The Human Homologue of the Yeast DNA Repair and TFIIH Regulator MMS19 Is an AF-1-specific Coactivator of Estrogen Receptor*

Steroid/nuclear hormone receptors are ligand-dependent transcriptional regulators that control gene expression in a wide array of biological processes. The transcriptional activity of the receptors is mediated by an N-terminal ligand-independent transcriptional activation function AF-1 and a C-terminal ligand-dependent transcriptional activation function AF-2. The nuclear receptor coactivator RAC3 (also known as AIB1/ACTR/pCIP/TRAM-1/SRC-3) is amplified in breast cancer cells, where it forms a complex with estrogen receptor (ER) and enhances AF-2 activity of the receptor. Here, we identify a putative human homologue of the yeast DNA repair and transcriptional regulator MMS19 as a RAC3-interacting protein. The human MMS19 interacts with the N-terminal PAS-A/B domain of RAC3 in vivo and in vitrothrough a conserved C-terminal domain. Interestingly, the human MMS19 also interacts with estrogen receptors in a ligand-independent manner but not with retinoic acid receptor or thyroid hormone receptor. Overexpression of the interacting domain of hMMS19 strongly inhibits ER-mediated transcriptional activation, indicating a dominant negative activity. In contrast, over expression of the full-length hMMS19 enhances ER-mediated transcriptional activation. We find that hMMS19 stimulates the AF-1 activity of ERα, but not the AF-2 activity, suggesting that hMMS19 may be an AF-1-specific transcriptional coactivator of estrogen receptor.

Steroid/nuclear hormone receptors (SNR) 1 are liganddependent transcriptional regulators that control gene expression in a wide array of biological processes such as development and reproduction (1,2). These receptors share a common domain structure: an N-terminal ligand-independent transcriptional activation function (AF-1, or A/B region), a centrally located DNA binding domain (or C region), and a C-terminal ligand-dependent transcriptional activation function (AF-2, LBD, or E/F domain). The AF-1 and AF-2 function together to regulate gene expression in a cell-type-specific and promoterspecific manner (3)(4)(5)(6). However, the exact mechanism is currently unclear.
Several transcriptional coactivators for SNR have been identified, including the steroid receptor coactivator (SRC)/p160 family (9), which contains SRC-1 (10), GRIP1/TIF2 (11)(12)(13), and RAC3/AIB1/ACTR/pCIP (14 -18). The SRC/p160 family of coactivators binds to a hydrophobic pocket on the surface of the receptor LBD in a ligand-dependent manner (19,20). The coactivator enhances transcriptional activation of the receptor via a mechanism involving histone acetylation and recruitment of additional coactivators such as CBP/p300 and P/CAF (21,22). Genetic studies demonstrate that SRC-1 and RAC3 (pCIP) are involved in regulating hormonal responses in mice (23)(24)(25). Interestingly, the SRC/p160 coactivators also interact with the AF-1 domain of several SNRs in a ligand-independent manner (6, 26 -30). For ER␤, such interaction is regulated by phosphorylation of the AF-1 region, suggesting that SRC/p160 coactivators may be involved in cross-talk between AF-1 and AF-2 of the receptor (26 -28). RAC3 (AIB1) is unique among the SRC/ p160 family of coactivators, because it is amplified and overexpressed in a subset of breast cancers (16). It has been demonstrated that RAC3 (AIB1) activators form a stable complex with ER␣ in MCF-7 breast cancer cells (31), suggesting that RAC3 may play an important role in regulating ER function in vivo.
The SRC/p160 coactivators share a common domain structure, including a highly conserved N-terminal basic-helix-loophelix (bHLH) and Per-Arnt-Sim (PAS) domains (9). The PAS domain can be subdivided into A and B regions. This domain is highly conserved among several Drosophila and mammalian proteins, including Per, Sim, and Arnt (32), as well as in many regulators of circadian rhythm and neural development (for a review, see Ref. 33). The bHLH-PAS domain of Drosophila proteins plays an important role in mediating protein-protein interaction (34 -36). However, the function of the bHLH-PAS domain of the SRC/p160 coactivators remains largely unknown. Here, we report the identification of a human homologue of the yeast MMS19 as an RAC3-interacting protein isolated in a yeast two-hybrid screening with the bHLH-PAS domain of RAC3 as bait. Interestingly, hMMS19 also interacts with ERs in a ligand-independent manner. Our data suggest that hMMS19 may function as an AF-1-specific transcriptional coactivator of ER.

