Intersection of Nuclear Receptors and the Proteasome on the Epigenetic Landscape

Abstract

Nuclear receptors (NRs) represent a class of transcription factors that associate with both positive and negative chromatin modifying complexes to activate or repress gene transcription. The 26S proteasome plays a major role in NR-regulated gene transcription by tightly regulating the levels of the receptor and coregulator complexes. Recent evidence suggests a robust nonproteolytic role for specific proteasome subunits in gene transcription mediated via alterations in specific histone modifications. The involvement of nuclear receptors and the proteasome with chromatin modifying complexes or proteins, particularly those that modify DNA and histone proteins, provides an opportunity to review two critical epigenetic mechanisms that control gene expression and heritable biological processes. Both nuclear receptors and the proteasome are targets of environmental factors including some which lead to epigenetic changes that can influence human diseases such as cancer. In this review, we will explore molecular mechanisms by which NR-mediated gene expression, under the control of the proteasome, can result in altered epigenetic landscapes.

BACKGROUND

Nuclear Receptors

Nuclear receptors (NR) are a large family of ligand dependent transcription factors with conserved structural and functional domains. In humans, the NR superfamily contains 48 members (NURSA, http://www.nursa.org/). The biological functions of NRs are manifested in a variety of physiological processes including, but not limited to development, reproduction, metabolism, and energy balance [Larsen et al., 2002]. These receptors can be subdivided into three major classes [Mangelsdorf et al., 1995]. Class I receptors which include estrogen (ER), progesterone (PR), glucocorticoid (GR), androgens (AR), and mineralocorticoid (MR) are the classic steroid hormone receptors (SRs) which, upon activation by ligands, bind to inverted repeat DNA sequences as homodimers [Yamamoto, 1985]. Class II nuclear receptors heterodimerize with other members of this class, primarily the 9-cis retinoic acid receptor (RXR) and are typically located in the nucleus and bind their cog-nate DNA elements in the absence of ligand. These are the receptors for thyroid hormone (TR), vitamin D (VDR), alltrans retinoic acid (RAR), and the peroxisome proliferators-activated receptor (PPAR). Class III nuclear receptors are termed “orphan receptors” because their cognate ligands are currently unknown. All members of the NR superfamily share a signature modular structure that includes an amino-terminal transcriptional activation function one (AF1) domain, a conserved zinc-finger DNA binding domain (DBD), a hinge region and a carboxyl-terminal ligand-binding domain (LBD) that overlaps with a second transcriptional activation function two (AF2) domain. To activate or repress transcription, NRs recruit both covalent and noncovalent chromatin modifying complexes to alter chromatin architecture at the target genes [Chen et al., 2006; Lonard and O'Malley, 2007]. Among prominent NR coregulators are the ATPase SWI/SNF family of chromatin remodeling complexes and histone-modifying complexes [Chen et al., 2006; Kishimoto et al., 2006; Lonard and O'Malley, 2007]. Nuclear receptors work with diverse histone-modifying complexes as both coactivators and corepressors. The coactivator complexes are represented by histone acetyltransferases (HATs) and some histone methyltransferases (HMTs), whereas complexes such as NCoR/SMRT that contain histone deacetylases (HDACS) or Co-REST which contains histone demethylases (DMTs) represent corepressor complexes [Berger, 2007; Garcia-Bassets et al., 2007; Lonard and O'Malley, 2007]. Because of their ability to recruit these chromatin modifying machines to specific DNA within chromatin, NR can impact on chromatin architecture that can be interpreted into epigenetic readouts as we will discuss later.

Nuclear receptors are of particular interest in this era of the environmental epigenome [Wade and Archer, 2006]. Nuclear receptors, especially steroid hormone receptors such as the estrogen, androgen, and glucocorticoid receptors, are activated by environmental chemicals that mimic their respective ligands. Accumulating evidence now shows that a number of these chemicals cause adverse effects on development and reproductive systems through epigenetic mechanisms that finally promote diseases like cancer (Reviewed in [Robertson, 2005; Jirtle and Skinner, 2007]).

The 26S Proteasome Holoenzyme

The proteasome holoenzyme, also known as the 26S proteasome is a multienzyme complex responsible for the ATP-dependent proteolytic degradation of cellular proteins [Coux et al., 1996; Voges et al., 1999]. The proteasome is highly conserved in eukaryotic evolution and consists of two major subcomplexes, the 19S regulatory particle and the 20S catalytic core particle [Coux et al., 1996; Voges et al., 1999]. The 19S complex has two subcomplexes, the lid and the base. The 19S base composed of six distinct AAA-family ATPases (ATPases Associated with various cellular Activities, Rpt1–Rpt6) and three non-ATPase subunits (Rpn1, 2, and 13) is positioned at both ends of the 20S, whereby the ATP-dependent interaction promotes opening of pores and provides an access channel for substrates entering the 20S [Voges et al., 1999]. The remaining non-ATPases subunits (Rpn 3−12) constitute a separate subcomplex known as the “lid.” The 20S core particle which harbors the catalytic unit of the 26S proteasome is sandwiched between two 19S regulatory particles. In eukaryotes, the 20S proteasome is composed of two copies of 14 different gene products (α1–α7 and β1–β7) arranged in four axially stacked heptameric rings (α1−7, β1−7, β1−7, α1−7) [Voges et al., 1999].

Link Between Nuclear Receptors and the 26S Proteasome

Protein Degradation

The 26S proteasome is an essential component of the ATP-dependent proteolytic pathway, which catalyzes the rapid degradation of poly-ubiquitinated protein targets like transcriptional regulators (e.g., p53) and cell cycle regulatory proteins (e.g., cyclins) [Coux et al., 1996]. The proteasome is also essential for the rapid elimination of highly abnormal proteins, which may arise by mutation or post synthetic damage [Coux et al., 1996]. Although the role of the 26S proteasome system in regulating many transcription factors such as p53 is well established, the system has only recently been linked to degradation of NRs (Reviewed in [Nawaz and O'Malley, 2004; Kinyamu et al., 2005]). Control of cellular protein levels by the ubiquitin–proteasome system is essential for various cellular functions; ultimately, dysregulation of the system is associated with many pathological conditions, some of which are tightly regulated by nuclear hormone receptors [Ciechanover, 2006].

