Regulation of cAMP-responsive element-binding protein-mediated transcription by the SNF2/SWI-related protein, SRCAP.

SRCAP (SNF2-related CPB activator protein) belongs to the SNF2 family of proteins whose members participate in various aspects of transcriptional regulation, including chromatin remodeling. It was identified by its ability to bind to cAMP-responsive-binding protein (CREB)-binding protein (CBP), and it increases the transactivation function of CBP. The phosphoenolpyruvate carboxykinase (PEPCK) promoter was used as a model system to explore the role of SRCAP in the regulation of transcription mediated by factors that utilize CBP as a coactivator. We show that transcription of a PEPCK chloramphenicol acetyltransferase (CAT) reporter gene activated by protein kinase A (PKA) is enhanced 7-fold by SRCAP. In the absence of PKA this SRCAP-mediated enhancement does not occur, suggesting that SRCAP functions as a coactivator for PKA-activated factors such as CREB. Replacing the PEPCK promoter binding site for CREB with a binding site for Gal4 (DeltaCRE (cAMP-responsive element) Gal4 PEPCK-CAT reporter gene) blocks the ability of SRCAP to activate transcription despite the presence of PKA. Expression of a Gal-CREB chimera restores the ability of PKA to regulate transcription of the DeltaCRE Gal4 PEPCK gene and restored the ability of SRCAP to stimulate PKA-activated transcription. In addition, SRCAP in the presence of PKA enhances the ability of the Gal-CREB chimera to activate transcription of a Gal-CAT reporter gene that contains only binding sites for Gal4. SRCAP binds to CBP amino acids 280-460, a region that is important for CBP to function as a coactivator for CREB. Overexpression of a SRCAP peptide corresponding to this CBP binding domain acts as a dominant negative inhibitor of CREB-mediated transcription. Structure-function studies were done to explore the mechanism(s) by which SRCAP regulates transcription. These studies indicate that the N-terminal region of SRCAP, which contains five of the seven regions that comprise the ATPase domain, is not needed for activation of CREB-mediated transcription. SRCAP apparently has several domains that participate in the activation of transcription.

The transcription factor cAMP-responsive element-binding protein (CREB) 1 regulates the transcription of a number of genes (for review, see Refs. 1 and 2). It stimulates a low level of basal transcription, which has been proposed to be mediated through action of CREB domains that bind directly to TAF 110, TAF 130, TFIIB and TBP (3)(4)(5)(6)(7). CREB also stimulates a much higher level of transcription when it is activated in response to diverse biological stimuli such as neural and hormonal signals. These signals stimulate phosphorylation of CREB within the kinase-inducible domain. Phosphorylation of serine 133 within the kinase-inducible domain by several kinases, including calmodulin kinase II and IV, protein kinase A, ribosomal S6 kinase 1-3, mitogen-and stress-activated protein kinase 1, and mitogen-activated protein kinase-activated protein kinase 2/3, leads to the activation of transcription (for review, see Ref. 8). Although phosphorylation of serine 133 has been implicated as the trigger for transcriptional activation, other studies have demonstrated that additional phosphorylation sites in the kinase-inducible domain are important for the regulation of the transcriptional activity of CREB. For example, phosphorylation of serine 133 creates a consensus site for phosphorylation of CREB at serine 129 by glycogen synthase kinase 3, and phosphorylation of this latter site is required for full activation of CREB (9,10). Repression of CREB-mediated transcription has also been reported to result from the phosphorylation of serine 142 by calmodulin kinase II (11).
The phosphorylation of CREB on serine 133 results in the activation of transcription because this modification promotes the association of CREB with CREB-binding protein (CBP) (12,13). The importance of the interaction of CREB with CBP for activation of transcription is supported by several studies. A CREB mutant in which serine 133 is changed to an alanine can neither activate transcription nor bind CBP (12,14). Shaywitz et al. (15) demonstrated that the magnitude of transcription activated by CREB is dependent on the strength of the interaction of CREB with CBP. In addition, studies by Cardinaux et al. (16) demonstrated that CREB modified to bind CBP constitutively also activates transcription constitutively.
