Download - An oligomeric form of simian virus 40 large T-antigen is immunologically related to the cellular tumor antigen p53

Vol. 50, No. 2JOURNAL OF VIROLOGY, May 1984, p. 457-4640022-538X/84/050457-08$02.00/0Copyright © 1984, American Society for Microbiology

An Oligomeric Form of Simian Virus 40 Large T-Antigen IsImmunologically Related to the Cellular Tumor Antigen p53

KEITH N. LEPPARDt AND LIONEL V. CRAWFORD*

Laboratory of Molecular Virology, Imperial Cancer Research Fund, Lincoln's Inn Fields, London WC2A 3PX, UnitedKingdom

Received 26 October 1983/Accepted 25 January 1984

The cellular tumor antigen p53 is bound to the simian virus 40 (SV40) large T-antigen in SV40-infected and-transformed cells. As a result, p53 can in general be immunoprecipitated by either monoclonal orpolyclonal antibodies that react with large T-antigen. Despite extensive immunological characterization ofboth these antigens, they have not previously been found to share any antigenic determinants. We haveisolated several monoclonal antibodies that bind to the human p53 protein (K. Leppard and L. V. Crawford,EMBO J. 2:1457-1464, 1983) and show here that antibody PAb11O4 has distinct, intrinsic activities towardsboth p53 and SV40 large T-antigen. Only a subset of T-antigen is bound by PAb1104. This subset is anoligomeric form of T-antigen, as judged by its sedimentation velocity in sucrose. In contrast, all of thedetectable p53 carries the PAb11O4-reactive determinant. The detection of a chance cross-reactivitybetween two antigens that are already well characterized and which associate with one another in vivo ishighly unlikely. It is possible therefore that the element of structural similarity between large T and p53 thatis implied by our results has some genuine functional significance.

Simian virus 40 (SV40) expresses two proteins, large T-and small t-antigens, from the early half of its genome (4, 15).Antibodies to both these proteins are produced by animalsbearing SV40-induced tumors (5, 15, 39, 44). Such antiseraimmunoprecipitate large T- and small t-antigens from lysatesof metabolically labeled SV40-transformed cells (15, 45). Inaddition, such immunoprecipitates contain a specificallyprecipitated protein of 45,000 to 55,000 (45k to 55k) molecu-lar weight, the precise size of which depends on cell type (7,9, 22, 37). This protein is now termed p53 and is known to becellular in origin (33, 42, 43). It is generally present inimmunoprecipitates of SV40-transformed cell extracts be-cause of its physical association with large T-antigen (28, 33,36). However, some anti-SV40 tumor sera, as well as someantisera from animals bearing tumors of diverse origin andfrom human cancer patients, contain intrinsic anti-p53 anti-bodies (17, 19, 28, 33). It is still not known why p53 becomesimmunogenic in a tumor-bearing animal, but since manyrodent and human tumor cell lines and tumors containelevated levels of p53 as compared with levels in controlcells and tissues (3, 14, 18, 40), it is possible that the normalprocess of self-tolerance is broken down by exposure to p53released from necrotic tumor tissue. Alternatively, p53 fromtumors and transformed cells may be structurally distinctfrom that found in normal cells, causing it to be recognizedas foreign by the immune system of the animal.

After the initial detection of the SV40 tumor antigens andp53 with antitumor sera, a large number of monoclonalantibodies have been raised to these proteins (10, 20, 23, 24,29, 32, 34, 41; L. Gooding, J. Pipas, and E. Harlow,manuscript in preparation). By use of these antibodies, atleast 14 distinct antigenic determinants on large T-antigen(24; Gooding et al., in preparation; E. Harlow, personalcommunication), 4 determinants on rodent p53 (3), and 3determinants on human p53 (32) have been defined. Howev-

* Corresponding author.t Present address: Department of Microbiology, Health Sciences

Center, State University of New York, Stony Brook, NY 11794.

er, despite this very extensive immunological characteriza-tion of SV40 large T and p53, there has previously been noevidence of any immunological relationship between thesetwo proteins. We report here the properties of a monoclonalantibody, PAb11O4, raised against human p53 (32) that bindsto both human p53 and SV40 large T-antigen, each in theabsence of the other. The PAbl1O4-reactive large T is asubset of the total non-p53-associated T-antigen in bothSV40-infected, permissive and SV40-transformed, nonper-missive cells. This subset migrates on a sucrose gradientwith a sedimentation coefficient of ca. 16S. Our results implythat p53 shares an element of structure with an oligomericform of SV40 large T-antigen.

MATERIALS AND METHODSCells, virus, and antibodies. The virus and cell lines used,

and the conditions employed for cell culture, have beenpreviously described (32). Antibodies of the PAb400 serieswere isolated by Harlow et al. (24), antibody PAb122 wasisolated by Gurney et al. (23), and all were kindly providedby Ed Harlow. Antibodies of the PAb200 series were isolat-ed and provided by David Lane (10, 29); those of the PAb600series were isolated and provided by Linda Gooding (Good-ing et al., in preparation). Antibody RA3-2C2 was isolatedand provided by Robert Coffman (12). Antibodies PAb122,PAb410, PAb421, PAb6O7, and RA3-2C2 have anti-p53 ac-tivity, restricted in the latter two cases to mouse p53. Otherantibodies of the PAb200, PAb400, and PAb600 series haveactivity towards SV40 large T-antigen. The isolation ofPAb11O4 has been previously described (32). Antibodieswere either diluted from 10-mg/ml stocks in phosphate-buffered saline, with the amounts quoted in micrograms ofprotein, or were used in the form of tissue culture superna-tant from hybridoma cells, with the amounts quoted inmicroliters. Tissue culture, fluid generally contains antibodyat around a 30-,ug/ml concentration.

