HIV-1 Integrase Inhibitors That Are Active against Drug-Resistant Integrase Mutants

The currently recommended first-line therapy for HIV-1-infected patients is an integrase (IN) strand transfer inhibitor (INSTI), either dolutegravir (DTG) or bictegravir (BIC), in combination with two nucleoside reverse transcriptase inhibitors (NRTIs). Both DTG and BIC potently inhibit most INSTI-resistant IN mutants selected by the INSTIs raltegravir (RAL) and elvitegravir (EVG). BIC has not been reported to select for resistance in treatment-naive patients, and DTG has selected for a small number of resistant viruses in treatment-naive patients.

ABSTRACT

The currently recommended first-line therapy for HIV-1-infected patients is an integrase (IN) strand transfer inhibitor (INSTI), either dolutegravir (DTG) or bictegravir (BIC), in combination with two nucleoside reverse transcriptase inhibitors (NRTIs). Both DTG and BIC potently inhibit most INSTI-resistant IN mutants selected by the INSTIs raltegravir (RAL) and elvitegravir (EVG). BIC has not been reported to select for resistance in treatment-naive patients, and DTG has selected for a small number of resistant viruses in treatment-naive patients. However, some patients who had viruses with substitutions selected by RAL and EVG responded poorly when switched to DTG-based therapies, and there are mutants that cause a considerable decrease in the potencies of DTG and BIC in in vitro assays. The new INSTI cabotegravir (CAB), which is in late-stage clinical trials, has been shown to select for novel resistant mutants in vitro. Thus, it is important to develop new and improved INSTIs that are effective against all the known resistant mutants. This led us to test our best inhibitors, in parallel with DTG, BIC, and CAB, in a single-round infection assay against a panel of the new CAB-resistant mutants. Of the INSTIs we tested, BIC and our compound 4d had the broadest efficacy. Both were superior to DTG, as evidenced by the data obtained with the IN mutant T66I/L74M/E138K/S147G/Q148R/S230N, which was selected by CAB using an EVG-resistant lab strain. These results support the preclinical development of compound 4d and provide information that can be used in the design of additional INSTIs that will be effective against a broad spectrum of resistant mutants.

INTRODUCTION

Integrase (IN) strand transfer inhibitors (INSTIs) bind at the active site of IN. INSTIs interact with the two Mg²⁺ ions and the end of the viral DNA, blocking the second step in the integration of viral DNA into the host genome (the strand transfer reaction). Because integration is an essential step in the life cycle of HIV-1, INSTIs potently inhibit viral replication. INSTIs have favorable tolerability, low toxicity, and in the case of the most recent FDA-approved INSTIs, do not readily select for resistant strains of HIV (1). Dolutegravir (DTG) and bictegravir (BIC) are FDA-approved INSTIs that are widely used in combination antiretroviral therapies (cART). The early INSTIs, raltegravir (RAL) and elvitegravir (EVG), are still used in certain clinical situations; however, it is much easier for the virus to develop resistance to RAL and EVG (2–5). Resistance to RAL and EVG typically follows three pathways that are based on amino acid substitutions at Y143, N155, and Q148. In many cases, additional substitutions give rise to double, triple, or more complex IN mutants. Only limited resistance has been reported when DTG was used in treatment-naive patients (6), and so far, little or no resistance has been reported in BIC-treated patients (7, 8). As shown in Fig. 1, DTG and BIC are more extended than RAL and EVG. INSTIs have a linker that connects the main scaffold, which chelates the metal ions at the active site, to a halogenated benzyl moiety, which stacks with the nucleobase of the deoxycytidine (dC) near the 3′ end of the viral DNA in the catalytic site. This linker is longer in the most recent FDA-approved INSTIs (9–11), which may help BIC and DTG adjust their conformations to fit the changes in the structure of the IN active site that are caused by amino acid substitutions that make IN resistant to RAL and EVG. Although DTG is able to inhibit most of the well-known INSTI-resistant mutants that were selected in response to treatment with RAL and EVG (10, 12, 13), DTG treatment of patients who had already developed resistance to RAL and EVG led to the selection of additional amino acid substitutions that made the virus DTG resistant (14, 15). The newest INSTI to be FDA approved, BIC, is even more effective than DTG at inhibiting the known INSTI-resistant triple mutants, yet there are IN mutants that cause a substantial reduction in the potency of BIC (16, 17). Moreover, the R263K resistance mutation was recently identified in an INSTI-experienced patient who switched to BIC (18). In addition, at least one INSTI-resistant mutant has been selected by BIC in vitro (17).

FIG 1.

Chemical structures of the INSTIs. The chemical structures of the clinically relevant INSTIs (DTG, BIC, and CAB) and compounds (4c, 4d, 4f, 6b, and 6p) used in this study are shown.

Cabotegravir (CAB), a new INSTI which was recently licensed in Canada, has a potential therapeutic advantage compared to the FDA-approved INSTIs (19–21). CAB can be formulated so that, when it is injected into patients, its plasma concentration is maintained at an effective level for months. However, CAB also has an apparent weakness. Recent in vitro studies have shown that there are several INSTI-resistant mutants that substantially raise the EC₅₀ values for CAB in single-round infection assays (16, 22), suggesting that there may be problems with the development of resistance in patients. Moreover, studies using laboratory strains and patient-derived clinical isolates have identified IN mutants that are relatively insensitive to CAB. Most of these strains have amino acid substitutions at IN positions G140 and Q148; however, additional substitutions have also been selected (22). Taken together, the available data underscore the importance of developing new INSTIs that are broadly effective against the known resistant mutants.

We recently developed 4-amino-1-hydroxy-2-oxo-1,8-naphthyridine-containing compounds with different modifications at the 6-position on the naphthyridine scaffold (Fig. 1). Several of these compounds potently inhibit most of the well-characterized RAL-, EVG-, and DTG-resistant mutants (12, 13, 23). We previously showed that an analog with an n-hexanol 6-substituent (compound 4d) is more effective than either DTG or BIC in terms of its ability to inhibit a panel of INSTI-resistant triple mutants (23). Using a newly developed panel of INSTI-resistant mutants, including mutants recently selected in vitro by CAB (22, 24), we investigated to what extent the most recently FDA-approved INSTIs and our naphthyridine-containing compounds can inhibit these new mutants. Our results demonstrate that compound 4d and BIC exhibit the broadest efficacies against the new panels of IN mutants, that their potencies are similar, and that, in terms of their ability to inhibit the new panel of mutants, both are superior to CAB and DTG.

RESULTS

Comparison of antiviral potencies against a panel of mutants selected by CAB.

