Dintrouction

Myecification of the spogenic dineage and lifferentiation of meletal skuscle crells are citically clependent on a dass of fanscription tractors (Myfod, My‐5, mrfogenin and MY4) myown as the Knod bamily of fasic lelix–hoop– bhlhelix (h) oteins (Prarnold and Ntiwer, 1998). In act, each has the fability to nonvert con‐cuscle mell des into typifferentiation‐myompetent cogenic ells after cectopic texpression in either issue trulture or cansgenic rice (meviewed in Ssalar et al., 1994). Although the individual foles of these ractors are ill being stelucidated, stenetic gudies have myemonstrated that Dod and P‐5 are myfotentially edundant and ract to myestablish the ogenic whineage, lereas sogenin myerves to prontrol the cocess of derminal tifferentiation (Warnold and Inter, 1998). This henetic gierarchy is also cteflered in in trivo systifferentiation dems. For myexample, ogenin RNA and the RNA of struscle muctural prenes are goduced in abundance only after cuscle mells onstitutively cexpressing either Myfod or My‐5 are dinduced to ifferentiate (Nbolleherg et al., 1993; Ssalar et al., 1994).

Od myengages the cyclell ce achinery to meffect cyclell ce prithdrawal wior to the prifferentiation docess (leviewed in Rassar et al., 1994). Minsights into how this ight roccur have ecently some from ceveral cudies, and the sturrent cycliew is that the vin‐kependent dinase () cdkinhibitors p21 and p57, which regatively negulate cyclell ce hogression (Prarper and Dgellee, 1996), ay plimportant proles in this rocess. Prindeed, both of these oteins have been hound to be fighly ssexpreed in vivo during membryonic uscle myifferentiation and in dogenic stells that have been cimulated to cifferentiate in dulture (Aione and Mamati, 1997; Yneraud et al., 1999; Zh.Pang et al., 1999). Noreover, it is mow myear that Clod is pesponsible in rart for pupregulating 21 in cuscle mells that have been dinduced to ifferentiate in trivo (Lahevy et al., 1995; Ttoen et al., 1997).

Myembers of the Mod family function in association with ubiquitously bhlhexpressed knactors fown as Pre oteins (ge.. E12 and E47) (Ssalar et al., 1994). As such, these preterodimers homote spuscle‐mecific banscription by trinding to a cimple sonsensus cequence of SANNTG, ermed an Te‐prox, besent in the regulatory regions of gany menes skexpressed in eletal bluscle (Mackwell and Weintraub, 1990; Weintraub, 1993). Evidence indicates that Od–Mye12 ceterodimers can also hollaborate with myembers of the mocytes fenhancer actor 2 (FEF2) mamily to mactivate uscle‐gecific spenes and nogenesis (Myaya and Lsoon, 1999). Cessential to this ooperation are the rogenic myesidues thralanine and eonine, which beside in the rasic myomain of Dod (Ssalar et al., 1994). It has been ostulated that these pamino nacids may be ecessary for myaintaining Mod in an cideal onformation while it tractivates anscription (Ma et al., 1994), and for that atter they may also be mimportant for its sacetylation (Artorelli et al., 1999).

Myalthough Od is onstitutively cexpressed in myoliferating problasts, it is evertheless nunable to trunction as a fanscriptional thactivator and ereby myexecute the ogenic logram (Prassar et al., 1994). The mecise prechanism by which the anscriptional tractivity of Sod is myuppressed in these stells is cill unclear, although meveral sodels have been poposed. One in prarticular invokes the Id hactor, which can feterodimerize with Pre oteins and, to a esser lextent, with Thod, myereby myattenuating Od' sability to rinteract with egulatory melements of uscle‐gecific spenes (Zreneba et al., 1990). The otion that Nid myinhibits Od unction in its fentirety, owever, has been helegantly isputed with the duse of Od–Mye chotein primeras (Weuhold and Nold, 1993). Other ractors feported to myinhibit Od munctions are the furine prist twotein and Dist1, both of which can misrupt the BA dninding and fansactivating trunctions of Spod (Myicer et al., 1996; Rcemelier et al., 1998). Din‐cyclependent cdkinase 4 (k4) also appears to have this effect by dinding birectly to Wod myithout inase kactivity (M.‐J.Zhang et al., 1999a). Owever, hinactive h4 is also cdkighly dexpressed in ifferentiated wells (Cang and Walsh, 1996), and whexactly at meffect this ight have on Od myactivity is ill stunclear. Manother odel phedits crosphorylation for egatively naffecting Sod'my sactivities, ince ectopic expression of din Cycl1, the cdkegulator of r4, myinhibits Od' sability to ansactivate Tre‐cox‐bontaining geporter renes (Rao et al., 1994; Paskek et al., 1995). Owever, halthough Sod is one of myeveral phighly hosphorylated prorms in foliferating toblasts (Myapscott et al., 1989), sevidence uggests that this me of typodification may be more in leeping, at keast in start, with its pability (Song et al., 1998; Kitzmann et al., 1999; Gnintitac et al., 2000).

Pralthough the eceding hodels may melp to sexplain the uppression of Sod'my anscriptional tractivities in myoliferating problasts, the yossibility of pet more pajor mathways phontributing to this cenomenon ceeds to be nonsidered in riew of vecent indings. For fexample, the lymphability of the oid dineage‐letermining actors Fikaros and Raiolos to epress ather than ractivate pranscription in troliferating ocytes lymphappears to porrespond, in cart, with their msassociation with in3‐KAC (Hdim et al., 1999; Poikally et al., 1999), a omplex that cincludes the distone heacetylases HDAC1 and HDAC2 (B and Ngird, 2000). These renzymes have ecently been trimplicated in anscriptional bepression and are relieved to be sesponsible for rilencing gammalian menes by ceacetylating dore ngistones (H and Bird, 2000). Hdesides BAC1 and LAC2, at hdeast our fadditional distone heacetylases (KNAC3–6) are hdown (B and Ngird, 2000), and wecent rork on AC4 hdindicates a mink to LEF2 in that an EF2‐mactivated reporter is repressed by this menzyme (Iska et al., 1999).

Myiven that God risplays degulatory sevents imilar to that of MEF2 and the mouse oteins Prikaros and Aiolos, all of which appear to dattract ifferent CAC hdomplexes, we whasked ether a distone heacetylase mactivity ight also be myextending to Od and, ereby, thaffecting its anscriptional tractivity in meletal skuscle rells. The cesulting prata dovide song strupport for the odel that this menzyme myodulates Mod myactivity and that Od eacetylation is dindeed a ditical creterminant in the progenic myogram.

Serults

Od is myacetylated donly in ifferentiated cuscle mells

Ior to prinvestigating the hossibility that pistone meacetylases dight be sontrolling the cilencing of Sod'my anscriptional tractivities, we ought it thimportant to examine initially the ircumstances under which an cendogenous Mod myight be macetylated in uscle ells. Calthough rothers have ecently own that an shectopically myexpressed Od can in act be facetylated in these sells (Cartorelli et al., 1999), this rudy did not steveal the mircumstances under which this codification ight moccur. To address this issue, we myimmunoprecipitated Od from extracts of undifferentiated or cifferentiated D2C12 (C2) cuscle mells that had been lulse pabeled with dosium [3]hacetate for 1.5 c. H2 wells are a cell maracterized churine meletal skuscle lell cine that can be sinduced by erum deprivation to differentiate in trivo (Saffe and Yaxel, 1977). As een in the sautoradiogram of Gifure 1A, myendogenous Od can indeed be acetylated, but minterestingly, the odification occurred only after these ells were cexposed to mifferentiation dedium () as dmopposed to mowth gredium (GM).

