CREB-binding Protein/p300 Co-activation of Crystallin Gene Expression*

Although some of the transcription factors that are required for expression of crystallins during lens development have been identified, the molecular interactions that contribute to enhanced crystallin expression are not yet well defined. In this study, we designed experiments to test whether the co-activators CREB-binding protein (CBP) and/or p300 interact with c-Maf, Prox-1, or Sox-1 to enhance transcription of crystallin genes. Promoter regions from the mouse (cid:1) A-, (cid:2) B2-, and (cid:3) F-crystallin genes were linked to a luciferase reporter. each these or p300 c-Maf was co-activate each promoter. CBP and p300 were less effective or in-effective at co-activation with Prox-1 or Sox-1. Co-immu-noprecipitation and mammalian two-hybrid experiments revealed that CBP and p300 bind to c-Maf and Prox-1 but not to Sox-1. The co-activation of c-Maf by CBP/p300 requires histone acetyltransferase activity. Our results suggest that c-Maf recruits CBP and/or p300 to crystallin promoters leading to up-regulation of crystallin gene expression through localized histone acetylation and consequent chromatin re-modeling. In a pro-moter-specific fashion, co-activation can be modulated by Prox-1 and/or upon of and/or p300 to specific crystallin promoters, by localized histone acetylation and, we predict, consequent chromatin remodeling.

During embryonic development, the ocular lens is formed as an invagination of the surface ectoderm. As differentiation proceeds, the lens forms a hollow sphere of epithelial cells termed the lens vesicle (1). The cells located at the posterior surface of the lens vesicle are induced to differentiate into the primary lens fiber cells that elongate to fill the lens vesicle. The lens subsequently is comprised of a monolayer of proliferative cuboidal epithelial stem cells on the anterior surface overlying a core of elongated fiber cells (2,3). Lens induction is accompanied by localized expression of several different transcription factors including Pax-6, Six-3, Sox-2, c-Maf, Prox-1, and Sox-1 (4). Within the developing and the mature lens there is a distinctive pattern of crystallin gene expression. The ␣-crystallins are expressed at the lens vesicle stage, and are produced by both lens epithelial and fiber cells. The ␤and ␥-crystallins are synthesized by the fiber cells, with the expression of most ␤-crystallins preceding the expression of the ␥-crystallins (5,6). Altogether, the crystallins constitute more than 90% of the soluble lens proteins (7,8).
Gene targeting studies have shown that c-Maf, Prox1, and Sox1 are required for lens fiber cell differentiation and crystallin gene expression (9 -11). c-Maf is required for expression of all three major crystallin gene families. Prox-1 and Sox-1 are specifically required for the expression of ␥-crystallins (10 -12). Maf consensus binding sequences (MAREs) 1 have been identified in mouse ␣A-, ␤B2-, and ␥F-promoter regions (9,10,13). Prox-1 and Sox-1 responsive elements are also present in the ␥F-crystallin promoter region (9,12). Recently, it has been reported that Prox-1 and Six-3 act antagonistically to regulate the ␥F-crystallin promoter (12). It is not yet known how c-Maf, Prox-1, and Sox-1 interact with each other to regulate crystallin gene expression. It is also not known how these transcription factors specify the remarkably high levels of crystallin gene expression seen in lens cells.
CBP (CREB-binding protein) and p300 are well known transcriptional co-activators that have histone acetyltransferase (HAT) activity (14,15). They interact with many transcription factors and are required for cell differentiation and tissue development (16). CBP/p300 can be recruited to promoters by direct interaction with DNA-binding transcription factors or as components of large complexes containing other cofactors such as P/CAF, SRC-1, and the ACTR/p/CIP group of proteins (17,18). In previous studies we showed that inhibition of CBP/p300 activity in vivo (by expression of adenoviral E1A proteins in the lens of transgenic mice) resulted in inhibition of fiber cell differentiation and loss of ␤and ␥-crystallin gene expression (19). In the present study, we used a cell culture system to investigate the possibility that CBP and/or p300 may interact directly with the transcription factors that regulate crystallin gene expression.
We generated three crystallin promoter-luciferase reporters (␣A, ␤B2, and ␥F), as well as a MARE-TK-luciferase reporter, and transfected COS-1 and human lens epithelial cells (HLEC-B3) with each reporter and with plasmids encoding c-Maf, Prox-1, and/or Sox-1 in the presence or absence of CBP or p300.