MATERIALS AND METHODS
Plasmids-The DNA fragments encoding various N-terminal domains of human RAC3 (14) were amplified by polymerase chain reac-* This work was supported by a National Institutes of Health Grant DK52888 (to J. D. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF357881.
Yeast Two-hybrid Screening-The yeast two-hybrid screening was conducted as described previously (14). Briefly, the pGBT-RAC3-N (amino acids 1-408) was used as bait to screen a human placenta cDNA library (CLONTECH, Inc.). Positive clones were isolated and tested for ␤-galactosidase expression. Library plasmids were recovered and tested with various baits to confirm specificity of the protein-protein interactions. The RAC3 screening clone 2.1 (RS2.1) showed homology with the yeast MMS19 protein and was further analyzed in this study.
Rapid Amplification of cDNA Ends-The human MMS19-specific primer (CAT AAG ATA GGA GAT CTG GCT GGG CAC CCA AGA CTG TC) was used to amplify the extreme 5Ј-end of the hMMS19 cDNA from a HeLa cell Marathon-ready cDNA along with an adaptor primer from the manufacturer (CLONTECH, Inc.). The resultant product was reamplified and subcloned into pBluescript II SKϩ and sequenced.
Construction of hMMS19 Expression Plasmids-The RS2.1 insert was released with SalI/BglII and subcloned into the pCMX-HA vector. pCMX-HA-hMMS19mt was constructed by assembling the SalI/BglII fragment of the longest RACE clone and the BglII/NsiI fragment of GenBank accession number AF007151 (Research Genetics) together with a SalI/NsiI fragment of the RS2.1 clone. The pCMX-HA-hMMS19 was obtained by replacing the Dra3/NsiI fragment of pCMX-HA-hMMS19mt with the same fragment from the expressed sequence tag clone BE206052 (GenBank accession number). All constructs were confirmed by DNA sequencing.
Northern Blot-Multiple Tissue Northern blots (CLONTECH) were probed with a 32 P-labeled NsiI/BglII 1.5-kb fragment of hMMS19. Hybridization was performed using the ExpressHyb solution (CLON-TECH). The blot was washed twice for 20 min in 2ϫ SSC/0.1% SDS at room temperature and subjected to autoradiography at Ϫ70°C.
GST Pull-down Assay-GST fusion proteins were expressed in E. coli BL-21 cells and purified by glutathione agarose beads (Amersham Pharmacia Biotech). 35 S-Labeled proteins were made by in vitro transcription/translation reactions using the T7-Quick reticulocyte lysate (Promega). For GST pull-down assay, 5 g of bead-conjugated fusion protein was incubated with 5 l of in vitro translated 35 S-labeled protein with moderate shaking at 4°C overnight in binding buffer as previously described (37). The pellet was washed four times with the binding buffer. Supernatant was removed, and bound protein was eluted in SDS sample buffer by boiling. The complex was then analyzed by SDS-PAGE and autoradiography.
Cell Culture and Transient Transfection Assays-HEK293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. One day before transfection, cells were seeded in 12-well plates at 50,000 cells per well in phenol-red free Dulbecco's modified Eagle's medium supplemented with 10% charcoal resinstripped fetal bovine serum. Transfection was performed using a standard calcium/phosphate method as described before (37). After transfection, cells were washed with phosphate-buffered saline and re-fed with fresh medium containing either vehicle alone or vehicle plus ligands. Cells were lysed 24 h after treatment and assayed for luciferase and ␤-galactosidase activities.