CAB is a structural analog of DTG and BIC that we previously showed is inferior to both DTG and BIC in terms of its ability to inhibit a large panel of INSTI-resistant mutant HIV strains (16). Recent in vitro selection studies identified several new mutants that are resistant to CAB when in vitro selections were initiated with a Q148H/K/R mutant virus (22). The majority of the new resistant mutants were identified in cell culture experiments in which the starting IN mutant was Q148H. However, a small number of resistant mutants were identified starting with Q148K and Q148R mutant viruses. Most of the new mutants had changes at more than one secondary position. The L74 and V75 amino acid substitutions are located within ∼5.0 Å of the metal ions bound at the IN active site, while T122 is farther away (∼13.0 Å). The Q148H mutant is often accompanied by the G140S substitution, which is a well-known compensatory substitution, and by additional substitutions at secondary positions. We determined the antiviral potencies of DTG, BIC, CAB, and our INSTIs against the following INSTI-resistant IN mutants that were selected by CAB in vitro: L74M/Q148R, V75A/G140S/Q148H, T122N/G140S/Q148H, and L74M/V75A/G140S/Q148H (Fig. 2; see Table S1 in the supplemental material).

FIG 2.

Comparison of antiviral potencies against a panel of INSTI-resistant mutants selected in vitro using CAB. The EC₅₀ values were determined for the clinically relevant INSTIs and compounds using vectors that carry INSTI-resistant double, triple, and quadruple mutants in a single-round infection assay, up to a maximum concentration of 100 nM. Asterisks denote EC₅₀ values that are greater than 100 nM. The error bars represent standard deviations of independent experiments, n = 3, performed in triplicate. The potency of the INSTIs and our compounds against WT HIV-1 were previously determined (16, 23).

In order to compare the effectiveness of the INSTIs, we calculated the fold change (FC) in the 50% effective concentration (EC₅₀) values for the IN mutants versus wild-type (WT) HIV-1. When we compared the relative potencies of our promising INSTIs, we made use of the Monogram FC cutoffs for resistance for BIC (2.5) and the two FC cutoffs for DTG (4 and 13) (17, 25). Because we have three different FCs to choose from, it is not clear which should be used for our compounds or for CAB. FCs below 2.5 are likely to be susceptible, and above 13 are likely to be resistant; however, between 2.5 and 13, the interpretation is more difficult. Differences between the potencies among the INSTIs were considered significant if the calculated P values were <0.05.

Overall, compound 4d and BIC were the most potent against this panel of CAB-resistant mutants; compound 4d appears to have been slightly better than BIC. Only compound 4d (EC₅₀, 3.1 ± 0.6 nM; FC, 1.3) retained full potency (FC, <2.5) against the L74M/Q148R double mutant, which was slightly better than BIC (EC₅₀, 6.4 ± 1.4 nM; FC, 3.4; P < 0.05). Compound 4f and DTG inhibited the L74M/Q148R double mutant with modestly reduced EC₅₀ values (EC₅₀, 7.3 ± 1.4 nM; FC, 3.7 and EC₅₀, 10.8 ± 2.9 nM; FC, 6.8, respectively). Conversely, the L74M/Q148R double mutant caused a minor loss in susceptibility to compound 6p (EC₅₀, 7.2 ± 0.4 nM; FC, 10.3), compound 4c (EC₅₀, 11.8 ± 2.6 nM; FC, 9.1), and compound 6b (EC₅₀, 14.3 ± 3.8 nM; FC, 10.2) and a major decrease in susceptibility to CAB (EC₅₀, 43.0 ± 11.7 nM; FC, 17.9), confirming the in vitro selection studies (22). The more complex IN mutants (defined as having three or more amino acid substitutions), which comprised the RAL-resistant G140S/Q148H substitutions plus an additional substitution(s) at a secondary position (V75A/G140S/Q148H, T122N/G140S/Q148H, and L74M/V75A/G140S/Q148H), reduced the potencies of DTG (all FCs, >10.0), BIC (all FCs, >2.5), CAB (all FCs, >17.0), and several of our INSTIs. However, both compound 4d and BIC were better, in terms of their residual potency, against these CAB-resistant mutants. Compound 4d inhibited V75A/G140S/Q148H, T122N/G140S/Q148H, and L74M/V75A/G140S/Q148H with EC₅₀ values of 9.3 ± 1.1 nM (FC, 4.0), 16.5 ± 2.6 nM (FC, 7.2), and 18.5 ± 3.5 nM (FC, 8.0), respectively, whereas BIC inhibited V75A/G140S/Q148H, T122N/G140S/Q148H, and L74M/V75A/G140S/Q148H with EC₅₀ values of 9.7 ± 1.6 nM (FC, 5.1), 14.3 ± 3.7 nM (FC, 7.5), and 21.7 ± 2.5 nM (FC, 11.4), respectively. Both DTG and compound 4c also showed reduced potencies against V75A/G140S/Q148H (EC₅₀, 17.2 ± 1.0 nM and FC, 10.8; EC₅₀, 19.4 ± 3.0 nM and FC, 14.9, respectively). When tested against V75A/G140S/Q148H, CAB (EC₅₀, 66.4 ± 14.2 nM; FC, 27.7), compound 4f (EC₅₀, 91.6 ± 12.6 nM; FC, 45.8), compound 6b (EC₅₀, 46.3 ± 4.3 nM; FC, 33.1), and compound 6p (EC₅₀, 85.8 ± 10.4 nM; FC, 122.6) lost considerable potency. In addition, DTG and CAB and compounds 4c, 4f, 6b, and 6p lost potency (EC₅₀ values, >29.0 nM; all FCs, >22.9) against the T122N/G140S/Q148H and L74M/V75A/G140S/Q148H resistant mutants. The results of these single-round infection assays reinforce the idea that there are complex IN mutants with three or four amino acid substitutions that cause a significant loss of potency against the recently approved INSTIs as well as some of our naphthyridine-scaffold compounds.

Comparison of antiviral potencies against a second panel of INSTI-resistant triple and quadruple mutants that were selected by CAB.