Proteasome and Nuclear Receptor-Regulated Transcription

Studies where proteasome activity is inhibited with drugs such as MG132 imply that proteasome activity is required for activation of transcription by some, but not all NRs [Lonard et al., 2000; Wallace and Cidlowski, 2001; Deroo et al., 2002; Dennis et al., 2005]. This has led to a widely accepted model whereby proteolytic activity of the proteasome is believed to be critical for promoting the exchange of transcriptional factors on chromatin, possibly facilitating multiple rounds of transcription initiation, hence controlling receptor-mediated gene expression [Reid et al., 2003; Nawaz and O'Malley, 2004; Collins and Tansey, 2006]. Consequently, receptor turnover is tightly linked to receptor-mediated transcription. On the other hand, it is clear that specific proteasome subunits like the19S ATPase thyroid interacting protein 1 (TRIP1/ Sug1) and the 20S beta subunit low molecular mass poly-peptide 2 (LMP2) or alpha subunit PMSA3 (HC8) occupy promoters on genes regulated by nuclear receptors. Further, it is apparent that they contribute to receptor-mediated gene transcription by yet unknown mechanisms that may be separate from the proteolytic function of the proteasome [Lee et al., 1995; Reid et al., 2003; Perissi et al., 2004; Zhang et al., 2006; Kinyamu and Archer, 2007].

An emerging theme in transcription is the contribution that epigenetics makes to both transcriptional competence and transcriptional regulation. In the course of this review, we will explore whether the newly attributed functions for the proteasome in transcription might function via epigenetic mechanisms. However, before elaborating on the evidence for an epigenetic explanation of the NR-proteasome effects on transcription, let us first develop a definition for epigenetics.

EPIGENETIC LANDSCAPE

What is Epigenetics?

Recently, a great effort has been dedicated to the understanding of epigenetics [Allis et al., 2007; Bird, 2007; Goldberg et al., 2007]. Concomitantly, there is a renewed interest in the phenomena and what the word itself means within the context of biological processes [Wade and Archer, 2006; Bird, 2007; Goldberg et al., 2007; Ptashne, 2007]. The term “epigenetics” was introduced in the 1940s by the developmental biologist Conrad H. Waddington, which he defined as the interactions of genes with their environment, or how genotype gives rise to phenotype during development [Van Speybroeck, 2002; Goldberg et al., 2007]. The original meaning has digressed to the current definition: “epigenetics constitutes mechanisms by which stable and heritable changes in gene expression can occur without altering the DNA sequence” [Van Speybroeck, 2002; Goldberg et al., 2007]. In a nutshell, gene expression changes in the absence of alterations in the DNA sequence are regulated by changes in chromatin modifications (modifications of DNA and histones) and interactions between noncoding RNAs with the genome [Bird, 2007; Goldberg et al., 2007]. Epigenetic mechanisms are exemplified in several systems where cellular phenotypes are transmitted from generation to generation without alteration in DNA sequence. These include embryonic development, X-chromosome inactivation, genomic imprinting, heterochromatic silencing of the mating loci in yeast and developmental reprogramming in Drosophila (reviewed in [Bird, 2007]). Studies within these systems established chromatin based modifications of DNA and histones that correlate with a particular phenotype and define the epigenetic landscape. Emerging evidence causally links epigenetic alterations of chromatin to perturbed biological functions, many of which lead to diseases like cancer, and are regulated by NRs and the proteasome.

Within the eukaryotic cell nucleus, DNA is highly compacted and organized into chromatin by both histone and nonhistone proteins. The fundamental unit of chromatin is the nucleosome which is composed of ∼146 base pairs of DNA wrapped around an octamer containing 2 molecules of each the core histones (H3, H4, H2A, H2B) [Kornberg, 1974]. The four histone proteins consist of distinct amino-terminal tails subject to numerous posttranslational modifications [Jenuwein and Allis, 2001; Berger, 2007; Kouzarides, 2007]. A combination of these modifications can alter biological functions by dictating chromatin architecture of a particular loci in two ways: First, the modification or combination of different modifications can directly alter chromatin packing, for example by changing the electrostatic charge or internucleosomal contacts, hence loosening chromatin to allow DNA binding chromatin modifiers (Reviewed in [Berger, 2007; Kouzarides, 2007]). Secondly, a specific modification permits recruitment of effector protein complexes that modify chromatin to modulate biological function [Daniel et al., 2005; Berger, 2007]. As such, the current view of epigenetics encompasses the activities of numerous enzymes and effector proteins that modulate chromatin architecture. Hence, many favor the idea that a combination of covalent modifications of DNA and histones can create an epigenetic landscape [Berger, 2007; Bernstein et al., 2007; Bird, 2007; Goldberg et al., 2007].

Chromatin Modifications Associated With Epigenetic Landscape

DNA Methylation

DNA methylation is perhaps the best characterized chemical modification of chromatin. In mammals, DNA is covalently modified by methylation of the cytosine base in the dinucleotide sequence “5-CpG-3” [Bird, 2002; Goll and Bestor, 2005)]. Cytosine methylation patterns are established and maintained by a conserved group DNA (cytosine 5) methyltransferases (DNMTs) together with methyl-CpG binding proteins which are involved in reading the methylation mark [Wade, 2001; Goll and Bestor, 2005]. After de novo establishment of methyl groups on genomic DNA by DNMT3A and DNMT3B in conjunction with the cofactor DNMT3L, the DNMT1 maintenance enzyme preserves the methylation at CpG sites during replication (reviewed in [Goll and Bestor, 2005]). Once established, the biological effects of the methyl cytosine modification are interpreted by a number of methyl CpG binding proteins, namely, Methyl CpG binding protein 1 (MeCP1), MeCP2, and methyl CpG binding domain (MBD), MBD1, MBD2, MBD3, and MBD4 [Bird and Wolffe, 1999; Wade, 2001; Klose and Bird, 2006]. In mammals, DNA methylation is implicated in a diverse range of cellular functions and pathologies, including tissue-specific gene expression, cell differentiation, genomic imprinting, X-chromosome inactivation, regulation of chromatin structure, carcinogenesis, and aging [Bird, 2002]. Dysregulation of DNA methylation patterns, including genome-wide loss or gain of DNA methylation is a characteristic of many pathologies including hormone receptor regulated cancers [Fraga et al., 2005; Robertson, 2005; Jirtle and Skinner, 2007; Jones and Baylin, 2007].