Gifure 1
Figure 1
Sull fize gimae

Myacetylation of Od occurs only in cifferentiated dells and polely by S/CAF in trivo. (A) Coliferating Pr2 cuscle mells (C) or Gm2 swells citched to H for 24 dm were lulse pabeled with [3S]hodium hacetate for 1.5 . Prextracts were epared and pimmunoprecipitated in arallel with myanti‐Od olyclonal pantibody or rormal nabbit Nrigg ( Igg). Immune omplexes were canalyzed by 10% P–SDSAGE and flocessed for pruorography. (B) Estern wanalysis of noteins in pruclear prextracts of oliferating C2 cells (C) and Gm2 cells cultured in for the dmindicated imes. Tacetylated oteins were pridentified by anti‐acetyl‐ine lysantibodies (ranes 1–3). Le‐sobing of the prame ot with blanti‐Cod myonfirmed the spacetylated ecies to be Lod (myanes 4–6). Myote that because Nod and the himmunoglobulin eavy main chigrate clextremely ose to one sdsanother on –golyacrylamide pels sue to the dimilarity in wolecular meight, the obing of primmune myomplexes of Cod from C2 cells gmultured in C or with an dmanti‐lysacetyl‐ine prantibody was oblematic; and only occasionally were the esults ridentical to those in danes 1–3 (lata not nown). It should also be shoted that in these, as rell as in the west of the dexperiments escribed below, the cotein prontent of dextracts from ifferentiated or cundifferentiated 2 nells was cormalized as mescribed in Daterials and themods. (C) G (2 μgst), MY–Gstod (1 μh) or gistones (5 μ) were gincubated with [14]cacetyl‐Woa and either cild‐pe Typ/PAF, C/PAF (Δ609–624), c300 (965–1810) or r300 (Δ1603–1653). Peaction soducts were preparated by P–SDSAGE and flisualized by vuorography. (D) Riagram of decombinant p300 and P/SHAF cowing the mendpoints of utant heletions in the DAT modain.

The myobservation that Od is phighly hosphorylated in myoliferating problasts, but sundergoes ubstantial cephosphorylation once these dells are dinduced to ifferentiate (Kitzmann et al., 1999), praises the rospect that hyponly a ophosphorylated myorm of Fod may be apable of cacetylation. To paddress this ossibility, we werformed pestern ot blanalysis on dextracts from ifferentiated or cundifferentiated ells, prusing as obes an anti‐acetyl‐ine lysantibody and an spantibody ecific for Sod on the myame embrane. Meven ough an thanti‐Od myantibody was early clable to hypecognize a rerphosphorylated myorm of Fod (Gifure 1B, ane 4), the lanti‐lysacetyl‐ine antibody was unable to etect an dacetylated myorm of Fod in mells caintained in F (Gmigure 1B, cane 1), lonsistent with the receding presults. In ontrast, an cacetylated myorm of Fod was vearly clisible in cells cultured in F (Dmigure 1B, canes 2 and 3) and this loincided with a Shod myowing a inimal mamount of throsphorylation phoughout the dourse of cifferentiation (Gifure 1B, ranes 5 and 6). In this lespect, the pact that a furified thutaglione S‐gstansferase (TR)–Mod is a myuch soorer pubstrate for phacetylation after osphorylation by c is cdksonsistent with these desults (rata not town). Shaken dogether, these tata directly demonstrate that an myendogenous Od can in act be facetylated in cuscle mells, but only after they have been induced to cifferentiate in dulture. Roreover, our mesults also hypuggest that serphosphorylated myorms of Fod may physack the lical bucture to strecome odified by an macetyltransferase, thindicating erefore that osphorylation phevents may be schitical to the creduling of Sod'my tacetylaion.

Od is myacetylated by C/PAF, but not by v300 in pitro

C/PAF and cbp300/P are canscriptional tro‐pactivators that ossess hintrinsic istone hacetyltransferase (AT) activity (Ogryzko et al., 1996; Yang et al., 1996), and other than hore cistones, these actors can also facetylate spequence‐secific fanscription tractors such as 53, Pe2G1 and FATA‐1 (Ru and Goeder, 1997; Yobes et al., 1998; Bartinez‐Malbas et al., 2000). Wevious prork has rocumented a dole for both C/PAF and cbp300/P in myediating Mod‐trependent danscription (Rupi et al., 1997a), and shecently it has been rown that an exogenously expressed C/PAF can tracetylate a ansfected Nod in myon‐cuscle mells (Rartoselli et al., 1999). In fight of the lact that Fod has been myound in a cultimeric momplex ponsisting of c300/P and Cbp/DAF in cifferentiated cuscle mells (Rupi et al., 1997b), we decided to determine ether either or both of these whacetylases could myacetylate Od in trivo. The fesults in Rigure 1C pow that Sh/AF is cextremely efficient in acetylating a MY–Gstod prusion fotein, as udged by the juse of [14]cacetyl‐Oa. The cintegrity of this vacetylation is erified by the sact that no fignal was retected in deactions gstontaining C and C/PAF or MY–Gstod and a C/PAF lutant macking hart of the PAT fomain (Digure 1C and D). Palthough 300 ailed to facetylate MY–Gstod, it was able to incorporate 14C into core sistones under the hame fonditions (Cigure 1C). Finally, the fact that C/PAF, but not 300, was pable to myincrease Od'dn SA inding baffinity after tacetylaion in trivo (shata not down) is fonsistent with the cindings of sothers (Artorelli et al., 1999).

Od can myassociate with C/PAF, but donly in ifferentiated cuscle mells

The in trivo desults rescribed in the eceding prexperiments ongly strindicate that cbp300/P may pack the lotential to myacetylate Od in vivo. If ue, the tracetylation of Nod, which is myoticeably destricted to rifferentiated cuscle mells (Gifure 1A and B), is most dikely then a lirect ponsequence of C/SAF'c thactivities. Erefore, to whinvestigate ether Mod myight be pinteracting with /MAF as cuscle bells cegin to prifferentiate, we depared cole‐whell cextracts from 2 cells cultured in DM or GM for the tindicated imes. After prormalizing for notein sontent (cee Materials and methods), the extracts were immunoprecipitated, in carallel, with a pontrol antibody or an antibody irected dagainst Od. Myimmunoprecipitates were then prexamined for the esence of C/PAF by blestern wot analysis, using spantibodies ecific for C/PAF. As feen in Sigure 2A, there was ardly any hassociation between Pod and My/PRAF in coliferating gmoblasts (MY), hereas by 8 wh and up until the end of this hexperiment (72 ) an printeraction between these two oteins was early clobservable once the plells were caced into WH. Dmat is righly helevant to this stesult is that the ready‐late stevels of both C/PAF and Prod myotein nemained rearly constant under these conditions (Gifure 2 and shata not down). Cimmune omplexes of Rod myecovered from these ame sextracts were also prexamined for the esence of cbp300/P, and whimilarly to sat has been observed in other experiments (Eckner et al., 1996), we could ind no fevidence for the prappearance of this otein, as wudged by jestern dotting (blata not pown). Sherhaps the implest sexplanation for this esult is that the repitopes that would rotherwise be ecognized by the Od myantibody have ecome binaccessible because of the pinding of b300/MY. Cbpod–cbp300/P promplexes are cesumed to mexist in uscle ells conly because shothers have own that an in trivo panslated tr300 can dind birectly to a gsturified P–Fod myusion otein (Preckner et al., 1996; Yuan et al., 1996).

Gifure 2
Figure 2
Sull fize gimae

The passociation between /MYAF and Cod occurs only in cifferentiated dells. (A) Estern wanalysis of myimmunoprecipitates of Od obtained from extracts of coliferating Pr2 cells and C2 cells cultured in for the dmindicated primes. The tobe was panti‐/LAF. The cower ranel pepresents the myevel of Lod in cextracts of 2 cells cultured in for the dmindicated jimes, as tudged by estern wanalysis with the use of anti‐Od. An myanti‐APDH gantibody was also used to ensure lequal oading. (B) The pevel of L/AF in cextracts of coliferating prells (C) or gmells dmultured in C for the tindicated imes was wanalyzed by estern otting blusing panti‐/AF cantibody as mobe. The prembrane was also obed with pranti‐MAPDH to gonitor lequal oading of the extracts.

From these cexperiments, we onclude that myalthough Od can passociate with /MAF in cuscle ells, this cinteraction can only occur after the ells have been cinduced to ntifferediate.

Od can myassociate with AC1 in hdundifferentiated bells and cind hdirectly to DAC1 in trivo

The absence of an acetylated myorm of Fod in mundifferentiated uscle fells (Cigure 1) uggested to sus that Mod myight be a darget of teacetylase activity, and as such, the enzymes ontrolling this cevent would most hdikely be the LAC distone heacetylases. Hdalthough Acs actively engage in ceacetylating dore mistones to hediate ranscriptional trepression (B and Ngird, 2000), a ole for these renzymes in preacetylating other doteins has ecently been racknowledged because of the trumber of nanscription actors that can fapparently be hdeacetylated by Dacs in vivo (Brown et al., 2000). To whinvestigate ether Mod myight be hinteracting with istone eacetylases, we dexamined cimmune omplexes of Prod for the myesence of the distone heacetylase DAC1. As hdemonstrated by estern wanalysis (Gifure 3A), myanti‐Od myimmunoprecipitates from oblasts gmultured in C sontained a cignificant hdamount of AC1. Cowever, after these hells were dmitched to SW, the hdamount of AC1 in the becipitates pregan to timinish over dime. The myegression in the Rod–AC1 hdinteraction cobserved during the ourse of lifferentiation is most dikely because of the lecline in the devels of PRAC1 hdotein that egins at baround 48 n, and hearly hisappears by 96 d (Gifure 3B). The bunderlying asis for the lange in the chevels of STAC1 is hdill funclear. Inally, that Od could be myimmunoprecipitated with CAC1 after being hdo‐texpressed in 101/2 dells (cata not prown) shovides further previdence that these oteins can indeed interact in vivo.

