Abstract

HIV protease is essential for processing the Gag polyprotein to produce infectious virions and is a major target in antiretroviral therapy. We have identified an unusual HIV-1 subtype C variant that contains insertions of leucine and asparagine (L38↑N↑L) in the hinge region of protease at position 38. This was isolated from a protease inhibitor naïve infant. Isothermal titration calorimetry showed that 10% less of L38↑N↑L protease was in the active conformation as compared with a reference strain. L38↑N↑L protease displayed a ±50% reduction in KM and kcat. The catalytic efficiency (kcat/KM) of L38↑N↑L protease was not significantly different from that of wild type although there was a 42% reduction in specific activity for the variant. An in vitro phenotypic assay showed the L38↑N↑L protease to be susceptible to lopinavir (LPV), atazanavir (ATV) and darunavir in the context of an unrelated Gag. However, in the presence of the related Gag, L38↑N↑L showed reduced susceptibility to darunavir while remaining susceptible to LPV and ATV. Furthermore, a reduction in viral replication capacity (RC) was observed in combination with the related Gag. The reduced susceptibility to darunavir and decrease in RC may be due to PTAPP duplication in the related Gag. The present study shows the importance of considering the Gag region when looking at drug susceptibility of HIV-1 protease variants.

Introduction

HIV remains a global health problem with 36.7 million people infected with it globally [1]. HIV is particularly a problem in Africa as 52% of people living with HIV are from southern Africa and 45% of all new infections occur in this region [1]. It was estimated that 13% of the South African population was infected with HIV in 2016. However, there have been great improvements in treatment in Africa, particularly South Africa. In 2015, more than any other country, South Africa had more people (3.4 million) on treatment. There are, however, still populations within South Africa that are at a high risk of HIV infection; these include sex workers, people who inject drugs, transgender people, prisoners, gay men and men who have sex with men. In 2015, these populations accounted for 20% of all new infections in sub-Saharan Africa [1].

HIV is of two types: HIV-1 (the main type) and HIV-2. HIV-1 is divided into groups M, N, O and P. Group M is the main group and is further divided into subtypes A, B, C, D, F, G, H, J and K [2]. Subtype C is found in sub-Saharan Africa, India, Brazil and China [2] and accounts for ∼50% of global infections [3]. The great diversity among HIV is attributed to the high replication rate as well as the low fidelity of reverse transcriptase [4]. Subtype C is of particular interest to this study as the variant in this study is of subtype C origin.

HIV protease, a homodimeric aspartyl protease, cleaves the Gag and Gag-Pol polyproteins to produce three enzymes (reverse transcriptase, integrase and protease) and the structural proteins (capsid, matrix, nucleocapsid, p6, gp120 and gp41) needed for capsid assembly [5]. The secondary structure of protease consists mainly of β-sheets and one α-helix per monomer. The active site contains the characteristic Asp-Gly-Thr sequence of an aspartyl protease [6]. The hydrophobic active site is covered by two β-turns, termed the flap region, which open up to allow substrate to bind to the active site and then close upon substrate binding. The movement of the flaps is aided by the hinge region (residues 32–42) of the protein, a region known to contain several polymorphisms in subtype C.

HIV-1 protease plays a critical role in viral replication since failure to cleave the Gag and Gag-Pol polyproteins results in immature virions that are non-infectious [7]. This vital step is one of the drug targets of second-line antiretroviral therapy (ART) that makes use of protease inhibitors (PIs) such as lopinavir (LPV), atazanavir (ATV) or darunavir (DRV), which can be ritonavir-boosted, in addition to two nucleot(s)ide inhibitors [8]. This combination therapy has proved to be highly effective as it extended the life expectancy of patients and reduced the spread of AIDS in regions where treatment was available. HIV is notorious for the development of drug resistance to the antiretrovirals (ARVs). Resistance to PIs is no exception. Mutations occur rapidly due to reverse transcriptase lacking proof-reading capabilities. Resistance to PIs is a result of amino acid substitutions in the substrate-binding sites as well as distal sites [9]. Mutations in the protease gene are divided into major and minor mutations. Major mutations occur first in response to drug pressure and these occur mainly in the active site, which affects drug binding. Mutations in the active site affect the efficiency of the enzyme and as a result minor mutations occur distal to the active site to restore efficiency of the enzyme [10].

Rarely, amino acid insertions in the protease gene are selected for during ART. Insertion mutations are not unusual in reverse transcriptase [11] but are rare in protease (∼0.1%) [12]. Most insertions arise due to duplications of neighbouring DNA sequences due to primer/template slippage during reverse transcription. Resistance mutations in protease have been well characterised [13]. However, little is known about the effect of amino acid insertions in the hinge region on drug binding. The present study shows the drug susceptibility to LPV, ATV and DRV of the L38↑N↑L protease as well as with its accompanying Gag sequence when compared with the wild-type subtype C protease and Gag.

