Effects of B-Myb on Gene Transcription PHOSPHORYLATION-DEPENDENT ACTIVITY AND ACETYLATION BY p300*

The transcription factor B-Myb is a cell-cycle regulated phosphoprotein involved in cell cycle progression through the transcriptional regulation of many genes. In this study, we show that the promoter of the fibroblast growth factor-4 ( FGF-4 ) gene is strongly activated by B-Myb in HeLa cells and it can serve as a novel diagnostic tool for assessing B-Myb activity. Specifically, B-Myb deletion mutants were examined and domains of B-Myb required for activation of the FGF-4 promoter were identified. Using phosphorylation-deficient mutant forms of B-Myb, we also show that phosphorylation is essential for B-Myb activity. Moreover, a mutant form of B-Myb, which lacks all identified phosphorylation sites and which has little activity, can function as a dominant-negative and suppress wild-type B-Myb activity. Acetylation is another post-translational modification known to affect the activity of other Myb family members. We show that B-Myb is acetylated by the co-activator p300. We also show that the bromo and histone acetyltransferase domains of p300 are sufficient to interact with and acetylate B-Myb. These data indicate that phosphorylation

. A MBS is not always necessary for B-Myb activity (2,7), which argues that B-Myb may also transactivate via proteinprotein interactions. A second conserved domain is the acidic region, which has been shown to contain transactivation potential (8). This domain is regulated by the carboxyl-terminal tail of B-Myb (9). In addition, the transactivation potential of B-Myb has been shown to be cell line-specific and dependent on binding of a putative co-factor to the carboxyl-terminal tail (10), indicating that other domains in B-Myb are involved in regulation of the transactivation domain. The third region of similarity is the conserved region (CR) of B-Myb. The CR shares homology with other Myb family members, but diverges in structure by the absence of the leucine zipper and the EVES motif found in c-Myb (11,12). These structural differences may account for the positive regulatory role of the CR in B-Myb (8,9,13,14); whereas, the CR in c-Myb has a negative role (1). Finally, the CR is flanked by two nuclear localization signals (15).
The expression and biological activity of B-Myb is regulated at multiple levels in a cell cycle-dependent manner. In quiescent cells, the expression of B-Myb mRNA is repressed, but upon entering the cell cycle, expression is up-regulated in late G 1 and maximally expressed during S-phase (16 -18). Although not well understood, B-Myb is also regulated at the protein level by phosphorylation. During S-phase, B-Myb can be hyperphosphorylated by cyclin-dependent kinase 2 in conjunction with cyclin A1, cyclin A2, or cyclin E (19 -24). The phosphorylation of B-Myb can affect DNA binding affinity and transactivation potential (9,(21)(22)(23)25). Tissue distribution of the Myb proteins may also play a role in their function. Although A-Myb and c-Myb have restricted tissue expression (3,26), B-Myb is expressed in all tissues and cell types examined thus far (3). Furthermore, recent reports indicate that B-Myb is the only Myb expressed in embryonic stem cells and it has been shown to be required for development of the early embryo (16,27,28).
MBS have been found in the promoters of numerous genes that can be activated by Myb proteins (reviewed in Ref. 29). Sequence analysis suggests that the fibroblast growth factor-4 (FGF-4) promoter may contain a potential MBS. The FGF-4 gene is controlled by a 5Ј promoter region and a powerful distal enhancer (30,31). The promoter region contains four identified cis-regulatory elements: a TATA box, which is bound by the basal transcription machinery; a CCAAT box, which is bound by the transcription factor NF-Y (32)(33)(34); and two GC boxes, which are bound by the transcriptions factors Sp1 and Sp3 (35,36). The enhancer, located in the untranslated region 3 kb downstream from the transcription start site, is also required for expression of the FGF-4 gene (30,31). The enhancer con-tains three identified cis-regulatory elements: a HMG motif, which binds Sox-2 (35,(37)(38)(39); a POU motif, which binds Oct-3 (31,35,37,39,40); and a GT box, which binds Sp1 and Sp3 (35)(36)(37). Recently, the co-activator p300 has been implicated in transcription of the FGF-4 gene where it may mediate the effects of Sox-2 and Oct-3 (47). p300 and the related family member CREB-binding protein (CBP) are co-activators that mediate transactivation of RNA polymerase II (41). p300/CBP presumably function by acting as a bridge between transcription factors and the general transcription initiation machinery (42,43). p300/CBP also possess the ability to acetylate histones and other proteins (44). Interestingly, the transcription factors Sp1, Sp3, and NF-Y have also been shown to interact with p300 (45)(46)(47).
In this study, we examined whether B-Myb influences the activity of the FGF-4 promoter and/or enhancer. We determined that B-Myb substantially activates the FGF-4 promoter in HeLa cells. By utilizing mutant forms of B-Myb, we demonstrate that specific domains are required for B-Myb activity. In a similar manner, we demonstrate that the phosphorylation of B-Myb plays an important and necessary role in B-Myb activity. Finally, we show that B-Myb interacts with and is acetylated by p300 in vivo, implicating acetylation as another means of B-Myb regulation. Our findings further demonstrate that the FGF-4 promoter provides a novel system to probe the mechanisms by which B-Myb activates transcription. In this regard, the FGF-4 promoter is stimulated as much as 35-fold by B-Myb, which is substantially above that observed with many other promoters (8,13,14,48,49).

