. and Kouzarides T. 2005. Reversing histone methylation. Nature 436: 1103-1106. Bannister A.J., Zegerman P., Partridge J.F., Miska E.A., Thomas rO., Allshire R.C, and Kouzarides T. 2001. Selective recognition of methylated lysine 9 on histone H3 by the HPI chromo domain. Nature 410: 120-124. Barton S.C, Surani M.A., and orris M.L. 1984. Role of paternal and maternal genomes in mouse development. Nature 311: 374-376. Bastow R., Mylne J.S., Lister C, Lippman Z., Martienssen R.A., and Dean C 2004. Vernalization requires epigenetic silencing of FLC by histone methylation. Nature 427: 164-167. Baxter J., Sauer 5., Peters A., John R., Williams R., Caparros M.L., Arney K., Otte A., Jenuwein T., Merkenschlager M., and Fisher A.G. 2004. Histone hypomethylation is an indicator of epigenetic plasticity in quiescent lymphocytes. EMBO ]. 23: 4462--4472. Berger S.L. 2002. Histone modifications in transcriptional regulation. Curro Opin. Genet. Dev. 12: 142-148. Bernard P., Maure J.F., Partridge J.F., Genier 5., Javerzat J.P., and Allshire R.C 2001. Requirement of heterochromatin for cohesion at centromeres. Science 294: 2539-2542. Bernstein B.E., Kamal M., Lindblad-Toh K., Bekiranov 5., Bailey D.K., Huebert D.J., McMahon 5., Karlsson E.K., Kulbokas E.J., III, Gingeras T.R., et al. 2005. Genomic maps and comparative analysis of histone modifications in human and mouse. Cell 120: 169-181. Bernstein RE., Mikkelsen T.S., Xie X., Kamal M., Huebert D.J., Cuff J., Fry B., Meissner A., Wernig M., Plath K., et al. 2006. A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 125: 315-326. Bernstein E. and Allis CD. 2005. RNA meets chromatin. Genes Dev. 19: 1635-1655. Bird A.P. 1986. CpG-rich islands and the function of DNA methylation. Nature 321: 209-213. Birky CW., Jr. 2001. The inheritance of genes in mitochondria and chloroplasts: Laws, mechanisms, and models. Annu. Rev. Genet. 35: 125-148. Bolland DJ, Wood A.L., Johnston CM., Bunting S.F., Morgan G., Chakalova L., Fraser P.J., and Corcoran A.E. 2004. Antisense intergenic transcription in V(D)J recombination. Nat. Immunol. 5: 630-637. Bourc'his D. and Bestor T.H. 2004. Meiotic catastrophe and retrotransposon reactivation in male germ cells lacking Dnmt3L. Nature 431: 96-99. Boyer L.A., Plath K., Zeitlinger J., Brambrink T., Medeiros L.A., Lee T.I., Levine 5.5., Wernig M., Tajonar A., Ray M.K., et al. 2006. Polycomb complexes repress developmental regulators in murine embryonic stem cells. Nature 441: 349-353. Bracken A.P., Dietrich N., Pasini D., Hansen K.H., and Helin K. 2006. Genome-wide mapping of Polycomb target genes unravels their roles in cell fate transitions. Genes Dev. 20: 1123-1136. Braig M., Lee 5., Loddenkemper C, Rudolph C, Peters A.H., Schlegelberger B., Stein H., Dorken B., Jenuwein T., and Schmitt CA. 2005. Oncogene-induced senescence as an initial barrier in lymphoma development. Nature 436: 660-665. Brownell J.E., Zhou J., Ranalli T., Kobayashi R., Edmondson D.G.,

o Roth S.Y., and Allis e.D. 1996. Tetrahymena histone acetyltransferase A: A hom*olog to yeast Gcn5p linking histone acetylation to gene activation. Cell 84: 843-851. Cairns B.R. 2005. Chromatin remodeling complexes: Strength in diversity, precision through specialization. Curro Opin. Genet. Dev. 15: 185-190. Campbell K.H., McWhir J., Ritchie WA., and Wilmut 1. 1996. Sheep cloned by nuclear transfer from a cultured cell line. Nature 380: 64--66. Cao R., Wang L., Wang H., Xia L., Erdjument-Bromage H., Tempst P., Jones R.S., and Zhang Y. 2002. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science 298: 1039-1043. Cavalli G. and Paro R. 1999. Epigenetic inheritance of active chromatin after removal of the main transactivator. Science 286: 955-958. Chakalova L., Debrand E., Mitchell J.A., Osborne e.S., and Fraser P. 2005. Replication and transcription: Shaping the landscape of the genome. Nat. Rev. Genet. 6: 669-677. Chalker D.L. and Yao M.e. 2001. Nongenic, bidirectional transcription precedes and may promote developmental DNA deletion in Tetrahymena thermophila. Genes Dev. 15: 1287-1298. Chan S.w., Zilberman D., Xie Z., Johansen L.K., Carrington J.e., and Jacobsen S.E.2004. RNA silencing genes control de novo DNA methylation. Science 303: 1336. Chen R.Z., Pettersson U., Beard e., Jackson-Grusby L., and Jaenisch R. 1998. D A hypomethylation leads to elevated mutation rates. Nature 395: 89-93. Cheng D., Yadav ., King R.W., Swanson M.S., Weinstein E.J., and Bedford M.T. 2004. Small molecule regulators of protein arginine methyltransferases.]. BioI. Chem. 279: 23892-23899. Cheng J.e., Yoo e.B., Weisenberger D.J., Chuang J., Wozniak e., Liang G., Marquez Y.E., Greer S., Orntoft TF., Thykjaer T, and Jones P.A. 2004. Preferential response of cancer cells to zebularine. Cancer Cell 6: 151-158. Cho K.S., Elizondo L.1., and Boerkoel e.F. 2004. Advances in chromatin remodeling and human disease. Curro Opin. Genet. Dev. 14: 308-315. Chuikov S., Kurash J.K., Wilson J.R., Xiao B., Justin N., Ivanov G.S., McKinney K., Tempst P., Prives e., Gamblin S.J., et al. Regulation of p53 activity through lysine methylation. Nature 432: 353-360. Clark-Adams e.D., Norris D., Osley M.A., Fassler J.S., and Winston F. 1988. Changes in histone gene dosage alter transcription in yeast. Genes Dev. 2: 150-159. Cohen EE. and Prusiner S.B. 1998. Pathologic conformations of prion proteins. Annu. Rev. Biochem. 67: 793-819. Cosgrove M.S., Boeke J.D., and Wolberger e. 2004. Regulated nucleosome mobility and the histone code. Nat. Struct. Mol. Bioi. 11: 1037-1043. Cremer T. and Cremer e. 2001. Chromosome territories, nuclear architecture and gene regulation in mammalian cells. Nat. Rev. Genet. 2: 292-301. Oaujat S., Zeissler 0., Waldmann T., Happel N., and Schneider R. 2005. HPI binds specifically to Lys26-methylated histone H1.4, whereas simultaneous Ser27 phosphorylation blocks HPI binding. f. BioI. Chem. 280: 38090-38095. Oellino G./., Schwartz Y.B., Farkas G., McCabe D., Elgin S.e., and Pirrotta Y. 2004. Polycomb silencing blocks transcription initiation. Mol. Cell 13: 887-893. Ohalluin C., Carlson J.E., Zeng L., He e., Aggarwal A.K., and Zhou M.M.1999. Structure and ligand of a histone acetyl transferase bromodomain. Nature 399: 491-496. Oi Croce L. 2005. Chromatin modifying activity of leukaemia associ-

V E R V lEW

AND

CON C E P T 5

ated fusion proteins. Hum. Mol. Genet. 14 Spec. No.1: R77-R84. Dou Y. and Gorovsky M.A. 2000. Phosphorylation of linker histone HI regulates gene expression in vivo by creating a charge patch. Mol. Cell 6: 225-231. Fan Y., Nikitina T, Zhao J., Fleury TJ., Bhattacharyya R., Bouhassira E.E., Stein A., Woodco*ck e.L., and Skoultchi A.!. 2005. Histone HI depletion in mammals alters global chromatin structure but causes specific changes in gene regulation. Cell 123: 1199-1212. Feinberg A.P. 2004. The epigenetics of cancer etiology. Semin. Cancer Bioi. 14: 427-432. Feinberg A.P. and Tycko B. 2004. The history of cancer epigenetics. Nat. Rev. Cancer 4: 143-153. Feinberg A.P. and Vogelstein B. 1983. Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature 301: 89-92. Felsenfeld G. and Groudine M. 2003. Controlling the double helix. Nature 421: 448-453. Fischle W., Wang Y., and Allis e.D. 2003a. Binary switches and modification cassettes in histone biology and beyond. Nature 425: 475-479. - - - . 2003b. Histone and chromatin cross-talk. Curro Opin. Cell BioI. 15: 172-183. Fischle W, Tseng B.S., Dormann H.L., Ueberheide B.M., Garcia B.A., Shabanowitz J., Hunt D.E, Funabiki H., and Allis e.D. 2005. Regulation of HPI-chromatin binding by histone H3 methylation and phosphorylation. Nature 438: 1116-1122. Fisher A.G. and Merkenschlager M. 2002. Gene silencing, cell fate and nuclear organisation. Curro Opin. Genet. Dev. 12: 193-197. Fodor B.D., Kubicek S., Yonezawa M., O'Sullivan R.J., Sengupta R., Perez-Burgos L., Opravil S., Mechtler K., Schotta G., and Jenuwein T 2006. Jmjd2b antagonizes H3K9 trimethylation at pericentric heterochromatin in mammalian cells. Genes Dev. 20: 1557-1562. Fraga M.E, Ballestar E., Paz M.F., Ropero S., Setien F., Ballestar M.L., Heine-Suner D., Cigudosa J.e., Urioste M., Benitez J., et al. 2005. Epigenetic differences arise during the lifetime of monozygotic twins. Proc. Natl. Acad. Sci. 102: 10604--10609. Francis N.J., Kingston R.E., and Woodco*ck e.L. 2004. Chromatin compaction by a polycomb group protein complex. Science 306: 1574-1577. f*ckagawa T, Nogami M., Yoshikawa M., Ikeno M., Okazaki T, Takami Y., Nakayama T, and Oshimura M. 2004. Dicer is essential for formation of the heterochromatin structure in vertebrate cells. Nat. Cell BioI. 6: 784-791. Garcia-Manero G. and Issa J.P. 2005. Histone deacetylase inhibitors: A review of their clinical status as antineoplastic agents. Cancer Invest. 23: 635-642. Gilbert N., Boyle S., Fiegler H., Wood fine K., Carter N.P., and Bickmore W.A. 2004. Chromatin architecture of the human genome: Generich domains are enriched in open chromatin fibers. Cell 118: 555-566. Gilfillan G.D., Dahlsveen 1.K., and Becker P.B. 2004. Lifting a chromosome: Dosage compensation in Drosophila melanogaster. FEBS Lett. 567: 8-14. Goll M.G., Kirpekar F., Maggert K.A., Yoder J.A., Hsieh e.L., Zhang X., Golic K.G., Jacobsen S.E., and Bestor TH. 2006. Methylation of tR AAsp by the D A methyltransferase hom*olog Dnmt2. Science 311: 395-398. Gonzalo S. and Blasco M.A. 2005. Role of Rb family in the epigenetic definition of chromatin. Cell Cycle 4: 752-755. Gottschling D.E., Aparicio O.M., Billington B.L., and Zakian Y.A. 1990. Position effect at S. cerevisiae telomeres: Reversible repression of Pol II transcription. Cell 63: 751-762.

58 • C HAP

T ER

Grant P.A., Duggan L., Cote J., Roberts S.M., Brownell J.E., Candau R., Ohba R., Owen-Hughes T, Allis C.D., Winston P., et al. 1997. Yeast Gcn5 functions in two multisubunit complexes to acetylate nucleosomal histones: Characterization of an Ada complex and the SAGA (Spt/Ada) complex. Genes Dev. 11: 1640-1650. Greiner D., Bonaldi T., Eskeland R., Roemer E., and Imhof A. 2005. Identification of a specific inhibitor of the histone methyltransferase SU(VAR)3-9. Nat. Chem. BioI. 1: 143-145. Grewal S.l. and Klar A.I. 1996. Chromosomal inheritance of epigenetic states in fission yeast during mitosis and meiosis. Cell 86: 95-101. Grozinger C.M. and Schreiber S.L. 2002. Deacetylase enzymes: Biological functions and the use of small-molecule inhibitors. Chem. BioI. 9: 3-16. Guarente L. and Picard P. 2005. Calorie restriction-The SIR2 connection. Cell 120: 473-482. Hall l.M., Shankaranarayana G.D., Noma K., Ayoub N., Cohen A., and Grewal S.l. 2002. Establishment and maintenance of a heterochromatin domain. Science 297: 2232-2237. Hanahan D. and Weinberg R.A. 2000. The hallmarks of cancer. Cell 100: 57-70. Harvey A.C. and Downs I.A. 2004. What functions do linker histones provide? Mol. Microbiol. 53: 771-775. Hassan A.H., Prochasson P., Neely K.E., Galasinski S.c., Chandy M., Carrozza M.J., and Workman J.L. 2002. Function and selectivity of bromodomains in anchoring chromatin-modifying complexes to promoter nucleosomes. Cell Ill: 369-379. Heard E. 2005. Delving into the diversity of facultative heterochromatin: The epigenetics of the inactive X chromosome. Curro Opin. Genet. Dev. 15: 482-489. Henikoff S. 2005. Histone modifications: Combinatorial complexity or cumulative simplicity? Proc. Natl. Acad. Sci. 102: 5308-5309. Henikoff S. and Ahmad K. 2005. Assembly of variant histones into chromatin. Annu. Rev. Cell Dev. BioI. 21: 133-153. Herr A.J., Jensen M.B., Dalmay T, and Baulcombe D.C. 2005. RNA polymerase IV directs silencing of endogenous DNA. Science 308: 118-120. Hirota T, Lipp J.J., Toh B.H., and Peters J.M. 2005. Histone H3 serine 10 phosphorylation by Aurora B causes HPI dissociation from heterochromatin. Nature 438: 1176-1180. Hochedlinger K., Blelloch R., Brennan c., Yamada Y., Kim M., Chin L., and Jaenisch R. 2004. Reprogramming of a melanoma genome by nuclear transplantation. Genes Dev. 18: 1875-1885. Holbert M.A. and Marmorstein R. 2005. Structure and activity of enzymes that remove histone modifications. Curro Opin. Struct. BioI. 15: 673-680. Holliday R. 1994. Epigenetics: An overview. Dev. Genet. 15: 453-457. Ishii K.J. and Akira S. 2005. TLR ignores methylated RNA? Immunity 23: 111-113. Jackson J.P., Lindroth A.M., Cao X., and Jacobsen S.E. 2002. Control of CpNpG DNA methylation by the KRYPTONITE histone H3 methyltransferase. Nature 416: 556-560. Jacobson R.H., Ladurner A.G., King D.S. and Tjian R. 2000. Structure and function of a human TAFII250 double bromodomain module. Science. 288: 1422-1425. Jaenisch R. and Bird A. 2003. Epigenetic regulation of gene expression: How the genome integrates intrinsic and environmental signals. Nat. Genet. (suppl.) 33: 245-254. Janicki S.M., Tsukamoto T, Salghetti S.E., Tansey W.P., Sachidanandam R., Prasanth KV., Ried T, Shav-Tal Y., Bertrand E., Singer R.H., and Spector D.L. 2004. From silencing to gene expression: Real-time analysis in single cells. Cell 116: 683-698. Jeddeloh J.A., Stokes TL., and Richards E.I. 1999. Maintenance of

genomic methylation requires a SWI2/SNF2-like protein. Nat. Genet. 22: 94-97. Jenuwein T and Allis C.D. 2001. Translating the histone code. Science 293: 1074-1080. Jones P.A. and Martienssen R. 2005. A blueprint for a Human Epigenome Project: The AACR Human Epigenome Workshop. Cancer Res. 65: 11241-11246. Kanellopoulou c., Muljo S.A., Kung A.L., Ganesan S., Drapkin R., Jenuwein T, Livingston D.M., and Rajewsky K. 2005. Dicer-deficient mouse embryonic stem cells are defective in differentiation and centromeric silencing. Genes Dev. 19: 489-501. Kayne P'S., Kim U.J., Han M., Mullen J.R., Yoshizaki P., and Grunstein M. 1988. Extremely conserved histone H4 N terminus is dispensable for growth but essential for repressing the silent mating loci in yeast. Cell 55: 27-39. Khochbin S. 2001. Histone HI diversity: Bridging regulatory signals to linker histone function. Gene 271: 1-12. Khorasanizadeh S. 2004. The nucleosome: From genomic organization to genomic regulation. Cell 116: 259-272. Kim J., Daniel J., Espejo A., Lake A., Krishna M., Xia L., Zhang Y., and Bedford M.T 2006. Tudor, MBT and chromo domains gauge the degree of lysine methylation. EMBO Rep. 4: 397-403. Kimmins S. and Sassone-Corsi P. 2005. Chromatin remodeling and epigenetic features of germ cells. Nature 434: 583-589. Klar A.I. 1998. Propagating epigenetic states through meiosis: Where Mendel's gene is more than a DNA moiety. Trends Genet. 14: 299-301. - - - . 2004. An epigenetic hypothesis for human brain laterality, handedness, and psychosis development. Cold Spring Harbor Symp. Quant. BioI. 69: 499-506. Knoepfler P.S., Zhang X.-Y., Cheng P.P., Gafken P.R., McMahon S.B., and Eisenman R.N. 2006. Myc influences global chromatin structure. EMBO f. 25: 2723-2734. Kornberg R.D. 1974. Chromatin structure: A repeating unit of histones and DNA. Science 184: 868-871. Lachner M., O'Sullivan R.J., and Jenuwein T. 2003. An epigenetic road map for histone lysine methylation. f. Cell Sci. 116: 2117-2124. Lachner M., Sengupta R., Schotta G., and Jenuwein T 2004. Trilogies of histone lysine methylation as epigenetic landmarks of the eukaryotic genome. Cold Spring Harbor Symp. Quant. BioI. 69: 209-218. Lachner M., O'Carroll D., Rea S., Mechtler K., and Jenuwein T 2001. Methylation of histone H3 lysine 9 creates a binding site for HP I proteins. Nature 410: 116-120. Langst G. and Becker P.B. 2004. Nucleosome remodeling: One mechanism, many phenomena? Biochim. Biophys. Acta 1677: 58-63. Lee D.Y., Teyssier c., Strahl B.D., and Stallcup M.R. 2005. Role of protein methylation in regulation of transcription. Endocr. Rev. 26: 147-170. Lee Tl., Jenner R.G., Boyer L.A., Guenther M.G., Levine S.S., Kumar R.M., Chevalier B., Johnstone S.E., Cole M.P., Isono K., et al. 2006. Control of developmental regulators by Polycomb in human embryonic stem cells. Cell 125: 301-313. Lengauer c., Kinzler KW., and Vogelstein B. 1998. Genetic instabilities in human cancers. Nature 396: 643-649. Li E., Bestor TH., and Jaenisch R. 1992. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69: 915-926. Lippman Z., Gendre! A.V., Black M., Vaughn M.W., Dedhia N., McCombie W.R., Lavine K, Mittal v., May B., Kasschau K.D., et al. 2004. Role of transposable elements in heterochromatin and epigenetic control. Nature 430: 471-476.

o V E R V lEW Litt M.D., Simpson M., Gaszner M., Allis c.D., and Felsenfeld G. 2001. Correlation between histone lysine methylation and developmental changes at the chicken beta-globin locus. Science 293: 2453-2455. Luger K., Mader A.W, Richmond R.K., Sargent D.E, and Richmond T.j. 1997. Crystal structure of the nucleosome core particle at 2.8 A resolution. Nature 389: 251-260. Lund A.H. and van Lohuizen M. 2004. Polycomb complexes and silencing mechanisms. Curro Opin. Cell BioI. 16: 239-246. Lyko E 2001. DNA methylation learns to fly. Trends Genet. 17: 169-172. Maher E.R. 2005. Imprinting and assisted reproductive technology. Hum. Mol. Genet. 14 Spec. No.1: RI33-RI38. Maison c., Bailly D., Peters A.H., Quivy j.P., Roche D., Taddei A., Lachner M., jenuwein T., and Almouzni G. 2002. Higher-order structure in pericentric heterochromatin involves a distinct pattern of histone modification and an RNA component. Nat. Genet. 30: 329-334. Marks P.A. and Jiang X. 2005. Histone deacetylase inhibitors in programmed cell death and cancer therapy. Cell Cycle 4: 549-551. Martens j.H., O'Sullivan R.j., Braunschweig U., Opravil S., Radolf M., Steinlein P., and jenuwein T. 2005. The prome of repeat-associated histone lysine methylation states in the mouse epigenome. EMBO ]. 24: 800-812. Maurer-Stroh S., Dickens N.j., Hughes-Davies L., Kouzarides T., Eisenhaber E, and Ponting c.P. 2003. The Tudor domain 'Royal Family': Tudor, plant Agenet, Chromo, PWWP and MBT domains. Trends Biochem. Sci. 28: 69-74. Mayer w., Niveleau A., Walter j., Fundele R., and Haaf T. 2000. Demethylation of the zygotic paternal genome. Nature 403: 501-502. McClintock B. 1951. Chromosome organization and genic expression. Cold Spring Harbor Symp. Quant. BioI. 16: 13-47. McGarvey K., Fahrner j., Green E., Martens j., jenuwein T., and Baylin S.B. 2006. Silenced tumor suppressor genes reactivated by DNA demethylation do not return to a fully euchromatic chromatin state. Cancer Res. 66: 3541-3549. McGrath j. and Solter D. 1984. Completion of mouse embryogenesis requires both the maternal and paternal genomes. Cell 37: 179-183. Metzger E., Wissmann M., Yin N., Muller j.M., Schneider R., Peters A.H., Gunther T., Buettner R., and Schule R. 2005. LSDI demethylates repressive histone marks to promote androgen-receptordependent transcription. Nature 437: 436-439. Meyer B.j., McDonei P., Csankovszki G., and Ralston E. 2004. Sex and X-chromosome-wide repression in Caenorhabditis elegans. Cold Spring Harbor Symp. Quant. BioI. 69: 71-79. Misteli T. 2004. Spatial positioning; a new dimension in genome function. CelllI9: 153-156. Mito Y., Henikoff j.G., and Henikoff S. 2005. Genome-scale proming of histone H3.3 replacement patterns. Nat. Genet. 37: 1090-1097. Mizuguchi G., Shen X., Landry j., Wu WH., Sen S., and Wu C. 2004. ATP-driven exchange of histone H2AZ variant catalyzed by SWR1 chromatin remodeling complex. Science 303: 343-348. Mochizuki K., Fine N.A., Fujisawa T., and Gorovsky M.A. 2002. Analysis of a piwi-related gene implicates small RNAs in genome rearrangement in tetrahymena. Cell lID: 689-699. Mohrmann L. and Verrijzer c.P. 2005. Composition and functional specificity of SWI2/SNF2 class chromatin remodeling complexes. Biochim. Biophys. Acta 1681: 59-73. Morgan H.D., Santos E, Green K., Dean W., and Reik W. 2005. Epigenetic reprogramming in mammals. Hum. Mol. Genet. 14 Spec. No.1: R47-R58.

AND

CON C E P T 5

Motamedi M.R., Verdel A., Colmenares S.U., Gerber S.A., Gygi S.P., and Moazed D. 2004. Two RNAi complexes, RITS and RDRC, physically interact and localize to noncoding centromeric RNAs. CelllI9: 789-802. Muller H.j. 1930. Types of visible variations induced by X-rays in Drosophila. f. Genet. 22: 299-334. Nakayama j., Rice j.c., Strahl B.D., Allis C.D., and Grewal S.l. 200l. Role of histone H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science 292: 110-113. Narita M., Nunez S., Heard E., Narita M., Lin A.W., Hearn S.A., Spector D.L., Hannon G.j., and Lowe S.W. 2003. Rb-mediated heterochromatin formation and silencing of E2F target genes during cellular senescence. Cell1l3: 703-716. Narlikar G.j., Fan H.Y., and Kingston R.E. 2002. Cooperation between complexes that regulate chromatin structure and transcription. Cell 108: 475-487. Nowak S.j. and Corces Y.G. 2004. Phosphorylation of histone H3: A balancing act between chromosome condensation and transcriptional activation. Trends Genet. 20: 214-220. Orphanides G., LeRoy G., Chang C.H., Luse D.S., and Reinberg D. 1998. FACT, a factor that facilitates transcript elongation through nucleosomes. Cell 92: 105-116. Pal-Bhadra M., Leibovitch B.A., Gandhi S.G., Rao M., Bhadra U., Birchler j.A., and Elgin S.c. 2004. Heterochromatic silencing and HP1 localization in Drosophila are dependent on the RNAi machinery. Science 303: 669-672. Paro R. and Hogness D.S. 1991. The Polycomb protein shares a hom*ologous domain with a heterochromatin-associated protein of Drosophila. Proc. Natl. Acad. Sci. 88: 263-267. Petersen-Mahrt S. 2005. DNA deamination in immunity. Immunol. Rev. 203: 80-97. Pirrotta Y. 1998. Polycombing the genome: PcG, trxG, and chromatin silencing. Cell 93: 333-336. Pontier D., Yahubyan G., Vega D., Bulski A., Saez-Vasquez j., Hakimi M.A., Lerbs-Mache S., Colot Y., and Lagrange T. 2005. Reinforcement of silencing at transposons and highly repeated sequences requires the concerted action of two distinct R . A polymerases IV in Arabidopsis. Genes Dev. 19: 2030-2040. Ratcliff E, Harrison B.D., and Baulcombe D.C. 1997. A similarity between viral defense and gene silencing in plants. Science 276: 1558-1560. Razin A. and Riggs A.D. 1980. DNA methylation and gene function. Science 210: 604-610. Rea S., Eisenhaber E, O'Carroll D., Strahl B.D., Sun Z.W, Schmid M., Opravil S., Mechtler K., Ponting c.P., Allis C.D., and jenuwein T. 2000. Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406: 593-599. Reinberg D., Chuikov S., Farnham P., Karachentsev D., Kirmizis A., Kuzmichev A., Margueron R., Nishioka K., Preissner T.S., Sarma K., et al. 2004. Steps toward understanding the inheritance of repressive methyl-lysine marks in histones. Cold Spring Harbor Symp. Quant. BioI. 69: 171-182. Reinhart B.j. and Bartel D.P. 2002. Small RNAs correspond to centromere heterochromatic repeats. Science 297: 183l. Rine j. 2005. Cell biology. Twists in the tale of the aging yeast. Science 310: 1124-1125. Ringrose L. and Paro R. 2004. Epigenetic regulation of cellular memory by the Polycomb and Trithorax group proteins. Annu. Rev. Genet. 38: 413-443. Ringrose L., Ehret H., and Paro R. 2004. Distinct contributions of histone H3 lysine 9 and 27 methylation to locus-specific stability of polycomb complexes. Mol. Cell 16: 641-653.

60 • C HAP

T ER

Robertson K.D. 2005. DNA methylation and human disease. Nat. Rev. Genet. 6: 597-610. RoloffT.C. and Nuber U.A. 2005. Chromatin, epigenetics and stem cells. Eur.]. Cell BioI. 84: 123-135. Roth S.Y., Denu J.M., and Allis C.D. 2001. Histone acetyltransferases. Annu. Rev. Biochem. 70: 81-120. Sanchez-Elsner T., Gou D., Kremmer E., and Sauer F. 2006. Noncoding RNAs of trithorax response elements recruit Drosophila Ashl to ultrabithorax. Science 311: 1118-1123. Santos-Rosa H., Schneider R., Bannister A.J., Sherriff J., Bernstein B.E., Emre N.C., Schreiber S.L., Mellor T., and Kouzarides T. 2002. Active genes are tri-methylated at K4 of histone H3. Nature 419: 407-411. Sarma K. and Reinberg D. 2005. Histone variants meet their match. Nat. Rev. Mol. Cell BioI. 6: 139-149. Scaffidi P., Gordon L., and Misteli T. 2005. The cell nucleus and aging: Tantalizing clues and hopeful promises. PLoS. BioI. 3: e395. Schalch T., Duda S., Sargent D.E, and Richmond T.T. 2005. X-ray structure of a tetranucleosome and its implications for the chromatin fibre. Nature 436: 138-141. Schmitt S., Prestel M., and Paro R. 2005. lntergenic transcription through a polycomb group response element counteracts silencing. Genes Dev. 19: 697-708. Schneider R., Bannister A.T., and Kouzarides T. 2002. Unsafe SETs: Histone lysine methyltransferases and cancer. Trends Biochem. Sci. 27: 396-402. Schreiber S.L. and Bernstein B.E. 2002. Signaling network model of chromatin. Cell 111: 771-778. Schwartz B.E. and Ahmad K. 2005. Transcriptional activation triggers deposition and removal of the histone variant H3.3. Genes Dev. 19: 804-814. Schwartz Y.B., Kahn T.G., and Pirrotta Y. 2005. Characteristic low density and shear sensitivity of cross-linked chromatin containing polycomb complexes. Mol. Cell BioI. 25: 432-439. Sekinger E.A., Moqtaderi Z., and Struhl K. 2005. Intrinsic histone-DNA interactions and low nucleosome density are important for preferential accessibility of promoter regions in yeast. Mol. Cell 18: 735-748. Seligson D.B., Horvath S., Shi T., Yu H., Tze S., Grunstein M., and Kurdistani S.K. 2005. Global histone modification patterns predict risk of prostate cancer recurrence. Nature 435: 1262-1266. Shen X., Mizuguchi G., Hamiche A., and Wu C. 2000. A chromatin remodeling complex involved in transcription and DNA processing. Nature 406: 541-544. Shi Y., Lan E, Matson c., Mulligan P., Whetstine T.R., Cole P.A., Casero R.A., and Shi Y. 2004. Histone demethylation mediated by the nuclear amine oxidase hom*olog LSD 1. Cell 119: 941-953. Shorter J. and Lindquist S. 2005. Prions as adaptive conduits of memory and inheritance. Nat. Rev. Genet. 6: 435-450. Sims R.T., Ill, Belotserkovskaya R., and Reinberg D. 2004. Elongation by RNA polymerase II: The short and long of it. Genes Dev. 18: 2437-2468. Smith c.L. and Peterson c.L. 2005. ATP-dependent chromatin remodeling. Curro Top. Dev. BioI. 65: 115-148. Squazzo S.L., O'Geen H., Komashko Y.M., Krig S.R., Tin V.x., Jang S.w., Margueron R., Reinberg D., Green R., and Farnham P.T. 2006. Suz12 binds to silenced regions on the genome in a cell-type-specific manner. Genome Res. 16: 890-900. Sterner D.E. and Berger S.L. 2000. Acetylation of histones and transcription-related factors. Microbiol. Mol. BioI. Rev. 64: 435-459. Strahl B.D. and Allis C.D. 2000. The language of covalent histone modifications. Nature 403: 41-45. Strahl B.D., Ohba R., Cook R.G., and Allis C.D. 1999. Methylation of

histone H3 at lysine 4 is highly conserved and correlates with transcriptionally active nuclei in Tetrahymena. Proc. Natl. Acad. Sci. 96: 14967-14972. Su G.H., Sohn T.A., Ryu B., and Kern S.E. 2000. A novel histone deacetylase inhibitor identified by high-throughput transcriptional screening of a compound library. Cancer Res. 60: 3137-3142. Sung S. and Amasino R.M. 2004. Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 427: 159-164. Surani M.A., Barton S.c., and Norris M.L. 1984. Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis. Nature 308: 548-550. Tagami H., Ray-Gallet D., Almouzni G., and Nakatani Y. 2004. Histone H3.1 and H3.3 complexes mediate nucleosome assembly pathways dependent or independent of DNA synthesis. Cell 116: 51-61. Tamaru H. and Selker E.U. 2001. A histone H3 methyltransferase controls DNA methylation in Neurospora crassa. Nature 414: 277-283. Tanaka E.M. 2003. Regeneration: If they can do it, why can't we? Cell 113: 559-562. Thomas T.O. 1999. Histone HI: Location and role. Curro Opin. Cell BioI. 11: 312-317. Tsukada Y., Fang T., Erdjument-Bromage H., Warren M.E., Borchers C.H., Tempst E, and Zhang Y. 2006. Histone demethylation by a family of TmjC domain-containing proteins. Nature 439: 811-816. Tsukiyama T., Daniel c., Tamkun T., and Wu C. 1995. ISWI, a member of the SWI2/SNF2 ATPase family, encodes the 140 kDa subunit of the nucleosome remodeling factor. Cell 83: 1021-1026. Turner B.M. 2000. Histone acetylation and an epigenetic code. BioEssays 22: 836-845. VaJk-Lingbeek M.E., Bruggeman S.W., and van Lohuizen M. 2004. Stem cells and cancer; the polycomb connection. Cell 118: 409-418. van Attikum H. and Gasser S.M. 2005. The histone code at DNA breaks: A guide to repair? Nat. Rev. Mol. Cell BioI. 6: 757-765. van der Heijden G.w., Dieker J.w., Derijck A.A., Muller S., Berden T.H., Braat D.D., van der Vlag J., and de Boer P. 2005. Asymmetry in histone H3 variants and lysine methylation between paternal and maternal chromatin of the early mouse zygote. Mech. Dev. 122: 1008-1022. Vaquero A., Loyola A., and Reinberg D. 2003. The constantly changing face of chromatin. Sci. Aging Knowledge Environ. 2003: RE4. Varga-Weisz ED., Wilm M., Bonte E., Dumas K., Mann M., and Becker P.B. 1997. Chromatin-remodeling factor CHRAC contains the ATPases ISWI and topoisomerase II. Nature 388: 598-602. Verdel A., Tia S., Gerber S., Sugiyama T., Gygi S., Grewal S.I., and Moazed D. 2004. RNAl-mediated targeting of heterochromatin by the RITS complex. Science 303: 672-676. Vidanes G.M., Bonilla c.Y., and Toczyski D.E 2005. Complicated tails: Histone modifications and the DNA damage response. Cell 121: 973-976. Vignali M., Hassan A.H., Neely K.E., and Workman J.L.. 2000. ATPdependent chromatin-remodeling complexes. Mol. Cell. BioI. 20: 1899-1910. Vire E., Brenner c., Deplus R., Blanchon L., Fraga M., Didelot c., Morey L., Van E.A., Bernard D., Vanderwinden J.M., et al. 2005. The Polycomb group protein EZH2 directly controls DNA methylation. Nature 439: 861-874. Volpe T.A., Kidner c., HaJI I.M., Teng G., Grewal S.I., and Martienssen R.A. 2002. Regulation of heterochromatic silencing and histone H3 lysine-9 methylation by RNAl. Science 297: 1833-1837. Waddington C.H. 1957. The strategy of the genes. MacMillan, New York. Wade EA., Gegonne A., Tones P.L., BaJlestar E., Aubry E, and Wolffe A.P. 1999. Mi-2 complex couples DNA methylation to chromatin remodeling and histone deacetylation. Nat. Genet. 23: 62-66.

o Walsh c.P., Chaillet J.R., and Bestor TH. 1998. Transcription of lAP endogenous retroviruses is constrained by cytosine methylation. Nat. Genet. 20: 116-117. Watanabe Y., Yokobayashi S., Yamamoto M., and Nurse P. 2001. Pre-meiotic S phase is linked to reductional chromosome segregation and recombination. Nature 409: 359-363. Watson J.D. 2003. Celebrating the genetic jubilee: A conversation with lames D. Watson. Interviewed by John Rennie. Sci. Am. 288: 66-69. Wei Y., Yu 1., Bowen J., Gorovsky M.A., and Allis CD. 1999. Phosphorylation of histone H3 is required for proper chromosome condensation and segregation. Cel/97: 99-109. Whetstine l.R., Nottke A., Lan E, Huarte M., Smolikov S., Chen Z., Spooner E., Li E., Zhang G., Colaiacovo M., and Shi Y. 2006. Reversal of histone lysine trimethylation by the JMJD2 family of histone demethylases. Cel/125: 467-481. Wolffe A.P. and Matzke M.A. 1999. Epigenetics: Regulation through repression. Science 286: 481-486.

V E R V lEW

AND

CON C E P T S

•

Wysocka J., Swigut T, Milne TA., Dou Y., Zhang X., Burlingame A.L., Roeder R.G., Brivanlou A.H., and Allis CD. 2005. WDR5 associates with histone H3 methylated at K4 and is essential for H3 K4 methylation and vertebrate development. Cel/121: 859-872. Yan Q., Huang J., Fan T., Zhu H., and Muegge K. 2003. Lsh, a modulator of CpG methylation, is crucial for normal histone methylation. EMBO f. 22: 5154-5162. Yu B., Yang Z., Li J., Minakhina S., Yang M., Padgett R.W, Steward R., and Chen X. 2005. Methylation as a crucial step in plant microRNA biogenesis. Science 307: 932-935. Zhang Y. and Reinberg D. 2001. Transcription regulation by histone methylation: Interplay between different covalent modifications of the core histone tails. Genes Dev. 15: 2343-2360. Zhang Y., LeRoy G., Seelig H.P., Lane WS., and Reinberg D. 1998. The dermatomyositis-specific autoantigen Mi2 is a component of a complex containing histone deacetylase and nucleosome remodeling activities. Cel/95: 279-289.

Epigenetics in Saccharomyces •

•

cerevlslae Michael Grunstein' and Susan M. Gasser2 I University of California, Los Angeles, California 90095-1570 2Priedrich Miescher Institute for Biomedical Research, 4058 Basel, Switzerland

CONTENTS 1. The Genetic and Molecular Tools of Yeast, 65 2. The Life Cycle of Yeast, 66 3. Yeast Heterochromatin Is Present at the Silent HM Mating Loci and at Telomeres, 67 4. Heterochromatin Is Distinguished by a Repressive Structure That Spreads through the Entire Silent Domain, 69

5. Distinct Steps in Heterochromatin Assembly, 70 5.1

HM Heterochromatin, 70

5.2

Telomeric Heterochromatin, 71

6. Histone Deacetylation by Sir2 Provides Binding Sites for the Spread of SIR Complexes, 71

8. Histone Acetylation in Euchromatin Restricts SIR Complex Spreading, 73

9. Telomere Looping, 73 10. Discontinuity of Repression at Natural Subtelomeric Elements by Telomere Looping, 74 11. Trans-interaction of Telomeres, and Perinuclear Attachment of Heterochromatin, 74 12. Inheritance of Epigenetic States, 75 13. Aging and Sir2: Linked by rONA Repeat Instability, 76 14. Summary, 77 References, 78

7. Sir2 Deacetylates Histone H4 at Lysine 16, 72

GENERAL SUMMARY The fraction of chromatin in a eukaryotic nucleus that bears its active genes is termed euchromatin. This chromatin condenses in mitosis to allow chromosomal segregation and decondenses in interphase of the cell cycle to allow transcription to occur. However, some chromosomal domains were observed by cytological criteria to remain condensed in interphase, and this constitutively compacted chromatin was called heterochromatin. With the development of new techniques, molecular rather than cytological features have been used to define this portion of the genome, and the constitutively compacted chromatin found at centromeres and telomeres was shown to contain many thousands of simple repeat sequences. Such heterochromatin tends to replicate late in S phase of the cell cycle and is found clustered at the nuclear periphery or near the nucleolus. Importantly, its characteristic nuclease-resistant chromatin structure can spread and repress nearby genes in a stochastic manner. In the case of the fly locus white, a gene that determines red eye color, epigenetic repression yields a red and white sectored eye due to a phenomenon called position-effect variegation (PEV). Mechanistically, PEV reflects the recognition of methylated histone H3K9 by heterochromatin protein 1 (HP1) and the spreading of this mark along the chromosomal arm. In Saccharomyces cerevisiae, also known as budding yeast, a distinct mechanism of heterochromatin formation has evolved, yet it achieves a very similar result. S. cerevisiae is a microorganism commonly used in making beer and baking bread. However, unlike bacteria, it is a eukaryote. The chromosomes of budding yeast, like those of more complex eukaryotes, are complexed with

histones, enclosed in a nucleus, and replicated from multiple origins during S phase of the cell cycle. Still, the yeast genome is tiny, with only 14 megabase pairs of genomic DNA divided among 16 chromosomes, some not much larger thi;ln certain bacteriophage genomes. There are approximately 6000 genes in the yeast genome, closely packed along chromosomal arms with generally less than 2 kb spacing between them. The vast majority of yeast genes are in an open chromatin state, meaning that they are either actively transcribed or can be very rapidly induced. This, coupled with a very limited amount of simple repeat DNA, makes the detection of heterochromatin by cytological techniques virtually impossible in budding yeast. Nonetheless, using molecular tools, it has been determined that yeast has distinct heterochromatin-like regions adjacent to the telomeres on all 16 chromosomes and at two silent mating loci on chromosome III. Transcriptional repression of these latter two loci is essential for maintaining a mating-competent haploid state. Both the subtelomeric regions and the silent mating-type loci repress integrated reporter genes in a position-dependent, epigenetic manner; they replicate late in S phase and are present at the nuclear periphery. Thus, these loci bear many of the characteristic features of heterochromatin, other than the cytologically visible condensation in interphase. Indeed, for the scientist studying heterochromatin, yeast combines the advantages of a small genome and the genetic and biochemical tools available in microorganisms with important aspects of higher eukaryotic chromosomes.

E PIG ENE TIC SIN

1 The Genetic and Molecular Tools of Yeast

Yeast provides a flexible and rapid genetic system for studying cellular events. With an approximate generation time of 90 minutes, colonies containing millions of cells are produced after just 2 days of growth. In addition, yeast can propagate in both haploid and diploid formsgreatly facilitating genetic analyses. Like bacteria, haploid yeast cells can be mutated to produce specific nutritional requirements or auxotrophic genetic phenotypes, and recessive lethal mutations can be maintained either in haploids bearing conditional lethal alleles (e.g., temperature-sensitive mutants) or in heterozygous diploids (bearing both wild-type and mutant alleles). The highly efficient system of hom*ologous recombination in yeast allows the alteration of any chosen chromosomal sequence at will. In addition, portions of chromosomes can be manipulated by recombinant means on plasmids that can be stably maintained in dividing yeast cells by including short sequences that provide centromere and origin of DNA replication function. Even linear plasmids, or minichromosomes, which carry telomeric repeats to cap their ends, propagate stably in yeast. PEV using the fly white gene as a reporter has been important in defining epigenetic gene regulation and the genes that affect this unique form of gene repression (see Chapter 5 for more detail). The discovery and characterization of a similar phenomenon near yeast telomeres, called telomere position effect (TPE), has been analogously aided by the use of Ura3 and Ade2 reporter genes (Fig. 1). In the presence of 5-fluoroorotic acid (5FOA), the Ura3 protein converts 5-FOA to 5-fluorouracil (5-FU), an inhibitor of DNA synthesis that causes cell death. However, when Ura3 is integrated into regions of heterochromatin, the Ura3 gene is repressed in some, but not all, cells, and only the cells that silence Ura3 are able to grow in the presence of 5-FOA. Thus, by scoring the efficiency of growth on 5-FOA with a serial dilution drop assay (Fig. 1a), one can quantify the repression of this reporter gene over a very large range (e.g., 1O-106 -fold). Moreover, mutations that disrupt TPE can be readily identified by monitoring for increased sensitivity to 5-FOA. Similarly, when the Ade2 gene is targeted for integration into a region of heterochromatin, the gene is repressed and a precursor in adenine biosynthesis accumulates in the cell, turning it a reddish color. Importantly, the epigenetic nature of Ade2 repression is visible within a single colony of genetically identical cells: The gene can be "on" in some cells and "off" in others, pro-

SAC C H A ROM Y C ESC ERE V I S I A E

TPE of URA3 expression in S.cerevisiae Telomere

URA3 Chr VIIL :: URA3-Tel No. of cells: sir2

10 6

10 5 10 4 10 3 10 2 10

.... ....

YPD

. . . . . .liliiii

u .....

yku70 . . . . . sir2 wt

yku70

'~ff

• • • 1$ , -

+ 5-FOA

TPE of AOE2 expression in S.cerevisiae Telomere

ADE2

.,.\YNINi ade2red and white sectors

ADE2

ade2-; ADE2-TeIVR

variegated repression

Figure 1. Silencing and TPE in Yeast (a) The Ura3 gene, inserted near the telomeric simple TG-rich repeat at the left arm of chromosome VII, is silenced by telomeric heterochromatin in this yeast strain. In normal rich medium (YPD), no growth difference can be detected between wild-type (wt) cells that repress the subtelomeric Ura3 gene and silencing mutants that lose telomeric heterochromatin and express Ura3. In media containing 5-FOA (lower panel), on the other hand, cells that repress Ura3 (e.g., wt cells) can grow, whereas cells that express it (sir2 and ykulO mutants) cannot. This is because the Ura3 gene product converts 5-FOA to the toxic intermediate 5-fluorouracil. The serial dilution/drop assay allows detection of silencing in as few as 1 in 10· cells. (b) Cells containing the wt Ade2 gene produce a colony that is "white," whereas those containing mutant ade2 appear red, due to the accumulation of a reddish intermediate in adenine biosynthesis. When the Ade2 gene is inserted near the telomere at the right arm of chromosome V, it is silenced in an epigenetic manner. The silent Ade2 state and the active Ade2 state in genetically identical cells are both inherited, creating red and white sectors in a colony (much like PEV).

C HAP T E R

ducing red sectors in a white colony background or vice versa (Fig. 1b). Unlike the Ura3 assay, there is no selection against cells that fail to repress Ade2, and therefore, the phenotype of the Ade2 reporter inserted in subtelomeric heterochromatin demonstrates the switching rate as well as the heritability of the epigenetic state. The Ade2 color assay provides a striking illustration of the semi-stable nature of both repressed and derepressed states. Combined with these genetic approaches, biochemical techniques are readily applied to protease-deficient strains grown either synchronously or asynchronously in large cultures. Recently, the battery of tools available has broadened to include sophisticated microarray and protein network techniques that easily accommodate the small genome of yeast. These methods have enabled genome-wide analyses of transcription, transcription factor binding, histone modifications, and protein-protein interactions. This broad range of sophisticated tools has allowed scientists to explore the mechanisms that regulate both the establishment of heterochromatin and its physiological roles in budding yeast. However, before describing these discoveries further, it is necessary to review the life cycle of yeast in more detail.

C0\

((§) Mitotis (haploid)

Conjugation

Mitotis (diploid)

@> ~

f-@ @4

CQt.LJUGATION

@>@ C9Q)

ha~c:@

~@:l

@?~l

~:T~ 0+

Meiosis

SPORULATION

GERMINATION

0G¥>

GERMINATION

2 The Life Cycle of Yeast

S. cerevisiae multiplies through mitotic division in either

a haploid or a diploid state, by producing a bud that enlarges and eventually separates from the mother cell (Fig. 2a). Haploid yeast cells can mate with each other (i.e., conjugate), since they exist in one of two mating types, termed a or a, reminiscent of the two sexes in mammals. Yeast cells of each mating type produce a distinct pheromone that attracts the cells of the opposite mating type: a cells produce a peptide of 12 amino acids called a factor, which binds to a membrane-spanning afactor receptor on the surface of an a cell. Conversely, a cells produce a 13 aa peptide that binds to the a-factor receptor on the surface of a cells. These interactions result in the arrest of the cells in mid-to-Iate G I phase of the cell cycle. The arrested cells assume "shmoo"-like shapes (named after the pear-shaped Al Capp cartoon character; Fig. 2b), and the shmoos of opposite mating type fuse at their tips, producing an a/a diploid. The mating response is repressed in diploid cells, which propagate vegetatively (i.e., by mitotic division) just like haploid cells. On the other hand, exposure to starvation conditions will induce a meiotic program that results in the formation of an ascus containing four spores, two of

Figure 2. The life Cycle of Budding Yeast (0) Yeast cells divide mitotically in both haploid and diploid forms. Sporulation is induced in a diploid by starvation, whereas mating occurs spontaneously when haploids of opposite mating type are in the vicinity of each other. This occurs by pheromone secretion, which arrests the cell cycle in G, of a cell of the opposite mating type, and after sufficient exposure to pheromone, the mating pathway is induced. The diploid state represses the mating pathway. (b) In response to pheromone, haploid cells distort toward cells of the opposite mating type. These are called shmoos. The nuclear envelope is visible as green fluorescence.

each mating type. Given sufficient nutrients, the haploid spores grow into cells that are again capable of mating, starting the life cycle over again. Although haploid yeast cells in the laboratory are usually designated as one mating type or the other, in the wild, yeast switch their mating type nearly each cell cycle (Fig. 3a). Mating-type switching is provoked by an endonuclease activity (HO) that induces a site-specific double-strand break at the MAT locus. A gene conversion event then transposes the opposite mating-type

E PIG ENE TIC SIN

5 Ace H A ROM Y C ESC ERE V I 5 I A E

Yeast life cycle

!! haploid

Mating Type Switching

p 0 0 ~ /t

!! haploid

l'-..

0000 ~

a factor

afacto~

@>@ CW)

@ b

Figure 3. Mating Type Switching in Yeast

Conjugation

ala diploid

Chromosome III

HO endonuclease

HMLa

MATa

HMRa

85%~~%

€}-

.....-e;7!-II-------~.t----{c::::J}-------- -----~~

MATa

HMLa

HMRa

Transcriptionally silent domains and silencer elements HMLa

MATa

HMRa

.,...M~ •

MATn

2.5kb

is 46

R = Rap1 binding site

a1 al

112 kbl

IRIRIRI

= Abf1

binding site

63~'

= ORC consensus

IRIRIRI

silenced chromatin region

information from a constitutively silent donor locus, HMLa or HMRa, to the active MAT locus. Such strains are called hom*othallic. This means that a vegetatively growing MATa cell will rapidly produce MATa progeny,

and vice versa. Because in the laboratory it is desirable to have cells with stable mating types, laboratory strains are usually constructed to contain a mutant HO endonuclease gene, which eliminates cleavage at the MAT locus. The loss of HO endonuclease activity prevents mating-type switching, producing a heterothallic strain. These strains contain silent HM loci and an active MAT locus whose mating type information is stably either a or a. Two silent mating loci (Fig. 3b), one for each "sex," are maintained constitutively silent in an epigenetic manner and have become a classic system for the study of heterochromatin.

(0) hom*othallic yeast strains are able to switch mating type after one division cycle. The switch occurs before DNA replication so that both mother and daughter cells assume the new mating type. (b) The position of the silent and expressed mating-type loci on chromosome III are shown here. The active MAT locus is able to switch through gene conversion roughly once per cell cycle, due to a double-strand break induced by the HO endonuclease. The percentages indicated show the frequency with which the gene conversion event replaced the MAT locus with the opposite mating-type information. The directionality of switching is guaranteed by the recombination enhancer (RE) on the left arm of chromosome III. (c) Repression at the silent mating-type loci HMR and HML is mediated by two silencer DNA elements that flank the silent genes. These silencers are termed E (for essential) or I (for important) (Brand et al. 1997) and provide binding sites for Rap1 (R), Abf1 (A), and ORC (0). Artificial silencers can be created using various combinations of the redundant binding sites, although their efficiency is less than that of the native silencers. HMLa and HMRa are 12 kb and 23 kb, respectively, from the telomeres of chromosome III. Telomeric heterochromatin domains at chromosome III are silenced independently from the HM loci in a process that is initiated at the telomeres through multiple binding sites for Rap1 (R).

3 Yeast Heterochromatin Is Present at the Silent HM Mating Loci and at Telomeres

The three mating-type loci, HMLa, MAT, and HMRa are located on chromosome III and contain the information that determines a or a mating type in yeast. HMLa (~11 kb from the left telomere) and HMRa (~23 kb from the right telomere; Fig. 3b,c) are situated between short DNA elements called E and I silencers. Only when either of the silent cassettes is copied and integrated into the active MAT locus is it capable of transcription in a normal cell. The transfer of HMLa information into MAT results in an a mating type (MATa) cell, whereas the transfer of HMRa information into MAT results in the a mating type (MATa)(Fig. 3b). This shows that the genes and promoters at the HM loci are completely intact, although they remain

68 • C HAP

T ER 4

stably repressed when they are positioned at HMR and HML. This is essential for the maintenance of mating potential, because the combined expression of a and ex transcripts in the same cell results in a non-mating sterile state. The scoring of sterility as a phenotype proved very useful for identifying mutations that impair silencing at the HM loci. In this manner, the silent information regulatory proteins, SIR], SIR2, SIR3, and SIR4, were identified as being essential for the full repression of silent HM loci (for review, see Rusche et al. 2003). Mutations in sir2, sirJ, or sir4 caused a complete loss of silencing, whereas in sir] mutants, only a fraction of MATa cells were unable to mate due to a loss of HM repression. Taking advantage of the partial phenotype of sirl-deficient cells, it could be shown that the two alternative states (mating and non-mating) are heritable through successive cell divisions in genetically identical cells (Pillus and Rine 1989). This provided a clear demonstration that mating-type repression displays the hallmark characteristic of epigenetically controlled repression. In addition, it was shown from other studies that the amino termini of histones H3 and H4, repressor activator protein 1 (Rapl), and the origin recognition complex (ORC) are also involved as structural components ofheterochromatin (for review, see Rusche et al. 2003). Heterochromatin is also present immediately adjacent to the yeast telomeric repeat DNA (C\_3A/TG\). As men-

tioned above, when reporter genes such as Ura3 or Ade2 were integrated adjacent to these telomeric repeats, they were repressed in a variegated and epigenetic manner (Gottschling et al. 1990). This TPE shared the HM requirement for Rap1, Sir2, Sir3, Sir4, and the histone amino termini (Kayne et al. 1988; Aparicio et al. 1991). Genetics argued strongly that with the exception of Sirl, similar mechanisms silence genes at the HM mating loci and at telomere-adjacent sites. Moreover, given that the subtelomeric reporters could switch at detectable rates between silent and expressed states, the gene repression appeared to be very similar to fly PEY. In yeast, the four Sir proteins that mediate repression share no extensive hom*ology among themselves, and the Sid, Sid, and Sir4 proteins appear to be conserved only in S. cerevisiae and closely related budding yeasts. Sir2, on the other hand, is the founding member of a large family of NAD-dependent histone deacetylases, which is conserved from bacteria to man (Fig. 4). A role for Sir2-like histone deacetylases in transcriptional repression is observed even in organisms such as fission yeast and flies, which lack the other Sir proteins. The Schizosaccharomyces pombe Sir2 activity is required for transcriptional silencing near telomeres, and Drosophila Sir2 affects the stability of PEV (for review, see Chopra and Mishra 2005). The coupling of NAD hydrolysis with deacetyla-

IV SirT6

Figure 4. Sir2 Family of Deacetylases

P.hor

Hst1

P.aby

Sir2

nuclear

.....

_--_ ...

Sir2 is the founding member of a large family of NAD-dependent deacetylases. The Sir2 family of proteins is unusually conserved and is found in organisms that range from bacteria to humans, and contains both nuclear and cytoplasmic branches of the evolutionary tree. This phylogenetic unrooted tree of Sir2 hom*ologs was generated using CLUSTAL ~ and TREEVIE~ programs to compare the core domain sequences of hom*ologs identified in eDNA and unique libraries. The six subclasses and unlinked group (U) are described in Frye (2000). The mammalian hom*ologs are labeled SirTl-7 and are in bold, and the budding yeast proteins are underlined. Other species are indicated by the species name. (Modified, with permission, from Frye 2000 [© Elsevier].)

E P f G ENE T f C SIN

tion by Sir2 produces O-acetyl ADP ribose, an intermediate that may have a function of its own (Tanner et al. 2000; also see Section 13). It is important to note that the Sir2 family of enzymes modifies many substrates other than histones, with a large branch of the Sir2 family actually being cytoplasmic enzymes (Fig. 4). The diversity of Sir2 functions is illustrated by the fact that mammalian Sir2 deacetylates the transcription factors FOXO and p53 in response to stress and DNA damage, altering their interaction. In budding yeast, Sir2 has an important role in addition to gene silencing, which is to suppress nonreciprocal recombination in the highly repetitive genes of the rDNA locus that is found within the nucleolus (Gottlieb and Esposito 1989).

5 A C C H A ROM Y C ESC ERE V I 5 I A E

•

matin-immunopreClpltation techniques, which showed that Sir2, Sir3, and Sir4 proteins interact physically with chromatin throughout the subtelomeric domain of silent chromatin (Hecht et al. 1996; Strahl-Bolsinger et al. 1997). Evidence that this induces a repressive, less accessible chromatin structure comes from other approaches. For instance, it was shown that the DNA of silenced chromatin was not methylated efficiently in yeast cells that express a bacterial dam methylase, although the enzyme readily methylated sequences outside the silent region. This suggested that heterochromatin can restrict access to macromolecules like dam methyltransferase (Gottschling 1992). Similarly, the approximately 3-kb HMR locus in isolated nuclei is preferentially resistant to certain restriction endonucleases (Loo and Rine 1994), and nucleosomes were shown to be tightly positioned between two silencer elements, creating nuclease-resistant domains at silent, but not active, HM loci (Weiss and Simpson 1998). Thus, yeast heterochromatin clearly assumes a distinct chromatin structure. The extent to which either yeast or metazoan heterochromatin is hyper-condensed, and condensation stericallY hinders access to transcription factors, is less certain. Surprisingly, the repressive complex formed by the interaction of Sir proteins and histones appears to be dynamic, because Sir proteins can be incorporated into HM silent chromatin even when cells are arrested at a stage in the cell cycle when heterochromatin assembly generally does not occur (Cheng and Gartenberg 2000).

4 Heterochromatin Is Distinguished by a Repressive Structure That Spreads through the Entire Silent Domain Repression of gene activity in euchromatin can occur due to the presence of a repressive protein or complex that recognizes a specific sequence in the promoter of a gene, thus preventing movement or engagement of the transcription machinery. Heterochromatic repression occurs through a different mechanism that is not promoter-specific: Repression initiates at specific sites, yet spreads continuously throughout the domain, silencing any and all promoters in the region (Fig. 5) (Renauld et al. 1993). This was most clearly demonstrated by the use of chro-

Telomeric heterochromatin TG'-3 repeats

Rap1

Ku70/Ku80

Rap1

RNA Pol III TBF B,E dam methylase HO endonuclease restriction enzymes

SIR complex Sir3

Rap1 ~

_ _-,I '-

Silencer

_ _~

~ / ,~

Repressed domain

HM heterochromatin

Silencer

Sir2 M W Sir4

~ acetyl group

nucleosome

Figure 5. Model for Yeast Heterochromatin at Telomeres and the HM loci The telomere and HM silencer mechanisms for nucleating SIR complex spreading both use Rapl, Sir2, Sir3, and Sir4, yet they differ in that telomeres also rely on yKu whereas the HM silencer elements use the factors aRC, Abfl, and Sirl. Telomeric heterochromatin is thought to fold back onto itself to form a cap that protects the telomere from degradation and whose condensation and folding silences genes. In the case of HM heterochromatin, the repressed domain between the silencer elements consists of closely spaced nucleosomes that form a condensed structure. Both the telomeric and HM silent regions are inaccessible to the transcription machinery and degradative enzymes.

70 • C HAP

T ER 4

This may explain why Sir-bound heterochromatin can serve as a binding site for certain transcription factors even in its repressed state (Sekinger and Gross 1999). Although such studies argue that heterochromatin does not act by sterically hindering all non-histone-protein access, no obvious transcription occurs, and engaged RNA polymerases cannot be detected experimentally. Experiments by Chen and Widom (2005) argue that the step which is specifically prevented by heterochromatin is the recruitment of complexes containing RNA polymerase II and the promoter-binding transcription factors TFIIB and TFIIE. Thus, although the silent yeast chromatin is dynamic, allowing SIR factors and possibly some transcription factors to exchange, it selectively impedes the binding of the basal transcription machinery, and thereby blocks the production of mRNA (for more detail, see Chapter 10).

5 Distinct Steps in Heterochromatin Assembly

The assembly of heterochromatin in all species involves a series of molecular steps, several of which have been identified in budding yeast. One of the best characterized is the site-specific nucleation of heterochromatin, an event that requires sequence-specific DNA-binding factors. Next, heterochromatin spreads from the initiation site. Its spreading is limited by specific boundary mechanisms that are well characterized in yeast. Finally, yeast has been useful for demonstrating a role for subnuclear compartments in heterochromatin-mediated repression. The assembly of heterochromatin at telomeres varies in some ways from its assembly at HM loci, but both reflect a very similar principle: the presence of specific DNA-binding factors that nucleate the spread of general repressors. We describe these mechanisms in detail below (Fig. 6). 5.1 HM Heterochromatin

The silent mating loci HML and HMR are bracketed by short DNA elements termed silencers, labeled E (for essential) and I (for important; Fig. 3b,c). Silencer elements provide binding sites for at least two multifunctional nuclear factors, namely Rap1 and Abfl, as well as the origin recognition complex (ORC) (Brand et al. 1987). Although the deletion of HMR-E, which has all three recognition sites, has a much stronger effect on silencing than deletion of HMR - I, which has two, each silencer at HMR and at HML can serve as a specific nucleation site for Sir silencing complex recruitment, to promote subsequent Sir protein spreading along the

intervening nucleosomes. Contact in vitro between Rap1 molecules at two separate binding sites argued that direct interaction may occur between the factors bound at E and I silencers through looping of the repressed domain, which would explain the cooperative effects by E and I on the initiation of repression (Hofmann et al. 1989). Redundancy of silencer element function is a hallmark of repression by heterochromatin and is also true within a silencer element. Rap1 and Abfl, which are general transcription factors, and ORC, which seeds the prereplication complex at origins of replication, function in a redundant manner. This was demonstrated by deletion experiments: DNA-binding sites for any two of these factors are sufficient to allow silencing (Brand et al. 1987), despite the fact that Rap1, Abfl, and ORC share little structural similarity. The explanation for their redundancy comes from an analysis of the proteins they recruit. For example, Rap1 recruits Sir4 both at HM silencers and at telomeres, whereas Abfl interacts with Sir3, and ORC has high affinity for Sirl, a SIR factor specific for HM repression (for review, see Rusche et al. 2003). Sirl itself interacts directly with the amino terminus of Sir4, providing the bridge between ORC and the SIR2-3-4 complex. Thus, the various silencer binding factors all lead to the recruitment of Sir4 and, in turn, the SIR2-3-4 complex, which is required in all cases for repression. The apparent redundancy among Rap 1, Abfl, and ORC (at silencers), as well as the Ku heterodimer (at telomeres, see below), can be attributed to the ability of each to nucleate repression by direct contact with different components of the SIR complex. It should be noted that Sirl is involved primarily in the establishment rather than the maintenance of heterochromatic repression. Once Sirl helps establish silencing, it is no longer needed for the stable maintenance of the repressed state (Pillus and Rine 1989). The important role played by Sirl in establishment was shown by tethering the protein artificially through a Gal4 DNA-binding domain to Gal4-binding sites, which replaced the HMRE silencer. In this context, GBD-Sirl can efficiently nucleate repression, rendering the silencer and its binding factors unnecessary (Chien et aI. 1993). Nonetheless, the Sirl-targeted repression still required all the other Sir proteins and intact histone tails. This argues that one of the primary roles of the silencer-binding factors is to attract Sirl, which in turn nucleates repression by recruiting the other Sir proteins to interact with adjacent nucleosomes. In support of this is the fact that, unlike the other Sir proteins, Sirl does not spread with the SIR complex beyond the silencers (Fig. 5) (Rusche et al. 2002).

E PIG ENE TIC 5

STEP 1)

Recruitment 01 Sir4, then Sir2 and Sir3 to telomere-bound Rapl

Rapl binding sites

subtelomeric regions

STEP 2) Sir2-mediated deacetylation 01 histone H4K16

STEP 3)

Spreading 01 the SIR complex along nucleosomes

STEP 4)

Folding 01 a silent telomere into a higher-order structure

Figure 6. Steps in the Assembly of Telomeric Heterochromatin

(Step 7) At telomeres, Rap1 and yKu recruit Sir4 even in the absence 01 Sir2 or Sir3. Only Sir4 can be recruited, in the absence 01 the other Sir proteins, and its binding is antagonized by Rill and Ril2 (Mishra and Shore 1999). (Step 2) Sir4-Sir2 and Sir4-Sir3 interact strongly, creating Sir complexes along the TG repeats. Sir2 NAD-dependent histone deacetylase activity is stimulated by complex formation, and Sir2 deacetylates the acetylated histone H4 K16 residue in nearby nucleosomes. (Step 3) SIR complexes spread along the nucleosomes, perhaps making use of the a-acetyl ADP ribose intermediate produced by NAD hydrolysis (Liou et al. 2005). Sir3 and Sir4 bind the deacetylated histone H4 tails. Although the deacetylated histone H3 amino-terminal tail also binds Sir3 and Sir4 proteins, it is not shown here. (Step 4) The silent chromatin "matures" at the end of M phase to create an inaccessible structure. This may entail higher-order folding and sequestering at the nuclear envelope.

5.2 Telomeric Heterochromatin

At telomeres, an RNA-based enzyme called telomerase maintains a simple but irregular TG-rich repeat of 300-350 bp in length, which provides 16-20 consensus sites for Rap 1. The array of Rap I-binding sites forms a non-nucleosomal cap on the chromosomal end and plays a critical role in telomere length maintenance (Marcand et al. 1997). Along the telomeric repeat, Rapl

INS Ace H A ROM Y C ESC ERE V I 5 I A E

binds its consensus through a core DNA-binding domain and binds Sir4 through its carboxy-terminal domain, even in the absence of the other SIR proteins or the H4 amino terminus. Since the disruption of Sir4 prevents other proteins from binding to telomeric chromatin (Luo et al. 2002), Sir4 appears to be a crucial link between nucleating events and the ensuing silent chromatin structure (Fig. 6). The DNA end-binding complex yKu70/yKu80 helps Rapl recruit Sir4 to chromosome ends. Indeed, loss of yKu strongly derepresses telomeric silencing, and a targeted GBD-yKu fusion can efficiently nucleate repression at silencer-compromised reporter genes. The requirement for yKu can be bypassed by elimination of the Rapl-binding factor Rifl, which competes for the interaction of Sir4 with the Rapl carboxy-terminal domain (Fig. 6) (Mishra and Shore 1999). The cooperative effects of yKu and Rapl in the nucleation of heterochromatin are demonstrated by the observation that 600 bp of telomeric repeat sequence, which provides more than 30 Rap I-binding sites, is not sufficient to nucleate repression at internal chromosomal loci, although insertion of 900 bp or about 45 Rapl-binding sites can (Stavenhagen and Zakian 1994). It should be noted that at promoters dispersed throughout the yeast genome, Rapl serves as a general transcription factor contributing to the activation of many genes, particularly those encoding ribosomal proteins. Why Rap 1 recruits activators to these promoters rather than nucleating heterochromatin, is presently unknown. 6 Histone Deacetylation by Sir2 Provides Binding Sites for the Spread of SIR Complexes

The molecular interactions of the SIR proteins have been well-characterized, with Sir4 playing a key scaffolding role for their assembly. Sir4 interacts strongly with Sir2 in vitro. Sir4 also interacts independently with Sir3, whereas Sir3 and Sir2 appear to interact very weakly (Moazed et al. 1997; Strahl-Bolsinger et al. 1997; Hoppe et al. 2002). Sid and Sir4 also hom*odimerize (Moretti et al. 1994). Nonetheless, when coordinately expressed in insect cells, Sir2, Sid, and Sir4 are readily isolated as a stable 360-kD complex with a 1:1:1 stoichiometry of SIR proteins (Cubizolles et al. 2006). Consistent with a functional heterotrimeric complex of SIR2-3-4, it was shown by chromatin immunoprecipitation that the three SIR components spread to equal extents throughout a heterochromatic domain (Hecht et al. 1996; Strahl-Bolsinger et al. 1997). It is evident,

72 •

CHAPTER

nonetheless, that Sir3 has a special role in this process, because Sir3 overexpression can extend the silent domain, coincident with the spreading of Sir3 from its normal ~3 kb to ~ 15 kb (Renauld et al. 1993; Hecht et al. 1996). Imbalanced expression of Sir2 or Sir4 alone, or even expression of subdomains of either protein, has precisely the opposite effect: Overexpression of either Sir2 or Sir4 disrupts TPE, although the coordinated ectopic expression of Sir3 and Sir4 counteracts this imbalance and again restores silencing (Maillet et al. 1996). This underscores the importance of dosage within the SIR complex for its repressive function, which is also true for Polycomb complexes in flies. Consistent with a unique ability of overexpressed Sir3 to spread along chromatin, it was demonstrated that Sir3 can form a stable multimer in vitro (Liou et al. 2005). The platform upon which the SIR complex spreads comprises nucleosomes with deacetylated histone H3 and H4 amino termini (Braunstein et al. 1996; Suka et al. 2001). The manner in which SIR proteins interact with histones helps explain how spreading occurs (Fig. 7). Sir3 and Sir4 proteins bind deacetylated histone H3 and H4 amino termini in vitro and in vivo (Hecht et al. 1995, 1996), and neither the H2A nor H2B tail is required for this interaction. The most important histone region in this regard is contained in residues 16-29 of histone H4, of which lysine 16 in particular must be deacetylated or positively charged for Sir3 to bind (Johnson et al. 1990, 1992). Unlike mutations at other acetylation sites, mutation of H4K16 by even a conservative change disrupts telomeric silencing completely. The histone H3/H4 tails, in particular the region of H4 (residues 16-24), also promote nucleosome array compaction in vitro, in which case the acetylation state of H4K16 is likely to regulate higher-order folding of the nucleosomal fiber. How is deacetylation of H4K16 regulated in vivo?

H Ctfl

tRNA genes

Rebl VP16 \ Tbfl

Sir2 deacetylase -¢-

/H3K79me /

(

Nuclear pore tethering

Sas2 acetyl transferase

~~w;~tm~tiiWiAtiilJ'iJlilJ~ Rapl binding sites

subtelomeric regions

Sir2 is an NAD-dependent histone deacetylase whose activity is enhanced by association with Sir4. Sir2 activity links deacetylation with the conversion of NAD to 0acetyl-ADP-ribose using an ADP-ribosyl transferase activity (Tanner et al. 2000). Given that a positively charged H4 lysine 16 is critical for forming heterochromatin, it is striking that Sir2 can deacetylate in vitro and in vivo H4 lysine 16, although the enzyme also deacetylates other lysines within the H4 amino terminus and H3 lysine 9 and lysine 14 as well (Imai et al. 2000; Suka et al. 2002; Cubizolles et al. 2006). All these target sites are within domains of H3 and H4 that are required for silencing. Interestingly, O-acetyl-ADP-ribose itself promotes not only the multimerization of Sid, but also the interaction of Sid with Sir4-Sir2 in vitro (Liou et al. 2005). Together these data argue that histone deacetylation by Sir2 promotes the formation and multimerization of the Sir complex, as well as preparing deacetylated binding sites on adjacent nucleosomes for SIR protein binding. We summarize the different steps for the initiation and spreading of heterochromatin in telomeric regions and at the HM loci in Figure 6. At telomeres, Rap 1 and yKu recruit Sir4, Sir4 recruits Sir2 to deacetylate histone H4 and H3 amino-terminal tails. Sir4 also recruits Sir3. The deacetylation of the histone tails produces Sir3/Sir4binding sites and nucleates binding of the SIR2-3-4 complex. The mutual interaction of Sir3 with Sir4, and of both with histone amino termini, is thought to stabilize the Sir complex on the nucleosomal fiber, allowing it to spread along the histone tails. Finally, the folding of the chromatin fiber (discussed below) may stabilize the repressed state. Most of these events are likely to be very similar at HM loci, although the initial recruitment of Sir4 is mediated by Rapl, Abfl, ORC, and Sirl. What then, causes spreading to stop?

Htzl Bdfl

..........

7 Sir2 Deacetylates Histone H4 at Lysine 16

Figure 7. Heterochromatin Boundary Function in Budding Yeast Spreading of heterochromatin through deacetylation of histone H4 K16 by Sir2 is limited by the competing activity of Sas2 histone acetyl transferase which acetylates H4K16 in adjacent euchromatin, thus preventing Sir3 binding. Methylation of H3K79 in adjacent euchromatin also affects the spreading of heterochromatin. In addition, factors such as Rebl, Tbfl, and mammalian or viral factors Ctfl or VP16, nuclear pore tethering, and the presence of tRNA genes may also mediate boundary function. It is conceivable that several of these factors function through the recruitment of histone acetyltransferases.

E P f G ENE T f C S

8 Histone Acetylation in Euchromatin Restricts SIR Complex Spreading

Since the deacetylation of histone H4 lysine 16 by Sir2 is crucial to the formation of heterochromatin, it is not surprising that modification of this site also plays a key role in providing a barrier to heterochromatin propagation. Interestingly, of all histone acetylation sites, only H4K16 is modified in monoacetylated H4 of euchromatin (Clarke et al. 1993). One of the enzymes that contributes to H4K16 acetylation in subtelomeric regions in yeast is Sas2, a member of the highly conserved MYST class of histone acetyltransferases (HATs). If Sas2 is deleted, preventing the acetylation ofH4K16, or ifH4K16 is changed to arginine to simulate the deacetylated state, the SIR complex spreads at low levels approximately fivefold farther at the right telomere of chromosome VI than in a wild-type cell. This argues that the spreading of subtelomeric heterochromatin is controlled, at least in part, by the opposing activities of Sir2 and Sas2 on H4K16 (Fig. 7) (Kimura et al. 2002; Suka et al. 2002). At the HM loci, restricting the spread of silent chromatin is perhaps even more critical than at telomeres, since genes important for growth are found along the arm of chromosome III, and silencers were shown to function bidirectionally, promoting repression of flanking DNA sequences. One boundary that prevents further spreading of silencing toward the telomere from HMR is a tRNA gene (Donze and Kamakaka 2001). This boundary function is likely to require the HAT activity that is associated with transcription or transcriptional potential of this locus. It is significant that one of these HATs is Sas2, although the histone H3 HAT, GenS, also promotes the boundary function of tRNA genes. This suggests that transcriptional activators can generally restrict SIR complex propagation by recruiting HATs. Consistently, boundary activity has also been attributed to the transcription factors, Reb1, Tbfl, to a mammalian factor CTCF, as well as to the acidic trans-activating domain of VP16 (Fourel et al. 1999,2001). Each of these may also promote hyperacetylation of histones, thereby attenuating SIR complex propagation by impairing its association with nucleosomes (Fig. 7). Finally, coupled with the mechanism described above, it was reported that the presence in euchromatin of the variant histone H2A.Z and the RNA polymerase-associated factor Bdfl (Meneghini et al. 2003), the methylation of histone H3lysine 79 (van Leeuwen et al. 2002), and the tethering of DNA to nuclear pores (Ishii et al. 2002) all help limit the spread of silent chromatin. Although the

f N

5 Ace H A ROM Y C ESC ERE V I 5 I A E

•

mechanisms by which these factors affect heterochromatin spreading are unknown, it is interesting to note that some active genes are associated with nuclear pores (Ishii et al. 2002; Brickner and Walter 2004). Thus, the common characteristic of boundary factors in yeast may be that of a strong transcriptional activator or nucleosome remodeler that directly or indirectly disrupts histone interactions with heterochromatin proteins. 9 Telomere Looping

Several lines of evidence support the notion that longrange interactions enable chromosomal ends to loop back, bypassing subtelomeric boundary elements and stabilizing repressed chromatin at subtelomeric genes (Figs. 5,6). For instance, despite the presence of Rap1binding sites only within the first ~300 bp of TG repeat DNA on the end of a telomere, chromatin immunoprecipitation showed that Rap1 is associated with nucleosomes as far as ~3 kb away from the TG repeat (Strahl-Bolsinger et al. 1997). Similarly, yKu is recovered for ~3 kb from the chromosomal end to which it binds (Martin et al. 1999). Furthermore, when silencing is disrupted by mutation of SIR genes, both Rap 1 and yKu are lost exclusively from the more internal sequences and not from the terminal TG repeats (Martin et al. 1999). To account for the recovery of the 3 kb of heterochromatin with Rap1 and/or yKu after shearing the chromatin into fragments of <500 bp, it was proposed that the truncated telomere folds back, enabling TG-bound Rapl and yKu to bind SIR proteins across the chromosome in trans (Figs. 5,6). This structure might contribute to the "capping" function of telomere-bound proteins. Supporting evidence for telomere looping comes from the work of de Bruin et al. (2001), who have exploited the inability of transcriptional activators such as Ga14 to function from a site downstream of the gene whose promoter they are meant to activate. Strains were constructed in which the Ga14 upstream activating sequence (UAS) element was placed downstream of the reporter, and the construct was inserted either at an internal chromosomal location or near a telomere. At an internal site, this construct could not support galactoseinducible transcription. However, in a subtelomeric context, the Ga14 UAS could activate the promoter from a site 1.9 kb downstream of the promoter, in a Sir3dependent manner. It was argued that the telomeric end can fold back in the presence, but not in the absence, of Sid to allow the Ga14 UAS to position itself proximal to the transcription start site (de Bruin et al. 2001).

74 • C HAP

TER 4

10 Discontinuity of Repression at Natural Subtelomeric Elements by Telomere looping

We have set forth here a simplistic view of continuous silent chromatin emanating from the telomeric Rap 1binding sites, yet the situation at native telomeres is significantly more complex, largely due to the presence of natural boundary elements found in subtelomeric repeat sequences. Generally, when reporter constructs for telomeric repression are integrated, the subtelomeric repeat elements called X and y' at telomeres are deleted, placing the reporter gene and unique sequence immediately adjacent to TG repeats. All native telomeres, on the other hand, contain a core subtelomeric repeat element, X, which is positioned between the TG repeat and the most telomere-proximal gene, and 50-70% of native telomeres also contain at least one copy of a larger subtelomeric element called Y' (Fig. 8). Both X and Y' elements contain binding sites for the transcriptional regulators Tbfl and Reb1, and these have been shown to reduce the spread of silent chromatin (Fourel et al. 1999). However, X elements also contain the consensus for ORC and binding sites for Abfl, which have the opposite effect: These reinitiate or boost the repression of reporters placed on the centromere-proximal side of these elements. The result is one of discontinuity in silencing at native telomeres, which differs from the

r~--_A,-

TG repeats STAR

••• I

Y' ORF1 Y' ORF2

model of continuous spreading outlined in Figure 6. To explain this, Pryde and Louis (1999) have also proposed that telomeres loop back to allow a region of unrepressed chromatin to intervene between two repressed domains, leading to discontinuity in silent domains without eliminating the need for nucleation and spreading from the TG repeats. 11 Trans-interaction of Telomeres, and Perinuclear Attachment of Heterochromatin

One of the most universally conserved aspects of heterochromatin is that it occurs in discrete nuclear subcompartments. This is also true in budding yeast, where telomeres cluster into groups during interphase, remaining closely associated with the nuclear periphery. This clustering was initially observed as prominent foci of Rap 1 and SIR proteins that were detected above a diffuse nuclear background of these factors by immunostaining (Fig. 9). Disruption of silencing by histone H4 K16 mutation, or interference in Rap1 or yKu function, led to the dispersion of the SIR proteins from these clusters (Hecht et al. 1995; Laroche et al. 1998). Later it was shown that not only telomeres, but also the HML and HMR loci, are closely associated with the nuclear envelope. This association is mediated by redundant pathways that depend either on the telomere-bound yKu factor, or on the forma-

r,...--_A,-_ _~,\

STR

core X

1~.4Blclol

XY'-ends Silencing profile

~-----.;------Figure 8. The Organization of Native Telomeres and Their Silencing Patterns STR

coreX

• •4bl ol B

X-ends

~----------

URA3 reporter

artificial telomeres

••• I

~--------

Subtelomeric elements are shown with their major protein-binding sites. Telomeres fall into the two general classes: X-containing or X+Y'-containing ends. The STAR and STR elements block the propagation of repression and leave a region of reduced repression within the Y' or X element. This is not the case at artificially truncated telomeres where there is a gradient of repression that extends 3-4 kb from the TG repeat. Looping similar to that in Fig. 6 is proposed for native telomeres so that repressed regions contact each other, leaving unrepressed chromatin in between areas of contact. (Adapted from Pryde and Louis 1999.)

E PIG ENE TIC 5

tion of silent chromatin itself (Hediger et al. 2002). Within silent chromatin, the anchoring function has been assigned to a subdomain of Sir4 that binds a nuclear envelope-associated protein called Escl (enhances silent chromatin 1; Taddei et al. 2004). Sir4-Esc1 interactions tether the SIR-repressed chromatin domain at the nuclear envelope, at sites distinct from pores. Even in the absence of a yKu anchoring pathway, the association of telomeres with the nuclear envelope can be achieved through the Sir4-Escl association as long as repression is maintained. Moreover, excised rings of silent chromatin separated from their adjacent telomeres by recombination remain tightly associated with the nuclear periphery in a SIRdependent manner (Gartenberg et al. 2004). The initial recruitment of telomeres to the nuclear envelope is probably mediated by yKu, since this functions even in the absence of silencing. This anchoring, together with interactions in trans between telomeres, allows a nuclear subcompartment to form that in turn sequesters SIR proteins (Fig. 10). This compartment is critical for creating a gradient of silencing at telomeres. Even silencer-flanked HM constructs repress more efficiently when they are integrated near telomeres (Thompson et al. 1994; Maillet et al. 1996) or when they are artificially tethered at the nuclear envelope by a transmembrane factor (Andrulis et al. 1998). Importantly, the ability to improve repression due to telomere proximity is lost when Sir3 and Sir4 are no longer sequestered in foci or are overexpressed (Maillet et al. 1996; Marcand et al. 1996). This argues that the SIR protein concentration gra-

INS Ace H A ROM Y C ESC ERE V I S I A E

dient is the feature of telomere clustering that is critical for promoting repression. Finally, it is proposed that the sequestering of general repressors which are in limiting concentrations helps a cell ensure epigenetic inheritance of the silent state as outlined in Figure 10. In brief, the model proposes that the assembly of newly replicated DNA into heterochromatin is likely to be favored if DNA is replicated within a subcompartment that is enriched for silencing factors.

12 Inheritance of Epigenetic States A universal characteristic of heterochromatin is that its silent state is passed from one generation to the next. This requires that the reassembly of a heterochromatic structure on daughter strands occur soon after replication of the DNA template. Pioneering work on the question of how the cell cycle affects the establishment or inheritance of chromatin states was performed by Miller and Nasmyth (1984), who studied the onset and loss of silencing with a temperature-sensitive sid" mutant. A shift from permissive temperature to nonpermissive temperature caused silencing to be lost immediately, indicating that Sid was required for maintenance of the repressed state. However, in the reciprocal experiment, shifting from nonpermissive temperature to permissive temperature did not lead to immediate restoration of repression: Passage through the cell cycle was required. They concluded that an event in S phase was required for establishment of heritably repressed chromatin. This requirement was later shown to

Figure 9. SIR Proteins and Rapl Are Found in Foci at the Nuclear Periphery In panel G, Rap1 (green) identifies 7 clusters representing all 64 telomeres in this diploid cell. They are either perinuclear or adjacent to the nucleolus (blue, anti-Nop 1). DNA is in red. In panel b, telomeric DNA (red) is identified by fluorescent in situ hybridization (FISH), and HML is visualized in green. The two colocalize in about 70% of the cases, and both are adjacent to the nuclear envelope (blue) (Heun et al. 2001). Panel c shows the focal distribution of Sir4 (green) adjacent to the nuclear envelope (Mab414, red). This pattern is lost in a yKu70 deletion strain, coincident with the loss of telomeric silencing (Laroche et al. 1998).

76 •

C HAP T E R

ff~'

,- /1\ .. ,

<===J

Silencer sites

c==:>

Figure 10. Spontaneous Formation of Silencing Subcompartments

1. nucleation = Sir4 recruitment

4. SIR spreading

2. anchoring

3. increased local SIR factor concentration yKu

Sir

'\ir Sir4

<===J

Sir4

Silencer sites

involve events in both Sand G/M phases (Lau et al. 2002). Initially, it was thought that origin firing from the silencer-linked ARS elements might be a critical event in the establishment or inheritance of silent chromatin, but because initiation could not be detected at the origins of the HML locus, this seemed an unlikely explanation. Indeed, the experiment showing that ORC can be efficiently replaced by a targeted GDB-Sirl fusion protein put to rest the notion that origin firing contributes in an essential manner to the inheritance of silent chromatin. On the other hand, several lines of evidence indicate that passage through S phase is necessary for heterochromatin assembly. This was widely interpreted as a requirement for DNA replication and its associated reassembly of nucleosomal structure, yet recent experiments have shown that establishment of repression can occur on DNA that does not replicate (Kirchmaier and Rine 2001; Li et al. 2001). Candidates for the missing factor(s) needed for establishing the repressed chromatin state are thus proteins that might be specifically activated in late S phase, or have a specific S-G z phase function. Among these may be de novo synthesized histones or histonemodifying enzymes. They may also include chaperones, such as the chromatin assembly factor l(CAF1) complex, which might ensure a critical histone assembly step. Other studies have shown that robust silencing is not achieved until telophase, well beyond the S-phase window of nucleosome assembly. It appears that a cohesin subunit, Sccl, inhibits stable repression unless it is

A simple model for the formation of subnuclear compartments is shown. (1) Sir4 is first recruited at the nucleation center by DNA-binding proteins that can bind Sir4. These include Rap1, ORC, Abfl, and yKu. (2) The presence of Sir4 at the locus will then bring it to the nuclear periphery through one of the two Sir4-anchoring pathways (yKu or Es(1). (3) At the nuclear envelope, the high local concentrations of SIR proteins will help silencing complexes assemble and spread. (4) The ability of silent loci to remain attached at the periphery increases the local concentration of SIR proteins and reinforces the silencing of other loci within this region. Importantly, telomere-bound yKu can independently recruit telomeres to the nuclear envelope just as Sir4 recruits silencer sequences.

destroyed at the metaphase/anaphase transition (Lau et al. 2002). This correlates with findings that the targeting of transcription factors can efficiently disrupt or compete with the establishment of silent chromatin in G/M phase, but not after cells have passed M and entered G} (Aparicio and Gottschling 1994). Together, these findings argue that in addition to a critical S-phase component or event, there is an additional step that requires passage through mitosis and depends either directly or indirectly on the loss of sister chromatid cohesion. 13 Aging and Sir2: linked by rDNA Repeat Instability

In Drosophila, highly active rDNA repeats are adjacent to centromeric heterochromatin, and in many higher eukaryotic species, nucleoli and condensed heterochromatin are spatially juxtaposed. It is significant, therefore, that in yeast, Sir2, independent of the other SIR proteins, is genetically and physically associated with the highly transcribed rDNA repeats (Gotta et al. 1997). Indeed, rDNA recombination is suppressed by Sir2 (Gottlieb and Esposito 1989), as is the repression of RNA pol-IIdependent reporters that are introduced into the rDNA array. Due to their tandemly repeated nature, the rDNA repeats are prone to unequal recombination events that can lead to either a reduction or increase of the rDNA array. Such instability has been also correlated with the accumulation of extrachromosomal rDNA circles (Fig.

