Tobacco Documents Online

3584 MAYER AND HANAFUSA J. Vraor 4 . FIG. 3. Inducuon of anch~nmge. independent growrh by v.rrA murant> CI-:F were transfected nrlh pCllO.denved DNA plus UR2AV helper vims DNA as marked, plated in soft.agar empen•ion 2 days posuransfecnonn and mcubmed at 37"C Pholomrcrographs werP. lakrn 28 days after plating. Magnification. x 25. agar suspension 2 days posttransfecuon (Fig. 3). At most, one round of vnal replication would have occurred by rhu time, so the effect of mutauon is expected lo hc negligible. The results of this assay were qualitatively similar to Ihose in which later-passage. fully infected cells were used (data nol shown). This colony formation assay is also designed to detect anchorage-independent growth of a small number of transformed cells in a large population of normal cells. Mutant v-erk proteins. We used immunoblotting to examine the level of mulant v-nA proteins expressed in tmns- fected cultures (Fig. 4A). In all ca.es except dSH3, rnutant protein of the expected molecular weight was observed. Fhe level of expression of nontransformmg mulanls was lower than that of wl or fully transforming mutanb, presumabh- owing to selection against repllcatron of the nontransformn€ mutant viruses. We conclude that the lower Ievelsof protein reflect wt levels of expression in a subpopulanon of cell., and not lower, perhaps "subthreshold" levels of expression in all cells. for the following reasons. First. inditecl rmmu- nofluorescence with an anticd, antisemm demonstrated that a lower percentage of cells expressed the nontransforming A B FIG 44 rrR protems and pho~phulvrosme-conlammg proteins of CF.F rransfectcd rrnh v nA mutants IAI AnbrrA immunohlnt: (B) anhptyr rmmunohlot Approxrmatclv 30 ng of prolern from CEF • transfected with mulanls as mdicared was separated on 8 W gets and rmmunoblolted Mzrkcr. (hnes on the Icft of the gel.) : 220'. 100. h8, 47- and 27 kDe. PUBL,CCNTIONS 016619 10335650

t VOL. 64. 1990 mutant protetns. but that the level of expressioh in these 0 cells was comparable to wt levels (data not shown). Second. we examined the steady-state levels of v-crk mRNA in mutant-infected cultures by Northern (RNA/ blo0#ng, and in all cases the level of mRNA was proportional lo th'e levels of mutant protein observed by immunoblotting (data not showN. Since the mutations that we introduced were in most cases minor, we do not expect large diBerences irv viral transcription rate or stability; we therefore assuofe that the reduced mRNA levels seen in cultures >ransfected with the nontransforming mutants indicate that a small number of cells were infected with the mutants. Since mRN.4 levels were proportional to protein levels. we also assume that there were no major differences in prolein half-hfe- although we have nol rigorously excluded this possibilily by pulse- chase experiments. When mutant protein was detected. we conclude that there was no gross defect in viral replication relative to wt CT10 and that the lack of biological activipobserved in the colony formation assav (Fig. 3) was due to lesions in the v-crk proteins of the mutant viruses. Srnce we saw no protein or RNA expression of the -1SH3 mutant, no conclusion could be drawn about the biologrcal properties of this mutant. Phosphotyrosine levels in mulant-transfected CEF were assayed by immunobloning with anti-ptyr 02. 351. Phospho- tyrosine levels correlated absolutely with transforming ac- tivity (Fig. 4B). The fully transforming mulants had phos- Fhotyrosine patterns similar to wl CT70, the pantally transforming mutanl BSP had phosphotyrosine levels only slightly above control levels- and the nomransformmg mu- tants had no detectable elevation in phosphotyrosine levels. These results are consistenl with a central role for elevated (~ phusphotyrosine levels in transformation by v-e rA. We, also `_ examined the association of mutant rrA proteins with cellular phosphoproteins. Wt P47°°' '•" binds to the major phospho- tyrosine-containing proteins of infected cells, of apparent molecular masses 135 to 155, 120. and 70 kDa (21a). The f6lly transforming mutants- SMH and MH, bound to these proteins as well as w•t p47°°°"" whereas the panially tFansforming mutant. BSP. bound much lower levels and the nontransforming mutants bound no detectable phosphppro- teins (Fig. 5). This experiment also demonstrates that the mutant rrk proteins, like wt P47°°I were hrghly phosphorylated- ConstrueUon of gag(-) crk mutants. To acsess the cqmn• bution of viral gag sequences to the biological activity of P47°°"'a we also constructed two gng-deleted IRng(-/l mutants. For these mutants we used a replicahon-compBtent RSV-derived vector (6, 16) in which the uvncated rrA gene is inserted at the posuion occupied by the .cre gene in wdd- type RSV. These constructs do nut require couansfectiun with helper virus to replicate, and should therefore express the mutant rrA proteins more efficienny than in the ca'se of the replrcauon-defecuve. CT10-denved mutants desnibed above. The two gag(-) construcls have different translalion start sites (Fig- 2)- A c-rrA cDNA has been cloned. and the sequence revealed that the cellular product initiates a9 the methionine homologous to methionine 236 of v-rrA. The region of v-rrA between the gag-rrA junction and methiomne 236 is derived from intron and 5• untranslated sequcncei and is not found in the c-rrl product I(-. T. Reichman. Yl. J- • Mayer. S. Keshav, and H. Hanafusa. suFmttled for putihcatronl. The mutant SHB was constructed to use the normal orrA initiator methiomne (amino acid 236 of v-rrAt and encode a:05-amino-acid product. The second mutant. NS(-. MUTAGENIC ANALYSIS OF v-crA 35g5 A B ? ; x r.' o r;v N x e x = k" u J M' S 1r1. vi ^ 2 i v 3 m 4 1a.. ~Fi. u i e wstt.attt.w+.