EXPERIMENTAL PROCEDURES
Cell Culture, Transient Transfection, and Promoter/Reporter Assays-HeLa and 293T cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum (Hyclone) and incubated at 37°C in 5% CO 2 . Chinese hamster ovary cells were maintained in F-12 supplemented with 10% fetal bovine serum and incubated at 37°C in 5% CO 2 . HeLa and 293T cells were seeded 24 h prior to transfection with the calcium phosphate precipitation method as described previously (31,50). 14 -16 h later, the cells were washed and refed with Dulbecco's modified Eagle's medium ϩ 10% fetal bovine serum. The following day, cell lysates were prepared and chloramphenicol acetyltransferase and ␤-galactosidase activities were determined as described previously (51). Null vectors were used to maintain the total amount of DNA transfected. To adjust for differences in transfection efficiency, all transfections were normalized with pCMV␤-galactosidase (CLONTECH, Palo Alto, CA). All transfections were performed in duplicate or triplicate with representative experiments shown. Plasmid DNA was purified by Qiagen tip-500 columns (Qiagen, Valencia, CA).
Promoter/Reporter Gene Constructs and Expression Plasmids-The promoter/reporter gene constructs pkFGF4 -427TϩE (referred to in this paper as Ϫ427TϩE), pBLCAT7, and Sp1DP have been described previously (31,36). Promoter/reporter construct Ϫ427T was created by removal of the FGF-4 enhancer with a SacI digest and religation. Ϫ61T was created by digesting Ϫ427T with SphI and SacII and religation. The promoter/reporter gene construct Ϫ427TϩE CCAATmut was generated by the QuikChange method of mutagenesis (Stratagene) using the FGF-4 promoter/reporter gene construct Ϫ427TϩE. The complimentary primers (synthesized by the Eppley Cancer Institute Molecular Biology Core Facility) were used to disrupt the CCAAT box and putative MBS (mutated bases in lowercase) while creating a BamHI site (underlined): 5Ј-CCCCCGGCGGTGggatcCAcGCGGCCTGCGCCC-3Ј. Positive clones were selected by digestion with BamHI. Ϫ427T CCAATmut, which lacks the FGF-4 enhancer, was made by removal of the enhancer from Ϫ427TϩE CCAATmut with a SacI digest and religation. To disrupt the TATA box and create Ϫ427T TATAmut, Ϫ427TϩE was digested with KasI and BssHII. This removed a small fragment containing the wild type TATA box. Next, two oliogonucleotides containing a mutated TATA box (lowercase), (5Ј-GCGCCGGC-agatctccCCACTGCTCCGGAGGGCTGGG-3Ј and 5Ј-CGCGCCCAGC-CCAGCCCTCCGGAGCAGTGGggagatctGCCG-3Ј), were annealed and ligated into the KasI and BssHII sites of the remaining plasmid. This created Ϫ427T TATAmutϩE. The enhancer was subsequently deleted by SacI digestion and religation to create Ϫ427T TATAmut.
The B-Myb deletion mutants were created by polymerase chain reaction (PCR) amplification and ligation into the parent vector, pCMV5 FLAG-B-Myb. MutO was created by a two-step mega-primer method, using pCMV5 FLAG-B-Myb as the template. The mega-primer contained an in-frame splice of B-Myb sequences at 1120 -1145 and 1556 -1579, deleting 411 bp 5Ј-GTAAATTTGACCTTCCTGAAGAACC-CCTGGAGAG CCCCTCACTGACATC-3Ј. The downstream primer hybridized to B-Myb sequence 2437-2460: 5Ј-CAGGAGTGGACCTTG-TTCTATTAG-3Ј. PCR was performed with these primers, and the product of this reaction was used as a primer in the second reaction, together with the upstream primer that hybridized to B-Myb at 646 -667: 5Ј-G-GACAATGCTGTGAAGAATCAC-3Ј. The second PCR product and parent vector were digested with SalI and BsmI and then religated. GACATGAGGC-3Ј. The reverse primer was the complement. The PCR product was digested with DpnI to remove the template DNA. Clones were screened by digestion with SacI, which is present in the deleted portion of B-Myb. Other expression plasmids utilized have been described previously: mut1 and mut2 (13), CMV12SE1a and CMV12S mE1a⌬2-36 (52), CMVp300 (53), and Gal4-p300(964 -1922) (54).
Identification of Additional B-Myb Phosphorylation Sites-The method for labeling, electrophoresis, excision, tryptic digestion, HPLC, and phosphoamino acid analysis of B-Myb has been described previously (22,55). The 10mut B-Myb tryptic phosphopeptides were generated as described previously, with the exception that the amount of protein obtained was scaled down 10-fold. Consequently, after the HPLC fractions were subjected to automated Edman degradation, there was not enough protein left to directly identify the amino acid sequence. Therefore, each cycle was counted by ␤-scintillation to determine which amino acid position contained the radiolabeled phosphoamino acid. In addition, phosphoamino acid analysis was performed on each HPLC fraction to determine whether phosphoserine or phosphothreonine was present. The remaining phosphorylation sites were predicted by comparing the phosphoamino acid analysis data to a computer-generated tryptic digest of B-Myb listing all serines and threonines followed by a proline. This allowed for the determination of five additional potential phosphorylation sites. These sites were mutated individually and together using the 10mut as template. The resulting B-Myb phosphorylation mutant proteins were then mapped by HPLC analysis to confirm that the mutations eliminated the targeted phosphorylation sites (data not shown).
