| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Minireview |
Chromatin and Gene Expression Section, Laboratory of Molecular Carcinogenesis, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina 27709
Address all correspondence and requests for reprints to: Trevor K. Archer, Chromatin and Gene Expression Section, Laboratory of Molecular Carcinogenesis, National Institute of Environmental Health Sciences, 111 Alexander Drive, P.O. Box 12233 (MD E4-06), Research Triangle Park, North Carolina 27709. E-mail: archer1{at}niehs.nih.gov.
 |
ABSTRACT |
|---|
 TOP ABSTRACT NUCLEAR RECEPTORS (NRs)ATP-DEPENDENT CHROMATIN...ATP-DEPENDENT CHROMATIN...CONCLUSIONS AND FUTURE...REFERENCES |
|---|
|
NUCLEAR RECEPTORS (NRs) |
|---|
|
TOPABSTRACTNUCLEAR RECEPTORS (NRs)ATP-DEPENDENT CHROMATIN...ATP-DEPENDENT CHROMATIN...CONCLUSIONS AND FUTURE...REFERENCES |
|---|
All members of the NR super family share a signature modular domain structure, which consists of an amino-terminal transcriptional activation function (AF1) domain, a conserved zinc-finger DNA binding domain, a hinge region and a carboxyl-terminal ligand binding domain that overlaps with a second transcriptional AF2 domain. Whereas the AF1 activity is constitutively active in most cell types, the AF2 activity is ligand dependent (5).
Chromatin
Within the eukaryotic nuclei, DNA is complexed with histones to form a dynamic polymer termed chromatin (6). The fundamental structural unit of chromatin is the nucleosome which consists of approximately 146 bp of DNA wrapped around an octamer of histones containing two copies each of four core histone proteins: H2A, H2B, H3, and H4 (7). In addition to the four core histones, a fifth class of histone, the linker histone H1, associates with DNA between nucleosomes and may facilitate the formation of higher order chromatin organization such as chromatin fibers (8). Packaging genomic DNA into chromatin presents a significant physical barrier for regulatory proteins such as transcription factors to access their DNA target sites (9). However, unlike many transcription factors, some NRs are capable of binding to their hormone response DNA elements (HRE) embedded in chromatin (10). Upon ligand activation, NRs bind to HRE and recruit many transcriptional cofactors including chromatin remodeling factors to the target gene promoters. The chromatin remodeling activities recruited by NRs alter the local chromatin structure and allow other essential transcription factors to bind and activate or repress transcription (11, 12).
Chromatin Remodeling Factors
Two distinct classes of chromatin remodeling protein complexes have been identified. The first class is ATP-dependent complexes that are capable of directly changing the chromatin structure. They use the energy derived from ATP hydrolysis to alter the position and/or stability of nucleosomes in a noncovalent manner (13). The ATP-dependent chromatin remodeling complexes contain a core ATPase catalytic subunit that belongs to the Swi2/Snf2 superfamily of DNA helicases (14). To this date, five major families of ATP-dependent remodeling machines have been described based on the identity of their distinct ATPase subunits, including SWI/SNF, ISWI, Mi-2/NuRD, INO80, and SWR1 (15). Among them, the SWI/SNF family was the first to be discovered and since then has been the best characterized with regard to structure, function, and enzymatic activity.
The second class of chromatin remodeling complex consists of the histone modifying enzymes including histone acetyltransferases, histone deacetylases (HDACs), histone methyltransferases, kinases, and ubiquitin ligases that catalyze various posttranslational modifications of histone such as acetylation, phosphorylation, methylation, and ubiquitination (16). Current experimental evidence supports the notion that histone modifications are epigenetic markers that facilitate the recruitment of chromatin binding proteins to dictate a distinct chromatin structure (histone code hypothesis) (17). Because the role of histone modifications in transcriptional regulation has been the subject of several recent reviews, we will discuss it here only briefly (16, 18). In this review, we focus our attention on the ATP-dependent chromatin remodeling complexes, with an emphasis on their established and putative roles in NR-mediated transcriptional regulation.
|
ATP-DEPENDENT CHROMATIN REMODELING COMPLEXES |
|---|
|
TOPABSTRACTNUCLEAR RECEPTORS (NRs)ATP-DEPENDENT CHROMATIN...ATP-DEPENDENT CHROMATIN...CONCLUSIONS AND FUTURE...REFERENCES |
|---|
Orthologs of SWI/SNF complexes closely related to those characterized in yeast were subsequently discovered in humans and Drosophila (27, 28). The Brahma protein or Brm is the only Drosophila ortholog of yeast Swi2/Snf2 ATPase and functions as the catalytic ATPase in Drosophila SWI/SNF like Brm complex (27). The Brm gene was first identified as a suppressor for the polycomb (Pc) mutation (29). Polycomb family proteins are transcriptional repressors and there is evidence to suggest that Pc family proteins repress transcription by facilitating the formation of a repressive, hetero-chromatin-like structure (29). Therefore, the Brm complex presumably antagonizes Pc family protein activity and activates transcription by loosening the chromatin structure (29). The Brm complex has also been shown to be important for homeotic genes expression in Drosophila development and deletion of Brm gene is embryonic lethal (30).
