Effect of HIV-1 Integrase Resistance Mutations When Introduced into SIVmac239 on Susceptibility to Integrase Strand Transfer Inhibitors

ABSTRACT

Studies on the in vitro susceptibility of SIV to integrase strand transfer inhibitors (INSTIs) have been rare. In order to determine the susceptibility of SIVmac239 to INSTIs and characterize the genetic pathways that might lead to drug resistance, we inserted various integrase (IN) mutations that had been selected with HIV under drug pressure with raltegravir (RAL), elvitegravir (EVG), and dolutegravir (DTG) into the IN gene of SIV. We evaluated the effects of these mutations on SIV susceptibility to INSTIs and on viral infectivity. Sequence alignments of SIVmac239 IN with various HIV-1 isolates showed a high degree of homology and conservation of each of the catalytic triad and the key residues involved in drug resistance. Each of the G118R, Y143R, Q148R, R263K, and G140S/Q148R mutations, when introduced into SIV, impaired infectiousness and replication fitness compared to wild-type virus. Using TZM-bl cells, we demonstrated that the Q148R and N155H mutational pathways conferred resistance to EVG (36- and 62-fold, respectively), whereas R263K also displayed moderate resistance to EVG (12-fold). In contrast, Y143R, Q148R, and N155H all yielded low levels of resistance to RAL. The combination of G140S/Q148R conferred high-level resistance to both RAL and EVG (>300- and 286-fold, respectively). DTG remained fully effective against all site-directed mutants except G118R and R263K. Thus, HIV INSTI mutations, when inserted into SIV, resulted in a similar phenotype. These findings suggest that SIV and HIV may share similar resistance pathways profiles and that SIVmac239 could be a useful nonhuman primate model for studies of HIV resistance to INSTIs.

IMPORTANCE The goal of our project was to establish whether drug resistance against integrase inhibitors in SIV are likely to be the same as those responsible for drug resistance in HIV. Our data answer this question in the affirmative and show that SIV can probably serve as a good animal model for studies of INSTIs and as an early indicator for possible emergent mutations that may cause treatment failure. An SIV-primate model remains an invaluable tool for investigating questions related to the potential role of INSTIs in HIV therapy, transmission, and pathogenesis, and the present study will facilitate each of the above.

INTRODUCTION

The drugs used in highly active antiretroviral therapy for the treatment of HIV infection can be classified into six classes: nucleoside/nucleotide reverse transcriptase (RT) inhibitors, non-nucleoside RT inhibitors, protease inhibitors, CCR5 antagonists, fusion inhibitors, and integrase strand transfer inhibitors (INSTIs) (1).

The use of animal models can help to elucidate mechanisms of pathogenesis and can also be useful for the development of vaccines and antiviral therapies. Macaques are physiologically and immunologically similar to humans (2) and can be used to study simian immunodeficiency virus (SIV) that causes a similar progressive persistent infection to that of AIDS, making the macaque SIV system a useful nonhuman primate model (3, 4).

Preliminary tissue culture selection experiments performed in our lab with elvitegravir (EVG) led to the emergence of the E92G and T97A substitutions in SIVmac239; mutations at these positions in HIV-1 have been shown to confer resistance to raltegravir (RAL) and EVG. Therefore, we undertook to determine what effect some of the other known HIV resistance mutations might possess if introduced into an SIV model. Previous studies on SIV have documented the antiviral activities of multiple antiretroviral drugs, including that of INSTIs such as L-870812, an INSTI that belongs to a chemical class distinct from RAL (5). Monotherapy of rhesus macaques infected with a simian-human immunodeficiency virus (SHIV) variant termed 89.6P with L-870812 led to the detection of a drug-resistant virus that contained a N155H mutation; this virus exhibited lower viral replication capacity and reduced pathogenicity compared to the wild type (WT) (5). The N155H mutation in integrase (IN) has also been documented in HIV in patients failing therapy with RAL (6). The susceptibility of SIVmac251 to INSTI drugs such as RAL encourages the preclinical testing of novel INSTIs in SIVmac-infected animals (7).

The INSTIs are the most recent class of antiretrovirals to be developed and include RAL (8), EVG (9) and, most recently, dolutegravir (DTG) (10). INSTIs inhibit the strand transfer step of integration that is critical in the replication cycle of retroviruses (11–14). Mutations that confer resistance to both RAL and EVG have been observed in treatment-naive individuals following treatment failure and major resistance pathways have been identified that involve substitutions at any of positions E92 (EVG), Y143 (RAL), Q148 (both drugs), and N155 (both drugs) (11, 15–20). After an initial loss of viral replicative fitness, secondary mutations at multiple positions may compensate for this deficit, while simultaneously increasing the overall levels of drug resistance. In contrast, a mutation at position R263K in integrase seems to confer low-level resistance against DTG, and it is uncertain whether this substitution, which is also associated with diminished replication competence in HIV, can be compensated (21–23). Although RAL and EVG can be compromised by extensive cross-resistance conferred by mutations within IN, DTG possesses an improved resistance profile with far less cross-resistance with other drugs (21, 24–28). Given the recent approval of DTG by the U.S. Food and Drug Association, it is important to better understand how viruses resistant to INSTIs in general and to DTG in particular affect HIV pathogenesis in vivo. The development of a predictive animal model may help to identify impending resistance mutations and to possibly inform treatment decisions.