CBP and its homolog p300 interact with a large number of transcription factors and thus have been implicated in regulating the transcription of a number of genes (for review, see Refs. 1, 2, and 17). The mechanism(s) underlying the ability of CBP/ p300 to activate transcription has not been elucidated com-pletely but is thought to occur, in part, by the interaction of CBP/p300 with general transcription factors such as TFIIB, TBP, and RNA helicase A (13,18,19). Efforts to understand how CBP activates transcription have led to the realization that CBP is a histone acetyltransferase capable of acetylating not only histones (20) but several transcription factors such as c-Myb, MyoD, GATA-1, and p53 (21)(22)(23)(24)(25). In addition, CBP binds to several proteins that also function as histone acetyltransferases such as P/CAF, p/CIP, and the p160 coactivators such as SRC-1 (26,27). Precisely how CBP interacts with these coactivators and other cellular factors to activate transcription at a specific promoter is not known. However, the notion that CBP interacts with a specific subset of factors at different promoters has been suggested by the work of Korzus et al. (28). These authors showed that CBP, in conjunction with P/CIP, SRC-1, and P/CAF, is required for activation of transcription of a retinoic acid response element reporter gene by the retinoic acid receptor, whereas only CBP, P/CAF, and p/CIP are needed for activation of a cAMP-responsive element (CRE) reporter gene by CREB.
SNF2-related-CBP activator protein (SRCAP) was identified by our laboratory (29) using a yeast two-hybrid assay. It binds to amino acids 227-460 of CBP, a region important for activation of CREB-mediated transcription (19). This region of CBP lies adjacent to, but does not overlap, the CREB binding domain, which is located between amino acids 586 and 666 (30). SRCAP belongs to the SNF2 family of proteins whose functions include remodeling of chromatin, DNA repair, and regulation of transcription (for review, see Refs. [31][32][33]. A hallmark of the SNF2 family is the presence of seven highly conserved regions that collectively comprise an ATPase domain. Previous studies indicate that SRCAP is an ATPase and that it enhances CBPactivated transcription (29).
In the present report, we have used the well characterized phosphoenolpyruvate carboxykinase (PEPCK) promoter as a model to study the role of SRCAP in CREB-mediated transcription. PEPCK catalyzes a rate-controlling step in hepatic gluconeogenesis. Expression of the protein is regulated primarily at the transcriptional level by the action of several hormones including glucagon, glucocorticoids, thyroid hormone, and insulin (34,35). Although it has been shown that this promoter contains multiple cis-acting elements that are required to mediate the full transcriptional response to the various hormones, the CRE is absolutely required for cAMP induction (36). CREB as well as C/EBP proteins bind to the CRE of this promoter (36). Our cotransfection experiments using a PEPCK-CAT reporter gene and plasmids for expression of CREB, PKA and SRCAP demonstrate that SRCAP enhances transcription in a CREB-dependent manner.
Studies with a Gal-CREB chimeric protein indicate that SRCAP has several features consistent with a role as a coactivator of CREB-mediated transcription. Increasing intracellular levels of SRCAP result in an increase in CREB-mediated transcription, whereas disruption of the interaction of SRCAP with CBP (using a dominant negative form of SRCAP) results in inhibition of CREB-mediated transcription. These studies also indicate that, although SRCAP functions as an ATPase, mutant SRCAP that lacks a large portion of this domain retains its ability to function as a CREB coactivator.