EPIGENETICS

11) (Sinclair and Guarente 1997). Sir2 is required both to repress reporter genes inserted in the rDNA array and to prevent aberrant recombination that leads to a loss of rDNA repeats, perhaps either by positioning nucleosomes (Fritze et al. 1997) or by aligning repeats between sister chromatids to preclude unequal exchange events (Kobayashi et al. 2004). The most surprising phenotype of the yeast sir2 null allele is a reduction in life span, which in yeast has no direct link to the cell's loss of telomeric repression or to the length of the TG tract. Indeed, the short-lived phenotype of sir2-deficient yeast cells means that these cells divide on average less than 12 times, rather than 20-25 times as observed for wild-type cells (Kaeberlein et al. 1999). It is now convincingly established that the production of extrachromosomal rDNA repeat circles (ERe) due to unequal recombination events in the rDNA, and their accumulation in mother cells, correlates with yeast senescence (Fig. 11). Importantly, life span can be not only shortened by the loss of Sir2, but also lengthened by Sir2 overexpression, which increases the amount of Sir2 bound to rDNA. Other mutations that reduce the efficiency of rDNA excision, for instance, elimination of the replication fork barrier protein Fob1 (Defossez et al. 1999), also extend life span in yeast, just as the artificial production of ERC is sufficient to cause cellular aging (Sinclair and Guarente 1997). Thus, rDNA instability is clearly correlated with aging in yeast, although its contribution to senescence may be indirect. One model suggests that the high levels of ERC titrate DNA repair or replication proteins from other genomic loci, leading to increased genomic damage or reduced replication of the rest of the yeast genome.

Virgin cell

SACCHAROMYCES

.77

Because Sir2 is an NAD-dependent deacetylase, and because NAD levels act as a metabolic thermostat, it was proposed that the effect of yeast Sir2 on life span might be related to the extension of life span by caloric restriction, a conserved pathway that attenuates replicative aging in many species. Although this view is supported by studies showing that caloric restriction increases Sir2 activity in yeast, flies, and mammals, yeast life span is extended by growth on low glucose (caloric restriction) in a manner that is independent of and additive to the role of Sir2 (Kaeberlein et al. 2004). Thus, Sir2 and caloric restriction increase life span through independent pathways. The accumulation of excised rDNA rings has not been detected in any other species, yet it has been proposed for both Caenorhabditis elegans and rodents that other types of genomic instability are associated with shortened life span in these species. Analogous to the events in budding yeast where loss of Sir2 leads to unequal inter-sister recombination, it is possible that the loss of heterochromatin at mammalian telomeres leads to end-to-end chromosomal fusions, which restrict the division potential of cells. Although it is not yet known whether mammalian Sir2 influences these mechanisms, it is nonetheless likely that genomic instability will be a common factor in aging, and that the loss of heterochromatin structure may well contribute specifically to these events. 14 Summary

Combined genetic, biochemical, and cytological techniques have been exploited in budding yeast to demonstrate fundamental principles at work during heterochromatin-mediated gene silencing. These princi-

-Q;~ Excision or inheritance of an ERG

CEREVISIAE

Replication Recombination Asymmetrical segregation

Nucleolar fragmentation Relocalization of SIR proteins Death

Sir2

Figure 11. rDNA Recombination Leads to Cellular Senescence in Yeast The rDNA is organized in an array of 140-200 direct repeats of a 9.1-kb unit (red block). These encode the 185, 5.85, 255, and 55 rRNAs, and contain two 5ir2-responsive elements downstream of the 55 gene and within the 185 gene. The rDNA repeats tend to be excised in aging yeast cells, and the circles accumulate in the mother cell (Kaeberlein et al. 1999). This correlates with premature senescence and can be antagonized by 5ir2, which helps suppress unequal recombination and ring excision.

78 •

C HAP T E R 4

pIes include the mechanism of initiation, spreading and barriers to spreading of heterochromatin; the balance of heterochromatin factors and their distribution within a subnuclear environment; and heterochromatin looping and cell cycle involvement in its formation. Moreover, in vitro systems are being developed for the reconstitution of yeast heterochromatin. These in vivo and in vitro studies provide a strong mechanistic basis for our understanding of the assembly of heterochromatin from chromatin fibers in all eukaryotes. References Andrulis E.D., Neiman A.M., Zappulla D.C, and Sternglanz R. 1998. Perinuclear localization of chromatin facilitates transcriptional silencing. Nature 394: 592-595. Aparicio O.M. and Gottschling D.E. 1994. Overcoming telomeric silencing: A trans-activator competes to establish gene expression in a cell cycle-dependent way. Genes Dev. 8: 1133-1146. Aparicio O.M., Billington B.L., and Gottschling D.E. 1991. Modifiers of position effect are shared between telomeric and silent mating-type loci in S. cerevisiae. Cell 66: 1279-1287. Boscheron C, Maillet 1., Marcand S., Tsai-Pflugfelder M., Gasser S.M., and Gilson E. 1996. Cooperation at a distance between silencers and proto-silencers at the yeast HML locus. EMBO f. 15: 2184-2195. Brand A.H., Micklem G., and Nasmyth K. 1987. A yeast silencer contains sequences that can promote autonomous plasmid replication and transcriptional activation. Cell 51: 709-719. Braunstein M., Sobel R.E., Allis CD., Turner B.M., and Broach J.R. 1996. Efficient transcriptional silencing in Saccharomyces cerevisiae requires a heterochromatin histone acetylation pattern. Mol. Cell. Bioi. 16: 4349-4356. Brickner J.H. and Walter P. 2004. Gene recruitment of the activated INOllocus to the nuclear membrane. PLoS BioI. 2: e342. Chen 1. and Widom J. 2005. Mechanism of transcriptional silencing in yeast. Celll20: 37-48. Cheng T.H. and Gartenberg M.R. 2000. Yeast heterochromatin is a dynamic structure that requires silencers continuously. Genes Dev. 14: 452-463. Chien CT., Buck S., Sternglanz R., and Shore D. 1993. Targeting of SIR1 protein establishes transcriptional silencing at HM loci and telomeres in yeast. Cell 75: 531-541. Chopra V.S. and Mishra R.K. 2005. To SIR with Polycomb: Linking silencing mechanisms. Bioessays 27: 119-121. Clarke D.J., O'Neill L.P., and Turner B.M. 1993. Selective use of H4 acetylation sites in the yeast S. cerevisiae. Biochem. f. 294: 557-561. Cubizolles E, Martino E, Perrod S., and Gasser S.M. 2006. A hom*otrimer-heterotrimer switch in Sir2 structure differentiates rDNA and telomeric silencing. Mol. Cell 21: 825-836. de Bruin D., Zaman Z., Liberatore R.A., and Ptashne M. 2001. Telomere looping permits gene activation by a downstream UAS in yeast. Nature 409: 109-113. Defossez P.A., Prusty R., Kaeberlein M., Lin S.J., Ferrigno P., Silver P.A., Keil R.L., and Guarente 1. 1999. Elimination of replication block protein Fob 1 extends the life span of yeast mother cells. Mol. Cell 3: 447-455. Donze D. and Kamakaka R.T. 2001. RNA polymerase III and RNA

polymerase II promoter complexes are heterochromatin barriers in S. cerevisiae. EMBO f. 20: 520-531. Fourel G., Revardel E., Koering CE., and Gilson E. 1999. Cohabitation of insulators and silencing elements in yeast subtelorneric regions. EMBO f. 18: 2522-2537. Fourel G., Boscheron C, Revardel E., Lebrun E., Hu Y.E, Simmen K.C, Muller K., Li R., Mermod N., and Gilson E. 2001. An activationindependent role of transcription factors in insulator function. EMBO Rep. 2: 124-132. Fritze CE., Verschueren K., Strich R., and Easton Esposito R. 1997. Direct evidence for SIR2 modulation of chromatin structure in yeast rDNA. EMBO f. 16: 6495-6509. Frye R.A. 2000. Phylogenetic classification of prokaryotic and eukaryotic Sir2-like proteins. Biochem. Biophys. Res. Comm. 273: 793-798. Gartenberg M.R., Neumann ER., Laroche T., Baszczyk M., and Gasser S.M. 2004. Sir-mediated repression can occur independently of chromosomal and subnuclear contexts. Cell 119: 955-967. Gotta M., Strahl-Bolsinger S., Renauld H., Laroche T., Kennedy B.K., Grunstein M., and Gasser S.M. 1997. Localization of Sir2p: The nucleolus as a compartment for silent information regulators. EMBO ]. 16: 3243-5503. Gottlieb S. and Esposito R.E. 1989. A new role for a yeast transcriptional silencer gene, SIR2, in regulation of recombination in ribosomal DNA. Cell 56: 771-776. Gottschling D.E. 1992. Telomere-proximal DNA in Saccharomyces cerevisiae is refractory to methyltransferase activity in vivo. Proc. Natl. Acad. Sci. 89: 4062-4065. Gottschling D.E., Aparicio O.M., Billington B.L., and Zakian V.A. 1990. Position effect at S. cerevisiae telomeres: Reversible repression of PolII transcription. Cell 63: 751-762. Hecht A., Strahl-Bolsinger S., and Grunstein M. 1996. Spreading Of transcriptional repressor SIR3 from telomeric heterochromatin. Nature 383: 92-96. Hecht A., Laroche T., Strahl-Bolsinger S., Gasser S.M., and Grunstein M. 1995. Histone H3 and H4 N-termini interact with SIR3 and SIR4 proteins: A molecular model for the formation of heterochromatin in yeast. Cell 80: 583-592. Hediger E, Neumann ER., Van Houwe G., Dubrana K., and Gasser S.M. 2002. Live imaging of telomeres: yKu and Sir proteins define redundant telomere-anchoring pathways in yeast. Curro Bioi. 12: 2076-2089. Heun P., Laroche T., Raghuraman M.K., and Gasser S.M. 2001. The positioning and dynamics of origins of replication in the budding yeast nucleus. ]. Cell Bioi. 152: 385-400. Hoppe G.J., Tanny J.C, Rudner A.D., Gerber S.A., Danaie S., Gygi S.P., and Moazed D. 2002. Steps in assembly of silent chromatin in yeast: Sir3-independent binding of a Sir2/Sir4 complex to silencers and role for Sir2-dependent deacetylation. Mol. Cell. Bioi. 12: 4167-4180. Imai S.l., Armstrong C, Kaeberlein M., and Guarente 1. 2000. Transcriptional silencing and longevity protein Sir2 is an NADdependent histone deacetylase. Nature 403: 795-800. Ishii K., Arib G., Lin C, Van Houwe G., and Laemmli U.K. 2002. Chromatin boundaries in budding yeast: The nuclear pore connection. Cell 109: 551-562. Johnson L.M., Fisher-Adams G., and Grunstein M. 1992. Identification of a non-basic domain in the histone H4 N-terminus required for repression of the yeast silent mating loci. EMBO]. 11: 2201-2209. Johnson L.M., Kayne P.S., Kahn E.S., and Grunstein M. 1990. Genetic evidence for an interaction between SIR3 and histone H4 in the

E PIG ENE TIC 5

repression of the silent mating loci in S. cerevisiae. Pmc. Natl. Acad. Sci. 87: 6286-6290. Kaeberlein M., McVey M., and Guarente L. 1999. The SIR2/3/4 complex and SIR2 alone promote longevity in S. cerevisiae by two different mechanisms. Genes Dev. 13: 2570-2580. Kaeberlein M., Kirkland K.T., Fields S., and Kennedy B.K. 2004. Sir2independent life span extension by calorie restriction in yeast. PLoS BioI. 2: 296--307. Kayne P.S., Kim u.J., Han M., Mullen J.R., Yoshizaki F., and Grunstein M. 1988. Extremely conserved histone H4 N terminus is dispensable for growth but essential for repressing the silent mating loci in yeast. Cell 55: 27-39. Kimura A., Vmehara T., and Horikoshi M. 2002. Chromosomal gradient of histone acetylation established by Sas2p and Sir2p functions as a shield against gene silencing. Nat. Genet. 3: 370-377. Kirchmaier A.L. and Rine J. 2001. DNA replication-independent silencing in S. cerevisiae. Science 291: 646-650. Kobayashi T., Horiuchi T., Tongaonkar P., Vu L., and Nomura M. 2004. Sir2 regulates recombination between different rDNA repeats, but not recombination within individual rRNA genes in yeast. Cell 117: 441-453. Laroche T., Martin S.G., Gotta M., Gorham H.e., Pryde F.E., Louis E.J., and Gasser S.M. 1998. Mutation of yeast Ku genes disrupts the subnuclear organization of telomeres. Curro BioI. 8: 653-656. Lau A., Blitzblau H., and Bell S.P. 2002. Cell-cycle control of the establishment of mating-type silencing in S. cerevisiae. Genes Dev. 16: 29352945. Li Y.e., Cheng T.H., and Gartenberg M.R. 2001. Establishment of transcriptional silencing in the absence of DNA replication. Science 291: 650-653. Liou G.G., Tanny J.e., Kruger R.G., Walz T., and Moazed D. 2005. Assembly of the SIR complex and its regulation by O-acetyl-ADP-ribose, a product of AD-dependent histone deacetylation. Cell 121: 515-527. Loo S. and Rine J. 1994. Silencers and domains of generalized repression. Science 264: 1768-1771. Luo K., Vega-Palas M.A., and Grunstein M. 2002. Rapl-Sir4 binding independent of other Sir, yKu, or histone interactions initiates the assembly of telomeric heterochromatin in yeast. Genes Dev. 12: 1528-1539. Maillet L., Boscheron e., Gotta M., Marcand S., Gilson E., and Gasser S.M. 1996. Evidence for silencing compartments within the yeast nucleus: A role for telomere proximity and Sir protein concentration in silencer-mediated repression. Genes Dev. 10: 1796--1811. Marcand S., Gilson E., and Shore D. 1997. A protein-counting mechanism for telomere length regulation in yeast. Science 275: 986--990. Marcand S., Buck S.W., Moretti P., Gilson E., and Shore D. 1996. Silencing of genes at nontelomeric sites in yeast is controlled by sequestration of silencing factors at telomeres by Rapl protein. Genes Dev. 10: 1297-1309. Martin S.G., Laroche T., Suka N., Grunstein M., and Gasser S.M. 1999. Relocalization of telomeric Ku and SIR proteins in response to DNA strand breaks in yeast. Cell 97: 621-633. Meneghini M.D., Wu M., and Madhani H.D. 2003. Conserved histone variant H2A.Z protects euchromatin from the ectopic spread of silent heterochromatin. Cell 112: 725-736. Miller A.M. and asmyth K.A. 1984. Role of DNA replication in the repression of silent mating type loci in yeast. Nature 312: 247-251. Mishra K. and Shore D. 1999. Yeast Ku protein plays a direct role in telo-

INS Ace H A ROM Y C ESC ERE V I S I A E

meric silencing and counteracts inhibition by Rif proteins. Curro BioI. 9: 1123-1126. Moazed D., Kistler A., Axelrod A., Rine J., and Johnson A.D. 1997. Silent information regulator protein complexes in S. cerevisiae: A SIR2/SIR4 complex and evidence for a regulatory domain in SIR4 that inhibits its interaction with SIR3. Pmc. Natl. Acad. Sci. 94: 2186--2191. Moretti P., Freeman K., Coodly L., and Shore D. 1994. Evidence that a complex of SIR proteins interacts with the silencer and telomerebinding protein RAP1. Genes Dev. 8: 2257-2269. Palladino F., Laroche T., Gilson E., Axelrod A., Pillus L., and Gasser S.M. 1993. SIR3 and SIR4 proteins are required for the positioning and integrity of yeast telomeres. Cell 75: 542-555. Pill us L. and Rine J. 1989. Epigenetic inheritance of transcriptional states in S. cerevisiae. Cell 59: 637-647. Pryde F.E. and Louis E.J. 1999. Limitations of silencing at native yeast telomeres. EMBO f. 18: 2538-2550. Renauld H., Aparicio O.M., Zierath P.D., Billington B.L., Chhablani S.K., and Gottschling D.E. 1993. Silent domains are assembled continuously from the telomere and are defined by promoter distance and strength, and by SIR3 dosage. Genes Dev. 7: 1133-1145. Rusche L.N., Kirchmaier A.L., and Rine J. 2002. Ordered nucleation and spreading of silenced chromatin in Saccharomyces cerevisiae. Mol. Bioi. Cell 7: 2207-2222. - - - . 2003. The establishment, inheritance, and function of silenced chromatin in Saccharomyces cerevisiae. Annu. Rev. Biochem. 72: 481-516. Sekinger E.A. and Gross D.S. 1999. SIR repression of a yeast heat shock gene: VAS and TATA footprints persist within heterochromatin. EMBO]. 18: 7041-7055. Sinclair D.A. and Guarente L. 1997. Extrachromosomal rDNA circles-A cause of aging in yeast. Cell 91: 1033-1042. Stavenhagen J.B. and Zakian V.A. 1994. Internal tracts of telomeric DNA act as silencers in Saccharomyces cerevisiae. Genes Dev. 8: 1411-1422. Strahl-Bolsinger S., Hecht A., Luo K., and Grunstein M. 1997. SIR2 and SIR4 interactions differ in core and extended telomeric heterochromatin in yeast. Genes Dev. 11: 83-93. Suka N., Luo K, and Grunstein M. 2002. Sir2p and Sas2p opposingly regulate acetylation of yeast histone H4 lysine16 and spreading of heterochromatin. Nat. Genet. 3: 378-383. Suka N., Suka Y., Carmen A.A., Wu J., and Grunstein M. 2001. Highly specific antibodies determine histone acetylation site usage in yeast heterochromatin and euchromatin. Mol. Cell 8: 473-479. Taddei A., Hediger F., Neumann F.R., Bauer e., and Gasser S.M. 2004. Separation of silencing from perinuclear anchoring functions in yeast Ku80 Sir4 and Escl proteins. EMBO]. 23: 1301-1312. Tanner KG., Landry J., Sternglanz R., and Denu J.M. 2000. Silent information regulator 2 family of NAD-dependent histone/protein deacetylases generates a unique product, 1-0-acetyl-ADP-ribose. Proc. NatlAcad. Sci. 97: 14178-14182. Thompson J.S., Johnson L.M., and Grunstein M. 1994. Specific repression of the yeast silent mating locus HMR by an adjacent telomere. Mol. Cell. BioI. 14: 446-455. van Leeuwen F., Gafken P.R., and Gottschling D.E. 2002. Dotlp modulates silencing in yeast by methylation of the nucleosome core. Cell 109: 745-756. Weiss K and Simpson R.T. 1998. High-resolution structural analysis of chromatin at specific loci: S. cerevisiae silent mating type locus HMLa. Mol. Cell. BioI. 18: 5392-5403.

Position-Effect Variegation, Heterochromatin Formation, and Gene Silencing in Drosophila Sarah C.R. Elgin 1 and Gunter Reute~ 'Department of Biology, Washington University, St. Louis, Missouri 63130 lInstitute of Genetics, Biologicum, Martin Luther University Halle, D-06120 Halle, Germany

CONTENTS 1. Genes Abnormally Juxtaposed with Heterochromatin Exhibit a Variegating Phenotype, 83 2. Screens for Suppressors and Enhancers of PEV Have Identified Chromosomal Proteins and Modifiers of Chromosomal Proteins, 85 3. Immunofluorescent Staining of Polytene Chromosomes Has Identified Proteins Specifically Associated with Heterochromatin, 88

6. How Is Heterochromatin Formation Targeted in Drosophila?, 93 7. Not All Heterochromatin Is Identical, 95 8. PEV, Heterochromatin Formation, and Gene Silencing in Different Organisms, 96 9. Summing Up: There Is Much That We Do Not Know about Heterochromatin, 97 Acknowledgments, 98 References, 98

4. Histone Modification Plays a Key Role in Heterochromatin Silencing, 88 5. Chromosomal Proteins Form Mutually Dependent Complexes to Maintain and Spread Heterochromatic Structure, 91

GENERAL SUMMARY Genes that are abnormally juxtaposed with heterochromatin, either by rearrangement or by transposition, exhibit a variegating phenotype, indicating that the gene has been silenced in some of the cells in which it is normally active (position-effect variegation, PEV). The silencing that occurs in PEV can be attributed to packaging of the reporter gene in a heterochromatic form, indicating that heterochromatin formation, once initiated, can spread to encompass nearby genes. Genetic, cytological, and biochemical analyses are all possible in Drosophila melanogaster, and in this chapter we show how these different approaches have converged to identify many potential contributors to this system, leading to characterization of several proteins that play key roles in establishing and maintaining heterochromatin. Heterochromatin formation depends critically on methylation of histone H3 at lysine 9, with concomitant association of. Heterochromatin Protein 1 (HP1) and

other interacting proteins, including H3K9 methyltransferases; the multiple interactions of these proteins are required for maintenance and spreading of heterochromatin. Targeting of heterochromatin formation, including accumulation of H3K9me, appears to involve the RNA interference (RNAi) machinery, although specific protein-DNA interactions may also playa role. Although heterochromatic regions (pericentromeric regions, telomeres, and the small fourth chromosome) share a common biochemistry, each is distinct, and the pericentromeric regions are mosaic. Heterochromatin in Drosophila is gene-poor, but it is not devoid of genes, and those genes that reside in heterochromatin are dependent on this environment for full expression. The final model for heterochromatin formation and maintenance (including targeting and spreading) will need to take into account the different responses of different genes to this chromatin environment.

PEV,

HETEROCHROMATIN

FORMATION,

Genes Abnormally Juxtaposed with Heterochromatin Exhibit a Variegating Phenotype

Large segments of the eukaryotic genome, primarily repetitious sequences, are packaged in a permanently inactive form as constitutive heterochromatin. This chromatin fraction was originally identified as that portion of the genome that remains condensed and deeply staining (heteropycnotic) as the cell makes the transition from metaphase to interphase; such material is generally associated with the telomeres and pericentromeric regions of the chromosomes. Heterochromatic regions tend to be late replicating and show little or no meiotic recombination. These regions are gene-poor, but they are not devoid of genes, and those genes that are present frequently are dependent on that environment for optimal expression. About one-third of the Drosophila genome is considered heterochromatic, including the entire Y chromosome, most of the small fourth chromosome, the pericentromeric 40% of the X chromosome, and the pericentromeric 20% of the large autosomes. During the last few decades, we have learned a great deal about the biochemistry of heterochromatin, and much of that understanding derives from our studies with Drosophila (Richards and Elgin 2002; Schotta et al. 2003). One of the first mutations identified in D. melanogaster was white, a mutation that results in a fly with a white eye, rather than the characteristic red pigmentation. Using X rays as a mutagen, Muller (1930) observed an unusual phenotype, in which the eye was variegating, with some patches of red and some patches of white facets (Fig. 1). This phenotype suggested that the white gene itself was not damaged-after all, some facets remained red, and flies with entirely red eyes could be recovered as revertants, again using X rays as the mutagen. However, the white gene had clearly been silenced in some of the cells in which it is normally expressed. Subsequent examination of the polytene chromosomes (shown below, see Fig. 4) indicated that such phenotypes were the consequence of an inversion or rearrangement, with one breakpoint within the pericentromeric heterochromatin and one breakpoint adjacent to the white gene (see Fig. 1). Because the variegating phenotype is caused by a change in the position of the gene within the chromosome, this phenomenon is referred to as position-effect variegation (PEV). In Drosophila, virtually every gene that has been examined in an appropriate rearrangement has been shown to variegate, and rearrangements involving the pericentromeric heterochromatin of any chromosome

AND

GENE

SILENCING

DROSOPHILA.

a white

Wild Type

Inversion

.E(var)

---.

Su(var)

Figure 1. Schematic Illustration of white Variegation in the XChromosome Inversion In(1 )w"'4 The white locus, normally located in the distal euchromatin (blue), is now placed within 25 kb of a breakpoint within the pericentromeric heterochromatin (pink) of the X chromosome due to an Xray-induced inversion. Spreading of heterochromatin packaging into the euchromatic domain results in silencing; loss of silencing in some cells during differentiation results in a variegating phenotype. Given a fly exhibiting PEV, one can select for second-site mutations that either suppress the phenotype (Su(var) mutations; resulting in a loss of silencing) or enhance the phenotype (E(var) mutations; causing an increase in silencing).

can lead to PEY. PEV has been observed in a variety of organisms, including yeasts, flies, and mammals; but has been used as a tool to study heterochromatin formation primarily in Drosophila. PEV indicates that such rearrangements allow packaging in a heterochromatic configuration to "spread" along the chromosome. Apparently, the rearrangement has removed a normally existing barrier or buffer zone. The consequence is an altered packaging and silencing of genes normally arranged in a euchromatic form. Visual inspection of the polytene chromosomes of larvae carrying such a rearrangement shows that the region carrying the reporter gene is now packaged in a dense block of heterochromatin, but only in the cells in which the gene is inactive (Zhimulev et al. 1986). Patterns observed as a consequence of rearrangement of white can vary in the number of pigmented cells, the size of the pigmented patches, and the level of pigment in the two different cell types observed (Fig. 1). In a system using an inducible lac-Z gene as a reporter, investigators observed that silencing occurs in embryogenesis, when heterochromatin is first observed

84 •

C HAP T E R 5

cytologically, and is epigenetically inherited in both somatic and germ-line lineages; the mosaic phenotype was determined during differentiation by variegated relaxation of silencing in third-instar larvae (Lu et al. 1996). However, not all variegating genes remain silent until after differentiation, and the balance of factors leading to the "ON/OFF" decision no doubt differs for different genes. (See Ashburner et al. 2005b, for a more detailed discussion.) Given a fly exhibiting a PEV phenotype, it is straightforward to screen for dominant second-site mutations (induced by chemical mutagens that cause point mutations or small insertions/deletions) that are either suppressors of PEV (denoted Suppressor of variegation, Su[varJ), resulting in a loss of silencing, or enhancers of PEV (denoted Enhancer of variegation, E[varJ), resulting in an increase in silencing (Fig. 1). About 30 modifiers of PEV have been isolated and characterized, but many more candidates are predicted from such screens. Where the gene has been cloned and the product characterized, one generally finds a chromosomal protein or a modifier of a chromosomal protein (see below). A small subset of these loci cause both a haplo-abnormal and an opposite triplo-abnormal phenotype; i.e., if one copy of the gene results in suppression of PEV, three copies result in enhancement of PEV. Identification of such loci has led to the suggestion that the protein products of these genes play a structural role in heterochromatin, and that the spread of heterochromatic packaging can be driven by the dosage of these proteins in a stochastic manner (Fig. 2) (Locke et al. 1988). However, "spreading" is a complex process, not a simple linear continuum-which most likely is dependent on the organization of the DNA in the region being silenced (see below). The results observed on rearrangement of chromosomes suggest that a euchromatic gene inserted into a heterochromatic domain by transposition will also show a variegating phenotype, and this has been found to be the case. The P element, a DNA transposon found in many strains of Drosophila in the wild, can be engineered for this purpose. A natural P element has distinctive inverted repeat sequences at each end, and codes for just one enzyme, the P-specific DNA transposase. Reporter constructs lacking the DNA transposase but containing other genes of interest can be inserted into the Drosophila genome in the presence of active transposase by co-injection into Drosophila embryos. A P-based transposable element such as that shown in Figure 3a, carrying an hsp70-driven copy of white, can be used in a fly with no endogenous copy of white to identify domains of heterochromatin. When the P element is inserted into euchro-

Effect of PEV modifiers on white variegation Euchromatin

Heterochromatin

white loss of silencing

white enhanced silencing

Figure 2. Dosage-dependent Effects of Some Modifiers of PEV The modifiers of PEV that have a dosage-dependent effect are thought to be structural proteins of heterochromatin. Whereas a variegating phenotype (exhibited here by a white reporter gene) is seen when the wild-type modifier gene is present in two copies (middle chromosome, middle fly eye), the presence of three wild-type copies of the modifier gene will drive more extensive heterochromatin formation, resulting in an enhancement of reporter gene silencing (lower chromosome, lower fly eye). Conversely, the presence of only one wild-type copy of the modifier gene will result in less heterochromatin formation and more expression from the reporter gene (upper chromosome, upper fly eye).

matin, the fly has a red eye. When this P is mobilized (by crossing in the gene encoding the transposase), approximately 1% of the lines recovered show a variegating eye phenotype. In situ hybridization shows that in these cases, the P element has jumped into the pericentromeric heterochromatin, the telomeres, or the small fourth chromosome (Wallrath and Elgin 1995). This identification of heterochromatic domains is in agreement with earlier cytological studies. The use of such P elements has allowed comparison of the packaging of the same reporter gene in heterochromatic and euchromatic environments. Heterochromatin is relatively resistant to cleavage by nucleases, whether nonspecific (e.g., DNase I) or specific (restriction enzymes), and is less accessible to other exogenous probes, such as dam methyltransferase. Analysis of the same hsp26 transgene (marked with a fragment of unique plant DNA, Fig. 3a) in euchromatin and pericentromeric heterochromatin using micrococcal nuclease (MNase) reveals a shift to a more ordered nucleosome array, indicating regular spacing of the nucleosomes in heterochromatin (Fig. 3b,c). The MNase cleavage fragments are well-defined, suggesting a smaller MNase target than usual in the linker region. The ordered nucleosome array extends across the 5' regulatory region of the gene, a shift

P E V,

H E T ERa C H ROM A TIN

FOR MAT ION,

Transposable Element

-'=-c:;:::.~;::::=.....=JI-=!L...,.~~~, P

hsp26-plant

Euchromatic insertion

hsp70-white

Heterochromatic insertion

39C-X

HS-2 _-====:"1::::1 MNase

....~ .... ....

.... ....

....

Gel electrophoresis

Heterochromatin

Figure 3. Heterochromatin Is Packaged into a Regular NUcleosome Array A transposable element such as that shown (a), carrying a marked copy of a heat shock gene for study and an hsp70-driven copy of white as a visual marker, can be used to examine the same gene in different chromatin domains. Nuclei from Drosophila embryos from a line carrying this transgene in a euchromatic domain (39C-X; red eye) and a line carrying the same transgene in a heterochromatic domain (HS-2; variegating eye) were digested with increasing amounts of MNase, the DNA purified and run out on an agarose gel, and a Southern blot hybridized with a probe unique to the transgene (b). Linker sites cleaved by MNase are marked with arrows. (c) Densitometer scans from the last lane of each sample are compared (top to bottom is left to right). An array of 9-10 nUcleosomes can be detected in heterochromatin (red line), compared to 5-6 in euchromatin (blue line), indicating more regular spacing in the former case. (d) A diagrammatic representation of the results. (b, (, Adapted, with permission, from Sun et al. 2001 [© American Society for Microbiology].)

AND

G ENE

5 f LEN C f N GIN

D R 0 SOP H I L A .

that no doubt contributes to the observed loss of 5' hypersensitive (HS) sites (Sun et al. 2001). Indeed, although the mechanism of silencing is as yet incompletely understood, there is abundant evidence of transcriptional repression of strongly variegating genes, including loss of binding of TFIID and other transcription factors (Cryderman et al. 1999b).

2 Screens for Suppressors and Enhancers of PEV Have Identified Chromosomal Proteins and Modifiers of Chromosomal Proteins

PEV can be modified by a variety of factors. The temperature during development and the amount of heterochromatin within the genome were the first factors shown to affect the extent of variegation. As a rule, an increase in developmental temperature (from 25°C to 29°C) results in suppression of variegation (loss of silencing), whereas lower temperatures (e.g., 18°C) cause enhancement of variegation (increase in silencing). Other changes in culture conditions that accelerate or slow the rate of development can have similar effects. Strong suppression is found in flies carrying an additional Y chromosome (XXY females and "XYY males), whereas strong enhancement is shown in males without a Y chromosome (XO). In general, duplication of heterochromatic material suppresses, whereas deletions of heterochromatic material enhance, variegation. These effects may be due to the titration of a fixed amount of key proteins required for heterochromatin packaging. The first second-site mutations to suppress or enhance PEV were identified by Schultz (1950) and Spofford (1967). At present, approximately 150 genes are implicated as modifiers of PEV loci. The Su(var) and E(var) mutations identify genes causally connected with the onset of heterochromatic gene silencing in PEV. Molecular analysis of these genes has been essential in developing an understanding of the mechanisms leading to heterochromatin formation and gene silencing. In most cases, the modifying effect of the mutations on PEV is dominant, and Su(var)/+ or E(var)/+ heterozygotes show a suppressed or enhanced PEV phenotype (Fig. 1). Efficient isolation and thorough genetic analysis of Survar) and E(var) mutations depend on the availability of an experimentally suitable PEV rearrangement. Although a large number of PEV rearrangements have been described (Flybase 2005), only a few can be readily used for efficient genetic screens to isolate dominant modifier mutations. One of the most useful PEV rearrangements for such experimental work is In(l)wnJ4 (Muller 1930). This rearrangement variegates for white, a

86 • C HAP

TER 5

phenotype easily recognizable in the eye of adult flies, as shown in Figure 1. Penetrance of white variegation in W"'4 is 100%, so every fly in the starting stock shows an eye with a white variegated phenotype. Inactivation of the white gene does not affect viability or fertility, allowing unlimited work with flies hom*ozygous for W"'4. In the W",4 rearrangement, an inversion results in juxtaposition of the white gene with heterochromatic material of the X chromosome located at the distal border of the nucleolus organizer (Cooper 1959). This region contains tandem arrays of Rl type mobile elements; the heterochromatic breakpoint of In(l)w,"4 is found within an Rl repeat unit (A. Ebert and G. Reuter, unpubl.). Phenotypic w+ revertants of W"'4 have been isolated after X-ray or EMS (ethane methyl sulfonate, a chemical mutagen) treatment (Tartof et al. 1984; Reuter et al. 1985). Analysis of a series of more than 50 of the w+ revertant chromosomes (all exhibiting reinversion or translocation of the white gene to a euchromatic neighborhood) has suggested that the heterochromatic material immediately flanking the breakpoint causes the inactivation of the white gene in W"'4. Most of the revertants show white variegation again if strong E(var) mutations are introduced, suggesting that some heterochromatic sequences remain associated with the white gene after relocation (Reuter et al. 1985), which is not surprising, given that the breakpoint in the flanking DNA is randomly introduced. These studies implicate repetitious DNA (in this case the Rl repeat units) as a target for heterochromatin formation. Most of the PEV modifier mutations known have been isolated using a sensitized genetic background. For isolation of dominant suppressor mutations, the t~st genotype contains a dominant enhancer, whereas a dominant suppressor is used in schemes for the isolation of enhancer mutations (Dorn et al. 1993b). If the test genotype contains an enhancer of variegation, all of the W",4 flies have white eyes, and exceptions with variegated or red eyes indicate newly induced dominant Su(var) mutations. Correspondingly, with a dominant suppressor in the test line, all W",4 flies have red eyes, and exceptional flies with a variegated phenotype indicate newly induced E(var) mutations. These sensitized genetic schemes favor isolation of strongly dominant Su(var) and E(var) mutations, which have been found to be very useful for detailed genetic analysis. More than one million flies have been inspected in different screens using this approach, and more than 140 Su(var) and 230 E(var) mutations have been isolated (Schotta et al. 2003). Mutations have been induced by EMS, by X-ray treatment, or by remobilization of P ele-

ments. Another set of Su(var) mutations has been isolated in a direct screen with W"'4 (Sinclair et al. 1983). Screens with a Df(l;f) chromosome, which shows strong variegation for the yellow gene, a body-color marker, resulted in isolation of 70 PEV modifier mutations (Donaldson et al. 2002). In addition, screens for dominant modifiers of transposon reporter gene expression have identified several mutations with a Su(var) effect (Birchler et al. 1994). A subset of critical regulatory genes is known to be downregulated by the Polycomb group (PeG) genes, and upregulated by the trithorax group (trxG) genes. In direct tests, relatively few mutations in PeG genes modify PEV (e.g., Sinclair et al. 1998). In contrast, many mutations in the trxG genes are enhancers of PEV (Dorn et al. 1993a; Farkas et al. 1994). Altogether, these screens have identified a large number of dominant Su(var) and E(var) mutations. Based on the genetic analysis performed to date, the total number of Su(var) and E(var) mutations can be estimated to be around 150. The large number of Su(var) and E(var) genes with almost identical phenotypic effects has sometimes resulted in inconsistencies in the genetic nomenclature. Most frequently, the Su(var) and E(var) gene symbols are combined with numbers indicating the chromosome where the mutation is located, the gene number, 17 and the number of the allele. Thus, Su(var)3_9 symbolizes allele 17 of the ninth Su(var) gene identified on the third chromosome. At present, only around 30 of the corresponding genes have been carefully mapped, and alleles have been identified (Table 1). Dosage-dependent effects have been observed for about one-third of the identified genes using a series of overlapping deficiencies and duplications. In these cases, reduction in the amount of the gene products, due to loss of one copy of the gene, consistently results in modification of the variegating phenotype. Deletions of these Survar) or E(var) loci suppress or enhance gene silencing, respectively. The duplication studies identified a few modifier loci that show an opposite (antipodal) effect on PEV if an extra copy of the gene is introduced by a duplication or by a transgene insertion. The total number of PEV modifier genes showing dosagedependent effects is estimated to be about 15-20 (Schotta et a12003). If loss of one copy of a gene results in suppression of PEV, and the presence of three copies of a gene leads to an enhancement of PEV, this suggests that the encoded gene product is required in stoichiometric amounts for the establishment of heterochromatin, with concomitant gene silencing (see Fig. 2). Three such loci, Su(var)2-5 (encoding HP1), Su(var)3-7 (encoding a zinc finger protein), and

PEV,

HETEROCHROMATIN

FORMATION,

AND

GENE

SILENCING

DROSOPHILA

Table 1. Genetically defined Su(var) and E(var) genes and their molecular functions Su(var)/ E(var) gene

Cytological position

Molecular function, protein distribution, and phenotypic effects

Suv4-20 [Su(var)]

X; 1B13-14

HKMT, histone H4K20 trimethylation

Su(z)5 [Su(var)]

2L; 21 B2

S-adenosylmethionine synthetase

chm (chameau)

2L; 27F3-4

Myst domain HAT; suppresses PEV but enhances Polycomb-group mutations

[Su(var)] Su(var)2-5 (HP1)

2L; 28F2-3

heterochromatin protein HP1, binding of di- and trimethyl H3K9; binding of SU(VAR)3-9

Su(var)2-HP2

2R; 51 B6

heterochromatin-associated protein, binds HP1

Su(var)2-70

2R; 45A8-9

PIAS protein, negative regulators of JAK/STAT pathway

Su(var)3-648 (HDAC1 =RPD3)

3L; 64B12

histone deacetylase HDAC1, deacetylation of H3K9

E(z) [Su(var)]

3L; 67E5

HKMT, H3K27 mono-, di-, and trimethylation; extra gene copy enhances PEV; in null mutation, all euchromatic and heterochromatic H3K27 methylation lost, H3K9 methylation not affected

SuUR [Su(var)]

3L; 68A4

suppresses heterochromatin underreplication; heterochromatin-associated protein

Su(var)3-7 (JILl)

3L; 68A5-6

antimorphic JIL 7 mutations, carboxy-terminal protein truncations do not affect kinase function; blocking of heterochromatin spreading

Dom (Domina)

3R; 86Bl-2

fork head winged-helix (FKH/WH) protein; heterochromatin-associated

3R; 87B9-10

PP1 protein serine/threonine phosphatase

Su(var)3-7

3R; 87E3

zinc-finger protein, heterochromatin-associated; interacts with HP1 and SU(VAR)3-9

Su(var)3-9

3R; 89E6-8

HKMT, histone H3K9 methylation, heterochromatin-associated, interaction with HP1

mod (modulo)

3R; 100E3

DNA- and RNA-binding protein, phosphorylated Mod binds rRNA

E(var)3-64E/ Ubp64'"''

3L; 64E5-6

putative ubiquitin-specific protease (Ubp46)

Tri (Trithorax-like)

3L; 70F4

GAGA factor, binding of repetitive DNA sequences

3R; 9307

transcription regulator, more than 20 protein isoforms produced by trans-splicing

3R; 93E9-F1

E2F transcription factor, haplo-enhancer and triplo-suppressor

[Su(var)] Su(var)3-6

[Su(var)]

[E(var)] Mod(mdg4)/

E(var)3-93D E(var)3-93E

See Flybase for original citations.

Su(var)3-9 (encoding a histone lysine methyltransferase), have been well characterized. Su(var)2-5 was cloned by screening a cDNA expression library with a monoclonal antibody that recognizes heterochromatin (James and Elgin 1986). The encoded heterochromatin-associated protein was consequently designated HP1, (heterochromatin protein 1). In situ hybridization analysis using the isolated cloned DNA identified a gene in region 28-29 of the polytene chromosomes, where Su(var)2-5 had been previously mapped. DNA sequence analysis of the mutant alleles confirmed that the Su(var)2-5Iocus at chromosome position 28Fl-2 encodes HP1 (Eissenberg et al. 1990). HP1 contains two conserved domains, an amino-terminal chromodomain and a carboxy-terminal chromoshadow domain (Paro and Hogness 1991), and interacts with several other chromosomal proteins. Su(var)3-7was first cytogenetically mapped (using a series of overlapping deletions and dupli-

cations) to region 87El-4 in the third chromosome. This region had been analyzed at the DNA level as part of the first chromosomal walk performed in Drosophila (Bender et al. 1983). Using a series of overlapping genomic clones, Su(var)3-7 was defined within a DNA fragment of 7.8 kb which had a triplo-enhancer effect on a variegating reporter (Reuter et. al. 1990). Su(var)3-7 encodes a protein with seven regularly spaced zinc fingers, domains that have been shown to function in DNA binding (Cleard and Spierer 2001). Su(var)3-9 was cloned by P-element transposon tagging (Tschiersch et al. 1994). The Su(var)3-9 gene in Drosophila (and in all other holometabolic insects studied to date) forms a bicistronic unit with the gene encoding eIF2y (Krauss and Reuter 2000). Because the Suevar)3-9 transcription unit has no introns, it is likely that Su(var)3-9was inserted into an intron of the eIF2ygene via retrotransposition. The SU(VAR)3-9 protein contains a

C HAP T E R 5

chromodomain in its amino-terminal region and the SET domain (identified first in the proteins SU(VAR)3-9, ENHANCER OF ZESTE [E(Z)], and TRITHORAX) (Jones and Gelbart 1993; Tschiersch et al. 1994) at its carboxyl terminus. This protein is a histone methyltransferase that specifically modifies histone H3 at lysine 9. Seven different mutant alleles of Su(var)2-5 have been described, including missense mutations in the chromodomain, premature stop codons, and splicing errors (Eissenberg et al. 1992). Su(var)3-7 mutations have been generated with the help of hom*ologous recombination (Seum et al. 2002); additional alleles have been recovered as suppressors of P-element-dependent silencing (Bushey and Locke 2004). Forty mutant alleles of Su(var)3-9 have been recovered and defined at the molecular level (Ebert et al. 2004). Immunocytological analyses using specific antibodies or transgene-expressed fusion proteins have demonstrated that all three proteins are preferentially associated with heterochromatin (see below and Fig. 4) (James et al. 1989; Cleard et al. 1997; Schotta et al. 2002). Strong colocalization is particularly evident for HP1 and SU(VAR)3-9. These proteins also bind to telomeres and at a number of euchromatic sites (Fanti et al. 1998; Schotta et al. 2002). Several P-element insertions carrying the w+ reporter gene into telomeric regions show white variegation. This phenomenon is called telomere position effect (TPE). Heterochromatin-like packaging is observed at telomereassociated satellite (TAS) sequences, clusters of repetitious DNA just proximal to the HeT-A and TART retroviral elements that make up Drosophila telomeres (Cryderman et al. 1999a). In general, TPE is not found to be modified by mutations in known modifier genes, although HP1 is important for telomere integrity. In cells deficient for this protein, the chromosomes frequently fuse at their telomeres (Fanti et al. 1998). No trans-acting dominant modifier of TPE was identified in Drosophila in a recent screen (Mason et al. 2004), suggesting that these regions are silenced by two (or more) independent mechanisms.

3 Immunofluorescent Staining of Polytene Chromosomes Has Identified Proteins Specifically Associated with Heterochromatin

One advantage of working with Drosophila is the ability to examine the polytene chromosomes, which provide a visual road map of the genome. During the larval stage, the chromosomes in many terminally differentiated cells are replicated but do not go through mitosis; rather, the chromatin strands remain paired, in perfect synapsis,

with all copies aligned. The most extreme case is found in the salivary glands, where the euchromatic arms of the chromosomes have undergone 10 rounds of replication, generating about 1000 copies. Replication is not uniform, however; many repetitious sequences are underreplicated, and satellite DNA sequences are not replicated at all. All of the chromosome arms fuse in a common chromocenter. Thus, in D. melanogaster, one observes five long arms (the X, second left [2LJ, second right [2RJ, third left [3L], third right [3R]), and the short fourth chromosome arm emanating from the condensed chromocenter made up of pericentromeric heterochromatin (see Fig. 4a) (for review, see Ashburner et al. 2005). Although genetic analysis has identified many of the loci required for heterochromatin formation, it does not, in itself, allow us to determine whether the product of a given locus plays a direct or indirect role. Specific association of a protein with heterochromatin was initially observed in a screen of monoclonal antibodies (generated using a fraction of tight-binding nuclear proteins), analyzing the distribution patterns on polytene chromosomes. Antibodies specific for a 22-kD protein subsequently designated HP1 resulted in immunofluorescent "staining" of the pericentromeric heterochromatin, the telomeres, and the banded portion of the small fourth chromosome, all known sites of heterochromatin (Fig. 4a) (James and Elgin 1986). Subsequent analysis (described above) demonstrated that the HP1 protein is encoded by Su(var)2-5, a known suppressor of PEV (Eissenberg et al. 1990). Examining chromosomal localization with specific antibodies, using either mitotic chromosomes (Fanti and Pimpinelli 2004) or polytene chromosomes (which give more resolution, but are deficient in centromeric heterochromatin) (Silver and Elgin 1976), remains the best demonstration that the product of a Su(var) locus encodes a chromosomal protein. Approximately 10 such heterochromatin-specific proteins have been identified; if mutations in the genes encoding these proteins are available, one often observes dominant suppression ofPEV (see Table 1) (Ashburner et al. 200Sb). These proteins, including the recently identified HP2 (Fig. 4a) (Shaffer et al. 2002), are candidates to be structural components of heterochromatin.

4 Histone Modification Plays a Key Role in Heterochromatin Silencing

Analysis of SU(VAR)3-9 has identified a key function required for heterochromatic gene silencing (Tschiersch et al. 1994). The protein contains a SET domain that enzymatically functions in histone H3K9 methylation.

P E V,

H E T ERa C H ROM A TIN

FOR MAT ION,

AND

G ENE

S I LEN C I N GIN

D R 0 SOP H I L A

DNA

HP1-EGFP

(J)

DNA

SU(VAR)3-9-EGFP

(J)

Figure 4. Immunofluorescent Staining of the Polytene Chromosomes Identifies Proteins Predominantly Associated with Heterochromatin (a) The polytene chromosomes, prepared by fixation and squashing of the larval salivary gland (shown by phase contrast microscopy, left) are "stained" by incubating first with antibodies specific for a given chromosomal protein, and then with a secondary antibody coupled to a fluorescent tag. HPl (right) and HP2 (center) have similar distribution patterns showing prominent association with the pericentromeric heterochromatin, small fourth chromosome (inset, arrow), and a small set of sites in the euchromatin arms. Note that the efficacy of any antibody can be affected by the choice of fixation protocol (see Stephens et al. 2003). (b, c) Association of HPl and SU(VAR)3-9 with pericentromeric heterochromatin is interdependent. Mutations in Su(var)3-9 result in a loss of HPl from the pericentromeric heterochromatin (but not the fourth chromosome; see text) (b), whereas mutations in Su(var)2-5 result in delocalization of SU(VAR)3-9 (c). (Adapted from Shaffer et al. 2002.)

That this protein is a histone lysine methyltransferase (HKMT) that targets H3K9 was first shown by characterization of the human SUV39Hl hom*olog (Rea et al. 2000). In Drosophila, SU(VAR)3-9 is the main, but not the only, H3K9 HKMT (Schotta et al. 2002; Ebert et al. 2004). SU(VAR)3-9 controls dimethylation of H3K9 in the bulk of the pericentromeric heterochromatin, but not at the fourth chromosome, the telomeres, or euchromatic sites. Trimethylation of H3K9, which in

Drosophila is observed primarily in the inner chromocenter, is also controlled by SU(VAR)3-9. Dimethylation of this inner region is independent of SU(VAR)3-9, as is monomethylation of H3K9 in pericentromeric heterochromatin (Ebert et al. 2004). The HKMTs responsible for these modifications are still unknown. The importance of H3K9 dimethylation in heterochromatic gene silencing is demonstrated by the strong dosage-dependent effect of SU(VAR)3-9 on PEV (dis-

90 •

C HAP T E R 5

cussed above), as well as by the finding that suppression of gene silencing by Su(var)3-9 mutations correlates with their HKMT activity. The enzymatically hyperactive Su(var)3-~/1I mutation is a strong enhancer of PEV and causes elevated H3K9me2 and H3K9me3 at the chromocenter, as well as generating prominent H3K9me2 signals at many euchromatic sites (ectopic heterochromatin). S-Adenosylmethionine functions as the methyl donor for all of these methylation reactions; consequently, mutations in the gene encoding S-adenosylmethionine synthase, Su(z)5, are dominant suppressors of PEV (Larsson et al. 1996). Studies using mutations in Su(var) genes have begun to reveal the sequence of molecular reactions required to establish heterochromatic domains. SU(VAR)3-9 binding at heterochromatic sequences depends on both its chromo and its SET domains (Schotta et al. 2002). How SU(VAR)3-9 binding is controlled is not yet understood. Methylation of H3K9 by SU(VAR)3-9 establishes binding sites for HPl. The HPI chromodomain specifically binds H3K9me2 and H3K9me3 (Bannister et al. 2001; Lachner et al. 2001). That SU(VAR)3-9 binds HPI has been shown by yeast two-hybrid tests and by immunoprecipitation (Schotta et al. 2002). The region of SU(VAR)3-9 aminoterminal to its chromodomain interacts with the chromoshadow domain of HPl. This region of SU(VAR)3-9 also interacts with the carboxy-terminal domain of SU(VAR)3-7. SU(VAR)3-7 interacts at three different sites with the chromoshadow domain of HP 1 (Delattre et al. 2000). Given this pattern of interactions, one can suggest that the three proteins-HP1, SU(VAR)3-7, and SU(VAR)3-9-physically associate in multimeric heterochromatin protein complexes. Association of SU(VAR)3-9 and HPI with pericentromeric heterochromatin is interdependent (Schotta et al. 2002). SU(VAR)3-9 causes H3K9 dimethylation, which is specifically recognized by the chromodomain of HP 1 (Bannister et al. 2001; Lachner et al. 2001). Consequently, in Suevar)3-9 null larvae, HP 1 binding to pericentromeric heterochromatin is impaired (see Fig. 4b). This reflects the specific activity of HP1, which binds to H3K9me2 but not to H3K9mel; monomethylation is not affected by SU(VAR)3-9 (Ebert et al. 2004). H3K9 dimethylation in the inner chromocenter, the fourth chromosome, at telomeres, and at euchromatic sites does not depend on SU(VAR)3-9, and consequently, HP 1 continues to be found at all of these sites in the mutant lines. SU(VAR)3-9 associates with these sites in wild-type cells, but appears to be inactive; an unknown HKMT controls H3K9 methylation in these regions.

Conversely, if HP 1 is not present (having been depleted by mutations), SU(VAR)3-9 is no longer associated with the pericentromeric heterochromatin, but is also found along the euchromatic chromosome arms (Fig. 4b). It is now seen at almost all bands, where it causes ectopic mono- and dimethylation of H3K9 (H3K9mel and H3K9me2) (Fig. 5). Thus, HPI is essential for the restricted binding of SU(VAR)3-9 to pericentromeric heterochromatin. These data suggest a sequence of reactions starting with SU(VAR)3-9 association with heterochromatic domains and consequent generation of H3K9me2. This mark is recognized by the chromodomain of HP 1; binding of SU(VAR)3-9 to the HPI chromoshadow domain ensures its association with heterochromatin. A chimeric HP 1-PC protein has been generated in which the chromodomain of HP 1 is replaced with the chromodomain of the Polycomb (PC) protein (Platero et al. 1996). The chromodomain of PC binds strongly to H3K27me3 (Fischle et al. 2003). The HPI-PC chimeric protein therefore recognizes H3K27me3 Polycomb-binding sites in the euchromatic arms; in the presence of such a chimeric HPI-PC protein, the SU(VAR)3-9 protein is also found at PC-binding sites, demonstrating its strong association with the chromoshadow domain of HP 1 (Schotta et al. 2002). In SU(VAR)3-9 null cells, another heterochromatinspecific methylation mark, H4K20 trimethylation (H4K20me3), is strongly reduced (Schotta et al. 2004). The interdependence between H3K9 dimethylation and H4K20 trimethylation in heterochromatin has been shown to reflect an interaction between the SU(VAR)3-9, HP1, and SUV4-20 proteins. SUV4-20 is a HKMT that controls H4K20 methylation in heterochromatin. This heterochromatin-specific methylation mark is strongly impaired in SU(VAR)3-9 as well as in HPI null cells, suggesting association of SU(VAR)3-9, HP1, and SUV4-20 in a mutually dependent protein complex. Mutations in the Suv4-20 gene cause strong suppression of PEV-induced gene silencing, indicating that the H4K20me3 mark is required for this process. A third histone methylation mark that is functionally connected with heterochromatin formation is H3K27 methylation catalyzed by the E(Z) HKMT. In Drosophila, E(Z) controls all mono-, di-, and trimethylation ofH3K27 in both euchromatin and heterochromatin. Consequently, in E(z) null cells, all H3K27 methylation is lost (Ebert et al. 2004). A function of H3K27 methylation in heterochromatic gene silencing is indicated by both the Su(var) effect of fez) loss-of-function mutations and the enhancer effect of additional fez) gene copies (Laible et al. 1997). It is not clear whether this effect is direct or indirect. H3K27

PEV,

HETEROCHROMATIN

FORMATION,

AND

methylation is critical for the Polycomb silencing system, which operates in euchromatic domains. Relatively little overlap has been observed between the distribution patterns, and functional roles, of PC and HPl. How H3K27 methylation might fit into the HP1-dependent heterochromatin complexes remains to be elucidated. The HP1

DNA

H3K9me2

GENE

SILENCING

DROSOPHILA

protein has a central linker function in heterochromatin formation and the associated gene silencing, binding H3K9me2 and H3K9me3, and interacting directly with SU(VAR)3-9 (the H3K9 HKMT), SUY4-20 (the H4K20 HKMT), and several additional proteins. Given the number of identified Su(var) loci, the model is certain to become more complex. In mammals and plants, histone H3K9 methylation and DNA methylation represent interrelated marks of repressed chromatin (Martienssen and Colot 2001; Bird 2002). Whether or not DNA methylation occurs at all in Drosophila has been a point of contention for many years. Recent reports of low levels of DNA methylation in the early embryo have renewed this discussion (Kunert et al. 2003). Analysis of the genome indicates that the only recognizable DNA methyltransferase present is Dnmt2. Mutations in this gene have little impact on the organism. Nonetheless, a role in early embryogenesis cannot be ruled out.

5 Chromosomal Proteins Form Mutually Dependent Complexes to Maintain and Spread Heterochromatic Structure

Figure 5. Interaction of SU(VAR)3-9 and HPl in Setting the Distribution Pattern of H3K9me (a) SU(VAR)3-9 is responsible for dimethylation of H3K9 (H3K9me2); loss of enzyme results in loss of this modification in the pericentromeric heterochromatin, as shown by loss of antibody staining of the polytene chromosomes (compare middle panel with top panel). Loss of HPl results in a loss of targeting of SU(VAR)3-9; high levels of H3K9me2 are now seen throughout the chromosome arms (bottom panel). (b) HPl interacts with H3K9me2 through its chromodomain, and with 5U(VAR)3-9 through its chromoshadow domain. By recognizing both the histone modification and the enzyme responsible for that modification, HPl provides a mechanism for heterochromatin spreading and epigenetic inheritance.

PEV reflects a change in gene expression, specifically a loss in expression of the reporter gene in some of the cells in which it is normally active, as a consequence of a genetic rearrangement. Several different models, not all mutually exclusive, have been suggested to explain PEY. One possibility originally considered was the random loss of the gene, perhaps as a consequence of late replication (Karpen and Spradling 1990). Quantitative Southern blot analysis has shown that this explanation is not generally applicable; variegating genes are generally fully replicated in diploid tissue (Wallrath et al. 1996). Other models have focused on the association of the variegating gene with a heterochromatic compartment in the nucleus, and/or on the spreading of heterochromatic structure from the newly adjacent heterochromatin. The spreading model, which is based on extensive genetic and cytological data, explains gene silencing as a consequence of heterochromatin packaging spreading across the breakpoint into normally euchromatic domains. In normal chromosomes, euchromatic and heterochromatic regions appear to be insulated from each other by specific sequences or buffer zones. Because these "insulating sequences" (never well-defined in Drosophila) are not present at the euchromatic-heterochromatic junction in PEV rearrangements (see Fig. 1), heterochromatinization of euchromatic sequences is variably induced. This heterochromatinization is cytologically visible in the polytene chromosomes

92 ,. C HAP T E R 5

as a shift from a banded to an amorphous structure at the base of the chromosome arms (Hartmann-Goldstein 1967); the extent of this change can be modified by Su(var) and E(var) mutations (Reuter et al. 1982). Inactivation of euchromatic genes over a distance along the chromosome can be genetically demonstrated (Demerec and Slizynska 1937). The affected regions become associated with HP1 (Belyaeva et al. 1993) and show H3K9me2, a typical mark of Drosophila heterochromatin (Ebert et al. 2004). Because the spreading model postulates a competition between packaging into euchromatin versus packaging into heterochromatin, PEV modifier genes could encode functions controlling either heterochromatin formation or euchromatin formation. The recovery of dosage-dependent modifiers, as discussed above, supports such a model (Locke et al. 1988; Henikoff 1996). Recently, Su(var) mutations controlling the balance between euchromatin and heterochromatin have indeed been identified (Ebert et al. 2004). PEV rearrangements have allowed us to visualize and study cases where heterochromatin packaging spreads into the flanking euchromatin domain. The spreading effect clearly depends on a series of molecular reactions within the euchromatic regions. Several histone modifications are known that are

mutually exclusive and that define these alternative chromatin states. Acetylation of H3K9, methylation of H3K4, and phosphorylation of H3SlO are typical marks of active euchromatin, whereas methylation of H3K9, H3K27, and H4K20 is a specific mark of silenced regions. Heterochromatinization of euchromatic regions therefore requires specific deacetylation, demethylation, and dephosphorylation reactions within euchromatin, as illustrated in Figure 6. This transition depends initially on H3K9 deacetylation by HDACl. Mutations in the rpd3 gene, encoding the histone H3K9-specific deacetylase HDAC1, are strong suppressors of PEV (Mottus et al. 2000), antagonizing the effect of SU(VAR)3-9 in gene silencing (Czermin et al. 2001). HDAC1 has been shown to be associated in vivo with the SU(VAR)3-9/HP1 complex; the two enzymes work cooperatively to methylate pre-acetylated histones. It has recently been observed that spreading of heterochromatin into euchromatin is completely blocked in Su(var)3-1 mutations (Ebert et al. 2004). Su(var)3-1 mutations are frameshift mutations within the gene encoding JIll kinase that result in expression of a truncated JIll protein, lacking the carboxy-terminal region. The JIll protein contains two kinase domains and catalyzes H3SlO phosphorylation in euchromatin. The fILls,,!v.r)]-1 muta-

b Removal of "active" histone marks Euchromatin

Addition of "repressive" histone marks

Heterochromatin

Euchromatin

Heterochromatin

H3 H3

Figure 6. The Transition from a Euchromatic State to a Heterochromatic State Requires a Series of Changes in Histone Modification (a) Active genes are marked by H3K4me2 and me3; if present, this mark must presumably be removed by LSDl (not yet characterized in Drosophila). H3K9 is normally acetylated in euchromatin; this mark must be removed by a histone deacetylase, HDAC1. Phosphorylation of H3S1 can interfere with methylation of H3K9; dephosphorylation appears to involve a phosphatase targeted by interaction with the carboxyl terminus of the ]IL1 kinase. These transitions set the stage for acquisition of the modifications associated with silencing, shown in b, including methylation of H3K9 by SU(VAR)3-9, binding of HP1, and subsequent methylation of H4K20 by SUV4-20, an enzyme recruited by HP1. Methylation of H3K27 by E(Z) may also occur.

PEV,

HETEROCHROMATIN

FORMATION,

AND

tions do not affect H3SlO phosphorylation, but probably impair dephosphorylation of H3S10, effectively inhibiting methylation ofH3K9. This suggests involvement of a phosphatase. Whether the PP1 enzyme (which has been identified with Survar)3-5 mutations) (Baksa et a1. 1993) is directly involved in this reaction is not known. Demethylation of H3K4 appears to be another prerequisite for heterochromatinization of euchromatic regions. Recent work has shown that the LSD1 aminoxidase functions in mammalian systems as an H3K4 demethylase (Shi et al. 2005). The putative Drosophila LSD1 hom*olog SU(VAR)3-3 facilitates spreading of heterochromatin into euchromatic regions in all PEV rearrangements tested (S. Lein et a1., unpub1.). Consistent with this, in Su(var)3-3 null cells, lacking LSD1, the acquisition of H3K9 methylation in the euchromatin flanking a breakpoint is eliminated, although constitutively heterochromatic regions are not affected. These findings demonstrate that the coordinated function of several enzymes is required to remove euchromatin-specific histone modification marks before the transition to heterochromatin packaging can take place (see Fig. 6). It seems likely that the required enzymes will be found to form complexes with SU(VAR)3-9/HP1, as has already been shown for HDACl. 6 How Is Heterochromatin Formation Targeted in Drosophila? Although we have learned about many mechanistic aspects and the biochemistry of heterochromatin structure, as discussed above, this leaves open the question of how heterochromatin formation is targeted to selected regions of the genome in its normal configuration. All heterochromatic domains share certain features, and two of these features have been considered as essential inputs for assembling heterochromatin on a given DNA sequence: the position of the locus relative to spatially distinct subdomains of heterochromatin in the nucleus, and the presence of repetitious DNA. In general, heterochromatic masses are seen at the nuclear periphery and around the nucleolus. In Drosophila embryos, this tendency is even more pronounced. Heterochromatic masses are first seen in early embryogenesis, as the nuclei move to the periphery of the egg. Early development in Drosophila is syncitial until nuclear division cycle 14, when cell walls form between the nuclei, creating the typical blastula, a ball of cells. The heterochromatic material (centromeres, chromosome four) is concentrated at one side of the nucleus, oriented to the exterior surface of the egg (Foe and Alberts 1985).

GENE

SILENCING

DROSOPHILA

Such spatial subdivision of the nucleus persists during development, leading to the concept of heterochromatin "compartments" within the nucleus. These compart.ments might maintain a high concentration of factors required for heterochromatin formation (such as HP1 and HKMTs) , while being depleted in factors required for euchromatin assembly and gene expression (such as HATs and RNA pol II). Indeed, proximity to heterochromatic masses, both in position along the chromosome and in three dimensions, has been shown to be a factor in PEY. Proximity to the mass of centric heterochromatin has been shown to have an impact on variegation both for euchromatic genes (of which white is an example, see above), and for heterochromatic genes, the best-studied examples being light and rolled. Heterochromatic genes, mapped to those domains, can be observed to variegate when a rearrangement places them in juxtaposition with euchromatin; generally they show the opposite dependencies, requiring normal levels of HP1 for full expression, and showing an enhancement of variegation when HP1 is depleted. Variegation of light depends not only on its juxtaposition to euchromatin, but also on the position of the breakpoint, specifically on the distance from heterochromatin measuring along the chromosome arm (Wakimoto and Hearn 1990). Similar results have been reported for rolled. Investigations of brown dominant (bwD), a euchromatic gene induced to variegate by insertion of repetitious DNA, have shown that a shift in proximity of the locus to the centric heterochromatin can result in enhancement of silencing (if closer) or suppression of silencing (if farther away) (Henikoff et al. 1995). Similarly, translocation of a fourth chromosome carrying a white reporter to the distal half of chromosome arm 2L or 2R results in a dramatic loss of silencing; this was correlated with a change in nuclear disposition, to frequent occupancy of sites distant from the chromocenter in the salivary gland nucleus (Cryderman et a1. 1999a). A recent study using high-resolution microscopy examined both gene activity (using antibodies specific for the product) and nuclear location of a reporter (using FISH, fluorescence in situ hybridization) in the same cell during the normal time frame of expression. A white variegating inversion, bwD , and a variegating lacZ transgene were studied in differentiating eye discs or adult eyes. This investigation found a strong correlation between the position of the reporter gene in the cell nucleus relative to pericentromeric heterochromatin and the level of expression, supporting the idea that a heterochromatic "compartment" exists, and that positioning within this

C HAP T E R 5

compartment is correlated with gene silencing (Harmon and Sedat 2005). However, the correlation is not absolute. This is not surprising, given that studies with a white reporter indicate the presence of both euchromatic and heterochromatic domains interspersed on the small fourth chromosome (which is always close to the mass of pericentromeric heterochromatin). These latter observations point to other local determinants that contribute to packaging chromatin into one form or the other. In D. melanogaster, it is estimated that one-third of the genome is heterochromatic by cytological criteria. This includes large blocks that flank the centromeres, smaller blocks associated with the telomeres, the whole of the Y chromosome, and most of the small fourth chromosome. The centromeric regions are made up of large (0.2-1 Mb) blocks of satellite DNA interspersed with "islands" of complex sequences, generally transposable elements (Le et al. 1995). Although gene-poor, these regions are not devoid of genes; the current estimate is that several hundred genes reside in the pericentromeric heterochromatin (Hoskins et al. 2002). The telomeres of Drosophila do not have the typical G-rich repeats seen elsewhere, but are composed of copies of HeT-A and TART retrotransposons. Telomere-associated sequences (TAS), blocks of 10 2-10 3 nucleotide repeats, are found just proximal, and white transgene reporters inserted in these regions display a variegating phenotype. Although the Y chromosome does carry the genes for a number of male fertility factors, the bulk of the chromosome is made up of satellite DNA, and it remains condensed in cells other than the male germ line. The small fourth chromosome is on the order of 4.3 Mb in size, with about 3 Mb made up of satellite DNA. The distal 1.2 Mb can be considered euchromatic in that it is polytenized in the salivary gland (see Fig. 4), but it appears heterochromatic by virtue of its late replication, its complete lack of meiotic exchange, and its association with HP1, HP2, and H3K9me2 (Fig. 4). This region has a six- to sevenfold higher density of transposon fragments than is found in the euchromatic arms, similar to regions at the junction of centric heterochromatin and euchromatin on the other chromosomes (Kaminker et al. 2002). Interestingly, an investigation of the fourth chromosome using the white reporter P element discussed above (Fig. 3a) found both euchromatic domains (resulting in a red-eye phenotype) and heterochromatic domains (resulting in a variegating phenotype) interspersed (Sun et al. 2004). This finding suggests the presence of local elements in the DNA that can drive the formation of heterochromatin or euchromatin. Genetic screens for a switch in phenotype

(from red to variegating or vice versa) have demonstrated that local deletions or duplications of 5-80 kb of DNA flanking a transposon reporter can lead to the loss or acquisition of variegation, pointing to short-range cis-acting determinants for silencing (see Fig. 7). This silencing is dependent on HPl and correlates with a change in chromatin structure, as shown by a change in nuclease accessibility, pointing to a shift from a euchromatic to a heterochromatic state. Mapping data in one region of the fourth implicate the 1360 transposon as a target for heterochromatin formation and indicate that once heterochromatin formation is initiated at dispersed repetitive elements, it can spread along the fourth chromosome for about 10 kb, or until it encounters competition from a euchromatic determinant (Sun et al. 2004). Short-range cis-acting determinants related to copy number are also implied by the observation that tandem or inverted repeats of reporter P elements will result in heterochromatin formation and gene silencing (Dorer and Henikoff 1994). Such cis-acting elements in the DNA might function by sequence-specific binding of a protein capable of triggering heterochromatin formation. Proteins that bind specifically to some of the satellite DNAs have been identified (e.g., Dl, Aulner et al. 2002). The importance of these interactions has been inferred from the impact of satellite-specific DNA-binding drugs, which can suppress PEV (Janssen et al. 2000). However, the findings in yeast and plants (Elgin and Grewal 2003; Matzke and Birchler 2005; see Chapters 8 and 9) suggest a second model, specifically that an RNAi-based mechanism could be used to target heterochromatin formation to repetitious elements. Work from several labs has demonstrated that the RNAi system is present in Drosophila and plays an important role in developmental regulation via posttranscriptional gene silencing (PTGS). D. melanogaster has two genes encoding DICER proteins and numerous genes (aubergine, AG01, AG02, spindleE [aka homeless], vasa intronic gene [VIG], armitage, Fmrl) encoding components or proteins required for assembly of the RNA-induced silencing complex (RISC) (Sontheimer 2005). The system has been implicated in the PTGS of repetitious sequences, notably the tandemly repeated Stellate genes, several retrotransposons, and Alcohol dehydrogenase (Adh) transgenes, and in the transcriptional gene silencing (TGS) of Adh transgenes (Aravin et al. 2001; Pal- Bhadra et al. 2002). In a direct test, Pal-Bhadra et al. (2004) found that mutations in piwi (a member of the PAZ domain family) and homeless (a DEAD box helicase) suppress the PEV associated with tandem arrays

P E V,

HE T ERa C H ROM A TIN

Heterochromatin

FOR MAT ION,

AND

G ENE

Euchromatin

S I LEN C I N GIN

0 R 0 SOP H I L A .

Heterochromatin

Mutant Lines 39C-12

____________________V

7..-.. M.2... 01 4-M382 4-M325 5-M340

V 4~ y;-------Figure 7. Possible Model for Heterochromatin Targeting dsRNA from repetitious sequences is processed through RiSe to generate a hypothetical "targeting complex," which directs either histone modification or HPl association as an initial step in assembling heterochromatin at the site identified by the small ssRNA. Data from the fourth chromosome suggest that 7360 DNA transposon fragments (orange bars) are a target for heterochromatin formation; local deletions or duplications that shift the position of the P element reporter (triangle) away from a 7360 element lead to loss of silencing (red triangle indicates a red eye), whereas proximity to 7360 leads to silencing (dotted triangle indicates a variegating eye). (Based on data in Sun et al 2004.)

of the white gene, and that mutations in piwi, aubergine, and homeless suppress silencing of the white transgene P[hsp70-w] in pericentromeric heterochromatin or the fourth chromosome. This suppression of PEV was associated with a significant decrease in the levels of H3K9 methylation. Repeat-associated small interfering RNAs (rasiRNAs) have been identified from 40% of the known transposable elements (including 1360) and other repeated sequences (Aravin et al. 2003). Put together, the results discussed above suggest that heterochromatin formation may be dependent both on nuclear location (perhaps providing an abundant pool of required proteins) and on specific targeting based on RNAi recognition and processing of double-stranded RNA from repetitious elements, particularly some of the DNA transposons. Such targeting via a RISe could bring either a histone H3 methyltransferase or a complex including HP 1 (or both) to a site to trigger the assembly process discussed in Section 4.

7 Not All Heterochromatin Is Identical

Although heterochromatin has been described above in general terms, it is clear that heterochromatic domains vary in detail. All heterochromatic domains are characterized by repetitious DNA (see above), but this can vary from a tandem array of short repeats (satellite DNA) found in blocks in centromeric regions, to a high density of interspersed repetitious sequences, as seen on the fourth chromosome. Whereas all heterochromatic regions appear to be associated with HPI and H3K9me2, it is clear that the protein complexes involved must differ in other ways. Examination of the impact of 70 different modifiers on different variegating genes (including W"'4, bwD, P-element reporters in pericentromeric heterochromatin or in a TAS array) showed that whereas there is substantial overlap in the targets of modifiers, there is also surprising complexity. This set of tests divided the modifiers into seven different groups in terms of their ability to affect

96 •

C HAP T E R 5

silencing in a given compartment (Donaldson et ai. 2002). Interestingly, the only modifer in this group to affect silencing in the TAS array was a new allele of Su(var)3-9. These differences no doubt reflect changes in the local biochemistry, or in the enzymes used to achieve it. For example, cytological results indicate that whereas H3K9me2 is highly concentrated along the fourth chromosome, the enzyme responsible is not SU(VAR)3-9 (Schotta et ai. 2002; K.A. Haynes et aI., unpubI.). Even within the pericentromeric heterochromatin, one should anticipate a mosaic, given the differences in the underlying blocks of DNA, which vary from satellite DNA to clusters of interspersed repeats (Le et ai. 1995), which might utilize a different mix of heterochromatin proteins. The consequences have been seen in studies that examine the impact of different blocks of pericentromeric heterochromatin on expression from a reporter, where one can observe that the severity of the phenotype does not depend simply on the amount of heterochromatin in cis, but varies depending on the local heterochromatin environment (Howe et ai. 1995). Heterochromatin-associated proteins that might playa role in specific subdomains include the AT-hook protein D1, preferentially associated with the 1.688 g/cm3 satellite III (Aulner et al. 2002), and DDP1, a multi-KH-domain protein hom*ologous to vigilin that binds the pyrimidine-rich C strand of the dodeca satellite (Cortes and Azorin 2000). 8 PEV, Heterochromatin Formation, and Gene Silencing in Different Organisms

The phenomenon of position-effect variegation was initially detected in Drosophila, simply because this was one of the first organisms for which X irradiation was used to induce mutations. X irradiation is much more likely than other commonly used mutagens to induce chromosomal rearrangements, which can result in PEY. Similar mutations have been isolated from the mouse, where variegating coat color indicates PEY. Genetic analysis revealed an insertion of the autosomal region carrying wild-type alleles of the fur-color genes into the X chromosome (Cattanach 1961; Russel and Bangham 1961). Variegation is only observed in females carrying this insertion combined with a hom*ozygous mutation in the original coat-color genes. In these females, the wild-type allele becomes inactivated as a consequence of X inactivation by heterochromatinization (see Chapter 17). In plants, the only unequivocal case of PEV that has been described was reported in Oenothera blandina (Catcheside 1939). In these cases, as in Drosophila, PEV silencing of euchro-

matic genes is connected with placement of those genes into a new heterochromatic neighborhood. Transcriptional gene silencing has also been observed for repeated sequences (RIGS; repeat-induced gene silencing), particularly in plants. Analysis of the affected sequences has revealed the appearance of similar epigenetic marks (histone and DNA methylation) as found in heterochromatin and in regions silenced by PEY. If DNA fragments containing tandemly arranged luciferase genes are introduced into Arabidopsis, variegated luciferase expression is seen. Again, heterochromatin formation is responsible for the gene silencing observed. The underlying molecular mechanisms are conserved in higher eukaryotic organisms. A central feature of heterochromatic gene silencing in Drosophila is the interaction of HP1 with H3K9me2 and the SU(VAR)3-9 HKMT. HP1 is conserved from the yeast Schizosaccharomyces pombe to man, and is consistently associated with pericentromeric heterochromatin. The human HP 1 genes can be used to rescue the deficiency in Drosophila (Ma et al. 2001). However, HP1 has not been identified in plants as such. SU(VAR)3-9 is even more widely represented, having been identified in fission yeast (Clr4), Neurospora (DIMS), Arabidopsis, and mammals (SUV39H). All of the SU(VAR)3-9 hom*ologs catalyze H3K9 methylation and function in heterochromatin formation. Again, a human SUV39H1 transgene can completely compensate for the loss of the endogenous Drosophila protein in mutant lines (Schotta et ai. 2002). In higher plants (rice, Arabidopsis, and maize), several SU(VAR)3-9 hom*ologous proteins (SUVH) are found (Baumbusch et al. 2001). The high number of HKMTs might reflect the plasticity of plant development or the need to respond to environmental factors (see Chapter 9 for further discussion). Four SUVH proteins, SUVHl, SUVH2, SUVH4 (KYP), and SUVH6, have been studied in detail (Jackson et ai. 2002; Naumann et al. 2005). All are histone H3K9 methyltransferases. SUVH2 plays a pivotal role in control of heterochromatin states, exhibiting dosage-dependent effects on heterochromatin formation similar to those reported for Drosophila SU(VAR)3-9 (Naumann et al. 2005). SUVH2 loss of function strongly suppresses repeat-dependent silencing, and overexpression causes significant enhancement of such silencing in plants with luciferase transgenes. Other genes identified by Drosophila Su(var) mutations encode proteins with conserved functions. The SUV4-20 HKMT has been characterized in mammals and in Drosophila (Schotta et al. 2004). In both organisms, it controls trimethylation of H4K20. Histone demethylases,

PEV,

HETEROCHROMATIN

FORMATION,

acetylases, and deacetylases are also conserved (T. Rudolph et al., unpubl.). The evolutionary conservation of many of the key enzymes controlling histone modification supports the idea of a histone code (Jenuwein and Allis 2001). However, examination of the heterochromatin-specific histone modification marks observed in Drosophila, mammals, and plants (Arabidopsis) also identifies some genus-specific elements. Significant hallmarks of constitutive heterochromatin in mammals include H3K9me3, H3K27mel, and H4K20me3 (Peters et al. 2003; Rice et al. 2003; Schotta et al. 2004). Drosophila heterochromatin is characterized by H3K9mel/me2, H3K27mel/me2/me3, and H4K20me3 (Schotta et al. 2002,2004; Ebert et al. 2004). In contrast to mammals, H3K9me3 is underrepresented in Drosophila. In mammals, H3K9mel is not a heterochromatic mark. In Arabidopsis, as in Drosophila, H3K9mel/me2 are heterochromatic marks, whereas H3K9me3 is euchromatic (Naumann et al. 2005). H3K27mel and H3K27me2 are heterochromatic marks in Arabidopsis, whereas these marks in Drosophila are found in euchromatin and heterochromatin. H3K27me3 is exclusively euchromatic in Arabidopsis. H4K20mel in Arabidopsis is heterochromatic, but H4K20me2 and H4K20me3 are euchromatic. Another striking difference between Arabidopsis and animals concerns the chromosomal distribution of H3SlO phosphorylation. This mark is heterochromatic in Arabidopsis (A. Fischer and G. Reuter, unpubl.) but euchromatic in Drosophila (Wang et al. 2001; Ebert et al. 2004). Similarities and differences in heterochromatin-specific histone modification marks between mammals, Drosophila, and Arabidopsis clearly indicate that the histone code is not completely universal, but rather exists in different dialects. 9 Summing Up: There Is Much That We Do Not Know about Heterochromatin

Although PEV has provided us with an extraordinary opportunity to study heterochromatin formation and gene silencing, the phenotype itself remains puzzling. Why do we observe a variegating pattern of silencing? What tips the balance, leading to a switch from the active to the silent state, or vice versa? Why does this appear to be clonally inherited? PEV is generally analyzed as a problem of maintaining the reporter gene "ON" or "OFF," but in many instances (particularly when using P-element-based reporters), one observes red facets on a yellow or pale orange background, suggesting that gene expression has been reduced uni-

AND

GENE

SILENCING

DROSOPHILA.

formly, but that that down-regulation has been lost in some cells. Careful analysis of such lines might lead to identification of chromatin states with an intermediate impact on gene expression. Although the data support a crude model for loss or maintenance of silencing based on mass action, the final model will be complex, involving numerous interacting proteins (see, e.g., the proposal by Henikoff 1996). One is tempted to consider the nucleosome as a summation device, collecting modifications and displaying the results in terms of both particular protein-binding patterns and facility for remodeling in that region. The chromatin state might then reflect the results of competition for achieving different modifications. Such a model could be useful in sorting out the effects noted above. It is also compatible with observations demonstrating that the frequency of silencing of a GAL4-dependent reporter is sensitive to GAL4levels (Ahmad and Henikoff 2001). The RNAi system provides a plausible mechanism for targeting heterochromatin formation, presumably by targeting a complex including HP1, an H3K9 HKMT, or both. However, many questions remain. What is the source of the dsRNA? Must it be produced in cis (as implied by the results in S. pombe), or can it operate in trans (as suggested by results in plants); i.e., can the production of dsRNA from one 1360 site result in targeting of all 1360 sites? Are all repetitious elements potential targets? This seems unlikely from the fourth-chromosome analysis described above. If a subset of repetitious elements plays a key role, what determines that choice? The results obtained on the fourth chromosome argue that the density and distribution of critical repetitious elements will affect expression of the genes in the vicinity. This argues for the need to ascertain this characteristic when sequencing a genome. How is spreading of heterochromatin accomplished, and what are the normal barriers to spreading? Note that there is no evidence for transitive RNAi in Drosophila; i.e., the spread of silencing to targets in a transcript that lie upstream of the dsRNA sequence (Celotto and Gravely 2002). This is in congruence with the lack of evidence for any RNA-dependent RNA polymerase in this system. An assembly system based on the interactions of HP1, H3K9me2, and an HKMT might well account for the spread of heterochromatin for approximately 10 kb, as observed on the fourth chromosome; this type of spreading could be limited by a site of histone acetylation. But what about the spreading that occurs in rearrangements, which has been found to extend for hundreds of kilobases? This form of spreading is not contiguous, but again

98 •

C HAP T E R 5

appears to depend critically on chromatin proteins, notably JIL-I, in a role that does not depend on its kinase activity. These and other questions remain unanswered. Acknowledgments

We thank Gabriella Farkas for creating the figures used here, Anja Ebert for immunocytological photos, and the members of our research groups for a critical review of this chapter. Our work is supported by Deutsche Forschungsgemeinschaft and the Epigenome Network of Excellence of the European Union (G.R.) and by grants from the National Institutes of Health (S.C.R.E.).