++ FIG ~~ Phospho?vrosrnc-eontarmng proteins assocrated with mutant v-rrA proteins. IA) CEF rnfected with mutant vrrus a-ere labeled m vivo with '=P,. Ivscd. and rmmunoprecrpilaled w•nh anu-rrd serum. Mutants are as marked:lane AV contains UR2AV helper virus alone. The gel was 8.5a polyacrylamide. Posrnons of molecular mass markers are indicated on the IeO: 220. 100. 68. 43. and ]7 kUa Bands of 130. 120. and 70kDa proierns are indicated by arrowheads on the righr. (B) Same geI as rn panel A. washed in I M KOH at 55°C for 2 h to enhance detecuon of phosphotyrosmc conlanmng proten0 (4). was constructed to inntale translation with the amino-ter- mnal 14 amino acids of p60"'- which are fused to the C-terminal 234 amino acids of v-<rA. These 14 srr-derived amino acids are snflicienl to direct mvristoylation of heter- ologous proteins (25). so this construct is expected to direct the synthesis of a myristoylated. 252-amino-acid product- Myristoylation is often sufficient to direct light association of proteins wilh the plasma membrane (25. 31); we hoped that by using the arr amino termmus wc could target this gag( -1 rrA mutant to membranes. as well as take advantage of a translauonnl sian site that is known to work efficiently in a retroviral conlcxt. Biological and biochemical properties of gag(-) mutants- The biological properties of these iwo mutants are summa- rized in Table 2. Both mutants had partially transforming phenotypes. The myrisroylated NSC mutant induced subtle morphological alteration in monolayer culture and very small colonies in suspension (Fig. 6). although in both cases n was much less pntent than wi CT10 vim.. When injected rnlo chickeru ir induced iumors with high frequency, hut when virus was isolated from tumors it was found to be as transforming as .rt CTIO, funhermore- the tumor-dtrived rrA protein could he shown by tmmunoblotting and rmmunoprecipitauon to have regained viral gag sequences (data not shown). Thts was presumably due to deletion of the sequences between gr,g and rr4, in some cases mediated by eight nucleoudes of viral gag sequence present in the NSC rrd gene. Owing to this recombinalion. we cannot assess the tumorigenic potential of the authentic NSC mutanl. although it musl necessarily be less than that of the gup-conlaining recombinants. The SHB mutant was somewhat more tmn+- formng than NSC in both monolayer culture and suspension (Frg- 6). although still less so than the wl strain. It indtrced tumors in chtckens, but the latency period was longer and the tumors wcre slower growing than in the case of wt CTIO. No recombination with gap was observed with this mulant- so we assume the tumors were induced by the authentic gngf-) mutant. PuBLrcArroNS 016620 10335651

3586 MAYER AND HINAFUS.A TABLE 2 Biological aclrvny bf gupl 1 c.A mulanl. -- A YrorNn Mon•hpAnrhnmpr - - Vamw espre•logreal mdepem Tnmongcmul•" vnn" aherannn" dence' ft10 + NR=00 + ++ r+ ++ 14.4 bt x dpl l NSC ++ . + ' Ihl6 b1 2? dpl I- SF18. ++ . . a(x%xt+v2tdpil' NR?011 - - /(pb hv _x dpr) ° Pnilnn lerel> .+ere suaped ti, immon~o6ha unh.,ni, .d.crum t r-W 7 ^ Assayed in :+ltar-o•-erlamd monoluocr cuhurc ahcr Ihv ercnnJ p.n•aFc /FrR 6) ' AssaYed in syar au•penaun IFrg 6r ° Assayed 6y nprcuon of ea<h w,nB ••ch i+f ntn hrnn chn\rm mumnrr af srtes xnh palw6lc rumnrs'Iwal number of inc. m)eaed. dp. da++ nunl ',Tummrr entuhrim were morc morpholnh¢all> ahcred rhan Ihr pamni.d NSC-cclla m monolxvcr colmrr- r.A prnrem had reunerN r,,r •ryucmc. 'TUrnoh wne relairvch alonrro.mf, ci,culmrr• urrr morphal.•plalis srmdai ro pmenlal SHH ceIt, the gaF( -) mwant protelns were eapre+,ed at relalirelq high levels. comparable lo w( P47°^w '''* IFIg. 7A) . The levek of phoxpholyrbsine in cells infected wuh lhe,c mul:mt,- ho<vever_ were reduced compared v,-rth those in wr-mfecled celOs IFig. 7B1. Although rn snme eNpenmems the tyroalne phosphorylauon of the 1?5- (o 155. and the 120kDa proteln, of NSCanfected cells was comparahle to that in the .vl stimin. the 70-kDa protein phosphol groslne level wa, ;d.+a} < lowcr (Fig 7B: dala nol .ho,vn) . It u'as generally drehcull lo see any elevation of i'_0- and 135- to 155.kDa proteln fIG 7 r rA prolcrn. amd pho•phnlyro.ine.conlarnmF protern• rn CEI' trar•,fccted .•nh x.,xdelcrcJ uA mulemn ('F.fwnr mm•fecled wJh vlr.J I>NA n, m;,rAed. h,:Je• wire prepJed. and prolcros werc ep.mnrJ on 1tY> IAI or 7 y7 IBI {u+h'ucryhlnudc gr•I. Protenn• ••erc n.m.ferrcd to rnnocelluln.c. and filler, wcrc probed wJh anL. rA .crum IAI or anu.plyr IBI Apparenl molcr'ular ma..e• of porcm •IanJ.u.l, nn kdodehon•1 are markcd on the left Pu,mnm of ( 11U. Nti('( and SHB /'rotrm• rA) and 110 . 1](1.-and 7(1-hU0 prnlem. IHl.rrr mdraalyd ht arroune:ul. on Ihe nphl F1G 6. Inducbon of anchoragerndepkndcnt piouih and mnrphologicul alleraLOn in C1f mfectoJ w,lh mux•delercd v r ri mulanl.. ('FI were tran,fecled wuh ( T10 plu, NR'/x71C71O1IA1. NS(' rRr. S1114 r(l. or NHMNI helper vrru• (l)1 In Ihe sup row. Iran.fccled cells were plated rn wspenslon 2 days pmn ran.fectiun a, fur F1F 1, and ph.uomrar•rFraph..+ere raACn ] N d:ry. aher plannyt Magmficauon. a 25 In Ihe bottom row. Iransfecled CEF xern malnlbmed in monnhryer cullure under •n0 eKar al SnT. and pha.c.cnnba•I phnmm,ungraph% were taken after theaecond pasnage Mugnr6canon, ._on PUBLIC19TIONS 016621 4 s 10335652

