Roles for the Human Sirtuin Sirt7 in Chromatin Regulation and Cancer

ROLES FOR THE HUMAN SIRTUIN SIRT7 IN CHROMATIN REGULATION AND CANCER

A DISSERTATION SUBMITTED TO THE DEPARTMENT OF BIOLOGY AND THE COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

Matthew Frederick Barber December 2011

This work is licensed under a Creative Commons Attribution- Noncommercial 3.0 United States License. http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/gn133dy2696

ii I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Or Gozani, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Katrin Chua

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Judith Frydman

Approved for the Stanford University Committee on Graduate Studies. Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file in University Archives.

iii ABSTRACT

Sirtuins are a conserved family of NAD+-dependent deacetylases that regulate diverse biological processes including genomic stability, metabolism, and aging. SIRT7 is a mammalian sirtuin whose enzymatic activity and physiologic functions have been unclear. Here I present work aimed at elucidating the biochemical, molecular, and cellular roles of human SIRT7. In Chapter 1, I report the discovery that SIRT7 possesses deacetylase activity specific for histone H3 acetylated at lysine 18 (H3K18Ac). In genome-wide binding studies, we find that SIRT7 associates with promoters of a specific set of gene targets, where it deacetylates H3K18Ac and promotes transcriptional repression. In Chapter 2, I discuss the identification of a novel SIRT7 interacting protein, the ETS transcription factor ELK4. Our work suggests that ELK4 plays an important role in defining the spectrum of SIRT7 target genes, by recruiting SIRT7 to a subset of promoters. In Chapter 3, I present our work establishing new functions for SIRT7 in cancer progression. Notably, selective hypoacetylation of H3K18Ac has recently been linked to oncogenic transformation, and in patients is associated with aggressive tumor phenotypes and poor prognosis. We find that deacetylation of H3K18Ac by SIRT7 is necessary for maintaining essential features of human cancer cells including anchorage independent growth and escape from contact inhibition. Moreover, SIRT7 is necessary for a global hypoacetylation of H3K18Ac associated with cellular transformation by viral oncoproteins. Finally, SIRT7 depletion markedly reduces the tumorigenicity of human cancer cell xenografts in mice. Together, this work establishes SIRT7 as the first known site-specific H3K18Ac deacetylase and demonstrates a pivotal role for SIRT7 in chromatin regulation, cellular transformation programs, and tumor formation in vivo.

iv TABLE OF CONTENTS

Abstract ...... iv List of Figures...... vi Preface ...... viii Acknowledgements ...... viii Introduction...... 1 The biology of chromatin ...... 1 Histones leave their mark...... 3 Histone modifications and disease...... 4 Sirtuins – silent information regulators ...... 5 Mammalian sirtuins...... 7 Summary ...... 9 Chapter 1: Discovery and Characterization of SIRT7 Enzymatic Activity ... 10 Overview...... 10 Results ...... 10 Discussion...... 13 Chapter 2: The ETS Protein ELK4 Recruits SIRT7 to Target Promoters ...... 24 Overview...... 24 Results ...... 24 Discussion...... 26 Chapter 3: SIRT7 Maintains the Transformed State of Cancer Cells...... 33 Overview...... 33 Results ...... 33 Discussion...... 34 Concluding Remarks ...... 43 Methods...... 46 Cell culture, RNAi and viral transduction ...... 46 Biochemical fractionation and IPs ...... 47 Histone deacetylation assays ...... 47 Quantitative mass spectrometry...... 47 ChIP-seq...... 48 ChIP and mRNA analysis...... 49 Flow cytometry ...... 49 Tumor xenograft experiments...... 50 References...... 51

v LIST OF FIGURES

Figure 1.1. SIRT7 is a chromatin-associated H3K18Ac-specific deacetylase.16 Figure 1.2. SIRT7 binds to gene promoters and couples H3K18 deacetylation to transcriptional repression ...... 18 Figure 1.3. Analysis of SIRT7 ChIP-seq...... 20 Figure 1.4. SIRT7 occupancy at target promoters...... 21 Figure 1.5. H3K18Ac changes in response to SIRT7 depletion...... 22 Figure 1.6. SIRT7 depletion results in transcriptional activation of targets. 23 Figure 2.1. SIRT7 is directed to target promoters by interaction with the ETS family transcription factor ELK4 ...... 27 Figure 2.2. SIRT7 does not interact with ELK1 or GABPα...... 28 Figure 2.3. ELK4 depletion does not globally alter SIRT7 chromatin association...... 29 Figure 2.4. ELK4 occupancy and H3K18 acetylation at subset of SIRT7 target promoters ...... 30 Figure 2.5. ELK4 recruits SIRT7 in prostate cancer cells...... 31 Figure 2.6. ELK1 and GABPα do not affect SIRT7-mediated transcriptional repression...... 32 Figure 3.1. SIRT7 depletion reverses cancer cell phenotypes and inhibits tumor growth in vivo...... 37 Figure 3.2. SIRT7 depletion reverses cancer cell phenotypes in multiple cell lines...... 39 Figure 3.3. SIRT7 deacetylase activity is required for cancer cell growth..... 39 Figure 3.4. SIRT7 depletion induces cell death and accumulation in G2/M phase...... 40 Figure 3.5. SIRT7 is over-expressed in patient tumor samples ...... 41 Figure 3.6. HDAC1 does not regulate SIRT7 target genes ...... 42

I dedicate this work to Nicola – my wife, friend, teacher, coach, partner-in-crime, and the absolute love of my life.

vii PREFACE

The work presented here constitutes the bulk of a manuscript currently undergoing revision for publication, presently titled “SIRT7 is a histone H3 lysine 18 deacetylase that regulates oncogenic transformation.” The list of authors includes Matthew F. Barber, Eriko Michishita, Yuanxin Xi, Mitomu Kioi, Zarmik Moqtaderi, Ruth I. Tennen, Nicolas L. Young, Kevin Struhl, Benjamin A. Garcia, Or Gozani, Wei Li, and Katrin F. Chua.

ACKNOWLEDGEMENTS

I am grateful to my mentors, Or Gozani and Katrin Chua, for their generous support and for sharing their scientific passion. I am thankful to my lab-mates, past and present, for all the laughs and great memories. I thank Dennis Bua for countless discussions and advice over the years. I thank my numerous collaborators, especially Eriko Michishita, Ruth Tennen, and Luisa Tasselli, who have contributed to this work. I am grateful to my undergraduate mentor, Ken Belanger, for imparting his excitement for biology and teaching. I thank my family, especially my brother Andrew and my parents John and Susan, for their love and encouragement. Finally I thank my wife, Nicola, who has supported me every step of the way.

viii INTRODUCTION

The biology of chromatin The genomic DNA of each cell in the human body is nearly identical, and yet these same cells are able to adopt an amazing array of forms and functions. How, then, are two cells with identical genomes able to achieve distinct fates? The answer lies in the ability of cells to dynamically control the expression of their genomes. The DNA of eukaryotes does not exist as a naked fiber in the cell nucleus. Rather, DNA is packaged by a number of proteins and protein complexes to form a material termed chromatin. The major component of chromatin is a repeating protein-DNA structure called the nucleosome. Each nucleosome is comprised of a protein octomer containing two copies each of four core histones (H2A, H2B, H3, and H4), around which roughly 146 base-pairs of DNA is wrapped1. Nucleosome-containing chromatin is able to assume various degrees of compaction, from the flexible 10-nm fiber to the level of a condensed mitotic chromosome. In the interphase nucleus, chromatin exists in one of two visibly distinguishable forms – densely compacted “heterochromatin”, and open “euchromatin2,3.” While heterochromatic regions tend to be gene-poor and transcriptionally silent, most active genes lie in areas of euchromatin. This observation reflects the physical constraints imposed by chromatin itself – increased compaction limits access to underlying DNA sequences for proteins like RNA polymerase. This fundamental distinction between active and silent chromatin domains demonstrates how chromatin structure can greatly influence genome function. Not surprisingly, factors that modulate chromatin structure have now been implicated in nearly all aspects of DNA metabolism and regulation, including transcription, replication, repair, recombination, and segregation4-6. The importance of chromatin in gene regulation is exemplified by studies of position-effect variegation (PEV). A classic study of PEV involves mutations in the white+ gene in D. melanogaster, which controls eye color7,8. The wildtype white+ gene confers red color, whereas null mutations result in white

1 color. In genetic screens performed over 80 years ago, H.J. Muller noticed that some mutants produced a variegated phenotype, with eyes containing patches of both red and white cells. It was later discovered that these mutations did not alter the white+ gene itself, but rather induced a translocation which placed the white+ gene near a boundary between heterochromatin and euchromatin. Thus, in cells where heterochromatin had spread over the white+ locus, the gene was silenced, and visa versa. Furthermore, mutants that were able to suppress or enhance this variegated phenotype – termed Su(var) and E(var), respectively – were also identified. It was not until years later that many of these Su(var) and E(var) genes were found to encode chromatin regulatory factors, which could influence the relative balance between heterochromatin and euchromatin9-11. This concept, that heritable changes in phenotype can be conferred without changes in genotype, is often termed “epigenetics.” Although the precise definition of epigenetics has evolved since Conrad Waddington first coined the term in 194212-14, the contribution of chromatin structure to epigenetic processes has become well established. Cells possess diverse means by which to modulate the properties of chromatin. DNA itself can be modified by cytosine methylation, which has primarily been linked to gene silencing15. In addition, several modes of chromatin regulation occur at the level of the nucleosome. ATP-dependent chromatin remodeling complexes, such as SWI-SNF, ISWI, and INO80, couple ATP hydrolysis to nucleosome sliding or eviction16. Because nucleosome- bound DNA is typically unrecognizable by transcription factors and other proteins, these remodeler complexes have the potential to either hide or expose sequences of regulatory DNA to trans-acting factors. Another mode of chromatin regulation occurs through the incorporation of histone variants. In addition to the four core histones, there are several conserved variants which confer unique properties to the nucleosome17. CENP-A, for example, is an H3 variant found at centromeres, where it is required for kinetochore assembly and accurate chromosome segregation18. H2AX is an H2A variant which is dispersed at low levels throughout the genome, and plays a critical role in the DNA damage response19. In many cases, the precise mechanisms by which

2 histone variants confer their specialized functions are still under investigation20.

