Loss of Protease Dimerization Inhibition Activity of Darunavir Is Associated with the Acquisition of Resistance to Darunavir by HIV-1

Abstract

Dimerization of HIV protease is essential for the acquisition of protease's proteolytic activity. We previously identified a group of HIV protease dimerization inhibitors, including darunavir (DRV). In the present work, we examine whether loss of DRV's protease dimerization inhibition activity is associated with HIV development of DRV resistance. Single amino acid substitutions, including I3A, L5A, R8A/Q, L24A, T26A, D29N, R87K, T96A, L97A, and F99A, disrupted protease dimerization, as examined using an intermolecular fluorescence resonance energy transfer (FRET)-based HIV expression assay. All recombinant HIV_NL4-3-based clones with such a protease dimerization-disrupting substitution failed to replicate. A highly DRV-resistant in vitro-selected HIV variant and clinical HIV strains isolated from AIDS patients failing to respond to DRV-containing antiviral regimens typically had the V32I, L33F, I54M, and I84V substitutions in common in protease. None of up to 3 of the 4 substitutions affected DRV's protease dimerization inhibition, which was significantly compromised by the four combined substitutions. Recombinant infectious clones containing up to 3 of the 4 substitutions remained sensitive to DRV, while a clonal HIV variant with all 4 substitutions proved highly resistant to DRV with a 205-fold 50% effective concentration (EC₅₀) difference compared to HIV_NL4-3. The present data suggest that the loss of DRV activity to inhibit protease dimerization represents a novel mechanism contributing to HIV resistance to DRV. The finding that 4 substitutions in PR are required for significant loss of DRV's protease dimerization inhibition should at least partially explain the reason DRV has a high genetic barrier against HIV's acquisition of DRV resistance.

INTRODUCTION

Currently available combination therapy or highly active antiretroviral therapy (HAART) for human immunodeficiency virus type 1 (HIV) infection and AIDS has been shown to potently suppress the replication of HIV and extend the life expectancy of HIV-infected individuals (32, 34). Recent analyses have revealed that life expectancy in HIV-infected patients treated with HAART has significantly increased, that mortality rates for HIV-infected persons have recently become close to that of general population, and that the appearance of the current first-line antiretroviral therapy with boosted protease inhibitor (PI)-based regimens has made the development of HIV resistance relatively less likely (2, 7, 18, 39). However, the ability to provide effective long-term antiretroviral therapy for HIV infection remains a complex issue since many of those who initially achieved favorable viral suppression to undetectable levels still suffer treatment failure (12, 18, 29).

Dimerization of HIV protease (PR) subunits is an essential process for the acquisition of proteolytic activity of HIV PR, which plays a critical role in the maturation and replication of the virus (28, 40). Thus, inhibition of PR dimerization by chemical reagents is likely to abolish proteolytic activity and intervene in HIV replication. We have recently developed an intermolecular fluorescence resonance energy transfer (FRET)-based HIV-expression assay that employs cyan fluorescent protein (CFP)- and yellow fluorescent protein (YFP)-tagged HIV PR monomers to detect and quantify PR dimerization (26). Using this assay, we identified a group of nonpeptidyl small molecule inhibitors of HIV PR dimerization. These inhibitors, including darunavir (DRV) and tipranavir (TPV) as well as a series of potent experimental antiretroviral agents such as TMC126 (41), blocked PR dimerization at concentrations of as low as 0.01 μM and potently blocked HIV replication in vitro (26).

DRV contains a structure-based and designed privileged nonpeptidic P2 ligand, 3(R),3a(S),6a(R)-bis-tetrahydrofuranylurethane (bis-THF) (14, 15, 27), which potently inhibits the enzymatic activity and dimerization of HIV PR (26) and has a high-level genetic barrier against HIV development of resistance to DRV (9, 10). Nevertheless, we have witnessed that HIV acquires significant levels of resistance against DRV among HIV-infected individuals who have received long-term combination chemotherapy (33, 38). Indeed, a variety of amino acid substitutions that are potentially related to HIV resistance to DRV have been reported (9, 24, 33, 38). Thus, the elucidation of the mechanism of the development of HIV drug resistance represents an urgent subject in the research area of HIV-1 infection/AIDS and therapy.

MATERIALS AND METHODS

Cells, viruses, and antiviral agents.

MT-4 cells were grown in RPMI 1640-based culture medium, and 293T and COS7 cells were propagated in Dulbecco's modified Eagle's medium. These media were supplemented with 10% fetal calf serum (FCS; PAA Laboratories GmbH, Linz, Austria) plus 50 U of penicillin and 50 μg of kanamycin per ml. The following HIV strains were used for the determination of 50% effective concentrations (EC₅₀s) against DRV and to construct plasmids for use in the FRET-based HIV expression assay, including HIV_NL4-3 and HIV_8MIX^P51. Three recombinant clinical HIV isolates (_rCLHIV_F16, _rCLHIV_T45, and _rCLHIV_T48) used in this study were produced using recombinant HIV_NL4-3-based infectious molecular clones generated by ligating patient-derived amplicons encompassing approximately 200 nucleotides of 3′ Gag (beginning at the unique ApaI restriction site), the entire protease, and the first 72 nucleotides of reverse transcriptase (RT) using the expression vector pNLPFB (a generous gift from Tomozumi Imamichi of the National Institute of Infectious Diseases and Allergy). The four clinical HIV isolates examined in the present study were chosen from 32 isolates that had been obtained from multi-PI-treated patients whose protease genotype contained prototypical patterns of PI resistance. The median duration of continuous PI treatment was 7.5 years (range, 6 to 10 years). The median number of PIs received (excluding the use of ritonavir for pharmacokinetic boosting) was 5 (range, 4 to 8). According to the PhenoSense assay, two isolates (_rCLHIV_T45 and _rCLHIV_F16) had high-level resistance to all PIs, including >90-fold decreased susceptibility to DRV and >8-fold decreased susceptibility to TPV, and one isolate (_rCLHIV_T48) had intermediate resistance to DRV and TPV and high-level resistance to the remaining PIs (see Table 2).

Table 2.