In this study, we constructed a series of recombinant SIVmac239 viruses carrying clinically relevant resistance-associated mutations for RAL, EVG, and DTG. We investigated the effects of these substitutions on SIV infectivity, replication fitness and drug susceptibility. Our results demonstrate that G118R, Q148R, N155H, and G140S/Q148R conferred resistance in SIV to all INSTIs tested, albeit to various degrees, whereas R263K only slightly diminished susceptibility to DTG and EVG. We also show that the G118R, Y143R, Q148R, R263K, and G140S/Q148R mutations decreased SIV infectivity and viral replication fitness in a manner similar to HIV. Finally, we present homology modeling of the active site of SIV integrase, based on the prototype foamy virus (PFV) strand transfer complex (STC) (29), and show that the integrase resistance profile of SIV should logically resemble that of HIV.

MATERIALS AND METHODS

Animal and blood collection.

Whole monkey blood (obtained from Primus Bio-Resources, Inc., Vaudreuil-Dorion, Québec, Canada) was collected from uninfected rhesus macaques (Macaca mulatta) and delivered in BD Vacutainer heparin tubes (Becton Dickinson). Peripheral blood mononuclear cells (PBMCs) were isolated from donor monkeys by Ficoll-Hypaque (GE Healthcare) gradient centrifugation from rhesus blood. Stimulation of PBMCs was facilitated with RPMI 1640 media (Gibco/Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Invitrogen), 2 mM l-glutamine, 50 U of penicillin/ml, 50 μg of streptomycin/ml, 5% (vol/vol) interleukin-2 (IL-2; Roche), and 10 μg of phytohemagglutinin (Invitrogen)/ml and maintained at 37°C under 5% CO₂. Prior to infection, rhesus PBMC cultures were stimulated for 48 to 72 h.

SIVmac239 plasmids.

A full-length infectious proviral DNA of SIVmac239 SpX was obtained from the National Institutes of Health (NIH) AIDS Research and Reference Reagent Program (catalogue number 12249).

Cells and antiviral compounds.

TZM-bl cells were obtained through the NIH AIDS Research and Reference Reagent Program. The 293T cell line was obtained from the American Type Culture Collection (CRL-11268). TZM-bl and 293T cells were used as described previously (21). Cells were subcultured every 3 to 4 days in Dulbecco minimal essential medium (DMEM) supplemented with 10% FBS, 2 mM l-glutamine, 50 U of penicillin/ml, and 50 μg of streptomycin/ml and maintained at 37°C under 5% CO₂.

RAL, EVG, and DTG were obtained from Merck, Gilead Sciences, and GlaxoSmithKline/ViiV Healthcare, respectively. Lamivudine (3TC) was obtained from the NIH AIDS Research and Reference Reagent Program. Prior to use, the compounds were solubilized in dimethyl sulfoxide and stored at −20°C. Compound stocks were appropriately diluted into DMEM immediately prior to experimentation.

Construction of site-directed IN mutants.

Single or double RAL, EVG, or DTG resistance-associated mutations were generated by using the QuikChange II XL site-directed mutagenesis kit (Stratagene). The following primers were used: H51Y primers (sense, 5′-TCATCAGAAAGGAGAGGCTATATATGGGCAGGCAA-3′; antisense, 5′-TTGCCTGCCCATATATAGCCTCTCCTTTCTGATGA-3′), G118R primers (sense, 5′-TACACATCTACACACAGATAATCGTGCTAACTTTGCTTCG-3′; antisense, 5′-CGAAGCAAAGTTAGCACGATTATCTGTGTGTAGATGTGTA-3′), G140S primers (sense, 5′-CAGGGATAGAGCACACCTTTAGCGTACCATACAATCCACAG-3′; antisense, 5′-CTGTGGATTGTATGGTACGCTAAAGGTGTGCTCTATCCCTG-3′), Y143R primers (sense, 5′-CACACCTTTGGGGTACCACGCAATCCACAGAGTCAGGG-3′; antisense, 5′-CCCTGACTCTGTGGATTGCGTGGTACCCCAAAGGTGTG-3′), Q148R primers (sense, 5′-ATACAATCCACAGAGTCGGGGAGTAGTGGAAGCAA-3′; antisense, 5′-TTGCTTCCACTACTCCCCGACTCTGTGGATTGTAT-3′), N155H primers (sense, 5′-GGAGTAGTGGAAGCAATGCATCACCACCTGAAAAATC-3′; antisense, 5′-GATTTTTCAGGTGGTGATGCATTGCTTCCACTACTCC-3′), and R263K primers (sense, 5′-CAGACATTAAGGTAGTACCCAGAAAAAAGGCTAAAATTATCAAAGATTA-3′; antisense, 5′-TAATCTTTGATAATTTTAGCCTTTTTTCTGGGTACTACCTTAATGTCTG-3′). After mutagenesis, the plasmids were digested using DpnI for 4 h at 37°C and transformed using MAX Efficiency Stbl2 competent cells [F⁻ mcrA Δ(mcrBC-hsdRMS-mrr) recA1 endA1 lon gyrA96 thi supE44 relA1 λ⁻ Δ(lac-proAB) (Invitrogen)]. The plasmids were purified by using a Plasmid Maxi kit (Qiagen) and quantified with a NanoDrop spectrophotometer. The presence of the mutations was confirmed by sequencing.