Cell labeling, immunoprecipitation, and gel electrophoresis.Cells growing in monolayers were labeled just before conflu-ence either on 90-mm NUNC dishes in 2.5 ml of Dulbecco

457

458 LEPPARD AND CRAWFORD

modified Eagle medium (E4) lacking phosphate with 1.0mCi of 32p, (carrier-free) (Amersham International) for 3 h,or in 1.5 ml of E4 lacking methionine with 1.0 mCi of[35S]methionine (>1,000 Ci/mmol) (Amersham Internation-al) for 3 h, or on 50-mm NUNC dishes in 1.0 ml of E4 lackingleucine with 0.5 mCi of [4,5-3H]leucine (130 to 190 Ci/mmol)(Amersham International) for 2 h. Hybridoma cells were

cultured in suspension in E4 medium lacking methionine at adensity of 5 x 106 cells per ml for 3 h in the presence of 0.5mCi of [35S]methionine per ml to label secreted antibodymolecules. Lysates of metabolically labeled cells were pre-

pared, labeled proteins were immunoprecipitated, and pre-

cipitated proteins were separated by sodium dodecyl sulfate(SDS)-polyacrylamide gel electrophoresis (27), all as previ-ously described (32). Lysis buffer for preparation of cellextracts was 1% Nonidet P-40-150 mM NaCI-50 mM Tris(pH 8.0). NET gel buffer was 150 mM NaCl-5 mM EDTA-50mM Tris (pH 7.4)-0.05% Nonidet P-40-0.02% NaN3-0.25%gelatin. A 10% suspension of Staphylococcus aureus Cowan1 (26) in NET gel buffer was 10% SAC. The proteinsemployed as molecular weight markers for SDS gel electro-

32P-SV80

| precleared preclea red cC of largeT of P53

Antibody PAb: 419 421 1104 419 421 1104 419 421:~~~~~~ezf ~~~~~~~~~rk. I2eO

Large T-

p53 =_ 4W

a b c d

-__ -"Large T

_p53

e f g h

FIG. 1. SV80 cells were labeled in vivo with 32Pi, and lysate wasimmunoprecipitated either directly (lanes a and h) or after preclear-ance of all large T-antigen (lanes b, c, and d) or p53 (lanes e, f, andg). Lysate was cleared of large T by diluting it with 9 volumes of 10%sucrose in lysis buffer and a further 10 volumes of NET gel bufferand leaving it at +10°C overnight to allow the T-p53 complex todissociate. Large T was then removed by the addition, three times,of 15 j±g of anti-T antibody PAb419 and the pellet from 0.5 ml of 10%SAC. Fresh lysate was cleared of p53 by similar additions of anti-p53 antibody, PAb421, and S. aureus cells. Aliquots of lysatederived from 106 cells were precipitated with 1.5 ,ug of PAb419(lanes a, d, and g), 1.5 ,ug of PAb421 (lanes b, e, and h), or 5 ,ul ofPAb11O4 (lanes c and f), and precipitated proteins were analyzed byelectrophoresis through a 10% polyacrylamide-SDS gel and detectedby overnight exposure to Kodak SB-5 film at room temperature.Protein molecular weight markers of 92k and 53k migrated to thepositions arrowed on the right of the diagram.

phoresis were "C-labeled phosphorylase b (92k), glutamatedehydrogenase (53k), carbonic anhydrase (30k), soybeantrypsin inhibitor (21k), cytochrome c (12.5k), and aprotinin(6.5k). 3H-labeled proteins were visualized by fluorography(6), and other labeled proteins were visualized by autoradi-ography of dried gels.

Sucrose gradient analysis. Labeled proteins were in somecases separated by velocity sedimentation through sucrosebefore immunoprecipitation. A 0.3-ml portion of labeled celllysate was layered over a 4.4-ml, linear, 5 to 20% sucrosegradient, formed in lysis buffer containing 1 mM dithiothrei-tol on a 0.3-ml cushion of 60% sucrose. Gradients were thenresolved by centrifugation in a Beckman SW55 rotor for 6 hat 55k rpm at +6°C. Gradients were collected in ca. 0.2-mlfractions, which were then divided as required for parallelimmunoprecipitation. Parallel gradients were calibrated withmarker proteins of sedimentation coefficients 26S (glutamatedehydrogenase), 13.5S (phosphorylase a), and 7S (immuno-globulin G).Radioimmunoassay. Procedures for determining the class

of antibody PAb11O4, and the relationship of the determinantit recognizes to others already characterized, were as de-scribed previously (13, 24, 32). Solution binding assays wereperformed as described by Harlow et al. (24), modified fromLane and Robbins (31).

