Relationship Among Tau Antigens Isolated from Various Lines of Simian Virus 40-Transformed Cells

In addition to the virus-specified tumor antigens, simian virus 40-transformed cells contain at least one other protein which can be immunoprecipitated with serum from animals bearing simian virus 40-induced tumors. This protein, which is designated Tau antigen, has an apparent molecular weight of 56,000 as determined by electrophoresis on acrylamide gels. The relationship among Tau antigens isolated from different lines of simian virus 40-transformed cells was examined by comparing the methionine-labeled tryptic peptides of these proteins by two-dimensional fingerprinting on thin-layer cellulose plates. In this fashion, we initially determined that the Tau antigens isolated from three different lines of transformed mouse cells were very similar. Second, we found that Tau antigen isolated from a line of rat transformants was closely related, but not identical, to the mouse cell Tau antigens. Approximately 70% of their methionine peptides comigrated in two dimensions. Finally, we showed that Tau antigen isolated from a line of transformed human cells was only partially related to the mouse and rat proteins. About 40% of the methionine peptides of the human protein were also contained in the Tau antigens from the other two species. These results strongly indicate that the Tau antigens isolated from these various simian virus 40-transformed cell lines contain common amino acid sequences.     

Demonstration of Infectious Deoxyribonucleic Acid in Transformed Cells I. Recovery of Simian Virus 40 from Yielder and Nonyielder Transformed Cells

Deoxyribonucleic acid (DNA) was extracted from virus-free simian virus 40 (SV40)-transformed hamster, mouse, and monkey cells and was inoculated into simian cells in the presence of diethylaminoethyl (DEAE)-dextran; infectious SV40 was recovered by using DNA from cell lines which fail to yield virus by the fusion technique as well as from cell lines which readily yield virus by fusion. The rescued virus was identified as SV40 by three methods: (i) neutralization of plaque formation by specific antiserum; (ii) induction of synthesis of viral-specific antigens detected by immunofluorescence; and (iii) presence of papovavirus particles seen by the electron microscope. Treatment of the transformed cell DNA with deoxyribonuclease or omission of the DEAE-dextran prevented the rescue of virus. Large amounts of transformed cell DNA were required (>10 mug/culture of 10(6) cells) to effect rescue of SV40 by passage through monkey cells. A linear response was obtained between the input of DNA with inocula between 10 and 45 mug of DNA/culture and the yield of SV40 recovered. Biological activity was demonstrable irregularly when the transformed cell DNA was assayed directly in the presence of DEAE-dextran. The DNA induced plaque formation in about 50% of the trials as well as the synthesis of SV40 tumor and viral antigens in rare simian cells. The infectious DNA appeared to be associated with cellular DNA. The infectivity was found in the pellet of precipitated DNA obtained by the Hirt technique and was inactivated by boiling for 15 min. These properties are characteristic of linear cellular DNA and not of free, circular SV40 DNA.     

Quantitation of a 55K cellular protein: similar amount and instability in normal and malignant mouse cells.

Quantitative expression of a specific 55,000 (55K)-molecular-weight cellular protein was studied in two groups of mouse embryo fibroblast (clonal) cells originating from two parent clones, one of which possessed high tumorigenicity and the other of which possessed very low tumorigenicity. From the clone with low tumorigenicity, tumor lines and clones were obtained by selecting rare spontaneously transformed highly tumorigenic (mutant) cells. Cells were labeled during exponential growth for 3 h at 37 degrees C, with [35S]methionine, and the cellular 55K protein was immunoprecipitated with a monoclonal antibody and quantitated. There were low and approximately equal amounts of 55K protein in cells (clones) with both low and high tumorigenicity from both groups of cells, and there was no correlation at all between quantitative expression of 55K protein and of cellular tumorigenicity. There was approximately 10- to 20-fold more 55K protein in all simian virus 40-transformed T antigen-positive derivative clones, as shown previously. The T antigen-negative revertant tumor lines and clones obtained by an immunological in vivo selection method had low amounts of 55K protein, similar to the parent cell before simian virus 40 transformation. In all of the T antigen-negative cells, including the highly tumorigenic cells, degradation (turnover?) of the 55K protein was rapid, and a half-life of 15 to 60 min was estimated from pulse-chase experiments. In all of the T antigen-positive cells the 55K protein was stable (half-life greater than 10 h). In primary cells established from the tumors induced by highly tumorigenic cells there was a very low or no detectable amount of the 55K protein. This is in contrast to the primary cells obtained from early murine embryos in which we have reported high amounts of (stable) 55K proteins.     