Histones leave their mark Arguably the most intensely studied area of chromatin biology in recent years has been histone post-translational modifications6. Histones are among the most heavily modified proteins in the cell, with dozens of different modifications identified to date21. In addition to their central fold domains, histones contain long, unstructured N-termini which are believed to protrude out from the nucleosome core. These “tails” are the predominant sites of modification, although several important marks have been identified in the globular regions as well. Histones are subject to a wide variety of modification types, including lysine and arginine methylation, acetylation, phosphorylation, ubiquitylation, sumoylation, ADP-ribosylation, and others21. Major breakthroughs in modern chromatin biology were made in 1996, when Allis and colleagues identified the first histone acetyltransferase (HAT) in Tetrahymena22. Moreover, this gene shared close homology to a transcriptional activator in the budding yeast Saccharomyces cerevisiae, suggesting a link between histone marks and transcriptional regulation. Later that year, Schreiber and colleagues identified the first histone deacetylases (HDACs), which had also been implicated as transcriptional co-repressors23. It became rapidly apparent that these histone marks possess variable distributions and functions. Methylation of histone H3 at lysine 4 (H3K4me), for example, is associated with promoters of active genes24, while the same modification on a nearby residue on the same histone (H3K9me) serves as a hallmark of silent heterochromatin25. Histone modifications can also promote or antagonize each other. For example, phosphorylation of H3S10 can prevent the recognition of H3K9 methylation26,27. While some modifications are believed to influence nucleosome dynamics or higher-order structure through electrostatic changes or physical displacement (such as with phosphorylation or ubiquitylation), modifications can also be recognized by specialized protein domains. Lysine

3 methylation is a prime example where the addition of mono-, di-, or tri-methyl groups does not confer changes in charge or significant steric bulk to the nucleosome tail. How, then, could lysine methylation be providing such an important function to chromatin? In 2001 it was found that the chromodomain of heterochromatin protein 1 (HP1) serves as a binding module for methylated H3K928,29. This binding was highly specific, as the nearby (but functionally distinct) modification H3K4me cannot be recognized by HP1. H3K9 methylation thus serves as a scaffold by which HP1 can bind and spread in order to promote the formation of silent heterochromatin domains. Over the years additional protein domains have been identified as binders for lysine methylation30-34 as well as acetylation35. By the early 2000’s it had become apparent that histone modifications, (1) possess distinct localizations throughout the genome, (2) contribute to diverse biological functions, (3) can promote or antagonize the addition or recognition of other marks, and (4) can be specifically recognized by protein domains to transduce diverse functional outputs. This emerging complexity led Jenuwein and Allis to propose the widely-publicized “histone code” hypothesis in 2001, which posited that specific combinations of histone modifications comprise a code that can be “read” by binding modules as a means for indexing the genome36. Although the details of the theory are still debated37, this work surmised what many in the field had also begun to realize – that histone marks constitute a highly conserved and complex means of regulating underlying the DNA sequence38. Intense study over the past decade has further expanded our understanding of the scale, function, and mechanisms by which histone modifications contribute to genome regulation.

Histone modifications and disease Almost as quickly as histone modifying enzymes and binding proteins were identified, it became apparent that many of these factors were linked to human genetic diseases39. In much the same way that DNA rearrangements

4 are known to produce various pathologic states, so might changes at the level of chromatin structure lead to similar defects. In no case is this dangerous potential more apparent than in cancer. To date, there are a huge array of studies implicating both chromatin modifying enzymes and modification binding proteins in cancer progression40,41. Large-scale nuclear alterations in cancer are also widely observed, again reflecting an epigenetic rather than genetic basis for disease42. Although non-histone substrates of these enzymes likely contribute to pathogenesis as well, there is little doubt that alterations in chromatin structure play an important role in various human pathologies. The potential for chromatin regulators to influence disease is also encouraging – because histone modifications are dynamic, they are also reversible and amenable to therapeutic intervention. Moreover, enzymes like histone modifiers, which contain a catalytic active site, are strong candidates for drug design. Efforts over the past several years have led to the development of inhibitors of class I/II/IV HDACs, which show promise as anti-cancer agents in the clinic43,44. In addition to cancer, several other disease states have emerged with links to chromatin modifications. RAG2, the enzyme responsible for V(D)J recombination, also contains a PHD finger which specifically recognizes H3K4me345,46. Mutations in this domain that abolish H3K4me3 binding, but retain recombinase function, have been identified in human patients with immunological disorders. PRDM9, a lysine methyltransferase, has been implicated in fertility and meiotic defects in mice47,48, and several chromatin modifiers have been implicated in mental retardation49,50. It is likely that future work will continue to expand our understanding of chromatin regulation in disease pathogenesis.

Sirtuins – silent information regulators One conserved protein family with links to chromatin regulation is the class III deacetylases, termed sirtuins. The founding member of this family was identified in a genetic screen for mutants with defects in gene silencing at

5 mating loci in the budding yeast Saccharomyces cerevisiae51. It was thus named Silent Information Regulator 2, or Sir2. Work over the next several years established that Sir2 mediates silencing not only at mating loci, but also telomeres and ribosomal DNA (rDNA) repeats52,53. Intriguingly, Sir2 was also found to promote yeast replicative lifespan, a model that parallels aging of dividing cells54. However, for over a decade after its initial discovery, the molecular mechanism of Sir2 function remained a mystery. Major breakthroughs were made from work showing that sirtuins could couple the cleavage of an oxidized metabolite, nicotinamide adenine dinucleotide (NAD+), to ADP-ribosylation55. Subsequent work found that Sir2 in fact possessed another enzymatic function: NAD+-dependent histone deacetylation56-58. Detailed mutational analysis indicated that it was in fact the deacetylase activity of Sir2 which likely imparted silencing functions56. The identification of an enzymatic activity for Sir2 was exciting on many fronts. First, it was known that acetylated histones are enriched in transcriptionally active euchromatin, while heterochromatic regions are largely devoid of acetyl marks. Thus, histone deacetylation by Sir2 provided a molecular link to its silencing functions. Furthermore, the unusual NAD+- dependent activity of Sir2 raised the fascinating possibility that it could function as a metabolic sensor. Since NAD+ levels are elevated under conditions of low nutrients, this suggested that Sir2 activity could be modulated by the energy state of the cell. The links between sirtuins and longevity have garnered both intense research focus as well as intense scientific debate. Shortly after the identification of its deacetylase activity, Sir2 was implicated in the response to calorie restriction, a dietary regimen that has been shown to extend lifespan from yeast to mammals59. However, this effect was later reported to depend upon both exact dietary conditions and genetic background60. In addition, while yeast Sir2 appears to promote replicative lifespan (the number of times a mother cell can divide), it can also inhibit chronological lifespan (the length of time a mother cell can sustain reversible quiescence)61. Sir2 homologues in both C. elegans and D. melanogaster were also found to promote lifespan

6 extension62,63, albeit through a distinct mechanism from yeast Sir2. These findings were also later called into question64, citing issues of proper controls and genetic background. Thus, despite significant interest in sirtuin biology, their precise role in longevity is at times uncertain.

Mammalian sirtuins Mammals possess seven Sir2 homologues, SIRT1-765. Among them, SIRT1 is the closest homolog to yeast Sir2 and by far the most well characterized. SIRT1 is a nuclear protein and an extremely robust deacetylase with dozens of reported substrates66,67. However, unlike yeast Sir2, SIRT1 appears to predominantly target non-histone proteins and does not associate strongly with chromatin. Sirt1-knockout mice display a variety of developmental defects (most notably in the retina and heart) as well as sterility68,69, suggesting pleiotropic defects in various tissues. SIRT1 has also been implicated in cancer, both as a tumor suppressor and oncogene70,71. Although mouse models suggest potential links between SIRT1 and aging- related processes72,73, a clear link between SIRT1 and mammalian lifespan has not emerged. While less studied, the remaining six mammalian sirtuins hint at diverse and fascinating roles in cell biology and physiology. SIRT2 is predominantly cytoplasmic and has been shown to catalyze tubulin deacetylation74. SIRT3 is a mitochondrial protein with numerous reported deacetylation substrates75. Mouse models for SIRT3 have also revealed that it plays an important role during the fasting response, providing yet another link between sirtuins and energy metabolism76. SIRT4 and SIRT5 are also mitochondrial proteins with links to metabolism77,78, although they are less promiscuous enzymes than SIRT3. The remaining sirtuins, SIRT6 and SIRT7, are both nuclear proteins. Interest in mammalian SIRT6 mounted several years ago following detailed analysis of Sirt6-deficient mice79. Although small, these mice appear normal at birth. Within several weeks, however, Sirt6-/- mice develop several

7 phenotypes synonymous with human aging, including osteoporosis, curvature of the spine, loss of subcutaneous fat, and colitis79. Sirt6-/- mice also experience a metabolic defect resulting in severe hypoglycemia, and typically die within their first month. Sirt6-/- mouse embryonic fibroblasts (MEFs) display genomic instability and defects in one or more DNA repair pathways. These striking phenotypes suggest that SIRT6 plays important roles in maintaining organismal health, and perhaps aging-related cellular processes. Despite this, the molecular function of SIRT6 remained a mystery, and numerous studied concluded that SIRT6 lacks deacetylase activity74,80,81. Mirroring to the case of yeast Sir2 almost a decade earlier, a breakthrough was made when Chua and colleagues identified the first physiologic enzymatic activity for human SIRT6 to be NAD+-dependent deacetylation of histone H3 lysine 9 (H3K9)82. Shortly thereafter, H3K56 was also identified as a SIRT6 substrate83,84. Unlike SIRT1, SIRT6 is not a promiscuous enzyme in vitro, which likely contributed to the difficulty in finding bona fide substrates. The identification of SIRT6 enzymatic activity set the stage for a number of subsequent investigations into its molecular functions. Human SIRT6 was initially found to play an important role at telomeres – the structures that protect chromosome ends. SIRT6 is required to deacetylate H3K9 and H3K56 at telomeres specifically during S-phase, and SIRT6 depletion leads to telomere shortening, chromosomal fusions, and premature senescence of human primary cells82,83. Over the past several years, SIRT6 has been implicated in a number of other important cellular processes including transcriptional silencing85,86, DNA double-strand break (DSB) repair87,88, and metabolism89,90. Future work will likely continue to expand our understanding of SIRT6 molecular and health-related functions. Mammalian SIRT7, like SIRT6, also localizes to the nucleus. In addition, it appears to be enriched in the nucleolus as well as the nucleoplasm80. An early investigation into SIRT7 function found that it interacts with components of the PolI transcriptional machinery, and promotes PolI transcription of pre-ribosomal RNA91. Subsequent work in mice suggested that Sirt7 deficiency leads to cardiac defects, including hypertrophy and

8 cardiomyopathy, as well as reduced lifespan92. However, these results appear to conflict with other documented cases of Sirt7-deficient mice93, potentially stemming from differences in genetic background. In addition, two studies have identified SIRT7 as a gene significantly up-regulated in breast94 and thyroid cancer95,96. Although it was initially reported that mouse Sirt7 is capable of deacetylating the tumor suppressor p5392, this activity was in fact only observed upon addition of the now defunct sirtuin-activating compound resveratrol, whose effect has since been shown to be the product of an experimental artifact97-99. Thus, of the seven mammalian sirtuins, SIRT7 remains the only member lacking a known physiologic enzymatic activity.

Summary Over the past two decades researchers have greatly expanded our understanding of the roles for chromatin structure in regulating the genome. Histone modifications, and the proteins that add, subtract, or recognize these marks, are critical players in modulating chromatin. The disruption of chromatin modifiers and modifications in numerous diseases also points the their potential as therapeutic targets or diagnostic tools for treatment. Among the various groups of histone modifying enzymes, the sirtuins have emerged as an important class with links to genomic stability, gene regulation, metabolism, and disease. Future studies of these enzymes, from the level of basic biochemistry to animal models, will likely expand our understanding of chromatin biology as well as provide new avenues for disease treatment. In the following chapters I will describe my efforts to characterize the human SIRT7 protein, beginning with the search for its enzymatic activity as well as molecular functions.

9 CHAPTER 1: DISCOVERY AND CHARACTERIZATION OF SIRT7 ENZYMATIC ACTIVITY

Overview Among the seven sirtuins present in mammalian genomes (SIRT1-7), SIRT7 is the only one for which a clear enzymatic activity and substrate have remained elusive. SIRT7 is a nuclear protein that is concentrated in nucleoli, where it is associated with RNA polymerase I activity at ribosomal DNA91. However, no direct deacetylase activity for SIRT7 on histones or other substrates has been detected in multiple studies, and the molecular mechanism of SIRT7 function remains largely unclear. In this chapter I describe our efforts to identify SIRT7 deacetylase substrates using biochemical approaches, followed by genomic and molecular studies aimed at understanding the sites of SIRT7 function at chromatin.

