Although the prevalence of benign JC virus infection of the kidney is very high, routes of transmission and events leading to brain pathology are not known. We sought a means of classifying and organizing viral isolates as an aid in studying such questions. Two major types of JCV have been identified. A 610-bp segment of the 3' ends of the T-antigen and VPl genes amplified and sequenced from the brains of 11 progressive multifocal leukoencephalopathy (PML) patients contains 20 sites of point mutation which allow reliable classification of JCV isolates. At these 2O type-determining sites the presence of alternate nucleotides creates two distinct patterns of substitution, allowing six of the isolates to be grouped into Type l and the other 5 into Type 2. The prototype strain Mad-l was found to be Type l, while the other previously sequenced strain, GS/B, is a Type 2. There are three additional type-determining sites to the early side of origin, which have allowed classification of many of the published regulatory region sequences. Only 4 of our 11 isolates had a 'crossover' to the opposite type consensus sequence, indicating a very high type specificity. These crossovers, along with random unique mutations, allow individual strains to be distinguished. No cases of multiple infecting strains were observed; however, one strain contained six crossovers, suggesting the possibility of genetic recombination in potential rare cases of multiple infection. Based on these stable point mutations, two sets of PCR primers were developed which type-specifically amplify shorter fragments that also contain type-specific restriction enzyme sites, permitting rapid and accurate classification of new isolates. Although amino acid sequence is conserved between types in the VPl gene, several sites in the T-antigen gene cause a type-specific amino acid substitution. Comparison of each type's consensus sequence at type-determining sites to those sites in BK virus suggests that Type 2 represents the ancestral JCV sequence from which Type 1 diverged during human evolution. The regulatory regions from these strains showed roughly defined patterns of rearrangement, with individual variation unique to each strain. The observed patterns are not specific to one or the other viral type, and probably are generated by the host cell. Specific blocks of sequence are consistently preserved, and several combinations of elements apparently are viable. Usually, only one rearranged promotor form is isolated from each individual brain.