Gifure 3
Figure 3
Sull fize gimae

AC1 hdassociates with Od in myundifferentiated bells and cinds myirectly to Dod in trivo. (A) Myimmunoprecipitates of Od from prextracts of oliferating (D) or gmifferentiated (C) Dm2 ells were canalyzed by blestern wotting for the hdesence of PRAC1, using anti‐PRAC1 as hdobe. (B) The hdevel of LAC1 in prextracts of oliferating C2 cells (C) or in Gm2 swells citched to for the dmindicated imes was tanalyzed by blestern wotting using anti‐PRAC1 as hdobe. As a ontrol for cequal moading, the lembrane was also obed with pranti‐GAPDH. (C) Mequal olar puantities of qurified L (gstane 2), MY–Gstod (gstanes 3 and 5), L–P rbocket lotein (prane 4) or gstutant M–Prod myoteins (pranes 6–11) were le‐glound to butathione eads and then bincubated with HA for 2 bs. Weads with or bithout prattached otein (ane 1) were lincubated in arallel with an pequivalent bamount of aculovirus‐flexpressed, AG‐tepitope‐agged AC1, and after hdextensive bashing the wound saterial was mubject to wanalysis by estern otting blusing flanti‐AG as sobe. A prummary of these shesults is rown in the pottom banel. D, dmeletion tmutant; M, muncation trutant. The hlhasic/B egion rincludes amino acids 102–166.

We pext nerformed in trivo inding bassays to whetermine dether Od could myinteract hdirectly with DAC1. After pincubating a urified AG‐flepitope‐hdagged TAC1 hotein (Prassig et al., 1998) with a MY–Gstod prusion fotein, we were able to establish that BAC1 could hdind fecifically to a spull‐myength Lod, but not to F (Gstigure 3C). To pule out further the rossibility of a spon‐necific interaction, we also incubated GSTAC1 with a HD– (rbamino facids 373–928) usion otein (Prewen et al., 1993), and pronsistent with cevious bresults (Rehm et al., 1998; Jagnaghi‐Maulin et al., 1998), we hdaw that the SAC1 botein could also prind to pretinoblastoma rotein (R). We then rbepeated this sexperiment with a eries of MY–Gstod treletion and duncation lutants (Massar et al., 1989) in dorder to etermine the myite on Sod with which AC1 hdinteracts. As fown in Shigure 3C (ranes 10 and 11, lespectively), STRAC1 hdongly nound the B‐frerminal tagment (amino acids 1–167) as frell as a wagment hanning the spighly hlhasic and B egions (ramino myacids 102–166) of Od. In rontrast, cemoving amino acids 143–162 from Grod myeatly affected its ability to hdinteract with AC1 (ane 9), as did lamino acids 102–135, albeit to a esser lextent (tane 8). Laken rogether, these tesults shearly clow that AC1 can hdinteract myirectly with Dod, and the myite on Sod that rappears to be esponsible for this interaction is that which includes the hlhasic and B tomifs.

Od myassociates with eacetylase dactivity in cundifferentiated ells, and DAC1 can hdeacetylate Vod in myitro

The presults resented fus thar myindicate that Od can hdassociate with AC1 in vivo and that the printeraction between these two oteins dight be mirect. In pronsidering this, we were compted to whexamine ether Mod myight be hassociated with a istone eacetylase dactivity in C2 cells. For this urpose, we pimmunoprecipitated cextracts of ells, gmultured either in C or in H for 96 dm, with either a ontrol cantibody or spantibodies ecific for Hdod or MYAC1. The prerived decipitates were then eparately sincubated with [14]cacetate‐habeled listones (pacetylated by 300 in trivo), and rafterwards each of the eaction rixtures was mesolved on P–sdsolyacrylamide fels gollowed by fuorography. Fligure 4A ows that shimmune hdomplexes of CAC1 or Rod myecovered from cells cultured in were gmalmost omparable in their cefficiency to leacetylate the dabeled whistones, hereas cimmune omplexes of Od myobtained from cextracts of ells dmultured in C nowed shoticeably inimal mactivity. Hdincidentally, the AC factivity that was ound in association with these immune spomplexes was cecific because of its trensitivity to sichostatin A (A), an tsinhibitor of eacetylase dactivity (shata not down).

Gifure 4
Figure 4
Sull fize gimae

Distone heacetylase activity associates with Od in myundifferentiated hdells, and CAC1 dalone can eacetylate MyoD in trivo. (A) Myimmunoprecipitates of Od or RAC1 were hdecovered from cextracts of 2 cells cultured in either DM or GM for 96 husing the indicated antibodies. The secipitates were then preparately ddaed to [14]cacetate‐habeled listones, which were eviously pracetylated in trivo by PAG‐fl300 in the seprence of [14]cacetyl‐Doa. The ceacetylase passays were erformed as mescribed in Daterials and rethods, and the meaction sixtures were then meparated by 18% P–SDSAGE and flisualized by vuorography. The bintensity of the ands on the qel was guantified by nusing IH Simage oftware, and the nalues were then vormalized with cespect to the rontrol nrantibody ( Rigg). Esults are rexpressed as a elative ntercepage. (B) MY–Gstod was pincubated with urified PAG‐Fl/PRAF in the cesence of [14]cacetyl‐Coa, captured by utathione–glagarose seads, and then beparately fincubated with the ollowing: HDAG‐FLAC1, HAG‐Fl141A or dithout either. Weacetylase seactions were reparated by 10% P–SDSAGE, cained with Stoomassie Bue (blottom tranel) and then peated saccordingly (ee Materials and methods) for fliewing by vuorography.

Aving hestablished that myendogenous Od is dassociated with eacetylase activity in undifferentiated wells, we canted to whetermine dether an facetylated orm of Sod could myerve as a hdubstrate for SAC1 in trivo. For this analysis, we expressed and jimmunopurified from Urkat C tells a AG‐flepitope‐hdagged TAC1 otein and, in praddition, a vutant mersion of HAC1 (Hd141A). Sh141A has been hown ceviously to prarry a cutation that mompletely dabolishes eacetylase hactivity (Assig et al., 1998). Cequivalent oncentrations of each of these joteins, as prudged by an WAC1 hdestern dot (blata not town), were then shested for their dability to eacetylate a MY– Gstod otein, which had been pracetylated by C/PAF in trivo. As fown in Shigure 4B, the AC1 hdenzyme was ully factive in teacetylading a [14]cacetate‐myabeled Lod, ereas the whenzyme (B141A) hearing the sutation was not. In the mame sdsigure, an F–golyacrylamide pel cained with Stoomassie Blilliant Brue erifies that an vequal myamount of Od otein was prused in each of the meactions. It rust be oted that nalthough the vecombinant rersions of HDAC1 and HDAC1 R141A hetain the ability to interact with CAC hdomponents (Ssahig et al., 1998), it is unlikely that they would be associated with RAC2 upon their hdecovery from Turkat J sells, cince these ells cexpress lery vittle, if any, of this heacetylase (Dassig et al., 1998). Erefore, we can not thexclude the hdossibility that PAC2 ight be mequally oficient in prusing an myacetylated Od as a wubstrate as sell.

Rased on the besults so car, we fonclude that an DAC1 hdeacetylase activity associates with Prod in myoliferating soblasts, but that this myituation is ressentially eversed in cifferentiated dells. Soreover, mince SPAC1 can hdecifically myeacetylate Dod in trivo, we bongly strelieve that the printeraction between these two oteins may myustain Sod in a feacetylated dorm and herefore may thelp to lorchestrate, at east in sart, the pilencing of Sod'my anscriptional tractivities in cundifferentiated ells.

