View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Cell, Vol. 96, 405–413, February 5, 1999, Copyright 1999 by Cell Press Regulation of Acetyltransferases p300 and PCAF by the bHLH Protein Twist and Adenoviral Oncoprotein E1A

Yasuo Hamamori,* Vittorio Sartorelli,* of Twist have been considerably delineated in myogen- Vasily Ogryzko,† Pier Lorenzo Puri,‡§ esis, it remains to be clarified how Twist can regulate cell Hung-Yi Wu,* Jean Y. J. Wang,‡ differentiation in such diverse lineages. The profound † Yoshihiro Nakatani, and Larry Kedes*k effects that Twist exerts on transcription and cellular *Institute for Genetic Medicine differentiation might result from the inhibition of crucial Department of Biochemistry and Molecular Biology molecules located at the crossroad of multiple cellular University of Southern California School of Medicine nodal points. Los Angeles, California 90033 Some of Twist’s properties are reminiscent of E1A † Laboratory of Molecular Growth Regulation behavior. In fact, E1A precludes differentiation of several National Institute of Child Health cell lineages, including muscle (Webster et al., 1988; and Human Development Braun et al., 1992; Caruso et al., 1993), neuronal cells Bethesda, Maryland 20892 (Maruyama et al., 1987; Boulukos and Ziff, 1993), and ‡ Department of Biology thyroid cells (Berlingieri et al., 1993). Whilst E1A affects Center for Molecular Genetics transcription in disparate ways, a recurrent finding is a University of California, San Diego net inhibitory effect (reviewed in Flint and Shenk, 1997). La Jolla, California 92093-0347 Several cellular proteins interact with and are inhibited § Laboratory of Expression by E1A. These include Rb, p107, p130, and p300/CBP Fondazione Andrea Cesalpino (reviewed in Moran, 1993; Flint and Shenk, 1997; Shi- Instituto I Clinica Medica kama et al., 1997). The two paralogous transcriptional Policlinico Umberto I coactivators p300 and CBP (CREB-binding protein) University of Rome, La Sapienza were originally identified as an E1A-binding protein and 00161 Rome a CREB-binding protein, respectively (Shikama et al., Italy 1997). p300 and CBP have common domains and func- tions and are collectively described here as p300/CBP unless otherwise specified. p300/CBP is ubiquitously expressed and regulates a broad spectrum of biological Summary activities such as cell differentiation, cell cycle control, apoptosis, and tumorigenesis. p300/CBP is recruited to Histone acetyltransferases (HAT) play a critical role in specific promoters through interactions with sequence- transcriptional control by relieving repressive effects specific activators and in turn recruits a second coacti- of chromatin, and yet how HATs themselves are regu- vator such as PCAF (p300/CBP–associated protein) lated remains largely unknown. Here, it is shown that (Yang et al., 1996). Among these activators are MyoD Twist directly binds two independent HAT domains (Eckner et al., 1996; Yuan et al., 1996; Puri et al., 1997a; of acetyltransferases, p300 and p300/CBP–associated Sartorelli et al., 1997), NF-␬B (Perkins et al., 1997), factor (PCAF), and directly regulates their HAT activi- GATA-1 (Boyes et al., 1998), and (Avantaggiati et ties. The N terminus of Twist is a primary domain inter- al., 1997; Gu and Roeder, 1997; Lill et al., 1997). These acting with both acetyltransferases, and the same do- molecules are critical for the execution of certain funda- main is required for inhibition of p300-dependent mental biological functions such as cell differentiation, transcription by Twist. Adenovirus E1A protein mimics immunological responses, and apoptosis. the effects of Twist by inhibiting the HAT activities of In eukaryotic cells, DNA is tightly packed in repressive p300 and PCAF. These findings establish a cogent chromatin that hampers the access of transcriptional argument for considering the HAT domains as a direct machinery. One mechanism to overcome such repres- target for acetyltransferase regulation by both a cellu- sion is believed to involve histone acetylation (reviewed lar transcription factor and a viral oncoprotein. in Grunstein, 1997; Wade and Wolffe, 1997; Struhl, 1998). The surprising findings that transcriptional coactivators Introduction p300/CBP and PCAF possess histone acetyltransferase (HAT) activities have begun to shed light on promoter- Twist is a basic helix-loop-helix (bHLH) protein ex- targeted derepression mechanisms (Bannister and Kouz- pressed specifically in mesodermal and cranial neural arides, 1996; Ogryzko et al., 1996; Yang et al., 1996). In crest cells during embryogenesis (Thisse et al., 1988; contrast to the accumulating data on functions of the Wolf et al., 1991; Fuchtbauer, 1995; Stoetzel et al., 1995) coactivators, however, little is known about the cellular whose activity has been implicated in the inhibition of regulation of p300/CBP itself. A lone insight is that viral differentiation of multiple cell lineages, including muscle oncoproteins E1A and SV40 T antigen exert trans- (Hebrok et al., 1994; Spicer et al., 1996; Hamamori et forming activities, at least in part by perturbing p300/ al., 1997), bone (Murray et al., 1992), and neuronal cells CBP functions (Yang et al., 1996; Puri et al., 1997b; (Lee et al., 1995). Although the inhibitory mechanisms reviewed in Shikama et al., 1997). Here, we demonstrate that Twist interacts with the HAT domains and inhibits acetyltransferase activities of p300 and PCAF. In addi- tion, we show that viral oncoprotein E1A mimics the k To whom correspondence should be addressed (e-mail: kedes@ hsc.usc.edu). effects of Twist in inhibiting the HAT activities. These Cell 406

findings provide evidence for a novel and unifying mech- anism to explain the effects of viral and cellular proteins on fundamental aspects of cell function through the regulation of acetyltransferases.