Histone Modifications

Apart from DNA methylation, changes in chromatin structure via histone modifications may contribute to epigenetic mechanisms. The core histones are subject to over 100 posttranslational modifications that are now proposed to regulate the epigenetic landscape [Berger, 2007; Kouzarides, 2007]. The amino-terminal tails of core histones are exposed on the nucleosome surface where specific residues are subject to posttranslational modifications, including lysine acetylation, ubiquitylation and sumoylation, lysine/arginine methylation, serine/threonine phosphorylation, ADP ribosylation, deimination, and proline isomerization [Berger, 2007; Kouzarides, 2007].

There is compelling evidence supporting the heritable potential of specific histone modifications, although the mechanisms by which the phenotypic effects are transmitted still remain obscure. A plausible idea to explain heritable epigenetic changes via histone modification would be to define a link between a specific histone modification and DNA methylation. High levels of histone acetylation are normally correlated with gene activity, whereas hypoacetylated histones are found in transcriptionally inactive euchromatic and heterochromatic regions [Wolffe, 1998; Li et al., 2007]. A key link between histone acetylation and deacetylation was established by the seminal studies showing that methyl DNA binding proteins bind to methylated CpG islands and recruit histone deacetylases (HDACs) to effect gene silencing [Jones et al., 1998; Nan et al., 1998]. Conversely, a recent study has shown that acetylation of histone H3-K9, an active mark, is inversely correlated with DNA methylation [Wu et al., 2007]. While the mechanism by which the chromatin state formed via these interactions is propagated is not clear, histone acetylation/deacetylation seems to be a part of the epigenetic network.

Histone lysine methylation is also implicated in epigenetic mechanisms. As shown by a number of studies the polycomb (PcG) and trithorax (Trx) systems that catalyze methylation of histone H3-K27 and H3-K4, respectively, maintain repressive and active states of chromatin, through several generations in order, to regulate lineagespecific gene expression in Drosophila melanogaster [reviewed in [Ringrose and Paro, 2007; Schuettengruber et al., 2007].

Posttranslational modification of chromatin by histone methylation has wide-ranging effects on various nuclear functions including transcription, genome integrity, and potentially epigenetic inheritance [Berger, 2007; Kouzarides, 2007; Martin and Zhang, 2007]. Histone ly-sine methylation is particularly complex since it occurs in multiple states, mono, di, or tri each providing an additional, and possibly unique, layer of epigenetic information, since each modification state is capable of generating different biological outcomes. Furthermore, whereas histone lysine acetylation almost always correlates with an open and active chromatin state, lysine methylation has divergent effects on chromatin state and function depending on the modified residue of the specific histone. Methylation at lysine 4, 36, or 79 of histone H3 is correlated with active transcription, whereas methylation at lysine 9 and 27 is correlated with gene repression [Martin and Zhang, 2005; Kouzarides, 2007]. The complexity further extends with the degree of methylation at the specific residue. For example, the trimethyl H3-K4 is found at the 5′ end of active genes whereas the di-methyl H3-K4 is found along the whole gene on inactive or poised genes [Schneider et al., 2004; Bernstein et al., 2005].

Consequently, a major focus of this review explores putative mechanistic links between histone lysine acetylation/methylation and epigenetic regulation of biological processes. This is particularly relevant because histone ly-sine acetylation and methylation play a significant role in NR regulated gene expression and function [Metzger et al., 2005; Mo et al., 2006; Aoyagi and Archer, 2007; Garcia-Bassets et al., 2007; Kinyamu and Archer, 2007; Wang et al., 2007; Wissmann et al., 2007]. These histone modifications are also of special interest because of their established links to the proteasome and gene transcription [Ezhkova and Tansey, 2004; Lee et al., 2005; Kinyamu and Archer, 2007; Laribee et al., 2007].

Evidence for a Link Between Histone H3-K4 Methylation and DNA Methylation

The methyl H3-K4 modification is a prominent mark with proven epigenetic relevance, since the trithorax (Trx) complex that regulates this mark was first identified as a regulator of Hox gene expression, transcription factors that specify cell-lineage during development in Drosophila melanogaster (reviewed in [Muller and Kassis, 2006; Schwartz and Pirrotta, 2007]). In mammals, H3-K4 methylation is accomplished by a number of SET domain enzymes, including mixed lineage leukemia family of proteins (MLL), SET, and MYND domain containing 3 (SMYD3), SET1, and SET7/9. Mice expressing a deletion of the SET domain in MLL show a dramatic decrease in histone H3-K4 methylation and DNA methylation defects at specific Hox loci [Terranova et al., 2006]. Other studies also find a strong positive correlation between the presence of histone H3-K4 methylation and the density of CpG dinucleotides in the DNA sequence [Bernstein et al., 2006; Weber et al., 2007]. The strongest link between DNA methylation and histone H3-K4 methylation comes from a recent study demonstrating that the histone octamer is a part of the DNMT3L complex. DNMT3L specifically binds to the unmethylated tail of histone H3-K4 to induce de novo DNA methylation by recruiting DNMT3A2 to the target DNA sequence [Ooi et al., 2007]. Another study supports the idea that methylation of histone H3-K4 protects gene promoter sequences from DNA methylation in somatic cells [Weber et al., 2007]. Proving that histone modification is a part of an epigenetic mechanism requires a link between the modifying enzyme and the biological response. This could be challenging in humans, for example where multiple enzymes are responsible for one modification and because these enzymes are also tissue-specific [Shi and Whetstine, 2007]. Additionally, until recently the existence of enzymes that remove the methyl mark (demethylases) was contentious. However, recent discovery of the demethylases, lysine specific demethylase 1 (LSD1), and the Jumonji domain family of proteins changes the idea that lysine methylation is a stable epigenetic mark [Klose and Zhang, 2007; Shi and Whetstine, 2007]. Still there is evidence suggesting that histone lysine methylation can regulate DNA methylation patterns and vise versa [Bernstein et al., 2006; Weber et al., 2007].