Density dependent inhibition of both growth and T-antigen expression in revertants isolated from simian virus 40-transformed mouse SVT2 cells.

Phenotypic revertants were isolated from simian virus 40-transformed cells in order to examine the relationship between simian virus 40 T-antigen expression and G1 arrest of growth. Revertant clones with increased adherence were selected from cultures of SVT2, a simian virus 40-transformed BALB/c mouse cell line, and screened to find arrestable revertant clones which inhibited DNA synthesis when crowded. The clones selected from untreated SVT2 were unstable and showed little or no inhibition of DNA synthesis when crowded. Stable revertants were found after treatment of SVT2 with Colcemid to increase ploidy. The stable revertants all lost most transformed growth properties tested, including tumorigenicity, but only a few showed the same degree of inhibition of DNA synthesis at high cell density as BALB/3T3. All revertant clones expressed T antigen at low cell density. Three revertants showed coordinate inhibition of DNA synthesis and apparent loss of T antigen at high cell density. We suggest that changes in gene dosage rather than mutations caused the altered properties of the new revertants and that continued DNA synthesis in confluent cultures may be the transformed phenotype that requires the least simian virus 40 T antigen.     

Demonstration of T antigens on the surface of cells transformed with primate polyoma viruses

Primate polyoma virus-transformed hamster, mouse, and rat cell lines were examined by indirect immunofluorescence staining for cell surface-associated T antigens, by using a rabbit antiserum prepared against sodium dodecyl sulfate-denatured large T antigen of simian virus 40 (anti-SV40-SDS-T serum). Positive surface staining was shown not only on SV40-transformed cells, but also on BK and JC virus-transformed cells. In contrast, normal cells and cells transformed with mouse polyoma-, human adeno-, and murine sarcoma viruses were negative. The data on SV40-transformed cells confirmed the reports of others demonstrating the cell surface location of SV40 large T antigen, and the data on BK and JC virus-transformed cells proved that these cells have cell-surface T antigens that cross-react with anti-SV40-SDS-T serum.     

Characterization of tau antigens isolated from uninfected and simian virus 40-infected monkey cells and papovavirus-transformed cells.

Tau antigens (also known as cellular or nonviral tumor antigens) were detected in uninfected and simian virus 40-infected monkey cells after immunoprecipitation with serum from hamsters bearing simian virus 40-induced tumours (anti-T serum). These two proteins (56,000 daltons) were digested to similarly sized peptides with various amounts of Staphylococcus aureus V8 protease. The Tau antigen isolated from infected monkey cells was closely related but was not identical to the corresponding protein from human cells transformed by simian virus 40, as determined by two-dimensional mapping of their methionine-labeled tryptic peptides. Hamster cells transformed by various primate papovaviruses (simian virus 40, BK virus, and JC virus) synthesized indistinguishable Tau antigens, as determined by two-dimensional peptide mapping. When tested by the same procedure, these proteins and the ones made in monkey and human cells were found to be related to the Tau antigens isolated from simian virus 40-transformed mouse and rat cells. Based on these results, an "evolutionary tree" was constructed to show the relationship among the methionine-containing tryptic peptides of all of these proteins.     

Surface T-antigen expression in simian virus 40-transformed mouse cells: correlation with cell growth rate.

Cell growth control appears to be drastically altered as a consequence of transformation. Because the cell surface appears to have a role in modulating cell growth and simian virus 40 (SV40)-transformed cells express large T antigen (T-Ag) in the plasma membrane, we investigated whether surface T-Ag expression varies according to cell growth rate. Different growth states were obtained by various combinations of seeding density, serum concentration, and temperature, and cell cycle distributions were determined by flow microcytofluorometry. Actively dividing SV40-transformed mouse cell cultures were consistently found to express higher levels of surface T-Ag and T-Ag/p53 complex than cultures in which cells were mostly resting. In addition, the T-Ag/p53 complex disappeared from the surface of tsA7-transformed cells cultured under restrictive conditions known to induce complete growth arrest (39.5 degrees C), although the surface complex did not disappear from other tsA transformants able to keep cycling at 39.5 degrees C. These results suggest that surface SV40 T-Ag or surface T-Ag/p53 complex, or both, are involved in determining the growth characteristics of SV40-transformed cells.     

