The HIV-1 protease (PR) enzyme cleaves the Gag and Gag-Pol polyproteins to produce the viral enzymes and structural proteins necessary for the release of mature virion particles.1 It is composed of 2 noncovalently associated sequence and structurally identical monomers and contains an active site, which includes an Asp-Thr-Gly motif at residue positions 25 to 27.1 Protease inhibitor (PI) resistance mutations have been described in the substrate cleft of the enzyme that impair the binding between the inhibitor and the mutant PR.2-4 Some major primary mutations are seen frequently in patients failing highly active antiretroviral therapy (HAART), including 1 or more PIs.5 In addition, secondary mutations in the PR gene and in Gag-Pol cleavage site regions are often selected, which seem to play a major role in the recovery of viral replicative fitness after the selection of primary mutations.6-11 Rare amino acid substitutions can also occur, and their role as natural polymorphic variants or as minor mutations required for resistance to emerge in vivo is a subject of investigation.12
We observed an unusual glycine-to-glutamate substitution at PR residue position 48 (G48E) in sequential isolates from an African patient infected with a subtype A1 HIV-1 strain. This amino acid substitution was observed after saquinavir (SQV) had been introduced in the treatment regimen of the patient. Resistance to SQV is usually associated with a glycine-to-valine substitution at position 48 (G48V), located on the tip of the PR flap, a region that extends into the substrate cleft. G48V frequently follows the development of a valine-to-alanine change at position 82 (V82A). Other mutations at position 48 are extremely rare, particularly in untreated patients.13
In this study, we performed a longitudinal analysis of the complete PR and Gag-Pol cleavage site regions throughout the course of infection of the patient. We showed the impact of G48E on viral replicative fitness and PI susceptibility on the basis of in vitro phenotypic evaluation. Finally, using a molecular dynamics (MD) simulation approach, we investigated whether G48E might have an impact on HIV-1 replicative capacity and how several polymorphisms contained in the PR flap regions of the patient's virus might compensate the fitness cost of G48E.
MATERIALS AND METHODS
Population Nucleotide Sequencing of the Patient's Virus
Viral RNA was extracted from plasma samples of the patient and directly sequenced on an automated ABIPrism 3130 (Applied Biosystems, Foster city, CA). For sequencing the PR and the gag and pol cleavage site regions, we used Viroseq technology (Abbott N.V. Diagnostics, Louvain-la-Neuve, Belgium) and BigDye Terminator v3.1 chemistry (PE Applied Biosystems, Foster City, CA) according to previously described protocols.14,15 Deduced amino acid sequences were compared with the subtype B HXB2 reference sequence. An additional comparison was made with a subtype A1 reference sequence for the Gag-Pol cleavage site sequences.
G48E, V82A, and G48E/V82A mutations were inserted into a plasmid encoding the polymerase gene of NL4 to 3 (pUC19-NL4 to 3, a gift from the Rega Institute, Katholieke Universiteit Leuven, Leuven, Belgium) by site-directed mutagenesis using the following primers: 5′-gatggaaaccaaaaatgatagagggaattgg-3′ (sense primer, positions 2374 to 2404 of the HXB2 genome) and 5′-actttgataaaacctccaattccctctatca-3′ (antisense primer, HXB2 positions 2419 to 2389) for G48E and 5′-gtaggacctacacctgccaacataattggaagaaatctg-3′ (sense primer, HXB2 positions 2481 to 2519) and 5′-cagatttcttccaattatgttggcaggtgtaggtcctac-3′ (antisense primer, HXB2 positions 2519 to 2481) for V82A. V82A mutant was created, because V82A is a common mutation and we were interested to investigate its fitness effects in the settings of our experiment. Mutagenesis products were transformed to One-shot Escherichia coli electrocompetent cells (Invitrogen, Merelbeke, Belgium) after digestion of the parental plasmid DNA with DpnI. After overnight incubation on Luria-Burtani (LB) agar plates, clones were cultured in 5 mL of LB medium and plasmid DNA was extracted using the QIAquick Minispin extraction kit (QIAGEN, KJ Venlo, The Netherlands). The presence of desired mutations was screened for by sequencing plasmids on an automated ABIPrism3100, using BigDye Terminator cycle-sequencing chemistry and RVP5 (sense, 5′-gggaagatctggccttcctacaaggg-3′, HXB2 positions 2092 to 2117) and RVP3 (antisense, 5′-ggcaaatactggagtattgtatgg-3′, HXB2 positions 2712 to 2735) as sequencing primers.