Results

Twist Interacts with p300 and PCAF In Vitro and In Vivo On the basis of functional similarities between E1A and Twist, we first tested whether Twist interacted with p300 in vitro and in vivo. Total cell extracts from C2C12 myo- tubes were incubated with equal amounts of either GST- Twist, a control GST-E1A, or GST alone. GST-Twist and GST-E1A, but not GST alone, interacted with cellular p300 (Figure 1A). To further determine whether this interaction occurs in vivo, we took a cotransfection approach, given a lack of appropriate Twist protein- expressing cell lines. COS7 cells were transiently trans- fected either with a vector for -tagged Twist (myc- Twist) alone (Figure 1B, lane 1) or a combination of the two vectors for myc-Twist and Flag-tagged p300 (lanes 2 and 3). Extracts from the transfected cells were first immunoprecipitated with either anti-Flag M2 antibody (lanes 1 and 2) or an unrelated HA antibody (lane 3). The associated proteins were subjected to Western blot analysis with anti-myc antibody (9E10.2). An intense band corresponding to myc-Twist was readily detected only when Flag-p300 was expressed and immunopre- cipitated by the Flag antibody (lane 2), indicating that Twist specifically interacted with p300 in vivo. Since PCAF colocalizes with p300 to form a multiprotein com- plex (Yang et al., 1996; Puri et al., 1997b), we asked whether PCAF might also be a target of Twist. Using essentially the same strategy, we found that GST-Twist specifically interacted with cellular PCAF, as detected by Western blot analysis using an anti-PCAF antibody (Figure 1C). The interaction was further confirmed in vivo by transiently expressing Flag-PCAF and myc- Twist. Myc-Twist was detected only when PCAF was coexpressed and immunoprecipitated (Figure 1D, lane 1). These findings demonstrate that Twist can interact either directly or indirectly with the two histone acetyl- transferases p300 and PCAF both in vitro and in vivo.

Twist Interacts Directly with the Two Independent HAT Domains through Its N Terminus A number of transcriptional activators, coactivators, and Figure 1. Twist Interacts with Endogenous and Recombinant p300 viral proteins have been shown to have functional inter- as well as PCAF In Vitro and In Vivo actions with distinct domains of p300/CBP and PCAF (A) Total cell extracts from C2C12 muscle cells were incubated with (summarized in Figure 2). As a first step toward under- GST-Twist bound to glutathione-agarose beads. Bound proteins standing the functional significance of the interactions were analyzed by Western blot using an anti-p300 antibody. (*) An ,.kDa is frequently observed (Lill et al 80ف unidentified protein of between Twist and p300 as well as PCAF, we wished 1997; Dallas et al., 1998). to delineate the domains engaged in their association. (B) U2OS cells were cotransfected with vectors expressing Flag- GST pull-down studies were performed using GST-Twist tagged p300 and myc-epitope-tagged Twist. Immunoprecipitation and a series of in vitro–translated [35S]p300 proteins (Fig- was performed on cell extracts with either the Flag- or control HA- ure 3). Twist strongly bound the C-terminal fragment antibody, followed by Western blot analysis to detect myc-Twist. (amino acids 1257–2414) of p300 spanning the HAT do- (C) Interaction between cellular PCAF and immobilized GST-Twist was detected by Western blot analysis with anti-PCAF antibody. main as well as the CH3 domain. Further deletion re- (D) In vivo interaction between transfected Twist and PCAF was vealed that this interaction requires the CH3 domain detected by Western blot analysis using a myc antibody. (compare 1572–2414 and 1869–2414), which is known to interact with other proteins. An isolated CH3 fragment Twist Inhibits Histone Acetyltransferases 407

Figure 2. Schematic Representations of p300, PCAF, and Twist and Their Interaction Do- mains Proteins known to interact with p300, PCAF, and Twist are indicated above each inter- acting domain. The bold lines indicate the interacting domains determined in the pres- ent study. CH3, cysteine-histidine-rich region 3; Bromo, ; HAT, histone ace- tyltransferase.

alone also supported the interaction with GST-Twist (not common motifs present in these HAT domains (Neuwald shown). Of particular interest, Twist retained an interac- and Landsman, 1997; Martinez-Balbas et al., 1998). The tion with a HAT domain even in the absence of the CH3 functional significance of Twist’s interaction with the domain (1257–1572) (Figure 3). Twist also bound the PCAF bromodomain remains to be studied. To deter- N terminus of p300 (1–566 and 1–744), although these mine which domains of Twist interact with p300 and interactions were 5- to 10-fold weaker than those with PCAF, equal amounts of various Twist protein fragments the CH3 and HAT domains. fused to GST were incubated with in vitro–translated Incubation of GST-Twist with affinity-purified PCAF p300 and PCAF (Figure 5). We found that the N terminus showed strong binding (Ͼ20% of the input) (Figure 4, of Twist is a primary interacting site for both p300 and top). Intriguingly, experiments using a series of PCAF PCAF, while a bHLH domain may also contribute to both internal deletion mutants revealed that this interaction interactions, albeit to a much lesser degree (Figure 5). required the presence of the intact HAT domain and This provides us with a new insight as to roles of the N bromodomain (Figure 4). Thus, Twist interacts indepen- terminus of Twist, since we have previously found that dently with the HAT domains of two different proteins, this domain is necessary for efficient inhibition of MyoD p300 and PCAF, suggesting that Twist may recognize transcription by Twist, and yet in contrast to the other