Amino acid substitutions associated with DRV resistance

DRV resistance-associated amino acid substitutions identifieda	Variantb	EC50, μM (fold change)c
L10I, I15V, K20R, L24I, V32I, L33F, M36I, M46L, I54 M, L63P, K70Q, V82I, I84V, L89 M	HIV8MIXP51	>1 (>333)
L10F, V11I, I13V, L19Q, K20 M, V32I, L33F, E35A, M36I, M46I, I47V, I54 M, R57K, I62V, L63P, I64V, G73T, T74A, I84V, L89V, L90 M	rCLHIVF16	0.30 (97)
L10F, V11I, T12P, I13V, I15V, L19P, K20T, V32I, L33F, E35G, M36I, I54V, I62V, L63P, K70T, A71I, G73S, P79A, I84V, L89V, L90 M	rCLHIVT45	0.33 (105)
L10I, I13V, I15V, L19V, L24I, V32I, L33F, K43E, M46L, I54L, D60E, L63P, A71V, I72V, V82A, I84V	rCLHIVT48	0.17 (54)
V11I, V32I, L33F, I47V, I50V, I54L/M, G73S, L76V, I84V, L89V
V32I, I50V, I54L, I54 M, L76V, V82F

^aAmino acid substitutions identified in the protease-encoding region of HIV are shown. The amino acid substitutions shown in the second row from the bottom were reported by De Meyer et al. (9) and Mitsuya et al. (33). Those in the bottom row were reported by Van Marck et al. (38) and were reported to have the greatest impact on HIV-1 resistance to DRV. Those for HIV_8MIX^P51 were reported by Koh et al. (24).

^bThree infectious clones (_rCLHIV_F16, _rCLHIV_T45, and _rCLHIV_T48) were derived from clinical strains isolated from patients who failed to respond to DRV-containing regimens.

^cShown are EC₅₀s of DRV for each infectious clone. Values in parentheses represent fold changes of EC₅₀s of DRV compared to EC₅₀s against a wild-type clinical strain, HIV-1_ERS104pre. All assays were conducted in triplicate, and the mean values are shown. The EC₅₀s of ritonavir and lopinavir against HIV_8MIX^P51, _rCLHIV_F16, _rCLHIV_T45, and _rCLHIV_T48 were all >1 μM.

DRV was synthesized by A. K. Ghosh as described previously (16, 27). GRL-0216 (37), GRL-98065 (1), and TMC-126 (41) were synthesized in a convergent manner by coupling an optically active P2 ligand and an (R)-hydroxyethylamino sulfonamide isostere (17). TPV was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, National Institutes of Health.

Generation of FRET-based HIV expression system.

Cyan fluorescent protein (CFP)- and yellow fluorescent protein (YFP)-tagged HIV PR constructs were generated using BD Creator DNA cloning kits (BD Biosciences, San Jose, CA). The basic concepts of the intermolecular FRET-based HIV expression assay (FRET-HIV assay) are illustrated in Fig. 1. In brief, XhoI and HindIII fragments from the pCR-XL-TOPO vector containing the HIV PR-encoding gene excised from pHIV_NL4-3 were inserted into pDNR-1r, the donor vector, which had been digested with XhoI and HindIII. In the transfer of the PR gene from the donor vector into pLP-CFP/YFP-C1 (acceptor vector), the Cre-loxP site-specific recombination method was used according to the manufacturer's instructions. Using Cre-recombinase with the loxP site, the PR gene from pDNR-1r was inserted into pLP-CFP-C1 or pLP-YFP-C1 through Cre-mediated recombination (19), generating a plasmid expressing CFP-tagged wild-type PR (PR_WT) and one expressing YFP-tagged PR_WT, with which HIV PR was successfully expressed as a fusion protein with CFP and YFP tagged at the C terminus, respectively. Western blot assay using anti-green fluorescent protein-specific rabbit polyclonal antibodies revealed that PR was correctly tagged with CFP or YFP (26).

Fig. 1.

FRET-based HIV expression system. Plasmids encoding full-length molecular infectious HIV (HIV_NL4-3) clones that produce CFP- or YFP-tagged PR were prepared using the PCR-mediated recombination method as described in Materials and Methods. A linker consisting of five alanines was inserted between PR and the fluorescent protein. A phenylalanine-proline site (F/P) that HIV PR cleaves was introduced between the fluorescent protein and RT. Shown are structural representations of PR monomers and dimer in association with the linker atoms and fluorescent proteins. FRET occurs when the two fluorescent proteins become 1 to 10 nm apart. If an agent that is capable of inhibiting the dimerization of PR monomer subunits is present when the CFP- and YFP-tagged PR monomers are produced within the cell upon cotransfection, no FRET occurs. If certain amino acid substitutions (AA) such as D29N (shown below) are introduced, PR subunits do not get dimerized and no FRET occurs.

For the generation of full-length molecular infectious clones containing CFP- or YFP-tagged PR, the PCR-mediated recombination (PMR) method was used (11). To this end, we amplified an upstream proviral DNA fragment containing an ApaI site and HIV PR (excised from pHIV_NL4-3) with primer pair Apa-PRO-F (5′-TTG CAG GGC CCC TAG GAA AAA GG-3′) plus PR-5Ala-R (5′-GGC TGC TGC GGC AGC AAA ATT TAA AGT GCA GCC AAT CT-3′), a middle proviral DNA fragment containing CFP (excised from pCFP-C1) or YFP (excised from pYFP-C1) (Clontech, Mountain View, CA) with primer pair CFPYFP-5Ala-F (5′-GCT GCC GCA GCA GCC GTG AGC AAG GGC GAG GAG CTG-3′) plus CFPYFP-FP-R (5′-ACT AAT GGG AAA CTT GTA CAG CTC GTC CAT GCC G-3′), and a downstream proviral DNA fragment containing the 5′-DNA fragment of RT and an SmaI site from pHIV_NLSma (13, 25), which had been created to have an SmaI site by changing two nucleotides (2590 and 2593) of pHIV_NL4-3, with primer pair FRT-F (5′-TTT CCC ATT AGT CCT ATT GAG ACT GTA-3′) plus NL4-3-RT263-R (5′-CCA GAA ATC TTG AGT TCT CTT ATT-3′). A linker consisting of five alanines was inserted between the PR and fluorescent protein. The phenylalanine-proline site that HIV PR cleaves was also introduced between the fluorescent protein and RT. Thus, the three DNA fragments obtained were subsequently joined by using the PMR reaction performed under the standard condition for ExTaq polymerase (Takara Bio, Inc., Otsu, Japan) with 10 pmol of Apa-PRO-F (5′-TTG CAG GGC CCC TAG GAA AAA GG-3′) and NL4-3-RT263-R (5′-CCA GAA ATC TTG AGT TCT CTT ATT-3′) and the three DNA fragments (100 ng each) in a 20-μl reaction solution. Thermal cycling was carried out at 94°C for 3 min, followed by 35 cycles of 94°C for 50 s, 53°C for 50 s, and 72°C for 2 min, and finally 72°C for 15 min. The amplified PCR products were cloned into pCR-XL-TOPO vector according to the manufacturer's instructions (Gateway cloning system; Invitrogen). PCR products were generated with pCR-XL-TOPO vector as templates, followed by digestion by both ApaI and SmaI, and the ApaI-SmaI fragment was introduced into pHIV_NLSma (13), generating pHIV-PR_WT^CFP and pHIV-PR_WT^YFP, respectively.

