The nuclear hormone receptors are transcriptional regulators that activate gene expression upon binding of their respective ligands. A new class of protein, termed coregulators, has emerged during the last few years. These proteins have the faculty to repress (corepressors) or to enhance (coactivators) the activity of genes regulated by nuclear hormone receptors in a ligand-dependent fashion. In this review we describe most of these coregulators and discuss their mode of action. In particular, we comment on the link between coregulators and histone acetylation, which is a crucial event in the transcriptional response within chromatin. We describe novel alternative pathways, which elicit the recruitment of coregulators independently of the presence of any ligand and speculate on how the convergence of ligand-dependent and -independent mechanisms might enhance the transcriptional response of target genes. Nuclear hormone receptors are ligand-inducible transcription factors that are involved in a number of physiological and cellular events (see Table 1 for nuclear receptor nomenclature). Together, they form a superfamily, which includes the classic steroid receptors (androgen, estrogen, glucocorticoid, mineralocorticoid, and progesterone receptors), the thyroid, vitamin D, and retinoid receptors, as well as many others that have been characterized more recently. All of them share common functional domains named A to F. The N-terminal A/B region is weakly conserved among the members of the superfamily, has a variable length, and contains an autonomous activation function (AF-1). The conserved C domain is the DNA-binding domain, which consists of two zinc-finger-like motifs. The D domain is a variable hinge. The multifunctional C-terminal half of the protein (domain E) encompasses the ligand-binding domain (LBD), a second activation function (AF-2), a dimerization domain, and a region involved in nuclear localization. The AF-2 autonomous activation domain (AF-2 AD) is composed of an amphipathic α-helix that is highly conserved among nuclear receptors and is critical for transcriptional activation (1–4). The most C-terminal region (domain F) is variable and has no known function. This domain is absent in some receptors such as the progesterone receptor (PR), peroxisome proliferator-activated receptors (PPAR), and retinoid receptors [retinoic acid receptor (RAR), retinoid X receptor (RXR)]. Nomenclature According to the Nuclear Receptors Nomenclature Committee, 1999 Nomenclature According to the Nuclear Receptors Nomenclature Committee, 1999 Transcriptional activation by both AF-1 and AF-2 of the estrogen receptor (ER) is cell type specific and relies on the promoter context of the hormone-response element (HRE) (5). This suggests the existence of different mediating or coactivating proteins, several of which have been identified to date (see below). These mediators interact with the LBD and some are capable of increasing the AF-2 response in a ligand-dependent fashion. On certain promoters, AF-1 and AF-2 must synergize to reach efficient transactivation. The observation of transcriptional interference or squelching between steroid hormone receptors provided evidence for the existence of limiting common transcriptional cofactors that mediate AF-2 function (6, 7). The subsequent biochemical identification of several nuclear receptor-interacting proteins in a ligand-dependent manner supported this hypothesis (8) (Table 2). These mediators or coactivators are required to achieve efficient transcription (reviewed in Refs. 9–11). Nuclear Receptor Cofactors—Part I Nuclear Receptor Cofactors—Part I Numerous potential receptor-interacting proteins were identified and described in the past few years (Table 2), and many others will certainly be discovered in the near future. This rapid increase has led to some confusion in the nomenclature and raised questions about the definition of a coactivator. A real coactivator must fulfill certain requirements. First it must interact directly with the activation domain of a nuclear receptor in an agonist-dependent manner (but not in the presence of an antagonist), leading to enhancement of the receptor activation function. Most of the potential cofactors meet this definition. A coactivator should also interact with components of the basal transcription machinery. Finally, coactivators should not enhance the basal transcriptional activity by their own, although they contain an autonomous activation function (12, 13). Indeed, in the absence of a nuclear hormone receptor, coactivators cannot be recruited to promoters and therefore cannot coactivate transcription. Here, we will first discuss some well characterized coactivators and then we will comment on proteins whose coactivator status is not clearly established. Among all the described coactivators to date, SRC-1 (steroid receptor coactivator 1) has attracted much attention. The human SRC-1 was first discovered as a ligand-dependent interacting protein for the