Generation of replication-competent genetically homogenous SIVmac239.

Genetically homogeneous viruses were produced by transfecting wild-type (WT) and mutated SIVmac239 proviral DNA into 293T cells. A plasmid (12.5 μg) coding for WT or mutant SIVmac239 was transfected into 293T cells using Lipofectamine 2000 (Invitrogen) as recommended by the manufacturer. Fresh medium was added at 4 h posttransfection. After 48 h, the culture supernatants were harvested, centrifuged, passed through a 0.45-μm-pore-size filter to remove cellular debris, treated with Benzonase (EMD) to degrade the transfected plasmids, and stored at −80°C for future use. Virion-associated reverse transcriptase (RT) activity was measured as previously described (30).

SIV infectivity.

SIV infectivity was evaluated using a noncompetitive short-term infectivity assay in TZM-bl cells as previously described (21, 28). Briefly, 30,000 cells per well were seeded in triplicate into 96-well culture plates (Corning). The cells were infected with the indicated amounts of virus normalized on the basis of RT activity. The luciferase activity was measured at 48 h postinfection using the luciferase assay system (Promega) and a MicroBeta2 luminometer (Perkin-Elmer). Whenever relevant, analyses of statistical significance of differences between data sets for two or more viruses were determined using a one-sample two-tailed t test. Probability values equal to or below 0.05 (P ≤ 0.05) were used to indicate statistically significant differences between different results.

SIV replication capacity.

Long-term infection was used to quantify SIV replication capacity in rhesus macaque PBMCs by measuring levels of RT activity (counts per minute [cpm]) over time. Two million cells per well were added into a 24-well culture plate in 2 ml of RPMI 1640 medium (Invitrogen) supplemented with 10% FBS, 2 mM l-glutamine, 50 U of penicillin/ml, 50 μg of streptomycin/ml, and 10 μg of IL-2/ml. The cells were infected with the indicated amounts of virus normalized on the basis of RT activity (cpm). Culture supernatants were collected at different times and analyzed for RT activity. Culture wells were refreshed with volumes of media equivalent to the volumes of supernatants that were removed.

Analysis of phenotypic drug susceptibility of SIV in TZM-bl cells.

Phenotypic susceptibilities of SIV WT and mutant IN viruses to the INSTIs RAL, EVG, and DTG and to the RT inhibitor 3TC were determined in TZM-bl cells at 48 h postinfection as described previously (21). Briefly, 30,000 cells per well were infected with WT and mutant viruses in the presence of serial dilutions of RAL, EVG, or DTG in 96-well culture plates. The amount of virus added to each well was normalized on the basis of RT activity. The fold changes in the 50% inhibitory concentrations (IC₅₀s) and standard errors of the mean (SEM) were calculated on the basis of at least two sets of experiments, each performed in triplicate, using GraphPad Prism 4.0 software (GraphPad Software, Inc., San Diego, CA). Drug susceptibility was expressed as the fold change in IC₅₀ ± the SEM.

Sequence alignment.

Multiple sequence alignments were conducted using ClustalW2 software (31, 32). Integrase amino acid sequences were studied from three HIV-1 isolates (subtypes B and C and a circulating recombinant form, A/G [CRF-02_A/G]), all belonging to group M, as well as viruses of HIV-1 group O, HIV-2, SIVmac251, SIVmac239, and PFV. The GenBank accession numbers or Protein Data Bank identifications (PDB ID) for the amino acid sequences are as follows: HIV-1 subtype B (AAK08484.2), HIV-1 subtype C (AB023804), HIV-1 subtype CRF-02_A/G (AAM74144), HIV-1 integrase group O (AAL98930), HIV-2 (DQ307022), SIVmac251 (PDB ID 1C6V), SIVmac239 (M33262), and PFV (PDB ID 3OY9).