References Ahmad K. and Henikoff S. 2001. Modulation of a transcription factor counteracts heterochromatin gene silencing in Drosophila. Cell 104: 839-847. Aravin A.A., Numova H. M., Tulin A. V, Vagin, VV., Rozovsky YM., and Gvozdev VA. 2001. Double-stranded RNA-mediated silencing of genomic tandem repeats and transposable elements in the D. melanogaster germline. Curro BioI. 11: 1017-1027. Aravin A. A., Lagos-Quintana M., Yalcin A., Zavolan M., Marks D., Snyder B., Gaasterland T, Meyer J., and Tuschl T 2003. The small RNA profile during Drosophila melanogaster development. Dev. Cell 5: 337-350. Ashburner M., Golic K.G., and Hawley R.S. 200Sa. Chromosomes. In Drosophila: A laboratory handbook, 2nd edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, pp. 24-57. - - - . 2005b. Position effect variegation. In Drosophila: A laboratory handbook, 2nd edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, pp.l007-1049. Aulner N., Monod C, Mandicourt G., Jullien D., Cuvier 0., Sall A., Janssen S., LaemmJi U.K, and Kas E. 2002. The AT-hook protein Dl is essential for Drosophila melanogaster development and is implicated in position-effect variegation. Mol. Cell. Bioi. 22: 1218-1232. Baksa K., Morawietz H., Dombradi V., Axton M., Taubert H., Szabo G., Torok I., Gyurkovics H., Szoor B., Gloover D., et aI. 1993. Mutations in the phosphatase 1 gene at 87B can differentially affect suppression of position-effect variegation and mitosis in Drosophila melanogaster. Genetics 135: 117-125. Bannister A.J., Zegermann P., Patridge J.E, Miska EA, Thomas J.O., Allshire TC, and Kouzarides T 2001. Selective recognition of methylated lysine 9 on histone H3 by the HPI chromo domain. Nature 410: 120-124. Baumbusch L.O., Thorstensen T, Krauss V, Fischer A., Naumann K., Assalkhou R., Schulz I., Reuter G., and Aalen R. 2001. The Arabidopsis thaliana genome contains at least 29 active genes encoding SET domain proteins that can be assigned to four evolutionary conserved classes. Nucleic Acids Res. 29: 4319-4333. Belyaeva E.S., Demakova O.V., Umbetova G.H., and Zhimulev I.E 1993. Cytogenetic and molecular aspects of position-effect variegation in Drosophila melanogaster. V. Heterochromatin-associated protein HP 1 appears in euchromatic chromosomal regions that are inactivated as a result of position-effect variegation. Chromosoma 102: 583-590. Bender w., Spierer P., and Hogness D.S. 1983. Chromosome walking and

jumping to isolate DNA from the Ace and rosy loci and the bithorax complex of Drosophila melanogaster. f. Mol. BioI. 168: 17-33. Birchler J.A., Bhadra u., Rabinow L., Linsk R., and Nguyen-Huyuh A.T 1994. Weakener of white (Wow), a gene that modifies the expression of white eye color locus and that suppresses position effect variegation in Drosophila melanogaster. Genetics 137: 1057-1070. Bird A. 2002. DNA methylation patterns and epigenetic memory. Genes Dev. 16: 6-21. Bushey D. and Locke J. 2004. Mutations in Su(var)205 and Su(var)3-7 suppress P-element-dependent silencing in Drosophila melanogaster. Genetics 168: 1395-1411. Catcheside D.G. 1939. A position effect in Oenothera. ]. Genet. 38: 345-352. Cattanach B.M. 1961. A chemically-induced variegated-type position effect in the mouse. Z. Vererbungsl. 92: 165-182. Celotto A.M. and Graveley B.R. 2002 Exon-specific RNAi: A tool for dissecting the functional relevance of alternative splicing. RNA 8: 718-724. Cleard E and Spierer P. 2001. Position-effect variegation in Drosophila: The modifier Su(var)3-7 is a modular DNA-binding protein. EMBO Rep. 21: 1095-1100. Cleard E, Delattre M., and Spierer P. 1997. SU(VAR)3-7 a Drosophila heterochromatin-associated protein and companion of HPI in the genomic silencing of position-effect variegation. EMBO ]. 16: 5280-5288. Cooper K.W. 1959. Cytogenetic analysis of major heterochromatic elements (especially Xh and Y) in Drosophila melanogaster and the theory of "heterochromatin". Chromosoma 10: 535-588. Cortes A. and Azorin E 2000. DDPl, a heterochromatin-asociated multiKH-domain protein of Drosophila melanogaster, interacts specifically with centromeric satellite DNA sequences. Mol. Cell. BioI. 20: 3860-3869. Cryderman D.E., Morris E.J., Biessmann H., Elgin S.CR., and Wallrath L.L. 1999a. Silencing at Drosophila telomeres: Nuclear organization and chromatin structure play critical roles. EMBO f. 18: 3724-3735. Cryderman D. E., Tang H., Bell C, Gilmour D.S., and Wallrath L.L. 1999b. Heterochromatic silencing of Drosophila heat shock genes acts at the level of promoter potentiation. Nucleic Acids Res. 27: 3364-3370. Czermin B., Schotta G., Hi.ilsmann B.B., Brehm A., Becker P.B., Reuter G., and Imhof A. 2001. Physical and functional interaction of SU(VAR)3-9 and HDACI in Drosophila. EMBO Rep. 2: 915-919. Delattre M., Spierer A., Tonka CH., and Spierer P. 2000. The genomic silencing of position-effect variegation in Drosophila melanogaster: Interaction between the heterochromatin-associated proteins Su(var)3-7 and HPJ. f. Cell Sci. 113: 4253-4261. Demerec M. and Slizynska H. 1937. Mottled white 258-18 of Drosophila melanogaster. Genetics 22: 641-649. Donaldson K.M., Lui A., and Karpen G.H. 2002. Modifiers of terminal deficiency-associated position effect variegation in Drosophila. Genetics 160: 995-1009. Dorer D.R. and Henikoff S. 1994. Expansion of transgene repeats causes heterochromatin formation and gene silencing in Drosophila. Cell 77: 993-1002. Dorn R., Krauss V, Reuter G., and Saumweber H. 1993a. The enhancer of position-effect variegation E(var)93D, codes for a chromatin protein containing a conserved domain common to several transcriptional regulators. Proc. Natl. Acad. Sci. 90: 11376-11380. Dorn R., Szidonya J., Korge G., Sehnert M., Taubert H., Archoukieh I., Tschiersch B., Morawietz H., Wustmann G., Hoffmann G., and Reuter G. 1993b. P Transposon-induced dominant enhancer muta-

P E V,

H E T E ROC H ROM A TIN

FOR MAT ION,

AND

tions of position-effect variegation in Drosophila melanogaster. Genetics 133: 279-290. Ebert A., Schotta G., Lein S., Kubicek S., Krauss Y., Jenuwein T., and Reuter G. 2004. Su(var) genes regulate the balance between euchromatin and heterochromatin in Drosophila. Genes Dev. 18: 2973-2983. Eissenberg J.C, Morris G.D., Reuter G., and Hartnett T. 1992. The hetero-chromatin-associated protein HP-1 is an essential protein in Drosophila with dosage-dependent effects on position-effect variegation. Genetics 131: 345-352. Eissenberg J.C, James T.C, Foster-Hartnett D.M., Hartnett T., gan Y., and Elgin S.CR. 1990. A mutation in a heterochromatin-specific chromosomal protein is associated with suppression of position effect variegation in Drosophila melanogaster. Proc. Natl. Acad. Sci. 87: 9923-9927. Elgin S.CR. and Grewal S.l.S. 2003. Heterochromatin: Silence is Golden. Curro BioI. 13: R895-R898. Fanti 1. and PimpineLJi S. 2004. Immunostaining of squash preparations of chromosomes of larval brains. Methods Mol. BioI. 247: 353-361. Fanti 1., Giovinazzo G., Berloco M., and Pimpinelli S. 1998. The heterochromatin protein 1 prevents telomere fusions in Drosophila. Mol. Cell 2. 527-538. Farkas G., Gausz J., GaLJoni M., Reuter G., Gyurkovics H., and Krach F. 1994. The trithorax-like gene encodes the Drosophila GAGA factor. Nature 371: 806-808. Fischle w., Wang Y, Jacobs S.A., Kim Y, Allis CD., and Khorasanizadeh S. 2003. Molecular basis for the discrimination of repressive methyl-lysine marks in histone H3 by Polycomb and HP1 chromodo mains. Genes Dev. 17: 1870-1881. Flybase 2005. The Drosophila database. Available from World Wide Web at the URLs http://morgan/harvard.edu and http://www.ebi.ac.ukIflybase/ Foe V.E. and Alberts B.M. 1985. Reversible chromosome condensation induced in Drosophila embryos by anoxia: Visualization of interphase nuclear organization. f. Cell BioI. 100: 1623-1636. Harmon B. and Sedat J. 2005. Cell-by-cell dissection of gene expression and chromosomal interactions reveals consequences of nuclear reorganization. PLoS BioI. 3: e67. Hartmann-Goldstein l.J. 1967. On the relationship between heterochromatization and variegation in Drosophila, with special reference to temperature sensitive periods. Genet. Res. 10: 143-159. Henikoff S. 1996. Dosage-dependent modification of position-effect variegation in Drosophila. BioEssays 18: 401-409. Henikoff S., Jackson J.M., and Talbert P.B. 1995. Distance and pairing effects on the brown D",,,;,,,,,,, heterochromatic element in Drosophila. Genetics 140: 1007-1017. Hoskins R.A., Smith CD., Carlson J.w., Carvalho A.B., Halpern A., Kaminker J.S., Kennedey C, Mungall CJ., Sullivan B.A., Sutton G.G., et al. 2002. Heterochromatic sequences in a Drosophila whole-genome shotgun assembly. Genome Bioi. 3: RESEARCH0085. Howe M., Dimitri E, Berloco M., and Wakimoto B.T. 1995. cis-effects of heterochromatin on heterochromatic and euchromatic gene activity in Drosophila melanogaster. Genetics. 140: 1033-1045. Jackson J.P., Lindroth A.M., Cao X., and Jacobsen S.E. 2002. Control of Cp pG D A methylation by the KRYPONITE histone H3 methyltransferase. Nature 416: 556-560. James T.C and Elgin S.CR. 1986. Identification of a nonhistone chromosomal protein associated with heterochromatin in Drosophila melanogaster and its gene. Mol. Cell BioI. 6: 3862-3872. James T.C, Eissenberg J.C, Craig C, Dietrich V., Hobson A., and Elgin

G ENE

5 I LEN C I N GIN

0 R 0 SOP H I L A

S.CR. 1989. Distribution patterns of HP1, a heterochromatinassociated nonhistone chromosomal protein of Drosophila. Eur. f. Cell Bioi. 50: 170-180. Janssen S., Cuvier 0., Muller M., and Laemmli U.K. 2000. Specific gainand loss-of-function phenotypes induced by satellite-specific DNA-binding drugs fed to Drosophila melanogaster. Mol. Cell 6: 1013-1024. Jenuwein T. and Allis CD. 2001. Translating the histone code. Science 293: 1074-1080. Jones R.S. and Gelbart W.M. 1993. The Drosophila Polycomb-group gene Enhancer ofzeste contains a region with sequence similarity to trithorax. Mol. Cell. BioI. 13: 6357-6366. Kaminker J.S. Bergman CM., Kronmiller B., Carlson J., Svirskas R., Patel S., Frise E., Wheeler D.A., Lewis S.E., Rubin G.M., et al. 2002. The transposable elements of the Drosophila melanogaster genome: A genomics perspective. Genome BioI. 3: RESEARCH0084. Karpen G.H. and Spradling A.C 1990. Reduced DNA polytenization of a minichromosome region undergoing position-effect variegation in Drosophila. Cell 63: 97-107. Krauss Y. and Reuter G. 2000. Two genes become one: The genes encoding heterochromatin protein SU(VAR)3-9 and translation initiation factor subunit eIF-2y are joined to a dicistronic unit in holometabolic insects. Genetics 156: 1157-1167. Kunert N., Marhold J., Stanke J., Stach D., and Lyko F. 2003. A Dnmt2Like protein mediates DNA methylation in Drosophila. Development 130: 5083-5090. Lachner M., O'Carroll D., Rea S., Mechtler K., and Jenuwein T. 2001. Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature 410: 116-120. Laible G., Wolf A., Dorn R., Reuter G., islow C, Lebesorger A., Popkin D., Pillus 1., and Jenuwein T. 1997. Mammalian hom*ologues of the Polycomb-group gene Enhancer ofzeste mediate gene silencing in Drosophila heterochromatin and at S. cerevisiae. EMBO f. 16: 3219-3232. Larsson J., Zhang J., and Rasmuson-Lestander A. 1996. Mutations in the Drosophila melanogaster S-adenosyLmethionine synthase suppress position-effect variegation. Genetics 143: 887-896.· Le M.H., Duricka D., and Karpen G.H. 1995. Islands of complex DNA are widespread in Drosophila centric heterochromatin. Genetics 141: 283-303. Locke J., Kotarski M. A., and TartofK.D. 1988. Dosage-dependent modifiers of position effect variegation in Drosophila and a mass action model that explains their effect. Genetics 120: 181-198. Lu B.Y., Bishop CP., and Eissenberg J.C 1996. Developmetal timing and tissue specificity of heterochromatin-mediated silencing. EMBO f. 15: 1323-1332. Ma J., Hwang K.K., Worman H.J., Courvalin J.C, and Eissenberg J.C 2001. Expression and functional analysis of three isoforms of human heterochromatin-associated protein HP 1 in Drosophila. Chromosoma 109: 536-544. Martienssen R.A. and Colot V. 2001. D A methylation and epigenetic inheritance in plants and filamentous fungi. Science 293: 1070-1074. Mason J.M., Ransom J., and Konev A.Y 2004. A deficiency screen for dominant suppressors of telomeric silencing in Drosophila. Genetics 168: 1353-1370. Matzke M.A. and Birchler J.A. 2005. RNAi-mediated pathways in the nucleus. Nat. Rev. Genet. 6: 24-35. Mottus R., Sobels R.E., and Grigliatti T.A. 2000. Mutational analysis of a histone deacetylase in Drosophila melanogaster: Missense mutations suppress gene silencing associated with position effect variegation. Genetics 154: 657-668.

100 •

C HAP T E R 5

Muller H.J. 1930. Types of visible variations induced by X-rays in Drosophila. J. Genet. 22: 299-334. Naumann K, Fischer A., Hofmann I., Krauss V., Phalke S., Irmler K., Hause G., Aurich A.C., Dorn R., Jenuwein T., and Reuter G. 2005. Pivotal role of AtSUVH2 in control of heterochromatic histone methylation and gene silencing in Arabidopsis. EMBO f. 24: 1418-1429. Pal-Bhadra M., Bhadra u., and Birchler J.A. 2002. RNAi related mechanisms affect both transcriptional and post-transcriptional transgene silencing in Drosophila. Mol. Cell 9: 315-327. Pal-Bhadra M., Leibovitch B.A., Gandhi S.G., Rao M., Bhadra u., Birchler J.A., and Elgin S.C.R. 2004. Heterochromatic silencing and HPI localization in Drosophila are dependent on the RNAi machinery. Science 303: 669-672. Paro R and Hogness D.S. 1991. The Polycomb protein shares a hom*ologous domain with a heterochromatin-associated protein of Drosophila. Proc. Natl. Acad. Sci. 88: 263-267. Peters A.H.EM., Kubicek S., Mechtler K., O'Sullivan J., Derijck A.A.H.A., Perez-Burgos 1., Kohlmaier A., Opravil S., Tachibana M., Shinkai Y., et al. 2003. Partitioning and plasticity of repressive histone methylation states in mammalian chromatin. Mol. Cell 12: 1577-1589. Platero J.S., Sharp E.J., Adler EN., and Eissenberg J.c. 1996. In vivo assay for protein-protein interaction using Drosophila chromosomes. Chromosoma 104: 393-404. Rea S., Eisenhaber E, O'Carroll D., Strahl B.D., Sun Z-W., Schmid M., Opravil S., Mechtler K, Ponting c.P., Allis C.D., and Jenuwein T 2000. Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406: 593-599. Reuter G., Werner W., and Hofmann H.J. 1982. Mutants affecting position-effect heterochromatinization in Drosophila melanogaster. Chromosoma 85: 539-551. Reuter G., Wolff I., and Friede B. 1985. Functional properties of the heterochromatic sequences inducing w"" position-effect variegation in Drosophila melanogaster. Chromosoma 93: 132-139. Reuter G., Giarre N., Farah J., Gausz J., Spierer A., and Spierer P. 1990. Dependence of position-effect variegation in Drosophila on dose of a gene encoding an unusual zinc-finger protein. Nature 344: 219-223. Rice J.c., Briggs S.D., Ueberheide B., Barber C.M., Shabanowitz J., HWlt D.E, Shinkai Y., and Allis C.D. 2003. Histone methyltransferases direct different degrees of methylation to define distinct chromatin domains. Mol. Cell 12: 1591-1598. Richards E.J. and Elgin S.C.R 2002. Epigenetic codes for heterochromatin formation and silencing: Rounding up the usual suspects. Cell 108: 489-500. Russel L.B. and Bangham J,W. 1961. Variegated type position effects in the mouse. Genetics 46: 509-525. Schotta G., Ebert A., Krauss v., Fischer A., Hoffmann J., Rea S., Jenuwein T., and Reuter G. 2002. Central role of Drosophila SU(VAR)3-9 in histone H3-K9 methylation and heterochromatic gene silencing. EMBO J. 21: 1121-1131. Schotta G., Ebert A., Dorn R., and Reuter G. 2003. Position-effect variegation and the genetic dissection of chromatin regulation in Drosophila. Semin. Cell Dev. BioI. 14: 67-75. Schotta G., Lachner M., Sarma K., EbertA., Sengupta R, Reuter G., Reinberg D., and Jenuwein T 2004. A silencing pathway to induce H3-K9 and H4-K20 trimethylation at constitutive heterochromatin. Genes Dev. 18: 1251-1262. Schultz J. 1950. Interrelations of factors affecting heterochromatininduced variegation in Drosophila. Genetics 35: 134.

Seum c., Pauli D., Delattre M., Jaquet Y., Spierer A, and Spierer P.2002. Isolation of Su(var)3-7 mutations by hom*ologous recombination in Drosophila melanogaster. Genetics 161: 1125-1136. Shaffer c.D., Stephens G.E., Thompson B.A., Funches 1., Bernat J,A., Craig C.A, and Elgin S.C.R. 2002. Heterochromatin protein 2 (HP2), a partner ofHPI in Drosophila heterochromatin. Proc. Natl. Acad. Sci. 99: 14332-14337. Shi Y., Lan E, Matson c., Mulligan P., Whetstine J.R., Cole EA., Casero RA., and Shi Y. 2005. Histone demethylation mediated by the nuclear amine oxidase hom*olog LSD1. Cell1l9: 941-953. Silver L.M., and Elgin S.C.R. 1976. A method for determination of the in situ distribution of chromosomal proteins. Proc. Natl. Acad. Sci. 73: 423-427. Sinclair DAR, Mottus R.C., and Grigliatti TA 1983. Genes which suppress position effect variegation in Drosophila melanogaster are clustered. Mol. Gen. Genet. 191: 326-333. Sinclair DAR., Clegg N.J, Antonchuk J, Milner TA., Stankunas K, Ruse c., Grigliatti T.A, Kassis J., and Brock H.W. 1998. Enhancer of Polycomb is a suppressor of position-effect variegation in Drosophila melanogaster. Genetics 148: 211-220. Sontheimer E,J. 2005. Assembly and function of RNA silencing complexes. Nat. Rev. Mol. Cell BioI. 6: 127-138. Spofford J.B. 1967. Single-locus modification of position-effect variegation in Drosophila melanogaster. I. white variegation. Genetics 57: 751-766. Stephens G.E., Craig C.A., Li Y., Wallrath L.L., and Elgin S.C.R. 2003. Immunofluorescent staining of polytene chromosomes: Exploiting genetic tools. Methods Enzymol. 376: 372-393. Sun E-L., Cuaycong M.H., and Elgin S.C.R. 2001. Long-range nucleosome ordering is associated with gene silencing in Drosophila melanogaster pericentromeric heterochromatin. Mol. Cell. Bioi. 21: 2867-2879. Sun E-L., Haynes K, Simpson c.L., Lee S.D., Collins 1., WuUer J., Eissenberg J.c., and Elgin S.C.R. 2004. cis-acting determinants of heterochromatic formation on Drosophila melanogaster chromosome four. Mol. Cell. BioI. 24: 8210-8220. Tartof K.D., Hobbs c., and Jolmes M. 1984. A structural basis of variegating position effects. Cell 37: 869-878. Tschiersch B., Hofmal1l1 A., Krauss v., Dorn R., Korge G., and Reuter G. 1994. The protein encoded by the Drosophila position effect variegation suppressor gene Su(var)3-9 combines domains of antagonistic regulators of homeotic gene complexes. EMBO f. 13: 3822-3831. Wakimoto B.T. and Hearn M.G. 1990. The effects of chromosome rearrangements on the expression of heterochromatic genes in chromosome 2L of Drosophila melanogaster. Genetics 125: 141-154. Wallrath L.L. and Elgin S.C.R. 1995. Position effect variegation in Drosophila is associated with an altered chromatin structure. Genes Dev.9: 1263-1277. Wallrath L.L., Gunter v.P., Rosman L.E., and Elgin S.C.R. 1996. DNA representation of variegating heterochromatic P element inserts in diploid and polytene tissue of Drosophila melanogaster. Chromosoma 104: 519-527. Wang Y., Zhang W., Jin Y., Johansen J., and Johansen K.M. 2001. The JIL-l tandem kinase mediates histone H3 phosphorylation and is required for maintenance of chromatin structure in Drosophila. Cell 105: 433-443. Zhimulev I.E, Belyaeva E.S., Fomina OV., Protopopov M.O., and Bolshkov V.N. 1986. Cytogenetic and molecular aspects of position effect variegation in Drosophila melanogaster. Chromosoma 94: 492-504.

Fungal Models for Epigenetic Research: Schizosaccharomyces pombe and Neurospora crassa Robin C. Allshire' and Eric U. Selker

'WeLlcome Trust Centre for Cell Biology, The University of Edinburgh, Edinburgh EH93JR, Scotland, United Kingdom 2Institute of Molecular Biology, University of Oregon, Eugene, Oregon 97403-1229

CONTENTS 1. Schizosaccharomyces pombe: The Organism, 103 7.7

Chromatin Silencing in S. pombe Is Different from That in S. cerevisiae, 703

7.8

Epigenetic Inheritance of the Functional . Centromere State, 770

7.9

Diverse Silencing Mechanisms in Fungi, 777

2. Neurospora crassa: History and Features of the Organism, 112

7.2

Genes Placed in Fission Yeast Centromeres Are Silenced, 704

7.3

Fission Yeast Centromeres Are Composed of Distinct Heterochromatin and Central Kinetochore Domains, 704

2.7

DNA Methylation in Neurospora, 773

2.2

RIp, a Genome Defense System with Both Genetic and Epigenetic Aspects, 775

7.4

Centromere Outer Repeats Alone Allow the Assembly of Silent Chromatin, 707

2.3

Studies of Relics of RIP Provided Insights into the Control of DNA Methylation, 776

7.5

RNA Interference Directs the Assembly of Silent Chromatin, 708

7.6

Centromere Repeat Transcription by RNA Polymerase /I Links RNAi to Chromatin Modification, 709

7.7

Silent Chromatin at Centromeres Is Required to Mediate Sister-Centromere Cohesion and Normal Chromosome Segregation, 7 70

2.4

Quelling, 777

2.5

Meiotic Silencing by Unpaired DNA (MSUD), 779

2.6

Probable Functions and Practical Uses of RIp, Quelling, and MSUD, 720

3. Concluding Remarks, 121 Acknowledgments, 122 References, 122

101

GENERAL SUMMARY Fungi provide excellent models for understanding the structure and function of chromatin both in actively transcribed regions (euchromatin) and in transcriptionally silent regions (heterochromatin). The budding yeast, Saccharomyces cerevisiae, has been an invaluable eukaryotic model for studying chromatin structure associated with transcription at euchromatic regions and for providing a paradigm for silent chromatin. The fission yeast, Schizosaccharomyces pombe, and the filamentous fungus, Neurospora crassa, on the other hand, have been instrumental for studying forms of silencing more closely related to those of higher eukaryotes. Heterochromatic regions are relatively small and not essential for viability in these fungi, making them easier to dissect and manipulate. Our understanding of heterochromatin around centromere and telomere regions is most advanced in the yeasts S. cerevisiae and S. pombe; how-

ever, the mechanism of chromatin silencing employed by S. pombe exhibits features that are conserved with heterochromatic regions of higher eukaryotes. Indeed, both fungi discussed in this chapter-No crassa and S. pombe-contrast with S. cerevisiae in that they employ RNA interference (RNAi), histone H3 methylation of lysine 9 and heterochromatin protein 1 (HP1) type proteins to bring about silent chromatin formation in a manner that is conserved or similar to that in plants and metazoa. In addition, N. crassa sports DNA methylation, which is a characteristic feature of heterochromatin in many higher eukaryotes and which is a classic epigenetic phenomenon. The nature and function of heterochromatin is first discussed in S. pombe after a brief introduction to the organism. We then turn to N. crassa to demonstrate how this filamentous fungus has contributed to epigenetics research.

FUN GAL

MOD E L S

FOR

E PIG ENE TIC

RES EAR C H

•

103

1 Schizosaccharomyces pombe: The Organism Fission yeast, S. pombe, is found in the fermentations involved in the production of beer in subtropical regions; "pombe" is in fact the Swahili word for beer. S. pombe is primarily a haploid (I N) unicellular organism. In medium rich with nutrients, wild-type cells undergo a mitotic division approximately every 2 hours. But a vari. ety of conditions, or conditional mutants, can be used to block cells at distinct stages of the cell cycle or to synchronize cell cultures in G/S, G2, or at metaphase. This is particularly useful, since G phase is very short in fast-growing cultures, and cells pass almost immediately into S phase following cytokinesis; the major portion of the cell cycle is spent in G2 (Fig. 1) (Egel 2004). Like S. cerevisiae, S. pombe can switch between opposite mating types, named Plus (+) and Minus (-). Mating types are equivalent to dimorphic sexes in higher eukaryotes, albeit they are haploid. The information for both mating types resides in the genome as epigenetically regulated silent cassettes-the mat2- P( +) or mat3-M(-) loci. These silent loci provide the genetic template for mating-type identity, but mating type itself is determined by which particular information (+ or -) resides at the active mati locus. Switching of information at the active locus matl, and hence mating type, occurs by recombination between a silent locus and the matl locus according to a strict pattern (Egel 2004). When starved of nitrogen, cells stop dividing and arrest in G J , which promotes the sexual phase of the life cycle through conjugation of pairs of + and - cells to form diploid zygotes (Fig. 1). After mating and nuclear fusion, premeiotic replication occurs (increasing DNA content from 2N to 4N), pairing and recombination of hom*ologous chromosomes then occurs, and this is followed by the reductional meiosis I division and the equational meiosis II division. This produces four separate haploid nuclei (IN) that become encapsulated into spores enclosed in an ascus. The subsequent provision of a rich nutrient source allows germination and resumption of vegetative growth and mitotic cell division. Non-switching derivatives have been isolated or constructed in which all cells are either + or - mating type. This facilitates controlled mating between strains of distinct genotypes. Although S. pombe is normally haploid, it is possible to select for diploid strains. Such diploid cells can then divide by vegetative mitotic growth until starved of nitrogen, when they too undergo meiosis and form "azygotic asci" (Fig. 1). J

conjugation

germination

E2J

.... nuclear fusion, .~ replication,

~ination

I@@@@ •

•

fmeiosis I

0,0]

sporulation •••••

~. meiosis II

[r0-0-0-0-"'-]

Figure 1. Life Cycle of the Fission Yeast, S. pombe Fission yeast has a short G, taking less than 10% of the cell cycle (stippled area is expanded to aid representation). In rich medium, G, cells proceed into 5 phase followed by a long G, (-70% of the cell cycle), mitosis, and cytokinesis. When starved of nitrogen, cells of opposite mating type (+ and -) conjugate, after which nuclei fuse in a process known as karyogamy. Premeiotic replication and recombination allow meiosis I and II to proceed, resulting in four haploid nuclei that are separated into four spores in an ascus. Provision of rich medium allows germination of spores and resumption of the vegetative cell cycle.

1.1 Chromatin Silencing in S. pombe Is Different from That in S. cerevisiae

Fission yeast is a particularly useful model organism in which to study silent chromatin and related epigenetic effects. It is unlike S. cerevisiae, but more akin to N. crassa, plants, and metazoa in the mechanisms that it employs to achieve silencing via heterochromatin. Telomeres and the mating-type loci regions of the genome are subject to silencing in both S. cerevisiae and S. pombe and, in addition,

104.

CHAPTER

centromeres exhibit silencing in S. pombe. However, as described in Chapter 4, S. cerevisiae employs a unique set of proteins-the Sir proteins-to achieve chromatin silencing. Fission yeast and other eukaryotes, on the other hand, appear to use a combination of distinct histone modifications (in particular, histone H3K9 methylation) and RNA interference (RNAi) proteins to silence chromatin. As discussed in detail below, studies in S. pombe have revealed that this silent chromatin is essential for the function of each specialized region of heterochromatin (i.e., centromeres and mating-type loci). At centromeres, heterochromatin is necessary to ensure normal chromosome segregation (Allshire et al. 1995; Ekwall et al. 1995) whereas at the matingtype loci, it facilitates and regulates mating-type switching (for review, see Egel 2004). In addition, silent chromatin is formed adjacent to telomeres, although a function has yet to be ascribed to this telomeric silent chromatin (Nimmo et al. 1994; Kanoh et al. 2005). Marker genes inserted within rDNA are also silenced (Thon and Verhein-Hansen 2000; Cam et al. 2005). In contrast to N. crassa and higher eukaryotes, S. pombe appears to lack any detectable DNA methylation (Wilkinson et al. 1995), a common mechanism used for silencing chromatin in many eukaryotes (see Chapters 9 and 18); thus, silencing in fission yeast is mediated primarily by chromatin modification. As discussed below, the establishment of these modifications, and thus silent chromatin, employs the RNAi machinery.

1.2 Genes Placed in Fission Yeast Centromeres Are 5i1enced

Heterochromatin formation at fission yeast centromeres is essential to allow normal chromosome segregation during nuclear division. Studies have shown that the centromere in fact consists of two distinct chromatin structures: heterochromatin and CENP-A containing kinetochore chromatin. This has been demonstrated by studying the variable silencing of reporter genes inserted into different centromere regions. At the DNA level, centromere regions in fission yeast are composed of outer repeats (subdivided into elements known as dg and dh, or K and L) which flank the central domain that includes the inner repeats (imr or B), and a central core (cnt or CC) (Fig. 2a). The three centromeres, cenl, cen2, and cen3, occupy -40, ~60, and ~ 120 kb on chromosomes I, II, and III, respectively (for reviews, see Ege12004; Pidoux and Allshire 2004). The repetitive nature of fission yeast centromere DNA resembles the larger, more complex repeated structures associated with many metazoan centromeres, but they are more amenable to manipu-

lation (Takahashi et al. 1992; Steiner et al. 1993; Ngan and Clarke 1997). Because repetitive DNA frequently correlates with the presence of heterochromatin in other eukaryotes (see also Chapter 5), the presence of repetitive DNA at fission yeast centromeres suggested that they might have heterochromatic properties such as the ability to hinder gene expression. As described below, two blocks of heterochromatin flank the central domain of each fission yeast centromere. The central domain itself is assembled in a distinct type of chromatin (CENP-A chromatin) that differs from the neighboring heterochromatin. It is well known that the type of chromatin surrounding a gene can strongly influence its expression. This was originally demonstrated in the fruit fly, Drosophila melanogaster, where chromosomal rearrangements that move the white gene close to centromeric heterochromatin lead to its variable expression in eye facets and, thus, variegation in eye coloration (see Fig. 1 in Chapter 5). It is now apparent that transgenes in many organisms can be influenced by the environment into which they are placed. In fission yeast, gene silencing can be monitored by phenotypic assays similar to those used in S. cerevisiae that assess expression of reporter genes. For example, when the ura4+ reporter is silenced, 5-fluoroorotic acid-resistant colonies are formed. Alternatively, silencing of the ade6' reporter results in red rather than white colony color (Fig. 3a). Placement of a normally expressed gene, such as ura4+ or ade6', within the centromere (as defined by the outer repeat and central domain elements) results in its transcriptional silencing. Silencing is robust in the outer repeats, such that most cells form colonies in which repression of markers is stably maintained (i.e., for the ade6' reporter, red colonies are formed). Within the central domain, however, silencing of ade6' is comparatively unstable, resulting in variegated colonies, manifested as either red, white, or red-white, sectored colonies (Fig. 3a). However, no silencing occurs just 1 kb distal to the outer repeats (Allshire et al. 1994, 1995), indicating that this transcriptional repression is confined to the centromere as defined by the central domain and flanking outer repeats. 1.3 Fission Yeast Centromeres Are Composed of Distinct Heterochromatin and Central Kinetochore Domains

The difference in the quality of silencing across fission yeast centromeres reflects the fact that repression of transcription is a result of different chromatin structures, including the associations of different non-histone proteins, at outer repeats and the central domain (Partridge et al. 2000). There are two distinct chromatin structures that have been characterized in centromeric regions of

FUN GAL

MOD E L S

FOR

E P f G ENE T feR ESE ARC H

outer repeat

central domain

outer repeat

Heterochromatin

CENP-A chromatin

Heterochromatin

105

Figure 2. Distinct Outer Repeat Heterochromatin and Central Kinetochore Domains at Fission Yeast Centromeres (a) Representation of a fission yeast centromere. The central domain (pink, kinetochore) is composed of imr and cnt elements, the outer repeats contain transcribed dg and dh repeats (green, heterochromatin). All three centromeres have a similar overall arrangement; however, the number of outer repeats differs: cen7 (40 kb) has two, cen2 (65 kb) has three, and cen] (110 kb) has approximately thirteen. Clusters of tRNA genes (double arrowheads) occur in the imr region and at the extremities of all three centromeres. Transcription of marker genes placed within the outer repeats or central domain is silenced. (b) Heterochromatin: Outer repeats are packaged in nucleosomes which are methylated on H3K9me2, allowing binding of the chromodomain proteins Chpl, Chp2, and Swi6. Central "kinetochore" chromatin: CENP-A is found in the central domain where it probably replaces the majority of H3 to form specialized nucleosomes (pink squares). In addition to CENP-A, several kinetochore proteins (those indicated) have been shown to associate with central domain sequences but not the outer repeats. Kinetochore assembly within the central domain mediates attachment to microtubules upon spindle formation and chromosome segregation. Mutation of heterochromatin components alleviates silencing of marker genes in the outer repeats but not the central domain. Defects in some kinetochore components allow expression of marker genes in the central domain but not the outer repeats.

fission yeast: heterochromatin over the outer repeat regions, and "CENP-A" chromatin coating the central domain where the kinetochore is assembled. Different proteins associate with, and are required for, the silencing of reporter genes in the two domains.

HETEROCHROMATIN

In chromatin, the amino-terminal tails of histones H3 and H4 are subject to a range of posttranslational modifications, which generally correlate with active or repressed states (see Chapter 3). Centromeric heterochromatin at outer repeats is associated with the histone H3 lysine 9 di- and trimethyl states (H3K9me2 and H3K9me3) (Nakayama et al. 2001; Yamada et al. 2005). The formation of centromeric heterochromatin requires the action of several proteins that modify chromatin and

thereby promote other factors to bind. Heterochromatin formation first requires the histone deacetylases (HDACs; such as Clr3, Clr6, and Sir2) to deacetylate histone H3. This subsequently allows the histone lysine methyltransferase (HKMT) Clr4 to methylate histone H3 on lysine 9 over the centromeric outer repeats. This modification creates a specific binding site that is recognized by the chromodomain motif present in Swi6 and Chp2 (hom*ologs of heterochromatin protein 1 [HP 1] described in Chapter 5) and another chromodomain protein Chp 1 (Fig. 2b). These proteins all contribute to the formation of silent chromatin over the outer repeats (Allshire et al. 1995; Cowieson et aI. 2000; Partridge et al. 2000; Bannister et al. 2001; Nakayama et al. 2001; Shankaranarayana et al. 2003; Sadaie et al. 2004). The chromatin associated with reporter gene insertions at the outer repeats (e.g., the ura4+ gene) is notably

106 •

C HAP T E R 6

ade6+

swi6A

ade6 + gene inserted in the outer repeats of centromere 1

ade6+gene inserted in the central domain of centromere 1 Figure 3. Variegation and Alleviation of Marker Gene Silencing

(0) Expression of ode6+ from the central domain is variegated, resulting in red, white, and sectored colonies. Cells express-

ing ode6+ form white colonies, whereas when ode6+ is repressed, red colonies are formed. (b) ode6+ inserted in the outer repeats is robustly silenced, resulting in uniform red colonies. This silencing requires heterochromatin proteins such as Swi6 (which binds methylated H3K9) and RNAi components such as Dcrl (Dicer, an RNase III ribonuclease).

enriched in H3K9me2 and Swi6 proteins. This indicates that chromatin modification and Swi6 can spread from neighboring centromeric repeat DNA into interposing sequences (Cowieson et al. 2000; Nakayama et al. 2001). Swi6 localization and silencing are dependent on H3K9 methylation, as illustrated by disruption of Swi6 localization in cells lacking the H3K9-specific HKMT, Clr4, or in H3 mutants where lysine 9 is replaced with arginine (Ekwall et al. 1996; Mellone et al. 2003). The Swi6 protein dimerizes via its chromoshadow domain (Cowieson et al. 2000), and this probably facilitates its spreading along chromatin fibers, aided by the sequential action of HDACs and the Clr4 H3K9 HKMT. In addition to Swi6, the Chp 1 and Chp2 chromodomain proteins also associate with outer repeat chromatin at centromeres by binding histone H3 methylated on lysine 9. Chp1 has been shown to be a component of the RNAi effector complex RITS (see Chapter 8) and is required for complete methylation of histone H3K9 over the outer repeats and inserted reporter genes (Partridge et al. 2002; Motamedi et al. 2004; Sadaie et al. 2004).

CENTRAL KINETOCHORE DOMAIN CHROMATIN

Before discussing the details of how heterochromatin is formed on the outer repeats at centromeres, it is important to appreciate that central domain chromatin, where the kinetochore is assembled, is very distinct from the flanking outer repeat heterochromatin, because this is

where the kinetochore is assembled. In contrast to silencing at the outer repeats, the silencing of reporter genes in the approximately lO-kb central domain of cen] is essentially independent of Clr4 and therefore does not involve methylation of histone H3 on lysine 9. In fact, the central domain has been shown to have a distinct chromatin composition. This was initially demonstrated by micrococcal nuclease analysis (for explanation of MNase digestion, see Chapter 5), which revealed a smear in contrast to the regular 150-bp ladder characteristic of flanking outer repeat chromatin (Polizzi and Clarke 1991; Takahashi et al. 1992). This distinct pattern differentiates central domain chromatin from heterochromatin and euchromatin, and is related to its assembly in distinctive CENP-A chromatin and the assembly of the kinetochore over this region. In all eukaryotes examined, a histone H3-like protein, known as CENP-A (or cenH3), associates specifically with active centromeres (Cleveland et al. 2003), and CENP-A chromatin is critical for specifying the site of kinetochore assembly (Fig. 2b). In the central domain chromatin of fission yeast centromeres, most histone H3 is replaced by the CENP-A ortholog, known as Cnp 1 (Fig. 4b) (Takahashi et al. 2000). CENP-ACnpl deposition can occur in a replicationdependent manner at S phase or in a replication-independent manner during G2 (for more detail, see Chapter 13). Kinetochore proteins themselves govern the localization and assembly of CENP-ACnpl specifically within the centromere central domain (Goshima et aI. 1999; Taka-

FUN GAL

hashi et a1. 2000; Pidoux et a1. 2003). Ams2 (a GATA factor), for instance, directs replication-coupled CENP-Nnp, deposition in S phase. In its absence, replication-independent deposition by Mis6 in G z can compensate, allowing the levels of CENP-Nnp, to be topped up in interphase (Chen et a1. 2003; Takahashi et a1. 2005). If CENP-Nnp, chromatin structure is disrupted (as in the case of cnpl, mal2, mis6, sim4, and other mutants), the specific smeared micrococcal nuclease digestion pattern reverts to a pattern more typical of bulk chromatin (i.e., a nucleosomal ladder). Mutants that affect central domain chromatin have no detectable effects on silencing in the outer repeats (Cowieson et a1. 2000; Jin et a1. 2002; Pidoux et al. 2003; Hayashi et al. 2004). Furthermore, the fact that CENP-ACnPl, Mal2, Mis6, Sim4, and other kinetochore proteins only associate with the central domain demonstrates that the central kinetochore domain is structurally complex and functionally distinct from outer repeat silent chromatin (Fig. 2b). Not all kinetochore domain proteins have been tested, but it appears that silencing within the central domain results from the assembly of an intact kinetochore which, as in S. cerevisiae, involves at least 50 proteins (Measday and Hieter 2004). This large complex of proteins presumably restricts access of RNA polymerase II to reporter genes placed within this region and thereby impedes their transcription. In mutants such as cnp l, mal2, mis6, and sim4, kinetochore integrity is clearly partially defective at the permissive temperature,

Figure 4. Silent Chromatin in 5. pombe Nuclei Two interphase nuclei with heterochromatin (centromeres, telomeres, and the silent mat2-mat3 loci) decorated by red fluorescent immunolocalization of Swi6, and kinetochore chromatin (centromeres only) decorated by green fluorescent immunolocalization of CENP_A,npl. Red signals not in close proximity to green represent telomeres or mat2-mat3. All centromeres are clustered at the nuclear periphery adjacent to the spindle pole body.

MOD E L S

FOR

E PIG ENE TIC

RES EAR C H

•

107

and this allows increased transcription of reporter genes. A spin-off of this is that a normally silent reporter gene has been used to assay for defects in central core chromatin, leading to the identification of novel kinetochore proteins (Pidoux et a1. 2003). Centromere regions are not completely devoid of genes and, intriguingly, several tRNA genes reside between the outer repeats and the central kinetochore domain (Fig. 2a) (Kuhn et al. 1991; Takahashi et a1. 1991). Recently, these have been shown to act as a barrier preventing heterochromatin from encroaching into the central domain (Scott et a1. 2006).

1.4 Centromere Outer Repeats Alone Allow the Assembly of 5i1ent Chromatin Clr4-dependent silent chromatin is assembled not only at centromeres, but also over a region of about 20 kb containing the silent mating-type loci (mat2-mat3) (Noma et a1. 2001) and adjacent to the terminal telomeric repeats (consensus TTACAGG) added by telomerase (Nimmo et al. 1994, 1998; Allshire et al. 1995; Kanoh et a1. 2005). The cenH (for centromere hom*ologous) region that resides between mat2 and mat3 shares a high degree of similarity, over ~ 7 kb, with the outer repeats found at centromeres (Grewal and Klar 1997). In addition, at least 0.5 kb of DNA sequences with >84% identity to cenH are located within the telomere-associated sequences (TAS) that occupy up to 40 kb proximal to the telomeres of chromosomes I and II (Kanoh et a1. 2005). This suggests that the outer centromeric and related cenH repeats might act in cis to bring about silent chromatin assembly. Assembly of Clr4-dependent silent chromatin occurs even on adjacent marker genes when centromeric outer repeat (dg) or mat2-mat3 (cenH) DNA sequences are inserted in regions of the genome where silencing does not normally occur ("ectopic" silencing; Fig. 5) (Ayoub et al. 2000; Partridge et al. 2002; Volpe et aI. 2003). A simple explanation would be that DNA-binding proteins recognize these repeats, and when bound, these proteins recruit HDACs and the HKMT Clr4, resulting in H3K9 methylation, the binding of chromodomain proteins, and the formation of heterochromatin. However, the situation is more complicated than this. It is now known that the centromeric outer repeats are transcribed. Remarkably, this transcription results in the production of a doublestranded RNA (dsRNA) substrate for the RNAi machinery, and this then recruits the Clr4 HKMT to trigger the assembly of silent chromatin (Volpe et aI. 2002; Sadaie et a1. 2004). RNAi also acts on the related cenH repeats

108

C HAP T E R 6

"-.tIIHt-H

domain, bind centromeric repeat DNA, and are required for effective silent chromatin formation (Irelan et al. 2001; Nakagawa et al. 2002). Other observations indicate that the Clr3 HDAC also acts independently of RNAi to maintain heterochromatin integrity (Yamada et al. 2005).

! !

1.5 RNA Interference Directs the Assembly of Silent Chromatin

OFF

Figure

s. Centromere

Repeat Sequences Mediate Silencing

Insertion of fragments (1-2 kb) adjacent to an expressed ura4+ gene at a locus that is not normally prone to silencing results in H3K9 methylation, binding of Swi6 and Chp1, and transcriptional silencing.

found in the mat2-mat3 region and in the TAS repeats at telomeres (Cam et al. 2005; Kanoh et al. 2005). However, the picture is further complicated by the demonstration that at mat2-mat3, RNAi acts to establish silent chromatin while the DNA-binding proteins Atfl and Pcrl bind near cenH and maintain silent chromatin over mat2-mat3 even in the absence of key RNAi components (Jia et al. 2004; Kim et al. 2004). Similarly, overlapping mechanisms of silencing also operate at telomeres; terminal repeats alone can recruit Clr4 HKMT and thus Swi6 via the telomere repeat-binding protein Tazl, but RNAi also acts via the cenH part of the TAS elements to form an extended region of silent chromatin at telomeres (Nimmo et al. 1994; Allshire et al. 1995; Kanoh et al. 2005). Is there a comparable overlapping mechanism to maintain silent chromatin at centromeres? Although silencing of outer repeat reporter genes and centromeric outer repeat transcripts themselves is defective in cells lacking RNAi, H3K9me2 is retained on the outer repeat chromatin in the absence of RNAi components (Sadaie et al. 2004). What factors are responsible for maintaining this methylation? One possibility is the CENP-B-related proteins (Abpl, Cbhl, Cbh2), which contain a conserved DNA-binding

The phenomenon of RNAi was first discovered in Caenorhabditis elegans, where it was found that expression of dsRNA results in loss of expression of a hom*ologous gene. It soon became apparent that this form of RNAi is related to the process of transcriptional gene silencing (TGS) described in plants and quelling in N. crassa (described below). It is known in plants and metazoa that dsRNA, when processed by RNAi machinery, yields small RNAs that bring about DNA and/or chromatin modifications on hom*ologous chromatin. The presence of fission yeast orthologs of the main components of the RNAi pathway, i.e., Argonaute (Agol), Dicer (Dcrl), and RNA-dependent RNA polymerase (Rdpl), provoked their investigation in fission yeast (Volpe et al. 2002), which has led to significant advances in understanding RNAi-mediated chromatin modification and silencing. The phenotypic consequences resulting from loss of Agol, Dcrl, or Rdpl function are similar to those of swi6 mutants: reduced H3K9me2 and loss of silencing over the outer repeats of centromeres (Fig. 3b). In RNAi mutants, however, overlapping noncoding RNA (ncRNA) transcripts of a discrete size were detected, originating from centromeric outer repeats. These ncRNAs are hom*ologous to naturally occurring small RNAs called short interfering RNAs (siRNAs; -21 nt) that have been isolated and sequenced from S. pombe (Reinhart and Bartel 2002; Cam et al. 2005). These siRNAs are generated by Dicer (an RNAi component) mediated by cleavage of the doublestranded derivatives of the long centromere repeat hom*ologous noncoding transcripts. The chromodomain protein Chpl also turns out to be a component of the RNAi machinery. Chp 1 binds to H3K9me2 chromatin and is required for reporter gene silencing at the outer repeats of centromeres, but is also required for the generation of siRNAs hom*ologous to centromere repeats (Noma et al. 2004) and for normal levels of H3K9 methylation on the centromere repeats (Partridge et al. 2002; Sadaie et al. 2004). The role of Chpl in RNAi was further supported by the identification of the RITS (RNA-induced transcriptional silencing) effector

FUN GAL

complex, which contains Chp1, Ago1, Tas3, and siRNAs hom*ologous to centromeric outer repeats (Motamedi et al. 2004). In addition to RlTS, a complex containing Rdp1 (RDRC: RNA-directed RNA polymerase complex) has been identified which contains a predicted RNA helicase (Hrrl) and a putative poly(A) polymerase (Cid12). These also appear to be integral to the process of RNAi and the assembly of intact silent chromatin on the outer repeats at centromeres (Fig. 6) (Motamedi et al. 2004). The presence of the RNAi-independent (AtfllPcr1dependent) silencing mechanism operating at the silent mating-type locus to maintain heterochromatin was discovered by treatment of cells lacking RNAi with the HDAC inhibitor trichostatin A (TSA). This resulted in the complete loss of silent chromatin from the mat2-mat3 region mimicking the effect of deleting atfl or perl in RNAi mutants (Hall et al. 2002; Jia et al. 2004). However, as noted above, Atfl and Pcr1 are not required for the formation of silent chromatin at the centromeric outer repeats (Kim et al. 2004). Transient treatment with TSA

hete

cohesin

RITS: Ago1 Tas3 Chp1 RDRC: Rdp1 Cid12 Hrr1 Figure 6. Centromere Repeat Transcription Links RNAi, Heterochromatin Formation, and Cohesion Transcription of outer repeats by RNA pol II provides an initial substrate for RNAi and Oicer-dependent siRNA generation. Loading of Ago1 in the RITS complex (Ago1, Tas3, Chpl) with siRNA allows targeting of the hom*ologous transcript. The action of the RORC (Rdpl, Cid12, and Hrrl) would allow dsRNA production providing more substrate for Ocrl to produce siRNA and perhaps amplify the signal. Interactions between the transcript, RNA pol II subunits, and RNAi components recruit Clr4, which methylates histone H3 on lysine 9, allowing binding of chromodomain proteins, recruitment of cohesin, and sister-centromere cohesion.

MOD E L S

FOR

E PIG ENE TIC

RES EAR C H

109

also causes loss of silencing at centromeres, resulting in hyperacetylation of histones coupled with defective centromere function (Ekwall et al. 1997). Although this TSAinduced "epistate" is metastable, it can be propagated through several rounds of cell divison and even meiosis. The most likely explanation is that heterochromatin is difficult to reestablish once this abnormal hyperacetylated state is attained. Epistates can also be established at a compromised mat2-mat3 locus, and these too are propagated through meiosis (Nakayama et al. 2000).

1.6 Centromere Repeat Transcription by RNA Polymerase 1/ Links RNAi to Chromatin Modification

The previous section demonstrated that RNAi and histone modifications are required to assemble centromeric heterochromatin, which can spread into adjacent inserted reporter genes. These observations raised obvious questions, such as, How is RNAi linked with the covalent modification of histones to form silent chromatin? and How are particular regions of the genome targeted for such RNAi-directed chromatin modification and silencing? In many organisms, the expression of a specific dsRNA, hom*ologous to a gene of interest, results in either transcriptional (DNA/chromatin modification) gene silencing (TGS) or posttranscriptional (mRNA degradation) gene silencing (PTGS). Can any dsRNA be processed to form siRNAs in fission yeast, and do such siRNAs induce only posttranscriptional silencing (RNA knockdown), or can they also bring about chromatin modifications (e.g., H3K9me2) that silence transcription from the hom*ologous gene? Expression of dsRNAs hom*ologous to a GFP reporter produces siRNAs that cause a reduction in GFP transcripts, but the transcriptional activity of the GFP reporter gene does not decline. Thus, although GFP-siRNAs reduce GFP mRNA levels, they are unable to bring about the chromatin modifications that result in transcriptional silencing. These GFPsiRNAs must therefore only act posttranscriptionally to destroy GFP mRNA (Sigova et al. 2004). It is unclear why GFP-siRNAs do not induce chromatin modification on the hom*ologous gene, but it may be related to the nature of the nascent transcript or the strength of the RNA pol II promoter driving GFP expression. Which RNA polymerase is responsible for transcription of centromere repeats? Mutation of either of two RNA pol II subunits (Rpb2 and Rpb7) results in defective centromere silencing (Djupedal et al. 2005; Kato et al. 2005), although these mutations show very different phenotypes. The rpb7-1 mutant shows reduced levels of cen-

110 •

CHAPTER

tromere repeat transcnptlOn, resulting in less ncRNA and, consequently, less siRNA production and a loss of silent chromatin. This implies that RNA pol II is required to transcribe centromere repeats to provide the primary substrate for RNAi. In contrast, in the rpb2-m203 mutant, centromeric transcripts are produced, but they are not processed to siRNA, and H3K9me at centromeres is reduced. These studies indicate that RNAi not only requires an RNA pol II transcript, but also that, like other RNA processing events, the production of centromeric siRNA may be coupled to transcription by interactions between RNAi machinery, chromatin, modifying enzymes, and RNA pol II (Fig. 6). It is possible that RITSassociated siRNAs may home in on nascent transcripts as they emerge from RNA pol II engaged with the hom*ologous locus. Once recognition has taken place, the RITS-siRNA complex might be stabilized on these transcripts, resulting in recruitment of chromatin-modifying activities such as Clr4. Surprisingly, centromeric siRNAs are also lost in cells lacking Clr4 HKMT activity (Noma et al. 2004; Hong et al. 2005). It is possible that the absence of Clr4 affects siRNA production by destabilizing associations between various components at the interface between transcription, RNAi, and chromatin modification (Motamedi et al. 2004). Alternatively, H3K9 methylation may be required to allow the generation of siRNAs in cis via the action of various RNA processing activities (e.g., RdRP) on primary centromeric transcripts (Fig. 6) (Noma et al. 2004; Sugiyama et al. 2005; see Chapter 8).

1.7 Silent Chromatin at Centromeres Is Required to Mediate Sister-Centromere Cohesion and Normal Chromosome Segregation

How do outer repeat centromeric heterochromatin and CENP-Nnpl kinetochore chromatin affect the overall function of chromosome segregation? Clr4-dependent silent chromatin assembles on outer repeats at centromeres, at a related repeat at the mating-type locus, and adjacent to telomeres. Experiments with naked DNA plasmid constructs have shown that outer repeats contribute in some way to the assembly of a functional centromere, imparting the ability for these Cen-plasmids to segregate on mitotic and meiotic spindles. But neither the outer repeats nor the central domain alone is sufficient to assemble a functional centromere (Clarke and Baum 1990; Takahashi et al. 1992; Baum et al. 1994; Ngan and Clarke 1997; Pidoux and Allshire 2004). Mutants that cause loss of silencing at the outer repeats (i.e., those defective in Clr4,RNAi components,

or Swi6) have elevated rates of mitotic chromosome loss and a high incidence of lagging chromosomes on late anaphase spindles (Fig. 7a) (Allshire et al. 1995; Ekwall et al. 1996; Bernard et al. 2001; Nonaka et al. 2002; Hall et al. 2003; Volpe et al. 2003). Cells lacking Swi6 are defective in cohesion at centromeres but retain cohesion along the chromosome arms (Bernard et al. 2001; Nonaka et al. 2002). The formation of a properly bioriented spindle requires that sister kinetochores attach to microtubule fibers emanating from opposite spindle poles. The forces exerted on bioriented kinetochores require that sister kinetochores be held together tightly (Fig. 7b). Swi6 is required to recruit cohesin to outer repeat chromatin and thereby mediate tight physical cohesion between sister centromeres (Fig. 6). Thus, one function of silent chromatin at centromeres is to mediate cohesion. Cohesin is also strongly associated with telomeres and the mat2-mat3 region (Bernard et al. 2001; Nonaka et al. 2002). In addition, cohesin is also recruited to silent chromatin formed on a ura4+ gene in response to an adjacent ectopic centromere repeat (Fig. 5), underscoring the link between silent chromatin and cohesion (Partridge et al. 2002). Thus, the recruitment of cohesin seems to be a general property of Swi6-associated silent chromatin. How Swi6 chromatin brings about cohesin recruitment is not known, but Swi6 does interact with the Psc3 cohesin subunit (Nonaka et al. 2002). In addition, Dfp1, the regulatory subunit of the conserved kinase Hsk1 (Cdc7), interacts with Swi6 and is required to recruit cohesin to centromeres (Bailis et al. 2003). This functional link between heterochromatin and chromatid cohesion appears to be conserved in other organisms, since depletion of the RNAi component, Dicer, appears to affect heterochromatin integrity and sister-centromere cohesion in vertebrate cells (f*ckagawa et al. 2004).

1.8 Epigenetic Inheritance of the Functional Centromere State

An interesting epigenetic phenomenon has been described with respect to the assembly of functional centromeres in fission yeast on plasmids containing minimal regions for centromere function. Although constructs retaining only part of an outer repeat and most of the central domain inefficiently assemble a functional centromere, surprisingly, once this functional centromere active state has been established, it can be propagated through many mitotic divisions and even through meiosis (Steiner and Clarke 1994; Ngan and

FUNGAL

.... ago1L1

dcr1L1

.... .... rdp1L1

Wildtvpe Bi-oriented sister centromeres

heterochromatin

.Defective heterochromatin Merotelically oriented single centromere

Figure 7. loss of Heterochromatin Results in Defective Chromosome Segregation (0) Cells lacking RNAi or heterochromatin components display elevated rates of chromosome loss and lagging chromosomes on late anaphase spindles. (b) lagging chromosomes in cells with defective heterochromatin may result from disorganized kinetochores so that one centromere can attach to microtubules from opposite poles. Such merotelic orientation could persist into anaphase; breakage of attachment with one pole or other would lead to random segregation and result in chromosome loss/gain events.

MODELS

FOR

EPIGENETIC

RESEARCH.

111

Clarke 1997). One interpretation is that the outer repeats provide an environment that is favorable for kinetochore assembly (Pidoux and Allshire 2005), but once assembled, CENP-ACnpl chromatin, and thus the kinetochore, is propagated at this position by a templating mechanism that may be coupled to replication (Takahashi et al. 2005). It is possible that heterochromatin somehow induces or aids the deposition of CENP-ACnpl in the central domain (Fig. 8) and that only one block of heterochromatin does not permit efficient kinetochore assembly. An alternative explanation is that one outer repeat is insufficient to recruit enough cohesin, and this leads to defective centromeric cohesion and elevated rates of chromosome loss. Such centromere constructs may stochastically accumulate sufficient cohesin after several cell divisions, resulting in increased mitotic stability. Once attained, this stabilized state must be somehow duplicated on daughter molecules to allow its propagation through subsequent divisions (Fig. 8). As mentioned above, TSA can also set up a centromere-defective state caused by loss of silencing (Ekwall et al. 1997). Given the connection between silent chromatin formation and cohesion at centromeres, it seems likely that TSA-induced hyperacetylation blocks the efficient reestablishment of silent chromatin via RNAi and that defective centromere function is propagated due to the loss of silencing and thus sister-centromere cohesion (for more detail, see Chapter 14). 1.9 Diverse Silencing Mechanisms in Fungi

In the first half of this chapter, we described how in a single, relatively simple eukaryote, the fission yeast S. pombe, reporter genes are silenced by two distinct types of chromatin at centromeres. CENP-A chromatin resides in the middle of the centromere and is flanked on both sides by blocks of heterochromatin. CENP-A chromatin marks the region over which the kinetochore forms and presumably attaches to microtubules. RNAi is utilized in the formation of flanking heterochromatin to target noncoding transcripts emanating from the outer repeats and deliver chromatin-modifying enzymes such as Clr4 histone H3K9 methyltransferase, which creates binding sites for chromodomain proteins and, thus, robust silencing. This heterochromatin contributes to centromere function by providing tight physical cohesion and perhaps aiding kinetochore organization. The use of RNAi in forming silent chromatin at centromeres may be derived from its role in genome defense against RNA viruses and transposable elements, as has been described in plants. Indeed,

112

CHAPTER

establiSh~

?? CENP_A

cnp1

isms, including the model fungi presented in this chapter, it is already clear that S. pombe and N. crassa will continue to serve as rich sources of information on epigenetic mechanisms operative in eukaryotes.

HETEROCHROMATIN

2 Neurospora crassa: History and Features of the Organism

Figure 8. Establishment and Maintenance of CENP-A Chromatin Central domain DNA alone is unable to establish a functional centromere; outer repeats are required. Loss of heterochromatin from established centromeres does not affect CENP-A,"pl or kinetochore assembly on the central domain. This suggests that heterochromatin may in some way direct the site of CENP-A,npl chromatin and thus kinetochore assembly. It is not known how CENP_A,npl is deposited in nucleosomes or how this specialized chromatin is maintained in the central domain.

it is possible that centromere repeats themselves are the remnants of ancient mobile elements. This possibility is strengthened by early studies of centromeres of the filamentous fungus, N. crassa, the subject of the second half of this chapter. Although there are clearly common epigenetic mechanisms at work in the diverse pair of fungi presented in this chapter, it is clear that these fungi show dramatic differences, as described below. For example, unlike the well-studied budding and fission yeasts, Neurospora uses DNA methylation, a classic epigenetic process common in higher eukaryotes such as mammals and flowering plants. In addition, studies in Neurospora have revealed several independent silencing systems operating at distinct stages of its life cycle. The first such mechanism, named repeat-induced point mutation (RIP), has both epigenetic and genetic aspects and clearly serves as a genome defense system. The second, named quelling, is an RNAi-based mechanism that results in silencing of transgenes and their native hom*ologs. The third, named meiotic silencing by unpaired DNA (MSUD), is also RNAi-based but is distinct from quelling in its time of action, targets, and apparent purpose. Although we are still in the early days of epigenetic studies in all organ-

The filamentous fungus Neurospora crassa (see Figs. 9 and 10) was first developed into an experimental organism by Dodge in the late 1920s and, about 10 years later, was adopted by Beadle and Tatum for their famous "one geneone protein" studies linking biochemistry and genetics (Davis 2000). Beadle and Tatum selected Neurospora in part because this organism grows fast and is easy to propagate on defined growth media and because genetic manipulations, such as mutagenesis, complementation tests, and mapping, are simple with Neurospora. Although not as widely studied as some model eukaryotes, Neurospora continues to attract researchers because of its moderate complexity and because it is well suited for a variety of genetic, developmental, and subcellular studies (Borkovich et al. 2004). Neurospora has been especially useful for studies of photobiology, circadian rhythm, population biology, morphogenesis, mitochondrial import, DNA repair and recombination, DNA methylation, and other epigenetic processes. The N. crassa life cycle is illustrated in Figure 9. The vegetative phase is initiated when either a sexual spore (ascospore) or an asexual spore (conidium) germinates, giving rise to multinucleate cells that form branched filaments (hyphae; Fig. lOC). The two mating types (A and a) are morphologically indistinguishable (Fig. lOB). Conveniently, ascospores require a heat shock to germinate, whereas conidia germinate spontaneously. The hyphal system spreads out rapidly (linear growth >5 mm per hour at 37°C) to form a "mycelium." After the mycelium is well established, aerial hyphae develop, leading to the production of the abundant orange conidia that are characteristic of the organism (Fig. lOA, B). The conidia, which contain one to several nuclei each, can either establish new vegetative cultures or fertilize crosses. If nutrients are limiting, Neurospora prepares to enter its sexual phase by producing nascent fruiting bodies ("protoperithecia"). When a specialized hypha ("trichogyne") projecting from the protoperithecium contacts tissue of the opposite mating type, a nucleus is picked up and transported back to the protoperithecium. The sexual phase of Neurospora and other filamentous ascomycetes differs from that of yeasts in having a prolonged heterokaryotic phase between fertilization

FUNGAL

MODELS

FOR

EPIGENETIC

RESEARCH.

113

DNA Methylation; Quelling

A conidia

A mycelium ?)L ~)L

mycelium

germinated ascospores

~conidia

8scogenous hypha with A and -nuclel-

prolnw.,lon of_lei

diploid nucleus -.Iosls MIl mlt~

.......

nuclear (CSryogMfy)

MSUD

RIP

Figure 9. life Cycle of N. crassa The stages in which epigenetic processes that are described in the text occur are indicated. (Adapted, with permission, from Shiu et al. 2001.)

and karyogamy (nuclear fusion). The heterokaryotic cells resulting from fertilization proliferate in the developing perithecium. In the final divisions, the cells are binucleate, containing one nucleus of each mating type. These cells bend to form hook-shaped cells ("croziers"), a final mitosis occurs to produce four nuclei, and septa are laid down to produce one binucleate cell at the crook of the crozier. This cell gives rise to an ascus. Genetic analyses have indicated that, in general, the ~ 100 or more asci of a perithecium are derived from a single maternal nucleus and a single paternal nucleus. When karyogamy occurs, the resulting diploid nucleus immediately enters into meiosis. Thus, the diploid phase of the life cycle is brief and limited to a single cell. The meiotic products undergo one mitotic division before being packaged as ascospores and undergo additional mitoses in the developing ascospores (see Fig. 9 and Davis 2000). The ~40-megabase N. crassa genome consists of seven chromosomes with approximately 10,000 predicted protein-coding genes (Galagan et al. 2003) and a total genetic map length of roughly 1000 map units (Perkins et al. 2001). Only about 10% of the genome consists of repeti-

tive DNA and, aside from a tandem array of ~ 70 copies of the ~9-kb rDNA unit encoding the three large rRNAs, most of the repetitive DNA consists of inactivated transposable elements. That most strains of Neurospora lack active transposons and have very few close paralogs almost certainly reflects the operation of RIP, the first hom*ologydependent genome defense system discovered in eukaryotes (Selker 1990). We know that Neurospora has at least three gene-silencing processes which should serve to conserve the structure of the genome: RIP, quelling, and MSUD (Borkovich et al. 2004). All of these processes have epigenetic aspects and have direct or indirect connections with DNA methylation, a basic epigenetic mechanism found in Neurospora and many other eukaryotes. We discuss DNA methylation and then RIP, quelling, and MSUD. 2.1 DNA Methylation in Neurospora

Since its discovery decades ago, DNA methylation in eukaryotes has remained remarkably enigmatic. Basic questions are still debated, such as, What determines which chromosomal regions are methylated? and What is

114

CHAPTER

Figure 10. Images of N.

crassa

(A) Vegetative growth in the wild on sugarcane (photo by D. jacobson; Stanford University). (8) Slants of vegetative cultures of N. crassa in the laboratory (photo by N. B. Raju, Stanford University). (C) Hyphae of N. crassa stained with DAPI to show abundant nuclei (photo by M. Springer, Stanford University). (0) Rosette of maturing asci showing ascospore patterns. (0, Reprinted, with permission, from Raju 1980 [©Elsevier].)

the function of DNA methylation? Neurospora revealed itself to be an excellent system to study the control and function of DNA methylation. Some model eukaryotes, including the nematode C. elegans and the yeasts S. cerevisiae and S. pombe, lack detectable DNA methylation, and reports of DNA methylation in another model organism, D. melanogaster, remain controversial. DNA methylation is essential for viability in some organisms, such as mammals, complicating certain analyses. In Neurospora DNA, about 1.5% of the cytosines are methylated, but this methylation is dispensable, facilitating genetic studies. Although one must be cautious when extrapolating from one system to another, at least some aspects of DNA methylation appear conserved. For example, all known DNA methyltransferases (DMTs), including those from both prokaryotes and eukaryotes, show striking hom*ology in their catalytic domains (Grace Goll and Bestor 2005). Findings from Neurospora, Arabidopsis, mice, and other systems in the last decade have revealed both important similarities and interesting differences in the control and function of DNA methylation, demonstrating the value of carrying out investigations in multiple model systems.

Discovery of DNA methylation in Neurospora initially attracted interest because it was not limited to symmetrical sites, such as CpG dinucleotides or CpNpG trinucleotides. Riggs, and Holliday and Pugh, had proposed an attractive model for the "inheritance" or "maintenance" of methylation patterns that relied on the symmetrical nature of methylated sites observed in animals. Although results of a variety of in vitro and in vivo studies have supported the "maintenance methylase" model (see Chapter 18), mechanisms for maintenance methylation that do not rely on faithful copying at symmetrical sites can be imagined and may be operative in a variety of organisms (see, e.g., Selker et al. 2002). The possibility that the observed methylation at asymmetric sites represented "de novo methylation" was exciting because mechanisms that blindly propagate methylation patterns can complicate determination of which sequences are methylated in the first place. Indeed, results of DNA-mediated transformation and methylation inhibitor studies with Neurospora demonstrated reproducible de novo methylation (see, e.g., Singer et al. 1995). Additional studies defined, in part, the underlying signals for de novo methylation (see, e.g., Tamaru and Selker 2003).

FUNGAL

~~ ~

fertilization

l®@(

.+ premeiotic DNA synthesis .+ karyogamy

.....--

~ ! ::::E5= ::::::::~:::

""'"'''''''''''''' ,,',,""""""""