/ y .. (st. 1990 - cno Rr- SR aN? r:Na -y5n2 dar,oaa PU6-339-naan •00 ffiMI-Sp awb ae eSH3 6~_F_~ 9r0--.na suo curwtovtt nVAc at3v nAB6roaH4o rro-8y MH U6S4rWwu YFWIII ESP ESta-EUCO w~sP SMN Pa+6-PP$IG Sm+Wn R294 i Hypa .n ram N2]J ~i Rl)l-N Nprr SHB ffzz)_~ ,nure.mn Mai]aE NSC myr.yroyunv. wwae~e ~'r zo>-aro I 0 FIG 2 ]n•itro-construnrdmulamsoflhevnRgene VrrjIFrrx sequences 1C7). SFt] and SH2domalns =/. and the SHI domam t=/ are indicated. Rectncuon stte• used for or crealed as a result of mulagenesvs arc idenufied (see Marerlals and Melhoda Abbreviations Sr. Srvl: RI. trnRl- Sp. Sp/rl. Ae. At rl. BI). Btdll. hfll. .M1f.ull: HI11. H.nd11L Sm. Srnnl. Ammo acid laal changes in the vannus mutants are mdtcated on the nght in the svngle-leuer amino and code. Amino acid numbers arc a+ in reference 20. w1 vrrA encodes 440 amino acid, I last few amino acids of SH2' L%SH3/ was deleted. When convenient restriction snes were crvailaMe- linker Insenlon mutants resulting in the in-frame msenton of four or flve amino acids were generated. These insertions were in the amino-terminal pan of SH3 IBSPI. in the region helween SH2' and SH3 (MH). within the SH2' block IESP), and in the region between SFl2 and SH2' (SMH 1. Since there are no convenient restnction sltes in SH2- two mumants with point mutations at well-conserved residues within this domam were made by oligonucleoude-duecled mulagenesis. The histidine at position 294 was changed to argnnne tR'94)- and the arginine at position ?73 was changed to asparagme (N273). The argininc at position 273 is absolutely cunserved in all known SH2 domains and is withln the very hlghly conserved PLVRXS hexapeptlde- The hlstldinc at position 294 is conserved in all but one SH: domain and ts at Ihe C-terminal boundary of the SH: homology. Biologienl eelivity of v-erk mutants. 'tAe biological activity of the mutants was assaycd by transfection of pCT10. denved DNA, along with UR2AV helper virus DNA, into CEF- In this system. expression of the mutant tr( gene product is dependent on rephcation of the mutant CT10 • virus. UR2AV DNA ie cntransfected to provide essential replicatrve gene producls in trans to the replreauon-defective C:T10-denved virus. Since all mulants contain the same long terminal repeat and ris-acnng replicat.lve sequences (except TABLE l. Biological activity of v-rrA mutants Vlru. CTIO + UR2AV ASH2 + UR?AV .1SH3 + 11R2AV BSI' - UR2AV MH + UR2AV ESP + UR2AV SMH + UR2AV R29J + UR2AV N273 + UR2AV UR2AV Prolem Morph..Anchorapc e~prc.. lopmal mdepen- Tumongemcdp° Aon" allerHnOn" d[rlee' +++ ++ ++ ++(4/4by13dpil + - - - (0/4 by 34 dpi/ - - - (0'4 by 34 dpi) ++ - + ~ (1/4by34dpd` +++ 4 + ++ ++(4/4by16dpr) + + - . ° ('J4 by 74 dpr)' +++ ++ ++ ++t4f4by13dpi1 (0/4 by 23 dpi) la4hya3dpil - - - - tOr4 by 34 dp0 ° Proran Ih-tIS asauyed by immonublor wuh anrrnR s m iF,p. 41 " Ae>ayed in arar-ovenaid mnnlaso cullure a0er Ihe thud passaac ' Assaped m apar auepenrron IF1g 11 ° As+sved hf iNe<non of each v.mg uvb of nvu-bom chickens mumber of sne, u,nh p:dpaWe rumoreholal numner of snee rNecled: ho. day, 1%nnNecIron l ` Onl, lumor mdueed -as small and slo..--grow'inrp cocultured tumor crlb urre nor morpholog¢ally elmred in monolaca cullure ' Tumoi cocahures ..ere morphnlogrcaly nnered in monolaytr eultun:. unhkc the parCnml 1n.E5P mmant for two Inulants that are expressed in an RSV-derived vector Isce heloyvl). the efficiency of replication of the mutants should be similar to Ihat of wt CTIU. Three crilena for the Iransforming activity of Ihe mutant rrd viruses were assayed: morpholo(tidal alteration and anchorage-independent growth of infected CEF and tumor- Igenicity in newborn chickens. Table I summanzes these resulls- The only mutants thal retained full transforming acnvny were SMH and MH. which had linkers Innened into the nonconscrved reglon, between the 5H2 and SH2" do- mains or lyetween Ihe SH]' and SH3 domains. These viruses were induungwshable from parental CTIO by all cntcria examined. 'ihe BSP mutant. containing a linker in the SH) dnmain. had a very weak transforming phenotype. Infected cells werc mdislmgmshahle from helper-infected eontrols /n monolayer cullure, but were able to form small colonies in sofl agar (Frg. 3). In addilion, one of two chickens lone of four snesl mlected with virus developed a small, slow- growing tumor. Vlrus recovered from this tumor had the propenles of the parental BSP mutant, suggesting that the lumor was Induced by Ihe authentic mutant. The remaining mulants had no detectable Iransforming activity in mono- layer culture or agar suspenslon. The tumorigenic activity of the ESP vlrus was shown lo he due to generation of fully Iransformmg variants. since vtrus recovered from tumors induced morphological aheratEoll and anchorage-indcpen- dent growth, unhke the parental mutant. Since retroviruses generate frequent mutants through rep. hcatton errurs. one often observes either the to.s of sc- quences that are not selected for (such as nonlransforming oncogenc sequences) or accumulalion of mutants thal are selected for (changes that activate the transforming potential of nonuansformng genesl (5- 15: BJ.M. and H.H., unpub- bvhed ohscrvations). This may be especially true with repli- catlon-defective viruses such as CTlO. since the defective sarcoma virus is at a selective disadvantage compared with rephcauon-compelenl helper virus unrll the culture becomes fully Infected. To mmimlzc the effecl of viral mutation and selection dunng culture. it is therefore useful to have a transformation assay that does not depend on viral spread and full Infectlon. For this reason we assayed the ability of mutanl-Iransfected fEF to form colomes when plated in PUBL ICATIONS 016618 MUTAGF:NIC ANALYSIS OF v-erA 3<83 10335649

