Introduction
HIV-1 is characterized by a high degree of genetic diversity[1-3]. The genetic variation is caused by the high mutation rate[4], the rapid turnover rate[5-7], and the relatively large viral population[8-11]. It is thought that the wild-type HIV-1 population is ideally fit and that suboptimal antiretroviral therapy will result in selection of drug resistant viruses with reduced fitness. This is supported by several studies demonstrating that suboptimal therapy selects for amino acid changes that confer a selective disadvantage during replication in a drug-free environment in comparison with the original wild-type virus population[12-18].
In this study, the effects of suboptimal antiretroviral therapy on the evolution and fitness of HIV-1 were investigated by using sequential samples obtained from a patient treated with ritonavir, an inhibitor of the viral protease[19,20]. Loss of antiviral effect of ritonavir is known to be associated with the accumulation of up to eight amino acid substitutions in the protease[21]. Only two of these residues, amino acids 82 and 84, are in direct contact with the inhibitor[22,23]. It has been suggested that most of the additional substitutions are selected because of their contribution to the function of the enzyme, either by reducing the affinity of the active site for the inhibitor in an indirect manner [23] or by enhancing the protease catalytic efficiency. In this study these two factors that contribute to the fitness of the viral protease, i.e. drug resistance and protease catalytic activity, were investigated both on enzymatic and virologic level.
We demonstrate that initially during protease therapy drug resistant viral variants are selected that display reduced protease catalytic efficiency and replicative capacity. Upon continued replication novel protease variants are generated that display no further increase in drug resistance but display such an increase in protease catalytic efficiency that they replicate better than viruses containing the wild-type protease.
Material and methods
Viral RNA analysis
Viral RNA isolation and quantitation
RNA was extracted from 100 μl of serum according to the method described by Boom et al. [24] and was quantified using the prototype Roche assay[25].
Amplification of viral protease
After viral RNA isolation an equivalent of 10 μl of serum was used to reverse transcribe and amplify the protease gene (nucleotides 2252-2548). A one-tube RT-PCR procedure, essentially as described by Nijhuis et al.[26], using 1 mM MgCl2 and 10 pmol of primer 5‚prot-1 (5‚-AGG CTA ATT TTT TAG GGA AGA TCT GGC CTT CC-3‚; nucleotides 2077-2108) and primer 3‚prot-1 (5‚-GCA AAT ACT GGA GTA TTG TAT GGA TTT TCA GG-3‚; nucleotides 2733-2702) (Pharmacia Biotechnology, Roosendaal, The Netherlands) was performed. After this procedure the amount of amplified product was increased further in a second (nested) amplification reaction, containing 12 pmol of primer 5‚prot-2 (5‚-TCA GAG CAG ACC AGA GCC AAC AGC CCC A-3‚; nucleotides 2135-2162) and 11 pmol of primer 3‚prot-2 (5‚-AAT GCT TTT ATT TTT TCT TCT GTC AAT GGC-3‚; nucleotides 2649-2620).
Amplified products of the protease gene were cloned using the TA cloning system (Invitrogen, Leek, The Netherlands). Plasmids were purified and sequenced using the Taq Dye Deoxy Terminator cycle sequencing kit (Applied Biosystems International, Foster City, California, USA) and oligonucleotides PR1 (5‚-AGG AGC CGA TAG ACA AGG-3‚; nucleotides 2215-2232) and PR2 (5‚-CTT TTG GGC CAT CCA TTC-3‚; nucleotides 2609-2592).
