Experimental tumour induction by SV40 transformed cells.

Although SV40 transforms cells from many species, transformed cells from species other than the Syrian hamster are rarely tumorigenic in immunocompetent animals. However, secondary manipulations of SV40-transformed cells can result in increased tumorigenicity. Some early observations on tumour induction by SV40 and transformed cells will be followed by a selected review of evidence suggesting that tumorigenicity of SV40-transformed cells involves serial mutations. An SV40-transformed rat cell model will be described to illustrate the changes in tumorigenic phenotype that can occur during tumour progression. This information will be used to propose that SV40 immortalization of cells (other than hamster cells) is only the first in a series of steps in the pathway toward tumorigenicity and that a complete understanding of the oncogenicity of SV40 will require definition of the secondary genetic events which complement SV40 immortalization to create the fully tumorigenic phenotype.

A 47-amino-acid fragment of SV40 T antigen represses transcription from human GSTalpha promoters.

SV40 T antigen downregulates the expression of an important detoxification enzyme, glutathione S-transferase alpha (GSTalpha). We show here that the target of this repression is a 14-bp element common to the human GSTA1 and GSTA2 promoters. This element, which we have named TAGR, is also critical for high-level, constitutive expression from these promoters. The TAGR element does not appear to contain a binding site for any transcription factor known to be present in fibroblasts, although the TAGR element does resemble the binding site for the Ikaros transcription factor found in hematopoietic cells. We also have identified a 47-amino-acid fragment of T antigen that includes amino acids 83-100 and 119-147, which is sufficient to repress transcription from the GSTalpha promoter in transient transcription assays. Thus, GSTalpha repression does not require binding of T antigen to pRb, p300, or p53, since the domains of T antigen required for binding these cellular proteins are missing from this T antigen fragment. We show, however, that this fragment does bind to three cellular proteins with approximate molecular weights of 54, 59, and 94 kDa.

SV40 and adenovirus may act as cocarcinogens by downregulating glutathione S-transferase expression.

We have discovered a novel function of the SV40 T antigen and the adenovirus E1A proteins: the ability to downregulate the endogenous expression of an important detoxification enzyme, glutathione S-transferase alpha (GST alpha). GST alpha mRNA is much less abundant in rat and human cells that express SV40 T antigen than in the parental cell lines. This GST alpha downregulation does not require expression of SV40 small t antigen or complex formation between large T antigen and p53, p300, or the pRb family of proteins. As might be predicted, cells that express SV40 T antigen are more sensitive than normal cells to alkylating drugs, which GST alpha is known to detoxify. Finally, GST alpha expression is also downregulated in cells that express the adenovirus E1A proteins. We propose that by downregulating GST alpha expression and inactivating p53 function, SV40 and adenovirus may contribute to the initiation of, or the progression toward, malignancy. Thus, in their quest to establish persistent infections, these viruses may inadvertently make the cellular environment more permissive for tumorigenesis.

Reduced levels of alpha1(XI) procollagen mRNA in SV40-transformed cells.

We have used mRNA differential display to search for cellular genes whose expression is modulated in SV40-transformed cells. We identified a gene, the alpha1 subunit of type XI procollagen (alpha1(XI)), whose mRNA level is reduced in most F111 rat cell lines that express either wild-type SV40 T antigen or an amino-terminal fragment of T antigen (T147D). alpha1(XI) mRNA expression is also reduced in rat embryo fibroblasts transformed by SV40. This phenomenon is not unique to rat cells, because alpha1(XI) mRNA expression is repressed in C3H10T1/2 mouse cells transformed by either T147D or wild-type SV40. Interestingly, alpha1(XI) mRNA levels are not reduced in F111 cells that express only the polyomavirus large T antigen, but are reduced in F111 cells transformed by polyoma virus (which expresses all three polyomavirus T antigens) and in F111 cells that express the adenovirus E1A proteins. Thus, repression of alpha1(XI) mRNA expression may be a common feature of transformation by DNA tumor viruses.

An amino-terminal fragment of SV40 T antigen induces cellular DNA synthesis in quiescent rat cells.

SV40 large T antigen can stimulate cellular DNA synthesis in quiescent cells. Here we report that an SV40 mutant, which expresses only the amino-terminal 147 amino acids of this protein and which does not make small t antigen, stimulates cellular DNA synthesis in quiescent rat cells at levels similar to those achieved with serum stimulation. Thus, the amino-terminal 147 amino acids of large T antigen define a domain of this protein that is sufficient for efficient stimulation of cellular DNA synthesis in the absence of small t antigen.