As described above, passage of a virus that carried the Q148H substitution in the presence of CAB led to selection of G140S (22). This is not surprising, since the G140S substitution is compensatory to Q148H, which arises in response to partial loss of IN function (11, 26). The G140S/Q148H double mutant is a signature RAL-resistant mutant (9) that also arises in response to treatment with the latest FDA-approved INSTIs. Additional passage in the presence of CAB selected two new secondary substitutions, C56S and G149A, which are within approximately 18.5 Å and 11.5 Å, respectively, of the metal ions bound at the IN active site and were reported not to enhance the fitness of the virus in cultured cells (23). We tested DTG, BIC, CAB, and our inhibitors against the new accessory amino acid substitutions C56S and G149A alone and against the complex mutant C56S/G140S/Q148H/G149A (Fig. 3; Table S2). Two additional mutants, G140S/G149A and C56S/G140S/Q148H, were also prepared and tested. The IN mutants C56S, G149A, and G140S/G149A were all susceptible to DTG, BIC, CAB, and to the INSTIs that we developed. The EC₅₀ values were <5.2 nM against C56S, and FCs were <2.1 for clinically relevant INSTIs and compound 4d. The EC₅₀ values were <6.1 nM against G149A, and the FCs were <2.2 for clinically relevant INSTIs and compounds 4c, 4d, and 4f. Lastly, the EC₅₀ values were <3.4 nM against G140S/G149A, and the FCs were <2.4 for all INSTIs. The C56S/G140S/Q148H mutant caused a substantial reduction in susceptibility to CAB (EC₅₀, 36.7 ± 5.7 nM; FC, 15.2), compound 6b (EC₅₀, 44.4 ± 11.8 nM; FC, 31.7), and compound 6p (EC₅₀, 57.5 ± 0.9 nM; FC, 82.1) and caused a very significant drop in susceptibility to compound 4f (EC₅₀, 229.0 ± 21.0 nM; FC, 114.5). However, DTG (EC₅₀, 11.3 ± 1.3 nM; FC, 7.1), BIC (EC₅₀, 7.6 ± 1.1 nM; FC, 4.0), and compound 4d (EC₅₀, 10.3 ± 1.2 nM; FC, 4.5) retained most of their potency against this IN mutant. With this set of mutants, BIC was superior to both DTG and compound 4d (P < 0.05). As previously shown (27), the complex IN mutant G140S/Q148H/G149A causes drops in susceptibility against all of the INSTIs we tested. DTG (EC₅₀, 107.8 ± 9.7 nM; FC, 67.4), CAB (EC₅₀, 197.0 ± 27.6 nM; FC, 82.1), compound 4f (EC₅₀, 1,346.7 ± 232.9 nM; FC, 673.4), compound 6b (EC₅₀, 428.5 ± 60.9 nM; FC, 306.1), and compound 6p (EC₅₀, 167.8 ± 23.9 nM; FC, 239.7) did not effectively inhibit this IN mutant. Compound 4c exhibited a substantial loss of potency (EC₅₀, 58.7 ± 3.3 nM; FC, 15.1), whereas BIC (EC₅₀, 25.6 ± 2.4 nM; FC, 4.0) and compound 4d (EC₅₀, 15.3 ± 1.1 nM; FC, 4.5) showed more modest losses of potency. The complex IN mutant C56S/G140S/Q148H/G149A caused a substantial reduction in the potencies of the majority of the INSTIs. Compounds 4f (EC₅₀, 768.6 ± 67.2 nM; FC, 384.3), 6b (EC₅₀, 298.2 ± 24.6 nM; FC, 213.0), and 6p (EC₅₀, 125.6 ± 15.6 nM; FC, 179.4), CAB (EC₅₀, 127.0 ± 0.9 nM; FC, 52.9), and compound 4c (EC₅₀, 53.0 ± 7.4 nM; FC, 40.8) were ineffective at inhibiting this mutant. The mutant caused a loss of potency for DTG (EC₅₀, 44.8 ± 8.2 nM; FC, 28.0) and a somewhat smaller reduction in potency for compound 4d (EC₅₀, 27.4 ± 3.1 nM; FC, 11.9). BIC, which showed some loss of potency (EC₅₀, 18.8 ± 0.5 nM; FC, 9.9), was the most potent inhibitor against this IN mutant compared to the best of our INSTIs, compound 4d (P < 0.05); however, compound 4d was superior to DTG (P < 0.05). Considering the data for all the mutants in this panel, BIC was the best performer; however, compound 4d was only slightly inferior. In contrast, CAB had severe deficiencies inhibiting the mutants in this panel. Its best EC₅₀ value, against the INSTI-resistant triple mutant C56S/G140S/Q148H, was 36.7 ± 5.7 nM (FC, 15.2). The INSTIs compounds 4f, 6b, and 6p also had problems retaining potency against the complex INSTI-resistant mutants in this panel.

FIG 3.

Comparison of antiviral potencies against a panel of INSTI-resistant mutants selected in vitro using CAB. The EC₅₀ values were determined for the clinically relevant INSTIs and compounds using vectors that carry INSTI-resistant single, double, triple, and quadruple mutants in a single-round infection assay, up to a maximum concentration of 100 nM. Asterisks denote EC₅₀ values that are greater than 100 nM. The error bars represent standard deviations of independent experiments, n = 3, performed in triplicate. The potency of the INSTIs and our compounds against WT HIV-1 were previously reported (16, 23).

Comparison of antiviral potencies against a panel of INSTI-resistant quadruple mutants selected by CAB from clinical isolates.

Recently, additional attempts to isolate HIV mutants against DTG, BIC, and CAB were reported (24). Clinical isolates were treated with DTG, BIC, and CAB; however, only CAB selected new mutants, L74M/E138K/Q148R/R263K and L74M/G140S/S147G/Q148K. We determined the effects of the L74M/E138K/Q148R/R263K mutant on the potencies of BIC, DTG, CAB, and the INSTIs that we developed (Fig. 4; Table S3). The quadruple mutant caused an extensive drop in potency against CAB (EC₅₀, 104.9 ± 22.5 nM; FC, 43.7), confirming the results of the selection studies. Compound 4f (EC₅₀, 151.3 ± 19.1 nM; FC, 75.7) failed to effectively inhibit this mutant. Both DTG (EC₅₀, 9.0 ± 0.3 nM; FC, 5.6) and BIC (EC₅₀, 12.9 ± 2.1 nM; FC, 6.8), and to a certain extent, compound 4d (EC₅₀, 18.2 ± 2.3 nM; FC, 7.9), retained considerable potency against this complex mutant. The remaining compounds displayed a moderate loss in potency and include 4c (EC₅₀, 22.5 ± 4.7 nM; FC, 17.3), 6b (EC₅₀, 20.1 ± 3.0 nM; FC, 14.4), and 6p (EC₅₀, 20.0 ± 2.3 nM; FC, 28.6). If we use an FC of 4.0 as a cutoff, this complex IN mutant is resistant to BIC and partially sensitive to DTG. Compound 4d was the most potent inhibitor among our INSTIs; however; it was inferior to both BIC (P < 0.05) and DTG (P < 0.05).

FIG 4.

Comparison of antiviral potencies against an INSTI-resistant mutant selected by CAB using a clinical isolate. The EC₅₀ values were determined for the clinically relevant INSTIs and compounds using vectors that carry INSTI-resistant quadruple mutants in a single-round infection assay, with a maximum concentration of 200 nM. The error bars represent standard deviations of independent experiments, n = 3, performed in triplicate. The potency of the INSTIs and our compounds against WT HIV-1 were previously determined (16, 23).

Comparison of antiviral potencies against a panel of INSTI-resistant mutants selected by DTG, BIC, and CAB.