The human genome contains genes encoding two yeast Swi2/Snf2 orthologs: the human Brm protein (hBrm) and the Brm-related gene 1 protein (BRG1) (31, 32). The human SWI/SNF complex contains either BRG1 or hBrm as the catalytic ATPase subunit and approximately 10–12 BRG1-associated factors (BAFs) as accessory factors (33). The human SWI/SNF complex is heterogeneous in terms of subunit composition. Most of the purified complexes contain subunits BAF170, BAF155, BAF60, BAF57, BAF53, and BAF47 (IN1/SNF5) besides BRG1 or hBrm (28). In addition, many other proteins including histone-modifying enzymes HDAC1/2, protein arginine methyltranferase 5, coactivator-associated arginine methyltransferase 1 (CARM1), transcription corepressor mSin3A, and tumor suppressor BRCA1 have been shown to copurify with human SWI/SNF complex when different purification schemes are used (34). The minimal catalytic core required for efficient chromatin remodeling in vitro consists of four subunits: BRG1, BAF155, BAF170, and BAF47 although BRG1 alone possesses some remodeling activity (35). BAF155 and BAF170 are highly homologous and probably form a heterodimer in the complex (33). In some tissues, either BAF155 or BAF170 is predominantly expressed and a homodimer may be formed (36). In humans, there are three BAF60 genes encoding three highly homologous proteins termed BAF60a, BAF60b, and BAF60c (33). Whereas BAF60a and BAF60b are expressed in many tissues, BAF60c is expressed mostly in the heart and muscle, suggesting a tissue-specific role for this subunit (37, 38). Two additional BAF subunits, BAF250 and BAF180, are present mutually exclusively in human SWI/SNF complex. The human BAF180-containing SWI/SNF complex (termed P-BAF complex) is proposed to be more related to the yeast RSC complex than the BAF250-containing human SWI/SNF complex (also known as BAF complex) in terms of subunit composition as BAF180 is homologous to three RSC-specific subunits (Rsc1, Rsc2, and Rsc4) (39, 40). The P-BAF complex has been shown to localize to kinetochores of mitotic chromosomes. The P-BAF, but not BAF complex, was able to support vitamin D receptor-dependent gene activation on a chromatinized template in vitro, suggesting a functional difference between these two variant SWI/SNF complexes. However, either BAF or P-BAF complex can support GR-dependent activation of the mouse mammary tumor virus promoter in vivo (41). Recently, the phenotype of BAF180 knockout mouse was reported (42). Deletion of BAF180 gene in mouse results in severe hypoplastic ventricle development and trophoblast placental defects. Because the expression of BAF250 and other BAF complex subunits are not affected in BAF180 null mice, these results indicate a BAF180/P-BAF complex-specific function in vivo. In the same study, a group of BAF180-regulated genes in the developmental heart was also identified by microarray analysis; however, it is not known currently whether these genes represent bona fide BAF180/P-BAF-specific targets.
The BRG1 and Brm-based complexes have slightly different chromatin remodeling activities in vitro, particularly when assayed with mononucleosome templates (43). In vivo, the two complexes may have quite distinct biological functions. Targeted disruption of BRG1 results in very early embryonic death in mouse, whereas Brm null mice are viable and fertile, although embryonic fibroblasts derived from these mice exhibit aberrant cell cycle regulation (44, 45). Indeed, recent in vitro and cellular studies indicated that Brm, rather than BRG1, was usually recruited to cell cycle regulated gene promoters (43). Targeted deletion of BAF155 subunit in mouse is also embryonic lethal (46). However, neither BRG1 nor hBrm seems to be critical for fundamental cellular functions because cells such as the human adrenal adenocarcinoma SW13, which lack both BRG1 and hBrm, survive in tissue culture (47).