Radioimmunoassay of the cellular protein p53 in mouse and human cell lines

We have developed quantitative radioimmunological solid phase assays for the host protein p53 from mouse cells and from human cells. The first assay, for mouse p53, depends on having two monoclonal antibodies reacting with different determinants on the p53 molecule. With this assay we have shown that SV40-transformed cells have approximately 100-fold more p53 than untransformed mouse cells and that other transformed cells have intermediate levels. Embryonal carcinoma cell lines have approximately 50-fold less p53 than SV40-transformed cells. This is in contrast to the high levels of incorporation of [35S]methionine into p53 in these cells and indicates that metabolic labelling is not a valid approach for measuring p53 levels. The second assay, for human p53, required a different approach and made use of the anti-p53 antibodies detected in the sera of some breast cancer patients. Human tumour cell lines contained amounts of p53 varying from the high level seen in SV40-transformed human fibroblasts down to less than one hundredth of this amount. Normal human cells showed low levels of p53. The data confirm that many, but not all, human tumour cell lines contain more p53 than normal cells.     

Identification and partial characterization of new antigens from simian virus 40-transformed mouse cells.

Two new species of antigens were detected in simian virus 40-transformed mouse cells, in addition to the large (94,000 daltons) and small (20,000 daltons) tumor antigens. These antigens were immunoprecipitated from cell extracts by using anti-T serum and not normal, nonimmune serum. One of these was a protein with a molecular weight of approximately 130,000 and was present in some but not all SV40-transformed mouse cells. The other, which we have named Tau antigen, has a molecular weight of 56,000 as estimated by electrophoresis through acrylamide gels and was found in all virus-transformed cells examined. The 13,000-daltons antigen contained about 15 methionine-tryptic peptides which were also present in the large SV40 tumor antigen as determined by ion-exchange chromatography. This strongly suggested that the protein was virus coded. The 56,000-dalton Tau antigen appeared to share only two methionine-tryptic peptides with the large species of SV40 tumor antigen, as determined by ion-exchange and paper chromatographies. Our results are compatible with a cellular origin for Tau antigen. However, our data do not exclude the possibility that this protein contains sequences specified by the virus DNA.     

Virogenic Properties of Bromodeoxyuridine-sensitive and Bromodeoxyuridine-resistant Simian Virus 40-transformed Mouse Kidney Cells

When simian virus 40 (SV40)-transformed mouse kidney cells (mKS) were grown in the presence of susceptible indicator cells, SV40 was readily recovered from: (i) 15 transformed cell lines, (ii) transformed cells subcultured 45 times over a 7-month period in medium containing antiviral serum and bromodeoxyuridine (dBU), (iii) 45 of 46 clonal lines isolated in the presence of antiviral serum, (iv) 19 of 19 secondary clones isolated from two clonal lines, and (v) dBU-resistant transformed cell lines. dBU-resistant SV40-transformed mouse kidney cell lines were selected and shown to contain the T antigen and to have normal levels of thymidylate kinase and deoxyribonucleic acid (DNA) polymerase, but to be deficient in thymidine (dT) kinase. Radioautographic and biochemical experiments demonstrated that very little (3)H-dT was incorporated into DNA of dBU-resistant cells during a 6-hr labeling period. After infection of dT kinase-deficient mKS cells with vaccinia virus, high levels of dT kinase were induced. The properties of SV40 recovered from dBU-sensitive and dBU-resistant cells were studied. SV40 recovered from transformed cells was shown to express in CV-1 cells at least six functions characteristic of parental virus: synthesis of capsid antigen, synthesis of T antigen, synthesis of viral DNA, induction of dT kinase, induction of DNA polymerase, and induction of host cell DNA synthesis. In addition, SV40 recovered from the transformed cells induced T antigen, dT kinase, deoxycytidylate deaminase, thymidylate kinase, and DNA polymerase in abortively infected mouse kidney cultures, and the virus was also capable of transforming primary cultures of mouse kidney cells.     

Simian virus 40 large T antigen and p53 are microtubule-associated proteins in transformed cells.