Methods

Expression and purification

The L38↑N↑L protease sequence data were obtained from Prof. Lynn Morris (AIDS Virus research Unit, NICD of Johannesburg, South Africa) [14]. The pET-11b plasmid encoding the HIV-1 wild-type subtype C protease was previously generated in our laboratory [15]. The pET-11a plasmid containing the L38↑N↑L gene was purchased from GenScript (Hong Kong). The plasmids containing the wild type and L38↑N↑L sequence were used to transform Escherichia coli BL21 (DE3) pLysS and E. coli Rosetta cells, respectively. The proteases were overexpressed by the addition of 1 mM IPTG at mid-log phase (A600nm = 0.6). Overexpression was allowed to continue for 4 h for wild type and 6 h for L38↑N↑L at 37°C. The cells were pelleted by centrifugation at 5000×g and resuspended in buffer A (10 mM Tris, 5 mM EDTA, pH 8). The bacterial cell membranes were disrupted by sonication, 12 V for 10 × 30 s intervals. The cell lysates were centrifuged at 23 000×g for 40 min to separate the soluble and insoluble fractions. The insoluble fractions were resuspended in buffer A+ (10 mM Tris, 2 mM EDTA and 2% Triton X-100 at pH 8). This was centrifuged at 23 000×g for 30 min and the process was repeated. The proteases were recovered from the inclusion bodies by resuspending the pellet in buffer B (8 M urea, 10 mM Tris and 2 mM DTT at pH 8) and incubation at 20°C for an hour. The unfolded proteases were centrifuged at 23 000×g and refolded by dialysis against 10 mM formic acid containing 10% glycerol (v/v) for 4 h at 4°C. The proteases were then dialysed against buffer D (10 mM sodium acetate and 2 mM DTT at pH 5) overnight and purified using a CM-Sepharose column. Elution was performed using a salt gradient of 0–1 M NaCl. Finally, the protease was dialysed against 10 mM sodium acetate (pH 5), and the purity was assessed using a 16% Tricine gel [16]. The concentration of the proteases was determined using the molar extinction coefficient of 25 480 M−1 cm−1 from absorbance spectra obtained on a Jasco V-630 spectrophotometer. The purified proteases were aliquoted and stored at −80°C until used.

Active site titration

To assess the percentage of protease in the active conformation, ITC was performed. Acetyl pepstatin (200 µM), an inhibitor of aspartyl proteases, was titrated against 17 µM wild-type protease and 18 µM L38↑N↑L protease in 10 µl injections at 293.8 K using a Nano-ITC instrument (TA Instruments, Delaware, U.S.A.). The heat due to the dilution of acetyl pepstatin was subtracted from the data set and the baseline adjusted using NITPIC [17]. The changes in heats were integrated and fitted using ITCsy [17]. The percentage of active sites was determined from the stoichiometry value, with a value of 1 indicating 100% of the protease was in the active conformation.

Steady-state and inhibition kinetics

The kinetic parameters KM, kcat, kcat/KM and the specific activity were determined in separate experiments. The hydrolysis of the HIV-1 protease fluorogenic substrate Abz-Arg-Val-Nle-Phe(NO2)-Glu-Ala-Nle-NH2 was monitored. For all kinetic measurements, an excitation wavelength of 337 nm and an emission wavelength of 425 nm were used at 1 min measurement intervals during steady state. All activity assays were performed in 50 mM sodium acetate and 1 M sodium chloride (pH 5.0) at 20°C. All kinetic experiments were performed in triplicate using a Jasco FP-6300 spectrofluorometer (Easton, MD, U.S.A.).

An active enzyme concentration of 50 nM and substrate concentrations ranging from 5 to 200 µM, which achieved saturation kinetics, were used to determine the KM. To determine the kcat/KM, a substrate concentration ranging from 1 to 10 µM was used with an active enzyme concentration of 50 nM. The specific activity and the kcat were determined using active enzyme amounts ranging from 1 to 10 pmol and a constant substrate concentration of 50 µM. All data were fit using SigmaPlot (Systat Software, San Jose, CA, U.S.A.).