Amplification and analysis of the cleavage sites
After viral RNA isolation an equivalent of 10 μl of serum was used in three different RT-PCR (I-II-III) to reverse transcribe and amplify the nine Gag and Gag-Pol protease cleavage sites. In addition to the amplification of the protease (PCR II), a 1.3 kb region which extended from the 5‚ end of p24 into p6 was amplified in PCR I, and a 1 kb region which extended from the 3‚ end of RT into integrase was amplified in PCR III. The amplification conditions used in PCR I were identical to those described for protease (PCR II), except for an annealing temperature of 50 °C and 14 pmol primer 1 (5‚-TAG TAT GGG CAA GCA GGG AGC TAG-3‚; nucleotides 889-912) and primer 2 (5‚-CCT TGT CTA TCG GCT CCT GCT TC-3‚; nucleotides 2232-2210) in the first PCR and primer 3 (5‚-TAG AGG AAG AGC AAA ACA AAA GTA-3‚; nucleotides 1099-1122) and primer 4 (5‚-TCT CTT CTG GTG GGG CTG TTG GCT C-3‚; nucleotides 2172-2148) in the second PCR. The amplification conditions used in PCR III were identical to those described for the amplification of the viral protease, except for the presence of 2 mM MgCl2 and 12 pmol of primer 9 (5‚-GAC AGC TGG ACT GTC AAT GAC ATA CAG-3‚; nucleotides 3296-3322) and primer 10 (5‚-TCT ACT TGT CCA TGC ATG GCT TC-3‚; nucleotides 4392-4370) in the first PCR and primer 11 (5‚-TTA GTG GGA AAA TTG AAT TGG GCA AGT C-3‚; nucleotides 3326-3353) and primer 12 (5‚-AGG TTA AAA TCA CTA GCC ATT GCT CTC C-3‚; nucleotides 4284-4311) in the second PCR. PCR products were sequenced using the Taq Dye Deoxy Terminator cycle sequencing kit and the nested PCR primers.
Fitness analysis
Enzymatic characterization
To generate patient-derived wild-type and mutant HIV-1 proteases for the analysis of the enzymatic characteristics, a bacterial cloning and expression system was used (pET system, Novagen, R&D system, Abingdon, UK). To insert the different proteases in frame in the expression vector pET24a(+), the proteases originally cloned in the TA cloning system were amplified using oligonucleotides HindIII-prot (5‚-CCC AAG CTT TTA AAA ATT TAA AGT GCA GC-3‚ ) and NdeI-prot (5‚-GGA ATT CCA TAT GCC TCA AAT CAC TCT TTG-3‚), digested with NdeI and HindIII, and inserted in the NdeI and HindIII sites of pET24a(+). The introduced protease encoding fragments included an initiation codon and termination codon flanking the 99 codons of the protease (confirmed by sequencing). After transformation and induction of expression, the enzymes were purified from inclusion bodies and refolded[27]. After refolding, wild-type and mutant proteases were more than 90% pure as judged by SDS-PAGE.
There are eight major protease cleavage sites in the Gag-Pol precursor. Evaluation of the kinetic constants for wild-type and mutant enzymes towards peptides corresponding to each of the cleavage sites usually represents the subject of full paper itself and is outside the scope of this paper. We chose the p6*-PR cleavage site as a representative based on the following reasoning. Autocatalytic processing of protease from Gag-Pol precursor is apparently the first step in viral maturation. The initial cleavage during this autoprocessing occurs at the N terminus of the protease; this is supported by studies of HIV PR model precursors in vitro[28,29]. Thus being the first step, the cleavage of p6*-PR junction is important as it can control the release of protease and overall speed of viral maturation. Finally, other groups have also proposed that this cleavage site is one of the most appropriate for kinetic simulation of intravirion activities[30].
The inhibition constants (Ki) for ritonavir were measured essentially as described[31], using the fluorogenic substrate 2-aminobenzoyl-Thr-Ile-Nle-p-nitroPhe-Asn-Arg-NH2 (Bachem, Torrance, California, USA). Kinetic parameters (Kcat and Km) for processing of the decapeptide corresponding to the patient p6*-protease cleavage site (Ile-Ser-Phe-Ser-Phe-Pro-Asn-Ile-Thr-Leu, P1 and P1‚ residues are shown in bold) were determined by competition with the fluorogenic substrate (Xie D, Suvorov L, Erickson, JW and Gulnic S, unpublished data).
Virologic characterization
Recombinant protease viruses were generated by introducing viral protease sequences derived from serum into a protease deleted HIV-1 clone (HXB2Δpro) [32]by homologous recombination. Analysis of the protease in this constant viral background results in variation at the p17-p24 cleavage site (VSQNY-PIVQN as present in HXB2 compared with VSQNF-PIVQN as present in the patient) and at the p2-p7 cleavage site (SATIM- MQRGN as present in HXB2 compared with SATIM-MQKGN as present in the patient). The cloned amplified viral RNAs (digested from the vector using EcoRI) were cotransfected with HXB2Δpro (linearized with BstEII) into SupT1 cells. The transfected cell cultures were subsequently monitored for the appearance of syncytia. When fully developed syncytia were observed, cell-free virus was harvested. The recombination sites located in the protease flanking regions were sequenced to control for correct recombination. The infectious virus titre (TCID50) was determined using end-point dilutions in MT2 cells[33]. Ritonavir susceptibility of the recombinant protease viruses was determined in duplicate using an MTT assay[34].