EXPERIMENTAL PROCEDURES
Plasmids-The pSRCAP plasmid encoding amino acids 1-2971 of SRCAP was generated by subcloning the 9,121-bp SRCAP cDNA into the pcDNA 3.1 Myc/His plasmid (Invitrogen) digested with the restriction enzymes NheI and BamHI. The plasmids pSRCAP 1-2302 , pSRCAP 1274 -2971 , pSRCAP 1274 -2309 , and pSRCAP 1380 -1670 were constructed by PCR and sequenced to confirm that no changes in DNA sequence occurred. Oligonucleotides used in the PCRs introduced an initiator methionine embedded in a consensus Kozak sequence at the 5Ј-end of the SRCAP cDNA (37). The pGal-SRCAP 1-1186 , pGal-SRCAP 1274 -2971 , and pGal-SRCAP 2316 -2971 plasmids were made by subcloning the corresponding SRCAP cDNA fragments into the pGal 1-147 plasmid as described previously by Johnston et al. (29). The pPKA plasmid encoding the catalytic subunit of PKA was a gift from Mike Uhler (University of Michigan, Ann Arbor). The pGal-CREB chimera was a gift from Robert Rehfuss (McGill University, Montreal). The plasmid encoding CREB (pRc/RSV-CREB) (38) and the PEPCK-CAT (pPL32) and p⌬CRE-Gal4 PEPCK-CAT reporter genes have been described previously (39,40). The pCAT-3 vector containing a minimal promoter from SV-40 was obtained from Promega.
Transfections-HeLa cells were maintained in Dulbecco's modified Eagle medium, and HepG2 cells were maintained in minimum essential medium with Earle's salts. Each was supplemented with 10% fetal bovine serum, 50 units/ml penicillin, and 50 g/ml streptomycin. Cells were seeded either at 1 ϫ 10 5 cells/35-cm dish or at 5 ϫ 10 5 cells/10-cm dish, 18 h prior to transfection. Each transfection utilized 100 ng of the pGal-CAT reporter plasmid (pGal4/E1b TATA) or 200 ng of the PEPCK-CAT reporter plasmid and the indicated amounts of each additional plasmid. Each transfection was adjusted to contain equal molar amounts of the CMV promoter using the pcDNA 3.1 Myc/His plasmid and was adjusted to contain the same amount of total DNA using salmon sperm DNA. The LipofectAMINE (Life Technologies, Inc.) transfection method was used for HeLa cells, and Fugene6 (Roche Biochemicals) was used for HepG2 cells. Each was used according to the manufacturer's directions. Cells were harvested 24 or 48 h after transfection and assayed for CAT activity as described (19) or Gal-CREB levels using Western blots. CAT activity reported was normalized for variation of transfection efficiency between samples (the variation in the amount of plasmid taken up by cells in each sample) using a statistical approach where each experimental point was repeated triplicate in at least three separate experiments. We have used this approach because SRCAP regulates (presumably through regulation of CBP) the transcription of common reporter genes used as internal controls. For example, we have found that SRCAP activates transcription of a CMV-␤-galactosidase reporter gene. In test experiments, using CMV-␤-galactosidase as an internal control we found that SRCAP increased CREB-mediated transcription about 6-fold versus 7-fold when corrected by the statistical approach (data not shown).
Western Blotting-48 h post-transfection HeLa cell nuclear extracts were prepared by the method of Dignam et al. (41). Equal amounts of each nuclear extract (66 g) were analyzed for Gal-CREB levels by Western blot analysis using an anti-Gal antibody (19).
Reverse Transcription-PCR-Total RNA was prepared using the Trizol reagent (Life Technologies, Inc.). 2 g of RNA was treated with 2 units of DNase I (Life Technologies, Inc.) and the reverse transcription reaction performed as described (42). PCRs were performed using Vent polymerase (New England Biolabs) with 30 cycles of amplification. The amplification of the ␤-actin cDNA was done using the sense primer 5Ј-GCTCGTCGTC-GACAACGGCTC-3Ј and the antisense primer 5Ј-CAAACATGATCT-GGGTCATCTTCTC-3Ј. The SRCAP cDNA was amplified using two distinct sets of primers. Used in reaction 1 (Fig. 3C, lane 3) were the sense primer (5Ј-AAGACCCCAACCTCCAGCCCA-3Ј) containing SRCAP nucleotides 7474 -7495 and the antisense primer (5Ј-CACCATGGACCCTC-CACTCTT-3Ј) containing SRCAP nucleotides 8692-8671. Used in reaction 2 (Fig. 3C, lane 4) were the sense primer (5Ј-GCGGATCCGTTCCA-GGGTTGAACTCAACCGTG-3Ј) containing SRCAP nucleotides 4033-4056 and the antisense primer (5Ј-GCGAATCTCAAGCCTCACTA-AGTTGGAAAATCCGTTC-3Ј) containing SRCAP nucleotides 5162-5137. Note that these last two primers contain an additional eight nucleotides on the 5Ј-ends coding for restriction sites. The expected size cDNA for ␤-actin is 353 bp and 1,218 bp for SRCAP in reaction 1 (Fig. 3C, lane 3) and 1,129 bp for SRCAP in reaction 2 (Fig. 3C, lane 4). A sample of each PCR was electrophoresed through a 1.5% agarose gel and visualized with ethidium bromide. Contaminating genomic DNA was not amplified because no PCR products were obtained without first reverse transcribing RNA preparations.

