Altered substrate specificity of drug-resistant human immunodeficiency virus type 1 protease

Abstract

Resistance to human immunodeficiency virus type 1 protease (HIV PR) inhibitors results primarily from the selection of multiple mutations in the protease region. Because many of these mutations are selected for the ability to decrease inhibitor binding in the active site, they also affect substrate binding and potentially substrate specificity. This work investigates the substrate specificity of a panel of clinically derived protease inhibitor-resistant HIV PR variants. To compare protease specificity, we have used positional-scanning, synthetic combinatorial peptide libraries as well as a select number of individual substrates. The subsite preferences of wild-type HIV PR determined by using the substrate libraries are consistent with prior reports, validating the use of these libraries to compare specificity among a panel of HIV PR variants. Five out of seven protease variants demonstrated subtle differences in specificity that may have significant impacts on their abilities to function in viral maturation. Of these, four variants demonstrated up to fourfold changes in the preference for valine relative to alanine at position P2 when tested on individual peptide substrates. This change correlated with a common mutation in the viral NC/p1 cleavage site. These mutations may represent a mechanism by which severely compromised, drug-resistant viral strains can increase fitness levels. Understanding the altered substrate specificity of drug-resistant HIV PR should be valuable in the design of future generations of protease inhibitors as well as in elucidating the molecular basis of regulation of proteolysis in HIV.

FIG. 1.

Substrate preferences for HIV PR. (A) Initial rates for HIV PR with the P1-diverse library indicate the relative preferences for residues at P1. Rates are expressed as relative fluorescent units per second. (B) Preferences at P2 through P4 were assayed using three P1-fixed libraries. Data bars are colored as specified in the key in the P2 panel. All values are reported as initial rates relative to either P3-Leu (for the P1-Met and P1-Nle libraries) or P3-Trp (for the P1-Leu library). Since a single experiment collects data on all three subsites simultaneously, the y-axis scale is consistent among the three panels.

FIG. 2.

(A) Alignment of all protease sequences used in this study. All sequences include NC/p1 and p1/p6 cleavage site sequences except for WT, the synthetic wild-type variant. Mutations associated with protease inhibitor resistance are shown against a green background; polymorphic mutations are shown against a blue background. The Q7K mutation in WT, highlighted in gray, was engineered to stabilize the protease to autoproteolysis. (B) In vivo exposure to protease inhibitors and phenotypic resistance profiles are shown for the drug-resistant protease panel. Inhibitor susceptibilities are expressed as 50% inhibitory concentrations relative to that of a reference control.

FIG. 3.

The substrate preferences of HIV PR and drug-resistant protease variants were compared using P1-Met library results. All experiments have been normalized to Nle at P3 in order to facilitate comparison.

FIG. 4.

Comparison of P1-Met data for the WT, mdr, and 954 proteases at position P4. Rates are shown only for residues with altered preferences.