Detection, characterization and regulation of antisense transcripts in HIV-1

Abstract

BACKGROUND: We and others have recently demonstrated that the human retrovirus HTLV-I was producing a spliced antisense transcript, which led to the synthesis of the HBZ protein. The objective of the present study was to demonstrate the existence of antisense transcription in HIV-1 and to provide a better characterization of the transcript and its regulation. RESULTS: Initial experiments conducted by standard RT-PCR analysis in latently infected J1.1 cell line and pNL4.3-transfected 293T cells confirmed the existence of antisense transcription in HIV-1. A more adapted RT-PCR protocol with limited RT-PCR artefacts also led to a successful detection of antisense transcripts in several infected cell lines. RACE analyses demonstrated the existence of several transcription initiation sites mapping near the 5' border of the 3'LTR (in the antisense strand). Interestingly, a new polyA signal was identified on the antisense strand and harboured the polyA signal consensus sequence. Transfection experiments in 293T and Jurkat cells with an antisense luciferase-expressing NL4.3 proviral DNA showed luciferase reporter gene expression, which was further induced by various T-cell activators. In addition, the viral Tat protein was found to be a positive modulator of antisense transcription by transient and stable transfections of this proviral DNA construct. RT-PCR analyses in 293T cells stably transfected with a pNL4.3-derived construct further confirmed these results. Infection of 293T, Jurkat, SupT1, U937 and CEMT4 cells with pseudotyped virions produced from the antisense luciferase-expressing NL4.3 DNA clone led to the production of an AZT-sensitive luciferase signal, which was however less pronounced than the signal from NL4.3Luc-infected cells. CONCLUSION: These results demonstrate for the first time that antisense transcription exists in HIV-1 in the context of infection. Possible translation of the predicted antisense ORF in this transcript should thus be re-examined.