Abstract Body

Background: Most mutations selected during antiretroviral therapy (ART) are non-synonymous (i.e. result in an amino acid change). However, silent mutations, K65K and K66K, in HIV-1 reverse transcriptase (RT) occur in over 35% of highly experienced patients and are among the most commonly selected mutations in subtype B strains. K65K/K66K are associated with thymidine analogue mutations (TAMs) including D67N+K70R and interrupt a homopolymeric nucleotide region introduced by D67N; however their impact in the context of HIV-1 replication is unknown. We hypothesise that silent mutations alleviate viral fitness defects conferred by D67N+K70R.

Methodology: The impact of K65K/K66K in the HIV-1 subtype B clone NL4.3 harboring the TAMs D67N+K70R on drug resistance and fitness was determined using in vitro drug susceptibility and growth competition assays. The mechanism of the fitness advantage conferred by K65K/K66K was assessed by analyzing virions produced by transfection and single-cycle infection assays, qPCR quantitation of intracellular reverse transcription products, and evaluation of cDNA synthesis efficiency by recombinant HIV-1 RT from RNA templates with D67N+K70R +/- K65K or K66K using a quantitative product enhanced reverse transcription (qPERT) assay.

Results: K65K/K66K did not alter susceptibility to zidovudine (ZDV), tenofovir (TFV), abacavir or nevirapine in MT-2 cells (p>0.05, n=4) or the infectivity of virus (normalized to p24) produced by transfection or from single-cycle infection assays. However, we found that K65K/K66K alleviated fitness defects conferred by D67N+K70R following multiple rounds of replication in the absence or presence of ZDV and TFV in MT-2 and in peripheral blood mononuclear cells. TAMs and silent mutations did not affect virion-associated RT activity or steady-state RT levels in virions produced by HIV plasmid transfection of MT-2 cells. No impact of K65K/K66K on early and late intracellular reverse transcription products in a single-cycle assay was detected. In contrast, using a highly sensitive qPERT assay, D67N+K70R in the RNA template resulted in an 8% decreased synthesis of a 111 nucleotide cDNA product compared to wild-type (WT) RNA (p=0.028, n=4) and this defect was restored by either K65K or K66K to WT levels.

Conclusions: Silent mutations in HIV-1 subtype B RT selected during ART alleviate fitness defects caused by TAMs and have no direct impact on drug susceptibility. The fitness advantage conferred by K65K/K66K is likely due to their ability to alleviate pausing by RT at homopolymeric stretches of nucleotides introduced by TAMs. Our study provides an explanation for why silent mutations are strongly selected in drug-treated individuals and supports their role as novel compensatory mutations in vivo.