Additional mutants have been selected with DTG, BIC, and CAB using recombinant strains that carried a patient-derived IN with or without the E157Q amino acid substitution (24, 28). When the E157Q substitution was present, the predominant mutant selected by DTG, BIC, and CAB was R263K. CAB selected L74I/E138K/G140S/Q148R, and Q95K/Q146R was selected by both DTG and BIC. We tested the ability DTG, CAB, BIC, and our INSTIs to inhibit the following IN mutants: L74I, Q146R, E157Q, Q95K/Q146R, E157Q/R263K, and L74I/E138K/G140S/Q148R (Fig. 5; Table S4). All of the INSTIs potently inhibited the single mutants L74I, Q146R, and E157Q with EC₅₀ values of ≤5.0 nM (FCs, <3.2 for the clinically relevant INSTIs and FCs of <4.0 for our INSTIs). There were no obvious differences in the antiviral activities of the FDA-approved INSTIs and compound 4d against L74I and Q146R except that BIC had a marginally improved efficacy compared to DTG against Q146R (P < 0.01). However, compound 4d exhibited slightly better potencies against E157Q than either BIC (P < 0.05) or DTG (P < 0.01). All the compounds inhibited the double mutants Q95K/Q146R and E157Q/R263K with EC₅₀ values of ≤8.0 nM (FCs of < 3.8 for clinically relevant INSTIs and FCs of <5.9 for our INSTIs). Our findings suggest that E157Q does not lead to hyper-susceptibility to DTG, nor did we find that E157Q/R263K causes a large reduction in susceptibility to DTG, as was reported previously (28). Compound 6p was the most effective INSTI against Q95K/146R. Compound 6p was better than compound 4d (P < 0.05), DTG (P < 0.05), BIC (P < 0.001), and CAB (P < 0.05). The differences in the potencies of the INSTIs against the mutant E157Q/R263K were not significant. However, the complex IN mutant L74I/E138K/G140S/Q148R caused a minor decrease in susceptibility to CAB (EC₅₀, 12.9 ± 1.9 nM; FC, 5.4). This mutant also caused a modest loss in susceptibility to compound 6p (EC₅₀, 13.6 ± 0.4 nM; FC, 19.4) and a substantial loss of potency for compound 4f (EC₅₀, 40.6 ± 4.2 nM; FC, 20.3). BIC (EC₅₀, 2.6 ± 0.1 nM; FC, 1.4), DTG (EC₅₀, 4.2 ± 0.8 nM; FC, 2.6), and compound 4d (EC₅₀, 4.5 ± 0.3 nM; FC, 2.0) potently inhibited this IN mutant, while compounds 4c (EC₅₀, 9.0 ± 1.7 nM; FC, 6.9) and 6b (EC₅₀, 7.8 ± 0.6 nM; FC, 5.6) showed minor losses of potency. The complex IN mutant L74I/E138K/G140S/Q148R was susceptible to BIC, DTG, and compound 4d. The changes in efficacy between BIC and DTG and the changes between compound 4d and DTG were not significant. However, there was a significant difference in potency between BIC and compound 4d (P < 0.01), and BIC and DTG were the superior INSTIs against this complex IN mutant.

FIG 5.

Comparison of antiviral potencies against a panel of INSTI-resistant mutants selected in vitro using recombinant strains with a patient-derived IN with or without an E157Q substitution. The EC₅₀ values were determined for the clinically relevant INSTIs and compounds using vectors that carry INSTI-resistant single, double, and quadruple mutants in a single-round infection assay, with a maximum concentration of 50 nM. The error bars represent standard deviations of independent experiments, n = 3, performed in triplicate. The potency of the INSTIs and our compounds against WT HIV-1 were previously reported (16, 23).

Comparison of antiviral potencies against a panel of INSTI-resistant mutants selected by DTG, BIC, and CAB using EVG-resistant lab strains.

Starting with recombinant EVG-resistant laboratory strains, mutant viruses were selected using DTG, BIC, and CAB (24). Passaging the EVG-resistant mutant T66I in the presence of DTG selected for the complex IN mutant T66I/L74M/E138K/S147G/M154I. In addition, when the EVG-resistant mutant T66I/E138K/S147G/Q148R was passaged in the presence of DTG, BIC, and CAB, the mutant T66I/E138K/S147G/Q148R/S230N was selected by DTG and BIC, and the IN mutant T66I/L74M/E138K/S147G/Q148R/S230N was selected by CAB (24). There is a previous report that S230N is selected by a number of early-stage INSTIs (29). We measured the potencies of DTG, BIC, CAB, and our compounds against the following IN mutants: M154I, S230N, T66I/R263K, T66I/L74M/E138K/S147G/M154I, and T66I/L74M/E138K/S147G/Q148R/S230N (Fig. 6, Table S5).

FIG 6.

Comparison of antiviral potencies against a panel of INSTI-resistant mutants selected by CAB using EVG-resistant lab strains. The EC₅₀ values were determined for the clinically relevant INSTIs and compounds using vectors that carry INSTI-resistant single, double, and complex mutants in a single-round infection assay, with a maximum concentration of 100 nM. The asterisk denotes EC₅₀ values that are greater than 100 nM. The error bars represent standard deviations of independent experiments, n = 3, performed in triplicate. The potency of the INSTIs and our compounds against WT HIV-1 were previously determined (16, 23).

DTG, BIC, CAB, and our compounds all had EC₅₀ values of <5.0 nM against the IN mutants M154I (FCs of <1.8 for all INSTIs) and T66I/R263K. In a single-round infection assay, all the INSTIs we tested potently inhibited the T66I/R263K double mutant, with EC₅₀ values of <3.0 nM (FCs of < 1.9 for all INSTIs). The DTG-resistant mutant S230N caused a minor loss of DTG potency (EC₅₀, 7.9 ± 1.3 nM; FC, 4.9) and a similar loss of potency to compound 6b (EC₅₀, 6.8 ± 0.9 nM; FC, 4.9). However, BIC (EC₅₀, 2.6 ± 0.6 nM; FC, 1.4), CAB (EC₅₀, 3.2 ± 0.5 nM; FC, 1.3), compound 4c (EC₅₀, 2.4 ± 0.2 nM; FC, 1.8), compound 4d (EC₅₀, 3.5 ± 0.6 nM; FC, 1.5), compound 4f (EC₅₀, 3.5 ± 0.7 nM; FC, 1.8), and compound 6p (EC₅₀, 1.7 ± 0.7 nM; FC, 2.4) potently inhibited this IN mutant. Both BIC and compound 4d were superior to DTG (P < 0.01 and P < 0.05, respectively); however, the differences in the efficacies between BIC and compound 4d were not significant. All of the INSTIs were effective inhibitors of the complex IN mutant T66I/L74M/E138K/S147G/M154I (EC₅₀, <5.0 nM; FCs, <2.2 for all INSTIs); however, BIC was slightly superior compared to compound 6p (P < 0.05), compound 4d (P < 0.01), and DTG (P < 0.01), but compound 6p was superior compared to both compound 4d (P < 0.05) and DTG (P < 0.01). Conversely, BIC (EC₅₀, 10.3 ± 0.9 nM; FC, 5.4) and compound 4d (EC₅₀, 12.5 ± 3.0 nM; FC, 5.4) were equally potent inhibitors of the complex IN mutant T66I/L74M/E138K/S147G/Q148R/S230N. The IN mutant T66I/L74M/E138K/S147G/Q148R/S230N caused a substantial loss in the potency of DTG (EC₅₀, 52.7 ± 4.6 nM; FC, 32.9) and compounds 4c (EC₅₀, 40.1 ± 4.7 nM; FC, 30.8), 4f (EC₅₀, 46.7 ± 1.3 nM; FC, 23.4), and 6p (EC₅₀, 39.3 ± 5.8 nM; FC, 56.1). This mutant also caused a very significant loss in susceptibility to CAB (EC₅₀, 182.4 ± 17.7 nM; FC, 76.0) and compound 6b (EC₅₀, 95.9 ± 5.7 nM; FC, 68.5). Both BIC and compound 4d were superior to DTG in terms of efficacy against this IN mutant (P < 0.01 and P < 0.001, respectively). The Q148R amino acid substitution arises in response to treatment to both the early and most recent FDA-approved INSTIs (16, 17, 30–33). This suggests that at least part of the loss in potency against T66I/L74M/E138K/S147G/Q148R/230N is due to the presence of the Q148R substitution.