Figure 3. Twist Interacts with the HAT and CH3 Domains of p300 In vitro–translated [35S]p300 fragments were incubated with GST-Twist immobilized on glutathione-agarose beads. The bound proteins were analyzed by autoradiography after SDS-PAGE (top). Summary of p300 domain mapping for the interaction with Twist (bottom). The bound p300 fragments were quantified by a phosphoimager using the fragment 1257–1572 as a standard (100%). ϩϩϩ, 81%–100% of interaction; ϩϩ, 61%–80%; ϩ, 41%–60%; Ϯ, 21%–40%; Ϫ, 0%–20%. Cell 408

other preparations of Twist, including GST-Twist and histidine-tagged Twist, but not by GST-MyoD or histi- dine-MyoD (not shown). The N terminus of Twist, which provides a primary binding site for both p300 and PCAF (Figure 5), was required for the inhibition of acetylation, since deletion of the N terminus (⌬N Twist) drastically diminished the inhibitory effects (Figure 6A). In addition to , p300 and PCAF are known to acetylate and modify the functional properties of several nonhistone substrates (Gu and Roeder, 1997; Imhof et al., 1997; Boyes et al., 1998; Waltzer and Bienz, 1998). We noted that p300 auto- (or cross-) acetylated itself (Figure 6A, middle left). Although the functional significance of this autoacetylation remains to be studied, it was inhibited by Twist in a manner similar to the acetylation of his- tones, and thus autoacetylation provides an excellent measure to monitor regulation of acetyltransferase ac- tivities by Twist. Because of a number of common features shared by Figure 4. Twist Interacts with the C-Terminal HAT- and Bromodo- E1A and Twist as well as the association of E1A with mains of PCAF p300, we hypothesized that E1A might also mimic the The affinity-purified Flag-PCAF deletion mutant proteins (1 pmol) inhibitory effects of Twist on histone acetyltransferases. were incubated with immobilized GST-Twist followed by Western blot analysis with anti-Flag antibody. Indeed, E1A efficiently inhibited the HAT activities of both p300 and PCAF (Figure 6B). However, the E1A inhibition was more substrate specific. In fact, whereas domains of Twist, a precise role of the N terminus has Twist caused a general reduction in histone acetylation, been undefined (see Figure 2). E1A selectively inhibited acetylation of H3 and H4. Unlike p300, PCAF is known to preferentially acetylate H3 and Twist and E1A Inhibit HAT Activities of p300 H4 among the four core histones (Yang et al., 1996), and and PCAF In Vitro E1A also inhibited the acetylation of both H3 and H4 by The finding that Twist directly binds the HAT domains PCAF. These findings illustrate how E1A mimics Twist by of both p300 and PCAF (Figures 3 and 4) prompted us inhibiting the HAT activities of p300 and PCAF, although to test the hypothesis that Twist might affect their HAT E1A appears to have additional mechanisms enabling activities through these direct interactions with the HAT E1A to distinguish different substrates. domains. Affinity-purified p300 and PCAF were able to acetylate histones in the presence of [14C]acetyl-CoA Twist and E1A Inhibit HAT Activities In Vivo (Figure 6A). However, such acetylation was completely To study whether Twist and E1A inhibit the acetyltrans- abrogated upon an addition of Twist. The inhibition was ferases in cells, we transiently transfected cells with specific to Twist, since it was also observed using a few expression vectors for Flag-p300 and PCAF together

Figure 5. The N Terminus of Twist Interacts with Both p300 and PCAF Twist fragments fused to GST were incubated with in vitro–translated [35S]p300 or [35S]PCAF. The bound proteins were analyzed by autora- diography after SDS-PAGE. The bound p300 or PCAF was quantified using a phospho- imager and the results shown using GST- Twist1-206 as a standard (100%) (bottom). ϩϩϩ, 81%–100% of interaction; ϩϩ, 61%– 80%; ϩ, 41%–60%; Ϯ, 21%–40%; Ϫ, 0%– 20%. N/D, not determined. Twist Inhibits Histone Acetyltransferases 409

with vectors for Twist or E1A and monitored the acetyla- tion of p300. When either p300 or PCAF alone was ex- pressed, only a slight p300 autoacetylation was ob- served (not shown). However, when p300 and PCAF were coexpressed, we observed a robust p300 acetyla- tion that allowed us to study possible inhibitory effects of Twist and E1A on acetyltransferase activities (Figure 6C). In line with the in vitro HAT inhibition (Figures 6A and 6B), this in vivo acetylation was completely inhibited in the presence of either Twist or E1A (Figure 6C). Each immunoprecipitate contained an equal amount of Flag- p300 protein as determined by Western blot analysis using an aliquot of the precipitates (Figure 6C, bottom). These results, together with in vitro inhibition of HATs by Twist and E1A supported by the direct interactions, demonstrate that both Twist and E1A target the acetyl- transferase activities in the cells.

Twist Requires the Same N-Terminal Domain for Inhibition of Both HAT Activities and p300-Dependent Transcription To further analyze functional implications of the interac- tions between Twist and the HAT coactivators, we stud- ied the effects of Twist on different types of transcrip- tional activators. We used plasmids encoding the Gal4 DNA-binding domain fused to various activators (Figure 7A). Some of these activators were selected because they are inhibited by E1A through mechanisms thought to involve p300. Both Twist and E1A inhibited all tested p300-dependent acidic activators (MyoD, VP16, p53) but not the glutamine-rich activation domain of Sp1, suggesting that Twist specifically targets p300-depen- dent transcription. The MyoD N-terminal activation do- main (amino acids 1–54) contains the FYD motif that is conserved among the MyoD family and is required for both physical interaction with and coactivation by p300 (Sartorelli et al., 1997). Although Twist did not directly bind this activation domain of MyoD (Hamamori et al., 1997), Twist efficiently inhibited transcription from Gal-