The FDA-approved drugs (LPV, ATV and DRV) competitively inhibit HIV protease with dissociation constants in the nanomolar and picomolar range. The inhibition constants (Ki) for LPV, ATV and DRV were determined using the equation for tight-binding inhibitors [18]: 
formula
where E is the active enzyme concentration (50 nM), S is the substrate concentration (50 µM) and IC50 is the concentration of inhibitor at which there is half-maximal activity of the protease. IC50 values were determined using inhibitor concentrations ranging from 0 to 200 µM. A final concentration of 2% (v/v) dimethyl sulfoxide was used during IC50 determinations for inhibitor solubility.

Phenotypic assessment of PI susceptibility

A phenotypic assay using a single infection event per virion was conducted as previously described [19]. Briefly, HEK293T cells were transfected with 300 ng of plasmid pDMG (encodes for vesicular stomatitis virus G protein for entry), 500 ng of plasmid pCSFLW (encodes for firefly luciferase for quantification) and 300 ng of the HIV expression vector (encodes HIV-1 Gag-Pol) using 3.3 µg of polyethylenimine (PEI, Polysciences, Inc., Warrington, PA, U.S.A.). The transfected cells were harvested after 18 h and seeded in the presence of serially diluted LPV (60 nM–3 pM), ATV (40 nM–2 pM) and DRV (60 nM–3 pM). The supernatants were transferred to corresponding 96-well culture plates that contained fresh HEK293T cells after 24 h. The degree of infection was determined 48 h later by measuring the expression of firefly luciferase using a BrightGlo luciferase assay system (Promega, Madison, U.S.A.) on the Victor3 multi-label plate reader (PerkinElmer, Waltham, U.S.A.). For each drug–virus combination, the IC50 was calculated using Microsoft Excel (Microsoft, Redmond, U.S.A.). The phenotypic susceptibility was expressed as the fold change in the IC50 relative to that of the wild-type virus. The assay-specific fold-change cut-off value for each drug was determined using the 99th percentile of the average IC50 for the wild-type pseudovirus assessed in multiple repeat screens for each drug. Two-way analysis of variance (ANOVA) and Bonferroni's post-test were used to identify significant differences in IC50 values between wild type, L38NL and WTGAG L38NL. The p8.MJ4GP wild-type pseudovirus contained the Gag-pol of an HIV-1 subtype C reference isolate while the p8.WTGAG L38NL contained the MJ4 Gag and L38↑N↑L protease. The L38NL pseudovirus contained the Gag-protease from L38↑N↑L. The drug-resistant control was a Gag and protease sequence found within a patient failing therapy and contained many drug resistance mutations (all protease mutations present: L10I, K20R, E35D, M46I, I54V,Q61H, I62V, L63P, T74S and V82A).

Replication capacity

To assess the RC of each of the pseudo virions, an in-house p24 enzyme-linked immunosorbent assay (ELISA) was used according to Aalto Bioreagents (Dublin, Ireland). Supernatants containing pseudovirions generated in the absence of PIs were used. Supernatants were submitted in triplicate for viral titration and p24 ELISA, and the ratio of relative light units (RLU) to p24 was calculated. The percentage replication capacity (%RC) was calculated relative to the wild-type control.

Results

Expression and purification

The protease variant in this study (Figure 1) was isolated from an infant whose mother was part of the Prevention of Mother-to-Child Transmission programme in South Africa [20] and was PI naïve. This sequence is a subtype C sequence, and the variant contains the following subset of mutations: K20R, E35D, R57K and V82I as well as a double insertion of leucine and asparagine at position 38. The variant will be referred to as L38↑N↑L, the upward arrows indicating the insertion of two amino acids at position 38.

Homology model of L38↑N↑L protease and sequence data.

Figure 1.
Homology model of L38↑N↑L protease and sequence data.

HIV-1 protease is a homodimer and contains mainly β-sheets and one α-helix per monomer. There are five regions defined within the structure: the flap region (green), the hinge region (red), the fulcrum region (cyan), the cantilever region (violet) and the dimer interface (orange). The catalytic Asp25 residues are shown in the active site (blue). The yellow spheres and the yellow boxes on the sequence represent the relative positions of the subset of mutations, K20R, E35D, R57K and V82I. The red spheres on the structure and the red box in the sequence alignment represent the double insertion of Leu and Asn. PyMOL was used to generate the homology model using data from the Protein Data Bank (PDB ID: 3U71). The sequence alignment was generated using the Clustal Omega tool (EMBL-EBI).

Figure 1.
Homology model of L38↑N↑L protease and sequence data.