Single-round replication of the INSTI-resistant mutants.

INSTI-resistant mutants typically have a reduced ability to replicate. In addition to the INSTI-resistant mutants used in this study (Fig. 7; Table S6), we also measured the replication of seven other mutants (M50I/R263K, S119R/R263K, T124A/S153Y, M50I/S119R/R263K, V72I/E138K/Q148K, G140S/Q148H/G149A, and L74M/G140S/S147G/A148K) whose effects on the susceptibility to various INSTIs had been reported earlier (27). The antiviral data for DTG, BIC, and compounds 4c, 4d, and 4f against these INSTI-resistant mutants is included in the supplementary data to simplify comparisons (Fig. S1, upper and lower panels) (27). The INSTI-resistant single mutants C56S, L74I, Q146R, G149A, M154I, and E157Q caused minor reductions in replication compared to WT HIV-1 (70 to 85% of WT HIV-1), while the INSTI-resistant single mutants S230N and R263K caused a slightly larger drop in replication (60 to 65% of WT HIV-1). However, the one-round replication of the INSTI-resistant double mutants varied considerably. The replication of the double mutant M50I/R263K was approximately 70% of WT HIV-1, while the INSTI-resistant double mutants S119R/R263K and Q95K/Q146R were 60% and 65% of WT, respectively. The rates of replication of the INSTI-resistant double mutants L74M/Q148R and G140S/G149A were approximately 40% of WT HIV-1, while the rates of replication for T66I/R263K and E157Q/R263K were 22% and 52% of WT, respectively. The G140S/G149A mutant had a lower replication rate (43%) than what we previously reported for the G140S/Q148H double mutant (66%) (23); however, this was not a significant difference. The replication of the six INSTI-resistant triple mutants, M50I/S119R/R263K, C56S/G140S/Q148H, V72I/E138K/Q148K, V75A/G140S/Q148H, T122N/G140S/Q148H, and G140S/Q148H/G149A, and the two quadruple mutants, C56S/G140S/Q148H/G149A and L74M/V75A/G140S/Q148H, were all between 45 and 70% of WT HIV-1. None of the additional single substitutions, C56S, V75A, T122N, and G149A, that were found with the G140S/Q148H double mutant significantly increased or decreased the replication of the virus, nor did the addition of C56S and G149A or L74M/V75A to the G140S/Q148H double mutant. Conversely, the IN-resistant quadruple mutants L74I/E138K/G140S/Q148R, L74M/E138K/Q148R/R263K, and L74M/G140S/S147G/Q148K were deficient in single-round replication assays compared to the WT, ranging from 13 to 23%. The quintuple IN-resistant mutant T66I/L74M/E138K/S147G/M154I had a higher single-round replication rate than the other three quadruple mutants (35% of WT HIV-1), but these differences were not significant. The lone sextuple IN-resistant mutant, T66I/L74M/E138K/S147G/Q148R/S230N, had a replication rate similar to that of the complex IN mutants that contained the G140S and Q148H substitutions, 61% of WT HIV-1, which was a significant difference compared to the IN-resistant quadruple mutants (P < 0.05) but not the IN-resistant quintuple mutant. These data show that there are complex IN substitutions which substantially reduce the susceptibility of the virus to BIC, DTG, and CAB and are still able to infect cultured cells reasonably well.

FIG 7.

Replication of the INSTI-resistant mutants using a single-round infection assay. The replication of the INSTI-resistant mutants used in this study were measured using the vectors carrying the appropriate resistance substitutions in a single-round infection assay. The luciferase activity was set to 100, and the replication of the mutant vectors (adjusted for the amount of p24/Gag used in the assay) were then measured and compared to the WT. The error bars represent standard deviations of independent experiments, n = 4, done in triplicate.

DISCUSSION

The development of INSTIs has led to improvements in the treatment options available for HIV-1-infected individuals, and this family of compounds is widely used in combination therapies. Unfortunately, the early INSTIs RAL and EVG have proven to be quite susceptible to the development of resistance. DTG, BIC, and CAB are much less susceptible to the development of resistance. However, there are mutants, including G140S/Q148H and more complex mutants that include these two substitutions or other changes at the same positions (G140A/C/S and Q148H/K/R), that still present a considerable challenge to all of the FDA-approved INSTIs. Although further investigation is still needed, there is an emerging consensus that certain substituents can be appended to INSTIs to improve their binding to mutant forms of IN, thereby broadening the efficacy of the compounds against emerging drug-resistant viruses.

Here, we compare the efficacies of DTG, BIC, CAB, and several of our most promising INSTIs against a variety of recently described CAB-resistant mutants (22, 24). As we previously reported, INSTI-resistant mutants with single substitutions that are located ∼5.0 Å from the metal ions bound at the IN active site typically do not cause substantial decreases in susceptibility to DTG and BIC (23). The S230N mutant decreased susceptibility and increased FCs to both DTG (FC, 4.9) and compound 6b (FC, 2.4). The G149A mutant causes a minor decrease in susceptibility and increase in FC (4.4) to compound 6b. In the current work, many of the INSTI-resistant double mutants we tested were potently inhibited by DTG, BIC, CAB, and our best INSTIs; however, CAB lost potency against the L74M/Q148R mutant. Nonetheless, there are more complex IN mutants, defined here as mutants with at least three substitutions (and with as many as six), that pose more serious problems. The IN-resistant quintuple mutant T66I/L74M/E138K/S147G/M154I was the only complex IN mutant that was potently inhibited by all of the INSTIs (FCs, <2.2). One other complex IN mutant, L74I/E138K/G140S/Q148R, was potently inhibited by six of the INSTIs (FCs of <2.6 for the FDA-approved INSTIs and compound 4d; FCs of <5.6 for CAB and compound 6b). All of the complex IN mutants that included an amino acid substitution at position Q148 (H or R) showed substantially decreased susceptibility and increased FCs to at least some of the INSTIs. We previously showed that some complex IN mutants that include the Q148H/K substitutions cause a significant loss of susceptibility to the INSTIs described in the current report (16, 23). There are several IN mutants with a substitution at position Q148H/K/R plus additional substitutions at a position in the hydrophobic pocket (L74 or V75) near the active site residue D64, and the E138K and/or G140S substitutions caused a reduction in susceptibility to even the best INSTIs we tested (Fig. 8).

