Sequence Analysis of In Vivo-Expressed HIV-1 Spliced RNAs Reveals the Usage of New and Unusual Splice Sites by Viruses of Different Subtypes.

HIV-1 RNAs are generated through a complex splicing mechanism, resulting in a great diversity of transcripts, which are classified in three major categories: unspliced, singly spliced (SS), and doubly spliced (DS). Knowledge on HIV-1 RNA splicing in vivo and by non-subtype B viruses is scarce. Here we analyze HIV-1 RNA splice site usage in CD4+CD25+ lymphocytes from HIV-1-infected individuals through pyrosequencing. HIV-1 DS and SS RNAs were amplified by RT-PCR in 19 and 12 samples, respectively. 13,108 sequences from HIV-1 spliced RNAs, derived from viruses of five subtypes (A, B, C, F, G), were identified. In four samples, three of non-B subtypes, five 3' splice sites (3'ss) mapping to unreported positions in the HIV-1 genome were identified. Two, designated A4i and A4j, were used in 22% and 25% of rev RNAs in two viruses of subtypes B and A, respectively. Given their close proximity (one or two nucleotides) to A4c and A4d, respectively, they could be viewed as variants of these sites. Three 3'ss, designated A7g, A7h, and A7i, located 20, 32, and 18 nucleotides downstream of A7, respectively, were identified in a subtype C (A7g, A7h) and a subtype G (A7i) viruses, each in around 2% of nef RNAs. The new splice sites or variants of splice sites were associated with the usual sequence features of 3'ss. Usage of unusual 3'ss A4d, A4e, A5a, A7a, and A7b was also detected. A4f, previously identified in two subtype C viruses, was preferentially used by rev RNAs of a subtype C virus. These results highlight the great diversity of in vivo splice site usage by HIV-1 RNAs. The fact that four of five newly identified splice sites or variants of splice sites were detected in non-subtype B viruses allows anticipating an even greater diversity of HIV-1 splice site usage than currently known.