Isolation of a Putative Human MMS19 in Yeast Two-hybrid
Screening-To identify proteins that interact with the bHLH-PAS domain of RAC3, we conducted yeast two-hybrid screening with the RAC3-N-terminal domain as bait. The RS2.1 clone showed strongest interaction. Sequence analysis and data base search revealed that RS2.1 encodes a C-terminal 164-amino acid polypeptide that shares homology with the yeast MMS19 (38). The full-length cDNA of RS2.1 was obtained, and it predicts an open reading frame of 1030 amino acids (Fig. 1A). The full-length RS2.1 is named hMMS19 because it is related in size and sequence to the yeast MMS19. We also found a mutant cDNA clone in the GenBank (accession number AF007151), which contains a single deletion at nucleotide 2240, resulting in a truncation mutant at 779 amino acids. We designate this truncation mutant as hMMS19-mt. Comparison of hMMS19 with the yeast homologue shows a 36% overall similarity (or 26% identity). The homology is most striking at the N-and C-terminal regions with 61% or 57% similarities, respectively. Related sequences were also predicted from the genomes of Drosophila melanogaster, Caenorhabditis elegans, and Arabidopsis thaliana. Alignment of these sequences indicates that MMS19 is conserved through evolution at the N-and C-terminal domains, whereas the central region has diverged significantly. We also found that the hMMS19 gene resides in a genomic clone (RP11-445I23, accession number AL359388), which maps to chromosome 10q24. Interestingly, loss of heterozygosity in this region was observed in several human cancers, including endometrial and prostate carcinomas (39,40).
Northern blot analysis of hMMS19 in various human cancer cells revealed a single transcript of about 4 kb (Fig. 1B, top). The size of this message is in agreement with the length of the hMMS19 cDNA. The hMMS19 transcript was detected in all cancer cell lines analyzed as well as in several human tissues such as heart, brain, placenta, liver, skeletal muscle, and kidney (data not shown), suggesting that hMMS19 is ubiquitously expressed. Furthermore, by indirect immunofluorescence, we found that hMMS19 is located in the cell nucleus (Fig. 1C). This nuclear localization is consistent with the hypothesis that MMS19 is involved in DNA repair and transcriptional regulation.
Interactions of hMMS19 with RAC3-To further analyze the interaction of hMMS19 with RAC3, we first tested the specificity of RS2.1 interactions with various baits in a yeast twohybrid assay ( Fig. 2A). As expected, RS2.1 alone did not activate reporter expression nor did the bait RAC3-N or other unrelated baits. Coexpression of RS2.1 with RAC3-N strongly activated reporter expression as expected. In contrast, RS2.1 did not interact with the bHLH-PAS domain of mouse Sim2 (Sim-N) nor did it interact with RAR␣ or p53. These data confirm that RS2.1 interacts specifically with RAC3-N in the yeast two-hybrid assay.
We then mapped the RS2.1-interacting region within the bHLH-PAS domain of RAC3 (Fig. 2B). The bHLH-PAS domain of RAC3 was subdivided into bHLH, PAS-A, PAS-B, or PAS-A/B (Fig. 2F) and tested for interactions with RS2.1 in a similar assay. As expected, none of these baits alone activated reporter expression (data not shown). Coexpression of RS2.1 with PAS-A/B or RAC3-N each strongly activated reporter expression, whereas coexpression with bHLH, PAS-A, or PAS-B did not activate reporter expression in this assay. These data suggest that the PAS-A and PAS-B domains are both required for strong interaction with RS2.1.
To confirm the interaction of RS2.1/hMMS19 with RAC3 in vitro, we performed a GST pull-down assay using GST fusion of various RAC3 N-terminal fragments to pull down the 35 Slabeled RS2.1 fragment or the full-length hMMS19. The RS2.1 fragment and full-length hMMS19 were transcribed/translated and labeled with [ 35 S]methionine in reticulocyte lysate. First, the 35 S-RS2.1 probe was incubated with GST-RAC3-N, and the bound probe was eluted and analyzed by SDS-PAGE and autoradiography. As expected, GST-RAC3-N, but not GST alone, pulled down a significant amount of 35 S-RS2.1 (Fig. 2C), confirming in vitro interaction of RS2.1 with RAC3-N. Second, 35 S-hMMS19 was incubated with GST or GST fusion of bHLH, PAS-A, PAS-B, PAS-A/B, or the bHLH-PAS domain of RAC3. The bound hMMS19 probe was analyzed by SDS-PAGE and autoradiography (Fig. 2D). As expected, GST or GST-bHLH did not bring down a detectable amount of 35  C-terminal RAC3 fragments, including the nuclear receptorinteracting domain and the transcriptional activation domain, showed no binding to hMMS19 (data not shown). Therefore, these data are consistent with the above yeast two-hybrid results in implicating the PAS-A/B domain of RAC3 as the binding surface for hMMS19.