To determine the relative replication efficiencies of recombinant viruses containing patient-derived viral protease sequences, viral mixtures were prepared and used to infect 5 3 106 phytohaemagglutinin (PHA; Sigma, Amsterdam, The Netherlands) stimulated peripheral blood mononuclear cells (PBMC) at a multiplicity of infection of 0.001. After 2 h infection the cells were washed twice with 10 ml RPMI1640 (Gibco BRL, Life Technologies Inc., Breda, The Netherlands), suspended in 10 ml PBMC culture medium (RPMI1640 supplemented with 10% foetal calf serum (Gibco BRL, Life Technologies Inc.), 10% Lymphocult IL-2 (Biotest, Soest, The Netherlands), antibiotics and polybrene (5 μg/ml; Sigma). One-half of the culture suspension was replaced by this medium twice weekly, and at weekly intervals the PBMC culture medium was supplemented with 2 3 106 fresh PHA stimulated PBMC. Viral replication was monitored by p24 antigen production [35,36]: cell-free virus was harvested at the moment 25-250 ng/ml p24 was present in the culture supernatant, and 200 μl cell-free virus was used to infect 5 3 106 PHA stimulated PBMC. At the same time-points cell-free virus was used for the isolation of viral RNA. Viral RNA was isolated, the viral protease was amplified and population sequencing was performed using the Taq Dye Deoxy Terminator cycle sequencing kit and oligonucleotides PR1 and PR2. Multiple mixture experiments were performed and the relative fitnesses of the protease variants were determined from the regression slopes of the logarithm ratios against time in PBMC culture. The 95% confidence intervals of the regression slopes were determined.
Phylogenetic analysis
Phylogenetic inference was performed on a gap-stripped alignment (297 base pairs) of RNA sequences using the PHYLIP package, version 3.52[37]. The programs DNAML (maximum likelihood), DNAPARS (maximum parsimony) as well as DNADIST together with NEIGHBOR (F84 model plus neighbor-joining) were used to create trees. Bootstrap analysis was carried out by using SEQBOOT (500 replicates), DNADIST, NEIGHBOR and CONSENSE.
Results
Evolution of genotypic protease resistanceDuring ritonavir therapy, HIV-1 RNA levels and CD4 cell counts were determined (Fig. 1). A 1.6 log10 decline in HIV-1 RNA concentration demonstrated after 17 days of ritonavir therapy was followed by a return to the baseline level after approximately 1 month. The CD4 cell count mirrored the inital changes in RNA levels. Genotypic analysis of the viral protease genes revealed that before the start of ritonavir therapy all five protease clones harboured identical sequences, except for substitutions at two non-conserved positions (codon 61 I/V and 77I/V) [38,39](Fig. 2). Subsequent selection of viral variants harbouring some of the previously described ritonavir resistance conferring substitutions (84V, 36I + 54V) was paralelled by a rebound in HIV-1 RNA concentration starting after 24 days of therapy. In addition to these viral variants an 82T substitution either alone or in combination with 36I + 54V was observed 4 days later, at the moment the HIV-1 RNA concentration had returned to baseline. Continued virus replication resulted in the acquisition of a 71V and 20R + 71V substitution in the background of the 36I + 54V + 82T mutant at day 82 and 115, respectively. Despite the appearance of the changes at codon 71 and 20 + 71 no further increase in HIV-1 RNA concentration was observed. However, these constant viral RNA concentrations were observed in the presence of gradually decreasing CD4 counts.
Evolution of the protease cleavage sites
Protease inhibitor therapy can select for substitutions in the cleavage sites as present in the Gag and Gag-Pol polyproteins. Analysis of nine cleavage site sequences at the start and after 115 days of ritonavir therapy revealed no significant changes. Some variation was observed in the p17-p24 cleavage site from VSQNF-PIVQN to a population containing a mixture of VSQNF-PIVQN and ASQNF-PIVQN and in the p1-p6 cleavage site from RPGNF-LQSRP to a population containing both RPGNF-LQSRP and RPGNF-LQNRP.
Fitness analysis
The effect of the observed protease substitutions on resistance and protease catalytic activity, both of which contribute to fitness of the viral protease was investigated enzymatically and virologically.