Immunoprecipitation and Western Blot Analysis-Cells were frozen on liquid nitrogen and solubilized in lysis buffer (50 mM HEPES, pH 7.8, 1% Triton X-100, 10 mM EDTA, 1 mM phenylmethylsulfonyl, 10 mM sodium fluoride, 1 mM sodium vanadate, 30 mM sodium pyrophosphate, 25 mM benzamidine, 20 g/ml aprotinin, pH 7.5). Insoluble matter was removed by centrifugation for 10 min at 15,000 rpm in a microcentrifuge. Clarified cell supernatants were immunoprecipitated overnight with the appropriate antibody conjugated to agarose beads with rotation at 4°C. Samples were washed 4 times with 50 mM HEPES (pH 7.8) before addition of sample buffer (20% glycerol, 2.5% SDS, 125 mM Tris (pH 6.8), 100 mM dithiothreitol, 0.001% bromphenol blue). The proteins were separated by 8% SDS-PAGE at 150 V until the dye-front was run off the gel. Proteins from the gel were transferred to polyvinylidene difluoride membranes in transfer buffer (15 mM Tris, 120 mM glycine, 20% methanol, pH 8.3) for 1 h at 40 V, 250 mA. The membrane was blocked for 1 h in TBST (10 mM Tris, pH 8, 150 mM NaCl, 0.05% Tween 20) with 3% bovine serum albumin, blotted using anti-FLAG M2 antibodies (1:3000 in TBST)(Sigma) and alkaline phosphatase-conjugated goat anti-mouse secondary antibodies (1:2000 in TBST) (Pierce, Rockford, IL), each followed by three 5-min washes in TBST. Proteins were detected by colormetric development with 33 l each of 50 mg/ml 5-bromo-4-chloro-3-indolyl phosphate and 50 mg/ml nitro blue tetrazolium in 10 ml of AP buffer (100 mM Tris, pH 9.5, 100 mM NaCl, 5 mM MgCl 2 ). Alternatively, proteins were detected using the enhanced chemifluorescence kit (Amersham Biosciences Inc.) following the manufacturers directions and visualized using a Storm PhosphorImaging cassette and the Storm PhosphorImager (Molecular Dynamics). The antibody for acetyl-lysine and the anti-FLAG antibody were obtained from Upstate Biotechnology (Lake Placid, NY) and Sigma, respectively. All Western blots were performed in duplicate or triplicate, with representative blots shown.
Cellular Immunofluorescence for Subcellular Localization-Chinese hamster ovary cells cultured on coverslips were transiently transfected with B-Myb or mutant B-Myb expression plasmids by LipofectAMINE (Invitrogen) according to the manufacturers directions. After 48 h, the cells were fixed with 2.5% paraformaldehyde for 10 min at 4°C, neutralized by 50 mM ammonium chloride, then permeabilized with 70% ethanol. The cells were washed with phosphate-buffered saline, the primary antibody (M2, 1:500 in phosphate-buffered saline) and secondary antibody (fluorescein isothiocyanate-conjugated anti-mouse 1:100) were added for 1 h each, and the cells were washed after each antibody. The coverslips were then mounted on slides using glycerol-gelatin with 2.5% 1,4-diazabicyclo-[2.2.2]octane (Sigma) or Vectashield (Vector Laboratories). Photographs were taken using a Nikon phase-contrast microscope utilizing a fluorescein isothiocyanate filter.

B-Myb Stimulates the FGF-4 Promoter-The transcription factor B-Myb is involved in the regulation of numerous genes.
Although B-Myb can bind and function through a consensus MBS, there is also evidence that B-Myb can function through protein-protein interactions in the absence of a MBS (2, 7). Sequence analysis of the FGF-4 promoter suggested that it may contain sequences that are able to bind B-Myb. We employed HeLa cells to examine the effects of B-Myb on the FGF-4 promoter, since these cells had been used previously to examine the function of FGF-4 regulatory regions (37,44,46,47). For this purpose, we employed the FGF-4 promoter/reporter construct, pk-FGF4 -427TϩE ( Fig. 1A) (referred to in this paper as Ϫ427TϩE), which contains 427 base pairs of the FGF-4 promoter driving expression of the chloramphenicol acetyltransferase gene along with the FGF-4 enhancer placed downstream. HeLa cells were co-transfected with Ϫ427TϩE and an expression plasmid coding for murine B-Myb (pCDNA3 FLAG-B-Myb). Interestingly, we observed a dramatic stimulation (Ͼ30-fold) of the FGF-4 promoter/reporter construct by B-Myb (Fig. 1B); whereas, B-Myb had no effect on the parent vector, pBLCAT7 (data not shown).