Human MMS19 Interacts with Estrogen Receptor-Because RAC3 interacts with nuclear receptors, we tested whether hMMS19 could interact with nuclear receptor in a similar GST pull-down assay. Intriguingly, we found that GST-RS2.1 pulled down a significant fraction (about 20% of input) of ER␣ in a ligand-independent manner (Fig. 3A). Similarly, GST-RS2.1 also pulled down a significant amount of ER␤ (Fig. 3B), suggesting that hMMS19 can interact with both ER␣ and ER␤. Similar results were obtained by GST fusion of full-length hMMS19 (data not shown), suggesting that the interaction also occur in the context of full-length protein. In contrast, little interactions of GST-RS2.1 with 35 S-hRAR␣ and hTR␤ were detected, consistent with the yeast two-hybrid data. These data suggest that hMMS19 may also interact with ER␣ and ER␤.
Human MMS19 Modulates Transcriptional Activity of ER-The above interactions of hMMS19 with RAC3 and ERs prompted us to investigate whether hMMS19 could modulate transcriptional activity of ER. To test this possibility, we performed reporter gene assay by transient transfection in HEK293 cells. First, we analyzed the effects of overexpression of RS2.1 on ER-mediated transcriptional activation of a luciferase reporter gene driven by ER-responsive elements (Fig. 4,  A and B). As expected, both ER␣ and ER␤ strongly activated luciferase expression in an E2-dependent manner. Cotransfection with RS2.1 significantly reduced the ligand-dependent transactivation in a dose-dependent manner without affecting reporter expression in the absence of E2. We found that, at the highest concentration of RS2.1, both ER␣ and ER␤ activities were reduced about 8-fold. Importantly, coexpression of RAC3 can reverse the inhibitory effect of RS2.1 on ER transcriptional activity (Fig. 4C). The inhibition of ER activity by RS2.1 is specific, for RS2.1 had no effect on an otherwise identical reporter lacking ER binding sites (data not shown). Furthermore, overexpression of RS2.1 had no effect on ligand-dependent transcriptional activation of RAR␣ or TR␤ (Fig. 4D), suggesting that RS2.1 may selectively affect ER transcriptional activity. These data are consistent with the preferential interaction of RS2.1 with ERs but not with RAR␣ or TR␤, implicating that hMMS19 may play a more important role in regulating ER activity than RAR or TR.
The inhibition of ER transcriptional activity by RS2.1 may be explained by the dominant negative effect of RS2.1, because this fragment contains the RAC3/ER-interacting domain. It is possible that RS2.1 interferes with the assembly of a functional ER-coactivator complex. If RS2.1 indeed acts as a dominant negative mutant, the full-length protein should have a positive effect on ER transcriptional activity. Indeed, overexpression of hMMS19 enhanced ER transcription activation in HEK293 cells (Fig. 5A). The moderate effect of this enhancement may be due to high levels of endogenous hMMS19 in these cells. In contrast, the hMMS19-mt mutant, which lacks the C-terminal interacting domain, failed to coactivate ER-mediated transcrip- tion. Furthermore, coexpression of hMMS19 with RAC3 enhanced ER transcriptional activation more strongly than the expression of each coactivator alone (Fig. 5B), suggesting that hMMS19 and RAC3 may synergize to stimulate transcriptional activity of ER.
The above findings suggest that hMMS19 may function as a transcriptional coactivator of ER. Intriguingly, the yeast MMS19 has been shown to regulate transcription by influencing TFIIH activity (38), and TFIIH has been shown to regulate the AF-1 function of human ER␣ (41). To shed light on the mechanism of MMS19 function in regulating ER activity, we examined whether hMMS19 affects AF-1 or AF-2 activity of hER␣. We subdivided ER␣ into the N-terminal AF-1 and the C-terminal AF-2 domains, each of which contains the DNA binding domain. The effect of overexpression of hMMS19 on the ER␣ AF-1 or ER␣ AF-2 was then tested by transient transfection (Fig. 5C). As expected, E2 strongly activated transcription mediated by the ER␣ AF-2 but to a lesser extent than the full-length ER␣, suggesting that AF-1 is required for maximal transcriptional activity of ER␣. Cotransfection of hMMS19 had no effect on this ligand-dependent transcriptional activity of ER␣ AF-2, suggesting that hMMS19 is not an AF-2-dependent coactivator. As expected, ER␣ AF-1 had a constitutive activity in the absence of ligand. Addition of E2 slightly enhanced the ER␣ AF-1 activity, presumably due to E2-mediated phosphorylation of the ER␣ AF-1 domain (28,42). Interestingly, cotransfection of hMMS19 strongly enhanced AF-1 activation in both the absence and presence of E2 by 6.0-and 2.5-fold, respectively, suggesting that hMMS19 can regulate ER transcriptional activity by stimulating the N-terminal AF-1 function of ER.

