585 WITHIN THE MAJOR a Mo group of HIV-1, phylogenetic analysis permits subclassification into the subtypes A to J. In 1996, a reference set of HIV-1 sequences representing the different subtypes was proposed and this list has been updated and modified as more full-length HIV-1 genomic sequences, including an increasing number of complex recombinant isolates, have become available.2 An analysis of four full-length HIV-1 genomes from Africa (DJ263, DJ264, HH8793, and SE6165) has proposed a new classification of strains previously defined as subtype G or as containing regions of subtype G.3 Isolates with a characteristic genomic pattern of recombination between subtypes A and G cluster together into a distinct clade exemplified by the strain IbNg. This recombinant virus shows an extensive geographic spread, having been isolated in Nigeria, Djibouti, and, by extrapolation of more limited sequence data from isolates that cluster closely with these examples, in Ivory Coast, the Democratic Republic of Congo (DRC), and Gabon.3,4 It is proposed that such circulating recombinant HIV-1 strains should be designated as a circulating recombinant formso (CRFs) of the virus, rather than as separate subtypes. Thus, sequences of the IbNg CRF type are termed a AG-IBNG.o By contrast, the two novel isolates HH8793 and SE6165, from Kenya and Congo, respectively, carry genomes that appear to be nonrecombinant and cluster together in subtype G. This sequence note presents the full-length genomic sequence of a subtype G HIV-1 originating in Kinshasa, DRC (termed DRCBL). The clinical history of the infected patient and initial subtyping of the virus in gag have been previously reported as isolate BL.5 It has become apparent that classification of HIV-1 on the basis of the analysis of small genomic regions is sometimes inadequateÐparticularly when dealing with isolates of African origin. Complex recombinant forms of HIV1 are increasingly being identified and accurate classification requires the examination of large, continuous regions of the genome. For this reason the entire genomic sequence of the virus was determined from patient peripheral blood mononuclear cells (PBMCs) by a previously described method.6 All reading frames were open and intact, apart from a 7-nucleotide (nt) deletion in pol at position 2948, resulting in a premature termination of the predicted protein product. This deletion was most probably the result of deriving the genomic sequence from a single integrated provirusÐmany of which contain defects. Its presence had no effect on the subsequent phylogenetic analysis. Initial phylogenetic analysis of p17gag, pol, the C4±V4 region of env and nef of DRCBL, by comparison with reference strain sequences, showed clustering in subtype G with a high degree of confidence in all regions (data not shown). Neighborjoining phylogenetic analysis was then performed using fulllength genomic sequences from reference strains of subtypes A to J. DRCBL clustered with subtype G isolates with 100% bootstrap value (Fig. 1). The LTR of DRCBL shows the described structural features of subtype G. Interestingly, the 3-nucleotide TAR bulge has the sequence UUUÐ in common with the isolates SE6165, HH8793, DJ263, DJ264, 92NG083.2,7 92NG003.1,7 and 92RU1318; but not IbNg, which has UCU. To examine DRCBL for evidence of intersubtype recombination, maximum parsimony bootscanning was performed as described with a sliding window of 300 nt advancing in 20-nt increments (Fig. 2). DRCBL was compared with the two reference subtype G strains SE6165 and HH8793 (mentioned above). One reference strain of each subtype was included, excepting E, because the A-type portions of this recombinant sub-
Read more