Phenotypic Analysis of PI Susceptibility and Replicative Fitness
After detection of clones incorporating the desired mutation, a polymerase chain reaction (PCR) assay was performed on mutant plasmid DNA using RVP5 and RVP3 to amplify 620-base pair (bp) PR fragments. PR fragments were then recombined into a proviral plasmid deleted of the PR gene, as described previously.16 Recombinant viruses were used to infect MT4 cells. After detection of cytopathic effects (CPEs) in culture, viral supernatants of G48E, V82A, and G48E/V82A mutants were titrated and analyzed by direct sequencing. Phenotypic resistance tests (ie, inhibitory concentrations [IC50] values, defined as the concentration of drug inhibiting 50% of HIV-1 replication in MT4 cells) were performed using a standard 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) procedure as described.17
In addition, for every mutant, pol recombinant viruses were generated by cotransfection of PCR-amplified pol gene (carrying G48E, V82A, or G48E/V82A mutations) with a pol-deleted proviral plasmid tagged with the enhanced green fluorescent protein (EGFP) or Discosoma spp. red fluorescent protein (DsRed2) as described.18 Viral production was monitored every 3 to 4 days for CPEs, and supernatant was analyzed using an in-house reverse transcriptase (RT) assay. Culture supernatants were harvested, cleared by low-speed centrifugation, filtered using a 0.45-μm sterile vacuum filtration system, and aliquoted. Tissue culture dose for 50% infectivity (TCID50) was determined as described previously.19
Viral replicative fitness was individually evaluated using viral growth kinetics in MT-4 cells and by growth competition experiments in peripheral blood mononuclear cells (PBMCs) against wild-type NL4-3 strain. To estimate growth kinetics 3 × 104 MT-4 cells were infected separately in 96-well plates in triplicate with each virus at a multiplicity of infection (MOI) of 0.01 IU/cell in 200 μL of RPMI 1640 and incubated for 2 hours at 37°C in 5% CO2. Infected cells were then washed 3 times with phosphate-buffered saline (PBS) and cultured in growth medium. RT activity in the supernatant was measured on day 2 after infection and every day thereafter. RT activity was also used for evaluating susceptibility of the G48E, V82A, or G48E/V82A mutants to PIs.
In growth competition experiments, EGFP- or DsRed2-tagged viruses competed with a wild-type NL4-3 strain in a 1:1 initial proportion using a MOI of 0.01 IU/cell. One milliliter of these virus mixtures was incubated with 1 × 106 PBMCs for 2 hours at 37°C, 5% CO2. Cells were subsequently washed 3 times with 1 × PBS and then resuspended in culture medium (1 × 106 cells/mL). Dual infections were monitored daily using a Leica DMIRB inverted, upright, wide-field fluorescence microscope (Heidelberg, Germany). After the detection of green and/or red fluorescent cells, aliquots of HIV-infected cultures were homogenized and diluted in 1 × PBS until reaching a single-cell monolayer in a 96-well plate. Each well was individually analyzed, and 4 representative pictures per well were obtained using a MicroMax 5-MHz charge-coupled device (CCD) digital camera (Princeton Instruments, Trenton, NJ). Images were processed with Image-Pro Plus 5.0 software (Media Cybernetics, Silver Springs, MD), allowing the quantification of green and red cells in each competition by measuring their area and mean intensity. To determine HIV-1 replicative fitness, the final proportion of the 2 viruses produced in each growth competition experiment was quantified by fluorescence microscopy after normalizing to viral production in the HIV-1 monoinfections as described elsewhere.18
Molecular Dynamics Simulation
All MD simulations were carried out using the GROMACS software package (University of Groningen, Groningen, the Netherlands).20 The simulations were performed in the presence of explicit water and counter ions. Periodic boundaries conditions were applied in all directions. The applied force field was the GROMOS96 force field. Initialization of the MD simulation included an energy minimization of 500 steps and a 20-picosecond MD simulation with harmonic positional restraints on the solute. The final production MD simulations were run for 2 nanoseconds at 300 K, keeping pressure constant at a reference of 1 bar. All calculations were done on an Apple X-serve cluster with 12 nodes.