Translating the Histone Modification Into an Epigenetic Signal: The Readers

Whereas DNA methylation patterns are inherited during replication and cell division, it is not clear how histone modifications are propagated through cell division to mediate a chromatin state that results in a specific biological process or phenotype [Bird, 2007]. If epigenetic memory is mediated by one or more of the histone modifications, then there should be a mechanism for the transmission of such modifications onto the chromatin of newly replicated DNA. One way histone modifications can be passed from generation to generation is by incorporation of histone-modifying enzymes and binding effector proteins that maintain the pattern of chromatin marks in much the same way heterochromatin spreading is maintained [Grewal and Moazed, 2003]. Determining how localized enzymatic activities recognize and propagate the patterns of covalent histone marks to nearby nucleosomes and from generation to generation is challenging, although progress is being made in identifying readers and interpreters of some of the histone modifications. For example, methylation at H3-K9 creates a binding site for the heterochromatin protein 1 (HP1) which according to a recent study targets DNMT1 enzymatic activity to euchromatic sites acting as a reader of DNA methylation to effect gene silencing [Smallwood et al., 2007]. Similar to HP1, the chromodomain of polycomb protein EZH2 which binds methyl H3-K27 ties together two essential epigenetic systems involved in heritable repression of gene activity (Polycomb group (PcG) proteins and the DNA methylation). A recent study finds that PcG protein EZH2 (Enhancer of Zeste homolog 2) interacts-within the context of the Polycomb repressive complexes 2 and 3 (PRC2/3)-with DNA methyltransferases (DNMTs) and associates with DNMT activity in vivo, leading to EZH2-depedent gene repression [Vire et al., 2006]. These studies link two repressive histone modifications (methyl H3-K9 and H3-K27) and DNA methylation.

The effectors of the methyl H3-K4 mark are diverse [Daniel et al., 2005; Sims and Reinberg, 2006; Ruthenburg et al., 2007]. Of interest is the ATP chromatin remodeling complex nucleosome remodeling factor (NURF) which recognizes H3K4me2/3 through the plant homeodomain (PHD) finger of its largest subunit the bromodomain PHD transcription factor (BPTF) [Wysocka et al., 2006]. Depletion of either WD40-repeat protein,WDR5 (a component of HMT MLL1), or the BPTF subunit cause similar phenotype, including axial deformities and gut patterning defects in Xenopus embryos. This observation links the MLL enzyme and the effector protein to a specific phenotype in vivo which is viewed as a prerequisite for epigenetic inheritance. The fact that the histone modification can be translated to a phenotypic effect interpreted by an effector protein suggests lysine methylation has heritable epigenetic potential. The issue that remains, however, is whether the modification pattern that is inherited by the daughter chromatin is sufficient to impose the correct chromatin structure originating from the mother cell. However, the recent studies demonstrating that methylation of H3-K4 protects DNA sequences from DNA methylation, whereas unmethylated H3-K4 interacts with DNMTs to induce de novo DNA methylation supports a role of methyl H3-K4 in epigenetic mechanisms [Ooi et al., 2007; Weber et al., 2007].

Can histone lysine methylation dictate the memory of chromatin structure? Until recently, lysine methyl marks were thought to be “stable” and “permanent”; however, this dogma has been challenged by the discovery of histone demethylase enzymes which catalyze the removal of the methyl group from specific residues [Klose et al., 2006; Shi and Whetstine, 2007]. A question arises as to whether the resulting modifications created by these enzymes can impact on chromatin structure and the epigenetic landscape.

Two classes of lysine demethylases have recently been identified. The first class is the lysine-specific demethylase 1(LSD1/BHC110), which demethylates H3-K4me1/2 [Shi and Whetstine, 2007]. A role of LSD1 in epigenetic regulation is suggested by recent studies showing that LSD1 can regulate specific euchromatic genes and also functions in the organization of higher-order chromatin structure, resulting in definite phenotypic states in various organisms [Di Stefano et al., 2007; Garcia-Bassets et al., 2007; Lan et al., 2007; Wang et al., 2007].

The second class is the Jumonji protein family of demethylases which is characterized by JmJC domain [Trewick et al., 2005]. Four of the seven Jumonji protein subfamilies JHDM1, JHDM2, JHDM3/JMJD2, and JARID possess demethylase activity [Klose et al., 2006; Shi and Whetstine, 2007]. Over expression of mammalian JMJD2A, JMJD2B, JMJD2C, and JMJD2D decreases global levels of H3K9me3 and H3K36me3 [Cloos et al., 2006; Fodor et al., 2006; Klose et al., 2006; Whetstine et al., 2006]. Recent studies in multiple organisms have identified the JARID subfamily of Jumonji proteins as demethylases for H3-K4me2/3 [Christensen et al., 2007; Iwase et al., 2007; Klose et al., 2007; Lee et al., 2007a,b; Liang et al., 2007; Secombe et al., 2007; Tahiliani et al., 2007; Yamane et al., 2007]. The mammalian JARID1 family members named Jarid1a (also called Rbp2), Jarid1b (also called Plu1), Jarid1c (also called Smcx), and Jarid1d (also called Smcy) all demethylate H3-K4me2/3 [Christensen et al., 2007; Iwase et al., 2007; Klose et al., 2007; Lee et al., 2007a,b; Liang et al., 2007; Secombe et al., 2007; Yamane et al., 2007].