RESULTSPAb11O4 recognizes both p53 and T-antigen. The isolation

of hybridoma PAb11O4 has already been described in detail(32). The splenocyte parent of this hybridoma was derivedfrom a BALB/c mouse immunized with p53 from the humanSV40-transformed cell line SV80 that was purified by im-munoprecipitation and SDS gel electrophoresis. PAb11O4was cloned by using the ability of its secreted antibody toimmunoprecipitate the large T-p53 complex from 32P-labeledSV80 cell lysate as a screen. It was found to produce animmunoglobulin M antibody that did not bind to S. aureusprotein A.To determine which of the two known components of the

SV80 T-p53 complex was recognized by PAb11O4, immuno-precipitation reactions were performed with SV80 lysatesthat had been precleared of either T-antigen or p53. Asshown in Fig. 1 (lane c), PAb11O4 immunoprecipitated thesame doublet of p53 as was precipitated by PAb421 (lane b)from SV80 lysate cleared of large T under conditions whereno indirect precipitation of p53 via large T and an anti-large-T antibody, PAb419, was observed (lane d). Equally,PAb11O4 (lane f) precipitated the same large-T speciesprecipitated by PAb419 (lane g) from SV80 lysate cleared ofp53, when no precipitation ofT via p53 and PAb421 was seen(lane e). It appeared, therefore, that the antibody secreted byhybridoma PAb11O4 had distinct anti-large-T and anti-p53activities.

Single antibody has anti-T and anti-p53 activities. Duringthe isolation of hybridoma PAb11O4, cells were cloned threetimes, until they were monoclonal. However, to exclude theremote possibility that the dual activity of PAb11O4 resultedfrom the secretion of a mixed population of antibodies by thehybridoma cells, we metabolically labeled PAb11O4 with[35S]methionine and determined which labeled species werecapable of binding to large T and p53 (Fig. 2). Since PAb11O4does not bind to protein A, it can only be precipitated on S.aureus cells indirectly, either via large T and a protein A-binding anti-T antibody, PAb416, or via p53 and a protein A-binding anti-p53 antibody, PAb421. The labeled immuno-

J. VIROL.

SHARED DETERMINANT ON SV40 LARGE T AND p53 459

35 S-PAb11

p53/T/odT od(p53

Ig9 -to"

Ig light -

a b c d e f

FIG. 2. Secreted antibody from PAbllbeled with [35S]methionine, and aliquots wof anti-T antibodies and then anti-p53 antilor vice versa (lanes g through 1). Large T-lysate of 2 x 107 CV1 monkey cells per ml,7.5 PFU of SV40 small-plaque strain SV-Sof all p53 by three successive additions ofpellet from 0.5 ml of 10% SAC before us(was a lysate of 2 x 107 C331 cells peicarcinoma cells (2). Each of lanes a thrcomprised the addition at 30-min intervPAb11O4 supernatant, diluted to 0.5 ml witEof cell extract, 1.5 ,ug of antibody, and 20and j omitted the addition of cell extract, athe addition of both cell extract and antibodused large T and PAb416; lanes f through

Immunoprecipitated proteins were separaamide-SDS gel and detected by exposure ffilm at room temperature. Proteins of kmigrated as indicated on the right of the d

globulin subunits precipitated inbehaved identically on SDS-polyacrlanes a and g). These same aliquotsfurther precipitated with either largcaureus (lanes b through e) or p53, Pi(lanes h through k) to remove all antib4of binding to large T and p53, respectiprecipitated with the alternative comband S. aureus in lane f and large T, PAlane 1). It was found that no anti-p53 adepletion of the anti-T activity in theand that only a very small amount of arafter anti-p53 depletion, probably reflpletion in the first phase. The distinct a

and anti-p53 activities in PAb11O4 !reside in the same population of antibPAbl1O4 recognition of a particular

antigen. We were interested to determpopulations of large T and p53 in infecells shared an immunological determ

04 was an attribute of only a subset of one or both proteins.Proteins in lysates prepared from SV-S-infected CV1 cells,labeled at 48 h postinfection with 32p;, were separated byfractionation on 5 to 20% sucrose gradients. Each fraction

p53/dp5dt from two gradients run in parallel was divided into two equalp53/c]p5d~parts, creating four sets of fractions, which were immuno-precipitated with PAb11O4 (Fig. 3A), PAb421 (B), PAb416(C), and PAb4O2 (D). Each panel shows, from left to right,the precipitation profile from bottom to top of the gradient.

-92 The association of monkey p53 with SV40 T-antigen isNE known to be very weak (25). As a result, any T-p53 complex

initially present in the extracts analyzed here had dissociated-53 before the immunoprecipitation of gradient fractions, shown

by the failure of anti-p53 antibody PAb421 to achieve anysignificant precipitation of T-antigen (Fig. 3B). The observedprecipitation of T-antigen by PAb11O4 from fractions of aparallel gradient (Fig. 3A) must therefore be attributed to thedirect anti-T activity of this antibody. Antibodies PAb416and PAb4O2 recognize determinants in the N- and C-terminalhalves of T-antigen, respectively (24). The latter recognizesthe p53-associated form of T-antigen poorly if at all (Harlow,personal communication), whereas the former recognizes allforms of T-antigen so far defined. Comparison of the T-g h ; j k I antigen precipitated by PAb11O4 with that brought down by