meiosis

~~~~ ~~~~ Figure 11. Repeat-induced Point Mutation (RIP)

For clarity, only two chromosomes are illustrated. The open box represents a gene, or any chromosomal segment, which when duplicated (e,g., in the strain indicated on the top right) is subject to RIP (symbolized by lightning bolt) between fertilization and karyogamy, Results of genetic experiments reveal that duplications can be repeatedly subjected to volleys of C-to-T transitions (symbolized by filled boxes) during this period of -10 mitoses, right up to the final premeiotic DNA synthesis (Selker et al. 1987; Watters et al. 1999), The four possible combinations of chromosomes in progeny are indicated, and the red "m" represents DNA methylation, which is frequently (although not always) associated with products of RIP.

The first methylated patch characterized in detail was the 1.6-kb S-11 region, which consists of a diverged tandem duplication of a 0.8-kb segment of DNA, including a 5S rRNA gene. Comparison of this region with the corresponding chromosomal region of strains lacking the duplication initially led to the idea that repeated sequences can somehow induce DNA methylation and ultimately led to the discovery of the genome defense system named RIP. Elucidation of RIP revealed that repeated sequences do not directly trigger DNA methylation, at least in Neurospora; instead, repeats trigger RIP, which is closely tied to DNA methylation, as described below. Both the S-11 region and the \jf63 region, the second methylated region discovered in Neurospora, are products of RIP. Moreover, subsequent

MODELS

FOR

EPIGENETIC

RESEARCH.

115

genome-wide analyses of DNA methylation revealed that the majority of methylated regions in Neurospora are relics of transposons inactivated by RIP (Galagan et al. 2003; Selker et al. 2003). Indeed, the only DNA methylation in Neurospora that may not have resulted from RIP is that in the tandemly arranged rDNA (Perkins et al. 1986).

2.2 RIp, a Genome Defense System with Both Genetic and Epigenetic Aspects

RIP was discovered as a result of a detailed analysis of progeny from crosses of Neurospora transformants (Selker 1990). It was noticed that duplicated sequences, whether native or foreign, and whether linked or unlinked, were subjected to numerous polarized transition mutations (G:C to A:T) in the haploid genomes of the special binucleate cells resulting from fertilization (Fig. 12). Experiments in which the stability of a gene was tested when it was unique in the genome or else combined with an unlinked hom*olog demonstrated that RIP is not simply repeat-associated; it is truly repeat-induced. In a single passage through the sexual cycle, up to about 30% of the G:C pairs in duplicated sequences can be mutated. Frequently (but not invariably), sequences altered by RIP become methylated de novo. It is likely that the mutations from RIP occur by enzymatic deamination of 5-meCs or by deamination of Cs followed by DNA replication (Selker 1990). Cytosine methylation involves a reaction intermediate that is prone to spontaneous deamination, suggesting that the putative deamination step of RIP might be catalyzed by a DMT or DMT-like enzyme. Consistent with this possibility, one of the two DMT hom*ologs predicted from the Neurospora genome sequence is involved in RIP (Freitag et al. 2002). Progeny from hom*ozygous crosses of mutants with defects in this gene, rid (RIP defective), do not show new instances of RIP. Rid mutants do not display any noticeable defects in DNA methylation, fertility, growth, or development. In contrast, the second Neurospora DMT hom*olog (DIM-2), which was identified genetically and shown to be necessary for all known DNA methylation, is not required for RIP (Kouzminova and Selker 2001). All indications are that every sizable duplication (greater than ~400 bp for tandem duplication or ~ 1000 bp for unlinked duplication) is subject to RIP in some fraction of the special binucleate cells. Nevertheless, duplications escape RIP at some frequency (typically less than 1% for a tandem duplication or ~50% for an unlinked duplication). Even duplications of chromosomal segments containing numerous genes are sensitive to RIP (Perkins et al.

116 •

CHAPTER

------+----H--

amR1P1 amR1P2 am R1P3 am R1P4 am R1P5 am RIP6

am R1P7 am RIP8 I

e - bp IB 100

B Figure 12. Mutations from RIP and Methylation Status of Eight am Alleles

Vertical bars indicate mutations. Alleles shown in black were not methylated, alleles in blue were initially methylated but, after loss of methylation was induced with 5-azacytidine, or by cloning and gene replacement, did not become remethylated. Alleles shown in red were not only initially methylated, but also triggered methylation de novo. (Adapted from Singer et al. 1995.)

1997; Bhat and Kasbekar 2001). Although RIP is limited to the sexual phase of the life cycle, the existence of this process raised the question of whether Neurospora can utilize gene duplications to evolve. The genome sequence revealed gene families, but tellingly, virtually all paralogs were found to be sufficiently divergent that they should not trigger RIP (Galagan and Selker 2004). We conclude that RIP may indeed limit evolution in Neurospora. Interestingly, some fungi show what appear to be milder genome defense systems that are similar to RIP. The most notable example is methylation induced premeiotically (MIP), a process that detects linked and unlinked sequence duplications during the period between fertilization and karyogamy, like RIP, but which relies exclusively on DNA methylation for inactivation; no evidence of mutations has been found in sequences inactivated by MIP (Rossignol and Faugeron 1994). 2.3 Studies of Relics of RIP Provided Insights into the Control of DNA Methylation NONCANONICAL MAINTENANCE METHYLATION

The finding that a single DMT, DIM-2, is responsible for all detected DNA methylation was as surprising as the initial finding of rampant methylation at nonsymmetrical sites in Neurospora because no previously identified DMT was

known to methylate cytosines in a variety of sequence contexts. An obvious but important question was, Does methylation at nonsymmetrical sites necessarily reflect the potential of the corresponding sequences to induce methylation de novo? Early transformation experiments were consistent with this possibility; methylated sequences that were stripped of their methylation (e.g., by cloning) regained their normal methylation when reintroduced into vegetative cells. A surprise came, however, when eight alleles of the am gene that were generated by RIP were tested for their capacity to induce methylation de novo (Singer et a1. 1995). Some products of RIP with relatively few mutations (Fig. 12, am RlP3 and am RlP4 ) did not become re-methylated, even at their normal locus, suggesting that the observed methylation represented propagation of methylation established earlier. Importantly, their methylation, like other observed methylation in Neurospora, was not limited to symmetrical sites, did not significantly spread with time, and was "heterogeneous" in the sense that the pattern of methylated residues was not invariant within a clonal population of cells. Thus, although dependent on preexisting methylation established in the sexual phase (perhaps by RIP), this methylation could not reflect the action of a "maintenance methylase" of the type envisioned in the original model for inheritance of methylation patterns. It is noteworthy that MIP in Ascobolus, which also

FUNGAL

results in heterogeneous methylation, provided the first evidence for propagation of DNA methylation in fungi (Rossignol and Faugeron 1994). The capacity of Neurospora to perform maintenance methylation was confirmed experimentally (Selker et al. 2002). Interestingly, propagation of methylation was found to be sequence-specific (i.e., it did not work on all sequences), adding a new dimension to the maintenance methylation concept. Although a number of potential schemes that would result in propagation of DNA methylation can be imagined, the actual mechanism operative in Neurospora remains unknown. In principle, maintenance of methylation at nonsymmetrical sites could depend on methylation of nearby symmetrical sites, but the observed heterogeneous methylation, including at CpG sites, renders this possibility unlikely. Feedback mechanisms involving proteins associated with the methylated DNA could result in methylation that depends on preexisting methylation, i.e., maintenance methylation. As discussed below, findings from Neurospora (and other organisms) implicate histone modifications in the control of DNA methylation, raising the possibility that histones playa role in maintenance methylation.

INVOLVEMENT OF HISTONES IN

DNA

METHYLATION

The first indication of a role of histones in DNA methylation came from the observation that treatment of Neurospora with the histone deacetylase inhibitor trichostatin A (TSA) reduced methylation in some chromosomal regions (Selker 1998). The selectivity of demethylation by TSA could reflect differential access to histone acetyltransferases, but this has not been thoroughly investigated (Selker et al. 2002). Chromatin was unambiguously tied to the control of DNA methylation through investigations with the Neurospora mutant dim-5, which, like dim-2 strains, shows a complete loss of DNA methylation. The SET domain protein DIM-5 was found to be a histone H3 methyltransferase that specifically trimethylates lysine 9 (Tamaru and Selker 2001; Tamaru et al. 2003). Confirmation that histone H3 is the physiologically relevant substrate of DIM-5 came from two demonstrations: (1) Replacement of lysine 9 in H3 with other amino acids caused loss of DNA methylation and (2) trimethyl-lysine 9 was found specifically at methylated chromosomal regions. The discovery that histone methylation controls DNA methylation, at least in Neurospora, led to two important questions: (1) What tells DIM-5 which nucleosomes to methylate? (2) What reads the trimethyl mark and transmits this information to the DMT, DIM-2? The

MODELS

FOR

EPIGENETIC

RESEARCH.

117

factors that control DIM-5 have not been definitively identified, but there are strong suggestions that it is controlled by one or more proteins that recognize products of RIP and by the modification state of amino acid residues in the neighborhood of its action (Selker et al. 2002).The latter should allow DIM-5 to integrate information relevant to whether DNA in a particular region should be methylated and provides a possible explanation for the observation that TSA can inhibit DNA methylation in certain regions. Findings in other systems led to discovery of what reads the trimethyl mark on lysine 9 of histone H3 in Neurospora. Knowledge that HPl, a protein first identified in Drosophila (see Chapter 5), binds methylated lysine 9 of histone H3 in vitro motivated a search for an HP 1 hom*olog in Neurospora. A likely hom*olog was found, and its involvement in DNA methylation was tested by gene disruption (Freitag et al. 2004a). The gene, named hpo (HPone) was indeed found to be essential for DNA methylation. As another test of whether Neurospora HP1 reads the mark generated by DIM-5, its subcellular localization was examined in wild type and dim-5 strains. In wild type, HP 1-GFP localized to heterochromatic foci, but this localization was lost in dim-5, confirming that Neurospora HP1 is recruited by the trimethyl-lysine 9 mark generated by DIM-5. Evidence that RNA interference (RNAi) is important for heterochromatin formation and maintenance in S. pombe raised the question of whether the RNAi machinery of Neurospora is involved in HP1 localization' and/or DNA methylation. Neurospora has hom*ologs of a variety of genes implicated in RNAi (Borkovich et al. 2004). Studies of mutants with null mutations in all three RNA-dependent RNA polymerase (RdRP) genes, in both Dicer genes, or in other presumptive RNAi genes revealed no evidence that RNAi is involved in methylation of H3K9, heterochromatin formation, or DNA methylation in Neurospora (Chicas et al. 2004; Freitag et al. 2004c). However, as discussed below, the Neurospora RNAi genes are involved in at least two other silencing mechanisms with epigenetic aspects, quelling and meiotic silencing. 2.4 Quelling

Soon after transformation techniques were established for Neurospora, researchers in several laboratories noticed that a sizable fraction (e.g., ~ 30%) of Neurospora transformants showed silencing of transforming DNA and, more surprisingly, silencing of native

118 •

CHAPTER

copies of transforming DNA in a tight array. Nuclei flow freely in hyphae of Neurospora, allowing "heterokaryosis" in which genetically distinct nuclei share a common cytoplasm. Thus, it was easy to demonstrate that quelling is "dominant"; i.e., a transformed nucleus can silence hom*ologous sequences in nearby nuclei (Cogoni et al. 1996). This implicated a cytoplasmic silencing factor, perhaps an RNA species. Consistent with this possibility, identification of genes involved in quelling revealed that quelling is closely related to RNAi in other systems (Pickford et al. 2002). Specifically, qde-l, qde-2, qde-3, dcl-l, and dcl-2 encode, respectively,

sequences p*rnologous to those of the transforming DNA. The latter form of vegetative phase silencing was named "quelling" by the Macino laboratory, which has carried out most of the research on this phenomenon to date (Pickford et al. 2002). Quelling is most apparent with visible markers such as the albino genes, which encode enzymes required for carotenoid biosynthesis (Fig. 13), and is thought to be comparable to "cosuppression" or posttranscriptional gene silencing (PTGS) in plants. Interestingly, genes seem to vary in their sensitivity to quelling. For genes that are sensitive, quelling seems most common in transformants bearing multiple

e ~@0@0

asex ",' spores

€) A

quelled untransformed nucleuy

",",f"~.d

transformation with al DNA '"

~",!-",I,"~ ~!no quelling

quelling

[@ @/

RIP mitotic proliferation of sexual dlkaryon

wild type

'W ti5 [@ @)! 'CJ

[@ @![@@![@ ----

----

RIP

~f @/~

----

sexual spores

Figure 13. Queiling For simplicity, only two of the seven chromosomes are diagrammed (straight line segments in gray circles representing nuclei). The native albino gene (af) is indicated by the dark orange rectangle on the top chromosome; the other (dark orange or yellow) rectangle represents ectopic al sequences introduced by transformation. Since transformed cells are often multinucleate, transformants are often heterokaryotic, as illustrated. Whether or not the transforming DNA includes the entire coding region, in some transformants it silences ("quells") the native gene in both transformed and nontransformed nuclei through an undefined trans-acting molecule (red lines emanating from the transforming DNA indicated by the yellow rectangle). This results in poorly pigmented or albino (An tissue in some transformants, as shown.

FUNGAL

an RdRP, an "Argonaute"-like protein thought to be involved in small interfering RNA (siRNA)-guided mRNA cleavage (Baumberger and Baulcombe 2005), a RecQ-like presumptive DNA helicase (qde-3) and two "Dicers," which are presumably involved in generating siRNAs (Catalanotto et al. 2004). Although critical details are lacking, the developing model is that, in some cases, transforming DNA generates an "aberrant" transcript that somehow triggers the RNAi machinery, leading to degradation of hom*ologous mRNAs. Although DNA methylation is frequently associated with transforming DNA, neither the DNA methyltransferase, DIM-2, nor the histone H3 lysine 9 methyltransferase, DIM-5, is required for quelling (Cogoni et al. 1996; Chicas et al. 2005).

MODELS

FOR

EPIGENETIC

Meiotic Prophase

paired wt alleles

deletion allele

ectopic allele

paired null allele

2.5 Meiotic Silencing by Unpaired DNA (MSUD)

MSUD, the most recent addition to the list of silencing mechanisms operative in Neurospora, was discovered by Metzenberg and colleagues (Aramayo and Metzenberg 1996; Shiu et al. 2001; Shiu and Metzenberg 2002) while they were investigating the curious observation that a deletion mutation in the Asm-l (ascospore maturation) gene is functionally dominant. An elegant series of experiments, cartooned in Figure 14, led to the conclusion that sequences that lack a pairing partner in meiotic prophase cause meiotic silencing of identical, or nearly identical, sequences. A possible explanation for the observation that a deletion of Asm-l is dominant (Fig. 14b) was that the remaining single dose of the gene produced inadequate gene product, but this possibility was rendered unlikely by the observation that a functional copy of the gene at an ectopic location failed to complement the defect (Fig. 14c). Conversely, only one functional copy of the gene was required in meiosis if the functional allele had a pairing partner; i.e., its partner could harbor mutations rendering it unable to produce a functional product (Fig. 14d). Normal meiotic expression of the gene was observed in a strain having paired ectopic alleles and deletions of the native gene on both hom*ologs, showing that the deletion was not itself "toxic" and that the ectopic copies were indeed functional (Fig. 14e). Interestingly, some alleles generated by RIP elicit MSUD if the strains are proficient for DNA methylation but fail to elicit MSUD in dim-2 strains, suggesting that the DNA methylation frequently associated with such alleles can inhibit pairing (compare d and f in Fig. 14) (Pratt et al. 2004). (Incidentally, this observation also provided the first evidence that DIM-2 is functional in the sexual phase of Neurospora.) Alleles with

RESEARCH.

e paired ectopic alleles

methylated allele

unmethylated heavily RIP'd allele

h extra ectopic allele

Expression?

® ®® ®®== ® -

® ®-

119

Yes

Figure 14. Meiotic Silencing by Unpaired DNA (MSUD) Cartoon of tests conducted by the Metzenberg and Aramayo laboratories (Aramayo and Metzenberg 1996; Shiu et al. 2001; Pratt et al. 2004) that defined the phenomenon. Only two of the seven Neurospora chromosomes are indicated, with their sequences illustrated in green and blue, respectively. The rectangular box signifies a gene normally functional during meiosis. Vertical lines indicate mutations, and DNA methylation is indicated by red coloring. Deletions are diagramed as gaps. See text for interpretations.

high densities of mutations by RIP were found to be dominant, like deletions (Fig. 14g) (Shiu et al. 2001; Shiu and Metzenberg 2002; Pratt et al. 2004), consistent with the hypothesis that they are unable to pair with the wild-type allele. To distinguish between the possibility that MSUD is due to absence of pairing and the possibility that it is due to presence of an unpaired allele, the researchers analyzed a cross in which the meiotic nucleus would have three copies of a gene: two wild-type alleles (which should pair)

120

C HAP T E R 6

and an ectopic copy (which should be unpaired). Silencing was observed, implying that MSUD results from the presence of unpaired alleles rather than from the absence of paired alleles (Fig. 14h). MSUD can be observed cytologically in crosses heterozygous for a gene encoding a GFPtagged protein, as illustrated in Figure 15. A hunt for mutants defective in MSUD resulted in the identification of a telling member of the MSUD machinery. The Sad-l (Suppressor of ascus dominance) gene, identified by selection for mutants that were able to pass through a cross in which asm-l is not paired, encodes an RdRP (Shiu et al. 2001). This suggested that MSUD is related to quelling in Neurospora, and to RNAi generally. Interestingly, the Neurospora genome contains genes predicted to encode three putative RdRPs (one required for quelling, one required for MSUD, and one of unknown function), two Argonaute-like proteins (one required for quelling and one required for MSUD), and two Dicer-like proteins. It will be fascinating to learn the detailed mechanism of MSUD.

2.6 Probable Functions and Practical Uses of RIp, Quelling, and M5UD

RIP seems custom-made to limit the expression of "selfish DNA" such as transposons that direct the production of copies of themselves in a genome. Consistent with this possibility, the vast majority of relics of RIP are recognizably similar to transposons known from other organisms,

and most strains of Neurospora lack active transposons (Galagan et al. 2003; Selker et al. 2003). Nevertheless, because RIP is limited to the premeiotic dikaryotic cells, this process should neither prevent the spread of a new (e.g., horizontally acquired) transposon in vegetative cells nor prevent the duplication of a single-copy transposon in meiotic cells. Quelling and MSUD should deal with such eventualities, however. Although quelling does not completely suppress the spread of transposons in vegetative cells, as evidenced by the proliferation of an introduced copy of the LINE-like transposon, Tad, it does appear to partially silence such transposons (Nolan et al. 2005). Information about the action of MSUD suggests that this process will silence any transposed sequence in meiotic cells, even if it is only present as a single copy in the genome (Shiu et al. 2001). In addition to dealing with errant transposons in meiosis, MSUD also appears to play an important role in the process of speciation, as shown by the observation that mutants defective in MSUD relieve the sterility of strains bearing large duplications of chromosome segments and allow closely related species to mate with N. crassa (Shiu et al. 2001). Although RIP, quelling, and MSUD can all be a nuisance for some genetic experiments, all have been exploited for research purposes. RIP provided the first simple method to knock out genes in Neurospora and is still the preferred method for generating partial-function mutants. Quelling has also been used to reduce, if not eliminate, gene function, much as RNAi is exploited in a

Figure 15. Neurospora crassa

Fluorescent image of a rosette of maturing asci from a heterozygous cross of a transformant engineered to express GFP-tagged histone Hl. Four ascospores of each ascus show glowing nuclei (Freitag et al. 2004b). (Photo courtesy of N. B. Raju, Stanford University.)

FUN GAL

variety of organisms. And MSUD provides a simple assay to test whether particular genes are required to function in (or immediately after) meiosis; if a gene is found to cause sterility when duplicated, or when at an ectopic location, and the sterility is rescued by a mutation blocking MSUD, it is safe to assume that it plays an important function in meiosis. In addition to the postulated evolutionary roles of RIP, quelling, and MSUD, and to their utility in the laboratory, it is worth considering the possibility that these processes serve in other ways. For example, the fact that Sad-l function is required for full fertility suggests that MSUD is directly or indirectly required for meiosis (Shiu and Metzenberg 2002). In the case of RIP, although this process is nonessential, the distribution of products of RIP in the Neurospora genome suggests that junked transposons can serve the organism as substrates for kinetochore formation, much as repeated sequences do in S. pombe and other organisms. Analyses of DNA sequences from chromosome 7 illustrate that sequences around the genetically mapped centromeres of Neurospora consist primarily of relics of transposons heavily mutated by RIP

MOD E L 5

FOR

E PIG ENE T feR ESE ARC H

(Fig. 16). Relics of RIP are also found adjacent to telomere sequences of Neurospora. Interestingly, transposons and relics of transposons are also commonly found in heterochromatic sequences of other organisms, such as Drosophila (Chapter 5), mammals (Chapter 17), plants (Chapter 9), and other fungi. 3 Concluding Remarks

The fungi S. pombe and N. crassa have emerged as powerful systems to discover and elucidate epigenetic phenomena. The field of epigenetics is still in its infancy, with epigenetic mechanisms continuing to come to light. Therefore, it is not surprising that the depth and breadth of our current understanding of epigenetic processes, such as those described in this chapter, vary between organisms. It is too early to know how general the various epigenetic mechanisms described are. For example, it is possible that some organisms, like S. pombe, rely primarily on RNAi for silencing and heterochromatin formation, whereas others, such as N. crassa, rely more heavily on DNA methylation. It is already clear that even these two model eukaryotes have

TA/AT

(CM TG)/(AC+GT) - - -

3 Ratio

o Genes Repeats

\~,V''',"III'oWYi',Y',I/''lllll//r/III~''I·'NN-'I,~\

\ l,'

r 1J.,-L ," r r r u"~\.w.l!JIIII/M'f'M.wYlIV\IIt\ffI\""~,.N."I~mr,'W'''YllrINA I

TE relics

' ••• 1111.'_'''•• [

0.2

0.4

121

''_.1

[

0.6

0.8

1.0

1.2

1.4

Mb Figure 16. Organization of Centromere 7 Region of N. crassa Contigs 249, 255, and 21 of genomic sequence release 7.0 (http://www.broad.mit.edu/annotation/fungi/ neurospora_crassa]/index.html), and all except the first 400 kb of sequence contig 10, were assembled, and the combined sequence file was analyzed in 200-bp increments for the "RIP indices" (TpA/ApT [blue] and CpA+TpG/ApC+GpT [red]) (Galagan et al. 2003; Selker et al. 2003; Galagan and Selker 2004). An -360-kb region with a high density of transposable elements (TE) inactivated by RIP (retrotransposon relics in blue; DNA transposon relic in violet) was found between markers flanking the centromere, which was mapped genetically. The -1 .5-Mb segment shown includes 383 annotated genes (above and below line to indicate genes in opposite orientations), of which only 20 short predicted genes are within the predicted centromere region. The sizes of sequence gaps between the contigs (positions 0.5466, 0.6956, and 0.9058 Mb in the figure) are unknown.

122 • C HAP

TER 6

both important differences and striking similarities. Both organisms utilize histone H3K9 methylation and RNAi, neither of which is found in the budding yeast, S. cerevisiae. Of these three fungi, however, only Neurospora sports DNA methylation. It is also noteworthy that a given process may be functionally rather different in two organisms. For instance, in Neurospora, RNAi components have been implicated in quelling and meiotic silencing, but not in heterochromatin formation, whereas in fission yeast, RNAi components contribute to heterochromatin formation, but other roles are not established. Finally, it is worth noting that even shared features, such as heterochromatin associated with centromeres of both fission yeast and Neurospora, may have important differences. An important goal for the future is to discover the extent to which information gleaned from one organism is applicable to others. Further exploration of epigenetic processes in various model organisms, including S. pombe and N. crassa, will provide this information. We anticipate that the richly diverse fungi will continue to serve as useful systems for epigenetic research for many years.

Acknowledgments

R.C.A. thanks Alison L. Pidoux and Sharon A. White for Figure 4 and Figure 3, respectively; and Alison L. Pidoux and Elizabeth H. Bayne for comments on the manuscript. Research in the Allshire lab is supported by the Wellcome Trust, MRC, and the EC FP6 Epigenome Network of Excellence. R.C.A. is a Wellcome Trust Principal Research Fellow. E.U.S. thanks N.B. Raju for help assembling figures of Neurospora, and Robert L. Metzenberg for comments on the manuscript. Research in the Selker laboratory is supported by U.S. Public Health Service grant GM35690 from the National Institutes of Health and by National Science Foundation grant MCB0131383.

References AlIshire R.C., Javerzat J.P., Redhead N.J., and Cranston G. 1994. Position effect variegation at fission yeast centromeres. Cell 76: 157-169. AlIshire R.C., Nimmo E.R., Ekwall K., Javerzat J.P., and Cranston G. 1995. Mutations derepressing silent centromeric domains in fission yeast disrupt chromosome segregation. Genes Dev. 9: 218-233. Aramayo R. and Metzenberg R.L. 1996. Meiotic transvection in fungi. Cell 86: 103-113. Ayoub N., Goldshmidt 1., Lyakhovetsky R., and Cohen A. 2000. A fis-

sion yeast repression element cooperates with centromere-like sequences and defines a mat silent domain boundary. Genetics 156: 983-994. Bailis J.M., Bernard P., Antonelli R., Allshire R.C., and Forsburg S.L. 2003. Hskl-Dfpl is required for heterochromatin-mediated cohesion at centromeres. Nat. Cell BioI. 5: 1111-1116. Bannister A.J., Zegerman P., Partridge J.F., Miska E.A., Thomas J.O., AlIshire R.C., and Kouzarides T. 2001. Selective recognition of methylated lysine 9 on histone H3 by the HPI chromo domain. Nature 410: 120-124. Baum M., Ngan V.K., and Clarke L. 1994. The centromeric K-type repeat and the central core are together sufficient to establish a functional Schizosaccharomyces pombe centromere. Mol. BioI. Cell 5: 747-761. Baumberger N. and Baulcombe D.C. 2005. Arabidopsis ARGONAUTEI is an RNA Slicer that selectively recruits microRNAs and short interfering RNAs. Proc. Natl. Acad. Sci. 102: 11928-11933. Bernard P., Maure J.F., Partridge J.F., Genier S., javerzat J.P., and Allshire R.C. 2001. Requirement of heterochromatin for cohesion at centromeres. Science 294: 2539-2542. Bhat A. and Kasbekar D.P. 2001. Escape from repeat-induced point mutation of a gene-sized duplication in Neurospora crassa crosses that are heterozygous for a larger chromosome segment duplication. Genetics 157: 1581-1590. Borkovich K.A., Alex L.A., Yarden 0., Freitag M., Turner G.E., Read N.D., Seiler S., Bell-Pedersen D., Paietta J., Plesofsky N., et al. 2004. Lessons from the genome sequence of Neurospora crassa: Tracing the path from genomic blueprint to multicellular organism. Microbiol. Mol. BioI. Rev. 68: 1-108. Cam H.P., Sugiyama T., Chen E.S., Chen X., FitzGerald P.c., and Grewal S.1. 2005. Comprehensive analysis of heterochromatin- and RNAi-mediated epigenetic control of the fission yeast genome. Nat. Genet. 37: 809-819. Catalanotto c., Pallotta M., ReFalo P., Sachs M.S., Vayssie L., Macino G., and Cogoni C. 2004. Redundancy of the two dicer genes in transgene-induced posttranscriptional gene silencing in Neurospora crassa. Mol. Cell. BioI. 24: 2536-2545. Chen E.S., Saitoh S., Yanagida M., and Takahashi K. 2003. A cell cycleregulated GATA factor promotes centromeric localization of CENP-A in fission yeast. Mol. Cell 11: 175-187. Chicas A., Cogoni c., and Macino G. 2004. RNAi-dependent and RNAi-independent mechanisms contribute to the silencing of RIPed sequences in Neurospora crassa. Nucleic Acids Res. 32: 4237-4243. Chicas A., Forrest E.C., Sepich S., Cogoni c., and Macino G. 2005. Small interfering RNAs that trigger posttranscriptional gene silencing are not required for the histone H3 Lys9 methylation necessary for transgenic tandem repeat stabilization in Neurospora crassa. Mol. Cell. BioI. 25: 3793-3801. Clarke, L. and M.P. Baum. 1990. Functional analysis of a centromere from fission yeast: A role for centromere-specific repeated DNA sequences. Mol. Cell. BioI. 10: 1863-1872. Cleveland D.W., Mao Y, and Sullivan K.F. 2003. Centromeres and kinetochores: From epigenetics to mitotic checkpoint signaling. Cell 112: 407-421. Cogoni c., Irelan J.T., Schumacher M., Schmidhauser T.J., Selker E. U., and Macino G. 1996. Transgene silencing of the al-l gene in vegetative cells of Neurospora is mediated by a cytoplasmic effector and does not depend on DNA-DNA interactions or DNA methylation. EMBO]. 15: 3153-3163.

FUN GAL

Cowieson N.P., Partridge J.P., Allshire R.C., and McLaughlin P.j. 2000. Dimerisation of a chromo shadow domain and distinctions from the chromodomain as revealed by structural analysis. Curro BioI. 10: 517-525. Davis R.H. 2000. Neurospora: Contributions of a model organism. Oxford University Press, New York. Djupedall., Portoso M., Spahr H., Bonilla c., Gustafsson C.M., Allshire R.C., and Ekwall K. 2005. RNA Pol II subunit Rpb7 promotes centromeric transcription and RNAi-directed chromatin silencing. Genes Dev. 19: 2301-2306. Egel R. 2004. The molecular biology of Schizosaccharomyces pombe: Genetics, genomics and beyond. Springer-Verlag, Heidelberg, Germany. Ekwall K., Javerzat J.P., Lorentz A., Schmidt H., Cranston G., and AlIshire R. 1995. The chromodomain protein Swi6: A key component at fission yeast centro meres. Science 269: 1429-1431. Ekwall K., Nimmo E.R., javerzat j.P., Borgstrom B., Egel R., Cranston G., and Allshire R. 1996. Mutations in the fission yeast silencing factors clr4+ and rikr disrupt the localisation of the chromo domain protein Swi6p and impair centromere function.]. Cell Sci. 109: 2637-2648. Ekwall K., Olsson T., Turner B.M., Cranston G., and AlIshire R.C. 1997. Transient inhibition of histone deacetylation alters the structural and functional imprint at fission yeast centromeres. Cell 91: 1021-1032. Freitag M., Hickey P.c., Khlafallah TK., Read N.D., and Selker E.U. 2004a. HP1 is essential for DNA methylation in Neurospora. Mol. Cell 13: 427-434. Freitag M., Hickey P.c., Raju N.B., Selker E.U., and Read N.D. 2004b. GFP as a tool to analyze the organization, dynamics and function of nuclei and microtubules in Neurospora crassa. Fungal Genet. BioI. 41: 897-910. Freitag M., Lee D.W., Kothe G.O., Pratt R.j., Aramayo R., and Selker E.U. 2004c. DNA methylation is independent of RNA interference in Neurospora. Science 304: 1939. Freitag M., Williams R.L., Kothe G.O., and Selker E.U. 2002. A cytosine methyltransferase hom*ologue is essential for repeat-induced point mutation in Neurospora crassa. Proc. Natl. Acad. Sci. 99: 8802-8807. f*ckagawa T., Nogami M., Yoshikawa M., Ikeno M., Okazaki T, Takami Y., Nakayama T, and Oshimura M. 2004. Dicer is essential for formation of the heterochromatin structure in vertebrate cells. Nat. Cell BioI. 6: 784-791. Galagan J.E., Calvo S.E., Borkovich K.A., Selker E.U., Read N.D., Jaffe D., FitzHugh W., Ma L.J., Smirnov S., Purcell S., et al. 2003. The genome sequence of the fJlamentous fungus Neurospora crassa. Nature 422: 859-868. Galagan J.E. and Selker E.U. 2004. RIP: The evolutionary cost of genome defense. Trends Genet. 20: 417-423. Goshima G., Saitoh S., and Yanagida M. 1999. Proper metaphase spinclIe length is determined by centromere proteins Mis12 and Mis6 required for faithful chromosome segregation. Genes Dev. 13: 1664-1677. Grace Goll M. and Bestor TH. 2005. Eukaryotic cytosine methyltransferases. Annu. Rev. Biochem. 74: 481-514. Grewal S.l. and K1ar A.j. 1997. A recombinationally repressed region between mat2 and mat3 loci shares hom*ology to centromeric repeats and regulates directionality of mating-type switching in fission yeast. Genetics 146: 1221-1238. HallI.M., Noma K., and Grewal S.l. 2003. RNA interference machinery regulates chromosome dynamics during mitosis and meiosis in fission yeast. Proc. Natl. Acad. Sci. 100: 193-198.

MOD E L 5

FOR

E PIG ENE TIC

RES EAR C H

•

123

Hall l.M., Shankaranarayana G.D., Noma K., Ayoub N., Cohen A., and Grewal S.l. 2002. Establishment and maintenance of a heterochromatin domain. Science 297: 2232-2237. Hayashi T, Fujita Y., Iwasaki 0., Adachi Y., Takahashi K., and Yanagida M. 2004. Mis16 and Mis18 are required for CENP-A loading and histone deacetylation at centromeres. Cell 118: 715-729. Hong E.-j.E., Villen J., Gerace E.L., Gygi S., and Moazed D. 2005. A cullin E3 ubiquitin ligase complex associates with Rikl and the Clr4 histone H3-K9 methyltransferase and is required for RNAimediated heterochromatin formation. RNA BioI. 2: 106-111. Irelan J.T., Gutkin G.I., and Clarke 1. 2001. Functional redundancies, distinct localizations and interactions among three fission yeast hom*ologs of centromere protein-B. Genetics 157: 1191-1203. jia S., Noma K., and Grewal S.l. 2004. RNAi-independent heterochromatin nucleation by the stress-activated ATF/CREB family proteins. Science 304: 1971-1976. jin Q.w., Pidoux A.L., Decker c., AlIshire R.C., and Fleig U. 2002. The mal2p protein is an essential component of the fission yeast centromere. Mol. Cell BioI. 22: 7168-7183. Kanoh j., Sadaie M., Urano T, and Ishikawa P. 2005. Telomere binding protein Taz1 establishes Swi6 heterochromatin independently of RNAi at telomeres. Curro BioI. 15: 1808-1819. Kato H., Goto D.B., Martienssen R.A., Urano T, Furukawa K., and Murakami Y. 2005. RNA polymerase II is required for RNAidependent heterochromatin assembly. Science 309: 467-469. Kim H.S., Choi E.S., Shin J.A., Jang Y.K., and Park S.D. 2004. Regulation of Swi6/HP1-dependent heterochromatin assembly by cooperation of components of the mitogen-activated protein kinase pathway and a histone deacetylase Clr6. ]. BioI. Chem. 279: 42850-42859. Kouzminova E.A. and Selker E.U. 2001. Dim-2 encodes a DNA-methyltransferase responsible for all known cytosine methylation in Neurospora. EMBO ]. 20: 4309-4323. Kuhn R.M., Clarke L., and Carbon J. 1991. Clustered tRNA genes in Schizosaccharomyces pombe centromeric DNA sequence repeats. Proc. Natl. Acad. Sci. 88: 1306-1310. Measday Y. and Hieter P. 2004. Kinetochore sub-structure comes to MIND. Nat. Cell BioI. 6: 94-95. Mellone B.G., Ball L., Suka N., Grunstein M.R., Partridge J.P., and AlIshire R.C. 2003. Centromere silencing and function in fission yeast is governed by the amino terminus of histone H3. Curr Bioi 13: 1748-1757. Motamedi M.R., Verdel A., Colmenares S.u., Gerber S.A., Gygi S.P., and Moazed D. 2004. Two RNAi complexes, RITS and RDRC, physically interact and localize to noncoding centromeric RNAs. Cell 119: 789-802. Nakagawa H., Lee J.K., Hurwitz J., AlIshire R.C., Nakayama J., Grewal S.l., Tanaka K., and Murakami Y. 2002. Fission yeast CENP-B hom*ologs nucleate centromeric heterochromatin by promoting heterochromatin-specific histone tail modifications. Genes Dev. 16: 1766-1778. Nakayama j., K1ar A.J., and Grewal S.1. 2000. A chromodomain protein, Swi6, performs imprinting functions in fission yeast during mitosis and meiosis. Cell 101: 307-317. Nakayama J., Rice J.c., Strahl B.D., Allis C.D., and Grewal S.l. 2001. Role of histone H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science 292: 110-113. Ngan Y.K. and Clarke 1. 1997. The centromere enhancer mediates centromere activation in Schizosaccharomyces pombe. Mol. Cell. BioI. 17: 3305-3314. Nimmo E.R., Cranston G., and AlIshire R.C. 1994. Telomere-associated

124

•

C HAP T E R 6

chromosome breakage in fission yeast results in variegated expression of adjacent genes. EMBO f. 13: 3801-3811. Nimmo E.R., Pidoux A.L., Perry P.E., and Allshire R.C 1998. Defective meiosis in telomere-silencing mutants of Schizosaccharomyces pombe. Nature 392: 825-828. Nolan T., Braccini L., Azzalin G., De Toni A., Macino G., and Cogoni C 2005. The post-transcriptional gene silencing machinery functions independently of DNA methylation to repress a LINEl-like retrotransposon in Neurospora crassa. Nucleic Acids Res. 33: 1564-1573. Noma K., Allis CD., and Grewal S.l. 2001. Transitions in distinct histone H3 methylation patterns at the heterochromatin domain boundaries. Science 293: 1150-1155. Noma K., Sugiyama T, Cam H., Verdel A., Zofall M., Jia S., Moazed D., and Grewal S.l. 2004. RlTS acts in cis to promote RNA interference-mediated transcriptional and post-transcriptional silencing. Nat. Genet. 36: 1174-1180. Nonaka N., Kitajima T, Yokobayashi S., Xiao G., Yamamoto M., Grewal S.l., and Watanabe Y. 2002. Recruitment of cohesin to heterochromatic regions by Swi6/HP 1 in fission yeast. Nat. Cell BioI. 4: 89-93. Partridge J.E, Borgstrom B., and Allshire R.C 2000. Distinct protein interaction domains and protein spreading in a complex centromere. Genes Dev. 14: 783-791. Partridge J.E, Scott K.S., Bannister A.J., Kouzarides T, and Allshire R.C 2002. cis-acting DNA from fission yeast centromeres mediates histone H3 methylation and recruitment of silencing factors and cohesin to an ectopic site. Curro BioI. 12: 1652-1660. Perkins D.O., Margolin B.S., Selker E.U., and Haedo S.D. 1997. Occurrence of repeat induced point mutation in long segmental duplications of Neurospora. Genetics 147: 125-136. Perkins D.O., Metzenberg R.L., Raju N.B., Selker E.U., and Barry E.G. 1986. Reversal of a Neurospora translocation by crossing over involving displaced rONA, and methylation of the rONA segments that result from recombination. Genetics 114: 791-817. Perkins D.O., Radford A., and Sachs M.S. 2001. The Neurospora compendium; Chromosomal loci. Academic Press, San Diego, Califorma. Pickford A.5., Catalanotto C, Cogoni C, and Macino G. 2002. Quelling in Neurospora crassa. Adv. Genet. 46: 277-303. Pidoux A.L. and Allshire R.C 2004. Kinetochore and heterochromatin domains of the fission yeast centromere. Chromosome Res. 12: 521-534. - - - . 2005. The role of heterochromatin in centromere function. Philos. Trans. R. Soc. Lond. B BioI. Sci. 360: 569-579. Pidoux A.L., Richardson W., and Allshire R.C 2003. Sim4: A novel fission yeast kinetochore protein required for centromeric silencing and chromosome segregation. f. Cell BioI. 161: 295-307. Polizzi C and Clarke L. 1991. The chromatin structure of centromeres from fission yeast: Differentiation of the central core that correlates with function.]. Cell BioI. 112: 191-201. Pratt R.J., Lee D.W., and Aramayo R. 2004. DNA methylation affects meiotic trans-sensing, not meiotic silencing, in Neurospora. Genetics 168: 1925-1935. Raju N.B. 1980. Meiosis and ascospore genesis in Neurospora. Eur. f. Cell BioI. 23: 208-223. Reinhart B.J. and Bartel D.P. 2002. Small RNAs correspond to centromere heterochromatic repeats. Science 297: 1831. Rossignol J.-L. and Faugeron G. 1994. Gene inactivation triggered by recognition between DNA repeats. Experientia 50: 307-317. Sadaie M., Iida T., Urano T, and Nakayama J. 2004. A chromodomain protein, Chpl, is required for the establishment of heterochromatin in fission yeast. EMBO ]. 23: 3825-3835.

Scott K.C, Merrett S.L., and Willard H.E 2006. A heterochromatin barrier partitions the fission yeast centromere into discrete chromatin domains. Curro BioI. 16: 119-129. Selker E.U. 1990. Premeiotic instability of repeated sequences in Neurospora crassa. Annu. Rev. Genet. 24: 579-613. - - - . 1998. Trichostatin A causes selective loss of DNA methylation in Neurospora. Proc. Natl. Acad. Sci. 95: 9430-9435. Selker E.U., Cambareri E.B., Jensen B.C, and Haack K.R. 1987. Rearrangement of duplicated DNA in specialized cells of Neurospora. Cell 51: 741-752. Selker E.U., Freitag M., Kothe G.O., Margolin B.S., Rountree M.R., Allis CD., and Tamaru H. 2002. Induction and maintenance of nonsymmetrical DNA methylation in Neurospora. Proc. Natl. Acad. Sci. (supp!. 4) 99: 16485-16490. Selker E.U., Tountas N.A., Cross S.H., Margolin B.S., Murphy J.G., Bird A.P., and Freitag M. 2003. The methylated component of the Neurospora crassa genome. Nature 422: 893-897. Shankaranarayana G.D., Motamedi M.R., Moazed D., and Grewal S.l. 2003. Sir2 regulates histone H3 lysine 9 methylation and heterochromatin assembly in fission yeast. Curro BioI. 13: 1240-1246. Shiu P.K. and Metzenberg R.L. 2002. Meiotic silencing by unpaired DNA: Properties, regulation and suppression. Genetics 161: 1483-1495. Shiu P.K., Raju N.B., Zick.ler D., and Metzenberg R.L. 2001. Meiotic silencing by unpaired DNA. Cell 107: 905-916. Sigova A., Rhind N., and Zamore P.D. 2004. A single Argonaute protein mediates both transcriptional and posttranscriptional silencing in Schizosaccharomyces pombe. Genes Dev. 18: 2359-2367. Singer M.T., Marcotte B.A., and Selker E.U. 1995. DNA methylation associated with repeat-induced point mutation in Neurospora crassa. Mol. Cell. BioI. 15: 5586-5597. Steiner N.C and Clarke L. 1994. A novel epigenetic effect can alter centromere function in fission yeast. Cell 79: 865-874. Steiner N.C, Hahnenberger K.M., and Clarke L. 1993. Centromeres of the fission yeast Schizosaccharomyces pombe are highly variable genetic loci. Mol. Cell. BioI. 13: 4578-4587. Sugiyama T, Cam H., Verdel A., Moazed D., and Grewal S.l. 2005. RNA-dependent RNA polymerase is an essential component of a self-enforcing loop coupling heterochromatin assembly to siRNA production. Proc. Natl. Acad. Sci. 102: 152-157. Takahashi K., Chen E.S., and Yanagida M. 2000. Requirement of Mis6 centromere connector for localizing a CENP-A-like protein in fission yeast. Science 288: 2215-2219. Takahashi K., Murakami S., Chikashige Y, Funabiki H., Niwa 0., and Yanagida M. 1992. A low copy number central sequence with strict symmetry and unusual chromatin structure in fission yeast centromere. Mol. BioI. Cell 3: 819-835. Takahashi K., Murakami S., Chikashige Y, Niwa 0., and Yanagida M. 1991. A large number of tRNA genes are symmetrically located in fission yeast centromeres.]. Mol. BioI. 218: 13-17. Takahashi K., Takayama Y, Masuda E, Kobayashi Y, and Saitoh S. 2005. Two distinct pathways responsible for the loading of CENP-A to centromeres in the fission yeast cell cycle. Philos. Trans. R. Soc. Lond. B Bioi. Sci. 360: 595-607. Tamaru H. and Selker E. U. 2001. A histone H3 methyltransferase controls DNA methylation in Neurospora crassa. Nature 414: 277-283. - - - . 2003. Synthesis of signals for de novo DNA methylation in Neurospora crassa. Mol. Cell. BioI. 23: 2379-2394. Tamaru H., Zhang X., McMillen D., Singh P.B., Nakayama J., Grewal S.l., Allis CD., Cheng X., and Selker E.U 2003. Trimethylated

FUN GAL

lysine 9 of histone H3 is a mark for DNA methylation in Neurospora crassa. Nat. Genet. 34: 75-79. Thon G. and Verhein-Hansen J. 2000. Four chromo-domain proteins of Schizosaccharomyces pombe differentially repress transcription at various chromosomal locations. Genetics 155: 551-568. Volpe T., Schramke v., Hamilton G.L., White S.A., Teng G., Martienssen R.A., and Allshire R.C 2003. RNA interference is required for normal centromere function in fission yeast. Chromosome Res. 11: 137-146. Volpe TA., Kidner C., Hall l.M., Teng G., Grewal 5.1., and Martienssen R.A. 2002. Regulation of heterochromatic silencing and histone H3

MOD E L S

FOR

E PIG ENE TIC

RES EAR C H

•

125

lysine-9 methylation by RNAi. Science 297: 1833-1837. Watters M.K., Randall T.A., Margolin B.S., Selker E.U., and Stadler D.R. 1999. Action of repeat-induced point mutation on both strands of a duplex and on tandem duplications of various sizes in Neurospora. Genetics 153: 705-714. Wilkinson CR., Bartlett R., Nurse P., and Bird A.P. 1995. The fission yeast gene pmtl + encodes a DNA methyltransferase hom*ologue. Nucleic Acids Res. 23: 203-210. Yamada T, Fischle w., Sugiyama T, Allis CD., and Grewal 5.1. 2005. The nucleation and maintenance of heterochromatin by a histone deacetylase in fission yeast. Mol. Ce1/20: 173-185.

Epigenetics of Ciliates Eric Meyer' and Douglas L. Chalker2 'Laboratoire de Genetique Moleculaire, CNRS UMR 8541 Ecole Normale Superieure, 75005 Paris, France 2Deptartment of Biology, Washington University, St. Louis, Missouri 63130

CONTENTS 1. Ciliates: Single Cells with Two Different Genomes, 129 2. Conjugation: Reciprocal Fertilization Reveals non-Mendelian Inheritance, 129 3. Cytoplasmic Inheritance in Ciliates, 130

4. Cortical Patterning: A Case of Structural Inheritance, 131

5. Macronuclei and Micronuclei: A Model for Active Versus Silent Chromatin, 132

5.7

5.2.

Separation of Micro- and Macronuclear Histones Reveals Distinct Roles for Histone Variants, 732 Chromatin Modifications Correlate with Activity States, 733

6. Genome-wide Rearrangements Occur during Macronuclear Development, 134

6.7

9.7. Experimental Induction of Specific Deletions in the Developing Macronucleus, 737 10. Rearrangement Patterns Are Likely Determined by a Comparison of Germ-line and Somatic Genomes, 138

70.7 The d48 Paradigm: Epigenetic Inheritance of Alternative Rearrangements, 738 70.2 Epigenetic Inheritance of Experimentally Induced Deletions, 739 70.3 "Spontaneous" Elimination of Foreign Sequences Introduced into the Micronuclear Genome, 740 70.4 Experimental Rescue of Inherited Macronuclear Deletions, 740 70.5 Experimental Inhibition of IES Elimination in the Developing Macronucleus, 740

Internal DNA Deletion: Precise (lntragenic lESs) and Imprecise (lntergenic Repeats) Events, 735 Chromosome Fragmentation, 735

11. A trans-Nuclear Comparison of Whole Genomes Mediated by RNA Interference, 142

7. Mechanisms of Genome Rearrangements, 136

77.2 Transporting RNA from Maternal to Zygotic Macronuclei in Paramecium, 744

6.2

8. hom*ology-dependent Gene Silencing in Ciliates, 136

8.7 8.2

Transgene-induced Silencing, 737 Silencing Is Induced by Double-stranded RNA, 737

9. Genome Rearrangements Are Guided by hom*ology-dependent Mechanisms, 137

77.7 Linking Short RNAs to DNA Elimination, 742

12. Conclusions: DNA Elimination as a Genome Defense Mechanism, 144

13. Future Contributions of Ciliate Research to Epigenetics, 146 References, 147 127

GENERAL SUMMARY Anyone watching ciliates under a microscope is sure to be fascinated by these complex little animals that use their hair-like cilia to swim, eat, and find a mate. Vegetative growing cells duplicate by simple binary fission; yet periodically, ciliates will mate with a partner or, in some species, undergo self-fertilization, resulting in sexual progeny with a different genotype. What uniquely distinguishes these single-celled eukaryotes is that they maintain two functionally distinct genomes, carried in separate nuclei, within a common cytoplasm. The smaller of these, the micronucleus, contains the germ-line genome. It is transcriptionally silent during growth, but stores the genetic information that is passed to progeny at each sexual generation. The larger macronucleus performs somatic functions because it is responsible for all gene expression and thus governs the cell's phenotype. It is discarded at the end of each vegetative cycle when a new macronucleus differentiates from the germ line. During macronuclear development, massive DNA rearrangements generate a streamlined version of the genome ready for expression. A portion of the germ-line genome, including all the repetitive DNA that has long been considered "junk," is eliminated, while all the genes needed for the organism's survival throughout the life cycle are amplified to a high ploidy level. The ciliates' genetic oddities have made them very useful model organisms with which to discover and understand epigenetic mechanisms. In some species, the two sibling progeny that develop upon mating begin with identical genomes, but their somatic nuclei differentiate within the context of two different parental cells. This permits easy detection of hereditary characters that are not solely determined by the nuclear genome. Genetic experiments conducted with Paramecium tetraurelia, primarily those of Tracy Sonneborn, provided some of the earliest descriptions of non-Mendelian inheritance in any eukaryote. Among cases in which genetically identical siblings expressed different variants of specific traits, some were true cases of cytoplasmic inheritance (maternal inheritance of organelle DNA), but others, such as the inheritance of an individual's mating type, were of a different kind. A progeny's mating type is not determined by its genotype, but rather is specified by the preexisting type of the parental cell in which its somatic genome developed. In simple terms, different environments (the parental cytoplasm) direct expression of alternative traits from identical DNA complements. This is the hallmark of epigenetics. The peculiar genetic organization of ciliates also implies mechanisms that differentially regulate hom*olo-

gous sequences contained within the distinct nuclei. Early studies aimed to elucidate the means by which the germ line was kept silent and the somatic genome transcriptionally active. The compartmentalization of gene expression states offered researchers an opportunity to investigate the role of chromatin proteins and their modifications in epigenetic regulation. They could readily correlate specific histones and their modification with transcriptional activity or cell cycle stage. For instance, by comparing chromatin proteins from germ-line and somatic nuclei of Tetrahymena thermophila, some of the first histone variants were identified. Furthermore, new chromatin regulators, such as the first histone acetyltransferase (HAT), were identified in this ciliate, in part by taking advantage of the fact that only the macronucleus contains acetylated histones. Although ciliate genetics may seem unconventional, the underlying mechanisms are widely used for epigenetic regulation in eukaryotes, as illustrated by the role of RNA interference in whole-genome rearrangements. The extent and form of these rearrangements are remarkably diverse among ciliate species, yet one common feature is that they normally direct the elimination of transposonlike elements and other repetitive sequences. In both Paramecium and Tetrahymena, short RNAs are generated from the germ-line genome during meiosis. The discovery of these small RNAs, together with the demonstration in Tetrahymena that Argonaute and Dicer hom*ologs are required for DNA rearrangements, has led to the realization that an RNAi-like mechanism is involved. The small RNAs are thought to target histone H3 lysine 9 methylation to hom*ologous sequences, marking them for elimination. Thus, ciliate DNA rearrangements are mechanistically similar to the more broadly used RNA-directed establishment of heterochromatin. The use of RNAi to eliminate transposable elements further underscores the importance of this pathway as a genome defense mechanism. Furthermore, many experiments have shown that DNA rearrangement patterns are not strictly determined by the germ-line genome, but are controlled, at least in part, by preexisting rearrangements within the parental somatic genome. The implication is that the germ-line and somatic genomes are compared to each other during nuclear differentiation, a comparison that is likely mediated by hom*ology-dependent interactions between germ-line and somatic RNAs. Fully understanding this process will undoubtedly provide new insight into the roles of RNA in the epigenetic programming of the genome.

E PIG ENE TIC S

1 Ciliates: Single Cells with Two Different Genomes

Ciliates, which comprise a monophyletic lineage that emerged about one billion years ago (Philippe et al. 2000), were among the first unicellular eukaryotes to be used as genetic models. In the late 1930s, when T.M. Sonneborn discovered the mating types of Paramecium aurelia (Sonneborn 1937), the chromosome theory of inheritance elaborated by T.H. Morgan was still unsatisfying to many researchers, in particular embryologists (for historical detail, see Chapter 2). Unable to envision how such static entities as genes could be the sole basis of heredity, they believed that the cytoplasm had to be involved, if only to coordinate gene action (see Harwood 1985). Whereas mainstream geneticists largely focused on gene action, Sonneborn's early genetic analyses showed that the transmission of many heritable characteristics could not be fully explained by Mendel's laws. Due to their unique biology, the study of ciliates revealed some of the first examples of cytoplasmic inheritance and continues to provide new insights into epigenetic mechanisms. One of the most distinctive features of ciliates is nuclear dimorphism: Each cell contains two kinds of nuclei that differ in structure and function. The diploid micronuclei are transcriptionally silent during vegetative growth but contain the germ-line genome. These nuclei undergo meiosis to produce gametic nuclei that transmit the Mendelian genome to the next sexual generation (Fig. 1). In contrast, the higWy polyploid macronuclei are responsible for gene expression during vegetative growth and thus govern the cell's phenotype, but they are lost during sexual development and can therefore be considered the equivalent of the soma (Fig. 1). The numbers of nuclei of each type vary in different species. For example, P. aurelia species have two micronuclei and one macronucleus, whereas Tetrahymena thermophila has just one of each. Macro- and micronuclei divide by separate mechanisms. Micronuclei divide via conventional closed mitosis. Macronuclei, in contrast, divide by a poorly understood amitotic mechanism that does not involve spindle formation or visible condensation of the centromere-less, somatic chromosomes. After DNA synthesis, the macronucleus simply splits into two rougWy equal halves. There does not appear to be any mechanism to ensure equal segregation of macronuclear chromosomes to the two daughter cells. Instead, it is likely that the high ploidy level (-800n in P. tetraurelia, -45n in T. thermophila) prevents lethal gene loss for a number of vegetative divisions. Most species have a finite vegetative life span, and clonal cell lines will eventually die if they do not engage in sexual reproduction before they become senescent.

0 F elL I ATE S

•

129

2 Conjugation: Reciprocal Fertilization Reveals non-Mendelian Inheritance

Ciliates are hermaphroditic species capable of conjugation, a mating process that involves cross-fertilization between two parent cells. Mature cells of appropriate clonal age will become sexually reactive upon mild starvation and pair with cells of compatible mating types to initiate conjugation. If no compatible partner is available, some species will undergo a self-fertilization process called autogamy. In both cases, nuclear reorganization ensues, starting with meiosis of micronuclei. The sequence of nuclear events is similar in all species with some variations, and is depicted in Figure 1 for the P. aurelia and T. thermophila species (see Sonneborn 1975). Postmeiotic development starts with the selection of a single haploid nucleus in each cell to pass on the genome. The selected nucleus undergoes an additional division that produces two genetically identical gametic nuclei. In the case of conjugation, the two mates exchange one of their two haploid nuclei, and subsequent karyogamy (i.e., the fusion of two haploid nuclei) therefore generates genetically identical zygotic nuclei in each conjugant (stages 3-5 in Fig. 1). In autogamy, the two gametic nuclei within the single cell fuse to produce an entirely hom*ozygous diploid genome. In both cases, the resulting diploid zygotic nucleus (stage 5) divides twice more, and the four products differentiate, two into new micronuclei and two into new macronuclei (stages 6 and 7). Upon completion of nuclear development, either both new micronuclei are maintained in the new vegetative clones as occurs in P. aurelia species, or one of the two is degraded as in T. thermophila. In both species, the two new macronuclei do not divide during the first cellular division (stage 9) but are distributed to the two daughter cells; they start dividing only at the second vegetative division. While the parental micronuclei give rise to the new micro- and macronuclei of the next generation, the parental macronucleus is lost. In P. aurelia, it is fragmented into about 30 pieces in which DNA replication is rapidly inhibited, although transcription continues actively throughout the differentiation of the new macronuclei. When vegetative growth resumes, the fragments are distributed randomly to daughter cells until none is left (stage 9). In T. thermophila, the parental macronucleus does not fragment, but becomes pycnotic and is degraded by an apoptosis-like mechanism before the first vegetative division (Davis et al. 1992) (stages 7 and 8).

1 30 •

C HAP T E R 7

3 Cytoplasmic Inheritance in Ciliates The biology of P. aurelia conjugation is quite favorable for the detection of epigenetic phenomena associated with cytoplasmic effects. Whereas reciprocal fertilization generates genetically identical zygotic nuclei in the two conjugants, almost no cytoplasm is exchanged between them, which effectively distinguishes the action of nuclear genes from that of the most influential of environments, the cytoplasm of the mother cell (Fig. 2). Each genetic cross is

Paramecium

therefore equivalent to a study of monozygotic twins being born to different mothers. Sonneborn's studies revealed that phenotypic differences between two parental cells can be maintained in their respective progeny even though the latter have identical genotypes. In a few cases, the phenotypic differences were later found to be determined by extranuclear genes and would not today qualify as epigenetic phenomena. For instance, the deadly properties of killer strains of P. aurelia are due to endosymbiotic bacteria of the genus Caedibacter, harbored in the cytoplasm, that release a toxin killing sensitive strains (for reviews, see Preer et al. 1974; Pond et al. 1989). Other cases, such as the maternal inheritance of serotype, which requires mutually exclusive expression of one of several paralogous surface antigen genes, are clear cases of cytoplasmic influence on gene activity. As with serotypes, the two complementary mating types of P. tetraurelia, which are called 0 and E, exhibit a cytoplasmic pattern of inheritance. The 0 and E traits are terminally differentiated phenotypes that are determined during development of the somatic macronucleus from a totipotent germ line. After conjugation, a vegetative clone descended from the 0 parent is almost always mating type 0, whereas one arising from the E parent is almost always of type E, even though both exconjugants develop from identical zygotic genomes (Fig. 2b). Furthermore, when a large cytoplasmic bridge forms between the two conjugating cells, allowing a significant exchange of cytoplasm, the progeny of both parents usually develop as type E (Sonneborn 1977). Thus, a cytoplasmic factor

Ciliate Life Cycles

@~. tQ). 8

·~@~~4@ (

formation

.g.

Tetrahymena

Figure 1. life Cycles of Paramecium (top) and Tetrahymena (bottom) (Stage 7) Vegetative cells multiply by binary fission. Sexual development, stages 2-8, will initiate upon conjugation of two cells or autogamy (in Paramecium only). (Stages 2-3) Micronuclear meiosis ends with selection of one

of the haploid products as the gametic nucleus and degeneration of those remaining. In Paramecium, the parental macronucleus starts forming lobes. (Stages 4-6) Zygote formation. An additional division of the selected nucleus produces two genetically identical haploid nuclei. During conjugation, one of the two identical gametic nuclei is exchanged between the two mates and subsequent karyogamy produces the diploid zygotic nucleus (red). During autogamy, the two identical gametic nuclei simply fuse together. Two additional postzygotic divisions (6) produce the undifferentiated micro- and macronuclei. (Stages 6-8) Nuclear differentiation. After the second postzygotic division, two of the resulting nuclei become the new micronuclei, while the other two begin differentiating into new macronuclei (pink). In Paramecium, the maternal macronucleus is fragmented. In Tetrahymena, it becomes pycnotic. Also in Tetrahymena, one of the new micronuclei degenerates. (Stage 9) Caryonidal division: This first vegetative division is special, as new macronuclei are distributed to the daughter cells without division while micronuclei are segregated to progeny by mitosis. Finally, fragments of the Paramecium parental macronucleus are nondividing, but remain until lost through random distribution during subsequent fissions.

EP/GENETICS

a. Mendelian segregation of a pair of alleles

CILIATES

131

b. Maternal inheritance of mating types

~x© I

Figure 2. Mendelian vs. Cytoplasmic Inheritance

conjugation \

(a) Conjugation and autogamy are illustrated by a cross between two Paramecium cells, each hom*ozy-

conjUgation\

@@ @ @ I\ I\ I\ F1

autogamy

autogamyl \

F2~@ ~© F2~~ ©© must exist that directs development of the E type. As introduced below, the germ-line genome undergoes extensive DNA rearrangements during macronuclear development, and although the putative mating-type gene has yet to be identified, one Mendelian mutation affecting mating-type determination has been shown to perturb these genome rearrangements at other loci (Meyer and Keller 1996). If regulated genome rearrangements determine mating type, the E-determining cytoplasmic factor should have the capacity to direct an alternative rearrangement of the mating-type gene, resulting in a macronuclear form of the gene that specifies type E. During conjugation, this form must also produce the E-determining cytoplasmic factor required for its further inheritance. As described in Section 10, alternative rearrangement patterns can be transmitted from maternal macronuclei to zygotic macronuclei, a transnuclear effect that appears to be mediated by RNA molecules acting in a hom*ology-dependent manner. The cytoplasmic factor responsible for non-Mendelian inheritance is thus likely to be an RNA molecule that controls the developmental "mutation" of the mating-type gene. 4 Cortical Patterning: A Case of Structural Inheritance

Studies of the complex architecture of the cell cortex revealed another form of non-Mendelian inheritance. The Paramecium cell is covered with about 4000 cilia arranged in longitudinal rows of anchored units (Fig. 3a).

gous for different alleles, M or m. Conjugation involves the reciprocal exchange of one of two identical gametic nuclei. This results in F, exconjugants with identical heterozygous genotypes. Autogamy, a self-fertilization process, generates an entirely hom*ozygous genotype in just one sexual generation such that these F, individuals have a 50% chance of becoming M/M or m/m. (b) Phenotypic difference between F, clones reveals cytoplasmically inherited characteristics. In Paramecium, mating type (0 or E) is irreversibly determined during the development of the somatic macronucleus (large circle) from the totipotent germ-line micronucleus (small circle); however, the parental macronucleus directs differentiation of each exconjugant toward maintaining the existing mating type.

Each cilium is rooted in a basal body or kinetosome, a complex structure with both antero-posterior and leftright asymmetries. As cells divide, the duplication of basal bodies is constrained by the structure of ciliary units, so that the new basal bodies remain in the same orientation (Fig. 3b). However, surgical grafting of a small patch of cortex in the reverse polarity will direct the eventual formation of a complete inverted ciliary row as the grafted basal bodies duplicate (Fig. 3c). This antero-posterior inversion will be propagated for an indefinite number of vegetative cell divisions and will be maternally inherited during conjugation (Beisson and Sonneborn 1965). These grafting experiments showed that genes are not responsible for the inheritance of such structural variation, and revealed the essential role of preexisting structures for the correct assembly of new structures. The oriented duplication of the centriole (Beisson and Wright 2003), and the propagation of flagellar shape upon cell division of trypanosomes (Moreira-Leite et al. 2001), represent other examples. Prions further exemplify self-propagating protein conformations that are responsible for cytoplasmic inheritance in yeast and mammals (Shorter and Lindquist 2005). Even the centromere of eukaryotic chromosomes behaves as a self-replicating protein complex that resides on DNA but is not determined by it (Cleveland et al. 2003; see Chapter 14). What these epigenetic phenomena tell us is that not all cellular structures can be assembled de novo by simply reading the information contained in genes. In a broader sense, replication of DNA itself is a case of structural inheritance, but the

1 32

C HAP T E R 7

~~ ~~

INNNN

Figure 3. Structural Inheritance of Cortical Unit Polarity in Paramecium (a) Immunolabeling of basal bodies and ciliary rootlets highlights the regular organization of parallel cortical rows of wildtype cells. Shown are ventral (left) and dorsal (right) views. (b) Dorsal view of a cell exhibiting disruption of the regular organization due to the reversed antero-posterior polarity of a few rows of cortical units. Basal bodies immunolabeled in red, and ciliary rootlets in green. (c) Enlargement of a patch of cortex shows the reversed orientation of ciliary rootlets in inverted rows (I) relative to normal rows (N). (d) Schematic of basal bodies (green circles) duplication during growth; each is shown flanked on its right side by an anteriorly oriented ciliary rootlet (purple) and two microtubular ribbons. Duplication occurs with a fixed geometry: Each new basal body is positioned anterior of its parent, ensuring identical polarity. (e) The repeated duplication of basal bodies within each row maintains hom*ogeneous orientation indefinitely. (Photographs courtesy of janine Beisson.)

genome is certainly not the only structure that dividing cells need to duplicate. Thus, far from being a rare curiosity, "epigenetic" structural inheritance may be viewed as one of the most fundamental mechanisms of life. 5 Macronuclei and Micronuclei: A Model for Active Versus Silent Chromatin One basic concept in epigenetics is that individual copies of a DNA sequence can possess different activities and that differential states can be stably maintained. The nuclear dimorphism of ciliates is a natural example of hom*ologous sequences that are maintained in a common cytoplasm, yet possess opposite activity states. The macronucleus serves as a model for the transcriptionally active state, the micronucleus for the repressed or silent state (Fig. 4). Early biochemical and immunohistochemi-

cal studies, primarily in Tetrahymena, aimed to compare the properties of these different nuclei, both in vegetative cells and during sexual development, in order to define how these different activity states might be determined, particularly at the level of chromatin structure. 5.1 Separation of Micro- and Maeronuclear Histones Reveals Distinct Roles for Histone Variants

Isolation of histone proteins separately from the macronucleus and micronucleus of Tetrahymena led to the discovery of specific histone variants. Histone variants hvl and hv2 that are now known to correspond to the H2A.Z and H3.3 variants of other eukaryotes, respectively (see Chapter 13), reside exclusively within macronuclei, which provided an early indication that these variants are important for maintaining transcrip-

EPIGENETICS

properties - silent - diploid - chromosome condensation - active - polyploid

CILIATES

133

chromatin modifications

histone composition

- H3S10ph - micH1ph

- H2A, H2B, H3, H4 - micH1

--'----_......

developing mac - H2A, H2B, H3, H4 - macH1

- variants hv1 (H2A.Z) hv2 (H3.3)

tional activity (Allis et a1. 1980; Hayashi et a1. 1984). The hv2 (H3.3) variant was shown to be constitutively expressed, a property critical for its deposition into chromatin outside of S phase, allowing this isoform of histone H3 to serve as a replacement histone (Yu and Gorovsky 1997). In addition to the different complement of core histone variants, the macro- and micronucleus contain different linker histones. The macro nuclear HI has similar amino acid composition and biochemical properties to the linker histones of other eukaryotes but lacks a central globular domain (Wu et a1. 1986; Hayashi et a1. 1987). Neither linker histone gene is essential for cell viability. Interestingly, gene knock-outs of either cause the volume of their respective nuclei to increase, indicating that both perform roles in overall chromatin compaction, possibly by stabilizing higher-order chromatin structure (Shen et a1. 1995). Loss of the macronuclear HI-like protein also leads to gene-specific changes in expression, implicating this linker histone in maintaining proper transcriptional regulation (Shen and Gorovsky 1996).

- H3K4me - H3K9me,H3K27me - (H2A,H2B,H3,H4)ac - H3K4me - (H2A,H2B,H3,H4)ac - (H2A, macH1)ph

Figure 4. Nuclear Dimorphism of Ciliates The germ-line micronucleus, the developing macronucleus, and the somatic macronucleus contain different histone complements and modifications. Those known to occur specifically in each or in the developing somatic genome are listed.

ity and followed their purification using an in-gel assay. For this assay, purified histones were polymerized into the polyacrylamide matrix of the gel used to separate purified protein fractions. After electrophoresis, the proteins were renatured and incubated with radiolabeled acetyl-CoA to reveal a polypeptide with an apparent molecular weight of 55 kD that could incorporate the label into the histone matrix. The real breakthrough came after microsequencing the purified protein and cloning the gene. This Tetrahymena HAT was found to be hom*ologous to a well-characterized transcriptional regulator of baker's yeast, the Gcn5 protein. Before this discovery, transcriptional activators were primarily thought to act by recruiting RNA polymerase to promoters, but this work established that transcriptional activators may also possess enzymatic activity, modifying chromatin or other transcriptional regulators, thus changing the state of the template. The door was opened, and many known regulators were quickly thereafter shown to act as HATs.

METHYLATION

5.2 Chromatin Modifications Correlate with Activity States ACETYLATION

The hyperacetylation of histones in the macronucleus and absence of this modification from the micronucleus have provided further evidence correlating this posttranslational modification with gene activation (Vavra et a1. 1982). The enzymes that performed histone acetylation in any organism remained unknown until the mid1990s, when C. David Allis and coworkers purified the first type A (nuclear) histone acetyltransferase (HAT) (Brownell and Allis 1995; Brownell et a1. 1996). These researchers started with highly purified macronuclei to separate this activity from type B cytoplasmic HAT activ-

The nuclear dimorphism of ciliates again proved advantageous in elucidating roles of histone methylation. This modification is restricted to the macronuclei in growing Tetrahymena cells (Fig. 4). The histone lysine methyltransferase (HKMT) activity purified from these nuclei specifically modified histone H3 at lysine 4 (H3K4me), providing an early correlation between this specific modification and transcriptional activity (Strahl et a1. 1999). Histone H3 lysine 9 methylation is absent from vegetatively growing cells but occurs specifically during macronuclear development on germ-line-limited sequences that are eliminated from the somatic genome (Taverna et a1. 2002). The developmentally regulated establishment of H3K9me2 (dimethylated) on these spe-

134 • C HAP

T ER

cific sequences provides a useful model with which to elucidate the targeting of this modification to the equivalent of heterochromatin (described in detail below). PHOSPHORYLATION

Purification oe 2P-radiolabeled histones from micro- and macronuclei showed that H3, H2A, and linker histones were highly phosphorylated (Allis and Gorovsky 1981). Multiple sites of macronuclear HI are phosphorylated, and this modification was shown to participate in the regulation of specific gene transcription (Mizzen et al. 1999). Using mutational analysis, Dou and Gorovsky found that this requirement for phosphorylation could be mimicked by the addition of charged amino acids into HI (Dou et al. 1999). The charged residues, however, did not need to be present in the corresponding positions of the phosphorylated amino acid, but the complementary effect required a cluster of charged sites (Dou and Gorovsky 2000, 2002). These studies indicated that phosphorylation per se was not required, but that a critical charge density promoted proper transcription. A single position, serine 10, is phosphorylated in histone H3 (H3SlOph) (Wei et al. 1998). This modification is cell-cycle-dependent and is correlated with mitosis in many eukaryotes. In Tetrahymena, it is restricted to micronuclei during mitosis and meiosis. Replacing the normal histone H3 gene with a mutant form containing an alanine substitution at serine 10 (SlOA) causes defects in micronuclear division resulting in lagging chromosomes and aneuploidy (Wei et al. 1999). Macronuclear amitotic division, however, is not affected. These results demonstrated that H3 phosphorylation plays an important role in chromosome condensation and/or segregation. The unique nuclear dimorphism of the ciliate again revealed key insight into the role of a chromatin modification.

at specific stages of macronuclear development (for reviews, see Prescott 1994; Coyne et al. 1996; Klobutcher and Herrick 1997; Jahn and Klobutcher 2002). Thus, the nuclear differentiation of the germ-line and somatic genomes involves a physical reorganization of chromosomes in addition to establishing the distinct chromatin states just described. Two types of rearrangements are commonly observed and are virtually unique to all ciliates that have been studied, internal DNA deletion and chromosome fragmentation (Fig. 5). Instances of this in other eukaryotes can be exemplified by localized VDJ locus recombination in the lymphocytes of mammals (see Chapter 21). Depending on the species, these events eliminate between 10% and 95% of the germ-line genome from each newly formed macronucleus. Virtually all repeated sequences, including transposable elements, are eliminated, which can in part account for the high macronuclear gene densities observed. The numbers of rearrangement sites also vary between species, from about 6,000 in Tetrahymena to perhaps as many as 100,000 in hypotrichous ciliates. The genome-wide distribution of these highly regulated and reproducible events provides an opportunity to investigate how cells identify and direct action on particular

Paramecium: precise and imprecise deletions

..-rA~A

...-

{::::::::::::::::::f·:·.; . G~:T3)" Tetrahymena: imprecise deletions and chromosome fragmentation

6 Genome-wide Rearrangements Occur during Macronuclear Development

The sequencing of the somatic (macronuclear) genomes of P tetraurelia and T. thermophila revealed very high gene numbers (-40,000 and -27,000, respectively) despite relatively small genome sizes (-72 Mb and -104 Mb, respectively). This organization is consistent with a genome that is optimized for efficient gene expression (see Fig. 15 of Chapter 3 for relative comparisons with other eukaryotic organisms). This "streamlining" of the somatic genome is achieved by massive DNA rearrangements of the germ-line-derived chromosomes that occur

Figure 5. DNA Rearrangements of Ciliates During development of a new macronucleus, extensive chromosome fragmentation and DNA elimination occur. (Top) DNA rearrangements occurring in Paramecium include both precise deletion of TA-bounded lESs (colored bars found in coding [arrow] and noncoding regions) and imprecise deletions (orange bar) that result alternatively in DNA deletion or fragmentation (G4T3\. (Bottom) In Tetrahymena imprecise deletion of lESs (colored bars) occurs at about 6000 loci, and chromosome fragmentation is specified by a conserved 15-bp sequence, the CBS (star).

E P f G ENE TIC S

DNA segments. Studies of ciliate DNA rearrangements have already revealed unique insights into epigenetic mechanisms, particularly regarding hom*ology-dependent processes. The following description of DNA rearrangement events should provide the background sufficient for the subsequent discussion of the associated epigenetic regulation.

F C f L I ATE S

•

135

on average) hom*ologous repeats, shared by cognate MDS ends, that are unrelated in sequence to those at other MDS ends. Although these long repeats certainly contribute to accurate unscrambling, it is not clear that they are sufficient, which has led to the proposal that a preexisting template may be involved (Prescott et al. 2003). IMPRECISE DELETIONS

6.7 Internal DNA Deletion: Precise (lntragenic lESs) and Imprecise (lntergenic Repeats) Events PRECISE DELETIONS

Precise deletions are those that occur at the same nucleotide positions in all copies of a macronuclear chromosome. The internal eliminated sequences (lESs) are short, single-copy DNA segments that are primarily removed from coding sequences, but are also found in intergenic or intronic regions of germ-line chromosomes (Fig. 5). Precisely excised lESs are bounded by short direct repeats, which typically vary in sequence between species. A prominent class, the so-called "TA" lESs found in Paramecium and some spirotrichs such as Euplotes crassus, are identified as having invariable 5'-TA-3' repeats at their boundaries, one copy of which remains within the macronuclear locus after excision (Betermier 2004). The few nucleotide positions internal to the TA dinucleotides are not random and form a degenerate consensus that resembles the ends of Tel/mariner transposons. Thus, these lESs may be evolutionarily derived from ancient insertions of such mobile elements (Klobutcher and Herrick 1997). Nevertheless, many IES ends conform poorly to the consensus sequence whereas many perfect matches can be found in macronuclear sequences that are not excised, indicating that the consensus does not contain sufficient information to specify excision of the approximately 60,000 lESs per haploid genome of P. tetraurelia, raising the question of how they are recognized. An amazing variation to precise excision occurs in the stichotrichs, a subgroup of the spirotrichs, in which IES removal occurs simultaneously with gene "unscrambling" (Prescott 1999). In the micronuclear version of scrambled genes, the macronuclear destined sequences (MDSs, i.e., the DNA "exons") are not only separated by lESs, but are also disordered relative to the linear arrangement found in the reorganized macronuclear sequence. In the germ line, two MDSs that will be joined to form the expressed gene can be located far apart, sometimes in unlinked loci (Landweber et al. 2000), and may even be in an inverted orientation relative to each other. The precision of reordering appears to be guided by relatively long (11 bp

Whereas lESs are efficiently and reproducibly removed from the somatic genome, in some cases, the deletion boundaries formed by independent excision events in the polyploid nucleus vary in position (Fig. 5). This heterogeneity extends over tens of base pairs in Tetrahymena and up to several kilobase pairs in Paramecium. Imprecise deletion is characteristic of all studied Tetrahymena lESs and is primarily responsible for the removal of repeated sequences such as intergenic transposons or minisatellites of Paramecium (Le MoueI et al. 2003). Like precise IES excision, these deletions typically occur between short direct repeats, one of which is maintained in the macronuclear sequence. In Tetrahymena, the repeat sequences are higWy variable among the known lESs, whereas in Paramecium, the repeats always contain at least one TA dinucleotide, suggesting that the mechanism involved may be related to that of precise IES excision. 6.2 Chromosome Fragmentation

During macronuclear development, the germ-linederived chromosomes are fragmented into shorter molecules that are capped by de novo addition of telomeric repeats (Fig. 5). The resulting macronuclear chromosomes apparently lack centromeres, and thus, chromosome fragmentation may facilitate the equal distribution of these molecules during amitotic divisions of the macronucleus. The extent of fragmentation varies widely among species. In spirotrichs it is carried out to an extreme, producing tiny "nanochromosomes" that typically contain single genes, whereas in Paramecium and Tetrahymena, macronuclear chromosomes range in size from 20 kb to over 1 Mb and contain many genes. The process is imprecise in most ciliates, so that the exact position of telomeric repeat addition is heterogeneous and often results in some loss of germ-line sequence. One exception is Euplotes crassus, in which telomeres are always added at the same nucleotide positions (Klobutcher 1999). In Tetrahymena, a conserved 15bp chromosome breakage sequence (CBS) that is found in an estimated 280 loci within germ-line chromosomes

1 36 •

C HAP T E R 7

(Cassidy-Hanley et al. 2005; Hamilton et al. 2005) is both necessary and sufficient to direct fragmentation and new telomere addition (Fan and Yao 1996). In contrast, no analogous CBS has been identified in P. aurelia species. Rather, all evidence indicates that chromosome fragmentation is an alternative outcome of imprecise deletion (Fig. 5) such that DNA elimination is healed either by the rejoining of flanking sequences or by telomere addition (Le MoueI et al. 2003). 7 Mechanisms of Genome Rearrangements

Although ciliate genome rearrangements have been studied for more than 30 years, the molecular mechanisms that carry out these events remain largely unknown (Yao et al. 2002). Part of the challenge for researchers has been to explain the diversity of events that take place among the different ciliates. Excision intermediates as well as circular excision (by)products have been described for several ciliates (Jaraczewski and Jahn 1993; Klobutcher et al. 1993; Williams et al. 1993; Saveliev and Cox 1995, 1996; Betermier et al. 2000; Gratias and Betermier 2003). The data do not allow a unified excision mechanism to be deduced, even for those IESs of different ciliates that have similar consensus sequences at their termini (see Klobutcher and Herrick 1995; Gratias and Betermier 2001; Betermier 2004). Despite the diversity of excision mechanisms, there still may exist significant overlap in the events that guide DNA rearrangements. One commonality is that transposon-like and repetitive sequences appear to be preferentially eliminated. As described elsewhere throughout this book, one role of epigenetic mechanisms is to suppress the activity of these potentially deleterious DNA elements. Allowing a transposon to escape from the silent germ line into the highly active macronucleus is potentially disastrous to the somatic genome. Many lines of evidence implicate mechanisms related to RNAi and heterochromatin formation in these processes of DNA rearrangement (Mochizuki and Gorovsky 2004; Yao and Chao 2005). In most eukaryotes, methylated histone H3 at lysine 9 (H3K9me) is widely associated with the repressed DNA that is partitioned in the nucleus as heterochromatin (see Section 7 of Chapter 3). In Tetrahymena, this modification is not found in the transcriptionally silent micronucleus as one might presume, but is exclusively found in developing macronuclei immediately preceding and concurrent with DNA rearrangement (Taverna et al. 2002). Chromatin immunoprecipitation experi-

ments have shown that this modification is enriched on the histones associated with IESs. DNA elimination and heterochromatin formation were originally linked by the identification of the chromodomain-containing Programmed DNA Degradation 1 (Pddl) protein, an abundant, developmentally expressed protein that colocalizes within foci containing germ-line-limited DNA to be eliminated (i.e., IESs) from the somatic genome (Madireddi et al. 1994, 1996). Chromodomains are protein motifs that have binding affinity to certain methylated histone residues. Perhaps the archetypal model of chromodomain protein involvement in chromatin regulation is the binding of heterochromatin protein 1 (HP 1-chromodomain containing) to methylated H3K9 in Drosophila, involved in the formation of heterochromatin domains (for more details, see Chapter 5). The chromodomains of Pdd1 and Pdd3, two proteins required for DNA rearrangement (Coyne et al. 1999; Nikiforov et al. 2000), bind to H3K9me2 peptides (Taverna et al. 2002). To demonstrate that this chromatin modification is required for DNA rearrangement, Liu, Mochizuki, and Gorovsky used hom*ologous gene replacement to substitute the major histone H3 genes with copies that contain a K9Q lysine 9 substitution mutation (Liu et al. 2004). These cells could not efficiently remove IESs during development despite the fact that Pdd1 localized appropriately within the precursors of the macronuclei, thus showing that H3K9me2 is required for DNA elimination. Because the establishment of chromatin states is a key determinant of ,epigenetic regulation, the important question to answer is, How is the H3K9me2 mark specifically targeted to the DNA segments destined for elimination? As described below, several experiments have demonstrated that hom*ologous RNAs are involved in guiding DNA rearrangements. 8 hom*ology-dependent Gene Silencing in Ciliates

hom*ology-dependent, RNA-mediated silencing mechanisms are widely used in eukaryotes for epigenetic regulation. Evidence that such mechanisms are active in ciliates was first observed in Paramecium after transformation of the vegetative macronucleus with nonexpressible transgenes that produced phenocopy of Mendelian mutants in the endogenous genes. Similar effects were then reproduced in Paramecium and in spirotrichs by feeding cells double-stranded RNA, suggesting the involvement of RNAi pathways. One of these pathways leads to the ultimate in gene silencing, DNA elimination.

E PIG ENE TIC 5

8.1 Transgene-induced Silencing

The Paramecium macronucleus is easy to transform by microinjection because any introduced DNA fragment can be maintained at a wide range of copy numbers, replicating autonomously without the need for any specific origin. Transformation with high-copy, nonexpressible transgenes can trigger posttranscriptional silencing of endogenous genes that possess sufficient sequence similarity (Ruiz et al. 1998; Galvani and Sperling 2001). Silencing is not observed if the 3' UTR of the gene is present in the transgene (Galvani and Sperling 2001), which suggests that regulatory signals present in the RNA influence a construct's silencing capacity. Subsequently, silencing was found to correlate with the accumulation of hom*ologous short RNAs approximately 23 nucleotides (nt) in length (Garnier et al. 2004), indicating that an RNAi pathway is involved. The -23-nt short RNAs appear to be responsible for the targeted degradation of hom*ologous mRNAs and may thus be called siRNAs (short interfering RNAs). A similar class of 23- to 24-nt RNAs has also been identified in vegetative Tetrahymena cells, and although there are no data about their possible roles, it is likely that they represent endogenous siRNAs (Lee and Collins 2006).

8.2 Silencing Is Induced by Double-stranded RNA

Double-stranded RNA (dsRNA) is likely the primary trigger for the transgene-induced silencing observed in Paramecium. The silencing efficiency of transgenes correlates with the production of aberrant RNA molecules that correspond to both the sense and the antisense strands of the injected sequence. Furthermore, the ability of dsRNA to promote gene silencing was demonstrated by feeding Paramecium cells Escherichia coli expressing dsRNA of a cloned gene using methodology developed for Caenorhabditis elegans (Timmons and Fire 1998; Timmons et al. 2001). Silencing of the endogenous gene can be observed phenotypically after as little as three vegetative divisions; i.e., less than 24 hours (Galvani and Sperling 2002). Feeding heat-killed E. coli to spirotrich species that normally feed on algae also promotes gene silencing (Paschka et al. 2003), suggesting that a wide variety of ciliates have this mechanism. In Paramecium, molecular analyses showed that feeding dsRNA leads to the accumulation of the same -23-nt siRNAs as observed upon transgene-induced silencing (Nowacki et al. 2005), indicating that both phenomena rely on a common RNAi pathway. Silencing induced by dsRNA feeding in Paramecium can be reversed immediately by replacing E. coli with the

0 F elL I ATE 5

•

137

normal food bacterium in the culture medium; similarly, direct microinjection of dsRNA into the cytoplasm induces only transient silencing of the hom*ologous genes, presumably because the injected dsRNA is rapidly diluted out during vegetative growth (Galvani and Sperling 2002). These observations suggest that dsRNA molecules cannot be amplified to any significant degree in the cytoplasm of vegetative cells, unlike the apparent fate of dsRNA in C. elegans that can lead to a heritable silent state. This further implies that RNAi in Paramecium does not lead to the establishment of stable transcriptional gene silencing in the macronucleus. Heritable silencing would likely require histone H3K9 methylation, and as mentioned above, this modification is apparently absent from the vegetative macronucleus, at least in T. thermophila.

9 Genome Rearrangements Are Guided by hom*ology-dependent Mechanisms

During ciliate development, three different genomes must be distinguished and channeled toward disparate fates: The germ-line micronuclear genome must be preserved intact while the developing somatic genome in the new macronucleus is directed to undergo extensive reorganization, and the maternal somatic genome is destined for destruction. Within this broader framework, ciliate researchers have aimed to understand the reproducibility of DNA rearrangement patterns. Initial efforts attempted to identify cis-acting DNA sequence motifs that could recruit the recombination proteins, but these, searches have had few clear successes (see Yao et al. 2002; Betermier 2004). Primary DNA sequence is clearly not the sole determinant guiding reorganization of the macronuclear genome, a conclusion that is supported by evidence that H3K9 methylation marks DNA segments for elimination. Furthermore, many rearrangement patterns are sensitive to hom*ology-dependent effects, as described below, that allow alternative patterns to be maternally inherited in subsequent sexual generations, independently from the Mendelian transmission of the wild-type germ-line genome. The inheritance of rearrangement patterns therefore satisfies the definition of an epigenetic phenomenon. 9.1 Experimental Induction of Specific Deletions in the Developing Macronucleus

Even before posttranscriptional gene silencing was described in Paramecium, the introduction of cloned sequences at high copy number into the vegetative macronucleus was found to alter the DNA rearrangements

138 •

C HAP T E R 7

in the sexual progeny of transformed clones. Strikingly, the sequences hom*ologous to the transgene were specifically deleted by the imprecise mechanism, while the micronuclear genome remained intact (Fig. 6) (Meyer 1992). This phenomenon appeared to be quite general because all tested DNA fragments could produce deletions (Meyer et al. 1997). These experiments indicated that sequence-specific information is transmitted through the cytoplasm during sexual events from the transformed maternal macronucleus to the developing zygotic macronucleus. The generality of the effect did not support interpretations that invoked a role for sequence-specific DNA-binding proteins produced from or titrated by the injected transgenes. The most parsimonious explanation that satisfies the observed specificity assumes that nucleic acids, presumably RNA molecules, are transferred between nuclei and recognize their targets by pairing interactions. The constructs that efficiently induce postzygotic deletions are nonexpressible transgenes, such as ones

a.Wild type autogamy Generation n

Generation n+ 1

mic

mac

b.Transgene-induced deletion upon autogamy

that contain frameshift mutations or truncations of 5' or 3' UTRs (Garnier et al. 2004). Those that can produce stable, translatable mRNAs do not promote the elimination of the hom*ologous genes from developing macronuclei. The same constructs that promote DNA deletion also caused the silencing of endogenous maternal genes during the vegetative growth of transformed clones. Thus, postzygotic deletions correlate with prezygotic silencing and with the accumulation of -23-nt siRNAs. The siRNAs were further shown to persist in the cells throughout development, suggesting that they may be responsible for triggering these deletions. If silencing-associated siRNAs direct DNA rearrangements, then introducing dsRNA prior to or during development should promote the elimination of the hom*ologous sequence within the descending progeny. To test this, Paramecium cells fed an E. coli strain producing dsRNA hom*ologous to the ND7 coding sequence, a gene involved in the regulated exocytosis of secretory vesicles called trichocysts, were allowed to undergo autogamy and develop new macronuclei from wild-type micronuclei. A number of postautogamous progeny showed an ND7 mutant phenotype due to elimination of the gene from their macronuclei (Garnier et al. 2004). Even phenotypically wild-type progeny showed partial elimination of ND7 gene copies. Like those associated with transgene-induced silencing, the -23-nt siRNAs associated with dsRNA feeding were shown to persist in the cells throughout autogamy (Nowacki et al. 2005), confirming their implication in the targeting of postzygotic deletions. Induced DNA deletion directed by dsRNA is not restricted to Paramecium. Microinjection of conjugating Tetrahymena with in vitrotranscribed sense and antisense RNA that is hom*ologous to genomic loci normally retained in the macronucleus resulted in the imprecise deletion of these sequences (Yao et al. 2003).

10 Rearrangement Patterns Are likely Determined by a Comparison of Germ-line and Somatic Genomes Figure 6. Transgene-induced Deletions in Paramecium (a) In the wild type, the A gene (red arrows) sits near the heterogeneous ends of a macronuclear chromosome; the pink boxes =

macronuclear telomeres. The rearrangement pattern of the micronuclear chromosome is faithfully reproduced from one generation to the next. (b) The introduction of a large copy number of A transgenes into the maternal macronucleus can induce the complete deletion of the endogenous A gene in sexual progeny, when a new macronucleus develops from the Wild-type germ line. The new macronuclear telomeres are positioned just upstream of the A gene.

10.1 The d48 Paradigm: Epigenetic Inheritance of Alternative Rearrangements

Although transgene and dsRNA-induced rearrangements illustrate the epigenetic nature of the DNA elimination process, what may be more striking are observations that induced rearrangement patterns can be inherited through subsequent rounds of macronuclear development. The first evidence for epigenetic regulation of rearrangements was uncovered when an aberrant deletion from the

