Extensive Intra-subtype Recombination in South African Human Immunodeficiency Virus Type 1 Subtype C Infections

Rousseau, C. M., G. H. Learn, T. Bhattacharya, D. C. Nickle, D. Heckerman, S. Chetty, C. Brander, P. J. R. Goulder, B. D. Walker, P. Kiepiela, B. T. Korber, J. I. Mullins

Department of Microbiology, University of Washington, Seattle, Washington 98195-8070 USA, Los Alamos National Laboratory, Los Alamos, New Mexico 87545 USA, Machine Learning and Applied Statistics Group, Microsoft Research, Redmond, Washington 98052 USA, HIV Pathogenesis Program, Doris Duke Medical Research Institute, University of KwaZulu Natal, Durban, 4015 South Africa, Partners AIDS Research Center, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02129 USA, Nuffield Department of Medicine, The Peter Medawar Building for Pathogen Research, Oxford University, Oxford, United Kingdom OX1 3SY, and Howard Hughes Medical Institute, Chevy Chase, Maryland 20815 USA

Recombinant human immunodeficiency virus type 1 (HIV-1) strains containing sequences from different viral genetic subtypes (intersubtype) and different lineages from within the same subtype (intrasubtype) have been observed. A consequence of recombination can be the distortion of the phylogenetic signal. Several intersubtype recombinants have been identified; however, less is known about the frequency of intrasubtype recombination. For this study, near-full-length HIV-1 subtype C genomes from 270 individuals were evaluated for the presence of intrasubtype recombination. A sliding window schema (window, 2 kb; step, 385 bp) was used to partition the aligned sequences. The Shimodaira-Hasegawa test detected significant topological incongruence in 99.6% of the comparisons of the maximum-likelihood trees generated from each sequence partition, a result that could be explained by recombination. Using RECOMBINE, we detected significant levels of recombination using five random subsets of the sequences. With a set of 23 topologically consistent sequences used as references, bootscanning followed by the interactive informative site test defined recombination breakpoints. Using two multiple-comparison correction methods, 47% of the sequences showed significant evidence of recombination in both analyses. Estimated evolutionary rates were revised from 0.51%/year (95% confidence interval [CI], 0.39 to 0.53%) with all sequences to 0.46%/year (95% CI, 0.38 to 0.48%) with the putative recombinants removed. The timing of the subtype C epidemic origin was revised from 1961 (95% CI, 1947 to 1962) with all sequences to 1958 (95% CI, 1949 to 1960) with the putative recombinants removed. Thus, intrasubtype recombinants are common within the subtype C epidemic and these impact analyses of HIV-1 evolution.