Results Although SIRT7 is known to localize to the nucleus, it was previously unclear whether it associates directly with chromatin. To investigate a potential role for SIRT7 in chromatin regulation, I performed biochemical fractionation studies in human cells (Fig. 1.1a). SIRT7 was almost exclusively detected in the chromatin-enriched fraction, co-purifying with histones. This suggested that histones themselves might be strong candidate SIRT7 substrates. To investigate this possibility, I initiated a mass spectrometry- based screen to test for NAD+-dependent SIRT7 deacetylase activity in vitro, using a collection of acetylated histone peptides as substrates. Each histone peptide is acetylated at single lysine residue, allowing deacetylation to be assessed on a site-specific basis. Peptides were incubated with NAD+ either in the presence or absence of recombinant SIRT7; a decrease in mass of 42 Daltons corresponds to deacetylation of lysine residues. Strikingly, recombinant SIRT7 protein showed highly specific deacetylase activity on histone H3 peptides acetylated on lysine 18 (H3K18Ac) (Fig. 1.1b). No activity

10 for SIRT7 was observed for any other peptide used in the screen (Fig. 1.1c). This pattern of activity lies in contrast to SIRT1, which displays almost no substrate selectivity in vitro100, as well as SIRT6, which has been reported to deacetylate H3K9 and H3K5682-84. To test for SIRT7 enzymatic activity on a full-length protein substrate, I next performed in vitro deacetylase reactions using purified calf thymus histone H3. However, despite extensive trouble- shooting, I was never able to detect significant deacetylation of H3K18Ac by SIRT7 using purified histones (data not shown). After some consideration, we reasoned that SIRT7 might require the presence of a native chromatin substrate for efficient activity. To test this hypothesis, I repeated deacetylation reactions using poly-nucleosomes purified from HeLa cells as substrate. In contrast to histones, SIRT7 displayed robust and specific NAD+-dependent H3K18Ac-deacetylase activity in the context of purified poly-nucleosomes (Fig. 1.1d). Similar to our observations using peptides, H3K18Ac was the only positive substrate detected using nucleosomes. This activity was also directly due to SIRT7 enzymatic activity, as it was abolished by substitution of a conserved histidine residue (H187→Y) in the predicted catalytic domain of SIRT7, and by the general sirtuin inhibitor nicotinamide (Fig. 1.1e). These first experiments indicated that SIRT7 is capable of deacetylating H3K18 in an in vitro setting. To assess SIRT7 activity in a more physiologic system, I over-expressed both wild-type SIRT7 and the SIRT7-HY catalytic mutant in human 293T cells. Whole cell extracts were prepared and analyzed by western blot. Consistent with in vitro observations, wild-type SIRT7 specifically deacetylated H3K18Ac when over-expressed in cells, whereas SIRT7-HY had no effect (Fig. 1.1f). One caveat of using westerns for assessing histone modification changes is that antibodies against these marks can display varying degrees of specificity. To assess SIRT7 enzymatic activity in cells using an antibody-independent approach, we collaborated with the lab of Ben Garcia at Princeton using quantitative mass spectrometry to measure relative levels of acetylated histone marks in both control and SIRT7 over- expressing 293T cells. This analysis revealed that SIRT7 induces a specific and dramatic depletion of H3K18Ac in cells (Fig. 1.1g). In contrast, changes in the

11 other acetylation marks, potentially due to downstream effects on chromatin structure, were modest or negligible. Together, our biochemical, cellular, and proteomic data demonstrate that SIRT7 is an NAD+-dependent deacetylase with specificity for the H3K18Ac mark. These results constitute the first identification of a clear enzymatic activity for SIRT7 at chromatin, and also make SIRT7 the first known deacetylase with specificity for the H3K18Ac mark. This activity is curious considering the broader context of mammalian deacetylases. Traditionally these enzymes have been shown to display little specificity in the sites that they target, especially when compared to other chromatin modifiers like lysine methyltransferases. Thus, the presence of “site-specific” deacetylases is quite unusual. SIRT6, the most closely related sirtuin to SIRT7, also displays strong substrate selectivity – thus far only two histone substrates (H3K9 and H3K56) have been reported for this enzyme. The specificity of SIRT7 activity prompted us to further investigate the significance of this modification, and the contexts in which SIRT7 might be regulating it. Previous reports indicate that H3K18Ac is enriched at gene promoters and correlates with transcriptional activation101. In addition, depletion of H3K18Ac has been associated with aggressive cancer cell phenotypes and poor patient prognosis102,103, and in cellular studies, has been linked to epigenetic reprogramming of primary human cells during cellular transformation by viral oncoproteins104,105. We therefore hypothesized that SIRT7 might deacetylate H3K18Ac at promoters to regulate gene expression. Depletion of SIRT7 from cells does not alter global levels of H3K18Ac (data not shown), suggesting that endogenous SIRT7 functions at specific genomic loci. We therefore collaborated with the labs of Wei Li at Baylor and Kevin Struhl at Harvard to perform chromatin immunoprecipitation coupled with high throughput sequencing (ChIP-seq) to determine the genome-wide occupancy of SIRT7. This analysis revealed that SIRT7 is highly enriched at a subset of gene promoters (Fig. 1.2a, b, 1.3a, b), consistent with the known positioning H3K18Ac. Using a high stringency cut-off (p-value < 1e-8), we identified ~250 target genes with significant enrichment of SIRT7 in their

12 promoters. Functional categorization of SIRT7 target genes revealed strong enrichment for factors involved in protein translation, RNA processing, and metabolism (Fig. 1.3c), with potential links to tumor suppressive activities (Fig. 1.3d). SIRT7 target genes exhibit higher expression levels than the genomic average (Fig. 1.3e), suggesting that SIRT7 generally binds to promoters that are actively transcribed. Notably, SIRT7 bound upstream of a significant number of ribosomal protein (RP) genes, as well as genes repressed in aggressive cancers (e.g., NME1) or identified in screens for tumor suppressor genes (e.g., COPS2)106. We next used direct quantitative RT-PCR (qPCR) to confirm that such promoter sequences are substantially enriched in the SIRT7 ChIPs. These signals are also specific, as they are abolished by siRNA-mediated depletion of SIRT7 (Fig. 1.2c, 1.4). I next asked whether SIRT7 deacetylates H3K18Ac at the promoters of these candidate target genes, using ChIP and qPCR. This analysis revealed that siRNA-mediated depletion of SIRT7 from HT1080 cells leads to significant hyperacetylation of H3K18 at the RPS20 (Ribosomal Protein S20), RPS7, RPS14, NME1, and COPS2 promoters, but not at several negative control promoters (Fig. 1.2d, 1.5). SIRT7 knockdown (S7KD) cells also showed specific increases in expression of several target genes at both the mRNA and protein levels, indicating that increased acetylation of H3K18 is coupled to transcriptional activation (Fig. 1.2e, f, 1.6). Together, these findings demonstrate that SIRT7 mediates gene-specific transcriptional repression by deacetylating H3K18Ac at promoters.

Discussion The identification of a physiologic enzymatic activity for SIRT7 provides new directions for studies of SIRT7 biology and molecular function. The observation that SIRT7 is capable of deacetylating nucleosomes, but not purified histone H3, further suggests that SIRT7 functions within the context of chromatin. The fact that histone peptides are also recognized by SIRT7 indicates that there may be some incompatibility between SIRT7 and the free

13 globular domain of H3 that inhibits substrate recognition, which is relieved in the case of a peptide substrate. The selectivity of SIRT7 for H3K18Ac is highly unusual – although other deacetylases are capable of modifying this site, none have been reported to possess this degree of selectivity. Previously, histone acetylation has been proposed to function primarily through negating the positive charges on lysine, which disrupts nucleosome-DNA interactions and prevents higher-order compaction107. This is in contrast to histone methylation, where distinct sites have been linked to diverse (and sometimes antagonistic) molecular functions. Recent studies have begun to reveal more unique functions for several histone acetylation sites as well. H3K56, for example, appears to be uniquely important for nucleosome deposition and the DNA damage response108,109. As mentioned above, selective regulation of H3K18Ac also appears to be linked to cancer progression103-105. The molecular nature of this link is not yet clear, but as H3K18Ac is known to be enriched in gene promoters101, it is likely that transcriptional regulation plays a major role. It is also not yet clear how modification of a single acetylation site in the H3 tail might be sufficient for regulating transcription. Based on previous reports that H3K18Ac influences deposition of H3R17 methylation110, as well as our own preliminary observations that SIRT7 can indirectly influence H3K4me2/3 (data not shown), it seems plausible that modulation of H3K18Ac might induce a cascade of histone modification changes which culminate in robust transcriptional phenotypes.

15 Figure 1.1. SIRT7 is a chromatin-associated H3K18Ac-specific deacetylase. a, Western analysis showing chromatin association of SIRT7. Shown are biochemical fractions of human 293T and HT1080 cells enriched for cytoplasm (S2), nucleoplasm (S3), or chromatin (P3). b, Mass spectra showing deacetylation of H3K18Ac peptide by SIRT7 (right panel). Control reactions were incubated with NAD+ alone (left panel). Predicted molecular weight of the acetylated peptide is 2650 Daltons (Da); arrows indicate predicted position of the deacetylated peptide. c, Summary of SIRT7 mass spectrometry results using a library of acetylated histone peptides. d, Western analysis showing deacetylation of H3K18Ac by SIRT7 in vitro using poly-nucleosomes as substrate. e, Deacetylase reactions were performed as in d, including the SIRT7-HY mutant or nicotinamide (NAM). f, Deacetylation of H3K18 by SIRT7 in 293T cells detected by western blot. Empty vector (control) and the SIRT7-HY mutant were used as negative controls. g, Changes in global histone acetylation levels in SIRT7 over-expressing versus control 293T cells, determined by quantitative mass spectrometry. Error bars represent standard error of the mean (S.E.M.) of three independent experiments.

17 Figure 1.2. SIRT7 binds to gene promoters and couples H3K18 deacetylation to transcriptional repression. a, Enrichment of SIRT7 in promoter regions determined by ChIP-seq. Promoter constitutes 1 kb upstream of the transcriptional start site (TSS). b, SIRT7 occupancy at the RPS20 gene determined by ChIP-seq. Arrow indicates position of the TSS. c, Direct qPCR (mean +/- S.E.) showing SIRT7 occupancy in control or SIRT7 knockdown (S7KD1, S7KD2) HT1080 cells. Negative control (IgG) and SIRT7 ChIPs were plotted relative to input. d, Direct qPCR (mean +/- S.E.) showing hyperacetylation of H3K18 in S7KD HT1080 cells. Negative control (IgG) and H3K18Ac ChIPs were plotted relative to input. e, Over-expression of SIRT7 target genes in S7KD HT1080 cells determined by quantitative RT-PCR (mean +/- S.E.). Signals were normalized to GAPDH expression. f, Western blots of cell extracts corresponding to cells in e.

19 Figure 1.3. Analysis of SIRT7 ChIP-seq. a, Enrichment of SIRT7 ChIP-seq peaks in promoters compared to the distribution over the whole genome (p- value 2.6e-188). b, Average SIRT7 ChIP-seq signal across all transcriptional start sites. c, Non-redundant gene ontology (GO) categories of SIRT7-occupied genes determined by ChIP-seq analysis. Categories were plotted and ranked based on the inverse log(10) of multiple testing corrected false discovery rate (FDR). d, Oncomine clusters containing significant overlap with SIRT7 target genes as determined by ChIP-seq analysis, plotted and ranked based on q- value. Each cluster represents strongly misregulated genes across patients in a single cancer study. e, SIRT7 occupied genes are highly expressed compared to whole-genome averages. Box plots were constructed and scaled by log(10) of fragments per kilobase per million sequence reads (FPKM).

Figure 1.4. SIRT7 occupancy at target promoters. SIRT7 occupancy in control and S7KD HT1080 cells determined by ChIP analysis (mean +/- S.E.). The GAPDH promoter was included as negative control.