0 JOVRNnI Or V1ROL0al'• Aug. 1990. p. 35g1_ac8q 002.-538JU90/003581-09502.0010 Copynght x(°' 1990. Amencan Socieqfor Msrobrulog) f ILE COP Y Va+l. 1.J• Na K Mutagenic Analysis of the v-crk Oncogene: Requirement for SH2 and SH3 Domains and Correlation between Increased Cellular Phosphotyrosine and Transformation BRUCE J- MAl'ER* AND HIDESABURO HANAFUSA' Lnhoralaq' of Alr+lecalar Orrrulax r. The Rr.rArlrllcr Lqrit <r:cih. 1230 firrk Arrmrc. Nen- Yrok. NenTrrrk lfN)ZL Recerred R FeCruay 19911'Accepted '_5 April N)'a) We have constructed a series of mutants with delelion, linker inser1ion, and point mulallons In 11Ae v-rrk oncogene of avfan sarcoma virus CTIO. The v-rrk gene cronteins no apparcnl calutrllc domain. but does contain two blocks of homology to putalive regulatory domains, termed SH2 and SH3, found in a varfeA,r of proteins implicated in signal Iransduc6on. Infection with CT10 causes a dramatir increase in the leccl of tyrvnlna• phosphorylolion of several cellular proteins. N'e found that mutation of eilher the SH2 or SH3 domain of v-rrk reduced or eliminated transforming actfvily, whereas mutation or regions oulside the consrnrd domains had no effect. Delelion of aminu-terminal gag sequences cnused a partial luss of Irunsfonning ac1i. ity and a chanpa• in subeellulnr distribution of the rrk protein. In all eases, there was an absolute correlation between Increased celluter phosphot-rrosine and Iransformatfon. The erk oncogene of avian sarcoma viruo CT10 encode, a 47-kilodalton (kDa) (tag fusion producl- termed P47'°'-" 120). A second crd-encoding avian ,arcoma viru, wa, re- cently described thal has a genetic structure very similar to CTIO (33). Although the v-rrA gene has no sequence simi- larity to any known catalytic domain. il containn two hlock. S of sequence. the SH2 and SH3 domarns• which are found in a wide variety of prolerns. These domams are thoughl to regulate the activity of proteins in which Ihey are found. perhaps by interacting with catalytic domaen, or wnh alhcr cellular proteins (21a. 241 The SH2 domain (the B and C boxes of Stabl et al. 129)) i• composed of one region of approxrmately 50 ammo aod. I Il box) and a second region of 10 amino aod. (SH]' or C hntl separated by a variable regron, It was originally identified a• a common N-ierminal domain in all nonrcccptor protcrn- tyrosine klnases 1_'S) and was subsequently found in two isoforms of phosphalidyhnositol-specrfic pho.pholipa,c C' (PLD.y and PLC-IV) (10. 29. 30) and thc ras GrPa.c activator prolein (32. 34). as well as v-rrk. The SH3 dumam (A box), consisling of approxtmalely 50 amino acid.- i• found in a v. ider variety of protems, including mosl nunrr. ceptor lyrosine kinases: PL.C-'y and PLC-IV; the (iTPasc activator protein: the cyloskeletal matris protem. fodrin. myosin-1, and acrtn-bmdmg protein ANPIh: the Sar, haro. myces rererisiae proleins Cdc2g and Fusl, and neutrophd cytosohc oxidase factor p47 (9. 18. 24• ?7) . Whcncvn enzymatic aclivily has been examined Irn the case uf the tyrosine kinases, PI.Cs, and the Gi'Pa.e activator protein). the SH2 and SH3 domain, have been shown to he unnece.sary (2, 3, 10, 17, 19)- Since the SH2 and SH3 dumain1 arc found independently in at ieaa one prorern and are pre,ent in various combinations in others, we presume that they have modular and independent functlons. CT30-transformed cells have greatly elevated phosphoty- rosine levels on at least three cellular protetns compared ' Corresponding amhor. + Present address. The Whitehead In.urure for Biumed.cal Rcsearch. Cambrrdge. MA 02142 .vtth normal c ontrol cell. 120) _ We havc shown that P4Y'°"" can lightly associate with Ihe.e pho.pholyrosrnc-conlaining protcros t71. 21a) and that tyro+inc kina.e and scr'inr Ihreumne kmasr activuy alx. coimmunoprecipitate with P47..•..' C(1. 21. 21a) . lhc>c da4r led us ur propose a model in which the v-rrA prodact increases lyrosme phpsphoryla- uon of ccllular >uMtratc prolein. hy mediating the formauon of tcrnnry complexc, conu.ting of P47°""'", substrntc protcm.• and endogcnou% protem kinu.c. ('1a). Wc havc constructcd u scrics of mutant e-crd genc% tu detcrmrnr which domenn, of the protein urc required for transforming activity and lo correlate tran.fornnng activity with hroehemical paramcrcrs rn muhmt-infeclcd ccll.. We pre.enl evidence that the SH2. SH: . and SH3 domiim tire all required for full oan•furming activity ;md that N-Icrmimd gag sequences also conlrihule lo tramfurming :rclivity, presumahly hy afTcclinE +ubccllular luculrzanun. M all c;rx.. mcrca.cd cclluhrr pho+photyro.mc correlated w•ith Iranv forminE activity, suggesting Ihal Ihi. rncrCn•c r. Cenlral In the mccham•m of nom•furmuuon by v-crA. MA'fl•:RIAIS ANI) METHUDS Cells and virus. (hrckcn embryo fihrohlasl% rt-P.1') were prcp;ircd and maintaincd cs•cnlinlly a. de•crihed frreviou.ly (14). Vrru% slock% were prepared by culling pC170 or lt• derivativc, woth Snrl, mixing with S..ncclcaved p(]R2AV hcipcr virus DNA 1221. Irgulinh hrir0y• and tran.fccunp. I µg of carh plasmid per M1-cm dish uf CEP h) calcium f+ho.phatc coprccrpmmon f h, ih). Stock, of Rous sarcoma vrru. ( RSV ). dcrivcd mulanl vou.c. were prep;aed as descrihed previuuay (b. )b). 1'r:.n.fcrted cell, werc routinely mninmined in medium containing n.77cr~i agar m 4fl"(-. ('hf Iran+fccted with wrdd-lypc rwu p('i-10 were generally fully tramfurmed by the third passage (12 to 14 day% po.itrandeclninl. 'Io asuay anchorage mdepcndence. CF1~ were tryp.mved : day• po.ltran.fection and I(Y' celks were phned per 11Lem hncicrial plastic dnh in ]) ml of medium containing 11^6crdf .erum• 1,4 chicken scrum. and t1.4'G agar un a l:ryer rrf ] 5 ml of the +ume medium cuntaining 0 7'A agur. The plales were incubated at 3T('. PUBLICATIONS 016616 10335647

