Selection of Drug-Resistant Feline Immunodeficiency Virus (FIV) Encoding FIV/HIV Chimeric Protease in the Presence of HIV-Specific Protease Inhibitors

An infectious chimeric feline immunodeficiency virus (FIV)/HIV strain carrying six HIV-like protease (PR) mutations (I37V/N55M/V59I/I98S/Q99V/P100N) was subjected to selection in culture against the PR inhibitor lopinavir (LPV), darunavir (DRV), or TL-3. LPV selection resulted in the sequential emergence of V99A (strain S-1X), I59V (strain S-2X), and I108V (strain S-3X) mutations, followed by V37I (strain S-4X). Mutant PRs were analyzed in vitro, and an isogenic virus producing each mutant PR was analyzed in culture for LPV sensitivity, yielding results consistent with the original selection. The 50% inhibitory concentrations (IC₅₀s) for S-1X, S-2X, S-3X, and S-4X were 95, 643, 627, and 1,543 nM, respectively. The primary resistance mutations, V99⁸²A, I59⁵⁰V, and V37³²I, are consistent with the resistance pattern developed by HIV-1 under similar selection conditions. While resistance to LPV emerged readily, similar PR mutations causing resistance to either DRV or TL-3 failed to emerge after passage for more than a year. However, a G37D mutation in the nucleocapsid (NC) was observed in both selections and an isogenic G37D mutant replicated in the presence of 100 nM DRV or TL-3, whereas parental chimeric FIV could not. An additional mutation, L92V, near the PR active site in the folded structure recently emerged during TL-3 selection. The L92V mutant PR exhibited an IC₅₀ of 50 nM, compared to 35 nM for 6s-98S PR, and processed the NC-p2 junction more efficiently, consistent with increased viral fitness. These findings emphasize the role of mutations outside the active site of PR in increasing viral resistance to active-site inhibitors and suggest additional targets for inhibitor development.     

Structural studies of FIV and HIV-1 proteases complexed with an efficient inhibitor of FIV protease

Three forms of feline immunodeficiency virus protease (FIV PR), the wild type (wt) and two single point mutants, V59I and Q99V, as well as human immunodeficiency virus type 1 protease (HIV-1 PR), were cocrystallized with the C2-symmetric inhibitor, TL-3. The mutants of FIV PR were designed to replace residues involved in enzyme-ligand interactions by the corresponding HIV-1 PR residues at the structurally equivalent position. TL-3 shows decreased (improved) inhibition constants with these FIV PR mutants relative to wt FIV PR. Despite similar modes of binding of the inhibitor to all PRs (from P3 to P3'), small differences are evident in the conformation of the Phe side chains of TL-3 at the P1 and P1' positions in the complexes with the mutated FIV PRs. The differences mimick the observed binding of TL-3 in HIV-1 PR and correlate with a significant improvement in the inhibition constants of TL-3 with the two mutant FIV PRs. Large differences between the HIV-1 and FIV PR complexes are evident in the binding modes of the carboxybenzyl groups of TL-3 at P4 and P4'. In HIV-1 PR:TL-3, these groups bind over the flap region, whereas in the FIV PR complexes, the rings are located along the major axis of the active site. A significant difference in the location of the flaps in this region of the HIV-1 and FIV PRs correlates with the observed conformational changes in the binding mode of the peptidomimetic inhibitor at the P4 and P4' positions. These findings provide a structural explanation of the observed Ki values for TL-3 with the different PRs and will further assist in the development of improved inhibitors.     