Figure 1.5. H3K18Ac changes in response to SIRT7 depletion. Hyperacetylation of H3K18 at SIRT7 target genes (RPS14, COPS2, RPS7) in S7KD HT1080 cells determined by ChIP analysis (mean +/- S.E.). The GAPDH, gamma-tubulin, TMEM71, TIMM9, and UBE2N promoters were included as negative controls.

Figure 1.6. SIRT7 depletion results in transcriptional activation of targets. Increased mRNA expression (mean +/- S.E.) of target genes in S7KD HT1080 cells, measured by quantitative RT-PCR. The data are normalized to GAPDH mRNA levels.

23 CHAPTER 2: THE ETS PROTEIN ELK4 RECRUITS SIRT7 TO TARGET PROMOTERS

Overview Sirtuins lack sequence-specific DNA binding domains, raising the question of how selective recruitment of SIRT7 to its target promoters is achieved. In the case of other sirtuins, this recruitment is often mediated by interactions with proteins which themselves confer targeting specificity. Yeast Sir2, for example, is known to associate with distinct protein complexes which are required for silencing functions111,112. In addition, mammalian SIRT6 can function as a co-repressor of both NF-κB and HIF-1α by interacting with these factors and deacetylating histones at cognate promoters85,86. We therefore hypothesized that SIRT7 might be recruited to specific genomic loci by interacting with sequence-specific DNA binding proteins. In the following chapter I will describe our identification of a novel SIRT7-interacting protein which serves to target SIRT7 to a subset of promoters.

Results If SIRT7 were being targeted to promoters by a particular class of DNA- binding proteins, then the identity of such proteins might be inferred by analyzing the sequence-specific DNA motifs to which they bind. To test this hypothesis, we first used MDModule to identify de novo DNA motifs that are highly enriched in SIRT7-bound promoter sequences. These motifs were then compared to curated transcription factor binding motifs in the JASPAR CORE database113. Of the 50 most significant SIRT7-associated motifs, 25 corresponded to consensus binding sites for the ETS (E26 transformed specific) family of transcription factors. This family is perhaps most well known for including numerous proteins linked to cellular transformation and tumor formation114. The SIRT7 consensus motif was highly similar to the DNA sequence recognized by the ETS protein ELK4, also known as SAP-1 (Fig. 2.1a). Although the molecular function of ELK4 has not been extensively

24 studied, reports indicate that it can act as both a transcriptional activator and repressor115. Of the 289 most significant SIRT7 binding sites determined by ChIP-seq, 159 (55%) contained at least one ELK4 consensus motif. To test whether SIRT7 interacts physically with ELK4, I first performed co- immunoprecipitation (IP) studies of epitope-tagged or endogenous SIRT7 and ELK4 proteins. Western analyses revealed that FLAG-tagged SIRT7 associates with HA-tagged ELK4 IPed from 293T cells, and conversely, HA-ELK4 co-IP’s with FLAG-SIRT7 (Fig. 2.1b). In contrast, SIRT7 does not interact with two other ETS proteins, ELK1 and GABPα (Fig. 2.2). Moreover, I also observed interaction between endogenous SIRT7 and ELK4 proteins in co-IP experiments in HT1080 cells using SIRT7 and ELK4-specific antibodies (Fig. 2.1c). These results indicate that SIRT7 is preferentially bound to promoters containing ETS binding sites, and that the ETS protein ELK4 specifically interacts with SIRT7 in cells. I next examined whether ELK4 might play a role in targeting SIRT7 to specific promoters. ELK4 was depleted with siRNAs (Fig. 2.1d) and SIRT7 ChIP signals were then examined at two SIRT7-bound promoters that contain an ETS binding motif (NME1 and COPS2) and at a SIRT7-bound promoter lacking such a motif (RPS20). This analysis revealed that ELK4 depletion reduces SIRT7 occupancy at the ETS motif-containing NME1 and COPS2 promoters, but has no effect on SIRT7 binding at the RPS20 promoter or general chromatin-association (Fig. 2.1e, 2.3, 2.5b). Thus, while depletion of ELK4 does not affect SIRT7 binding at all target promoters, it does appear to be required for targeting to at least a subset of sites. Direct ChIP further confirmed that ELK4 is indeed present at these promoters (Fig. 2.4a). In addition, depletion of ELK4 resulted in a modest but reproducible increase in H3K18Ac at affected promoters, consistent with a moderate loss of SIRT7 occupancy (Fig. 2.4b). Depletion of ELK4 in SIRT7 over-expressing cells also attenuated SIRT7-mediated transcriptional repression of NME1 and COPS2, but not RPS20 (Fig. 2.1f, 2.5c), whereas depletion of ELK1 or GABPα had no effect (Fig. 2.6). Thus, ELK4 stabilizes the association of SIRT7 at a subset of its

25 target promoters and is required for SIRT7-mediated repression of these genes.

Discussion The identification of ELK4 as a SIRT7-interacting protein sheds new light on the molecular basis for SIRT7 function, as well as raises a number of novel questions. The ELK4 gene was recently shown to undergo a recurrent fusion in prostate cancer116, and has also been implicated in other cancer types as well117. As discussed further in Chapter 3, multiple lines of evidence have led us to speculate that SIRT7 itself possesses oncogenic functions. This would lead us to hypothesize that SIRT7 and ELK4 may perform overlapping functions that contribute to cancer progression. Although ETS proteins are often associated with transcriptional activation, ELK4 as well as several other ETS family members have also been linked to transcriptional repression115. In these instances, it could be that SIRT7 normally contributes to gene repression important for ELK4 function. Although our preliminary studies failed to detect interaction between SIRT7 and either ELK1 or GABPα, it is entirely possible that SIRT7 interacts with other ETS factors in addition to ELK4. These proteins have also been reported to display functional redundancy as well as diverse tissue-specific expression118, further suggesting that SIRT7 may interact with addition ETS proteins under certain physiologic conditions. In summary, these findings identify ELK4 as a novel SIRT7-interacting protein that serves to target SIRT7 to a subset of gene promoters.

Figure 2.1. SIRT7 is directed to target promoters by interaction with the ETS family transcription factor ELK4. a, Comparison of the SIRT7 consensus motif to the ELK4 consensus motif from JASPAR. b, FLAG-SIRT7 and HA- ELK4 interact in 293T cells. Immunoprecipitations (IPs) were analyzed by

27 western blot. c, Co-IP of endogenous SIRT7 using ELK4-specific antibodies in HT1080 cells. IPs were analyzed by western blot. d, Western blot showing depletion of ELK4 in co-IP samples. Input samples were probed for loading control. e, Depletion of SIRT7 at target promoters in ELK4 KD HT1080 cells determined by ChIP (mean +/- S.E.). Control (IgG) and SIRT7 IPs were normalized to input DNA. f, ELK4 depletion attenuates SIRT7-mediated transcriptional repression in HT1080 cells determined by quantitative RT-PCR. Error bars represent S.E.M. of three independent experiments.

Figure 2.2. SIRT7 does not interact with ELK1 or GABPα. Either FLAG- SIRT7 and HA-ELK1 or FLAG-SIRT7 and HA-GABPα were co-expressed in 293T cells. Immunoprecipitations (IPs) were performed from whole cell extracts and analyzed by western blot.

Figure 2.3. ELK4 depletion does not globally alter SIRT7 chromatin association. Biochemical fractions enriched for cytoplasm (S2), nucleoplasm (S3), or chromatin (P3) were isolated from control or ELK4-depleted HT1080 cells and analyzed by western blot.

Figure 2.4. ELK4 occupancy and H3K18 acetylation at a subset of SIRT7 target promoters. a, ELK4 occupancy at target promoters. ChIP qPCR analysis of control or ELK4 KD HT1080 cells. ELK4 and control (IgG) IPs were measured relative to input DNA. b, ChIP qPCR analysis of H3K18Ac levels at target promoters. RPS20, which is bound by SIRT7 but not ELK4, and GAPDH, which is bound by neither, were included as control. H3K18Ac and control (IgG) IPs were measured relative to input DNA.

Figure 2.5. ELK4 recruits SIRT7 in prostate cancer cells. a, Extracts from DU145 prostate cancer cells were analyzed by western blot. b, ELK4 promotes SIRT7 promoter targeting in prostate cells. Control and ELK4-depleted DU145 cells were used for ChIP qPCR analysis (mean +/- S.E.). Negative control (IgG) and SIRT7 IPs were plotted relative to input. c, DU145 cells were treated with either control or ELK4-specific siRNAs and transfected with either empty vector or SIRT7-expressing vector. Transcript levels were quantified by qPCR (mean +/- S.E.).

Figure 2.6. ELK1 and GABPα do not affect SIRT7-mediated transcriptional repression. a, ELK1 and GABPα do not affect SIRT7-mediated transcriptional repression. Human HT1080 cells were treated with either control siRNA, or siRNA directed against ELK1 or GABPα. Cells were then transfected with a control vector or vector expressing SIRT7. RNA was subsequently purified from cells and mRNA levels were analyzed by qPCR. b, Extracts from HT1080 cells treated with indicated siRNAs were analyzed by western blot.

32 CHAPTER 3: SIRT7 MAINTAINS THE TRANSFORMED STATE OF CANCER CELLS

Overview Several lines of evidence led us to hypothesize that SIRT7 might play a role in cancer progression. First, analysis of SIRT7-occupied genes revealed a strong correlation with genes whose expression is altered in various cancer types (Fig. 1.3d). Second, as previously mentioned, depletion of the H3K18Ac mark has been associated with cancer malignancy103-105. Finally, the observation that SIRT7 interacts with a putative oncogene, ELK4, suggests that they could be playing similar or redundant roles in cancer. All of these observations are consistent with the hypothesis that aberrant H3K18 deacetylation and transcriptional repression by SIRT7 contributes to cancer progression. In the follow chapter, I describe our efforts to determine the role of SIRT7 in regulating various cancer-associated cellular phenotypes, as well as implications for SIRT7 deacetylase activity in mediating these processes.

Results To investigate the role of SIRT7 in malignant cancer phenotypes, I first monitored anchorage-independent growth – a hallmark of transformed cells. Depletion of SIRT7 in multiple cancer cell lines using short hairpin RNAs (shRNAs) severely impaired colony formation in soft agar (Fig. 3.1a, b, 3.2a, b). SIRT7 depletion also inhibited the ability of DU145 prostate cancer cells to proliferate in low serum, another feature of oncogenic transformation (Fig. 3.2c). To determine the requirement for SIRT7 deacetylase activity in regulating these phenotypes, I first depleted endogenous SIRT7, then reconstituted with either wild-type SIRT7 or the SIRT7-HY mutant (Fig. 3.1a). While reconstitution with wild-type SIRT7 largely rescued colony formation in soft agar, the SIRT7-HY mutant did not (Fig. 3.1b). Similar observations were made for cell growth in low serum conditions (Fig. 3.3). Annexin staining and flow cytometry revealed that depletion of SIRT7 results in both increased

33 levels of cell death, as well as accumulation in G2/M phase (Fig. 3.4), which likely contributes to the observed growth phenotypes. Together, these findings demonstrate an important role for SIRT7 deacetylase activity in maintaining the transformed state of human cancer cells. Epigenetic reprogramming and oncogenic transformation by adenovirus is an established model for cancer development. Intriguingly, the adenoviral E1A oncoprotein induces a specific decrease in H3K18 acetylation that is important for transformation104,105. From my own work, we observed that SIRT7 depletion in HT1080 cells severely inhibits the E1A-dependent decrease in H3K18Ac levels (Fig. 3.1c). In addition, expression of E1A in non- dividing, contact-inhibited primary human fibroblasts triggers cell-cycle re- entry and escape from contact inhibition, another hallmark of oncogenic transformation119, and siRNA depletion of SIRT7 abolished this effect (Fig. 3.1d). These data indicate that SIRT7 is required for both the global H3K18Ac deacetylation and escape from contact inhibition that are induced by the E1A oncoprotein. To investigate the importance of SIRT7 in tumor growth in vivo, we collaborated with Mitomu Kioi, a former post-doc in Martin Brown’s lab at Stanford, to perform subcutaneous injections of control and S7KD U251 glioma cells in immune-compromised mice. Xenografts were subsequently measured over several weeks. Tumor formation was severely impaired in two independent SIRT7-depleted cell lines relative to the control lines (Fig. 3.1e, f), demonstrating that SIRT7 is required to maintain the tumorigenicity of cancer cells in vivo. Taken together, our results demonstrate that SIRT7 is required to maintain the oncogenic capacity of human cancer cells as well as tumor formation.