E PIG EN E T f C S

macronucleus of the gene encoding the A surface antigen of P. tetraurelia was found to be cytoplasmically inherited in crosses with the wild type (Fig. 7) (Epstein and Forney 1984). In wild-type macronuclei, the A gene is located near a chromosome end either 8, 13, or 26 kb away from the telomere, dictated by three alternative fragmentation sites. A variant cell line called d48 was found to lack Agene expression because the gene itself was lost along with all downstream sequences as its telomere formed at the 5' end of the gene (Forney and Blackburn 1988). Nuclear transplantation experiments confirmed that the d48 germ-line micronucleus carried the wild-type A gene. For example, replacement of the d48 micronucleus with one from a wild-type strain did not prevent maternal trans-

0 F C f L I ATE S

139

mission of the A-gene deletion to sexual progeny; and conversely, the d48 micronucleus, when transplanted into a wild-type cell, gave rise after autogamy to a new macronucleus that contained the A gene (Harumoto 1986; Kobayashi and Koizumi 1990). Similar experiments that focused on the maintenance or deletion of another telomere-proximal surface antigen gene, the B gene, showed that such maternal effects can be observed in other genomic regions (Scott et al. 1994). Together, these studies strongly suggest that a genomic region must be present in the maternal macronucleus to be effectively retained and amplified to a wild-type copy number in the developing macronucleus. 10.2 Epigenetic Inheritance of Experimentally Induced Deletions

a. Wild type Generation n mic

[A~

Generation n+ 1 A

mac

b. d48

c. Rescued d48

Figure 7. Epigenetic Inheritance and Experimental Rescue of Macronuclear A-gene Deletions (0) Wild-type strain. (b) The d48 strain lacks the A gene in its

macronucleus, but has a wild-type micronucleus. The A gene is reproducibly deleted during macronuclear development in each generation. (c) Transformation of the macronucleus of the d48 strain with A-gene sequences will specifically restore amplification of the germline A gene in the developing macronucleus of sexual progeny.

Maternal influence on DNA rearrangements as observed for d48 appears to control the development of many, and possibly all, regions of the macronuclear genome of Paramecium. Indeed, the macronuclear deletions that are created experimentally in sexual progeny can be "spontaneously" reproduced in further sexual generations, following a maternal pattern of inheritance. This has been observed both for high-copy transgene-induced deletions of the G gene in P. primaurelia, another subtelomeric surface antigen gene (Meyer 1992), and for the macronuclear deletions of the ND7 gene that were initially induced by dsRNA feeding (Garnier et al. 2004). In either case, the inducing transgene or introduced dsRNA was no longer needed to propagate the deletion to sexual progeny. In d48 and in these induced variant cell lines, the state of the somatic genome can occasionally revert to the wild-type rearrangement pattern after autogamy, confirming that the gene is still present in the micronuclear genome and highlighting the epigenetic mode of inheritance. It is remarkable that gene silencing and the resulting developmental DNA deletion can be "remembered" in subsequent generations with the targeted genomic region treated like transposons and IESs during macronuclear development. The recurrent deletion of a gene in each sexual generation does not appear to be induced by -23-nt siRNAs, since these have only been detected when silencing is experimentally induced in the first generation. Furthermore, genetic analyses of such cell lines have shown that the micronuclear gene does not carry any permanent imprint, since it can be normally amplified during macronuclear development when it is transferred by conjugation into a cell line with a wild-type macronucleus. It therefore appears that the

140 • C HAP

TER 7

gene is deleted in the developing macronucleus simply because it is absent from the maternal macronucleus. 10.3 //Spontaneous// Elimination of Foreign Sequences Introduced into the Micronuclear Genome

The propagation of maternal rearrangement states suggests that the germ-line genome to be rearranged is compared to the existing rearranged genome, and the sequences. that were absent in the previous generation are targeted for elimination from the newly forming somatic genome. If this were indeed the case, transgenes introduced into the germ-line micronucleus would be predicted to be frequently deleted during new macronuclear development. In Tetrahymena, where micronuclear transformation has been achieved, researchers have observed that integrated drug resistance markers used for gene disruption studies can be deleted from the macronuclear genome during successive rounds of conjugation (Yao et al. 2003; Liu et al. 2005). The genomic location of the transgene, as well as the number of copies introduced into the germ-line genome, significantly altered the efficiency of elimination (Liu et al. 2005), which indicates that the phenomenon in this ciliate is less generally induced. Nevertheless, these results demonstrate that a foreign sequence, the bacterial neo gene, could be recognized by the cell as an IES. 10.4 Experimental Rescue of Inherited Macronuclear Deletions

Is the mere absence of a genomic region from the maternal somatic macronucleus sufficient to direct its future elimination? If so, then reintroducing the A gene into the d48 macronucleus should rescue the defect in A-gene propagation during development. It was first shown that injection of either wild-type macronucleoplasm into the d48 vegetative macronucleus (Harumoto 1986), or cytoplasm from autogamous, wild-type cells into d48 cells early in development (Koizumi and Kobayashi 1989), resulted in a permanent reversion of d48 to wild type. Subsequently, direct microinjection of several nonoverlapping A-gene fragments, spread over most of the -8-kb coding sequence, into the d48 macronucleus demonstrated that the A-gene sequence itself was sufficient to restore the wild-type macronuclear rearrangement pattern (Fig. 7) (Koizumi and Kobayashi 1989; Jessop-Murray et al. 1991; You et al. 1991). The maternal genome's influence on DNA rearrangements shows marked sequence specificity. The A and B gene coding sequences are 74% identical overall. Nevertheless, injection of the A -gene sequences into the

macronucleus of a cell line that carried macronuclear deletions of both genes could only prevent A-gene deletion from the new macronucleus; and similarly, injection with the B gene could only prevent its own deletion (Scott et al. 1994). In addition, the d48 macronuclear deletion could not be rescued by transformation with the G gene from P. primaurelia, which shares 78% identity with the A gene. On the other hand, the macronuclear deletion of the A gene could be rescued by transformation with a different allele of the A gene, showing 97% identity (Forney et al. 1996). Thus, the maternal rescue of macronuclear deletions is a hom*ology-dependent process that does not require any specific sequence within the genes, but requires a minimum level of sequence identity. The d48 maternal effect rescue, at first glance, appears to be at odds with the transgene-induced deletions, since in these experiments injection of A-gene sequences into the maternal macronucleus had exactly opposite consequences on the zygotic A gene. This apparent paradox was solved when it was shown that the postzygotic deletion effect depends on the establishment of hom*ologydependent silencing in the transformed clones (Garnier et al. 2004). Conversely, the rescue effect is observed only in transformation conditions that do not elicit silencing (i.e., only moderate copy numbers for nonexpressible constructs, or any copy number for expressible transgenes) (Garnier et al. 2004). Thus, it appears that the accumulation of the -23-nt siRNAs can prevent the rescue effect of A-gene sequences in the maternal macronucleus, which promote A-gene amplification in the devdoping macronucleus. This strongly suggests that the cytoplasmic factor mediating the trans-nuclear effect is an A-gene transcript. However, it is not necessarily the full-length mRNA, because even fragments of the coding sequence were shown to have rescue activity. Furthermore, clones containing the entire gene often express the mRNA throughout vegetative growth, whereas production of the rescuing cytoplasmic factor was shown to be restricted to the period of nuclear reorganization. 10.5 Experimental Inhibition of 1£5 Elimination in the Developing Macronucleus hom*oLOGY-DEPENDENT INHIBITION OF

IES

ELIMINATION IN PARAMECIUM

If the deletion or maintenance of cellular genes is controlled by a comparison of somatic and germ-line genome content, then even the normally efficient excision of IESs could perhaps be perturbed when copies are present in the maternal macronucleus. Examination of a Mendelian

EPIGENETICS

mutation, mtpE, which has pleiotropic effects on macronuclear development, including an effect on mating-type determination (Brygoo and Keller 1981a,b), provided the opportunity to test this prediction. This mutation, when hom*ozygous, was found to abolish the excision of an IES located within the G surface antigen gene. Surprisingly, when the wild-type allele of the mtP gene was reintroduced into the mutant strain by conjugation, the IES was still not excised during the subsequent macronuclear development (Meyer and Keller 1996). The genetic analysis of the resulting variant strain, which was called the IES+ strain, confirmed that it is genetically wild type and that the specific retention of this IES is maternally inherited in sexual progeny (Duharcourt et al. 1995). Is the excision of the IES within developing macronuclei induced by maternal copies of the correctly rearranged G gene or inhibited by maternal copies of the IES-retaining G gene? To answer this question, the macronuclei of IES+ or IES- cells were transformed by direct microinjection of plasmids containing a fragment of the G coding sequence in either its micronuclear (IES+) or macronuclear (IES-) versions (Fig. 8) (Duharcourt et al. 1995). After autogamy, transformed clones that contained the IES+ plasmid produced progeny lines that retained the IES in their newly formed macronuclei, while IES+ cells transformed with IES- plasmid proved unable to induce excision, and their progeny remained in the IES+ state. A plasmid containing only the IES, without any flanking sequences, also caused IES retention, showing that the maternal IES copies alone inhibit excision from zygotic macronuclei (Duharcourt et al. 1995). "MATERNALLY CONTROLLED" VERSUS "NON MATERNALLY CONTROLLED"

lESs

IN PARAMECIUM

Can the excision of other IESs be controlled by similar maternal effects? This question was addressed by transforming the macronuclei of wild-type cells with large seg-

Generation n

[u- ....J

Generation n+ 1

Generation t

CILIATES

141

ments of micronuclear DNA (IES+) containing either the G or A surface antigen genes (Duharcourt et al. 1998). The injected segments contained 6 and 9 IESs, respectively. Excision of 13 of these IESs was examined, and 5 IESs were found to be retained in the macronuclei of postautogamous progeny of the transformed cells. The injection of plasmids containing single IESs showed that inhibition was strictly specific: Each of these 5 IESs induced the retention of only the hom*ologous zygotic IES, but did not affect any other. A control DNA fragment containing most of the macronuclear (IES-) G gene had no effect on any of the IESs tested. Among the 13 Paramecium IESs tested, there is no obvious difference in size, base composition, or position within the genes between the 5 that show the maternal effect and the 8 that do not (Duharcourt et al. 1998). hom*oLOGY-DEPENDENT INHIBITION OF

IES

ELIMINATION IN TETRAHYMENA

Experiments analogous to those performed in Paramecium revealed that the DNA content of the Tetrahymena parental macronucleus can regulate the elimination of the hom*ologous sequences from the developing somatic genome. Two well-characterized IESs, the M and R deletion elements, were microinjected into the macronuclei of wild-type strains such that they were maintained on high-copy vectors. When these cells were induced to conjugate, the progeny of the cells containing maternal copies of the M element failed to efficiently eliminate the corresponding IESs during macronuclear development (Chalker and Yao 1996). Likewise, the cells whose parental macronuclei contained copies of the R element failed to excise the hom*ologous IES. Significant inhibition of excision of nonhom*ologous elements was not observed. Thus, the inhibition of DNA elimination was sequence-specific. Sequences hom*ologous to the IES itself were sufficient for this inhibition, and the immediately flanking DNA had no effect. Importantly, this

[u- ....J

6'0,,6

Generation t+1

E-

lost

II'.

Figure 8. hom*ology-dependent Inhibition of IES Excision by the Maternal Macronucleus During normal development, lESs (red and green bars) are excised efficiently. However, transformation of the maternal macronucleus with a given IES (initial transformants = Generation t) can inhibit the elimination of the hom*ologous IES during the subsequent (Generation t + 1) (and future) rounds of new macronuclear differentiation.

142 •

C HAP T E R 7

induced failure of DNA elimination was heritable, as subsequent generations also retained genomic copies of the IES in their macronuclei. Because Tetrahymena conjugating pairs extensively share cytoplasm during development, researchers were able to observe that the inhibition was transmitted through the cytoplasm such that only one mating partner need carry the IES in its parental somatic genome for the hom*ologous sequences in all developing nuclei within a pair, including those in a wild-type partner, to be affected. Therefore, the DNA rearrangement state of the parental nuclei is transmitted through the cytoplasm to regulate the events that occur during the formation of the new somatic genome of the next generation. No exchange of germ-line nuclei is required to transmit the state of one mating partner to another, an observation that rules out that transmission occurs by imposing an imprint on the germ line during normal cell growth prior to entering development (Chalker et al. 2005). In fact, by physically separating mating pairs consisting of one wild-type cell and a partner containing copies of the M element in its maternal somatic nucleus, it was shown that transmission occurs after meiosis and very near the time that developing nuclei first begin differentiating into the new macronuclei and micronuclei. Therefore, the influence of the maternal somatic genome is actively established during development.

BIOLOGICAL IMPLICATIONS OF MATERNAL CONTROL

The ability of ciliates to alter DNA rearrangements to reflect a maternal pattern provides cells with a simple way to transmit alternative somatic versions of the genome to sexual progeny. This dynamic regulation implies that a stable equilibrium between two alternative genetic states, e.g., IES+ (100% excision) and IES(0% excision), can be reached and maintained over the course of many sexual generations. This has been shown for at least one Paramecium IES for which the maternal influence is sufficient to drive a smaller fraction of macronuclear IES+ copies to a greater fraction in the macronucleus of the following generation (Duharcourt et al. 1995). Such influence parallels the stable, maternal inheritance of 0 and E mating types in P. tetraurelia. Comparable asymmetry is found in both systems, because both IES excision and O-type determination appear to be the default developmental pathways, whereas both IES retention and E-type determination require a cytoplasmic signal from the maternal macronucleus to alter their pathways. These systems may very well be related, because the Mendelian mtP

mutation, which impairs excision of one G gene IES, regardless of the parental state, similarly makes determination for E constitutive, an observation that links the alternative states of both examples. Elucidating the mechanism(s) that underlies these phenomena will undoubtedly reveal novel modes for maternal inheritance of epigenetic information. 11 A trans-Nuclear Comparison of Whole Genomes Mediated by RNA Interference

The hom*ology-dependent effects described above demonstrate that a cross talk occurs between the maternal somatic and germ-line genomes during nuclear differentiation, which can profoundly alter DNA rearrangement patterns. The observation that only highly hom*ologous sequences are affected suggests strongly that this cross talk is mediated by nucleic acids. This regulatory mechanism could possibly involve transcripts from the maternal macronucleus that are exported to the developing macronucleus where, if they contain an IES, they would prevent the elimination of the hom*ologous sequences. However, it was shown that putative maternal transcripts do not participate as donor templates during the repair of double-stranded breaks induced by constitutive IES excision (Duharcourt et al. 1995). If protective maternal transcripts exist, the -23-nt siRNAs that dictate gene silencing may target DNA deletion indirectly by promoting the destruction of those with sequence identity. This role for siRNAs alone cannot explain the maternal inheritance of rearrangement patterns in subsequent generations. As discussed below, it is likely the interplay between maternal macronuclear transcripts and a novel class of short RNAs derived from the meiotic micronucleus that ultimately directs genome reorganization. 11.1 Linking Short RNAs to DNA Elimination

In Tetrahymena, the canonical heterochromatin modification, H3K9 methylation, marks the chromatin associated with lESs just prior to their excision. How is this modification specifically targeted to these DNA segments destined for elimination? An RNAi-like pathway has been described that employs RNA guides to direct it to the proper loci (for review, see Mochizuki and Gorovsky 2004; Yao and Chao 2005). An initial indication that DNA deletion utilizes hom*ologous RNA molecules was the observation that Tetrahymena lESs are bidirectionally transcribed early during conjugation (Chalker and Yao 2001). A real breakthrough came when Mochizuki et. al.

E PIG ENE TIC 5

(2002) identified a PIWIIArgonaute family protein encoded by TWIl that is developmentally expressed and was required for DNA rearrangement (Mochizuki et al. 2002). Because Argonaute proteins are key players in RNAi-triggered processes, these researchers looked for and found a species of endogenous small (~28 nt) RNAs that are preferentially complementary to germ-line-limited sequences. Disruption of the TWIl gene destabilized these short RNAs and abolished H3K9 methylation in the developing somatic nucleus. This work, published in 2002, together with experiments in Schizosaccharomyces pombe (see Chapters 6 and 8), established a new paradigm that heterochromatin is generated by the action of short hom*ologous RNAs targeting the H3K9me2 silencing mark to specific loci. Characterization of the gene, DCLl, encoding the Dicer ribonuclease that processes the bidirectional transcripts into the ~28-nt small RNAs, has provided additional insight into the overall regulation of this process (Malone et al. 2005; Mochizuki and Gorovsky 2005).

=.--

~ ..............

-:::;-

0 F elL I ATE 5

•

143

The protein is expressed at high levels early in conjugation and localizes to the premeiotic micronuclei, which indicates that generation of the small RNAs is temporally and spatially compartmentalized within this germline nucleus. Disruption of the DCLl gene caused loss of small RNA production, accumulation of germ-line transcripts, and ultimately, failure of IES excision. Intriguingly, loss of the small RNAs did not abolish H3K9 methylation as was observed in the TWIl knockout lines. Nevertheless, chromatin immunoprecipitation analysis showed that this modification is no longer enriched on IESs; thus, the small RNAs are required to target this chromatin modification to the proper loci (Malone et al. 2005). To explain the role of the maternal genome in regulating these events, the scan RNA model was proposed (Mochizuki et al. 2002). In the variation of this model shown in Figure 9, the 28-nt "scan" (scn)RNAs, which are generated in the micronucleus, assemble with a Twil containing RISC-like complex in the cytoplasm and are ini-

'=.--

~ -<01IIII••••••••••••••••••••••••••

-:-:scan RNA production

n(A2C~

-:,-

! I

genome comparison

H3K9 methylation

DNA deletion

Figure 9. The Scan RNA Model for Control of DNA Deletion Bidirectional transcription of a large portion of the germ-line genome occurs early in development and leads to the production of scnRNAs. These are then transported into the maternal macronucleus where any encounter with a hom*ologous sequence will trigger their removal from the active pool. The remaining, micronucleus-specific RNAs are redirected to the developing macronucleus, where they target H3K9 methylation to hom*ologous sequences, signaling their removal from the genome. Model adapted from Mochizuki et al. (2002).

144 •

C HAP T E R 7

tially channeled to the maternal macronucleus. There they scan the existing rearranged genome for hom*ology. scnRNAs that pair with maternal sequences are removed from the pool of active complexes. The remaining Twilassociated scnRNAs are then transported to the developing macronucleus where they target H3K9 methylation to the hom*ologous sequences, marking them for excision by the DNA rearrangement machinery. This model is further supported by the observation that Twil localizes in the maternal macronucleus early in development after the bulk of small RNA production, but before the appearance of the new macronuclear precursors. Thus, the regulated trafficking of the small RNA protein complexes facilitates the comparison of somatic and germ-line genomes.

11.2 Transporting RNA from Maternal to Zygotic Macronuclei in Paramecium

The genome-wide comparison of germ-line and somatic sequences would require a highly sophisticated machinery, both to ensure the massive transport of RNA molecules between nuclei and to effect the very large number of pairing interactions implied by the scanning model. The novel nucleic-acid-binding protein Nowal appears to participate in this trans-nuclear cross talk in Paramecium (Nowacki et al. 2005). Nowal is synthesized shortly before meiosis and first accumulates in the maternal macronucleus and, like the Tetrahymena Twil, relocalizes later to the developing zygotic macronucleus. Tagging Nowal with a photoactivatable GFP allowed researchers to conclusively demonstrate that this protein is transported from one nucleus type to the other. One domain of the protein can bind RNA, and a second domain is necessary and sufficient for internuclear transport; thus Nowal may be an RNA transporter. Nowal is essential for the development of a viable new macronucleus, including the elimination of germline transposons and of a subset of IESs. Strikingly, only those lESs that are subject to maternal control are affected by Nowal knockdown, suggesting that the protein is involved in trans-nuclear genome comparison. IESs that are not sensitive to the presence of hom*ologous sequences in the maternal macronucleus do not depend on Nowal for their excision, confirming the existence of mechanistically distinct classes of IESs in Paramecium. Although the nucleic acids bound by Nowal in vivo are not yet known, the effects of Nowal depletion are consistent with the scanning model proposed (Fig. 9) and suggest that the protein may carry RNAs that have been

selected to target the elimination of maternally controlled lESs and transposons in the developing macronucleus. 12 Conclusions: DNA Elimination as a Genome Defense Mechanism

The study of RNAi, particularly in plants and nematodes, has led to the hypothesis that this pathway evolved as a defense mechanism that allows cells to control the proliferation of viruses and transposons by degrading mRNAs and targeting the formation of heterochromatin on these genomic parasites (Matzke and Birchler 2005). As already mentioned, transposable elements present in the germ-line genome of ciliates are eliminated during development of the somatic macronucleus, which will effectively negate their impact. The observation in Tetrahymena that H3K9 methylation marks genomic regions for developmental DNA deletion suggests that the use of RNAi in ciliates is fundamentally similar to its role in establishing heterochromatin in other eukaryotes; ciliates just go one step farther and eliminate heterochromatin from their somatic genome. Nevertheless, one original contribution of ciliate studies is the idea that the specific sequences that are targeted by RNAi for heterochromatin formation/elimination during early development are selected by a global comparison of maternal germline and somatic genomes, starting during meiosis. This mechanism would efficiently protect against the deleterious effects of transposition in the germ line: Any new transposon integrating into the maternal germ line would be recognized as alien by comparison with the somatic genome during sexual reproduction, leading to its removal from the transcribed somatic genome of progeny, thereby limiting its future spread. The Tetrahymena scnRNAs that target DNA elimination likely mediate the trans-nuclear cross talk between the germ-line and somatic genomes. It is unclear whether the whole micronuclear genome produces scnRNAs, but unpublished evidence obtained in Paramecium suggests that at least a large fraction of it does. Northern blot analyses have revealed meiosis-specific, endogenous short RNAs that correspond to cellular genes, as well as to transposons and IESs (M. Nowacki et al., unpubl.). These short RNAs are -25 nt in length and are clearly distinct from the -23-nt siRNAs. They appear to be functionally equivalent to the Tetrahymena scnRNAs, because the inactivation of specific Paramecium Dicer-like genes suppresses their production during meiosis and abolishes the

EPIGENETICS

145

CILIATES

1. Genomic transcription (wi siRNA PTGS)

.......... ..............................

~...........~

~I:=

•••••• at •••••••••• ,• • • • • • • • • • • • • • • • ,1

. ,".

••• I •••••••••• I ••••• ~

I •••• I •• , •••• '

....

..I ,

2. Germline sRNAIMaternal transcript comparison

-00IIIIII

AA b -

•••-

=====:==~;, =~==: =:===

• • • • • • • • • • • 11. . . • • • • • • • • • • • • • • • 11

[Jo-. .~ _

fiI

--=': =-==-==-=_=.= = : - : -=.- == -:: - : - : - = - - : , - :. :. =---=--- - --=--- .....................,

• • • 11 • • • • • • • • • • • • :• • • • ,

~~ 1111.1

.... W

..,

0..

---

•• I

2. Germline sRNAIMaternal transcript comparison /

...................... ,

rw •••••• III • • • • • • ~I • • • • • • • • • • • •11

•

-= ==== =="::.=~ =====:==:~=:==:===== :.:~-- ..:;::::::::::::::::::, ::::::::::::::!;.,,:~:, . . . . . :.:.:= .,,==-===_ =_ _ __ .. '-"'~" - -----,"W" _ _

3. Germline-specific sRNA transport

===' =..===....:. ==:-:. ................ ..................... '

-==--...... -....

::===

"1:... :7I1~~ ••••

•• 1.1.111

-........ . . . . ." . "-..,==-==_ .. ..

...

:- : -

=====:==:~

::=:=

1, •• " ••• " •••••••••• ' ~ '

III-.........-

...;;... ::-..~~

....~....-

--0

0--

Figure 10. The Transcriptome Scanning Model for the Comparison of Germ-line and Somatic Genomes

(a) Default regulation. (7) Upon initiating meiosis, most or all of the macronuclear and micronuclear genomes is transcribed; the dashed lines in the macronucleus represent uncharacterized transcripts. In the micronucleus, transcription is bidirectional, resulting in the production of scnRNAs (short double-stranded molecules) for all types of sequences (cellular genes, light purple arrows; transposons, orange double arrow; IES, green boxes). (2) The scnRNAs are exported to the maternal macronucleus, where they may pair with hom*ologous somatic transcripts. Pairing may also occur in the cytoplasm. (3) scnRNAs that pair with hom*ologous transcripts are sequestered or destroyed, while the micronucleus-specific ones are re-exported to the developing zygotic macronucleus, where they pair with hom*ologous sequences (DNA or nascent transcripts), thus targeting H3K9 methylation to micronuclear-specific sequences (transposons and lESs). (4) The marked sequences are eliminated. (b) Effects of posttranscriptional silencing of a gene in the maternal macronucleus. Experimental induction of posttranscriptional silencing by high-copy transgenes or dsRNA results in the production of double-stranded siRNAs hom*ologous to that gene (in red). These siRNAs degrade the hom*ologous maternal somatic transcripts, so that hom*ologous scnRNAs will not be inactivated and will be free to target the deletion of the gene in the developing zygotic macronucleus.

146 •

C HAP T E R 7

elimination of transposons and maternally controlled IESs, as well as the maternal inheritance of macronuclear gene deletions (Y. Serrano et al., unpubl.). The original genome-scanning model proposed that scnRNAs are compared to the rearranged DNA in the maternal macronucleus, inactivating those matching macronuclear sequences and selecting for the micronucleus-specific pool (Fig. lOa). However, several lines of evidence suggest that scnRNAs may be compared with macronucleus-derived transcripts rather than with DNA itself. This would dispense with the need to open the DNA duplex along the entire genome to test for complementarity. IESs introduced into the maternal macronucleus of Tetrahymena that interfere with DNA deletion are transcribed, allowing for this possibility (Chalker and Yao 2001; Chalker et al. 2005). Transcripts arising from an IES in the maternal macronucleus would protect the zygotic IES from elimination, as initially postulated to explain IES inhibition in Paramecium, by titrating the hom*ologous scnRNAs. The evidence for a cytoplasmic factor that can rescue the deletion of the A gene in the d48 strain also supports the existence of protective transcripts, and further suggests that scnRNA selection could occur in the cytoplasm as well as in the maternal macronucleus. Finally, this would explain how the posttranscriptional silencing of a given gene in Paramecium can cause its deletion in the next generation: If maternal macronuclear transcripts are degraded by the 23-nt siRNAs, hom*ologous scnRNAs will be free to target deletions in the new macronucleus despite the presence of the gene in the maternal macronucleus. One interesting aspect of this modified hypothesis, which may be called the transcriptome-scanning model, is that all pairing interactions may occur between RNA molecules. If scnRNAs finally target heterochromatin formation by interacting with nascent transcripts at the hom*ologous locus, as is thought to occur in S. pombe and plants, these later pairing interactions need not be fundamentally different from those involved in the selection step. scnRNAs may simply pair with any available RNA after they leave the meiotic micronucleus. Pairing with the abundant macronuclear transcripts present in the cell early in development would efficiently remove the hom*ologous scnRNAs from the pool by sequestering or destroying them. By the time the zygotic macronucleus forms and starts transcribing its unrearranged genome, only micronucleus-specific scnRNAs would be left to pair with nascent transcripts. These pairing interactions may lead to DNA deletion because they occur in the developing macronucleus, or because they occur at the correct developmental stage.

In this model, the recognition of self (macronuclear sequences) versus non-self (genomic parasites in the germ line) is achieved by a simple developmental switch that alters the outcome of similar pairing interactions. This is conceptually similar to the process by which the vertebrate immune system learns the distinction between self and non-self. A huge repertoire of lymphocytes expressing different antibodies is initially generated, but in early development all that recognize available antigens (likely self-antigens) are eliminated from the future pool. Once past this stage, the recognition of a cognate antigen (then likely to be alien) leads to the clonal expansion of the corresponding lymphocytes. The ciliate genomic immune system that utilizes RNAi thus has striking parallels to the cellular immune system of vertebrates. 13 Future Contributions of Ciliate Research to Epigenetics

The recent sequencing of the macronuclear genomes of Tetrahymena and Paramecium will sustain the utility of these facile unicellular models for investigation of epigenetic mechanisms. The biology of nuclear dimorphism has facilitated the discovery of chromatin regulators and will continue to provide an advantageous system for novel findings. With two distinct RNAi pathways, one used for posttranscriptional gene silencing and another for epigenetic modification of the genome, ciliates are also uniquely poised to unravel the complexity of RNA-guided, hom*ology-dependent regulatory processes involved in development. Dicer-related ribonucleases, Argonaute-like proteins, and their partners have been or will be identified, and current studies promise to decipher their functional specialization. Given the phylogenetic position of ciliates on the tree of life, such information will undoubtedly shed light on the evolution of these mechanisms in eukaryotes. It is the investigation of multiple epigenetic phenomena that led to an appreciation of the widespread use of RNA-mediated regulation in eukaryotes. Yet, even prior to this understanding, the observation that the DNA content of one nucleus imparts effects on hom*ologous sequences in another forced ciliate researchers to postulate the existence of trans-acting RNA molecules directing hom*ology-dependent cross talk. The discovery of RNAi and its role in DNA rearrangements has validated this speculation. What future insight of general relevance might be gained from uncovering the mechanisms that mediate trans-nuclear cross talk? If the "transcriptome scanning" hypothesis (Fig. 10) is correct, maternal somatic transcripts protect hom*ologous zygotic DNA sequences from elimination. If one equates

EPIGENETICS

DNA elimination with heterochromatin formation, then the maternal transcripts would, by blocking elimination, enforce euchromatin-specific modifications on hom*ologous sequences. The proposed positive role of these transcripts on hom*ologous genes is thus indirect, arising from their capacity to inactivate hom*ologous scnRNAs that would otherwise target the formation of heterochromatin. Nevertheless, this effect of maternal transcripts, which some experiments suggest are not necessarily protein-coding mRNAs, would represent a novel mechanism given that all known RNA-mediated, hom*ology-dependent mechanisms lead to the down-regulation of the target gene. The degradation or sequestration of short RNAs by long transcripts is, in essence, the reverse of the demonstrated RNAi-based mechanisms, whereby short RNAs inactivate long ones. If long and short RNAs can antagonize each other's action, their interaction must steer the system toward one of two alternative states, depending on their relative abundance. This suggests that the inheritance of the expression status of genes could depend on a constant feedback from an RNA pool, and not exclusively on semiconservative chromatin replication mechanisms. It is tempting to believe that such an RNA-mediated, trans-acting mechanism is responsible for the stability of surface antigen gene expression during vegetative growth in Paramecium, because this would also explain how, after sexual reproduction, the developing macronucleus can inherit the maternal serotype via the cytoplasm. Thus, the epigenetic inheritance of genome rearrangement patterns may be only one particular aspect of a comprehensive RNA-based inheritance system in ciliates. As future ciliate research illuminates the players (RNAs and proteins) mediating this regulation, its existence in other organisms and its further connection to known epigenetic mechanisms should become apparent. It is not implausible that RNA-based inheritance, in its basic form, occurs widely. Recent studies of the transcriptome output of multicellular eukaryotes has revealed an unexpected abundance and complexity of noncoding transcripts (Mattick 2004; Meyers et al. 2004; Suzuki and Hayashizaki 2004; Cheng et al. 2005; Claverie 2005), as well as the frequent occurrence of short RNAs matching all types of genomic regions (Lu et al. 2005). Many hom*ology-dependent effects, including meiotic silencing in fungi and paramutation in plants (see Chapters 6 and 9), remain incompletely understood. It will be the combined insight provided by future experiments in all eukaryotes, including ciliates, that will expose the full scope of epigenetic processes.

CILIATES

147

References Allis CD. and Gorovsky M.A. 1981. Histone phosphorylation in macro- and micronuclei of Tetrahymena thermophila. Biochemistry 20: 3828-3833. Allis CD., Glover CY, Bowen J.K., and Gorovsky M.A. 1980. Histone variants specific to the transcriptionally active, amitotically dividing macronucleus of the unicellular eucaryote, Tetrahymena thermophila. Cell 20: 609-617. Beisson J. and Sonneborn T.M. 1965. Cytoplasmic inheritance of the organization of the cell cortex in Paramecium aurelia. Proc. Natl. Acad. Sci. 53: 275-282. Beisson J. and Wright M. 2003. Basal body/centriole assembly and continuity. Curro Opin. Cell BioI. 15: 96-104. Betermier M. 2004. Large-scale genome remodelling by the developmentally programmed elimination of germ line sequences in the ciliate Paramecium. Res. Microbiol. 155: 399-408. Betermier M., Duharcourt S., Seitz H., and Meyer E. 2000. Timing of developmentally programmed excision and circularization of Paramecium internal eliminated sequences. Mol. Cell. Bioi. 20: 1553-1561. Brownell J.E. and Allis CD. 1995. An activity gel assay detects a single, catalytically active histone acetyl transferase subunit in Tetrahymena macronuclei. Proc. Natl. Acad. Sci. A 92: 6364-6368. Brownell J.E., Zhou J., Ranalli T., Kobayashi R., Edmondson D.G., Roth S.Y, and Allis CD. 1996. Tetrahymena histone acetyltransferase: A hom*olog to yeast Gcn5p linking histone acetylation to gene activation. Cell 84: 843-851. Brygoo Y and Keller A.-M. 1981a. A mutation with pleitropic effects on macronuclearly differentiated functions in Paramecium tetraurelia. Dey. Genet. 2: 23-34. - - - . 1981 b. Genetic analysis of mating type differentiation in Paramecium tetraurelia. HI. A mutation restricted to mating type E and affecting the determination of mating type. Dey. Genet. 2: 13-22. Cassidy-Hanley D., Bisharyan Y, Fridman Y, Gerber J., Lin C, Orias E., Orias J.D., Ryder H., Yong L., and Hamilton E.P. 2005. Genome-wide characterization of Tetrahymena thermophila chromosome breakage sites. H. Physical and genetic mapping. Genetics 170: 1623-1631. Chalker D.L. and Yao M.C 1996. Non-Mendelian, heritable blocks to DNA rearrangement are induced by loading the somatic nucleus of Tetrahymena thermophila with germ line-limited DNA. Mol. Cell. BioI. 16: 3658-3667. - - - . 2001. Nongenic, bidirectional transcription precedes and may promote developmental DNA deletion in Tetrahymena thermophila. Genes Dey. 15: 1287-1298. Chalker D.L., Fuller P., and Yao M.C 2005. Communication between parental and developing genomes during Tetrahymena nuclear differentiation is likely mediated by hom*ologous R As. Genetics 169: 149-160. Cheng J., Kapranov P., Drenkow J., Dike S., Brubaker S., Patel S., Long J., Stern D., Tammana H., Helt G., et al. 2005. Transcriptional maps of 10 human chromosomes at 5-nucleotide resolution. Science 308: 1149-1154. Claverie J.M. 2005. Fewer genes, more noncoding R A. Science 309: 1529-1530. Cleveland D.W., Mao, Y, and Sullivan K.F. 2003. Centromeres and kinetochores: From epigenetics to mitotic checkpoint signaling. CelllI2: 407-421. Coyne R.S., Chalker D.L., and Yao M.C 1996. Genome downsizing during ciliate development: Nuclear division of labor through chromosome restructuring. Annu. Rev. Genet. 30: 557-578.

148 •

C HAP T E R 7

Coyne R.S., Nikiforov M.A., Smothers J.E, Allis CD., and Yao M.C 1999. Parental expression of the chromodomain protein Pddlp is required for completion of programmed DNA elimination and nuclear differentiation. Mol. Cell 4: 865-872. Davis M.C, Ward J.G., Herrick G., and Allis CD. 1992. Programmed nuclear death: Apoptotic-like degradation of specific nuclei in conjugating Tetrahymena. Dev. BioI. 154: 419-432. Dou Y. and Gorovsky M.A. 2000. Phosphorylation of linker histone HI regulates gene expression in vivo by creating a charge patch. Mol. Cell 6: 225-231. - - - . 2002. Regulation of transcription by HI phosphorylation in Tetrahymena is position independent and requires clustered sites. Proc. Natl. Acad. Sci. 99: 6142-6146. Dou Y., Mizzen CA., Abrams M., Allis CD., and Gorovsky M.A. 1999. Phosphorylation of linker histone HI regulates gene expression in vivo by mimicking HI removal. Mol. Cell 4: 641-647. Duharcourt S., Butler A., and Meyer E. 1995. Epigenetic self-regulation of developmental excision of an internal eliminated sequence on Paramecium tetraurelia. Genes Dev. 9: 2065-2077. Duharcourt S., Keller A.M., and Meyer E. 1998. hom*ology-dependent maternal inhibition of developmental excision of internal eliminated sequences in Paramecium tetraurelia. Mol. Cell. BioI. 18: 7075-7085. Epstein L.M. and Forney J.D. 1984. Mendelian and non-Mendelian mutations affecting surface antigen expression in Paramecium tetraurelia. Mol. Cell. Bioi. 4: 1583-1590. Fan Q. and Yao M.-C 1996. New telomere formation coupled with sitespecific chromosome breakage in Tetrahymena thermophila. Mol. Cell. BioI. 16: 1267-1274. Forney J.D. and Blackburn E.H. 1988. Developmentally controlled telomere addition in wild-type and mutant paramecia. Mol. Cell. BioI. 8: 251-258. Forney J.D., Yantiri E, and Mikami K. 1996. Developmentally controlled rearrangement of surface protein genes in Paramecium tetraurelia. f. Eukaryot. Microbiol. 43: 462-467. Galvani A. and Sperling L. 2001. Transgene-mediated post-transcriptional gene silencing is inhibited by 3' non-coding sequences in Paramecium. Nucleic Acids Res. 29: 4387-4394. - - - . 2002. RNA interference by feeding in Paramecium. Trends Genet. 18: 11-12. Garnier 0., Serrano v., Duharcourt S., and Meyer E. 2004. RNA-mediated programming of developmental genome rearrangements in Paramecium tetraurelia. Mol. Cell. Bioi. 24: 7370-7379. Gratias A. and Betermier M. 2001. Developmentally programmed excision of internal DNA sequences in Paramecium aurelia. Biochimie 83: 1009-1022. - - - . 2003. Processing of double-strand breaks is involved in the precise excision of Paramecium internal eliminated sequences. Mol. Cell. BioI. 23: 7152-7162. Hamilton E., Bruns P., Lin C, Merriam v., Orias E., Yong L., and Cassidy-Hanley D. 2005. Genome-wide characterization of Tetrahymena thermophila chromosome breakage sites. 1. Cloning and identification of functional sites. Genetics 170: 1611-1621. Harumoto T. 1986. Induced change in a non-Mendelian determinant by transplantation of macronucleoplasm in Paramecium tetraurelia. Mol. Cell. BioI. 6: 3498-3501. Harwood ]. 1985. The erratic career of cytoplasmic inheritance. Trends Genet. I: 298-300. Hayashi T., Hayashi H., Fusauchi Y., and Iwai K. 1984. Tetrahymena histone H3. Purification and two variant sequences. ]. Biochem. 95: 1741-1749.

Hayashi T., Hayashi H., and Iwai K. 1987. Tetrahymena histone HI. Isolation and amino acid sequence lacking the central hydrophobic domain conserved in other HI histones.]. Biochem. 102: 369-376. Jahn CL. and Klobutcher L.A. 2002. Genome remodeling in ciliated protozoa. Annu. Rev. Microbiol. 56: 489-520. ]araczewski ].W. and ]ahn CL. 1993. Elimination of Tec elements involves a novel excision process. Genes Dev. 7: 95-105. Jessop-Murray H., Martin L.D., Gilley D., Preer J.R., and Polisky B. 1991. Permanent rescue of a non-Mendelian mutation of Paramecium by microinjection of specific DNA sequences. Genetics 129: 727-734. Klobutcher L.A. 1999. Characterization of in vivo developmental chromosome fragmentation intermediates in E. crassus. Mol. Cell 4: 695-704. Klobutcher L.A. and Herrick G. 1995. Consensus inverted terminal repeat sequence of Paramecium lESs: Resemblance to termini of Tel-related and Euplotes Tec transposons. Nucleic Acids Res. 23: 2006-2013. - - - . 1997. Developmental genome reorganization in ciliated protozoa: The transposon link. Prog. Nucleic Acid Res. Mol. BioI. 56: 1-62. Klobutcher L.A., Turner L.R., and LaPlante ]. 1993. Circular forms of developmentally excised DNA in Euplotes crassus have a heteroduplex junction. Genes Dev. 7: 84-94. Kobayashi S. and Koizumi S. 1990. Characterization of non-Mendelian and Mendelian mutant strains by micronuclear transplantation in Paramecium tetraurelia. f. Protozool. 37: 489-492. Koizumi S. and Kobayashi S. 1989. Microinjection of plasmid DNA encoding the A surface antigen of Paramecium tetraurelia restores the ability to regenerate a wild-type macronucleus. Mol. Cell. BioI. 9: 4398-4401. Landweber L.E, Kuo T.C, and Curtis E.A. 2000. Evolution and assembly of an extremely scrambled gene. Proc. Natl. Acad. Sci. 97: 3298-3303. Le Mouel A., Butler A., Caron E, and Meyer E. 2003. Developmentally regulated chromosome fragmentation linked to imprecise elimination of repeated sequences in Paramecia. Eukaryot. Cell 2: 1076-1090. Lee S.R. and Collins K. 2006. Two classes of endogenous small RNAs in Tetrahymena thermophila. Genes Dev. 20: 28-33. Liu Y., Mochizuki K., and Gorovsky M.A. 2004. Histone H3 lysine 9 methylation is required for DNA elimination in developing macronuclei in Tetrahymena. Proc. Natl. Acad. Sci. 101: 1679-1684. Liu Y., Song X., Gorovsky M.A., and Karrer K.M. 2005. Elimination of foreign DNA during somatic differentiation in Tetrahymena thermophila shows position effect and is dosage dependent. Eukaryot. Cell 4: 421-431. Lu C, Tej S.S., Luo S., Haudenschild CD., Meyers B.C, and Green P.J. 2005. Elucidation of the small RNA component of the transcriptome. Science 309: 1567-1569. Madireddi M.T., Davis M., and Allis 0. 1994. Identification of a novel polypeptide involved in the formation of DNA-containing vesicles during macronuclear development in Tetrahymena. Dev. BioI. 165: 418-431. Madireddi M.T., Coyne R.S., Smothers J.E, Mickey K.M., Yao M.C, and Allis CD. 1996. Pddlp, a novel chromodomain-containing protein, links heterochromatin assembly and DNA elimination in Tetrahymena. Cell 87: 75-84. Malone CD., Anderson A.M., Motl J.A., Rexer CH., and Chalker D.L. 2005. Germ line transcripts are processed by a Dicer-like protein that is essential for developmentally programmed genome rearrangements of Tetrahymena thermophila. Mol. Cell. BioI. 25: 9151-9164.

EPfGENETICS

Mattick ).S. 2004. RNA regulation: A new genetics? Nat. Rev. Genet. 5: 316-323. Matzke M.A. and Birchler ).A. 2005. RNAi-mediated pathways in the nucleus. Nat. Rev. Genet. 6: 24-35. Meyer E. 1992. Induction of specific macronuclear developmental mutations by microinjection of a cloned telomeric gene in Paramecium primaurelia. Genes Dev. 6: 211-222. Meyer E. and Keller A.M. 1996. A Mendelian mutation affecting matingtype determination also affects developmental genomic rearrangements in Paramecium tetraurelia. Genetics 143: 191-202. Meyer E., Butler A., Dubrana K., Duharcourt S., and Caron E 1997. Sequence-specific epigenetic effects of the maternal somatic genome on developmental rearrangements of the zygotic genome in Paramecium primaurelia. Mol. Cell. BioI. 17: 3589-3599. Meyers B.C, Vu T.H., Tej S.S., Ghazal H., Matvienko M., Agrawal v., Ning )., and Haudenschild CD. 2004. Analysis of the transcriptional complexity of Arabidopsis thaliana by massively parallel signature sequencing. Nat. Biotechnol. 22: 1006-1011. Mizzen CA., Dou Y., Liu Y., Cook R.G., Gorovsky M.A., and Allis CD. 1999. Identification and mutation of phosphorylation sites in a linker histone. Phosphorylation of macronuclear HI is not essential for viability in Tetrahymena. J. BioI. Chem. 274: 14533-14536. Mochizuki K. and Gorovsky M.A. 2004. Small RNAs in genome rearrangement in Tetrahymena. Curro Opin. Genet. Dev. 14: 181-187. - - - . 2005. A Dicer-like protein in Tetrahymena has distinct functions in genome rearrangement, chromosome segregation, and meiotic prophase. Genes Dev. 19: 77-89. Mochizuki K., Fine N.A., Fujisawa T., and Gorovsky M.A. 2002. Analysis of a piwi-related gene implicates small RNAs in genome rearrangement in Tetrahymena. Cell 110: 689-699. Moreira-Leite EE, Sherwin T., Kohl L., and Gull K. 2001. A trypanosome structure involved in transmitting cytoplasmic information during cell division. Science 294: 610-612. Nikiforov M.A., Gorovsky M.A., and Allis CD. 2000. A novel chromodomain protein, pdd3p, associates with internal eliminated sequences during macronuclear development in Tetrahymena thermophila. Mol. Cell. BioI. 20: 4128-4134. Nowacki M., Zagorski-Ostoja W., and Meyer E. 2005. Nowa1p and Nowa2p: Novel putative RNA binding proteins involved in transnuclear crosstalk in Paramecium tetraurelia. Curro Bio/. 15: 1616-1628. Paschka A.G., Jonsson E, Maier v., Mollenbeck M., Paeschke K., Postberg )., Rupprecht S., and Lipps H.). 2003. The use of RNAi to analyze gene function in spirotrichous ciliates. Eur. f. Protistol. 39: 449-454. Philippe H., Germot A., and Moreira 0. 2000. The new phylogeny of eukaryotes. Curro Opin. Genet. Dev. 10: 596-601. Pond ER., Gibson 1., Lalucat )., and Quackenbush R.L. 1989. R-bodyproducing bacteria. Microbial. Rev. 53: 25-67. Preer ).R., )r., Preer L.B., and )urand A. 1974. Kappa and other endosymbionts in Paramecium aurelia. Bacterial. Rev. 38: 113-163. Prescott D.M. 1994. The DNA of ciliated protozoa. Microbial. Rev. 58: 233-267. - - - . 1999. The evolutionary scrambling and developmental unscrambling of germline genes in hypotrichous ciliates. Nucleic Acids Res. 27: 1243-1250. Prescott D.M., Ehrenfeucht A., and Rozenberg G. 2003. Templateguided recombination for IES elimination and unscrambling of genes in stichotrichous ciliates. f. Theor. Bio/. 222: 323-330. Ruiz E, Vayssie L., Klotz C, Sperling L., and Madeddu L. 1998. hom*ol-

CILIATES.

149

ogy-dependent gene silencing in Paramecium. Mol. BioI. Cell 9: 931-943. Saveliev S.V. and Cox M.M. 1995. Transient DNA breaks associated with programmed genomic deletion events in conjugating cells of Tetrahymena thermophila. Genes Dev. 9: 248-255. ---.1996. Developmentally programmed DNA deletion in Tetrahymena thermophila by a transposition-like reaction pathway. EMBO f. 15: 2858-2869. Scott ).M., Mikami K., Leeck CL., and Forney ).D. 1994. NonMendelian inheritance of macronuclear mutations is gene specific in Paramecium tetraurelia. Mol. Cell. Bio/. 14: 2479-2484. Shen X. and Gorovsky M.A. 1996. Linker histone HI regulates specific gene expression but not global transcription in vivo. Cell 86: 475-483. Shen X., Yu L., Weir ).W., and Gorovsky M.A. 1995. Linker histones are not essential and affect chromatin condensation in vivo. Cell 82: 47-56. Shorter ). and Lindquist S. 2005. Prions as adaptive conduits of memory and inheritance. Nat. Rev. Genet. 6: 435-450. Sonneborn T.M. 1937. Sex, sex inheritance and sex determination in Paramecium aurelia. Proc. Nat/. Acad. Sci. 23: 378-385. - - - . 1975. Paramecium aurelia. In Handbook of genetics (ed. R. King), pp. 469-594. Plenum, New York. ---.1977. Genetics of cellular differentiation: Stable nuclear differentiation in eucaryotic unicells. Annu. Rev. Genet. 11: 349-367. Strahl B.D., Ohba R., Cook R.G., and Allis CD. 1999. Methylation of histone H3 at lysine 4 is highly conserved and correlates with transcriptionally active nuclei in Tetrahymena. Proc. Nat/. Acad. Sci. 96: 14967-14972. Suzuki M. and Hayashizaki Y. 2004. Mouse-centric comparative transcriptomics of protein coding and non-coding RNAs. Bioessays 26: 833-843. Taverna S.D., Coyne R.S., and Allis CD. 2002. Methylation of histone h3 at lysine 9 targets programmed DNA elimination in Tetrahymena. Cell 110: 701-711. Timmons L. and Fire A. 1998. Specific interference by ingested dsRNA. Nature 395: 854. Timmons L., Court D.L., and Fire A. 2001. Ingestion of bacterially expressed dsRNAs can produce specific and potent genetic interference in Caenorhabditis elegans. Gene 263: 103-112. Vavra K.)., Allis CD., and Gorovsky M.A. 1982. Regulation of histone acetylation in Tetrahymena macro- and micronuclei. f. BioI. Chem. 257: 2591-2598. Wei Y., Yu L., Bowen )., Gorovsky M.A., and Allis Co. 1999. Phosphorylation of histone H3 is required for proper chromosome condensation and segregation. Cell 97: 99-109. Wei Y., Mizzen CA., Cook R.G., Gorovsky M.A., and Allis CD. 1998. Phosphorylation of histone H3 at serine 10 is correlated with chromosome condensation during mitosis and meiosis in Tetrahymena. Proc. Natl. Acad. Sci. 95: 7480-7484. Williams K., Doak T.G., and Herrick G. 1993. Developmental precise excision of Oxytricha trifa/lax telomere-bearing elements and formation of circles closed by a copy of the flanking target duplication. EMBO J. 12: 4593-4601. Wu M., Allis CD., Richman R., Cook R.G., and Gorovsky M.A. 1986. An intervening sequence in an unusual histone HI gene of Tetrahymena thermophila. Proc. Natl. Acad. Sci. 83: 8674-8678. Yao M.C and Chao ).L. 2005. RNA-guided DNA deletion in Tetrahymena: An RNAi-based mechanism for programmed genome rearrangements. Annu. Rev. Genet. 39: 537-559. Yao M.C, Duharcourt S., and Chalker D.L. 2002. Genome-wide

HAP

RNAi and Heterochromatin Assembly Robert Martienssen 1 and Danesh Moazed 2 ICoid Spring Harbor Laboratory, Cold Spring Harbor, New York 11724 2Department of Cell Biology, Harvard Medical School, Boston, Massachusetts 02115-5730

CONTENTS 1. Overview of the RNAi Pathway, 153 2. Early Evidence Implicating RNA as an Intermediate in Transcriptional Gene Silencing, 154 3. RNAi and Heterochromatin Assembly in S. pombe, 155

6. RNA-RNA Versus RNA-DNA Recognition Models, 159 7. How Does RNAi Recruit Chromatin-modifying Enzymes?, 159 8. RNAi-mediated Chromatin and DNA Modifications in Arabidopsis, 160 9. Conservation of RNAi-mediated Chromatin Modifications in Animals, 162

4. Small RNAs Initiate Heterochromatin Assembly in Association with an RNAi Effector Complex, 156

10. Concluding Remarks, 164

5. dsRNA Synthesis and siRNA Generation, 157

References, 164

151

GENERAL SUMMARY The intersection of RNA interference (RNAi) and heterochromatin formation brought together two areas of gene regulation that had previously been thought to operate by different, perhaps even unrelated, mechanisms. Using cytological staining methods, heterochromatin was originally defined nearly 80 years ago as those chromosome regions that retained a condensed appearance throughout the cell cycle. Early investigators studying the relationship between chromosome structure and gene expression noticed that certain chromosome rearrangements resulted in the spreading of heterochromatin into adjacent genes, which then became silent. But the stochastic nature of spreading gave rise to genetically identical populations of cells that had different phenotypes, providing a striking example of epigenetic regulation. The term RNAi was first used to describe gene silencing when hom*ologous antisense or double-stranded RNA (dsRNA) is introduced into the nematode Caenorhabditis elegans. It was soon recognized that a related mechanism accounted for posttranscriptional transgene silencing (PTGS) described earlier in petunia and other plants. In contrast, heterochromatin was widely believed to operate directly at the chromatin level to cause transcriptional repression, by a mechanism referred to as transcriptional gene silencing (TGS). This chapter focuses on the relationship between the RNAi pathway and the formation of epigenetically heritable heterochromatin at specific chromosome regions. It draws on recent examples that demonstrate this relationship in the fission yeast Schizosaccharomyces pombe and the mustard plant Arabidopsis thaliana.

The fission yeast nuclear genome is composed of three chromosomes that range from 3.5 Mb to 5.7 Mb in size. Each chromosome contains large blocks of repetitive DNA, particularly at centromeres, which are packaged into heterochromatin. In addition, the mating-type loci (which control cell type) and subtelomeric DNA regions also contain repetitive sequences that are packaged into heterochromatin. We now know that the assembly of DNA into heterochromatin plays both regulatory and structural

roles. In the case of the mating-type loci in yeast, regulation of gene transcription by heterochromatin is important for cell-type identity. In the case of telomeres and centromeres, heterochromatin plays a structural role that is important for proper chromosome segregation during cell division. Moreover, repetitive DNA sequences and transposable elements account for a large fraction, in some cases more than half, of the genomes of many eukaryotic cells. Heterochromatin and associated mechanisms playa critical role in maintaining genome stability by regulating the activity of repeated sequences. Recent studies have uncovered a surprising requirement for components of the RNAi pathway in the process of heterochromatin formation in fission yeast and have provided insight into how these two pathways can work together at the chromatin level. Briefly, small interfering RNA (siRNA) molecules, which are a signature of RNAi and other dsRNA silencing mechanisms, assemble into the RNA-Induced Transcriptional Silencing (RITS) complex and direct epigenetic chromatin modifications and heterochromatin formation at complementary chromosome regions. RITS uses siRNAdependent base-pairing to guide association with either DNA or nascent RNA sequences at the target locus destined to be silenced, an association that is stabilized by direct binding to methylated histone H3. The presence of these two activities in RITS triggers heterochromatin formation in concert with well-known heterochromatin-associated factors and directly links RNA silencing to heterochromatin modification. In A. thaliana and other eukaryotes (with the exception of Saccharomyces cerevisiae), centromeric DNA regions are also composed of repetitive elements. These and other repeat sequences, such as retroelements and other transposons, are the source of siRNAs, attracting histone H3K9 and DNA methylation. Here again, several components of the RNAi pathway are required for the initiation and maintenance of these repressive methylation events. In this chapter, we discuss how heterochromatic siRNAs are produced and mediate DNA and/or chromatin modifications in fission yeast and A. thaliana.

RNA I AND

1 Overview of the RNAi Pathway

Although the term RNAi was originally used to describe silencing that is mediated by exogenous dsRNA in C. elegans (Fire et al. 1998), it now broadly refers to gene silencing that is triggered by some kind of dsRNA. The steps involved in RNAi include the generation of dsRNA (which can be endogenous or exogenous such as viral RNA), processing into siRNA, and targeting of these molecules to either mRNAs (PTGS) or chromatin regions (TGS) to effect silencing. Therefore, before introducing the components of the RNAi machinery specific to TGS, we discuss the source of dsRNA that harnesses the RNAi machinery into action. dsRNA may originate from bidirectional transcription of repetitive DNA elements, or transcription of RNA molecules that can base-pair internally to form dsRNA segments (see Fig. 1, a and b, respectively). For example, transcription through inverted repeat regions produces RNA molecules that fold back on themselves to produce hairpin structures. dsRNAs are then cleaved by Dicer, an RNase III class ribonuclease, which generates siRNAs. These are complementary duplexes, 21-27 nucleotides (nt) in size, that have a characteristic 2-nt overhang at each 3' end of the duplex (Hamilton and Baulcombe 1999; Zamore et al. 2000; Bernstein et al. 2001; Elbashir et al. 2001; Hannon 2002; Zamore 2002; Bartel 2004; Baulcombe 2004). These duplexes are unwound into single-stranded siRNA to act as guides, through base-pairing interactions with complementary target sequences. They are therefore specificity factors and playa central role in all RNAi-mediated silencing mechanisms. To date, two related complexes have been identified that incorporate siRNA: RISC and RITS. In the RNAInduced Silencing Complex (RISC), siRNAs recognize

(a) bidirectional transcription

(b) inverted repeat transcription

HE T ERa C H ROM A TIN

......

====J

1 siRNAs

1 ..

153

target mRNAs and initiate their degradation by endonucleolytic cleavage within the mRNA region that is basepaired to the siRNA (Hannon 2002; Bartel 2004). The RNase H domain of the Argonaute/PIWI family protein (a subunit of RISC) carries out this initial mRNA cleavage event. In the nuclear RNA-Induced Transcriptional Silencing (RITS) complex (similar to the RISC), siRNAs target the complex to chromosome regions for chromatin modification (Verdel et al. 2004; Buhler et al. 2006). It is the RITS-mediated RNA pathway that is the focus of this chapter. The central Argonaute and Dicer proteins are required for an additional type of RNA silencing mechanism involving microRNAs (miRNA). RNA, transcribed from endogenous noncoding genes that initially form hairpin RNA structures, due to extended dsRNA regions, is processed into miRNA through a series of steps (Bartel 2004; Filipowicz et al. 2005). Like siRNAs, miRNAs are 21-24 nt in size and form part of the RISC via the Argonaute proteins, to target specific mRNAs. This targeting can result in mRNA cleavage via the PIWI/RNAse H domain and translational repression involving interactions with the 7meG cap at the 5'-end of the mRNA. This may be coupled to sequestration of the mRNA to cytoplasmic RNA-processing organelles known as P bodies (Processing bodies). Thus, at least two different dsRNA-processing pathways result in the generation of siRNA or miRNA, yet these RNAs use a similar machinery to inactivate cognate mRNAs. The miRNA pathway distinguishes itself because miRNAs are all produced by endogenous noncoding genes that are largely developmentally regulated and, in turn, generally target and developmentally regulate the silencing of hom*ologous genes. Although dsRNAs can form by the annealing of forward and reverse RNAs that result from bidirectional

dsRNA

ASS EM B L Y

RdRP

Figure 1. Sources of dsRNA, Which Act as a Substrate for Generation of siRNAs by the Dicer Ribonuclease, and Are the Trigger for RNA Silencing (0) Bidirectional transcription has been

observed at the s. pombe centromeric repeats and the cenH region of the silent mating-type locus. (b) Transcription through inverted repeats found in many plant and animal cells can potentially produce dsRNA. (c) Transcription of aberrant RNAs that may lack proper processing signals may trigger dsRNA synthesis by RdRPs.

1 54

C HAP T E R 8

transcription or are present in hairpin structures, in some cells, RNAi requires an additional enzyme to make dsRNA. This is the RNA-directed RNA polymerase (RdRP) found in plants and C. elegans (Dalmay et al. 2000; Sijen et al. 2001). It uses siRNAs as primers to generate more dsRNA, which can then be processed into additional siRNA by Dicer. The primary function of RdRP is thus thought to be in amplification of the RNAi response, but, as discussed later, RdRPs may have more specific roles in initiating dsRNA synthesis (see Section 5). Indeed, it seems to be involved in a process adapted for producing a better host defense response to the introduction of exogenous dsRNA. This idea is strengthened by the fact that RdRPs are not involved in the miRNA silencing pathways (Sijen et al. 2001). Interestingly, insects (including Drosophila) and vertebrates (including mammals) lack recognizable RdRP-like sequences in their genomes, but it remains possible that other polymerases carry out dsRNA synthesis in these organisms. What then is the function of the various RNA silencing mechanisms? They are widely conserved in organisms ranging from fission yeast to plants to human, and they play central roles in the regulation of gene expression and genome stability (through stable heterochromatin formation at centromeres and telomeres). In addition, these silencing mechanisms are involved in defense against transposons and RNA viruses through degradation of their RNA transcripts (Plasterk 2002; Li and Ding 2005). Finally, transcription from some transposons generates aberrant RNAs that trigger RNAi by a mechanism thought to involve the conversion of aberrant transcripts to dsRNA by RdRPs (Fig. 1) (Baulcombe 2004). 2 Early Evidence Implicating RNA as an Intermediate in Transcriptional Gene Silencing

Before discussing the better-understood examples of RNAi-based chromatin modifications in fission yeast and Arabidopsis, we briefly discuss early experiments that suggested a role for RNA in mediating chromatin and DNA modifications. The earliest evidence for the role of an RNA intermediate in TGS came from studies of plant viroids. The potato spindle tuber viroid (PSTV) consists of a 359-nt RNA genome and replicates via an RNA-RNA pathway. The introduction of PSTV into the tobacco genome results in the DNA methylation of hom*ologous nuclear sequences, albeit transgenic in origin (Wassenegger et al. 1994). However, these and integrated copies of PSTV DNA only become methylated in plants that sup-

port viroid RNA transcription, suggesting the involvement of an RNA intermediate that directs DNA methylation (Wassenegger et al. 1994). Furthermore, in Arabidopsis, the production of aberrant transcripts somehow results in the DNA methylation of all hom*ologous promoter regions and transcriptional gene silencing (Mette et al. 1999). This, together with the finding that the replication of viral genomes in plants leads to the production of small RNAs that are 22 nt in size, suggests that RNAi-reiated mechanisms mediate DNA methylation (Mette et al. 2000). These observations, as well as repeatinduced silencing by transgenes, which was first discovered in petunia and in tobacco, are now widely recognized as the earliest examples of silencing by RNAi (Napoli et al. 1990; discussed in Chapter 9). Further evidence for a link between RNAi and TGS comes from studies of repeat-induced gene silencing in Drosophila (see Chapter 5). The introduction of multiple tandem copies of a transgene results in the silencing of both the transgene and the endogenous copies (PalBhadra et al. 1999). This silencing requires the chromodomain protein Polycomb, which is also involved in the packaging of homeotic regulatory genes into heterochromatin-like structure outside of their proper domains of action (Francis and Kingston 2001). In addition, this repeat-induced gene silencing requires Piwi, a Drosophila Argonaute family member required for RNAi (PalBhadra et al. 2002). In Tetrahymena, another Piwi protein family member, Twil, is required for small RNA accumulation and the massive DNA elimination that is observed in the somatic macronucleus of the protozoa (see Chapter 7). These and more recent results discussed in Section 8 suggest that the RNAi pathway is involved in the assembly of repressive chromatin structures in flies. Other repeat-induced silencing mechanisms have been described in filamentous fungi, including RepeatInduced Point mutation (RIP) in Neurospora crassa and Methylation Induced Pre-meiotically (MIP) in Ascobolus immersus, that do not appear to involve an RNA intermediate since they occur independently of the transcriptional state of the locus (Galagan and Selker 2004). Instead, RIP and MIP involve paired loci, where (for example) two out of three gene copies are silenced, suggesting some kind of DNA-DNA interaction mechanism involving hom*ologous loci to induce silencing. Conversely, silencing of unpaired DNA in meiosis (MSUD), which also occurs in Neurospora, requires the RNAi pathway (Shiu et al. 2001; discussed in Chapter 6) and may have parallels in other organisms, including C. elegans (Maine et al. 2005; see Chapter 15).

RNA I AND

3 RNAi and Heterochromatin Assembly in 5. pombe S. pombe chromosomes contain extensive heterochromatic DNA regions that are associated with underlying repetitive DNA elements at the centromeres and the silent mating-type loci (mat2/3) (Grewal 2000; Pidoux and Allshire 2004). Each fission yeast centromere contains a unique central core region (ent) that is flanked by two types of repeats, called the innermost (imr) and outermost (otr) repeats (Fig. 2). The otr region itself is composed of dh and dg repeats. Heterochromatin formation in S. pombe involves the concerted action of a number of trans-acting factors. These include histone deacetylases (HDACs), Clr4, a histone H3 lysine 9 methyltransferase (HKMT), and the histone H3K9-methyl binding proteins, Swi6 (an HPI hom*olog) and Chpl. The initial recruitment of Swi6 and Clr4 to chromatin has been proposed to result in the spreading of H3K9 methylation and heterochromatin formation through sequential cycles of Clr4-catalyzed

H E T ERa C H ROM A TIN

ASS E M B L Y

•

155

H3K9 methylation coupled to Swi6-mediated spreading to adjacent nucleosomes through its self-association (Grewal and Moazed 2003). Mutation in components of the RNAi pathway surprisingly results in a loss of centromeric heterochromatin and the accumulation of noncoding forward and reverse transcripts from bidirectional promoters within each dg and dh repeat (Fig. 2) (Volpe et al. 2002). Fission yeast contains a single gene for each of the RNAi proteins, Dicer, Argonaute, and RdRP (derl+, agol+, and rdpl+, respectively). Deleting any of these genes results in the loss of histone H3K9 methylation, and mutants display defects in chromosome segregation, which are generally associated with defects in heterochromatin assembly (Provost et al. 2002; Volpe et al. 2003). Moreover, sequencing of a library of fission yeast small RNAs identified ~22-nt RNAs that mapped exclusively to centromeric repeat regions and ribosomal DNA repeats, suggesting that een RNAs can produce dsRNAs that are processed into siRNAs (Reinhart and Bartel 2002). Thus, it was suggested that the RNAi pathway

S.pombe centromere

otrL

imrL

cnt1

imrR

otrR

• dsRNA====

S.pombe silent mating type

cenH Figure 2. Organization of Heterochromatic Chromosome Regions in 5. pombe and A. thaliana

dsRNA

A.thaliana centromere 180-bp repeats • • • • • • • • • • LTR

180-bp repeats • • • • • • • • • • LTR 4

••••••••••••• !

dsRNA=====

The centromere of S. pombe chromosome 1 is shown as an example. The unique central core (ent1) region is flanked by innermost (imrL and imrR) and outermost (otrL and otrR) repeats. The otr region is transcribed in both directions, giving rise to forward (blue) and reverse (red) transcripts. The region between the mat2 and mat3 genes contains a domain that is hom*ologous to the centromeric dg and dh repeats (eenH) and is also bidirectionally transcribed. Atfl and Pcrl are DNA-binding proteins that act in parallel with RNAi in mating-type silencing. Arabidopsis centromeres are composed of 180-bp repeats (green) interspersed with retrotransposable elements (yellow). Forward transcripts initiating within the long terminal repeat (LTR) of the retroelement and reverse transcripts initiating within the 180-bp repeats are indicated.

156 • C HAP

T ER

could recruit Swi6 and Clr4 to chromatin to initiate and/or maintain heterochromatin formation at each of the above loci (Fig. 3) (Hall et al. 2002; Volpe et al. 2002). Interestingly, both TGS and PTGS mechanisms appear to contribute to the down-regulation of een RNAs. The forward strand transcript is primarily silenced at the transcriptional level, as demonstrated in RNAi mutants (Volpe et al. 2002). The reverse strand of een transcripts, however, is not affected by Swi6 mutants (Volpe et al. 2002), and silencing of this cen-reverse transcript occurs primarily at the posttranscriptionallevel. RNAi also plays a role in silencing the mating-type locus (mat2/3) (Hall et al. 2002). mat2/3 is interrupted by a region of DNA that is highly hom*ologous to centromeric repeats (called eenH, eenhom*ology) (Fig. 2). Like the een repeats, the eenH region is divergently transcribed to produce forward and reverse RNA (Noma et al. 2004). These eenH transcripts accumulate to high levels in RNAi mutants. In contrast, RNAi is not necessary for silencing of a reporter transgene inserted at mat2/3 if silencing is not first somehow compromised. The reason for this difference is that a partially redundant silencing mechanism, involving two DNA-binding proteins, Pcrl and Atfl, which bind to mat2/3, recruits the heterochromatin

active chromatin

RNAi machinery

HDAC

CD 4

silent chromatin

Figure 3. Assembly of Heterochromatin Involves the Concerted Action of Histone-modifying Enzymes (HDACs and Clr4) and Histone-binding Proteins (e.g., Swi6) and Can Be Directed by the RNAi Machinery Deacetylation by HDACs is followed by recruitment of Clr4 and histone H3K9 methylation. Swi6 binds to H3K9-methylated histone tails, and spreading results from sequential cycles of H3K9 methylation that are coupled to Swi6 oligomerization.

machinery independently of RNAi (Jia et al. 2004). This is sufficient for silencing the reporter gene in the absence of RNAi but not for preventing the accumulation of noncoding eenH transcripts (Fig. 2). 4 Small RNAs Initiate Heterochromatin Assembly in Association with an RNAi Effector Complex

The discovery that the RNAi pathway is involved in heterochromatin formation in fission yeast and in transcriptional gene silencing in other systems raised the question of how it could directly regulate chromatin structure. Purification of Chpl, a chromodomain protein that is a structural component of heterochromatin, led to the identification of the RITS complex (Verdel et al. 2004). RITS contains the fission yeast Agol protein and Tas3, a protein of unknown function, in addition to Chp 1. It also contains centromeric siRNAs, which are produced by the Dicer ribonuclease, and importantly, RITS associates with centromeric repeat regions in an siRNA-dependent fashion. RITS has therefore been proposed to use centromeric siRNAs to target specific chromosome regions for inactivation, and this provides a direct link between RNAi and heterochromatin assembly (Fig. 4). Like RISC, which mediates PTGS, RITS uses siRNAs for target recognition. Unlike RISC, however, RITS associates with chromatin and initiates heterochromatin formation as opposed to mRNA inactivation. How can siRNAs target specific chromosome regions? Two possible mechanisms have been proposed. In the first model, .siRNAs bound to Agol in the RITS complex must somehow basepair with an unwound DNA double helix. In the second model, RITS-associated siRNAs base-pair with noncoding RNA transcripts at the target locus (Fig. 4). According to either model, the association of RITS with chromatin via siRNA results in the recruitment of the Clr4 HKMT and subsequent histone H3K9 methylation. This is followed by Swi6 binding and the spreading of H3K9 methylation and heterochromatin. However, Clr4 is also required for the association of RITS with chromatin, suggesting that it provides methylated H3K9 to which the RITS complex can bind, thereby stabilizing its association with chromatin. The chromodomain of Chpl was already known to bind specifically to methylated H3K9 residues (Partridge et al. 2002), and mutations in Clr4 or the chromodomain of Chp 1 that are involved in this interaction result in a loss of RITS binding to chromatin (Partridge et al. 2002; Noma et al. 2004). Moreover, RITS can also bind to chromatin domains that are coated with methylated H3K9 through the chromo-

RNA I

dsRNA

siRNAs

1 -

nascent transcript

Figure 4. RNAi and siRNA-directed Assembly of Heterochromatin in 5. pombe Both Dicer and RDRC are required for siRNA generation. Initial targeting is proposed to involve RITS and siRNA-mediated recognition of cognate transcripts. The binding of RITS is stabilized by association of the chromodomain of Chpl with H3K9 methylated histone. The recruitment of Clr4 and Swi6 mediates the spreading of H3K9 methylation.

domain of Chpl at the rnat2/3 and telomeric regions in the absence of siRNAs (Noma et al. 2004; Petrie et al. 2005). In summary, the RITS complex shows affinity to chromatin via Chpl binding to methylated H3K9 and

AND

HE T E ROC H ROM A TIN

ASS E M 8 L Y

157

through base-pairing of siRNA with either DNA or RNA transcripts. Recent evidence strongly supports a role for RITS and siRNAs in the initiation of heterochromatin assembly. BuWer et al. (2006) used a site-specific RNA-binding protein to artificially tether the RITS complex to the RNA transcript of the normally active ura4+ gene. Remarkably, this tethering results in the generation of ura4+ siRNAs and silencing of the ura4+ gene in a manner that requires both RNAi and heterochromatin components. In addition, this system allowed a direct evaluation of the ability of newly generated siRNAs to initiate H3K9 methylation and Swi6 binding, which are molecular markers for heterochromatin formation. Interestingly, the newly generated ura4+ siRNAs were found to be under negative control by the conserved siRNA ribonuclease, Eril, which restricts them to the locus where they are produced. However, when the gene encoding Eril is deleted, ura4+ siRNAs are able to act in trans to silence a second copy of the ura4+ gene, which is inserted on a different chromosome in the same cell. This experiment therefore demonstrates that siRNAs can act as specificity factors that direct RITS and heterochromatin assembly to a previously active region of the genome. The ability of siRNAs to initiate silencing in S. p*rnbe has also been examined using a different method, which relies on the expression of a hairpin RNA to produce siRNAs hom*ologous to a GFP transgene (Sigova et al. 2004). In this system, hairpin siRNAs promoted silencing of the GFP reporter gene at the PTGS, but not TGS, level (Sigova et al. 2004). It is unclear why the hairpin siRNAs cannot induce TGS and heterochromatin assembly at the chromosomal copy of GFP. One possible explanation is that heterochromatin is assembled at specific subnuclear locations, and assembly outside these locations occurs inefficiently (Gasser et al. 2004; Chapter 4). 5 dsRNA Synthesis and siRNA Generation

Bidirectional transcription of centromeric DNA repeats could in principle provide the initial source of dsRNA in fission yeast (Volpe et al. 2002). dsRNA resulting from the annealing of forward and reverse transcripts could then be a substrate for the Dicer ribonuclease. However, RNA-directed RNA polymerase (Rdpl) and its associated cofactors, as well as the Clr4 HKMT, are also required for siRNA production by Dicer (Hong et al. 2005; Li et al. 2005; Buhler et al. 2006). These observations indicate that the generation of heterochromatic siRNAs by Dicer is coupled to chromatin and Rdpldependent events (Fig. 4).

158 •

C HAP T E R 8

The Rdpl enzyme resides in a m\.lltiprotein complex that also contains Hrrl, an RNA helicase, and Cid12, a member of the ~ family of DNA polymerases which includes poly(A) polymerase enzymes (Motamedi et al. 2004). This complex has been termed RNA-directed RNA polymerase complex (RDRC), and all of its subunits are required for heterochromatin formation at centromeric DNA regions (Motamedi et al. 2004). As expected from the presence of Rdpl, RDRC has RNA-directed RNA polymerase activity in vitro, and mutations that abolish this activity also abolish RNAi-dependent silencing in vivo (Motamedi et al. 2004; Sugiyama et al. 2005). The in vitro RNA synthesis activity ofRDRC does not require an siRNA primer (Motamedi et al. 2004). RITS may therefore provide in vivo specificity by recruiting RDRC to selected RNA templates via siRNA. Consistent with this hypothesis, subunits of the RDRC are required for siRNA generation, and RITS complexes purified from cells that lack any subunit of the RDRC are devoid of siRNAs (Motamedi et al. 2004; Li et al. 2005; Sugiyama et al. 2005; Buhler et al. 2006). The presence of Cid12 in the RDRC is intriguing and raises the possibility that another polymerase activity participates in chromosome-associated RNA silencing.

Because some members of this family have poly(A) polymerase activity, one possibility is that adenylation of Rdpl-produced dsRNA may be important for their further processing. Interestingly, Cid12-like proteins are conserved throughout eukaryotes (Table 1); mutations in Rde-3, a C. elegans member of this family, result in defective RNAi (Chen et al. 2005), corroborating a conserved role for these enzymes in the RNAi pathway. There is evidence for dsRNA synthesis and processing associated with the generation of heterochromatic siRNAs occurring on the chromosome, at sites of transcription of noncoding centromeric RNAs (Fig. 4). Evidence includes, first, that Rdpl can be cross-linked to centromeric DNA repeats (Volpe et al. 2002; Sugiyama et al. 2005), and to the forward and reverse RNA transcripts that originate from these regions (Motamedi et al. 2004). As is the case with cross-linking to DNA, cross-linking to centromeric RNAs requires Dicer and Clr4, and is therefore siRNA- and chromatin-dependent. Second, siRNA generation requires chromatin components, including Clr4, Swi6, and the HDAC Sir2 (Hong et al. 2005; Li et al. 2005; Buhler et al. 2006). Finally, the association of RDRC with RITS is dependent on siRNAs as well as Clr4, suggesting that it occurs on chromatin (Motamedi et al.

Table 1. Conservation of RNAi and heterochromatin proteins

5. pombe

A. thaliana

C. elegans

Drosophila

Dcr1

DCLl to 4

Dcr-1

Dcr1 and 2

DCR-1

Ago1

AG01 to 10

Rde-1, Alg-1 and -2

Ago1 to 3, Piwi

AGO-1 to AGO-4

PRG-1 and 2, and 19 others

Aubergine/Sting

Piwi-1 to Piwi-4

PIWI-1 to PIWI-4

Chp1'

CMT3

H. sapiens

Tas3 b Rdp1

RDR1 to 6

Ego-1, Rrf-1 to -3

Hrr1

SGS2/SDE3

ZK1067.2

GH20028p

KlAA1404

Rde-3, Trf-4'

CGl1265'

POLS'

Cid12 Swi6

LHP1 (TFL2)

Clr4

SUVH2 to 6

Hpl-1, Hpl-2, F32E10.6

HP1

HP1a, ~,y

Su(var)3-9

SUV39H1 and 2

Rik1 e

DDB1

M18.5

Ddb1

DDB1

Cul4

CUL4

Cul4

CUL4

Sir2

SIR2

Sir2-1

Sir2

SIRT1

Eri1

ERl1

Eri-1

CG6393

THEX1

, An obvious ortholog of the chromodomain protein, Chp1, has not been identified in the other model organisms listed here, but most eukaryotic cells contain multiple chromodomain proteins. CMT3 in Arabidopsis is a chromodomain DNA methyl transferase, which acts in the same pathway as AG04 and may be analogous to Chp1. b

No obvious orthologs of Tas3 have been identified, but it shares weak sequence similarity with a mouse ovary testis specific protein (NP_035152).

'Cid12 belongs to a large family of conserved proteins that share sequence similarity with the classic poly(A) polymerase as well as 2'-5'-oligoadenylate enzymes. d

C. elegans have about 20 SET domain proteins, but an H3K9 HKMT has not yet been identified in this organism .

• S. pombe contains another Rik1-like protein, Ddb1, which is involved in DNA damage repair. Metazoans and plants appear to contain only a single Rik1-like gene, called Ddb1, which has been shown to be involved in DNA damage repair, but it is unknown whether it also participates in heterochromatin formation.

RNA fAN 0

2004). Thus, the generation of dsRNA and heterochromatic siRNAs may involve the recruitment of RDRC to chromatin-associated nascent pre-mRNA transcripts as illustrated in Figure 5 (Martienssen et al. 2005; Verdel and Moazed 2005). The fact that transcription and siRNA generation are likely to occur simultaneously reinforces the difference between RNA silencing mechanisms that mediate chromatin modifications and PTGS. However, this distinction is unlikely to be absolute. For example, in C. elegans, mutations in several chromatin components, similar to S. p*rn be, result in defects in RNAi and transposon-induced RNA silencing (see Table 1) (Sijen and Plasterk 2003; Grishok et al. 2005; Kim et al. 2005), raising the possibility that in some cases dsRNA synthesis and processing may occur on the chromosome regardless of whether silencing occurs at the TGS or PTGS level. 6 RNA-RNA Versus RNA-DNA Recognition Models

An outstanding question in working out the role of RNAi in heterochromatin assembly is whether RlTS/RDRC associates with DNA or nascent RNA. The observation that tethering components of the RNAi machinery to a gene transcript can induce heterochromatin-dependent gene silencing in cis clearly demonstrates that this process can be promoted via initial interactions with nascent RNA transcripts (Buhler et al. 2006). Importantly, cis-restriction rules out the possibility that the initial events of dsRNA synthesis and siRNA generation occur on mature transcripts where mRNA

- Figure 5. Model for Co-transcriptional dsRNA and siRNA Generation, and Recruitment of the Clr4-Rikl-CuI4 Histone Methyltransferase Complex in S. pombe

H E T E ROC H ROM A T f N

ASS E M B L Y

•

159

products from different alleles cannot be distinguished. Furthermore, a direct prediction of the RNA-RNA interaction model is that transcription at the target locus should be required for RNAi-mediated heterochromatin assembly. Although the requirement for transcription has not been directly tested, mutations in two different subunits of RNA polymerase II (RNA pol II), denoted Rpb2 and Rpb7, have specific defects in siRNA generation and heterochromatin assembly, but not on general transcription (Djupedal et al. 2005; Kato et al. 2005). This is reminiscent of Rbpl mutants, which have defects in histone modifications (i.e., H3K4 methylation and H2B ubiquitination) coupled to transcriptional elongation (Hampsey and Reinberg 2003), and provides a precedent for the hypothesis that RNAimediated H3K9 methylation and heterochromatin formation could be coupled to transcriptional elongation via the association of RNAi complexes with RNA pol II. In fact, contrary to the widely held view that heterochromatin is an inaccessible structure that inhibits transcription, RNAi-mediated heterochromatin assembly has little or no effect on the association of RNA pol II with S. p*rn be centromeric repeats (Volpe et al. 2002; Djupedal et al. 2005; Kato et al. 2005; Buhler et al. 2006). Therefore, nascent RNA transcripts, which act as templates for RITS in the RNA-RNA recognition model, are present in heterochromatic domains (Fig. 4). The RNA-RNA targeting model is also supported by the observation that components of both the RITS and RDRC complexes can be localized to noncoding centromeric RNAs using in vivo cross-linking experiments (Motamedi et al. 2004). This localization is siRNA-dependent, which suggests that it involves base-pairing interactions with the noncoding RNA. In addition, it requires the Clr4 HKMT, suggesting that it is coupled to binding of RITS to methylated H3K9 and occurs on chromatin. Nonetheless, the possibility that siRNAs can also recognize DNA directly through base-pairing interactions cannot be ruled out. For example, in plants, siRNAs that are complementary to promoter regions that are (presumably) not transcribed can still direct DNA methylation, another modification which takes place during heterochromatin formation within these regions (see Chapter 9). 7 How Does RNAi Recruit Chromatin-modifying Enzymes?

The recruitment of Clr4 and Swi6 is a key step in initiating histone H3K9 methylation and heterochromatin assembly, through an autoregulatory modification-

1 60 • C HAP T E R 8

binding model (Figs. 3 and 4) (Grewal and Moazed 2003). However, because RITS association to chromatin and Clr4-catalyzed histone H3K9 methylation are interdependent processes, it has been difficult to determine the event that provides the initial trigger for RNAidependent heterochromatin assembly. One solution to this chicken-and-egg problem is that siRNA-dependent base-pairing interactions could provide the initial signal for heterochromatin assembly (Fig. 4). Consistent with this hypothesis, de novo generation of ura4+ siRNAs promotes silencing of a previously active copy of the ura4+ gene that is coupled to the recruitment of RITS and Swi6 to chromatin (Buhler et al. 2006). The initial binding of RITS may, however, be transient and difficult to detect; stable binding of RITS to chromatin would require dual interactions between (1) RITS-bound siRNAs and the nascent transcript and (2) the chromodomain binding of Chpl to methylated H3K9. In this model, RITS itself directly recruits Clr4. Alternatively, Clr4 may be recruited by a parallel pathway that involves one or more DNA-binding proteins, as is the case at the silent mating type and telomeric regions (Jia et al. 2004; Kanoh et al. 2005). In either scenario, Clr4-mediated H3K9 methylation would be required to stabilize RITS association with chromatin, which then leads to the recruitment ofRDRC, dsRNA synthesis, and siRNA generation (Fig. 5). Clr4 has recently been found to be a component of a multiprotein complex that contains the heterochromatin protein Rikl, a Cullin E3 ubiquitin ligase, Cul4, and several other proteins (Hong et al. 2005; Horn et al. 2005; Jia et al. 2005; Li et al. 2005). These Clr4-associated proteins further strengthen the link between RNA and heterochromatin formation. The Rikl protein is a member of a large family of ~ propeller WD repeat proteins that have been implicated in RNA or DNA binding. Members of this protein family include the Cleavage Polyadenylation Specificity Factor A (CPSF-A) involved in pre-mRNA splicing, and the DNA damage binding 1 (Ddbl) protein involved in binding UV-damaged DNA. CPSF-A is of particular interest because Rikl shares sequence similarity with its putative RNA-binding domain involved in the recognition of mRNA polyadenylation sequences (Barabino et al. 2000). The Ddbl protein, like Rikl, is a component of a Cul4 E3 ubiquitin ligase complex and is involved in the recognition and repair of UV-damaged DNA (Higa et al. 2003; Zhong et al. 2003). An exciting possibility is that Rikl acts in a fashion that is similar to CPSF-A and Ddbl, binding to an RNAi-generated product during heterochromatin assembly (Fig. 5).

8 RNAi-mediated Chromatin and DNA Modifications in Arabidopsis

The mechanism by which RNAi guides heterochromatic modifications in plants is similar to the mechanism in fission yeast, but there are also many differences. The most important difference is that plants have methylated DNA at many repressive heterochromatin regions: In this respect they resemble vertebrates, but differ from worms and Drosophila (Lippman and Martienssen 2004). Four genetic screens for mutants that relieve RNA-mediated TGS have recovered mutants in H3K9-specific HKMTs, and in RNAi components, but they have also uncovered the required function of DNA methyltransferases, SWIISNF remodeling complexes, and a novel RNA polymerase (Baulcombe 2004). These screens are described in detail in Chapter 9, but here we briefly compare the mechanism in fission yeast and plants. Each of the silencing mutant screens used inverted repeats introduced in trans to induce the silencing of endogenous or transgenic reporter genes. Relief from silencing indicates a mutation that has arisen in a necessary component of the silencing pathway. The endogenous genes used were PAl2 (involved in amino acid biosynthesis) (Mathieu and Bender 2004) and SUPERMAN (a transcription factor that regulates flower development) (Chan et al. 2004), and the reporter genes used were driven by either a strong viral promoter or a strong seed-specific promoter (Matzke et al. 2004). In each case, the promoter was targeted for silencing, in some cases along with the rest of the gene. A number of the genes found through these screens are illustrated in Figure 6 (see also Table 1 of Chapter 9). Only one RNAi mutant was identified, in only one of the screens, and this was the argonaute gene AG04. However, three of the screens recovered mutants in DNA methyltransferases, including METl and CMT3. A third DNA methyltransferase related to the mammalian DNMT3 was identified by reverse genetics, as this activity is encoded by DRMl and DRM2, two redundant genes unable to be determined in single mutant screens (see Chapter 9). Indeed, redundancy may account for the failure to recover additional components of the RNAi apparatus: for example, although DCL3 (DICER-LIKE 3) and RNA-DEPENDENT RNA POLYMERASE 2 (RDRP) are predominantly required for production of the 24-nt siRNA associated with transposons and repeats, at least two other DCL genes in Arabidopsis can substitute for DCL3 to some extent (Gasciolli et al. 2005).

RNA I

Figure 6. Summary of RNAi and Chromatin Proteins Required for RNAi-mediated DNA and Histone Methylation in Arabidopsis Synthesis of dsRNA from repeated DNA elements provides a substrate for Dicer-mediated cleavage and siRNA generation (DCL3 and other Dicers). RNA-directed RNA polymerases (RdRP, RdR2) and RNA polymerase IV (RNA Pol IV) may be directly involved in the synthesis of dsRNA or its amplification. siRNAs then load onto Argonaute proteins (e.g., AG04), which is likely to help target cognate repeat sequences for DNA and H3K9 methylation in association with other factors.

Other mutants found in the PAI2 screen included mutants in the H3K9 methyltransferase gene KYP/SUVH4 and the chromodomain-containing DNA methyltransferase gene CMT3. The parallels with fission yeast in this case are striking, as the RITS complex contains both an Argonaute protein and the chromodomain protein Chpl, which depends on H3K9 methylation for its association with the chromosome. Unlike fission yeast, however, loss of CMT3 or of H3K9me2 does not result in loss of siRNA in Arabidopsis (Lippman et al. 2003), and it is not yet clear whether these proteins form a complex with AG04. There are, however, several other H3K9-specific HKMTs in Arabidopsis, at least three of which have genetic function, so redundancy may be part of the explanation here as well (Ebbs et al. 2005). Mutants in the other DNA methyltransferases, METl and DRM1I2, in contrast to mutants in CMT3, do result in loss of siRNA accumulation, at least from a subset of transposons and from tandem repeats, which generally

AND

HE T E ROC H ROM A TIN

ASS EM B L Y

•

161

produce siRNA if they are transcribed (Martienssen 2003). In these cases, loss of siRNA is correlated with the loss of H3K9me2 (Cao et al. 2003; Lippman et al. 2003). Mutants in the SWI2/SNF2 chromatin-remodeling ATPase DDMI (decreased DNA methylation) also abolish siRNA and H3K9me2 accumulation from a wide range of transposons, although when siRNA is retained, so is H3K9me2 (Lippman et al. 2003). It is possible, therefore, that siRNA in plants is bound to the chromosome via methylated DNA instead of, or in addition to, binding via methylated histones as is the case in S. pombe (Fig. 7). DDMI has an exquisite specificity for transposons and repeats, and must somehow recognize these as being different from genes. siRNA, perhaps bound to the chromosome by methyl-binding proteins, would have the required specificity to make this distinction. Transposons and repeats in Arabidopsis are a major source of 24-nt, and some 21-nt, siRNA, consistent with this idea (Lippman et al. 2004). Centromeric satellite repeats, which are arranged in tens of thousands of tandem copies on either side of each centromere, are also transcribed and processed by RNAi (Fig. 6). This processing depends on DCL3, RDR2, and DDMl. Silencing also depends on H3K9me2 and CMT3. However, silencing is more complex than in fission yeast, as retrotransposon insertions into the repeats can silence them, and this depends on other mechanisms including METl, DDMl, and the histone deacetylase HDA6 (May et al. 2005). As mentioned earlier, in fission yeast, subunits of RNA pol II are required for silencing and siRNA production, supporting the idea that the RNAi- and chromatin-modification apparatus is recruited to the chromosome by nascent transcripts (Fig. 5). In Arabidopsis, two subunits of a novel RNA polymerase (RNA pol IV) were recovered in one of the four screens mentioned above (Kanno et al. 2005) but were first isolated as weak mutants in PTGS, along with mutants in RNA-dependent RNA polymerase (Herr et al. 2005). It is not yet known what template is used by RNA pol IV, but both methylated DNA (Onodera et al. 2005) and double-stranded RNA have been proposed (Vaughn and Martienssen 2005). Only the largest subunits are unique to RNA pol IV, which presumably uses the same complement of small subunits as RNA pol II. Additional SWI2/SNF2 chromatin remodelers that were also recovered in these screens may alter local chromatin structure to facilitate processivity of RNA polymerases. It is therefore likely that they facilitate transcription by RNA pol IV (Kanno et al. 2004). A similar role can be proposed for DDMl, although the require-

162 •

C HAP T E R 8

d,RNA