Enzymatic characterization
Wild-type and mutant HIV-1 proteases were cloned, expressed, purified and used to determine ritonavir inhibition constants (Ki). In addition, protease catalytic efficiency was determined by investigating the kinetic parameters (Kcat and Km) for processing of a decapeptide corresponding to the patient p6*protease cleavage site. These parameters were used to determine the vitality (kcat/Km 3 Ki mut/kcat/Km 3 Ki wild-type) of the protease mutants. Vitality values can be considered as a measure of biochemical or enzymatic fitness[27]. The protease variants observed initially during ritonavir therapy, harbouring one or two substitutions (36I + 54V and 82T) displayed an increase in Ki to ritonavir as compared with the wild-type protease (9 and 17-fold, respectively) (Table 1). A subsequently observed protease variant combining the 82T substitution with the 36I + 54V displayed a further increase in Ki to ritonavir (411-fold). This combination of mutations also altered the protease catalytic activity, as the triple mutant displayed a threefold reduction in catalytic efficiency as compared with wild-type protease. Addition of the 71V substitution did not lead to a further increase in Ki, but instead increased the catalytic efficiency of the quadruple mutant 10-fold. Interestingly, this quadruple mutant displayed a threefold increase in protease activity as compared with the original wild-type protease of the patient. The increase in ritonavir resistance and protease catalytic efficiency led to increased vitality of the quadruple mutant as compared with the wild-type protease.
Virological characterization
Patient-derived wild-type and mutant protease genes were introduced into a fixed viral background to determine viral ritonavir resistance. In addition, protease catalytic activity, as revealed by the relative replication capacity of the virus, was investigated by performing competition experiments between wild-type and mutant viruses.
The two protease variants present before ritonavir therapy, with variation at codons 62 and 77, exhibited similar ritonavir susceptibility (IC50, 0.02 μM) and replication efficiency (relative replication efficiency 1.0) (Fig. 3a, Table 1). The protease variants observed initially during ritonavir therapy (84V, 36I + 54V, 82T and 36I + 54V + 82T), expressed increased resistance to ritonavir (25, 11, 8 and 48-fold, respectively). This correlates well with the observed increases in Ki for ritonavir. These substitutions also changed the replication capacity, as all early protease variants displayed a reduced replication efficiency as compared with wild-type protease in the absence of the drug (relative replication efficiency of 0.6, ≤ 0.6, 0.9 and ≤ 0.6, respectively) (Fig. 3b,c,d,e and Table 1). The diminished replication efficiency of the triple mutant is in agreement with the enzymatically determined reduction in the protease catalytic efficiency. Direct competition experiments between the two poorly replicating mutants (36I + 54V + 82T and its progenitor 36I + 54V) revealed that addition of the 82T reduced the replication capacity of the virus further (data not shown). As a consequence of continued replication during antiretroviral drug pressure, novel protease variants which acquired the 71V and 20R + 71V substitution were selected. Addition of the 71V mutation did not result in a further increase in resistance, but instead improved the replication efficiency to the extent that the resulting virus replicated better than the original wild-type (relative replication efficiency of 1.2) (Fig. 3f, Table 1). The increased replication efficiency of the quadruple mutant is in agreement with the enzymatically determined increase in the protease catalytic efficiency. This increased replication capacity was maintained after the acquisition of the 20R substitution (20R + 36I + 54V + 71V + 82T; relative replication efficiency of 1.1) (Fig. 3g, Table 1). The increase in replicative capacity of the two variants over wild-type virus was statistically significant (36I + 54V + 71V + 82T relative replication efficiency 1.2, 95% confidence interval 1.20-1.22; 20R + 36I + 54V + 71V + 82T relative replication efficiency 1.1, 95% confidence interval 1.07-1.23).
Phylogenetic analysis of the protease sequences
Phylogenetic tree analyses were performed to study the evolution of the protease gene during ritonavir therapy. Fig. 4 shows the tree obtained with the maximum likelihood method. The bootstrap values for some of the branches in the tree were relatively low, but trees obtained with the maximum parsimony and neighbour-joining methods had very similar topology indicating that the tree presented is a good estimate of the true evolutionary relationship of the protease sequences. The analyses revealed that the protease population before the start of therapy was relatively homogeneous. Interestingly, the genetic diversity of the population had increased considerably after 24 days of ritonavir therapy, as demonstrated by the appearance of three new sequence clusters (72T, 84V and 36I + 54V) in the tree. Two of these clusters were associated with the appearance of known ritonavir resistance mutations (84V, 36I + 54V). Appearance of the 82T substitution in two different contexts, i.e. on the wild-type backbone sequence as well as on the 36I + 54V backbone increased genetic diversity further. Separation of these two clusters in the tree suggests that they represent two independent events, although recombination cannot be excluded. The progressive pattern of protease gene evolution continued as the 36I + 54V + 82T variant acquired additional amino acid substitutions (71V and 20R + 71V). These variants rapidly dominated the population resulting in a highly homogeneous population, as indicated by the presence of one major cluster in the tree demonstrating almost no genetic diversity.