An amino-terminal fragment of SV40 T antigen transforms REF52 cells.

An SV40 mutant, T147D, encodes only the amino-terminal 147 amino acids of large T antigen and does not make small t antigen. We show here that a retrovirus which expresses this mutant T antigen transforms rat REF52 cells as efficiently as a retrovirus that expresses both the wild-type large and small T antigens. This cell line had previously been refractory to transformation by mutants that make short, amino-terminal fragments of T antigen.

The amino-terminal 147 amino acids of SV40 large T antigen transform secondary rat embryo fibroblasts.

T147D is an SV40 mutant that encodes only the amino-terminal 147 amino acids of large T antigen and does not make small t antigen. We have constructed a recombinant retrovirus that expresses the T147D mutant protein. We show here that this virus can transform the established rat cell line, F111, in an agar assay with high efficiency. More importantly, we demonstrate that this retrovirus transforms secondary rat embryo fibroblasts to anchorage independence as efficiently as a recombinant retrovirus that expresses both wild-type large and small T antigens. These data indicate that in rat cells, the amino-terminal 147 amino acids of T antigen are sufficient for transformation. Further, since the T147D protein does not bind p53, we conclude that the association between T antigen and p53 is not required for the transformation of rat cells to anchorage-independent growth.

Method to identify genomic targets of DNA binding proteins.

We have devised a cyclical immunoprecipitation protocol that can be used to identify and clone a specific DNA sequence that is recognized by a DNA binding protein, even if that sequence is present in only one copy in the genome of a mammal. As an example, we have used this procedure to purify mouse genomic sequences to which the simian virus 40 tumor (T) antigen binds.

An SV40 mutant oncoprotein has a nuclear location.

T147 is an SV40 mutant that makes a normal small t antigen and a large T antigen that is only 147 amino acids long. We have introduced a second mutation into the genome of T147 which eliminates its ability to encode small t antigen. We show that this double mutant is able to transform C3H10T1/2 mouse cells in a focus assay and F111 rat cells in an agar suspension assay, demonstrating that the transforming domain of T antigen is located within its amino-terminal 147 amino acids. We also show that the T147 mutant T antigen, like wild-type T antigen, has a nuclear location. However, in contrast to wild-type T antigen, which is also found in the plasma membranes of wild-type transformed cells, we fail to detect any mutant T antigen associated with the plasma membranes of T147 transformants.

A new SV40 mutant that encodes a small fragment of T antigen transforms established rat and mouse cells.

We have constructed a new SV40 mutant, T147, that makes a large T antigen that is only 147 amino acids long. We show that the T147 T antigen is a phosphoprotein that is as stable as wild-type T antigen and that the SV40 viral origin binding activity of the T147 T antigen is reduced at least 100-fold relative to that of wild-type T antigen. Most importantly, we demonstrate that cloned T147 DNA transforms rat F111 cells to anchorage independence as efficiently as cloned wild-type SV40 DNA and that cloned T147 DNA also efficiently transforms C3H10T1/2 mouse cells in a focus assay.

An SV40 mutant T antigen does not bind the SV40 viral origin.

F8dl is an SV40 deletion mutant that lacks over 60% of the coding sequences for large T antigen and yet is able to immortalize early passage rat cells, to transform established cell lines, and to cause tumors in animals. We report here on the further characterization of this mutant and show that (a) transformation by F8dl is protein mediated but does not require the action of the SV40 small t antigen; (b) the F8dl T antigens have, or are associated with, an ATPase activity; (c) the 34-kDa mutant T antigen of F8dl is localized in nuclei and cell membranes of F8dl transformants and binds to double-stranded DNA; (d) the 20-25 kDa forms of the mutant T antigen are cytoplasmic; and (e) the F8dl T antigens do not bind with high affinity to the SV40 origin of viral DNA replication.

The simian virus 40 sequences between 0.169 and 0.423 map units are not essential to immortalize early-passage rat embryo cells.

F8dl is a simian virus 40 early-region deletion mutant that lacks the sequences between 0.169 and 0.423 map units. We show that cloned F8dl DNA immortalized early-passage Fisher rat embryo cells with an efficiency that was about 20% of that of cloned wild-type simian virus 40 DNA. In contrast, we detected no immortalized colonies when we transfected the cells with DNA of five other early-region deletion mutants that do not make stable truncated forms of T antigen. Since all five of these mutants have intact early- and late-region control sequences, we conclude that these control sequences are not sufficient for immortalization. Three of the mutants that did not immortalize did make a normal small t antigen, suggesting that the expression of this protein alone is not sufficient for immortalization of early-passage Fisher rat embryo cells.