L04 was metabolically la- either of these well-characterized anti-T antibodies shows,ere sequentially depleted that PAb11O4 recognizes only a discrete subset of T-antigenbodies (lanes a through f) which migrates with a velocity of ca. 16S relative to marker-containing extract was a proteins. Identical experiments with SV80 cell lysate, pre-at 48 h postinfection with cleared of p53 with PAb421 before sucrose gradient analysis,per cell, that was cleared showed that PAb11O4 recognized a similarly restricted size15 g of PAb421 and the class of T-antigen in these cells (data not shown).e. p53-containing extract The size class specificity of PAb11O4 anti-T activity couldrml of human cervicalrough c and f through in principle result from the antibody having a low affinity forals to 50

and

of labeled T-antigen, which would favor the precipitation of oligomersh NET gel buffer of100,u rather than monomers. However, a more interesting, and,u1 of 10% SAC. Lanes d equally plausible, explanation of the data would be that T-ind lanes e and k omitted antigen oligomerization is necessary to create the PAb11O4ly. Lanes a through d and determinant, by either a direct contribution from two orij used p53 and PAb421. more monomers or allosteric changes consequent on oligo-Lted on a 10% polyacryl- merization. It is also possible that the PAbl1O4-reactive T-cow days to KodakSB-w antigen is associated with other proteins in a heterooligomer,nown molecular weight but no such proteins have been detected by metaboliciagram. labeling with 32p;, [35S]methionine, or [3H]leucine. The

PAbl1O4-reactive material is, therefore, likely to be a ho-mooligomer of SV40 large T-antigen.PAbllO4 binds the same size class of p53 as does PAb421.

these two reactions Preparation of an SV80 cell extract in which T-p53 complexylamide gels (Fig. 2, has been dissociated and all T-antigen removed depends onof PAb11O4 were then dilution of the initial extract so as to make its use in sucrosee T, PAb416, and S. gradient analysis impossible. To study the size distributionAb421, and S. aureus of PAbl1O4-reactive p53, we therefore employed a humanody molecules capable cervical carcinoma cell line, C33I (2), which has been shownively, and were finally to express high levels of p53 (3). C331 cells were labeled with)ination (p53, PAb421, 32p;, and proteins in a cell lysate were separated by velocity,b416, and S. aureus in sedimentation through sucrose. Each gradient fraction wasLctivity remained after divided into two parts, and each set was immunoprecipitat-PAb11O4 supernatant ed, one with PAb11O4 and the other with PAb421 (Fig. 4Aiti-T activity remained and B). This analysis revealed that the p53 recognized bylecting incomplete de- these two antibodies was indistinguishable by the criterion ofnd independent anti-T sucrose gradient mobility. 32P-labeled p53s derived fromsupernatant therefore similar gels were also indistinguishable when either their V8)ody molecules. protease partial digestion products (11) or their trypticsize class of large T- phosphopeptides were compared (data not shown). It ap-

ine whether the entire pears, therefore, that PAb11O4 recognizes the same spec-cted and transformed trum of p53 molecules as is recognized by PAb421, althoughinant, or whether this it recognizes only a subset of large T-antigen.

VOL. 50, 1984


PAb11O4 is highly specific for large T and p53. It waspossible that PAb11O4 was a widely cross-reactive antibody,recognizing a very simple determinant, and that the appar-ently specific immunoprecipitation of large T and p53 byPAb11O4 was due to the highly selective labeling of proteinsachieved with 32p;. In this regard, we had already noted thespecific precipitation by PAb11O4 of phosphoproteins ofmolecular weight 60k and 55k (60K and 55K phosphopro-teins) (Fig. 3A and 4A). Titration experiments suggested thatthe affinity of PAb11O4 for these cellular proteins, whichseemed from their comigration on sucrose gradients to beassociated, was considerably lower than for either large T orp53, since precipitation of the 60K-55K complex was onlyobserved at high antibody inputs; no similarity between p53,60K, and 55K could be detected by V8 protease partialdigestion analysis. To test the possibility that PAb11O4 mightbe widely cross-reactive, SV80 cells were labeled with[3H]leucine, which should label many more proteins than

would 32p;, and extracts were immunoprecipitated withPAb11O4 and control antibodies (Fig. 5). RA3-2C2 is anegative control in this system since it does not bind tohuman p53. The only proteins precipitated by PAb11O4 andnot by RA3-2C2 (apart from large T and p53) were, asindicated, a very high-molecular-weight protein and proteinsof 21.5k and 20k. The 21.5k protein was also precipitated byPAb419, which additionally precipitated a 14k protein. RA3-2C2 precipitated a protein of ca. 6k. There were no qualita-tive differences in the immunoprecipitation observed whenlabeling was carried out for 30 min, with or without priordepletion of leucine pools for 30 min. The 60K-SSK phospho-protein complex, which cross-reacts with PAb11O4 as notedabove, was not precipitated in this experiment, probablybecause inputs of PAb11O4 were below the amounts requiredfor its detection. Although one cannot exclude the possibilitythat other specifically precipitated proteins might have beendetected under different labeling conditions, we believe

C.

A.

large T-

PA b 416

PAb 1104

Z4N46S*a I'.

60K/55K =

Bottom

B.

