PURINE LOADING 0F EBNA-1 mRNA AVOIDS SENSE-ANTISENSE "COLLISIONS". A. D. Cristillo, T. P. Lillicrap and D. R. Forsdyke. Dept. Biochem., Queen's University, Kingston, Ontario, Canada K7L3N6

   In the 1960's Szybalski showed that mRNA-synonymous strands of DNA have purine-rich clusters, and Chargaff presented his second parity rule that, to a close approximation, %A=%T and %C=%G for single DNA strands. We report that small deviations from the rule (%R>%Y in mRNA-synonymous strands) allow determination of transcription direction for the majority of genes in many species (Bell & Forsdyke, unpublished work; Dang et al. 1998, Biochem. Cell Biol. in press). Purine clusters in an mRNA correspond to the loop domains of potential stem-loop structures, which, by virtue of being enriched with non-complementary bases, should avoid loop-loop "kissing" interactions with other mRNAs in the same cell.  However, most genes of CG-rich viruses with a high commitment to latency (HTLV-1, Epstein-Barr) disobey the transcription direction rule (%Y>%R). Thus, the viruses have pyrimidine clusters in the loop domains of potential RNA stem-loop structures. By "driving on the wrong side of the road" they invite sense-antisense RNA "collisions" with host mRNAs, thus triggering the interferon response and increasing MHC protein expression. However, the only gene transcribed during the "EBNA-1 only program" of viral latency is "polite" (%R>%Y). This purine-loading requires the protein to carry a non-functional (GlyAla) simple-sequence domain.