RESULTS
We demonstrated previously that SRCAP binds CBP and enhances its ability to activate transcription (29). This suggested that SRCAP might regulate the activity of transcription factors such as CREB which utilize CBP as a transcriptional coactivator. To test this hypothesis we asked whether SRCAP could enhance the transcriptional activity of the PEPCK gene promoter, which is known to be regulated by CREB (34). For this purpose we transfected the PEPCK-CAT reporter gene (39) into the HepG2 human hepatoma cell line. The transcriptional activity of this reporter gene, as shown in Fig. 1, is minimally activated by transfection with a plasmid encoding CREB in the absence of activated PKA. A 10-fold activation of transcription of the PEPCK promoter was observed following transfection of a plasmid encoding the catalytic subunit of PKA. This result is consistent with the activation of endogenous CREB in HepG2 cells (43). A similar activation was achieved when plasmids encoding both CREB and PKA were cotransfected. Cotransfection of the plasmid encoding SRCAP (pSRCAP) with the plasmid encoding PKA or with both the plasmids encoding CREB and PKA resulted in an additional 7-fold increase in the activation of transcription. Cotransfection of pSRCAP with only the plasmid encoding CREB (in the absence of PKA) did not activate transcription.
The observation that SRCAP strongly activates transcription only in the presence of PKA suggested that SRCAP is a coactivator for CREB or C/EBP␣, either of which mediates PKAactivated transcription of the PEPCK promoter (43). To test this hypothesis, we made use of a PEPCK gene promoter in which the binding site for CREB (the CRE) was replaced by a Gal4 binding site (40). Transfection of either a plasmid encoding a Gal-CREB chimera or a plasmid encoding the catalytic subunit of PKA resulted in a weak activation of the modified reporter gene, ⌬CRE-Gal4 PEPCK. In contrast, if plasmids expressing both Gal-CREB and PKA were cotransfected, SR-CAP transcription was increased by more than 11-fold. In contrast to the effects on the PEPCK promoter, we found that SRCAP had a much smaller effect on the transcriptional activity (2.5-fold) of a control reporter gene pCAT-3 whose expression is driven by the basal SV40 promoter (data not shown).
To test more directly the hypothesis that SRCAP acts as a coactivator for CREB-mediated transcription, we asked whether SRCAP could regulate Gal-CREB-mediated transcription from a simple reporter gene containing only binding sites for Gal4 (pGal-CAT). As shown in Fig. 2, cotransfection of plasmids encoding pSRCAP, Gal-CREB, and the catalytic subunit of PKA activates transcription by about 9-fold in HeLa cells. SRCAP also enhances transcription mediated by the Gal-CREB chimera by 25-fold in HepG2 cells (data not shown). As shown in Fig. 2, the cotransfection of pSRCAP had only a small effect on transcriptional activation by a Gal-VP16 chimeric protein. This chimera was chosen as a control because the VP16 protein has not been reported to use CBP as a coactivator, but rather appears to regulate transcription by direct contact with TBP (44).