flowering. No such phenotype is observed in mutants of RNAi, even when siRNA is lost. Instead, RNAi may playa role in initiating FWA silencing, because FWA transgenes are rapidly silenced when first introduced into a plant, and this silencing depends on DeL3, RDR2, and AG04 (Chan et al. 2004). Silencing might then be maintained by DNA methylation, regulated by MET1. Similarly, transposons that lose siRNA in metl mutants cannot be resilenced in backcrosses, but those that do not lose siRNA can be re-silenced, implicating siRNA in reestablishing silencing in cis rather than in trans (Lippman et al. 2003). Similarly, late-flowering FWA alleles are stably inherited in backcrosses after being removed from metl or ddml mutant backgrounds because maintenance of epialleles is heritable (Soppe et al. 2000). Finally, it is possible that miRNA may guide DNA methylation of genes in some circ*mstances. mRNA from the PHABULOSA gene is targeted for cleavage by miRNA 165 and 166 in Arabidopsis, and the gene itself is methylated downstream from the region that matches the miRNA. Interestingly, this match spans an exon junction, so that the spliced RNA must interact with the miRNA if this guides methylation (Bao et al. 2004). However, other members of the same gene family are not methylated in this way, and neither are most other miRNA target genes (Martienssen et al. 2004; Ronemus and Martienssen 2005). Conversely, several other genes are methylated in the Arabidopsis genome, and typically at their 3' end, in a mechanism that requires METl but not DDM1 (Lippman et al. 2004; Tran et al. 2005). It remains to be seen whether RNA is involved in these cases.

Figure 7. Hypothetical Models for the Role of RNA Pol IV in RNAi-directed DNA and/or Histone H3 Methylation (0) RNA pol IV transcribes methylated DNA; RdR2 synthesizes dsRNA

using the RNA pol IV product. siRNAs then direct a methyltransferase complex to the chromosome. (b) RNA pol IV uses a dsRNA template synthesized by RdR2 to produce more ssRNA template.

ment for DDM1 (also a chromatin remodeler) in silencing transposons is far more severe than that of RNA pol IV or the other SWI2/SNF2 proteins. Genes can be silenced epigenetically by nearby transposons, and an important example in Arabidopsis is the imprinted homeobox gene FWA (Kinosh*ta et al. 2004). The first two exons of this gene are noncoding and form a tandem repeat due to the ancient integration of a SINE element at this site (Lippman et al. 2004). siRNAs from the SINE element are lost in metl (i.e., DNA methyltransferase) mutants, and the gene becomes strongly up-regulated in the inflorescence meristem, resulting in late

9 Conservation of RNAi-mediated Chromatin Modifications in Animals

Perhaps the most widely studied examples of epigenetic silencing are found in animals, including Drosophila and C. elegans, as well as the mouse. The role of RNA and RNA interference in transcriptional silencing and heterochromatic modifications appears to be conserved in some model animals as well as in protists and plants. In Drosophila, both PIWI and the PIWI class Argonaute hom*olog, Aubergine (Sting), are required for epigenetic and heterochromatic silencing (see also Chapter 5). Gypsy retrotransposons are the target of silencing in ovary follicle cells and female gonads by PIWI itself (Sarot et al. 2004). This is mediated by the heterochromatic gene Flamenco (with as-yet-unknown function), and requires the 5'UTR of the Gypsy polyprotein gene. The detection of 25-27-nt small RNAs from this region suggests it occurs

RNA fAN 0

via an RNAi-mediated mechanism. Cut-and-paste DNA transposons are also affected by RNAi. For example, certain telomeric P elements (a type of DNA transposon) can suppress transposition to elsewhere in the genome when inherited through the female germ line, resulting in a strongly repressive "cytotype." This repression is completely dependent on the PIWI hom*olog, Aubergine, as well as the Swi6 hom*olog HPI (Reiss et al. 2004). However, not all P-repressive cytotypes such as those mediated by other, nontelomeric P elements are dependent on Aubergine or HPl. Unlinked transgenes in Drosophila are silenced posttranscriptionally when present in many copies (PalBhadra et al. 1997, 2002). Silencing is associated with large amounts of 21-nt siRNA and depends on PIWI. Transgene fusions can also silence each other transcriptionally, in a manner that requires the Polycomb chromatin repressor. This silencing is not associated with increased levels of siRNA from the transgene transcript but is (largely) dependent on PIWI. Involvement of Polycomb in this example, and HPI in other examples, of PIWI-dependent silencing, implicates the RNAi pathway and histone methylation in the silencing process. Tandem transgene arrays also exhibit position-effect variegation in Drosophila, and this variegation is strongly suppressed by mutants in HPI as well as in piwi, aubergine, and the putative RNA helicase Spindle-E (homeless) (Pal-Bhadra et al. 2004). Transgenes inserted within centric heterochromatin are also affected, and heterochromatic levels ofH3K9me2 are reduced in spindle-E mutant cells. These observations strongly support a role for both chromatin proteins and components of the RNAi pathway in gene silencing within Drosophila heterochromatin. In the Drosophila male germ line, the heterochromatic Suppressor of Stellate repeats (Su(ste)), located on the Y chromosome, are transcribed first on the antisense strand, and then on both strands during spermatocyte development, possibly following the insertion of a nearby transposon (Aravin et al. 2001). These nuclear transcripts are required to silence sense transcripts of the closely related X-linked Stellate gene, whose overexpression results in defects in spermatogenesis. Although heterochromatic sequences are involved, silencing in this case appears to be posttranscriptional, is associated with 25-27-nt siRNA, and depends on both Aubergine and Spindle-E. In C. elegans, examples of TGS in somatic cells have been reported. This depends on the RNAi pathway genes rde-l, dcr-l, rde-4, and rrf-l, as well as HP 1 hom*ologs and the histone modification apparatus (Grishok et al. 2005).

H E T ERa C H ROM A T f N

ASS EM B L Y

•

163

Somatic heterochromatin is not widespread in C. elegans, but an example of naturally occurring RNAi-dependent heterochromatic silencing has been described in the germ line (Sijen and Plasterk 2003). During meiosis, unpaired sequences, such as the X chromosome in males, are silenced via H3K9me2, and this silencing depends on RNA-dependent RNA polymerase (Maine et al. 2005; see Chapter 15), reminiscent of meiotic silencing of unpaired DNA (MSUD) in Neurospora (see Shiu et al. 2001; Chapter 6). However, other components of the RNAi apparatus have not yet been implicated in this process, and it is not known whether it is related mechanistically to RNAimediated heterochromatin assembly in fission yeast. Finally, like Drosophila, mammalian cells lack genes related to RNA-dependent RNA polymerases found in plants, worms, and fungi. Nonetheless, antisense RNA has been implicated in the most widely studied epigenetic phenomena of all, imprinting and X inactivation (see Chapters 19 and 17, respectively). In the case of X inactivation, a 17-kb spliced and polyadenylated noncoding RNA known as Xist is required to silence the inactive X chromosome from which it is expressed. Conversely, Xist itself is silenced on the active X chromosome, a process that depends in part on the antisense RNA Tsix. Silencing is accompanied by modification of histones associated with upstream chromatin regions, which are marked with H3K9me2 and H3K27me3 (see Chapter 17). Silencing of other imprinted loci in the mouse, including Igf2r and the Dlkl-Gtl2 region, is also maintained by antisense transcripts from the paternal or maternal allele, respectively. In the case of Dlkl-Gtl2, this noncoding RNA is specifically processed into miRNA that targets the antisense transcript from the paternal allele, encoding a sushi (gypsy) class retrotransposon (Davis et al. 2005). Although the parallels with forward and reverse transcription from heterochromatic repeats in S. pombe are many, a role for RNAi itself in imprinting and X inactivation has so far proved elusive. Nonetheless, introduction of siRNA into cancer cell lines can result in chromatin being marked with H3K9me2 at hom*ologous promoters (Ting et al. 2005). In some cases, it can also result in DNA methylation (Morris et al. 2004), perhaps mediated by direct binding of small RNA with DNA methyltransferases and DNA methylation binding proteins (Jeffery and Nakielny 2004). Finally, Dicer knockout vertebrate cell lines have chromosome segregation defects reminiscent of those found in fission yeast mutants, accompanied by changes in heterochromatic morphology, expression of satellite repeats, and mislocalization of cohesin (f*ckagawa et al. 2004; Kanellopoulou et al. 2005).

164 •

C HAP T E R 8

10 Concluding Remarks

The possibility that genes may be regulated by small RNA molecules was suggested over 40 years ago (Jacob and Monod 1961), as well as the notion that "control RNA" might be related to repeats (Britten and Davidson 1969). Since the identification of the lambda and lac repressors as site-specific DNA-binding proteins in Escherichia coli and the infecting bacteriophage lambda (Gilbert and Muller-Hill 1966; Ptashne 1967), studies of gene regulation have focused almost exclusively on the role of nucleic-acid-binding proteins as specificity factors. The discovery of small RNA molecules as specificity agents in diverse RNA silencing mechanisms now clearly establishes a role for RNA as a sequence-specific regulator of genes and their RNA products. Studies in fission yeast, Arabidopsis, and other model organisms have revealed a surprisingly direct role for small RNAs in mediating epigenetic modifications of the genome that direct gene silencing and contribute to heterochromatic domains necessary for genome stability and nuclear division. Many important mechanistic questions remain at large, and future studies are likely to provide more surprises about how RNA regulates gene expression.

References Aravin A.A., Naumova N.M., Tulin A.V., Vagin VV, Rozovsky Y.M., and GvozdevVA. 2001. Double-stranded RNA-mediated silencing of genomic tandem repeats and transposable elements in the D. melanogaster germline. Curro BioI. 11: 1017-1027. Bao N., Lye K.W., and Barton M.K. 2004. MicroRNA binding sites in Arabidopsis class III HD-ZIP mRNAs are required for methylation of the template chromosome. Dev. Cell 7: 653-662. Barabino S.M., Ohnacker M., and Keller W. 2000. Distinct roles of two Ythlp domains in 3'-end cleavage and polyadenylation of yeast pre-mRNAs. EMBO f. 19: 3778-3787. Bartel D.P. 2004. MicroRNAs: Genomics, biogenesis, mechanism, and function. Cell 116: 281-297. Baulcombe D. 2004. RNA silencing in plants. Nature 431: 356-363. Bernstein E., Caudy A.A., Hammond S.M., and Hannon G.J. 2001. Role for a bidentate ribonuclease in the initiation step of RNA interference. Nature 409: 363-366. Britten R.J. and Davidson E.H. 1969. Gene regulation for higher cells: A theory. Science 165: 349-357. Buhler M., Verdel A., and Moazed D. 2006. Tethering RlTS to a nascent transcript initiates RNAi- and heterochromatin-dependent gene silencing. Cell 125: 873-886. Cao X., Aufsatz W., Zilberman D., Mette M.E, Huang M.S., Matzke M., and Jacobsen S.E. 2003. Role of the DRM and CMT3 methyltransferases in RNA-directed DNA methylation. Curro BioI. 13: 2212-2217. Chan S.W., Zilberman D., Xie Z., Johansen L.K., Carrington, J.e., and Jacobsen S.E. 2004. RNA silencing genes control de novo DNA methylation. Science 303: 1336. Chen e.e., Simard M.J., Tabara H., Brownell D.R., McCollough J.A.,

and Mello e.e. 2005. A member of the polymerase ~ nucleotidyltransferase superfamily is required for RNA interference in C. elegans. Curro BioI. 15: 378-383. Dalmay T., Hamilton A., Mueller E., and Baulcombe D.e. 2000. Potato virus X amplicons in Arabidopsis mediate genetic and epigenetic gene silencing. Plant Cell 12: 369-379. Davis E., Caiment E, Tordoir X., Cavaille J., Ferguson-Smith A., co*ckett N., Georges M., and Charlier e. 2005. RNAi-mediated allelic transinteraction at the imprinted RtlllPeglllocus. Curro BioI. 15: 743-749. Djupedal I., Portoso M., Spahr H., Bonilla e., Gustafsson e.M., Allshire R.e., and Ekwall K. 2005. RNA Pol II subunit Rpb7 promotes centromeric transcription and RNAi-directed chromatin silencing. Genes Dev. 19: 2301-2306. Ebbs M.L., Bartee L., and Bender J. 2005. H3 lysine 9 methylation is maintained on a transcribed inverted repeat by combined action of SUVH6 and SUVH4 methyltransferases. Mol. Cell. BioI. 25: 10507-10515. Elbashir S.M., Lendeckel W., and Tuschl T. 2001. RNA interference is mediated by 21- and 22-nucleotide RNAs. Genes Dev. 15: 188-200. Filipowicz W., Jaskiewicz L., Kolb EA., and Pillai R.S. 2005. Post-transcriptional gene silencing by siRNAs and miRNAs. Curro Opin. Struct. BioI. 15: 331-341. Fire A., Xu S., Montgomery M.K., Kostas S.A., Driver S.E., and Mello e.e. 1998. Potent and specific genetic interference by doublestranded RNA in Caenorhabditis elegans. Nature 391: 806-811. Francis N.J. and Kingston R.E. 2001. Mechanisms of transcriptional memory. Nat. Rev. Mol. BioI. 2: 409--421. f*ckagawa T., Nogami M., Yoshikawa M., Ikeno M., Okazaki T., Takami Y., Nakayama T., and Oshimura M. 2004. Dicer is essential for formation of the heterochromatin structure in vertebrate cells. Nat. Cell BioI. 6: 784-791. Galagan J.E. and Selker E.U. 2004. RIP: The evolutionary cost of genome defense. Trends Genet. 20: 417--423. Gasciolli V, Mallory A.e., Bartel D.P., and Vaucheret H. 2005. Partially redundant functions of Arabidopsis DICER-like enzymes and a role for DCL4 in producing trans-acting siRNAs. Curro BioI. 15: 1494-1500. Gasser S.M., Hediger E, Taddei A., Neumann ER., and Gartenberg M.R. 2004. The function of telomere clustering in yeast: The circe effect. Cold Spring Harbor Symp. Quant. BioI. 69: 327-337. Gilbert W. and Muller-Hill B. 1966. Isolation of the Lac repressor. Pmc. Natl. Acad. Sci. 56: 1891-1898. Grewal S.L 2000. Transcriptional silencing in fission yeast. f. Cell. Physiol. 184: 311-318. Grewal S.l. and Moazed D. 2003. Heterochromatin and epigenetic control of gene expression. Science 30: 798-802. Grishok A., Sinskey J.L., and Sharp P.A. 2005. Transcriptional silencing of a transgene by RNAi in the soma of C. elegans. Genes Dev. 19: 683-696. Hall l.M., Shankaranarayana G.D., Noma K., Ayoub N., Cohen A., and Grewal S.L 2002. Establishment and maintenance of a heterochromatin domain. Science 297: 2232-2237. Hamilton A.J. and Baulcombe D.e. 1999. A species of small antisense RNA in posttranscriptional gene silencing in plants. Science 286: 950-952. Hampsey M. and Reinberg D. 2003. Tails of intrigue: Phosphorylation of RNA polymerase II mediates histone methylation. Cell 113: 429-432. Hannon G.J. 2002. RNA interference. Nature 418: 244-251. Herr A.J., Jensen M.B., Dalmay T., and Baulcombe D.e. 2005. RNA polymerase IV directs silencing of endogenous DNA. Science 308: 118-120.

RNA fAN 0

Higa L.A., Mihaylov l.S., Banks D.P., Zheng j., and Zhang H. 2003. Radiation-mediated proteolysis of CDT! by CVL4-ROCI and CSN complexes constitutes a new checkpoint. Nat. Cell BioI. 5: 1008-1015. Hong E.E., Villen j., Gerace E.L., Gygi, S.P., and Moazed D. 2005. A Cullin E3 ubiquitin ligase complex associates with Rikl and the Clr4 histone H3-K9 methyltransferase and is required for RNAimediated heterochromatin formation. RNA BioI. 2: 106-111. Horn P.j., Bastie J.N., and Peterson c.L. 2005. A Rikl-associated, cullindependent E3 ubiquitin ligase is essential for heterochromatin formation. Genes Dev. 19: 1705-1714. jacob E and Monod J. 1961. Genetic regulatory mechanisms in the synthesis of proteins.]. Mol. BioI. 3: 318-356. jeffery L. and Nakielny S. 2004. Components of the DNA methylation system of chromatin control are RNA-binding proteins. ]. BioI. Chem. 279: 49479-49487. jia S., Kobayashi R., and Grewal S.l. 2005. Vbiquitin ligase component Cul4 associates with Clr4 histone methyltransferase to assemble heterochromatin. Nat. Cell BioI. 7: 1007-1013. Jia S., Noma K., and Grewal S.l. 2004. RNAi-independent heterochromatin nucleation by the stress-activated ATF/CREB family proteins. Science 304: 1971-1976. Kanellopoulou c., Muljo S.A., Kung A.L., Ganesan S., Drapkin R., jenuwein T., Livingston D.M., and Rajewsky K. 2005. Dicer-deficient mouse embryonic stem cells are defective in differentiation and centromeric silencing. Genes Dev. 19: 489-501. Kanno T., Huettel B., Mette M.E, Aufsatz W., Jaligot E., Daxinger L., Kreil D.P., Matzke M., and Matzke A.J. 2005. Atypical RNA polymerase subunits required for RNA-directed DNA methylation. Nat. Genet. 37: 761-765. Kanno T., Mette M.E, Kreil D.P., Aufsatz W., Matzke M., and Matzke A.j. 2004. Involvement of putative SNF2 chromatin remodeling protein DRDI in RNA-directed DNA methylation. Curl'. BioI. 14: 801-805. Kanoh j., Sadaie M., Vrano T., and Ishikawa E 2005. Telomere binding protein Tazl establishes Swi6 heterochromatin independently of RNAi at telomeres. Curl'. BioI. IS: 1808-1819. Kata H., Goto D.B., Martienssen R.A., Vrano T., Furukawa K., and Murakami Y. 2005. RNA polymerase II is required for RNAidependent heterochromatin assembly. Science 309: 467-469. Kim j.K., Gabel H.W, Kamath R.S., Tewari M., Pasquinelli A., Rual J.E, Kennedy S., Dybbs M., Bertin N., Kaplan j.M., et al. 2005. Functional genomic analysis of RNA interference in C. elegans. Science 308: 1164-1167. Kinosh*ta T., Miura A., Choi Y., Kinosh*ta Y., Cao X., Jacobsen S.E., Fischer R.L., and Kakutani T. 2004. One-way control of FWA imprinting in Arabidopsis endosperm by DNA methylation. Science 303: 521-523. Li E, Goto D.B., Zaratiegui M., Tang X., Martienssen R., and Cande W.Z. 2005. Two novel proteins, Dosl and Dos2, interact with Rikl to regulate heterochromatic RNA interference and histone modification. Curl'. BioI. IS: 1448-1457. Li H.W. and Ding S.W. 2005. Antiviral silencing in animals. FEBS Lett. 579: 5965-5973. Lippman Z., Gendrel AV, Black M., Vaughn M.W, Dedhia N., McCombie W.R., Lavine, K., Mittal V., May B., Kasschau K.D., et al. 2004. Role of transposable elements in heterochromatin and epigenetic control. Nature 430: 471-476. Lippman Z. and Martienssen R. 2004. The role of RNA interference in heterochromatic silencing. Nature 431: 364-370. Lippman Z., May B., Yordan c., Singer T., and Martienssen R. 2003. Distinct mechanisms determine transposon inheritance and methyla-

H E T E ROC H ROM A T f N

ASS E M B L Y

•

165

tion via small interfering RNA and histone modification. PLoS BioI. I: E67. Maine E.M., Hauth J., RatliffT., Vough V.E., She X., and Kelly WG. 2005. EGO-I, a putative RNA-dependent RNA polymerase, is required for heterochromatin assembly on unpaired dna during C. elegans meiosis. Curl'. BioI. IS: 1972-1978. Martienssen R.A. 2003. Maintenance of heterochromatin by RNA interference of tandem repeats. Nat. Genet. 35: 213-214. Martienssen R.A., Zaratiegui M., and Goto D.B. 2005. RNA interference and heterochromatin in the fission yeast Schizosaccharomyces pombe. Trends Genet. 21: 450-456. Martienssen R., Lippman Z., May B., Ronemus M., and Vaughn M. 2004. Transposons, tandem repeats, and the silencing of imprinted genes. Cold Spring Harbor Symp. Quant. BioI. 69: 371-379. Mathieu O. and Bender j. 2004. RNA-directed DNA methylation.]. Cell Sci. 117: 4881-4888. Matzke M., Aufsatz W., Kanno T., Daxinger L., Papp 1., Mette M.E, and Matzke A.j. 2004. Genetic analysis of RNA-mediated transcriptional gene silencing. Biochim. Biophys. Acta 1677: 129-141. May B.P., Lippman Z.B., Fang Y., Spector D.L., and Martienssen R.A. 2005. Differential regulation of strand-specific transcripts from Arabidopsis centromeric satellite repeats. PLoS Genet. 1: e79. Mette M.E, van der Wind en J., Matzke M.A., and Matzke A.j. 1999. Production of aberrant promoter transcripts contributes to methylation and silencing of unlinked hom*ologous promoters in trans. EMBO]. 18: 241-248. Mette M.E, Aufsatz W., van der Winden J., Matzke M.A., and Matzke A.j. 2000. Transcriptional silencing and promoter methylation triggered by double-stranded RNA. EMBO]. 19: 5194-5201. Morris K.Y., Chan S.w., jacobsen S.E., and Looney D.j. 2004. Small interfering RNA-induced transcriptional gene silencing in human cells. Science 305: 1289-1292. Motamedi M.R., Verdel A., Colmenares S.u., Gerber SA, Gygi S.P., and Moazed D. 2004. Two RNAi complexes, RITS and RDRC, physically interact and localize to noncoding centromeric RNAs. Cell 119: 789-802. Napoli c., Lemieux c., and Jorgensen R. 1990. Introduction of a chimeric chalcone synthase gene into Petunia results in reversible co-suppression of hom*ologous genes in trans. Plant Cell 2: 279-289. Noma K., Sugiyama T., Cam H., Verde] A., Zofall M., jia S., Moazed D., and Grewal S.l. 2004. RITS acts in cis to promote RNA interferencemediated transcriptional and post-transcriptional silencing. Nat. Genet. 36: 1174-1180. Onodera Y., Haag j.R., Ream T., Nunes P.c., Pontes 0., and Pikaard C.S. 2005. Plant nuclear RNA polymerase IV mediates siRNA and DNA methylation-dependent heterochromatin formation. Cell 120: 613-622. Pal- Bhadra M., Bhadra V., and Birchler J.A. 1997. Cosuppression in Drosophila: Gene silencing of Alcohol dehydrogenase by white-Adh transgenes is Polycomb dependent. Cell 90: 479-490. - - - . 1999. Cosuppression of nonhom*ologous transgenes in Drosophila involves mutually related endogenous sequences. Cell 99: 35-46. - - - . 2002. RNAi related mechanisms affect both transcriptional and posttranscriptional transgene silencing in Drosophila. Mol. Cell. 9: 315-327. Pal-Bhadra M., Leibovitch B.A., Gandhi S.G., Rao M., Bhadra u., Birchler J.A., and Elgin S.c. 2004. Heterochromatic silencing and HPI localization in Drosophila are dependent on the RNAi machinery. Science 303: 669-672. Partridge j.E, Scott K.S., Bannister A.J., Kouzarides T., and Allshire R.C. 2002. cis-acting DNA from fission yeast centromeres mediates his-

166 •

C HAP T E R 8

tone H3 methylation and recruitment of silencing factors and cohesin to an ectopic site. Curro BioI. 12: 1652-1660. Petrie v.J., Wuitschick J.D., Givens CD., Kosinski A.M., and Partridge J.E 2005. RNA interference (RNAi)-dependent and RNAi-independent association of the Chp 1 chromodomain protein with distinct heterochromatic loci in fission yeast. Mol. Cell. BioI. 25: 2331-2346. Pidoux A.L. and Allshire R.C 2004. Kinetochore and heterochromatin domains of the fission yeast centromere. Chromosome Res. 12: 521-534. Plasterk R.H. 2002. RNA silencing: The genome's immune system. Science 296: 1263-1265. Provost P., Silverstein R.A., Dishart D., Walfridsson J., Djupedal 1., Kniola B., Wright A., Samuelsson B., Radmark 0., and Ekwall K. 2002. Dicer is required for chromosome segregation and gene silencing in fission yeast cells. Proc. Natl. Acad. Sci. 99: 16648-16653. Ptashne M. 1967. Specific binding of the lambda phage repressor to lambda DNA. Nature 214: 232-234. Reinhart B.J. and Bartel D.E 2002. Small RNAs correspond to centromere heterochromatic repeats. Science 297: 1831. Reiss D., Josse T., Anxolabehere D., and Ronsseray S. 2004. aubergine mutations in Drosophila melanogaster impair P cytotype determination by telomeric P elements inserted in heterochromatin. Mol. Genet. Genomics. 272: 336-343. Ronemus M. and Martienssen R. 2005. RNA interference: Methylation mystery. Nature 433: 472--473. Sarot E., Payen-Groschene G., Bucheton A., and Pelisson A. 2004. Evidence for a piwi-dependent RNA silencing of the gypsy endogenous retrovirus by the Drosophila melanogaster flamenco gene. Genetics 166: 1313-1321. Shiu P.K., Raju N.B., Zickler D., and Metzenberg R.L. 2001. Meiotic silencing by unpaired DNA. Cell 107: 905-916. Sigova A., Rhind N., and Zamore P.D. 2004. A single Argonaute protein mediates both transcriptional and posttranscriptional silencing in Schizosaccharomyces pombe. Genes Dev. 18: 2359-2367. Sijen T and Plasterk R.H. 2003. Transposon silencing in the Caenorhabditis elegans germ line by natural RNAi. Nature 426: 310-314. Sijen T, Fleenor J., Simmer E, Thijssen K.L., Parrish S., Timmons 1., Piasterk R.H., and Fire A. 2001. On the role of RNA amplification in dsRNA-triggered gene silencing. Cell 107: 465-476. Soppe W.J., Jacobsen S.E., Alonso-Blanco C, Jackson J.P., Kakutani T,

Koornneef M., and Peeters A.J. 2000. The late flowering phenotype of fwa mutants is caused by gain-of-function epigenetic alleles of a homeodomain gene. Mol. Cell. 6: 791-802. Sugiyama T, Cam H., Verdel A., Moazed D., and Grewal S.L 2005. RNAdependent RNA polymerase is an essential component of a selfenforcing loop coupling heterochromatin assembly to siRNA production. Proc. Natl. Acad. Sci. 102: 152-157. Ting A.H., Schuebel K.E., Herman J.G., and Baylin S.B. 2005. Short double-stranded RNA induces transcriptional gene silencing in human cancer cells in the absence of DNA methylation. Nat. Genet. 37: 906-910. Tran R.K., Zilberman D., de Bustos C, Ditt R.E, Henikoff J.G., Lindroth A.M., Delrow J., Boyle T, Kwong S., Bryson TD., et al. 2005. Chromatin and siRNA pathways cooperate to maintain DNA methylation of small transposable elements in Arabidopsis. Genome BioI. 6: R90. Vaughn M.W. and Martienssen R.A. 2005. Finding the right template: RNA Pol IV, a plant-specific RNA polymerase. Mol. Cell 17: 754-756. Verdel A., Jia S., Gerber S., Sugiyama T., Gygi S., Grewal S.L, and Moazed D. 2004. RNAi-mediated targeting of heterochromatin by the RITS complex. Science 303: 672-676. Verdel A. and Moazed D. 2005. RNAi-directed assembly of heterochromatin in fission yeast. FEBS Lett. 579: 5872-5878. Volpe TA., Kidner C, Hall LM., Teng G., Grewal S.L, and Martienssen R.A. 2002. Regulation of heterochromatic silencing and histone H3 lysine-9 methylation by RNAi. Science 297: 1833-1837. Volpe T, Schramke V., Hamilton G.L., White SA, Teng G., Martienssen R.A., and Allshire R.C 2003. RNA interference is required for normal centromere function in fission yeast. Chromosome Res. 11: 137-146. Wassenegger M., Heimes S., Riedel 1., and Sanger H.L. 1994. RNAdirected de novo methylation of genomic sequences in plants. Cell 76: 567-576. Zamore P.D. 2002. Ancient pathways programmed by small RNAs. Science 296: 1265-1269. Zamore P.D., Tuschl T, Sharp P.A., and Bartel D.P. 2000. RNAi: Doublestranded RNA directs the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell 101: 25-33. Zhong w., Feng H., Santiago EE., and Kipreos E.T 2003. CUL-4 ubiquitin ligase maintains genome stability by restraining DNA-replication licensing. Nature 423: 885-889.

HAP

•

Epigenetic Regulation In Plants Marjori Matzke and Ortrun Mittelsten Scheid Gregor Mendel Institute of Molecular Plant Biology, Austrian Academy of Sciences, A -1030 Vienna, Austria

CONTENTS 1. Benefits of Plants in Epigenetic Research, 169 7.7 Plants and Mammals Are Similar in Terms of (Epi)Genome Organization, 769

3.7 Elaboration of RNA i-media ted Silencing in Plants, 780

7.2 Plants Provide Additional Topics for Epigenetic Research, 769

3.2 Pathway 7: Transgene-related Posttranscriptional and Virus-induced Silencing (PTGS/VIGS), 787

7.3 Plants Tolerate Methodological Approaches That Are Difficult in Mammals, 770

3.3 Pathway 2: Regulation of Plant Development by RNAs and Trans-acting siRNAs, 783

7.4 Plants Have a Proven Record of Contributing to Epigenetic Research, 774

3.4 Pathway 3: Transgene-related Transcriptional Silencing, RNA-directed DNA Methylation, and Heterochromatin Formation, 784

2. Molecular Components of Chromatin in Plants, 175 2.7 Regulators of DNA Methylation in Plants, 775 2.2 Histone-modifying Enzymes, 777 2.3 Other Chromatin Proteins, 778

4. Epigenetic Regulation without RNA Involvement, 186 5. Outlook, 186 References,187

3. Molecular Components of RNAi-mediated Gene Silencing Pathways, 180

167

GENERAL SUMMARY Plants are masters at epigenetic regulation. All major epigenetic mechanisms present in eukaryotes are used by plants and often elaborated to a degree unsurpassed in other kingdoms. DNA methylation, commonly associated with gene silencing, is found in CpG, CpNpG, and CpNpN nucleotide groups in plant genomes and relies on a number of plant-specific proteins, including several that might be specialized for active demethylation. Histonemodifying enzymes that modulate chromatin structure in plants are generally conserved within the catalytic domains, but they are frequently encoded by comparatively large gene families, which allows more extensive diversification or redundancy of gene function. RNAimediated gene-silencing pathways have also diversified in plants to combat viruses, tame transposons, orchestrate development, and organize the genome. Although the interplay between DNA methylation and histone modifications has been recognized for some time, the recent discovery that these modifications can be targeted to specific regions of the genome by the RNAi machinery has added a new dimension to epigenetics research. The intersections and overlaps among these silencing pathways provide plants with a multilayered and robust epigenetic circuitry. The prominence of epigenetic regulation in plants reflects their evolutionary history, mode of development, and "Iifestyle." Polyploidization-an increase in the number of sets of chromosomes-is a recurring

event in plant lineages, amplifying gene families and fostering functional specialization of duplicated genes. Unlike mammals, where organ and tissue formation is largely completed during embryonic development, plants grow by continuously producing new aerial and underground parts from self-sustaining stem cell populations called meristems. Consequently, postembryonic development of plants is shaped by environmental influences and is characterized by a high degree of plasticity and variability. Because plants are unable to escape their surroundings, they are forced to cope with changeable and often unfavorable growth conditions. The inherent flexibility of epigenetic regulatory mechanisms can facilitate rapid changes in gene activity and fine-tune gene expression patterns, enabling plants to survive and reproduce successfully in unpredictable environments. Historically, plants have provided excellent systems for discovering and analyzing epigenetic phenomena. A change from bilateral to radial symmetry in some variants of the plant Linaria vulgaris (see title figure), observed by Carl von Linne in the 18th century, was pinpointed to an epigenetic modification of the eye/aidea gene, regulating flower development. Progress has been particularly impressive in the past 5 years, owing to the availability of the genome sequence of Arabidapsis thaliana-a "useful weed" that is highly amenable to genetic analyses-and to the synergy created by parallel studies of epigenetic phenomena in animal and fungal systems.

E PIG ENE TIC

1 Benefits of Plants in Epigenetic Research 1.1 Plants and Mammals Are Similar in Terms of (Epi)Genome Organization

Soon after biology was established as an independent scientific discipline, animals and plants were grouped into separate kingdoms, and this view became traditionally manifest by training biologists in either zoology or botany. Of course, there are good arguments for this partition, including heterotrophic (i.e., requiring organic matter for growth) versus mainly autotrophic (i.e., selfsustaining) energy generation, mobile versus sessile lifestyle, potentially migrating and flexible cells versus motionless and rigid cells. However, geneticists and molecular biologists have uncovered in recent decades a degree of congruence between animals and plants that was surprising in the light of their long evolutionary separation. Common principles include sexual propagation via meiosis and fertilization, the regulation of individual development by a few master genes, the control of cell division and proliferation by related factors, and the reception of environmental factors through similar signaling cascades. This similarity extends to many aspects of genome and epigenome .organization. The resemblance is particularly striking between plants and mammals, which have comparable genome sizes, genome complexities, and ratios of heterochromatin. As in many other eukaryotes, euchromatin and heterochromatin are characterized by specific acetylation and methylation of histones, but heterochromatin of plants and mammals is specified additionally by significant DNA cytosine methylation. A comparison of components participating in genome organization and epigenetic regulation across different model systems reveals that there are more common features between plants and mammals than there are within the animal kingdom itself (Table 1). Therefore, even if interest is driven by an anthropocentric focus, similar questions can be addressed in plants and mammals, and basic information can often be shuttled between both systems.

1.2 Plants Provide Additional Topics for Epigenetic Research

In addition to elements shared with mammals, plants have acquired some specialties that are potentially relevant for epigenetic phenomena. Whereas in mammals fertilization is achieved by fusion of two haploid cells that are direct products of the preceding meiosis, plants have a haploid (gametophyte) growth stage between meiosis

REG U L A T ION

I N

P LAN T 5

169

and fertilization (Fig. 1). The gametophytes correspond to the germinating pollen grain (male) and the embryo sac (female), each with several haploid nuclei that originate from the meiotic products by two or three subsequent mitotic divisions, respectively. In the gametophytic phase, any loss of genetic or epigenetic information cannot be compensated by information on hom*ologous chromosomes or alleles. Although extensive studies have not yet been undertaken, there is no evidence for a massive programmed erasure of epigenetic marks during plant gametogenesis as occurs in mammals, and this might explain why epigenetic changes are often transmissible through meiosis in plants. Another distinctive feature of plants is the less welldefined germ line and its separation from somatic cells only late in development (Fig. 1). Plants have apical meristems, which are growth points at shoot and root tips that generate new tissues and organs. The shoot apical meristem eventually forms the flowers that generate the gametes for sexual propagation, but additional lateral meristems can also grow out and form flowers, and many plants have developed specialized organs like rhizomes, tubers, or bulbs that contain meristems. These mechanisms of vegetative propagation can be even more common or successful than seed dissemination. Embryos can be formed not only by development of a fertilized egg, but also from somatic tissue (somatic embryogenesis). Upon manipulation in tissue culture, some differentiated somatic cells can undergo dedifferentiation and be reprogrammed toward alternative differentiation. This means that somatic cloning, still with low success rates in mammals, is routine in many plant species, and countless "green Dollies" have been produced over the years. Nevertheless, a surprising amount of phenotypic variability has been observed in supposedly genetically uniform populations of cloned plants. This so-called "somaclonal variation" has a strong epigenetic basis and is potentially useful for plant breeding and adaptation (for review, see Kaeppler et al. 2000). Another plant-specific feature is the existence of plasmodesmata, cytoplasmatic bridges between individual cells, which are permeable to small molecules, some proteins, and RNAs, and viral genomic information. Despite the high degree of interconnection, plant shoots can be cut and grafted as scions on top of genetically different stocks (Fig. 1). This permits the production of chimeras in which vegetative tissue, and tissue that gives rise to progeny, are genetically different. Therefore, whereas epigenetic marks are transmitted through the germ line, they seem to be more flexible and reversible in plants relative to animals.

170

•

CHAPTER

Table 1. Compilation of genomic and epigenetic components in epigenetic model systems Feature Genome size Number of genes

Saccharomyces Schizosaccharomyces cerevisiae pombe 12 Mb 14 Mb

Neurospora crassa 40 Mb

Caenorhabditis elegans 100 Mb

Drosophila melanogaster 180 Mb

Mammals 3,400 Mb

Plants 1505,000 Mb

6,000

5,000

10,000

20,000

14,000

25,00030,000

25,00040,000

1.45 kb

1.7

2 kb

5 kb

35-46 kb

2 kb

Average number of introns/gene

s 1 (4% of genes with introns)

2 (40% of genes with introns)

6-8

4-5

% Genome as protein coding

4 (Hs)

26 (At) 10(05)

Transposon silencing

(+)

+ (+ RIP)'

Average size of genes

Imprinting

H3K9 (+)

H3K9 + H3K27 +

RNAi mechanisms Repressive histone methylation DNA methylation at CG at CNG/CNN

+ (+)

(-)

Genes related to DNA methylation and recognition

HP1-like protein

Polycomb proteins

(At) Arobidopsis tha/iana, (Cb) Caenorhabditis briggsae, (Ce) Caenorhabditis e/egans, (Dm) Drosophila melanogoster, (Hs) hom*o sapiens, (Os) Oryzo sativa, (Pp) Pristionchus pacificus. • Repeat-induced point mutation, see Chapter 6. b

Chromosome- or genome-wide rather than gene-specific.

Mutated Dnmt2.

Dnmt2 (Pp) and MBD-domain proteins (Ce, Cb, Pp) .

• Dnmt2 and MBD-domain proteins (Dm).

Plants have a higher tolerance toward polyploidy (the multiplication of the whole-chromosome complement) than mammals. The numerous wild polyploid species and cultivated polyploid plants-such as wheat, cotton, potato, peanut, sugarcane, and tobacco-suggest that polyploidy offers certain competitive advantages. Inspection of many plant genome sequences, including the small genome of Arabidopsis thaliana, provided clear evidence for ancient genome and gene duplication events. Even diploid plants can contain polyploid cells, which arise much more frequently than the few examples of highly specialized polyploid cells in mammals. The formation of polyploids is often associated with significant genomic and epigenetic changes (for review, see Adams

and Wendel 2005). Some of these changes occur within one or a few generations and can contribute to rapid adaptation and evolution in plants.

1.3 Plants Tolerate Methodological Approaches That Are Difficult in Mammals

In mammals, genetic approaches are limited by demanding procedures for generating mutations and by the requirement for mating in order to establish hom*ozygous genotypes, which are mandatory for revealing recessive traits. In contrast, plants allow efficient mutagenesis, either by chemical or physical treatments or by largely random insertion of sequence-tagged transgenes

E PIG ENE TIC

SOMATIC PROPAGATION endoreduplication polyploidy

REG U L A T ION

P LAN T 5

1 71

SEXUAL LIFE CYCLE

chimera

by grafting

sporophyte (2n)

~_ _--:m~eiosis

Figure 1. Specialties of the Plant life Cycle

••

•

~ mitosis 1 ~

9 ~

mitosis 2

l a

•

mitosis

protoplast

tissue explant gametophytes (1 n)

tuber somatic embryogenesis

double

\I~

fertilization

or transposable elements. Individuals with hom*ozygous mutations are easily generated in Mendelian ratios in the following generation by self-pollination. Screens for mutations in epigenetic regulators were based on the recovery of gene expression from epigenetically inactivated marker genes or for epigenetic down-regulation of stably active reporter genes. In addition to such forwarddirected unbiased methodology, the rapidly growing collections of insertion mutants within defined genomic integration sites permit reverse genetic approaches by analyzing the effects of mutations in chosen, defined genes orthologous to epigenetic regulators in other model organisms (Table 2).

Plants can propagate sexually (gametogenesis, fertilization, and seed formation, right) as well as somatically (vegetative sprigs, de- and re-differentiation or embryogenesis, left). The body of higher plants, with roots, stem, leaves, and flowers, is the diploid sporophyte. During meiosis, the chromosome number is reduced to half. Whereas in animals the meiotic products form the gametes without further division and fuse directly to produce the diploid embryo, plants form haploid male or female gametophytes by two or three mitotic divisions, respectively. The pollen tube ultimately contains one vegetative (white) and two generative (black) nuclei. The two generative nuclei fertilize the egg cell (black) and the central cell, which has a diploid nucleus derived from fusion of the two polar nuclei (yellow). This double fertilization gives rise to the diploid embryo and the triploid endosperm, which provides a nutrient source for the developing embryo. After seed germination, the embryo will grow into a new sporophyte. In addition, most plants have the potential for vegetative propagation through activation of quiescent lateral meristems, outgrowth of specialized root structures such as tubers, amplification in tissue culture, and even regeneration from individual somatic cells after removal of the cell wall (protoplasts). Endoreduplication is frequent in plants, producing polyploid cells or tissues. Plants can be grafted to produce chimeras. In summary, genetic and epigenetic information in plants therefore passes a much less well-defined germ line than in animals.

The number of members within families of hom*ologous genes can differ significantly between plants and mammals. As a consequence of functional redundancy, some mutations are less severe in either plants or mammals, and this can be important if a complete loss of function would eliminate the corresponding individuals prior to analysis in early development. onessential genes in pathways that determine coloration of plant tissues permit easy and inexpensive gene expression readouts in vivo (Fig. 2a-e). Other epigenetic changes can be followed by scoring for morphological defects, tolerated by many plants without lethal consequences (Fig. 2£). Additionally, thousands of individual plants can be

172 • C HAP T E R 9

Table 2. Components of epigenetic regulation in the model plant Arabidopsis thaliana identified in forward or reverse genetic screens Gene or mutant acronym

Type of protein, confirmed or putative function

Gene or mutant name

Mutant screen

DNA methylation

MEn DDM2

Methyltransferase Decreased DNA methylation

DNA methyltransferase (CG)

reactivation of endogenous repeats reactivation of transgenes interference with RdDM hypomethylation of centromeric repeats

CMTJ

Ch romomethyltra nsferase

DNA methyltransferase (non CG)

reactivation of SUP-elk reactivation of PAl

DRMI

Domain-rearranged methyltransferase

DRM2

Domain-rearranged methyltransferase

HOGI

de novo DNA methyltransferase

search for insertion mutants

de novo DNA methyltransferase

search for insertion mutants

hom*ology-dependent gene silencing

S-adenosyl-L-hom*ocysteine hydrolase

reactivation of repetitive transgene

ROSI

Repressor of silencing

DNA glycosylase-domain protein

inactivation of transgene

DME

Demeter

DNA glycosylase-domain protein

seed abortion

HDAI

Histone deacetylase

histone deacetylase

search for insertion mutants antisense expression

HDA6 SiLl AXEl RTSI

Histone deacetylase Modifier of silencing Auxin-gene repression RNA-mediated transcriptional silencing

histone deacetylase

reactivation of repetitive transgene

SUVH2

Su(var)3-9 hom*olog

histone methyltransferase

search for insertion mutants antisense expression

SUVH4 KYPI

Su(var)3-9 hom*olog Kryptonite

histone methyltransferase

reactivation of PAl reactivation of SUP-elk

DDMI SaM

Decreased DNA methylation Somniferous

SWI2/SNF2 ATPase

hypomethylation of centromeric repeats reactivation of repetitive transgene

DRDI

Defective in RNA-directed DNA methylation

SWI2/SNF2 ATPase

interference with RdDM

SPD

Splayed

SWI2/SNF2 ATPase

altered meristem maintenance

PIE

Photoperiod-independent early flowering

ATP-dependent chromatinremodeling protein

change of flowering time

PKL

Pickle

CHD3 chromatin-remodeling factor

abnormal root development

FASI

Fasciated

chromatin assembly factor subunit

altered morphology

FAS2

Fasciated

chromatin assembly factor subunit

altered morphology

BRUI

Brushy

uncharacterized protein

DNA damage sensitivity

LHPI

Like heterochromatin protein

formation of repressive chromatin

early flowering and altered morphology

MOMI

Morpheus' molecule

incomplete SWI2/SNF2 ATPase

reactivation of repetitive transgene

RPA2

Replication protein A

subunit of the ssDNA-binding replication protein complex

suppressor screen in ras 1 reactivation of repetitive transgene

Histone modification

interference with RdDM

Chromatin formation/remodeling

RNAi-mediated silencing

DCLl CAF1 SINI EMB76 SUSI

Dicer-like Carpel factory Short integuments Embryo-defective Suspensor

RNase III (dsRNase)

search for insertion mutants abnormal flower development abnormal ovule development arrested embryo development suspensor proliferation

E PIG ENE TIC

REG U L A T ION

P LAN T S

•

1 73

Table 2. (continued) Gene or mutant acronym

Gene or mutant name

OCL2

Dicer-like

RNase III (dsRNase)

search for insertion mutants

oeu

Dicer-like

RNase III (dsRNase)

search for insertion mutants

AG07

Argonaute

PAZ-PIWI domain protein

altered morphology reactivation of transgenes

AG04

Argonaute

PAZ-PIWI domain protein

reactivation of SUP-c1k

AGOl ZIP

Argonaute Zippy

PAZ-PIWI domain protein

timing and type of trichome development

AG070 PNH/ZLL

Argonaute

PAZ-PIWI domain protein

meristem defects

ROR7

RNA-dependent RNA polymerase

search for insertion mutants

ROR2

RNA-dependent RNA polymerase

search for insertion mutants

ROR6 SOE7 SGS2

RNA-dependent RNA polymerase Silencing-defective Suppressor of gene silencing

RNA-dependent RNA polymerase

reactivation of transgene altered trichome development

NRP07a SOE4

RNA polymerase Silencing-defective

RNA polymerase IV subunit

reactivation of transgene

NRP07b OR03

RNA polymerase Defective in RNA-directed DNA methylation

RNA polymerase IV subunit

interference with RdDM

NRP02a OR02

RNA polymerase Defective in RNA-directed DNA methylation

RNA polymerase IV subunit

search for insertion mutants interference with RdDM

SOE3

Silencing-defective

RNA helicase

reactivation of transgene

SGS3

Suppressor of gene silencing

coiled-coil protein

reactivation of transgene altered trichome development

HEN7

HUA enhancer

dsRNA binding, methyltransferase

reactivation of transgene altered morphology

HYL7

Hyponastic leaves

nuclear dsRNA-binding protein

altered morphology search for insertion mutants

WEX

Werner syndrome-like exonuclease

RNase D exonuclease

search for insertion mutants

XRN4

XRN hom*olog

exoribonuclease

search for insertion mutants

HST

Hasty

miRNA export receptor

timing and type of trichome development

Type of protein, confirmed or putative function

Mutant screen

RNAi-mediated silencing (continued)

Polycomb group proteins Medea Fertilization-independent seeds

Polycomb group protein

FlS7

seed abortion fertilization-independent seed set

CLF

Curly leaf

Polycomb group protein

altered leaf morphology

FIE FlS3

Fertilization-independent endosperm Fertilization-independent seeds

Polycomb group protein

fertilization-independent seed set

MSI7

Multicopy suppressor of IRA hom*olog

Polycomb group protein chromatin assembly factor subunit

search for insertion mutants defects in endosperm patterning

SWN

SWinger

Polycomb group protein

search for insertion mutants

EMF2

Embryonic flower

Polycomb group protein

lack of vegetative stage

VRN2

Vernalization

Polycomb group protein

late flowering in spite of vernalization

MEA

erochromatin was first made after cytological analysis in mosses (Heitz 1928). The observation of heritable but reversible changes in gene expression after allelic interaction in tomato and maize, later termed paramutation, was early evidence for non-Mendelian genetics (for review, see Chandler and Stam 2004), now also apparent in mammalian systems. Likewise, parental imprinting of individual genes in plants was first observed in maize (for review, see Alleman and Doctor 2000). The repeated occurrence of individuals with altered flower symmetry, already described by Carl von Linne as "peloria" (monster) (see title figure), could now be explained by the formation of an epiallele, a stable epigenetic modification of a regulatory gene with the same sequence as the expressed version (Cubas et al. 1999). Cytological analysis in plants revealed changes in secondary chromosome constrictions that later were linked with nucleolar dominance, the silencing of one parental set of rRNA genes in interspecific hybrids, md shown to depend on epigenetic regulation (for review, see Pikaard 2000). The pioneering work on transposable elements in maize, by Barbara McClintock and e>ther workers, revealed numerous links between their senetic behavior and epigenetic regulation (for review, ,ee Fedoroff and Chandler 1994). Indeed, extant transposons and their degenerate remains provide the foundation for establishing epigenetic modifications throughout plant genomes (Section 3.4). More recently, when transgenic technology became routine in the late 1980s for plants such as tobacco, petunia, and Arabidopsis, a major advance in epigenetic research arose from the unexpected results obtained in the course of introducing marker genes (for review, see rorgensen 2003; Matzke and Matzke 2004). The concept e>fhom*ology-dependent gene silencing was formulated as it became evident that silencing was often correlated with multiple copies of linked or unlinked transgenes. Differ~nt cases of hom*ology-dependent gene silencing were due :0 either enhanced turnover of mRNA (posttranscripjonal gene silencing, PTGS) or repression of transcripjon (transcriptional gene silencing, TGS), both of which iVere correlated with increased cytosine methylation of ;ilenced genes. A striking example of PTGS in transgenic Jetunia was initially termed "cosuppression": Attempts to modify floral coloration by overexpression of chalcone iynthase (CHS) genes that condition purple petals often Jroduced variegated or even completely white flowers. rhe lack of pigmentation was shown to result from coorfinate gene silencing of both the CHS transgene and the ~ndogenous CHS gene (Jorgensen 2003). PTGS is now :onsidered the plant equivalent of RNA interference

E PIG ENE TIC

REG U LA T ION

I N

P LAN T 5

•

175

(RNAi) later described in Caenorhabditis elegans and other organisms (see Section 3.2). By the mid-1990s, links between PTGS and virus resistance had been forged. PTGS was shown to naturally protect plants from unchecked replication of viruses, which can be both inducers and targets of PTGS. This principle was exploited in plants to experimentally down-regulate plant genes by constructing viral vectors containing sequences hom*ologous to a target gene, resulting in virusinduced gene silencing (VIGS; for review, see Burch-Smith et al. 2004). In addition, RNA-directed DNA methylation (RdDM) was discovered in viroid-infected plants, providing the first demonstration that RNA could feed back on DNA to elicit epigenetic modifications (Wassenegger et al. 1994). This principle has been successfully used to transcriptionally silence and methylate promoters by intentionally generating hom*ologous double-stranded RNA (see Section 3.4, RNA-directed DNA methylation). 2 Molecular Components of Chromatin in Plants

A number of molecular components of epigenetic regulation in plants were identified by the mutational approaches in Arabidopsis mentioned above (Table 2). However, mutant screens have probably not yet revealed a complete list of epigenetic modifiers because of either functional redundancy in large gene families or the lethal consequences of losing essential components. 2.1 Regulators of DNA Methylation in Plants

Methylation of carbon 5 of cytosines in DNA is a hallmark of epigenetic inactivation and heterochromatin in both plants and mammals (Table 1) (Chapter 18). In plants, however, DNA methylation has a number of unique features with respect to the pattern of methylation, proteins of the methylation machinery, and the possibility to reverse methylation in nondividing cells (for review, see Chan et al. 2005). In this section, we discuss the proteins required to establish, maintain, interpret, and erase DNA methylation. Special components needed for the process of RNA-directed DNA methylation are presented in Section 3.4. DNA

METHYLTRANSFERASES

DNA methylation can be divided into two steps: de novo methylation and maintenance methylation. De novo methylation refers to the modification of a previously unmethylated DNA sequence (Fig. 1) (Chapter 18). In plants, de novo methylation can alter CpG, CpNpG, and

176

C HAP T E R

CpNpN nucleotide groups (where N is A, T, or C). In contrast, methylation in mammals is largely restricted to CpG dinucleotides, and there is no evidence for extensive methylation in asymmetric CpNpN nucleotide groups. Although the signals that trigger de novo methylation are largely unknown, double-stranded RNA can fulfill this role in plants (Section 3.4). Maintenance methylation perpetuates methylation patterns during DNA replication and occurs most efficiently at CpG and CpNpG nucleotide groups with their palindromic symmetry. Maintenance of methylation occurs on a hemimethylated substrate after replication or repair, guided by the modification still present on the parental DNA strand. Although it is usually assumed that distinct DNA cytosine methyltransferase enzymes contribute to either de novo or maintenance methylation, an emerging view in plants is that enzymes with different site specificities (CpG or non-CpG) frequently cooperate to catalyze both steps. The three conserved families of DNA methyltransferase are all present in plants. Members of the methyltransferase (METl) family, which are hom*ologs of the mammalian Dnmtl type (see Chapter 18), are considered CpG maintenance methyltransferases, although one has also been assigned a role in CpG de novo methylation in the RdDM pathway (Section 3.4). The Dnmt2 class, of which one member is encoded in the Arabidopsis genome, comprises the most widespread and highly conserved DNA methyltransferase family (Table 2), but its function remains obscure. The plant Domainsrearranged methyltransferases (DRM) and their mammalian hom*ologs, the Dnmt3 group, are usually considered de novo methyltransferases. The DRM enzymes catalyze methylation of cytosines in all sequence contexts and are prominent in the RdDM pathway (Section 4.4). As their name implies, the DRM proteins have rearranged domains (VI-X, followed by I-V) compared to Dnmt3 (I-X). This might give them the ability to methylate asymmetric CpNpN nucleotide groups, which are not methylated in mammalian cells. The plant-specific chromomethylase CMT3 modifies CpNpG trinucleotides. Similarly to METl, CMT3 has been implicated in both de novo and maintenance methylation. The exact function of CMT3 is not entirely clear, although loss-of-function mutants reactivate certain silent transposons (for review, see Chan et al. 2005). In contrast to mammals, where dnmtl and dnmt3 mutants die during embryonic development or shortly after birth, metl, emt3, and drm mutants are viable and usually fertile. The nonlethality of DNA methyltransferase mutations in plants has permitted more extensive

analyses of deficiency mutants during development and sexual reproduction than is possible in mammals (for review, see Chan et al. 2005).

ACTIVE (pG DEMETHYLATION AND

DNA

GlYCOSYLASES

Epigenetic regulation implies that marks corresponding to active or inactive genetic states are potentially reversible. DNA methylation permits such reversibility, because it can be lost through passive or active means. Passive loss occurs when methylation fails to be maintained during multiple rounds of DNA replication. In contrast, active demethylation can occur in nondividing cells and requires enzymatic activities. Early reports from animal systems suggested that active demethylation can result from the action of DNA glycosylases, which are normally involved in base excision repair (for review, see Kress et al. 2001). Interest in this idea has been rekindled by the discovery in Arabidopsis of Demeter (DME) and Repressor of silencing (ROSl), which are large proteins containing DNA glycosylase domains. The ROSl gene was identified in a screen for epigenetic down-regulation and hypermethylation of a stably expressed reporter gene (Gong et al. 2002). The ROSI protein displays nicking activity on methylated but not unmethylated DNA, which is consistent with a role in removing methylated cytosines from DNA in a pathway related to base excision repair. ROSI is expressed constitutively and hence could potentially contribute to loss of DNA methylation in nondividing cells at all stages of development (Kapoor et al. 2005a). In contrast, DME activity is restricted to the female gametophyte, where it activates the imprinting factor Medea (MEA) in a manner that is dependent on a functional DNA glycosylase domain (Choi et al. 2002). The CG methyltransferase METl acts antagonistically to DME, suggesting that DME is indeed required for demethylation of CG dinucleotides (Hsieh and Fisher 2005). In Arabidopsis, there are two additional uncharacterized members of the DME/ROSl family that are unique to plants. The expansion of this gene family suggests that reversible gene silencing by active demethylation is important for plant physiology, development, or adaptation to the environment.

METHYl-DNA-BINDING PROTEINS

Methyl-CG-Binding Domain (MBD) proteins are thought to provide a means to transduce DNA methylation patterns into altered transcriptional activity. In mammals, MBD proteins bind methylated DNA and per-

EPIGENETIC

form various functions, such as recrmtmg histone deacetylases, to reinforce transcriptional silencing. Arabidopsis has 12 MBD-containing genes, compared to 11 in mammals, 5 in Drosophila, 2 in C. elegans, and none in sequenced fungal genomes (Hung and Shen 2003). Little is known about the functions of Arabidopsis MBD proteins, although RNAi-knockdown of one, AtMBDll, was associated with pleiotropic effects on development (Springer and Kaeppler 2005). None of the Arabidopsis MBD proteins has been identified in forward genetic screens, perhaps because of functional redundancy. In addition, despite the amino acid conservation of DNA methyltransferases among plants and mammals, the MBD-containing proteins in the two kingdoms diverge completely outside of the methyl-CG-binding domain. Thus, even though plants and mammals establish and maintain DNA methylation patterns using related enzymes, they might have evolved different ways of interpreting these patterns by means of distinct MBD proteins (Springer and Kaeppler 2005).

REGULATION

PLANTS.

177

histone hypo acetylation and CpG methylation, the latter of which can potentially be actively removed by DNA glycosylases (Section 2.1, Active CG demethylation and DNA glycosylases). Arabidopsis has 18 putative HDACs and 12 putative HATs (Pandey et al. 2002), which is around the same number found in mammals, but more than in other non-plant eukaryotes. The putative HDACs are generally conserved in all eukaryotes, but there is one plant-specific family, HD2, whose function remains obscure. Genetic screens have identified only two members of a conserved family: HDAl and HDA6 (Table 2). HDA6 has roles in maintaining CpG methylation induced by RNA and in repeated sequences, but contributes minimally to development, as indicated by the normal phenotype of deficiency mutants. In contrast, reduced expression of HDAI results in pleiotropic effects on development. None of the Arabidopsis HATs has been identified in forward genetic screens, which might reflect functional redundancy or the direction of most screens toward activation of silent genes.

COMPONENTS OF THE METHYL GROUP DONOR SYNTHESIS

Methylating enzymes require an activated methyl group, usually in the form of S-adenosyl-methionine. Therefore, it is surprising that the biochemical pathways providing this cofactor were not linked with epigenetic regulation earlier. Only recently, however, has a mutation (hogl) in the Arabidopsis gene encoding S-adenosyl-L-hom*ocysteine hydrolase been found to be responsible for epigenetic defects (Rocha et al. 2005). 2.2 Histone-modifying Enzymes

Like other organisms (Table 1), plants contain enzymes that posttranslationally modify the amino-terminal tails of histones, thus establishing a putative histone code (for review, see Loid! 2004). In plants, histone-modifying enzymes are often encoded by comparatively large gene families. Functional information about most family members is still limited. The two most common modifications are histone acetylation/deacetylation and histone methylation.

HISTONE DEACETYLASES AND HISTONE ACETYLTRANSFERASES

The opposing functions of histone acetyltransferases (HATs) and deacetylases (HDACs) ensure reversibility of this epigenetic mark. The potential for reversibility is reinforced by the frequent coexistence at silent genes of

HISTONE METHYLTRANSFERASES

Proteins that are able to methylate lysine residues in histones (referred to in this book as histone lysine methyltransferases or HKMTs) and other proteins contain a common SET domain (SU(VAR)/E(Z)/TRX). Through their ability to methylate histone H3 or H4 at various lysine residues, different complexes containing SET domain proteins play roles in promoting or inhibiting the transcription of specific genes and in forming heterochromatin. Some SET domain proteins are members of the Polycomb group (PcG) or trithorax group (trxG), which maintain transcriptionally repressed or active states, respectively, of homeotic genes during plant and animal development (see Chapters 11 and 12). Other SET domain proteins, such as SU(VAR)3-9, participate in maintaining condensed heterochromatin, often in repetitive regions, by methylating H3 at lysine 9 (H3K9). The Arabidopsis genome encodes 32 SET domain proteins, 30 of which are expressed. They can be grouped into four conserved families: E(Z), TRX, ASH1, and SU(VAR)3-9, as well as a small fifth family present only in yeast and plants (Baumbusch et al. 2001; Springer et al. 2003). The number of expressed SET domain proteins in Arabidopsis is relatively high compared to the 14 in Drosophila and 4 in fission yeast, although there are 50 SET domain proteins in mice. In addition to expansion of the SET domain protein family by polyploidy, retro-

178 •

C HAP T E R 9

transposition has also played a role in the amplification of SU(VAR)3-9 members in Arabidopsis. Outside of the SET domain, the plant and animal proteins are not always well conserved. The divergent regions are predicted to mediate protein-protein interactions, suggesting that plant SET domain proteins might act in complexes distinct from those in animals. Although incomplete, the functional information available for Arabidopsis SET domain proteins implicates them in chromatin regulation and epigenetic inheritance. The first two SET domain proteins to be identified in genetic screens were Curly leaf (CLF) and Medea/Fertilization independent seed formation (MEA/FISl), which are negative regulators related to Drosophila E(Z). In addition to being SET domain proteins, MEA, CLF, and E(Z) are also PcG proteins (Section 2.3, Other polycomb proteins). Mutations in eLF result in altered leaf morphology and homeotic changes in flower development. MEA/PIS 1 regulates gametophyte-specific gene expression and is an imprinting factor that inhibits endosperm development in the absence of fertilization (for review, see Schubert et al. 2005). In contrast, the TRX family member Arabidopsis trithoraxl (ATXl) acts as an activator of floral homeotic genes, presumably by means of its ability to catalyze histone H3 lysine 4 (H3K4) methylation, a mark often associated with transcriptionally active chromatin (for review, see Hsieh and Fischer 2005). Kryptonite/Suppressor of variegation 3-9 hom*olog 4 (KYP/SUVH4) was identified in screens for suppressors of epigenetic silencing at two endogenous genes (Jackson et al. 2002; Malagnac et al. 2002). KYP/SUVH4 catalyzes mono- and dimethylation of H3 at lysine 9 (H3K9me2/me3) and acts together with CMT3 to maintain CpNpG methylation of a subset of sequences in Arabidopsis. KYP/SUVH4 appears to play only a minor role in heterochromatin formation (Chan et al. 2005). In contrast, Suppressor of Variegation 3-9 hom*olog 2 (SUVH2), identified in a screen for reactivation of a silent transgene, appears to be the major activity responsible for methylation ofH3 at lysines 9 (H3K9) and 27 (H3K27) in heterochromatin in Arabidopsis (Naumann et al. 2005). Lysines in histones H3 and H4 can be mono-, di-, or trimethylated, which increases the combinatorial complexity of these modifications. Specific states define heterochromatin in different organisms. For example, H3K9me3 is a prominent feature of heterochromatin in animals and fungi, whereas this epigenetic mark is associated with euchromatin in Arabidopsis. Conversely, H3K9mel and H3K9me2 are the predominant marks for

silenced heterochromatin in Arabidopsis, whereas they are euchromatic modifications in mammals. The origins of these differences and how they relate to the postulated histone code remain to be determined. In addition, the intricate relationships between specific histone modifications and DNA methylation patterns in plants remain to be fully elucidated (Tariq and Paszkowski 2004). In contrast to histone acetylation, which can be dynamically regulated by the opposing activities of HDACs and HATs, histone methylation was thought until recently to be a more permanent epigenetic mark. Recent work in mammals, however, has identified a lysine demethylase, LSDl, that can remove H3K4mel and H3K4me2 but not H3K4me3 (see Chapter 10). Four putative LSD hom*ologs are encoded in the Arabidopsis genome, suggesting that at least some histone methylation is reversible in plants. 2.3 Other Chromatin Proteins OTHER POLYCOMB PROTEINS

PcG proteins were initially identified in Drosophila as factors required to maintain repression of homeotic genes (see Chapter 11). In animals, structurally disparate PcG proteins act together in multiprotein complexes to repress gene expression. The PRCl complex is absent in plants and C. elegans but present in Drosophila and mammals. The PRC2 complex, however, is found in plants and animals, where it has been shown to methylate predominantly H3 at lysine 27 (H3K27) through the histone methyltransferase activity of the SET domain and PcG protein E(Z). Arabidopsis hom*ologs of the core components of PRC2 have been identified in mutant screens designed to dissect various developmental pathways. In Drosophila, PRC2 components are encoded by single-copy genes. In contrast, genes encoding these proteins in Arabidopsis show functional diversification of at least three PRC2 complexes-PIS (fertilization independent seeds), EMF (embryonic flower), and VRN (vernalization)-that differ in their target gene specificity (Schubert et al. 2005; see also Fig. 2 in Chapter 11). PIS genes were identified in screens for mutants showing partial seed development in the absence of fertilization. A major target is the MADS-box transcription factor PHERES (Kohler et al. 2005). Components of the EMF complex were identified by their common role in repressing floral homeotic genes, such as Agamous and Apetala3. A member of the VRN complex, VRN2, was identified on the basis of its contribution to epigenetic memory of vernalization, which is defined as the break of seed dor-

E PIG ENE TIC

mancy by cold treatment. Plants have to program their reproduction to occur during the proper season, and they do this in temperate climates by flowering only after extended periods of cold temperatures. The epigenetic memory of winter requires VRN2, which maintains coldinduced transcriptional repression of the gene encoding the flowering inhibitor FLC during later periods of growth at warmer temperatures. H3K27me2 is lost from FIC in vrn2 mutants, which is consistent with a role for PRC2 complexes in facilitating histone methylation (Schubert et al. 2005).

COMPONENTS OF IMPRINTING

Flowering plants and mammals are the only groups of organisms that have parental imprinting (Table 1), an epigenetic phenomenon in which a gene is differentially expressed depending on the parent from which it was inherited. In view of the parental conflict theory for the evolution of imprinting (for further discussion, see Chapter 19), the occurrence of parental imprinting in flowering plants and mammals likely reflects the fact that both taxa have a special maternal tissue that provides a nutrient source for the developing embryo. In mammals, this tissue is the placenta, and in plants it is the triploid endosperm, a terminally differentiated tissue that contains one paternal and two maternal genomes (Fig. 1). Indeed, the first example of parental imprinting of a single gene in any organism was observed in maize endosperm (for review, see Alleman and Doctor 2000). In Arabidopsis, two genes expressed in the endosperm, MEA and FWA (a flowering time control gene), are imprinted. In these cases, the two maternal copies are activated, presumably by DME-catalyzed active demethylation of CpGs in the female gametophyte (see Section 2.1, Active CpG demethylation and DNA glycosylases), whereas the paternal copy remains silent (for review, see Autran et al. 2005). Intriguingly, even though imprinting evolved independently in plants and mammals, DNA methylation and PcG proteins are required in both cases (Kohler et al. 2005).

CHROMATIN-REMODELING PROTEINS

Switch2/Sucrose Non-Fermentable2 (SWI2/SNF2) chromatin-remodeling factors constitute a conserved family of ATP-dependent chromatin remodelers that are able to displace nucleosomes or loosen histone/DNA contacts. Genetic screens have provided functional information for only a handful of the approximately 40 SWI2/SNF2 hom*o logs encoded in the Arabidopsis

REG U L A T ION

I N

P LAN T S

•

1 79

genome (Plant Chromatin Database). So far, only twoDecreased DNA methylation 1 (DDMl, Jeddeloh et al. 1999) and Defective in RNA-directed DNA methylation 1 (DRD1, Kanno et al. 2004)-have been implicated in regulating DNA methylation. Deficiency mutants of DDMl, which undergo genome-wide reduction of DNA methylation and transcriptionally reactivate a number of silent transposons and repeats, display severe developmental and morphological defects. These appear only after several generations of inbreeding hom*ozygous ddml plants and appear to be due to the accumulation of epimutations and to insertional mutagenesis by transposons that are reactivated in the mutant. DDMl has an ortholog in mammals, Lymphoid-Specific Helicase (ISH), which is likewise important for global CpG methylation and embryonic development. In contrast, DRDl is unique to the plant kingdom and probably has a specialized role in RdDM (Section 3.4). No phenotypic alteration other than a release of certain repetitive targets from silencing is caused by mutations of Morpheus' Molecule (MOM, Amedeo et al. 2000), a plant-specific gene with an incomplete ATPdependent helicase motif. MOM acts synergistically with, but independently of, the DDMl/DNA methylation pathway, indicating multiple layers of transcriptional regulation in plants (Tariq and Paszkowski 2004). Three more proteins with putative chromatin-remodeling function, Splayed (SPD), Photoperiod-independent early flowering (PIE), and Pickle (PKL), which were each identified by developmental effects in deficiency mutants, have not yet been implicated in specific chromatin modifications (Wagner 2003).

CHROMATIN ASSEMBLY FACTORS

Whereas the SWI2/SNF2 proteins probably act on assembled chromatin, other components are required to reestablish chromatin after replication and repair-associated DNA synthesis. The Chromatin Assembly Factor (CAF) complex, composed of three subunits, helps to bring semi-assembled nucleosomes to the replication fork. Mutations in genes of the two larger CAF subunits in Arabidopsis (lasl, fas2) cause characteristic morphological anomalies (fasciation, Fig. 2F), deficiencies in DNA repair, and derepression of repetitive targets (Takeda et al. 2004). This suggests that correct nucleosome deposition is essential for development and epigenetic control. Whereas the lack of CAF subunits does not interfere with maintenance of DNA methylation, it could lead to the erasure of other epigenetic marks, such as histone modifications. Reduced

180 •

C HAP T E R 9

levels of the third CAF unit MSIl do not reiterate fasciation but lead to distorted seed development and several morphological changes (for review, see Hennig et al. 2005). A mutation in the BRU gene that is unrelated to any known chromatin assembly protein, but results in a phenotype very similar to that of the fas mutants, makes it likely that additional factors are involved in maintaining the epigenetic information and genetic integrity during postreplicative chromatin assembly (Takeda et al. 2004). Finally, lack of RPA2, a subunit of the Replication Protein A complex, results in DNA damage sensitivity and release of transcriptional silencing, changing histone modification marks but not DNA methylation patterns (Elmayan et al. 2005; Kapoor et al. 2005b). HETEROCHROMATIN-LIKE PROTEINS

HP1 (heterochromatin protein 1) in Drosophila and mammals, and their hom*ologs in fungi, are important components of silenced heterochromatin. The binding of HP1 through its chromodomain to methylated histone H3 at lysine 9 (H3K9me) promotes spreading of the silenced state to establish heterochromatic domains. The Arabidopsis genome encodes a single protein with hom*ology to Drosophila HPl. Mutations in this gene, termed Like heterochromatin protein (LHP 1) (Gaudin et al. 2001) or Terminal flower 2 (TFL2) (Kotake et al. 2003), result in changes in plant architecture, altered leaf development, and early onset of flowering. Although this mutant phenotype suggests an important role in regulating plant gene expression, it is unlikely that LHP1 acts through the formation of repressive chromatin complexes similarly to HP1 in other organisms. Instead, LHP 1 in Arabidopsis regulates loci in euchromatin that are not targets of DNA methylation (Kotake et al. 2003; Tariq and Paszkowski 2004). Thus, LHP1 in plants and HP1 in other organisms appear to have evolved different modes of action.

3 Molecular Components of RNAi-mediated Gene Silencing Pathways

Modern epigenetics research has traditionally focused on DNA methylation and histone modifications. During the past several years, it has become evident that these alterations can be targeted to specific regions of the genome by the RNA interference pathway. Indeed, it is impossible nowadays to consider epigenetic regulation in many eukaryotes without integrating components of the RNAi machinery (Matzke and Birchler 2005). This is particu-

larly true for plants, where the proliferation of RNAimediated gene-silencing pathways exceeds that present in any other type of organism.

3.1 Elaboration of RNAi-mediated 5i1encing in Plants

RNAi and related types of gene silencing represent cellular responses to double-stranded RNA (dsRNA). In these pathways, the dsRNA is processed by the RNase III-like endonuclease, Dicer, to produce small RNAs which determine the specificity of silencing by base-pairing to complementary target nucleic acids. Small RNAs incorporate into multiprotein silencing effector complexes to direct mRNA degradation, repress translation (PTGS), or guide chromatin modifications (TGS) in a sequence-specific manner. A key component of silencing effector complexes is an Argonaute protein, which binds small RNAs through its PAZ domain. Individual members of the Argonaute protein family, which comprises the largest group of proteins important for RNAi-mediated silencing, confer functional specificity to different silencing effector complexes (for review, see Carmell et al. 2002). In addition to participating in viral defense and transposon control, RNAi-mediated gene silencing plays essential roles in plant and animal development. The elaboration of RNAi-mediated silencing in plants reflects in part their co-evolution with pathogens that generate dsRNA during replication, such as RNA viruses and viroids. Indeed, together with transgenes-another type of "foreign" nucleic acid-these RNA pathogens have been invaluable for detecting and studying various forms of RNAi-mediated gene silencing in plants. The proliferation of RNAi-mediated gene-silencing pathways in plants is illustrated by 1. the expansion and functional diversification of gene families encoding core components of RNAi: the Arabidopsis genome encodes four DICER-LIKE (DCL) proteins and ten Argonaute (AGO) proteins 2. the heterogeneity in length and functional diversity of small RNAs, including the 21-nucleotide short interfering RNAs (siRNA) derived from transgenes and viruses, and several types of endogenous small RNAs, such as 21- to 24-nucleotide microRNAs; 21-nucleotide trans-acting siRNAs, and 24- to 26nucleotide heterochromatic siRNAs 3. the various modes of gene silencing elicited by different small RNAs: PTGS involves mRNA degradation or repression of translation, and TGS is

E PIG ENE T / eRE G U LA T ION

associated with epigenetic modifications such as DNA cytosine methylation and histone methylation 4. the importance of PTGS in antiviral defense, which can be countered by a variety of plant viral proteins that repress silencing at different steps of the pathway 5. the existence of processes, such as non-cellautonomous silencing and transitivity (see Section 3.2, Non-ceil-autonomous silencing and transitivity), that rely on RNA-dependent RNA polymerases, six of which are encoded in the Arabidopsis genome These aspects will be discussed in the framework of three major pathways of RNAi-mediated gene silencing in plants (Fig. 3a-c). However, it should be kept in mind

RNAi-mediated gene silencing induced by transgenes and viruses appears to function primarily as a host defense to foreign or invasive nucleic acids, including viruses, transposons, and transgenes. ORIGIN AND PROCESSING OF DsRNA

Transgene constructs can be introduced into plant genomes in sense or antisense orientations or as inverted DNA repeats. Viruses can have single-stranded

I miRNA gene I

ITAS gene I

==~

TGS/RdDM/heterochromatin

'--,-__--' Itarget gene/repeats I

POL IVa

pri_miRNAl SGS3 RDR6j

DCLl

SDE3

HYLl

WEX HENl

RD:2

...

DCL3

vRdRP

HEN1

pre-miRN.FO

== ==

DCLl

24-26 nl

DCL4

HENl

miR/miR· ....

SDE3

RDR6 amplification transitivity mobile signals

== ~==

21 nt

== I==

24-26 nt

~ 11 ~mRNA ~

mRNA cleavage

RdDM, mobile signals

181

3.2 Pathway 1: Transgene-related Posttranscriptional and Virus-induced Gene 5i1encing (PTG5N/G5)

microRNAItrans-acting siRNA

aRNA

P LAN T 5

that the pathways feed into each other at various points. Components with assigned functions are listed in Table 2.

b transgene PTGSIVGS

I N

miRNA 21 nl

_ _ lasiRNA

1-----'1 -

21nl

===:::::::::::_= ~~~31 ~~:~ DDMl

METl

RdDM

DNA

histone modifications

- _.......;,,;;;;iii. . . mRNA

/ mRNA cleavage or translational block

RdDM mRNA cleavage

Figure 3. RNA-mediated Silencing Pathways in Plants Although there are some overlaps and shared components, three major pathways can be distinguished by the source of dsRNA, class of small RNA, nature of the target sequence, and the mode of silencing evoked. Silencing effector complexes containing an Argonaute protein are shown as light gray spheres. Yellow boxes mark processes known to occur within the nucleus. See text for details and Table 2 for the names of regulatory components. Plant-specific proteins are labeled in green. (PTGS) Posttranscriptional gene silencing, (VIGS) virus-induced gene silencing, (TGS) transcriptional gene silencing, (RdDM) RNA-directed DNA methylation, (IR) inverted repeats, (AS) antisense, (vRdRP) virally encoded RNA-dependent RNA polymerase, (a RNA) aberrant RNA, (siRNA) short interfering RNA, (RISe) RNA-induced silencing complex. (Modified from Meins et al. 2005.)

182

•

C HAP T E R 9

or double-stranded DNA or RNA genomes. Therefore, in this pathway, dsRNA can be produced by a variety of routes. In principle, antisense transcripts can base-pair directly to target mRNAs to form dsRNA. Transcription through inverted DNA repeats can produce hairpin RNAs. RNA viruses, which encode their own RNAdependent RNA polymerase (vRdRP) and replicate via dsRNA intermediates, enter the pathway directly at the level of dsRNA. In contrast, sense transgenes and DNA viruses, such as geminiviruses, require the cellular RNAdependent RNA polymerase RDR6 for dsRNA synthesis as well as several other factors identified genetically (SDE3, SGS3, and WEX; Table 2). To render them substrates for RDR6, transcripts of sense transgenes and DNA viruses are presumed to be aberrant in some way; for example, by lacking a 5' cap or a polyadenylated tail (for review, see Meins et al. 2005). The DCL activity required to process dsRNA into siRNAs in the PTGS pathway has not yet been identified (DCLX). Tests of dell partial loss-of-function mutants indicated that DCLl is unlikely to be involved in this processing step. The plant-specific protein HENl adds a methyl group to the 3'-most nucleotide of small RNAs, thus protecting them from uridylation and subsequent degradation (Li et al. 2005). DCL2 has been implicated in generating siRNAs from some, but not all, RNA viruses (Xie et al. 2004). PTGS and VIGS result in the production of two distinct size classes of siRNA, 21-22 nucleotides and 24-26 nucleotides, that have been implicated in diverse functions (Baulcombe 2004). In general, the 2l-nucleotide siRNAs are thought to guide mRNA cleavage, whereas the 24- to 26-nucleotide size class, termed heterochromatic siRNA, directs epigenetic modifications to hom*ologous DNA sequences (i.e., TGS; see Section 3.4). Following DCL processing, the siRNA duplex is unwound and the antisense strand associates with a member of the Argonaute protein family, as part of the assembly into the RNA-induced silencing complex (RISC). The siRNA-programmed RISC can then direct endonucleolytic cleavage of target mRNAs at a single site near the center of siRNA-mRNA complementarity. For the mammalian equivalent, cleavage is catalyzed by the Ago2 "slicer" activity (see Chapter 8). The Arabidopsis protein carrying out this function in the transgene PTGS pathway is AGOl (Baumberger and Baulcombe 2005). Following endonucleolytic cleavage, the severed 3' segment of the mRNA is degraded in the 5' to 3' direction by the exonuclease AtXRN4 (Souret et al. 2004); the 5' portion is probably degraded by the exosome in a 3' to 5' direction.

NON-CELL-AUTONOMOUS SILENCING AND TRANSITIVITY

PTGS in plants has two special properties that rely on the activity of the RNA-dependent RNA polymerase RDR6: non-ceIl-autonomous silencing and transitivity (Fig. 3a). In the former, RNA signals that induce PTGS move from the cell of origin into neighboring cells through plasmodesmata or-as originally shown in grafting experiments-through the vascular system to induce sequence-specific gene silencing at distant sites (for review, see Voinnet 2005). Mobile small RNAs, providing a systemic silencing signal, thus might play the dual function of influencing plant development by facilitating communication between cells, and coordinating activities in remote parts of the plant. This proposal is supported by the finding of microRNAs (miRNAs; important for development, Section 3.3) and a small RNA-binding protein in phloem sap, which is the main transporter of metabolites through the plant vascular system (Yoo et al. 2004). Transitivity refers to the generation of secondary siRNAs corresponding to sequences located outside the primarily targeted regions. To make these, RNAdependent RNA polymerase catalyzes synthesis of secondary dsRNAs from transgene or viral template RNAs using primary siRNAs as primers. Dicer processing yields secondary siRNAs, which amplify the silencing reaction and, when viral RNAs are involved, strengthen virus resistance (Voinnet 2005). The only other organism in which both non-cellautonomous silencing and transitivity have been observed is C. elegans (see Chapter 8), which has putative RNA-dependent RNA polymerase activities that are absent in mammals and Drosophila. VIRAL SUPPRESSORS OF SILENCING

Plant viruses are not only inducers and targets of silencing; they also encode proteins that can suppress silencing (for review, see Voinnet 2005). This reinforces the idea that PTGS is a natural defense to viruses, since these suppressor proteins constitute a counter-defense "strategy" of the pathogen. Most plant viruses encode at least one silencing suppressor protein that acts at a distinct step of the PTGS pathway, typically downstream of dsRNA processing. Suppression of PTGS by a virus is strikingly revealed in mottled soybeans, where the dark color is the result of reversal of natural PTGS (i.e., reactivation) of a pigment gene (Fig. 2e) (Senda et al. 2004). Viral suppressors of RNAi have recently also been found in an insect virus and a mammalian retrovirus (Lecellier et al. 2005; Voinnet 2005).

E PIG ENE TIC

3.3 Pathway 2: Regulation of Plant Development by miRNAs and Trans-acting siRNAs

The discovery of endogenous populations of miRNAs in plants and animals opened a new era in research of developmental biology and RNAi-mediated gene silencing (for review, see Bartel 2004). miRNAs silence gene expression by base-pairing to target messenger RNAs (mRNAs) and inducing either mRNA cleavage or translation repression. The importance of miRNAs in plant development is illustrated by the fact that many genes needed for miRNA biogenesis and silencing-including DeLl, AG01, HEN1, HYLl, and HST-were identified in screens for developmental mutants and only later shown to be important for miRNA accumulation. The phenotypes of mutants defective in these proteins suggest diverse roles for miRNAs in meristem function, organ polarity, vascular development, floral patterning, and stress/hormone responses (for review, see Kidner and Martienssen 2005). miRNAs have recently been implicated in the biogenesis of a new type of small RNA, the trans-acting siRNAs. ROLES AND BIOGENESIS OF MIRNAs

miRNAs were initially recovered by cloning size-fractionated small RNAs ranging from about 18 to 28 nucleotides in length. Their high degree of complementarity to target mRNAs in plants facilitated identification of additional miRNAs by computational approaches. So far, 92 loci in Arabidopsis that encode 27 distinct miRNAs have been discovered, and there are a similar number in rice. The expression of many miRNA genes is developmentally or environmentally regulated. About 50% of their known targets in Arabidopsis are transcription factors, many of which were known modulators of meristem formation and identity, prior to their identification as miRNA targets. In contrast, animal miRNAs do not preferentially target transcription factors but regulate diverse genes that operate at many levels in the cell. Two essential proteins of the miRNA pathway in Arabidopsis, DCLl and AGO 1, are themselves regulated by miRNAs, providing a means for negative modulation by feedback control (Kidner and Martienssen 2005). Many miRNAs are evolutionarily conserved among eukaryotes (Axtell and Bartel 2005), in some cases over extended periods of time. Remarkably, in flowering plants, gymnosperms, and more primitive plants, mRNAs of a group of transcription factors that regulate meristem formation and lateral organ asymmetry have maintained perfect complementarity to the cognate miRNA. This

REG U L A T ION

I N

P LAN T S

•

183

indicates conservation of function for at least 400 million years (Floyd and Bowman 2004). miRNAs are encoded in regions between protein-coding genes or in introns. They originate from imperfect RNA hairpin precursors, ranging from 70 bp to more than 300 bp in length, that are transcribed by DNA-dependent RNA polymerase II. Processing of plant miRNA precursors occurs in multiple steps in the nucleus. First, the ends of the pri-miRNA are removed by nuclear DCLI. This step requires the dsRNA-binding protein HYLl, originally identified by the hormone response defects of its mutant phenotype (Han et al. 2004; Vasquez et al. 2004a). The second step involves release of the miRNA duplex (miRlmiR*, Fig. 3b), again by DCLl, and 3'-end methylation by HENI (see Section 3.2, Origin and processing of dsRNA). Transport of the miR/miR* duplex from the nucleus to the cytoplasm requires HASTY (HST), a hom*olog of mammalian Exportin 5 (Park et al. 2005). Mature miRNAs are also found in nuclear fractions, suggesting that some may function in the nucleus to direct epigenetic modifications. Indeed, a miRNA that is complementary to the spliced, nascent transcript of a transcription factor induces cytosine methylation of DNA sequences downstream of the target gene, by an unknown mechanism (Schubert et al. 2005). miRNA biogenesis differs somewhat in mammals, which have a single Dicer that is located in the cytoplasm and a second RNase III-type activity, Drosha, in the nucleus. Drosha, together with the dsRNA-binding protein Pasha-neither of which has a hom*olog in plants-cleaves the ends of the pri-miRNA. The resulting pre-miRNA is then transported to the cytoplasm by an Exportin5-mediated pathway to undergo final processing to mature miRNAs by Dicer (Du and Zamore 2005; Kim 2005).

PLANT MIRNAs GUIDE MRNA CLEAVAGE

In general, animal miRNAs show imperfect complementarity to target mRNAs and repress translation by binding to multiple sites in 3'UTRs. In contrast, the nearly perfect complementarity of plant miRNAs to the coding regions of target mRNAs favors mRNA cleavage, presumably in a manner similar to siRNAs. However, there are increasing exceptions to both of these "rules." For example, plant miRI72 is able to block translation, and certain mammalian miRNAs may direct cleavage of target mRNAs (for review, see Du and Zamore 2005). AGOI is the founding member of the Argonaute family of proteins and the mRNA "slicer" component of

184 •

C HAP T E R 9

miRNA-programmed RISC in Arabidopsis (Baumberger and Baulcombe 2005). AG01 was identified prior to the discovery of miRNAs in a screen for Arabidopsis mutants defective in leaf development (for review, see Carmell et al. 2002). The name Argonaute was inspired by the phenotype of agol mutants, which resemble a small squid because of their narrow, filamentous leaves. Agol mutants display shoot apical meristem defects similar to mutants deficient in PNH/ZLL/AG010 (Table 2), which is similar to AGO 1 but not yet shown to be needed for PTGS (Vaucheret et al. 2004). The essential function of AGO proteins in plant meristems is consistent with a conserved function of these proteins in stem cell maintenance (Carmell et al. 2002; Kidner and Martienssen 2005).

3.4 Pathway 3: Transgene-related Transcriptional Silencing, RNA-directed DNA Methylation, and Heterochromatin Formation

Current concepts of RNAi-mediated transcriptional gene silencing grew out of early plant work on hom*ologydependent gene silencing triggered by multiple copies of promoter regions and on RNA-directed DNA methylation (for review, see Matzke and Matzke 2004). More recent studies on RNAi-mediated heterochromatin formation in fission yeast (see Chapters 6 and 8) and on siRNA-mediated TGS in mammalian cells have expanded the phylogenetic scope of this process and confirmed mechanistic overlaps to RNAi. RNA-DIRECTED DNA METHYLATION

TRANS-ACTING SIRNAs

Endogenous trans-acting siRNAs (ta-siRNAs) are a new type of small RNA that have been discovered recently in Arabidopsis. The ta-siRNAs, which elicit cleavage of their target mRNAs, share features with both siRNAs and miRNAs. Similarly to siRNAs, the synthesis of the dsRNA precursor of ta-siRNAs depends on RDR6 and SGS3. Similarly to miRNAs, ta-siRNAs originate from genomic regions that have little overall resemblance to their target mRNA. To ensure formation of the correct ta-siRNA with complementarity to the target mRNA, a miRNA sets the phased cleavage of the dsRNA precursor by DCL4 (Fig. 3b) (Allen et al. 2005; Gasciolli et al. 2005). ta-siRNAs have been assigned a role in developmental timing. During development, the shoot of flowering plants undergoes two phase changes: the vegetative phase change, comprising the juvenile-to-adult transition, and the reproductive phase change, which results in growth of flower-containing branches instead of vegetative shoots (see Fig. 1). Although genetic analysis of floral induction is well advanced, less is known about the vegetative phase change. An Argonaute protein, ZIPPY/AGO?, however, has a specialized role in this transition. A screen for mutants undergoing precocious vegetative phase change similar to zip/ago7 mutants identified RDR6 and SGS3, two genes important for PTGS (Fig. 3b) (Peregrine et al. 2004). Further analysis showed that several genes that are up-regulated in rdr6 and sgs3 mutants are silenced posttranscriptionally by ta-siRNAs (Vazquez et al. 2004b). These findings imply that components of the PTGS machinery are important not only for viral defense and transgene silencing, but also for temporal control of developmental switches. It is not yet known whether ta-siRNAs have counterparts in animals.

RdDM was first observed in tobacco plants infected with viroids (Wassenegger et al. 1994). Viroids are minute plant pathogens consisting solely of a non-protein-coding, circular RNA several hundred bases in length. In the original experiments, replicating viroids were found to trigger de novo methylation of viroid cDNAs integrated as transgenes into the tobacco genome. Transgene systems were subsequently used to establish that RdDM requires a dsRNA that is processed to small RNAs, a hallmark of RNAi. RNA viruses that replicate exclusively in the cytoplasm were shown to elicit methylation of hom*ologous nuclear DNA, indicating that small RNAs produced in the cytoplasm as a consequence of PTGS are able to enter the nucleus and induce epigenetic changes. dsRNAs containing promoter sequences can direct DNA methylation and transcriptional silencing of cognate target promoters (for review, see Mathieu and Bender 2004; Matzke et al. 2005). In plants, RNA induces a distinctive pattern of de novo methylation that is typified by the modification of cytosines in all sequence contexts, largely within the region of RNA-DNA sequence identity. This characteristic pattern hints that RNA-DNA base-pairing provides a substrate for de novo methylation, but this remains to be experimentally verified. Whereas asymmetrical CpNpN methylation is not efficiently retained after withdrawing the trigger RNA, symmetrical CpG and CpNpG methylation can be maintained to varying extents without RNA at different promoters. Differences in the efficiency of maintenance methylation might reflect differences in sequence composition or patterns of histone modifications. Combined data from genetic screens using transgene and endogenous gene systems are revealing the molecular components needed for RdDM and TGS. Transgene sys-

E PIG ENE TIC

terns rely on transcribed inverted repeats or viruses to produce dsRNA that is hom*ologous to target DNA loci. Endogenous genes that have been informative in forward genetic screens include the phosphoribosyl anthranilate isomerase (PA!) gene family (Mathieu and Bender 2004) and the SUPERMAN gene (Chan et al. 2005). These genes have features that render them targets or inducers of RdDM and TGS. For example, the PAl gene family contains four members, two of which are arranged as an inverted repeat. Transcription through the inverted repeats from an unrelated upstream promoter produces a dsRNA that targets the singlet copies of the PAl gene for methylation and silencing. PLANT-SPECIFIC MACHINERY FOR

RDDM

For the most part, conserved DNA methyltransferases and histone-modifying enzymes are required for RdDM (Sections 2.1 and 2.2). De novo methylation of cytosines in all sequence contexts is catalyzed by the conserved DRM class of DNA methyltransferase. The conserved METl and plant-specific CMT3 function primarily to maintain methylation of CpG and CpNpG nucleotide groups, respectively, although minor contributions to de novo methylation have been reported. The conserved histone deacetylase HDA6 and the SWI2/SNF2 protein DDMI help to maintain CpG methylation at some loci. The histone methyltransferase KYP/SUVH4 is involved in locus-specific maintenance of CpNpG methylation induced by RNA (for review, see Chan et al. 2005). A recent, surprising finding is that RdDM requires a plant-specific RNA polymerase, termed pol IV. In all eukaryotes examined so far, there are three DNAdependent RNA polymerases-pol I, pol II, and pol IIIthat contain multiple subunits encoded by distinct genes. The first hint of the existence of pol IV came from analyzing the Arabidopsis genome sequence, which revealed genes encoding the largest and second-largest subunits of an atypical RNA polymerase unique to plants. There appear to be two functionally diversified pol IV complexes that are specified by unique largest subunits that each act with a common second-largest subunit. pol IVa is needed to generate siRNAs, presumably by initially transcribing target genes (Herr et al. 2005; Onodera et al. 2005). The initial transcript is used as a template by RDR2 to synthesize dsRNA, which is processed by DCL3, a nuclear activity that is specialized for producing 24nucleotide heterochromatic siRNAs from transposons and repeats (Xie et al. 2004). Downstream of siRNA production, pol IVb acts together with the plant-specific

REG U L A T ION

I N

P LAN T 5

•

185

SWI2/SNF2-like protein DRDI to signal DNA methylation (Kanno et al. 2005), probably in cooperation with AG04 (Fig. 3c) (Chan et al. 2005). Whether pol IVb actually transcribes RNA is not known, but its net effect is to create a chromatin structure that permits DNA methyltransferases to catalyze de novo cytosine methylation at the siRNA-targeted site. Even though other eukaryotes do not contain pol IV subunits, two subunits of pol II, which transcribes mRNA precursors, are required for RNAi-mediated heterochromatin formation in fission yeast (see Chapters 6 and 8). Although promoter-directed siRNAs can induce TGS in human cells, there are conflicting reports about whether this is accompanied by detectable DNA methylation (Kawasaki et al. 2005; Ting et al. 2005). Many proteins required for RdDM in plants are found only in that kingdom (Fig. 3c). Thus, if RdDM occurs regularly in mammals, the mechanism or protein machinery differs from those in plants. SILENCING OF ENDOGENOUS GENES BY RNAI-MEDIATED

