SV40 late replacement vectors containing the polyoma middle-T coding sequences have been constructed. Mixed hybrid virus stocks have been obtained through complementation with a defective SV40 helper genome (dl 1055) following DNA transfection into CV-1 cells. Middle-T antigen is expressed in the infected simian cells at about 5–10 fold higher levels than in polyoma virus-infected mouse cells and has the pp60c-sec-associated tyrosine-specific protein kinase activity in vitro. However, the ‘specific activity’ of the kinase in extracts of the infected CV-1 cells is lower than that observed in polyoma infected 3T6 cell extracts. The half-life of middle-T antigen in the CV-1 cells is about 4 h but the in vitro kinase activity associated with middle-T has a half-life of at least 8 h and hence appears to be stabilized. The in vivo phosphorylated species of middle-T has been shown by sucrose gradient analyses to be largely distinct from the middle-T with associated protein kinase activity in vitro.

The ability of polyoma virus to transform cells results primarily from the action of one of the virus-coded early proteins, called middle-T antigen. Middle-T has an associated tyrosine-specific protein kinase activity that can be measured in vitro and results in the phosphorylation of middle-T itself. Almost all mutants so far tested that lack the ability to transform cells, also lack associated kinase activity. Attempts to map within middle-T the tyrosine residue(s) that are phosphorylated in vitro suggest that a likely site of phosphorylation is tyrosine 315 (refs 8-10 and unpublished results). The amino acid sequence preceding Tyr 315 includes a tract of six contiguous glutamic acid residues and bears some homology with that preceding the tyrosine phosphorylated in vivo in pp60v-src, the transforming protein of Rous sarcoma virus, and with a region in the polypeptide hormone, gastrin, preceding a tyrosine that is sulphated. Furthermore, although surprisingly large tracts of middle-T may be removed without affecting its transforming activity, mutants that lack the sequences corresponding to amino acids 311-318 inclusive are transformation defective. Because the likely site of phosphorylation, the homology with pp60v-src and gastrin and the sequence apparently required for transformation all overlap, it has generally been accepted that this region of middle-T may form part of an essential region, possibly an active site on the protein. Here we have used techniques of site-directed and site-specific mutagenesis to probe the sequence requirements in more detail. Contrary to expectation, the results obtained strongly suggest that Tyr 315 and conservation of the surrounding amino acid sequence are not essential for transformation.

Deletion loop mutagenesis is a new, general method for site-directed mutagenesis that allows point mutations to the introduced within a sequence of DNA defined by a previously isolated deletion mutant. Wild type and deletion mutant DNA are cloned into a bacterial plasmid and each is cleaved with a different single cut restriction enzyme. Heteroduplexes are formed between the two DNAs to produce circular molecules containing a nick in each strand and a single-stranded deletion loop. The deletion loops are mutagenised using sodium bisulphite and the DNA transfected directly into a uracil repair deficient strain of Escherichia coli. Up to half of the resultant clones contain DNA produced by replication of the wild-type length strand and bear mutations exclusively within the target area. An example is given in which a deletion mutant lacking 21 nucleotides from the region coding for SV40 large-T was used. Eight of the possible nine target cytosine residues were mutagenised. The method described is specific, efficient and simple.