The c‐MYC oncoprotein regulates various aspects of cell behaviour by modulating gene expression. Here, we report the identification of the cAMP‐response‐element‐binding protein (CBP) as a novel c‐MYC binding partner. The two proteins interact both in vitro and in cells, and CBP binds to the carboxy‐terminal region of c‐MYC. Importantly, CBP, as well as p300, is associated with E‐box‐containing promoter regions of genes that are regulated by c‐MYC. Furthermore, c‐MYC and CBP/p300 function synergistically in the activation of reporter‐gene constructs. Thus, CBP and p300 function as positive cofactors for c‐MYC. In addition, c‐MYC is acetylated in cells. This modification does not require MYC box II, suggesting that it is independent of TRRAP complexes. Instead, CBP acetylates c‐MYC in vitro, and co‐expression of CBP with c‐MYC stimulates in vivo acetylation. Functionally, this results in a decrease in ubiquitination and stabilization of c‐MYC proteins. Thus, CBP and p300 are novel functional binding partners of c‐MYC.
The expression of genes is controlled by DNA‐sequence‐specific transcriptional regulators that recruit cofactors to modulate the activity of the polymerase complex (Narlikar et al., 2002). Cofactors function at least in part by affecting chromatin structure through their associated enzymatic activities, including ATPases/helicases, histone acetyl‐transferases (HATs) and histone deacetylases (HDACs) (Jenuwein & Allis, 2001; Sudarsanam & Winston, 2000). Targets of the latter include histone tails that, depending on their modification status, function at least partly as interaction surfaces for proteins involved in the regulation of chromatin structure and gene transcription (Jenuwein & Allis, 2001).
MYC proteins, including c‐MYC, N‐MYC and L‐MYC, are transcriptional regulators that were first identified as transforming factors. Subsequently, it was shown that MYC stimulates cell proliferation, inhibits differentiation and induces apoptosis in many cell types. Structure–function analyses have revealed several regions that are important for the ability of MYC to control cell behaviour. These include the amino‐terminal transactivation domain (TAD) and the carboxy‐terminal bHLHZip (basic helix–loop–helix zipper) domain, which is responsible for dimerization with its essential partner, MAX, and for sequence‐specific DNA binding, respectively (Lüscher, 2001; Oster et al., 2002).
Little detailed information is available as to how c‐MYC affects the activity of the polymerase II (Pol II) complex (Amati et al., 2001; Lüscher, 2001). Among the many proteins that have been shown to interact with c‐MYC (Oster et al., 2002), TRRAP, a component of HAT‐containing complexes, attracted particular interest (McMahon et al., 1998, 2000; Park et al., 2001). TRRAP binds to the MYC box II (MBII) motif in the N‐terminal TAD of c‐MYC, and is recruited to c‐MYC‐regulated promoters (Bouchard et al., 2001; Frank et al., 2001). The activation of some genes, including cyclin D2, telomerase reverse‐transcriptase and nucleolin, by c‐MYC requires MBII (Bouchard et al., 2001; Frank et al., 2001; Nikiforov et al., 2002). However, several other genes can be activated by c‐MYCΔMBII, that is, in the absence of TRRAP recruitment (Nikiforov et al., 2002). These findings suggest that TRRAP is important for c‐MYC‐dependent regulation of some MYC target genes, whereas others may depend on additional MYC‐associated activities.
c‐MYC‐dependent activation of target genes is associated with an increase in histone H3 and histone H4 acetylation (Bouchard et al., 2001; Frank et al., 2001). At least three distinct HATs, GCN5, PCAF and TIP60, can associate with TRRAP complexes. For two of these, GCN5 and TIP60, an interaction with c‐MYC has been demonstrated (McMahon et al., 2000; B. Amati, personal communication). Thus, recruitment of these HATs to TRRAP‐containing complexes is probably significant for gene regulation by c‐MYC.