TGS

Many transposons and repetitive DNA sequences, such as the 5S rDNA arrays and the transposon-rich heterochromatic knob on Arabidopsis chromosome 4, are transcriptionally silenced and methylated by an RNAi-mediated mechanism (Lippman and Martienssen 2004; Chan et al. 2005). The endogenous DNA targets reflect the natural roles of RNAi-mediated TGS in repressing transposition and in packaging repeats into heterochromatin. However, plant genes containing transposon insertions can themselves become targets of RNAi-mediated silencing and methylation. For example, transposon-derived repeats in the promoter of the Arabidopsis floral gene FWA are targeted for methylation by cognate siRNAs (Lippman and Martienssen 2004), thus silencing the gene in vegetative tissues where it is not required. In some Arabidopsis accessions, a Mu element in an intron of the FIC gene, a repressor of flowering, renders the gene susceptible to repressive chromatin modifications directed by siRNAs originating from dispersed copies of Mu (Liu et al. 2004). The resulting lowered expression of FIC can accelerate flowering time, which might have adaptive significance in certain environments. Since many plant genes have transposon insertions in the vicinity of promoters or in introns, this mode of regulation might be common in the plant kingdom. Indeed, as more is learned about epigenetic regulation, McClintock's idea of transposons acting as elements that control host genes and development is gaining increasing support (McClintock 1956).

186 •

C HAP T E R 9

4 Epigenetic Regulation without RNA Involvement

Despite the specificity provided by small RNAs, they probably do not induce all epigenetic modifications in plants. For example, MOM, a protein with a partial SNF2 domain, has not yet been implicated in RNAi-mediated TGS. There is also no evidence that PcG proteins in plants are directed to their target genes by small RNAs. Other types of signal, such as hom*ologous pairing of non-transcribed repetitive sequences or special sequence compositions, might nucleate heterochromatin formation or attract DNA methyltransferases. The RNAi machinery, for instance, is dispensable for DNA methylation and histone methylation in Neurospora, where TArich segments are preferentially targeted for modification (see Chapter 6). Moreover, there are pathways for heterochromatin formation in fission yeast that are independent of RNAi (see Chapters 6 and 8). An unusual epigenetic phenomenon in plants that has not yet been shown to involve RNAi is paramutation. Paramutation occurs when certain alleles, termed paramutagenic, impose an epigenetic imprint on susceptible (paramutable) alleles. The epigenetic imprint is inherited through meiosis and persists even after the two interacting alleles segregate in progeny. Paramutation represents a violation of Mendel's law, which stipulates that alleles segregate unchanged from a heterozygote. Paramutation was first observed decades ago in maize and tomato, but the mechanism(s) has remained enigmatic. Paramutation-like phenomena have been observed recently in mammals, suggesting that it is not limited to the plant kingdom (for review, see Chandler and Starn 2004). The B locus in maize, one of the most intensively studied cases of paramutation, contains a series of direct repeats almost 100 kb from the transcription start site that mediate paramutation in an unknown manner. Although RNA-based silencing has not been fully ruled out, alternate mechanisms relying on pairing of alleles are still under consideration (for review, see Starn and Mittelsten Scheid 2005). 5 Outlook

In this chapter, we have discussed what is known about basic epigenetic principles in plants and their relationship to epigenetic regulation in other organisms. Plants clearly share a number of features of epigenetic control with other organisms, yet they have also evolved a number of plant-specific variations and innovations. These likely underpin the unique aspects of plant development

and their extraordinary ability to survive and reproduce successfully in unpredictable environments. Prominent among the plant-specific innovations is a built-in system for reversible epigenetic modifications, which likely makes a key contribution to plant developmental plasticity and adaptability. The capacity to induce or erase repressive modifications in nondividing cells-the former through RdDM and histone modifications, and the latter through the activity of DNA glycosylases such as DME and ROS1-allows epigenetic reprogramming without intervening cycles of DNA replication. The facile erasure of epigenetic marks from plant genomes probably accounts for the relative ease of cloning whole plants from single somatic cells. Nevertheless, induction as well as removal of epigenetic marks is likely neither perfect nor uniform throughout an individual, which creates epigenetic variability in populations of supposedly genetically identical cloned plants. Such somaclonal variation can be exploited in plant breeding programs. Similarly, the differential inheritance of epigenetic marks during sexual reproduction can lead to epigenetic variation in natural populations. Selection can act on this variability by fixing specific epialleles that might have adaptive significance. As we have described for the process of vernalization, environmental cues can trigger epigenetic modifications in plants and alter physiological responses. Thus, plants can "learn" if environmentally or stress-induced epigenetic modifications in shoot meristem cells enter the germ line and are adaptive. Defining the full range of conditions under which epigenetic changes are likely to occur spontaneously or are programmed will reveal more about the biological functions of these modifications. Likewise, unraveling the mechanisms of meiotic inheritance of epigenetic marks in plants could eventually permit scientists to manipulate this feature for improvements in horticulture and agriculture. In addition to responding appropriately to environmental stimuli, plants have confronted a variety of genomic challenges during their evolutionary and breeding histories. Polyploidization or hybridization can have a significant impact on epigenetic modifications owing to the still ill-defined process of genome shock, a response to an unusual stress leading to widespread mobilization of transposons (McClintock 1983). The origin of heterosis, the superior performance of hybrids compared to that of inbred parent lines, is still unknown, but it is likely to involve epigenetic alterations triggered by combining two related but distinct

E PIG ENE TIC

genomes. Similarly, polyploidization combines and/or multiplies whole genomes, with innumerable possibilities for epigenetic changes. Learning the epigenetic consequences of polyploidization in plants would also help to understand our own evolutionary history, which is increasingly thought to involve two whole-genome duplications (Furlong and Holland 2004). Clearly, even at this scale of inquiry, plant epigenetics can be informative for human biology, justifying their reputation as "masters of epigenetic regulation." References Adams K.L. and Wendel J.E 2005. Polyploidy and genome evolution in plants. Curro Opin. Plant BioI. 8: 135-141. Alleman M. and Doctor J. 2000. Genomic imprinting in plants: Observations and evolutionary implications. Plant Mol. BioI. 43: 147-161. Allen E., Xie Z., Gustafson A.M., and Carrington J.c. 2005. microRNAdirected phasing during trans-acting siRNA biogenesis in plants. Cell 121: 207-221. Amedeo P., Habu Y, Afsar K., Mittelsten Scheid 0., and Paszkowski J. 2000. Disruption of the plant gene MOM releases transcriptional silencing of methylated genes. Nature 405: 203-206. Autran D., Huanca-Mamani W., and Vielle-Calzada J.P. 2005. Genomic imprinting in plants: The epigenetic version of an Oedipus complex. Curro Opin. Plant BioI. 8: 19-25. Axtell M.J. and Bartel D.P. 2005. Antiquity of microRNAs and their targets in land plants. Plant Cell 17: 1658-1673. Bartel D.P. 2004. MicroRNAs: Genomics, biogenesis, mechanism, and function. Cell 116: 281-297. Baulcombe D.C. 2004. RNA silencing in plants. Nature 431: 356-363. Baumberger N. and Baulcombe D.C. 2005. Arabidopsis ARGONAUTEI is an RNA slicer that selectively recruits microRNAs and short interfering RNAs. Proc. Natl. Acad. Sci. 102: 11928-11933. Baumbusch L.O., Thorstensen 1., Krauss V, Fischer A., Naumann K., Assalkhou R., Schulz 1., Reuter G., and Aalen R.B. 2001. The Arabidopsis thaliana genome contains at least 29 active genes encoding SET domain proteins that can be assigned to four evolutionarily conserved classes. Nucleic Acids Res. 29: 4319-4333. Burch-Smith 1.M., Anderson J.c., Martin G.B., and Dinesh-Kumar S.P. 2004. Applications and advantages of virus-induced gene silencing for gene function studies in plants. Plant f. 39: 734-746. Carmell M.A., Xuan Z., Zhang M.Q., and Hannon G.J. 2002. The Argonaute family: Tentacles that reach into RNAi, developmental control, stem cell maintenance, and tumorigenesis. Genes Dev. 16: 2733-2742. Chan S.W.-L., Henderson l.R., and Jacobsen S.E. 2005. Gardening the genome: DNA methylation in Arabidopsis thaliana. Nat. Rev. Genet. 6: 351-360. Chandler VL., and Stam M. 2004. Chromatin conversations: Mechanisms and implications of paramutation. Nat. Rev. Genet. 5: 532-544. Chandler V.L" Eggleston W.B., and Dorweiler J.E. 2000. Paramutation in maize. Plant Mol. BioI. 43: 121-145. Choi Y, Gehring M., Johnson L., Hannon M., Harada J.J., Goldberg R.B., Jacobsen S.E., and Fischer R.L. 2002. DEMETER, a DNA glycosylase domain protein, is required for endosperm gene imprinting and seed viability in Arabidopsis. Cell 110: 33-42.

REG U L A T ION

I N

P LAN T S

•

187

Cubas P., Vincent c., and Coen E. 1999. An epigenetic mutation responsible for natural variation in floral symmetry. Nature 401: 157-161. Du 1. and Zamore P.D. 2005. microRNAPrimer: The biogenesis and function of microRNA. Development 132: 4645-4652. Elmayan 1., Proux E, and Vaucheret H. 2005. Arabidopsis RPA2: A genetic link among transcriptional gene silencing, DNA repair, and DNA replication. Curro BioI. 15: 1919-1925. FedoroffN.V and Chandler V 1994. Inactivation of maize transposable elements. In hom*ologous recombination and gene silencing in plants (ed J. Paszkowski), pp. 349-385. KJuwer Academic Publishers, Dordrecht, The Netherlands. Floyd S.K. and Bowman J.L. 2004. Gene regulation: Ancient microRNA target sequences in plants. Nature 428: 485-486. Furlong R.E and Holland P.W. 2004. Polyploidy in vertebrate ancestry: Ohno and beyond. BioI. ]. Linnean Soc. 82: 425-430. Gasciolli V, Mallory A.C., Bartel D.P., and Vaucheret H. 2005. Partially redundant functions of Arabidopsis DICER-like enzymes and a role for DCL4 in producing trans-acting siRNAs. Curro BioI. 15: 1494-1500. Gaudin V, Libault M., Pouteau S., Juul T., Zhao G., Lefebvre D., and Grandjean O. 2001. Mutations in LIKE HETEROCHROMATIN PROTEIN 1 affect flowering time and plant architecture in Arabidopsis. Development 128: 4847-4858. Gong Z., Morales-Ruiz 1., Ariza R.R., Roldan-Arjona 1., David L., and Zhu J.K. 2002. ROSl, a repressor of transcriptional gene silencing in Arabidopsis, encodes a DNA glycosylase/lyase. Cell Ill: 803-814. Han M.H., Goud S., Song L., and Fedoroff N. 2004. The Arabidopsis double-stranded RNA-binding protein HYLl plays a role in microRNA-mediated gene regulation. Proc. Natl. Acad. Sci. 101: 1093-1098. Heitz E. 1928. Das Heterochromatin der Moose. 1. ]ahrb. Wiss. Bot. 69: 762-818. Hennig L., Bouveret R., and Gruissem W. 2005. MSIl-like proteins: An escort service for chromatin assembly and remodeling complexes. Trends Cell BioI. 15: 295-302. Herr A.J., Jensen M.B., Dalmay 1., and Baulcombe D.C. 2005. RNA polymerase IV directs silencing of endogenous DNA. Science 308: 118-120. Hsieh 1.-E and Fischer R.L. 2005. Biology of chromatin dynamics. Annu. Rev. Plant BioI. 56: 327-351. Hung M.-S. and Shen C.-K. 2003. Eukaryotic methyl-CpG-binding domain proteins and chromatin modification. Eukaryot. Cell 2: 841-846. Jackson J.P., Lindroth A.M., Cao X., and Jacobsen S.E. 2002. Control of CpNpG DNA methylation by the KRYPTONITE histone H3 methyltransferase. Nature 416: 556-556. Jeddeloh J.A., Stokes 1.L., and Richards E.]. 1999. Maintenance of genomic methylation requires a SWI2/SNFs-like protein. Nat. Genet. 22: 94-97. Jorgensen R.A. 2003. Sense cosuppression in plants: Past, present, and future. In RNAi: A guide to gene silencing (ed. G.]. Hannon), pp. 5-22. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Kaeppler S.M., Kaeppler H.E, and Rhee Y 2000. Epigenetic aspects of somacionaI variation in plants. Plant Mol. BioI. 43: 59-68. Kanno 1., Mette M.E, Kreil D.P., Aufsatz W., Matzke M., and Matzke A.J. 2004. Involvement of putative SNF2 chromatin remodeling protein DRD1 in RNA-directed DNA methylation. Curro BioI. 14: 801-805. Kanno 1., Huettel B., Mette M.E, Aufsatz W., ]aligot E., Daxinger L., Kreil D.P., Matzke M., and Matzke A.J.M. 2005. Atypical RNA poly-

188 •

C HAP T E R

me rase subunits required for RNA-directed DNA methylation. Nat. Genet. 37: 761-765. Kapoor A., Agius E, and Zhu J.K. 2005a. Preventing transcriptional gene silencing by active DNA demethylation. FEBS Lett. 579: 5889-5898. Kapoor A., Agarwal M., Andreucci A., Zheng X., Gong Z., Hasegawa P.M., Bressan R.A., and Zhu J.K. 2005b. Mutations in a conserved replication protein suppress transcriptional gene silencing in a DNA-methylation-independent manner in Arabidopsis. Curro BioI. 15: 1912-1918. Kawasaki H., Taira K., and Morris K.V. 2005. siRNA induced transcriptional gene silencing in mammalian cells. Cell Cycle 4: 442-448. Kidner C.A. and Martienssen R.A. 2005. The developmental role of microRNA in plants. Curro Opin. Plant Bioi. 8: 1-7. Kim V.N. 2005. MicroRNA biogenesis: Coordinated cropping and dicing. Nat. Rev. Genet. 6: 376-385. Kohler c., Page D.R., Gagliardini v., and Grossniklaus U. 2005. The Arabidopsis thaliana MEDEA polycomb group protein controls expression of PHERESI by parental imprinting. Nat. Genet. 37: 28-30. Kotake T., Takada S., Nakahigashi K., Ohto M., and Goto K. 2003. Arabidopsis TERMINAL FLOWER 2 gene encodes a heterochromatin protein 1 hom*olog and represses both FLOWERING LOCUS T to regulate flowering time and several floral homeotic genes. Plant Cell Physiol. 44: 555-564. Kress c., Thomassin H., and Grange T. 2001. Local DNA demethylation in vertebrates: How could it be performed and targeted? FEBS Lett. 494: 135-140. Lecellier C.-H., Dunoyer P., Arar K., Lehmann-Che J., Eyquem S., Himber c., Saib A., and Voinnet O. 2005. A cellular microRNA mediates antiviral defense in human cells. Science 308: 557-560. Li J., Yang Z., Yu B., Liu J., and Chen X. 2005. Methylation protects miRNAs and siRNAs from 3' -end uridylation activity in Arabidopsis. Curro BioI. 15: 1501-1507. Lippman Z. and Martienssen R. 2004. The role of RNA interference in heterochromatic silencing. Nature 431: 364-370. Liu J., He Y., Amasino R., and Chen X. 2004. siRNAs targeting an intronic transposon in the regulation of natural flowering behavior in Arabidopsis. Genes Dev. 18: 2873-2878. Loidl P. 2004. A plant dialect of the histone language. Trends Plant Sci. 9: 84-90. Malagnac E, Bartee L., and Bender J. 2002. An Arabidopsis SET domain protein required for maintenance but not establishment of DNA methylation. EMBO f. 21: 6842-6852. Mathieu O. and Bender J. 2004. RNA-directed DNA methylation. f. Cell Sci. 117: 4881-4888. Matzke M.A. and Birchler J.A. 2005. RNAi-mediated pathways in the nucleus. Nat. Rev. Genet. 6: 24-35. Matzke M. and Matzke A.J.M. 2004. Planting the seeds of a new paradigm. PLoS BioI. 2: 528-586. Matzke M., Kanno T., Huettel B., Jaligot E., Mette M.E, Kreil D.P., Daxinger L., Rovina P., Aufsatz W., and Matzke A.J.M. 2005. RNA-directed DNA methylation. In Plant epigenetics (ed. P. Meyer), pp. 69-96. Blackwell Publishing, Oxford, United Kingdom. McClintock B. 1956. Intranuclear systems controlling gene action and mutation. Brookhaven Symp. BioI. 8: 58-74. - - - . 1983. The significance of responses of the genome to challenge. Nobel lecture. http://nobelprize.org/medicine/laureates/ 1983/mcclintock-lecture.pdf Meins E, Si-Ammour A., and Blevins T. 2005. RNA silencing systems

and their relevance to plant development. Annu. Rev. Cell Dev. BioI. 21: 297-318. Naumann K., Fischer A., Hofmann 1., Krauss v., Phalke S., Irmler K., Hause G., Aurich A.-C., Dorn R., Jenuwein T., and Reuter G. 2005. Pivotal role of AtSUVH2 in heterochromatic histone methylation and gene silencing in Arabidopsis. EMBO f. 24: 1418-1429. Onodera Y., Haag J.R., Ream T., Nunes P.c., Pontes 0., and Pikaard C. 2005. Plant nuclear RNA polymerase IV mediates siRNA and DNA methylation-dependent heterochromatin formation. Cell 120: 613-622. Pandey R., Muller A., Napoli C.A., Selinger D.A., Pikaard C.S., Richards E.J., Bender J., Mount D.W., and Jorgensen R.A. 2002. Analysis of histone acetyltransferase and histone deacetylase families of Arabidopsis thaliana suggests functional diversification of chromatin modification among multicellular eukaryotes. Nucl. Acids Res. 30: 5036-5055. Park M.Y., Wu G., Gonzalez-Sulser A., Vaucheret H., and Poethig R.S. 2005. Nuclear processing and export of microRNAs in Arabidopsis. Proc. Natl. Acad. Sci. 102: 3691-3696. Peregrine A., Yoshikawa M., Wu G., Albrecht H.L., and Poethig R.S. 2004. SGS3 and SGS2/SDElIRDR6 are required for juvenile development and the production of trans-acting siRNAs in Arabidopsis. Genes. Dev. 18: 2368-2379. Pikaard C.S. 2000. The epigenetics of nucleolar dominance. Trends Genet. 16: 495-500. Rocha P.S., Sheikh M., Melchiorre R., fa*gard M., Boutet S., Loach R., Moffatt B., Wagner c., Vaucheret H., and Furner 1. 2005. The Arabidopsis hom*oLOGY-DEPENDENT GENE SILENCINGl gene codes for an S-adenosyl-L-hom*ocysteine hydrolase required for DNA methylation-dependent silencing. Plant Cell 17: 404-417. Schubert D., Clarenz 0., and Goodrich J. 2005. Epigenetic control of plant development by Polycomb-group proteins. Curro Opin. Plant BioI. 8: 553-561. Senda M., Masuta c., Ohnishi S., Goto K., Kasai A., Sano T., Hong J.S., and MacFarlane S. 2004. Patterning of virus-infected Glycine max seed coat is associated with suppression of endogenous silencing of chalcone synthase genes. Plant Cell 16: 807-818. Souret EE, Kastenmayer J.P., and Green P.J. 2004. AtXRN4 degrades mRNA in Arabidopsis and its substrates include selected miRNA targets. Mol. Cell 15: 173-183. Springer N.M. and Kaeppler S.M. 2005. Evolutionary divergence of monocot and dicot methyl-CpG-binding domain proteins. Plant Physiol. 138: 92-104. Springer N.M., Napoli C.A., Selinger D.A., Pandey R., Cone K.C., Chandler v.L., Kaeppler H.E, and Kaeppler S.M. 2003. Comparative analysis of SET domain proteins in maize and Arabidopsis reveals multiple duplications preceding the divergence of monocots and dicots. Plant Physiol. 132: 907-925. Starn M. and Mittelsten Scheid O. 2005. Paramutation: An encounter leaving a lasting impression. Trends Plant Sci. 10: 283-290. Takeda S., Tadele Z., Hofmann I., Probst A.V., Angelis K.J., Kaya H., Araki T., Mengiste T., Mittelsten Scheid 0., Shibahara K., et al. 2004. BRUl, a novel link between responses to DNA damage and epigenetic gene silencing in Arabidopsis. Genes Dev. 18: 782-793. Tariq M. and Paszkowski J. 2004. DNA and histone methylation in plants. Trends Genet. 20: 244-251. Ting A.H., Schuebel K.E., Herman J.G., and Baylin S.B. 2005. Short double-stranded RNA induces transcriptional gene silencing in human cancer cells in the absence of DNA methylation. Nat. Genet. 37: 906-910. Vaucheret H., Vazquez E, Crete P., and Bartel D.P. 2004. The action of

E P ! G ENE TIC

ARGONAUTEl in the miRNA pathway and its regulation by the miRNA pathway are crucial for plant development. Genes Dev. 18: 1187-1197. Vazquez E, Gasciolli V., Crete P., and Vaucheret H. 2004a. The nuclear dsRNA binding protein HYLl is required for microRNA accumulation and plant development, but not posttranscriptional transgene silencing. Curro BioI. 14: 346-345. Vazquez E, Vaucheret H., Rajagopalan R., Lepers e., Gasciolli v., Mallory A.C., Hilbert J.-L., Bartel D.P., and Crete P. 2004b. Endogenous trans-acting siRNAs regulate the accum ulation of Arabidopsis mRNAs. Mol. Cell 16: 69-79. Voinnet O. 2005. Induction and suppression of RNA silencing: Insights from viral infections. Nat. Rev. Genet. 6: 206-220. Wagner D. 2003. Chromatin regulation of plant development. Curro Opin. Plant BioI. 6: 20-28. Wang X.J., Gaasterland T., and Chua N.H. 2005. Genome-wide prediction and identification of cis-natural antisense transcripts in Arabidopsis thaliana. Genome BioI. 6: R30. Wassenegger M., Heimes S., Riedel 1., and Sanger H.L. 1994. RNA-

REG U L A T ION

I N

P LAN T 5

•

189

directed de novo methylation of genomic sequences in plants. Cell 76: 567-576. Xie Z., Johansen L.K., Gustafson A.M., Kasschau K.D., Lellis A.D., Zilberman D., Jacobsen S.E., and Carrington J.e. 2004. Genetic and functional diversification of small RNA pathways in plants. PLoS BioI. 2: EI04. Yoo B.-e., Kragler E, Varkonyi-Gasic E., Haywood V., Archer-Evans S., Lee Y.M., Lough T.J., and Lucas W.J. 2004. A systemic small RNA signalling system in plants. Plant Cell 16: 1979-2000.

WWW Resources http://asrp.cgrb.oregonstate.edu. Arabidopsis thaliana small RNA project http://mpss.dbi.udel.edu. MPSS (Massively parallel signature sequencing) http://www.arabidopsis.org/abrc. Arabidopsis Biological Resource Center Stocks http://www.chromdb.org. Plant Chromatin Database

HAP

OFF

Chromatin Modifications and Their Mechanism of Action Tony Kouzarides' and Shelley L. Berger I

The Gurdon Institute, University of Cambridge, Cambridge, United Kingdom 'The Wistar Institute, Philadelphia, Pennsylvania 19104

CONTENTS 1. Histones and Acetylation Are Regulatory to Transcription, 193

6. Ubiquitylation/Deubiquitylation and Sumoylation, 203

2. Acetylation and Deacetylation, 194

7. Themes in Modifications, 204

3. Phosphorylation, 196 4. Methylation, 197 4.7

Methylation of Lysines, 797

4.2

Demethylation of Lysines, 207

4.3

Methylation of Arginines, 207

5. Deimination, 202

7.7

Histone Code, 204

7.2

Modification Patterns, 204

7.3

Changes in Chromatin Structure Associated with Transcription Activation and Elongation, 205

8. Concluding Remarks, 206 References, 206

191

GENERAL SUMMARY Histones are the building blocks of nucleosomes, making an octameric structure that packages DNA in eukaryotes forming a structure known as chromatin. Chromatin is not a uniform structure, however, and in recent years, an explosion in our knowledge of the variations in chromatin structure has occurred. This, in turn, has enhanced our understanding of the mechanisms that regulate genome templated processes, the posttranslational modifications of histone proteins (a central feature of this genomic regulation). There are, in fact, a large number of histone posttranslational modifications (HPTMs), and they divide into two groups. First, there are the small chemical groups, including acetylation, phosphorylation, and methylation. Second, there are the much larger peptides, including ubiquitylation and sumoylation. How are HPTMs thought to affect genome regulation and function? Three mechanisms are commonly considered, and it is helpful to keep these mechanisms in mind as the wealth of information and history of HPTMs is presented in this chapter. First, HPTMs may somehow affect the structure of chromatin; for example, by preventing crucial contacts that facilitate certain chromatin conformations or higher-order structures (which can be considered as cis-modifying effects). In contrast, two other mechanisms are considered to operate in trans. HPTMs

may disrupt the binding of proteins that associate with chromatin or histones. Alternatively, HPTMs may provide altered binding surfaces that attract certain effector proteins. This third mechanism has been characterized in the most detail, and such recruitment of proteins is defining with regard to the functional consequence: That is, it may have an activating or repressive outcome on transcription. The large number of HPTMs that have been discovered and their various combinations have led to the idea that HPTMs regulate via combinatorial patterns, in temporal sequences, and can be established over short- and long-range distances. These varied mechanisms establish different functional outcomes-some transient, others stable and epigenetically heritable. It was during the 1960s that Vincent Allfrey identified acetylation, methylation, and phosphorylation of histones purified from many eukaryotes. Histones were also the first recognized ubiquitylated protein substrates. However, although Allfrey observed certain correlations between modifications and transcriptionally active chromatin sources, genetic and functional evidence to support a role for HPTMs in gene regulation did not emerge until much later. In fact, many scientists studying the biochemistry and genetics of gene regulation during the 1980s and 1990s were skeptical that HPTMs had a causal role in gene regulation.

C H ROM A TIN

MOD I Fie A T ION 5

1 Histones and Acetylation Are Regulatory to Transcription As shown in this chapter, histones are subject to many different posttranslational modifications (PTMs), and time will undoubtedly reveal new, as-yet-unknown HPTMs. The known modifications can be categorized as either small chemical groups discussed in Sections 2-4 of this chapter, or as larger peptide changes to histones as discussed in Section 5 (see Table 1). The mechanism by which HPTMs affect the chromatin template and related processes such as gene transcription or repression are considered in the context of three conceptual models illustrated in Figure 1. Model 1 proposes that posttranslationally modified histones may, in some way, alter chromatin structure. In Model 2, an HPTM may inhibit the binding of a factor to the chromatin template, whereas Model 3 proposes that an HPTM creates a binding site for a particular protein (see also Section 5 of Chapter 3). From a historical perspective, what changed the mainstream view that chromatin was largely inert packing material for DNA? Early evidence that HPTMs regulated transcriptional activation and silencing came from experiments in Saccharomyces cerevisiae during the late 1980s. This budding yeast provides an efficient organism to carry out genetic experiments (both forward and reverse genetics) to examine the importance of histones. The reason is that, unlike higher eukaryotes where there are multiple copies of each histone gene, the single-copy yeast

Table 1. Types of covalent histone posttranslational modifications Role in transcription

Histone-modified sites

Acetylation

activation

H3 (K9,K14,K18,K56) H4 (K5,K8K12,K16) H2A H2B (K6,K7,K16,K17)

Phosphorylation

activation

H3 (510)

Methylation

activation

H3 (K4,K36,K79)

repression

H3 (K9,K27) H4 (K20)

activation

H2B (K123)

repression

H2A (Kl19)

repression

H3 (7) H4 (K5,K8,K12,K16) H2A (K126) H2B (K6,K7,K16,K17)

GROUP 1

GROUP 2 Ubiquityiation

5umoylation

HPTMs are categorized into two groups: Group 1 represents small chemical group modifications, whereas Group 2 includes larger chemical modifications.

AND

THE I R

M E C HAN ISM

0 F ACT ION

193

Model 1: Chromatin structural change

.• ~~~

---

Model 2: Inhibit binding of negative-acting factor

Model 3: Recruit positive-acting factor

mod

binder mod

Figure 1. Models Showing How Histone Posttranslational Modifications Affect the Chromatin Template

Modell proposes that changes to chromatin structure are mediated by the cis effects of covalent histone modifications, such as histone acetylation or phosphorylation. Model 2 illustrates the inhibitory effect of an HPTM for the binding of a chromatin-associated factor (CF), as exemplified by H351 0 phosphorylation occluding HP1 binding at methylated H3K9. In Model 3, an HPTM may provide binding specificity for a chromatin-associated factor. A classic example is HP1 binding through its chromodomain to methylated H3K9.

histone genes can easily be genetically manipulated. For instance, in a background where all the histone genes have been deleted, a copy of each gene can be introduced, encoded on an episome that carries a selectable marker, such as the URA3 gene, to maintain the episome. A second copy of the histones can be introduced on a second episome, carrying a different selectable marker. This second copy can be mutated by site-directed mutagenesis, and then the first, wild-type copy can be selectively lost from the cell using the 5-FOA (5-fluoroorotic acid) drug, which causes the URA3 gene product to be toxic to the cell. The end result is that the only copy present in the cell is the altered second episomal copy, which contains any number of mutations to be tested. In S. cerevisiae, the histone genes are arranged in pairs of H3/H4 and H2A/H2B, and their transcription is highly coordinated to coincide with S phase. Each nucleosome is assembled from an H3/H4 tetramer and two dimers of H2A/H2B; when one pair of either duo is under- or overtranscribed, nucleosomes are depleted. This alteration of histone dosage by genetic means provided some of the initial evidence that chromatin structure is crucial for regulating expression. One such

194

C HAP T E R 7 0

approach utilized "forward" genetics, where random mutations were selected that result in gene activation of a marker (Clark-Adams et al. 1988). These mutations were found to alter the amount of histone pairs. A second approach used "reverse" genetics, where directed depletion of histone genes provided clear evidence that histones regulate gene transcription (Han and Grunstein 1988). The next step was to direct deletion of only the histone amino-terminal tails (the sites of many HPTM) or to carry out substitution mutations of acetylation sites in histones. These more surgical changes also caused decreases in gene activation, suggesting that acetylation is required for gene transcription (Durrin et al. 1991). Other approaches investigated whether nucleosomes are naturally altered during gene activation. Biochemical experiments had shown that nucleosomes were repressive to transcription on DNA templates in vitro (Workman and Roeder 1987), but whether this was true in vivo was under debate. Some promoters have naturally positioned nucleosomes upstream of transcriptional start sites, and these positioned nucleosomes became altered when the gene was activated (Svaren et al. 1994; Shim et al. 1998). In the case of PROS, nucleosome alteration required an activator, showing that without transcription the nucleosomes were not changed. However, it was unclear whether this alteration was a cause or an effect of transcription. To address this, the TATA box was deleted, which abrogated transcription in yeast. Nucleosome position nevertheless changed, strongly suggesting that the alteration of nucleosomes preceded transcription. Taken together, these experiments began to provide strong evidence that both nucleosome repositioning and acetylation of specific residues within the histone tails may be required for transcriptional activation. 2 Acetylation and Deacetylation

Additional experimental approaches continued to provide evidence that acetylation (versus the absence of acetylation) correlates with transcription. Regions that are transcriptionally active, or are poised for transcription, tend to have an "open" chromatin configuration and therefore are accessible to enzymes such as DNase and MNase, which, when added to isolated but intact nuclei, can digest DNA. In the early 1990s, researchers began to use chromatin immunoprecipitation (ChIP), a powerful technique for analyzing what proteins are bound to particular DNA sequences in vivo. This involves cross-linking proteins that are bound to DNA using a cell-permeable chemical such as formaldehyde, followed by sonication to break up the

DNA:protein complexes into smaller fragments. The DNA:protein complexes of interest are then immunoprecipitated using a specific antibody as a probe. The crosslinks are then reversed in order to isolate and identify the DNA sequences that associated with the antibody-bound protein, by analysis using either radioactively labeled DNA probes or PCR. One group used this method to investigate the correlation between DNA sites around the active globin genes that are hypersensitive to DNase digestion and associate with acetylated histones in chicken erythrocytes; the correlation was remarkably close (Hebbes et al. 1994). In S. cerevisiae, similar approaches were employed within transcriptionally silenced regions of the genome, and they showed very low levels of histone acetylation (Braunstein et al. 1993). Conversely, genetic disruption of silencing correlated with increased acetylation. All of these experiments were slowly revealing that histones and, in particular, sites of reversible acetylation play a role in gene regulation. However, it was not until the mid-1990s that the first nuclear histone acetylation and deacetylation enzymes were identified, and these provided the "smoking gun"-the most direct evidence that these enzymes playa role during transcription. The first nuclear histone acetyltransferase (HAT) was isolated from the socalled macronucleus (the very large transcriptionally active nucleus, as distinct from the meiotic micronucleus) from Tetrahymena, which has high transcription rates (Brownell et al. 1996). The key approach was the "in-gel" assay to detect HAT activity: A complex mixture of proteins from cell extracts was separated on a histone-permeated SDS gel, the peptides were then subjected to renaturation, and proteins with HAT enzymatic activity labeled the gel by the transfer of radiolabeled cofactor, acetyl coenzyme A, onto localized histones. This allowed further biochemical fractionation and purification of the polypeptide. The HAT enzyme that was identified was hom*ologous to a previously isolated transcriptional coactivator in S. cerevisiae, called GenS, known to interact with transcriptional activators. Contemporaneously, the first histone deacetylase (HDAC) enzyme was isolated via biochemical purification (Taunton et al. 1996). In this case, the enzyme was purified from cell extracts using an inhibitor bound to an insoluble matrix, which physically bound to the catalytic site of the enzyme. The enzyme was hom*ologous to a previously isolated gene, which has a cofactor role in gene repression. These remarkable parallel findings for the first enzymes found to metabolize acetyl groups on histones led to a model that is now the paradigm for gene-specific histone PTMs: DNA-bound activators recruit HATs to acetylate nucleosomal histones, while

C H ROM A TIN

MOD I Fie A T ION 5

repressors recruit HDACs to deacetylate histones. These changes lead to alterations of the nucleosome and up- or down-regulation of the gene, respectively (Fig. 2). Many other well-known coactivators and corepressors were shown to possess HAT or HDAC activity, or to associate with such enzymes (Sterner and Berger 2000; Roth et al. 2001). Moreover, the enzymatic activities of the HATs and HDACs are critical for their role in gene activation and repression. The enzymes are often components of large complexes that are modular in structure and function; histone-modifying enzymatic activity is just one function, and others include, for instance, the recruitment of the TATA-binding protein (TBP) (Grant et al. 1998). Interestingly, certain nuclear hormone receptors function both as DNA-binding transcriptional repressors (when not bound to hormone ligand) and as transcriptional activators (when bound to hormone ligand); the receptors do this partly through the PTM of target chromatin regions, by recruiting HDACs when unliganded and HATs when liganded (Baek and Rosenfeld 2004). HAT proteins can acetylate lysine residues on all four core histones, but different enzymes possess distinct specificities in their substrate of choice (Fig. 3; Table 1), although each enzyme rarely targets just a single site. One major HAT family-GNAT (for GcnS related acetyltrans-

Gene activator recruits histone acetyltransferase

Gene repressor recruits histone deacetylase

Figure 2. Histone-modifying Enzymes Are Recruited to Promoters by DNA-binding Transcription Factors Histone acetyltransferases (HAT) are recruited by activators that bind to specific upstream activating sequences (UAS). This enzyme catalyzes the acetylation of local histones, known to contribute to transcriptional activation. Histone deacetylases (HDAC) are recruited by repressors of transcription that bind to upstream repressive sequences (URS) and deacetylate local histones. This contributes to transcriptional repression.

AND

THE I R

ME C HAN ISM

0 F ACT ION

195

ferase)-targets histone H3 as its main substrate. A second major HAT family, the MYST family, targets histone H4 as its main substrate. A third major familyCBP/p300-targets both H3 and H4, and is the most promiscuous. Structural analyses have been carried out for the catalytic domains of the first two major families (GNAT and MYST), and they are distinct; the structure of the CBP Ip300 family has not yet been solved. Incidentally, each of these acetyltransferase families is also able to acetylate non-histone substrates (Glozak et al. 2005). As discussed above, there are three models for the role of HPTMs in regulating chromatin structure (Fig. 1). The first model considers structural changes to chromatin induced by the direct effects of HPTMs, such as changes in charge. In this case, the neutralization of positively charged lysine by acetylation reduces the strength of binding of the strongly basic histones or histone tails to negatively charged DNA, and thus opens DNA-binding sites (VetteseDadey et al. 1996). Still focusing on the first model, there is also evidence that acetylation can decompact nucleosome arrays, consistent with a role in opening chromatin for gene activation (Shogren-Knaak et al. 2006). The third model proposes that HPTMs provide a binding surface for proteins to associate with chromatin and regulate DNAtemplated processes; this was first shown for acetylation. A specialized protein domain called a bromodomain, commonly found in chromatin-associated proteins, specifically binds to acetylated lysines (Fig. 3) (Dhalluin et al. 1999). Bromodomains are present in many HATs, such as GenS and CBP/p300. Proteins with this motif, when part-oflarge chromatin-associating/altering complexes such as the ATPdependent remodeling complex, SwilSnf, promote its binding to chromatin (Hassan et al. 2002). Other examples of proteins containing bromodomains with binding specificity to acetylated histone include Tafl and Bdfl in the TFIID complex, Rsc4 in the Rsc remodeling complex, and Brd2 in a large family of bromodomain proteins. There are numerous HDAC enzymes that remove acetyl groups (Kurdistani and Grunstein 2003; Yang and Seto 2003). They fall into three catalytic groups, which are conserved through evolution from S. cerevisiae to mammals, and referred to as Type I, Type II, and Type III or Sir2-related enzymes. Type I and Type II have a related mechanism of deacetylation, which does not involve a cofactor, whereas the Sir2-related enzymes require the cofactor NAD as part of their catalytic mechanism. The structures of representatives for all three families have been solved. Many of the HDACs are found within large multisubunit complexes, components of which serve to target the enzymes to genes, leading to transcriptional

196 •

C HAP T E R

." Swil

\ Histone

Sot

~H~3~~~....J....-1

Histone H4

Figure 3. Characterized Sites of Histone Acetylation Histones are mostly acetylated at lysine residues located in the amino termini of H3 and H4, with the exception of H3K56 localized in the globular domain. The proteins that express binding specificity to acetylated histones are shown.

repression. For example, Rpd3 is part of a large complex including the HDAC Sin3, which interacts with DNAbound repressors (Kurdistani and Grunstein 2003; Yang and Seto 2003). Rpd3 is also part of a small complex, which is targeted to gene open reading frames (ORFs) via chromodomain association with H3K36me (see Section 4 for further discussion of chromodomains). This results in histone deacetylation, in part to suppress internal RNA polymerase II (pol II) initiation, and also to regulate different steps during the transcription cycle (Carrozza et al. 2005; Joshi and Struhl 2005).

3 Phosphorylation Phosphorylation is the most well known PTM because it has long been understood that kinases regulate signal transduction from the cell surface, through the cytoplasm, and into the nucleus, leading to changes in gene expression. Histones were among the first proteins found to be phosphorylated. By 1991, it was discovered that when cells were stimulated to proliferate, the so-called "immediate-early" genes were induced to become transcriptionally active and to function in stimulating the cell cycle. This increased gene expression correlates with histone H3 phosphorylation (Mahadevan et al. 1991). The histone H3 Serine 10 residue (H3SlO) has turned out to be an important phosphorylation site for transcription from yeast to humans, and appears to be especially important in Drosophila (Nowak and Corces 2004). Many kinases have been identified that target this site, including Mskl/2 and the related Rsk2 in

mammals, and SNFI in s. cerevisiae (Sassone-Corsi et al. 1999; Lo et al. 2001; Soloaga et al. 2003). Studies of linker histone HI in Tetrahymena have revealed that phosphorylation of this histone may also affect transcriptional control. Perhaps counterintuitively, phosphorylation of certain residues correlates with chromosome condensation, during both mitosis and meiosis. It is unclear how phosphorylation contributes to the process, but recently, H3S10 phosphorylation acts like a temporal switch, ejecting HPI bound to the adjacently methylated H3K9 residue, referred to as the methyl-phos binary switch (FiscWe et al. 2005; Hirota et al. 2005). It remains to be seen whether this, perhaps in concert with the phosphorylation of H3S28 and H3Tl1, may effect chromatin condensation by recruiting the condensin complex and the mitotic spindle (Nowak and Corces 2004). Less is known about the precise mechanistic role of histone phosphorylation. There is evidence to support all three models for the role of HPTMs. First, histone phosphorylation alters chromatin compaction in vivo (Model 1). Indeed, work in Tetrahymena demonstrated that the collective negative "charge patch" resulting from phosphorylation of clusters of nearby residues within linker histone HI affects the affinity of its binding to DNA, positively increasing the transcriptional potential of the local chromatin environment (Dou and Gorovsky 2002). In support of Model 2, proteins bound to chromatin can be dislodged by phosphorylation. This was recently demonstrated by the lowered binding affinity of HPI during mitosis subsequent to mitosis-specific H3SlO phosphory-

C H ROM A TIN

MOD I Fie A T ION 5

lation (FiscWe et a1. 2005; Hirota et a1. 2005). In support of Model 3, the 14-3-3 adapter protein, a known phospho-binding protein, recognizes H3S10ph at promoters of inducible genes (Macdonald et a1. 2005). 4 Methylation

Methylation as a histone covalent modification is more complex than any other, since it can occur on either lysines or arginines. Additionally, unlike any other modification in Group 1, the consequence of methylation can be either positive or negative toward transcriptional expression, depending on the position of the residue within the histone (Table l). A further level of complexity lies in the fact that there can be multiple methylated states on each residue. Lysines can be mono- (mel), di- (me2), or tri(me3) methylated, whereas arginines can be mono- (mel) or di- (me2) methylated. Given that there are at least 24 identified sites of lysine and arginine methylation on H3, H4, H2A, and H2B, the number of distinct nucleosomal methylated states is enormous. Such combinational potential of methylated nucleosomes may be necessary, at least partly, to allow for the regulation of complex and dynamic processes such as transcription, which requires sequential and precisely timed events (Jenuwein and Allis 2001; y. Zhang and Reinberg 2001; Lee et a1. 2005; Martin and Zhang 2005; Wysocka et al. 2006a).

TRANSCRIPTION REGULATION Silent heterochromatin Transcriptional

activation ' "

AND

THE I R

ME C HAN ISM

0 F ACT ION

197

4.1 Methylation of Lysines The fact that lysine residues within histones are methylated has been known for many decades. The biological significance of this modification has only come to light recently, however, following the identification of the first lysine methyltransferase that uses histones as its substrate (Rea et a1. 2000). Now, more histone lysine methyltransferases (HKMTs) have been identified, and their sites of modification on histones are defined (Martin and Zhang 2005). All of these enzymes, except Dot 1, share the SET domain, which contains the catalytically active site and allows binding to the S-adenosyl-Lmethionine cofactor. Of the many known methylated sites, six have been well characterized to date: five on H3 (K4, K9, K27, K36, K79) and one on H4 (K20). Methylation at H3K4, H3K36, and H3K79 has, in general, been linked to activation of transcription, and the rest to repression (Table 1). In addition, two of these sites-H3K79me and H4K20me-have been implicated in the process of DNA repair. Specific protein binders have been identified that recognize each of the six characterized methylation sites (Fig. 4). These proteins have one of three distinct types of methyl lysine recognition domains: the chromo, tudor, and PHD repeat domains. Below, each of these characterized modifications is discussed in more detail.

DNA REPAIR

Transcriptional elongation

Transcriptional / memory

1 PC

Histone

HI~3~:--~K~9~-""'~ Figure 4. Sites of Histone Methylation, Their Protein Binders, and Functional Role in Genomic Processes Methylation of histones occurs at lysine residues in histones H3 and H4. Certain methylated lysine residues are associated with activating transcription (green Me flag), whereas others are involved in repressive processes (red Me flag). Proteins that bind particular methylated lysine residues are indicated.

198 • C HAP T E R 7

H3K4

METHYLATION

Methylation of H3K4 is associated with euchromatin and, specifically, with genes that are active or destined to be so. The demonstration that H3K4 methylation correlates with active chromatin came from analysis of the chicken ~-glo bin locus and the budding yeast mating-type loci (Litt et al. 2001; Noma et al. 2001). ChIPs using antibodies specific for methylated H3K4 indicated that islands of the modified histones track active genes. Subsequent work in yeast established that different methyl states are important for activity and that the trimethyl state (H3K4me3) appears during the process of active transcription (Santos-Rosa et al. 2002). H3K4me3 is observed at the 5' ends of genes in yeast during activation of transcription. Three components of the transcriptional machinery are thought to be responsible for this mark. First, RNA pol II that has been phosphorylated at Ser-5 of the carboxy-terminal domain (CTD) can recruit the Setl HKMT that methylates H3K4 in the vicinity of promoters (Fig. 5). Such phosphorylation normally releases RNA pol II from the transcription initiation complex into an early elongating complex (often referred to as promoter clearance or escape). The second component that recruits H3K4me3 is the PAF complex, which regulates different steps of RNA metabolism and also interacts with Setl. The third component important for the establishment of H3K4me3 is monoubiquitylation of H2B at Lys-l23 (H2BK123ub1, or H2BK120ubl in humans, discussed further in Section 6). What remains unclear is what transcriptional elongation processes H3K4me3 controls (Hampsey and Reinberg 2003); however, factors that bind specifically to methylated H3K4me3 are beginning to reveal its role. Mechanistically, H3K4 methylation can lead to the recruitment of specific factors such as the CHD 1 protein, shown to bind to H3K4me2 and me3 (Fig. 4), and the

NURF complex, known to mobilize nucleosomes at active genes in Drosophila. The domains that mediate association with methylated H3K4 are a tandem set of chromodomains in Chd1 (Sims et al. 2006) and a PHD finger within NURF (Li et al. 2006). Other proteins recruited by H3K4 methylation include the ISWI ATPase, which binds indirectly via other protein(s). Conversely, there is evidence that the mammalian NuRD repressor complex no longer binds to methylated H3K4 tails (D.Y. Lee et al. 2005; Martin and Zhang 2005). Methylation at H3K4 seems to communicate with other modifications. For instance, methylation of H3K9 by the SUV39H HKMT is prevented in vitro if H3K4 is methylated and H3S10 is phosphorylated. This may well be a way to occlude the repressive H3K9 modification on actively transcribed genes. In a more elaborate "trans-tail" form of communication, the mono ubiquitylation of H2BK123 affects levels of H3K4me3. How this comes about is unclear, but one suggestion is that the Setl complex cannot tri-methylate H3K4 unless the nucleosome(s) is in a certain conformational state defined by ubiquitylation of H2B (Zhang and Reinberg 2001; D.Y. Lee et al. 2005) . The Setl/MLL/ALLl/HRX protein, which is the human hom*olog of Setl, can be recruited to HOX gene promoters. A distinct H3K4 HKMT, SMYD3, has been linked to transcriptional activation. Methylation by SMYD3 has also been linked to the induction of cell proliferation. Indeed, limited analysis of the human H3K4 methylating enzymes suggests that they are implic.ated in the genesis of cancer (D.Y. Lee et al. 2005). H3K36

METHYLATION

Evidence has led to a proposal that H3K36 methylation is necessary for efficient elongation of RNA pol II through

Figure 5. Role of Histone Lysine Methylation in Transcriptional Elongation

H3K36

PROMOTER ESCAPE

H3K36

TRANSCRIPTION ELONGATION

RNA polymerase II recruits distinct types of HKMTs, depending on the phosphorylation state of its carboxy-terminal domain (CTD). RNA pol II is activated for transcriptional initiation in the vicinity of the promoter, when Ser-5 is phosphorylated. This recruits the Setl HKMT to methylate H3K4. Phosphorylation of Ser-2 occurs during transcriptional elongation, prompting H3K36 methylation as a result of Set2 HKMT recruitment to the chromatin template.

C H ROM A TIN

MOD I Fie A T ION SAN 0

the coding region. This modification is highly enriched on the coding region of active genes, in contrast to the 5' location of H3K4 methylation. The Set2 protein is the HKMT capable of methylating H3K36. The Set2 enzyme binds preferentially to RNA pol II that has been phosphorylated within its CTD at Ser-2 (Fig. 5). This form of RNA pol II, which, incidentally, is different from the phosphorylated state associated with promoter clearance, tends to accumulate within the transcribed regions as well as at the 3' ends of the genes. This is consistent with the finding that H3K36me3 peaks at the 3' ends of genes that are actively transcribed. The recruitment of Set2 to active genes also requires components of the PAF complex, as in the case for the recruitment of Setl. However, H2B monoubiquitylation has a negative repressive role on H3K36 methylation (Zhang and Reinberg 2001; Martin and Zhang 2005). Indeed, the SAGA complex, recruited to transcribed genes in yeast, contains Ubp8, a deubiquitinase that is specific for H2BK123. Further studies have suggested that ubiquitylation and deubiquitylation of H2BK123 is an active process during transcription elongation. Processivity of RNA pol II through coding regions requires acetylation of nucleosomes. Transcriptional regulation also needs to suppress inappropriate internal initiation of transcription from cryptic start sites that occur inside coding regions. To suppress this process, methylation at H3K36 by Set2 creates a recognition site for the EAF3 protein through its chromodomain, which in turn mediates the recruitment of the Rpd3S HDAC complex. The deacetylase activity of Rpd3S then removes histone acetylation associated with elongation, thus suppressing internal initiation (Carrozza et al. 2005; Joshi and Struhl 2005; Keogh et al. 2005). Methylation of H3K36 has also been found at much lower levels in the promoter of inducible genes, but in this case, its effect appears to be repressive (Zhang and Reinberg 2001). H3K79

METHYLATION

Methylation at H3K79 is unusual because the modification lies within the core of the nucleosome rather than in the tail, where most other characterized methylation sites are found. Global analysis has shown that H3K79 is methylated in euchromatic regions of yeast and associates primarily with the coding region of active genes. Limited analysis in higher eukaryotes shows the same profile. The mammalian enzyme that methylates H3K79, hDOTlL, has been shown to mediate the leukemogenic functions of the MLL-AFI0 fusion protein. There is, however, no protein to date that binds H3K79me and links it to

THE I R M E C HAN ISM

0 F ACT ION

199

transcriptional events. The only mechanistic evidence of how H3K79 methylation functions in transcriptional activation comes from work in budding yeast. This shows that this modification somehow limits repressive proteins such as Sir2 and Sir3 at euchromatin, thus contributing to the regulation and maintenance of silent heterochromatin by enhancing their concentration at repressive chromatin regions. A distinct function ascribed to the H3K79 HKMT DotI in yeast is the mediation of DNA repair checkpoint. Consistent with this latter finding, a protein has been identified in human cells-P53BPl-that can bind to methylated H3K79 and has a role in DNA repair checkpoint function (Martin and Zhang 2005). H3K9

METHYLATION

This has been the most studied of histone modifications to date, primarily because the enzyme that methylates H3K9-SUV39Hl-was the first HKMT to be identified (Rea et al. 2000). The Drosophila hom*olog, Su(var)3-9, was initially identified as a suppressor of variegation, indicating that it was involved in the silencing mechanism of position-effect variegation (PEV), which involves the spreading of heterochromatin into adjacent euchromatic genes (for more detail, see Chapter 5). The realization that SUV39Hl had sequence similarity to a plant methyltransferase which had Rubisco as its substrate led to the identification of the Suv39 SET domain as the catalytic domain capable of methylating H3K9. Progress has been made in defining the function of H3K9 methylation in pericentromeric heterochromatin formation, which is also discussed extensively in other chapters (for studies on Drosophila, see Chapter 5; for studies in S. pombe, see Chapter 6; and for studies on RNAi-mediated heterochromatin formation, see Chapter 8). The results have come largely from studies in fission yeast and mammals, where heterochromatic structures are thought to be reasonably well conserved (but note, H3K9 methylation has not been detected in budding yeast). To summarize, the first stage of our understanding emerged from studies on the factors involved in the establishment of heterochromatin. This involves the cooperation of two proteins: SUV39H (or Clr4 in fission yeast) and its binding partner HPI (or Swi6 in fission yeast [Nakayama et al. 2001; Noma et al. 2001]). A model has been proposed whereby SUV39H methylates H3K9, creating a binding platform for HPl, through its chromodomain (Bannister et al. 2001; Lachner et al. 2001). Once HPI binds, it can spread onto adjacent nucleosomes by its association with SUV39H, which further catalyzes neigh-

200

C HAP T E R

boring histone methylation (Nakayama et a1. 2001). In addition, HP1 self-associates via the chromoshadow domain facilitating the spread of heterochromatin. How HP1 spreading dictates the formation of the densely packed heterochromatic structures is, however, unknown. The above model predicts that there should be a specific heterochromatin-based recruitment mechanism for the SUV39H HKMT enzyme, before HP1 can spread. The clue as to what this may be came from a series of experiments in fission yeast which showed a link between heterochromatin formation and the production of short interfering RNAs (siRNAs) (Hall et a1. 2002; Volpe et a1. 2002). These RNAs come from the bidirectional transcription of centrometric repeats which are processed into siRNAs by the enzyme dicer. The siRNAs are then packaged into the RITS complex, which contains the chromodomain-containing protein, Chp1, which binds methylated H3K9. Thus, the targeting of the RITS complex to chromatin forms the initiation stage of heterochromatin formation. The spreading and maintenance of heterochromatin over a 20-kb region, as described above, requires the methylation of histone H3K9 by the Clr4 HKMT and the binding of Swi6 to H3K9 methylated chromatin (Martin and Zhang 2005; for more detail, see Chapter 8). The interdependence of different repressive epigenetic mechanisms has emerged from studies first in Neurospora crassa, but also in plants, notably demonstrating a link between H3K9 methylation and the process of DNA methylation (see Chapters 6 and 8). H3K9 methylation is necessary for DNA methylation to take place, and the reciprocal connection seems to be operational, whereby H3K9 methylation is dependent on DNA methylation. Moreover, recent studies in mammalian cancer cells lacking DNA methyltransferase enzymes (Dnmts) show reduced levels of H3K9 methylation, and this can be attributed to the fact that the methyl-CpG-binding protein 1 (MBD1) associates with the H3K9 HKMT, SETDB1 (Zhang and Reinberg 2001; Martin and Zhang 2005; for more detail, see Chapter 18). Methylation at H3K9 also functions in the repression of euchromatic genes. ChiPs have detected this methylation at the promoter of mammalian genes when the genes are silent. The mechanism of this repression at euchromatic sites appears to be slightly different from those encountered at heterochromatic regions. The RB repressor protein delivers the SUV39H 1 HKMT and HP 1 to euchromatic genes such as the E2F-regulated cyclin E gene. Unlike heterochromatin, however, HP1 occupancy appears to be restricted to one or a few nucleosomes

around the initiation site, even though H3K9 methylation occurs elsewhere on the promoter. In another example, the KAP 1-repressor brings the ESET/SETD B1 HKMT to the promoter of KAP1 regulated genes and silences transcription by methylation of H3K9 and HP1 recruitment. The special restriction of HP1 on these euchromatic promoters, and the prevention of spreading, suggest a distinct mechanism of action for HP1 relative to its heterochromatic role. One possible mode of action for HP1, which has some support, is that it acts as an anchor into heterochromatin-rich nuclear compartments. Movements have been observed during the repression of euchromatic genes which show that a silenced gene is displaced into a heterochromatic region, a movement which is dependent on the gamma isoform of HP1 (Martin and Zhang 2005). Heterochromatin formation at telomeres, although involving HP1 and H3K9 methylation, varies from the aforementioned pericentromeric and silent euchromatic regions. In Drosophila, HP 1 is not recruited to telomere ends through its chromo- or chromoshadow domain, and H3K9 methylation is catalyzed by an unknown HKMT. In mammals, distinct HP1 hom*ologs, CBX1, CBX3, and CBXs, are involved in binding to methylated H3K9, transduced by the SUV39H1 and SUV39H2 proteins to form repressive chromatin domains at chromosome ends (for more detail, see Chapter 14). H3K27

METHYLATION

H3K27 methylation is a repressive modification found in three distinct places in the cell: (1) euchromatic gene loci, predominantly where there are Polycomb Response Elements (PREs) in the case of Drosophila, (2) at pericentromeric heterochromatin, and (3) at the inactive X in mammals. The enzyme that mediates H3K27 methylation is EZH2 in human cells, a hom*olog of the Drosophila ENHANCER OF ZESTE [E(Z)] protein. The EZH2 enzyme is found in a number of distinctive Polycomb repressive complexes (PRCs) which associate with specific repressive Polycomb DNA elements in promoters in Drosophila (see also Chapter 11). What targets the EZH2-containing complexes to specific genes in mammals is unknown, as Polycomb repressive elements have not been identified. However, targeting of these EZH2 complexes may be mediated by a variety of transcription factors, including GAGA and MYc. The mechanisms of repression by EZH2 involve methylation of H3K27 and the recruitment of the Polycomb (Pc) protein to this modified site (as in Model 3 in Fig. 1). An

C H ROM A TIN

MOD I Fie A T ION 5

important aspect of the pathway that leads to H3K27 methylation is that it is implicated in cancer. The EZH2 H3K27 HKMT is found overexpressed in a number of cancer tissues, including breast and prostate (Martin and Zhang 2005). H4K20

METHYlATION

Very little is known mechanistically about the role of this modification in transcriptional control. What is clear is that H4K20me2 and H4K20me3 are present at pericentromeric heterochromatin and that the HKMT enzymes that mediate these modifications are SUV4-20Hl and SUV4-20H2. Methylation of H3K9 seems to be required for methylation of H4K20. Another enzyme that can mono-methylate H4K20 in higher eukaryotes is PR-Set7, which has been implicated in mitotic events. Last, there is functional evidence that H4K20 methylation has been linked to DNA repair via the binding of the DNA damage checkpoint protein CrB2 in budding yeast (Martin and Zhang 2005). 4.2 Demethylation of Lysines

Until recently, it was unclear whether histone lysine demethylation was taking place in the cell. The search for such enzymes had been fruitless, and evidence existed that methyl groups can be quite stable on heterochromatin regions. The discovery of LSDI changed all that (Shi et al. 2004). This protein was shown to be an enzyme that removes methyl groups specifically from H3K4 and is involved in repression. LSD 1 is present in a number of different repressor complexes, and some of these allow it to more efficiently demethylate lysine residues from nucleosomal histone H3 (M.G. Lee et al. 2005; Shi et al. 2005). The specificity of LSDI can be changed if it binds a partner such as the androgen receptor (AR). An LSD 1AR complex demethylates H3K9 instead of H3K4, and under these conditions, activates, rather than represses, transcription (Metzger et al. 2005). Recently, five new demethylases were identified that possess a common catalytic structure distinct from LSD 1, called the JmjC-domain (Fig. 6). This domain was previously predicted to possess enzymatic activity (Trewick et al. 2005). These new demethylases are found to demethylate distinct methyl states of H3K9 and H3K36. JHDMI demethylates H3K36mel and me2, and JHDM2A demethylates H3K9mel and me2 (Tsukada et al. 2006; Yamane et al. 2006). The tri-methyl state of these two modified residues is removed by a distinct set of enzymes. JHDM3A and JMJD2A can act on both H3K36me3 and

AND

THE I R M E C HAN ISM

0 F ACT ION

•

201

H3K9me3 (Cloos et al. 2006; Fodor et al. 2006; Klose et al. 2006; Tsukada et al. 2006; Whetstine et al. 2006). It is perhaps surprising that enzymes exist which can simultaneously demethylate an active (e.g., H3K36me) and a repressive (e.g., H3K9me) mark. This may be explained by the recent finding that H3K9 methylation also associates with actively transcribed genes (Vakoc et al. 2005). In a more classic mechanism, the JHDM2A enzyme is recruited to promoters by AR, where it is involved in activating transcription via demethylation of H3K9 (Yamane et al. 2006). Structural analysis of JMJD2A has revealed that four distinct domains UmjN, JmjC, an unusual Zing finger, and a carboxy-terminal domain) come together to form the catalytic core. A deep cleft is formed by these domains coming together, which demands a conformational change in the enzyme or substrate to accommodate the methyl group for demethylation. Such a conformational shift may explain the specificity of demethylation (Chen et al. 2006). It is interesting to note that one of the newly discovered demethylases, JMJD2C, was previously known as GASCl, a gene amplified in squamous carcinoma. Consistent with a causative role for this enzyme in cancer development, the overexpression of GASCI was shown to induce cell proliferation (Cloos et al. 2006). These results together imply that demethylases as well as HKMTs may be targets for anticancer drug development (see also Chapter 24). 4.3 Methylation of Arginines

The importance of histone arginine methylation in transcriptional control came after the identification of CARMI, an enzyme that can methylate arginines within H3 in vitro (Chen et al. 1999). In vivo, arginine methylation was subsequently demonstrated in experiments using specific antibodies to arginine-methylated sites (Strahl et al. 2001; Wang et al. 2001; Bauer et al. 2002). Arginine methylation has been implicated in the positive and negative regulation of transcription. Two methyltransferases, PRMTl (protein arginine methyltransferase) and PRMT4/CARMl, have been linked to transcriptional activation. PRMTl has the ability to methylate H4R3 (Strahl et al. 2001; Wang et al. 2001), whereas PRMT4/CARMI can catalyze the methylation of H3R2, H3R17, and H3R26 (Schurter et al. 2001; Bauer et al. 2002). Specific transcription factors (NR, p53, YYl, NFKB) recruit these enzymes to specific promoters where they activate transcription. In contrast, PRMT5 (which can methylate H3R8 and H4R3) acts as a repressor of numerous genes, including some regulated by MYC (Fabbrizio et al. 2002; Pal et al. 2003).

202 •

C HAP T E RIO

. .\.~~v. ~ '\ Histone H3

???

Histone H4

Figure 6. Histone lysine Demethylases and Their Sites of Demethylation on Histone H3 Sites of histone lysine methylation may be mono-, di-, or tri-methylated. Known histone lysine demethylases show different specificities in demethylating histone residues or methylated states, as illustrated.

Most of our knowledge regarding arginine methylation comes from the analysis of the estrogen-signaling pathway that regulates the pS2 gene. ChIPs have indicated that a complex and cyclic set of events follows the stimulation of this gene by estrogen. The estrogen receptor is first recruited to the pS2 promoter within minutes of the stimulus and brings with it many protein complexes and enzymes that modify histones (Metivier et al. 2003). Relevant here is the recruitment of CARMI and PRMTl, which can methylate arginine residues of histones H3 and H4 (Ma et al. 2001; Bauer et al. 2002). This methylation is detected very soon after the arrival of the enzymes and coincides with the appearance of active RNA pol II on the promoter. Surprisingly, however, minutes after these events, methylation at arginines is no longer detected by specific antibodies, and RNA pol II disappears. Soon after that, methylation of arginines and RNA pol II reappears (Metivier et al. 2003; Cuthbert et al. 2004; Wang et al. 2004). The reason for these cyclic events is not known. One possibility is that it provides a mechanism for rapid shutoff of transcription if estrogen signaling fails. Experiments done on reconstituted chromatin templates have helped establish a direct role for arginine methylation in gene expression. Analysis of p53-mediated activation of transcription in vitro has shown that there is a synergistic effect of methylation transduced by PRMTl and CARMI, and acetylation by CBP/p300 (An et al. 2004). Furthermore, these assays have confirmed the in vivo observations on the pS2 gene that a specific order of events takes place during activation in which the sequen-

tial activity of PRMTl, CBP/p300, and CARMI is necessary (Metivier et al. 2003). Given that arginine methylation is such a dynamic process, several ways have been described in which the effectiveness of arginine methyltransferase is controlled. First, the interaction of the enzyme with another protein can control its substrate specificity. Second, there is potential for competition between enzymes for a given arginine substrate. Both PRMTI and PRMT5 can methylate H4R3, but the first enzyme is an activator and the second is a repressor of transcription. A third level of regulation of the methyl state may come from arginine demethylation. Such an activity has not yet been isolated, but there are clear indications of the rapid disappearance of methyl groups from arginines, making such an activity a very attractive possibility (Zhang and Reinberg 2001; D.Y. Lee et al. 2005; Wysocka et al. 2006a).

5 Deimination The lack of an arginine demethylase prompted the suggestion that other types of enzymatic reactions may antagonize arginine methylation (Bannister et al. 2002). One such reaction is deimination, a process by which an arginine can be converted to citrulline via the removal of an imine group. If the arginine is mono-methylated, removal of methylamine would effectively result in the removal of the methyl group from the arginine. The presence of citrulline in histones has now been demonstrated, and the enzyme, PADI4, has been identified that can convert arginines within histones into citrulline (Cuthbert et

C H ROM A TIN

MOD I Fie A T ION SAN 0

al. 2004; Wang et al. 2004). Moreover, the appearance of citrulline on histones H3 and H4 correlates with the disappearance of arginine methylation in vivo. Additionally, analyses of estrogen-regulated promoters, where arginine methylation coincides with the active state of transcription, have shown that citrulline appears when the promoter is shut off. Many unanswered questions remain regarding this modification. Is the citrulline acting to suppress active methylation at arginines, or does it repress transcription by actively recruiting proteins? What about the reversal of citrulline deposition? This clearly takes place on the promoters at a very rapid pace, but is this an enzymatically driven reaction or is it merely due to the replacement of nucleosomal histones by histone variants, which contain arginine in place of citrulline?

6 Ubiquitylation/Deubiquitylation and Sumoylation Ubiquitin and SUMO are quite distinct PTMs compared to acetylation, phosphorylation, and methylation. Whereas the latter PTMs are small chemical groups, Ub and SUMO are large polypeptides, which increase the size of the histone by approximately two-thirds. Ub and SUMO are 18% identical in sequence and share a threedimensional structure, but are dissimilar in surface charge. Histones were the first proteins shown to be monoubiquitylated, although precise positions of Ub were not identified until relatively recently (Robzyk et al. 2000; Wang et al. 2004). Like methylation, and unlike acetylation and phosphorylation (and, possibly, sumoylation), ubiquitylation can be either repressive or activating, depending on the specific sites. H2A and H2B are

TRANSCRIPTIONAL REPRESSION

THE I R M E C HAN ISM

0 F ACT ION

•

203

monoubiquitylated, which contrasts with proteolysisassociated polyubiquitylation. The effects of monoubiquitylation on each core histone are opposite (Fig. 7). H2B monoubiquitylation is activating to transcription, transduced by Rad6/Brel (and the human counterpart RNF20/RNF40 + UbcH6) (Wood et al. 2003; Kim et al. 2005; Zhu et al. 2005), and leads to H3K4 methylation, as described in the previous section and in the next section (Henry et al. 2003; Kao et al. 2004). This sequence of events, although as yet not understood mechanistically, is conserved from yeast to human (Kim et al. 2005; Zhu et al. 2005). H2AK119ubl, on the other hand, is repressive to transcription in mammals and catalyzed by the Polycomb group Bmil/RinglA protein (Wang et al. 2004). There is no evidence for evolutionary conservation of repressive H2Aub in yeast. To date, no histone-specific ubiquitin-binding proteins have been identified. However, because numerous ubiquitin interaction domains have been documented as binding to non-histone ubiquitylated substrates, it seems highly likely that effectors for ubiquitylated histones will be found. However, they may interact in a different manner than the chromatin interacting domains for acetylation and methylation; i.e., there are likely to be two simultaneous binding interactions, one to a surface on ubiquitin and a second interaction within histone sequences, to provide specificity of interaction. Deubiquitylation of the H2BK123 site is involved in both gene activation and maintenance of heterochromatic silencing, through the action of two distinct proteases, Ubp8 and Ubp 10. Ubp8 is a subunit of the SAGA histone acetylation complex (Sanders et al. 2002) and acts following ubiquitylation by Rad6 (Henry et al. 2003;

TRANSCRIPTIONAL ACTIVATION Ub

Histone H2A

.. Histone H2B

~~=-_ K123

Figure 7. Sites of Histone Ubiquitylation and Their Consequence for Transcriptional Regulation Ubiquitylation of H2A at Lys-119 is correlated with transcriptional repression. H2BK123 ubiquitylation is conversely associated with transcriptional activation.

204

C HAP T E RIO

Daniel et al. 2004). This sequence of H2B ubiquitylation followed by deubiquitylation is required to establish the appropriate levels ofH3K4 (H2Bub required) and H3K36 methylation (H2Bub not required) (Henry et al. 2003). UbpIO functions at silenced regions to maintain low levels of H3K4me and H3/H4 lysine acetylation, and thus assists in preventing transcription (Emre et al. 2005; Gardner et al. 2005). Sumoylation is the only HPTM described in yeast as repressive and is conserved in mammals (Shiio and Eisenman 2003). Its role may be generally negative-acting to prevent activating HPTMs. The inhibition of active HPTMs may occur through two mechanisms. First, SUMO-histone may directly block lysine substrate sites that are alternatively acetylated or sumoylated (as in Model 2 in Fig. 1). Second, sumoylated histones may recruit HDACs both to chromatin (Model 3) and via a SUMO group that occurs on DNA-bound repressors.

2004). Similarly, H3K9me has recently been shown to increase during gene induction (Vakoc et al. 2005), in addition to its well-characterized role in heterochromatic silencing. Finally, many of the same HPTMs occur in both transcription and DNA repair, which are mechanistically distinct processes. Based on some of these considerations, a more general hypothesis has been proposed where HPTMs serve as a nuclear DNA-associated signal transduction pathway, similar to cytoplasmic signal transduction that is generated and propagated largely through Ser/Thr phosphorylation (Schreiber and Bernstein 2002). In this model, there is not a strict histone code for specific processes, but rather HPTM recognition and binding via a plethora of proteinbinding motifs. This model explains how any site could be both activating and repressing and involved in more than one process, because different binding effector proteins are cognates for the same HPTM for distinct processes.

7 Themes in Modifications

7.2 Modification Patterns

The preceding discussion of the numerous types and sites of histone PTM occurring in transcription might lead to the conclusion that there are few overarching guiding principles or ideas. However, there do appear to be a number of broad themes that occur repeatedly, although the specifics may change depending on the histone, the sites of HPTM, and the binding proteins. Indeed, chromatin regulation may vary between promoters and distinct pathways.

Some experimental evidence points to the structural alteration of chromatin with certain HPTMs (Model 1). This can result from altering the charge of single or cluster of histone residues. This is particularly true when residues are acetylated or phosphorylated, which reduces the positive charge of histone regions (see Section 5 of Chapter 3). Such cis alterations can alter internucleosomal spacing and reduce the affinity of histones to negatively charged DNA, as exemplified by the. negative charged patches that occur on linker histone (Dou and Gorovsky 2002). These types of HPTMs may be cumulative in their effect on, for example, transcriptional activation or for creating higher-order chromatin structures, rather than producing a binary ON/OFF effect (Kurdistani et al. 2004; Henikoff 2005). Another model for the "output" of the myriad of HPTMs is that the code is complex and is read in patterns and often in temporal sequences. In this view, the intricacy of the patterns in three-dimensional space and over time during a process requiring many chromatin-associated steps, such as transcription, yields a meaningful mechanistic result. Two types of HPTM patterns have been identified. First, there are patterns on the same histone tail, or in "cis;' and second, patterns on different histone tails, or in "trans:' The most well-characterized cis pattern is between H3SIOph and H3KI4ac on the H3 amino-terminal tail (Cheung et al. 2000; Lo et al. 2000), where H3SIOph leads to H3KI4ac. The mechanism underlying the establishment of this pattern is understood in structural detail: The

7.1 Histone Code

One key question emerges after this lengthy discussion of the intricacies of HPTMs: Why are there so many modifications? Clearly, many of them correlate with transcription, and others occur during different DNA-templated processes. Thus, one hypothesis is that there is a histone "code," linking specific modifications with individual processes (Strahl and Allis 2000; Turner 2000). The simplest code would be a binary relationship between HPTMs and either gene activation or repression, and distinct HPTMs for other processes. The evidence supporting such a code is the observed tendency, as described above, for certain HPTMs to be positive-acting and others negative-acting. However, there are observations that are inconsistent with a simple binary code. For example, phosphorylation ofH3SlO is both activating to transcription, which presumably involves opening the chromatin, and involved in chromosome condensation, making chromatin even more inaccessible (Nowak and Corces

C H ROM A TIN

MOD I Fie A T ION SAN 0

enzyme that acetylates binds to the previously phosphorylated H3 tail with increased affinity due to a greater number of amino acid side-chain contact points (Clements et al. 2003). Other cis patterns are H3K23 acetylation and H3R17 methylation (Daujat et al. 2002) and H4R3 methylation and H4K8 acetylation (Wang et al. 2001). As described above, one trans tail pattern has been identified, where initial ubiquitylation of H2BK123 leads to methylation of H3K4 (Briggs et al. 2002; Dover et al. 2002; Sun and Allis 2002). The mechanism linking these HPTMs has not been elucidated, although several possible hypotheses exist. Because the link is from one large modification (ubiquitin) to a nonadjacent HTPM, one model is that ubiquitin wedges the chromatin open, like a crowbar, to allow the methylating enzyme access to its site. A second general model is that H2BK123ubi functions to recruit effector proteins, similar to the role of the other HPTMs. The noncatalytic portion of the proteosome requires H2B ubiquitylation for chromatin association (Ezhkova and Tansey 2004), and the function of the elongation complex FACT is stimulated by H2B ubiquitylation (Pavri et al. 2006), although neither complex has yet been shown to directly bind to ubiquitylated H2B.

THE I R M E C HAN ISM

205

F ACT ION

REPRESSED (DEFAULT) promoter

gene body

-~ DEREPRESSED I ACTIVATED

Model 2:

Detamar eviction

Model 3: Histone variant

exchange H2A

7.3 Changes in Chromatin 5tructure Associated with Transcription Activation and Elongation

The transcriptionally active euchromatic regions contain nucleosomes, but in an "unfolded" state, denoted as "beads-on-a-string" or II-nm fiber. The nucleosomes in this state still impose an intrinsic inhibition to the transcription machinery. Some transcription factors, be they activators or repressors, can gain access to their sites when contained in nucleosomes, but others cannot. Moreover, the machinery recruited by the DNA-bound regulators and responsible for delivering RNA pol II to promoters is constrained by the presence of nucleosomes. A number of distinct mechanisms serve to reconfigure the chromatin, poising genes for subsequent transcription, or promoting initiation or elongation. Some of these mechanisms are illustrated in Figure 8. The nucleosome problem during transcription is solved in part by the recruitment of protein complexes to mobilize and/or alter the structure of the nucleosome. These complexes fall into two different families, one represented by SNF2H (or ISWI and ISW2 in yeast), and one by the Brahma-Swi/Snf family (Narlikar et al. 2002; Peterson 2002; Flaus and Owen-Hughes 2004). The first family mobilizes nucleosomes, whereas Swi/Snf also transitorily alters the structure of the nucleosomes. Acetylated nucle-

H2AZ

wr1 H2AlH2B

H2AlH2B

e-Le-L

Model 4: FACT recruitment and H2NH2B dimer loss H2AZ

H2AZ

Figure 8. Models for the Involvement of Chromatin Remodeling and Histone Exchange in Transcriptional Processes In Modell, the Swi/Snf family of ATPase binds chromatin through bromodomain recognition of acetylated histones and acts to alter the local chromatin structure. Model 2 depicts the reported octamer eviction that occurs at certain loci such as PH05 by an unknown mechanism. In Model 3, the ATPase SWRl catalyzes the replacement of histone H2A with H2AZ, which poises chromatin for transcription. Model 4 focuses on the involvement of FACT in transcriptional elongation, assisting in nucleosome unraveling by the displacement of an H2A/H2B dimer. Concomitantly, histone H3 may be exchanged with H3.3 during the process.

osomes are recognized by the Swi/Snf complex through bromodomain interaction (Model 1 in Fig. 8). A second mechanism involved in gene activation is selective octamer loss at promoters. For example, histone octamers are evicted at the promoter of the PROS gene in S. cerevisiae during transcriptional induction (Model 2 in Fig. 8) (Boeger et al. 2003; Reinke and Horz 2003). In

206

•

C HAP T E RIO

addition, promoters of S. cerevisiae have a constitutively low density of nucleosomes, which allows access for transcription factors (Sekinger et al. 2005). It is not yet known whether or how ATP-dependent remodeling complexes assist in generating and maintaining this low occupancy. A third major mechanism involved in setting up transcriptional states is the presence of histone variants. There are two types of histone variants associated with gene activity. First, a variant of H2A called H2AZ is found in nucleosomes around the promoter gap, and poises the gene for activation (Santisteban et al. 2000; Raisner et al. 2005; Zhang et al. 2005); a specific ATP-dependent remodeling complex, called Swrl, replaces H2A with H2AZ (Model 3 in Fig. 8) (Mizuguchi et al. 2004; for more detail, see Chapter 13). Second, one H3 isoform, called H3.1, is incorporated into chromatin during replication, whereas isoform H3.3 is incorporated in a replication-independent manner (Ahmad and Henikoff 2002) with the aid of the HIRA (histone regulator A) chaperone. This variant is predominantly found within gene ORFs (Mito et al. 2005), suggesting that its deposition is a transcription-coupled process. There are additional mechanisms to overcome the nucleosomal barrier to elongating RNA pol II (and RNA pol I). A large number of factors have been isolated that affect transcription elongation (Sims et aI. 2004). One of these factors was found to allow the RNA pol II to traverse nucleosomes. This factor is known as FACT (for FAcilitate Chromatin Transcription). Importantly, FACT functions exclusively through nucleosomes, binds to them, and then promotes the displacement of one H2A/H2B dimer (Model 4 in Fig. 8) (Belotserkovskaya et al. 2003). As transcription ceases, FACT also promotes the reconstitution of the nucleosome. Interestingly, FACT performs its functions in the absence of energy, but physically interacts with CHDl, a protein that hydrolyzes ATP to mobilize nucleosomes and bind to the active H3K4me mark. Moreover, FACT also interacts with NuA4, a complex that contains HAT activity. Although FACT can promote displacement of the H2A/H2B dimer in vitro in an ATPindependent manner, it is possible that this is promoted by its interaction with factors such as CHDl, which mobilize or alter the structure of nucleosomes in vivo and also the interplay with HPTMs (Reinberg and Sims 2006). 8 Concluding Remarks

We have come a long way in this "modern era" of histone modifications which covers the last 10 years. In this time, there have been six distinct types of histone modification pathways characterized and numerous sites of modifications identified. Yet this is clearly still the beginning of our

understanding. Mechanistically we know that modifications affect the binding of proteins, but we are still unaware precisely how these proteins result in reorganization of chromatin structure. We still do not know whether there is a code or whether modifications are simply part of a signaling pathway. In addition, our knowledge is lacking on the many cellular processes, other than transcription, that modifications are involved in. Thus, in short, we have become aware of the complexity of the system, but we are a long way from making sense of the complexity. One thing is for sure, it is worth the effort to find out, since histone modifications playa fundamental role in both normal and diseased processes. References Ahmad K. and Henikoff S. 2002. The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9: 1191-1200. An W., Kim J., and Roeder R.G. 2004. Ordered cooperative functions of PRMTl, p300, and CARMI in transcriptional activation by p53. Cell1l7: 735-748. Baek S.H. and Rosenfeld M.G. 2004. Nuclear receptor coregulators: Their modification codes and regulatory mechanism by translocation. Biochem. Biophys. Res. Commun. 319: 707-714. Bannister A.J., Schneider R., and Kouzarides T 2002. Histone modification: Dynamic or static? Cell 109: 801-806. Bannister A.J., Zegerman P., Partridge J.P., Miska E.A., Thomas J.O., Allshire R.C, and Kouzarides T 2001. Selective recognition of methylated lysine 9 on histone H3 by the HPI chromo domain. Nature 410: 120-124. Bauer U.M., Daujat S., Nielsen S.J., Nightingale K., and Kouzarides T. 2002. Methylation at arginine 17 of histone H3 is linked to gene activation. EMBO Rep. 3: 39-44. Belotserkovskaya R., Oh S., Bondarenko V.A., Orphanides G., Studitsky V.M., and Reinberg D. 2003. FACT facilitates transcriptiondependent nucleosome alteration. Science 301: 1090-1093. Boeger H., Griesenbeck J., Strattan J.S., and Kornberg R.D. 2003. Nucleosomes unfold completely at a transcriptionally active promoter. Mol. Cellll: 1587-1598. Braunstein M., Rose A.B., Holmes S.G., Allis CD., and Broach J.R. 1993. Transcriptional silencing in yeast is associated with reduced nucleosome acetylation. Genes Dev. 7: 592-604. Briggs S.D., Xiao T, Sun Z.W., Caldwell J.A., Shabanowitz J., Hunt D.P., Allis CD., and Strahl B.D. 2002. Gene silencing: Trans-histone regulatory pathway in chromatin. Nature 418: 498. Brownell J.E., Zhou J., Ranalli T, Kobayashi R., Edmondson D.G., Roth S.Y, and Allis CD. 1996. Tetrahymena histone acetyltransferase A: A hom*olog to yeast Gcn5p linking histone acetylation to gene activation. Cell 84: 843-851. Carrozza M.J., Li B., Florens L., Suganuma T, Swanson S.K., Lee K.K., Shia W.J., Anderson S., Yates J., Washburn M.P., and Workman J.L. 2005. Histone H3 methylation by Set2 directs deacetylation of coding regions by Rpd3S to suppress spurious intragenic transcription. Cell 123: 581-592. Chen D., Ma H., Hong H., Koh S.S., Huang S.M., Schurter B.T, Aswad D.W., and Stallcup M.R. 1999. Regulation of transcription by a protein methyltransferase. Science 284: 2174-2177. Chen Z., Zang J., Whetstine J., Hong X., Davrazou P., Kutateladze TG.,

C H ROM A TIN

MOD I Fie A T ION SAN 0

Simpson M., Mao Q., Pan CH., Dai S., et al. 2006. Structural insights into histone demethylation by JMJD2 family members. Cell 125: 691-702. Cheung P., Tanner K.G., Cheung W.L., Sassone-Corsi P., Denu J.M., and Allis CD. 2000. Synergistic coupling of histone H3 phosphorylation and acetylation in response to epidermal growth factor stimulation. Mol. CellS: 905-915. Clark-Adams CD., Norris 0., Osley M.A., Fassler J.S., and Winston E 1988. Changes in histone gene dosage alter transcription in yeast. Genes Dev. 2: 150-159. Clements A., Poux A.N., Lo W.S., Pill us L., Berger S.L., and Marmorstein R. 2003. Structural basis for histone and phosphohistone binding by the GCN5 histone acetyltransferase. Mol. Cell 12: 461--473. Cloos P.A., Christensen J., Agger K., Maiolica A., Rappsilber J., Antal T., Hansen K.H., and Helin K. 2006. The putative oncogene GASCI demethylates tri- and dimethylated lysine 9 on histone H3. Nature 442: 307-311. Cuthbert G.L., Daujat S., Snowden A.W., Erdjument-Bromage H., Hagiwara T, Yamada M., Schneider R., Gregory P.D., Tempst P., Bannister A.J., and Kouzarides T 2004. Histone deimination antagonizes arginine methylation. CellllS: 545-553. Daniel J.A., Torok M.S., Sun Z.W, Schieltz D., Allis CD., Yates J.R., Ill, and Grant P.A. 2004. Deubiquitination of histone H2B by a yeast acetyltransferase complex regulates transcription. ]. BioI. Chern. 279: 1867-1871. Daujat S., Bauer U.M., Shah v., Turner B., Berger S., and Kouzarides T 2002. Crosstalk between CARMI methylation and CBP acetylation on histone H3. Curro BioI. 12: 2090-2097. Dhalluin C, Carlson J.E., Zeng L., He C, Aggarwal A.K., and Zhou M.M. 1999. Structure and ligand of a histone acetyl transferase bromodomain. Nature 399: 491--496. Dou Y. and Gorovsky M.A. 2002. Regulation of transcription by HI phosphorylation in Tetrahymena is position independent and requires clustered sites. Proc. Natl. Acad. Sci. 99: 6142-6146. Dover J., Schneider J., Tawiah-Boateng M.A., Wood A., Dean K., Johnston M., and Shilatifard A. 2002. Methylation of histone H3 by COMPASS requires ubiquitination of histone H2B by Rad6.]. BioI. Chern. 277: 28368-28371. Durrin L.K., Mann R.K., Kayne P.S., and Grunstein M. 1991. Yeast histone H4 N-terminal sequence is required for promoter activation in vivo. Cell 65: 1023-1031. Emre N.C, Ingvarsdottir K., Wyce A., Wood A., Krogan N.J., Henry K.W., Li K., Marmorstein R., Greenblatt J.E, Shilatifard A., and Berger S.L. 2005. Maintenance of low histone ubiquitylation by Ubpl0 correlates with telomere-proximal Sir2 association and gene silencing. Mol. CellI?: 585-594. Ezhkova E. and Tansey W.P. 2004. Proteasomal ATPases link ubiquitylation of histone H2B to methylation of histone H3. Mol. Cell 13: 435--442. Fabbrizio E., El Messaoudi S., Polanowska J., Paul C, Cook J.R., Lee J.H., Negre v., Rousset M., Pestka S., Le Cam A., and Sardet C 2002. Negative regulation of transcription by the type II arginine methyltransferase PRMT5. EMBO Rep. 3: 641-645. Fischle W., Tseng B.S., Dormann H.L., Ueberheide B.M., Garcia B.A., Shabanowitz J., Hunt D.E, Funabiki H., and Allis CD. 2005. Regulation of HP I-chromatin binding by histone H3 methylation and phosphorylation. Nature 438: 1116--1122. Flaus A. and Owen-Hughes T 2004. Mechanisms for ATP-dependent chromatin remodelling: Farewell to the tuna-can octamer? Curro Opin. Genet. Dev. 14: 165-173. Fodor B.D., Kubicek S., Yonezawa M., O'Sullivan R.J., Sengupta R.,

THE I R M E C HAN ISM

0 F ACT ION

•

207

Perez-Burgos L., Opravil S., MechtJer K., Schotta G., and Jenuwein T 2006. Jmjd2b antagonizes H3K9 trimethylation at pericentric heterochromatin in mammalian cells. Genes Dev. 20: 1557-1562. Gardner R.G., Nelson Z.W, and Gottschling D.E. 2005. Ubpl0/Dot4p regulates the persistence of ubiquitinated histone H2B: Distinct roles in telomeric silencing and general chromatin. Mol. Cell. Bioi. 25: 6123-6139. Glozak M.A., Sengupta N., Zhang X., and Seto E. 2005. Acetylation and deacetylation of non-histone proteins. Gene 363: 15-23. Grant P.A., Sterner D.E., Duggan L.J., Workman J.L., and Berger S.L. 1998. The SAGA unfolds: Convergence of transcription regulators in chromatin-modifying complexes. Trends Cell BioI. 8: 193-197. Hall I.M., Shankaranarayana G.D., Noma K., Ayoub N., Cohen A., and Grewal S.l. 2002. Establishment and maintenance of a heterochromatin domain. Science 297: 2232-2237. Hampsey M. and Reinberg D. 2003. Tails of intrigue: Phosphorylation of RNA polymerase II mediates histone methylation. Cell 113: 429--432. Han M. and Grunstein M. 1988. Nucleosome Joss activates yeast downstream promoters in vivo. Cell 55: 1137-1145. Hassan A.H., Prochasson P., Neely K.E., Galasinski S.C, Chandy M., Carrozza M.J., and Workman J.L. 2002. Function and selectivity of bromodomains in anchoring chromatin-modifying complexes to promoter nucleosomes. Celllll: 369-379. Hebbes TR., Clayton A.L., Thorne A.W., and Crane-Robinson C 1994. Core histone hyperacetylation co-maps with generalized DNase I sensitivity in the chicken beta-globin chromosomal domain. EMBO]. 13: 1823-1830. Henikoff S. 2005. Histone modifications: Combinatorial complexity or cumulative simplicity? Proc. Natl. Acad. Sci. 102: 5308-5309. Henry K.W, Wyce A., Lo WS., Duggan L.J., Emre N.C, Kao CE, Pilius L., Shilatifard A., Osley M.A., and Berger S.L. 2003. Transcriptional activation via sequential histone H2B ubiquitylation and deubiquitylation, mediated by SAGA-associated Ubp8. Genes Dev. 17: 2648-2663. Hirota T, Lipp J.J., Toh B.H., and Peters J.M. 2005. Histone H3 serine 10 phosphorylation by Aurora B causes HPI dissociation from heterochromatin. Nature 438: 1176-1180. Jenuwein T and Allis Co. 2001. Translating the histone code. Science 293: 1074-1080. Joshi A.A. and Struhl K. 2005. EaD chromodomain interaction with methylated H3-K36 links histone deacetylation to Pol II elongation. Mol. Cell 20: 971-978. Kao C.E, Hillyer C., Tsukuda T, Henry K., Berger S., and Osley M.A. 2004. Rad6 plays a role in transcriptional activation through ubiquitylation of histone H2B. Genes Dev. 18: 184-195. Keogh M.C, Kurdistani S.K., Morris S.A., Ahn S.H., Podolny v., Collins S.R., Schuldiner M., Chin K., Punna T, Thompson N.J., et al. 2005. Cotranscriptional set2 methylation of histone H3 lysine 36 recruits a repressive Rpd3 complex. Cell 123: 593-605. Kim J., Hake S.B., and Roeder R.G. 2005. The human hom*olog of yeast BREI functions as a transcriptional coactivator through direct activator interactions. Mol. Cell 20: 759-770. Klose R.J., Yamane K., Bae Y., Zhang D., Erdjument-Bromage H., Tempst P., Wong J., and Zhang Y. 2006. The transcriptional repressor JHDM3A demethylates trimethyl histone H3 lysine 9 and lysine 36. Nature 442: 312-316. Kurdistani S.K. and Grunstein M. 2003. Histone acetylation and deacetylation in yeast. Nat. Rev. Mol. Cell. BioI. 4: 276--284. Kurdistani S.K., Tavazoie S., and Grunstein M. 2004. Mapping global histone acetylation patterns to gene expression. Ce1l1l7: 721-733. Lachner M., O'Carroll D., Rea S., MechtJer K., and Jenuwein T 2001.

208

C HAP T E R

Methylation of histone H3 lysine 9 creates a binding site for HPI proteins. Nature 410: 116-120. Lee D.Y, Teyssier C, Strahl B.D., and Stallcup M.R. 2005. Role of protein methylation in regulation of transcription. Endocr. Rev. 26: 147-170. Lee M.G., Wynder C, Cooch N., and Shiekhattar R. 2005. An essential role for CoREST in nucleosomal histone 3 lysine 4 demethylation. Nature 437: 432-435. Li H., llin S., Wang W, Duncan E.M., Wysocka )., Allis CD., and Patel D.). 2006. Molecular basis for site-specific read-out of histone H3K4me3 by the BPTF PHD finger of NURE Nature 442: 91-95. Litt M.D., Simpson M., Gaszner M., Allis CO., and Felsenfeld G. 2001. Correlation between histone lysine methylation and developmental changes at the chicken beta-globin locus. Science 293: 2453-2455. Lo W.S., Duggan L., Emre N.C, Belotserkovskya R., Lane W.S., Shiekhattar R., and Berger S.L. 2001. Snfl-A histone kinase that works in concert with the histone acetyl transferase GenS to regulate transcription. Science 293: 1142-1146. Lo WS., Trievel R.C, Rojas ).R., Duggan L., Hsu ).Y, Allis CO., Marmorstein R., and Berger S.L. 2000. Phosphorylation of serine 10 in histone H3 is functionally linked in vitro and in vivo to Gcn5mediated acetylation at lysine 14. Mol. Cell 5: 917-926. Ma H., Baumann CT, Li H., Strahl B.D., Rice R., Jelinek M.A., Aswad D.W., Allis CD., Hager G.L., and Stallcup M.R. 2001. Hormonedependent, CARM I-directed, arginine-specific methylation of histone H3 on a steroid-regulated promoter. Curro BioI. 11: 1981-1985. Macdonald N., Welburn ).P., Noble M.E., Nguyen A., Yaffe M.B., Clynes D., Moggs ).G., Orphanides G., Thomson S., Edmunds ).W, et al. 2005. Molecular basis for the recognition of phosphorylated and phosphoacetylated histone H3 by 14-3-3. Mol. Cell 20: 199-211. Mahadevan L.C, Willis A.C, and Barratt M.). 1991. Rapid histone H3 phosphorylation in response to growth factors, phorbol esters, okadaic acid, and protein synthesis inhibitors. Cell 65: 775-783. Martin C and Zhang Y. 2005. The diverse functions of histone lysine methylation. Nat. Rev. Mol. Cell. BioI. 6: 838-849. Metivier R., Penot G., Hubner M.R., Reid G., Brand H., Kos M., and Gannon E 2003. Estrogen receptor-a directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 115: 751-763. Metzger E., Wissmann M., Yin N., Muller j.M., Schneider R., Peters A.H., Gunther T., Buettner R., and Schule R. 2005. LSDI demethylates repressive histone marks to promote androgen-receptordependent transcription. Nature 437: 436-439. Mito Y, Henikoff ).G., and Henikoff S. 2005. Genome-scale profiling of histone H3.3 replacement patterns. Nat. Genet. 37: 1090-1097. Mizuguchi G., Shen X., Landry)., Wu W.H., Sen S., and Wu C 2004. ATP-driven exchange of histone H2AZ variant catalyzed by SWRI chromatin remodeling complex. Science 303: 343-348. Nakayama )., Rice ).C, Strahl B.D., Allis CO., and Grewal S.l. 2001. Role of histone H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science 292: 11 0-113. Narlikar G.)., Fan H.Y., and Kingston R.E. 2002. Cooperation between complexes that regulate chromatin structure and transcription. Cell 108: 475-487. Noma K., Allis CD., and Grewal S.l. 2001. Transitions in distinct histone H3 methylation patterns at the heterochromatin domain boundaries. Science 293: 1150-1155. Nowak S.). and Corces v.G. 2004. Phosphorylation of histone H3: A balancing act between chromosome condensation and transcriptional activation. Trends Genet. 20: 214-220.

