Mutants of polyomavirus middle-T antigen

Polyomavirus middle-T antigen induces the transformation of established cell lines in culture and is known to interact with and/or modulate the activity of several enzymes (pp60^c.src, protein kinase C and phosphatidylinositol kinase) in vitro. This review is a compilation of the reported mutants of middle-T antigen and their biochemical and biological properties as they relate to the transformation event. The mutants of polyomavirus middle-T antigen have been previously classified phenotypically. Given the now large number of mutants, the classification presented here is based upon the position within the molecule. A model of middle-T is presented in which the protein is considered as consisting of three domains: a hydrophobic domain (the putative membrane-binding domain), the amino-terminal half of the molecule (the putative pp60^c.src-binding domain) and the intervening amino acids (the putative modulatory domain). A current model for the induction of transformation by polyomavirus middle-T is presented.     

Mapping of the amino-terminal half of polyomavirus middle-T antigen indicates that this region is the binding domain for pp60c-src.

The majority of the carboxy-terminal half of polyomavirus middle-T antigen has been variously mutated and, with the exception of the putative membrane-binding domain (amino acids 394 to 415), was found to be largely dispensible for the transforming activity of the protein. A comparison of the small-T antigen amino acid sequences (equivalent to the region of middle-T encoded by exon 1) of simian virus 40, BK virus, polyomavirus, and a recently described hamster papovavirus highlighted regions of potential interest in mapping functions to the amino-terminal half of polyomavirus middle-T antigen. The regions of interest include amino acids 168 to 191 (previously investigated by this group [S. H. Cheng, W. Markland, A. F. Markham, and A. E. Smith, EMBO J. 5:325-334, 1986]), two cysteine-rich clusters (amino acids 120 to 125 and 148 to 153), and amino acids 92 to 117 (within the limits of the previously described hr-t mutant, SD15). Point mutations, multiple point mutations, and deletions were made by site-specific and site-directed mutagenesis within the cysteine-rich clusters and residues 92 to 117. Studies of the transforming ability of the altered middle-T species demonstrated that this activity is highly sensitive to amino acid changes. All four regions (as defined above) within the amino-terminal half of middle-T have now been studied in detail. The phenotype of the mutants is predominantly transformation defective, and the corresponding variant middle-T species are characterized by being either totally or severely handicapped in the ability to associate actively with pp60c-src. Whether the mutations affect the regions of interaction between middle-T and pp60c-src or simply interfere with the overall conformation of this domain is not known. However, there would appear to be a conformational constraint on this portion of the molecule with regard to its interaction with pp60c-src and by extension to the ability of the middle-T species to transform.     

Two-Dimensional Polyacrylamide Gel Analysis of Plodia interpunctella Granulosis Virus

Polyomavirus middle-T antigen contains a contiguous sequence of 22 hydrophobic amino acids near the carboxyl terminus, which is the putative membrane-binding domain of the protein. The DNA encoding this region was mutated to form a series of deletions, insertions, and substitutions called RX mutants. The phenotypes of these mutants fall into three groups based on the transforming and biochemical properties of their encoded proteins. The first group, with deletions outside but proximal to the hydrophobic domain, displayed an essentially wild-type phenotype. A second group, with extensive deletions within the region encoding the hydrophobic domain, expressed middle-T species which did not fractionate with cellular membranes or associate with pp60c-src and which were defective in their ability to transform. A third group of mutants with more subtle predicted alterations in the hydrophobic domain were wild type for the biochemical parameters investigated but were unable to transform cultured rodent cells. These observations are consistent with previous findings that membrane association plays an important role in transformation by middle-T and that, whereas association between middle-T and pp60c-src is a necessary correlate of transformation, it is not sufficient. A comparison of murine polyomavirus middle-T and a newly described hamster papovavirus putative middle-T revealed a strong homology between their respective hydrophobic-domain amino acid sequences. This homology is not observed in the anchorage domains of other model proteins, and this may imply that the middle-T hydrophobic domain is important in transformation for reasons other than simple membrane association.     