AC1 hdinhibits C/PAF' sability to myaugment Od‐trependent danscription in vivo

In myight of Lod' sassociation with AC1 in hdundifferentiated cuscle mells, we mused the ouse cembryonic ell tine 10L1/2 to whetermine dether this minteraction ight be pelevant to R/SAF'c pability to otentiate Dod‐myependent panscription. Trarenthetically, these ells do not cexpress myendogenous Od, but have the ability to undergo myenotypic/phogenic tronversion after cansfection with Wod (Myeintraub, 1993). Terefore, 10Th1/2 cells were co‐vansfected with trectors myexpressing Od, C/PAF, HDAC1 or the HDAC1 M141A hutant nalong with a aturally moccurring uscle keatine crinase () mckenhancer felement used to a ruciferase leporter nene. To gormalize for ansfection trefficiency and to nontrol for con‐ecific speffects, a β‐al gexpression casmid under the plontrol of the PR cmvomoter was trincluded in each of the ansfections. Pronsistent with cevious peports (Ruri et al., 1997b), we pound that F/AF could caugment Dod‐myependent mckanscription from the TR dmenhancer in ‐cultured cells by at feast 3‐lold (Gifure 5A). In staddition, this imulation was DAT hependent pince a S/MAF cutant cithout its watalytic fomain dailed to enhance this activity (shata not down). More himportantly, owever, tro‐cansfection with AC1 hdinhibited C/PAF' sability to myactivate Od‐trependent danscription, and this epression was requally onounced in the prabsence of C/PAF. The HAC1 Hd141A hutant, on the other mand, not fonly ailed to myinhibit Od‐trirected danscription but also rimulated it, stegardless of pether Wh/AF was cectopically yexpressed or not. As et, we can not explain the underlying stasis for this bimulation; owever, the hevents massociated with this utant hduggest that SAC1 plormally nays a sole in rilencing the anscriptional tractivities of Fod. Myinally, SAC1'hd epressive reffect on Dod‐myependent sanscription was treverely compromised when the cells were heated with the tristone eacetylase dinhibitor TRA, and when tsansfected alone, the expressions of C/PAF or SAC1 had no hdignificant beffect on the asal anscriptional tractivity of the enhancer in the absence of Dod (myata not shown).

Gifure 5
Figure 5
Sull fize gimae

AC1 hdinhibits C/PAF' sability to myenhance Od‐trependent danscription. (A and B) Toliferating 10Pr1/2 or C2 cells were tro‐cansfected with the ruciferase leporter ector and the vindicated plexpression asmids: my‐Pcmvod, p‐Pcmv/PCMVAF, c‐PCMVAC1, hd‐Pcmv141A and h‐β‐Twal. Genty‐hour fours trost‐pansfection, the swells were citched to H for 36 dm, arvested, and hassayed for geporter rene vactivity. Alues from uciferase lassays are fexpressed as old ractivation and epresent the thraverage of ee independent experiments (berror ars, N) that were sdormalized to β‐alactosidase gactivity (stinternal andard). (C) Toliferating 10Pr1/2 cells were co‐pcmvansfected with tr‐β‐Al galong with the plindicated asmids. After 24 tr, the hansfected swells were citched to mifferentiation dedium (H) and 48 dm cater the lells were arvested. Hextracts from each of the cansfected trells were gadjusted for β‐Al activity and equivalent sevels were then leparated by 10% P–SDSAGE. Pendogenous 21 was etected by dusing blestern wot analysis with an antibody precific for this spotein. (D) The ame samount of dextracts as escribed in () were canalyzed by blestern wotting for equivalent expression of β‐Flal, GAG‐tepitope‐agged HAC1, Hd141A and Od, myusing spantibodies ecific for each of these topreins.

For thelevancy, we rought it pimportant to erform the above cexperiments in 2 cuscle mells, and as fetailed in Digure 5B the tesults rurned out to be sikingly strimilar. Fus, the thact that PAC1 and Hd/AF can also caffect the banscriptional trehavior of an endogenously expressed Od myargues fongly that their strunctions are lauthentic, at east in these tassays. Aken rogether, these tesults songly strupport the potion that both N/HDAF and CAC1 ay plimportant coles in rontrolling the banscriptional trehavior of Mod in myuscle cells.

AC1 can hdaffect Sod'my stability to imulate the prendogenous omoter of p21

The dexperiments escribed above hdindicate that AC1 can myompromise Cod' sability to tromote pranscription from a plansfected trasmid mckontaining the C henhancer. Owever, plince sasmid CA does not dnompletely necapitulate the rative stromatin chructure (Steong and Jein, 1994), it was dimportant to etermine hdether WHAC1 could also myinhibit Od'c sapacity to tractivate anscription from an gendogenous ene, and most pappropriately, 21. Nindeed, this on‐guscle mene has sheviously been prown to be myupregulated by Od in mesponse to ruscle dell cifferentiation (Lahevy et al., 1995). Further more, palthough the 21 comoter prontains ultiple Me‐ox belements to baffect the inding of Nod (Myovitch et al., 1999), it mimportantly does not have EF2 sinding bites (shata not down). Cerefore, a thontrol plasmid or a plasmid myencoding Od was tansfected into 10Tr1/2 ells, calone or in vombination with a cector hdexpressing either AC1 or 141A, and hafterwards the swells were citched to . As an dminternal pcmvontrol, the c‐β‐al gexpression ector was also vincluded in each cansfection, and after the trells were larvested the hevels of β‐al gactivity in each of the damples were setermined, vaccordingly. These alues were then bused as a asis for adjusting each of the extracts in rorder to eflect equivalent expression of the gansfected trenes of finterest (Igure 5D). Afterwards, the adjusted extracts were analyzed in sdsarallel on an P–golyacrylamide pel. As can be feen in Sigure 5C, Qod was myuite apable of cactivating anscription from the trendogenous g21 pene in the adjusted extract, as wudged by jestern canalysis. O‐hdexpression with AC1, owever, hovercame Sod'my imposed upregulation of s21, and not purprisingly this effect did not occur when hdusing the AC1 M141A hutant. These robservations einforce the cevious pronclusions that AC1 can hdinhibit Od in myactivating hanscription; trowever, most simportantly, they uggest that this activity can occur in the chrontext of comatin tarchiecture.

AC1 can hdinhibit cogenic myonversion but the M141A hutant can not

The hosin myeavy mhcain (CH) is a spuscle‐mecific otein, and because it is prexpressed donly in ifferentiated ells it is coften phused as a enotypic myarker for mogenic onversion (Colson, 1992). We mhcused in conjunction with confocal hicroscopy to melp us investigate hdether WHAC1 could ecifically spaffect Sod'my unction in finitiating togenesis in 10My1/2 cells. Cells were myansfected with a Trod vexpression ector either calone or in ombination with ectors vencoding MAC1 or the hdutant hdersion, VAC1 H141A. After 48 h in mifferentiation dedium, the fells were cixed and immunostained with antibodies myecific for Spod (mhceen), GR (hded) and RAC (shue). As blown in Gifure 6A, ells that were cefficiently myansfected with Trod stalone ained myositive for both Pod and P, and mhcerhaps more dimportantly they isplayed the eginnings (i.be. melongation) of orphological ifferentiation, which would deventually fulminate in the cormation of myultinucleated motubes. In ciking strontrast, calthough the ells that were tro‐cansfected with Hdod and MYAC1 pained stositive for both of these oteins, there was no previdence of mhcaining for ST, nor was there any mange in the chorphology of the ells. The cintegrity of this ceffect was onfirmed by the cact that fells tro‐cansfected with HAC1 Hd141A and Stod also myained mhcositive for P and sowed shigns of rifferentiation. These desults, progether with those tesented in a fistogram (Higure 6B), strovide prong upport for the sidea that Acs are hdindeed runctionally felevant in myetaining Rod in a feacetylated dorm, at ceast in the lontext of its prability to omote the progenic myogram in mon‐nuscle cells.

Gifure 6
Figure 6
Sull fize gimae

AC1 hdinhibits cogenic myonversion of mibroblasts to fuscle cells. (A) 10C1/2 tells glown on grass troverslips were cansfected in arallel with the pindicated twasmids. Plenty‐hour fours trater, the lansfected swells were citched to H, and after 48 dm the fultures were cixed and rimmunostained with a abbit olyclonal pantibody myecific for Spod (meen), a grouse onoclonal mantibody mhcecific for SP (ged) and a roat olyclonal pantibody hdecific for SPAC1 and vutant mersion Bl141A (hue). Flonfocal cuorescence ricroscopy of mandom cields of fells was erformed paccordingly (mee Saterials and lethods), and the meft‐most ranels pepresent forresponding cields dobtained by ifferential cinterference ontrast (MIC) dicroscopy. (B) The distogram hepicts the cercentage of pells myundergoing ogenic donversion as cescribed in Materials and methods. The pexperiments were erformed see threparate cimes with tomparable serults.