In silico studies of SIV integrase.

A three-dimensional homology model of the full-length SIV integrase strand transfer complex (STC) was created based on the published structure of the STC (33) of the prototype foamy virus (PFV). The amino acid sequence of the SIV integrase was submitted to the three-dimensional protein structure and function prediction server termed I-TASSER (34). For modeling, the PFV crystal structure (PDB ID 3OY9) (29) was used as a lead template. Rotamer orientations for key residues in the present study were carefully examined, and the best backbone-dependent rotamers were selected (35, 36). The YASARA energy minimization server (37) was used to remove any inconsistencies and structural errors. The HIV-1 integrase structure (PDB ID 1EX4) was also used for comparison with SIV integrase; interaction hints were obtained by overlaying the SIV homology model with the crystal structure of the PFV integrase cocrystallized with Mn²⁺ and viral LTR DNA (PDB ID 3OY9). A molecular visualization program (PyMOL [http://pymol.org/]; The PyMOL Molecular Graphics System, version 1.3 [Schrödinger, LLC]) was used for structural visualization and image processing.

RESULTS

Sequence alignment of integrases from different retroviruses.

The sequence alignments of SIVmac239, HIV-2, and HIV-1 group M (subtypes B and C and CRF-02_A/G) and group O INs were based on amino acid identity, and all INs display high sequence similarity (Fig. 1A). The IN of SIVmac239 shares 86% sequence homology with that of HIV-2, while SIVmac239 shares ∼57% sequence homology with other HIV-1 INs. The homologies between SIVmac239, SIVmac251, and PFV INs were also evaluated (Fig. 1B). SIVmac239 shares 98 and 10% sequence homology with SIVmac251 and PFV INs, respectively. Since SIVmac239 IN also shares significant sequence similarity (>50%) with other INs in the alignment (except PFV), this indicates that these INs are orthologous, inherited from a shared ancestor.

FIG 1.

Multiple sequence alignments of the integrase gene in different viruses. (A) Alignment of HIV-1 group M (subtypes B and C and the CRF-02_A/G recombinant form), HIV-1 group O, HIV-2, and SIVmac239 INs was based on amino acid identity. Every stretch of 20 amino acids is indicated with a black tick and the corresponding amino acid number. (B) Sequence alignment of the integrase catalytic core and C-terminal domains of SIVmac251, SIVmac239, and PFV. The mutations that were investigated in the present study are shown above the alignments. Fully conserved residues conserved across both sequence alignments are marked with an asterisk (*); residues similar to each other are marked with a colon (:); those that are still less similar are marked with a period (.). The catalytic triad “D₆₄D₁₁₆E₁₅₂” residues are fully conserved and are marked in bold red type. Mutated residues that were examined in the present study are also fully conserved and are marked in bold blue type. The multiple sequence alignment was performed using ClustalW2 software.

As with HIV-1 IN, SIVmac239 IN consists of three functional domains: an N-terminal domain (NTD), a catalytic core domain (CCD), and a C-terminal domain (CTD), all of which are important for integration (38). The NTD, CCD, and CTD comprise residues 1 to 50, 51 to 212, and 213 to 293, respectively. The CCD shows higher conservation than the NTD and CTD among INs, while the CTD possesses the lowest degree of homology (Fig. 1) (39). The low degree of identity in the CTD may reflect differences in the specificity and/or stringency in its residues that are needed to bind DNA (40).

The catalytic triad of Asp64, Asp116, and Glu152 that is found in the CCD, also known as the D₆₄D₁₁₆E₁₅₂ motif, is conserved among all INs (Fig. 1). Amino acid positions G118, Y143, N155, and R263K are conserved among the different retroviral INs in the two alignments (Fig. 1). Positions H51, G140, and Q148 were conserved in the SIV and HIV IN sequences but not in PFV (Fig. 1B). Instead, the amino acids corresponding to H51, G140, and Q148 are P51, S140, and S148 in PFV. With the exception of R263 that is found in the CTD, all of the amino acid positions (H51, G118, G140, Y143, Q148, and N155) that have been examined here are within the CCD. Although the CTD possesses a relatively low degree of homology, position R263 is invariant and is located within a conserved region (residues 256 to 272). The amino acids found in SIVmac239 examined in the present study correspond to those found in HIV-1, HIV-2, and SIVmac251 INs.

Impact of SIVmac239 IN substitutions on infectivity.