The nature of the remodeled chromatin state generated by SWI/SNF complex has been extensively studied in vitro (13). Using in vitro assembled chromatin template and either purified SWI/SNF complex or the ATPase subunit alone, it was demonstrated that the SWI/SNF complex increases the mobility of the nucleosomes on DNA templates. When the mobility of a nucleosome is physically restrained, SWI/SNF complex can create a DNA loop or twist on the surface of the nucleosome. The remodeled state generated by SWI/SNF is not a static one but a mixture of several distinct nucleosome states. The energy derived from ATP hydrolysis is used to transiently disrupt DNA-nucleosome interactions (48). These observations provide a mechanistic explanation for the increased accessibility of nucleosome DNA to regulatory factors after remodeling by SWI/SNF.
ISWI Family
The ISWI (Imitation of SWI) family of ATP-dependent chromatin remodeling complexes consists of a group of heterogeneous complexes all containing the ISWI ATPase as the catalytic core (49). In addition to the highly conserved ATPase domain, the ISWI protein contains a SANT (SW13, ADA2, N-CoR, and TFIIB) domain, a potential nucleosome interaction module (50). The ISWI-based complexes were initially purified from Drosophila and at least three ISWI complexes are known in Drosophila: the NURF (nucleosome-remodeling factor) complex, the ACF (ATP-utilizing chromatin assembly and remodeling factor) complex and the CHRAC (chromatin assembly complex) complex (51). Each complex displays different accessory subunits and may have distinctive functions. The human genome encodes two ISWI orthologs, namely SNF2H and SNF2L. SNF2H is the catalytic subunit of numerous protein complexes biochemically purified from human cells. They include the hACF/WCRF (Williams syndrome transcription factor-related chromatin remodeling factor) complex, the WICH (Williams syndrome transcription factor-ISWI chromatin remodeling complex) complex, the RSF (remodeling and spacing factor) complex, the NoRC (nucleosome remodeling complex) complex, the huCHRAC complex, and the SNF2H-cohesin complex (51). In vitro, most of ISWI-based complexes have nucleosome spacing activity that results in the formation of regularly ordered nucleosome arrays (52).
Genetic studies in Drosophila suggest that ISWI based complexes can activate gene expression both in vitro and in vivo (53). However, the observation that ISWI is preferentially associated with transcriptionally silent chromatin regions in Drosophila polytene chromosome suggests that in vivo ISWI is primarily involved in transcriptional repression and gene silencing (54). The role of human ISWI based complexes in transcriptional regulation has not been extensively studied. Recently, it was found that human SNF2L is the catalytic core of a newly purified human NURF like complex that facilitates the transcriptional activation of human Engrailed, a homeodomain protein that regulates neuronal development in the brain (55).
Mi-2/NuRD Family
The Mi-2/NuRD (nucleosome remodeling and deacetylase) complexes are also highly heterogeneous in subunit composition but they all contain the dermatomyositis-specific autoantigen Mi-2 as the core ATPase (56). The Mi-2 protein contains the chromodomain, a potential methyl-lysine binding motif (57). In humans, there are two Mi-2 genes coding for two homologous proteins, namely Mi-2
and Mi-2ß (58). In addition to the Mi-2 or Mi-2ß ATPase, the complex also contains the histone modification enzyme HDAC1/2, methylated DNA binding proteins, histone H4-interacting protein RbAp46/48, and members of metastasis-associated protein gene family MTA1, MTA2, and MTA3 (56). The presence of histone-modifying enzymes in Mi-2 complex provides a unique link between Mi-2/NuRD complex and histone modification enzymes. The MTA proteins were first discovered due to their increased expression in metastatic breast cancer tissues (59). The three MTA proteins exist mutually exclusively in the Mi-2 complexes (60). Because both HDAC activity and methylated DNA are usually associated with gene repression and silencing, it is not surprising that Mi-2/NuRD complex has been involved predominantly in transcriptional repression (56).
INO80 Family and SRW1 Family
Search of the yeast genome for genes encoding proteins containing the conserved ATPase domain of Swi2/Snf2 identified two new members of the Swi2/Snf2 superfamily: the INO80 protein and SWR1 protein (61, 62). Biochemical purification revealed that both INO80 and SWR1 are the ATPases of distinct multisubunit complexes. Limited data suggest that INO80 based complexes are involved in transcriptional activation and DNA damage repair in yeast (61, 63). The SWR1-based complex is responsible for the incorporation of the histone H2A variant H2A.Z into nucleosome in a replication-independent manner (62). Histone H2A.Z is enriched at the border of heterochromatin and euchromatin in yeast and may play a role in suppressing the spread of heterochromatin (64). Thus, it appears that SWR1 complex may help establish and/or maintain transcriptionally competent chromatin domains within yeast genome. Although both the INO80 and SWR1 protein have orthologs in mammals, no biochemical characterization or functional study has been carried out on the putative mammalian INO80 or SWR1 complexes (34).