Evolutional Characteristics of the PR and Gag-Pol Cleavage Site Regions of the Subtype A1 Patient's Virus
Variation of the PR and Gag-Pol cleavage site regions is shown Table 1. Natural PR non-B polymorphisms (L10I, I13V, E35D, M36I, R57K, H69K, and L89M) were present at baseline and throughout the course of infection. The PR region exhibited increasing variability over time. Although several mutations were attributable to PI-based regimen pressure (M46I, I54V, Q58E, I62V, L63P, V82F, and L90M), other polymorphisms were classically not related to resistance to PI (K20I, E35N, K55Q, Q61E, and L89I). The G48E mutation was first detected 3 months after the introduction of SQV in the patient's treatment regimen and persisted after removal of SQV (see Table 1). At baseline, Gag-Pol cleavage site regions were comparable to a subtype A1 reference sequence. Several mutations were later observed, among which were an A-to-V substitution at residue P2 of the conserved NC/p1-NC/TFP cleavage site before the development of G48E in the patient's virus and a P-to-L change at residue P5′ of the p1/p6 cleavage site after the emergence of G48E.
Replication and Drug Susceptibility of Viruses Carrying the G48E, V82A, and G48E/V82A Mutation in the PR Gene
Replication characteristics of each site-directed mutant using the standard MTT assay are summarized in Table 2. After 8 days of viral growth, CPEs were observed for the V82A mutant. The valine-to-alanine substitution at residue position 82 was confirmed after sequencing of the viral supernatant. The virus was titrated at 2795 TCID50/mL. The G48E mutant was cultured for 7 days until CPEs were observed. The titer was determined at 2795 TCID50/mL. Direct sequencing of the viral supernatant revealed the appearance of a G/E mix at position 48, however, and the presence of a second mutation, K20M. Three G48E/V82A double-mutant viruses, namely, G48E/V82A-1, G48E/V82A-2, and G48E/V82A-3, were grown in vitro for 36, 46, and 60 days, respectively, until CPEs were observed. For all mutants, the V82A mutation was conserved at the time of titration. The titer of G48E/V82A-1 was calculated at 559 TCID50/mL. Similar to the G48E mutant, direct sequencing of the viral supernatant revealed a G/E mix at position 48, which was confirmed after an additional 5-day virus culture in vitro. Furthermore, a mixed viral population was observed at residue positions 20 and 34 (20K/M and 34E/K, respectively). The titer of G48E/V82A-2 was 559 TCID50/mL, and the sequence analysis revealed the presence of mixed viral populations at residue positions 20 (K/M), 48 (G/E), and 71 (A/V). G48E/V82A-3 exhibited a slower growth in vitro and a lower titer of 22 TCID50/mL. Interestingly, the G48E mutation was predominant in the viral supernatant and was associated with the emergence of an E21K mutation. When G48E/V82A-3 was grown for 15 additional days, however, the virus exhibited progressive loss of the E at position 48, in parallel with a higher titer in culture (559 TCID50/mL).
Finally, the susceptibility of V82A and G48E mutants to nelfinavir (NFV), lopinavir (LPV), ritonavir (RTV), indinavir (IDV), amprenavir (APV), and SQV was determined (Fig. 1). As described elsewhere in this article, the G48E mutant virus was characterized by impaired replication as compared with the NL4-3 virus control (ie, low RT activity in the absence of PI) and exhibited slightly decreased susceptibility to SQV (2-fold when compared with the wild-type virus). The G48E substitution showed no effect on the susceptibility to other PIs in vitro, however.