DISCUSSION
MMS19 was initially identified in a yeast genetic screen for mutations that render cells sensitive to the alkylating agent methyl methanesulfonate (MMS) (43) and later to UV and DNA cross-linking agents (44,45). Analysis of MMS19 mutation in yeast suggests that it plays a role in nucleotide excision repair and transcription regulation by affecting TFIIH activity (38,46,47). Here, we isolated a human MMS19 homologue as a RAC3-interacting protein. Intriguingly, hMMS19 interacts with both the coactivator RAC3 and the receptor ER and functions as an AF-1-specific transcriptional coactivator of ER.
The bHLH-PAS domain is the most conserved region in the SRC/p160 coactivators (14); however, the function of this domain is largely unknown. The isolation of hMMS19 as a bHLH-PAS domain-interacting protein suggests that this domain may be involved in protein-protein interaction. Further analysis of the hMMS19-RAC3 interactions reveals that interaction is mediated through the PAS-A/B region. Interestingly, strong interaction occurs only in the presence of both PAS-A and PAS-B, suggesting that there may be two interacting interfaces that together stabilize the interaction between hMMS19 and RAC3. The involvement of PAS domain in protein-protein interaction can also be found in the Drosophila protein Period, which interacts with Timeless through its PAS domain (48). Because the PAS domain is highly conserved among the SRC/p160 coactivators, hMMS19 might also interact with SRC-1 and/or TIF2, but this remains to be tested. However, despite the conservation of PAS domains in different proteins, hMMS19 does not interact with the PAS domain of Sim, suggesting the different PAS domains may be involved with different interacting proteins. The interacting surface in hMMS19 maps to its C-terminal conserved domain. Structural prediction suggests that this region contains multiple potential amphipathic helices, which might mediate the interaction with the PAS domain of RAC3.
Analysis of the interactions of hMMS19 with nuclear receptors shows strong binding with ER␣ and ER␤ in a ligandindependent manner. Because ER also interacts with RAC3 (49), several possibilities may explain the relevance of these interactions. For instance, hMMS19 may interact with ER first in the absence of ligand. This interaction may modulate, for instance, the phosphorylation states of ER by targeting TFIIH activity to ER. Upon ligand binding, ER undergoes a conformational change that allows the receptor to bind the LXXLL motifs of RAC3 via a coactivator-binding pocket in the LBD (9). RAC3 then interacts with hMMS19 via the N-terminal PAS domain to stabilize the complex or to signal a subsequent event in transcription. Alternatively, RAC3 may form a complex with hMMS19 first, which is then recruited to the liganded receptor. It is clear that RAC3 uses separate domains for interactions with hMMS19 and liganded ER. However, it is currently unknown whether hMMS19 can interact with RAC3 and ER simultaneously. The fact that the C-terminal domain of hMMS19 interacts with both RAC3 and ER suggests that the interactions of hMMS19 with RAC3 and ERs may be mutually exclusive. However, it is equally possible that different motifs in the C-terminal domain of hMMS19 interact with RAC3 and ER independently.
The physiological relevance of hMMS19 in ER signaling is demonstrated by overexpression of the C-terminal RS2.1 fragment and the full-length hMMS19, each of which significantly affect transcriptional activity of ER in transfection assays. Overexpression of RS2.1 inhibits ligand-dependent transcriptional activation of an ER reporter gene due to a dominant negative effect. Overexpression of RS2.1 has no effect on basal promoter activity in the absence of ligand, consistent with the observation that ER binds to chromatin in response to ligand (50,51). As expected, the full-length hMMS19 enhances liganddependent transcription of the ER reporter gene, whereas the truncation mutant lacking the RAC3/ER-interacting domain has no effect on ER activity. Intriguingly, hMMS19 strongly A, dose-dependent inhibition of ligand-induced transcriptional activity of ER␣ by RS2.1. Increasing concentrations of RS2.1 (Ϫ, 0 ng; ϩ, 100 ng; ϩϩ, 200 ng; ϩϩϩ, 400 ng) were transfected into HEK293 cells together with 20 ng of ER␣ and 0.5 g of the ERE-E1A-Luc reporter. After transfection, cells were treated with solvent or 25 pM of E2, and relative luciferase activity was determined as described under "Materials and Methods." B, dosedependent inhibition of ligand-induced transcriptional activity of ER␤ by RS2.1 conducted as described in A except that ER␤ is used. C, coexpression of RAC3 reverses RS2-1-mediated repression of ER␣ transcriptional activity. D, RS2.1 had no effect on transcriptional activity of hRAR␣ or hTR␤ on ␤-RARE-tk-Luc or TRE-tk-Luc reporter, respectively.