Analysis of inter- and intramolecular interactions of PR subunits.

Analysis of inter- and intramolecular interactions of PR subunits was conducted by examining the crystal structure of DRV with HIV PR (Protein Data Bank identification no. [PDB ID no.] 2IEN). Hydrogens were added and minimized using the OPLS2005 force field with constraints on heavy atom positions. The calculation was performed using MacroModel 9.1 from Schrödinger, LLC. Hydrogen bonds were assigned when the following distance and angle cutoff was satisfied: 3.0 Å for H-A distance, with a D-H-A angle of >90° and an H-A-B angle of >60°, where H is the hydrogen, A is the acceptor, D is the donor, and B is a neighbor atom bonded to the acceptor. The representative distance between the termini of two monomers was determined by analyzing the PR-DRV crystal structure (PDB ID no. 2IEN). The distance between the α carbons at the N termini and C termini is around 0.5 nm, whereas the distance between the α carbons of the N termini ends of two monomers is around 1.8 nm.

FRET procedure.

COS7 cells plated on an EZ view cover glass bottom culture plate (Iwaki, Tokyo) were transfected with the indicated plasmid constructs, using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions in the presence of various concentrations of each compound, cultured for 72 h, and analyzed under a Fluoview FV500 confocal laser scanning microscope (Olympus Optical Corp., Tokyo) at room temperature. When the effect of each compound was analyzed by FRET, test compounds were added to the culture medium simultaneously with plasmid transfection. The results of FRET were determined by quenching of CFP (donor) fluorescence and an increase in YFP (acceptor) fluorescence (sensitized emission), because part of the energy of CFP is transferred to YFP instead of being emitted. This phenomenon can be measured by bleaching YFP, which should result in an increase in CFP fluorescence. This technique, also known as acceptor photobleaching, is a well-established method of determining the occurrence of FRET (5, 6, 35, 36). Dequenching of the donor CFP by selective photobleaching of the acceptor YFP was performed by first obtaining YFP and CFP images at the same focal plane, followed by illuminating the same image for 3 min at a wavelength of 488 nm with a laser power set at the maximum intensity to bleach YFP and then recapturing the same CFP and YFP images. The changes in the CFP and YFP fluorescence intensity in the images of selected regions were examined and quantified using the Olympus FV500 Image software system (Olympus Optical Corp.). Background values were obtained from the regions where no cells were present and were subtracted from the values for the cells examined in all calculations. For each chimeric protein, the data were obtained from at least three independent experiments. Digitized image data obtained from the experiment were prepared for presentation using Photoshop 6.0 (Adobe Systems, Mountain View, CA). Ratios of intensities of CFP fluorescence after photobleaching to CFP fluorescence prior to photobleaching (CFP^A/B ratios) were determined. It is well established that CFP^A/B ratios of >1.0 indicate that association of CFP- and YFP-tagged proteins occurred and were interpreted to indicate that the dimerization of PR subunits occurred. CFP^A/B ratios of <1 indicated that the association of the two subunits did not occur and were interpreted to indicate that PR dimerization was inhibited. The difference in the CFP^A/B ratios determined in the presence or absence of test drugs was evaluated using the nonparametric Mann-Whitney U statistic test.

Replication kinetics of various NL-PR_mutant^YFP strains.

MT-4 cells (10⁵) were exposed to each infectious HIV-PR^YFP clone (100 ng of p24 Gag protein/ml) for 6 h, washed twice with phosphate-buffered saline (PBS), and cultured in 7 ml of complete medium with some modification as described previously (3, 13). Culture supernatants (50 μl) were harvested every other day, and virus replication was monitored by the amounts of p24 Gag produced in the culture supernatants.

RESULTS

DRV inhibits the dimerization of HIV PR expressed as a single protein.

The basic concepts of the intermolecular FRET-based HIV-expression assay (FRET-HIV expression assay) to assess PR dimerization are illustrated in Fig. 1 (26). Using the FRET-based HIV expression assay, we previously identified a group of PR dimerization inhibitors (PDIs), including DRV and TPV, although other conventional PR inhibitors (PIs), such as amprenavir (APV), failed to block dimerization (Fig. 2A) (26). In the FRET-based HIV expression assay, YFP- or CFP-tagged PR should be primarily expressed as a part of Pr160^gag-pol polyprotein, and it was assumed that DRV blocks the dimerization of the PR subunit within the polyprotein. Thus, it remained to be determined whether DRV also blocks the dimerization of PR in the form of a single PR molecule. We, therefore, generated a pair of plasmids encoding wild-type HIV_NL4-3 PR tagged with YFP and CFP in the 3′ terminus (pPR_WT^YFP and pPR_WT^CFP, respectively), transfected COS7 cells with the pair, and determined whether DRV blocked the dimerization of PR_WT^YFP and PR_WT^CFP. As shown in Fig. 2B, the average value of CFP^A/B ratios obtained in the absence of drug was 1.20 ± 0.24, which indicated that the dimerization between PR_WT^YFP and PR_WT^CFP occurred. The average value of the ratios determined in the presence of 1 μM DRV was 0.74 ± 0.18 (P = 0.000003), signifying that DRV clearly blocked the dimerization of PR_WT^YFP and PR_WT^CFP, while the value with APV was 1.17 ± 0.27 (P = 0.60), indicating that APV failed to block the dimerization, in line with our previous data (26). These results strongly suggest that DRV blocks dimerization of the PR monomer subunit in the form of Pr160^gag-pol polyprotein as well as in the form of a single molecule.