Our results indicate that the crystallin promoters are synergistically co-activated by c-Maf⅐CBP and c-Maf⅐p300 complexes. CBP and p300 bind directly to c-Maf and co-activation is dependent upon HAT activity. We also demonstrate that Prox-1 can enhance, and Sox-1 can inhibit, this co-activation. Thus, lens-specific transcription factors interact with CBP and/or p300 to provide differential regulation of individual crystallin gene expression.
Plasmids for Immunoprecipitation-A start codon followed by two "Myc" tag sequences (EQKLISEEDL) was placed in-frame and upstream from c-Maf, Prox-1, or Sox-1 coding sequences in the pSG5 (Stratagene) vector. For the vector pSG5-p300-F2, two "FLAG" tag sequences (DYKDDDDK) followed by a stop codon were placed in-frame and downstream from the human p300 cDNA sequence corresponding to amino acids 1-2378.

Cell Culture, Transient Transfections, and Mammalian
Two-hybrid Assays COS-1 and HLEC-B3 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum overnight at 37°C before transfection. For transient transfections, 200 ng of a luciferase reporter were co-transfected with 50 ng of pCMX-␤gal control plasmid plus 20 -320 ng of one or more of the transcripition factor plasmids per well of a 24-well tissue culture dish unless otherwise indicated. For the mammalian two-hybrid assays, 200 ng of pG5E1b-Luc reporter plasmid, 50 ng of TK-growth hormone control plasmid, and 100 ng each of the pCMX-Gal4-N and pCMX-VP16-N chimeric expression plasmids were used per well of a 24-well tissue culture dish. Plasmid DNAs were mixed with 2.5 M CaCl 2 and then added to BESbuffered saline. The DNA mixtures were added to the cells in 1 ml of fresh Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. After incubation for 8 -12 h, the cells were washed with phosphate-buffered saline, then incubated with Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum for 24 -36 h prior to harvesting and assaying for luciferase activity on a Dynex MLX luminometer as described (27). The ␤-gal activity was assayed using the Tropix ␤-galactosidase kit according to the manufacturer's instructions (Tropix). Growth hormone secretion was assayed using the human growth hormone transient gene expression kit according to the manufacturer's instructions (Nichols Institute). Luciferase expression was normalized to the ␤-gal or growth hormone control. Transfections were done in triplicate and were replicated with similar results in at least three independent experiments. Results are the mean Ϯ S.D. from three individual experiments. All statistical analyses were performed with two-tailed Student's t tests using Microsoft Excel software. Data were considered to be significantly different for p Ͻ 0.05.

Transactivation of Crystallin Promoters by c-Maf, Prox-1,
and Sox-1-Our previous transgenic studies showed that lensspecific expression of adenoviral E1A proteins that bind to CBP/p300 causes inhibition of lens fiber cell differentiation and nearly complete loss of crystallin expression (19). Although the lens phenotypes resembled the c-maf knockout mice (9), the expression of c-maf in lens fiber cells expressing the E1A transgene was not affected (19), suggesting that c-Maf-mediated transactivation of crystallin genes may require CBP/p300 coactivation. Previous studies have shown that Prox-1 and Sox-1, like c-maf, are expressed at early stages of lens development, and are required for specific crystallin (␥-crystallin) expression (10,11). We therefore postulated that expression of lens-specific crystallin genes might require interactions of ocular transcription factors with the co-activators CBP and p300.