Since B-Myb can strongly stimulate the FGF-4 promoter/ reporter construct, we examined whether B-Myb was exerting its effect on the FGF-4 promoter and/or enhancer. To address this question, we used a second FGF-4 promoter/reporter construct, Ϫ427T, which lacks the FGF-4 enhancer (Fig. 1A). Transient transfections carried out in HeLa cells resulted in a dose-dependent stimulation in the absence of the FGF-4 en-hancer, which was equal to that of Ϫ427TϩE (Fig. 1B). This indicates that B-Myb does not require the FGF-4 enhancer to exert its effect. To further narrow the region through which B-Myb was exerting its effect, we examined the four regulatory sequences identified in the 5Ј-flanking region of the FGF-4 gene: a TATA box, a CCAAT box motif, and two GC boxes (Sp1 sites). Initially, we examined whether B-Myb required the CCAAT box motif of the FGF-4 promoter. This was of particular interest, because a potential MBS overlaps the CCAAT box motif. Previously, we determined that the CCAAT box is bound by the transcription factor NF-Y both in vitro and in vivo, and that NF-Y plays an important role in the regulation of the FGF-4 gene (34). The CCAAT box and potential MBS were abolished by site-directed mutagenesis (as described under "Experimental Procedures") to produce the plasmid Ϫ427T CCAATmut (Fig. 1A). Upon co-transfection of this plasmid along with a B-Myb expression construct, substantial stimulation of the promoter was observed (Fig. 1C). The stimulation of Ϫ427T CCAATmut by B-Myb was slightly less than that of Ϫ427T, but deletion of the CCAAT box has been shown to decrease the basal activity of the FGF-4 promoter (32,33). 2 This finding indicates that the CCAAT box and potential MBS are not required for B-Myb to exert its affect on the FGF-4 promoter.
Using the same approach, we examined the two Sp1 sites in the FGF-4 promoter. Previous studies have shown that B-Myb 2 L. R. Johnson and A. Rizzino, unpublished data. can interact indirectly with Sp1 and function through Sp1 cis-regulatory elements (7). To test the role of the Sp1 sites, we employed an FGF-4 promoter/reporter construct (Sp1DP) in which both Sp1 sites were scrambled by site-directed mutagenesis (36). Upon co-transfection of Sp1DP with increasing amounts of the B-Myb expression plasmid, a dose-dependent stimulation was observed (data not shown). The stimulation of Sp1DP by B-Myb was slightly less than that of Ϫ427T, but this may be due to the fact that mutation of the Sp1 sites causes a 25% decrease in basal activity of Ϫ427T (36). Thus, it appears that the Sp1 sites in the FGF-4 promoter are not required for B-Myb stimulation. Similarly, disruption of the TATA box by site-directed mutagenesis did not prevent B-Myb from stimulating the expression of the reporter gene (Fig. 1D). B-Myb continues to stimulate the FGF-4 promoter over 20-fold even though disruption of the TATA box caused a significant decrease in the basal activity of the promoter. Hence, none of the known cis-regulatory elements in the FGF-4 promoter are required for stimulation by B-Myb.
To help identify the minimal region of the FGF-4 promoter needed to respond to B-Myb, we used a promoter/reporter construct, Ϫ61T. This construct contains only 61 bp upstream of the FGF-4 transcription start site and does not contain a putative MBS. Transfection of HeLa cells with Ϫ61T along with increasing amounts of the B-Myb expression plasmid resulted in a substantial stimulation of the FGF-4 promoter/reporter construct (Fig. 1E). This finding indicates that only a small region of the FGF-4 promoter is required for the effect of B-Myb.
Identification of B-Myb Phosphorylation Sites and Their Importance for Activity-It has been shown that B-Myb is phosphorylated at the onset of S-phase and that phosphorylation may control the biological activity of B-Myb (9, 20 -23, 25). Previously, we identified 10 phosphorylation sites on B-Myb and mutation of these sites was observed to affect DNA binding, B-Myb transactivation potential, and the phosphorylation of other B-Myb sites (22). More recently, five additional phosphorylation sites (Ser 343 , Ser 424 and Thr 443 , Thr 447 , and Thr 490 ) were identified by tryptic digestion and phosphoamino acid analysis of the 10mut (see "Experimental Procedures"). The phosphorylation sites Thr 443 , Thr 447 , and Thr 490 have been described previously and mutation of these sites individually results in decreased transactivation potential (24,25). Ser 343 and Ser 424 are novel phosphorylation sites. The five additional phosphorylated serines and threonines were replaced with alanines and valines, respectively, by site-directed mutagenesis of 10mut to create 15mut B-Myb, which lacks all 15 identified phosphorylation sites ( Fig. 2A). Mutation of the sites was confirmed by the loss of the specific HPLC peaks when comparing the 15mut to the 10mut (data not shown).
To determine whether the phosphorylation of B-Myb influences its ability to stimulate the FGF-4 promoter, we utilized a panel of FLAG-tagged B-Myb phosphorylation mutant constructs ( Fig. 2A) (22). These mutant constructs were co-transfected into HeLa cells along with the FGF-4 promoter/reporter construct, Ϫ427T. We observed lower transcriptional activity with 10mut and 8mut than that observed with wild type B-Myb. However, the most dramatic difference was observed with 15mut, which barely stimulated the activity of the promoter/ reporter gene (Fig. 2B). These findings argue that phosphorylation of B-Myb is required for stimulation of the FGF-4 promoter. To confirm this possibility, we tested the effect of a dominant-negative form of Cdk2 (Cdk2DN), which was reported recently to decrease the phosphorylation of B-Myb and its ability to transactivate (56). At high concentrations, Cdk2DN inhibited the basal expression of the FGF-4 promoter/ reporter gene construct. However, at lower concentrations Cdk2DN selectively reduced the response of the promoter/reporter construct to B-Myb (data not shown). Interestingly, B-Myb mutants 267-455, 396 -455, 497-3TV, and 267ϩ283 exhibited transcriptional activity that was equal to or greater than wild-type B-Myb (Fig. 2B). Moreover, Western blot analysis determined that B-Myb and the B-Myb phosphorylation mutant proteins were expressed at nearly equal levels (Fig.  2C). Thus, these findings argue that several phosphorylation sites play a negative role in the transcriptional activity of Extract amounts were normalized based on ␤-galactosidase activity. Extracts were subjected to Western blot analysis with M2 (anti-FLAG antibody). Molecular mass marker is indicated on the right. Experiments were repeated at least twice and representative results are shown.
wild-type B-Myb. Furthermore, our findings indicate that there is not one specific phosphorylation site or region of phosphorylation sites responsible for B-Myb activity, but rather phosphorylation has an essential and complex effect on B-Myb transcriptional activity.