Fig. 2.

DRV blocks the dimerization of both pHIV-PR_WT-encoded PR and pPR_WT-encoded PR. (A) COS7 cells were cotransfected with pHIV-PR_WT^CFP plus pHIV-PR_WT^YFP in the absence or presence of 1 μM DRV or APV. On day 3 after transfection, CFP^A/B ratios were determined using an FV500 confocal laser microscope. When the average value of CFP^A/B ratios was greater than 1.0, it was judged that the dimerization of PR occurred, whereas when it was less than 1.0, it was judged that the dimerization did not occur. (B) COS7 cells were cotransfected with a pair of wild-type PR-expressing plasmids (pPR_WT^CFP plus pPR_WT^YFP) in the absence or presence of 1 μM DRV or APV, and CFP^A/B ratios were determined as described above. Note that DRV inhibited the dimerization of PR when it was expressed as HIV virions and virion-free PR. The results of statistical evaluation of the changes in the CFP^A/B ratios, determined in the presence or absence of DRV or APV, using the nonparametric Mann-Whitney U test, are as follows. (A) For the CFP^A/B ratios in the absence of drug (CFP^{A/B_{No Drug}}) versus the CFP^A/B ratios in the presence of 1.0 μM DRV (CFP^{A/B_{1.0 DRV}}), P = 0.00001, and for CFP^{A/B_{No Drug}} versus CFP^{A/B_{1.0 APV}}, P = 0.42. (B) For CFP^{A/B_{No Drug}} versus CFP^{A/B_{1.0 DRV}}, P = 0.000003, and for CFP^{A/B_{No Drug}} versus CFP^{A/B_{1.0 APV}}, P = 0.60.

Dimerization profiles of single PR mutants in the presence of DRV.

Certain amino acids in the termini and active site interfaces, both of which are critical for the dimerization of PR monomer subunits (28, 40), do not significantly affect the dimerization process of PR. Such amino acids include Pro-1, Gln-2, Thr-4, Asp-25, Ala-28, Asp-30, Thr-96, and Asn-98 (26). It is assumed that DRV blocks PR dimerization by binding to a certain structural domain or domains within or in the proximity of either or both of the two interfaces (4, 22, 23). We, therefore, examined whether amino acid substitutions at positions 1, 3, 5, 25, 28, 30, 96, and 98, which allow PR to dimerize, affected the PR dimerization disruption by DRV. We reasoned that if any of the amino acid substitutions at these positions would affect PR dimerization inhibition by DRV, such amino acids could possibly be associated with the binding of DRV to the PR subunit. However, 1 μM DRV effectively blocked the dimerization of all of the mutated PR species, except that of the species with the A28S substitution (Fig. 3 A). These data suggest that all amino acid residues examined except A28S were not associated with the binding of DRV to the PR monomer subunit.

Fig. 3.

Dimerization profiles of single PR mutants in the presence of DRV. (A) COS-7 cells were cotransfected with pHIV-PR_WT^CFP plus pHIV-PR_WT^YFP (shown as WT^CFP/WT^YFP) or mutated pairs such as pHIV-PR_P1A^CFP plus pHIV-PR_P1A^YFP (shown as P1A^CFP/P1A^YFP) in the absence or presence of 1 μM DRV. On day 3 after transfection, CFP^A/B ratios were determined. (B) COS7 cells were cotransfected with plasmid pair pHIV-PR_A28S^CFP and pHIV-PR_A28S^YFP in the absence or presence of an agent (1 μM GRL-0216, DRV, GRL-98065, TPV, or TMC126), and CFP^A/B ratios were determined as described above. (A) The statistical evaluation of all the changes in the CFP^A/B ratios determined in the presence or absence of DRV using the nonparametric Mann-Whitney U test, gave P values ranging 0.000037 to 0.044, except for the P value for the pair A28S^CFP and A28S^YFP, which was 0.57. (B) The differences between the CFP^A/B ratios in the absence of drug (CFP^{A/B_{No Drug}}) and the CFP^A/B ratios in the presence of 1.0 μM DRV (CFP^{A/B_{1.0 DRV}}) were statistically insignificant, indicating that all of the agents examined failed to block the dimerization of A28S^CFP/A28S^YFP.

We have previously shown that, in addition to DRV and TPV, the three compounds GRL-0216 (37), GRL-98065 (1), and TMC126 (41) effectively blocked PR dimerization in the FRET-based HIV expression assay (26). Since the structures of these five compounds differ from each other, it was thought that the binding profiles of each compound also differed. We, therefore, examined if the four compounds other than DRV disrupted the dimerization of the A28S-carrying PR subunit. As shown in Fig. 3B, all four compounds failed to block protease dimerization, suggesting that Ala-28 is likely involved directly or indirectly in the binding of all four compounds to the PR monomer subunit.

Replication kinetics of HIV variants with failed PR dimerization.

The failure of PR dimerization should completely block or significantly compromise the replication of HIV. In order to confirm that the observed dimerization failure elicited by a single amino acid substitution (26) causes replication failure of HIV, we generated a panel of HIV variants carrying YFP-tagged PR with a single amino acid substitution and examined the replicative capability of each variant. When the amount of p24 antigen produced into culture medium following transfection of COS7 cells with each plasmid was quantified, an HIV variant containing I3A, L5A, L24A, D25A, T26A, A28S, D29N, R87K, T96A, L97A, or F99A produced no or a significantly small amount of p24 (Fig. 4 A). Among these variants, we have previously shown that I3A, L5A, T26A, D29N, R87K, T96A, L97A, and F99A disrupted PR dimerization (26). Figure 5 shows that R8A (P = 0.000099), R8Q (P = 0.000084), and L24A (P = 0.0000014) also disrupted PR dimerization, as examined in the FRET-based HIV expression assay (Fig. 5). When fresh MT-4 cells were exposed to each cell-free culture supernatant of the transfected COS7 cells, as described above, no further replication was seen over the 11-day period of the culture (Fig. 4B). The HIV variant carrying A28S also failed to replicate; however, this failure was explained by an observation that protease with A28S has almost no enzymatic activity, as reported by Hong et al. (20, 21). In contrast, recombinant HIV clones containing either of P1A, Q2A, T4A, D30N, or N98A continued to replicate (Fig. 4B).

Fig. 4.