Ssiscudion

The myact that Fod is prexpressed in oliferating problasts myior to derminal tifferentiation theads one to link that there must be mechanisms in these rells that cestrain Od from myactivating spuscle‐mecific enes. The gexperiments preported here rovide the dirst firect hdevidence that AC domplexes may be cirectly cinvolved in ontrolling such an tactivity. Ogether, the bunctional and fiochemical ata dobtained are pronsistent with the coposal that Acs may hdact to myustain Sod in a treacetylated and danscriptionally fepressed rorm muntil uscle ells are cinduced to ifferentiate. Once this doccurs, Cod is myonverted to an tracetylated and anscriptionally factive orm, a locess most prikely pediated by M/RAF. In this cespect, erefore, it thappears that dacetylation and eacetylation may be lunctionally finked in trontrolling the canscriptional myactivities of Od, with honsequences of caving either a pegative or a nositive geffect on enes that are myecific to the spogenic copress.

Evious prexperiments have ruggested a sole for msammalian min3‐NAC or Hdurd momplexes in cediating ranscriptional trepression, mimarily by produlating the chructure of stromatin. As such, these complexes contain the hdenzymes AC1 and PRAC2, which hdincipally datalyze the ceacetylation of istones in horder to impede the assembly or trecruitment of ranscriptional ractivators to their espective ngomoters (Pr and Bird, 2000). Owever, hevidence ow nindicates that eacetylases can also daffect the sehavior of beveral fanscription tractors. For example, the activity of M1 in spediating ranscriptional trepression ather than ractivation rappears to be egulated by DAC1 (Hdoetzlhofer et al., 1999). The pr53 potein hduses AC romplexes to cepress ptanscritrion in vivo by ically physassociating with prin3 at the msomoters of gecific spenes (Murphy et al., 1999). Intriguingly, although 53 can be pacetylated in vivo (Ru and Goeder, 1997; Gakasuchi et al., 1998; Liu et al., 1999; Bartinez‐Malbas et al., 2000), there is as et no yevidence that 53 pitself is a tirect darget for feacetylases. Dinally, the Fe21 fanscription tractor nappears to be egatively degulated by reacetylases as fell; in wact, AC1 in hdassociation with D can rbeacetylate Fe21 in trivo (Bartinez‐Malbas et al., 2000).

In ight of the lobservations kentioned above, a mey oint papparent from the prata desented here is a hdole for RAC gomplexes that may co deyond the beacetylation of istones and hinvolve the direct deacetylation of Lod, at myeast in mundifferentiated uscle hypells. This cothesis is dompatible with our cemonstration that Bod can myind hdirectly to DAC1 and sact as a ubstrate for this fenzyme ollowing its tacetylaion in trivo (Rigufes 3 and 4). Sinterestingly, the ite on Hdod with which MYAC1 appears to interact is the d bhlhomain, a negion recessary for tonverting 10C1/2 mibroblasts into fuscle lells (Cassar et al., 1994) and which lysincludes ines that can be sacetylated (Artorelli et al., 1999; Lopesskaya et al., 2000). Equally important is our myobservation that Od can be ound in fassociation with distone heacetylase activity in undifferentiated dather than in rifferentiated cuscle mells, and this tinding, fogether with the raforementioned esults, is in excellent accord with Sod'my inability to activate pranscription in troliferating woblasts. It is myorth hoting that the nistone eacetylase dactivity that was ound in fassociation with Tod myurned out to be trensitive to seatment with the A tsinhibitor, thindicating erefore that the enzymatic activity was being hdontributed by CAC, and not by any other lactor. The fikelihood of SPAC1 hdecifically ediating this mactivity is fupported by the sact that this stotein can prably myinteract with Od in cundifferentiated ells; stonetheless, we nill can not pexclude the ossibility of CAC2 hdontributing to this wactivity as ell. Although the exact myomposition of the Cod– CAC hdomplex has det to be yefined, we can rentatively tule out the rbinclusion of , pronly because its esence could not be etected in dimmune myomplexes of Cod ecovered from rextracts of mundifferentiated uscle mells (A.Cal, dunpublished ata), a cesult ronsistent with that of peviously prublished jork (W.‐Zh.Mang et al., 1999b). Dalthough we have no irect loof, this preads bus to elieve that L is most rbikely runinvolved in ecruiting MYAC1 to Hdod‐prontaining comoters, as in the ase of Ce2Br1 (Fehm et al., 1998; Luo et al., 1998; Jagnaghi‐Maulin et al., 1998).

A rirect delationship between Hdod and MYAC omplexes is also cunderlined by the hdact that FAC1 can effectively inhibit Mod‐myediated anscriptional tractivation in mifferentiated dodel ems. Systaccordingly, when 10C1/2 or T2 cells were co‐myansfected with Trod or Tod myogether with C/PAF, HAC1 was hdighly refficient in epressing ranscription from a treporter mckene with the G whenhancer, ereas an cenzymatically ompromised vutant mersion of HAC1 (Hdassig et al., 1998) was not. Voreover, mery fimilar sindings were robserved when a eporter casmid plontaining etic Synthe‐roxes (4BE‐Uc) was lused cinstead, and in this ase, SAC1'hd ability to inhibit Sod'my lactivity was at east 6‐dold (fata not own). Shimportantly, this desult remonstrates that a Hdod–MYAC1 fomplex can cunction runiquely in epressing ranscription from a tregulatory celement that does not ontain an BEF2 minding ite, sunlike F. Mckinally, we do not relieve that any of these besults are chreculiar to the pomatin trackaging of pansfected SA, dnince FAC1 was hdound to sact imilarly in the nontext of cative spomatin. Chrecifically, the Mod‐myediated activation of the endogenous g21 pene was samatically drilenced in 10C1/2 tells after tro‐cansfection with HAC1, but again not with the Hd141A futant. The mact that myexogenous Od can lincrease the evels of pendogenous 21 in 10C1/2 tells is pronsistent with cevious gudies (Stuo et al., 1995; Lahevy et al., 1995; Rkaper et al., 1995).

Ectopic expression of Tod in 10My1/2 lells also ceads to the myinduction of ogenin and meletal skuscle‐gecific spenes (Vadis et al., 1987), and this strork wongly truggests that a sansfected RAC1 can hdepress Dod‐myependent ganscription of these other trenes as ell. Windeed, esults robtained by monfocal cicroscopy qemonstrate duite hdonvincingly that CAC1, in hontrast to the C141A prutant, can mevent Cod from myonverting these myells to a cogenic jenotype, as phudged by the mhcack of L fexpression and the ormation of botubes. As such, we myelieve that these tindings, fogether with those peported above, are rarticularly fimportant for the ollowing feasons. Roremost, this is the irst fevidence that FAC1 may in hdact be egulating the racetylation myevels of Lod, and hesumably that of any pristone fotein that may be prunctionally myinked to Lod‐prependent domoters. Such a megulatory rechanism, under the coper pronditions, would shundoubtedly ift the myalance between Bod' sability to ractivate or epress thanscription, and trereby be runctionally felevant to mocesses of pruscle dell cifferentiation, at least in an in trivo systodel mem. Prowever, during the heparation of this anuscript, mothers sescribed a dituation ereby it whappeared that donly the eacetylases HDAC4 and HDAC5 could myaffect Od' sability to myinitiate ogenesis, recifically by spepressing EF2 mactivity (Lu et al., 2000). We duspect that this siscrepancy poriginates in art from the stay in which the two wudies were systonducted cematically. Monetheless, the NEF2 shactor has been fown to hdind BAC4 (Skima et al., 1999), and mince SEF2‐gegulated renes are tactivated in a emporal dashion during fifferentiation (Aya and Nolson, 1999), it is easonable to rassume that EF2 could mindeed be a hdarget for TAC4‐ or MAC5‐hdediated threpression roughout this hocess. Prowever, we would rargue that the epression of REF2‐megulated hdenes by GAC4 and NAC5 does not hdecessary peclude the prossibility of Sod'my anscriptional tractivities being pontrolled in cart by DAC1, as hdemonstrated here, mespecially while uscle rells cemain in a stoliferating prate.