Discussion
In this study, the effects of suboptimal antiretroviral therapy on fitness and evolution of HIV-1 were investigated by studying sequential viral populations derived from a patient who received ritonavir monotherapy.
It seems that two phases in the evolution of fitness can be distinguished. In the first phase drug resistant variants (84V, 82T, 36I + 54V) are selected which have a reduced protease catalytic efficiency as revealed by a reduction in the replication capacity. Because these variants differ by only one or two nucleotides from the wild-type sequence, it is most likely that they preexisted in the viral quasispecies and were selected as a consequence of antiretroviral therapy[40]. This is in line with the presence of drug resistant variants encoding one, two and even three substitutions in individuals who have never been exposed to antiviral therapy[38,39,41-43]. The clear reduction in replication capacity explains why they are represented as a minority population in untreated individuals.
A subsequently observed protease variant combining the 82T with the 36I + 54V substitutions displayed a further increase in ritonavir resistance. These substitutions also changed the protease catalytic activity, as the triple mutant displayed a threefold reduction in protease activity as compared with wild-type protease. This reduction in protease catalytic efficiency resulted in an inferior replication capacity of this mutant. These findings are in agreement with several investigations that demonstrate that during protease inhibitor monotherapy initially drug resistant variants are selected with a reduced replication capacity as compared with wild-type protease[15-18]. Phylogenetic analysis revealed a considerable increase in genetic diversity during the first phase. Coexistence of all different protease variants indicates that none of them had gained sufficient increases in fitness to dominate the population.
In the second phase, which was characterized by an increase in protease catalytic activity, mutations at codon 71V and 20R + 71V appeared. Selection of these mutations in the background of the 36I + 54V + 82T did not give any increase in resistance. However, analysis of the protease catalytic efficiency revealed that the 71V substitution was a clear compensatory mutation as this quadruple mutant had a 10-fold increase in catalytic efficiency as compared with its parent, the triple mutant. The protease catalytic activity of the quadruple mutant was threefold higher than that of the wild-type protease; this gave rise to an increased replication capacity of the quadruple mutant as compared with wild-type protease. This increase in fitness of the quadruple mutant observed in vitro is in agreement with the phylogenetic analysis of all protease sequences. The phylogenetic tree revealed that the quadruple mutant rapidly dominated the viral population, indicating a significant increase of the in vivo fitness of this protease variant.
Compensatory mutations were also observed during in vitro selection experiments [44-46] and clinical trials using protease inhibitors[16,18]. In the clinical trials compensatory mutations were selected in the substrate of the viral protease, i.e. Gag and Gag-Pol cleavage sites, and the impairment of viral replication was only partially compensated for. Analysis of all Gag and Gag-Pol cleavage sites present in the viral population in our patient demonstrated no changes during ritonavir therapy, except for minor polymorphisms in the p24-p17 and the p1-p6 cleavage sites. As the observed cleavage site variations coexisted with the original wild-type cleavage site sequences it seems that the changes did not confer a selective advantage for the mutant proteases.
This study demonstrates for the first time that suboptimal antiretroviral therapy initially results in the selection of drug resistant variants with impaired replication efficiencies but that continued evolution in the presence of the drug may result in the generation and selection of novel viral variants with increased fitness as compared with the original wild-type population. Increased fitness has been demonstrated in vitro for other RNA viruses and for bacteria[47-51]. Our observation is in line with Wright‚s concept of an adaptive fitness landscape[52]. According to this model, natural selection tends to drive a population to a local optimum, which is not necessarily a global optimum. Thus, a population may be trapped at a state with a suboptimal adaptation to its environment because natural selection does not allow it to pass through a trough of maladapted intermediate variants, even though a better solution may exist.