Replication kinetics of HIV-PR^YFP with wild-type or mutated PR. (A) 293T cells were transfected with pHIV-PR_WT^YFP or mutated pHIV-PR^YFP (if pHIV-PR_P1A^YFP was used, it is shown as P1A^YFP), and the amounts of p24 Gag in the culture supernatants were determined 48 h after transfection. (B) MT-4 cells (10⁵) were exposed to the harvested supernatant of each infectious HIV-PR^YFP clone shown in panel A (100 ng of p24 Gag protein/ml) for 6 h, washed twice with phosphate-buffered saline (PBS), and further cultured in 7 ml of complete medium. Culture supernatants (50 μl) were harvested every other day, and virus replication was monitored by the amounts of p24 Gag produced in the culture supernatants. Replication kinetics of various HIV-PR^YFP mutants are shown over 11 days. In subpanels a, b, and c, the replication kinetics of infectious clones carrying mutations in the active site, N ternimus, and C terminus, respectively, are shown. Note that recombinant HIV clones, whose replication rates were relatively poor, are illustrated in subpanel a. The experiments that generated data in subpanels a and b were performed on the same occasion. Thus, two controls (HIV-1_NL4-3sma and HIV-PR_WT^YFP) in subpanel b serve as controls in subpanel a as well.

Fig. 5.

Dimerization inhibition profiles of selected HIV-1 PR mutants. COS7 cells were cotransfected with a pair of HIV-PR^CFP and HIV-PR^YFP strains either wild type or carrying single amino acid (AA) substitutions, such as the R8A, R8Q, or L24A, in the absence of drug. The CFP^A/B ratios were determined at the conclusion of the 3-day period of culture. The differences between the CFP^A/B ratios of the WT and the CFP^A/B ratios of the mutant had P values of 0.000099 for R8A, 0.000084 for R8Q, and 0.0000014 for L24A.

DRV resistance profiles of PR species carrying DRV-resistance-associated amino acid substitutions.

Using the standardized protocol for selection of drug-resistant HIV variants, we along with others have experienced difficulty in selecting DRV-resistant HIV variants in vitro (8, 24). The emergence of DRV-resistant HIV variants was substantially slower than that of variants resistant to other FDA-approved PIs when a single HIV strain was employed as a starting viral strain (8, 24). In this respect, we have recently succeeded in selecting a highly DRV-resistant HIV variant by using a mixture of 8 highly multi-PI-resistant, DRV-susceptible clinical HIV strains (HIV_8MIX strains HIV_A, HIV_B, HIV_C, HIV_G, HIV_TM, HIV_MM, HIV_JSL, and HIV_SS), which were originally isolated from patients with AIDS, who had failed then-existing anti-HIV regimens after receiving 9 to 11 anti-HIV drugs over the previous 32 to 83 months in the late 1990s and contained 9 to 14 amino acid substitutions in the PR-encoding region (42). By passage 39 in the selection with DRV, HIV_8MIX (HIV_8MIX^P39) became highly resistant to DRV, with an EC₅₀ ∼333-fold greater than that against HIV_NL4-3. HIV_8MIX at passage 39 (HIV_8MIX^P39) was capable of replicating in the presence of 1 μM DRV with a replication fitness comparable to that of HIV_NL4-3 (24). HIV_8MIX at passage 51 (HIV_8MIX^P51), which was also capable of replicating in the presence of 5 μM DRV, was found to contain the following 14 mutations: L10I, I15V, K20R, L24I, V32I, L33F, M36I, M46L, I54M, L63P, K70Q, V82I, I84V, and L89M (24). As illustrated in Table 1, when HIV_8MIX^P51 was propagated in the presence of 0.1 and 1.0 μM DRV in CD4⁺ MT-4 cells, the virus replicated comparably to HIV_NL4-3 during the 9-day period of culture, while HIV_NL4-3 completely failed to replicate in the presence of 0.1 or 1.0 μM DRV, as examined according to the amounts of Gag protein produced in the culture supernatant, indicating that HIV_8MIX^P51 had acquired a high-level resistance against DRV, while it maintained its robust replication fitness.

Table 1.

HIV DRV-resistant strain HIV_8MIX^P51 is capable of replicating in the presence of DRV^a

Virus	DRV (μM)	Replication (ng/ml) at day postexposure:
		1	3	5	7	9
HIVNL4-3 (WT)	0	0	33 ± 9.9	955 ± 9.9	993 ± 57	1152 ± 127
	0.1	0	0	0	0	0
	1.0	0	0	0	0	0
HIV8MIXP51	0	0	701 ± 45	734 ± 68	771 ± 19	877 ± 88
	0.1	1.5 ± 2.1	590 ± 103	682 ± 199	729 ± 3	909 ± 178
	1.0	0.5 ± 0.7	270 ± 10	886 ± 117	936 ± 18	1,201 ± 170

^aCD4⁺ MT-4 cells were exposed to HIV_NL4-3 or HIV_8MIX^P51 (a highly DRV-resistant HIV variant derived from the mixture of 8 highly-PI-resistant clinical HIV isolates exposed to increasing concentrations of DRV up to 1 μM) (24), cultured in the absence or presence of 0.1 or 1.0 μM DRV. Viral replication was monitored by the amounts of p24 Gag protein (ng/ml) produced in the culture supernatant.

We, therefore, asked if the dimerization of the PR of HIV_8MIX^P51 was blocked by DRV, exploiting the FRET-based HIV expression system by using a pair of newly generated plasmids encoding a full-length molecular infectious HIV clone containing CFP- or YFP-tagged PR with all 14 amino acid substitutions. As shown in Fig. 6, DRV significantly blocked the dimerization of the wild-type PR of HIV_NL4-3 at concentrations of 0.1 and 1 μM. However, DRV failed to block the dimerization of the PR of HIV_8MIX^P51 at 0.1 μM (P = 0.42). These data suggested that all amino acid substitutions present in the PR of HIV_8MIX^P51 or subsets of them were associated with the HIV_8MIX^P51 strain's acquisition of DRV resistance.

Fig. 6.

