The R262K substitution combined with H51Y in HIV-1 subtype B integrase confers low-level resistance against dolutegravir

Abstract

Clinical studies have shown that integrase strand transfer inhibitors (INSTIs) can be used effectively against HIV-1 infection. To date, no resistance substitution has been found in INSTI-naive patients treated with the new integrase inhibitor dolutegravir (DTG). In a recent selection study with DTG, using a virus bearing the H51Y substitution in integrase, the emergence of an R to K substitution at position 262 (R262K) was observed. We characterized this double mutant with respect to integrase strand transfer activity and susceptibility to DTG both biochemically and in tissue culture. We showed that the addition of R262K to H51Y decreased recombinant integrase strand transfer activity but improved integrase DNA-binding affinity, compared to wild-type or H51Y-containing enzymes. The defect in strand transfer activity did not translate into a decrease in HIV-1 infectivity. The combination of H51Y and R262K substitutions slightly decreased susceptibility to DTG (fold change = 1.87) in cell-based resistance assays. Although viral replication was not affected and enzyme efficiency was impaired by the addition of R262K to H51Y, there was an overall increase in the level of biochemical drug resistance against DTG. Our findings suggest that the R at position 262 plays an important role in DNA binding.

FIG 1

Effects of the addition of R262K to H51Y on integrase strand transfer activity. (A) Relative strand transfer activities with various recombinant protein concentrations. Columns represent the means and standard errors of three independent experiments. (B) Relative strand transfer activities in the presence of various target DNA concentrations. Points represent the means and standard errors of seven independent experiments. (C) K_m values. Columns represent the means and standard errors of seven independent observations. (D) Recombinant enzyme efficiencies, as determined by dividing V_max by K_m. Columns represent the means and standard errors of seven independent experiments.

FIG 2

Effects of the addition of R262K to H51Y on integrase 3′ processing activity. (A) 3′ processing activity with various concentrations of viral LTR. Points represent the means and standard errors of two independent experiments. (B) Recombinant enzyme efficiency, as determined by the ratio of V_max to K_m. Columns represent the means and standard errors of two independent experiments.

FIG 3

Effects of the addition of R262K to H51Y on viral infectivity. The effects of the WT, H51Y, and H51Y/R262K substitutions on HIV infectivity in TZM-bl cells were measured with various amounts of virus. Points represent the means and standard errors of two independent experiments.

FIG 4

Effects of the addition of R262K to H51Y on susceptibility to INSTIs. (A) Resistance curves for WT, H51Y, and H51Y/R262K viruses with various concentrations of RAL. Points represent the means and standard errors of two independent experiments. (B) Resistance curves for WT, H51Y, and H51Y/R262K viruses with various concentrations of DTG. Points represent the means and standard errors of two independent experiments.

FIG 5

Homology model of HIV-1 integrase subtype B. A homology model was generated based on the crystal structure of the target capture complex of PFV (PDB identification no. 4E7K). Secondary structure overlays and catalytic triad orientations and positions were examined for WT and mutant HIV-1 integrases. (A) Global overlay of the IN_WT, IN_H51Y, and IN_R262K/H51Y models. (B and C) Close-up overlays, showing the relative positions of the catalytic triad of D64, D116, and E152 in IN_WT (B and C), IN_H51Y (B), and IN_R262K/H51Y (C) coordinating Mn²⁺ ions. Mn²⁺ ion positions were simulated by inserting their coordinates from the PFV structure (PDB identification no. 4E7K) in the active site, and the ions are represented as small black spheres. Coordinates of the viral LTR and target DNA were inserted in the active site of the integrase homology models by overlap of the PFV crystal structure (PDB identification no. 4E7K). All images were processed using PyMOL software. Diagrams and carbon atom coloration differentiate the IN_WT (white), IN_H51Y (black), and IN_R262K/H51Y (gray) homology models.