Top Bottom

D.PAb421

Top

PAb 402

es_I - -largeT

p53 -

Bottom Top Bottom Top

FIG. 3. Proteins in lysate from SV40-infected CV1 cells, labeled at 48 h postinfection with 32p,, were separated by velocity sedimentationthrough a 5 to 20% sucrose gradient, and gradient fractions were immunoprecipitated with 20 ,u1 of PAb11O4 (A), 20 ,u1 of PAb421 (B), 20 p.1 ofPAb416 (C), or 20 ,u1 of PAb402 (D). Each panel displays proteins derived from about 5 x 106 infected cells. Precipitated proteins wereseparated on 10% polyacrylamide-SDS gels and detected by exposure for 3 days to Kodak SB-5 film at room temperature. Protein molecularweight markers of 92k and 53k migrated to the positions arrowed in each panel.

. Irv large T

J. VIROL.


from metabolically labeled SV80 cells by immunoprecipita-tion, gel electrophoresis, and passive elution and used theseas antigens in a solution binding assay (24, 31). The positivecontrol antibodies rebound 75 and 36% of the large-T andp53 probes, respectively (Fig. 6). In parallel experiments,PAb11O4 bound only 5% of large T and 17% of p53, whereasnegative control antibodies bound only 5% of each probe.Thus the PAb11O4 determinant on large T was completelydestroyed by SDS denaturation, as one would expect sincePAb11O4 binds only to an oligomeric form of large T. Incontrast, a proportion of the p53 probe was rebound byPAb11O4, suggesting that the relevant structure elements inthis protein can at least partly refold under the assayconditions employed.

DISCUSSIONWe have presented evidence that a monoclonal antibody,

PAb11O4, has dual reactivity towards p53 and an oligomericform of SV40 large T-antigen, suggesting that these antigenshave some element of structure in common. There are,however, at least two alternative explanations of the datathat, although less likely, cannot yet be excluded. First, it is

- 3L

le -S 8

7% gel-p53

High M.W. -- -__ --

protein

Bottom Top

FIG. 4. Proteins in lysates of C331 cells labeled with 32Pi were

separated by velocity sedimentation through a 5 to 20% sucrose

gradient, and gradient fractions were immunoprecipitated with 20 ,u1of PAb11O4 (A) or 20 ,u1 of PAb421 (B). Precipitated proteins were

separated on 10% polyacrylamide-SDS gels and detected by expo-

sure for 5 days to preflashed Fuji RX film at -700C with fasttungstate screens. Protein molecular weight markers of 92k and 53kmigrated to the positions arrowed in each panel. Positive andnegative control immunoprecipitations of unfractionated lysate withantibodies PAb421 (421) and PAb419 (419) are shown at the right ofeach panel.

large T--

._ Um

p53=43 as

12% gel

O a" /large

_wo as am_ p/53

.,21.5

-20 5

- 14 0

-60

a b c d e

these data indicate that PAb11O4 is not widely cross-reactiveand that its dual reactivity for large T and p53 shows that a

significant element of structural similarity exists betweenthese two proteins.A characteristic of all the anti-T monoclonal antibodies

that have so far been shown to cross-react with cellularproteins is that they recognize determinants that surviveSDS denaturation (16), judged by their ability to rebind[35S]methionine-labeled antigen after elution from an SDS-polyacrylamide gel. To test the ability of PAb11O4 to binddenatured antigen, we prepared 35S-labeled large T and p53

f g h i j

FIG. 5. SV80 cells were labeled with [3H]leucine for 2 h, andextracts were prepared. Aliquots were immunoprecipitated with 20

of PAb421 (lanes a and f), PAb122 (lanes b and g), RA3-2C2 (lanesc and h), PAb419 (lanes e and j), or 10 p.l of PAb11O4 (lanes d and i).After resuspension in sample buffer, each precipitate was dividedinto two parts, and one part was separated on a 7% polyacrylamide-SDS gel (lanes a through e), and the other was separated on a 12%polyacrylamide-SDS gel (lanes f through j). Precipitated proteinswere detected by fluorography (6), PPO (2,5-diphenyloxazole)-impregnated gels being exposed for 3 days to preflashed Fuji RXfilm at -70°C. The positions to which protein molecular weightmarkers migrated are indicated by arrows, 92k and 53k on the left,and 92k, 53k, 30k, 21k, 12.5k, and 6.5k on the right of the diagram.

32P- C331

A. PAb 1104

5

..

TopBottom

B. PAb 421VW, -.

VOL. 50, 1984


60

bound

T-antigen,0/ of 40

input

20-

50 100 150

antibody input/ pI

40r

bound

p53,0/0 of

input

50 100 150antibody input/pi

FIG. 6. [35S]methionine-labeled lysate from 5was immunoprecipitated with 200 ,u of PAb42PAb419, and proteins were separated by SDS gelGel slices containing large T and p53 were excisedM ammonium bicarbonate-0.05% SDS at 25°C fwere filtered to remove gel fragments, and 5-,ul sa

2,200 cpm of large T or 1,500 cpm of p53 were allcvarying inputs of antibodies PAb421, PAb11O4, ar

,ul of NET gel buffer overnight at +4°C. Rabbit antglobulin antiserum (0.5 pJ) was added to each reactcollecting immune complexes with 20 ,ul of 10,pellets were counted in 1.0 ml of Aquasol (NewCorp.).