Our data not only indicate that phosphorylation of B-Myb is needed for transactivation, it appears that 15mut and, to a lesser extent, 10mut behave as dominant-negatives. To examine this possibility further, HeLa cells were transfected with the FGF-4 promoter/reporter construct Ϫ427T along with the B-Myb expression plasmid and increasing amounts of the 15mut expression plasmid. We observed a dose-dependent inhibition of B-Myb stimulation (Fig. 3A). Although this data suggested that 15mut could act as a dominant-negative, we considered the possibility that its expression might interfere with the expression of wild-type B-Myb. To test this possibility, HeLa cells were transfected with a truncated form of B-Myb (mutT), which lacks the final 245 amino acids of its COOH terminus. MutT was used in this case, because it can be readily distinguished from 15mut, due to its smaller size; whereas, B-Myb and 15mut migrate together during SDS-PAGE. Like B-Myb, mutT strongly stimulates the FGF-4 promoter. Importantly, 15mut reduced the stimulation of the reporter gene by mutT without reducing the protein level of mutT (Fig. 3B). Together, our findings argue that 15mut can function as a dominant-negative and block the action of B-Myb. Currently, the reason why 15mut behaves as a dominant-negative is unclear. It may exhibit an altered conformation, which may alter its ability to interact with other factors. Alternatively, phosphorylated residues may serve as docking sites for interacting proteins.
Domains of B-Myb Needed for Functional Activity in the Context of the FGF-4 Promoter-Domains of B-Myb have been characterized functionally in the context of natural and heterologous promoters (8,13,22,49,57). Based on our evidence that B-Myb can stimulate FGF-4 promoter constructs over 30-fold, we sought to determine the domains of B-Myb required for this effect. A panel of B-Myb mutants that have specific domains deleted was used for this purpose (Fig. 4A). Mut1 has the second and third repeat domains (R2 and R3) of the DNAbinding domain deleted. Mut2 has the acidic transactivation domain deleted. MutN lacks the amino-terminal portion of B-Myb along with the R1 domain of the DNA-binding domain. The R1 domain is not required for DNA binding (6). MutO lacks the region between the acidic region and the CR. MutC lacks the CR and the carboxyl-terminal domains. MutT has the carboxyl-terminal tail deleted, and ⌬CR lacks only the conserved region. Each of the B-Myb deletion mutant expression plasmids was co-transfected into HeLa cells along with the FGF-4 promoter/reporter construct Ϫ427T (Fig. 3B). Of this panel, only the mutT was able to stimulate the FGF-4 promoter to the same degree as full-length B-Myb, indicating that the region deleted in mutT is not necessary for the B-Myb stimulatory activity. The rest of the deletion mutants had substantially decreased ability to stimulate the FGF-4 promoter. Mut1 and mut2 had little to no activity when compared with full-length B-Myb. MutO, ⌬CR, and the 15mut had substantially decreased activity (7, 25, and 12% of full-length B-Myb, respectively). MutN and mutC had slightly higher activity than the other deletion mutants, but they were still decreased as compared with full-length B-Myb (32 and 40% of full-length B-Myb, respectively). Western blot analysis indicated that the proteins were expressed at roughly the same levels with the exception of mutC, mutT, and ⌬CR, which were expressed at slightly lower levels (Fig. 4C). Furthermore, similar results were obtained when this study was repeated with the FGF-4 promoter/reporter gene construct Ϫ61T (data not shown). Taken together, these findings suggest that each of the B-Myb domains examined, other than the carboxyl-terminal tail, are essential for stimulation of the FGF-4 promoter and/or are needed for proper folding of the protein.
Another explanation for the decrease in activity of the B-Myb mutants could be altered subcellular localization of B-Myb when specific domains are deleted. To examine this possibility, the subcellular localization of the B-Myb deletion mutants was analyzed by immunofluorescence. Chinese hamster ovary cells grown on coverslips and transiently transfected with the various B-Myb deletion mutants were fixed and stained with the anti-FLAG M2 antibody and a secondary fluorescein isothiocyanate-labeled antibody, then analyzed by fluorescence microscopy. Cells were scored as either exclusively nuclear or nuclear plus cytoplasmic (Fig. 5A). All mutants, with the exception of mutC and mutT, were predominantly nuclear (Fig.  5B). Interestingly, the 15mut was predominantly nuclear, which indicates that the nuclear localization of B-Myb does not require phosphorylation. However, mutC and mutT exhibited an altered distribution with the majority of cells scoring nu- clear plus cytoplasmic. These data argue that deletion of the carboxyl-terminal tail, which also removes the nuclear localization signal 2 (NLS2), alters its ability to localize primarily to the nuclear compartment. Nonetheless, it should be stressed that mutC and mutT are not excluded from the nucleus, since they stimulate the expression of the FGF-4 promoter/reporter gene construct (Fig. 4B). Our data also indicate that nuclear localization signal 1 (NLS1) is non-functional in this context, because mutO, which lacks NLS1, is predominantly localized in the nucleus. Furthermore, the CR may play a strong role in nuclear localization in conjunction with NLS2, as shown by the difference in the number of cells staining predominantly nuclear between mutC and mutT.