The JARID1 family members can facilitate alterations in an epigenetic landscape in various ways. Genetic and biochemical experiments have shown that depletion of JARID1a (RBP2) is important in regulation of Hox genes during ES cell differentiation and vulva formation in C.elegans, implicating RBP2 in a specific cell lineage [Christensen et al., 2007]. Another member of the Jarid family, JARID1C is required for rapid loss of H3-K4 trimethyl mark during X-chromosome inactivation. Mutation of JARID1C (SMCX) interferes with REST-mediated regulation of neuronal genes, resulting in X-linked mental retardation [Iwase et al., 2007; Tahiliani et al., 2007]. Furthermore, SMCX gene escapes X-chromosome inactivation and it is well known that X-chromosome inactivation is a hallmark of epigenetic mechanisms [Wu et al., 1994]. Together these seminal studies demonstrate that histone lysine demethylases can impact the epigenetic landscape. Importantly, besides having a potential to create an epigenetic landscape, both classes of demethylases (LSD1 and Jumonji) are involved in NR-mediated gene expression providing a mechanistic basis for understanding how epi-genetic mechanisms mediate NRs regulation of distinct biological processes.

Intersection With Nuclear Receptors and Epigenetic Landscape

Given the working definition of the epigenetic landscape as “chromatin based modifications of DNA and histones that correlate with a particular phenotype,” there are at least two ways NRs can intersect with this epigenetic landscape. Nuclear receptors interact with coregulator complexes which, in many cases, are chromatin modifying enzymes [Chen et al., 2006; Kishimoto et al., 2006; Lonard and O'Malley, 2007]. Histone modifications can generate short and long term effects on NR response. For example, histone acetylation is a hallmark of dynamic active transcription events that are short term and accompanied by multiple transcription initiation and reinitiation cycles at steroid hormone receptor regulated promoters [Metivier et al., 2003; Aoyagi and Archer, 2007]. Long term effects of histones can be exerted by defining and maintaining chromatin structure throughout the cell cycle. Because histone lysine methylation is a more stable mark than acetylation, long term effects of histone modifications may be exerted by histone lysine methylation. The methyl H3-K9 marks silenced regions and methyl H3-K4 generally marks transcriptionally active genes [Berger, 2007; Kouzarides, 2007]. HMTs specific for these modification are also coregulators of a number of nuclear receptors [Dreijerink et al., 2006; Lee et al., 2006; Mo et al., 2006]. For example, G9a (also known as Euchromatic HMT 2- EHMT2) and MLL2 (Mixed Lineage Leukemia 2), HMTs targeting H3-K9 and H3-K4, respectively, act as transcriptional coactivators for nuclear receptors [Lee et al., 2006; Mo et al., 2006]. In contrast, SETDB1 (SET Domain binding protein 1) another HMT for H3-K9 associates with the KAP1 (KRAB-zinc finger–associated protein 1) repressor, a component of NCoR (Nuclear Receptor Corepressor) to mediate HP1 binding on euchromatic regions which subsequently leads to transcription repression by unliganded nuclear receptors [Schultz et al., 2002; Ayyanathan et al., 2003].

The histone lysine demethylases responsible for H3-K9 and H3-K4 demethylation also modulate nuclear receptor function. Recent studies demonstrate a role of LSD1 in estrogen regulated gene expression and pituitary development [Garcia-Bassets et al., 2007; Wang et al., 2007]. Studies with LSD1 knockout mice demonstrate a role of LSD1 in cell lineage determination and differentiation during pituitary organogenesis regulated by the estrogen receptor [Wang et al., 2007]. In another study, mutation of LSD1 in Drosophila disrupts methylation of H3-K4, resulting in tissue specific defects during development; the mutants are sterile and have defects in ovary development, processes tightly regulated by steroid hormone receptors in higher organisms [Di Stefano et al., 2007].

The Jumonji family of proteins is also involved in NR-mediated gene regulation. A number of JmJC domain-containing proteins JMJD2A, JMJD2D, and JMJD1C are important for androgen receptor mediated gene regulation [Shin and Janknecht, 2007; Wolf et al., 2007]. Even more interesting is that when tethered to androgen receptor, LSD1 cooperates with the Jumonji demethylase JMJD2C [also called GASC1 (gene amplified in squamous cell carcinoma 1)] to demethylate H3-K9 and promote androgen receptor-dependent gene expression [Metzger et al., 2005; Wissmann et al., 2007]. Another member of the Jumonji family, JMJD2A is within the NCoR repressor complex and might have a role in receptor mediated transcriptional repression [Zhang et al., 2005]. Interestingly, JMJD2A is implicated in epigenetic regulation of ASCL2 (achaetescute complex homolog 2) a basic helix-loop-helix transcription factor whose mouse homolog is encoded by an imprinted gene highly expressed during the development of extra-embryonic trophoblast lineages but repressed in other tissues and is essential for proper placental development, a process regulated by steroid hormone receptors [Zhang et al., 2005]. Finally, the JARID1A (RBP2) is a coactivator of a number of nuclear hormone receptors and is of particular interest since deletion of the rbp2 homolog in C. elegans interferes with vulva development, a function tightly regulated by NRs [Chan and Hong, 2001; Christensen et al., 2007]. The involvement of histone and DNA modifying enzymes in NR-mediated gene regulation is an excellent platform for understanding epigenetic mechanisms that regulate many biological processes which when perturbed can lead to diseases like cancer. The 26S proteasome is an integral regulator of NR function and whether epigenetic modifications can play apart in this regulation is currently not clear.

Where Does the Proteasome Intersect With Histone Modifications that Form the Epigenetic Landscape?