DRV fails to inhibit the dimerization of the protease of a highly DRV-resistant HIV_8MIX^P51 variant. COS7 cells were transfected with a pair of plasmids encoding a full-length molecular infectious HIV-1 clone (HIV_8MIX^P51) containing CFP- or YFP-tagged PR with 14 amino acid substitutions (L10I, I15V, K20R, L24I, V32I, L33F, M36I, M46L, I54M, L63P, K70Q, V82I, I84V, and L89M) in the presence or absence of 0.1, 1, or 10 μM DRV. On day 3 after transfection, CFP^A/B ratios were determined as described in the legend to Fig. 2. HIV_NL4-3 served as a reference. Note that 0.1 and 1 μM DRV failed to block the dimerization of the protease of HIV_8MIX^P51, while the same concentration of DRV blocked protease dimerization in HIV_NL4-3. The differences between the CFP^A/B ratios in the absence of drug (CFP^{A/B_{No Drug}}) and the CFP^A/B ratios in the presence of 0.01 μM DRV (CFP^{A/B_{0.01 DRV}}), between the CFP^A/B ratios in the presence of 0.01 μM DRV (CFP^{A/B_{0.01 DRV}}) and 0.1 μM DRV (CFP^{A/B_{0.1 DRV}}), and between the CFP^A/B ratios in the presence of 0.1 μM DRV (CFP^{A/B_{0.1 DRV}}) and 1.0 μM DRV (CFP^{A/B_{1.0 DRV}}) had P values of 0.32, 0.0025, and 0.34 for HIV_NL4-3, respectively. The differences between the CFP^{A/B_{No Drug}} and the CFP^{A/B_{0.1 DRV}}, between the CFP^{A/B_{0.1 DRV}} and CFP^{A/B_{1.0 DRV}}, and between the CFP^{A/B_{1.0 DRV}} and the CFP^{A/B_{10.0 DRV}} had P values of 0.42, 0.022, and 0.26, respectively, for HIV_8MIX^P51.

Effects of V32I, L33F, I54M/L, and/or I84V substitutions on HIV susceptibility to DRV and PR dimerization inhibition by DRV.

When we examined the sequence of the PR-encoding gene in HIV_8MIX^P51 and three clinical HIV variants isolated from individuals with AIDS who did not respond to DRV-containing antiviral regimens (Table 2) (33), four amino acid substitutions (V32I, L33F, I54M, and I84V) were found to be mostly in common and thought to be relatively unique in such DRV-resistant HIV variants. The locations of the four amino acid substitutions are illustrated in Fig. 7A. This notion was further confirmed when we examined reports by others (9, 24, 33, 38) regarding the sequence of the PR-encoding region of DRV-resistant variants, as illustrated in Table 2. We consequently examined whether the notion described above was plausible by incorporating one of the four amino acid substitutions or subsets of them.

Fig. 7.

Amino acid changes conferring DRV resistance on HIV. (A) Locations of amino acid substitutions V32I, L33A/F, I54M, and I84V associated with HIV's DRV resistance. The location of Asp29 (D29), which is known to be an essential amino acid for dimerization, is also shown. (B) Profiles of DRV's dimerization inhibition of PR carrying a single amino acid substitution. COS7 cells were cotransfected with a pair of HIV-PR^CFP and HIV-PR^YFP variants carrying wild-type PR or a single amino acid substitution such as V32I, L33F, I54M, or I84V, each of which was found to be associated with the development of HIV resistance to DRV, in the presence of 1 μM DRV, further cultured, and the CFP^A/B ratios were determined. Note that none of the amino acid substitutions introduced blocked the dimerization of PR. The statistical evaluation of all the changes in the CFP^A/B ratios determined in the presence or absence of DRV, conducted using the nonparametric Mann-Whitney U test, showed P values ranging from 0.00000034 (3.4E−07) to 0.0026. (C) Profiles of DRV's dimerization inhibition of PR carrying combined amino acid substitutions. COS7 cells were cotransfected with a pair of HIV-PR^CFP and HIV-PR^YFP variants carrying combined amino acid substitutions such as V32I and I84V, V32I, L33F, and I84V, V32I, L33F, and I54M, or V32I, L33F, I54M, and I84V. The COS7 cells were further cultured in the continuous presence of 0, 0.1, and 1 μM DRV, and the CFP^A/B ratios were determined at the conclusion of the 3-day period of culture. The differences between the CFP^A/B ratios in the absence of drug (CFP^{A/B_{No Drug}}) and the CFP^A/B ratios in the presence of 0.1 μM DRV (CFP^{A/B_{0.1 DRV}}) and between the CFP^A/B ratios in the presence of 0.1 μM DRV (CFP^{A/B_{0.1 DRV}}) and 1.0 μM DRV (CFP^{A/B_{1.0 DRV}}) had P values of 0.0015 and 0.42 for V32I and I84V, 0.0047 and 0.15 for V32I, L33F, and I84V, 0.033 and 0.07 for V32I, L33F, and I54M, and 0.3 and 0.0000073 for V32I, L33F, I54M, and I84V, respectively.

When we introduced each of the four substitutions into the wild-type strain, HIV_NL4-3, there was no increase observed in the EC₅₀s of DRV against such infectious recombinant clones, with resistance ranging from 0.07- to 1.0-fold, as shown in Table 3. Introduction of the combinations of two or three amino acid substitutions did not increase the EC₅₀s of DRV against such clones either, with resistance ranging from 0.09- to 0.9-fold. However, when we introduced all four amino acid substitutions (V32I, L33F, I54M, and I84V), into HIV_NL4-3, generating HIV_NL4-3^{V32I/L33F/I54M/I84V}, the EC₅₀ of DRV against HIV_NL4-3^{V32I/L33F/I54M/I84V} was as high as 0.64 ± 0.02 μM with resistance of 205-fold (Table 3). These data suggested that the four amino acid substitutions are associated with the high-level resistance to DRV seen in HIV_8MIX^P51.

Table 3.

