HIV Evolution: CTL Escape Mutation and Reversion After Transmission

A J Leslie; K J Pfafferott; P Chetty; R Draenert; M M Addo; M Feeney; Y Tang; E C Holmes; T Allen; J G Prado; M Altfeld; C Brander; C Dixon; D Ramduth; P Jeena; S A Thomas; A St John; T A Roach; B Kupfer; G Luzzi; A Edwards; G Taylor; H Lyall; G Tudor-Williams; V Novelli; J Martinez-Picado; P Kiepiela; B D Walker; P J R Goulder


Nat Med. 2004;10(3) 

In This Article


The 128 C-clade-infected subjects analyzed in this study were collected from Durban, South Africa, and consisted predominantly of Zulu and Xhosa women recruited from the Cato Manor antenatal clinic. All subjects were ART-naive. The median viral load was 14,892 HIV RNA copies per ml plasma (range <400-698,000), and the median absolute CD4 count was 449 cells/mm3 (range 101-1,215 cells/mm3). The 97 B-clade-infected subjects were collected from diverse sources encompassing Europe, the Caribbean and North America. Of these subjects, 17% were receiving ART at the time of analysis. The median viral load in those not receiving ART was 24,600 RNA copies per ml plasma (range <400 to >750,000), and the median CD4 count for those not receiving ART was 420 cells/mm3 (range 131-1,280 cells/mm3). This study was approved by institutional review boards, and all subjects gave written informed consent.

We extracted genomic DNA from peripheral blood mononuclear cell (PBMC) pellets using the Puregene DNA isolation kit (Gentra). We then amplified HIV gag sequences using a nested PCR as described,[38] using the following gag-specific primers: 5'-CTAGCAGTGGCGCCCGAACA-3' and 5'-ACAGTCTTTCATTTGGTGTCCTTC-3' for first-round outer PCR, and 5'-TCTCTCGACGCAGGACTC-3' and 5'-TTTCCACATTTCCAACAGCC-3' for second-round inner PCR. The PCR product was purified by PEG precipitation and either directly sequenced (referred to in the text as 'population sequencing'), or cloned as previously described[38] using a TOPO TA cloning kit (Invitrogen). We recovered viral RNA from plasma samples using the Nucleospin RNA extraction kit (Mackery-Nagel). During this step, we added RNA-free DNase (Qiagen) to ensure removal of proviral DNA. We then synthesized a cDNA library from the RNA using the Reverse-iT 1st Strand Synthesis kit (ABgene), using random oligonucleotides to prime the reaction. We then amplified and sequenced gag fragments from the cDNA as described above. All sequencing was done with BigDye Terminator v3.0 Ready Reaction mix (Applied Biosystems), using the two inner PCR primers listed above and four additional primers (5'-CTGCACTATAGGATAATTTTGAC-3', 5'-GACACCAAGGAAGCCTTAG-3', 5'-CTCCCACTGGAACAGGTG-3' and 5'-GGAACAAATAGCATGGATGAC-3'). Sequences were analyzed on the ABI 3700 DNA analyzer. All residue numbers were taken against the HXB reference sequence.

HLA class I typing was performed on extracted genomic DNA by PCR single-strand conformation polymorphism.[39]

We confirmed TW10 (TSTLQEQIAW) as the optimal B57-restricted epitope in C-clade infection as previously described,[14] using fresh PBMCs from the C-clade-infected subject A-005M (HLA-B57/7). TSTLEQIGW was confirmed as the optimal epitope for B-clade infection in the same way, using PBMCs from the B-clade-infected subject SWS (HLA-B57/13). In both cases, we tested fresh PBMCs for recognition of synthetic TW10 peptides in ELISpot assays as previously described.[40]

A measure of the selection pressures operating at individual codons in the Gag protein was obtained using a maximum-likelihood method,[20] which was recently shown to be a powerful way to determine selection pressures in HIV.[41] This method works by assessing the fit to the data of various models of codon evolution, which differ in how dN/dS varies across the sequence and takes into account the phylogenetic relationships of the sequences in question.[20] Two models of codon evolution were used: M7, which specifies ten categories of dN/dS in the alignment, none of which may be <1 so that evolution is entirely neutral; and M8, which only differs from M7 in that it incorporates an extra (eleventh) class of codons that can take on any value of dN/dS, including those supporting positive selection (dN/dS > 1). Because M7 and M8 are nested, they may be compared using a standard likelihood ratio test. Consequently, if model M8 significantly rejects model M7 and includes a class of codons where dN/dS > 1, we can conclude that positive selection has acted. Individual positively selected amino acid sites can then be identified using a Bayesian approach. All analyses were undertaken using the CODEML program from the PAML package,[42] with input phylogenetic trees inferred using the maximum-likelihood method available in the PAUP* package.[43] All other phylogenetic trees were also inferred using the maximum-likelihood method available in PAUP*, assuming the GTR + + I model of nucleotide substitution (all parameter values are available from the authors on request). To assess the support for individual nodes on the tree, a bootstrap resampling analysis was undertaken, using the same substitution model but with 1,000 replicate neighbor-joining trees.

Note: Supplementary information is available on the Nature Medicine website.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.