In this study, we identified the co‐activator CREB (cAMP response element)‐binding protein (CBP) as a novel c‐MYC interaction partner. CBP and its close relative p300 are important transcriptional mediators for many transcription factors. CBP functions partly through its intrinsic HAT activity, as well as through associated HAT enzymes. In addition, CBP and p300 have been shown to interact with components of the Pol II complex, which may also contribute to co‐activator function. CBP/p300 has several functions in the control of cell behaviour, including cell proliferation (Chan & La Thangue, 2001; Goodman & Smolik, 2000). Our findings show that CBP interacts directly with c‐MYC and stimulates its function. Furthermore, CBP and p300 are recruited to MYC‐regulated genes. Last, CBP acetylates c‐MYC, thereby affecting its ubiquitination and half‐life. Together, these findings identify CBP and p300 as novel functional binding partners of c‐MYC.
Results and Discussion
In vivo and in vitro interaction of c‐MYC with CBP
In our attempt to identify novel c‐MYC interacting proteins, we metabolically labelled human embryonic kidney (HEK) 293 cells with [35 S]methionine, and immunoprecipitated c‐MYC and associated proteins under low stringency conditions. We saw several labelled protein bands that were not immunoprecipitated with control antibodies. One of these bands co‐migrated with CBP/p300, which have been identified previously as binding partners of the adenoviral protein E1A (Fig. 1A). We tested the possibility that c‐MYC interacts with CBP/p300. CBP was detected in the eluate of c‐MYC‐specific but not MAD1‐specific immunoprecipitates (Fig. 1B). Furthermore, CBP binding to c‐MYC was seen in co‐immunoprecipitation and western blot analysis in human Jurkat T cells and in HL60 and U937 myeloid cells. Importantly, CBP was also detected in MAX‐specific immunoprecipitates, showing that the c‐MYC/MAX heterodimeric complex interacts with CBP (Fig. 1C,D; and data not shown).
The interaction of c‐MYC with CBP was further confirmed using a mammalian two‐hybrid system. Transactivation by Gal4–CBP was strongly activated by co‐expressing a c‐MYC–VP16‐TAD fusion protein (Fig. 2A). In addition, glutathione‐S‐transferase (GST) pull‐down experiments showed a direct interaction between c‐MYC and CBP (Fig. 2B,C). In CBP, the regions encompassing amino acids 451–721, which includes the KIX domain, and to a significantly lesser degree, amino acids 1–451, interacted with c‐MYC. These CBP domains have been implicated in binding to several different transcription factors (Chan & La Thangue, 2001). Conversely, CBP binds to the C‐terminal domain of c‐MYC, but not to the TAD (Fig. 2B,C).
CBP stimulates c‐MYC‐dependent transactivation
Next, we tested whether CBP can modulate c‐MYC‐specific transactivation. c‐MYC weakly activated a reporter gene construct containing c‐MYC/MAX binding sites (Fig. 3A). However, this reporter was activated co‐operatively by c‐MYC and CBP, whereas a reporter with no binding sites was not regulated (Fig. 3A). In addition, p300 also co‐operated efficiently with c‐MYC (Fig. 3A), although the interaction of c‐MYC with p300 appeared weaker than with CBP, as inferred from co‐immunoprecipitation experiments (data not shown). The HAT domain of CBP was partly responsible for the co‐operativity with c‐MYC (Fig. 3B). As CBP bound to the C‐terminal half of c‐MYC, we tested whether CBP could stimulate MYC proteins lacking the N‐terminal TAD. MYCΔN147 and MYCΔN177 were unable to activate a reporter construct, but co‐operated efficiently with CBP (Fig. 3C). Thus, the C‐terminal domain of c‐MYC is sufficient to mediate a functional interaction with the co‐activator CBP.