progesterone receptor (14). It appeared, however, that the original cDNA clone was truncated at the N terminus (15, 16). In addition to the full-length SRC-1 (mSRC-1a, NCoA-1), several splice variants have been described, e.g. SRC-1b, -c, -d, and -e (15, 17). The isoform SRC-1e is a more potent coactivator for ER than SRC-1a (13). For instance, the estrogen-regulated rat oxytocin promoter (−363/+16) is coactivated by SRC-1e but not by SRC-1a, as analyzed by transient transfection assay in Cos-1 cells. On the other hand, both SRC-1 isoforms stimulate ER-mediated transcription from an artificial ERE-containing promoter. Thus, coactivation by SRC-1a appears to rely on the promoter context of the receptor target gene. Both isoforms contain three nuclear receptor-interacting motifs (LXXLL) found in many co-factors (18). SRC-1a however possesses a fourth LXXLL motif at its C terminus (13). The function of this additional motif is unclear since its mutation does not affect transcription. The difference in activity results most likely from the presence of two distinct activation domains in SRC-1. The first domain interacts with the mediator CREB-binding protein (CBP)/p300, whereas the second domain activates transcription independently of CBP/p300. It seems that the extra C-terminal portion of SRC-1a, which is not present in SRC-1e, represses this CBP/p300-independent activation domain. The fact that the promoter context influences the ability of SRC-1a to coactivate ER suggests strongly that the recruitment of p300/CBP by SRC-1 is not always sufficient on some promoters. The target factor of the second activation domain is not known to date. The interaction of SRC-1 with the estrogen receptors depends on ligand and the integrity of helix 12 within the LBD and requires the presence of two functional AF-2 domains in a receptor dimer (13). The ligand-dependent interaction between SRC-1 and TR was analyzed in detail (19). Five independent mutations within the LBD of TR abolished SRC-1a binding. These mutations include residues from helix 3, 5, and 12, which form a small interaction surface encircling a hydrophobic cleft. A similar mutation (K366 in helix 3) in the mouse ER was shown to interfere with SRC-1 recruitment (20). More recently, a complex containing the liganded PPAR-γ LBD (homodimer) and a portion of human SRC-1(623–710) was resolved at 2.3 Å (21). The crystal structure showed that each member of the receptor dimer interacts with a single and different LXXLL motif of the same SRC-1 molecule. The hydrophobic face of the LXXLL helix packs into a hydrophobic pocket formed by helices 3, 4, 5, and 13 (H12 in other receptors) of PPAR-γ. The nuclear hormone receptors contain similar LXXLL motifs within their own AF-2. Surprisingly, the crystal structure of the unliganded PPAR-γ homodimer indicates that the AF-2 helix of one receptor can interact with the LBD of a second receptor (21). This suggests that the ligand-dependent activation leads to the displacement of the AF-2 helix from the LBD of the other receptor in favor of the recruitment of an LXXLL motif of SRC-1. This model was also proposed for the RXR/RAR heterodimer (22). SRC-1 is also capable of interacting with both the A/B and D/E regions of PR and ER through multiple receptor-interaction sites (23, 24). Furthermore, the binding of SRC-1 to steroid receptors is more efficient when both AF-1 and AF-2 are present. This could potentially explain the transcriptional synergy observed between AF-1 and AF-2 (5). The ligand-dependent interaction between SRC-1 and nuclear receptors is established, but the way the transcriptional activation signal is transmitted to the transcriptional machinery remains obscure. One possibility is the direct binding of SRC-1 to the basal transcription machinery through TFIIB or TATA-binding protein (TBP) (17). Alternatively, SRC-1 may be part of a larger coactivator complex. Hence, upon estrogen binding, ER becomes associated with numerous proteins, including SRC-1 and p300 together with proteins of 140 (ERAP140), 100, 90, and 30 kDa (25). However, there is no clear evidence that these proteins are part of the same complex. Nevertheless, it was not surprising when SRC-1 was shown to interact directly with a conserved region in the C terminus of p300 and its homolog CBP (15, 16). Moreover, CBP/p300 is a coactivator that binds to nuclear hormone receptor in a ligand-dependent manner (26) and enhances steroid-dependent transcription in synergy with SRC-1 (27). However, there is increasing evidence indicating that the limiting CBP/p300 factor serves a broader function, i.e. as an integrator of many different activation pathways (28–30). Indeed, CBP/p300 has been shown to interact with an increasing number of other DNA-binding factors and with components of the basal transcription machinery. p300/CBP-associated factor (P/CAF) and p300/CBP cointegrator-associated protein (p/CIP) are two other nuclear hormone receptor coactivators that can associate with CBP/p300 (31–33). Both CBP/p300 and p/CIP, together with SRC-1 (NCoA-1), are required to allow full ligand-activated gene transcription in several cell lines (32). Finally, p/CIP and SRC-1 can bind P/CAF (34). Despite all the described potential interactions between all these cofactors, there is little biochemical evidence of the existence of such a complex in vivo. Some interactions may be mutually exclusive. Alternatively, various combinations of subsets of these coactivators may coexist in the cell, giving rise to a number of possibilities in term of specificity of regulation. In an attempt to isolate such complexes, cells were recently subjected to biochemical fractionation (35). This study indicates that the different cofactors cofractionate in various stable subcomplexes. These data also suggest that the liganded progesterone receptor recruits a preformed complex that contains SRC-1 and TIF2. Although many receptors can bind to a given coactivator, it is possible that they compete with each other and that each has a different cofactor affinity (36). Interestingly, P/CAF, CBP/p300, and SRC-1 present histone acetyltransferase activity (HAT) (33, 37, 38). Since histone acetylation correlates with promoter activation (reviewed in Ref. 58), it may explain how these cofactors increase the transcriptional activation by nuclear receptors. But are all the different HATs required for the coactivation or do they have some specificity? It appears that inactivation of the HAT domains of CBP or SRC-1 has no influence on the coactivation of RAR (34). However, the HAT domain of P/CAF is indispensable for nuclear receptor activation. On the other hand, CREB (CRE-binding protein) function needs CBP-HAT activity and not P/CAF-HAT. This suggests that there is a selectivity in the specific HAT activity required for the action of different classes of transcription factors. In addition, P/CAF acetylates preferentially nucleosomal histone H3, whereas p300/CBP acetylates all nucleosomal core histones (SRC-1 and ACTR have a specificity for histones H3 and H4) (33, 37, 38, 40, 41). The presence of multi-HAT activities within a given complex may lead to various patterns of histone acetylation that are specific for a particular transactivator or for a promoter context. Interestingly, P/CAF and p300/CBP have the property to acetylate nonhistone proteins such as TFIIEβ, TFIIF and and protein was and as a new coactivator complex Despite the of HAT is a potent coactivator of the vitamin D receptor in a context. activity to or has not been identified to date. Interestingly, some are to components of mediator complex that are found associated with the complex as This a as to how may target the to the promoter. Surprisingly, and most its most of the with The however, is a coactivator for transcription factors such as and within chromatin. It appears likely that there is a convergence in the coactivation pathways of many transcriptional the in the of coactivator or subcomplexes. According to the definition in this a coactivator must interact directly with the activation domain of a nuclear receptor in an agonist-dependent manner but not in the presence of an was shown to interact in with numerous receptors independently of the presence of any or but no functional have been protein, activation of transcription by and PR although no direct interaction with receptors was Interestingly, in one there is ligand-dependent of a coactivator. The receptor is in absence of its Surprisingly, the addition of or of the of the steroid receptor 1 and leads to transcriptional for to the coactivator is the ability to enhance receptor function. This is not observed with which by and in Cos-1 cells It is however, that of an limiting nuclear protein required for AF-2 such as of a mutation in the transcriptional activation domain of the and protein 1) interact with several nuclear hormone receptors in a ligand-dependent as well as with The fact that was proposed to be a of the its as a coactivator However, is a of the and it is likely that these proteins are not coactivators but are involved in receptor A is the for a direct between the cofactor and the basal transcription machinery in of the This is to and was not for all potential One can also that cofactors are part of a larger limiting the for a direct interaction with basal transcription factors. Although interacts with several nuclear receptors in and enhances weakly ER function in it is not to associate with TFIIB or this it as a nuclear hormone receptor coactivator It is possible that it interacts with other basal transcription factors. Moreover, the fact that transcription upon in favor of the for factor Finally, coactivators should not enhance the basal transcriptional activity on their own, although they contain an autonomous activation function (12, 13). Indeed, in the absence of a nuclear hormone receptor, coactivators cannot be recruited to promoters and therefore cannot coactivate transcription. The first described nuclear regulators