Combining mutations in HIV-1 protease to understand mechanisms of resistance

HIV-1 develops resistance to protease inhibitors predominantly by selecting mutations in the protease gene. Studies of resistant mutants of HIV-1 protease with single amino acid substitutions have shown a range of independent effects on specificity, inhibition, and stability. Four double mutants, K45I/L90M, K45I/V82S, D30N/V82S, and N88D/L90M were selected for analysis on the basis of observations of increased or decreased stability or enzymatic activity for the respective single mutants. The double mutants were assayed for catalysis, inhibition, and stability. Crystal structures were analyzed for the double mutants at resolutions of 2.2-1.2 A to determine the associated molecular changes. Sequence-dependent changes in protease-inhibitor interactions were observed in the crystal structures. Mutations D30N, K45I, and V82S showed altered interactions with inhibitor residues at P2/P2', P3/P3'/P4/P4', and P1/P1', respectively. One of the conformations of Met90 in K45I/L90M has an unfavorably close contact with the carbonyl oxygen of Asp25, as observed previously in the L90M single mutant. The observed catalytic efficiency and inhibition for the double mutants depended on the specific substrate or inhibitor. In particular, large variation in cleavage of p6(pol)-PR substrate was observed, which is likely to result in defects in the maturation of the protease from the Gag-Pol precursor and hence viral replication. Three of the double mutants showed values for stability that were intermediate between the values observed for the respective single mutants. D30N/V82S mutant showed lower stability than either of the two individual mutations, which is possibly due to concerted changes in the central P2-P2' and S2-S2' sites. The complex effects of combining mutations are discussed.     

Structural and kinetic analysis of drug resistant mutants of HIV-1 protease.

Mutants of HIV-1 protease that are commonly selected on exposure to different drugs, V82S, G48V, N88D and L90M, showed reduced catalytic activity compared to the wild-type protease on cleavage site peptides, CA-p2, p6pol-PR and PR-RT, critical for viral maturation. Mutant V82S is the least active (2-20% of wild-type protease), mutants N88D, R8Q, and L90M exhibit activities ranging from 20 to 40% and G48V from 50 to 80% of the wild-type activity. In contrast, D30N is variable in its activity on different substrates (10-110% of wild-type), with the PR-RT site being the most affected. Mutants K45I and M46L, usually selected in combination with other mutations, showed activities that are similar to (60-110%) or greater than (110-530%) wild-type, respectively. No direct relationship was observed between catalytic activity, inhibition, and structural stability. The mutants D30N and V82S were similar to wild-type protease in their stability toward urea denaturation, while R8Q, G48V, and L90M showed 1.5 to 2.7-fold decreased stability, and N88D and K45I showed 1.6 to 1.7-fold increased stability. The crystal structures of R8Q, K45I and L90M mutants complexed with a CA-p2 analog inhibitor were determined at 2.0, 1.55 and 1.88 A resolution, respectively, and compared to the wild-type structure. The intersubunit hydrophobic contacts observed in the crystal structures are in good agreement with the relative structural stability of the mutant proteases. All these results suggest that viral resistance does not arise by a single mechanism.     

Molecular basis for the relative substrate specificity of human immunodeficiency virus type 1 and feline immunodeficiency virus proteases

We have used a random hexamer phage library to delineate similarities and differences between the substrate specificities of human immunodeficiency virus type 1 (HIV-1) and feline immunodeficiency virus (FIV) proteases (PRs). Peptide sequences were identified that were specifically cleaved by each protease, as well as sequences cleaved equally well by both enzymes. Based on amino acid distinctions within the P3-P3' region of substrates that appeared to correlate with these cleavage specificities, we prepared a series of synthetic peptides within the framework of a peptide sequence cleaved with essentially the same efficiency by both HIV-1 and FIV PRs, Ac-KSGVF/VVNGLVK-NH(2) (arrow denotes cleavage site). We used the resultant peptide set to assess the influence of specific amino acid substitutions on the cleavage characteristics of the two proteases. The findings show that when Asn is substituted for Val at the P2 position, HIV-1 PR cleaves the substrate at a much greater rate than does FIV PR. Likewise, Glu or Gln substituted for Val at the P2' position also yields peptides specifically susceptible to HIV-1 PR. In contrast, when Ser is substituted for Val at P1', FIV PR cleaves the substrate at a much higher rate than does HIV-1 PR. In addition, Asn or Gln at the P1 position, in combination with an appropriate P3 amino acid, Arg, also strongly favors cleavage by FIV PR over HIV PR. Structural analysis identified several protease residues likely to dictate the observed specificity differences. Interestingly, HIV PR Asp30 (Ile-35 in FIV PR), which influences specificity at the S2 and S2' subsites, and HIV-1 PR Pro-81 and Val-82 (Ile-98 and Gln-99 in FIV PR), which influence specificity at the S1 and S1' subsites, are residues which are often involved in development of drug resistance in HIV-1 protease. The peptide substrate KSGVF/VVNGK, cleaved by both PRs, was used as a template for the design of a reduced amide inhibitor, Ac-GSGVF Psi(CH(2)NH)VVNGL-NH(2.) This compound inhibited both FIV and HIV-1 PRs with approximately equal efficiency. These findings establish a molecular basis for distinctions in substrate specificity between human and feline lentivirus PRs and offer a framework for development of efficient broad-based inhibitors.     