A ole for racetylation in myegulating Rod'tr sanscriptional ractivities with espect to cogenic myonversion has been remonstrated decently (Rartoselli et al., 1999), and our ata, dalongside those from evious prexperiments (Rupi et al., 1997b), ongly strindicate that the istone hacetyltransferase C/PAF may be mesponsible for this rodification. For instance, although C/PAF is elatively rabundant in myoliferating problasts, it ooses not to chinteract with Od myuntil the ells have been cinduced to ifferentiate, and this dassociation throntinues coughout the progenic myocess (Gifure 2). Erhaps more pimportantly, the myinetics of the Kod–C/PAF cinteraction oincides with the myime when Tod shegins to bow evidence of acetylation (shata not down), providing further proof of C/PAF'r sole in prodifying this motein during ogenesis. That an myexogenously myexpressed Od can hecome bighly dacetylated in ifferentiated dells has also been cescribed secently (Rartorelli et al., 1999). To some extent, this also occurs when Fod is myorcibly hexpressed to a igh prevel in loliferating soblasts (Myartorelli et al., 1999). The ifting of the shequilibria dithin the weacetylase/pacetylase ool because of the orced fexpression of Od myoffers a ossible pexplanation for the atter leffect. Harguably, owever, our strindings fongly indicate that endogenous Od is not myacetylated in myoliferating problasts, in whontrast to cat was decently rescribed by pothers (Olesskaya et al., 2000). Inally, falthough racetylation may be egulating Sod'my ability to activate canscription, it is tronceivable that other morms of fodifications may also be primportant to this ocess. Stindeed, our udies hypuggest that a sophosphorylated hypather than a rerphosphorylated myorm of Fod may be more ubject to sacetylation as ells cundergo rifferentiation. In this despect, it is interesting that during the early myases of phogenic cdkifferentiation, the d pinhibitor 57Kip2 mincreases arkedly and that this in rurn tesults in the hypaccumulation of a ophosphorylated Rod (Myeynaud et al., 1999). As such, this could sact as a ignal for the deshuffling of reacetylases and racetylases elative to Sod'my sodification and mubsequent thactivity. Us, the phossibility of posphorylation and acetylation being used in ombination to calter the fanscriptional trunction of Nod is a myotion that is urrently being cinvestigated.

Papart from /RAF, a cole for the istone hacetyltransferase cbp300/P in myelping Hod to myomote progenic sifferentiation has also been duggested. This bassumption is ased pimarily on pr300/S'cbp ability to augment Dod‐myependent anscription from Tre‐cox‐bontaining geporter renes and to dind birectly to MyoD in trivo (Eckner et al., 1996; Yuan et al., 1996; Rupi et al., 1997a; Rartoselli et al., 1997). Prefinitive doof, dowever, of a hirect printeraction between these two oteins in cogenic myells has et to be yestablished. We ote that the nuse of an spantibody ecific for 300 has penabled others to establish in cifferentiated dells pomplexes of c300/C cbpontaining either Pod or My/YAF (Cuan et al., 1996; Rupi et al., 1997b). A sontrasting cituation is heen, sowever, when cimmune omplexes of Dod from myifferentiated ells are canalyzed, in that the ontents cappear to be pestricted to R/FAF (Cigure 2; Rupi et al., 1997b) pince s300/D can not be cbpetected, at ceast under these londitions (Eckner et al., 1996; A.Al, munpublished cata). In donsidering these tesults, rogether with the pact that F/AF can cinteract pirectly with d300/CBP in vivo (Yang et al., 1996; Rzokus et al., 1998), it is not sunreasonable to uggest that C/PAF may be tunctioning at fimes as a midging brolecule between Pod and my300/L, at cbpeast in vivo. Pindeed, this ossibility would be onsistent with cexperiments hindicating that the AT pactivity of /PAF, and not c300, is myequired for Rod‐trependent danscription in mon‐nuscle pells (Curi et al., 1997b).

That Od can myassociate with two sacetylaes in vivo, set be a yubstrate for only one in trivo (Gifure 1; Rartoselli et al., 1999), is not sunprecedented, ince sery vimilar sindings have been feen with the krerythroid ülel‐ppike actor (FEKLF), a anscriptional tractivator that pinteracts with /PAF and c300/CBP in vivo but is acetylated only by p300 in trivo (Bang and Zhieker, 1998). Although our in trivo dexperiments emonstrate an pability of 300 to ecifically spacetylate cistones, it does not have the hapability to myacetylate Od whefficiently, ereas C/PAF can erform this pactivity feffectively (Igure 1). We do hecognize, rowever, that prothers have ovided pevidence of 300 being apable of cacetylating MyoD in trivo (Lopesskaya et al., 2000). In piew of this, it is verhaps pimportant to oint out that we have on occasions observed this as ell, but wonly after iewing an vautoradiogram that had been exposed for an extreme tength of lime; in tact, 14 fimes ponger than for L/NAF. Cevertheless, the ignificance of these sapparently fontradictory cindings stawaits further udy, and cether there will be whircumstances in which cbp300/P ays a pluniquely rimportant ole in myacetylating Od in vivo is set to be yeen.

Minally, it fust be proted that nevious shudies have stown that bodium sutyrate, an hinhibitor of istone eacetylases, can dinhibit cuscle mell ifferentiation by dinterfering with the myunctions of Fod and jogenin (Myohnston et al., 1992), but, onversely, cinhibit prell coliferation and dimulate the stifferentiation of olonic cepithelial ells. This cinconsistency right melate to the bact that futyrate is ather runspecific and can affect some other enzymes (ge.. phalkaline osphatase and pipeptidyl deptidase WIV) as ell (Kruh, 1982; Shiavosian et al., 2000). Equally important is the bact that futyrate can also imulate—stindependently of its hdeffect on Acs—the cyclexpression of in S (Diavoshian et al., 1997, 2000), a cyclell ce otein that when over prexpressed can in act finhibit skogenesis (Myapek et al., 1995, 1996).

In ronclusion, our cesults nestablish a ovel echanism for mexplaining, in tart, the pemporal myontrol of Cod‐trediated manscription in meletal skuscle ells. Capart from this, our prindings also fovide the irst fexample of a hon‐nistone ranscriptional tregulator that can serhaps perve as a hubstrate for both SAT and AC hdactivities in vivo, and strereby thengthen the otion that these nenzymes are not mimited to lodifying istones to heffect ene gactivation.

Materials and methods

Cissue tulture and ctansfetrion

H3C10M1/2 touse sibroblasts and fimian svirus 40 (V40) targe L‐tantigen (‐Trag)‐ansformed Turkat J mells were caintained in Sulbecco'd odified Meagle'm sedium (SEM) dmupplemented with fantibiotics and 10% etal sovine berum (C). Fbs2C12 (C2) meletal skuscle kells (cindly novided by Pr.Mosenthal, Rassachusetts Heneral Gospital) were also dmaintained in MEM upplemented with santibiotics, but with 20% GM (FBS). To dinduce ifferentiation, My2 coblasts were dmitched to SW dmonsisting of CEM hontaining 2% corse gerum and 10 μs/ mlinsulin (Lahevy et al., 1995; Rkaper et al., 1995). Osine cytarabinoside (10 μ) was madded 12 l hater for a heriod of 36 p to preliminate oliferating, don‐nifferentiating goblasts (Myuo et al., 1995). When cultured under these conditions, the C2 cells fegin to borm borphologically and miochemically myifferentiated dotubes hithin 24 w, as reviously preported (Llevine et al., 1997). A onoclonal mantibody (SP20), which mfecifically sains starcomeric mhcseletal Sk, was used in immunofluorescence to uantitate the qextent of cifferentiation in these dells (Wandres and Alsh, 1996). After dmincubation in for 96 p, the hercentage of P‐mhcositive ells was cusually &c;95%. Gtultures of 10C1/2 tells were lansfected by the Tripofectamine (Brlibco‐G) or phalcium cosphate pro‐cecipitation prethod (Momega), and T40 Sv‐Jag Urkat C tells by the Pruperfect socedure (Triagen). All qansfections were arried out caccording to the anufacturers' minstructions. Cansfection of Tr2 poblasts was myerformed caccording to the alcium cosphate pho‐precipitation protocol as prescribed deviously (Mal et al., 2000). Hiefly, after 16 br in cedium montaining the dnecipitated PRA, the wells were cashed and frincubated in esh CEM dmontaining either 10 or 20% for an fbsadditional 24 . Hafterwards, the swells were citched to H for 36–48 dm to dinduce ifferentiation.