Phylogenetic analysis of all protease sequences demonstrated that the protease variants containing four and five amino acid changes (five or six nucleotide substitutions, respectively) are quite distant from the wild-type protease. These data reflect that they were generated during the treatment period. This observation raises an important question: why are these viral variants displaying increased fitness as compared with wild-type HIV-1 not present in the population before the onset of treatment? This can be explained by the large genetic distance between wild-type and mutant, i.e., at least five nucleotide changes have to be selected in order to generate the protease variants with an increased fitness. Moreover, generation of these later viral variants from the wild-type population in the absence of the drug is hindered because the intermediate protease variants have an inferior replication capacity and therefore are counter selected in the absence of antiviral therapy.
In our virological in vitro fitness determinations we analysed the patient-derived viral proteases in the background of a reference virus (HXB2). It cannot be ignored that variation in other regions of the viral genome may also influence viral fitness. The observed fitness changes should therefore be interpreted in a relative manner and no absolute values can be calculated from this type of measurement. Despite these limitations, the virological in vitro fitness determinations agree well with our enzymatic in vitro fitness data and with measurements of in vivo viral fitness, as demonstrated by the phylogenetic analysis of the evolution of the viral protease genes. Another indication for changes of in vivo fitness was deduced by analysis of viral virulence. This parameter was estimated by the analysis of the relationship between HIV RNA concentration and CD4 cell numbers (an indicator for the amount of target cells) over time. Interestingly, after 1 month of therapy while the HIV RNA value had returned to baseline level, the CD4 cell numbers were still elevated. In line with the host-parasite model[53-55], which assumes that the number of target cells contributes to the amount of virus production, this return to baseline RNA level in the presence of an increased number of CD4 cells suggests a reduced viral fitness. In line with this model, subsequent appearance of protease variants with an increased fitness coincided with a decline in CD4 cells to the baseline level, while not further increasing viral RNA levels.
The clinical significance of these findings remains to be determined and will require further studies on larger numbers of patients.
In conclusion, incomplete inhibition of viral replication creates an environment in which preexisting drug resistant variants with an impaired replication potential are selected. Continued replication may drive the population through a trough of maladapted variants towards a new fitness peak which may be higher than that of the original wild-type population.
Acknowledgments
We thank S. Danner for providing us with the serum samples and the clinical information of the patient treated with ritonavir. We thank D. Xie and A. Noest for helpful discussion.
References
1. Domingo E, Escarmis C, Sevilla N, et al. Basic concepts in RNA virus evolution. FASEB J 1996, 10:859-864.
2. Coffin JM. Genetic diversity and evolution of retroviruses. Curr Top Microbiol Immunol 1992, 176:143-164.
3. Domingo E, Holland JJ. RNA virus mutations and fitness for survival. Annu Rev Microbiol 1997, 51:151-178.
4. Mansky LM, Temin HM. Lower in vivo mutation rate of human immunodeficiency virus type 1 than that predicted from the fidelity of purified reverse transcriptase. J Virol 1995, 69:5087-5094.
5. Perelson AS, Neumann AU, Markowitz M, Leonard JM, Ho DD. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science 1996, 271:1582-1586.
6. Wei X, Ghosh SK, Taylor ME, et al. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 1995, 373:117-122.
7. Ho DD, Neumann AU, Perelson AS, Chen W, Leonard JM, Markowitz M. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 1995, 373:123-126.
8. Piatak Jr M, Saag MS, Yang LC, et al. High levels of HIV-1 in plasma during all stages of infection determined by competitive PCR [see comments]. Science 1993, 259:1749-1754.
9. Embretson J, Zupancic M, Ribas JL, et al. Massive covert infection of helper T lymphocytes and macrophages by HIV during the incubation period of AIDS [see comments]. Nature 1993, 362:359-362.
10. Pantaleo G, Graziosi C, Demarest JF, et al. HIV infection is active and progressive in lymphoid tissue during the clinically latent stage of disease. Nature 1993, 362:355-358.
11. Chun TW, Carruth L, Finzi D, et al. Quantification of latent tissue reservoirs and total body viral load in HIV-1 infection. Nature 1997, 387:183-187.
12. Back NKT, Nijhuis M, Keulen W, et al. Reduced replication of 3TC-resistant HIV-1 variants in primary cells due to a processivity defect of the reverse transcriptase enzyme. EMBO J 1996, 15:4040-4049.
13. Goudsmit J, de Ronde A, de Rooij E, de Boer R. Broad spectrum of in vivo fitness of human immunodeficiency virus type 1 subpopulations differing at reverse transcriptase codons 41 and 215. J Virol 1997, 71:4479-4484.
14. Goudsmit J, de Ronde A, Ho DD, Perelson AS. Human immunodeficiency virus fitness in vivo: calculations based on a single zidovudine resistance mutation at codon 215 of reverse transcriptase. J Virol 1996, 70:5662-5664.
