Molecular Footprint of Drug-Selective Pressure in a Human Immunodeficiency Virus Transmission Chain

Lemey P, Derdelinckx I, Rambaut A, Van Laethem K, Dumont S, Vermeulen S, Van Wijngaerden E & Vandamme A-M

(2005) J Virol 79, 11981-11989.

Known human immunodeficiency virus (HIV) transmission histories are invaluable models for investigating the evolutionary and transmission dynamics of the virus and to assess the accuracy of phylogenetic reconstructions. Here we have characterized an HIV-1 transmission chain consisting of nine infected patients, almost all of whom were treated with antiviral drugs at later stages of infection. Partial pol and env gp41 regions of the HIV genome were directly sequenced from plasma viral RNA for at least one sample from each patient. Phylogenetic analyses in pol using likelihood methods inferred an evolutionary history not fully compatible with the known transmission history. This could be attributed to parallel evolution of drug resistance mutations resulting in the incorrect clustering of multidrug-resistant virus. On the other hand, a fully compatible phylogenetic tree was reconstructed from the env sequences. We were able to identify and quantify the molecular footprint of drug-selective pressure in pol using maximum likelihood inference under different codon substitution models. An increased fixation rate of mutations in the HIV population of the multidrug-resistant patient was demonstrated using molecular clock modeling. We show that molecular evolutionary analyses, guided by a known transmission history, can reveal the presence of confounding factors like natural selection and caution should be taken when accurate descriptions of HIV evolution are required.

Andrew Rambaut, 2007