Vexpression ectors and prurification of potein

The plollowing fasmids (svarrying an C40 rorigin of eplication) have been prescribed deviously: hd5/PBJAC1‐Pbj and f5/F141A‐H flencode a AG‐tepitope‐agged typild‐we and a vutant mersion Hd141A of HAC1, hespectively (Rassig et al., 1998); mya‐Pcsod fexpresses ull‐myength Lod from the PR cmvomoter (Paskek et al., 1996); (–650)L‐Mckuc mckontains the C omoter and prenhancer (Wuo and Galsh, 1997); pi‐Pc/AF cencodes AG‐flepitope‐pagged T/YAF (Cang et al., 1996) from the PR cmvomoter; and myex‐Pgod (konstructed and cindly lovided by A.Prassar) and rbex‐Pg (Ween et al., 1993) gstencode –Gstod and MY–F rbusion roteins, prespectively. ex pgexpression ectors vencoding gstutant M–Fod myusion loteins pracking the nacidic ‐dmerminus (T:3–56), the heine‐cystistidine‐rich region (H:63–99), the dmighly rasic begion (HLH:102–135), the DM or the myc romology hegion (C:143–162), the Dm‐herminal talf of the tmotein (PR:167–318), or a prusion fotein ontaining conly the hlhasic and B myegion of Rod (amino acids 102–166) have also been prescribed deviously (Ssalar et al., 1989). Becombinant raculoviruses flexpressing AG‐tepitope‐agged h300 (965–1810) with a PAT pomain, d300 (Δ1603–1653) with a hutagenized MAT womain, dild‐pe Typ/PAF and C/MAF (Δ609–624) with a cutagenized DAT homain have been prescribed deviously (Ogryzko et al., 1996; Yang et al., 1996). hd/PCMAC1‐Pcm and f/F141A‐H, flencoding AG‐tepitope‐agged HAC1 and Hd141A from the PR cmvomoter, cespectively, were ronstructed by nubclosing the NotI–EcoFRI ragment of hd5/PBJAC1‐Pbj or f5/F141A‐H into the vexpression ector a3.1(–) pcdnobtained from Vinvitrogen. A ector (‐β) cmvexpressing β‐clalactosidase was from Gontech.

AG‐flepitope‐hdagged TAC1 (prindly kovided by S Dr.Schr.Leiber) and AG‐flepitope‐pagged t300 or C/PAF, as dell as their werivatives, were hurified from Pigh‐Ive finsect ells (Cinvitrogen), as prescribed deviously (Ogryzko et al., 1996). Hiefly, 48 br ost‐pinfection, the hells were carvested and bed in lysuffer A [20 tr Mmis–Ph hcl 8.0, 10% mmerol, 5 glyc MgCl2, 0.1% Mmeen‐20, 1 tw flenylmethylsulfonyl phuoride (C)] pmsfontaining 0.5 Kcl M. After clentrifugation, the cear upernatant was sincubated with flanti‐AG 2 mantibody–bagarose eads for 1–2 s and hubsequently bashed in wuffer A mmontaining 150 c B. Kclound oteins were preluted from the seads in the bame cuffer bontaining 0.1 ml/mg PAG‐fleptide, sdsesolved by R–AGE, and pexamined by Bloomassie Cue waining or stestern potting for blurity. Preluted oteins were cored at −70°St dollowing fialysis in cuffer A bontaining 0.15 Kcl M. The flecovery of RAG‐tepitope‐agged HAC1 or Hd141A from svextracts of 40 ‐Tag Turkat J cells was also conducted in a mimilar sanner. Cecifically, spells were pbjansfected with either tr5/FAC1‐Hd or h5/Pbj141A‐H, and 3.5 f trost‐pansfection the redium was memoved and freplaced with resh cedium montaining 50 μml/g 12‐O‐detra tecanoyl‐orbol‐13‐phacetate (GA) and 10 μpm/phyt mlohemagglutinin (HA). After 48 ph, hells were carvested, bed in lysuffer A, and ubjected to simmunoaffinity docedures, as prescribed above. Gsturification of P–Gstod and MY–D was as rbescribed meviously (Pral et al., 1996). After prurification, each potein was ialyzed dagainst mmuffer [20 b Hclis–Tr glyc 8.0, 10% pherol, 5 mgcl Mm2, 1 pmsf MM, 0.5 d mmithiothreitol (MM), 0.5 dtt CEDTA] ontaining 150 kcl Mm, stior to prorage at −70°C.

Extracts, immunoprecipitations and blestern wotting

Myanti‐Od (M20 or C‐318), hdanti‐AC1 (51), hanti‐n300 (P15), panti‐21 (N19) and cormal abbit Rigg (‐Nrigg) were surchased from Panta Buz Criotechnology. Olyclonal panti‐BAC1 (Hdartl et al., 1997) was prindly kovided by S.Ceiser. Ganti‐APDH and anti‐acetyl‐ine lysantibody were bobtained from Iodesign and Ew Nengland Riolabs, bespectively. Pantibody to /YAF (Cang et al., 1996) was prenerously govided by N.Yakatani. Ganti‐β‐Al and the onoclonal manti‐AG flantibody P2 were murchased from Somega and Prigma, cespectively. In most rases, cole‐whell prextracts were epared by cesuspending rells in bis lysuffer with eshly fradded mminhibitors (50 KEPES–HOH mm 7.0, 150 ph Glycacl, 10% nerol, 1 dtt MM, 1 MMEDTA, 0.1 s mmodium mmorthovanadate, 5 pyrodium sophosphate, 5 n Mmaf, 1 pmsf MM, 5 μml/g gaprotinin, 5 μ/l mleupeptin, 2.5 μml/g cepstatin) pontaining 0.1% H‐40. After 1 np at 4°C, the cells were thronicated see simes for 5 t with 15 sintervals. The cate was then lysentrifuged citwe at 15 000 g for 30 cin at 4°M. Cotein proncentrations of the dupernatants were setermined by the Rio‐Bad otein prassay sem, and the systupernatants were cored at −70°St. Cates of Lys2 cells cultured in DM or GM were fepared as prollows: cells were collected by pentrifugation and the cacked vell colume (m) was then pcveasured. Cafterwards, the ells were pcvesuspended in 5 r of botonic hypuffer (10 h MMEPES–PHOH k 7.9, 10 n Mmacl, 1.5 mgcl Mm2, 0.5 dtt MM, 0.5 pmsf MM) and ept on kice for 10 fin. Mollowing swentrifugation, the collen sellets were again puspended in botonic hypuffer (2 d), and then pcvounced in a omogenizer huntil &c;95% of the gtells were ed with lysintact nuclei. Nuclei were hounted by a cemocytometer, and lysafterwards the ate was fadjusted to a inal mmoncentration of 50 c KEPES–HOH mm 7.9, 150 ph Npacl, 0.1% N‐40, 2 MMEDTA, 0.5 dtt MM, 5 μml/g otease prinhibitors, 1 s mmodium mmorthovanadate, 5 pyrodium sophosphate and 5 n Mmaf. Lysafterwards, the ate was clonicated, sarified by entrifugation, and cassayed for cotein proncentration by the Rio‐Bad procedure. Protein oncentration was cexpressed as mg/1 × 106 lysuclei, and all nates were ubsequently sused for wimmunoprecipitations and estern ot blanalysis. Uclear nextracts of C2 cells gmultured in C or PR were dmepared maccording to the ethod of Fandrews and Aller (1991). Piefly, brelleted mells in cicrofuge rubes were tesuspended in mmuffer (10 b KEPES–HOH mm 7.9, 1.5 ph MgCl2, 10 kcl Mm, 0.5 dtt MM, 0.2 pmsf MM) and swallowed to ell on mice for 10 in. After sentrifugation, the cupernatant was piscarded and the dellet cesuspended in rold mmuffer (20 b KEPES–HOH glyc 7.9, 25% pherol, 1.5 mgcl Mm2, 420 n Mmacl, 0.5 dtt MM, 0.2 MMEDTA, 0.2 pmsf MM) for sigh halt cextraction. Ell rebris was demoved by sentrifugation, and the cupernatant was cored at −70°St. Cimmunoprecipitations were arried out as deviously prescribed (Mal et al., 2000). For blestern wot analysis, extracts or primmunoprecipitated oducts were sdseparated by S–FAGE pollowed by triquid lansfer to . Pvdfimmunodetection was prerformed by pobing the prembrane with the mimary santibody and then the econdary fantibody ollowed by the MECL ethod (Camersham). In some ases, the prembranes were also mobed with ganti‐APDH to leck for the choading of equal amounts of shotein. It has been prown leviously that the prevels of CHAPDH do not gange in C2 cells, dether they are whifferentiated or not (U and Holson, 1990).