These results indicate that SRCAP functions as a positive regulator of CREB-mediated transcription and suggest that the cellular level of SRCAP protein might be a limiting factor for activation of CREB-mediated transcription. To test this latter hypothesis, we asked whether disruption of the CBP-SRCAP interaction in HeLa cells (which express SRCAP; see Fig. 3C) would result in the disruption of CREB-mediated transcription. A peptide that corresponds to the CBP binding domain of SRCAP was expressed in cells by transfection using the plasmid pSRCAP 1380 -1670 to disrupt CPB-SRCAP interaction. SRCAP 1380 -1670 acts as a dominant negative inhibitor and blocks CREB-mediated transcription by 95% but has only a slight effect on the transcriptional activity of the Gal-VP16 chimera (Fig. 3A).
Western blot analysis was used to examine the level of Gal-CREB protein in transfected cells to ensure that the alterations in CREB-mediated transcription observed were the result of changes in the transcription activity of the Gal-CREB chimera rather than changes in levels of the Gal-CREB protein. The transfection of plasmids encoding either SRCAP or the dominant negative SRCAP peptide did not alter Gal-CREB protein expression (Fig. 3B). In addition to these plasmids, 2,000 ng of a plasmid encoding SRCAP (pSRCAP) was also cotransfected, as indicated. The relative CAT enzymatic activity was determined by dividing the CAT enzymatic activity of each sample by the transcriptional activity induced by the Gal-CREB or Gal-VP16 chimeras (the activity of these chimeras was assigned a relative value of 1). Values represent the means Ϯ S.E. from three separate experiments in which each sample was assayed in triplicate.
The function of several SNF2 family members is dependent on their ability to hydrolyze ATP. This prompted us to determine whether the SRCAP ATPase domain is required for coactivation with CREB. For these studies, we tested the activity of an N-terminal truncated SRCAP that lacks five of the seven conserved regions that collectively make up the ATPase domain of SRCAP (see Fig. 4A). The transcriptional activation afforded by SRCAP 1275-2971 is not significantly different from that provided by wild-type SRCAP (Fig. 4B).
The observation that the ATPase domain of SRCAP is not needed for activation of CREB-mediated transcription suggests that SRCAP must provide some additional function. To identify regions within SRCAP which might contain this function the transcriptional activity of several Gal-SRCAP chimeras was assessed. A Gal-SRCAP chimera encoding a portion of the N-terminal end of SRCAP (Gal-SRCAP 1-1186 ), or a portion of the

FIG. 3. Disruption of SRCAP-CBP interaction represses CREBmediated transcription.
In panel A HeLa cells were transiently transfected with 100 ng of the pGal-CAT reporter gene and 100 ng of the plasmid encoding a Gal-CREB chimera (pGal-CREB) and 25 ng of a plasmid encoding the catalytic subunit of PKA (pPKA). As indicated, 2000 ng of a plasmid encoding the minimal CBP binding domain of SRCAP (pSRCAP 1380 -1670 ) designated as DN was also cotransfected. The relative CAT enzymatic activity was determined by dividing the CAT enzymatic activity of each sample by the transcriptional activity induced by the Gal-CREB (which was assigned a relative value of 1). Values represent the means Ϯ S.E. from three separate experiments in which each sample was assayed in triplicate. In panel B nuclear extracts were analyzed by Western blot to determine the level of Gal-CREB protein expressed using an anti-Gal antibody. Nuclear extracts were isolated from nontransfected HeLa cells as indicated by NT, from control cells transfected with the pGal-CAT, pGal-CREB, and pPKA plasmids as indicated by CT, from cells transfected with the SRCAP CBP binding domain plasmid pSRCAP 1380 -1670 (and the pGal-CAT, pGal-CREB, and pPKA plasmids) as indicated by DN, and cells transfected with SRCAP plasmid pSRCAP (and the pGal-CAT, pGal-CREB, and pPKA plasmids) as indicated by FL. Equal amounts of nuclear extract (66 g) from each test condition were analyzed by Western blot using an anti-Gal antibody, and the amount of the Gal-CREB chimera protein detected in each nuclear extract is shown. The Gal-CREB protein is denoted by the arrow, and a nonspecific protein is denoted by the star. In panel C expression of SRCAP mRNA in HeLa cells is detected using reverse transcription-PCR as described under "Experimental Procedures." Lane 1 contains the 100-bp ladder molecular weight marker (Life Technologies, Inc.). Lane 2 contains the PCR product generated using ␤-actin-specific primers. Lane 3 contains the PCR product generated using SRCAP-specific primers giving a predicted product of 1,218 bp. Lane 4 contains the PCR product generated using a second set of SRCAP-specific primers giving a predicted product of 1,129 bp.