Infectivity was assessed by performing single cycle infection of TZM-bl cells with increasing quantities of WT or IN mutant viruses. The measured luciferase activity from TZM-bl cells was directly proportional to the amount of virus used for infection, measured as the cpm of RT activity (Fig. 2A). Cells were harvested, and the mutant luciferase activities were normalized and expressed relative to the signal detected for WT, arbitrarily set at 100% (Fig. 2B). The H51Y and N155H IN mutant viruses did not impair viral infectivity and were able to mediate infection at WT levels (P = 0.52 and 0.64, respectively). In contrast, the IN mutants G118R, Y143R, Q184R, R263K, and G140S/Q148R all significantly reduced SIV infectivity to levels of ∼20% for Y143R, ∼16% for G140S/Q148R, and ∼10% for each of G118R, Q148R, and R263K (P < 0.0001) (Fig. 2B).

FIG 2.

Effects of integrase mutations on SIV infectivity. (A) The infectiousness of SIVmac239_IN(WT), SIVmac239_IN(H51Y), SIVmac239_IN(G118R), SIVmac239_IN(Y143R), SIVmac239_IN(Q148R), SIVmac239_IN(N155H), SIVmac239_IN(R263K), and SIVmac239_{IN(G140S/Q148R)} was assessed by quantifying the luciferase activity in relative luminescent units (RLU) at 48 h after infection of TZM-bl cells with increasing concentrations of virus (as determined by the cpm of the RT activity). (B) Infectivity of WT and mutant viruses represented by the means ± SEM for each of two independent TZM-bl infectivity assays (statistical significance was calculated for individual pairs of data using a one-sample two-tailed test with a statistical cutoff of P ≤ 0.05) normalized against the WT signal that was arbitrarily set at 100%.

Effect of integrase mutations on SIV replicative fitness in rhesus PBMCs.

To determine whether the impairment in single cycle infectiousness of some of the viruses in the TZM-bl system (Fig. 2B) also applied to replication capacity in rhesus PBMCs, we assessed viruses that were either WT or contained the following indicated mutations: SIVmac239_IN(WT), SIVmac239_IN(H51Y), SIVmac239_IN(G118R), SIVmac239_IN(Y143R), SIVmac239_IN(Q148R), SIVmac239_IN(N155H), SIVmac239_IN(R263K), and SIVmac239_{IN(G140S/Q148R)}. Long-term infection studies confirmed the results obtained for single cycle infection in TZM-bl cells. Although the H51Y and N155H substitutions did not negatively impact SIV replication capacity (Fig. 3), each of the G118R, Y143R, Q148R, R263K, and G140S/Q148R mutations decreased viral replication over 20 days of infection (Fig. 3). In agreement with the infectivity results, these experiments showed that all of the mutated viruses, except H51Y and N155H, were impaired in viral replication capacity relative to WT virus.

FIG 3.

Impact of integrase mutations on SIV replicative capacity in rhesus PBMCs. The RT activity was measured as cpm in the culture fluids of rhesus PBMCs infected with SIV viruses. Error bars indicate ± the SEM.

Susceptibility of IN mutant viruses to antiretroviral drugs.

We first studied 3TC as a control and showed that all of the IN-mutated SIV variants that were tested retained susceptibility to this drug (Table 1). In contrast, the mutant SIVmac239 viruses containing G118R or R263K displayed resistance to DTG (11- and 9.3-fold, respectively) (Table 2). The levels of resistance observed with Y143R, Q148R, N155H, and G140S/Q148R against DTG (i.e., 1.4-, 0.94-, 2.0-, and 1.52-fold, respectively) were not significant. H51Y did not confer resistance to DTG in comparison with WT and DTG was potent against all viruses examined in our panel except for those containing G118R and R263K. In the case of RAL and EVG, the H51Y mutant virus was sensitive to both of these drugs, whereas G118R conferred a moderate level resistance to both RAL and EVG (3.1- and 10-fold, respectively) (Table 2). Each of the Y143R, Q148R, and N155H mutations conferred low-level resistance to RAL (2.6-, 3-, and 2.4-fold, respectively), while each of Y143R (2.7-fold), Q148R (∼36-fold), and N155H (∼62-fold) diminished susceptibility to EVG (Table 2). In the case of the G140S/Q148R mutated viruses, high-level resistance to RAL and EVG was observed (>300 and 286-fold, respectively). The fold change in susceptibility to RAL observed with R263K (0.89-fold) was not significant, although R263K did confer moderate level resistance to EVG (∼12-fold).

TABLE 1.

Effects of SIVmac239 IN mutations on IC₅₀s for lamivudine^a

WT or mutation	IC50 (nM)		Fold change
	Mean	95% CI
WT	616.6	405.7–937.2	1
H51Y	636.9	452.4–897.0	1.03
G118R	591.2	486.3–901.1	0.96
Y143R	505.8	343.1–745.8	0.82
Q148R	476.1	288.0–787.1	0.77
N155H	670	496.2–904.7	1.09
R263K	518.1	302.2–888.1	0.84

^aControl studies confirmed that the M184V mutation in reverse transcriptase conferred >500-fold resistance to lamivudine (3TC; data not shown). CI, confidence interval.