To test further the interaction of c‐MYC with CBP and p300, we performed chromatin immunoprecipitations (ChIPs) and analysed the cyclin D2 promoter. We have previously demonstrated binding of c‐MYC/MAX and MAD1/MAX in a differentiation‐specific pattern, correlating with high and low levels of histone acetylation and promoter activity, respectively (Bouchard et al., 2001). This is consistent with the interpretation that c‐MYC/MAX and MAD1/MAX complexes recruit HAT and HDAC activities, respectively, to the promoter. CBP bound to the cyclin D2 promoter in exponentially growing HL60 cells, but not in 12‐O‐tetra‐decanoyl‐phorbol 13‐acetate (TPA)‐treated cells (Fig. 4A). This correlated with a reduction in c‐MYC binding, whereas MAD1 binding was induced. Low and constant signals were obtained with control antibodies against cytochrome c (Fig. 4A). In addition, CBP was localized to the promoter of the c‐MYC‐regulated ODC gene in HL60 cells (data not shown). Furthermore, p300 was crosslinked to the cyclin D2 promoter, but not to its 3′ untranslated region (UTR) in U937 cells (Fig. 4B). Proliferation of P493‐6 B cells is dependent on exogenous c‐MYC that is expressed from a tetracycline‐regulated transgene (Schuhmacher et al., 1999). Induction of c‐MYC results in an increase in its recruitment to the eIF2‐α and ODC promoters (Fig. 4C). This is accompanied by an increase in both CBP and p300 binding. Together, our data show that there is recruitment of CBP/p300 to target genes by c‐MYC/MAX complexes.
Acetylation of c‐MYC by CBP
CBP/p300 acetylate not only all four core histones, but also several transcriptional regulators, cofactors and general transcription factors (Chan & La Thangue, 2001; Sterner & Berger, 2000). Therefore, we tested whether c‐MYC or MAX are substrates for CBP/p300. Recombinant CBP acetylated a maltose‐binding protein (MBP)– c‐MYC fusion protein, but did not acetylate either MBP or GST–MAX (Fig. 5A,B; and data not shown). Acetylation was detected by the incorporation of radioactive acetate or by immunoblotting using antibodies specific for acetylated lysine residues (anti‐Ac–Lys). To determine whether acetylation also occurs in cells, FLAG‐tagged MYC proteins were immunoprecipitated from transiently transfected COS7 cells, and the proteins were detected with anti‐Ac–Lys (Fig. 5C). Acetylation of c‐MYC, MYCΔMBI and MYCΔMBII was detected. This suggests that HAT activities associated with TRRAP are not responsible for MYC acetylation. Additional evidence for a role of CBP/p300 in c‐MYC acetylation was obtained from co‐expression experiments. Acetylation of c‐MYC was increased when c‐MYC was co‐expressed with CBP, as compared with the control or with co‐expressed CBPΔHAT (Fig. 5D). These results show that CBP is responsible for c‐MYC acetylation in cells.
To address the functional relevance of c‐MYC acetylation, we tested whether DNA binding is affected, as acetylation of several transcription factors, including c‐MYB, GATA1 and E2F, has been shown to stimulate DNA binding (Chan & La Thangue, 2001). However, no effect was detectable in electrophoretic gel mobility‐shift assays (data not shown). Another possibility is that c‐MYC acetylation affects protein half‐life. c‐MYC is degraded in the ubiquitin–proteasome pathway (Lüscher, 2001). As ubiquitination and acetylation both occur at Lys residues, competition between the two types of modification is possible. Indeed, coexpression of c‐MYC with CBP or p300, but not with CBPΔHAT, reduced the ubiquitination and turnover of c‐MYC (Fig. 5E,F). Similarly, other transcription factors, including E2F1 and SMAD7, were shown to be stabilized on acetylation by CBP/p300 (Gronroos et al., 2002; Martinez‐Balbas et al., 2000).