|
ATP-DEPENDENT CHROMATIN REMODELING COMPLEXES AND NR-MEDIATED TRANSCRIPTIONAL REGULATION |
|---|
|
TOPABSTRACTNUCLEAR RECEPTORS (NRs)ATP-DEPENDENT CHROMATIN...ATP-DEPENDENT CHROMATIN...CONCLUSIONS AND FUTURE...REFERENCES |
|---|
Activation of MMTV Promoter by GR
When stably integrated into mammalian cells, the MMTV promoter forms an array of six positioned nucleosomes (nucleosomes A–F) over the regulatory region including the HREs (67). Nucleosomes A and B represent the hormone sensitive region encompassing four glucocorticoid response elements and binding sites for other transcription factors such as nuclear factor 1 (NF1), octamer transcription factor (Oct1), and the TATA binding protein (TBP). In the absence of hormone, the organized chromatin structure blocks the binding of transcription factors NF1, Oct1, and TBP to the promoter (Fig. 1A
) (68). Glucocorticoid activation of the MMTV promoter is a bimodal process (69). Within 5 min of glucocorticoid treatment, GR binds to glucocorticoid response elements within chromatin and induces rapid changes in nuclear architecture at the nucleosome B region of the promoter as shown by restriction enzyme hypersensitivity assay (70). The remodeled chromatin allows NF1, OTF, and TBP to bind their respective binding sites (Fig. 1C). This series of events, initiated by the GR binding on the promoter, eventually leads to formation of the preinitiation complex and activation of the MMTV promoter (Fig. 1D) (71). In contrast, a transiently transfected MMTV promoter does not adopt an organized nucleosome structure and allows the binding of NF1 in the absence of glucocorticoids (69, 72).
|
) (41, 66). First, type II antiprogestins can inhibit GR-induced remodeling and activation of MMTV promoter in T47D cells, but the inhibition is reversed by overexpression of BRG1 protein in these cells, suggesting that PR and GR compete to recruit BRG1 to the MMTV promoter (66). Secondly, GR can coimmmunoprecipitate with BRG1 complex components in the presence of hormone, suggesting a physical interaction between GR and BRG1 complex. Experiments using a stably integrated MMTV promoter in the BRG1/Brm null SW13 cells further demonstrated unequivocally that GR-induced chromatin remodeling at the MMTV promoter is solely dependent on BRG1 complex. Neither ISWI nor Mi-2 based remodeling complexes are able to substitute for the BRG1 complex in remodeling of MMTV chromatin (41). In the same study, it was also shown that the ATPase activity of BRG1 is essential for its remodeling function. Recently, data from this and other laboratories provided additional insights into the nature of GR-initiated recruitment of the BRG1 complex to the MMTV promoter. Using MMTV promoter assembled as chromatin in vitro, it was demonstrated that the recruitment of BRG1 by GR is transient as the remodeling event actually leads to displacement of GR from the promoter (73). Because the cycling of GR has been observed on MMTV promoter in vivo (74), these results suggest a role for BRG1 complex in GR cycling on MMTV promoter. In addition, chromatin immunoprecipitation (ChIP) assay showed that GR and BRG-1 cannot stably associate with MMTV promoters in vivo in the absence of NF1 binding (72). This observation raises the possibility that chromatin structure is not a passive landing site for transcription factors but by combinatorial and cooperative mechanisms defines a range of potential responses (75). These results point to a dynamic rather than a static relationship between GR binding, chromatin remodeling and NF1 binding on the MMTV promoter (76).
The interaction of GR with BRG1 complex appears to be mediated by at least two subunits of this complex (Fig. 1B) (40, 77). GR interacts with BAF60a independent of GR ligand status (77). The GR-BAF60a interaction is required for efficient GR transactivation and chromatin remodeling at the MMTV promoter in vivo. Expression of a dominant-negative form of BAF60a disrupts GR and BRG1 complex interaction and leads to decreased chromatin remodeling and MMTV activation (77). In contrast, GR interacts with BAF250 in a ligand-dependent manner (40). BAF250 has also been shown to potentiate the transcriptional activity of GR in transient transfection assays (78). However, BAF250 seems to be dispensable for GR-mediated chromatin remodeling at the MMTV promoter given that the remodeling process takes place efficiently in the BAF250-negative, BAF180-positive T47D cells (79). Surprisingly, BAF180 is also not required for GR-induced chromatin change at the same promoter because the BAF250-positive, BAF180-negative SW13 cells support the remodeling process as long as BRG1 is present (41). Given the fact that BAF250 and BAF180 are the signature components of the BAF and P-BAF complexes, respectively, these results suggest that GR may not discriminate between BAF and PBAF complexes and can use either one of them to remodel chromatin.