15. Eastman PS, Mittler J, Kelso R, et al. Genotypic changes in human immunodeficiency virus type 1 associated with loss of suppression of plasma viral RNA levels in subjects treated with ritonavir (norvir) monotherapy. J Virol 1998, 72:5154-5164.
16. Zhang Y-M, 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 1997, 71:6662-6670.
17. Zennou V, Mammano F, Paulos S, Mathez D, Clavel F. Loss of viral fitness associated with multiple gag and Gag-Pol processing defects in human immunodeficiency virus Type 1 variants selected for resistance to protease inhibitors in vivo. J Virol 1998, 72:3300-3306.
18. Mammano F, Petit C, Clavel F. Resistance-associated loss of viral fitness in human immunodeficiency virus type 1: phenotypic analysis of protease and gag coevolution in protease inhibitor-treated patients. J Virol 1998, 72:7632-7637.
19. Markowitz M, Saag M, Powderly WG, et al. A preliminary study of ritonavir, an inhibitor of HIV-1 protease, to treat HIV-1 infection. New Engl J Med 1995, 333:1534-1539.
20. Danner SA, Carr A, Leonard JM, et al. A short-term study of the safety, pharmacokinetics, and efficacy of ritonavir, an inhibitor of HIV-1 protease. New Engl J Med 1995, 333:1528-1533.
21. Molla A, Korneyeva M, Gao Q, et al. Ordered accumulation of mutations in HIV protease confers resistance to ritonavir. Nature Med 1996, 2:760-766.
22. Erickson JW, Burt SK. Structural mechanism of HIV drug resistance. Annu Rev Pharmacol Toxicol 1996, 36:545-571.
23. Rose RB, Craik CS, Stroud RM. Domain flexibility in retroviral proteases: structural implications for drug resistant mutations. Biochemistry 1998, 37:2607-2621.
24. Boom R, Sol CJA, Salimans MMM, Jansen CL, Wertheim-van Dillen PME, van der Noordaa J. Rapid and simple method for purification of nucleic acids. J Clin Microbiol 1990, 28:495-503.
25. Mulder J, McKinney N, Christopherson C, Sninsky J, Greenfield L, Kwok S. Rapid and simple PCR assay for quantification of human immunodeficiency virus type 1 RNA in plasma: application to acute retroviral infection. J Clin Microbiol 1994, 32:292-300.
26. Nijhuis M, Boucher CAB, Schuurman R. Sensitive procedure for the amplification of HIV-1 RNA using a combined reverse-transcription and amplification reaction. Biotechniques 1995, 19:178-180.
27. Gulnik S, Suvorov LI, Liu B, et al. Kinetic characterization and cross-resistance patterns of HIV-1 protease mutants selected under drug pressure. Biochemistry 1995, 34:9282-9287.
28. Louis JM, Nashed NT, Parris KD, Kimmel AR, Jerina DM. Kinetics and mechanism of autoprocessing of human immunodeficiency virus type 1 protease from an analog of the Gag-Pol polyprotein. Proc Natl Acad Sci USA 1994, 91:7970-7974.
29. Co E, Koelsch G, Lin Y, Ido E, Hartsuck JA, Tang J. Proteolytic processing mechanisms of a miniprecursor of the aspartic protease of human immunodeficiency virus type 1. Biochemistry 1994, 33:1248-1254.
30. Ermolieff J, Lin X, Tang J. Kinetic properties of saquinavir- resistant mutants of human immunodeficiency virus type 1 protease and their implications in drug resistance in vivo. Biochemistry 1997, 36:12364-12370.
31. Kageyama S, Mimoto T, Murakawa Y, et al. In vitro anti-human immunodeficiency virus (HIV) activities of transition state mimetic HIV protease inhibitors containing allophenylnorstatine. Antimicrob Agents Chemother 1993, 37:810-817.
32. 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 1995, 69:5431-5436.
33. Reed LJ, Muench H. A simple method for estimating fifty percent endpoints. Am J Hyg 1983, 493-497.
34. Boucher CAB, Keulen W, van Bommel T, et al. Human immunodeficiency virus type 1 drug susceptibility determination by using recombinant viruses generated from patient sera tested in a cell-killing assay. Antim Ag Chemoth 1996, 40:2404-2409.