To test this prediction, we designed experiments to compare the abilities of c-Maf, Prox-1, and Sox-1 to transactivate crystallin promoters. c-Maf has previously been shown to transactivate the ␥F-crystallin promoter in a cell culture system (29). An electrophoretic mobility shift assay showed that c-Maf could also bind to DNA fragments from the mouse ␤B2and ␣Acrystallin promoters (9). Putative MAREs have been identified in the mouse ␣A-(position Ϫ110/Ϫ98), ␤B2-(position Ϫ110/ Ϫ98), and ␥F-(position Ϫ47/Ϫ35) crystallin promoters (9, 30 -32). Based on these studies, we subcloned mouse ␣A, ␤B2, and ␥F promoter DNA fragments into the pGL3 basic luciferase vector ( Fig. 1) for transient transfection studies. Monkey kidney cells (COS-1) were co-transfected with each crystallin reporter ( Fig. 1) and with plasmids encoding the different transcription factors. As predicted, c-Maf transactivated the three promoters in a dose-dependent fashion (Fig. 2, A-C). Sox-1 transactivated the ␥F-crystallin promoter in a dose-dependent fashion, similar to c-Maf (Fig. 2C). Sox-1 also stimulated the ␤B2 promoter, but less strongly than c-Maf (Fig. 2B). Prox-1 (80 ng) stimulated the ␤B2and ␥F-crystallin promoters only about 2-fold (Fig. 2, B and C). To determine whether these transcription factors can collaborate to transactivate crystallin promoters, co-transfections were done using 60 ng of each transcription factor. We observed cooperative transactivation using Prox-1 plus c-Maf, especially for the ␤B2 reporter (Fig.  2D, lanes 5, 11, and 17) (p Ͻ 0.05). Co-expression of Sox-1 with c-Maf significantly increased ␥F promoter activation (Fig. 2D,   FIG. 3. CBP/p300 co-activation. A-C, COS-1 cells were co-transfected with each crystallin reporter and 60 ng each of the plasmids encoding c-Maf, Prox-1, or Sox-1 in the presence or absence of CBP or p300. Expression of CBP or p300 significantly increased (i.e. co-activated) c-Maf-dependent activation of each crystallin promoter (about 3-fold) (lanes 7 and 8 in A-C). None of the other transcription factors gave more than a 2-fold increase in promoter activity, even in the presence of CBP or p300. D, human lens epithelial cells (B3) were co-transfected with the ␣A-crystallin reporter and 60 ng each of the plasmids encoding c-Maf, Prox-1, or Sox-1 in the presence or absence of CBP. We observed stimulation of the ␣A promoter by c-Maf (3-fold) (lane 2). Co-activation was seen with Prox-1 plus c-Maf and with CBP plus c-Maf co-transfections (lanes 5 and 8) (p Ͻ 0.05). CBP did not significantly co-activate Prox-1 or Sox-1 activity (p Ͼ 0.05) (lanes 9 and 10). Co-expression of Sox-1 with c-Maf repressed the c-Maf-mediated ␣A promoter activation (p Ͻ 0.001) (lane 6). E, COS-1 cells were co-transfected with ␣A reporter and c-Maf, CBP, or p300, plus or minus two E1A constructs (E1A⌬N or E1A⌬CR2). Co-expression of E1A⌬CR2, which binds to CBP/p300, blocks CBP or p300 co-activation of the ␣A promoter (p Ͻ 0.001) (lanes 9 and 12). Fold activation is expressed relative to the Luc/␤-gal ratio obtained after co-transfection of the specific reporter (␣A, ␤B2, or ␥F) with the empty pCMX-D expression vector, which is arbitrarily set at 1. CBP and p300 Enhance c-Maf-mediated Transactivation of Crystallin Genes-To assay for CBP/p300 co-activation of crystallin promoter activity, COS-1 cells were co-transfected with each of the three crystallin reporters and plasmids encoding c-Maf, Prox-1, or Sox-1 in the presence or absence of CBP or p300. Although CBP or p300 alone did not stimulate any of the crystallin promoters (Fig. 3, A-C, lanes 5 and 6, p Ͼ 0.05), both proteins enhanced (or co-activated) c-Maf-dependent crystallin promoter activity for all three promoters (lanes 7 and 8, p Ͻ 0.01). In contrast, expression of CBP or p300 did not significantly alter the promoter activation induced by Prox-1 or Sox-1 (p Ͼ 0.05) (Fig. 3, A-C). Human lens epithelial cell line B3 (HLEC-B3) was also used for transient transfection assays (Fig. 3D). Synergistic co-activation of the ␣A-crystallin promoter was seen upon transfection with CBP plus c-maf (p Ͻ 0.01, lane 8). Stimulation of the promoter also occurred with c-Maf plus Prox-1 (p Ͻ 0.05, lane 5). Co-activation was minimal for the other combinations of factors. In addition, COS-1 cells were co-transfected with the ␣A-crystallin promoter and c-maf, CBP, or p300, plus two versions of the E1A gene of adenovirus: E1A⌬N (which does not bind to CBP/p300) and E1A⌬CR2 (which binds to and inhibits both CBP and p300) (19). Expression of E1A⌬CR2 completely blocked co-activation of the ␣Acrystallin promoter (p Ͻ 0.001, Fig. 3E, lanes 9 and 12).