The complex of polyoma virus middle-T antigen and pp60c-src.

We have recently proposed that the transforming protein of polyoma virus, middle-T antigen, forms a complex with pp60c-src. Here we provide additional evidence for the existence of the complex using both monoclonal antibodies specific for middle-T and antibodies raised against synthetic peptides corresponding to sequences from both middle-T and pp60c-src. The complex was retained during partial purification of middle-T and was stable to incubation under various conditions. A survey of a number of mutants of middle-T antigen showed that there was a complete correlation between the ability of middle-T to accept phosphate in the in vitro kinase reaction and the presence of a middle-T: pp60c-src complex. This result is in accord with our hypothesis that middle-T itself is not a protein kinase but rather that pp60c-src phosphorylates middle-T. All mutant forms of middle-T antigen capable of transformation had associated pp60c-src. The middle-T of two non-transforming mutants (hr-t mutants) did not have associated pp60c-src, whereas other non-transforming middle-T species did associate with pp60c-src. We propose that the complex plays an essential role in transformation by polyoma virus, but that the existence of the complex per se may not be sufficient.     

Myristylation of pp60c-src is not required for complex formation with polyomavirus middle-T antigen.

Middle-T antigen (middle-T), the transforming gene product of polyomavirus, associates with several cellular tyrosine kinases, such as pp60c-src. Complex formation leads to kinase activation and is essential for cell transformation. Middle-T-associated as well as uncomplexed pp60c-src is predominantly found in the plasma membrane. We transfected mouse 3T3 fibroblasts with a mutated c-src gene (2Ac-src), allowing the expression of a protein containing alanine instead of glycine in position 2 of the primary translation product. Contrary to the wild-type c-src gene product, pp60c-src(2A) was not myristylated and accumulated in the cytoplasm instead of being transferred to cellular membranes. The mutant protein was able to associate with middle-T and was activated similarly to the wild-type c-src gene product. Both wild-type and 2A mutant protein were membrane associated upon complex formation with middle-T. This finding suggests that the putative carboxy-terminal membrane anchor sequence of middle-T is sufficient to hold middle-T-associated pp60c-src(2A) in the plasma membrane.     

Stoichiometry of cellular and viral components in the polyomavirus middle-T antigen-tyrosine kinase complex.

Our results indicate that only one type of tyrosine kinase is present within each middle-T antigen-tyrosine kinase complex, suggesting that middle-T antigen forms separate complexes with different tyrosine kinases. Furthermore, we determined that there is only one molecule of middle-T antigen within any one of these complexes. We interpret this to mean that in any given cell, polyomavirus transformation involves, at least in part, the simultaneous deregulation of a number of separate pathways controlling cellular proliferation. Finally, we also demonstrate that the separate middle-T:pp60c-src and middle-T:pp59c-fyn complexes are each able to interact with the same cellular p81/85-kDa phosphoprotein, a possible component of the phosphatidylinositol kinase.     

Peptide antibodies to the human c-fyn gene product demonstrate pp59c-fyn is capable of complex formation with the middle-T antigen of polyomavirus.

The c-fyn proto-oncogene is a member of a family of closely related genes of which c-src is the prototype. Using peptide antibodies which had been raised against sequences predicted to be specific for the human c-fyn gene product, the c-fyn protein was identified. It is a tyrosine kinase with apparent mol. wt of 59 kd that is also phosphorylated and myristylated. Like pp60c-src and pp62c-yes, pp59c-fyn is able to form a stable complex with middle-T antigen, the transforming protein of polyomavirus. The transformation-defective middle-T mutant NG59, which is unable to associate stably with pp60c-src does not associate with pp59c-fyn. In contrast to pp60c-src, complex formation with middle-T antigen does not lead to a significant increase in the tyrosine kinase activity of pp59c-fyn. These findings lead us to suggest that middle-T mediated transformation may be a consequence of the deregulation of several members of the src-family of protein tyrosine kinases.     