Replicative fitness of protease inhibitor-resistant mutants of human immunodeficiency virus type 1

The relative replicative fitness of human immunodeficiency virus type 1 (HIV-1) mutants selected by different protease inhibitors (PIs) in vivo was determined. Each mutant was compared to wild type (WT), NL4-3, in the absence of drugs by several methods, including clonal genotyping of cultures infected with two competing viral variants, kinetics of viral antigen production, and viral infectivity/virion particle ratios. A nelfinavir-selected protease D30N substitution substantially decreased replicative capacity relative to WT, while a saquinavir-selected L90M substitution moderately decreased fitness. The D30N mutant virus was also outcompeted by the L90M mutant in the absence of drugs. A major natural polymorphism of the HIV-1 protease, L63P, compensated well for the impairment of fitness caused by L90M but only slightly improved the fitness of D30N. Multiply substituted indinavir-selected mutants M46I/L63P/V82T/I84V and L10R/M46I/L63P/V82T/I84V were just as fit as WT. These results indicate that the mutations which are usually initially selected by nelfinavir and saquinavir, D30N and L90M, respectively, impair fitness. However, additional mutations may improve the replicative capacity of these and other drug-resistant mutants. Hypotheses based on the greater fitness impairment of the nelfinavir-selected D30N mutant are suggested to explain observations that prolonged responses to delayed salvage regimens, including alternate PIs, may be relatively common after nelfinavir failure.     

Molecular mechanics analysis of drug-resistant mutants of HIV protease.

Drug-resistant mutants of HIV-1 protease limit the long-term effectiveness of current anti-viral therapy. In order to study drug resistance, the wild-type HIV-1 protease and the mutants R8Q, V32I, M46I, V82A, V82I, V82F, I84V, V32I/I84V and M46I/I84V were modeled with the inhibitors saquinavir and indinavir using the program AMMP. A new screen term was introduced to reproduce more correctly the electron distribution of atoms. The atomic partial charge was represented as a delocalized charge distribution instead of a point charge. The calculated protease-saquinavir interaction energies showed the highly significant correlation of 0.79 with free energy differences derived from the measured inhibition constants for all 10 models. Three different protonation states of indinavir were evaluated. The best indinavir model included a sulfate and gave a correlation coefficient of 0.68 between the calculated interaction energies and free energies from inhibition constants for nine models. The exception was R8Q with indinavir, probably due to differences in the solvation energy. No significant correlation was found using the standard molecular mechanics terms. The incorporation of the new screen correction resulted in better prediction of the effects of inhibitors on resistant protease variants and has potential for selecting more effective inhibitors for resistant virus.     

Altered gag polyprotein cleavage specificity of feline immunodeficiency virus/human immunodeficiency virus mutant proteases as demonstrated in a cell-based expression system

We have used feline immunodeficiency virus (FIV) protease (PR) as a mutational system to study the molecular basis of substrate-inhibitor specificity for lentivirus PRs, with a focus on human immunodeficiency virus type 1 (HIV-1) PR. Our previous mutagenesis studies demonstrated that discrete substitutions in the active site of FIV PR with structurally equivalent residues of HIV-1 PR dramatically altered the specificity of the mutant PRs in in vitro analyses. Here, we have expanded these studies to analyze the specificity changes in each mutant FIV PR expressed in the context of the natural Gag-Pol polyprotein ex vivo. Expression mutants were prepared in which 4 to 12 HIV-1-equivalent substitutions were made in FIV PR, and cleavage of each Gag-Pol polyprotein was then assessed in pseudovirions from transduced cells. The findings demonstrated that, as with in vitro analyses, inhibitor specificities of the mutants showed increased HIV-1 PR character when analyzed against the natural substrate. In addition, all of the mutant PRs still processed the FIV polyprotein but the apparent order of processing was altered relative to that observed with wild-type FIV PR. Given the importance of the order in which Gag-Pol is processed, these findings likely explain the failure to produce infectious FIVs bearing these mutations.     