are members of the of transcriptional enhancement of or ER in requires several gene such as and which are part of a complex The human of termed or were also shown to coactivate and RAR in cells It has not been established, however, or not the described interaction between and requires and is direct The that contains nuclear hormone motifs present in many cofactors is and might suggest that it is potentially a coactivator However, the of these LXXLL motifs was not for A of or can activate the ER including factor factor I and many Since all these protein it is likely that of the receptors associated is a event in the activation. Moreover, an of protein 1 and is also to activate transcription of the steroid hormone receptors has been known for a (reviewed in Refs. The ER is on residues in the A/B domain although of a in the domain was also The of events to ER has been analyzed more activates the through its receptor and leads to of on and to enhancement of transcription However, the functional between a particular and transcriptional activation recently. of two residues and within the AF-1 the the recruitment of SRC-1 in the absence of estrogen 1) were with the nuclear receptor factor enhances the recruitment of both a coactivator protein 1 and a mediator for retinoid and hormone receptor to In this particular the functional of in transcriptional activation appears The ER recruitment of coactivators by the The of residues within the AF-1 domain the functional interaction with SRC-1 Alternatively, the for a ligand is abolished by the presence of which as a protein between the ER domain and SRC-1 and or P/CAF In the the described between and might from their in the recruitment of The presence of SRC-1 P/CAF suggests that other components of a coactivation complex might be present as well recruitment of coactivators by the The presence of the ligand a in the ER ligand binding domain that leads to the recruitment of a coactivation complex also containing protein such as P/CAF, p/CIP, and many others It is possible that ligand-dependent and -independent mechanisms to transcriptional to the receptor is not the event that transactivation. has the property to the activity of the ER in a Interestingly, is to interact with SRC-1 through a region that the receptor coactivator binding motif in AF-2 1) is for recruitment of coactivators to unliganded ER and as a factor between the receptor and SRC-1. have shown that P/CAF with Thus, a crucial in ER activation by HAT activities in the absence of any these results that the activity of a receptor can be in multiple The of various mechanisms could elicit to different cellular Transcriptional activation is by the recruitment of coactivators by the However, nuclear hormone receptors can repress transcription various (reviewed in Refs. in the absence of a ligand or when an is to the In the the the activation. Transcriptional several mechanisms It may from the binding of a directly to leading to a for the same element the binding of the to an interference with the function binding to a or to the direct of the basal transcription machinery of the presence or absence of the Alternatively, may be the recruitment of a limiting to the promoter by interaction with the 2). In this the is not to bind to on its We will on the by the recruitment of a (Table between A Nuclear Receptor with a or a The nuclear hormone receptor is associated with a which in recruits a histone through its interaction with of histone leads to transcriptional of the ligand this complex in favor of the of a coactivation complex P/CAF, and These proteins a histone acetyltransferase activity that through histone The interaction between the nuclear hormone receptor AF-2 domain and the coactivation complex through the LXXLL motif found in many The coactivator and are with lines since their in is not Nuclear Receptor Nuclear Receptor A must fulfill several First it has to interact directly with the unliganded receptor, leading to enhancement of basal transcription A should also interact with components of the basal transcription machinery and an autonomous domain. The by some members of the nuclear receptor in the absence of ligand has attracted a of The unliganded which is to bind is not but as a The that is by of the unliganded RAR or the C terminus of the TR that TR might within the cell such were then identified (Table receptor and nuclear These proteins have the to interact with the unliganded TR or RAR associated with their on The C terminus of interacts with TR and RAR in a region the domain and a portion of the ligand-binding domain Interestingly, this interaction region is conserved between and RAR but not among the receptors that do not associate with such as ER (see below). is abolished upon ligand-dependent of the from the have that the of from TR is by the of the LBD upon hormone binding the is associated with of the presence or absence of a The of that the of is a for transcriptional activation. Interestingly, the which is of but its