The four amino acid substitutions V32I, L33F, I54M, and I84V confer on HIV_NL4-3 variants high-level resistance to DRV^a

Wild-type strain or recombinant HIV variant	Amino acid substitution(s) in PR	Mean EC50 ± SD (μM)	Fold resistance
HIV-1NL4-3	None (wild type)	0.0031 ± 0.0002	1
HIV-1NL4-3V32I	V32I	0.00022 ± 0.00006	0.07
HIV-1NL4-3L33F	L33F	0.0028 ± 0.0008	0.9
HIV-1NL4-3I54M	I54M	0.0026 ± 0.0001	0.8
HIV-1NL4-3I84V	I84V	0.0035 ± 0.0001	1
HIV-1NL4-3V32I/I54V	V32I, I54M	0.0017 ± 0.0002	0.5
HIV-1NL4-3V32I/I84V	V32I, I84V	0.00028 ± 0.00008	0.09
HIV-1NL4-3V32I/L33F/I54V	V32I, L33F, I54M	0.0019 ± 0.0006	0.6
HIV-1NL4-3V32I/L33F/I84V	V32I, L33F, I84V	0.0030 ± 0.0004	0.9
HIV-1NL4-3V32I/L33F/I54MV/82Ab	V32I, L33F, I54M, V82A
HIV-1NL4-3V32I/L33F/I54 M/V82I	V32I, L33F, I54M, V82I	0.034 ± 0.018	11
HIV-1NL4-3V32I/L33F/I54V/I84V	V32I, L33F, I54M, I84V	0.64 ± 0.02	205

^aThe data shown represent mean values derived from the results of three independent experiments conducted in triplicate. The EC₅₀s were determined by employing MT-4 cells exposed to each infectious recombinant HIV-1 clone (50 TCID₅₀) and cultured in the presence of various concentrations of DRV, using the inhibition of p24 Gag protein production by DRV by 50% as an endpoint.

^bNote that there are no EC₅₀ or fold resistance values for HIV-1_NL4-3^{V32I/L33F/I54MV/82A} because this variant was replication incompetent.

We next examined the effects of one of the four amino acid substitutions or subsets of them on PR dimerization inhibition by DRV using the FRET-based HIV expression assay. Figure 7B shows that each single mutation of the four mutations allowed PR to undergo dimerization, and 1.0 μM DRV effectively blocked the dimerization (P values ranging from 3.4E−07 to 0.0026). We next determined the effects of various combinations of the four amino acid mutations on dimerization inhibition by DRV (Fig. 7C). Two mutations such as V32I and I84V still allowed PR to undergo dimerization, and DRV effectively blocked the dimerization at 0.1 and 1 μM. When combinations of three amino acid substitutions such as V32I, L33F, and I84V and V32I, L33F, and I54M were introduced, the mean CFP^A/B ratios were close to 1.0, the threshold for indication of the occurrence of dimerization, with 0.99 and 0.98, respectively, in the absence of DRV, suggesting that possibly such groups of three substitutions somewhat compromised PR dimerization, although DRV still significantly blocked the dimerization at 0.1 and 1 μM, giving mean ratios ranging from 0.75 to 0.87. PR dimerization still occurred with all four substitutions, giving a mean CFP^A/B ratio of 1.11; however, 0.1 μM DRV clearly failed the dimerization, giving a ratio of 1.16 (P = 0.3), while 1.0 μM DRV blocked dimerization (P = 0.00000073) (Fig. 7C).

V82I, not V82A, contributes to the loss of DRV's activity to inhibit PR dimerization.

Since V82A and V82I were seen in _rCLHIV_T48 and HIV_8MIX^P51, respectively (Table 2), we examined whether V82A and V82I had effects on the loss of DRV's protease dimerization inhibition activity. Figure 8 shows that with V82A alone, PR dimerization occurred and 1 μM DRV effectively blocked the dimerization. When V82A was combined with the three amino acid substitutions V32I, L33F, and I54M, PR dimerization still occurred, which DRV effectively blocked at both 0.1 and 1 μM, suggesting that V82A substitution plays no significant role in conferring on HIV reduced sensitivity to the PR dimerization inhibition by DRV. We also examined the effects of the V82I substitution, which was identified in HIV_8MIX^P51, with or without the three amino acid substitutions, on inhibition of PR dimerization by DRV. Interestingly, the four combined amino acid substitutions V32I, L33F, I54M, and V82I abrogated DRV's PR dimerization inhibition activity at both 0.1 and 1.0 μM (Fig. 8), strongly suggesting that V82I with other three amino acid substitutions contributes to the loss of DRV's activity to inhibit PR dimerization.

Fig. 8.

Effects of V82A and V82I substitutions on DRV's activity to inhibit PR dimerization. COS7 cells were cotransfected with the pHIV-PR^CFP and pHIV-PR^YFP pair of plasmids carrying a single V82A or V82I substitution or four combined mutations (V32I, L33F, and I54M plus V82A or V82I). The COS7 cells were further cultured in the continuous presence of 0.1 and 1 μM DRV, and the CFP^A/B ratios were determined at the conclusion of the 3-day period of culture. Note that combined with other three substitutions V32I, L33F, and I54M, V82A did not have a significant effect on DRV's dimerization inhibition activity, while V82I compromised the dimerization inhibition of DRV at 0.1 and 1.0 μM. The differences between the CFP^A/B ratios in the absence of drug (CFP^{A/B_{No Drug}}) and the CFP^A/B ratios in the presence of 1.0 μM DRV (CFP^{A/B_{1.0 DRV}}) had P values of 0.00017 for V82A and 0.0027 for V82I. The differences between the CFP^A/B ratios in the absence of drug (CFP^{A/B_{No Drug}}) and the CFP^A/B ratios in the presence of 0.1 μM DRV (CFP^{A/B_{0.1 DRV}}) and between the CFP^A/B ratios in the presence of 0.1 μM DRV (CFP^{A/B_{0.1 DRV}}) and 1.0 μM DRV (CFP^{A/B_{1.0 DRV}}) were 0.000055 and 0.38 for V32I, L33F, I54M, and V82A and 0.026 and 0.91 for V32I, L33F, I54M, and V82I, respectively.

DISCUSSION

In the FRET-based HIV expression assay we previously reported, YFP- or CFP-containing PR should be initially expressed as a part of Pr160^gag-pol polyprotein, and it was assumed that PDIs block the dimerization of PR monomer subunit within the polyprotein. In the present study, we examined whether one of the PDIs, DRV, blocked the dimerization of HIV PR expressed as a single protease molecule. As expected, DRV effectively blocked the dimerization (Fig. 2B), indicating that PDIs can block the dimerization of the PR monomers in the form of polyprotein as well as after autolysis (30, 31). The dimerization of two identical PR monomers is required for the acquisition of PR catalytic activity (28, 40), and the failure of PR dimerization should result in the loss of viral replication or compromised viral replication.