The cellular proteins that interact with simian virus 40 large T antigen (T-ag) must be identified in order to understand T-ag effects on cellular growth control mechanisms. A protein extraction procedure utilizing single-phase concentrations of 1-butanol recovered a complex composed of T-ag, p53, and other Mr 35,000-60,000 proteins from suspension cultures of the simian virus 40-transformed mouse cell line mKSA. Partial protease mapping showed each of the associated proteins to be unique. Automated microsequence analysis of the NH2-terminal 30 amino acids of the Mr 56,000 protein purified after coprecipitating with T-ag and p53 identified it as the beta subunit of mouse tubulin. The existence of a complex containing tubulin, T-ag, and p53 was confirmed by reciprocal immunoblotting experiments. Both T-ag and p53 were coprecipitated by three different monoclonal antibodies directed against tubulin, and conversely, monoclonal antibodies specific for T-ag or p53 coprecipitated tubulin. Mixing experiments and extractions in the presence of purified tubulin indicated that the complex existed in situ prior to cell lysis. Both p53 and T-ag copurified with microtubules through two cycles of temperature-dependent disassembly and assembly. Both T-ag and p53 were localized to microtubules in the cytoplasm of mKSA cells by immunoelectron microscopy. Treatment of mKSA cells with 10 microM colchicine followed by lysis in 0.1% Nonidet P-40 resulted in increased amounts of solubilized T-ag and p53. Both T-ag and p53 were also associated with microtubules in three other simian virus 40-transformed mouse cell lines growing as monolayers, confirming the generality of the association. An interaction of T-ag and p53 with microtubules may be important in the intracellular transport of these proteins and may affect cellular signal transduction or growth control.     

Subcellular Localization of Simian Virus 40 Large Tumor Antigen

The distribution of simian virus 40 large tumor antigen in subcellular fractions from simian virus 40-transformed hamster (H-50) and mouse (VLM) cells and from simian virus 40-infected monkey cells was determined. Solubilized [(35)S]-methionine- or (32)P(i)-labeled surface membrane and nuclear fractions were prepared, immunoprecipitated with hamster anti-T serum, and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Tumor antigen with an apparent molecular weight of approximately 96,000 was detected in both subcellular fractions. Minor components of approximately 68,000 and approximately 56,000 with anti-T reactivity which labeled with [(35)S]methionine were also detected in both fractions from H-50 cells, as were components of approximately 140,000 and approximately 56,000 from VLM cells. The 56,000 component appeared to be greatly reduced in (32)P(i)-labeled surface membrane fractions. Normal cells or cells transformed with a heterologous agent, such as polyoma virus or a chemical carcinogen, lacked immunoprecipitable tumor antigen. Cell fractionation was monitored by [(3)H]thymidine labeling, NADH-diaphorase activity, and Na(+)-K(+)-dependent ATPase activity. These analyses revealed only trace contamination of surface membranes by nuclei, extremely low levels of nuclear rupture during homogenization, and an approximate 10-fold enrichment of surface membrane. Reconstruction experiments demonstrated that soluble tumor antigen failed to associate or copurify with surface membranes during fractionation procedures. These results indicate the presence of a protein in the plasma membrane of cells transformed or infected by simian virus 40 that is immunologically indistinguishable from nuclear tumor antigen.     

Rearrangement of integrated viral DNA sequences in mouse cells transformed by simian virus 40.

The organization of viral DNA sequences in several cell lines derived from a primary colony of simian virus 40 (SV40)-transformed mouse cells was analyzed to examine the origin of the various distinctive patterns of SV40 sequence arrangement present in transformed cells. This analysis revealed a complex arrangement of viral sequences in the uncloned transformed cells but simplified arrangements in cloned derivatives of the primary transformant. The cell lines studied had certain SV40 sequence arrangements in common, but the cloned lines had lost some parental arrangements and acquired new arrangements. These results indicate that the arrangement of viral sequences in some SV40-transformed cells is not fixed but that alterations occur after integration, creating a heterogeneous population of transformants. In the process, expression of viral genes may be altered. Possible causes for and implications of this genetic instability are discussed.     

Establishment of rat pancreatic endocrine cell lines by infection with simian virus 40.