i s • Vot.. 64, 1990 78:1624-1628. 18. Lomax. K. J., T. H. Leto, H. Nunol, J- I. Ga18n, and H. L. Malecb. 1989. Recombdnunt 47.kilodalton cytosol factor rc- stores NADPH oxidase in chronic gsanulomatous disease. Set- enee 245:4(lY-A12. 19. Marshall, M. S., W. 5: Hill, A. S. Ng, U. S. Vogel. M. D. Schaber. E. M. Scaldck: R. A. F. Dlxon, 1. S. $igal. and J. B. Gibbs. 1989. A C-terminal domain of GAP is su6incnt to stimulate ras GTPase aclivity. EMBO J. 8:1105-1110. 20. Mqyer, B. J-, M. H.maiucbl. and H. H^..an,sa. 1968. A novel viral oncogene with structural similarity to phospholipase C. Nature (London) 332:272-275. 21. Mayer. B. J., M. Hamaguchl, and H. Hanafusa. 1988. Charac- Ierivation of P47°'""`-a novel oncogene product wnh sequence similanly to a putative dtodulatory domain of piotein-tyrosine kinases and phospholipase C. Cold Spnng Harbor Symp. Quant. Biol. 53:907-914. 21a.Mayar, B. J., and H. HanaRna. 1990. Assoctaluon of the v-rrR oncogene product with phosphotyrosine-contaming proteins and protein kinase activity. Proc. Nad. Acad. Sci. USA 87: 2638-2642- 22. Neckameyer, W. S., ar,d L.-H. Weng. 1984. Molecular cloning and charactenvetiun of avian sarcoma virus UR2 and compan- son of its transforming sequence uith those of other avian sarcoma viruses. J. ViroL 50:914-921. 23. Ol.son, E. N., and G. Sphm- 1986. Fany acylafion of cellular proteins- temporal and subcellular diD-erences belween paimr late and mynslatt acylation. J. Biol. Chem 261:2458-2466. 24. Pawenn, T. 1988. Non•eatalytta domains of cytoplasmic prolein- tyrosine kmases regulatory elementa in signal transducuon Oncogene 3:491-495. 25. Pellman, D., E. A. Garber. F. R. Crosa. and H. HAnalLsa. 1985- An N-terminal peplide from p60"' can dnect rnynsqiatwn and plasma membrane localitation when fused to heteroloEous proteins. Nature d.ondon) 314:37-t-377. 26. Prywes, R., J. Hoag. N. Rosenberg. and D. Baltimore. 198~. Protein stabilization explains the gaR requirement for uamfar. MUTAGENIC ANALYSIS OF v.rrA 3589 mmion of lymphord cells by Ab<laon murine leukemia vims. J. Vuol 84:123-132. 27. Rodarrgv, A. R. F.. M. J. E. StRnber(t. and D. L. Bendey 1989. Similarity in membrane proteins. Nature (London) 342:624. 28. Sadowski, I.. J. C. Stone, and T. Pawson. 1986. A noncalalytic domain conserved rimong cytoplasmic protein-tyrosine kinases modifies the kmase function and tmnsforming activity of FLj: nami sarcoma vints,p130"°°r"'. Mol. Cell. Biol. 6W396440g. 29. Stahl, M. L., C. R. Fetena, K. L. KeHeher, R. W. Kda, and J. L. Knupf. 1988. Sequence similarity of phospholipose C with the non•catalyrc region of srr. Nature (London) 3J2t269-272. 30. Suh, P: G., S. H. Ryu. K. H. Mona. H. W. Sub. and S. G. Rhee. 1988. Inosirol phuspho8pid-speci8c phospholipase C; complete cDNA and protein sequences and sequence homology to ly- rosine kinase-nJuted oncogcnc produas. Proc. Nall. Acad. Sei. USA 85:5419-5423. 31. Towler, D. A., J. 1. Gordon, S. P. Adams, and L. Glaser. 1988. The biology and enzymology of eukeryotic protein aeylulion. Annu. Rev. D&ochem. 57:69-99. 32. Trahe,v, M., G. W6ng. R. Halenbeck, B. RubloOeld, G. A. Marlln, bf. Ladnrr, C. M. Long. W. J. Crosier, K. Wart, K. Kolhs, and F. MeCormick. 1988. Molecular cloning of lwo types of GAP complementary DNA from human placenta. Science 242:1697-1700. 3). Tsuehle, H.. C. H. W. CAeng,111. Vashlda, and P. K. VoR1. 1989. A neuiy isolated avuan sarcoma virus. ASV.1, carries the frA oncogene. Oncogene 4:1281-1284. 34. Vogd. U. S., R. A. F. Dixon. M. D. Srhaber, R. E. Dkhl. M. S. Marshall, E- M. Scolnick. 1. S. SigaL and J. ID. Gibbs. 1988. Cloning of bovine GAP and its interaction with oncogenie ras p21 Nature tl.ondory/ 3Ya:91L9J 3A Wang, J. y. J. 1985. Bsolavon of antibodies for phosphotyrosine by immunization with a vo4f oncogene-enended prc+tein. Mol. Cell. Biol. 5:3640.3613. 36. Wigler, M., A. PelBvef. S. SlAerstdn, R. Axel, G. Urlaub, and L. Chasin. 1979. DNA-medialed transfer of the adenine phospho• nbosyhranmfera>e locus into mammalian cdl.. Prnc. Natl. Acad. Sut. USA 76c137i1376. PUBLICATIONS 0 16s2a 10335655

3582 MAYER AND HANAFUSA Tumodigenicdy was assayed by injection of 0.1 ml of vims in each wing web of newborn 12 to 5 days posthatching) white leghorn chickens (SPAFAS. Inc.). Tococvlture Tumor cells. tumor tissue was minced and then incubated with 0.25% trypsin in buffered isotonic saline at 37'C with gentle shaking for 20 min. Dissociated cells and small clumps were poured dff. pelleted, and added to a feeder layer of freshly plated CtF. Cultures were transferred once to allow virus sprcad. Conslr+retfoer of pCTIO. A nonpermuted CTIO viral ge- nome was constructed from the origihal permured molecular clone. p]0-282 (20). as outlined in Fig. 1. Ftrst. the 0.3- kilobace (kb) Arrl fragment of p10-28'_ t:as deleted by restriction digestion, end filhng. and rehganon to generate p10-82A..The 3.2-kb Scal.EcoR1 fragment of p10-82,1 was combined in a three-pan ligation with the 1.3-kb £roR1-PsA and the L8-kb Psrl-Scal fragments of p10-?82 to generate pCTIO. This construct contains th'e entire nonpermuled CTIOgenume with twolong terminal Cepeatsin a pBluescnpi SK(-) vector (Stratagene Inc.). It rekarns Ihe full biologicai activity of The original molecular clones. Construction of v-crk mutants. p1O--NSH2 was constructed by ligating an Sphl 8-met linker to, the end-filled 5.9-kb EcoRl-Srvl fragment of pCTIO. p1O-aSH3 was constructed by biunt-eetd self ligation of Ihe end-filled 6.0-kb Acrl-ErnRl fragment of pCTlO. plO-BSP was made by ligation of an SpJtl 8-mer linker to BRHI-cut, end-filled pCTIO DNA. Similarly, p10-ESP was made by insenron of an SPfrl 8-mer linker into the end-filled ErnRl site of pCTIO. pl0-MH was made by insenion of a Hind111 12-mer linker into the end-filled A?srll site of pCT1O, and p10-SMH was made by insertion of a Hindill 12-mer linker tntb one of the two Smnl sites of pCTIO. Point mutants were generated by using the hacterial strain and protocols provided with the Mutagene kit (Bio-Rad Laboratories). The 0.23-kb PsA-EcoRl fragment of pCT10 was cloned into M13mp18 and M13mp19. and recombinant. uracil-subsiituted bacteriophages were prepared in £sritr- rirhip roli CJ236. Thu DNA was used as template for oligonucleotide-directed mutagenesis. The mutagenic obgo. nucleotides were 5'-CTTCTTGTTTAACGACTCCGG-3'. which creaAes a novel Hpal site and, changes Arg~273 of v-crk to Asn. and 5'-GACGAT(iTAPCGCGAGACGC-3'. which creates a novel Nne site and changes His-294 to Arg. The nucleo8ides in italics are differen8 from the wild-type CT10 sequence. Progeny clones were grown up. sequenced to confirm rhutations, and reinsened imo pCTIO by ligatuon of the PxtLFroRI fragment to the 6.1-kb EroRl-partial P.rrl fragment of,pCT10f gng-deleted mutants were constnrcted by using a replica. lion-competent RSV vector system (6) ) pl0-SHB was con structed by ligating the 0.7-kb Sfrt-Hur4.l fragmenl of pCTIO to a BmnH11 adaptor (5'-CGGATCCCiGCG(--3' plus t'-CG GATCCGTCA-3'), digesting with Bm»HI. and ligating to Bglll-cleavetl pNR200 (161. p10-NSC was made in two steps First, hybrid vector pNR111, consisting of the 5.8-kb Bx!