Multimerization of polyomavirus middle-T antigen.

The oncogenic protein of polyomavirus, middle-T antigen, associated with cell membranes and interacts with a variety of cellular proteins involved in mitogenic signalling. Middle-T antigen may therefore mimic the function of cellular tyrosine kinase growth factor receptors, like the platelet-derived growth factor or epidermal growth factor receptor. Growth factor receptor signalling is initiated upon the binding of a ligand to the extracellular domain of the receptor. This results in activation of the intracellular tyrosine kinase domain of the receptor, followed by receptor phosphorylation, presumably as a consequence of dimerization of two receptor molecules. Similar to middle-T antigen, phosphorylation of growth factor receptors leads to recruitment of cellular signalling molecules downstream in the signalling cascade. In this study, we investigated whether middle-T antigen, similar to tyrosine kinase growth factor receptors, is able to form dimeric signalling complexes. We found that association with cellular membranes was a prerequisite for multimerization, most likely dimer formation. A chimeric middle-T antigen carrying the membrane-targeting sequence of the vesicular stomatitis virus G protein instead of the authentic polyomavirus sequence still dimerized. However, mutants of middle-T antigen unable to associate with 14-3-3 proteins, like d18 and S257A, did not form dimers but were still oncogenic. This indicates that both membrane association and binding of 14-3-3 are necessary for dimer formation of middle-T antigen but that only the former is essential for cell transformation.     

The carboxy terminus of pp60c-src is a regulatory domain and is involved in complex formation with the middle-T antigen of polyomavirus.

A large number of mutations were introduced into the carboxy-terminal domain of pp60c-src. The level of phosphorylation on Tyr-416 and Tyr-527, the transforming activity (as measured by focus formation on NIH 3T3 cells), kinase activity, and the ability of the mutant pp60c-src to associate with the middle-T antigen of polyomavirus were examined. The results indicate that Tyr-527 is a major carboxy-terminal element responsible for regulating pp60c-src in vivo. A good but not perfect correlation exists between lack of phosphorylation at Tyr-527 and increased phosphorylation at Tyr-416, between elevated phosphorylation on Tyr-416 and activated kinase activity, and between activated kinase activity and transforming activity. Phosphorylation of Tyr-527 was insensitive to the mutation of adjacent residues, indicating that the primary sequence only has a minor role in recognition by kinases or phosphatases which regulate it in vivo. Three mutants which have in common a modified Glu-524 residue were phosphorylated on Tyr-416 and Tyr-527 and were weakly transforming. This suggests that other mechanisms besides complete dephosphorylation of Tyr-527 can lead to increased phosphorylation of Tyr-416 and activation of the transforming activity of pp60c-src. Furthermore, the residues between Asp-518 and Pro-525 were required to form a stable complex with middle-T antigen. The proximity of these sequences to Tyr-527 suggests a model in which middle-T activates pp60c-src by binding directly to this region of the molecular and thereby preventing phosphorylation of Tyr-527. Alternatively, middle-T binding may mediate a conformational change in this region, which in turn induces an alteration in the level of phosphorylation at Tyr-527 and Tyr-416.     

Mutations around the NG59 lesion indicate an active association of polyoma virus middle-T antigen with pp60c-src is required for cell transformation.