known that rapidly sedimenting T-antigen cawith p53. The evidence that the PAb11O4-reais not a T-p53 complex rests on the fact thatform can be precipitated under conditions whreactive with anti-p53 antibody PAb421 remaiimmunoreactivity with PAb421 defines p53, texcluded that this material is only a subset cthat PAb11O4 defines a wider subset. HowevcPAb11O4-reactive, PAb421-unreactive p53 cbeled either with phosphate or with leucine,min and chasing for up to 20 h. We therefexplanation as unlikely. Second, it is possiban oligomeric subset of T-antigen have in comto bind to some nonprotein component ofligand or macromolecule. The available datplained by PAb11O4 having direct activitycellular component, although this could no

very high levels in cells, otherwise the precand T-antigen would require much higher inpthan of genuine anti-T or anti-p53 antibodiesMonoclonal antibodies frequently show vai

cross-reactivity to other proteins that havetionship to the principle antigen in any oth4

Among examples of this phenomenon are a number ofantibodies that react with SV40 large T-antigen and variouscellular proteins (16, 29). However, in no case have antibod-ies reacting with demonstrably different determinants onlarge T cross-reacted with the same cellular protein. Al-though not excluding the possibility that one or more ofthese cross-reactivities reflected a genuine functional simi-larity between the two proteins, it was argued (16) that withantibodies such as these, which recognize denaturation-resistant determinants that probably comprise segments ofthe primary amino acid chain, an appreciable frequency of

200 250 cross-reaction with essentially unrelated proteins should beexpected.Two features of PAb1104 distinguish it from the general

case of cross-reactive monoclonal antibodies. First, sincePAb11O4 can only recognize an oligomer of T-antigen, thedeterminant involved cannot consist solely of a portion ofthe primary amino acid chain. This conclusion is supportedby the observed sensitivity of the determinant to SDSdenaturation. Second, whereas other cross-reactive anti-Tantibodies react with cellular proteins that were not definedin advance, PAb1104 binds to two already characterizedproteins that are known to associate in vivo. Clearly theprobability of obtaining an anti-T antibody that cross-reactswith any particular cellular protein is vanishingly low unless

200 250 that antibody is widely cross-reactive. Our data show thatPAb11O4 is not widely cross-reactive. Apart from p53 andlarge T, it binds to only one or two cellular proteins, at least

x 107 SV80 cells one of which, the 60K-55K phosphoprotein complex, is1 and 200 ,ul of bound only at high antibody inputs. For these reasons, weI electrophoresis. believe that PAb11O4 defines a genuine element of similarityIand eluted in 0.1 between the secondary and tertiary structures of p53 andFor 48 h. Eluates oligomeric T-antigen.mples containing The significance of the similarity between p53 and oligo-

nd PAb416 in 100 meric T-antigen depends on its extent. Certainly, an anti-i-mouse immuno- body-combining site (epitope) covers only a small area on

ion 30 mimnbefore the surface of a protein (1), and since no other anti-T% SAC. Washed antibody with properties similar to PAb11O4 has been de-England Nuclear scribed, we would not expect the extent of the similarity to

be very great. Nevertheless, the fact that the PAb11O4determinant on large T is apparently formed by an associa-tion of large T monomers gives us some confidence that thisrelationship between T and p53 may have functional signifi-

,n be associated cance. Protein monomers often oligomerize via interactionsctive T-antigen between homologous regions on their surfaces, and it isthis oligomeric possible that such a mechanism is responsible for large T-p53iere no material association. Clearly, if the PAb11O4 determinant lies in theins. In practice, region of association between T and p53, then such regionsbut it cannot be must also remain exposed on some of the monomers of the)f total p53 and T-p53 complex to allow precipitation by the antibody. Stud-er, this putative ies on the maturation of the T-p53 complex in wild-typeould not be la- SV40 and tsA58-transformed mouse cells suggest that as-labeling for 30 sembly of a large-T oligomer may be an essential intermedi-ore regard this ate step in generating the T-p53 complex (8, 21), whereasile that p53 and studies of in vitro complex formation with the purified T-imon the ability antigen from adeno-SV40 hybrid virus Ad2+D2 show thatthe cell, either p53 already in an oligomeric form, probably a tetramer, canLa could be ex- bind to the T-antigen produced by Ad2+D2 virus (35),against such a although in this case p53 was thought to be able to bind tot be present at monomeric D2 protein. There are, therefore, some groundsipitation of p53 for expecting the p53 combining site to be present only on'uts of PAb11O4 oligomeric T-antigen and for some unoccupied combining

sites on p53 to be present in the complex.rying degrees of Since the PAb11O4 determinant on T-antigen is not formedno known rela- solely by a segment of the polypeptide chain, the scope forer respect (30). mapping this determinant on the T-antigen molecule is

J. VIROL.

8r-8o


probably limited. Preliminary data suggest that PAb11O4 canbind to D2 protein, which lacks the N-terminal 82 aminoacids of large T, whereas, in blocking studies with anti-Tantibodies of the PAb200 series (10, 29), the PAb400 series(24), and the PAb600 series (Gooding et al., in preparation)with T-antigen from a lysate of SV-S-infected CV1 cells asantigen, only one antibody, PAb427, showed blocking ofPAb11O4 binding. The binding site of this antibody wasmapped to the C-terminal portion of large T by its ability tobind the Ad2+ ND1 28K hybrid protein (24). Unfortunately,the PAb427 hybridoma line has since been lost, so thatfurther experiments with it are not possible. Partial blockingof PAb11O4 binding to C331 p53 by the anti-p53 antibodyPAb11O3 (32) was also observed.