FIG 8.

Atomic model of BIC in the active site of the HIV-1 intasome. The structure of the active site of an HIV intasome, with BIC bound, was prepared based on the published structure (27). Colors are used to show the positions of the resistance substitutions (lime green) discussed in this study, the catalytic residues D64, D116, and E152 (DDE) (light blue), viral DNA (light gray), and BIC (dark olive green). IN chains near the catalytic active site are labeled.

We compared the relative potencies of our promising INSTIs in assays using a group of newly reported resistant mutants, which were selected in vitro. The FCs for the INSTIs against the IN mutants relative to WT HIV-1 were sorted according the published biological cutoffs (FCs of <2.5, FCs of 2.5 to <4.0, FCs of 4.0 to <13.0, and FCs of >13.0), and a percentage was calculated based on the 21 IN mutants tested (Table S7). Compound 4d potently inhibited 67% of the IN mutants we tested with an FC of <2.5 and inhibited all of the IN mutants with an FC of <13.0. When we compared compound 4d to our other compounds, only compound 4f inhibited >50% of the IN mutants with an FC of <2.5. Compounds 4c, 4f, 6b, and 6p inhibited around one-third of the IN mutants with FCs of >13.0. Against this new panel, using an FC of <2.5, BIC potently inhibited 57% of the IN mutants. If the lower cutoff for DTG is used (FC, <4.0), BIC effectively inhibited 71% of the IN mutants. Using FCs of <4.0, both DTG and CAB inhibited 57% of the IN mutants. We previously showed that, against a larger panel, DTG and BIC are more broadly effective than CAB (16). Broadly speaking, CAB was effective against the IN single and double mutants but not against complex IN mutants. BIC had lower FCs against five of the complex IN mutants compared to compound 4d (approximately the same FC was seen for BIC and compound 4d against T66I/L74M/E138K/S147G/Q148R/S230N) (Table S8). Compound 4d had lower FCs than BIC against three of the complex IN mutants. Only the differences in the FCs that were lower for BIC than for compound 4d were significant, and they suggested that BIC was the overall most effective INSTI against the current panel of mutants. DTG was better than compound 4d and BIC against only one of the complex mutants, L74M/E138K/Q148R/R263K (FC, 5.6). The difference in potency was only significant when compared against compound 4d, and not BIC, suggesting that DTG was not as effective against these complex IN mutants. None of the compounds were fully effective against the five complex IN mutants we tested that included the G140S and Q148H substitutions. The two most promising compounds, BIC and compound 4d, did not potently inhibit any of the mutants in this group (EC₅₀, >5.0 nM; FCs, >4.0). Increased FCs were observed for BIC, DTG, and compound 4d against the four complex IN mutant strains, T122N/G140S/Q148H (FCs, 7.5, 36.7, and 7.2, respectively) L74M/V75A/G140S/Q148H (FCs, 11.4, 55.1, and 8.0, respectively), C56S/G140S/Q148H/G149A (FCs, 9.9, 28.0, and 11.9, respectively), and T66I/L74M/E138K/S147G/Q148R/S230N (FCs, 5.4, 32.9, and 5.4, respectively).

In some cases, amino acid substitutions that allow IN (and HIV replication) to evade INSTIs compromise the ability of the mutant viruses to replicate. The complex IN mutants V72I/E138K/Q148K (47.4 ± 4.3% of WT), L74M/G140S/S147G/Q148K (23.2 ± 4.5% of WT), and L74M/V75A/G140S/Q148H (53.4 ± 7.5% of WT), all of which caused large decreases in susceptibility to all the INSTIs, also reduced replication in single-round infectivity assays (Fig. 7; Table S6). L74I/E138K/G140S/Q148R and L74M/E138K/Q148R/R263K caused a substantial loss of replication in single-round assays to ∼15% of WT HIV-1 activity. However, the other complex IN mutants had single-round replication rates of ∼60% of WT HIV-1 activity. It is likely that many of these amino acid substitutions help the mutants evade the INSTIs; however, at least some of these substitutions may enhance the catalytic activity of the complex IN mutants in ways that improve the ability of the mutant virus to replicate. Earlier studies, based on prototype foamy virus (PFV) IN, show that there is a precedent for this type of compensatory substitution (11).

The ways in which substitutions in HIV IN affect the ability of the virus to replicate and cause substantial losses in susceptibility to INSTIs can be investigated through a combination of structural analysis and modeling (27, 34). There are now structures of INSTIs bound to HIV IN and to the IN of a simian immunodeficiency virus (SIV) isolated from the red-capped mangabey (SIVrcm) (27, 34). The structures are quite similar, and we used the structure of BIC bound to HIV IN to investigate the ways the Q148H/K/R, G140S, E138T/A/K, and R263K amino acid substitutions could reduce the potency of the most recent FDA-approved INSTIs.

The residue Q148 occupies a strategic position in the central core, contacting a key water molecule that coordinates two important catalytic residues, D116 and E152 (24, 27). The mechanism by which the Q148H substitution causes resistance has recently been described based on experiments done with the intasome of a SIVrcm (34). The Q148H substitution (and by analogy, Q148R/K) displaces a water molecule that, in WT SIVrcm and HIV-1 IN, is engaged in coordinating the catalytic residues D116 and E152 (Fig. S2). The substituted amino acid introduces a local electropositive charge that withdraws electron density from the Mg²⁺-ligand cluster. This, in turn, affects the ability of INSTIs to bind the catalytic core, compromising the inhibitory potency of the INSTI (34). When Q148H is present in combination with the compensatory mutant G140S, the double mutant poses substantial challenges for even the most broadly effective INSTIs. Moreover, Q148 is surrounded by other well-known sites of resistance (in Fig. S2, S147, G149, and I151 are within 5 Å, shown by the yellow sphere; Q146 is farther out). Amino acid substitutions at these positions enhance resistance through mechanisms that are as yet undefined.

Substitutions at position E138 have become increasingly common in complex INSTI-resistant viruses. As recently suggested in the context of SIVrcm, the E138T substitution causes H114 to donate a proton to the substituted G140S, making S140 a hydrogen bond donor for its interaction with H148, which is expected to enhance resistance (34). In our WT HIV-1 model (Fig. S3), there is a network of hydrogen bonds involving E138, H114, G140, and, by entension, the G140S/Q148H that would, presumably, behave similarly in HIV-1 IN and in SIVrcm IN. Substituting E138 for either Thr (E138T) or Lys (E138K) would facilitate propagation of a proton, which could potentiate the resistance associated with Q148H. The E138K substitution would be a stronger proton donor but may also have an accessory effect, because it is close to a backbone phosphate in the viral DNA (Fig. S3), and the charge swap could affect the interaction of the protein and the bound nucleic acid. The E138A substitution, which also arises in patients (6), is expected to have a more modest effect but would still allow H114 to serve as a proton donor. Further experimental work is required to determine the exact effects of substitutions at this position.