Functional relevance of CBP binding to c‐MYC
The finding that CBP binds to the C‐terminal domain of c‐MYC was unanticipated. Previous mapping using Gal4 fusion proteins identified a TAD within the N‐terminal one‐third of c‐MYC, whereas the C‐terminal two‐thirds of the protein showed no transactivating activity (Kato et al., 1990). However, several findings have indicated that the C‐terminal region, in addition to its crucial function in interacting with MAX and binding to DNA, has other functions relevant to the regulation of gene transcription. The recruitment of the SWI/SNF complex has been suggested to stimulate c‐MYC transcriptional activity (Cheng et al., 1999). The binding of other factors, including YY1, may interfere negatively with c‐MYC transcriptional activity (Austen et al., 1998; Shrivastava et al., 1993). MIZ1 and SMAD2/SMAD3 are other factors that bind to the C terminus of MYC. Although the way in which these proteins affect MYC function has not been determined, both are repressed by c‐MYC, suggesting that when MYC is bound to MIZ1 or SMAD2/SMAD3, the MYC TAD is not active (Feng et al., 2002; Staller et al., 2001). These findings suggest that both positive and negative functions are associated with the C‐terminal domain of c‐MYC. This may explain why, at first, no transactivating activity was found to be associated with the C‐terminal half of c‐MYC (Kato et al., 1990). Thus, the binding of CBP/p300 to the c‐MYC C terminus provides an additional link from this region of c‐MYC to the control of gene transcription.
In addition to GCN5 and TIP60, which preferentially acetylate core histones, we provide evidence for a third HAT activity that is recruited by c‐MYC to promoters. Although CBP and p300 can modify core histones in vitro, it is not known whether this occurs in cells (Chan & La Thangue, 2001; Goodman & Smolik, 2000). Indeed, it has been argued that the CBP/p300 HAT activity is not involved in regulating chromatin by modifying histones (Agalioti et al., 2002; Li et al., 1999). Importantly, CBP/p300 acetylate other substrates, including transcription factors and components of the Pol II complex (Chan & La Thangue, 2001; Goodman & Smolik, 2000). Thus, whereas the primary targets of GCN5 and TIP60 are probably histones associated with c‐MYC‐regulated genes, our findings demonstrate that CBP/p300 affect c‐MYC itself. Acetylation seems to be linked to ubiquitination, and may, in addition to altering c‐MYC stability, influence its interaction with other factors, and thus modulate transcription. Furthermore, CBP/p300 may also function by contacting basal transcription factors, and thus perform a bridging function (Chan & La Thangue, 2001). In summary, the functions of CBP/p300 as c‐MYC cofactors seem to be complex, and involve direct modification of c‐MYC and, potentially, communication with the Pol II complex and/or chromatin remodelling.
Cell culture and transient transfection assays.
Culture and differentiation of cells and transient transfections have been described previously (Bouchard et al., 2001; Lüscher‐Firzlaff et al., 1999). P493‐6 cells (provided by D. Eick) were grown for 3 days in the presence of tetracycline to arrest the cells in G0/G1 phase. To induce c‐MYC expression, tetracycline was washed out and the cells were cultured for 5 h before harvesting.
Plasmids and recombinant proteins.
pVR1012–Gal4, pVR1012– Gal4–CBP and pVR1012–Gal4–p300 were provided by N. Perkins; pSV–FLAG–VP16, pSV–FLAG–c‐MYC–VP16, pMBP–Pre and pMBP– Pre‐MYC were provided by H. Ariga; GST–CBP–HAT and pcDNA3–Gal4–CBPΔHAT were provided by T. Kouzarides; a recombinant baculovirus expressing His–CBP was provided by D. Thanos; pCMV–His6–ubiquitin was provided by D. Bohmann.
GST–CBP fragments (Janknecht & Nordheim, 1996), p(Gal4)4–mintk–luc (a fusion of Gal4, the minimal thymidine‐kinase promoter and luciferase; Lüscher‐Firzlaff et al., 1999), pEQ176P2– c‐MYC, tk–luc, M4–tk–luc (a fusion of four MYC/MAX binding sites (M4) to tk and luc) and the β‐galactosidase expression vector pEQ176 (Austen et al., 1998) have been described previously.
For in vivo interaction assays, whole‐cell lysates were prepared in F buffer (10 mM Tris‐HCl, pH 7.05, 50 mM NaCl, 30 mM Na4 P2 O7, 50 mM NaF, 5 μM ZnCl2, 100 μM Na3 VO4, 1% Triton X‐100, 1 mM phenylmethylsulphonyl fluoride, 5 units ml−1 α2‐macroglobulin, 2.5 units ml−1 pepstatin A, 2.5 units ml−1 leupeptin, 0.15 mM benzamidin). Co‐immunoprecipitated proteins were detected either by western blot analysis or by eluting the proteins in AB buffer (20 mM Tris–HCl, pH 7.5, 50 mM NaCl, 0.5% NP‐40, 0.5% deoxycholate, 0.5% SDS, 0.5% aprotinin, 1 mM EDTA) and performing a second immunoprecipitation (Krippner‐Heidenreich et al., 2001; Lüscher‐Firzlaff et al., 1999).