Further insights into the mechanisms by which the GR and SWI/SNF cooperate to alter chromatin structure came from recent experiments that investigated the function of the ATPase subunit in the complex (80). In this study, the ATPase domains were swapped between BRG1 and hSNF2L and the activities of resultant chimeras were evaluated in vitro and in vivo. In vitro chromatin remodeling assays revealed that both chimeras were active but surprisingly, the source of the ATPase domain determined the mechanism of remodeling (13, 80). Both wild-type BRG1 and the BRG1/hSNF2L-ATPase chimera were active in vivo as evidenced by the profiles of genes induced when introduced into cells lacking BRG1. However, with respect to GR-mediated activation the chimera was unable to substitute for the wild-type BRG1 in transcriptional activity or chromatin remodeling assays, demonstrating that BRG1 and ISWI ATPase domains are not functionally interchangeable in vivo.
Activation of MMTV Promoter by Other Steroid Hormone Receptors
The MMTV promoter can also be activated by progestins, androgens, and mineralocorticoids (12). An initial report suggested that an ISWI-based complex is responsible for PR-induced chromatin remodeling in vitro (81). However, recent studies indicated that BRG1 complex, rather than ISWI complex, plays a critical role in chromatin remodeling and transcriptional activation by these receptors. In one study, the PR competes with the GR for the BRG1 complex in T47D breast cancer cells treated with dexamethasone in the presence of type II antiprogestins, suggesting that BRG1 is required for PR-dependent activation of MMTV promoter (66). In another study, ChIP analysis in T47D cells demonstrated that BRG1 is recruited to the MMTV promoter by ligand bound PR in these cells, again supporting a role for BRG1 in PR-dependent activation of the promoter (82).
BRG1 is also involved in AR-dependent chromatin remodeling and activation of the MMTV promoter. Studies aimed at dissecting the molecular mechanism of AR recruitment of BRG1 complex have employed the MMTV promoter assembled as chromatin in Xenopus oocytes as a model system (83). In this case, BRG1 is recruited to the MMTV promoter in the presence of the AR ligand R1881. In addition, ligand-dependent activation of the MMTV promoter by AR is impaired in SW13 cells, supporting a role for BRG1 complex in AR-dependent activation of MMTV promoter (83). However, in contrast to GR, a direct interaction between AR and BRG1 complex was not detected by coimmunoprecipitation assay in this study. In fact, the recruitment of BRG1 complex seems to be dependent on the p160 family of steroid hormone receptor coactivators and the histone acetyltransferase cAMP response element binding protein-binding protein (CBP).
The precise nature of GR- and PR-induced chromatin remodeling, as assessed by nuclease accessibility, remains unclear at this point. That it mostly occurs at the nucleosome B region in vivo is more consistent with a DNA looping out vs. sliding mechanism (13). An in vitro study suggested GR can induce rearrangement of the histone octamer at the nucleosome B region without measurable loss of histones (73). In addition, previous studies from our laboratory have shown that in human breast cancer cells, PR-mediated increases in MMTV transcription can occur in the absence of any structural rearrangement of the promoter as measured by restriction enzyme hypersensitivity assay (84). However, a recent report has provided some evidence to argue for a loss of histone H2A/H2B dimers on the MMTV promoter upon PR-mediated activation in a SWI/SNF complex-dependent manner (85). Currently, it is not known whether the seemingly divergent observations on the same promoter are receptor dependent or perhaps due to different experimental systems used in the studies. However, the availability of purified and well-characterized reagents and well-defined chromatin targets should allow these issues to be resolved.
Activation of pS2 Promoter by ER
The human pS2 gene promoter is estrogen responsive and the HRE as well as TATA box sequence are covered by two precisely positioned nucleosomes in human mammary epithelial cells (Fig. 2A) (86). The nucleosome structure blocks TBP binding to the TATA box (87). In MCF7 cells with endogenous ER, estrogen treatment induced two major deoxyribonuclease I (DNase I)-hypersensitive sites, one of which is located in the proximal promoter, suggesting the formation of an open chromatin structure in response to ligand (87).