HIV-1 protease is a homodimer and contains mainly β-sheets and one α-helix per monomer. There are five regions defined within the structure: the flap region (green), the hinge region (red), the fulcrum region (cyan), the cantilever region (violet) and the dimer interface (orange). The catalytic Asp25 residues are shown in the active site (blue). The yellow spheres and the yellow boxes on the sequence represent the relative positions of the subset of mutations, K20R, E35D, R57K and V82I. The red spheres on the structure and the red box in the sequence alignment represent the double insertion of Leu and Asn. PyMOL was used to generate the homology model using data from the Protein Data Bank (PDB ID: 3U71). The sequence alignment was generated using the Clustal Omega tool (EMBL-EBI).

The wild-type subtype C and the L38↑N↑L proteases were successfully purified from the insoluble fraction, using cation exchange chromatography, as shown in Figure 2. Both proteases were purified to >95% purity and a yield 2.7 mg per 28 g of wet weight was obtained for L38↑N↑L protease.

SDS–PAGE analysis of HIV protease purification.

Figure 2.
SDS–PAGE analysis of HIV protease purification.

Electrophoretogram showing the purification profile of (A) wild-type protease and (B) L38↑N↑L protease. The molecular mass is shown in the first lane of each gel. The purification steps, from cell lysis to final product, are shown from left to right.

Figure 2.
SDS–PAGE analysis of HIV protease purification.

Electrophoretogram showing the purification profile of (A) wild-type protease and (B) L38↑N↑L protease. The molecular mass is shown in the first lane of each gel. The purification steps, from cell lysis to final product, are shown from left to right.

Steady-state and inhibition kinetics

HIV-1 protease has the ability to autolyse. The peptides produced contribute towards the concentration determined spectroscopically but are not catalytically active. Therefore, to determine the percentage of protease in the active conformation, isothermal titration calorimetry (ITC) was conducted. Acetyl pepstatin was titrated against each protease. The stoichiometry of binding was used to determine the percentage active sites. Since a binding ratio of 1 : 1 is expected, a stoichiometry of 1 would represent a sample that has 100% of the protease molecules in the active conformation. The percentage of active protein in each sample preparation was 90% and 80% for wild-type subtype C protease and L38↑N↑L, respectively. This percentage was used to adjust the concentrations obtained spectroscopically.

Using a fluorogenic substrate that mimics the capsid/p2 cleavage site, the steady-state kinetics of each protease was determined (Table 1). A fluorogenic substrate was used rather than a chromogenic substrate due to the increased sensitivity. The KM, specific activity and kcat for L38↑N↑L were approximately half that of the wild type, while the Vmax was similar between the two enzymes. The catalytic efficiency (kcat/KM) of the L38↑N↑L protease was not significantly different from that of the wild-type protease. The L38↑N↑L-specific activity is ∼2-fold lower than that of the wild type.

Table 1
The steady-state kinetic parameters determined for wild-type and L38↑N↑L protease
Parameter Wild-type PR L38↑N↑L 
KM (µM) 14 ± 1.7 7 ± 0.9 
Vmax (µmol min−10.01 ± 0.0003 0.01 ± 0.0006 
Specific activity (µmol min−1 mg−121.0 ± 1.1 12.1 ± 1.1 
kcat (s−17.7 ± 0.4 4.5 ± 0.4 
kcat/KM (s−1 µM−11.6 ± 0.4 1.0 ± 0.004 
Parameter Wild-type PR L38↑N↑L 
KM (µM) 14 ± 1.7 7 ± 0.9 
Vmax (µmol min−10.01 ± 0.0003 0.01 ± 0.0006 
Specific activity (µmol min−1 mg−121.0 ± 1.1 12.1 ± 1.1 
kcat (s−17.7 ± 0.4 4.5 ± 0.4 
kcat/KM (s−1 µM−11.6 ± 0.4 1.0 ± 0.004 

IC50 values were used to calculate the inhibition constants for each drug (Table 2). The Ki for L38↑N↑L for LPV was 10-fold less than that for the wild-type protease and 3-fold less than that for DRV. The Ki for ATV was, however, 3-fold higher for L38↑N↑L than for the wild-type protease.

Table 2
The Ki (nM) values obtained for wild-type and L38↑N↑L protease against the drugs LPV, ATV and DRV
Protease LPV ATV DRV 
Wild type 2.1 ± 0.2 1.2 ± 0.1 1.4 ± 0.2 
L38↑N↑L 0.2 ± 0.02 3.5 ± 0.7 0.4 ± 0.02 
Protease LPV ATV DRV 
Wild type 2.1 ± 0.2 1.2 ± 0.1 1.4 ± 0.2 
L38↑N↑L 0.2 ± 0.02 3.5 ± 0.7 0.4 ± 0.02 