The transforming activity of polyoma virus middle-T antigen is believed to be dependent on its ability to form a complex with the cellular tyrosine protein kinase, pp60c-src. This hypothesis is based on observations of mutants of middle-T which demonstrated a correlation between these two activities. To investigate further the significance of pp60c-src association in transformation by middle-T, a series of deletion and point mutants were constructed around the NG59 lesion since this region has been implicated in pp60c-src binding. Analysis of the middle-T variants revealed a complete correlation between the presence of associated activated pp60c-src and the ability to transform. Further, this ability of pp60c-src to associate with middle-T may depend on the presence of a beta-turn between amino acids 177 and 180. The results indicate the NG59 phenotype results from the introduction of an isoleucine residue between amino acids 177 and 178 rather than the transition mutation at 179. The mutant MG1 is a single point mutation (at residue 180) and represents the smallest change in the middle-T which abolishes both the transformating and kinase activity of middle-T. Taken together, the data suggest the region surrounding the NG59 lesion is involved in the formation of an active complex between middle-T and pp60c-src and strongly suggest that this association is an absolute requirement for polyoma virus-induced transformation.     

Polyomavirus middle-T antigen associates with the kinase domain of Src-related tyrosine kinases.

Middle-T antigen of mouse polyomavirus, an oncogenic DNA virus, associates with and activates the cellular tyrosine kinases c-Src, c-Yes, and Fyn. This interaction is essential for polyomavirus-mediated transformation of cells in culture and tumor formation in animals. To determine the domain of c-Src directing association with middle-T, mutant c-Src proteins lacking the amino-terminal unique domain and the myristylation signal, the SH2 domain, the SH3 domain, or all three of these domains were coexpressed with middle-T in NIH 3T3 cells. All mutants were found to associate with middle-T, demonstrating that the kinase domain of c-Src, including the carboxy-terminal regulatory tail, is sufficient for association with middle-T. Moreover, we found that Hck, another member of the Src kinase family, does not bind middle-T, while chimeric kinases consisting of the amino-terminal domains of c-Src fused to the kinase domain of Hck or the amino-terminal domains of Hck fused to the kinase domain of c-Src associated with middle-T. Hck mutated at its carboxy-terminal regulatory residue, tyrosine 501, was also found to associate with middle-T. These results suggest that in Hck, the postulated intramolecular interaction between the carboxy-terminal regulatory tyrosine and the SH2 domain prevents association with middle-T. This intramolecular interaction apparently also limits the ability of c-Src to associate with middle-T, since removal of the SH2 or SH3 domain increases the efficiency with which middle-T binds c-Src.     

The amino terminus of polyomavirus middle T antigen is required for transformation.

In polyomavirus-transformed cells, pp60c-src is activated by association with polyomavirus middle T antigen. These complexes have a higher tyrosine kinase activity compared with that of unassociated pp60c-src. Genetic analyses have revealed that the carboxy-terminal 15 amino acids of pp60c-src and the amino-terminal half of middle T antigen are required for this association and consequent activation of the tyrosine kinase. To define in greater detail the borders of the domain in middle T antigen required for activation of pp60c-src, we constructed a set of unidirectional amino-terminal deletion mutants of middle T antigen. Analysis of these mutants revealed that the first six amino acids of middle T antigen are required for it to activate the kinase activity of pp60c-src and to transform Rat-1 fibroblasts. Analysis of a series of insertion and substitution mutants confirmed these observations and further revealed that mutations affecting the first four amino acids of middle T antigen reduced or abolished its capacity to activate the kinase activity of pp60c-src and to transform Rat-1 cells in culture. Our results suggest that the first four amino acids of middle T antigen constitute part of a domain required for activation of the pp60c-src tyrosyl kinase activity and for consequent cellular transformation.     

Site-directed mutagenesis of polyomavirus middle-T antigen sequences encoding tyrosine 315 and tyrosine 250.