Binding of CBP/p300 to c-Maf and Prox-1-Since CBP and p300 have been shown to bind to numerous sequence-specific transcription factors (16), we tested whether c-Maf, Prox-1, or Sox-1 can bind directly to CBP/p300 by co-immunoprecipitation. When COS-1 cells were co-transfected with Myc-tagged c-Maf and FLAG-tagged p300, antibodies to either Myc or FLAG were able to co-precipitate the two proteins (Fig. 5A, lane  4). Similar results were obtained when COS-1 cells were cotransfected with Myc-tagged c-Maf and full-length CBP (Fig.  5B). When COS-1 cells were co-transfected with Myc-tagged Prox-1 or Myc-tagged Sox-1 plus FLAG-tagged p300, the anti-FLAG antibody was able to co-precipitate Prox-1 but not Sox-1 (Fig. 5C, bottom panel, lanes 6 and 5, respectively). To investigate whether c-Maf normally is bound to CBP/p300 in the lens, co-immunoprecipitations were performed using newborn mouse lens lysates. As shown in Fig. 5D, anti-p300 antibody precipitated a c-Maf protein.
To further analyze the binding of c-Maf and Prox-1 to p300, a mammalian two-hybrid assay was established (Fig. 6). The DNA-binding domain of Gal4 was fused to full-length p300 and to several subregions of p300. At the same time, full-length c-Maf, Prox-1, and Sox-1 were linked to the transactivation domain of the herpes simplex virus protein VP16. The reporter construct (G5E1b-Luc) contains an E1b promoter cloned down-  5 and 6), whereas co-expression of Sox-1 repressed the co-activation (p Ͻ 0.01) (lanes 7 and 8). B, diagram of the MARE-TK-Luc construct. C, COS-1 cells were co-transfected with this reporter and 60 ng each of the plasmids encoding c-Maf, Prox-1, or Sox-1 in the presence or absence of CBP or p300. Prox-1 was found to increase co-activation by c-Maf⅐CBP (p Ͻ 0.05) (lane 11) or c-Maf⅐p300 (p Ͻ 0.05) (lane 12). In contrast, Sox-1 inhibited co-activation (p Ͻ 0.01) (lanes 13  and 14). Fold activation is expressed relative to the Luc/␤-gal ratio obtained after co-transfection of the specific reporter (␤B2 or MARE-TK) with the empty pCMX-D expression vector, which is arbitrarily set at 1. D, diagrammatic representation of MARE-TK-Luc regulation.
By comparison, when COS-1 cells were co-transfected with VP16-Prox-1 or VP16-Sox-1 plus Gal4-p300 (full-length), expression of Prox-1 (Fig. 7, lane 8) but not Sox-1 (Fig. 7, lane 9) dramatically increased promoter activation, confirming that Prox-1 but not Sox-1 can directly bind to full-length p300. When the cells were co-transfected with VP16-Prox-1 and various subregions of p300 (1-1235, 1256 -2415, and 1190 -1966) fused to Gal4, we did not observe significant changes in promoter activation (Fig. 7, lanes 13-15), suggesting that the binding of Prox-1 to p300 may be dependent upon the configuration of full-length p300. CBP/p300 Co-activation Requires HAT Activity-To ascertain whether the HAT activity of CBP/p300 is essential for co-activation of the crystallin genes, we constructed plasmids encoding two different mutations of p300. The p300 mut1504 clone has a point mutation in the HAT domain (alanine codon substituted for a tyrosine codon at amino acid 1504), whereas ⌬1430 -1504 has a deletion in the HAT domain (from amino acids 1430 to 1504). Mut1504 corresponds to a point mutation in CBP (mutation 1541) found to eliminate HAT activity (33). The ⌬1430 -1504 deletion removes several important domains of p300 that correspond to motifs required for CBP HAT activity (33). Co-expression of c-Maf with either p300 mut1504 or ⌬1430 -1504 produced no co-activation of the ␥F-crystallin promoter or the MARE-TK promoter (Fig. 8, A and B, last two  lanes). To test whether the HAT activity of CBP/p300 is sufficient to co-activate c-Maf, we constructed a plasmid encoding a c-Maf-HAT fusion protein (c-maf coding region fused with the HAT domain of p300). We found that expression of the c-Maf-HAT fusion protein activated the ␥F-crystallin and MARE-TK promoters to the same extent as c-Maf plus p300 (Fig. 8, C and  D, two lanes on the right), indicating that histone acetylation is sufficient to enhance c-Maf activation of crystallin gene expression. DISCUSSION We have used a cell culture system to begin to study the mechanism(s) by which CBP and p300 become involved in crystallin promoter activation. We demonstrate that representative crystallin promoters are transactivated by the lens-specific c-Maf transcription factor, and that the promoter activity can be significantly enhanced by the co-activators CBP and/or , coexpression of VP16-Prox-1 but not VP16-Sox-1 together with full-length Gal4-p300 dramatically increased the luciferase activity in the mammalian two-hybrid assay (lanes 8 and 9). However, co-expression of VP16-Prox-1 with various subregions of p300 (1-1235, 1256 -2415, and 1189 -1966) fused to Gal4 did not activate the reporter (lanes 13-15). Fold activation is expressed relative to luciferase/growth hormone activity obtained after co-transfection of the specific reporter (G5E1b reporter) and empty expression vector. p300. Co-activation is mediated by binding of c-Maf to a specific domain of p300 (and presumably also of CBP). In addition, co-activation involves histone acetyltransferase activity and can be influenced by Prox-1 and Sox-1. Our results indicate that high levels of crystallin gene expression in the lens are dependent upon binding of c-Maf plus CBP and/or p300 to specific crystallin promoters, followed by localized histone acetylation and, we predict, consequent chromatin remodeling.