TABLE 2.

Fold changes in IC₅₀s against DTG, RAL, and EVG for viruses mutated in IN as studied in TZM-bl cells^a

Mutation	Mean fold change in IC50 ± SD from WT virus
	Dolutegravir	Raltegravir	Elvitegravir
WT	1.0	1.0	1.0
H51Y	1.40 ± 0.33	0.71 ± 0.27	1.40 ± 0.22
G118R	11.0 ± 6.70	3.10 ± 2.10	10.0 ± 5.20
Y143R	1.40 ± 0.56	2.60 ± 0.12	2.70 ± 1.60
Q148R	0.94 ± 0.31	3.00 ± 1.90	36.0 ± 24.0
N155H	2.00 ± 0.70	2.40 ± 0.55	62.0 ± 11.5
R263K	9.30 ± 1.60	0.89 ± 0.11	12.0 ± 2.90
G140S/Q148R	1.52 ± 0.35	>300	286 ± 91

^aThe average IC₅₀s for WT viruses in regard to each of DTG, RAL, and EVG were 1.2 nM, 8.9 nM, and 0.047 nM, respectively. Significant fold changes are indicated in boldface.

In silico studies of WT SIV integrase.

Although the structure of the full-length SIVmac239 IN structure remains unsolved, the crystal structure of the full-length PFV IN, complexed with DNA and cofactor cations, is available (29), and the PFV structure has been used as a template to generate homology models of HIV-1 integrase (41–43). We used the available PFV IN structure with PDB ID 3OY9 as a lead template, together with the I-TASSER server (34), to generate a SIV IN homology model (Fig. 4). The modeled structure shows good global agreement with PFV and HIV-1 (Fig. 4A). Rotamer interrogation showed that the catalytic triad in SIV IN possessed the proper orientation to coordinate Mn²⁺ in a manner similar to that for PFV and HIV-1 (Mn²⁺ and LTR DNA coordinates were derived from the PFV structure). A closeup of the active site shows insignificant differences in the relative positions of the D₆₄D₁₁₆D₁₅₂, catalytic triad between the model and both the PFV and HIV structures (Fig. 4B).

FIG 4.

Homology modeling of SIV integrase. A homology model of SIV integrase was created based on the strand transfer complex of PFV together with viral LTR DNA (PDB ID 3OY9); the location of various integrase residues and their distances to the integrase active site were examined. (A) Detailed view of the overlay of the catalytic core domains of the integrases of the SIV model, HIV-1 (PDB ID 1EX4) and PFV. (B) Closeup overlay showing the relative positions of the D₆₄D₁₁₆E₁₅₂ catalytic core in the WT enzyme. (C) The amino acid residues of SIV IN (labeled and shown as green sticks) that were investigated in the present study (except for residue E92); catalytic triads are shown as yellow sticks. All residues shown with the exception of H51and R263 are within 10 Å of the active site. All images were processed using PyMOL software. Secondary structures are represented as diagram structures, with key residues labeled and shown as stick structures with standard atomic coloration. Diagram and carbon atom coloration differentiates the SIV integrase homology model (yellow), the PFV STC (green) and HIV-1 (blue). Mn²⁺ cations are shown as red spheres. Mn²⁺ coordinates were inserted from the 3OY9 PDB structure into the active site of the SIV and HIV-1 integrases. The viral DNA substrate sugar-phosphate backbone and nitrogenous bases are shown in brown and blue, respectively. Acidic (red) and basic (blue) moieties are identified by standard atomic coloration.

Further analysis that examined the distances of mutated IN residues to the catalytic triad revealed that the H51 and R263 residues lie outside the active site in SIV, while D64 and E152 surround N155 (Fig. 4C). All three acidic residues of the catalytic triad and the divalent cations are close to Q148. Although residues G140 and Y143 lie slightly outside the active site, G140 and Y143 are present on the same catalytic flexible loop as Q148 and are close to the latter residue. G118 is located one residue away from and therefore proximal to the catalytic residue D116. R263 in the SIV integrase model was within 4 Å of the phosphodiester backbone of the LTR DNA in the STC (Fig. 4C). Overall, this model highlights key functional components of the catalytic D₆₄D₁₁₆D₁₅₂ motif within IN and suggests how the R263K mutation, among others, may be unique, while conferring low-level resistance to DTG.