p300 Interacts with B-Myb-Based on previous reports that c-Myb and A-Myb can interact with the co-activator p300, or a related protein, CBP (58, 59), we examined whether p300 and B-Myb interact. Initially we used the viral protein E1a as a diagnostic tool, since it has been used to implicate p300/CBP in the transcription of other promoters (60), including the FGF-4 promoter (40). For this purpose, we utilized E1a (CMV12SE1a) and a mutant form of E1a (CMV12SE1a⌬2-36, referred to as mE1a in the remainder of this report), which cannot interact with p300 (52). HeLa cells were transiently co-transfected with the FGF-4 promoter/reporter construct, the B-Myb expression plasmid, and increasing amounts of the expression plasmid coding for E1a or mE1a. The stimulation of the FGF-4 promoter by B-Myb was substantially inhibited by E1a, whereas mE1a was far less inhibitory (Fig. 6). This data suggests that B-Myb may function through p300 to stimulate the FGF-4 promoter. To directly probe for an interaction between B-Myb and p300, we performed a co-immunoprecipitation assay. B-Myb and the 15mut, which are both FLAG-tagged, were individually cotransfected into 293T cells with or without an expression plasmid coding for p300 epitope-tagged with hemagglutinin (Fig.  7A). Cell lysates were immunoprecipited with an antibody against HA and subjected to Western blot analysis with the M2 antibody to visualize the presence of B-Myb or the 15mut. Both B-Myb and the 15mut co-immunoprecipitated with p300 (Fig.  7B), which indicates that these proteins interact with p300 either directly or as part of a complex. Although the 15mut appears to have greater affinity for p300 than wild-type B-Myb, this may not be the case, since p300 was expressed at a lower level in that sample containing wild-type B-Myb (Fig. 7B, bottom panel). This indicates that phosphorylation of B-Myb is not needed for interaction with p300.
In Vivo Acetylation of B-Myb by p300 -p300/CBP possess histone acetyltransferase (HAT) activity (44,61) and CBP has been shown previously to acetylate c-Myb (62). Therefore, we examined whether B-Myb or the B-Myb deletion mutants could be acetylated by p300. 293T cells were co-transfected with expression plasmids for p300 and B-Myb or the B-Myb deletion  Fig. 2 for abbreviations). The amino acids deleted in each mutant are labeled on the right. B, transfection into HeLa cells with 12 g of the indicated promoter/reporter gene construct Ϫ427T alone or in combination with the indicated amount of an expression plasmid for B-Myb or deletion mutant. Schematic of 15mut is found in Fig. 2. Results of the transfections are presented as fold increase over the promoter/reporter construct alone. C, expression of the FLAG-tagged B-Myb and B-Myb deletion mutants in transfected HeLa cells from Fig. 4B. Extract amounts were normalized based on ␤-galactosidase activity. Extracts were subjected to Western blot analysis with an antibody to FLAG (M2), or in the case of mut1 and mut2, an antibody against human B-Myb. Molecular mass markers (in kDa) are indicated on the left. The asterisk (*) indicates the expected band. Experiments were repeated at least twice and representative results are shown.

FIG. 5. Subcellular localization of B-Myb and B-Myb mutants.
A, Chinese hamster ovary cells grown on coverslips were transfected with expression plasmids for B-Myb or one of the B-Myb mutants. After 48 h, the cells were fixed with paraformaldehyde and the proteins were detected using an antibody to the FLAG epitope, and a fluorescein isothiocyanate-labeled secondary antibody. Photographs were taken using a phase microscope. B-Myb is a representation of predominantly nuclear staining and mutC represents nuclear ϩ cytoplasmic staining. B, cells from each slide were counted and scored as expressing B-Myb or B-Myb mutant either mostly nuclear, or as nuclear and cytoplasmic. Experiments were repeated and representative results are shown.
mutants. The various forms of B-Myb were immunoprecipitated using either the M2 antibody or a human B-Myb antibody (specific for non-FLAG-tagged mut1 and mut2) (22) and subjected to Western blot analysis. For this purpose, an antibody against acetyl-lysine was used to detect acetylation. With the exceptions of mut2 and mutC, B-Myb and the B-Myb mutants exhibited an increase in acetylation in the presence of p300 (Fig. 8A). Mut2 and mutC showed little to no acetylation by p300 compared with wild-type B-Myb. p300 interacts with the transactivation domain of other proteins, including c-Myb (58), p53 (63), BRCA-1 (64), and Ets-1 (65), therefore, it is likely that mut2, which lacks the transactivation domain, cannot interact with p300 and, thus, cannot be acetylated by p300. Alternatively, the region deleted in mut2 contains the majority of acetylation sites. Interestingly, 15mut exhibited high levels of acetylation in the presence of p300, which indicates that acetylation of B-Myb is not dependent on phosphorylation.