The 26S proteasome seems to have proteolytic and nonproteolytic roles in gene regulation [Collins and Tansey, 2006]. Its role in nuclear receptor degradation and receptor-mediated gene regulation is now well characterized [Nawaz and O'Malley, 2004; Kinyamu et al., 2005]. However, there is a dearth of information as to the nonproteolytic role of the proteasome in gene regulation and the associated biological consequences, particularly in mammalian cells. What we know from other organisms, mainly yeast, is that the specific proteasome subunits are involved in various stages of transcription [Ferdous et al., 2001, 2002; Gillette et al., 2004; Lee et al., 2005; Sikder et al., 2006]. To date, only the 19S ATPases have been implicated in the regulation of chromatin [Kinyamu et al., 2005; Collins and Tansey, 2006]. A global role for 19S ATPases in modulating chromatin structure is consistent with several reports showing widespread distribution of proteasome subunits associated with chromatin at gene promoters including those regulated by steroid hormone receptors [Gonzalez et al., 2002; Perissi et al., 2004; Sikder et al., 2006; Zhang et al., 2006; Kinyamu and Archer, 2007]. Whereas many of these interactions might result in proteolytic functions for the 26S proteasome, differential occupancy of 19S and 20S proteasome subunits on DNA indicate that others represent nonproteolytic functions of the proteasome [Gonzalez et al., 2002; Perissi et al., 2004; Sikder et al., 2006; Zhang et al., 2006; Kinyamu and Archer, 2007]. Chromatin modification could represent a nonproteolytic function of the proteasome. 19S ATPases modulate chromatin modifications by regulating nucleosomal histone modifications, specifically histone H3-K4 methylation and H2B ubiquitylation [Ezhkova and Tansey, 2004; Laribee et al., 2007]. The mechanism by which the 19S regulates chromatin modification is not entirely clear, but recent studies provide some clues. The 19S ATPase components interact physically with the SAGA complex and stimulate SAGA/activator interactions thereby facilitating SAGA recruitment and histone H3 acetylation at active promoters [Lee et al., 2005]. Yeast expressing sug1−25 mutant which lack the ATPase domain show decreased levels of H3-K9 and H3-K14 acetylation. Furthermore, the ATPase domain of SUG1 required for interaction with SAGA and histone H3-actetylation is also required for histone H3-K4 dime-thylation, thereby linking regulation of these two histone modifications to 19S subunits [Ezhkova and Tansey, 2004]. Another recent study supports a role of the 19 S ATPase in the regulation of histone H3-K4 methylation. The CCR4/NOT mRNA processing complex and 19S proteasome subunits physically interact and loss of the CCR4/NOT complex leads to a loss of proteasome subunit Rpt6 (SUG1) and a decrease in H3-K4 tri-methyl marks [Laribee et al., 2007]. In yeast the 19S seems to set up a “trans histone tail” regulatory pathway that modulates histone H3-K9/14 acetylation and H3-K4 methylation, associating the 19S at least with active transcription. In mammalian cells, the story is only beginning to emerge. The question remains as to whether the 19S or the 20S subunits can set up a trans histone tail regulatory pathway similar to that in yeast in mammalian cells. In recent experiments, we demonstrate that proteasome inhibition and knockdown of specific 19S and 20S subunits in breast cancer cells results in an increase in MLL expression, H3-K4 (me3), and steroid hormone- dependent gene expression [Kinyamu and Archer, 2007] (Fig. 1). It is worth noting that in yeast, while SET1 is responsible for all of the histone H3-K4 methylation, humans have multiple enzymes that regulate methylation of this residue and our studies show that effects of proteasome inhibition are HMT specific [Kinyamu and Archer, 2007; Kouzarides, 2007]. Thus, it will hardly be surprising if the effects of the 19S subcomplex on chromatin modifications could be different in humans compared to yeast. For example, the human CCR4/NOT complex was in fact shown to be a repressor of nuclear receptor-mediated transcription [Winkler et al., 2006]. This type of outcome portends an exciting era of discoveries as we seek to define the precise role of proteasome subunits in regulating chromatin modifications in mammalian cells.

Fig. 1.

How does the proteasome intersect with the epigenetic landscape? A model for enhanced glucocorticoid receptor mediated transcriptional output by regulating chromatin modifications upon proteasome inhibition. (A) In absence of hormone or proteasome inhibition (CONTROL) the 19S proteasome subunit associates with model gene loci with very little to no transcription. Addition of hormone (the synthetic glucocorticoid, dexamethasone (DEX) evicts the 19S proteasome subunit, by stimulating the GR to recruit chromatin remodeling complexes (e.g., BRG-1) and receptor coactivators (e.g., p300 and SRC 1−3) to remodel chromatin, enhancing hormone dependent transcription. (B) CONTROL (similar to panel A). Blocking proteasome activity, synthetic proteasome inhibitor (MG132) in the presence of hormone (DEX) increases transcriptional output at least in part by increasing the expression of specific histone methyltransferases (HMTs, e.g., MLL) that target histone H3-lysine 4 methylation to modify chromatin. This occurs in concert with increased occupancy of other histone-modifying enzymes that correlate with transcriptionally active genes and the recruitment of the 20S proteasome subunit.

How Does the Proteasome Intersect the Epigenetic Landscape?

Current evidence suggests that the 19S ATPases are clearly involved in histone modifications (acetylation and methylation) that converse with other epigenetic systems. Perhaps, the proteasome regulates turnover of specific modifying enzyme or effector factors or specific subunits directly regulate gene expression of such proteins. Our recent studies did not separate the function of the 26S proteasome and those of particular subunits. However, we took a genomic approach where we did genomic profiling of breast cancer cells treated with proteasome inhibitor alone or in combination with the hormone dexamethasone. Among the genes profoundly altered by proteasome inhibition are those encoding HMTs, demethylases and DNMTs [Kinyamu and Archer, 2007]. Our studies reveal that proteasome inhibition modulates mRNA expression of HMTs and demethylases, particularly targeting H3-K4, H3-K9 and H3-K36. These preliminary observations lead us to believe that there is a potential link between the proteasome and epigenetic chromatin modifications in human cells. Our observations are particularly interesting because they suggest that one way proteasome inhibitors work in therapy is by disrupting chromatin modifying enzymes. Indeed, proteasome inhibitors are in clinical trials for treatment of leukemia, multiple myeloma, and breast cancer [Roccaro et al., 2006].