MyoD1–54, and this inhibition was completely rescued by p300 and PCAF (Figure 7B). Notably, this inhibition required the N terminus of Twist, a domain required for the physical interaction with p300 and PCAF as well as for the inhibition of their HAT activities (Figure 7C). In these experiments, the truncated Twist proteins were Figure 6. The HAT Activities of Both p300 and PCAF Are Inhibited equally well expressed in the transfected cells (Hama- by Twist and E1A In Vitro and In Vivo mori et al., 1997). Taken together, these findings support (A) In vitro inhibition of HATs by Twist on two different substrates. the view that Twist suppresses the functions Purified p300 or PCAF was preincubated with Twist before adding of p300 and PCAF through physical interactions medi- purified histones and [14C]acetyl-CoA. Acetylated histones were de- ated by the N terminus of Twist. Since the same N termi- tected by autoradiography after SDS-PAGE. p300 autoacetylation nus of Twist is also required for the HAT inhibition, and was similarly inhibited by Twist (middle left). The relative amounts of acetylated histones were quantitated using a phosphoimager and transiently transfected plasmids are rapidly chroma- expressed as percentage HAT inhibition (bottom graph). tinized (Reeves et al., 1985; Pullner et al., 1996; Stanfield- (B) E1A abolishes the HAT activities of p300 and PCAF. Assays Oakley and Griffith, 1996), the inhibition of HAT activities similar to those described in (A) were performed using purified E1A. may be one of the mechanisms for the observed inhibi- The data shown is a composite of two different exposure periods tion of p300-dependent transcription by Twist. (3 days for p300 and 7 days for PCAF) from the same film. Acetylated histone isoforms (H3, H2A/H2B, and H4) were individually quanti- tated by a phosphoimager (bottom graph). (C) Both Twist and E1A abolish the HAT activities in vivo. C2C12 cells were transfected with the indicated vectors and pulsed with radiography (top). Ten percent of each immunoprecipitate was sub- [3H]sodium acetate. Flag-p300 was immunoprecipitated with the jected to Western blot analysis to ascertain whether an equal Flag-antibody, and [3H]acetylated Flag-p300 was detected by auto- amount of p300 was analyzed (bottom). Cell 410

Discussion

Histone Acetyltransferase Domains as a Direct Target of Twist Twist employs several specific mechanisms to regulate cellular differentiation. For instance, Twist inhibits mus- cle differentiation by E protein sequestration and by interactions with MyoD and MEF2 proteins (Spicer et al., 1996; Hamamori et al., 1997). However, Twist also inhibits transcription from the specific activation do- mains independent of the above mechanisms (Figure 7A), and therefore a new mechanism must be involved. We provide here a unifying explanation of the more gen- eral and profound effects exerted by Twist on multiple and critical cellular functions such as cell differentiation and proliferation. Here, we demonstrate that the cellular transcription factor Twist directly binds p300 and PCAF acetyltransferases using its N terminus, whose function has been so far undetermined. The binding of Twist to the coactivators is directed to the HAT domains of both p300 and PCAF and results in the inhibition of their acetyltransferase activities. These HATs belong to a large N-acetyltransferase superfamily whose members share several conserved motifs (Neuwald and Lands- man, 1997; Martinez-Balbas et al., 1998). An amino acid sequence alignment of N-acetyltransferases from nu- merous species has revealed four conserved motifs (A-D) within the acetyltransferase domain. Motifs A and B have been thought to be the binding site of acetyl- CoA by several mutagenesis experiments (Tercero et al., 1992; Coleman et al., 1996; Lu et al., 1996). In addition to motifs A and B, p300/CBP and PCAF have another motif, E, that appears to be more unique to them (Marti- nez-Balbas et al., 1998). The finding that Twist binds to the two HAT domains that belong to this superfamily and modifies their enzymatic activities suggests a new general scheme for acetyltransferase regulation by the cellular transcription factor Twist. One attractive model for the HAT inhibition by Twist is to prevent acetyl-CoA binding to motifs A and B. Twist might regulate HAT activities of not only p300 and PCAF but also other members of the N-acetyltransferase family by inter- acting with common motifs such as A and B. Alterna- tively, Twist might specifically regulate p300 and PCAF by binding to the more unique E motif. The N terminus of Twist is the primary binding site for both p300 and PCAF (Figure 5), and yet ⌬N Twist still shows weak HAT inhibition in vitro (Figure 6A). One pos- sible explanation for this is that the N terminus of Twist may not be the only domain responsible for the HAT inhibition, and ⌬N Twist (a combination of the bHLH domain and C terminus) might also carry an inhibitory domain. In line with this interpretation, ⌬N Twist that contains an intact bHLH domain retains weak but signifi- Figure 7. Twist Specifically Inhibits p300- and PCAF-Dependent cant binding activities to both p300 and PCAF (Figure Transcription, and the Inhibition Requires the N Terminus of Twist (A) Both Twist and E1A inhibit activities of several p300-dependent transcription factors but not that of the glutamine-rich activator Sp1. U2OS cells were transfected with G4M2-luciferase (1 ␮g) and (C) The N terminus of Twist is required for transcriptional inhibition. expression vectors encoding a Gal4 DNA-binding domain fused to Cells were transfected with Gal-MyoD1–54 and Twist deletion mutant the indicated activators (1 ␮g) and either Twist or E1A. vectors (top). A summary of transcriptional inhibition by Twist on

(B) Rescue by p300 and PCAF of the Twist inhibition of transcription. Gal-MyoD1–54 is shown at the bottom. Wild-type (wt) Twist is set as