Phenotypic susceptibility and replication capacity

A phenotypic viral assay was conducted to determine the susceptibility of L38↑N↑L to PIs in the presence of a Gag sequence. The Gag sequence that was isolated from the patient was tested in the pseudovirion termed L38NL. A wild-type subtype C Gag sequence was included with the variant protease to determine whether the effects seen were due to the Gag or the variant protease. This pseudovirion is denoted as WTGAG L38NL. The wild-type pseudovirion refers to a wild-type subtype C Gag and protease sequence. A drug-resistant control was included to demonstrate a phenotype of resistance. It was found that the L38↑N↑L protease was susceptible to LPV (L38NL: 4.0 ± 0.8 nM and WTGAG L38NL: 4.4 ± 1.2 nM) and ATV (L38NL: 4.2 ± 0.7 nM and WTGAG L38NL: 4.4 ± 0.4 nM) with both Gag sequences. The IC50 fold-change was not 2.9-fold greater than the wild-type (LPV and ATV: 1.4 ± 0.5 nM), indicating no reduction in susceptibility (Figure 3). The L38↑N↑L protease showed reduced susceptibility to DRV with both wild-type and mutated Gag (Figure 3). In the presence of the mutated Gag, the L38↑N↑L protease was less susceptible to DRV than when in the presence of the wild-type subtype C Gag. The DRV IC50 for L38NL (1.6 ± 0.2 nM) was 5-fold higher than that of wild type (0.3 ± 0.05 nM), above the 1.3-fold cut-off. WTGAG L38NL IC50 for DRV (1.0 ± 0.2 nM) was 3-fold higher than that of wild type. The resistant control showed high-level resistance (FC >10) for all three PIs.

Phenotypic susceptibility of pseudoviruses to approved PIs.

Figure 3.
Phenotypic susceptibility of pseudoviruses to approved PIs.

The phenotypic susceptibility of the wild-type control, L38NL sample with the L38↑N↑L protease and the accompanying Gag sequence (L38NL), and the L38↑N↑L protease with a wild-type subtype C Gag (WTGAG L38NL) is shown. ***P < 0.001, ****P < 0.0001, ns, not significant.

Figure 3.
Phenotypic susceptibility of pseudoviruses to approved PIs.

The phenotypic susceptibility of the wild-type control, L38NL sample with the L38↑N↑L protease and the accompanying Gag sequence (L38NL), and the L38↑N↑L protease with a wild-type subtype C Gag (WTGAG L38NL) is shown. ***P < 0.001, ****P < 0.0001, ns, not significant.

The replication capacity (RC) of each pseudovirion was determined because mutations in protease often result in reduced replication of the virus. The wild-type pseudovirion was used as the reference. An analysis of the RC (Figure 4) showed the L38NL pseudovirion to have a reduced RC (24%), while the WTGAG L38NL pseudovirion showed an increased RC (120%) when compared with the wild-type sample. The drug-resistant control showed the lowest RC (13%).

Relative RC of the pseudovirions.

Figure 4.
Relative RC of the pseudovirions.

The relative RC of each pseudovirus was compared with the wild type. The L38NL virus showed a reduction in RC while WTGAG L38NL showed an increase in 20%. The drug-resistant control had a RC of 13%.

Figure 4.
Relative RC of the pseudovirions.

The relative RC of each pseudovirus was compared with the wild type. The L38NL virus showed a reduction in RC while WTGAG L38NL showed an increase in 20%. The drug-resistant control had a RC of 13%.

Analysis of L38↑N↑L Gag

The accompanying Gag of L38↑N↑L was sequenced to determine if there were any drug-resistant mutations present. The following mutations were found in the protease cleavage sites: T370A, M374V, R376G in the p2/NC cleavage site, E424G in the NC/p1 cleavage site and N448S in the p1/p6 cleavage site. A duplication of the PTAPP motif was seen downstream of p1/p6 cleavage site. Multiple polymorphisms were also observed in the non-cleavage site regions as seen in Figure 5.

Alignment of Gag sequences of wild-type subtype C and the L38↑N↑L variant.

Figure 5.
Alignment of Gag sequences of wild-type subtype C and the L38↑N↑L variant.

The cleavage sites are indicated in red and PTATPPAE insertion indicated in blue. The sequence alignment was generated using the Clustal Omega tool (EMBL-EBI).

Figure 5.
Alignment of Gag sequences of wild-type subtype C and the L38↑N↑L variant.

The cleavage sites are indicated in red and PTATPPAE insertion indicated in blue. The sequence alignment was generated using the Clustal Omega tool (EMBL-EBI).