We evaluated the replicative fitness of viruses carrying G48E, V82A, or G48E/V82A mutations in the PR gene using a novel assay based on recombinant viruses expressing the EGFP or the DsRed2 protein in an HIV-1 NL4-3 backbone.18 EGFP- and DsRed2-tagged viruses with G48E and V82A mutations were generated. Attempts to create recombinant virus with G48E/V82A mutations resulted in virus carrying only the V82A mutation. Replicative fitness was initially evaluated using a viral growth kinetics assay (Fig. 2A). Results showed that the G48E mutant virus replicated with reduced kinetics compared with the wild-type NL4-3 strain and V82A mutant virus (at day 3 and day 4; P < 0.001, t test).
To confirm the effect of the G48E substitution on HIV replicative fitness, we performed growth competition experiments in PBMCs using 4 fluorescent viruses: G48E-EGFP versus NL4-3-DsRed2 and G48E-DsRed2 versus NL4-3-EGFP, respectively. The mean replicative fitness of the G48E mutant virus (1.14 ± 0.67 [mean ± SD], n = 6) was reduced to 55% compared with the wild-type NL4 to 3 (2.07 ± 1.07; n = 6; P = 0.041, Mann-Whitney rank sum test; see Fig. 2B).
Molecular Dynamics Simulation of the G48E Mutant
Root-mean-square deviation (RMSD) plots were obtained for the PR sequences of residues 35 to 63 (corresponding to the PR flap) of (1) the wild-type NL4-3 reference strain (based on the protein data bank (PDB) code: 5HVP, in blue in Fig. 3), (2) the site-directed mutant G48E mutant in the context of the wild type (in red in Fig. 3), and (3) the subtype A1 clinical isolate of the patient (in green in Fig. 3). RMSD plots (represented on the y axis) are an indication of the flexibility of the PR flap, on the tip of which the 48 residue is located. Over 2-nanosecond simulations, the wild type is much more flexible, compared with the G48E mutant, when site-directed in an NL4-3 backbone. The patient isolate, corresponding to a sample from April 11, 1999 (see Table 1 for complete amino acid composition), displays an intermediate flexibility (see Fig. 3).
We have demonstrated that G48E, when introduced in the context of a subtype B HIV-1 strain, is associated with moderate resistance to SQV but is also highly unstable, giving rise to viruses with diminished replicative fitness in vitro. Mixed viral subpopulations at position 48 appeared after 7 days in the viral cell culture of the single G48E mutant. Viral incubations were particularly long for the double mutant G48E/V82A viruses until CPEs were observed. Furthermore, for all investigated double mutants, viral supernatants were similarly characterized by the progressive loss of the G48E substitution, which could result from pure reversal to the wild-type form or from wild-type overgrowth in viral cultures. Replicative fitness experiments confirmed that the ability of the G48E virus to replicate in the absence of drug pressure was reduced to 55% as compared with the parental NL4-3 virus. Despite the reduced replicative fitness observed with G48E in our site-directed mutagenesis experiments with a subtype B backbone, however, the presence of G48E in clinical isolates attests that the mutant is viable in subtype A1 HIV-1 under the treatment conditions present in the patient investigated in our study.
Unlike other residue positions linked to PR inhibitor resistance, the codon 48 seems critical for the viability of the virus. Although G48V alone is sufficient for significant SQV resistance, it markedly affects viral infectivity by more than 50%.21 The low-replication kinetics observed when G48E is introduced in the prototype B subtype NL4-3 strain may be explained by the absence of adequate mutations with compensatory effect. In rare circumstances, G48E has been observed in subtype B HIV-1 clinical isolates, with this mutation always being accompanied by other mutations, possibly compensating for replication defects.22-29 Fast reversal to a wild-type form because of the low genetic barrier (only 1 transition from G [ggg] to E [gag] at position 48) might explain the lack of stability of G48E in the absence of compensatory mechanisms.