|
binds to and leaves the promoter in a cycling manner. The cyclic turnover of ER on pS2 promoter is dependent on proteasome-mediated degradation of liganded receptor and has been suggested as a mechanism for allowing cells to respond continuously to changes in extracellular hormone concentration (89). A large number of transcription factors and cofactors including ATP-dependent chromatin remodeling complexes are sequentially recruited to pS2 promoter by liganded ER. The BRG1 complex is the first of these enzymatic factors to be recruited by ER, followed by the p160 family of steroid receptors coactivators, histone-modifying enzymes, or general transcription factors (Fig. 2, B and C). Interestingly, BRG1 but not Brm complex is recruited at this stage, suggesting a specific requirement for a BRG1 based complex. The BRG1 complex presumably generates a permissive chromatin structure to allow other factors to bind. Near the end of this initial activation process, BRG1 or Brm complex containing HDAC is recruited to the promoter to facilitate promoter clearance and ER departure from the promoter (Fig. 2D). After ER and the general transcription factors including TBP leave the promoter, Mi-2/NuRD complex is recruited by an unknown mechanism to generate a repressive chromatin structure within the pS2 promoter (Fig. 2E). As such, at least two distinct ATP-dependent chromatin remodeling complexes participate in this process. Whereas the Mi-2/NuRD complex is only involved in reestablishing a repressive state of the pS2 promoter, the BRG-1 complex has a dual role: opening chromatin to allow transcription factors to bind and terminating transcription via receptor clearance (88). ChIP analysis also demonstrated that BRG1 is recruited to several other endogenous ER-responsive promoters in vivo (90). Because chromatin structure at these promoters has not been fully characterized, under the experimental conditions used for the ChIP assays, the structural impact of BRG1 mediated remodeling on these promoters is not clear. Importantly, the BRG1 complex subunit BAF57 interacts with the ER and enhances ER-dependent transcriptional activation on pS2 promoter (91). This result may imply that ER recruits BRG1 complex through the BAF57 subunit although the evidence is indirect.
Receptor-Dependent and Promoter-Specific Recruitment of ATP-Dependent Chromatin Remodeling Complexes
Although not as extensively studied as the GR and ER, many other NRs use ATP-dependent chromatin remodeling complexes to regulate transcription. In contrast to GR activation of MMTV promoter, which does not distinguish between BAF and P-BAF complexes, the VDR specifically requires P-BAF complex (92). In vitro transcription assays using a chromatinized artificial VDR-responsive promoter and purified remodeling complexes indicated that P-BAF complex is essential for VDR to activate promoter embedded in chromatin. Neither the highly analogous BAF complex nor the ISWI-based ARF complex can support transcriptional activation by VDR (92). Interestingly, biochemical purification of proteins associated with VDR in a ligand-dependent manner resulted in isolation of a variant human SWI/SNF complex termed WINAC complex, which contains the Williams syndrome transcription factor (WSTF) and elongation factors in addition to the core subunits of human SWI/SNF complex (93). Surprisingly, neither the P-BAF complex-specific BAF180 nor human ISWI complex components were found in this complex. Chromatin-immunoprecipitation and transient transfection assays indicated that this complex is recruited to VDR-responsive promoters and essential for VDR-mediated transcriptional regulation. Defects in this complex negatively affect VDR signaling and account for at least some phenotypes of Williams syndrome, a rare autosomal genetic disease. Recently, BAF60a and BAF60c, two homologous subunit of human SWI/SNF complex, have been shown to differentially interact with several NRs (37, 77, 94). For example, BAF60c but not BAF60a interacts with PPAR
. Transient transcription assays indicated that BAF60c enhances PPAR-mediated transcriptional activation of its target promoters. This observation suggests that PPAR specifically recruits BAF60c containing SWI/SNF complex to activate gene expression (37).
A role for Mi-2/NURD complex in repressing NR-dependent transcription was first suggested by a study showing that the corepressor Krab-associated protein 1 (KAP-1), together with HDAC-1/2 and SWI/SNF are components of the NR corepressor complex (NCoR) (95). KAP-1 has been shown to associate with an isoform of Mi-2 protein, an integral component of the Mi-2/NURD complex (96). In another study using artificial thyroid hormone-responsive promoter assembled as chromatin in Xenopus oocytes, the Mi-2/NuRD complex was found to associate with this promoter independent of thyroid hormone receptor status (97), suggesting a global transcription repression function for the Mi-2/NuRD complex. In contrast to TR, however, a direct interaction between ER and the MTA1 subunit of Mi-2/NuRD complex has been demonstrated (98). ER can recruit MTA1-containing Mi-2/NuRD complex to its target promoters, and lead to down-regulation of ER transcriptional activity. Recently, the Mi-2ß subunit was shown to directly interact with the retinoid-related orphan receptor (ROR) and suppress ROR activity in a transient transfect assay (99). Taken together, these observations support the hypothesis that Mi-2 has both promoter-specific repression activity and a global gene silencing function.