Tyrosine residues of middle-T and tyrosine phosphorylation are thought to be important in the transformation of cultured rodent cells by polyomavirus. Of the potential tyrosine sites in the carboxyl-terminal half of middle-T, tyrosines 297, 315, and 322 have been studied previously, whereas tyrosine 250 has not. Two mutant plasmids, XD121 and pT250, encode polyomavirus middle-T species in which the tyrosine 250 residue is affected. XD121 is a deletion mutant in which the region encoding tyrosine 250, together with three adjacent amino acids, is deleted, whereas pT250 is a point mutant in which the tyrosine 250 codon has been converted to a phenylalanine codon. The plasmids were handicapped in transforming ability, as judged by focus formation on a monolayer of Rat-1 cells. Both demonstrated a reduction in the number of foci produced and a lag in the time of appearance of foci when compared with wild-type plasmid. The importance of residue 250 in this phenotype was indicated by the observation that plasmids containing multiple mutations proximal to the tyrosine 250 codon were wild type in their transforming ability. Furthermore, a revertant of pT250 (pT250-w.t.), which utilized the alternative tyrosine codon of TAC, was shown to regain full transforming activity. A combined-mutant plasmid, pTH, encodes a middle-T species in which both tyrosines 250 and 315 are converted to phenylalanine. This plasmid was totally defective in the transformation of rodent cells in a focus formation assay; however, it did impart a small measure of anchorage-independent growth when the encoded protein was expressed in NIH 3T3 cells. The in vitro kinase activity and pp60c-src association of the mutant middle-T antigens were examined. These assays demonstrated a reduction in phosphate acceptor activity for the middle-T species encoded by pT250 and pTH. Quantitative kinase assays showed that all of the tyrosine-mutant middle-T species, encoded by pAS131 (containing the tyrosine 315 codon-to-phenylalanine codon mutation), pT250, and pTH, were able to enhance pp60c-src kinase activity but only at levels which were intermediate and which reflected their transforming abilities relative to wild type.     

pp60c-src variants containing lesions that affect phosphorylation at tyrosines 416 and 527.

The biological and biochemical properties of pp60c-src are regulated, in part, by phosphorylation at Tyr-416 and Tyr-527. The tyrosine kinase and transforming activities of pp60c-src are suppressed by phosphorylation at Tyr-527, whereas full activation of pp60c-src requires phosphorylation at Tyr-416. To test specifically the significance of the negatively charged phosphate moieties on these tyrosine residues, we have substituted the codons for both residues with codons for either Glu or Gln. A negatively charged Glu at position 527 was unable to mimic a phosphorylated Tyr at this position, and, in consequence, the mutated pp60c-src was activated and transforming. Similarly, substitution of Tyr-416 with Glu was unable to stimulate the activities of the enzyme. However, mutagenesis of Tyr-416 to Gln (to form the mutant 416Q) activated the kinase activity approximately twofold over that observed for wild-type pp60c-src. When introduced into the mutant 527F (containing Phe-527 instead of Tyr), the double mutant 416Q-527F exhibited weak transforming activity. This is in contrast to the other double mutants 416E-527F and 416F-527F, which were nontransforming. The biochemical basis by which 416Q activates pp60c-src is not understood but probably involves some local conformational perturbation. Deletion of residues 519 to 524 (RH5), a region previously shown to be necessary for association with middle-T antigen, led to loss of phosphorylation at Tyr-527 and activation of the enzymatic and focus-forming activities of pp60c-src. Hence, the sequences necessary for complex formation with middle-T antigen may also be required by the kinase(s) which phosphorylates Tyr-527 in vivo. This suggests that normal cells contain cellular proteins which are analogous to middle-T antigen and whose action regulates the activity of pp60c-src by controlling phosphorylation or dephosphorylation at residue 527.     

Altered sites of tyrosine phosphorylation in pp60c-src associated with polyomavirus middle tumor antigen.

We characterized the tyrosine phosphorylation sites of free pp60c-src and of pp60c-src associated with the polyomavirus middle tumor antigen (mT) in transformed avian and rodent cells. The sites of tyrosine phosphorylation in the two populations of pp60c-src were different, both in vitro and in vivo. Free pp60c-src was phosphorylated in vitro at a single site, tyrosine 416. pp60c-src associated with mT was phosphorylated in vitro on tyrosine 416 and on one or more additional tyrosine residues located in the amino-terminal region of the molecule. Free pp60c-src in polyomavirus mT-transformed cells was phosphorylated in vivo on tyrosine 527. In contrast, pp60c-src associated with mT was phosphorylated in vivo on tyrosine 416 and not detectably on tyrosine 527. Thus, the in vivo phosphorylation sites of pp60c-src associated with mT in transformed cells are identical to those of pp60v-src, the Rous sarcoma virus transforming protein. The results suggest that altered phosphorylation of pp60c-src associated with mT may play a role in the enhancement of the pp60c-src protein kinase activity and in cell transformation by polyomavirus.     

