Abstract

Mapping of transcripts of the human foamy virus genome was carried out in permissive human embryonic fibroblast cells by Northern blot hybridization and S1 nuclease analysis. Since several splice sites that are localized within a relatively narrow genomic region were detected, the polymerase chain reaction (PCR) was employed, and cloning and sequencing of the splice site junctions of the corresponding viral cDNAs were subsequently performed. All spumavirus transcripts have a common but relatively short leader RNA. Genomic, singly spliced env mRNAs and several singly and multiply spliced subgenomic transcripts were identified. The multiply spliced viral mRNAs consist of various exons located in the central or 3' part of the viral genome. At least four novel gene products, termed Bet, Bes, Beo, and Bel3, are predicted to exist. The poly(A) addition site that defines the boundary of the R and U5 region in the 3' long terminal repeat was determined. The pattern of spumavirus splicing is more complex than that of oncoviruses and more similar to that of lentiviruses. One of the characteristic features of spumavirus transcription is the existence of singly spliced bel1 and bel2 mRNAs that alternatively are multiply spliced, thereby generating a complexity comparable to, but different from, that of lentiviruses and from that of other known retroviruses. The complex spumavirus transcriptional pattern of human spumavirus and the coding potential of the 10 exons identified are discussed.