Cells were cotransfected with the Gal-MyoD1–54, Twist, p300, or a standard at 100% inhibition. ϩϩϩ, 81%–100% of inhibition; ϩϩ, PCAF expression vectors as indicated. 61%–80%; ϩ, 41%–60%; Ϯ, 21%–40%; Ϫ, 0%–20%. Twist Inhibits Histone Acetyltransferases 411

5). In the in vitro studies where relatively high concentra- Analogy between Twist and E1A tions of Twist can be used, such weak binding might Throughout this paper, we have provided an intriguing be sufficient to observe marginal HAT inhibition (Figure analogy between the functions of Twist and E1A: both 6A). In contrast, the high affinity binding provided by Twist and E1A inhibit MyoD transcription and myogen- the N terminus may be critical for efficient inhibition in esis (Webster et al., 1988; Braun et al., 1992; Caruso et vivo (Figure 7C) where the concentration of Twist is al., 1993; Hebrok et al., 1994; Spicer et al., 1996; Hama- unlikely to reach the level achieved in the in vitro studies. mori et al., 1997), bind to p300 and PCAF (Reid et al., It will be necessary to determine specific interactions 1998; this paper), inhibit the HAT activities of these co- between the motifs of HAT domains and Twist domains activators (this paper), and can antagonize p53-induced in order to better understand the detailed mechanisms cell cycle arrest and apoptosis (Lill et al., 1997; Maestro of how Twist modulates HAT activities. A similar regula- et al., submitted). Furthermore, we have more recently tory mechanism might also be shared by other Twist- noted that Twist expression is upregulated in human related bHLH proteins such as paraxis and dermo-1 that rhabdomyosarcoma tissues, making it likely that Twist have extensive sequence similarity with Twist (Querter- is an oncoprotein (Maestro et al., submitted). The avail- mous et al., 1994; Burgess et al., 1995; Cserjesi et al., ability of such Twist-expressing cells would allow us 1995; Li et al., 1995). to analyze physical interactions between endogenous The molecular mechanism by which E1A regulates Twist and acetyltransferases as well as a role of such HAT activity remains to be elucidated. E1A, by binding interactions in tumorigenesis. Given the common fea- to the CH3 domain, might modify activities of the adja- tures between E1A and Twist, many of the reported cent HAT domain through conformational changes. Al- biological activities of E1A might in fact reflect some of ternatively, E1A might interact directly with the HAT do- the physiological functions exerted by a cellular tran- mains of p300 and PCAF as does Twist. scriptional factor, Twist. E1A has recently been shown to activate the HAT activity of CBP in G1/S phase transition (Ait-Si-Ali et A Dual Mechanism for Coactivator Regulation al., 1998). This is in contrast to our observation of the E1A has been shown to bind the CH3 domain of p300/ inhibition of HAT activities of p300 by E1A and to other CBP and to displace PCAF from this domain (Yang et al., reports describing repression of p300 by E1A (Arany et 1996). Previously, we and others interpreted the effect of al., 1995; Flint and Shenk, 1997; reviewed in Shikama et E1A as a simple competition between E1A and PCAF al., 1997). Although several differences in experimental for the CH3 domain. Our present study adds a further conditions are readily noted, the exact reasons for the level of complexity by demonstrating that E1A and Twist apparent discrepancy remain to be clarified. One possi- may exert their inhibition not only by physically dis- ble explanation is that E1A differentially regulates p300 rupting the p300-PCAF complex formation but also and CBP. In fact, several contrasting properties of p300 through suppression of their enzymatic activities. The and CBP have been recently described, including their interaction of Twist at the CH3 domain raises the in- effects on F9 cell differentiation and expression of cell triguing possibility that Twist might also prevent PCAF cycle inhibitors p21 and p27 (Kawasaki et al., 1998). association with p300/CBP by competing with PCAF for Furthermore, p300 and CBP may be differentially phos- the common CH3 domain. These two mechanisms may phorylated (Ait-Si-Ali et al., 1998; P. L. Puri et al., unpub- not necessarily work simultaneously, and cells would lished results), which may be directly relevant to the have exquisite control mechanisms that determine how observed discrepancy, since phosphorylation of CBP these two mechanisms of p300 and PCAF regulation has been proposed to increase its HAT activity and E1A may be differentially utilized in a given situation. We mimics the effect of phosphorylation of CBP to enhance have previously demonstrated that individual histone its HAT activity (Ait-Si-Ali et al., 1998). Thus, the HAT acetyltransferases have distinct roles (Puri et al., 1997b). domains of p300 and CBP might be differentially regu- For instance, myogenic transcription and differentiation lated by E1A. are dependent on the HAT activity of PCAF but not on that of p300/CBP. Similar observations are made in Genetic Interaction between Twist and p300 other systems, indicating that the transcriptional activi- Mouse knockout models for Twist and p300 show ties of the HAT domains of p300 and PCAF are highly both similar and dissimilar phenotypes (Chen and Beh- promoter dependent (Korzus et al., 1998). The dual inhib- ringer, 1995; Yao et al., 1998). Although mice homozy- itory mechanisms involving the HAT inhibition as well as gous for either p300 or Twist gene disruption show open the competitive displacement of cofactors would allow neural tube defects, heterozygous mice show different E1A and possibly Twist to regulate a broad range of phenotypes: p300ϩ/Ϫ mice show exencephaly, while transcriptional activators that are differentially depen- twistϩ/Ϫ mice have craniosynostosis. Although our bio- dent on p300 and PCAF and their HAT activities. chemical data do not provide immediate explanations for these phenotypes, these phenotypes indicate a ge- Experimental Procedures netic link between Twist and p300. The expression of Twist and p300 in multiple overlapping regions of mouse Plasmids and Transient Transfections embryos (Thisse et al., 1988; Wolf et al., 1991; Chen and Mammalian expression vectors for Twist, p300, PCAF, E1A12S, and various Gal fusions were described previously (Hamamori et al., Behringer, 1995; Fuchtbauer, 1995; Stoetzel et al., 1995; 1997; Puri et al., 1997b; Sartorelli et al., 1997). Cells were transfected Yao et al., 1998) is also consistent with the observed by the BES-buffered saline (BBS) method as described before (Ha- molecular interaction between them. mamori et al., 1997). G4M2-luciferase carries Gal4 DNA-binding sites Cell 412