The restoration of the flexibility of the PR flaps is one of those mechanisms. We were interested in exploring whether polymorphisms that were present in the PR of the subtype A1 patient's virus could contribute to such a compensatory mechanism, when added to a single G48E mutation. To avoid multiple syntheses of site-directed mutants and time-consuming viral culture experiments, however, we used an in silico approach for testing such an effect. The MD simulations suggest that the single G48E mutation drastically alters such flexibility when introduced in a wild-type subtype B backbone. Previous MD work had already indicated that PR flap movements were central to the function of the enzyme, and hence to the viability of the virus. The tip of those flaps (including PR residues 48 to 52) must curl in and bury the hydrophobic tip, Ile-50, against the inside wall of the active site so as to open enough space for PR substrates.30 Using MD simulations, we observed that the loss of PR flap flexibility caused by the G48E mutation was partially restored when polymorphisms present in the patient's virus (namely, E35D, M36I, M46I, I54V, R57K, Q58E, and L63P) were added to the initial G48E mutant sequence. Both mutations selected in the context of PI utilization regardless of subtype (eg, M46I, I54V, Q58E, L63P) or natural polymorphisms specific to subtype A1 (in this case, E35D, M36I, and R57K) might contribute to this compensating effect. Although their respective contribution was difficult to evaluate in the settings of our study, it is interesting to note that E35D, a natural subtype A1 polymorphism, has already been identified as a critical component of a gag-pol substitution complex responsible for improving the fitness of NFV-resistant isolates.31 Similarly, M36I is a polymorphism that naturally occurs in 12% of subtype B viruses and more than 80% of non-B viruses and has been shown to increase the replication rate of HIV-1 in vitro.32
Other PR substitutions that were not investigated in MD simulation but that were present in the patient's virus might also have modified the fitness of the G48E mutant. In this respect, L10I can naturally occur in the context of non-B subtype (which was the case in our study patient) and is known to compensate for the loss of replication capacity conferred by G48V and V82A in vitro.21 In the context of drug pressure, particularly SQV, which selects for mutations at residue position 48, K20I provides a significant replication benefit in vitro.32 Although not present at baseline, K20I emerged simultaneous to G48E in the patient's virus (see Table 1). Variation at residue position 20 is more common in subtype non-B versus subtype B under PI pressure,33 which suggests a non-B-specific compensatory pathway for fixing G48E in the patient.
The genetic make-up of Gag-Pol cleavage site regions also contributes to rescue the fitness of HIV-1 PR mutants. In functional analyses of the HIV-1 PR, Moody and colleagues34 demonstrated that a synthetic G48E mutant had altered substrate specificity and that such G48E phenotype could be reversed by changing the P3′ aspartic acid of the RT/RTp66 cleavage site to glycine or asparagine. Although such specific change was not identified in our patient's virus, other substitutions were observed in cleavage site regions, before or after the emergence of G48E. An A-to-V substitution at residue P2 of the conserved NC/p1-NC/TFP cleavage site (A431V) appeared after G48E had emerged and, as already illustrated in previous studies,9-11 likely compensated for the partially defective PR in the patient's virus. A P-to-L change at residue P5′ of the p1/p6 cleavage site (P453L) was observed after the emergence of G48E. P453L, which is commonly selected with the I50V PR mutation under APV pressure, improves the catalytic activity of PR.35 Other substitutions were observed in less well-conserved cleavage site regions (eg, p2/NC, TFP/p6, p6/PR); however, their role in restoring fitness is currently unknown. Whether the natural variability of gag-pol cleavage site regions in subtype A1 favors the emergence of A431V and P453L (or other less well-characterized polymorphisms) during treatment remains to be established. Nevertheless, non-B subtypes are naturally more variable at NC/p1-NC/TFP and p1/p6 cleavage sites than subtype B HIV-1,36 and one could expect that such intersubtype difference in gag-pol cleavage sites might affect fitness more easily over time. In addition, a practical implication of such intersubtype difference may be that PR-recombinant virus assays that do not include cleavage sites in the recombination process are not optimal for evaluating drug sensitivities of non-B viruses in vitro. In particular, most gag cleavage sites, including all amino acids N-terminal to and P1′ and P2′ amino acids of the NC/p1 cleavage site as well as cleavage sites of the RT and integrase genes were not represented in the recombinant virus used for the MTT assay in our study.