Functional interaction between the SH2 domain of Fyn and tyrosine 324 of hamster polyomavirus middle-T antigen.

Middle-T antigen of mouse polyomavirus (MomT) associates with the cellular tyrosine kinases c-Src, c-Yes, and Fyn, while middle-T antigen of hamster polyomavirus (HamT) exclusively binds Fyn. This interaction is essential for polyomavirus-mediated transformation of cells in culture and tumor formation in animals. Here we show that the kinase domain of Fyn is sufficient for association with MomT but not for binding of HamT. We further demonstrate that a Fyn mutant lacking the SH2 domain is able to bind MomT but fails to associate with HamT, indicating that the SH2 domain of Fyn is essential for stable association with HamT. HamT, but not MomT, contains a tyrosine residue, Tyr-324, in the sequence context YEEI. Mutation of Tyr-324 to phenylalanine led to a drastic reduction of associated Fyn and abolished the oncogenicity of HamT. This suggests that Tyr-324 is the major phosphotyrosine residue mediating the binding of HamT to the SH2 domain of Fyn. These findings show that mouse and hamster polyomaviruses use different strategies to target Src-related tyrosine kinases.     

Transformation of NIH 3T3 cells by cotransfection with c-src and nuclear oncogenes.

pp60c-src, the cellular homolog of the Rous sarcoma virus transforming protein, does not completely transform cells even when present at high levels, but has been shown to be involved in polyomavirus-induced transformation when activated by polyomavirus middle T (pmt)-antigen binding. Here we show that cotransfection, but not solo transfection, of expression plasmids for c-src and either adenovirus E1A, v-myc, c-myc, or the 5' half of polyomavirus large T (pltN) antigen into NIH 3T3 cells induces anchorage-independent growth, enhanced focus formation, and, for pltN cotransfection, tumorigenicity in adult NFS mice. Enhancement of transformation was not observed with polyomavirus small t (pst) antigen. Cotransfection of c-src with pltN induced modification of pp60c-src that altered its electrophoretic mobility and in vivo phosphorylation state and stimulated its in vitro kinase activity. Similar alterations were not seen after c-src-E1A cotransfection, suggesting that at least two different mechanisms of enhancement are involved.     

An antibody to a synthetic peptide recognizes polyomavirus middle-T antigen and reveals multiple in vitro tyrosine phosphorylation sites.

Antibodies were raised against three synthetic peptides corresponding to sequences surrounding tyrosine 315, a putative in vitro phosphorylation site in polyomavirus middle-T antigen. Only one of the peptides (called C and corresponding to residues 311 to 330) elicited antibodies that recognized middle-T efficiently. Middle-T present in immunoprecipitates formed with purified anti-C serum still accepted phosphate on tyrosine in an in vitro kinase reaction. This implies that tyrosines other than 315 and 322 that lie within the antibody binding region are phosphorylated under these conditions. This conclusion was supported by the altered partial V8 proteolysis fingerprint of the labeled middle-T. Two-dimensional tryptic fingerprint analysis of 32P-labeled middle-T showed that several tryptic peptides identified as including tyrosine 315 and 322 were missing from middle-T labeled in anti-C immunoprecipitates compared with middle-T labeled in immunoprecipitates made by using anti-tumor cell serum. However, one major labeled peptide remained. This peptide was also present in fingerprints of 32P-labeled middle-T coded by M45, dl23, pAS131, and dl1013, but a peptide with altered mobility was present in dl8 middle-T. This identified the peptide as including tyrosine 250. We deduce from these data that (i) the presence of the antibody against peptide C inhibits phosphorylation of tyrosines 315 and 322; (ii) middle-T labeled in the kinase reaction after immunoprecipitation with anti-C serum is phosphorylated on tyrosine 250; and (iii) when anti-tumor cell serum is used in the in vitro kinase reaction, middle-T is phosphorylated at multiple sites, including residues 250, 315, and 322.     