Discussion

Insertion mutations in protease occur rarely and thus have not been well characterised, especially in subtype C viruses. Here, we show the drug susceptibility of a subtype C protease containing a double insertion, which displays reduced susceptibility to DRV using phenotypic assays.

The L38↑N↑L and wild-type proteases were expressed and purified from inclusion bodies, with 80% and 90% of the protease in the active conformation, respectively. An analysis of the KM indicated that the L38↑N↑L variant had an increased affinity for CA/p2 cleavage site-mimicking substrate. The catalytic efficiency (kcat/KM) of L38↑N↑L was not significantly different from that of wild type and is comparable to findings of Velázquez-Campoy et al. [21]. This is unusual as Kozísek et al. [12] showed that amino acid insertions in the hinge region of protease reduce the catalytic efficiency. The affinities of LPV and DRV were increased for L38↑N↑L as displayed by the decrease in inhibitory constant (Ki). However, in the presence of ATV, a 3-fold decrease in affinity was observed for the L38↑N↑L, although this was not large enough to indicate resistance. This indicates that these PIs would inhibit variant protease without a Gag region present.

A phenotypic assay was conducted to determine the effect of the L38↑N↑L insertion mutation on the in vitro drug susceptibility to PIs. This was performed in conjunction with an unrelated wild-type control Gag, as well as the related Gag, to assess the phenotypic impact of Gag on drug susceptibility in the context of the L38↑N↑L insertion mutation. For LPV and ATV, no significant reduction in phenotypic susceptibility was observed for either Gag-containing version of the L38↑N↑L protease. However, for DRV, both Gag-containing versions of L38↑N↑L protease showed a small but significant decrease in DRV susceptibility relative to the wild type. A reduced susceptibility to DRV for an insertion mutation at position 35 was also observed [12]. Furthermore, the decrease in DRV susceptibility was significantly greater for the related Gag-containing L38↑N↑L protease than for the unrelated Gag-containing L38↑N↑L protease, implicating Gag as a contributor towards drug resistance in the context of the L38↑N↑L insertion mutation. This underlines the importance of including the Gag region in PI-resistance screening, as was also shown by Giandhari et al. [22].

Resistance to PIs occurs with the accumulation of mutations in the protease, which may lead to a reduced RC of the virus as the affinity for its natural substrate has decreased [9,19,23]. There are, however, mutations that occur in Gag that will partially restore the replicative capacity of the virus, and thus, Gag mutates to compensate for mutations in protease [19,24]. These mutations can occur in cleavage sites or elsewhere [19]. Resistance to PIs is not only due to the protease itself but also due to amino acid substitutions in the Gag cleavage sites [19,25]. However, when patients are failing therapy, drug-resistance mutations in Gag are typically not considered.

Mutations in Gag cleavages sites have been linked to drug resistance [25], and several cleavage site mutations were observed in the related Gag sequence, including mutations in the p2/NC, NC/p1 and p1/p6 cleavage sites. Mutations in the non-cleavage site regions of Gag have also been linked to drug resistance; these include H219Q and R409K [26] and K437V and K436E [27]. The mutations reported here, to the best of our knowledge, have not been reported to be linked with drug resistance. The PTAPP motif within p6-Gag is a proline-rich domain that is responsible for recruiting Tsg101, a cellular factor involved budding of the virus. The PTAPP duplication, a common polymorphism in Gag, is more common in subtype C viruses [28]. This could be due to the absence of the Alix-binding YPXnL motif in p6gag, which aids in budding, and so, a duplication of the PTAPP is a compensatory mechanism for viral budding [29]. The Alix motif is mutated in more than 95% of subtype C viruses [30]. PTAPP duplications are selected for in viral isolates from HAART treatment [31], and PTAPP duplications accumulate in subtype C isolates with a smaller number of protease resistance mutations when compared with subtype B [29]. It has been suggested that PTAPP duplications may play a role in subtype C virus susceptibility to PIs due to its position close to the cleavage site as it may be responsible for an important secondary structure conferred by prolines [29]. The PTAPP duplication has been linked to poor virological response to the PI amprenavir [32]. The loss of susceptibility to PIs or any ARV drug may be due to enhanced budding in the presence of drugs [32].

The reduction in viral fitness is not uncommon in viruses that are evolving to escape drug pressure [23,33]. The selection of drug resistance mutations in HIV-1 protease often results in a loss of RC, which is then compensated for by the selection of additional compensatory mutations in Gag [34]. We have observed a reduction in the relative replicative capacity of the L38↑N↑L protease in the context of its related Gag. Kozísek et al. [12] have also shown that a protease containing amino acid insertions in the hinge region display reduced replicative capacity in the presence of a mutated Gag. It is unclear whether the mutations/polymorphisms observed in the related Gag were selected for as a consequence of the L38↑N↑L insertion mutation in the protease, especially since our RC studies have shown a debilitating effect of the related Gag on the RC of the L38↑N↑L protease. However, the possibility exists that changes in the related Gag occurred independent of the L38↑N↑L insertion mutation. In fact, it has been shown that mutations can develop in Gag without any mutations being present in the protease [27].