In conclusion, we showed that G48E severely impairs the replicative fitness of subtype B HIV-1 in vitro, although being viable in a clinical subtype A1 context. Several non-B-specific compensatory mechanisms could account for the fitness recovery observed in the patient's virus investigated in our study. Those mechanisms, linked to the natural or drug-induced variability in the PR and Gag-Pol cleavage sites, might ultimately have an impact on the PI-based treatment response of patients infected with non-B viruses. Addressing this issue has now become a crucial task if one considers that HIV-1 subtype B viruses are responsible for just more than 12% of the global HIV pandemic37 and that the number of persons with non-B viruses starting therapy are likely to increase dramatically in the coming years. In this respect, our findings illustrate that when confronted with too many mutations to evaluate in vitro, MD simulations are helpful for elucidating how polymorphisms can compensate for the potential fitness cost of resistance mutations.
1. Erickson JW, Burt SK. Structural mechanisms of HIV drug resistance. Annu Rev Pharmacol Toxicol
2. Velazquez-Campoy A, Muzammil S, Ohtaka H, et al. Structural and thermodynamic basis for resistance to HIV-1 protease inhibition: implications for inhibitor design. Curr Drug Targets Infect Disord
3. Kagan RM, Shenderovich MD, Heseltine PNR, et al. Structural analysis of an HIV-1 protease I47A mutant resistant to the protease inhibitor lopinavir. Protein Sci
4. King NM, Prabu-Jeyabalan M, Nalivaika EA, et al. Structural and thermodynamic basis for the binding of TMC114, a next-generation human immunodeficiency virus type 1 protease inhibitor. J Virol
5. Shafer RW. Genotypic testing for human immunodeficiency virus type 1 drug resistance. Clin Microbiol Rev
6. Berkhout B. HIV-1 evolution under pressure of protease inhibitors: climbing the stairs of fitness. J Biomed Sci
7. Nijhuis M, Deeks S, Boucher C. Implications of antiretroviral resistance on viral fitness. Curr Opin Infect Dis
8. Quiñones-Mateu ME, Aerts EJ. HIV-1 fitness: implications for drug resistance, disease progression and global epidemic evolution. In: Kuiken C, Foley B, Hahn B, et al, eds. HIV Sequence Compendium 2001
. Los Alamos, NM: Los Alamos National Laboratory/Theoretical Biology and Biophysics; 2001:134-170.
9. Doyon L, Croteau G, Thibeault D, et al. Second locus involved in human immunodeficiency virus type 1 resistance to protease inhibitors. J Virol
10. Feher A, Weber IT, Bagossi P, et al. Effect of sequence polymorphism and drug resistance on two HIV-1 Gag processing sites. Eur J Biochem
11. Zhang YM, Imamichi H, Imamichi T, et al. Drug resistance during indinavir therapy is caused by mutations in the protease gene and in its Gag substrate cleavage sites. J Virol
12. Svicher V, Ceccherini-Silberstein F, Erba F, et al. Novel human immunodeficiency virus type 1 protease mutations potentially involved in resistance to protease inhibitors. Antimicrob Agents Chemother
14. Van Laethem K, Schrooten Y, Dedecker S, et al. A genotypic assay for the amplification and sequencing of gag and protease from diverse human immunodeficiency virus type 1 group M subtypes. J Virol Methods
15. Snoeck J, Riva C, Steegen K, et al. Optimization of genotypic assay applicable to all human immunodeficiency virus type 1 protease and reverse transcriptase subtypes. J Virol Methods
16. Maschera B, Furfine E, Blair ED. Analysis of resistance to human immunodeficiency virus type 1 protease inhibitors by using matched bacterial expression and proviral infection vectors. J Virol
17. Vandamme AM, Witvrouw M, Pannecouque C, et al. Evaluating clinical isolates for their phenotypic and genotypic resistance against anti-HIV drugs. In: Kinchington D, Schinazi RF, eds. Methods in Molecular Medicine, vol. 24. Antiviral Methods and Protocols
. Totowa, NJ: Humana Press Inc., 2000:223-258.