The SV40 T-antigen gene can have two introns.

F8dl is an SV40 early-region mutant that lacks over 60% of the DNA sequences normally used to encode large T antigen. This mutant employs a novel splice donor junction at nucleotide 4425 to produce a family of doubly spliced messages. A similar splicing pattern with wild-type SV40 mRNA has been observed, indicating that the wild-type gene for T antigen can also have two introns. A single G-to-T base change at nucleotide 4425 is sufficient to eliminate this novel donor splice junction.

A fragment of the simian virus 40 early genome can induce tumors in nude mice.

Cell lines transformed by simian virus 40 mutant F8dl (deleted from 0.168 to 0.424 map units, corresponding to the carboxy-terminal 62% of the wild-type simian virus 40 large tumor antigen) are tumorigenic in nude mice. Four of five C3H10T1/2 cell lines transformed by F8dl were tumorigenic in nude mice, whereas two of two wild-type transformants were tumorigenic.

Less than 40% of the simian virus 40 large T-antigen-coding sequence is required for transformation.

F8dl is a simian virus 40 early-region deletion mutant that lacks the simian virus 40 DNA sequences between 0.168 and 0.424 map units. Despite this large deletion, cloned F8dl DNA transforms Fisher rat F111 cells and BALB/3T3 clone A31 mouse cells as efficiently as does cloned simian virus 40 wild-type DNA. These results indicate that less than 40% of the large T-antigen-coding sequence is required for efficient transformation.

Simian virus 40 sequences between 0.168 and 0.424 map units are not required for abortive transformation.

We have isolated a simian virus 40 deletion mutant, F8dl, that lacks the sequences from 0.168 to 0.424 map units. The deleted sequences represent over 60% of the coding region for large T antigen. Despite this deletion, F8dl abortively transformed rat cells as efficiently as wild-type simian virus 40. From this result, we conclude that the region of the simian virus 40 genome between 0.168 and 0.424 map units is not essential for abortive transformation. Since abortive transformation requires the expression of the simian virus 40 maintenance functions, we also infer that the sequences deleted from F8dl are not required to maintain transformation.

A simian virus 40 dl884/tsA58 double mutant is temperature sensitive for abortive transformation.

We have constructed a dl884/tsA58 double mutant and a t+T- early-region deletion mutant and have used these mutants to study the roles of the simian virus 40 tumor antigens (T and t) in transformation. Our major conclusions are that (i) although the mutant tsA58 is not temperature sensitive for the maintenance of transformation, the dl884/tsA58 double mutant is; (ii) small t antigen can provide at least one, but not all, of the functions required for the maintenance of transformation; and (iii) at least two different functions are required for the maintenance of simian virus 40 transformation.

Simian virus 40 deletion mutants that transform with reduced efficiency.

We have constructed two simian virus 40 (SV40) early-region deletion mutants that lack a significant portion of the sequences normally used to encode the SV40 large tumor antigen. Despite these deletions, the mutants were able to transform mouse cells in a focus assay, although with a frequency that was drastically reduced relative to wild-type SV40. Cell lines expanded from the mutant-transformed foci contained integrated mutant DNA, expressed an SV40 tumor antigen (small-t), and exhibited a range of transformed phenotypes, which included the ability to grow while suspended in soft agar. We also present evidence that these mutants are defective for abortive transformation in an assay that tested the transient loss of anchorage dependence. Their ability to stably transform, contrasted with their inability to abortively transform at detectable levels, raises the possibility that the mechanism by which these mutants transform may be different from that of wild-type SV40.

Stabilization of the 53,000-dalton nonviral tumor antigen is not required for transformation by simian virus 40.

We have isolated a simian virus 40 deletion mutant, F8dl, that lacks the sequences from 0.168 to 0.424 map units. The deleted sequences represent about one-half of the coding region for large T antigen. We present evidence here that F8dl is able to transform mouse cells in a focus assay and that cell lines derived from these foci exhibit fully transformed phenotypes, have integrated mutant genomes, and express mutant-encoded proteins. This result implies that the region of the simian virus 40 genome between 0.168 and 0.424 map units is not essential for the maintenance of transformation. In addition, we have found that cells fully transformed by F8dl produce a 53,000-dalton nonviral tumor antigen (p53) that is as unstable as the p53 of untransformed cells. From this result we infer that transformation by simian virus 40 does not require the stabilization of p53.