Virtually full-length sequences of HIV type 1 subtype J reference strains.

Abstract

293 THE RAPID VIRAL TURNOVER and high mutation rate of HIV1 result in remarkable genetic variation among strains obtained from different parts of the world. This has led to the development of evolutionary lineages (genetic subtypes) that can be identified by phylogenetic analysis of gene sequences.2,3 Eight genetic subtypes, named A through H in group M (major) and a group of other strains, group O (outlier), have previously been determined by phylogenetic analysis, mainly of gag and env genes. The genetic subtype classification is based on consistent clustering of the sequences in different parts of the genome, and approximate equidistance to strains belonging to other subtypes. Two additional subtypes, I and J, have been proposed, on the basis of analysis of gene subfragments. Sequences from these strains were shown to group as independent, monophyletic lineages in phylogenetic comparisons to representative reference strains of previously described subtypes A through H.2,3 However, the definition of a genetic subtype is dependent on consistent clustering of sequences independent of the genome fragment analyzed (provided that enough informative variation is present in the region). This means that the classification results must be similar regardless of the genome region examined. Since HIV-1 strains of different subtypes have been shown to recombine, a classification based on a single subgenomic fragment does not necessarily indicate the genetic subtype of the complete genome. When a putative new genetic subtype is found the complete genome should be analyzed to verify that all regions of the virus cluster independently. In this study we have cloned and sequenced the first virtually complete proviruses of two independently obtained, epidemiologically unlinked primary isolates of HIV-1 subtype J. They form an independent cluster, approximately equidistant to the other subtypes over all of the genome, verifying their classification as a new subtype. Blood samples SE9173 and SE9280 for virus isolation were collected in Sweden from two previously identified HIV-1-infected individuals carrying the putative new subtype J viruses.8 Both individuals were immigrants from the Democratic Republic of Congo (formerly Zaire). A thorough epidemiological investigation, including contact tracing, did not reveal any direct or indirect contact between the two individuals. The initial identification of subtype J was based on direct sequencing and phylogenetic analysis of the gag p17 and env V3 regions amplified from uncultured patient lymphocytes. Samples SE9173 and SE9280 were obtained from the same individuals as the previously described samples SE7022 and SE7887, respectively. A seminested extended polymerase chain reaction (PCR) was used to amplify the virtually complete provirus from peripheral blood mononuclear cell (PBMC) cocultures essentially as has been previously described. Briefly, primers MSF12, binding to the tRNA-binding site located in the U5-gag leader junction of the 59 long terminal repeat (LTR), and MSR5, binding to the mRNA-polyadenylation signal site located in the R-U5 junction of the 3 9 LTR10 (Fig. 1), were used for first-round amplification. For second-round amplification one-tenth of the reaction was reamplified using primers MSF14 and MSR5. Second-round amplicons were cloned with a TA-cloning kit (pCR2.1 cloning kit; Invitrogen, San Diego, CA). Clones were sequenced on both strands by primer walking, using an ABI 377 automated sequencer and dye-terminator chemistry (Perkin-Elmer Applied Biosystems Division, Foster City, CA). The sequences have been submitted to GenBank with accession numbers AF082394 and AF082395. The lengths of the clones from strains SE9173 and SE9280 were 8953 and 8943 bp, respectively. The genome structure was similar to all other subtypes of HIV-1, with the gag, pol, and env structural genes, and reading frames for vif, vpr, vpu, and nef single exon regulatory/accessory genes (Table 1). The tat and rev exons 1 and 2 were also discernible at their usual positions. All reading frames appeared open and of complete length. The newly obtained SE9173 and SE9280 sequences were 94.2 and 97.6% similar to the SE7022 and SE7887 sequences in the V3 region, respectively.