Discussion Our findings that SIRT7 is required for various cancer-associated cellular phenotypes suggest that SIRT7 may also contribute to development of human cancers. Consistent with this hypothesis, previous reports indicate that

34 SIRT7 is over-expressed in both breast and thyroid cancers94-96, and we have observed similar trends in other tissues as well (Fig. 3.5). Although depletion of SIRT7 impairs growth of transformed cells, we have not found that overexpression of SIRT7 is sufficient to induce transformation in primary cells (data not shown). This may reflect that SIRT7 is critical for maintenance, rather than initiation, of the transformed cell state. Such a role would also make SIRT7 an attractive candidate for therapeutic drug targeting. While inhibitors of class I/II/IV HDACs have been shown to effectively inhibit tumor growth in patients43,44, the potential for sirtuins as therapeutic targets in cancer has not been fully explored. While other cancer-associated deacetylases have previously been shown to target H3K18Ac in addition to other sites, the downstream molecular functions of SIRT7 appear to be distinct from at least some class I HDACs (Fig. 3.6), suggesting diverse means by which deacetylases contribute to cancer cell phenotypes. The observation that SIRT7 represses several RP genes is intriguing, because mutations in RP genes have been linked to cancer progression in both humans and zebrafish120,121. Notably, the SIRT7 target gene RPS14 was recently identified as a disease gene of the human 5q- syndrome, a myelodysplastic disorder that frequently progresses to acute myeloid leukemia121. The fact that such syndromes predispose – rather than induce – cancer in patients is further consistent with models where SIRT7 might be important for stabilizing rather than initiating tumor growth. In addition, several SIRT7-bound genes including NME1 (non- metastatic cells 1, or NM23A) and COPS2 (also known as CSN2 or ALIEN) possess known tumor suppressing functions106,122, suggesting that their repression by SIRT7 may contribute to cancer.

36 Figure 3.1. SIRT7 depletion reverses cancer cell phenotypes and inhibits tumor growth in vivo. a Western blots showing SIRT7 depletion in stable cell lines used in b. b, Reduced anchorage-independent growth of SIRT7 knockdown cells when plated in soft agar. Data represent averages and S.E.M. of three independent experiments. c, Western analysis showing impaired H3K18 deacetylation induced by E1A in S7KD HT1080 cells. Rel. H3K18Ac: relative levels of H3K18Ac in mock-treated versus E1A expressing cells, normalized to total H3 levels. d, SIRT7 depletion impairs E1A-mediated loss of contact inhibition in primary IMR90 fibroblasts determined by flow cytometry. DNA content (2N or 4N) is indicated. e, Representative imaging of tumors in mice at 16 days post injection. Control cells were implanted on the left, and S7KD cells on the right back of five individual RAG knockout mice. f, Tumor volume (mean +/- S.E.; n=5) as in e, was measured using a caliper over 35 days.

Figure 3.2. SIRT7 depletion reverses cancer cell phenotypes in multiple cell lines. a, SIRT7 was depleted from HT1080, U2OS, and DU145 cells. Retroviral transduction was performed for HT1080 and U2OS, while lentiviral transduction was used for DU145. Extracts were subsequently analyzed by

38 western blot. b, SIRT7 depletion impairs anchorage independent growth. Control and SIRT7 depleted cells were plated in soft-agar containing media and allowed to grow for two weeks before colonies were scored. Data represent average and S.E.M. of three independent experiments. c, SIRT7 regulates DU145 cell growth in low serum. SIRT7 was depleted from DU145 cells using lentiviral shRNA transduction. Cell growth in 1% serum was subsequently measured over four days. Data represent averages and S.E.M. of three independent experiments.

Figure 3.3. SIRT7 deacetylase activity is required for cancer cell growth. SIRT7-depletion in HT1080 cells impairs proliferation in low (1%) serum. Cells were transduced with retrovirus expressing either control or SIRT7 shRNA (S7KD2), as well as empty vector, wildtype (S7), or mutant (S7HY) SIRT7 expressing vectors. Error bars represent S.E.M. of three independent experiments.

Figure 3.4. SIRT7 depletion induces cell death and accumulation in G2/M phase. a, SIRT7 depletion increases basal cell death. Cells transduced with control or SIRT7 shRNA-expressing lentivirus were stained with annexin and analyzed by flow cytometry. Relative annexin-positive cells were plotted relative to control conditions. Cells were cultured in 1% serum. b, SIRT7 depletion leads to accumulation in G2/M phase. HT1080 cells were stained with propidium iodide and cell-cycle phase was determined by flow cytometry. Each cell cycle phase was plotted relative to control cells. Cells were cultured using 1% serum.

Figure 3.5. SIRT7 is over-expressed in patient tumor samples. Over- expression of SIRT7 in patient-matched tumor relative to unaffected control tissues measured by quantitative RT-PCR (mean +/- S.E.) and normalized to GAPDH expression.

Figure 3.6. HDAC1 does not regulate SIRT7 target genes. a, HDAC1 was depleted from HT1080 cells using siRNA. Whole cell extracts were analyzed by western blot. b, RNA was purified from control and HDAC1 KD HT1080 cells, and transcript levels were measured by qPCR (mean +/- S.E.).

42 CONCLUDING REMARKS

Together, our work has identified SIRT7 as a promoter-associated H3K18Ac deacetylase that mediates transcriptional repression and stabilizes cancer cell phenotypes. The discovery of physiologic SIRT7 enzymatic activity should provide a stepping-stone for future studies of SIRT7 function. While my work has focused on SIRT7 activity in the context of a pathologic state, we still have much to learn about the normal physiologic functions of SIRT7. The use of animal models to elucidate additional roles of SIRT7 will be critical. Although some studies have been initiated, discrepancies regarding the phenotype of Sirt7-deficient mice will need to be resolved before a clearer picture of SIRT7’s role can be produced. Based on the categories of genes occupied by SIRT7 as determined by ChIP-seq, we have postulated that SIRT7 could be playing roles in metabolism. Several other mammalian sirtuins have also been linked to metabolic control, both through transcriptional regulation in the nucleus86, as well as by deacetylating substrates in the mitochondria76,123. The fact that SIRT7 is enriched in nucleoli also suggests links to translation or ribosome biogenesis. A previous report indicates that SIRT7 promotes ribosomal RNA transcription91. Whether this activity functions in parallel to, or as a consequence of SIRT7’s role in RP gene repression remains to be elucidated, but may suggest a broad role for SIRT7 in coordinating the translation machinery. Interestingly, RP gene deletions and inhibition of translation have also been linked to lifespan extension in numerous model organisms, including mammals124,125, suggesting that gene repression by SIRT7 might also be linked to aging-related cellular processes. Consistent with this hypothesis, previous work reported that Sirt7-deficient mice possess a drastically shortened lifespan92, although independent data indicate that this phenotype is only observed in certain genetic backgrounds93. Future studies should assist in elucidating the potential role of SIRT7 in aging-associated pathologies and lifespan determination. It has been believed that histone acetylation functions primarily by negating positive charges on histone tails, reducing affinity of histones for the

43 negatively-charged DNA backbone and preventing higher-order chromatin compaction107. This is consistent with high levels of histone acetylation observed in euchromatin, where reduced compaction allows for DNA binding by polymerases and other proteins. Emerging evidence for specialized functions of acetyl-histone marks, including H3K18Ac, raises the question of how these modifications are transduced into functional outcomes. In the case of histone methylation, binding of specialized protein domains to specific modifications can account for a lack of charge or steric bulk of the modification itself. Likewise, bromodomains have been reported to possess a certain degree of selectivity for acetyl-histone marks126, suggesting that recognition of H3K18Ac by one or more bromodomain-containing proteins may contribute to the downstream functions of this mark. In addition, previous work has suggested that H3K18Ac enhances methylation of the nearby residue H3R17 by CARM1110. Thus, modulating H3K18Ac levels may also influence the addition or removal of other histone modifications which together results in transcriptional phenotypes. While SIRT7, like SIRT6, appears to be a highly selective enzyme, it is likely that other substrates exist aside from H3K18Ac. Although I tested a wide panel of known acetylation sites in histones, I cannot exclude the possibility that SIRT7 deacetylates additional histone residues, with either overlapping or independent functions. In addition, it is possible that SIRT7 possesses non-histone substrates that contribute to important biological processes. Future biochemical studies should provide a wider view of SIRT7 substrates, which will assist in understanding the role of SIRT7 in mammalian biology. In Chapter 2, I provided evidence that ELK4 is required to target SIRT7 to a subset of gene promoters. However, it is likely that SIRT7 interacts with additional proteins that provide similar function. Although we did not detect interaction between SIRT7 and either ELK1 or GABPα, there are over 20 ETS family members that recognize a similar DNA motif. It is also possible that an additional factor might be compensating for the function of ELK4, since depletion of ELK4 does not completely abolish SIRT7 binding at promoters

44 (Fig. 2.1e, 2.5b). The identification of additional interacting proteins could therefore provide novel insights into the biological functions of SIRT7. In this work I have also reported that SIRT7 depletion in transformed cells can greatly reduce tumorigenicity. However, in preliminary studies I have not observed that over-expression of SIRT7 is sufficient to induce malignancy of non-transformed cells (data not shown). This may reflect that increased levels of SIRT7 can only promote cancer in certain tissues or in combination with other oncogenes. It may also suggest that SIRT7 is critical for the maintenance, rather than initiation, of the transformed cell state. Future work should help to elucidate the precise role for SIRT7 in cancer progression, as well as its potential as a target for therapeutic intervention. In conclusion, the discovery of SIRT7 enzymatic activity and its connection to cancer-related cellular processes establishes a novel link between sirtuins, chromatin-based gene repression, and neoplastic disease.