FIG. 5. The human MMS19 is an AF-1-specific transcriptional coactivator of ER. A, the full-length hMMS19, but not hMMS19-mt, consistently enhances ER␣-mediated transcriptional activation. B, coexpression of hMMS19 and RAC3 further enhances transcriptional activation of ER␣. C, the hMMS19 enhances transcriptional activation by the ER␣ AF-1 domain in the absence or presence of E2. The ER␣ AF-1 contains the N-terminal A/B region and the DNA binding domain of ER␣. The ER␣ AF-2 contains the DNA binding domain and the D-F region. Note that cotransfection of hMMS19 significantly enhances transcriptional activation by ER␣ AF-1 but has no effect on ER␣ AF-2. enhances the AF-1 activity of ER in both the absence and presence of ligand and has no effect on the activity of a separated AF-2 domain. Therefore, we suggest that hMMS19 functions as an AF-1-specific coactivator of ER. Currently, only a few potential AF-1 coactivators were reported, such as the p68 RNA helicase (7), and the mechanism of AF-1 coactivator is largely unknown. The human MMS19 is unique among other coactivators in that it interacts with both ERs and RAC3. Because hMMS19 interacts with the PAS domain of RAC3 and regulates AF-1 activity of ER, it is intriguing to speculate that the PAS domain of RAC3 may be important for regulating AF-1 function. However, the mechanisms by which AF-1 activates transcription and synergizes with AF-2 are largely unknown, and further study is required to fully understand the mechanism of transcriptional activation by nuclear receptors. Our finding that hMMS19 is an AF-1 coactivator that also interacts with AF-2 coactivator RAC3 suggests that hMMS19 may bridge the functions of AF-1 and AF-2.
During preparation of this manuscript, another group reported that hMMS19 interacts directly with the XPD and XPB subunits of TFIIH, confirming a regulatory role of hMMS19 on TFIIH activity (8). TFIIH is a multiprotein complex consisting of nine subunits, each participating in transcription and nucleotide excision repair pathways. TFIIH harbors several enzymatic activities such as a DNA-dependent ATPase linked to XPD and XPB and a cdk-activating protein kinase, a cyclin-dependent kinase that phosphorylates the polymerase II Cterminal tail domain. Recently, it was shown that the AF-1 domain of hER␣ is phosphorylated by the TFIIH cyclin-dependent kinase in a ligand-dependent manner (41). Furthermore, it was shown that phosphorylation of polymerase II C-terminal tail domain precedes the dissociation of ER/p160 coactivator complex from target promoters (50). Therefore, the identification of hMMS19 as an ER-and RAC3-interacting protein suggests a novel mechanism by which TFIIH may be recruited to ER target promoters. In light of the important role of TFIIH in regulating the transcriptional activity of ER (41), we suggest that hMMS19 might be involved in regulation of ER activity by bridging TFIIH with ER. Possibly, the interaction of hMMS19 with RAC3 might be involved in ligand-dependent recruitment of TFIIH to ER (41). Alternatively, hMMS19 might facilitate TFIIH-mediated phosphorylation of ER in specific promoters and cell types.
In summary, we have identified and characterized the human homologue of the yeast DNA repair and TFIIH regulator MMS19 as an RAC3/ER-interacting protein. The human MMS19 interacts with the N-terminal PAS-A/B domain of RAC3 through an evolutionarily conserved C-terminal domain. Overexpression of hMMS19 modulates ER-mediated transcriptional activation by enhancing the AF-1 function of ER. These data reveal a novel function of hMMS19 as an AF-1-specific transcriptional coactivator of estrogen receptor.