The feasibility of infection and transformation by SV40 (simian virus 40) of primary cell cultures derived from newborn-rat pancreas was investigated. As judged by the presence of intranuclear SV40 T-antigen, exposure to the virus resulted specifically in infection and transformation of epithelioid (predominantly endocrine) cells. The transformed cells were subcultured (more than 64 passages) and cloned. Culture medium and acid/ethanol extracts of the cells did not contain detectable amounts of immunoreactive insulin after the third subculture. However, inoculation of such SV40-transformed pancreatic cells into immunodeficient rats results in tumours in which insulin production was partially restored through the passage in vivo, since the tumour cells contained and synthesized small amounts of immunoreactive insulin which co-migrated with an insulin marker on gel chromatography. Interestingly, the transformed cells maintained under tissue-culture conditions produced a protein immunologically related to insulin, soluble in aqueous buffer but insoluble in acid/ethanol. This 3000-dalton protein is too large to be a translation product of the rat preproinsulin 9S mRNA. SV40-transformed pancreatic cells might prove useful in the investigation of the factors controlling and maintaining insulin biosynthesis.     

Thymidine Kinase from Normal, Simian Virus 40-Transformed and Simian Virus 40-Lytically Infected Cells

Simian virus 40 (SV40) infection of human diploid cells failed to cause an enhanced production of thymidine kinase during the first 10 days after infection. Thymidine kinase activities from extracts of SV40-transformed cultures (human or simian) were considerably higher than the activity levels in extracts from the normal cells of origin. In addition, whereas the kinase activities obtained for human diploid cultures decreased as the cell sheet became confluent, the kinase activities for SV40-transformed human cells remained high after confluence was reached. Antisera obtained from hamsters bearing SV40 or adeno-7-SV40 hybrid virus tumors selectively inhibited enzyme from transformed sources (human or simian). Also, the antisera selectively inhibited enzyme extracted from SV40-lytically infected monkey cells. Sera from normal animals or from hamsters bearing polyoma tumors failed to inhibit enzymes from normal, SV40-transformed, or SV40-lytically infected cells. The Michaelis constant of partially purified enzyme from SV40-transformed cells was two to five times as high as that obtained for partially purified enzyme from human diploid cell cultures.     

Genetic and biochemical analysis of transformation-competent, replication-defective simian virus 40 large T antigen mutants.

To study the role of the biochemical and physiological activities of simian virus 40 (SV40) large T antigen in the lytic and transformation processes, we have analyzed DNA replication-defective, transformation-competent T-antigen mutants. Here we describe two such mutants, C8/SV40 and T22/SV40, and also summarize the properties of all of the mutants in this collection. C8/SV40 and T22/SV40 were isolated from C8 and T22 cells (simian cell lines transformed with UV-irradiated SV40). Early regions encoding the defective T antigens were cloned into a plasmid vector to generate pC8 and pT22. The mutations responsible for the defects in viral DNA replication were localized by marker rescue, and subsequent DNA sequencing revealed missense and one nonsense mutation. The T22 mutation predicts a change of histidine to glutamine at residue 203. C8 has two mutations, one predicts lysine224 to glutamamic acid and the other changes the codon for glutamic acid660 to a stop codon; therefore, C8 T antigen lacks the 49 carboxy-terminal amino acids. pC8A and pC8B were constructed to contain the C8 mutations separately. Plasmids pT22, pC8, pC8A, and pC8B were able to transform primary rodent cell cultures. T22 T antigen is defective in binding to the SV40 origin. C8B (49-amino-acid truncation) is a host-range mutant defective in a late function in CV-1 but not BSC cells. Analysis of T antigens in mutant SV40-transformed mouse cells suggests that the replicative function of T antigen is important in generating SV40 DNA rearrangements that allow the expression of "100K" variant T antigens in the transformants.     