A differential requirement for human SWI/SNF complex in AR-mediated transactivation was revealed in studies using two AR activated promoters, prostate-specific antigen (PSA) and probasin (100). Like in the case of MMTV promoter, AR activation of PSA is strictly dependent on SWI/SNF complex. Furthermore, BAF57 has been shown to directly interact with AR and is responsible for AR-dependent recruitment of SWI/SNF complex to the PSA promoter and gene activation (101). In contrast, SWI/SNF activity is not absolutely required for activating the probasin promoter. These results clearly indicate that the requirement for a specific chromatin remodeling complex in activation of AR-responsive genes is dependent on promoter architecture.
The Interplay between ATP-Dependent Chromatin Remodeling Complexes and Histone Modifying Enzyme
The presence of protein motifs recognizing histone modifications in some subunits of ATP-dependent chromatin remodeling complex implicates an intimate functional relationship between ATP-dependent remodeling complexes and histone modifying enzymes (14). This relationship is illustrated in several NR-mediated gene activation processes. The occupation of MMTV promoter by cAMP response element binding protein-binding protein (CBP)/p300 histone acetyltransferases appears to be a prerequisite for subsequent recruitment of BRG1 complex to this promoter in Xenopus oocyte (83). In addition, the presence of bromodomain in the BRG1 protein suggests that acetylated lysines may strengthen BRG1 complex binding to this promoter, although acetylated lysines alone are not sufficient for recruiting BRG1 complex (83). In the case of Mi-2 complex, the existence of HDACs in this complex suggests that histone deacetylation is inherently connected to Mi-2/NuRD-mediated chromatin remodeling. Finally, functional and mechanistic links between histone arginine methylation and ATP chromatin remodeling have been uncovered in a recent study showing association of CARM1, a histone arginine methyltransferase, with BRG1 complex components in a novel complex termed NUMAC (nucleosomal methylation activator complex) (102). In this complex, the histone methyltransferase activity of the NUMAC complex on nucleosomal histones is much higher than that of purified CARM1 alone. In addition, CARM1 stimulates the ATPase activity of BRG1. Therefore, CARM1 and BRG1 enhance each other’s enzymatic activity by forming the NUMAC complex. As a result, NUMAC complex may support ER-mediated transcriptional activation more efficiently when compared with isolated CARM1 protein or SWI/SNF complex alone (102).
Histone H1 can stabilize the nucleosome structure by limiting the mobility of nucleosomes. Phosphorylation status of histone H1 on the MMTV promoter is intimately linked to the ability of GR to activate the promoter (103). Upon prolonged hormone treatment, the promoter becomes refractory to hormone induction. This refractory state is correlated with a global dephosphorylation of histone H1 and reformation of the repressive chromatin structure at MMTV promoter (Fig. 1E). Removal of hormone restores the transcriptional competence of the MMTV promoter and reappearance of phosphorylated histone H1 accompanies this reactivation. The decreased activation upon prolonged hormone treatment is unique to the chromatin MMTV promoter because the transiently transfected MMTV promoter is not affected (103). Prolonged hormone treatment leads to a selective dephosphorylation of specific histone H1 isoforms (104). The mechanism for histone H1 de-phosphorylation induced by long-term glucocorticoid treatment is not known, and one possibility is that the hormone inhibits a histone H1 kinase either directly or indirectly (103). The cdc2/cyclin B and cdk2 kinases have been shown to phosphorylate histone H1 in vitro (105). Cdk2 is most likely the enzyme responsible for phosphorylation of histone H1 in GR-mediated transactivation of the MMTV promoter (105). Indeed, Cdk2-specific inhibitors can block GR-mediated transcriptional activation of the promoter and mimic the inhibitory effect of prolonged glucocorticoid treatment (105).
Modulation of Subunit Composition of ATP-Dependent Chromatin Remodeling Complex in NR-Mediated Transcription
As argued above, an ATP-dependent chromatin remodeling complex can achieve functional specificity by incorporation of highly specific subunits. Results from several recent studies suggest that NRs exploit this mechanism to influence transcription through regulating the expression of specific subunits. For example, the MTA3 subunit of Mi-2/NuRD complex was found to be positively regulated by estrogen in MCF-7 cells (106). The MTA3 but not MTA2 or MTA1 containing Mi-2/NuRD complex is then recruited to the promoter of transcription repressor snail, a master regulator of epithelial to mesenchymal transitions, to repress snail expression (60). Thus, estrogen signaling determines the metastasis potential of breast cancer cells by controlling snail expression through an indirect mechanism. Likewise, other transcription factors or signaling pathways can also exploit this mechanism to affect NR signaling. For example, MTA1 expression can be up-regulated by heregulin, an activator of the heregulin/HER2 pathway (98). The recruitment of MTA1-containing Mi-2/NuRD complex by ER to its target genes leads to diminished ER transactivation activity. Therefore, heregulin/HER2 signaling can antagonize ER transcription activity by increasing MTA1 expression. In another case, the expression of mouse BAF155 is down-regulated by several signaling pathways in lymphocytes (107). Loss of BAF155 expression results in resistance to GR-mediated apoptosis in these cells. In summary, ATP-dependent chromatin remodeling complexes provide a previously unsuspected platform for cross talk between NR and other signaling pathways.