35. Moore JP, McKeating JA, Weiss RA, Sattentau QJ. Dissociation of gp120 from HIV-1 virions induced by soluble CD4. Science 1990, 250:1139-1142.
36. McKeating J, Balfe P, Clapham P, Weiss RA. Recombinant CD4-selected human immunodeficiency virus type 1 variants with reduced gp120 affinity for CD4 and increased cell fusion capacity. J Virol 1991, 65:4777-4785.
37. Felsenstein J. PHYLIP: Phylogenetic inference package 3.4. Seattle: University of Washington; 1991.
38. Lech WJ, Wang G, Yang YL, et al. In vivo sequence diversity of the protease of human immunodeficiency virus type 1: presence of protease inhibitor-resistant variants in untreated subjects. J Virol 1996, 70:2038-2043.
39. Kozal MJ, Shah N, Shen N, et al. Extensive polymorphism observed in HIV-1 clade B protease gene using high-density oligonucleotide arrays. Nature Med 1996, 2:753-759.
40. Coffin JM. HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science 1995, 267:483-489.
41. de Jong MD, Schuurman R, Lange JMA, Boucher CAB. Replication of a pre-existing resistant HIV-1 subpopulation in vivo after introduction of a strong selective drug pressure. Antiviral Ther 1996, 1:33-41.
42. Najera I, Holguin A, Quinones-Mateu ME, et al. Pol gene quasispecies of human immunodeficiency virus: mutations associated with drug resistance in virus from patients undergoing no drug therapy. J Virol 1995, 69:23-31.
43. Havlir DV, Eastman S, Gamst A, Richman DD. Nevirapine- resistant human immunodeficiency virus: kinetics of replication and estimated prevalence in untreated patients. J Virol 1996, 70:7894-7899.
44. Rose RE, Gong YF, Greytok JA, et al. Human immunodeficiency virus type 1 viral background plays a mojor role in development of resistance to protease inhibitors. Proc Natl Acad Sci USA 1996, 93:1648-1653.
45. Doyon L, Croteau G, Thibeault D, Poulin F, Pilote L, Lamarre D. Second locus involved in human immunodeficiency virus type 1 resistance to protease inhibitors. J Virol 1996, 70:3763-3769.
46. Borman AM, Paulos S, Clavel F. Resistance of human immunodeficiency virus type 1 to protease inhibitors: selection of resistance mutations in the presence and absence of the drug. J Gen Virol 1996, 77:419-426.
47. Novella IS, Clarke DK, Quer J, et al. Extreme fitness differences in mammalian and insect hosts after continuous replication of vesicular stomatitis virus in sandfly cells. J Virol 1995, 69:6805-6809.
48. Novella IS, Duarte EA, Elena SF, Moya A, Domingo E, Holland JJ. Exponential increases of RNA virus fitness during large population transmissions. Proc Natl Acad Sci USA 1995, 92:5841-5844.
49. Bouma JE, Lenski RE. Evolution of a bacteria/plasmid association. Nature 1988, 335:351-352.
50. Schrag SJ, Perrot V, Levin BR. Adaptation to the fitness costs of antibiotic resistance in Escherichia coli. Proc Royal Soc Lond B 1997, 264:1287-1291.
51. Lenski RE, Travisano M. Dynamics of adaptation and diversification: A 10,000-generation experiment with bacterial populations. Proc Natl Acad Sci USA 1994, 91:6808-6814.
52. Wright S. The role of mutation, inbreeding, crossbreeding and selection in evolution. Proc VI Intern Cong Genet 1932, 1:356-366.
53. McLean AR, Emery VC, Webster A, Griffiths PD. Population dynamics of HIV within an individual after treatment with zidovudine. AIDS 1991, 5:485-489.
54. de Jong MD, Veenstra J, Stilianakis NI, et al. Host-parasite dynamics and outgrowth of virus containing a single K70R amino acid change in reverse transcriptase are responsible for the loss of human immunodeficiency virus type 1 RNA load suppression by zidovudine. Proc Natl Acad Sci USA 1996, 93:5501-5506.
55. de Boer RJ, Boucher CAB. Anti-CD4 therapy for AIDS suggested by mathematical models. Proc Royal Soc Lond B 1996, 263:899-905.
56. Roos M, Miedema F, Eeftinck Schattenkerk JKM, et al. Cellular and humoral immunity in various cohorts of male homosexuals in relation to infection with human immunodefiency virus. Neth J Med 1989, 34:132-141.
© 1999 Lippincott Williams & Wilkins, Inc.