Regulation of ␣A-Crystallin-In the mouse, ␣-crystallins are first expressed at the lens vesicle stage. Both ␣Aand ␣Bcrystallins are synthesized by lens epithelial and early fiber cells. At later stages ␣A-crystallin continues to be expressed in both cell types (although more strongly in early fiber cells), but ␣B-crystallin persists only in epithelial cells (34). Previous studies indicated that transcriptional regulation of the mouse ␣A-crystallin gene is dependent upon a cAMP-responsive element (DE1/CRE) and a Pax-6-binding site (30), indicating that Pax-6 plays a critical role in transactivation of the ␣A-crystallin gene. More recently, c-maf null mice were found to have defects of ␣A-crystallin expression (9), and a Maf responsive element (13) was found to be present in the ␣A-crystallin promoter region. Consistent with these later findings, our results indicate that expression of c-Maf significantly transactivates the ␣A-crystallin promoter. We also found that c-Maf can recruit the co-activators CBP/p300 to the ␣A-crystallin promoter. Although Pax-6 is required for ␣A-crystallin expression (30), the requirement may be indirect since recent evidence shows that c-maf expression can be up-regulated by Pax-6 (35). In our tissue culture system, we co-expressed Pax-6 plus c-Maf and CBP or p300 but did not see a significant further increase in promoter activity (data not shown). Therefore, we predict that Pax-6 does not directly interact with the c-Maf⅐CBP/p300 complex. Currently, there is no evidence that the 400-bp ␣A-crystallin promoter that we tested here contains Prox-1-or Sox-1binding sites. We did not observe significant transactivation of the ␣A-crystallin promoter by Prox-1 or Sox-1. Interestingly, expression of Sox-1 significantly repressed c-Maf transactivation and c-Maf⅐CBP/p300 co-activation of the ␣A promoter (p Ͻ 0.001, Figs. 2D and 3D). It is unknown whether Sox-1 competes with c-Maf for binding to the ␣A-crystallin promoter.
Regulation of ␤B2-crystallin-The ␤-crystallins can be divided into ␤A-(A1, A2, A3, and A4) and ␤B-(B1, B2, and B3) crystallins, and are considered as early markers of lens fiber cell differentiation (7). Similar to the ␣A-crystallin promoter, a putative MARE has been identified in the ␤B2-crystallin promoter in both mouse and rat (9,36). A recent study (37) indicates that c-Maf is not essential for the activity of the ␤B2crystallin promoter in the rat lens. Instead, a putative Soxbinding site at Ϫ164/Ϫ159 and a positive element at Ϫ14/Ϫ7 seem to be the primary regulatory elements (37). In our transfection study, we observed weaker stimulation of the ␤B2 promoter by Sox-1 than by c-Maf. In fact, Sox-1 was found to block c-Maf⅐CBP/p300 mediated co-activation, suggesting that Sox-1 may not be a significant positive regulator of the mouse ␤B2crystallin gene. We did find that c-Maf-mediated transactivation of the mouse ␤B2-crystallin promoter activity can be en- FIG. 8. CBP/p300 co-activation requires HAT activity. A and B, COS-1 cells were co-transfected with the ␥F-crystallin reporter or the MARE-TK-luc reporter plus plasmids encoding c-Maf, full-length p300, p300 with a point mutation (mut1504), or p300 with a deletion (⌬1430 -1504) in the HAT domain. The p300 mutants did not co-activate either reporter. C and D, COS-1 cells were co-transfected with the ␥F-crystallin reporter or the MARE-TK-luc reporter and plasmids encoding c-Maf, full-length p300, or c-Maf-HAT (encoding c-Maf fused with the HAT domain of p300). Expression of the c-Maf-HAT fusion protein was sufficient to co-activate both reporters (p Ͻ 0.001). Fold activation is expressed relative to luc/␤gal activity obtained after co-transfection of the specific reporter (␥F or MARE-TK) and an empty expression vector. hanced by Prox-1 (Fig. 2D, lane 11). Since this cooperative stimulation was significant with the MARE-TK-luc reporter (p Ͻ 0.05, Fig. 4C, lane 7), which does not have a Prox-1binding site, and since Prox-1 can bind to CBP/p300, it suggests that Prox-1 may be recruited indirectly to the mouse ␤B2crystallin