DISCUSSION

This is the first study of SIVmac239 in which mutations within IN were purposely introduced in order to determine susceptibility to INSTIs as well as infectivity. This study is important because it is crucial to validate SIVmac239 as a model to measure the effects of such mutations on resistance to INSTIs as well as viral pathogenesis. It is also important to verify whether the same mutations that are associated with drug resistance in HIV can also play this role in regard to SIV.

We show here that (i) SIVmac239 is susceptible to various INSTIs (RAL, EVG, and DTG) with IC_50s in a nanomolar range similar to those reported by others (44); (ii) the viral replicative fitness of SIV, as measured by relative infectiousness, was reduced by the introduction of certain mutations into the IN of SIV; (iii) the results of long-term infectivity studies in rhesus PBMCs resemble those observed with HIV-1; (iv) the IN mutant viruses studied displayed resistance profiles similar to those of HIV-1 in regard to each of RAL, EVG, and DTG; and (v) homology modeling revealed that key residues associated with IN-drug resistance are in similar proximal locations relative to the IN active site in both SIV and HIV-1.

In HIV-1, three main resistance pathways have been identified for RAL that involve mutations at positions Y143, Q148, and N155 within IN (18). Only N155 and Q148 are reported to confer cross-resistance to EVG (19, 45–47), while Y143 is reported to be specific for RAL (48). Primary substitutions at positions G118 and R263 and a secondary substitution at position H51 in integrase were previously shown in cell culture selections to confer resistance to INSTIs, including RAL, MK-2048, and DTG (21–23, 49). Our goal in the present study was to evaluate these six mutations in SIV, in view of the fact that no group has yet documented the selection of INSTI resistance mutations in SIV. We observed that each of G118R, Y143R, Q148R, R263K, and G140S/Q148R mutations in IN impaired viral infectivity, while H51Y and N155H were relatively innocuous in this regard. These infectivity data correlate well with previous reports that Y143R, Q148R, and G140S/Q148R in HIV-1 significantly decreased viral fitness in the HeLa-P4 reporter cell system at 48 h postinfection, while N155H exhibited similar viral fitness to WT viruses (50, 51). The relative infectivities of H51Y and R263K SIV reported here compared to WT are also in agreement with results reported for HIV-1 that documented an effect of R263K on infectiousness of ≈20% (21, 27, 28), although the effect in SIVmac239 was more pronounced, i.e., ∼90%. The G118R mutation in HIV-1 also resulted in a dramatic reduction in viral infectivity in SIV compared to WT (22). The infectivity data presented are in agreement with what was observed regarding the replication capacity of SIV in rhesus PBMCs (Fig. 3). In HIV-1, the relative viral fitness of IN mutant viruses in competition with WT showed that Q148R resulted in a reduction in viral fitness to ∼59% of WT levels, while N155H caused only a 24% reduction in relative fitness (52–54). The addition of the G140S mutation to Q148R diminished replication capacity to levels similar to those seen for Q148R-mutated HIV-1 (59% of WT levels) (54). In contrast, the addition of G140S to Q148H restored replication capacity to WT levels (99%) (54). Similar to the infectivity observations, H51Y and R263K modestly diminished HIV replication (27, 28). It has also been reported that G118R-containing HIV-1 were significantly impaired in replication capacity (1% of the WT virus) (22). The introduction of a Y143R mutation in IN also reduced the replication capacity of HIV-1 (50, 55).

Substitutions that are associated with resistance against RAL and EVG in HIV had no effect or minimal effect on SIVmac239 susceptibility to DTG, which retained full potency against all IN resistance mutants (including G140S/Q148R), except for G118R and R263K. These results are also in agreement with previous findings for HIV-1 that showed that DTG was active against most RAL- and EVG-resistant viruses that contained only one or two INSTI mutations (25, 52), as well as with results that both R263K and G118R can be selected in tissue culture with DTG (21, 28). The G140S/Q148R mutations in tandem conferred moderate levels of resistance, i.e., ∼8.4-fold, to DTG in HIV-1 (25) but only low-level resistance in SIV, ∼1.52-fold. In contrast, Q148R, N155H, and G140S/Q148R conferred high-level resistance against EVG (∼36-, ∼62-, and ∼286-fold, respectively), whereas R263K displayed moderate level resistance against EVG (12-fold). These results are expected since the latter substitutions result in high-level resistance against EVG in HIV-1 (25, 28, 52). Y143R viruses demonstrated low levels of resistance to EVG (2.7-fold), which was expected, since Y143R does not display significant resistance to EVG in HIV-1 (25, 52). In contrast, Y143R, Q148R, and N155H mutated SIVmac239 showed low levels of resistance to RAL (2.6-, 3-, and 2.4-fold, respectively). These results were unexpected since Y143R and Q148R display high-level resistance against RAL in HIV-1 (25, 52). These differences may be due to slight structural disparities between the active sites of SIV and HIV. The result that G140S/Q148R confers high-level resistance to RAL (>300-fold) is in agreement with what has been observed for HIV-1 (200-fold) (25). The R263K data that indicate susceptibility to RAL are also in accord with results obtained in HIV-1 (28). RAL has been reported to be potent against SIVmac251 and to inhibit replication in human T-cell lines in a manner similar to HIV-1 (7). RAL also decreased viral load in SIVmac251 rhesus macaques and stably suppressed viral loads in the presence of the RT inhibitors tenofovir disoproxil fumarate and emtricitabine (7). The H51Y substitution did not affect susceptibility to DTG, RAL, or EVG, compared to WT, which is also consistent with results obtained with HIV (28).