Pal S., Yun R., Datta A., Lacomis L., Erdjument-Bromage H., Kumar )., Tempst P., and Sif S. 2003. mSin3A/histone deacetylase 2- and PRMT5-containing Brgl complex is involved in transcriptional repression of the Myc target gene cad. Mol. Cell. BioI. 21: 7475-7487. Pavri R., Zhu B., Li G., Trojer E, Mandal S., Shilatifard A., and Reinberg D. 2006. Histone H2B monoubiquitination functions cooperatively with FACT to regulate elongation by RNA polymerase II. Cell 125: 703-717. Peterson CL. 2002. Chromatin remodeling: Nucleosomes bulging at the seams. Curro BioI. 12: R245-R247. Raisner R.M., Hartley P.D., Meneghini M.D., Bao M.Z., Liu CL., Schreiber S.L., Rando 0.)., and Madhani H.D. 2005. Histone variant H2A.Z marks the 5' ends of both active and inactive genes in euchromatin. Cell 123: 233-248. Rea S., Eisenhaber E, O'Carroll D., Strahl B.D., Sun Z.W., Schmid M., Opravil S., Mechtler K., Ponting CP., Allis CD., and )enuwein T 2000. Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406: 593-599. Reinberg D. and Sims R.)., III. 2006. de FACTo nucleosome dynamics. /. BioI. Chem.(in press). Reinke H. and Horz W. 2003. Histones are first hyperacetylated and then lose contact with the activated PHOS promoter. Mol. Cell 11: 1599-1607. Robzyk K., Recht )., and Osley M.A. 2000. Rad6-dependent ubiquitination of histone H2B in yeast. Science 287: 501-504. Roth S.Y, Denu ).M., and Allis CD. 2001. Histone acetyltransferases. Annu. Rev. Biochem. 70: 81-120. Sanders S.L., Jennings )., Canutescu A., Link A.)., and Weil P.A. 2002. Proteomics of the eukaryotic transcription machinery: Identification of proteins associated with components of yeast TFIID by multidimensional mass spectrometry. Mol. Cell. BioI. 22: 4723-4738. Santisteban M.S., Kalashnikova T, and Smith M.M. 2000. Histone H2A.Z regulates transcription and is partially redundant with nucleosome remodeling complexes. Cell 103: 411-422. Santos-Rosa H., Schneider R., Bannister A.)., Sherriff )., Bernstein B.E., Emre N.C, Schreiber S.L., Mellor )., and Kouzarides T 2002. Active genes are tri-methylated at K4 of histone H3. Nature 419: 407-411. Sassone-Corsi P., Mizzen CA., Cheung P., Crosio C, Monaco L., Jacquot S., Hanauer A., and Allis CD. 1999. Requirement of Rsk-2 for epidermal growth factor-activated phosphorylation of histone H3. Science 285: 886-891. Schreiber S.L. and Bernstein B.E. 2002. Signaling network model of chromatin. Cell 111: 771-778. Schurter B.T, Koh S.S., Chen D., Bunick G.)., Harp ).M., Hanson RL., Henschen-Edman A., Mackay D.R., Stallcup M.R., and Aswad D.W. 2001. Methylation of histone H3 by coactivator-associated arginine methyltransferase 1. Biochemistry 40: 5747-5756. Sekinger E.A., Moqtaderi Z., and Struhl K. 2005. Intrinsic histone- DNA interactions and low nucleosome density are important for preferential accessibility of promoter regions in yeast. Mol. Cell 18: 735-748. Shi Y., Lan E, Matson C, Mulligan E, Whetstine ).R., Cole EA., Casero R.A., and Shi Y. 2004. Histone demethylation mediated by the nuclear amine oxidase hom*olog LSD1. Cell 119: 941-953. Shi Y.)., Matson C, Lan E, Iwase S., Baba T, and Shi Y 2005. Regulation of LSD 1 histone demethylase activity by its associated factors. Mol. Cell. 19: 857-864. Shiio Y and Eisenman R.N. 2003. Histone sumoylation is associated with transcriptional repression. Proc. Natl. Acad. Sci. 100: 13225-13230.

C H ROM A TIN

MOD I Fie A T ION 5

Shim E.Y, Woodco*ck C, and Zaret K.S. 1998. Nucleosome positioning by the winged helix transcription factor HNF3. Genes Dey. 12: 5-10. Shogren-Knaak M., Ishii H., Sun T.M., Pazin M.J., Davie T.R., and Peterson CL. 2006. Histone H4-K16 acetylation controls chromatin structure and protein interactions. Science 311: 844-847. Sims. R.T., Belotserkovskaya R., and Reinberg D. 2004. Elongation by RNA polymerase II: The short and long of it. Genes Dev. 18: 2437-2468. Sims R.T., Chen CE, Santos-Rosa H., Kouzarides T., Patel 5.5., and Reinberg D. 2006. Human but not yeast CHD 1 binds directly and selectively to histone H3 methylated at lysine 4 via its tandem chromodomains.]. BioI. Chern 51: 41789-41792. Soloaga A., Thomson 5., Wiggin G.R., Rampersaud N., Dyson M.H., Hazzalin CA., Mahadevan L.C, and Arthur T.S. 2003. MSK2 and MSKI mediate the mitogen- and stress-induced phosphorylation of histone H3 and HMG-14. EMBO]. 22: 2788-2797. Sterner D.E. and Berger S.L. 2000. Acetylation of histones and transcription-related factors. Microbial. Mol. BioI. Rev. 64: 435-459. Strahl B.D. and Allis CD. 2000. The language of covalent histone modifications. Nature 403: 41-45. Strahl B.D., Briggs S.D., Brame CT., Caldwell T.A., Koh 5.5., Ma H., Cook R.G., Shabanowitz J., Hunt D.E, Stallcup M.R., and Allis CD. 2001. Methylation of histone H4 at arginine 3 occurs in vivo and is mediated by the nuclear receptor coactivator PRMT1. Curro BioI. 26: 996-1000. Sun Z.W. and Allis CD. 2002. Ubiquitination of histone H2B regulates H3 methylation and gene silencing in yeast. Nature 418: 104-108. Svaren J., Schmitz T., and Horz W. 1994. The transactivation domain of Ph04 is required for nucleosome disruption at the PH05 promoter. EMBO]. 13: 4856-4862. Taunton T., Hassig CA., and Schreiber S.L. 1996. A mammalian histone deacetylase related to the yeast transcriptional regulator Rpd3p. Science 272: 408-411. Trewick S.C, McLaughlin P.T., and Allshire R.C 2005. Methylation: Lost in hydroxylation? EMBO Rep. 6: 315-320. Tsukada Y., Fang ]., Erdjument-Bromage H., Warren M.E., Borchers CH., Tempst P., and Zhang Y 2006. Histone demethylation by a family of TmjC domain-containing proteins. Nature 439: 811-816. Turner B.M. 2000. Histone acetylation and an epigenetic code. Bioessays 22: 836-45. Vakoc CR., Mandat S.A., Olenchock B.A., and Blobel G.A. 2005. Histone H3 lysine 9 methylation and HPly are associated with transcription elongation through mammalian chromatin. Mol. Cell 19: 381-391. Vettese-Dadey M., Grant P.A., Hebbes T.R., Crane-Robinson C, Allis CO., and Workman J,L. 1996. Acetylation of histone H4 plays a primary role in enhancing transcription factor binding to nucleo-

AND

THE I R

M E C HAN ISM

0 F ACT ION

•

209

somal DNA in vitro. EMBO]. 15: 2508-2518. Volpe T.A., Kidner C, Hall I.M., Teng G., Grewal 5.1., and Martienssen R.A. 2002. Regulation of heterochromatic silencing and histone H3 lysine-9 methylation by RNAi. Science 297: 1833-1837. Wang H., Wang L., Erdjument-Bromage H., Vidal M., Tempst P., Tones R.S., and Zhang Y. 2004. Role of histone H2A ubiquitination in Polycomb silencing. Nature 431: 873-878. Wang H., Huang Z.Q., Xia L., Feng Q., Erdjument-Bromage H., StraW B.D., Briggs S.D., Allis CD., Wong T., Tempst P., and Zhang Y. 2001. Methylation of histone H4 at arginine 3 facilitating transcriptional activation by nuclear hormone receptor. Science 293: 853-857. Wang Y, Wysocka T., Sayegh T., Lee YH., Perlin T.R., Leonelli L., Sonbuchner L.S., McDonald CH., Cook R.G., Dou Y., et al. 2004. Human PAD4 regulates histone arginine methylation levels via demethylimination. Science 306: 279-283. Whetstine T.R., Nottke A., Lan E, Huarte M., Smolikov 5., Chen Z., Spooner E., Li E., Zhang G., Colaiacovo M., and Shi Y 2006. Reversal of histone lysine trimethylation by the TMTD2 family of histone demethylases. Cel/125: 467-481. Wood A., Krogan N.T., Dover J., Schneider T., Heidt T., Boateng M.A., Dean K., Golshani A., Zhang Y., Greenblatt T.E, et al. 2003. BreI, an E3 ubiquitin ligase required for recruitment and substrate selection of Rad6 at a promoter. Mol. Cel/11: 267-274. Workman T.L. and Roeder R.G. 1987. Binding of transcription factor TFIID to the major late promoter during in vitro nucleosome assembly potentiates subsequent initiation by RNA polymerase II. Cel/51: 613-622. Wysocka J., Allis CD., and Coonrod S. 2006a. Histone arginine methylation and its dynamic regulation. Front. Biosci. 11: 344-355. Yamane K., Toumazou C, Tsukada Y, Erdjument-Bromage H., Tempst P., Wong T., and Zhang Y 2006. THDM2A, a ]mjC-containing H3K9 demethylase, facilitates transcription activation by androgen receptor. CeLL 125: 483-495. Yang X.T. and Seto E. 2003. Collaborative spirit of histone deacetylases in regulating chromatin structure and gene expression. Curro Opin. Genet. Dey. 13: 143-153. Zhang H., Roberts D.N., and Cairns B.R. 2005. Genome-wide dynamics of Htzl, a histone H2A variant that poises repressed/basal promoters for activation through histone loss. CeLL 123: 219-231. Zhang Y and Reinberg D. 2001. Transcription regulation by histone methylation: Interplay between different covalent modifications of the core histone tails. Genes Dey. 15: 2343-2360. Zhu B., Zheng Y, Pham A.D., MandaI 5.5., Erdjument-Bromage H., Tempst P., and Reinberg D. 2005. Monoubiquitination of human histone H2B: The factors involved and their roles in HOX gene regulation. Mol. Cel/20: 601-611.

Transcriptional Silencing by Polycomb Group Proteins Ueli Grossniklaus 1 and Renata Paro 2 IInstitute of Plant Biology and Zurich-Basel Plant Science Center, University of Zurich, CH-8008 Zurich, Switzerland 2ZMBH, University of Heidelberg, D-69120 Heidelberg, Germany

CONTENTS 1. Introduction, 213

3.2

Targeting PRCl to Silenced Genes, 222

1. 1

The Concept of Cellular Memory, 213

3.3

1.2

Genetic Identification of the Polycomb Group, 213

Establishment of Repressive Functions by PRCl,223

3.4

Preventing Heritable Repression by Anti-silencing, 224

2. Establishing Silencing Marks on Chromatin, 215 2.1 2.2 2.3

Components and Evolutionary Conservation of PRC2, 215

4.1

Chromatin-modifying Activity of PRC2, 219

From Gene to Chromosome Repression, 224

4.2

Dynamic Function of PRO during Development, 21 9

Consequences of Aberrant Transcriptional Activation 225

4.3

Maintaining Stem Cell Fate, 226

3. Maintaining Transcriptional Silencing, 220 3.1

4. PcG Repression in Mammalian Development, 224

Components of PRCl, 220

5. Conclusion and Outlook, 227 References, 228

211

GENERAL SUMMARY The organs of humans, animals, and plants are constructed from a large pool of distinct cell types, each performing a specialized physiological or structural function. With very few exceptions, all cell types contain the same genetic information encoded in their DNA. Thus, the distinctiveness of a given cell type is achieved through specific gene expression programs. However, cell lineages need to have these programs of gene expression maintained during growth and cell division. This implies the existence of a memory system that ensures a faithful transmission of information for which gene has to be active or repressed from mother to daughter cells. The existence of such a system is illustrated by the fact that cultured tissues of plants and animals usually maintain their differentiated characters even if grown in a foreign environment. By way of example, ivy plants regenerated after tissue culture produce the type of leaf corresponding to the phase of development from which the original tissue was taken (i.e., juvenile or adult leaf). The major question to be addressed in this and the following chapter concerns the molecular identity of factors contributing to the mechanism(s) which maintains determined states over many cell divisions (a process termed "cellular memory" or "transcriptional memory"). Genetic analyses in Drosophila have identified regulators crucial in maintaining the fate of individual body segments that are determined by the action of the HOX genes. In Drosophila males, the first thoracic segment has legs with sex combs. Legs on the second and third thoracic segment lack these structures (see the left panel of the title images). In the 1940s, Drosophila mutants were identified (Polycomb and extra sex combs) where males had sex combs on all legs (see the right panel of the title images). They correspond to homeotic transformations

of the second and third leg identities into the first leg identity. Genetic and molecular studies showed that these mutations did not affect the products of the HOX genes themselves, but rather the way HOX gene activity was spatially controlled. Over the years, a large number of such regulatory genes were identified, which could be classified into two antagonistic groups, the Polycomb (PcG) and trithorax (trxG) groups. Whereas the PcG proteins are required to maintain the silenced state of developmental regulators such as the HOX genes, the trxG proteins are generally involved in maintaining the active state of gene expression. Thus, PcG and trxG proteins form the molecular basis for cellular memory. Proteins of the PcG and trxG are organized into large multimeric complexes that act on their target genes by modulating chromatin structure. In this chapter, we focus on the molecular nature and function of two major Polycomb Repressive Complexes, PRC1 and PRC2; the molecular nature of the trxG complexes is described in the next chapter. PcG complexes are recruited to target genes through a DNA sequence called a PcG Response Element (PRE) in Drosophila. Once recruited, they establish a silent chromatin state that can be inherited over many cell divisions. Members of PRC2 are conserved between plants and animals, whereas PRC1 proteins are only present in Drosophila and vertebrates. This implies conservation but also diversity in the basic building blocks of the cellular memory system. In addition to their function in the maintenance of cell types, PcG complexes may also play important roles in stem cell plasticity. Their deregulation can lead to neoplastic transformation and cancer in vertebrates. Thus, PcG proteins are crucial for many fundamental processes of normal development and disease in multicellular eukaryotes.

214

C HAP T E R

/ /

Figure 2. Homeotic Transformations in PeG Mutants of Various Species (a-d) Drosophila melanogaster, (e, f) Mus musculus, (g, h) Arabidopsis thaliana. (a, b) Leg imaginal discs undergoing a transdetermination event as indicated by the expression of the wing-specific gene vestigial (which is marked by GFP). (c, d) Cuticles of a wild-type (c) and a Su(z)/2 mutant embryo (d). In the Su(z)/2 mutant embryo, all abdominal, thoracic, and several

head segments (not all visible in this focal plane) are homeotically transformed into copies of the eighth abdominal segment due to misexpression of the Abd-B gene in every segment. (e, f) Axial skeleton of newborn wild-type (e) and Ring/A-/mice (f). Views of the thoracic regions of cleared skeletons showing bone (red) and cartilage (blue). The mutant displays anterior transformation of the eighth thoracic vertebra as indicated by the presence of eight (1-8) vertebrosternal ribs, instead of seven (1-7) as in the wild type. (g, h) Wild-type (g) and c1f-2 mutant (h) flowers. The wild-type flower shows the normal arrangement of sepals, petals, stamens, and carpels. In the c1f-2 flower, petals are absent or reduced in number. (a,b, Courtesy of N. Lee and R. Paro; c,d, reprinted, with permission, from Birve et al. 2001 [©Companyof Biologists Ltd.]; e,f, reprinted, with permission, from del Mar Lorente et al. 2000 [©Company of Biologists Ltd.]; h, courtesy of J. Goodrich.)

Drosophila, the actIvIty of maternally (i.e., inherited through the oocyte) and zygotically produced transcription factors generates a specific combination of HOX expression required for each segment. This segmentspecific profile of HOX gene activity is maintained throughout the development of the fly, long after the early transcriptional regulators have disappeared. When the function of HOX genes was genetically characterized, many trans-acting regulators were isolated. Among them, Polycomb (Pc) was identified and genetically analyzed by Pam and Ed Lewis (Lewis 1978). Heterozygous Pc mutant males have additional sex combs on the second and third legs, whereas wild-type males only carry sex combs on the first leg (see title figure). hom*ozygous mutants are embryonic lethal, exhibiting a transformation of all cuticular segments toward the most posterior abdominal segment (Fig. 2c,d). These classic PcG phenotypes were interpreted as being caused by ectopic expression of HOX genes. Thus, Pc and the other genes with similar phenotypes were defined as repressors of HOX gene activity. Detailed analyses subsequently uncovered the fact that the PcG proteins are only required for the maintenance of HOX repression, rather than the position-specific establishment of HOX activity

during pattern formation. This latter task is performed by the transcription factors encoded by the early acting segmentation genes. Based on their repressing or activating effect on HOX gene expression, these newly identified trans-acting regulators were divided into two antagonistic classes, the PcG and trxG, respectively (Fig. 1) (Kennison 1995). The molecular isolation of Drosophila PcG genes has made it possible to study the function of vertebrate orthologs in mice, where they were also found to be key regulators of HOX gene expression (van der Lugt et al. 1994; Core et al. 1997). In mammals, mutations in PcG genes lead to homeotic transformations of the vertebrae (Fig. 2e,f). In addition, PcG genes playa crucial role in the control of cell proliferation, stem cell maintenance, and cancer (see Sections 4.2 and 4.3). The remarkable conservation ofPcG genes between flies and mammals has facilitated biochemical analyses and led to the identification of some novel members of PcG complexes, e.g., the RING 1 protein (Satijn and Otte 1999). Targeted mutation of RlNG1a in the mouse, for instance, led to the classic homeotic transformation phenotype. Only subsequently was it found to correspond to the PcG gene Sex combs extra in Drosophila.

T RAN S C RIP T ION A LSI LEN C I N G

In two other model organisms, namely the worm Caenorhabditis elegans and the flowering plant Arabidopsis thaliana, the molecular characterization of mutants isolated in various genetic screens revealed the existence of other PcG protein orthologs. In C. elegans, PcG members were identified in screens for maternal effect sterile (mes) mutants and were shown to be involved in X-chromosome silencing in the hermaphrodite germ line (Fong et a1. 2002; see Chapter 15). In Arabidopsis, PcG genes were identified in several genetic screens investigating distinct developmental processes (Hsieh et a1. 2003). The first PcG gene in plants, CURLY LEAF (CLF) , was identified as a mutant with homeotic transformations of floral organs (Fig. 2g,h) (Goodrich et a1. 1997). Mutations in the FERTILIZATIONINDEPENDENT SEED (PIS) class of genes were found in screens for mutants showing maternal-effect seed abortion (Grossniklaus et a1. 1998), or allowing aspects of seed development to occur in the absence of fertilization (Luo et a1. 1999; Ohad et a1. 1999). Finally, PcG genes were identified in screens for flowering time mutants, e.g., mutants that flower directly after germination (Yoshida et a1. 2001) or that disrupt the vernalization response, i.e., the process rendering plants competent to flower after prolonged exposure to cold (Gendall et a1. 2001). The variety of processes regulated by PcG proteins illustrates the importance of maintaining the repressed state of key developmental regulators in different organisms. On the one hand, there is an amazing conservation of some biological functions from plants to mammals, e.g., the regulation of key developmental regulators such as homeotic genes or involvement in the tight regulation of cell proliferation. On the other hand, PcG complexes appear to be versatile and dynamic molecular modules that have been employed to control a large variety of developmental and cellular processes. 2 Establishing Silencing Marks on Chromatin

PcG proteins fall into two biochemically characterized classes, which form the Polycomb repressive complexes 1 and 2 (PRCI and PRC2). The two complexes are required for consecutive steps in the repression of gene expression. First, PRC2 has histone-modifying activity and methylates H3K27 and/or H3K9 at genes targeted for silencing. PRCI components can then recognize and bind to such modifications and induce appropriate structural changes in chromatin. Whereas PRC2 proteins are present in all multicellular model species, PRCI components have not been identified in C. elegans and Arabidopsis.

B Y POL

yeo

M B G R0 U P

PRO TEl N S

•

215

2.1 Components and Evolutionary Conservation of PRC2

Several variants of PRC2 have been purified from Drosophila embryos, but all of these complexes contain four core proteins (Levine et a1. 2004): the SET histonemethyltransferase Enhancer of Zeste (E(Z)), the WD40 protein ESC, the histone-binding protein p55, and Suppressor of Zestel2 (SU(Z)12) (Table 1 and Fig. 3). Based on this composition, PRC2 was originally referred to as the E(Z)-ESC complex. This section highlights the molecular and biochemical details known about the different PRC2 components identified to date in different model organisms. The E(z) gene encodes a 760-amino acid protein, containing a SET domain that confers histone lysine methyltransferase (HKMT) activity. The SET domain is preceded by a CXC or Pre-SET domain (Tschiersch et a1. 1994), which contains nine conserved cysteines that bind three zinc ions and is thought to stabilize the SET domain. Such a structural role is supported by the fact that several temperature-sensitive E(z) alleles affect one of the conserved cysteines (Carrington and Jones 1996). In addition, E(z) contains SANT domains implicated in histone binding, and a C5 domain required for the physical interaction with SU(Z)12. ESC is a short protein of 425 amino acids that contains five WD40 repeats, shown to form a ~ propeller structure. This serves as a platform for protein-protein interactions, hence giving ESC a central role in PRC2, to physically interact with both E(z) and p55 in all model systems analyzed. The SU(Z)12 protein is 900 amino acids long and characterized by a C2 H 2 -type zinc finger and a carboxyterminal VEFS domain. The VEFS domain was identified as a conserved region between SU(Z) 12 and its three hom*ologs in plants: VRN2, EMF2, FIS2 (see Fig. 3). Several mutant Su(z)12 alleles alter this domain, showing it is required for the interaction with the C5 domain of E(Z) (Chanvivattana et a1. 2004; Yamamoto et a1. 2004). The p55 protein was not identified as a PcG member by genetic approaches, possibly because it takes part in a multitude of other protein complexes associated with chromatin (Hennig et a1. 2005). The p55 protein was, however, identified biochemically as part of PRC2. It is 430 amino acids long and contains six WD40 repeats, which physically interact with ESC or its orthologs in mammals and plants (Tie et a1. 2001; Kohler et a1. 2003a). In addition to the core PRC2 proteins, some variants of the complex contain the RPD3 histone deacetylase

216 •

C HAP T E R 1 1

(HDAC) or the Polycomb-like (PCL) protein. The interaction with RPD3 is noteworthy, because histone deacetylation is correlated with a repressed state of gene expression (see Chapter 10). The different compositions of PRC2 likely reflect dynamic changes during development or tissue-specific variants. PRC2 is highly conserved in invertebrates, vertebrates, and plants (Fig. 3). In C. elegans, only hom*ologs of E(Z) and ESC are present: MES-2 and MES-6. Together with another nonconserved protein, MES-3, they form a small complex of about 230 kD required for X-chromosome silencing in the hermaphrodite germ line (see Chapter 15). In plants and mammals, all four core proteins of PRC2 are present. As in Drosophila, the mammalian complex is about 600 kD and plays a role not only in regulating homeotic gene expression, but also in the control of cell proliferation, X-chromosome inactivation, and

imprinted gene expression (see Section 4 and relevant chapters for more detail). In plants, several genes encoding PRC2 components have undergone duplications such that they now are present as small gene families. In Arabidopsis there is only one hom*olog of ESC, FERTILIZATION-INDEPENDENT ENDOSPERM (FIE), but three hom*ologs of E(Z), three hom*ologs ofSU(Z) 12, and five hom*ologs of p55 (referred to as MSIl-5) (Table 1). Varying combinations of these proteins form at least three distinct complexes that control specific developmental processes (Figs. 3 and 4) (Reyes and Grossniklaus 2003; Chanvivattana et al. 2004). The best studied of these complexes is formed by members of the FERTILIZATION-INDEPENDENT SEED (PIS) class, which playa crucial role in the control of cell proliferation in the seed (Grossniklaus et al. 2001). This FIS or MEA-FIE complex contains MEDEA, FIE, FIS2,

Table 1. Core PcG genes in model systems

M. musculus

D. melanogaster

A. thaliana

C. elegans

PcG DNA-binding proteins phD

Pleiohomeotic

zinc finger

phol

Pleiohomeotic-like

zinc finger

Psq

Pipsqueak

BTB-POl domain

Dsp1

Dorsal Switch Protein 1

HMG domain protein

esc

Extra sex combs

WD 40 repeats

Eed

FIE

MES-6

E(z)

Enhancer of zeste

SET domain

Ezh1/Enx2, Ezh2/Enxl

CLF MEA SWN

MES-2

Yy1

HMGB2

PRC2 core complex

Su(z) 12

p55

Suppressor of zeste12

p55

zinc finger VEFS box

mSU(Z) 12

histone-binding domain

RbAp48 RbAp46

FIS2 VRN2 EMF2 MSI1 (MS/2,3,4,5)

PRC1 core complex Pc

Polycomb

chromodomain

Cbx2/M33 Cbx4/MPc2 Cbx6 Cbx8/MPc3 Cbx7

Polyhomeotic

zinc finger SAM/SPM domain

Edr1/Mph I/Rae28 Edr2/Mph2

Psc

Posterior Sex Combs

zinc finger HTH domain

Bmil Rnf11 O/Zfp 144/ Me/18

dRing / Sce

dRing / Sex combs extra

RING zinc finger

Ring1/Ringla Rnf2/ Ring 1b

SOP-2

T RAN 5 C RIP T ION A LSI LEN C I N G

O. melanogaster

M. musculus

A. thaliana

B Y POL

yeo

M B G Ra U P

PRO TEl N 5

21 7

C. elegans

FIS complex

Figure 3. Conserved PRC2 Core Complexes

EMF complex

VRN complex

and MSIl. The FIS complex was found to regulate the genes encoding PHERES1 (PHE1), a MADS domain transcription factor; and MEIDOS, a hom*olog of Skp1, which in yeast plays a crucial role in the control of cell proliferation (Kohler et al. 2003b). Interestingly, the paternal allele of PHEl is expressed at higher levels than the maternal allele. This regulation of gene expression by genomic imprinting is under the control of the FIS complex, which specifically represses the maternal allele (Kohler et al. 2005). Thus, as outlined below, the FIS complex shares with its mammalian counterpart functions in regulating cell proliferation as well as imprinted gene expression. The EMF complex contains CLF and EMBRYONIC FLOWER2 (EMF2) (Chanvivattana et al. 2004). Mutations in either of these show weak homeotic transformations and an early flowering phenotype. The EMF complex is required to repress homeotic genes, whose combinatorial action determines the identity of floral organs (Goodrich et al. 1997). Thus, the EMF complex has a similar function in maintaining the repressed state of homeotic genes as PRC2 in Drosophila and vertebrates (Fig. 2). However, homeotic genes in plants do not encode homeodomain proteins, but rather other transcription factors belonging to the MADS-domain and the plant-specific AP2-domain families. Strong mutants of EMF2, however, have more severe phenotypes where their seedlings produce flowers directly after germination, bypassing the vegetative phase of development (Yoshida et al. 2001). Thus, the EMF complex plays a role both early in development, where it prevents immediate flowering, and later in floral organogenesis (Chanvivattana et al. 2004). At both stages, the

The core members of PRC2 in D. melanogaster, M. musculus, A. thaliana, and C. elegans are shown. In A. thaliana, an ancestral complex is proposed to have diversified into three variants with discrete functions in development. In C. elegans, the PRC2 core complex contains only three proteins: MES-3 does not have hom*ology with any other identified PRC2 protein. The colors indicate hom*ology, the contacts indicate interactions. (Adapted from Reyes and Grossniklaus 2003 and Chanvivattana et al. 2004.)

EMF complex represses floral homeotic genes such as AG and APETALA3 (AP3) (Fig. 4). The FIS class proteins, FIE and MSIl, have also been implicated in the control of homeotic gene expression (Figs. 3 and 4). Because mutations in both cause maternal-effect embryo lethality, this function was only revealed when partialloss-of-function alleles could be studied at later stages of development (Kinosh*ta et al. 2001; Hennig et aL 2003). Finally, the VRN complex plays a key role in a wellknown epigenetic process: vernalization (extended exposure to low temperature). Vernalization induces flowering in winter annuals, but the effect is only seen after many cell divisions (Fig. 4). A plant cell will remember that it was vernalized for many months, or even years, after the cold period. This cellular memory is maintained through passages in cell culture but not from one generation to the next (Sung and Amasino 2004a). The VERNALIZATION (VRN) genes mediate the response to vernalization. VRN2 was found to encode a SU(Z) 12 hom*olog (Gendall et al. 2001), which interacts with the plant E(Z) hom*ologs CLF and SWiNGER (SWN) in yeast two-hybrid assays (Chanvivattana et al. 2004). The transition to flowering is not only controlled by vernalization, but involves the perception of endogenous (developmental stage and age) as well as exogenous factors (day length, light conditions, temperature). Four pathways have been defined by genetic analyses: (1) The autonomous pathway constitutively represses flowering, (2) the photoperiod pathway accelerates flowering under long days, (3) the vernalization pathway induces flowering in response to exposure to cold temperature, and (4) gibberellins promote flowering. The flowering time

218

C HAP T E R

7 7

~:~e~:OPhyte ~opment

AGIAP3

ge,m;~ Figure 4. Involvement of Distinct PRC2s at Various Stages of Plant Development During the plant life cycle, distinct variants of PRC2 (see Fig. 3) control developmental progression. (A) A cleared wildtype ovule harboring the female gametophyte in its center. The FIS complex represses target genes that control proliferation of the central cell; as in all fis class mutants, this cell proliferates in the absence of fertilization. Around fertilization, MEA is also required to maintain a low level of MEA m expression, but this activity is independent of other components of the FIS complex. (B) Section of a wild-type seed harboring the embryo and endosperm, enclosed by the seed coat. After fertilization, the FIS complex is involved in the control of cell proliferation in embryo and endosperm. It maintains a low level of expression of PHE7 and is required to keep the paternal MEAP allele silent. (C) Wild-type (right) and emf2 mutant (left) seedling 21 days after germination. The emf2 seedling produced a flower with homeotic transformations but no leaves. The EMF complex prevents flowering and represses floral homeotic genes such as AG, AP3, and others. (0) Vernalized (right) and non-vernalized (left) plants, the latter being characterized by a prolonged vegetative phase and the production of many leaves. During the vegetative phase of development, exogenous and endogenous signals induce flowering. Vernalization leads to the repression of the floral repressor FLC and thus promotes flowering. The maintenance of this repression depends on the VRN complex. (E) Wild-type Arabidopsis flower. During flower organogenesis, the EMF complex regulates floral homeotic genes that determine the identity of floral organs. (A, Courtesy of j.M. Moore and U. Grossniklaus; B, courtesy of J.-P. Vielle-Calzada and U. Grossniklaus; C, reprinted, with permission, from Moon et al. 2003 [©ASPB]; 0, reprinted, with permission, from Sung and Amasino 2004a [©Elsevier]; E, reprinted, with permission, from Page and Grossniklaus 2002 [©Macmillan].)

T RAN S C R / P T / 0 N A L S / LEN C / N G

gene FIC, which contains a MADS box, is a key integrator of the flowering response: It represses flowering. FIC expression is reduced by both the vernalization and the autonomous pathway. Whereas the initial repression of FIC is independent of the VRN complex, the maintenance of repression requires VRN2, which alters chromatin organization at the FIC locus (Gendall et al. 2001). Note that one of the components of the autonomous pathway is a p55 hom*olog, FVE (or MSI4), which affects flowering time response but does not act in the vernalization pathway (Ausin et al. 2004; Kim et al. 2004). Because no biochemical studies on the VRN complex have been reported, its exact composition is currently unknown (Figs. 3 and 4). 2.2 Chromatin-modifying Activity of PRCl

How does PRC2 mediate its repressive effect? Several proteins of the PcG and trxG have SET domains, including the PRC2 component E(Z). The discovery that SET domain proteins possess HKMT activity (Rea et al. 2000) suggested an involvement of histone methylation in PcG function. Indeed, mammalian and Drosophila PRC2 complexes were shown to methylate histone H3 at lysine 27 (H3K27) and, to a lesser extent, H3K9 both in vivo and in vitro (Cao et al. 2002; Czermin et al. 2002; Kuzmichev et al. 2002; Muller et al. 2002). These histone marks are usually associated with a transcriptionally silent state. Furthermore, H3K9 and H3K27 methylation has been associated with repressed homeotic genes of the bithorax complex (Miiller et al. 2002). However, only H3K27 methylation was lost in E(z) mutants, stressing the importance of H3K27 methylation in PcG silencing. Unlike the SU(VAR)3-9 protein, which methylates H3K9 on its own, E(Z) proteins on their own do not have H3K27 HKMT activity. The smallest complex acting as a HKMT also requires ESC and SU(Z) 12, which may have modulating functions. It was recently shown that PRC2 complexes can also methylate H1K26 (Kuzmichevet al. 2004). Distinct isoforms of the mammalian ESC hom*olog, Eed, determine the specificity of mammalian PRC2 for H1K26 versus H3K27 methylation (Kuzmichev et al. 2004). However, the functional relevance of H1K26 methylation for PcG silencing remains unclear. In plants, the HKMT activity of PRC2 complexes has not yet been demonstrated in vitro. However, studies of FIC regulation have shown that vernalization induces a loss of acetylation and an increase of H3K9 and H3K27 methylation, mainly in the first intron of the gene (Bastow et al. 2004; Sung and Amasino 2004b). Both methylation marks were lost in vrn2 mutants, implicating the VRN complex in setting these repressive histone methyla-

POL

yeo

M B

G R 0 UP

PRO T E / N S

•

219

tion marks. In two other mutants, vrnl and vernalization insensitive3 (vin3), only the H3K9me2 mark is missing. VRNl and VIN3 encode transcription factors of the B3domain and homeodomain families, respectively, but the exact molecular mechanism of their involvement in modifying chromatin is currently unclear. From numerous studies to date, the main function of PRC2 seems to involve HKMT activity, but there are other chromatin-modifying activities present in some PRC2 variants. The Rpd3 gene encodes a HDAC that has been implicated in PcG silencing (Tie et al. 2001). However, although rpd3 mutations enhance PcG phenotypes, they do not show the typical homeotic transformations by themselves. The fact that RPD3 is not present in all PRC2 preparations may thus reflect either a weak overall interaction, or a tissue- and stage-specific interaction with the PRC2 core components. The interaction of RPD3 with PRC2 represents an interesting partnership, as both HKMT and HDAC activities associate with silent chromatin, and in combination may reinforce transcriptionally silent states. 2.3 Dynamic Function of PRC2 during Development

As pointed out above, the PRC1 and PRC2 core complexes are associated with distinct factors that may playa role in recruiting PcG complexes to tissue-specific target loci or in modulating their activity (Otte and Kwaks 2003). The different steps of PcG repression shown in Figure 5 illustrate the stage-specific compositions.of PcG complexes during Drosophila embryogenesis. So far, it has been difficult to characterize differences with respect to distinct tissues or cell types in flies because whole embryonic extracts are usually used for biochemical purifications. Studies performed in mammals and plants, however, clearly show that PcG complexes have distinct memberships in specific tissues and that their composition changes during cellular differentiation (Chanvivattana et al. 2004; Kuzmichev et al. 2005; Baroux et al. 2006). In mammals, expression levels of PcG genes differ tremendously from one cell line to the next. PcG complexes may even differ between target genes in the same cell, suggesting a highly dynamic behavior at different developmental stages. In Drosophila, PcG proteins maintain repressed states of homeotic genes, established during early embryogenesis, thereby fixing developmental decisions. Once the silent state of a PcG target has been fixed, it will remain in that state for the remainder of an individual's life span. In plants, a similar situation may occur with the VRN com-

220 •

C HAP T E R

1 1

Activation of PRE regulated gene

PRE

b PRE

plex: Once vernalized, the target gene(s) will be permanently inactivated and only reset in the next generation. Other plant PRC2 complexes, however, seem to respond more quickly to developmental or environmental stimuli. For instance, one function of the PIS complex is to repress cell proliferation in the absence of fertilization. Upon fertilization, however, cell proliferation is rapidly induced, presumably through the derepression of PcG target genes. This indicates that PcG repression is the default state, which has to be overcome by some unknown mechanism to allow normal developmental progression. The inactivation of PcG complexes as part of the normal plant life cycle may explain the absence of PRC1 proteins in plants (Fig. 4). PRC1 plays an important role in the permanent, stable, and long-lasting inactivation of target genes. Such permanent inactivation would be detrimental to plant development, where often PcG repression is released upon appropriate stimuli.

3 Maintaining Transcriptional Silencing 3.1 Components of PRC1

e Figure 5. Sequence of Events Leading to the PcG-dependent Repressed State of Gene Expression in Drosophila Embryos The original gene expression state of a PRE-regulated gene is determined by the activity of transcriptional regulators, either transcriptional repressors (TR) or activators (TA). Transcription through the PRE prevents the establishment of the "OFF" state and leads to the trxG-dependent "ON" state (for details, see Fig. 8 in Chapter 12). (a-b) A nontranscribed PRE binds specific DNA-binding proteins (e.g., PHO, PHOL, DSP1, or GAF) that are involved in the recruitment of the early PcG complex containing proteins of both PRC1 and PRC2. (e) This early PeG complex marks chromatin by E(Z)dependent histone methylation. (d) Maintenance of the silent state occurs through interactions of the two distinct complexes, PRC1 and PRC2, in the absence of the original transcriptional repressor. Maintenance of PRC1 is stabilized through binding of H3K27me3 via the chromodomain of Pc. (e) PRC1 can compact chromatin, further establishing tightly condensed, silent chromatin.

The molecular analysis of the PcG gene products has revealed a structurally diverse group of chromatin-associated proteins. PRC1 contains four PcG proteins; Polycomb (PC), Polyhomeotic (PH), Posterior Sex Combs (PSC), and Ring 1 (dRing1/SCE) (see Table 1) (Francis et al. 2001). They occur in stoichiometric amounts, and additional partner proteins have been identified depending on the material used for purification. A related complex has been purified from mammalian cells, suggesting that these four subunits form the core of PRC1 (Levine et al. 2002). Immunostaining of Drosophila polytene chromosomes, using antibodies directed against PRe1 proteins, showed overlapping localization patterns, which indicated that these proteins cooperate at a defined and common set of target genes (Fig. 6a). Additionally, the approximately 100 bands observed on the chromosomes provided evidence that the HOX genes are just part of a larger regulatory network, including other gene targets subject to PcG silencing. The PC gene encodes a 390-amino acid protein containing a chromodomain at its amino-terminal end. This conserved motif has hom*ology with HP1, a Drosophila protein required for heterochromatin formation (Paro and Hogness 1991; see also Chapter 5). The chromodomain was subsequently found to bind to methyl moieties at H3K27 and H3K9 (Bannister et al. 2001; Fischle et al. 2003). Another conserved domain is present at the carboxy-terminal end. The conservation, as well as the occur-

T RAN 5 C RIP T ION A LSI LEN C I N G

8 Y POL

yeo

G R0 U P

PRO TEl N 5

221

x 2L

ANT-C

• PC binding sites • predicted PREs

Figure 6. Targeting of PRCl to PREs on Polytene Chromosomes (a) Immunostaining of Drosophila polytene chromosomes to visualize the distribution of the PC protein. (b) Alignment of chromosome arms showing the overlap between predicted PRE sites on the Drosophila genome and the cytologically mapped PC-binding sites on polytene chromosomes. The two HOX gene clusters (ANT-C and BX-C) are prominent bind-

ing sites for PRCl s.

rence of several aberrations in mutant alleles, suggests an important but as-yet-unknown regulatory function in this part of the protein. The carboxyl terminus of PC is dispensable for targeting the protein to silenced genes (fulfilled by the chromodomain) but was found to interact in vitro with nucleosomes (Breiling et al. 1999). Whether this indicates an undiscovered recognition motif for another histone modification remains to be seen. For human Pc2, a SUMO E3 ligase activity has been demonstrated, pointing to SUMO modifications as important marks in the PcG silencing process (Kagey et al. 2003). The amino-terminal part of the PSC protein is conserved in the vertebrate proto-oncogene bmi-l and the tumor suppressor gene mel-18. This region contains a C3HC 4 ring finger motif, which may mediate protein-protein interactions. The ring finger motif has been implicated in subnuclear localization of Bmi-l/Mel-18, which is correlated with cellular transformations. In Drosophila, the polyhomeotic (ph) locus is duplicated, consisting of a proximal (ph-p) and a distal (ph-d) gene sharing extensive hom*ology. hom*ologous mouse PH proteins have been identified. All share a conserved single zinc finger and a SAM (also known as SEP or SPM) domain. This domain is also found in another PcG protein, Sex Combs on the Midleg (SCM). SAM domains are involved in protein-protein interactions, as it has been demonstrated that they participate in hom*o- or heterotypic interactions with other proteins. These findings support a possible function in generating large nuclear

complexes, required for silencing. Indeed, PcG proteins have been localized in subnuclear foci called PcG bodies, which might function as silencing compartments (Saurin et al. 1998). As mentioned above, dRING 1 was not initially recognized as a PcG member. Only biochemical purification uncovered the presence of this factor with a RING finger motif in PRCl, in which it is thought to playa structural role (Francis et al. 2001; Lavigne et al. 2004). The Ring1A and Ring1B proteins of mammalian PRC1 have been found to be associated with ubiquitylated H2A on the inactive X chromosome, and the maintenance of this histone mark was dependent on the Ring1 proteins (de Napoles et al. 2004; Fang et al. 2004; Cao et al. 2005; for more detail, see Section 4.1 and Chapter 17). These four proteins comprise the core structure of PRCl. However, other PcG proteins like SCM or the Zeste protein were found to be associated with the complex (OUe and Kwaks 2003). Their molecular function in PRCl remains unclear, as they seem to have additional roles in the nucleus; e.g., a transcriptional activator function of Zeste. Still other PcG genes were identified by virtue of their role as transcriptional regulators of the core PcG genes (Ali and Bender 2004). Namely, three PcG genes are upstream regulators of genes encoding PRC1 components. Negative feedback loops among PRC1 components, as well as positive regulation of PRC1 components by PRC2, further suggest a complicated cross-regulatory network among the PcG genes to ensure the fine-tuning of protein

222 • C HAP T E R

J J

levels in the complexes (Fig. 7a). Similarly, complex regulatory interactions have been described for the genes of the FIS complex in Arabidopsis (Baroux et al. 2006). 3.2 Targeting PRC1 to Silenced Genes

Transgene analyses of Drosophila homeotic gene clusters uncovered regulatory elements that are required for the maintenance of appropriate segment-specific expression of the HOX genes. These DNA elements-called Polycomb Response Elements or PREs-maintain the segment-specific expression of HOX genes beyond the embryonic ini-

Su(z)2-??

dRing1

E(PC)1t PSC Asx

Pcl

tlatlOn phase. PREs attract proteins of the PRCI when integrated at ectopic sites in the polytene chromosomes, suggesting that they define sequence specificity for the recognition and anchoring of PRCIs to target genes. However, the issue of PcG targeting appears to be a complex one. The size of functionally characterized PREs ranges from a few hundred to several thousand base pairs, containing consensus binding sites for many different DNAbinding proteins, and usually two or more PREs are found at a given target locus. So far, all characterized PREs come from Drosophila, and no PREs have yet been defined in mammals or plants. Despite the complexity of PREs, four

tHOXgenes

PRE

( esc/E(z)

PRE 2 - "ON"

• +~ • +cell

division

PRE 1 - "OFF"

PRE 2 - "ON"

cell cycle

PRE 1 - "OFF"

progression

Figure 7. PRCl Regulation and Function during Cell Division (0) Cross-regulatory interactions among the PcG genes, as suggested from genetic evidence. E(Pc), Pel, and Asx are positive regulators of the core PRCl members acting upstream. PRC2 members Esc and E(z) act as positive regulators of Pc transcription. A negative feedback by core PRCl members on Psc and dRingJ, as well as on Su(z)2, is observed. The finetuning of gene product level is probably required for well-balanced processes based on chemical equilibrium. (b) Sequence-specific transcription factors (TF) tether components of PRCl to a PRE. A stable silencing complex requires anchoring of PRCl via the chromodomain of PC to neighboring methylated histone tails. (c) Possible model for how differential gene expression states can be inherited. The process of intergenic transcription places positive epigenetic marks (e.g., acetylated histone tails, histone variants) at PREs that control active genes (PRE 2). All other PREs are silenced by default (PRE l). During DNA replication and mitosis, only the positive epigenetic signal needs to be transmitted to the daughter cells, ensuring that in the next interphase intergenic transcription is restarted at PRE 2 before default silencing is reestablished at all other PREs. (0, Adapted from Ali and Bender 2004.)

T RAN S C RIP T ION A LSI LEN C I N G

consensus sequence motifs could be identified and were shown to playa role in Drosophila PRE function. One of these motifs (GCCAT) is bound by both the Pleiohomeotic (PHO) and Pleiohomeotic-like (PHOL) proteins, which have partially redundant functions. PHO and PHOL function in PcG targeting, as they are found in PcG complexes isolated from early embryonic extracts, coimmunprecipitate with members of both PRCI and PRC2, and bind PREs in vitro (Fig. Sa) (Poux et al. 2001). Recently, a role in PcG recruitment was also demonstrated for DSPI, which binds the GAAA motif found in many PREs (Dejardin et al. 2005). Finally, the trxG proteins Zeste and GAF (encoded by the Trithorax-like gene) may help to recruit PcG proteins to their targets. A newly developed algorithm, based on the finding that clustered pairs of GAF, Zeste, and PHO/PHOL sites characterize a PRE, predicts known PREs with high probability and thus can identify new potential PcG target genes in the Drosophila genome (Fig. 6b) (Ringrose et al. 2003). The family of PRE-controlled genes ranges from the wellknown developmentally important transcription factor genes required for pattern formation to genes encoding factors involved in cell cycle regulation and senescence. PRCI, once bound, interacts with neighboring histones to generate stable silencing complexes at PREs (Fig. 7b). The H3K27me3 marks provided by the PRC2 act as additional binding sites for the chromodomain of PC (Fig. 7c). In their absence, as shown by competition with a soluble methylated histone tail peptide, the PRCls dissociate from their target genes (Czermin et al. 2002; Ringrose et al. 2004; Wang et al. 2004). The discovery of the HKMT activity of PRC2 and the associated histone marks typical of silent chromatin has suggested a new mechanism for the establishment of PcG repression. Following PRC2-catalyzed modification of H3K27me3, PRCI binds through the chromodomain of the PC protein to stabilize silencing. This is corroborated by the findings that (1) H3K27me3 marks and PC colocalize on polytene chromosomes and (2) PC binding is lost in £(z) mutants, which lack HKMT activity that modifies H3 with H3K27me3 marks at PREs, serving to recruit PC to its targets (Fig. 5). Although such a model is certainly attractive, the situation at PREs seems more complex because PRC2s and PRCls do not act sequentially, but rather are present together on PREs in early embryogenesis (Fig. 5b, c). Thus, it seems likely that H3K27 methylation is a downstream event after PcG recruitment, but plays a crucial role in establishing the silenced state. The model described above shows parallels to heterochromatin formation, where the Heterochromatin Pro-

B Y POL

yeo M B G R 0

PRO TEl N S

•

223

tein I (HP1) is recruited VIa its chromodomain to H3K9me marks generated by SU(VAR)3-9 (see Chapter 5). Thus, a productive silencing complex is targeted by transcription factors to defined DNA sequence elements but requires, in addition, an appropriately modified histone layer in the vicinity to generate a higher-order repressive chromatin structure (Fig. 5). During evolution, PREs have retained remarkably little sequence conservation. Even within closely related Drosophila species, the number, position, and composition of PREs vary substantially (L. Ringrose and R. Paro, unpubl.). This suggests that the sequence requirements as well as the position of the PREs are flexible and may be adapted to species-specific requirements. Nevertheless, the components of PRCI are highly conserved, and they presumably utilize the same basic molecular mechanism(s) to induce higher-order chromatin changes at silenced target genes.

3.3 Establishment of Repressive Functions by PRC1 The way in which PRE-bound PRCls interact with the promoter to prevent transcription is still unknown. The anchoring of paused RNA polymerase complexes at promoters, preventing initiation, has been attributed to PRE-PRCI interactions described for reporter constructs (Dellino et al. 2004). Additionally, PRCl was shown to counteract remodeling of nucleosomes in vitro and to induce a compact chromatin structure. Thus, PRCl potentially blocks the accessibility to DNA of transcription factors and other complexes required for transcription (Francis et al. 2004). Using the algorithm described above, PRE-like sequences are predicted to exist at almost all promoters of PcG-controlled Drosophila target genes. This suggests that PRCI occupation at both promoter and regulatory sites might foster interactions between PREs and promoters, and establish stably repressed chromatin structures unfavorable for transcription (Ringrose et al. 2003). The stability of silencing complexes, as demonstrated by anchoring via methylated histone tails, appears to be a hallmark of the long-term repressive function of the PcG proteins. However, when analyzed in vivo at the cellular level, a remarkably dynamic behavior is observed. PcG proteins cluster in PcG bodies, which vary in size and composition between cells, suggesting an interaction of silencing complexes in the nucleus in a developmentally regulated manner. Furthermore, dynamic in vivo analyses of GFPmarked PC and PH proteins uncovered a very high exchange rate of unbound proteins with their complexes at

224

•

C HAP T E R

7 7

silenced targets (Ficz et al. 2005). These results suggest that long-term repression is primarily based on a chemical equilibrium between bound and unbound proteins rather than on high-affinity protection of DNA-binding sites. 3.4 Preventing Heritable Repression by Anti-silencing

The binding of PRCls to PREs appears to be induced by default, as many of the anchoring PcG components and DNA-binding proteins are expressed in all cells, and PREs globally silence reporter genes in transgenic constructs. The counteracting proteins of the trxG do not, in fact, function as activators, but rather as anti-repressors (Klymenko and Mi.iller 2004; see Chapter 12). Thus, to maintain active transcription of a PRE-controlled gene, the silencing at that PRE has to be prevented in a tissue- and stage-specific manner. In Drosophila, for example, the activation of HOX genes is controlled by the early cascade of transcription factors encoded by the segmentation genes. Interestingly, these factors induce transcription not only of the HOX genes, but also of intergenic, noncoding RNAs that are transcribed through the associated PREs often found upstream or downstream (Fig. 5). It was demonstrated that transcription through PREs is required to prevent silencing and to maintain the active state of a reporter gene using transgenic constructs (Schmitt et al. 2005). The process of transcription most probably remodels PRE chromatin to generate an active state characterized, for instance, by a lack of repressive histone methylation and the presence of histone acetylation. Thus, even though the DNA-binding proteins attract PRC1 to this particular activated PRE, the histone environment does not allow anchoring of PC via H3K27me3, and no stable silencing will be established. Since silencing is induced by default in the PcG system, epigenetic inheritance of a differential gene expression pattern only requires the transmission of the active PRE state during DNA replication and mitosis (Fig. 7c). How this is achieved at the molecular level, and which epigenetic mark(s) is responsible for maintaining an active PRE state, are still open questions. Interestingly, it was recently shown that at a Drosophila PRE of the homeotic Ubx gene, noncoding RNAs produced at the PREs stay associated with chromatin and recruit the trxG regulator Absent Small or Homeotic discs 1 (ASH1). Destruction of these RNAs by RNAi attenuates ASH1 recruitment to the PRE, suggesting that this interaction plays an important role in the epigenetic activation of the homeotic genes, by overriding default PcG-induced silencing (SanchezElsner et al. 2006).

4 PeG Repression in Mammalian Development 4.1 From Gene to Chromosome Repression

Mutations in members of the murine PRC1 exhibit homeotic transformations of the axial skeleton. This can cause the appearance of additional vertebrae as a consequence of a derepression of HOX genes (Fig. 2e,f) (Core et al. 1997). In addition, the mutant mice display severe combined immunodeficiencies, caused by a lack of proliferative responses of hematopoietic cells (Raaphorst 2005). The role of PcG proteins has been particularly well studied in blood cells, in light of the fact that most bloodcell lineages are characterized by their well-described celltype-specific transcription programs. However, lineage commitment and restriction somehow need to be faithfully maintained through cell division. It turns out that in PcG knockout mice, B- and T-cell precursor populations are produced normally, indicating that PcG control is not involved in establishing lineage-specific gene expression patterns. PcG proteins, however, contribute to the irreversibility of the lineage choice, rather than being involved in the decision to follow a particular developmental pathway. Besides the control of the HOX genes, whose expression patterns characterize different blood-cell lineages, PcG proteins playa major role in controlling projiferation. The bmil gene, an ortholog of Drosophila Pse, was initially identified as an oncogene that, in collaboration with mye, induces murine lymphomagenesis (van Lohuizen et al. 1991). The Bmi1 protein controls the cell cycle regulators p161NK4a and p19 ARF (Jacobs et al. 1999). Both Bmi1 and the related protein Mel-18 are negative regulators of the INK4c-ARF locus required for normal lymphoid proliferation control. Misregulation of this important cell cycle checkpoint affects apoptosis and senescence in mice (Akasaka et al. 2001). Mammalian PcG proteins are also associated with the classic epigenetic phenomenon of X-chromosome inactivation (see Chapter 17). The inactivation of one X chromosome in female XX cells is accompanied by a series of chromatin modifications that involve PcG proteins (Heard 2004). In particular, components of PRC2, like the ESC hom*olog, Eed (Embryonic ectoderm expression), or the E(Z) hom*olog, Enx1 (Table 1), playa major role in the establishment of histone marks associated with transcriptional silencing. Transient association of this PRC2 with the X chromosome, coated by Xist RNA, is accompanied by H3K27 methylation. In contrast, eed mutant mouse embryos show no recruitment of the Enx1 HKMT, nor can any H3K27 methylation be

T RAN 5 C RIP T ION A LSI LEN C I N G

observed. However, the absence of these PRC2 components does not lead to a complete derepression of the entire inactive X chromosome; rather, the sporadic reexpression of X-linked genes and an increase in epigenetic marks associated with an active state (H3K9ac and H3K4me3) are observed in some cells. This is likely because other partially redundant epigenetic mechanisms are in place to ensure the maintenance of one inactive X chromosome. Recruitment of PRC2 to the inactive X chromosome appears to be dependent on Xist RNA. Because association of PRC2 to the inactive X is only transient, it appears that the complex is only required to set epigenetic marks (i.e., H3K27me3) for the maintenance of silencing. Currently, it is not known whether PRCI directly recognizes these marks and is required for the permanent silencing of the inactive X chromosome, but PRCI components are found to be associated with the inactive X chromosome. However, DNA methylation is known to accompany the maintenance phase and is required for permanent X inactivation. PRC2 is specifically involved in the regulation of monoallelic expression of the X chromosome both in the embryo, where X-chromosome inactivation is random, and in extraembryonic tissues, where the paternally inherited X chromosome is always inactivated (imprinted X-chromosome inactivation). In addition, it was recently found that PRC2 is involved in the regulation of some autosomal imprinted genes. For instance, an analysis of 14 imprinted loci from six unlinked imprinting clusters showed that four of these were biallelically expressed in eed mutant mice (Mager et al. 2003; for more detail, see Chapter 19). Interestingly, all loci that lost imprinted expression were normally repressed when paternally inherited, whereas none of the maternally repressed loci was affected. Because it was recently shown that Ezh2 directly interacts with the mammalian DNA methyltransferases and is required for their activity (Vin~ et al. 2006), it is possible that PRC2 plays a role in the regulation of these imprinted genes via DNA methylation (see Chapter 18). An involvement of PRC2 in the regulation of imprinted gene expression has also been reported in Arabidospis, where the PRE] locus is expressed at much higher levels from the paternal allele (Kohler et al. 2005). In mutants for the E(z) hom*olog MEA, the maternal PRE] allele is specifically derepressed. Similarly, MEA also regulates its own imprinted expression: Early in reproductive development, the maternal MEA allele is strongly derepressed in mea mutants. This effect, how-

B Y POL

yeo

M B

GR0 UP

PRO TEl N 5

•

225

ever, is independent of the other components of the FIS complex (Baroux et al. 2006). In contrast, later in development, the FIS complex ensures the stable repression of the paternal MEA allele (Baroux et al. 2006; Gehring et al. 2006; Jullien et al. 2006). In this latter case, the PIS complex is involved in the silencing of paternally repressed imprinted genes similar to the situation in mammals. In addition, MEA also has a role in keeping expression of the maternal PRE] and MEA alleles at low levels as described above (Fig. 4). Because PRC2 components are present in plants, invertebrates, and mammals, PRC2 represents an ancient molecular module suitable for gene repression that was already present in the unicellular ancestor of plants and animals, prior to the evolution of multicellularity. Thus, these examples suggest that PRC2 was recruited independently for the regulation of imprinted gene expression in plants and mammals, the two lineages where genomic imprinting evolved (Grossniklaus 2005). 4.2 Consequences of Aberrant Transcriptional Activation

The finding that Emil misregulation causes malignant lymphomas in mice raises the question of whether human BMIl (a PRCI component) itself contributes to the development of cancer in a similar fashion. There is accumulating evidence that altered PcG gene expression is widespread in human malignant lymphomas (Raaphorst 2005). For instance, the level of BMIl overexpression in B-cell lymphomas correlates with the degree of malignancy, suggesting that PRCI components do play a role in the development of human cancer. However, the target genes of BMIl in human cells appear to be different from those of mouse lymphocytes, as no obvious down-regulation of p161NK4a could be correlated to the overexpression of the oncogenes. PcG gene overexpression is not only observed in hematological malignancies, but is also found in solid tumors, including medulloblastomas, and tumors originating from liver, colon, breast, lung, penis, and prostate (Fig. 8). The high expression of a PRC2 marker, Ezh2, is often found in early stages of highly proliferative lung carcinomas. This suggests that the well-known cascade of PRC2 initiation and PRCI maintenance (Fig. 5) might also accompany the development of a tumor cell lineage. Interestingly, PRC2 components also play a crucial role in the control of cell proliferation in Arabidopsis. Although aberrant growth does not lead to cancer and death in plants, a strict control of cell proliferation is

226

C HAP T E R

7 7

type is shared with rbr1 mutants, providing a link to the Rb pathway. Remarkably, a connection between the Rb pathway and PRC2 has also been reported in mammals (Bracken et a1. 2003), illustrating conserved regulatory networks between plants and mammals. 4.3 Maintaining Stem Cell Fate

Figure 8. PRC2 Regulates Cell Proliferation in Mammals and Plants

(a, b) Plant embryos derived from wild-type and mea mutant egg cells. MEA encodes a protein of the FIS complex and regulates cell proliferation. The giant mea embryo is much larger than the corresponding wild-type embryo at the same stage of development (late heart stage). Mutant embryos develop more slowly and have approximately twice the number of cell layers. (c, d) Normal and cancerous prostate epithelium. In the cancerous epithelium, Ezh2 expression is highly increased (labeled with an anti-Ezh2 antibody). Thus, both loss of E(Z) function in plants and overexpression of E(Z) function in humans can lead to defects in cell proliferation. (e, f) Control and RING1 overexpressing rat 1a fibroblast cells. Overexpression of RING1 leads to anchorage-independent growth in soft agar, typical of neoplastically transformed cells. (a,b, Courtesy of J.-P. Vielle-Calzada and U. Grossniklaus; c,d, reprinted, with permission, from Kuzmichev et al. 2005 [©National Academy of Sciences]; e,f, reprinted, with permission, from Satijn and Otte 1999 [©American Society for Microbiology].)

essential for normal development. In mutants of the fis class, the two fertilization products of flowering plants, the embryo and endosperm, overproliferate, and the resulting seeds abort (Fig. 8) (GrossnikJaus et al. 2001; Hsieh et a1. 2003; Guitton and Berger 2005). Effects on cell proliferation are also observed in double mutants of elf and swn, two of the plant £(z) hom*ologs. Such plants undergo normal seed development after germination but produce a mass of proliferating, undifferentiated tissue (callus) rather than leaves (Chanvivattana et al. 2004). Although it is not known how PRC2 controls cell proliferation in plants, it is likely to involve interactions with RBR1, the plant hom*olog of the Retinoblastoma (Rb) protein (Ebel et a1. 2004; Mosquna et a1. 2004). Mutants in the PIS class of genes not only show proliferation defects during seed development after fertilization, but are also required to prevent proliferation of the endosperm in the absence of fertilization. This latter aspect of the pheno-

Stem cells play an ever-increasing role in medicine. Their potential to provide progenitors for the healing of damaged tissue places them into a well-treasured toolbox of regenerative medicine. Not surprisingly, it is in the very well characterized blood-cell lineage where we know most about the identity and location of stem cells. Hematopoietic stem cells (HSCs) maintain the pool of blood cells by self-renewing as well as by producing daughter cells that differentiate into the lymphoid, myeloid, and erythroid lineages. The stem cell niche in the adult bone marrow provides the cells with specific external signals to maintain their fate. On the other hand, cell-intrinsic cues for the maintenance of the "stem cellness" state seem to rely on the PcG system. Mouse mutants affecting PRCI genes (e.g., bmi1/mel18, mphl/rae28, and m33; see Table 1) suffer from various defects in the hematopoietic system, such as hyperplasia (i.e., increased cell proliferation) in spleen and thymus, reduction in Band T cells, and an impaired proliferative response of lymphoid precursors to cytokines. The requirements for Bmi1 and Mel-I8 in stem cell selfrenewal during different stages of development suggest a changing pool of target genes between embryonic and adult stem cells. The PcG system is also required for neural stem cells (NSCs) as indicated by the neuronal defects observed in bmi1 mouse mutants (Bruggeman et a1. 2005; Zencak et a1. 2005). In particular, the mice are depleted of cerebral NSCs postnatally, indicating an in vivo requirement of Bmil in NSC renewal. As was found for the hematopoietic system, it appears that embryonal NSC maintenance is under a different PcG network control than adult NSC self-renewal. External signals like the sonic-Hedgehog signaling cascade modulate the Bmil response in NSCs and ensure a proliferative/self-renewal capacity (Leung et a1. 2004). The identification of these external cues controlling PcG repression came through the analysis of the development of cerebellar granule neuron progenitors (CGNPs). A postnatal wave of proliferation is induced by the signaling factor Sonic hedgehog (Shh), secreted by the Purkinje cells. The Shh signal branches to control N-Myc and Bmi1 levels (Fig. 9). Thus, Bmi1-deficient CGNPs have a

T RAN 5 C RIP T faN A LSI LEN C I N G

Shh

cerebellar granule neuron

~/oo~ N-Myc

B Y POL yeo M B

G Ra U P

PRO TEl N 5

•

227

Conceivably, however, the reprogramming of plant cells, which are totipotent and have the potential to form a complete new organism under appropriate conditions, could involve PcG regulation. Indeed, plants lacking the £(z) hom*ologs eLF and SWN produce a mass of undifferentiated cells after germination, suggesting that PcG genes are required to maintain a differentiated state (Chanvivattana et al. 2004).

5 Conclusion and Outlook It has been remarkable to follow the development of our

Cyc-01 102

proliferation 1 self-renewal pathway in stem cells Figure 9. Sonic Hedgehog Signaling Maintains Proliferation/Self-renewal of Cerebellar Progenitor Cells The Shh signaling cascade regulates both the Rb pathway and the p53 pathway via Bmil control of the p16/p19 proliferation checkpoint. Inhibition of Smoothened (Smoh) by the Shh receptor Patched (Ptch) results in downstream signaling in the nucleus. One part of the signal induces N-Myc, Cyclin Dl, and D2, whereas the other part activates Bmil via the Gli effectors. (Adapted from ValkLingbeek et al. 2004.)

defective proliferative response upon Shh stimulation. The Shh signal is able to control proliferation of these stem cells ultimately by modulating both the downstream Rb pathway (via N-myc and Bmillpl6 INK4 ') and the p53 pathway (via Bmillpl9 ARF ). This mechanism explains why hyperactivation of Shh signaling leads to the development of medulloblastomas (Leung et al. 2004). HSCs are regulated by a similar Indian hedgehog-controlled pathway. In NSCs, expression of the Hoxd8, Hoxd9, and Hoxc9 loci is under the control of Bmil. The appropriate HOX expression profile confers the necessary stem cell fate. Indeed, because stem cells represent a defined and committed cellular fate, it is not surprising that the PcG system maintains this particular fate in a mitotically heritable fashion. In the future, it will be interesting to identify the pool of targets of the PcG system in the different stem cell populations, and to learn how to influence the maintenance system to allow a controlled reprogramming of stem cell fates. At the moment, it is not clear whether the PcG plays a role in stem cell maintenance in plants.

understanding of PcG epigenetic regulation from the initial genetic identification of a Drosophila mutant possessing additional sex combs on the second and third leg. This eventually led to the discovery of a new class of regulators found to be required for fundamental epigenetic processes such as vernalization in plants and silencing of the mammalian X chromosome. Control of genetic information is highly influenced by chromatin structure and composition of histones in their various modified forms. The proteins of the PcG are directly involved in generating epigenetic marks, for instance, H3K27me3, as a consequence of developmental decisions. The same group "reads" (i.e., shows high affinity to), through the action of the PRCl proteins, these epigenetic marks and translates them into a stable, transcriptionally repressed state. In the model organism Drosophila, we have a relatively clear picture of how PcG complexes are anchored at PREs, for a defined group of target genes that are subject to longterm repression. However, to date, no PREs have been identified in other organisms. Although the basic function of PcG proteins remains the same, it is unclear which part of the plant and vertebrate genomes is subjected to their repression and how they are targeted to their site of action. Additionally, we need to get a better understanding of how an apparently dynamic group of proteins can impose a stable state of transcriptional repression through a chemical equilibrium. The other major question of the PcG research focuses on the heritability of the repressed state, the very essence of epigenetics. What is the identity of the molecular marks required to transmit a state of gene expression through DNA replication and mitosis? We know that the cooperation of trxG and PcG proteins maintains active or silent states of gene expression. Do both states need a corresponding epigenetic mark that is transmitted to daughter cells, or is only one sufficient, while the other represents the default state? The mechanism by which PcG proteins impose silencing on transcription during

228 •

C HAP T E R

1 1

the interphase of the cell cycle has become increasingly clear. In the future, the focus of research will be on how the information regarding a state of gene expression endures the DNA replication process and is faithfully transmitted to the daughter cells following mitosis.

References Akasaka T., van Lohuizen M., van der Lugt N., Mizutani-Koseki Y., Kanno M., Taniguchi M., Vidal M., Alkema M., Berns A., and Koseki H. 2001. Mice doubly deficient for the Polycomb Group genes Me1l8 and Bmil reveal synergy and requirement for maintenance but not initiation of HOX gene expression. Development 128: 1587-1597. Ali J.Y. and Bender W. 2004. Cross-regulation among the Polycomb group genes in Drosophila melanogaster. Mol. Cell. BioI. 24: 7737-7747. Ausin 1., Alonso-Blanco c., Jarillo J.A., Ruiz-Garcia 1., and MartinezZapater J.M. 2004. Regulation of flowering time by FVE, a retinoblastoma-associated protein. Nat. Genet. 36: 162-166. Bannister A.J., Zegerman E, Partridge J.E, Miska E.A., Thomas J.O., Allshire R.C., and Kouzarides T 2001. Selective recognition of methylated lysine 9 on histone H3 by the HPI chromodomain. Nature 410: 120-124. Baroux c., Gagliardini v., Page D., and Grossniklaus U. 2006. Dynamic regulatory interactions of Polycomb group genes: MEDEA autoregulation is required for imprinted gene expression in Arabidopsis. Genes Dev. 20: 1081-1086. Bastow R., Mylne J.S., Lister c., Lippman Z., Martienssen R.A., and Dean C. 2004. Vernalization requires epigenetic silencing of FLC by histone methylation. Nature 427: 164-167. Birve A., Sengupta A.K., Beuchle D., Larsson J., Kennison J.A., Rasmuson-Lestander A., and MUller J. 2001. Su(z)12, a novel Drosophila Polycomb group gene that is conserved in vertebrates and plants. Development 128: 3371-3379. Bracken A.P., Pasini D., Capra M., Prosperini E., Colli E., and Helin K. 2003. EZH2 is downstream of the pRB-E2F pathway, essential for proliferation and amplified in cancer. EMBO f. 22: 5323-5335. Breiling A., Bonte E., Ferrari S., Becker P.B., and Paro R. 1999. The Drosophila Polycomb protein interacts with nucleosomal core particles in vitro via its repression domain. Mol. Cell. BioI. 19: 8451-8460. Bruggeman S.W.M., Valk-Lingbeek M.E., van der Stoop P.P.M., Jacobs J.J.L., Kieboom K., Tanger E., Hulsman D., Leung c., Arsenijevic Y., Marino S., and van Lohuizen M. 2005. Ink4a and Arf differentially affect cell proliferation and neural stem cell self-renewal in Bmil-deficient mice. Genes Dev. 19: 1438-1443. Cao R., Tsukada Y., and Zhang Y. 2005. Role of Bmi-l and RinglA in H2A ubiquitylation and HOX gene silencing. Mol. Cell 20: 845-854. Cao R., Wang 1.J., Wang H.B., Xia 1., Erdjument- Bromage H., Tempst P., Jones R.S., and Zhang Y. 2002. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science 298: 1039-1043. Carrington E.A. and Jones R.S. 1996. The Drosophila Enhancer ofzeste gene encodes a chromosomal protein: Examination of wild-type and mutant protein distribution. Development 122: 4073-4083. Chanvivattana Y., Bishopp A., Schubert D., Stock c., Moon Y.H., Sung Z.R., and Goodrich J. 2004. Interaction of Polycomb-group pro-

teins controlling flowering in Arabidopsis. Development 131: 5263-5276. Core N., Charroux B., McCormick A., Vola c., Fasano 1., Scott M.P., and Kerridge S. 1997. Transcriptional regulation of the Drosophila homeotic gene teashirt by the homeodomain protein Fushi tarazu. Mech. Dev. 68: 157-172. Czermin B., Melfi R., McCabe D., Seitz v., Imhof A., and Pirrotta V. 2002. Drosophila enhancer of Zeste/ESC complexes have a histone H3 methyltransferase activity that marks chromosomal Polycomb sites. Cell Ill: 185-196. Dejardin J., Rappailles A., Cuvier 0., Grimaud c., Decoville M., Locker D., and Cavalli G. 2005. Recruitment of Drosophila Polycomb group proteins to chromatin by DSP1. Nature 434: 533-538. Dellino G.1., Schwartz Y.B., Farkas G., McCabe D., Elgin S.c., and Pirrotta V. 2004. Polycomb silencing blocks transcription initiation. Mol. Cell 13: 887-893. del Mar Lorente D., Marcos-Gutierrez c., Perez c., Schoorlemmer J., Ramirez A., Magin T, and Vidal M. 2000. Loss- and gain-of-function mutations show a Polycomb group function for RinglA in mice. Development 127: 5093-5100. de Napoles M., Mermoud J.E., Wakao R:, Tang Y.A., Endoh M., Appanah R., Nesterova TB., Silva J., Otte A.P., Vidal M., et al. 2004. Polycomb group proteins RinglA/B link ubiquitylation of histone H2A to heritable gene silencing and X inactivation. Dev. Cell 7: 663-676. Ebel c., Mariconti 1., and Gruissem W. 2004. Plant retinoblastoma hom*ologues control nuclear proliferation in the female gametophyte. Nature 429: 776-780. Fang J., Chen TP., Chadwick B., Li E., and Zhang Y. 2004. Ringl b-mediated H2A ubiquitination associates with inactive X chromosomes and is involved in initiation of X inactivation. f. BioI. Chem. 279: 52812-52815. Ficz G., Heintzmann R., and Arndt Jovin D.J. 2005. Polycomb group protein complexes exchange rapidly in living Drosophila. Development 132: 3963-3976. Fischle w., Wang Y., Jacobs S.A., Kim Y., Allis C.D., and Khorasanizadeh S. 2003. Molecular basis for the discrimination of repressive methyl-lysine marks in histone H3 by Polycomb and HPI chromodomains. Genes Dev. 17: 1870-1881. Fong Y., Bender 1., Wang W., and Strome S. 2002. Regulation of the different chromatin states of autosomes and X chromosomes in the germ line of C. elegans. Science 296: 2235-2238. Francis N.J., Kingston R.E., and Woodco*ck c.1. 2004. Chromatin compaction by a Polycomb group protein complex. Science 306: 1574-1577. Francis N.J., Saurin A.J., Shao Z., and Kingston R.E. 2001. Reconstitution of a functional core Polycomb repressive complex. Mol. CellS: 545-556. Gehring M., Huh J.H., Hsieh TE, Penterman J., Choi Y., Harada J.J., Goldberg R.B., and Fischer R.1. 2006. DEMETER DNA glycosylase establishes MEDEA Polycomb gene self-imprinting by allele-specific demethylation. CeIl 124: 495-506. Gendall A.R., Levy Y.Y., Wilson A., and Dean C. 2001. The VERNALIZATION2 gene mediates the epigenetic regulation of vernalization in Arabidopsis. Cell 107: 525-535. Goodrich J., Puangsomlee E, Martin M., Long D., Meyerowitz E.M., and Coupland G. 1997. A Polycomb-group gene regulates homeotic gene expression in Arabidopsis. Nature 386: 44-51. Grossniklaus U. 2005. Genomic imprinting in plants: A predominantly maternal affair. In Annual plant reviews: Plant epigenetics (ed. P. Meyer), pp. 174-200. Blackwell, Sheffield, United Kingdom. Grossniklaus u., Spillane c., Page D.R., and Kohler C. 2001. Genomic

T RAN 5 C R f P T f 0 N A LSI LEN C f N G

imprinting and seed development: Endosperm formation with and without sex. Curro Opin. Plant Bioi. 4: 21-27. Grossniklaus U., Vielle-Calzada J.E, Hoeppner M.A., and Gagliano W.B. 1998. Maternal control of embryogenesis by MEDEA, a Polycomb group gene in Arabidopsis. Science 280: 446-450. Guitton A.E. and Berger E 2005. Control of reproduction by Polycomb Group complexes in animals and plants. Int. ]. Dev. BioI. 49: 707-716. Hackett W.P., Cordero R.E., and Sinivasan C 1987. Apical meristem characteristics and activity in relation to juvenility in Hedera. In Manipulation of flowering (ed. J.G. Atherton), pp. 93-99. Butterworth, London. Hadorn E. 1968. Transdetermination in cells. Sci. Am. 219: 110. Heard E. 2004. Recent advances in X-chromosome inactivation. Curro Opin. Cell Bioi. 16: 247-255. Hennig L., Bouveret R., and Gruissem W. 2005. MSll-like proteins: An escort service for chromatin assembly and remodeling complexes. Trends Cell BioI. 15: 295-302. Hennig L., Taranto E, Walser M., Schonrock N., and Gruissem W. 2003. Arabidopsis MSIl is required for epigenetic maintenance of reproductive development. Development 130: 2555-2565. Hsieh T.E, Hakim 0., Ohad N., and Fischer R.L. 2003. From flour to flower: How Polycomb group proteins influence multiple aspects of plant development. Trends Plant Sci. 8: 439-445. Jacobs J.J.L., Scheijen B., Voncken J.W., Kieboom K., Berns A., and van Lohuizen M. 1999. Bmi-l collaborates with c-Myc in tumorigenesis by inhibiting c-Myc-induced apoptosis via INK4a/ARE Genes Dev. 13: 2678-2690. Jullien P.E., Katz A., Oliva M., Ohad N., and Berger E 2006. Polycomb group complexes self-regulate imprinting of the Polycomb group gene MEDEA in Arabidopsis. Curro BioI. 16: 486-492. Kagey M.H., Melhuish T.A., and Wotton D. 2003. The Polycomb protein Pc2 is a SUMO E3. Cell 113: 127-137. Kennison J.A. 1995. The Polycomb and trithorax group proteins of Drosophila: Trans-regulators of homeotic gene function. Annu. Rev. Genet. 29: 289-303. Kim H.J., Hyun Y., Park J.Y., Park M.J" Park M.K., Kim M.D., Kim H.J., Lee M.H., Moon J" Lee I., and Kim J. 2004. A genetic link between cold responses and flowering time through FVE in Arabidopsis thaliana. Nat. Genet. 36: 167-171. Kinosh*ta T., Harada J,J., Goldberg R.B., and Fischer R.L. 2001. Polycomb repression of flowering during early plant development. Proc. Natl. Acad. Sci. 98: 14156-14161. Klebes A., Sustar A., Kechris K., Li H., Schubiger G., and Kornberg T.B. 2005. Regulation of cellular plasticity in Drosophila imaginal disc cells by the Polycomb group, trithorax group and lama genes. Development 132: 3753-3765. Klymenko T. and Muller J. 2004. The histone methyltransferases Trithorax and Ashl prevent transcriptional silencing by Polycomb group proteins. EMBO Rep. 5: 373-377. Kohler C, Page D.R., Gagliardini v., and Grossniklaus U. 2005. The Arabidopsis thaliana MEDEA Polycomb group protein controls expression of PHERESI by parental imprinting. Nat. Genet. 37: 28-30. Kohler C, Hennig L., Bouveret R., Gheyselinck J" Grossniklaus u., and Gruissem W. 2003a. Arabidopsis MSIl is a component of the MEA/FIE Polycomb group complex and required for seed development. EMBO f. 22: 4804-4814. Kohler C, Hennig L., Spillane C, Pien S., Gruissem W., and Grossniklaus U. 2003b. The Polycomb group protein MEDEA regulates seed development by controlling expression of the MADS-box gene PHERESI. Genes Dev. 17: 1540-1553.

8 Y

POL

yeo

GR0 UP

PRO TEl N 5

•

229

Kuzmichev A., Jenuwein T., Tempst P., and Reinberg D. 2004. Different EZH2-containing complexes target methylation of histone HI or nucleosomal histone H3. Mol. Cell 14: 183-193. Kuzmichev A., Nishioka K., Erdjument-Bromage H., Tempst P., and Reinberg D. 2002. Histone methyltransferase activity associated with a human multiprotein complex containing the Enhancer of Zeste protein. Genes Dev. 16: 2893-2905. Kuzmichev A., Margueron R., Vaquero A., Preissner T.S., Scher M., Kirmizis A., Ouyang X., BrockdorffN., Abate Shen C, Farnham P., and Reinberg D. 2005. Composition and histone substrates of Polycomb repressive group complexes change during cellular differentiation. Proc. Natl. Acad. Sci. 102: 1859-1864. Lavigne M., Francis N.J., King I.E, and Kingston R.E. 2004. Propagation of silencing; recruitment and repression of naive chromatin in trans by Polycomb repressed chromatin. Mol. Cell 13: 415-425. Lee N., Maurange C, Ringrose L., and Paro R. 2005. Suppression of Polycomb group proteins by JNK signalling induces transdetermination in Drosophila imaginal discs. Nature 438: 234-237. Leung C, Lingbeek M., Shakhova 0., Liu J" Tanger E., Saremaslani E, van Lohuizen M., and Marino S. 2004. Bmil is essential for cerebellar development and is overexpressed in human medulloblastomas. Nature 428: 337-341. Levine S.S., King I.E, and Kingston R.E. 2004. Division of labor in Polycomb group repression. Trends Biochem. Sci. 29: 478-485. Levine S.S., Weiss A., Erdjument Bromage H., Shao Z., Tempst E, and Kingston R.E. 2002. The core of the Polycomb Repressive Complex is compositionally and functionally conserved in flies and humans. Mol. Cell. Bioi. 22: 6070-6078. Lewis E.B. 1978. A gene complex controlling segmentation in Drosophila. Nature 276: 565-570. Luo M., Bilodeau P., Koltunow A., Dennis E.S., Peaco*ck W.J., and Chaudhury A.M. 1999. Genes controlling fertilization-independent seed development in Arabidopsis thaliana. Proc. Natl. Acad. Sci. 96: 296-301. Mager J., Montgomery N.D., de Villena EEM., and Magnuson T. 2003. Genome imprinting regulated by the mouse Polycomb group protein Eed. Nat. Genet. 33: 502-507. Marx J. 2005. Developmental biology-Combing over the Polycomb group proteins. Science 308: 624-626. Moon Y.H., Chen L., Pan R.L., Chang H.S., Zhu T., Maffeo D.M., and Sung Z.R. 2003. EMF genes maintain vegetative development by repressing the flower program in Arabidopsis. Plant Cell 15: 681-693. Mosquna A., Katz A., Shochat S., Graft G., and Ohad N. 2004. Interaction of FIE, a Polycomb protein, with pRb: a possible mechanism regulating endosperm development. Mol. Genet. Genomics 271: 651-657. Muller J., Hart CM., Francis N.J" Vargas M.L., Sengupta A., Wild B., Miller E.L., O'Connor M.B., Kingston R.E., and Simon JA 2002. Histone methyltransferase activity of a Drosophila polycomb group repressor complex. Cell 111: 197-208. Ohad N., Yadegari R., Margossian L., Hannon M., Michaeli D., Harada J,J., Goldberg R.B., and Fischer R.L. 1999. Mutations in FIE, a WD Polycomb group gene, allow endosperm development without fertilization. Plant Cell 11: 407-415. Otte A.E and Kwaks T.H. 2003. Gene repression by Polycomb group protein complexes: A distinct complex for every occasion? Curro Opin. Genet. Dev. 13: 448-454. Page D.R. and Grossniklaus U. 2002. The art and design of genetic screens: Arabidopsis thaliana. Nat. Rev. Genet. 3: 124-136. Paro R. and Hogness D.S. 1991. The Polycomb protein shares a hom*ologous domain with a heterochromatin-associated protein of Drosophila. Proc. Natl. Acad. Sci. 88: 263-267.

230 •

C HAP T E R 7 7

Poux S., McCabe D., and Pirrotta V. 2001. Recruitment of components of Polycomb group chromatin complexes in Drosophila. Development 128: 75-85. Raaphorst EM. 2005. Deregulated expression of Polycomb-group oncogenes in human malignant lymphomas and epithelial tumors. Hum. Mol. Genet. 14: R93-100. Rea S., Eisenhaber E, O'Carroll N., Strahl B.D., Sun Z.W., Schmid M., Opravil S., MechtJer K., Ponting e.P., Allis e.D., and Jenuwein T 2000. Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406: 593-599. Reyes J.e. and Grossniklaus U. 2003. Diverse functions of Polycomb group proteins during plant development. Semin. Cell Dev. BioI. 14: 77-84. Ringrose 1., Ehret H., and Paro R. 2004. Distinct contributions of histone H3 lysine 9 and 27 methylation to locus-specific stability of Polycomb complexes. Mol. Cell 16: 641-653. Ringrose 1., Rehmsmeier M., Dura J.M., and Paro R. 2003. Genomewide prediction of Polycomb/Trithorax response elements in Drosophila melanogaster. Dev. CellS: 759-771. Sanchez-Elsner T, Gou D., Kremmer E., and Sauer E 2006. Noncoding RNAs of trithorax response elements recruit Drosophila Ashl to Ultrabithorax. Science 311: 1118-1123. Satijn D.P. and Otte A.P. 1999. RINGI interacts with multiple Polycomb-group proteins and displays tumorigenic activity. Mol. Cell. BioI. 19: 57-68. Saurin A.J., Shiels e., Williamson J., Satijn D.P.E., Otte A.P., Sheer D., and Freemont P.S. 1998. The human polycomb group complex associates with pericentromeric heterochromatin to form a novel nuclear domain. f. Cell Bioi. 142: 887-898. Schmitt S., Prestel M., and Paro R. 2005. Intergenic transcription through a Polycomb group response element counteracts silencing. Genes Dev. 19: 697-708. Sung S. and Amasino R.M. 2004a. Vernalization and epigenetics: How plants remember winter. Curro Opin. Plant BioI. 7: 4-10. - - - . 2004b. Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 427: 159-164. Tie E, Furuyama T, Prasad-Sinha J., Jane E., and Harte P.J. 2001. The Drosophila Polycomb group proteins ESC and E(Z) are present in a complex containing the histone-binding protein p55 and the histone deacetylase RPD3. Development 128: 275-286. Tschiersch B., Hofmann A., Krauss v., Dorn R., Korge G., and Reuter

G. 1994. The protein encoded by the Drosophila position-effect variegation suppressor gene Su(var)3-9 combines domains of antagonistic regulators of homeotic gene complexes. EMBO f. 13: 3822-3831. Valk-Lingbeek M.E., Bruggeman S.W.M., and van Lohuizen M. 2004. Stem cells and cancer: The Polycomb connection. Cell 118: 409-418. van der Lugt N.M., Domen J., Linders K., van Roon M., RobanusMaandag E., te Riele H., van der Valk M., Deschamps J., Sofroniew M., van Lohuizen M., et al. 1994. Posterior transformation, neurological abnormalities, and severe hematopoietic defects in mice with a targeted deletion of the bmi-l proto-oncogene. Genes Dev. 8: 757-769. van Lohuizen M., Verbeek S., Scheijen B., Wientjens E., van der Gulden H., and Berns A. 1991. Identification of cooperating oncogenes in E~-myc transgenic mice by provirus tagging. Cell 65: 737-752. Vire E., Brenner e., Deplus R., Blanchon 1., Fraga M., Didelot c., Morey 1., van Eynde A., Bernhard D., Vanderwinden J.M., et al. 2006. The Polycomb group protein EZH2 directly controls DNA methylation. Nature 439: 871-874. Wang 1., Brown J.L., Cao R., Zhang Y, Kassis J.A., and Jones R.S. 2004. Hierarchical recruitment of Polycomb group silencing complexes. Mol. Cell 14: 637-646. Yamamoto Y, Girard E, Bello B., Affolter M., and Gehring W.J. 1997. The cramped gene of Drosophila is a member of the Polycombgroup, and interacts with mus209, the gene encoding proliferating cell nuclear antigen. Development 124: 3385-3394. Yamamoto K., Sonoda M., Inokuchi J., Shirasawa S., and Sasazuki T. 2004. Polycomb group Suppressor of zeste 12 links Heterochromatin Protein la and Enhancer of zeste 2. f. Bioi. Chem. 279: 401-406. Yoshida N., Yanai Y, Chen L.J., Kato Y., Hiratsuka J., Miwa T, Sung Z.R., and Takahashi S. 2001. EMBRYONIC FLOWER2, a novel Polycomb group protein hom*olog, mediates shoot development and flowering in Arabidopsis. Plant Cell 13: 2471-2481. Zencak D., Lingbeek M., Kostic e., Tekaya M., Tanger E., Hornfeld D., Jaquet M., Munier EL., Schorderet D.E, van Lohuizen M., and Arsenijevic Y. 2005. Bmil loss produces an increase in astroglial cells and a decrease in neural stem cell population and proliferation. J. Neurosci. 25: 5774-5783.

Transcriptional Regulation by Trithorax Group Proteins Robert E. Kingston 1 and John W. Tamkun 2 1Department of Molecular Biology, Massachusetts General Hospital, Boston, Massachusetts 02114 2Department of Molecular, Cell and Developmental Biology, University of California, Santa Cruz, California 95064

CONTENTS 1. Introduction, 233 7.7

Identification of Genes Involved in the Maintenance of the Determined State, 233

7.2

trxG Proteins in Other Organisms, 236

7.3.

trxG Proteins Play Diverse Roles in Eukaryotic Transcription, 237

2. Connections between trxG Proteins and Chromatin, 237 2.7.

trxG Proteins Involved in 238

2.2.

trxG Proteins That Covalently Modify Nucleosomal Histones, 242

3. Connections between trxG Proteins and the General Transcription Machinery, 243 4. Biochemical Functions of Other trxG Proteins, 244 5. Functional Interactions between trxG Proteins, 244 6. trxG Proteins: Activators or Anti-repressors?, 244 7. Conclusion and Outlook, 245 References, 246

231

GENERAL SUMMARY All cells in an organism must be able to "remember" what type of cell they are meant to be. This process, referred to as "cellular memory" or "transcriptional memory," requires two basic classes of mechanisms. The first class, discussed in the previous chapter, functions to maintain an "OFF" state for genes that, if turned on, would specify an inappropriate cell type. The Polycomb-Group (PcG) proteins have as their primary function this repressive role in cellular memory. The second class of mechanisms are those that are required to maintain key genes in an "ON" state. Any cell type requires the expression of master regulatory proteins that direct the specific functions required for that cell type. The genes that encode these master regulatory proteins must be maintained in an "ON" state throughout the lifetime of an organism in order to maintain the proper cell types within that organism. The striking multiple-winged fly in the left title figure illustrates the dramatic phenotypes that can result from the failure to maintain the "ON" state of a master regulatory gene. The proteins that are involved in maintaining the "ON" state are called trithorax-Group (trxG) proteins in honor of the trithorax gene, the founding member of this group of regulatory proteins. A large group of proteins with diverse functions make up the trxG. The roles these proteins play in the epigenetic mechanisms that maintain the "ON" state appear more complex at this juncture than the roles for PcG proteins in repression. The first complexity is that a very large number of proteins and mechanisms are needed to actively transcribe RNA from any gene. Thus, in contrast to repression, which might be accomplished by comparatively simple mechanisms that block access of all proteins, activation of a gene requires numerous steps, any

of which might playa role in maintaining an "ON" state. Thus, there are numerous possible stages at which a trxG protein might work. A second complexity in thinking about trxG proteins is that proteins which function in activation can also, in different contexts, function in repression. This might appear counterintuitive, but, depending on the precise architecture of a gene, the same protein carrying out its function might in one case help a gene become activated, and in another case help a different gene become repressed. At this time, it does not appear that trxG proteins are dedicated solely to the maintenance of gene expression, but that these proteins can also play multiple roles in the cell. These complexities evoke several interesting unanswered questions. Why are only some of the proteins needed to activate transcription also critical for maintenance of transcription? Do these proteins have functions that are uniquely suited to maintaining the active state? Or are some of these proteins needed for maintenance, solely due to an evolutionary accident that made them key regulators of a gene(s) particularly important to development? As shown below, some of the trxG proteins are involved in regulating chromatin structure in opposition to the mechanisms used by the PcG proteins. trxG proteins can place covalent modifications on chromatin or can alter chromatin by changing the structure and position of the nucleosomes that are the building blocks of chromatin. Other trxG proteins function as part of the transcription machinery. Thus, these proteins are found in a wider variety of complexes than the PcG proteins and are likely to play more complicated roles in epigenetic mechanism.

T R f THO R A X

1 Introduction

Numerous developmental decisions-including the determination of cell fates-are made in response to transient positional information in the early embryo. These decisions are dependent on changes in gene expression. This allows cells with identical genetic blueprints to acquire unique identities and to follow distinct pathways of differentiation. The changes in gene expression underlying the determination of cell fates are heritable; a cell's fate rarely changes once it is determined, even after numerous cell divisions and lengthy periods of developmental time. Understanding the molecular mechanisms underlying the maintenance of the determined state has long been a goal of developmental and molecular biologists. Many of the regulatory proteins involved in the maintenance of heritable states of gene expression were identified in studies of Drosophila homeotic (Hox) genes. Hox genes encode homeodomain transcription factors that regulate the transcription of batteries of downstream target genes, which in turn specify the identities of body segments (Gellon and McGinnis 1998). In Drosophila, Hox genes are found in two gene complexes: the Antennapedia complex (ANT-C), which contains the Hox genes labial (lab), Deformed (Dfd), Sex combs reduced (Scr), and Antennapedia (Antp); and the bithorax complex (BX-C), which contains the Hox genes Ultrabithorax (Ubx), abdominalA (abdA), and AbdominalB (AbdB) (Duncan 1987; Kaufman et al. 1990). Each Hox gene specifies the identity of a particular segment, or group of segments, along the anterior-posterior axis of the developing fly. For example, Antp specifies the identity of the second thoracic segment, including the second pair of legs, whereas Ubx specifies the identity of the third thoracic segment, including the balancer organs located behind the wings. Thus, the transcription factors encoded by Hox genes function as master regulatory switches that direct the choice between alternative pathways of development. The transcription of Hox genes must be regulated precisely, because dramatic alterations in cell fates can result from their inappropriate expression (Simon 1995; Simon and Tamkun 2002). For example, the derepression of Antp in head segments transforms antennae into legs, and the inactivation of Ubx in thoracic segments transforms balancer organs into wings. In Drosophila, the initial patterns of Hox transcription are established early in embryogenesis by transcription factors encoded by segmentation genes. The proteins encoded by segmentation genesincluding the gap, pair-rule, and segment polarity genessubdivide the early embryo into 14 identical segments.

U P PRO T E f N 5

•

233

These proteins also establish the initial patterns of Hox transcription, the first step toward the development of segments with distinct identities and morphology. Once established, the segmentally restricted patterns of Hox transcription must be maintained throughout subsequent embryonic, larval, and pupal stages in order to maintain the identities of the individual body segments. Because the majority of segmentation genes are transiently expressed during early development, this function is carried out by two other groups of regulatory proteins: the Polycomb group of repressors (PcG) and the trithorax group of transcriptional regulators (trxG) (Fig. 1). The re