18. Weber J, Weberova J, Carobene M, et al. Use of a novel assay based on intact recombinant viruses expressing green (EGFP) or red (DsRed2) fluorescent proteins to examine the contribution of pol
genes to overall HIV-1 replicative fitness. J Virol Methods
19. Quiñones-Mateu ME, Ball SC, Marozsan AJ, et al. A dual infection/competition assay shows a correlation between ex vivo human immunodeficiency virus type 1 fitness and disease progression. J Virol
20. Van Der Spoel D, Lindhal E, Hess B, et al. GROMACS: fast, flexible, and free. J Comput Chem
21. Mammano F, Trouplin V, Zennou V, et al. Retracing the evolutionary pathways of human immunodeficiency virus type 1 resistance to protease inhibitors: virus fitness in the absence and in the presence of drug. J Virol
22. Baxter JD, Shapiro JM, Boucher CA, et al. Genotypic changes in human immunodeficiency virus type 1 protease associated with reduced susceptibility and virologic response to the protease inhibitor tipranavir. J Virol
23. Janini LM, Pieniazek D, Peralta JM, et al. Identification of single and dual infections with distinct subtypes of human immunodeficiency virus type 1 by using restriction fragment length polymorphism analysis. Virus Genes
24. Kantor R, Fessel WJ, Zolopa AR, et al. Evolution of primary protease inhibitor resistance mutations during protease inhibitor salvage therapy. Antimicrob Agents Chemother
25. Monno L, Saracino A, Scudeller L, et al. HIV-1 phenotypic susceptibility to lopinavir (LPV) and genotypic analysis in LPV/r-naive subjects with prior protease inhibitor experience. J Acquir Immune Defic Syndr
26. Rhee SY, Fessel WJ, Zolopa AR, et al. HIV-1 protease and reverse-transcriptase mutations: correlations with antiretroviral therapy in subtype B isolates and implications for drug-resistance surveillance. J Infect Dis
27. Rousseau CM, Birditt BA, McKay AR, et al. Large-scale amplification, cloning and sequencing of near full-length HIV-1 subtype C genomes. J Virol Methods
28. Shulman NS, Zolopa AR, Passaro DJ, et al. Efavirenz- and adefovir dipivoxil-based salvage therapy in highly treatment-experienced patients: clinical and genotypic predictors of virologic response. J Acquir Immune Defic Syndr
29. Wu TD, Schiffer CA, Gonzales MJ, et al. Mutation patterns and structural correlates in human immunodeficiency virus type 1 protease following different protease inhibitor treatments. J Virol
30. Scott WRP, Schiffer CA. Curling of flap tips in HIV-1 protease as a mechanism for substrate entry and tolerance of drug resistance. Structure
31. Matsuoka-Aizawa S, Sato H, Hachiya A, et al. Isolation and molecular characterization of a nelfinavir (NFV)-resistant human immunodeficiency virus type 1 that exhibits NFV-dependent enhancement of replication. J Virol
32. Holguin A, Suñe C, Hamy F, et al. Natural polymorphisms in the protease gene modulate the replicative capacity of non-B HIV-1 variants in the absence of drug pressure. J Clin Virol
33. Gonzales MJ, Machekano RM, Shafer RW. Human immunodeficiency virus type 1 reverse-transcriptase and protease subtypes: classification, amino acid mutation patterns, and prevalence in a northern California clinic-based population. J Infect Dis
34. Moody MD, Pettit SC, Shao W, et al. A side chain at position 48 of the human immunodeficiency virus type-1 protease flap provides an additional specificity determinant. Virology
35. Maguire MF, Guinea R, Griffin P, et al. Changes in human immunodeficiency virus type 1 gag at positions L449 and P453 are linked to I50V protease mutants in vivo and cause reduction of sensitivity to amprenavir and improved viral fitness in vitro. J Virol
36. de Oliveira T, Engelbrecht S, Janse van Rensburg E, et al. Variability at human immunodeficiency virus type 1 subtype C protease cleavage sites: an indication of viral fitness? J Virol
37. Osmanov S, Patton C, Walker N, et al. Estimated global distribution and regional spread of HIV-1 genetic subtypes in the year 2000. J Acquir Immune Defic Syndr