Are there other indications that the proteasome can cross talk with histone-modifying enzymes in mammalian cells to form an epigenetic landscape? Two recent studies showed that both knockdown of SMYD3 and LSD1 alter gene expression of proteasome subunits PSMD9 and PSMB9 (LMP2), respectively [Hamamoto et al., 2004; Scoumanne and Chen, 2007]. These observations are intriguing since LSD1 and LMP2 proteasome subunits play a role in estrogen receptor signaling [Zhang et al., 2006; Garcia-Bassets et al., 2007; Wang et al., 2007]. SMYD3 methyltransferase activity is regulated by HSP90 and this may be very well be associated with the trans-generational epigenetic effects mediated by environmental estrogens such as DES [Hamamoto et al., 2004; Ruden et al., 2005].

There are other indications that the proteasome can intersect with the epigenetic landscape. One of the new classes of JmJC demethylases (FBXL10/ JHDM1B and FBXL11/ JHDM1A) has a substrate recognition component of the SCF (SKP1-CUL1-F-box protein)-type E3 ubiquitin ligase and components of the ubiquitin-proteasome pathway [Gearhart et al., 2006; Shi and Whetstine, 2007]. These demethylases target the methyl H3-K36 mark [Shi and Whetstine, 2007]. Interestingly, another study linked the huntingtin interacting protein (HIP1), a putative H3-K36 HMT, to the degradation of the androgen receptor again connecting a histone-modifying enzyme to the Ub-proteasome system [Mills et al., 2005]. How such crosstalk works is not known, although the 19S ATPases directly communicate with ubiquitin moieties and it is possible epigenetic marks can result in this way as it is known for ubiquitylated H2B [Ezhkova and Tansey, 2004]. The fact that some of the histone-modifying enzymes are ubiquitin ligases extends the connection between the proteasome and histone modifications with potential to be epigenetic marks.

Another intriguing observation that provides a link between the proteasome and the epigenetic landscape comes from evidence demonstrating that the proteasome is required to prevent the expansion of active chromatin regions at specific loci that designate developmental and cell lineages in embryonic stem cells [Szutorisz et al., 2006]. Interestingly, these loci harbor imprinted genes, which are normally regulated by epigenetic mechanisms. Additionally, as discussed earlier, developmental and cell lineage cues are dictated by active (H3-K4 (me) and repressive marks [H3-K27 (me) and H3-K9 (me)] which are hallmarks of epigenetic regulatory mechanisms [Bird, 2007]. Although the aforementioned study did not look at the impact of the proteasome on histone modifications at these chromatin regions in ES cells, the proteasome, particularly the 19S regulates specific histone modifications at gene loci [Ezhkova and Tansey, 2004; Lee et al., 2005; Kinyamu and Archer, 2007; Laribee et al., 2007]. Additional evidence regarding the impact of the proteasome on epigenetic marks that regulate developmental and cell lineages in ES cells can strengthen the possibility that the proteasome intersects with the epigenetic landscape. The impact of the proteasome on cues that regulate developmental and cell specific lineages in ES cells may be an important factor in the pathology of many human diseases, including cancer.

The proteasome is a well-known target of environmental stimuli [Halliwell, 2006]. Systemic exposure to environmental toxins that mimic proteasome inhibitors cause neurological disorders such as Parkinson's, Alzheimer's, and Huntington's disease [McNaught et al., 2004]. Recent studies show that some of histone-modifying enzymes were increased in these disorders. For example, ESET expression and tri-methyl histone H3-K9 is markedly increased in Huntington's disease patients [Ryu et al., 2006].

An increasing body of evidence implicates the proteasome in the regulation of gene transcription through a variety of mechanisms including transcriptional activator turnover and chromatin modifications, as discussed earlier. Through such mechanisms, the proteasome can modulate expression NR dependent and independent genes, including DNA and histone-modifying enzymes. The involvement of nuclear receptors and the proteasome with chromatin modifying complexes or proteins, particularly those that modify DNA and histone proteins, creates a chromatin state with an additional layer of information that contributes to epigenetic mechanisms that control gene expression and heritable biological processes (Fig. 2). Additionally, the final phenotypic outcomes of these mechanistic liaisons are tightly dictated by numerous environmental stimuli.

Fig. 2.

The intersection of nuclear receptors and the proteasome on the epigenetic landscape. Nuclear receptors (NR) under the influence of environmental cues or natural ligands bind directly to chromatin and recruit various cofactors, chromatin remodeling complexes (CRC), histone methyltransferases (HMTs), histone acetyltransferases (HATs), histone deacetylases (HDACs), histone demethylases (DMTs), the proteasome (26S and 19S), and ubiquitin enzymes (UBE) (yellow arrows). Some of these cofactors modify DNA and specific histone residues to form the epigenetic landscape. Two systems direct epigenetic programs: DNA methylation, controlled by DNA methyltransferases (DNMTs) and readers of this mark, the methyl domain binding proteins (MBDs) and histone modifications at particular residues, established by specific HMTs, DMTs, HATs, HDACs, and propagated by effector proteins (EPa-d). A repressive chromatin state (green nucleosome) is established by the conversation between methylated DNA (CpG me) and enzymes and complexes (HMTs, DMTs,HDACs, CRC) that methylate histone H3-K27 and H3-K9 (black arrows). Regions of the genome with unmethylated DNA generally exhibit an active chromatin state (red nucleosome) marked by H3-K9/14 acetylation and H3-K4 methylation. The active chromatin state blocks DNA methylation (stop sign) and repressive chromatin state (green arrows). The proteasome (26S) plays a role in degradation NR and cofactor complexes and importantly the 19S regulatory particle forms an epigenetic landscape by establishing histone H3 and H2B trans tail communication, whereby H2B ub directs H3-K4 (me) and H3-K9/14 acetylation (blue arrows, see text for details). It is important to recall that chromatin states are dynamic, although the model is static.