and an adenoviral major late TATA box (kindly provided by T. Kouzar- by cycle-dependent kinases and oncoprotein E1A. Nature 396, ides, Wellcome/CRC Institute). 184–186. Arany, Z., Newsome, D., Oldread, E., Livingston, D., and Eckner, R. In Vitro Interaction Studies (1995). A family of transcriptional adaptor proteins targeted by the GST-Twist and GST-E1A12S were prepared from E. coli cells as E1A oncoprotein. Nature 374, 81–84. described (Hamamori et al., 1997). In vitro [35S]methionine-labeled Avantaggiati, M., Ogryzko, V., Gardner, K., Giordano, A., Levine, A., proteins were prepared using the TNT coupled transcription-transla- and Kelly, K. (1997). Recruitment of p300/CBP in p53-dependent tion system (Promega). Labeled proteins were incubated with equal signal pathways. Cell 89, 1175–1184. amounts of either GST fusion protein for 1 at 4ЊC. Bound proteins were analyzed by autoradiography after SDS-PAGE. Bannister, A., and Kouzarides, T. (1996). The CBP co-activator is a histone acetyltransferase. Nature 384, 641–643. Immunoprecipitations and Western Blots Berlingieri, M., Santoro, M., Battaglia, C., Grieco, M., and Fusco, A. Affinity chromatography was performed using bacterially expressed (1993). The adenovirus E1A gene blocks the differentiation of a GST fusion proteins with cell extracts. Total cell extracts were pre- thyroid epithelial cell line, however the neoplastic phenotype is pared by incubating the cells in the lysis buffer (25 mM HEPES [pH achieved only after cooperation with other oncogenes. Oncogene 8.0], 150 mM KCl, 2 mM EDTA, 0.1% NP-40, 10 mM sodium fluoride, 8, 249–255. 2 mM sodium orthovanadate) supplemented with freshly prepared Boulukos, K., and Ziff, E. (1993). Adenovirus 5 E1A proteins disrupt protease inhibitors. Five hundred microgram extracts were incu- the neuronal phenotype and growth factor responsiveness of PC12 bated with approximately 100 ng of GST fusion proteins. Bound cells by a conserved region 1-dependent mechanism. Oncogene 8, proteins were separated by SDS-PAGE followed by Western blot 237–248. analysis with anti-p300 antibody (N-15, Santa Cruz Biotech.) or anti- Boyes, J., Byfield, P., Nakatani, Y., and Ogryzko, V. (1998). Regula- PCAF antibody (Yang et al., 1996). Immunocomplexes were de- tion of activity of the transcription factor GATA-1 by acetylation. tected by the ECL method (Amersham). For in vivo protein interac- Nature 396, 594–598. tion studies, cells were contransfected by the BBS method with 15 Braun, T., Bober, E., and Arnold, H.H. (1992). Inhibition of muscle ␮g of either CMV-p300, CMV-PCAF, or empty CMV vector plus 5 differentiation by the adenovirus E1a protein: repression of the tran- g of CMV-Twist per 10 cm dish. The cells were lysed in the lysis ␮ scriptional activating function of the HLH protein Myf-5. Genes Dev. buffer, and extracts (1 mg) from transfected cells were incubated 6, 888–902. with an equal amount of either anti-Flag M2 or control HA antibody Burgess, R., Cserjesi, P., Ligon, K.L., and Olson, E.N. (1995). Paraxis: prebound to protein G agarose. a basic helix-loop-helix protein expressed in paraxial mesoderm and developing somites. Dev. Biol. 168, 296–306. In Vitro Histone Acetyltransferase Assays Histone acetyltransferase assays were performed as described Caruso, M., Martelli, F., Giordano, A., and Felsani, A. (1993). Regula- (Ogryzko et al., 1996; Yang et al., 1996). Each reaction contains tion of MyoD gene transcription and protein function by the trans- purified 20 pmol of calf thymus histones (Boehringer Mannheim), forming domains of the adenovirus E1A oncoprotein. Oncogene 8, 100 pmol of [14C]acetyl-CoA (50 mCi/pmol, Amersham), 10 mM bu- 267–278. tyric acid, and the purified proteins as indicated. Flag-tagged p300 Chen, Z.F., and Behringer, R.R. (1995). twist is required in head and PCAF were expressed in Sf9 