promoter. Developmental regulation of the ␤B2-crystallin promoter may involve an additional level of complexity that is not reflected by our tissue culture assays. Previous studies have shown that expression of mouse ␤B2-crystallin mRNA is undectable at embryonic day 11.5 (E11.5) (9), and that mouse (and rat) ␤B2-crystallin protein levels are low prior to birth, but are up-regulated postnatally (38). It is not yet known whether protein levels reflect transcriptional or posttranscriptional changes during development. Although our current and previous studies (9,19) indicate that c-Maf and CBP/ p300 are necessary for high level mouse ␤B2 promoter activity, it is important to remember that the endogenous gene is likely to have additional important regulatory domain(s) that are not present in the promoter region that we tested here.
Regulation of ␥F-crystallin-The ␥-crystallins are encoded by a cluster of six genes (␥A-␥F) (7). In the mouse, expression of ␥-crystallins begins around embryonic day 13 (E13), and is restricted to maturing fiber cells in the lens (39). Previous loss-of-function studies showed that mice lacking Sox-1 have an almost complete absence of ␥-crystallin expression (10). c-maf and Prox-1 null mice also showed reduced expression of ␥-crystallins (9,11). A MARE and a Sox-1-binding site are present in the ␥F-crystallin promoter region (10,29). Our experimental results show that either c-Maf or Sox-1 can significantly transactivate the ␥F-crystallin promoter, which is consistent with previous findings. In addition, we found that c-Maf and Sox-1 can have a cooperative interaction at this promoter (Fig. 2D). A recent study indicates that the ␥F promoter also contains a Prox-1 responsive element, and that expression of Prox-1 in CD5A (human lens epithelial) cells significantly enhances promoter activation (12). In our transfection studies using COS-1 and HLEC-B3 cell lines, we observed a slight, but significant (p Ͻ 0.05), transactivation of the ␥F-crystallin promoter by Prox-1. We did not observe CBP/p300 co-activation with either Prox-1 or Sox-1 (Fig. 3C), suggesting that Prox-1-or Sox-1mediated transactivation of the ␥F-crystallin gene might use an alternative co-activator system. Given the fact that Sox-1 cooperates with c-Maf to enhance ␥F-crystallin but repress ␣A-crystallin promoter activity, our studies suggest that the patterns of crystallin gene expression in the lens are modulated by interactions between the lens-specific transcription factors and their co-activators.
CBP/p300 Co-activation-CBP and p300 are transcriptional co-activators that are often required for cell differentiation and tissue development. Since CBP and p300 have HAT activity (14, 15,17), histone acetylation and consequent chromatin remodeling are considered to be important for gene activation (40,41). Our previous transgenic studies have shown that c-Maf expression in the lens is not sufficient to activate ␤and ␥-crystallin expression in the absence of CBP/p300 function (19). Our current observations provide an explicit molecular model for the role of CBP/p300 in crystallin gene regulation, which is consistent with previous models for CBP/p300 function. Specifically, crystallin gene expression appears to be activated by sequence-specific binding by the c-Maf transcription factor, recruitment of the co-activators CBP/p300, and subsequent localized histone acetylation. Prox-1 and Sox-1 probably act to differentially modulate this activity for specific crystallin genes. Our studies do not establish the mechanism(s) by which Sox-1 can repress CBP/p300 co-activation. One possible explanation is that Sox-1 can recruit repressors to block c-Maf⅐CBP/ p300 binding to the crystallin promoters. Another possible explanation is that Sox-1 may indirectly inhibit the HAT activity of CBP/p300. Although additional features of crystallin gene regulation surely exist, our studies pinpoint the interaction between c-Maf and CBP/p300 as being essential for the high levels of crystallin gene expression that are seen in the fiber cells of the ocular lens.