ligand-dependent cannot indicating that is to the recruitment of The for a receptor to a to activate transcription is well with the It is well that and RAR activate transcription from direct when by but not when by one This from the of acid to from the heterodimer on a element and therefore to It that and RAR the and of a whereas the is on a element The is likely to on the of However, the of a or response element by has no on the ligand-dependent recruitment of coactivators It was first that ER and PR are to interact with or in the absence of any ligand It appeared, however, that their respective and such an Interestingly, these into when the receptor activation function is by the This activation is to the of the and to the recruitment of components of the coactivator complex This may explain for to a small coactivator or for protein for has been recently identified and whose enhances transcription of ER and Surprisingly, has no on agonist-dependent transcription by these receptors. In of these it is possible that the cellular between and coactivators such as might an receptor be or The interfere directly with transcriptional activation. Transcriptional can also be by nuclear receptor from and are such proteins whose binding to the of TR and RAR or to the DNA-binding domain of and RAR with their binding However, these proteins should not be as real to its definition Indeed, and transcriptional activation by with receptor binding but not by basal transcription The protein was as a of the estrogen and progesterone receptors It to repress the activity of It is not clear should as a nuclear hormone it AF-1 but not AF-2. In addition, there is no study that could all the of the and since the steroid hormone receptors are not in it is unclear a similar in Interestingly, is involved in gene and requires a to achieve full has been shown to mediate by its ability to interact directly with histones H3 and This suggests that some have that and are components of a cellular complex containing the proteins and histone The N terminus domain of interacts with which in with the histone through one of its two domains evidence of a direct interaction between and was that as a between and the complex. These that at part of the by nuclear hormone the of histones through the recruitment of a histone complex 2). The of histone associated to has been recently in human of are by that proteins between RAR and or Both and the complex through RAR in a fashion. These interactions are abolished with However, is also to associate and with and through the of the presence of the This with acid but not These data strongly suggest that by and is from through acetylation cannot for the of transcription by unliganded receptors. is observed in that are of such as transient and in transcription alternative pathways must and function independently of the recruitment of any histone results have that TR is by its direct interaction with the transcription factor TFIIB and that hormone is to this interaction In with these TFIIB was recently to interact with the and as well as with It appears that TFIIB binds in to the same domain of as does (see It is not clear to date the binding of TFIIB and to are mutually exclusive. Interestingly, of the transcriptional activity of TFIIB to a promoter indicating that their interaction is In was shown to and with the transcription factors and In this the binding of TFIIB with can in the presence of and The of TFIIB and by the functional interactions of the two which is crucial for transcriptional contains two domains within its and but interacts with or TFIIB possesses two one of which interacts with the histone Moreover, the histone has no on the ability to repress transcription. These results suggest in addition to the recruitment of TFIIB or other alternative pathways may The increasing number of described cofactors to the of the transcriptional by nuclear hormone receptors. One of the will be to the of the is evidence that coregulators do not the activity of all nuclear hormone receptors. For instance, it is known that represses activity although they interact in In the heterodimer to these to at at the gene promoter. More but not indicating that these two do not the recently described of estrogen receptor appears to be for the liganded ER Thus, the first of specificity might be by the recruitment of a given We that some coregulators are part of such as and P/CAF The presence of various proteins within these or alternative will likely influence the specificity of transcription. We have also that some coactivators a HAT Finally, of coregulators or of other components within their complex may as well to be for regulation. All these potential of increase not the but also the number of possibilities for a of transcriptional The in the nuclear hormone receptor during the last has the of the of receptor action. More are likely to in the near future.
No takes yet. Share an insight, caveat, or question.
Daniel Robyr (2000) studied this question.