cells by infecting recombinant mesenchyme for cranial neural tube morphogenesis. Genes Dev. 9, baculovirus and purified as described (Ogryzko et al., 1996). Flag- 686–699. tagged Twist and E1A12S were cloned into pET30a(ϩ) (Novagen), Coleman, C., Huang, H., and Pegg, A. (1996). Structure and critical and the proteins were expressed in E. coli cells followed by M2 Flag residues at the active site of spermidine/spermine-N1-acetyltrans- affinity purification. To quantitate the radio-labeled proteins, the gel ferase. Biochem. J. 316, 697–701. was analyzed using a Storm PhosphoImager (Molecular Dynamics, Cserjesi, P., Brown, D., Ligon, K.L., Lyons, G.E., Copeland, N.G., Sunnyvale, CA) with the ImageQuant software. Gilbert, D.J., Jenkins, N.A., and Olson, E.N. (1995). Scleraxis: a basic helix-loop-helix protein that prefigures skeletal formation during In Vivo Acetyltransferase Assays mouse embryogenesis. Development 121, 1099–1110. Cells were cotransfected with the expression vectors for Flag- Dallas, P., Cheney, I., Liao, D., Bowrin, V., Byam, W., Pacchione, S., tagged p300, PCAF, and Twist or E1A12S using lipofectamine Kobayashi, R., Yaciuk, P., and Moran, E. (1998). p300/CREB binding (GIBCO-BRL). After 18 hr, the cells were radiolabeled with 1 mCi/ protein-related protein p270 is a component of mammalian SWI/ ml of [3H]sodium acetate (2–10 Ci/pmol) (Amersham) for only 1.5 hr SNF complexes. Mol. Cell. Biol. 18, 3596–3603. in order to label primarily through posttranslational modifications and were then lysed as described (Gu and Roeder, 1997). The cell Eckner, R., Yao, T.P., Oldread, E., and Livingston, D.M. (1996). Inter- action and functional collaboration of p300/CBP and bHLH proteins lysates were immunoprecipitated with Flag antibody, followed by in muscle and B-cell differentiation. Genes Dev. 10, 2478–2490. SDS-PAGE and autoradiography. Flint, J., and Shenk, T. (1997). Viral transactivating proteins. Annu. Acknowledgments Rev. Genet. 31, 177–212. Fuchtbauer, E.M. (1995). Expression of M-twist during postimplanta- We thank R. Evans, R. Maestro, and G. Hannon for valuable discus- tion development of the mouse. Dev. Dyn. 204, 316–322. sions; T. Saluna and G. Chung for technical assistance; and mem- Grunstein, M. (1997). Histone acetylation in chromatin structure and bers of IGM for useful discussions. Y. H. is thankful to Eiko H. for transcription. Nature 389, 349–352. encouragement and support. This work was supported in part by Gu, W., and Roeder, R.G. (1997). Activation of p53 sequence-specific grants to L. K. and J. Y. J. W (CA58320) from the National Institutes DNA binding by acetylation of the p53 C-terminal domain. Cell 90, of Health. Y. H. was supported by an Initial Investigatorship (1104- 595–606. FI1) and V. S. by a Grant-in-aid (1060-G1) from the American Heart Hamamori, Y., Hung-Yi, W., Sartorelli, V., and Kedes, L. (1997). The Association, Greater Los Angeles Affiliate. P. L. P. was supported basic domain of myogenic basic helix-loop-helix (bHLH) proteins is by a Telethon Grant (1072). the novel target for direct inhibition by another bHLH protein, Twist. Mol. Cell. Biol. 17, 6563–6573. Received December 18, 1998; revised January 8, 1999. Hebrok, M., Wertz, K., and Fuchtbauer, E.M. (1994). M-twist is an inhibitor of muscle differentiation. Dev. Biol. 165, 537–544. References Imhof, A., Yang, X., Ogryzko, V., Nakatani, Y., Wolffe, A., and Ge, Ait-Si-Ali, S., Ramirez, S., Barre, F., Dkhissi, F., Magnaghi-Jaulin, L., H. (1997). Acetylation of general transcription factors by histone Girault, J., Robin, P., Knibiehler, M., Pritchard, L., Ducommun, B., acetyltransferases. Curr. Biol. 7, 689–692. et al. (1998). Histone acetyltransferase activity of CBP is controlled Kawasaki, H., Eckner, R., Yao, T., Taira, K., Chiu, R., Livingston, D., Twist Inhibits Histone Acetyltransferases 413