/ll. CIaI fragment of pNR200 ligated to the 1.8-kh (7nl-Bgftl fragment of pXDl l-I (6). was consiructed. Then the BamnHl- linkered 0.8kb Styl-Hnell fragment of frCT10 was inserted into the Bgfll site of pNR111 to generate p10-NSC. Prolein anxlysis. Isotopic labeling, cell lys4s. immunopre- op)iation, ared immunoblotting have been described prevr ously (12. 20- 21a). Polyclonal rabbit anti-phosphotyrosine serum (anti-ptyr) (12. 35) and anli-rrk serum (21a) have been previously described. Crude membrane (P10D) and cytosolic (5100) fraction. i ta 6c3kb 52 sca 24 si J. Vinor.. FIG. 1. Consiructmn of pCTIO Symbols -. pBluescnpt SK <-) vector sequences. M. cellular proto-oncogene•denved se- quences M - p+v sequencer. [.J- viral long lerminal repeal. Numbers denote kiloha.e pairs from the onginal Arrl site in the vector. Abbrevrauons fnr resmcuon enzymes: Ace. Ar<i; CAcc. dest royed Arrl sue-. ISr- Pnrl -. Eco. E'r nRl. Sty. Sni; Sea, Sr nl. were prepared as follows. A 10-cm dish of infected cells was Scraped in hypotomc buffer (10 mM NaCl. 20 mM Tris IpH 7.4), 1 mM drsodmm EDTA. 1 mM phenylmclhylsulfonyl Pluonde, I mM Na,VO,- 0.1 ruM Na.Mo04, 1% trasylol). gently pelleted- resuspended in 0.5 ml of fresh hypotonic buffer- and incubated on ice for 10 min. Cells wcre broken in a Dounce homogrnner and briefly spun at low speed to pellet nuclei and unbroken celln. Supernatant was spun at 100.000 x g for 30 min at 4"C The pellet (P100) was Qissolvcd in 0.5 ml cdhypotonic buffer containing 1% sodium dodecyl sulfate: sodium dodecyl sulfale was added to the supernatanl (S70(h l0 17. Fractions were boiled for 5 mrn and spun in a mtcrocentrifuge at room lemperature for 10 mtn. and 20 pl of each fraction supematant was analyzed bsudium dodecyl sulfate-polyacrylamrde gel electrophoresis and immunoblotung. ! RESULTS Construclion of v-crk mutants. We have used molecularly cloned ('TIO DNA to construct mutantb with deletion, linker rnsernon, and oligonuclcoude-dtrecled point mutations in the v-rrd gene. Since the original molecular clones were ctrcularly permuted at the ErnRl site within the SH?' domain of v-crA, to facilitate the generation of mutants we cdnstructed a planmid containing the nonpermuted CT10 genome with two long termrnul repeats (Fig. 1). This plas. mid. pCT10, is the parental plasmid of all the mutants deseribed here. The mulants that were constructed are diagrammed in Fig- ~ 2. The rnlire N-termrnal half of the c-rrL-derived sequences. which encodes the SH2 and SH2' domains (ASH2). or the entire C-terminal portion encoding the SH3 domain and the PuBL rcHrIONS 016617 10335648

vot.. 64. 1990 CrrO NSC 5.8 N4]00 P s P S P S P S ~ 411111111110 r . B FIG. 8. Subcellular fractionation of CEF nansfected wirh gae- deleted crA murams. CEF vansfecled wuh vrrus as marked were fracuonated into P100 crude membrane fraction IP) and 5100 cytosalic fraction IS). as desenbed in Maternal, and Methods. and separated on 10% (A) or 7 S'R (B) polyacrYlamnde gels. Protems were transferred to nirroceliulose filters and probed with ann.uA serum (A) or anu-pryr (BI. Positions of,molecular maxs markers are indicated on the left~ 220. )00. 68. 43. 27, and ]8 kDa Posunsn, of CT70. NSC, and SHS protems (A) rond ]30-. 120.. and 70-kpa pmleins (B) are indicated by anowheads on the nght phosphotyrosine content in SHB-mfected cells, although The 70-kDa protein clearly had elevated pho'sphotyrosine levels . _ relative to controls. . /~ Subkellular distribution of gagf-) mutant proleim. Since \~ inutanis NSC and SHB were identical tn wl v.(,rA in the c-crA-derived portion of the gene and since rmmunoblotnng demotistrated that the Iruncaled proteins were expressed at relatively high levels. it was surprPsing that these mutants were not fully transfonning. A IikeEy function for viral Aug sequences is to direct the subcellulat localization of proteins to which it is fused. We therefore examined the subcellular dislribntion of wt P47e°•"r and the Iwo Raxl-) mutants. The majority of wt P47`'°•"'' partitions in the crude mem- brane (P100) fraction. although a srgnificanl proportion is fbund in The cylosobc IS100) fraclimn (Fig- SA). The NSC protein is found almost exclusively in the P]OO fracuon- similar to many other myristoylated proteins 03. 