FIG. 4. SRCAP protein lacking the N-terminal domain retains the ability to activate CREB-mediated transcription. Panel A
shows a diagram of SRCAP indicating the position of the seven conserved regions that collectively make up the ATPase domain and the position of the CBP binding domain. Also shown are several SRCAP deletion mutants. In panel B HeLa cells were transiently transfected with 100 ng of the pGal-CAT reporter gene, 100 ng of a plasmid encoding a Gal-CREB chimera (pGal-CREB), and 25 ng of a plasmid encoding the catalytic subunit of PKA (pPKA). In addition to these plasmids, 2,000 ng of a plasmid encoding SRCAP (pSRCAP) or a plasmid encoding the C-terminal end of SRCAP (pSRCAP 1275-2971) was also cotransfected where indicated. The relative CAT enzymatic activity was determined by dividing the CAT enzymatic activity of each sample by the transcriptional activity induced by the Gal-CREB (which was assigned a relative value of 1). Values represent the means Ϯ S.E. from three separate experiments in which each sample was assayed in triplicate.
C-terminal end of SRCAP (Gal-SRCAP 2316 -2971 ) activated transcription of the Gal-CAT reporter gene 10 -20-fold more than Gal 1-147 (Fig. 5). These two chimeras also activate transcription to levels similar to those observed with a Gal-SRCAP 1275-2971 chimera, which activates transcription to the same extent as a Gal-CBP chimera (29). Neither the Gal-SRCAP 1-1186 chimera nor the Gal-SRCAP 2316 -2971 chimera contains the CBP binding domain of SRCAP, suggesting that they activate transcription through a CBP-independent mechanism. DISCUSSION Several coactivators are essential for activation of CREBmediated transcription. These include CBP (and its homolog p300) and a number of factors that bind CBP, such as p/CIP and PCAF. Studies with these coactivators suggest that they are present in limiting amounts for activation of transcription because when they are introduced into cells, they enhance transcription (13,28,45). In similar studies, we find that by increasing the cellular levels of SRCAP, we can increase the ability of CREB to activate transcription. One possible explanation for this observation is that activities associated with both CBP and SRCAP are needed for activation of CREBmediated transcription (see model in Fig. 6). Similar observations have been reported for other coactivator proteins that interact with CBP, such as p/CIP and p/CAF. This has been demonstrated through the use of approaches that either limit the activity of the coactivator, such as microinjection of antibodies (28,46), or antisense constructs that decrease the level of the coactivators (47,48).
We show here that SRCAP significantly enhances CREB-dependent transcription from the PEPCK promoter. Although there are multiple cis-acting elements involved in PEPCK gene regulation (33,34), SRCAP does not activate this promoter in the absence of PKA. This suggests that SRCAP is not involved in basal transcription, but rather it functions as a coactivator of a primed promoter. The activation observed from the ⌬CRE-Gal4 PEPCK-CAT reporter by the Gal-CREB chimeric protein indicates that CREB is involved in SRCAP function. This conclusion is also supported by experiments demonstrating that SRCAP enhances CREB-mediated transcription of a simple reporter gene, Gal-CAT, which only contains binding sites for Gal4. Furthermore, SRCAP activates the transcription of other cAMP-and CREB-regulated promoters, such as those contained in somatostatin and enkephalin reporter genes (ϳ18-fold increase; data not shown). Therefore, we conclude that SRCAP may be a general enhancer of CREB-mediated transcription.
