Interplay between single resistance-associated mutations in the HIV-1 protease and viral infectivity, protease activity, and inhibitor sensitivity

Abstract

Resistance-associated mutations in the HIV-1 protease modify viral fitness through changes in the catalytic activity and altered binding affinity for substrates and inhibitors. In this report, we examine the effects of 31 mutations at 26 amino acid positions in protease to determine their impact on infectivity and protease inhibitor sensitivity. We found that primary resistance mutations individually decrease fitness and generally increase sensitivity to protease inhibitors, indicating that reduced virion-associated protease activity reduces virion infectivity and the reduced level of per virion protease activity is then more easily titrated by a protease inhibitor. Conversely, mutations at more variable positions (compensatory mutations) confer low-level decreases in sensitivity to all protease inhibitors with little effect on infectivity. We found significant differences in the observed effect on infectivity with a pseudotype virus assay that requires the protease to cleave the cytoplasmic tail of the amphotropic murine leukemia virus (MuLV) Env protein. Additionally, we were able to mimic the fitness loss associated with resistance mutations by directly reducing the level of virion-associated protease activity. Virions containing 50% of a D25A mutant protease were 3- to 5-fold more sensitive to protease inhibitors. This level of reduction in protease activity also resulted in a 2-fold increase in sensitivity to nonnucleoside inhibitors of reverse transcriptase and a similar increase in sensitivity to zidovudine (AZT), indicating a pleiotropic effect associated with reduced protease activity. These results highlight the interplay between enzyme activity, viral fitness, and inhibitor mechanism and sensitivity in the closed system of the viral replication complex.

Fig 1

(A) Silver staining showing that p24 ELISA measures equal amounts of Gag products from the wild-type and protease mutant virion. An SDS-polyacrylamide gel was loaded with equal masses of p24, as determined by antigen-capture ELISA. The band signals were compared for the parental NL4-3 virus (lanes 1 to 3) and the poorly infectious G48V mutant (lanes 4 to 6). Threefold dilutions of each mutant were shown (lanes 1 and 4, 150 ng of p24; lanes 2 and 5, 50 ng of p24; lanes 3 and 6, 17 ng of p24). (B) Correlation analysis between SpIn values measured using the p24 ELISA and SpIn values measured using real-time RT PCR. The r² value was obtained from the statistical analysis using nonparametric (Spearman) correlation.

Fig 2

Infectivities of HIV-1 bearing single resistance-associated mutations in pro measured by two separate single-cycle assays. (A) SpIn assay, in which the amount of viruses used for infection was normalized by measuring either p24 mass (ELISA) or viral RNA (real-time RT-PCR). All the values are normalized to that for the parental NL4-3 strain, which is given an infectivity of 1.0. Data from either two or more than two infection values were averaged, and the infectivity data are shown as means ± standard deviations. Standard deviation values for L10I, K20R, and I50V are not determined. (B) RC assay in which the viruses were pseudotyped with the MuLV amphotropic Env protein. Since a luciferase expression cassette is inserted within a deleted region of the HIV-1 env gene, the ratio of luciferase activity in the infected cells to that in the transfected cells was used to normalize viral infectivity. The variability of the RC assay has been validated to be ±0.2 log.

Fig 3

Western blot analysis of the selected protease mutant virus particles pseudotyped with the MuLV amphotropic Env protein. Each infectious HIV-1 clone containing the indicated single resistance-associated mutation in the pro region was used along with the MuLV amphotropic Env plasmids to transfect 293 cells. Culture supernatants were harvested at 48 h after transfection and were subjected to ultracentrifugation to concentrate the viral particles. The pelleted viral particles were detected using either anti-MuLV p15E antibody or anti-HIV-1 CA antibody as the primary antibody. For comparison, MuLV particles and HIV-1 protease active-site mutant (D25A) are shown. Unprocessed p15E protein was not detected from the MuLV particles, and in contrast, only unprocessed p15E protein was detected from the D25A mutant.

Fig 4

FC EC₅₀s of mutants to protease inhibitors. The EC₅₀ was measured for each mutant virus with each of the seven approved protease inhibitors (APV, ATV, IDV, LPV, NFV, RTV, and SQV) and compared to the EC₅₀ of the parental virus to calculate the FC EC₅₀. Values are presented on a log scale as the average FC EC₅₀ (filled circles), with error bars indicating the standard deviation of these values. Those FC EC₅₀ values which are greater than 2 standard deviations from the mean of all seven FC EC₅₀s are indicated by open circles (K20I, NFV; D30N, NFV; G48V, SQV; V82A, RTV; V82T, RTV; and N88S, APV), and they are excluded from the calculation of the average FC EC₅₀ and standard deviation.

Fig 5

Comparison of average FC EC₅₀ between class 2 and class 3 mutations. The solid horizontal lines represent the median FC EC₅₀ in each group. The mean FC EC₅₀ values between class 2 and class 3 mutations were compared by statistical analysis using a Mann-Whitney test (nonparametric). The outliers shown as open circles in Fig. 4 are not included in the analysis.

Fig 6

Effect of nonspecific loss of protease activity on viral infectivity and inhibitor sensitivity. (A) Protease activity was reduced in a nonspecific way by transfecting the infectious molecular clone with an isogenic clone containing an inactivating mutation at the protease active site (D25A). Both the fraction of the D25A mutant and the calculated remaining protease activity in each preparation are indicated. The relative infectivity of each mixture was measured using the SpIn (filled circles) and RC (open circles) assays. (B) The effect of nonspecific loss of protease activity on sensitivity to inhibitors of HIV replication was studied by measuring the EC₅₀ for nucleoside reverse transcriptase inhibitors (tenofovir, AZT), nonnucleoside reverse transcriptase inhibitors (EFV), and protease inhibitors (LPV, darunavir [DRV]). Increasing the amount of mutant protease up to 50% (equivalent to 25% residual protease activity) had a systematic effect on the EC₅₀ value for the three classes of antiviral drugs tested. (C) The FC EC₅₀ of 50% D25A relative to the FC EC₅₀ of 0% D25A is shown for all the inhibitors tested. The values of three classes of inhibitors were compared by statistical analysis using a nonparametric Kruskal-Wallis test. The open arrowhead indicates the value from AZT. (D) Western blot analysis showing the extent of processing of p24 and RT in virus particle preparations containing either 100% WT or 50% inactive mutant protease.