|
CONCLUSIONS AND FUTURE PERSPECTIVES |
|---|
|
TOPABSTRACTNUCLEAR RECEPTORS (NRs)ATP-DEPENDENT CHROMATIN...ATP-DEPENDENT CHROMATIN...CONCLUSIONS AND FUTURE...REFERENCES |
|---|
). The recruitment of a specific complex to a particular promoter depends on the NR and promoter architecture. So far, the SWI/SNF complex has been associated with steroid hormone receptors and nonsteroid hormone receptors, whereas the Mi-2/NuRD is mostly connected to transcriptional regulation by the ER. Recruitment of ATP-dependent chromatin remodeling complexes to the promoter can lead to an open permissive chromatin structure that allows other essential transcription factors to bind, as demonstrated by the requirement for the SWI/SNF complex-mediated chromatin remodeling on MMTV promoter. Alternatively, a closed repressive chromatin structure may be generated so that transcription factors are excluded from the promoter as a result of Mi2/NuRD complex’s remodeling activity on pS2 promoter. Whereas the principles governing the relationship between NRs and ATP-dependent chromatin remodeling complex is being established, major research challenges still lie ahead. At present, only limited knowledge is available on the chromatin architecture of many NR-responsive, physiologically and/or pathologically relevant gene promoters. Lack of this knowledge impairs a mechanistic understanding of the potentially differential requirement of ATP-dependent chromatin remodeling complex at distinct promoters by otherwise identical NRs. Furthermore, we do not yet understand the structural biology underlying the interaction between a NR and the ATP-dependent chromatin remodeling complex. Equally important, we only have limited understanding of the role of ATP-dependent chromatin remodeling complexes in NR-mediated gene repression except in the case of ER and Mi-2/NuRD complex. Is Mi-2/NuRD complex also involved in gene repression by other steroid receptors such as GR, PR, or AR? Can SWI/SNF complex be recruited by NRs for the purpose of repression at some hormone-responsive promoters? Because yeast and human SWI/SNF complexes have been shown to mediate gene repression (108), it will not be surprising if future experiments reveal a repressive function of SWI/SNF complex in NR mediated transcriptional regulation. Finally, with two newly discovered ATP-dependent chromatin remodeling complexes (INO80 complex and SWR1 complex) and novel subunits of known complexes being discovered, it will be exciting and challenging to determine whether they are involved in NR-mediated transcriptional regulation.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: AF, Activation function; AR, androgen receptor; BAF, BRG1-associated factor; BRG1, Brm-related gene 1 protein; CARM1, coactivator-associated arginine methyltransferase 1; ChIP, chromatin immunoprecipitation; ER, estrogen receptor; GR, glucocorticoid receptor; hBrm, human Brm protein; HDAC, histone deacetylase; MMTV, mouse mammary tumor virus; NF, nuclear factor; NRs, nuclear receptors; OTF, octamer transcription factor; Pc, polycomb; PPAR, peroxisome proliferator-activated receptor; PR, progestin receptor; PSA, prostate-specific antigen; RSC, remodels the structure of chromatin; SNF, sucrose nonfermenting; SWI, switching independent; TBP, TATA binding protein; TR, thyroid hormone receptor; VDR, vitamin D receptor.
Received for publication May 13, 2005. Accepted for publication June 27, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTNUCLEAR RECEPTORS (NRs)ATP-DEPENDENT CHROMATIN...ATP-DEPENDENT CHROMATIN...CONCLUSIONS AND FUTURE...REFERENCES |
|---|
). EMBO J 17:6979–6991[CrossRef][Medline]-H2AX interaction links ATP-dependent chromatin remodeling to DNA damage repair. Cell 119:767–775[CrossRef][Medline] directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 115:751–763[CrossRef][Medline] on responsive promoters is an integral feature of estrogen signaling. Mol Cell 11:695–707[CrossRef][Medline]