It is clearly important to define as far as possible thebinding sites for PAb11O4 on large T and p53 and todetermine whether or not this antibody defines the region ofassociation between these two proteins. An important steptowards this goal will be the definition, by mutant analysis,of the region of T-antigen that is required for p53 binding.Such information can then be correlated with the bindingdata for PAb11O4 and other antibodies that react with largeT. The recent isolation of a large number of both deletionand point mutations in large T-antigen (38; K. Peden,personal communication) brings the definition of the p53-binding site on T-antigen that much closer. The combinedapplication of deletion mutant and monoclonal antibodyanalyses will provide the best approach to determining thefunctional significance of the large T-p53 complex in theprocess of SV40-mediated transformation.

ACKNOWLEDGMENTSWe thank Kit Osborn for her help with tissue culture and Ed

Harlow, David Lane, and Linda Gooding for their generous gifts ofmonoclonal antibodies. We are indebted to Margaret Hollis for herhelp in preparing this manuscript and to Audrey Symons for herwork on the figures.

LITERATURE CITED1. Atassi, M. Z. 1975. Antigenic structure of myoglobin: the

complete immunochemical anatomy of a protein and conclu-sions relating to antigenic structures of proteins. Immunochem-istry 12:423-438.

2. Auersperg, N. 1964. Long-term cultivation of hypodiploid hu-man tumor cells. J. Natl. Cancer Inst. 32:135-163.

3. Benchimol, S., D. C. Pim, and L. V. Crawford. 1982. Radio-immunoassay of the cellular protein p53 in mouse and humancell lines. EMBO J. 1:1055-1062.

4. Berk, A., and P. A. Sharp. 1978. Spliced early mRNA's ofsimian virus 40. Proc. Natl. Acad. Sci. U.S.A. 75:1274-1278.

5. Black, P., W. P. Rowe, H. C. Turner, and R. J. Huehner. 1%3. Aspecific complement-fixing antigen present in SV40 tumor andtransformed cells. Proc. Natl. Acad. Sci. U.S.A. 50:1148-1156.

6. Bonner, W. M., and R. A. Laskey. 1974. Detection method fortritium-labeled proteins and nucleic acids in polyacrylamidegels. Eur. J. Biochem. 46:83-88.

7. Carroll, R. B., S. M. Goldfine, and J. A. Melero. 1978. Antiser-um to polyacrylamide gel-purified simian virus 40 T antigen.Virology 87:194-198.

8. Carroll, R. B., and E. G. Gurney. 1982. Time-dependent matu-ration of the simian virus 40 large T antigen-p53 complex studiedby using monoclonal antibodies. J. Virol. 44:565-573.

9. Chang, C., D. T. Simmons, M. A. Martin, and P. T. Mora. 1979.Identification and partial characterization of new antigens fromsimian virus 40-transformed mouse cells. J. Virol. 31:463-471.

10. Clark, R., D. P. Lane, and R. Tjian. 1981. Use of monoclonalantibodies as probes of simian virus 40 T antigen ATPaseactivity. J. Biol. Chem. 256:11854-11858.

11. Cleveland, D. W., S. G. Fischer, M. W. Kirschner, and U. K.Laemmli. 1977. Peptide mapping by limited proteolysis in sodi-um dodecyl sulfate and analysis by gel electrophoresis. J. Biol.Chem. 252:1102-1106.

12. Coffman, R. L., and I. L. Weissman. 1981. A monoclonalantibody that recognizes B cells and B cell precursors in mice. J.Exp. Med. 153:269-279.

13. Crawford, L. V. 1982. The origins of p53 in relation to transfor-mation. Adv. Viral. Oncol. 2:3-21.

14. Crawford, L. V. 1983. The 53,000 dalton cellular protein and itsrole in transformation. Int. Rev. Exp. Pathol. 25:1-50.

15. Crawford, L. V., C. N. Cole, A. E. Smith, E. Paucha, P.Tegtmeyer, K. Rundell, and P. Berg. 1978. Organization andexpression of early genes of simian virus 40. Proc. Natl. Acad.Sci. U.S.A. 75:117-121.

16. Crawford, L. V., K. Leppard, D. Lane, and E. Harlow. 1982.Cellular proteins reactive with monoclonal antibodies directedagainst simian virus 40 T-antigen. J. Virol. 42:612-620.

17. Crawford, L. V., D. C. Pim, and R. D. Bulbrook. 1982.Detection of antibodies against the cellular protein p53 in serafrom patients with breast cancer. Int. J. Cancer 30:403-408.

18. Crawford, L. V., D. C. Pim, E. G. Gurney, P. Goodfellow, and J.Taylor-Papadimitriou. 1981. Detection of a common feature inseveral human tumor cell lines-a 53,000-dalton protein. Proc.Natl. Acad. Sci. U.S.A. 78:41-45.

19. DeLeo, A. B., G. Jay, E. Appella, G. C. Dubois, L. W. Law, andL. J. Old. 1979. Detection of a transformation-related antigen inchemically induced sarcomas and other transformed cells of themouse. Proc. Natl. Acad. Sci. U.S.A. 76:2420-2424.