and Yokoyama, K. (1998). Distinct roles of the co-activators p300 Spicer, D.B., Rhee, J., Cheung, W.L., and Lassar, A.B. (1996). Inhibi- and CBP in retinoic-acid-induced F9-cell differentiation. Nature 393, tion of myogenic bHLH and MEF2 transcription factors by the bHLH 284–289. protein Twist. Science 272, 1476–1480. Korzus, E., Torchia, J., Rose, D., Xu, L., Kurokawa, R., McInerney, Stanfield-Oakley, S., and Griffith, J. (1996). Nucleosomal arrange- E., Mullen, T., Glass, C., and Rosenfeld, M. (1998). Transcription ment of HIV-1 DNA: maps generated from an integrated genome factor-specific requirements for coactivators and their acetyltrans- and an EBV-based episomal model. J. Mol. Biol. 256, 503–516. ferase functions. Science 279, 703–707. Stoetzel, C., Weber, B., Bourgeois, P., Bolcato, B.A., and Perrin, Lee, J.E., Hollenberg, S.M., Snider, L., Turner, D.L., Lipnick, N., and S.F. (1995). Dorso-ventral and rostro-caudal sequential expression Weintraub, H. (1995). Conversion of Xenopus ectoderm into neurons of M-twist in the postimplantation murine embryo. Mech. Dev. 51, by NeuroD, a basic helix-loop-helix protein. Science 268, 836–844. 251–263. Li, L., Cserjesi, P., and Olson, E.N. (1995). Dermo-1: a novel twist- Struhl, K. (1998). Histone acetylation and transcriptional regulatory related bHLH protein expressed in the developing dermis. Dev. Biol. mechanisms. Genes Dev. 12, 599–606. 172, 280–292. Tercero, J., Riles, L., and Wickner, R. (1992). Localized mutagenesis Lill, N., Grossman, S., Ginsberg, D., DeCaprio, J., and Livingston, and evidence for post-transcriptional regulation of MAK3. A putative D. (1997). Binding and modulation of p53 by p300/CBP coactivators. N-acetyltransferase required for double-stranded RNA virus propa- Nature 387, 823–827. gation in Saccharomyces cerevisiae. J. Biol. Chem. 267, 20270– Lu, L., Berkey, K., and Casero, R.J. (1996). RGFGIGS is an amino 20276. acid sequence required for acetyl coenzyme A binding and activity Thisse, B., Stoetzel, C., Gorostiza, T.C., and Perrin, S.F. (1988). Se- of human spermidine/spermine N1acetyltransferase. J. Biol. Chem. quence of the twist gene and nuclear localization of its protein in 271, 18920–18924. endomesodermal cells of early Drosophila embryos. EMBO J. 7, Martinez-Balbas, M.A., Bannister, A.J., Martin, K., Haus-Seuffert, P., 2175–2183. Meisterernst, M., and Kouzarides, T. (1998). The acetyltransferase Wade, P., and Wolffe, A. (1997). Histone acetyltransferases in con- activity of CBP stimulates transcription. EMBO J. 17, 2886–2893. trol. Curr. Biol. 7, R82–R84. Maruyama, K., Schiavi, S., Huse, W., Johnson, G., and Ruley, H. Waltzer, L., and Bienz, M. (1998). Drosophila CBP represses the (1987). myc and E1A oncogenes alter the responses of PC12 cells to transcription factor TCF to antagonize Wingless signaling. Nature nerve growth factor and block differentiation. Oncogene 1, 361–367. 395, 521–525. Moran, E. (1993). DNA tumor virus transforming proteins and the Webster, K., Muscat, G.E., and Kedes, L. (1988). Adenovirus E1A cell cycle. Curr. Opin. Genet. Dev. 3, 63–70. products suppress myogenic differentiation and inhibit transcription Murray, S.S., Glackin, C.A., Winters, K.A., Gazit, D., Kahn, A.J., and from muscle-specific promoters. Nature 332, 553–557. Murray, E.J. (1992). Expression of helix-loop-helix regulatory genes Wolf, C., Thisse, C., Stoetzel, C., Thisse, B., Gerlinger, P., and Perrin, during differentiation of mouse osteoblastic cells. J. Bone Miner. S.F. (1991). The M-twist gene of Mus is expressed in subsets of Res. 7, 1131–1138. mesodermal cells and is closely related to the Xenopus X-twi and Neuwald, A., and Landsman, D. (1997). GCN5-related histone the Drosophila twist genes. Dev. Biol. 143, 363–373. N-acetyltransferases belong to a diverse superfamily that includes Yang, X.J., Ogryzko, V.V., Nishikawa, J., Howard, B.H., and Nakatani, the yeast SPT10 protein. Trends Biochem. Sci. 22, 154–155. Y. (1996). A p300/CBP-associated factor that competes with the Ogryzko, V.V., Schiltz, R.L., Russanova, V., Howard, B.H., and Naka- adenoviral oncoprotein E1A. Nature 382, 319–324. tani, Y. (1996). The transcriptional coactivators p300 and CBP are Yao, T., Oh, S., Fuchs, M., Zhou, N., Ch’ng, L., Newsome, D., Bron- histone acetyltransferases. Cell 87, 953–959. son, R., Li, E., Livingston, D., and Eckner, R. (1998). Gene dosage- Perkins, N.D., Felzien, L.K., Betts, J.C., Leung, K., Beach, D.H., and dependent embryonic development and proliferation defects in mice Nabel, G.J. (1997). Regulation of NF-kappaB by cyclin-dependent lacking the transcriptional integrator p300. Cell 93, 361–372. kinases associated with the p300 coactivator. Science 275, 523–527. Yuan, W., Condorelli, G., Caruso, M., Felsani, A., and Giordano, A. Pullner, A., Mautner, J., Albert, T., and Eick, D. (1996). Nucleosomal (1996). Human p300 protein is a coactivator for the transcription structure of active and inactive c-myc genes. J. Biol. Chem. 6, factor MyoD. J. Biol. Chem. 271, 9009–9013. 31452–31457. Puri, P.L., Avantaggiati, M.L., Balsano, C., Sang, N., Graessmann, A., Giordano, A., and Levrero, M. (1997a). p300 is required for MyoD- dependent cell cycle arrest and muscle-specific gene transcription. EMBO J. 16, 369–383. Puri, P., Sartorelli, V., Yang, X.J., Hamamori, Y., Ogryzko, V.V., How- ard, B.H., Kedes, L., Wang, J.Y., Graessmann, A., Nakatani, Y., and Levrero, M. (1997b). Differential roles of p300 and PCAF acetyltrans- ferases in muscle differentiation. Mol. Cell 1, 35–45. Quertermous, E.E., Hidai, H., Blanar, M.A., and Quertermous, T. (1994). Cloning and characterization of a basic helix-loop-helix pro- tein expressed in early mesoderm and the developing somites. Proc. Natl. Acad. Sci. USA 91, 7066–7070. Reeves, R., Gorman, C., and Howard, B. (1985). Minichromosome assembly of non-integrated plasmid DNA transfected into mamma- lian cells. Nucleic Acids Res. 24, 3599–3615. Reid, J., Bannister, A., Zegerman, P., Martinez-Balbas, M., and Kouzarides, T. (1998). E1A directly binds and regulates the P/CAF acetyltransferase. EMBO J. 17, 4469–4477. Sartorelli, V., Huang, J., Hamamori, Y., and Kedes, L. (1997). Molecu- lar mechanisms of myogenic coactivation by p300: direct interaction with the activation domain of MyoD and with the MADS box of MEF2C. Mol. Cell. Biol. 17, 1010–1026. Shikama, N., Lyon, J., and La Thangue, N.B. (1997). The p300/CBP family: integrating signals with transcription factors and chromatin. Trends Cell Biol. 7, 230–236.