To narrow down the region of p300 required for interaction with and acetylation of B-Myb, co-immunoprecipitation experiments were carried out on cell lysates of 293T cells co-transfected with an expression plasmid for B-Myb and an expression plasmid for either p300 or Gal4p300-(964 -1922). Gal4p300-(964 -1922) (Fig. 7A) contains the bromo and HAT domain of p300 fused to the Gal4 DNA-binding domain (54). We demonstrate that the bromo plus HAT domains of p300 are sufficient for the acetylation of B-Myb (Fig. 8B). Finally, we examined whether p300 or Gal4p300-(964 -1922) would enhance the ability of B-Myb to stimulate the expression of the FGF-4 promoter/ reporter gene construct (Ϫ427T). Neither p300 nor Gal4p300-(964 -1922) increased the activity of the reporter gene. However, Gal4p300-(964 -1922) with a mutant HAT domain appears to function as a dominant-negative, since its expression inhibited the response of Ϫ427T to B-Myb by ϳ50% (data not shown). Although this suggests that acetylation plays a role in the activation of the FGF-4 promoter, it does not demonstrate that acetylation of B-Myb is required. This will require mapping each of the acetylation sites of B-Myb and determining their roles in transactivation by B-Myb. DISCUSSION In this study, we show that B-Myb can stimulate the FGF-4 promoter in promoter/reporter gene constructs in HeLa cells. We have identified five additional phosphorylation sites in B-Myb, two of which are novel. By mutation of all 15 identified sites, we determined that lack of phosphorylation renders B-Myb inactive. This phosphorylation mutant also functions as a dominant-negative of B-Myb. Furthermore, we show that specific domains of B-Myb are necessary for the stimulation of the FGF-4 promoter. Finally, we show that the co-activator p300 not only interacts with B-Myb, but it can acetylate B-Myb and many of the B-Myb deletion mutants. Moreover, only the bromo and HAT domains of p300 are necessary for interaction with and acetylation of B-Myb. These results argue that p300 is a co-activator of B-Myb and that p300 can modify B-Myb posttranslationally. Furthermore, our findings indicate that the FGF-4 promoter can serve as a novel diagnostic tool for assessing B-Myb activity. In previous reports where B-Myb was shown to promote gene transcription, the amount of stimulation by B-Myb alone has generally been relatively low. In many cases, only a 5-fold or lower stimulation was observed (8,13,14,48,49). Our studies indicate that the FGF-4 promoter provides an excellent model to study B-Myb transcriptional activity, because of the large stimulation observed (Ͼ30-fold). Hence, this promoter can be used to gain a better understanding of B-Myb function, because one should be able to readily detect small differences in transactivation potential when more than one domain is modified.  (21), SV40 minimal promoter (21), and itself (7). Although we demonstrate that B-Myb can stimulate the FGF-4 promoter in HeLa cells, the mechanism responsible for this effect is only partially understood. We demonstrate that each of the previously identified cis-regulatory elements in the FGF-4 promoter can be inactivated without disrupting the response to B-Myb (Fig. 1). Moreover, only a relatively small region of the promoter, extending just 61 bp upstream of the transcription start site, is required for a strong response to B-Myb. Although sequence analysis of this region does not identify a putative MBS, this small promoter region may contain a novel MBS. Initially, we considered the possibility that the TATA box in our shortest FGF-4 promoter/reporter gene construct (Ϫ61T) might be responsible for the response to B-Myb. Others have shown that B-Myb can stimulate transcription through the TATA box of the HSP70 promoter (66). However, disruption of the TATA box in the FGF-4 promoter did not modify the response to B-Myb. Interestingly, our finding that mut1 does not stimulate the FGF-4 promoter (Fig. 4) raises the possibility that the effect of B-Myb may involve direct binding to DNA. In this regard, mut1 lacks the repeat domains R2 and R3, which are required for DNA binding (6). However, other studies have shown that several proteins, including C/EBP-␤, p100, HSF3, c-Maf, RAR, Cyp40, and D-cyclins, interact with the DNA-binding domain of the Myb proteins to affect transactivation (reviewed in Ref. 26). Hence, B-Myb may exert its effects on the FGF-4 promoter by protein-protein interactions rather than binding directly to DNA.
It is evident from the work reported in this study that several domains of B-Myb are required for its effect on the FGF-4 promoter (Fig. 4B). Transient transfection of the B-Myb deletion mutant expression plasmids show that all but amino acids 560 to 704 (the very carboxyl-terminal tail that is deleted in mutT) are required for full stimulation (Fig. 4B). In addition to mut1, mut2 and mutO exhibit very little activity, which suggests that the deleted domains are required for the effect on the FGF-4 promoter. The lack of response to mut2 is not surprising, because it lacks the acidic transactivation domain of B-Myb (13). MutO, which lacks the region between the acidic transactivation domain and the conserved region, had 10% of activity of wild type. The region deleted in mutO contains NLS1 and a glutamine-rich region. The NLS1 does not appear to function in this system because the mutO is predominantly localized to the nucleus (Fig. 5B). Glutamine-rich regions have been shown to function as transcriptional activators in some systems (69), therefore, this region in B-Myb may have a similar role. MutC, which lacks the carboxyl-terminal tail of B-Myb, including NLS2, exhibited only 40% of wild-type B-Myb activity. The decreased activity of this mutant is likely to be due to reduced nuclear localization (Fig. 5B), but lower protein expression could also be a factor (Fig. 4C). MutN, which has 30% of wild type activity, lacks the amino terminus including the R1 repeat of the DNA-binding domain, which is dispensable for DNA binding (6,70). MutT had activity equal to wild-type B-Myb, but was expressed at a lower level, which may underestimate its effect on the stimulation of the FGF-4 promoter. Our data corresponds well with previous findings that the carboxyl-terminal region is involved in the negative regulation of B-Myb (9). Finally, the ⌬CR mutant of B-Myb had only 40% of the wild-type B-Myb activity. Previous studies have shown that many proteins interact with the conserved domain in B-Myb (10). Therefore, deletion of this domain may result in the disruption of specific protein-protein interactions required for promoter activation.