The Environment and Epigenetic Mechanisms

Nuclear receptors are direct targets for modulation by environmental factors. A number of environmental toxicants are endocrine disrupting chemicals including, but not limited to, bisphenol A, DES, methoxycholor, and vinclozolin [Jirtle and Skinner, 2007]. These compounds have estrogenic or anti-androgenic activity and affect estrogen and androgen receptor target gene expression and biological function. Exposures to such chemicals have adverse effects on development of the reproductive system and can influence subsequent disease outcome. The classic example is perinatal exposure to DES and its detrimental effects on the developing reproductive system including neoplasia [Newbold et al., 2000; Huang et al., 2005]. In addition, these effects are also observed in the offspring of exposed female mice supporting the idea that these changes may be transmitted through the germ line [Newbold et al., 1998, 2000]. While the mechanism of this transmission of disease remains unclear, recent studies have revealed some clues. It is clear that these effects are mediated by the estrogen receptor (ERα), since ERα knock out mice do not exhibit these effects [Couse et al., 2000]. In addition, several studies have shown that exposure to environmental factors during development can impact the epigenetic landscape and lead to changes in gene expression that result in disease phenotypes such as cancer [Esteller, 2007; Jirtle and Skinner, 2007]. In addition, studies have now connected exposure to such compounds that mimic estrogen and androgen to phenotypic changes directed by DNA methylation of the germ line loci [Anway et al., 2005; Dolinoy et al., 2007].

Many studies suggest that prenatal and early postnatal exposure to various environmental factors, including nutritional supplements, behavioral cues, and endocrine disruptors result in altered epigenetic programming and subsequent changes in the risk of developing disease [Wade and Archer, 2006; Jirtle and Skinner, 2007; Jones and Baylin, 2007]. A recent report shows that prenatal exposure to vinclozolin and methoxychlor caused adverse effects on testis morphology and male fertility, and these effects were transmitted to subsequent generations [Anway et al., 2005]. In another recent study, maternal exposure to bisphenol A (BPA) an estrogen receptor-active compound shifted the coat color distribution of viable yellow agouti mouse offspring toward yellow by decreasing CpG (cytosine-guanine dinucleotide) methylation in an intracisternal A particle retrotransposon upstream of the Agouti gene [Dolinoy et al., 2007]. Furthermore, similar hypomethyalation effects were observed for another metastable locus, the CDK5 activator-binding protein. In support for a role of DNA methylation in the distribution of color among the offspring, maternal dietary supplementation, with either methyl donors like folic acid or the phytoestrogen genistein, negates the DNA hypomethylating effect of BPA to preserve agouti color [Dolinoy et al., 2007]. These reports provide evidence that environmental exposures can cause epigenetic alterations in the DNA, specifically hyper- and hypo-methylation and that these alterations are observed in subsequent generations [Anway et al., 2005; Dolinoy et al., 2007].

Epigenetic alterations by environmental factors can be inherited transgenerationally. An epigenetic transgenerational effect would require that epigenetic modifications in the germ line cause the inheritance of a phenotype through at least 3 generations after embryonic exposure [Anway et al., 2005]. Not all epigenetic multigenerational effects are transmitted through the germ line; epigenetic changes can be inherited mitotically in somatic cells. A number of studies in mice have demonstrated that the generation-to-generation acquisition of the maternal nurturing behaviors is passed on to the offspring directly from the mother during the first week of postnatal life [Szyf et al., 2007]. Mouse pups that receive extended nurturing (licking) from their mothers have reduced stress responses later in life. The epigenetic alterations due to maternal behavior were also shown to involve DNA methylation and histone modifications of the nerve growth factor inducible protein A binding motif in the promoter region of the brain-specific GR [Weaver et al., 2004]. These epigenetic changes are correlated with alterations in GR expression and behavioral responses to stress. Furthermore, HDAC inhibitors are able to reverse GR expression at the hypothalamus-pituitary axis and reverse the stress effects in adulthood, suggesting a dynamic interplay between histone modification and phenotypic changes or epigenetic alteration.

These data taken together suggest the complexity of the role of epigenetic modifications on overall phenotype, transmission to subsequent generations, and the potential to reverse effects once established. The importance of these chromatin modifications is only just beginning to be realized and the need for a more complete understanding of the mechanisms involved in this process is critical for our understanding of cancer and disease.

CONCLUDING REMARKS

In mammalian cells, it remains to be seen whether the potential for the proteasome to regulate chromatin architecture and the epigenetic landscape is dependent on proteolytic activity or specific proteasome subunits independent of proteolysis. Current analyses of this area have been limited to either pharmacological intervention and siRNA technologies or evaluation of a limited number of specific transcriptional targets by ChIP assays. A genomic localization analysis of specific proteasome subunits (19S and 20S) is a critical next step to deepen our understating of the proteasome in transcription. Would genes occupied by specific proteasome subunits present a specific chromatin state? Would genes encoding DNA and histone-modifying enzymes be among genes occupied by proteasome subunits? Additionally, how do environmental cues that affect proteasome or NR function impact on the epigenetic landscape and biological processes in vivo, including in mouse models? Does proteasome activity change in environmental exposure models? If there are changes in proteasome activity in these models, do these changes impact on histone and DNA modifications/chromatin structure? Understanding the crosstalk between the proteasome and histone modifications will be vital in evaluating the impact of the proteasome on the epigenetic regulation of gene expression. The potential of proteasome modulation of modifying enzymes, histone and DNA, as well as the proteins that interpret these modifications offers a broad range of epigenetic outcomes. Given the importance of the proteasome in human disease such as cancer, leukemia, neurodegenerative disorders, and viral infections, more studies on the influence of the proteasome on NRs and the epigenetic landscape will be useful in understanding environmental instructions to the epigenome.