In summary, catalytically active L38↑N↑L variant protease was successfully purified. L38↑N↑L protease is susceptible to LPV, ATV and DRV without a Gag sequence present. Phenotypic assays with a Gag sequence showed that the L38↑N↑L variant has reduced susceptibility to DRV. The L38↑N↑L protease was less susceptible to DRV when the mutated Gag was present, indicating that Gag plays a role in drug susceptibility here. The duplication of the PTAPP motif may also play a role. The mutated Gag lowered the RC, indicating that mutations may have occurred in this region before the protease, implicating an alternative mechanism for drug resistance developing in this patient.

Abbreviations

     
  • ART

    antiretroviral therapy

  •  
  • ARV

    antiretroviral

  •  
  • DRV

    darunavir

  •  
  • ELISA

    enzyme-linked immunosorbent assay

  •  
  • ITC

    isothermal titration calorimetry

  •  
  • L38↑N↑L

    HIV-1 subtype C protease containing leucine at position 38 followed by a double insertion of asparagine and leucine

  •  
  • L38NL

    pseudovirion containing L38↑N↑L protease and its related Gag

  •  
  • LPV

    lopinavir

  •  
  • PIs

    protease inhibitors

  •  
  • RC

    replication capacity

  •  
  • WTGAG L38NL

    pseudovirion containing L38↑N↑L protease and a wild-type subtype C Gag

Author Contribution

A.W. performed all the experimental work, analysed the data and prepared the manuscript. A.B. assisted in the analysis of phenotypic data and manuscript revision. I.A. assisted in experimental design and manuscript revision. H.W.D. assisted in manuscript revision. L.M. assisted in the analysis of phenotypic data and manuscript revision. Y.S. supervised the project and assisted in data analysis and interpretation.

Funding

The South African Medical Research Council supported this research under a Self-Initiated Research Grant to Y.S. The research was also supported by a grant to H.W.D. via the DST/NRF South African Research Chair Initiative Programme (grant no. 64788). The views and opinions expressed are those of the authors and do not necessarily represent the official views of the SA MRC and the DST/NRF.