The following antibodies were used for immunoprecipitations and/or western blot analyses (as indicated in the figure legends): monoclonal antibodies 6A10 and 9E10 (human c‐MYC); monoclonal antibody 5C9 (human MAD1; Sommer et al., 1997); monoclonal antibody M73 (against E1A, provided by P. Whyte); anti‐c‐MYC N262; anti‐MAX C17; anti‐MAD1 C20; anti‐CBP A22; anti‐p300 C20; anti‐cytochrome c SC7159; anti‐Gal4 SC577 (Santa Cruz); anti‐CBP AC238 (NeoMarkers); anti‐acetyl‐lysine (New England Biolabs); and anti‐FLAG M2–agarose‐slurry (Sigma).
Fusion proteins were prepared as described previously (Kitaura et al., 2000; Lüscher‐Firzlaff et al., 1999). For binding reactions, 3 μg of GST fusion protein was bound to glutathione–agarose beads and incubated with [35 S]methionine‐labelled in vitro transcribed and translated proteins in binding buffer (Oelgeschlager et al., 1996).
Purification of His–CBP and acetyltransferase assays.
Purification of His–CBP was performed as described previously (Chen et al., 2001). Purified MBP or MBP–MYC (3 μg) was incubated in 30 ml of HAT buffer (50 mM Tris‐HCl, pH 8.0, 2 mM EDTA, 1 mM dithiothreitol, 1 mM PMSF, 10 mM sodium butyrate) with 50 ng of purified His–CBP and 1 μl [14 C]acetyl‐coenzyme A (55 mCi mol−1; NEN) for 1 h at 30 °C. Alternatively, acetylation was detected by immunoblotting with an acetyl‐lysine‐specific antibody.
Chromatin crosslinking and immunoprecipitation (ChIP).
ChIP assays were performed as described previously (Bouchard et al., 2001; Strutt & Paro, 1999). The following PCR primers were used: hCycD23‐UTRf, 5′‐ATCAGACCCTATTCTCGGCTCA GG‐3′ hCycD2‐3‐UTRr, 5′‐CAGTCAGTAAGGCACTTTATTTCCCC‐3′ hCycD2prom1, 5′‐CCCCTTCCTCCTGGAGTGAAATAC‐3′ hCycD2prom2, 5′‐CGTGCTCTAACGCATCCTTGAGTC‐3′ eIF2a‐h1, 5′‐TTCTCGGAGGACCCAGACTCTATG‐3′ eIF2a‐h2, 5′‐TCA CAGAGACCAGACTTGCTTCCC; hODCprom‐h1, 5′‐GAGCAGAGC GCACCGGGATCA; and hODCprom‐h2, 5′‐CAGTACCTCGTGCC CGAGAGC.
Ubiquitination assays were performed as described in Campanero & Flemington (1997). Transiently transfected HEK293 cells were treated with 25 mM MG132 for 2 h before lysis in 8 M urea, 0.1 M sodium phosphate, pH 8.0, 10 mM imidazole. His‐tagged proteins were bound to Talon beads and eluted in SDS sample buffer containing 200 mM imidazole. Modified c‐MYC proteins were detected using monoclonal antibody 9E10.
We thank H. Ariga, D. Bohmann, D. Eick, T. Kouzarides, N. Perkins and P. Whyte for reagents, B. Amati for communicating unpublished results, and H. Burkhardt for excellent technical assistance. Discussions with L.‐G. Larsson are particularly appreciated. This work was supported by grants from the Deutsche Forschungsgemeinschaft and from the Fonds der Chemischen Industrie to B.L.
- Copyright © 2003 European Molecular Biology Organisation