20. Dippold, W. G., G. Jay, A. B. DeLeo, G. Khoury, and L. J. Old.1981. p53 transformation-related protein: detection by monoclo-nal antibody in mouse and human cells. Proc. Natl. Acad. Sci.U.S.A. 78:1695-1699.

21. Greenspan, D. S., and R. B. Carroll. 1981. Complex of simianvirus 40 large tumor antigen and 48,000-dalton host tumorantigen. Proc. Natl. Acad. Sci. U.S.A. 78:105-109.

22. Griffin, J. D., S. Light, and D. M. Livingston. 1978. Measure-ments of the molecular size of the simian virus 40 large Tantigen. J. Virol. 27:218-226.

23. Gurney, E. G., R. 0. Harrison, and J. Fenno. 1980. Monoclonalantibodies against simian virus 40 T antigens: evidence fordistinct subclasses of large T antigen and for similarities amongnonviral T antigens. J. Virol. 34:752-763.

24. Harlow, E., L. V. Crawford, D. C. Pim, and N. M. Williamson.1981. Monoclonal antibodies specific for simian virus 40 tumorantigens. J. Virol. 39:861-869.

25. Harlow, E., D. C. Pim, and L. V. Crawford. 1981. Complex ofsimian virus 40 large-T antigen and host 53,000-molecular-weight protein in monkey cells. J. Virol. 37:564-573.

26. Kessler, S. W. 1975. Rapid isolation of antigens from cells with astaphylococcal protein A-antibody adsorbent: parameters of theinteraction of antibody-antigen complexes with protein A. J.Immunol. 115:1617-1624.

27. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

28. Lane, D. P., and L. V. Crawford. 1979. T-antigen is bound to ahost protein in SV40-transformed cells. Nature (London)278:261-263.

29. Lane, D. P., and W. K. Hoeffier. 1980. SV40 large T shares anantigenic determinant with a cellular protein of molecularweight 68,000. Nature (London) 288:167-170.

30. Lane, D. P., and H. Koprowski. 1982. Molecular recognition andthe future of monoclonal antibodies. Nature (London) 296:200-202.

31. Lane, D. P., and A. K. Robbins. 1978. An immunochemicalinvestigation of SV40 T antigens. I. Production properties andspecificity of a rabbit antibody to purified simian virus 40 large-T antigen. Virology 87:182-193.

32. Leppard, K., and L. V. Crawford. 1983. Monoclonal antibodiesdisplaying a novel species specificity for the primate transfor-mation-related protein, p53. EMBO J. 2:1457-1464.

33. Linzer, D. I. H., and A. J. Levine. 1979. Characterization of a

VOL. 50, 1984


54K dalton cellular SV40 tumor antigen present in SV40-transformed cells and uninfected embryonal carcinoma cells.Cell 17:43-52.

34. Martinis, J., and C. M. Croce. 1978. Somatic cell hybridsproducing antibodies specific for the tumor antigen of simianvirus 40. Proc. Natl. Acad. Sci. U.S.A. 75:2320-2323.

35. McCormick, F., R. Clark, E. Harlow, and R. Tjian. 1981. SV40T antigen binds specifically to a cellular 53K protein in vitroNature (London) 292:63-65.

36. McCormick, F., and E. Harlow. 1980. Association of a murine53,000-dalton phosphoprotein with simian virus 40 large-T anti-gen in transformed cells. J. Virol. 34:213-224.

37. Melero, J. A., D. T. Stitt, W. F. Mangel, and R. B. Carroll. 1979.Identification of new polypeptide species (48-55K) immunopre-cipitable by antiserum to purified large T antigen and present inSV40-infected and -transformed cells. Virology 93:466-480.

38. Pipas, J. M., K. W. C. Peden, and D. Nathans. 1983. Mutationalanalysis of simian virus 40 T antigen: isolation and characteriza-tion of mutants with deletions in the T-antigen gene. Mol. Cell.Biol. 3:203-213.

39. Prives, C., E. Gilboa, M. Revel, and E. Winocour. 1977. Cell-

free translation of simian virus 40 early messenger RNA codingfor viral T-antigen. Proc. Natl. Acad. Sci. U.S.A. 74:457-461.

40. Rotter, V. 1983. p53, a transformation-related cellular-encodedprotein, can be used as a biochemical marker for the detectionof primary mouse tumor cells. Proc. Natl. Acad. Sci. U.S.A.80:2613-2617.

41. Rotter, V., 0. N. Witte, R. Coffman, and D. Baltimore. 1980.Abelson murine leukemia virus-induced tumors elicit antibodiesagainst a host cell protein, P50. J. Virol. 36:547-555.

42. Simmons, D. T., M. A. Martin, P. T. Mora, and C. Chang. 1980.Relationship among Tau antigens isolated from various lines ofsimian virus 40-transformed cells. J. Virol. 34:650-657.

43. Smith, A. E., R. Smith, and E. Paucha. 1979. Characterization ofdifferent tumor antigens present in cells transformed by simianvirus 40. Cell 18:335-346.

44. Tegtmeyer, P. 1974. Altered patterns of protein synthesis ininfection by SV40 mutants. Cold Spring Harbor Symp. Quant.Biol. 39:9-15.

45. Tegtmeyer, P., M. Schwartz, J. K. Collins, and K. Rundell. 1975.Regulation of tumor antigen synthesis by simian virus 40 geneA. J. Virol. 16:168-178.

J. VIROL.