Fibronectin gene expression in proliferating, quiescent, and SV40-infected mouse kidney cells

To study alterations in cellular gene expression in mouse kidney cell cultures infected with simian virus 40 (SV40) or polyomavirus, we performed a differential screening of a mouse kidney cDNA library with probes prepared from mRNAs of virus-infected and mock-infected cells. We isolated and characterized cDNA recombinant pKT13 which detected increased mRNA levels in infected cells. Sequence analysis of pKT13 revealed close to 100% homology with the 3'-end of mouse fibronectin (FN) mRNA. Since primary cultures of baby mouse kidney cells have been extensively characterized in our laboratories, we studied FN gene expression at different stages of uninfected and virus-infected cultures. High levels of FN and of its mRNA were found in the kidneys of suckling mice, while in primary cultures of proliferating epithelial kidney cells the expression of FN was very low until the cultures became confluent. Thereafter FN increased and reached high levels in cells which were irreversibly arrested in phase Go and which had apparently exhausted their finite division potential. Infection of confluent cultures with polyomavirus or SV40 resulted in a further stimulation of FN gene expression. However, during abortive infection with SV40, FN mRNA and FN levels decreased with emergence of transformed cells and were low in an established SV40-transformed mouse kidney cell line. These changes in FN gene expression suggest that high levels of FN might be indicative in vivo for terminal differentiation and in vitro for cellular senescence.     

Search for mammalian type-C retrovirus polypeptides on infected animal cells and human tumor cell lines using interspecies-reacting antibodies

Cells producing endogenous and exogenous type-C retroviruses of murine, feline and primate origin were evaluated for expression of those virus-specific cell-surface antigens which cross-react with antibodies to interspecies determinants of mammalian type-C viral polypeptides. Surface polypeptides of cells replicating endogenous and exogenous type-C retroviruses were iodinated by the lactoperoxidase method. Labelled antigens were immunoprecipitated and analyzed in polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. This method detected gp71 and less frequently p15E/p12E on the surface of virus-producing cells; in addition, p27 was found on F422 cells replicating feline leukemia virus. Antibodies to the membrane protein p15E displayed the broadest cross-reactivity but only antibodies to gp71 mediated complement-dependent interspecies cell lysis. The pattern of cross-reactivity reflected known genetic relationships among these mammalian viruses. Although antiserum to the simian sarcoma virus complex (SSV) was strongly cytotoxic to some human tumor cell lines, this reactivity could not be attributed to antibodies cross-reacting with SSV gp71.     

Antigenic and Immunogenic Characteristics of Nuclear and Membrane-Associated Simian Virus 40 Tumor Antigen

Antisera were prepared in syngeneic hosts against subcellular fractions of simian virus 40 (SV40)-transformed cells (MoalphaPM, MoalphaNuc), glutaraldehydefixed SV40-transformed cells (HaalphaH-50-G, MoalphaVLM-G), and electrophoretically purified denatured SV40 tumor antigen (T-ag) (RaalphaT). Immune sera were also collected from animals bearing tumors induced by SV40-transformed cells (HaalphaT, MoalphaT, HAF) and from SV40-immunized animals that had rejected a transplant of SV40-transformed cells (HaalphaS, MoalphaS). Immunological reagents prepared against cell surface (MoalphaPM, HaalphaS, MoalphaS, HaalphaH-50-G, MoalphaVLM-G) reacted exclusively with the surface of SV40-transformed cells by indirect immunofluorescence or protein A surface antigen radioimmunoassay. Immunological reagents prepared against the nuclear fraction (MoalphaNuc) or whole-cell determinants (HaalphaT, MoalphaT, HAF, RaalphaT) reacted with both the nuclei and surface of SV40-transformed or -infected cells. All reagents were capable of immunoprecipitating 96,000-molecular weight large T-ag from solubilized whole cell extracts of SV40-transformed cells. The exclusive surface reactivity of HaalphaS exhibited in immunofluorescence tests was abolished by solubilization of subcellular fractions, which then allowed immunoprecipitation of T-ag by HaalphaS from both nuclear and plasma membrane preparations. Specificity was established by the fact that all T-reactive reagents failed to react in serological tests against chemically transformed mouse cells, and sera from mice bearing transplants chemically transformed mouse cells (MoalphaDMBA-2) failed to react with SV40-transformed mouse or hamster cells. Reagents demonstrating positive surface immunofluorescence and protein A radioimmunoassay reactions against SV40-transformed cells were capable of blocking the surface binding of RaalphaT to SV40-transformed cells in a double-antibody surface antigen radioimmunoassay. This blocking ability demonstrated directly that a component specificity of each surface-reactive reagent is directed against SV40 T-ag. A model is presented which postulates that the differential detection of T-ag by the various serological reagents is a reflection of immunogenic and antigenic differences between T-ag polypeptides localized in nuclei and plasma membranes.