Structure/function mapping of amino acids in the N-terminal zinc finger of the human immunodeficiency virus type 1 nucleocapsid protein: residues responsible for nucleic acid helix destabilizing activity

The nucleocapsid protein (NC) of HIV-1 is 55 amino acids in length and possesses two CCHC-type zinc fingers. Finger one (N-terminal) contributes significantly more to helix destabilizing activity than finger two (C-terminal). Five amino acids differ between the two zinc fingers. To determine at the amino acid level the reason for the apparent distinction between the fingers, each different residue in finger one was incrementally replaced by the one at the corresponding location in finger two. Mutants were analyzed in annealing assays with unstructured and structured substrates. Three groupings emerged: (1) those similar to wild-type levels (N17K, A25M), (2) those with diminished activity (I24Q, N27D), and (3) mutant F16W, which had substantially greater helix destabilizing activity than that of the wild type. Unlike I24Q and the other mutants, N27D was defective in DNA binding. Only I24Q and N27D showed reduced strand transfer in in vitro assays. Double and triple mutants F16W/I24Q, F16W/N27D, and F16W/I24Q/N27D all showed defects in DNA binding, strand transfer, and helix destabilization, suggesting that the I24Q and N27D mutations have a dominant negative effect and abolish the positive influence of F16W. Results show that amino acid differences at positions 24 and 27 contribute significantly to finger one's helix destabilizing activity.     

Structural insights into the mechanisms of drug resistance in HIV-1 protease NL4-3

The development of resistance to anti-retroviral drugs targeted against HIV is an increasing clinical problem in the treatment of HIV-1-infected individuals. Many patients develop drug-resistant strains of the virus after treatment with inhibitor cocktails (HAART therapy), which include multiple protease inhibitors. Therefore, it is imperative that we understand the mechanisms by which the viral proteins, in particular HIV-1 protease, develop resistance. We have determined the three-dimensional structure of HIV-1 protease NL4-3 in complex with the potent protease inhibitor TL-3 at 2.0 A resolution. We have also obtained the crystal structures of three mutant forms of NL4-3 protease containing one (V82A), three (V82A, M46I, F53L) and six (V82A, M46I, F53L, V77I, L24I, L63P) point mutations in complex with TL-3. The three protease mutants arose sequentially under ex vivo selective pressure in the presence of TL-3, and exhibit fourfold, 11-fold, and 30-fold resistance to TL-3, respectively. This series of protease crystal structures offers insights into the biochemical and structural mechanisms by which the enzyme can overcome inhibition by TL-3 while recovering some of its native catalytic activity.     

Structure-activity studies of FIV and HIV protease inhibitors containing allophenylnorstatine

The interaction of P1 and P3 side chains with the combining S1 and S3 hydrophobic subsites of HIV and FIV proteases has been explored using asymmetric competitive inhibitors. The inhibitors evaluated contained (2S,3S)-3-amino-2-hydroxy-4-phenylbutyric acid (allophenylnorstatine) as the hydroxymethylcarbonyl isostere, (R)-5,5-dimethyl-1, 3-thiazolidine-4-carbonyl as P1', Val as P2 and P2' residues, and a variety of amino acids at the P3 and P3' positions. All inhibitors showed competitive inhibition of both enzymes with higher potency against the HIV protease in vitro. Within this series, 31 (VLE776) is the most effective inhibitor against FIV protease, and it contains Phe at P3, but no P3' residue. VLE776 also exhibited potent antiviral activities against the drug-resistant HIV mutants (G48V and V82F) and the TL3-resistant HIV mutants. Explanation of the inhibition activities was described. In addition, a new strategy was described for development of bifunctional inhibitors, which combine the protease inhibitor and another enzyme inhibitor in one molecule.