SIV genome-wide pyrosequencing provides a comprehensive and unbiased view of variation within and outside CD8 T lymphocyte epitopes

Abstract

Deep sequencing technology is revolutionizing our understanding of HIV/SIV evolution. It is known that acute SIV sequence variation within CD8 T lymphocyte (CD8-TL) epitopes is similar among MHC-identical animals, but we do not know whether this persists into the chronic phase. We now determine whether chronic viral variation in MHC-identical animals infected with clonal SIV is similar throughout the entire coding sequence when using a sensitive deep sequencing approach. We pyrosequenced the entire coding sequence of the SIV genome isolated from a unique cohort of four SIVmac239-infected, MHC-identical Mauritian cynomolgus macaques (MCM) 48 weeks after infection; one MCM in the cohort became an elite controller. Among the three non-controllers, we found that genome-wide sequences were similar between animals and we detected increased sequence complexity within 64% of CD8-TL epitopes when compared to Sanger sequencing methods. When we compared sequences between the MHC-matched controller and the three non-controllers, we found the viral population in the controller was less diverse and accumulated different variants than the viral populations in the non-controllers. Importantly, we found that initial PCR amplification of viral cDNA did not significantly affect the sequences detected, suggesting that data obtained by pyrosequencing PCR-amplified viral cDNA accurately represents the diversity of sequences replicating within an animal. This demonstrates that chronic sequence diversity across the entire SIV coding sequence is similar among MHC-identical animals with comparable viral loads when infected with the same clonal virus stock. Additionally, our approach to genome-wide SIV sequencing accurately reflects the diversity of sequences present in the replicating viral population. In sum, our study suggests that genome-wide pyrosequencing of immunodeficiency viruses captures a thorough and unbiased picture of sequence diversity, and may be a useful approach to employ when evaluating which sequences to include as part of a vaccine immunogen.

Figure 1. Variation in CD8-TL epitopes in virus populations from non-controller M3/M3 MCM and an SIVmac239 stock.

Amino acid variation detectable by pyrosequencing in 12 CD8-TL epitopes in viruses isolated from three M3/M3 MCM (CY0163, CY0164, and CY0166) is compared to variation previously reported by Sanger sequencing bulk PCR amplicons . Amino acid variants detectable within an SIVmac239 stock were characterized here and elsewhere . Sequences that match the inoculum are represented with a “.”. Mixed populations are represented with an “X”. The number of high quality reads with each specific sequence is shown. The total number of reads for each epitope can be found in Table S2. The frequency of reads with a specific amino acid sequence is shown as “% Var.” The sequences labeled as “other” include variants that were individually present at less than 1% and variants that were also detected in the inoculum, which included the A2V mutation in Gag_221–229PR9. The relative shading of each box reflects the frequency of reads, as indicated in the legend.

Figure 2. Sequence variants in CD8-TL epitopes are detectable by directly pyrosequencing virus populations isolated from CY0166.

Amino acid variation detected by directly pyrosequencing the virus population in CY0166 is shown for all 12 CD8-TL epitopes. Sequences that match the inoculum are represented with a “.”. The number of high quality reads with each specific sequence is shown. The total number of reads for each epitope can be found in Table S2. The frequency of reads with a specific amino acid sequence is shown as “% Var.” The sequences labeled as “other” include variants that were individually present at less than 1% and variants that were also detected in the inoculum. The relative shading of each box reflects the frequency of reads, as indicated in the legend.

Figure 3. Variation in CD8-TL epitopes in a virus population from an M3/M3 elite controller.

Amino acid variation detected by pyrosequencing in 12 CD8-TL epitopes in viruses isolated from CY0165 is compared to variation previously reported by Sanger sequencing bulk PCR amplicons . Sequences that match the inoculum are represented with a “.”. The number of high quality reads with each specific sequence is shown. The total number of reads for each epitope can be found in Table S2. The frequency of reads with a specific amino acid sequence is shown as “% Var.” The sequences labeled as “other” include variants that were individually present at less than 1% and variants that were also detected in the inoculum. The relative shading of each box reflects the frequency of reads, as indicated in the legend.