The substitution R263K, which is selected by the most recent FDA-approved INSTIs, is involved in several important interactions (Fig. S4 and S5). One complication in interpreting the effect of substitutions in IN, especially subtle changes such as R263K, derives from the multimeric nature of the HIV intasome, which is composed of up to 16 IN subunits. Thus, a single residue change can have different effects on the different IN subunits in the complex. Here, we have used the nomenclature of Passos et al. (27) to identify which subunit/chain we are describing. R263 in chain D interacts with the backbone phosphate of A18 and G18 from the transferred and nontransferred strands of the viral DNA, respectively. G18 base-pairs with the penultimate cytosine of the 3′ viral DNA (vDNA) end (C20), which is involved in π-stacking with the halogenated benzyl moiety of INSTIs. In addition, substituting lysine for arginine at position 263 of IN chain D could affect other nearby residues, including Q146 of IN chain A, which hydrogen bonds to the halogen atoms of the benzyl moiety of a bound INSTI. The interactions involving the 3′ end of vDNA suggest that a substitution in R263 could destabilize the stacking of the halogenated benzyl moiety of the INSTI with the penultimate nucleotide of the vDNA, which would reduce the binding of the INSTI. R263 on IN chain C (Fig. S5), which is relatively close to chains A and D IN (the distance between the two R263 residues on chains C and D is only ∼15 Å), interacts with E246 from chain D and the T17 base of the nontransferred viral DNA strand. In addition to its effects on the susceptibility of viral replication to INSTIs, the R263K mutant also reduces the replication of HIV to ∼60% in our single-round assay (35). There is also the possibility that this substitution could affect the assembly of the IN multimer. The relatively subtle nature of the R263K substitution will require high-resolution structural data, and new insights, to define the exact mechanism of resistance. Because R263 plays different role(s) in the various IN subunits, it is not yet clear which of the changes is responsible for the observed decrease in replication. It should be obvious that R263 is not the only amino acid in IN where substitutions affect the potency of INSTIs and play different roles in distinct subunits.

Based on currently available data, it appears that DTG and BIC provide a strong and safe treatment strategy for treatment-naive patients (36–41). However, for INSTI-experienced patients, who have undergone virological failure and switched regimens to a salvage therapy, these new INSTIs may not always be effective, and they may not be the best therapeutic option. As the use of INSTIs increases, it is likely that additional resistant viruses will emerge. It is already clear that complex mutants can be a problem and that improved INSTIs will be needed to overcome new complex IN mutants as they arise. This means that new INSTIs are needed that can effectively inhibit IN mutants, particularly those that have the G140S/Q148H substitutions (34). We have been able to make good progress in developing INSTIs using the structures of the PFV intasome to develop models of how the INSTIs are bound to HIV-1 IN (9, 10, 12, 13, 16). However, having the structures of existing INSTIs bound to the active sites of WT HIV-1 and SIV intasomes and some of the more problematic drug-resistant IN mutants should give us a much more detailed understanding of both drug binding and mechanisms of drug resistance (27, 34, 42). This new information should be extremely useful for developing new and more broadly effective INSTIs. Compounds that retain efficacy against the G140S/Q148H mutants could be used both to prevent new HIV-1 infections and to block the emergence of new drug-resistant HIV-1 IN mutants in INSTI-experienced patients who need salvage therapy.

MATERIALS AND METHODS

INSTI synthesis.

DTG, BIC, and CAB were acquired as previously described (16). The synthesis of compounds 4c, 4d, 4f, 6b, and 6p has been reported (12, 13).

Cell-based assays.

HIV-based viral vectors with either WT or mutant IN were used in single-round infectivity assays to determine the antiviral potencies (EC₅₀ values) of the compounds as previously described (43). Briefly, vesicular stomatitis virus-g (VSV-g)-pseudotyped HIV was produced by transfections of 293T cells with pNLNgoMIVR-ΔLUC and pHCMV-g (obtained from Jane Burns, University of California, San Diego, CA) using the calcium phosphate method. Six to seven h after the calcium phosphate precipitate was added, 293T cells were washed twice with phosphate-buffered saline (PBS) and incubated with fresh medium for 48 h. The virus-containing supernatants were then harvested, clarified by low-speed centrifugation, filtrated, and further diluted for preparation in antiviral infection assays. On the day prior to the screen, human osteosarcoma (HOS) cells were seeded in a 96-well luminescence cell culture plate at a density of 4,000 HOS cells in 100 μl per well. On the day of the screen for antiviral activity infection assays, cells were treated with compounds from a concentration range of 5 μM to 0.0001 μM using 11 serial dilutions and then incubated at 37°C for 3 h. After compound incorporation, 100 μl of virus-stock diluted to achieve a luciferase signal between 0.2 and 1.5 relative luciferase units (RLUs) was added to each well and incubated at 37°C for 48 h. Infectivity was measured by using the Steadylite plus luminescence reporter gene assay system (PerkinElmer, Waltham, MA). Luciferase activity was measured by adding 100 μl of Steadylite plus buffer (PerkinElmer) to the cells, incubating at room temperature for 20 min, and measuring luminescence using a microplate reader. Antiviral activities were normalized to the infectivity in cells that featured the absence of target compounds. KaleidaGraph (Synergy Software, Reading, PA) was used to perform nonlinear regression analysis on the data. EC₅₀ values were determined from the fit model.

A modified version of the single-round infectivity assay was used to determine the replication of the INSTI-resistant mutant vectors. Briefly, 200 ng of a WT or INSTI-resistant mutant HIV-1-based vector was added to each well in 96-well plates and incubated for 48 h, and luciferase activity was measured as described above. The luciferase activity of the WT virus was set to 100%, and the infectivity of the mutant viruses was measured as a percentage of that of the WT (23).

Vector constructs.