In our previous study (26) and in the present study, we demonstrated that a single amino acid substitution, such as I3A, L5A, R8Q, L24A, T26A, D29N, R87K, T96A, L97A, or F99A, effectively disrupts PR dimerization, as examined in the FRET-based HIV expression assay. In the present work, we constructed plasmids containing HIV_NL4-3-based recombinant clones with one of those amino acid substitutions and examined if such clones replicated in MT-4 cells. To this end, we attempted to propagate such recombinant clones in MT-4 cells using the supernatants obtained through transfection of 293T cells; however, all clones failed to replicate (Fig. 4B). In contrast, similarly generated recombinant clones containing a single amino acid substitution such as P1A, Q2A, T4A, D30N, and N98A (26), none of which disrupted PR dimerization, as examined with the FRET-based HIV expression assay (26), continued to replicate in MT-4 cells (Fig. 4B). These data further confirmed the validity of the results obtained in the FRET-based HIV expression assay.

It is of note that HIV-1 protease containing A28S has virtually no enzymatic activity, as previously published by Hong et al. (20), and as expected, recombinant HIV-1_NL4-3 with A28S (HIV_NL4-3^A28S) was totally replication incompetent, as shown in Fig. 4B. The D25A substitution is also known to render HIV-1 protease virtually enzymatically inactive; however, that substitution does not disrupt the dimerization of D25A protease as we previously published (26). Yet, 1 μM DRV clearly blocks the dimerization of D25A protease (Fig. 3A), suggesting that D25 is not significantly associated with the putative DRV binding site in monomer subunit. These observations have led us to conclude that A28 is likely involved directly or indirectly in binding of DRV to the protease monomer subunit.

In recent clinical studies, a set of mutations, including V32I, L33F, I47V, I54L, and L89V has been identified in HIV strains isolated from patients failing DRV-containing regimens and has been associated with a diminished virological response to a regimen containing DRV boosted with ritonavir (DRV/r) (9, 33). More recently, Van Marck et al. reported that a set of amino acid substitutions (V32I, I50V, I54L/M, L76V, and V82F) in the PR-encoding region of HIV might be associated with the failure of HIV-1 infected individuals receiving DRV-containing antiviral regimens (38). We have recently selected a highly DRV-resistant HIV variant (HIV_8MIX^P51) by using a mixture of eight highly multi-PI-resistant (but DRV-sensitive) clinical HIV strains, isolated from those who had received various PIs and failed to respond to PI (not DRV)-containing regimens, by propagating the mixture in the presence of increasing concentrations of DRV in phytohemagglutinin-activated peripheral blood mononuclear cells (PHA-PBMCs) followed by MT-4 cells. HIV_8MIX^P51 proved to have an EC₅₀ of DRV ∼333-fold greater than that against a wild-type HIV strain. HIV_8MIX^P51 was highly resistant to amprenavir, indinavir, nelfinavir, ritonavir, lopinavir, and atazanavir (all EC₅₀s of >1 μM) and moderately resistant to saquinavir and TPV (EC₅₀s 33- and 18-fold greater, respectively) and replicated as rapidly as the wild-type HIV_NL4-3 strain in the presence of 1 μM DRV (Table 1). The amino acid substitutions identified in HIV_8MIX^P51 were L10I, I15V, K20R, L24I, V32I, L33F, M36I, M46L, I54M, L63P, K70Q, V82I, I84V, and L89M (24). When we determined the amino acid sequences of three highly DRV-resistant clinical HIV strains, _rCLHIV_F16, _rCLHIV_T45, and _rCLHIV_T48, which had EC₅₀s of DRV of 0.30 μM (97-fold), 0.33 μM (105-fold), and 0.33 μM (105-fold), respectively, we recognized that all three clinical isolates contained V32I, L33F, I54V, and I84V in common.

Thus, we hypothesized that PR dimerization inhibition by DRV might contribute to the antiviral activity of DRV and that the loss of PR dimerization inhibition by DRV might be associated with the decreased antiviral activity of DRV. When we introduced single, double, or triple amino acid substitutions of the four substitutions into pHIV-PR_WT^CFP and pHIV-PR_WT^YFP and cotransfected COS7 cells with such two plasmids, the dimerization of HIV-PR_WT^CFP and HIV-PR_WT^YFP clearly occurred in the absence of DRV, giving CFP^A/B ratios ranging from 1.04 to 1.17 (Fig. 7B). DRV at 1 μM effectively blocked the dimerization of the PR containing a single one of the four amino acid substitutions (Fig. 7B), giving average CFP^A/B ratios ranging from 0.88 to 0.93. DRV at 0.1 and 1.0 μM also blocked dimerization when two amino acid substitutions (V32I and I84V) and three substitutions (V32I, L33F, and I84V or V32I, L33F, and I54M) were introduced. However, when all four combined substitutions were introduced, DRV at 0.1 μM failed to block the dimerization (Fig. 7C), suggesting that the combination of the four substitutions compromised the activity of DRV to block the dimerization, presumably through altering the conformation of the monomer's putative binding site for DRV.

Considering that (i) the four amino acid substitutions in PR V32I, L33F, I54M, and I84V, were identified in the highly DRV-resistant HIV variant (HIV_8MIX^P51) and various DRV-resistant clinical HIV strains, such as _rCLHIV_F16, _rCLHIV_T45, and _rCLHIV_T48, isolated from individuals failing DRV-containing regimens (9, 33); that (ii) multi-PI-resistant, but DRV-sensitive HIV variants (24) do not contain the combination of the four substitutions; and (iii) that with the four substitutions combined, the activity of DRV to block PR dimerization is compromised, it is strongly suggested that the loss of PR dimerization activity is associated with the acquisition of resistance of HIV against DRV. More critically, the fact that the emergence of the four combined mutations is often seen in individuals failing DRV-containing regimens strongly suggests that the protein dimerization inhibition activity of DRV is in operation for the drug to exert its anti-HIV activity in a clinical setting. Our finding that four combined amino acid substitutions are required for the loss of PR dimerization inhibition activity of DRV should explain at least in part the reason why DRV has a high level of genetic barrier to HIV acquisition of DRV resistance. It is noteworthy that the crystallographic data of HIV PR monomer complexed with DRV are as yet unavailable; however, if the results of such structural analysis are obtained, it should help our understanding of the dynamics of the dimerization process of HIV PR, the mechanisms of the activity of DRV and other PDIs to block the dimerization, and the mechanisms of the emergence of HIV resistance against PDIs.