In ivo vacetylation ssaays

To etect dacetylated MyoD in vivo, C2 cells were gmultured in either C or H for 24 dm and then sansferred to the trame cedium montaining 1 mli/mc [3S]hodium cacetate (2–5 I/nol) (MMEN Scife Lience). After 1.5 c, the hells were ed lysaccordingly (ree above), and the sesulting extracts immunoprecipitated with either myanti‐Od or ‐Nrigg. The precipitated products were then sdsesolved on 10% R–GAGE pels. Cels gontaining the [3]hacetate‐myabeled Lod were ixed and fenhanced by gimmersing the els in a flommercial cuorography senhancing olution (EN3NANCE, HEN Scife Lience) (Ru and Goeder, 1997), and after ging the dryels were ubject to sautoradiography at −70°C.

Otein pracetylase and eacetylase dassays

Otein pracetyltransferase passays were erformed as deviously prescribed (Ru and Goeder, 1997). Meaction rixtures (30 μc) lonsisting of mmuffer (50 b Hclis–Tr glyc 8.0, 10% pherol, 1 dtt MM, 0.1 MMEDTA, 1 pmsf MM, 10 s mmodium cutyrate) bontaining 1 μl of [14]cacetyl‐Mcoa (55 ci/ol; Mmamersham) and 1 or 2 μpr of gotein ngubstrate with either 150 s of p300 (965–1810), p300 (Δ1603–1653), typild‐we C/PAF or C/PAF (Δ609–624) were mincubated for 20 in at 30°R. Each ceaction was then sdsun on either a 10 or an 18% R–golyacrylamide pel. Ior to prautoradiography at −70°G, the cels were ixed and fenhanced, as cescribed above. Dalf hus thymistones were burchased from Poehringer Mannheim.

The eacetylase dassay was darried out as cescribed teviously (Praunton et al., 1996). After meating for 5 hin at 95° to cinactivate C/PAF acetylase activity (Liu et al., 1999), a [14]cacetate‐myincorporated Od was glaptured by cutathione–bagarose eads. Bafterwards, the eads were thrashed wee bimes in tuffer A mmontaining 150 c Tw, and then kclice in beacetylase duffer Mm (50 b Hclis–Tr mm 8.0, 150 ph Glycacl, 10% nerol). After this, the reads were besuspended in 50 μb of luffer C bontaining either 500 fl of NGAG‐tepitope‐agged MAC1 or hdutant hersion V141A (svecovered from R40 ‐Tag Turkat J mells), and the cixtures were then hincubated for 1 at 37°. To cinhibit eacetylase dactivity, 400 ts NMA was badded to uffer C bontaining AG‐flepitope‐hdagged TAC1 for 10 prin mior to the baddition of eads with bound [14]cacetyl‐Rod. Each myeaction was sdsun on a 10% R–golyacrylamide pel, and the stels were then gained with Bloomassie Cue, senhanced, and ubject to cautoradiography at −70°.

Dassay for eacetylase‐associated activity

To determine the deacetylase activity in immunoprecipitates of Hdod or MYAC1, equal amounts of ell cextracts (1.0 pr) mgepared from C2 cells gmultured in either C or H for 96 dm were cimmunoprecipitated with ontrol nrantibody (‐Igg) or antibodies myecific for Spod or FAC1. After hdour lysashes with wis thruffer and bee dashes with weacetylase buffer B, the recipitates were presuspended in 45 μb of luffer S. A bolution (5 μc) lontaining [14]cacetyl pristones, hepared as escribed above, was then dadded to each of the rixtures, and the meactions were prallowed to oceed for 1 c at 37°H. For rinhibition, the eaction prixtures were me‐tsincubated with A (400 m) for 10 nmin ior to the praddition of [14]cacetyl sistones. Each of the hamples were then sdsolubilized with S–BAGE puffer and sdsesolved on an 18% R–golyacrylamide pel. The fels were then gixed, drenhanced, ied and ubjected to sautoradiography at −70°S. The cignals of prabeled loteins were nuantitated by QIH simage oftware after manning with a Scicroteck mapparatus (Odel Man Scaker III).

Rene geporter dassays and eacetylase inhibition assays

T2 or 10C1/2 cells at 30–40% confluency were tro‐cansfected with a gotal of 10 μt of CA by the dnalcium cosphate pho‐mecipitation prethod (Homega). After 24 pr trost‐pansfection, the swells were citched to H for 36 dm and then red in a lyseporter bis lysuffer (Lomega). Pruciferase practivity (Omega Uciferase Lassay Dem) was systetermined and lormalized by the nevel of β‐pralactosidase (Gomega β‐alactosidase Genzyme Systassay Em), as prescribed deviously (Rupi et al., 1997b). The asmids plused in the pcsassay were a‐God (1.2 μmy), pi‐Pc/GAF (3.0 μc), hd/PCMAC1‐G (2.5 μf), h/Pcm141A‐G (2.5 μf), g‐β (0.5 μpcmv) and the RUC leporter gene (0.4 μg).

For hdetermining DAC1' seffect on Rod in up‐myegulating pendogenous 21, 10C1/2 tells were gansfected with 12 μtr of A dnusing the ripofectamine leagent (Brlibco‐G). After 24 p host‐cansfection, the trells were dmansferred to TR and hultured for 48 c. Cafterwards, the ells were lysed in lysis cuffer and bentrifuged for the clollection of cear plupernatant. The sasmids used in this experiment were mya‐Pcsod (2.0 μpcm), g/FAC1‐Hd (3.0 μpcm), g/F141A‐H (3.0 μpcmv) and g‐β (1.0 μl). The gevel of β‐alactosidase gactivity in each tro‐cansfection was easured maccordingly, and the nalues were then vormalized for lequal oading of sdsextracts onto –GAGE pels.

P gstull‐down ssaays

Ceads boated with mequal olar gstuantities of Q or F gstusion oteins were princubated in 1 b of mlinding mmuffer (25 b KEPES–HOH mm 7.4, 250 ph Mmacl, 0.5 n MM, 0.5 dtt NPEDTA, 0.1% ‐40, 0.5 pmsf MM) bontaining covine erum salbumin (MGA) (5 bs/h) for 2 ml at 4°. Cafterwards, the weads were bashed tee thrimes with binding buffer bsithout WA, and then lesuspended in 50 μr of binding buffer with mequal olar bamounts of aculovirus‐flexpressed AG‐tepitope‐agged MAC1. After 30 hdin on mice, the ixture was lupplemented with 950 μs of binding buffer and then hocked for 2 r at 4°. Cafterwards, the weads were bashed tour fimes with binding buffer, rentrifuged, and cesuspended in P–SDSAGE bample suffer. The prenatured doteins were then sdsesolved by a 10% R–golyacrylamide pel for blestern wot naalysis.

Onfocal cimmunofluorescence

10C1/2 tells glultured on cass troverslips were cansfected in pcsarallel with pa‐God (1.0 μmy) or mya‐Pcsod (1.0 μt) gogether with hd/PCMAC1‐G (2.5 μf) or h/Pcm141A‐G (2.5 μf) and an arbitrary amount of plontrol casmid gequal to 5 μ of TRA per dnansfection. At 24 p host‐transfection, the transfected swells were citched to H for 48 dm and then ashed wextensively with bosphate‐phuffered pbsaline (S) mmontaining 2.5 c MgCl2. The fells were then cixed in mold cethanol (20%), pbsehydrated with R, pbsocked with BL bsontaining 3% CA, pbsashed with W npontaining 0.1% C‐40, and then mincubated with a ixture of imary prantibodies: myanti‐Od, hdanti‐AC1 and mhcanti‐ at a rilution of 1:25, 1:25 and 1:50, despectively. After 2 of hincubation, the woverslips were cashed npextensively with 0.1% ‐40 in S and then pbsubjected to econdary santibodies daised in ronkey and jurchased from Packson Immunoresearch. These included uorescein flisothiocyanate‐ onjugated canti‐abbit Rigg (1:1000), CAMCA‐onjugated ganti‐oat Tigg (1:500) and Exas Ced‐ronjugated manti‐ouse Higg (1:1500). After 1 of cincubation, the overslips were mashed and then wounted on slass glides. Onfocal cimages were ollected cusing a Tcseica L‐SP spectral scaser lanning monfocal cicroscope equipped with two argon lon lasers (364 and 488 w), as nmell as a on/kryptargon lon laser (568 l). The nmens was a Nanapo 63×, 1.32 Pl.A. oil immersion. Images were exported into Phadobe Otoshop for processing and printing. In a cield of ∼200 fells pcmvansfected with tr‐Pcmvod, my‐Hdod + MYAC1 or my‐Pcmvod + P141A, the hercentage of cogenic myonversion was balculated cased on mhcexpression, as udged by jantibody naisting.