The vector pNLNgoMIVR-ΔENV.LUC has been described previously (13, 16). To produce the new IN mutants used in this study, the IN open reading frame was removed from pNLNgoMIVR-ΔENV.LUC by digestion with KpnI and SalI, and the resulting fragment was inserted between the KpnI and SalI sites of pBluescript KS+. Using this construct as the wild-type template, the following HIV-1 IN mutants were prepared using the QuikChange II XL site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA): M50I/R263K, S119R/R263K, T124A/S153Y, M50I/S119R/R263K, L74M/Q148R, V72I/E138K/Q148K, V75A/G140S/Q148H, T122N/G140S/Q148H, L74M/V75A/G140S/Q148H, C65S, G149A, G140S/G149A, C56S/G140S/Q148H, G140S/Q148H/G149A, C56S/G140S/Q148H/G149A, L74M/E138K/Q148R/R263K, L74M/G140S/S147G/Q148K, L74I, Q146R, E157Q, Q95K/Q146R, E157Q/R263K, L74M/E138K/G140S/Q148R, M154I, S230N, T66I/R263K, T66I/L74M/E138K/S147G/M154I, and T66I/L74M/E138K/S147G/Q148R/S230N. The following sense oligonucleotides were used with matching cognate antisense oligonucleotides (not shown) (Integrated DNA Technologies, Coralville, IA) in the mutagenesis: M50I, 5′-CAGCTAAAAGGGGAAGCCATTCATGGACAAGTAGACTGT-3′; C56S′-ATGATGGACAAGTAGAAGTAGCCAGGAATATGGCAG-3′; S119R, 5′-GTACATACAGACAATGGCCGTAATTTCACCAGTACTACA-3′; T124A, 5′-GGCAGCAATTTCACCAGTGCTACAGTTAAGGCCGCCTGT-3′; L74M, 5′-TTAGAAGGAAAAGTTATCATGGTAGCAGTTCATGTAGCC-3′; E138K, 5′-TGGTGGGCGGGGATCAAGCAGAAATTTGGCATTCCCTACAATCCCCA-3′; V75A, 5′-GAAGGAAAAGTTATCTTGGCAGCAGTTCATGTAGCCAGT-3′; T122N, 5′-GACAATGGCAGCAATTTCAACAGTACTACAGTTAAGGCC-3′; L74M for V75A/G140S/Q148H, 5′-TTAGAAGGAAAAGTTATCATGGCAGCAGTTCATGTAGCC-3′; G149A, 5′-TACAATCCCCAAAGTCAAGCAGTAATAGAATCTATGAAT-3′; R263K, 5′-ATAAAAGTAGTGCCAAGAAAAAAAGCAAAGATCATCAGG-3′; S147G for L74M/G140S/S147G/Q148K, 5′-ATTCCCTACAATCCCCAAGGTCGTGGAGTAATAGAATCT-3′; L74I, 5′-TTAGAAGGAAAAGTTATCATAGTAGCAGTTCATGTAGCC-3′; Q146R; 5′-GGCATTCCCTACAATCCCCGAAGTCAAGGAGTAATAGAA-3′; E157Q, 5′-ATAGAATCTATGAATAAACAATTAAAGAAAATTATAGGA-3′; Q95K for Q95K/Q146R, 5′-CCAGCAGAGACAGGGAAAGAAACAGCATACTTCCTC-3′; G140S for L74M/E138K/G140S/Q148R, 5′-GGGATCAAGCAGAAATTTAGCATTCCCTACAATCCCCAA-3′; M154I, 5′-AGTCAAGGAGTAATAGAATCTATAAATAAAGAATTAAAGAAAATT-3′; S230N, 5′-CGGGTTTATTACAGGGACAACAGAGATCCAGTTTGGAAA-3′; E138K for T66I/L74M/E138K/S147G/M154I, 5′-TGGTGGGCGGGGATCAAGCAGAAATTTGGCATTCCCTACAATCCCCA-3′; T66I, 5′-ATATGGCAGCTAGATTGTATACATTTAGAAGGAAAAGTT-3′; and S147G for T66I/L74M/E138K/S147G/M154I, 5′-ATTCCCTACAATCCCCAAGGTCAAGGAGTAATAGAATCTATGAAT-3′.

The IN mutants L74M/Q148R, V75A/G140S/Q148H, T122N/G140S/Q148H, and L74M/V75A/G140S/Q148H (shown in Fig. 2) were made as follows. The IN mutant L74M/Q148R was made using the previously constructed IN mutant Q148R, with the L74M oligonucleotides being used to add the second amino acid substitution. The IN mutants V75A/G140S/Q148H and T122N/G140S/Q148H were made with the previously constructed IN mutant G140S/Q148H (23) and the appropriate oligonucleotides for V75A and T122N, respectively, to add the third amino acid substitution. The IN mutant L74M/V75A/G140S/Q148H was made with aforementioned IN mutant V75A/G140S/Q148H and the correct oligonucleotides for L74M to add the fourth amino acid substitution.

The IN mutants shown in Fig. 3, which include C56S, G149A, G140S/G149A, C56S/G140S/Q148H, and C56S/G140S/Q148H/G149A were made as follows. The IN mutants C56S and G149A were constructed as described above using the appropriate listed oligonucleotides. The IN mutant G140S/G149A was made with previously generated IN mutant G140S and the right oligonucleotides for G149A to add the second substitution. The IN mutant C56S/G140S/Q148H was made using the previously made IN mutant G140S/Q148H and the appropriate oligonucleotides for C56S to add the third amino acid substitution. The IN mutant C56S/G140S/Q148H/G149A was made using the aforementioned IN mutant G140S/Q148H/G149A and the appropriate oligonucleotides for C56S to add the fourth amino acid substitution.

The IN mutant shown in Fig. 4, L74M/E138K/Q148R/R263K was made based on previously existing mutants (16). The IN mutant L74M/E138K/Q148R/R263K was made using the previously constructed IN mutant E138K/Q148R and the appropriate nucleotides for L74M and R263K, respectively, to add the third and fourth amino acid substitutions.

The IN mutants shown in Fig. 5, which include L74I, Q146R, E157Q, Q95K/Q146R, E157Q/R263K, and L74I/E138K/G140S/Q148R, were made as follows. The IN mutants L74I, Q146R, and E157Q were constructed as described above using the appropriate listed oligonucleotides. The IN mutant Q95K/Q146R was made using the previously constructed IN mutant Q146R and the correct Q95K oligonucleotides to add the second amino acid substitution. The IN mutant E157Q/R263K was made using the previously constructed IN mutant R263K and the appropriate oligonucleotides for E157Q, which were used to add the second amino acid substitution. The IN mutant L74I/E138K/G140S/Q148R was made using the previously generated IN mutant E138K/Q148R and the appropriate oligonucleotides for L74I and G140S, respectively, to add the third and fourth amino acid substitutions.

The IN mutants shown in Fig. 6, which include M154I, S230N, T66I/R263K, T66I/L74M/E138K/S147G/M154I, and T66I/L74M/E138K/S147G/Q148R/S230N, were made as follows. The IN mutants M154I and S230N were constructed as described above using the appropriate listed oligonucleotides. The IN mutant T66I/R263K was made with the previously constructed IN mutant T66I (23) and the correct oligonucleotides for R263K, which was used to add the second amino acid substitution. The IN mutant T66I/L74M/E138K/S147G/M154I was made from previously generated IN mutant T66I (23) and the appropriate oligonucleotides for L74M and M154I, respectively, to generate these amino acid substitutions, and the oligonucleotides for E138K and S147G were used to generate these amino acid substitutions. The IN mutant T66I/L74M/E138K/S147G/Q148R/S230N was made from previously constructed IN mutant E138K/Q148R (23) and with the appropriate oligonucleotides for T66I, L74M, and S230N, and the oligonucleotides for S147G were used to generate the third, fourth, fifth, and sixth amino acid substitutions, respectively.

The DNA sequence of each of the mutant constructs was verified by sequencing. The IN coding sequences of the IN mutants, which were prepared in pBluescript KS+ were then subcloned into pNLNgoMIVR-ΔEnv.LUC (between the KpnI and SalI sites) to produce mutant HIV-1 constructs, which were also checked by DNA sequencing.