31). Thc tipper bands in the NSC lanes are rrA protetro from variant viruses that had recombined with helper and recovered N-terminal gag sequences. As expecled, labeling of infected cells with f'Hlmynstic acid demonslrated Ihai the authentic NSC rrA protein was myrisloylaled (data not shoum. The SHB prolein, in contrast. is almost exclusively cytosolrc in dostribution (Fig. gAL Unlike wt P4T'°•-•", Iherefore. nei- ther of the two gaX(-) mutants provides srgnrficani amounts of rrk protein in boyh the soluble and the membrane- associated compartments. When the P100 and S1f10 fracliom of Cf10-mfecled cells were analyzed by anti-plyr rmmunksblotung. most of the phosphotyrosine-containing proteins were found in The membrane fraction (Fig. SB), consiStent with previous re. sulu with lyrosine kinase oncogenes (13). The majonly of the 7pkDa phosphotyrosinc-contauling prntein. however, • was found in the cytosolic fraction. If, as we suspect. binding to rrA protein is required For tyrosme phosphoryla. lion of these proteins, these data suggest a tole for rrrA protein in both the membrane and cytosolic compartments MUTAGENIC ANAI.YSIS OF v-rrA 3587 The results with the two gag(-) mutants were hmbiguous. The sum of c-vtosolic plus membrane phosphoproleins did not give the same paltern as when mutant-infected cells were lysed directly by boiling in sodium dodecyl sulfate (Fig. 7B): these results might be due to the lengthy manipulations required for fractionation, since similar results have been observed in several experiments. However, therk are some indicalions (for example, t he higher 70-kDa protein phospho- t)rosine in the cytosol of SHB-infected cells compared with NSC-infected cells) that phosphotyrcisine levels are more elevated in the cellular compartment in which the'rrA protein is found. When the data of Fig. 7 and S are taken together, our results demonstrate that rrA mutarots lacking gag have a different suhcellular localization than the wt protein and that these mutants do not induce the fu11 extent of tyrosine phosphorylation seen with the wt strain: the presence of r A protein in The cylosol or membrane fraction correlates with increased phosphoryrosine in that companment. DISCUSSION In the work described in this paper. we analy2ed a series of in vitro-constructed v-rrA mutants to determine what sequences are required for transformation by this novel oncogene. The v-.rA gene has sequence homology to two domains. termed SH2 and SH3. found in a variety of proteins implicated in signal transduction (24). We found that bolh domains are required for full Iransforthation by v-r.A: in contrast, mutation nf the spacer region: between these domains had no effect on transforming activity. This result highlights the importance of SH2 and SH3 domains. which may play a central role in mediating the assembly of multiprotem complexes during normal signal transduction (21a. 24: M. Matsuda and H. Hanafusa. submitted for publication). We have proposed that v-rrA might elevate the ihtracellu- lar phosphotyrostne content of crrtrcal cellular proteins by binding in a ternary complex with substrate proteins and endogenous protein-tyrosme kinases (21a). In this model. it is reasonable to assume that the SH2 and SH3 domains independently brnd to er[her kmascs or their substrates- but thal both domams are required for a'ssembly of temary complexes This is eonsislent wtlh the mulagenic data pre- sented here and wilh the sequence data for a number of other proterns. which implies thai the SH2 plus SH2' domairn and the SH3 domain mu<t have modular and independent func- trons (2..). Our results are incompatible with a model in whtch either SH2 or SH3 is sole]y responsible for the biological activity of v-rrA. The fact that the SH3 linker inscnion mutanl. BSP. relains some transforming activity could imply that this domain rs less cnbcal than SH2 plus SH7 for transformalron How- ever. the linker in this mutant is inserted inlo a region of the SH3 domain that ts extremely variable when other se- quences arc compared: in fact. Ihe S. rrirvisJar r-dr25 gene product has a five-amino-aetd insertion precisely at the homologoux spot in its SH3 domain (l, 27). This mutanon may therefore alter the conformation insuthciently to com- plelley abolish SH3 function. Further mutation of highly conserved SH 3 residues i. requued ro resolve this question, In contrast, the total lack of biological activity of the two SH2 point mulant. (N273 and Ii294) mphes a critical func- non for the mutated restdues or an extreme sensniviiy to conformational change. We have found that in all cases the degree of increased cellular phosphotyrosine levelN in mulant-infecled cells cor- PuaLrcArroNS 016622 10335653

3588 MAYER AND HANAFUSA related with the degree of transformahon. This is consislenl with involvement of increased lyrosinephosphorylation in a mechanism whereby v-rrk can alter cellular growth control. The most potemially useful result is that partially transform- ing mutants increase phosphotyrosine in a partial manner and never give the full extent of lyrosine phosphorylation seen in wl CT18-transfonned cells- Since only three major phosphotyrosine-tontaining proteins are observed, even with wt v-crk, it may be possible to cnrrelate the degree of phosphorylation df individual proteins with various transfor- mation parameters. The requirement for retroviral gag sequences for full transfotming activity is surprising, since the RaR-deleled mutant viruses express high levels of the c-rrA-derived portion of v-crA. Hetroviral gag has previously been shown to be important for the transforming activity of other onco- genes, including abl and fps. which encode tyrosine kinases (7. 11, 26Tn deregulalion of enzymatic activity. protein stabi- lization, and altered subcellular localization have been pro- posed to explain tfns requirement. We have found that avian gag (which, unlike mammalian gag, is nol mynstoylated 181) results in narlitioning of P471"`""' into both the membrane and cytosolic fractions, although the majonty of the protein is found on the membrane. Our expenments show that v-r.L prolein that is localized almost exclusively fo enher the membrane (NSCI or cytosol (SHB) cannot fully transform cells. PhosphotyrOsinc-eontaining rrk-bmding proteins are found in both compartments (21a; see above)- We assume that increased phosphotyrosine levels on thesc proteins are due to inleraction with the v-rrd protein and that the gag-deleted mutanis cannot interacl produclively with all of the cellular substrate proteins owing lo their subcellular locallzation- A problem with the replicating retroviral expression sys. tem used in these Studies is the high viral mutation rate and apparently strong selection for transforming viruses. Thts leads to a lower percentage of cells expressing nontran.