Catalytically Inactive Protein Phosphatase 2A Can Bind to Polyomavirus Middle Tumor Antigen and Support Complex Formation with pp60c-src

Interaction between the heterodimeric form of protein phosphatase 2A (PP2A) and polyomavirus middle T antigen (MT) is required for the subsequent assembly of a transformation-competent MT complex. To investigate the role of PP2A catalytic activity in MT complex formation, we undertook a mutational analysis of the PP2A 36-kDa catalytic C subunit. Several residues likely to be involved in the dephosphorylation mechanism were identified and mutated. The resultant catalytically inactive C subunit mutants were then analyzed for their ability to associate with a cellular (B subunit) or a viral (MT) B-type subunit. Strikingly, while all of the inactive mutants were severely impaired in their interaction with B subunit, most of these mutants formed complexes with polyomavirus MT. These findings indicate a potential role for these catalytically important residues in complex formation with cellular B subunit, but not in complex formation with MT. Transformation-competent MT is known to associate with, and modulate the activity of, several cellular proteins, including pp60(c-src) family kinases. To determine whether association of MT with an active PP2A A-C heterodimer is necessary for subsequent association with pp60(c-src), catalytically inactive C subunits were examined for their ability to form complexes containing pp60(c-src) in MT-expressing cells. Two catalytically inactive C subunit mutants that efficiently formed complexes with MT also formed complexes that included an active pp60(c-src) kinase, demonstrating that PP2A activity is not essential in cis in MT complexes for subsequent pp60(c-src) association.     

Cooperation of middle and small T antigens of polyomavirus in transformation of established fibroblast and epithelial-like cell lines.

We have reported recently that small T antigen of polyomavirus stimulates the growth of NIH 3T3 cells beyond their saturation density and induces weak anchorage-independent growth (T. Noda, M. Satake, T. Robins, and Y. Ito, J. Virol. 60:105-113, 1986). We examined whether small T antigen would cooperate with middle T antigen in the in vitro transformation of NIH 3T3 (fibroblasts) and NRK-52E (epitheliallike) cells. The small-T-antigen gene, when cotransfected with the middle-T-antigen gene, had no additional effect on the efficiency or size of dense foci formation induced by the middle-T-antigen gene on a monolayer of NIH 3T3 cells. However, the small-T-antigen gene dramatically increased the rate of growth of NIH 3T3 cells transformed by middle T antigen in semisolid medium. Introduction of the small-T-antigen gene into middle-T-antigen-transformed cells did not disturb the integrated middle-T gene, alter expression of the middle-T gene, or enhance middle-T-antigen-associated tyrosine protein kinase activity. For NRK-52E cells, the expression of middle T antigen alone resulted in small, slow-growing foci on a monolayer. These cells did not show anchorage-independent growth, despite the fact that middle-T-antigen-associated tyrosine protein kinase activity was clearly detected in these cells. NRK-52E cells expressing both middle and small T antigens formed faster growing foci on a monolayer than middle-T-antigen-expressing cells did and grew in semisolid medium, even when the amounts of middle T antigen and its associated kinase activities were lower than those of middle-T-antigen-expressing cells. We conclude that small T antigen cooperates with middle T antigen in the in vitro transformation of established cell lines of fibroblast and epitheliallike cells, that it does not share the middle-T-antigen function even though they are structurally related, and that it has a significantly more important role in the transformation of NRK-52E cells than that of NIH 3T3 cells.