45 METHODS

Cell culture, RNAi and viral transduction Cell lines were acquired from the ATCC. Human 293T, HT1080, DU- 145, U251 and U2OS cells were cultured in Advanced DMEM (Invitrogen) supplemented with penicillin-streptomycin (Invitrogen), GlutaMAX-1 (Invitrogen), and 10% newborn calf serum. IMR90 cells were cultured in DMEM/F12 (Invitrogen) containing penicillin-streptomycin, GlutaMAX-1, and 10% fetal bovine serum. Retroviral transduction was performed as previously described82. SIRT7 knockdown target sequences are as follows: S7KD1, 5’-CACCTTTCTGTGAGAACGGAA-3’; S7KD2, 5'- TAGCCATTTGTCCTTGAGGAA-3’, S7KD3, 5’- GCCTGAAGGTTCTAAAGAA-3’, S7KD4, 5’-GAACGGAACTCGGGTTATT- 3’, ELK4 siRNA target sequences are as follows: ELK4 KD1, 5’- CGACACAGACATTGATTCATT-3’; ELK4 KD2, 5’- GAGAATGGAGGGAAAGATATT-3’, as previously described117. Double- stranded siRNAs were purchased from Thermo Scientific. For retroviral packaging, 293T cells were co-transfected with pVPack-VSV-G, pVPack-GP (Stratagene) and expression or RNAi constructs, and viral supernatant was harvested after 48 hours. For transduction, cells were incubated with virus- containing supernatant in the presence of 8 µg/mL polybrene. After 48 hours, infected cells were selected for 72 hours with puromycin (2 µg/mL) or hygromycin (200 µg/mL). Adenovirus expressing the small E1A gene alone was generated and used to infect IMR90 cells using the Virapower Adenovirus System (Invitrogen) per the manufacturer’s instructions. Anchorage-independent growth was measured as previously described127. Cell cycle phase analysis was performed as previously described82. For analysis of H3K18Ac in E1A expressing cells, HT1080 cells were treated with control or SIRT7 siRNAs for 24 hours, then transfected with control (empty vector) or E1A-expressing vectors. 48 hours after siRNA transfection, extracts were prepared and analyzed by western blot. Relative levels of H3K18Ac (rel.

46 H3K18Ac) were determined by quantifying H3K18Ac western blot band intensities using ImageJ software, and normalizing to total H3 band intensities. Samples expressing E1A were set relative to their matched control.

Biochemical fractionation and IPs Samples enriched for cytoplasmic, nucleoplasmic, and chromatin fractions were prepared as previously described128. IP’s were performed as previously described129, except that one 150mm plate of cells was used per IP, Protein A/G beads (Sigma) were used instead of FLAG-resin, and elution was performed by boiling beads in western blot sample loading buffer.

Histone deacetylation assays In vitro histone deacetylation assays were performed as previously described82. Purification of human SIRT7 protein from baculovirus-infected insect cells was described previously80. Calf thymus histones were obtained from Roche, and poly-nucleosomes were purified from HeLa cells as previously described30. Histone peptides were synthesized at the Yale W. M. Keck peptide synthesis facility, and liquid chromatography mass spectrometry was performed at the Stanford University Vincent Coates Foundation Mass Spectrometry Laboratory. To determine histone acetylation levels in cells, 293T cells were transiently transfected with pcDNA 3.1 vectors containing FLAG-tagged wild-type SIRT7, the SIRT7-HY catalytic mutant, or an empty vector. Whole-cell lysates were harvested after 48 hours. Western blot analysis of histone acetylation levels was performed with modification-specific antibodies.

Quantitative mass spectrometry Acid-extracted total histones were subjected to chemical derivatization using D0-proionic anhydride and digestion with trypsin at a substrate to enzyme ratio of 10:1 for 6 hours at 37oC as previously described130. An

47 additional round of propionylation was performed on the digested peptides, with one sample being derivatized with the same D0-propionic anhydride reagent, and the other being derivatized with D10-propionic anhydride for 131 quantitative proteomics as previously described . D10-propionic anhydride introduces a 5 Da shift by derivatization of the free N-termini of all peptides generated from the trypsin digest. Equal amounts of both samples as quantified earlier by a Bradford assay were mixed together, and digested peptides were desalted using homemade STAGE tips as reported earlier132. Desalted peptides were loaded onto fused silica microcapillary column (75 µm) packed with C18 resin constructed with an ESI tip through an Eksigent AS-2 autosampler (Eksigent Technologies Inc., Dublin, CA) at a rate of ~200 nL/minute. Peptides were eluted using a 5-35% solvent B in 60 minute gradient (solvent A= 0.1 M acetic acid, solvent B = 70% MeCN in 0.1 M acetic acid) Nanoflow LC-MS/MS experiments were performed on an Orbitrap mass spectrometer (ThermoFisher Scientific, San Jose, CA) taking a full mass spectrum at 30,000 resolution in the Orbitrap and seven data-dependent MS/MS spectra in the ion trap. All MS and MS/MS spectra were manually verified and quantified.

ChIP-seq ChIP for ChIP-seq analysis was performed as previously described133. Four ChIP samples were sequenced using Illumina Solexa Genome Analyzer II single end sequencing protocol, including two SIRT7 replicates and two control replicates. The raw reads were generated and aligned to human reference genome hg18 by GAII data processing pipeline, allowing up to 2 mismatches. The uniquely mapped reads from replicates of SIRT7 and control samples were pooled respectively and processed by MACS (version 1.3.6)134 to generate the whole-genome ChIP-seq profiles, with the “--diag” option enabled for the sequencing depth saturation test. Clonal reads were automatically removed by MACS. The SIRT7 binding sites were called at p-

48 value cut off 1e-8 by comparing the SIRT7 signal with the control signal in MACS. The SIRT7 target gene list was generated by ranking the binding scores in descending order for all genes with SIRT7 binding sites detected within 3000 bp upstream of the TSS. The gene ontology (GO) analysis was performed using the DAVID database. (http://david.abcc.ncifcrf.gov)135,136. The cancer gene association study was performed using the Oncomine database (http://www.oncomine.org). Searches for de novo motifs (6 and 15 nucleotides in length) within SIRT7 binding sites (as determined by MACS analysis) were performed using MDModule137, with repetitive regions masked and running parameters “-s 100 -t 50”. The top 50 detected de novo motifs (top 5 of each motif size) were recorded and compared with JASPAR motif database using STAMP with default settings138.

ChIP and mRNA analysis Cells were prepared for ChIP as previously described139, with the exception that DNA was washed and eluted using a PCR purification kit (Qiagen) rather than by phenol-chloroform extraction. Whole mRNA was purified from cells using the RNEasy Mini Kit (Qiagen). Quantitative RT-PCR was performed using the Roche Universal ProbeLibrary System with a LightCycler 480 II (Roche), or using Taqman Gene Expression Assays (Applied Biosystems) on a 7300 Real Time PCR machine (Applied Biosystems). RNA from patient-matched tumor and unaffected control tissues was purchased from Ambion.

Flow cytometry Lentivirus-mediated SIRT7 depletion was performed in HT1080 cells. Cells were plated at constant density (0.26x10^6 cells in 6 cm plates), in triplicate, in low serum (1% new calf serum) for 84 hours. For annexin V analysis, cells were stained with annexin V and propidium iodide (FITC

49 Annexin Apoptosis Detection Kit, BD Pharmingen), samples data were acquired using a FACS LSRFortessa flow cytometer and FACS Diva software (BD Biosciences), and analyzed with CellQuest-Pro software (BD Biosciences). For cell cycle analysis, cells were pulsed with 33 µM BrdU and fixed in 75% ethanol in PBS. Prior to flow cytometry analysis, cells were stained with FITC mouse anti-BrdU (BD Pharmingen). Cells were then resuspended in a propidium iodide solution (2.5 µg/ml in PBS) with 250 µg/ml RNase A and incubated in the dark overnight at 4oC. Sample data was acquired and analyzed as for annexin analysis.

Tumor xenograft experiments Equal numbers of U251 cells expressing luciferase and either control (pSR) or SIRT7 KD vectors (upper quadrants: 4 x 106 pSR or S7KD1 cells; lower quadrants: 8 x 106 pSR or S7KD2 cells) were implanted on the backs of RAG knockout mice. Tumor growth was monitored using calipers and visualized using a bioluminescence-based IVIS system (Caliper LifeSciences).

50 REFERENCES

1 Luger, K., Mader, A. W., Richmond, R. K., Sargent, D. F. & Richmond, T. J. Crystal structure of the nucleosome core particle at 2.8 A resolution. Nature 389, 251-260, doi:10.1038/38444 (1997). 2 Grewal, S. I. & Moazed, D. Heterochromatin and epigenetic control of gene expression. Science 301, 798-802, doi:10.1126/science.1086887 (2003). 3 Woodcock, C. L. & Ghosh, R. P. Chromatin higher-order structure and dynamics. Cold Spring Harb Perspect Biol 2, a000596, doi:10.1101/cshperspect.a000596 (2010). 4 Li, G. & Reinberg, D. Chromatin higher-order structures and gene regulation. Curr Opin Genet Dev 21, 175-186, doi:10.1016/j.gde.2011.01.022 (2011). 5 Morrison, A. J. & Shen, X. Chromatin remodelling beyond transcription: the INO80 and SWR1 complexes. Nat Rev Mol Cell Biol 10, 373-384, doi:10.1038/nrm2693 (2009). 6 Kouzarides, T. Chromatin modifications and their function. Cell 128, 693-705, doi:10.1016/j.cell.2007.02.005 (2007). 7 Wakimoto, B. T. Beyond the nucleosome: epigenetic aspects of position-effect variegation in Drosophila. Cell 93, 321-324 (1998). 8 Henikoff, S. Position-effect variegation after 60 years. Trends Genet 6, 422- 426 (1990). 9 Rea, S. et al. Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406, 593-599, doi:10.1038/35020506 (2000). 10 James, T. C. & Elgin, S. C. Identification of a nonhistone chromosomal protein associated with heterochromatin in Drosophila melanogaster and its gene. Mol Cell Biol 6, 3862-3872 (1986). 11 Schotta, G. et al. A silencing pathway to induce H3-K9 and H4-K20 trimethylation at constitutive heterochromatin. Genes & development 18, 1251- 1262, doi:10.1101/gad.300704 (2004). 12 Russo, V. E. A., Martienssen, R. A. & Riggs, A. D. Epigenetic mechanisms of gene regulation. (Cold Spring Harbor Laboratory Press, 1996). 13 Jablonka, E. & Lamb, M. J. The changing concept of epigenetics. Ann N Y Acad Sci 981, 82-96 (2002). 14 Martin, C. & Zhang, Y. Mechanisms of epigenetic inheritance. Curr Opin Cell Biol 19, 266-272, doi:10.1016/j.ceb.2007.04.002 (2007). 15 Law, J. A. & Jacobsen, S. E. Establishing, maintaining and modifying DNA methylation patterns in plants and animals. Nat Rev Genet 11, 204-220, doi:10.1038/nrg2719 (2010). 16 Clapier, C. R. & Cairns, B. R. The biology of chromatin remodeling complexes. Annu Rev Biochem 78, 273-304, doi:10.1146/annurev.biochem.77.062706.153223 (2009). 17 Malik, H. S. & Henikoff, S. Phylogenomics of the nucleosome. Nat Struct Biol 10, 882-891, doi:10.1038/nsb996 (2003).