• forming mutants than the wt and to frequent generation of fully transfolming revenants or pseudorevertants from less- trnnsforming parental mutants. The small percentage of cells expressing nontransforming or weakly transforming mutanl proleins made it impossible lo assay these mutanls for rrA-associated tyroslne and scrine-threonme kma>e achvi. ttes as previously observed with the wt stram (20. ?1. 21a). since even with highly expressed wt v-crA Ihese activll ics arc relatively low and difficult to detecl. These problems With the retroviral system can be encumvenled in lhe future by expression of mutant genec in mammalian cell hnes and by in vitro reconsbluuon experi. menls wnh mutant rrd proleirn. wi v-rr4 can transform murine 3T3 cells and rat 3Y 1 cells when highly expressed via an integrated selectable vector. and theserrA-expressmg cell lines have similar biochemical properties to CT1U-mfecled CEF (C. Marshall and H- Hanafusa. unpublished data) . Mutant rrk genes could be expressed at levels comparable to wt v-r.R in mammalian cell lines. allowing a direct compar- mon. More promising is the finding that v-rrA protein ex- pressed in bactena, or in insect cells via a baculovins. vector, can brnd to the phosphotyrosrne-contaimng proteins and to the tyrosine kinase activity of v-rr0ransformed cells (21a; Matsuda and Hanafusa, submitred ). The nontransform- ing mutants descnbed here should prove extremely useful tn such reconstitution expenments to elucidate the functions of the SH2 and SH3 domains and to identify the protelns whose binding to rrA is required for transforming activity. 1. vmot ACKNOWLEDGMENTS We thank Lu.Har Wang for the UR2AV clone and for many yearn • of helpful advice. and we gratefully 3cknowledge the significam conrnbmmns of M. Matsuda and C. Marshall lo this work. We ah,- thank S. Kombluih and Y. Fukm for cntrcally reading the mano scnpt. This work was supported by Public Health Servrce gram CA44356-03 from the National Insulutes of Health and grant 2317 from the Council for Tobacco Research. B.J.kf. was supponed b. training gmnl A1072 3 3-14 from the National Instimtes of Health. LITERATURE CITED l. Broek. D., T- Tods, T. M/chae7l, L. Levin, C. Bfrchaneler, hi. Zoller, S. Poners, and M. WIPJer. 1987. The S. eerevisiac CDC75 gene pmduct regulates the RASradenylate cyclnse path. way Cell48:789-799. ?. Bristol, A.. S. M. Hall. R. W. Knz, M. L. Stahl, V. S. Fan. k1. G. Blrn. R. L. Eddy, T- B. ShoW s, and J. L. Knopf. 1988. Phosphohpase C.148r chromosomal location and delebon map- ptng of functional dnmarrts. C'old Spnng Harbor Symp. Quant Btol. 53:915-920. 3 Brugge, J. S., and D. Darros.. 1984. Analysis of the catalync domain of the phosphrrtransferase activity of two avian sarcoma vlms tmnsforming prorerns. l. Brol. Chem. 259c4550-n557. 4. Cooper. J. A., and T. Hunter. 1981. C-hanges in protein plios phorylauon in RSV-transforrned chicken embryo cells. Mol Cell Bml ta65--ll0 5. Cross, F. R., E. A. Garber. and H. Hanallw. 1985. N-termmal deletions in Rous sarcoma vrms p60"_ effects on ryrosme kmase and bolo6rcal acbvdres and on recombrnauon in tissue cunure with the cellular srr genc Mol Cell. Brol. S.?789-2795 6 Cross- F. R., and H. Hanafusa. 198? Local mutagcnesis of Rous sarcoma vrms% the major sues of tyrosme and senne phosphorv ytauon are dispensable for transformauon. Cell 3d:597-607. 7. Dnlry, G- Q., J. Mcf.nughlin. O. N. Witte. and D. Baltimore. 1987. The C M1.-specrfic P210 hrr/rrbf protein. unlike v-abL does ~ not nansform NIHr3TJ fibroblaets. Science i17:sq?-535. 8. Dlckson, C., R. Eisenman, and H. Fan. 1985. Protein blosynthe. srs and assembly. p 13S-146 In R. Wens. N- Teich. H Varmus.andJ Comnted.I.RNAtumorvlnues2.Supplemems and appendnes Cold SpnnR Harbor laboratory. Cold SpnnR Harbor. N V 9. Drubin. D. G.. J. Mulholland. Z. 2hu. and D. Batstrin- 1990- Humulogy of a veasr aenn-brndmg ptotem to signal lransducrinn proterm and mrnsm-l. Nature (f.rrodon) 343:288-290 10 Emori, V., 1'. Homma, H. Sorimarhi, H. KawaSakL O. Naka- nishi. K. Suzuki. and T. Takanaaa. 1969- A second lypc of rat phosphomosiUdnspecrfic phospholmase C containing a srr. relared sequence not essential for phosphomovude. hydrolyzing acuvay. J Bml Chem. 264:21885-21890 ] I. Foster. D. A., M. Shibu.a, and H. Hsnafusa. )98s. Activation of the transformmg potenual of the cellular JPs gene. Cell 42: ]O5. 115 12. Hamaguch/. M.. C. Orandori. and H. Hanafusn. 1989 . Phosphor ytauon of cellular protems in Rous sercoma vrrusanfected eells: analysis by u.e of anlrphosphotyrosme antibodres. Mol Cell. Brol. 8:3035..}0,f7 13 Hamaguchi, M., and H. Hunafusa. 1989 l.ocalrzauon of major pNenual substrate% of p6o' - k:naee in the plasma membrane matru fracuon. Oncogene Hes 4:29-37 14 Hannfusa, H. 1969. Rapid transformation of cells by Rous sarcoma virus Proc Natl Acad. Sn. USA 63:318-i2c. 15 . 1bu, H., F. R. Crosa. T. Hanafuaa-and H. Hanafuss. 1984. Rous sarcoma vuus vanants Ihal carry the cellular +rr gene instead of the vital rrr gene cannot transform chicken embryofibroblasts Proc Natl Acad. Sn. USA 81:4424-.-r428 16 Kurnbluth, S., F. R. ('ross. M. Harbison, and H. Hanafusa. ]986 Transformetron of chicken embryo fibroblasts and mnror induction by the middle T anhgen of polynmavrcus camed in an avran reuovrrus vector. Mol. Cell. Bml 6:1545_1551- • 17 Let1nsun, A. D.. S. A. ('ourtne{dge. and J. M. Bishop. 19R1 Suucrural and functional domains of Ihe Rous sarcoma vlm+ uansformmg protein Ipp6u"'). Proc Nail Acad. Set . (JSA PUBLICATIONS 016623 10335654