Amino acid sequence alignment of the IN regions of various HIV-1 isolates, HIV-2, and SIV shows high sequence conservation. In regard to individual domains, the identity/similarity percentages for the NTD are 55% when HIV-1 is compared to SIV, whereas they are 61 and 53% for the CCD and CTD, respectively. SIVmac239 and SIVmac251 possess nearly identical sequences, with ∼98% sequence homology, as expected since SIVmac239 is derived from SIVmac251 by animal passage and tissue culture proviral DNA cloning (56). It is clear that the CTD is the least conserved domain and that the CCD is the most conserved. Thus, the similarity of the SIV resistance profile to INSTIs to that of HIV-1 is probably due to the high degree of conservation of the CCD (57). All of the residues studied here are located within the CCD with the exception of R263 that is found within the CTD. Although both SIVmac239 and SIVmac251 share only ∼10 to 12% sequence identity/similarity with PFV, the conservation of the catalytic triad “D₆₄D₁₁₆E₁₅₂” and residues, G118, Y143, and R263 validated the use of PFV as a lead template for generating the homology model. Despite a difference in amino acid position Q148 in PFV, the latter retained susceptibility to RAL, further supporting the use of PFV in our model (58). Sequence alignment showed that HIV-2 and SIV also share a high degree of sequence homology (∼88%), which is not surprising, since it is believed that HIV-2 and SIV share a common ancestor (59–66). The high degree of sequence similarity between HIV-2 and SIV helps to explain the similar phenotypes of Y143R/C in regard to susceptibility to RAL in both viruses. In contrast, the introduction of the Y143C mutation into HIV-2 IN only led to moderate resistance to RAL in vitro, i.e., 3- to 4-fold (67–69). It is possible that Y143R may not be a primary RAL resistance mutation in SIV IN. Others have shown that Y143C is usually accompanied by E92Q in order to confer resistance to RAL (67).

Homology modeling based on the PFV crystal structure has provided some insight into the residues that form the IN active site, and one group solved a crystal structure of SIVmac IN strain 251 with a resolution of 3 Å (PDB ID 1C6V) after introduction of a solubility mutation (39). This crystal structure, however, lacked both the NTD and bound viral LTR DNA and was derived from SIV 251. It also contained amino acid substitutions in all three IN domains; therefore, this structure was not used in our analysis. Retroviral INs possess high conservation of key secondary structural elements as well as the three functional domains CTD, CCD, and NTD. The PFV crystal structure that contained viral DNA helped to reveal the architecture of the intasome and to clarify mechanisms of resistance to INSTIs in the presence of IN mutations (29, 42, 70). Caveats of the PFV structure are that the interdomain linkers are longer than those found in HIV and SIV and that some of the residues in SIV and HIV are not identical to those in PFV, e.g., Q148 in SIV/HIV is S148 in PFV.

Based on homology modeling, the SIVmac239 model proposes that residue R263 in the CTD is implicated in viral DNA binding (Fig. 4C), which is in agreement with studies that showed that R263K in HIV-1 IN impaired IN-DNA binding (21). In addition, the resistance profiles described here are attributable to the close proximity of residues (within 10 Å) to the active site (where INSTIs bind); hence, mutations at any of these positions may perturb the active site and its ability to bind to INSTIs (Fig. 4C).

We are currently in the process of studying these mutations among others, individually or in combination in cell-free assays, to determine whether effects similar to those seen here occur in SIV and are still performing tissue culture selection experiments with SIVmac239 and various INSTIs. Recently, a long-acting form of a new INSTI termed S/GSK-1265744, a DTG analogue, was shown to protect macaques against repeated vaginal and rectal exposures of SHIV_SF162p3 (71, 72). In another promising study, a vaginal topical gel containing the INSTI L-870812 was able to protect two of three macaques when applied 30 min before SHIV_SF162p3 challenge (44). Our study is relevant in that it is the first to show the effect of INSTI resistance mutations in SIV. An SIV-primate model remains an invaluable tool for investigating questions related to the potential role of INSTIs in HIV therapy, transmission, and pathogenesis, and our work will hopefully facilitate this research.