We introduced a plasmid encoding the peptide corresponding to the CBP binding domain within SRCAP into cells to test the hypothesis that a SRCAP-CBP interaction is needed for activation of CREB-mediated transcription. This peptide, when expressed in cells, functioned as a dominant inhibitor of CREBmediated transcription. Thus, disruption of the SRCAP-CBP interaction leads to a decrease in CREB-mediated transcription. Collectively, the results suggest that at least in some circumstances, contact between SRCAP and CBP is required for activation of CREB transcription. In support of this hypothesis, structure-function studies by Swope et al. (19) indicate that deletion of the SRCAP binding domain of CBP (amino acids 272-460) prevents CBP from acting as a coactivator for CREB. Furthermore, overexpression of a CBP peptide (amino acids 1-460) that overlaps with the SRCAP binding domain also prevents CBP from functioning as a CREB coactivator (19).
The mechanism by which the interaction of SRCAP with CBP leads to activation of CREB-mediated transcription remains unclear. However, the results of our structure-function studies indicate that deletion of the N-terminal end of SRCAP (amino acids 1-1274) does not prevent SRCAP from enhancing CREB-mediated transcription. Because this deleted portion contains five of the regions that comprise the ATPase domain, one conclusion that can be drawn from this result is that at least in transient transfection assays, the ATPase domain of SRCAP is not required for activation of CREB-mediated transcription. Although this result was somewhat surprising, similar findings have been reported for human SNF2. For example, a mutant SNF2 that cannot hydrolyze ATP and does not remodel chromatin retains the ability to activate transcription when tethered to DNA through a heterologous DNA binding domain (49). These findings suggest that activation of transcription by SNF2 occurs through at least two distinct mechanisms, one requiring the ATPase function (e.g. the chromatin remodeling activity) and one that does not. Multiple functions have also been reported for the SNF2 family member Cockayne Syndrome B. Mutation of the ATPase domain of Cockayne Syndrome B also prevents it from remodeling chromatin but does not inhibit its ability to function as a topoisomerase (50). Our data suggest that SRCAP may also have multiple functions. One activity is independent of the ATPase function and allows SRCAP to activate CREB-mediated transcription, and a second function (not yet determined) requires ATPase activity.
Other proteins, most notably CBP and p300, have multiple domains with distinct activities which collectively contribute to their ability to activate transcription. Early studies of the function of CBP using Gal-CBP chimeras found that several domains of CBP were able to activate transcription independently (19,51). Subsequent studies with CBP have determined that these domains have distinct functions. For example, several Nand C-terminal domains have the ability to bind proteins that contribute to the ability of CBP to activate transcription. These include transcription factors, coactivators (p/CIP, P/CAF, and SRC-1), general transcription factors (such as TAF110, TAF 130, TBP, TFIIB, and RNA helicase A), the histone binding protein RbAp48 (52), and nucleosome assembly proteins (53). Similar to the early studies on CBP, our studies with SRCAP using Gal-SRCAP chimeras provide an indication that SRCAP has several domains that can activate transcription independently. Two of these domains (amino acids 1-1186 and 2316 -2971) do not contain the CBP binding site (located at amino acids 1380 -1670) and therefore may activate transcription through contacts with additional associated proteins. In support of this notion, we have found in immunoprecipitation studies with an anti-SRCAP antibody that SRCAP is part of a large multiprotein complex which, in addition to containing CBP, contains at least 20 other proteins (54).
The studies presented in this paper indicate that SRCAP functions as a coactivator for CREB-mediated transcription and are consistent with the hypothesis that SRCAP functions through an interaction with the CREB coactivator CBP. Whether SRCAP activates transcription by modification of a function of CBP, e.g. regulation of the acetylation activity of CBP (an ability that has been reported recently for a number of factors; see Ref. 55) or whether it is because of distinct intrinsic activities directly associated with SRCAP is unknown. Studies to determine whether all or a subset of promoters that utilize CREB and CBP to activate transcription also utilize SRCAP are currently under way. These studies should also determine whether SRCAP serves a role as coactivator for other transcription factors that utilize CBP.