51 18 Allshire, R. C. & Karpen, G. H. Epigenetic regulation of centromeric chromatin: old dogs, new tricks? Nat Rev Genet 9, 923-937, doi:10.1038/nrg2466 (2008). 19 Stucki, M. & Jackson, S. P. gammaH2AX and MDC1: anchoring the DNA- damage-response machinery to broken chromosomes. DNA Repair (Amst) 5, 534-543, doi:10.1016/j.dnarep.2006.01.012 (2006). 20 Black, B. E. & Cleveland, D. W. Epigenetic centromere propagation and the nature of CENP-a nucleosomes. Cell 144, 471-479, doi:10.1016/j.cell.2011.02.002 (2011). 21 Bannister, A. J. & Kouzarides, T. Regulation of chromatin by histone modifications. Cell Res 21, 381-395, doi:10.1038/cr.2011.22 (2011). 22 Brownell, J. E. et al. Tetrahymena histone acetyltransferase A: a homolog to yeast Gcn5p linking histone acetylation to gene activation. Cell 84, 843-851 (1996). 23 Taunton, J., Hassig, C. A. & Schreiber, S. L. A mammalian histone deacetylase related to the yeast transcriptional regulator Rpd3p. Science 272, 408-411 (1996). 24 Strahl, B. D., Ohba, R., Cook, R. G. & Allis, C. D. Methylation of histone H3 at lysine 4 is highly conserved and correlates with transcriptionally active nuclei in Tetrahymena. Proc Natl Acad Sci U S A 96, 14967-14972 (1999). 25 Nakayama, J., Rice, J. C., Strahl, B. D., Allis, C. D. & Grewal, S. I. Role of histone H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science 292, 110-113, doi:10.1126/science.1060118 (2001). 26 Hirota, T., Lipp, J. J., Toh, B. H. & Peters, J. M. Histone H3 serine 10 phosphorylation by Aurora B causes HP1 dissociation from heterochromatin. Nature 438, 1176-1180, doi:10.1038/nature04254 (2005). 27 Fischle, W. et al. Regulation of HP1-chromatin binding by histone H3 methylation and phosphorylation. Nature 438, 1116-1122, doi:10.1038/nature04219 (2005). 28 Lachner, M., O'Carroll, D., Rea, S., Mechtler, K. & Jenuwein, T. Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature 410, 116- 120, doi:10.1038/35065132 (2001). 29 Bannister, A. J. et al. Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature 410, 120-124, doi:10.1038/35065138 (2001). 30 Shi, X. et al. ING2 PHD domain links histone H3 lysine 4 methylation to active gene repression. Nature 442, 96-99, doi:10.1038/nature04835 (2006). 31 Pena, P. V. et al. Molecular mechanism of histone H3K4me3 recognition by plant homeodomain of ING2. Nature 442, 100-103, doi:10.1038/nature04814 (2006). 32 Wysocka, J. et al. A PHD finger of NURF couples histone H3 lysine 4 trimethylation with chromatin remodelling. Nature 442, 86-90, doi:10.1038/nature04815 (2006).

52 33 Li, H. et al. Molecular basis for site-specific read-out of histone H3K4me3 by the BPTF PHD finger of NURF. Nature 442, 91-95, doi:10.1038/nature04802 (2006). 34 Bua, D. J. et al. Epigenome microarray platform for proteome-wide dissection of chromatin-signaling networks. PLoS One 4, e6789, doi:10.1371/journal.pone.0006789 (2009). 35 Dhalluin, C. et al. Structure and ligand of a histone acetyltransferase bromodomain. Nature 399, 491-496, doi:10.1038/20974 (1999). 36 Jenuwein, T. & Allis, C. D. Translating the histone code. Science 293, 1074- 1080, doi:10.1126/science.1063127 (2001). 37 Gardner, K. E., Allis, C. D. & Strahl, B. D. Operating on chromatin, a colorful language where context matters. J Mol Biol 409, 36-46, doi:10.1016/j.jmb.2011.01.040 (2011). 38 Turner, B. M. Histone acetylation and an epigenetic code. Bioessays 22, 836- 845, doi:10.1002/1521-1878(200009)22:9<836::AID-BIES9>3.0.CO;2-X (2000). 39 Bhaumik, S. R., Smith, E. & Shilatifard, A. Covalent modifications of histones during development and disease pathogenesis. Nature structural & molecular biology 14, 1008-1016, doi:10.1038/nsmb1337 (2007). 40 Gal-Yam, E. N., Saito, Y., Egger, G. & Jones, P. A. Cancer epigenetics: modifications, screening, and therapy. Annu Rev Med 59, 267-280, doi:10.1146/annurev.med.59.061606.095816 (2008). 41 Wang, G. G., Allis, C. D. & Chi, P. Chromatin remodeling and cancer, Part I: Covalent histone modifications. Trends Mol Med 13, 363-372, doi:10.1016/j.molmed.2007.07.003 (2007). 42 Misteli, T. Higher-order genome organization in human disease. Cold Spring Harb Perspect Biol 2, a000794, doi:10.1101/cshperspect.a000794 (2010). 43 Ellis, L., Atadja, P. W. & Johnstone, R. W. Epigenetics in cancer: targeting chromatin modifications. Mol Cancer Ther 8, 1409-1420, doi:10.1158/1535- 7163.MCT-08-0860 (2009). 44 Ganesan, A., Nolan, L., Crabb, S. J. & Packham, G. Epigenetic therapy: histone acetylation, DNA methylation and anti-cancer drug discovery. Curr Cancer Drug Targets 9, 963-981 (2009). 45 Matthews, A. G. et al. RAG2 PHD finger couples histone H3 lysine 4 trimethylation with V(D)J recombination. Nature 450, 1106-1110, doi:10.1038/nature06431 (2007). 46 Ramon-Maiques, S. et al. The plant homeodomain finger of RAG2 recognizes histone H3 methylated at both lysine-4 and arginine-2. Proc Natl Acad Sci U S A 104, 18993-18998, doi:10.1073/pnas.0709170104 (2007). 47 Hayashi, K., Yoshida, K. & Matsui, Y. A histone H3 methyltransferase controls epigenetic events required for meiotic prophase. Nature 438, 374-378, doi:10.1038/nature04112 (2005). 48 Mihola, O., Trachtulec, Z., Vlcek, C., Schimenti, J. C. & Forejt, J. A mouse speciation gene encodes a meiotic histone H3 methyltransferase. Science 323, 373-375, doi:10.1126/science.1163601 (2009).

53 49 Iwase, S. et al. The X-linked mental retardation gene SMCX/JARID1C defines a family of histone H3 lysine 4 demethylases. Cell 128, 1077-1088, doi:10.1016/j.cell.2007.02.017 (2007). 50 Qi, H. H. et al. Histone H4K20/H3K9 demethylase PHF8 regulates zebrafish brain and craniofacial development. Nature 466, 503-507, doi:10.1038/nature09261 (2010). 51 Rine, J. & Herskowitz, I. Four genes responsible for a position effect on expression from HML and HMR in Saccharomyces cerevisiae. Genetics 116, 9-22 (1987). 52 Aparicio, O. M., Billington, B. L. & Gottschling, D. E. Modifiers of position effect are shared between telomeric and silent mating-type loci in S. cerevisiae. Cell 66, 1279-1287 (1991). 53 Gottlieb, S. & Esposito, R. E. A new role for a yeast transcriptional silencer gene, SIR2, in regulation of recombination in ribosomal DNA. Cell 56, 771- 776 (1989). 54 Kaeberlein, M., McVey, M. & Guarente, L. The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms. Genes & development 13, 2570-2580 (1999). 55 Tanny, J. C., Dowd, G. J., Huang, J., Hilz, H. & Moazed, D. An enzymatic activity in the yeast Sir2 protein that is essential for gene silencing. Cell 99, 735-745 (1999). 56 Imai, S., Armstrong, C. M., Kaeberlein, M. & Guarente, L. Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase. Nature 403, 795-800, doi:10.1038/35001622 (2000). 57 Landry, J. et al. The silencing protein SIR2 and its homologs are NAD- dependent protein deacetylases. Proc Natl Acad Sci U S A 97, 5807-5811, doi:10.1073/pnas.110148297 (2000). 58 Smith, J. S. et al. A phylogenetically conserved NAD+-dependent protein deacetylase activity in the Sir2 protein family. Proc Natl Acad Sci U S A 97, 6658-6663 (2000). 59 Lin, S. J., Defossez, P. A. & Guarente, L. Requirement of NAD and SIR2 for life-span extension by calorie restriction in Saccharomyces cerevisiae. Science 289, 2126-2128 (2000). 60 Kaeberlein, M., Kirkland, K. T., Fields, S. & Kennedy, B. K. Sir2-independent life span extension by calorie restriction in yeast. PLoS biology 2, E296, doi:10.1371/journal.pbio.0020296 (2004). 61 Fabrizio, P. et al. Sir2 blocks extreme life-span extension. Cell 123, 655-667, doi:10.1016/j.cell.2005.08.042 (2005). 62 Tissenbaum, H. A. & Guarente, L. Increased dosage of a sir-2 gene extends lifespan in Caenorhabditis elegans. Nature 410, 227-230, doi:10.1038/35065638 (2001). 63 Rogina, B. & Helfand, S. L. Sir2 mediates longevity in the fly through a pathway related to calorie restriction. Proc Natl Acad Sci U S A 101, 15998- 16003, doi:10.1073/pnas.0404184101 (2004).

54 64 Burnett, C. et al. Absence of effects of Sir2 overexpression on lifespan in C. elegans and Drosophila. Nature 477, 482-485, doi:10.1038/nature10296 (2011). 65 Frye, R. A. Phylogenetic classification of prokaryotic and eukaryotic Sir2-like proteins. Biochem Biophys Res Commun 273, 793-798, doi:10.1006/bbrc.2000.3000 (2000). 66 Yu, J. & Auwerx, J. The role of sirtuins in the control of metabolic homeostasis. Ann N Y Acad Sci 1173 Suppl 1, E10-19, doi:10.1111/j.1749- 6632.2009.04952.x (2009). 67 Brunet, A. et al. Stress-dependent regulation of FOXO transcription factors by the SIRT1 deacetylase. Science 303, 2011-2015, doi:10.1126/science.1094637 (2004). 68 McBurney, M. W. et al. The mammalian SIR2alpha protein has a role in embryogenesis and gametogenesis. Mol Cell Biol 23, 38-54 (2003). 69 Cheng, H. L. et al. Developmental defects and p53 hyperacetylation in Sir2 homolog (SIRT1)-deficient mice. Proc Natl Acad Sci U S A 100, 10794-10799, doi:10.1073/pnas.1934713100 (2003). 70 Deng, C. X. SIRT1, is it a tumor promoter or tumor suppressor? Int J Biol Sci 5, 147-152 (2009). 71 Fang, Y. & Nicholl, M. B. Sirtuin 1 in malignant transformation: friend or foe? Cancer Lett 306, 10-14, doi:10.1016/j.canlet.2011.02.019 (2011). 72 Chen, D., Steele, A. D., Lindquist, S. & Guarente, L. Increase in activity during calorie restriction requires Sirt1. Science 310, 1641, doi:10.1126/science.1118357 (2005). 73 Chen, D. et al. Tissue-specific regulation of SIRT1 by calorie restriction. Genes & development 22, 1753-1757, doi:10.1101/gad.1650608 (2008). 74 North, B. J., Marshall, B. L., Borra, M. T., Denu, J. M. & Verdin, E. The human Sir2 ortholog, SIRT2, is an NAD+-dependent tubulin deacetylase. Mol Cell 11, 437-444 (2003). 75 Lombard, D. B. et al. Mammalian Sir2 homolog SIRT3 regulates global mitochondrial lysine acetylation. Mol Cell Biol 27, 8807-8814, doi:10.1128/MCB.01636-07 (2007). 76 Hirschey, M. D. et al. SIRT3 regulates mitochondrial fatty-acid oxidation by reversible enzyme deacetylation. Nature 464, 121-125, doi:10.1038/nature08778 (2010). 77 Haigis, M. C. et al. SIRT4 inhibits glutamate dehydrogenase and opposes the effects of calorie restriction in pancreatic beta cells. Cell 126, 941-954, doi:10.1016/j.cell.2006.06.057 (2006). 78 Nakagawa, T., Lomb, D. J., Haigis, M. C. & Guarente, L. SIRT5 Deacetylates carbamoyl phosphate synthetase 1 and regulates the urea cycle. Cell 137, 560- 570, doi:10.1016/j.cell.2009.02.026 (2009). 79 Mostoslavsky, R. et al. Genomic instability and aging-like phenotype in the absence of mammalian SIRT6. Cell 124, 315-329, doi:10.1016/j.cell.2005.11.044 (2006).