Phosphorylation Is Essential to B-Myb Activity-Phosphorylation is an important post-translational modification that can regulate DNA binding, transactivation, nuclear localization, and protein-protein interaction. In fact, many cell cycle-regulated proteins, including B-Myb, are regulated by phosphorylation. More specifically, the phosphorylation of B-Myb has been shown to play a role in the transcriptional activity of B-Myb, including its ability to bind to DNA (22,24,25). Our work, and that of others, have previously identified at least 10 phosphorylation sites on murine B-Myb that consist of the consensus sequence Ser(P)/Thr(P)-Pro and are conserved in the Myb family of proteins. These phosphorylation sites are targeted by the proline-directed kinase Cdk2 bound to cyclin E or cyclin A (20,(22)(23)(24)56). In this study, we identified five additional phosphorylation sites, which are localized to the carboxyl-terminal tail of B-Myb. The phosphorylation sites Ser 343 , Ser 424 and Thr 443 , Thr 447 are located in the region between the acidic transactivation domain and the conserved region and Thr 490 is in the conserved region. Our studies demonstrate that the 15mut protein, which lacks all phosphorylation sites, has little activity and functions as a dominant-negative when co-expressed with B-Myb (Fig. 3A). The phosphorylation of B-Myb at the carboxyl terminus may regulate B-Myb function by multiple mechanisms. One proposed mechanism is the autoregulation of B-Myb by the carboxyl-terminal tail folding back and interacting with the amino-terminal portion, blocking DNA binding and/or transactivation (29). A second possible mechanism is the phosphorylation-dependent interaction of proteins with the carboxyl-terminal tail of B-Myb. Importantly, the carboxyl-terminal tail of the Myb proteins has been shown to modulate biological function and interact with a number of proteins. Another potential mechanism is phosphorylation-dependent nuclear localization of B-Myb. Although Ser 581 lies within NLS2 and Ser 424 lies near NLS1, phosphorylation of B-Myb does not appear to be involved in the nuclear localization of B-Myb based on our finding that the 15mut is predominantly nuclear (Fig. 5B). Our findings also suggest that the phosphorylation of B-Myb is not involved in the interaction of B-Myb with p300 based on the co-immunoprecipitation of the 15mut with p300.
Acetylation of B-Myb by p300 -Acetylation is another important post-translational modification that can affect the activity of transcription factors. One example is the acetylation of p53 by p300/CBP and p300/CBP-associated factor (p/CAF) in response to DNA damage, which increases the affinity of p53 for DNA (71). Previous reports demonstrated that c-Myb and A-Myb interact with p300/CBP via the acidic transactivational domain of Myb and the KIX domain of p300/CBP (58,59,72) and that c-Myb is acetylated by p300 (73). More recently, B-Myb has been shown to interact with CBP to synergistically activate transcription (56). Our study demonstrates that B-Myb is acetylated by p300 in vivo. Furthermore, we show that the bromo and HAT domains of p300 are sufficient for interaction with and the acetylation of B-Myb. We also determined that specific domains of B-Myb are necessary for acetylation. Specifically, mut2 was not acetylated by p300, possibly because it lacks the acidic transactivation domain of B-Myb, which is necessary for the interaction of p300 with A-Myb and c-Myb. In addition, mutC does not appear to be acetylated by p300. Since mutC does not localize predominantly to the nucleus, it may have a decreased propensity to interact with p300. Nonetheless, the fact that mutC can stimulate the FGF-4 promoter demonstrates that it is able to enter the nucleus. Hence, we suspect that mutC lacks a domain required for acetylation.
An interesting issue that remains to be resolved is whether acetylation of B-Myb is required for stimulation of the FGF-4 promoter. Although transfection of HeLa cells with an expression vector for Gal4p300-(964 -1922) with a mutant HAT domain can inhibit the response of the FGF-4 promoter to B-Myb, it does not demonstrate that acetylation of B-Myb itself is required. To resolve this issue, the acetylation sites of B-Myb will need to be mapped and their roles in B-Myb transactivation determined. Finally, recent reports indicate that the functions of phosphorylation and acetylation can be interdependent. For example, a recent study observed that acetylation of p53 is regulated by amino-terminal phosphorylation (71,74). However, unlike p53, acetylation of B-Myb does not appear to be phosphorylation dependent, because 15mut is substantially acetylated in the presence of p300 (Fig. 7A).
In conclusion, we provide evidence that B-Myb interacts physically with p300 and is acetylated by the bromo plus HAT domains of p300 in vivo. Also, we provide strong evidence that phosphorylation is key to B-Myb stimulation of the FGF-4 promoter and that a phosphorylation defective mutant functions as a dominant-negative of B-Myb. Finally, we suggest that the FGF-4 promoter can be utilized as a diagnostic tool to assay for B-Myb activity. Further work is needed to decipher how phosphorylation and acetylation affect B-Myb activity.