55 80 Michishita, E., Park, J. Y., Burneskis, J. M., Barrett, J. C. & Horikawa, I. Evolutionarily conserved and nonconserved cellular localizations and functions of human SIRT proteins. Mol Biol Cell 16, 4623-4635 (2005). 81 Liszt, G., Ford, E., Kurtev, M. & Guarente, L. Mouse Sir2 homolog SIRT6 is a nuclear ADP-ribosyltransferase. J Biol Chem 280, 21313-21320, doi:10.1074/jbc.M413296200 (2005). 82 Michishita, E. et al. SIRT6 is a histone H3 lysine 9 deacetylase that modulates telomeric chromatin. Nature 452, 492-496, doi:10.1038/nature06736 (2008). 83 Michishita, E. et al. Cell cycle-dependent deacetylation of telomeric histone H3 lysine K56 by human SIRT6. Cell Cycle 8, 2664-2666 (2009). 84 Yang, B., Zwaans, B. M., Eckersdorff, M. & Lombard, D. B. The sirtuin SIRT6 deacetylates H3 K56Ac in vivo to promote genomic stability. Cell Cycle 8, 2662-2663 (2009). 85 Kawahara, T. L. et al. SIRT6 links histone H3 lysine 9 deacetylation to NF- kappaB-dependent gene expression and organismal life span. Cell 136, 62-74, doi:10.1016/j.cell.2008.10.052 (2009). 86 Zhong, L. et al. The histone deacetylase Sirt6 regulates glucose homeostasis via Hif1alpha. Cell 140, 280-293, doi:10.1016/j.cell.2009.12.041 (2010). 87 McCord, R. A. et al. SIRT6 stabilizes DNA-dependent protein kinase at chromatin for DNA double-strand break repair. Aging (Albany NY) 1, 109-121 (2009). 88 Kaidi, A., Weinert, B. T., Choudhary, C. & Jackson, S. P. Human SIRT6 promotes DNA end resection through CtIP deacetylation. Science 329, 1348- 1353, doi:10.1126/science.1192049 (2010). 89 Schwer, B. et al. Neural sirtuin 6 (Sirt6) ablation attenuates somatic growth and causes obesity. Proc Natl Acad Sci U S A 107, 21790-21794, doi:10.1073/pnas.1016306107 (2010). 90 Xiao, C. et al. SIRT6 deficiency results in severe hypoglycemia by enhancing both basal and insulin-stimulated glucose uptake in mice. J Biol Chem 285, 36776-36784, doi:10.1074/jbc.M110.168039 (2010). 91 Ford, E. et al. Mammalian Sir2 homolog SIRT7 is an activator of RNA polymerase I transcription. Genes Dev 20, 1075-1080 (2006). 92 Vakhrusheva, O. et al. Sirt7 Increases Stress Resistance of Cardiomyocytes and Prevents Apoptosis and Inflammatory Cardiomyopathy in Mice. Circ Res (2008). 93 Lombard, D. B., Schwer, B., Alt, F. W. & Mostoslavsky, R. SIRT6 in DNA repair, metabolism and ageing. J Intern Med 263, 128-141 (2008). 94 Ashraf, N. et al. Altered sirtuin expression is associated with node-positive breast cancer. British journal of cancer 95, 1056-1061, doi:10.1038/sj.bjc.6603384 (2006). 95 Frye, R. "SIRT8" expressed in thyroid cancer is actually SIRT7. British journal of cancer 87, 1479, doi:10.1038/sj.bjc.6600635 (2002). 96 De Nigris, F. et al. Isolation of a SIR-like gene, SIR-T8, that is overexpressed in thyroid carcinoma cell lines and tissues. British journal of cancer 87, 1479, doi:10.1038/sj.bjc.6600636 (2002).

56 97 Kaeberlein, M. et al. Substrate-specific activation of sirtuins by resveratrol. J Biol Chem 280, 17038-17045, doi:10.1074/jbc.M500655200 (2005). 98 Borra, M. T., Smith, B. C. & Denu, J. M. Mechanism of human SIRT1 activation by resveratrol. J Biol Chem 280, 17187-17195, doi:10.1074/jbc.M501250200 (2005). 99 Beher, D. et al. Resveratrol is not a direct activator of SIRT1 enzyme activity. Chem Biol Drug Des 74, 619-624, doi:10.1111/j.1747-0285.2009.00901.x (2009). 100 Blander, G. et al. SIRT1 shows no substrate specificity in vitro. J Biol Chem 280, 9780-9785, doi:10.1074/jbc.M414080200 (2005). 101 Wang, Z. et al. Combinatorial patterns of histone acetylations and methylations in the human genome. Nat Genet 40, 897-903 (2008). 102 Manuyakorn, A. et al. Cellular histone modification patterns predict prognosis and treatment response in resectable pancreatic adenocarcinoma: results from RTOG 9704. J Clin Oncol 28, 1358-1365. 103 Seligson, D. B. et al. Global levels of histone modifications predict prognosis in different cancers. Am J Pathol 174, 1619-1628 (2009). 104 Ferrari, R. et al. Epigenetic reprogramming by adenovirus e1a. Science 321, 1086-1088 (2008). 105 Horwitz, G. A. et al. Adenovirus small e1a alters global patterns of histone modification. Science 321, 1084-1085 (2008). 106 Leal, J. F. et al. Cellular senescence bypass screen identifies new putative tumor suppressor genes. Oncogene 27, 1961-1970, doi:10.1038/sj.onc.1210846 (2008). 107 Wang, X., He, C., Moore, S. C. & Ausio, J. Effects of histone acetylation on the solubility and folding of the chromatin fiber. J Biol Chem 276, 12764- 12768, doi:10.1074/jbc.M100501200 (2001). 108 Adkins, M. W., Carson, J. J., English, C. M., Ramey, C. J. & Tyler, J. K. The histone chaperone anti-silencing function 1 stimulates the acetylation of newly synthesized histone H3 in S-phase. J Biol Chem 282, 1334-1340, doi:10.1074/jbc.M608025200 (2007). 109 Masumoto, H., Hawke, D., Kobayashi, R. & Verreault, A. A role for cell- cycle-regulated histone H3 lysine 56 acetylation in the DNA damage response. Nature 436, 294-298, doi:10.1038/nature03714 (2005). 110 Daujat, S. et al. Crosstalk between CARM1 methylation and CBP acetylation on histone H3. Curr Biol 12, 2090-2097 (2002). 111 Moazed, D., Kistler, A., Axelrod, A., Rine, J. & Johnson, A. D. Silent information regulator protein complexes in Saccharomyces cerevisiae: a SIR2/SIR4 complex and evidence for a regulatory domain in SIR4 that inhibits its interaction with SIR3. Proc Natl Acad Sci U S A 94, 2186-2191 (1997). 112 Straight, A. F. et al. Net1, a Sir2-associated nucleolar protein required for rDNA silencing and nucleolar integrity. Cell 97, 245-256 (1999). 113 Bryne, J. C. et al. JASPAR, the open access database of transcription factor- binding profiles: new content and tools in the 2008 update. Nucleic Acids Res 36, D102-106 (2008).

57 114 Galang, C. K., Muller, W. J., Foos, G., Oshima, R. G. & Hauser, C. A. Changes in the expression of many Ets family transcription factors and of potential target genes in normal mammary tissue and tumors. J Biol Chem 279, 11281-11292, doi:10.1074/jbc.M311887200 (2004). 115 Kaikkonen, S., Makkonen, H., Rytinki, M. & Palvimo, J. J. SUMOylation can regulate the activity of ETS-like transcription factor 4. Biochim Biophys Acta 1799, 555-560, doi:10.1016/j.bbagrm.2010.07.001 (2010). 116 Maher, C. A. et al. Transcriptome sequencing to detect gene fusions in cancer. Nature 458, 97-101, doi:10.1038/nature07638 (2009). 117 Makkonen, H. et al. Identification of ETS-like transcription factor 4 as a novel androgen receptor target in prostate cancer cells. Oncogene 27, 4865-4876, doi:10.1038/onc.2008.125 (2008). 118 Hollenhorst, P. C., McIntosh, L. P. & Graves, B. J. Genomic and biochemical insights into the specificity of ETS transcription factors. Annu Rev Biochem 80, 437-471, doi:10.1146/annurev.biochem.79.081507.103945 (2011). 119 Braithwaite, A. W. et al. Adenovirus-induced alterations of the cell growth cycle: a requirement for expression of E1A but not of E1B. J Virol 45, 192-199 (1983). 120 Amsterdam, A. et al. Many ribosomal protein genes are cancer genes in zebrafish. PLoS Biol 2, E139 (2004). 121 Ebert, B. L. et al. Identification of RPS14 as a 5q- syndrome gene by RNA interference screen. Nature 451, 335-339 (2008). 122 Steeg, P. S. et al. Evidence for a novel gene associated with low tumor metastatic potential. J Natl Cancer Inst 80, 200-204 (1988). 123 Hirschey, M. D. et al. SIRT3 Deficiency and Mitochondrial Protein Hyperacetylation Accelerate the Development of the Metabolic Syndrome. Mol Cell 44, 177-190, doi:10.1016/j.molcel.2011.07.019 (2011). 124 Hansen, M. et al. Lifespan extension by conditions that inhibit translation in Caenorhabditis elegans. Aging Cell 6, 95-110 (2007). 125 Harrison, D. E. et al. Rapamycin fed late in life extends lifespan in genetically heterogeneous mice. Nature 460, 392-395 (2009). 126 Mujtaba, S., Zeng, L. & Zhou, M. M. Structure and acetyl-lysine recognition of the bromodomain. Oncogene 26, 5521-5527, doi:10.1038/sj.onc.1210618 (2007). 127 Chua, K. F. et al. Mammalian SIRT1 limits replicative life span in response to chronic genotoxic stress. Cell Metab 2, 67-76, doi:10.1016/j.cmet.2005.06.007 (2005). 128 Mendez, J. & Stillman, B. Chromatin association of human origin recognition complex, cdc6, and minichromosome maintenance proteins during the cell cycle: assembly of prereplication complexes in late mitosis. Mol Cell Biol 20, 8602-8612 (2000). 129 Hung, T. et al. ING4 mediates crosstalk between histone H3 K4 trimethylation and H3 acetylation to attenuate cellular transformation. Mol Cell 33, 248-256, doi:10.1016/j.molcel.2008.12.016 (2009).

58 130 Garcia, B. A. et al. Chemical derivatization of histones for facilitated analysis by mass spectrometry. Nat Protoc 2, 933-938 (2007). 131 Plazas-Mayorca, M. D. et al. One-pot shotgun quantitative mass spectrometry characterization of histones. J Proteome Res 8, 5367-5374 (2009). 132 Rappsilber, J., Ishihama, Y. & Mann, M. Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics. Anal Chem 75, 663-670 (2003). 133 Moqtaderi, Z. et al. Genomic binding profiles of functionally distinct RNA polymerase III transcription complexes in human cells. Nat Struct Mol Biol 17, 635-640. 134 Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137 (2008). 135 Dennis, G., Jr. et al. DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol 4, P3 (2003). 136 Huang da, W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4, 44-57 (2009). 137 Liu, X. S., Brutlag, D. L. & Liu, J. S. An algorithm for finding protein-DNA binding sites with applications to chromatin-immunoprecipitation microarray experiments. Nat Biotechnol 20, 835-839 (2002). 138 Mahony, S. & Benos, P. V. STAMP: a web tool for exploring DNA-binding motif similarities. Nucleic Acids Res 35, W253-258 (2007). 139 Dahl, J. A. & Collas, P. Q2ChIP, a quick and quantitative chromatin immunoprecipitation assay, unravels epigenetic dynamics of developmentally regulated genes in human carcinoma cells. Stem Cells 25, 1037-1046 (2007).