Acknowledgements

We thank Johanna Ledwaba and Dr Gillian Hunt for identifying this rare protease sequence and Dr Jennifer Giandhari for the help with Gag-Pro sequencing.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
UNAIDS
. (
2016
)
Global Report: UNAIDS Report on Global AIDS Epidemic 2016
,
UNAIDS
,
Geneva, Switzerland
2
McCutchan
,
F.E.
(
2006
)
Global epidemiology of HIV
.
J. Med. Virol.
78
,
S7
S12
3
Hemelaar
,
J.
,
Gouws
,
E.
,
Ghys
,
P.D.
and
Osmanov
,
S.
(
2011
)
Global trends in molecular epidemiology of HIV-1 during 2000–2007
.
AIDS
25
,
679
689
4
Roberts
,
J.D.
,
Bebenek
,
K.
and
Kunkel
,
T.A.
(
1988
)
The accuracy of reverse transcriptase from HIV-1
.
Science
242
,
1171
1173
5
Turner
,
B.G.
and
Summers
,
M.F.
(
1999
)
Structural biology of HIV
.
J. Mol. Biol.
285
,
1
32
6
Wlodawer
,
A.
,
Miller
,
M.
,
Jasklski
,
M.
,
Sathyanarayana
,
B.K.
,
Baldwin
,
E.
,
Weber
,
I.T.
et al.  (
1989
)
Conserved folding in retroviral proteases: crystal structure of a synthetic HIV-1 protease
.
Science
245
,
616
621
7
Kohl
,
N.E.
,
Emini
,
E.A.
,
Schleif
,
W.A.
,
Davis
,
L.J.
,
Heimbach
,
J.C.
,
Dixon
,
R.A.
et al.  (
1988
)
Active human immunodeficiency virus protease is required for viral infectivity
.
Proc. Natl Acad. Sci. U.S.A.
85
,
4686
4690
8
World Health Organization (WHO)
. (
2016
)
Consolidated Guidelines on the use of Antiretroviral Drugs for Treating and Preventing HIV Infection: Recommendations for A Public Health Approach
,
WHO Press
,
Geneva, Switzerland
9
Condra
,
J.H.
,
Schleif
,
W.A.
,
Blahy
,
O.M.
,
Gabryelski
,
L.J.
,
Graham
,
D.J.
,
Quintero
,
J.C.
et al.  (
1995
)
In vivo emergence of HIV-1 variants resistant to multiple protease inhibitors
.
Nature
374
,
569
571
10
Ridky
,
T.W.
,
Kikonyogo
,
A.
,
Leis
,
J.
,
Gulnik
,
S.
,
Copeland
,
T.
,
Erickson
,
J.
et al.  (
1998
)
Drug-resistant HIV-1 proteases identify enzyme residues important for substrate selection and catalytic rate
.
Biochemistry
37
,
13835
13845
11
Masquelier
,
B.
,
Race
,
E.
,
Tamalet
,
C.
,
Descamps
,
D.
,
Izopet
,
J.
,
Buffet-janvresse
,
C.
et al.  (
2001
)
Genotypic and phenotypic resistance patterns of human immunodeficiency virus type 1 variants with insertions or deletions in the reverse transcriptase (RT): multicenter study of patients treated with RT inhibitors
.
Antimicrob. Agents Chemother.
45
,
1836
1842
12
Kozísek
,
M.
,
Sasková
,
K.G.
,
Rezácová
,
P.
,
Brynda
,
J.
,
van Maarseveen
,
N.M.
,
De Jong
,
D.
et al.  (
2008
)
Ninety-nine is not enough: molecular characterization of inhibitor-resistant human immunodeficiency virus type 1 protease mutants with insertions in the flap region
.
J. Virol.
82
,
5869
5878
13
Clavel
,
F.
and
Hance
,
A.J.
(
2004
)
HIV drug resistance
.
N. Engl. J. Med.
350
,
1023
1035
14
Kuhn
,
L.
,
Hunt
,
G.
,
Technau
,
K.-G.
,
Coovadia
,
A.
,
Ledwaba
,
J.
,
Pickerill
,
S.
et al.  (
2014
)
Drug resistance among newly-diagnosed HIV-infected children in the era of more efficacious antiretroviral prophylaxis
.
AIDS
28
,
1673
1678
15
Mosebi
,
S.
,
Morris
,
L.
,
Dirr
,
H.W.
and
Sayed
,
Y.
(
2008
)
Active-site mutations in the South African human immunodeficiency virus type 1 subtype C protease have a significant impact on clinical inhibitor binding: kinetic and thermodynamic study
.
J. Virol.
82
,
11476
11479
16
Schägger
,
H.
(
2006
)
Tricine-SDS-PAGE
.
Nat. Protoc.
1
,
16
22
17
Brautigam
,
C.A.
,
Zhao
,
H.
,
Vargas
,
C.
,
Keller
,
S.
and
Schuck
,
P.
(
2016
)
Integration and global analysis of isothermal titration calorimetry data for studying macromolecular interactions
.
Nat. Protoc.
11
,
882
894
18
Copeland
,
R.A.
,
Lombardo
,
D.
,
Giannaras
,
J.
and
Decicco
,
C.P.
(
1995
)
Estimating KI values for tight binding inhibitors from dose-response plots
.
Bioorg. Med. Chem. Lett.
5
,
1947
1952
19
Parry
,
C.M.
,
Kohli
,
A.
,
Boinett
,
C.J.
,
Towers
,
G.J.
,
McCormick
,
A.L.
and
Pillay
,
D.
(
2009
)
Gag determinants of fitness and drug susceptibility in protease inhibitor-resistant human immunodeficiency virus type 1
.
J. Virol.
83
,
9094
9101
20
South African Department of Health
. (
2010
)
Clinical Guidelines: PMTCT (Prevention of Mother-to- Child Transmission)
,
National Department of Health
,
South Africa
21
Velázquez-Campoy
,
A.
,
Vega
,
S.
,
Fleming
,
E.
,
Bacha
,
U.
,
Sayed
,
Y.
,
Dirr
,
H.W.
et al.  (
2003
)
Protease inhibition in African subtypes of HIV-1
.
AIDS Rev.
5
,
165
171
PMID:
[PubMed]
22
Giandhari
,
J.
,
Basson
,
A.E.
,
Sutherland
,
K.
,
Parry
,
C.M.
,
Cane
,
P.A.
,
Coovadia
,
A.
et al.  (
2016
)
Contribution of gag and protease to HIV-1 phenotypic drug resistance in paediatric patients failing protease-inhibitor based therapy
.
Antimicrob. Agents Chemother.
60
,
2248
2256
23
Croteau
,
G.
,
Doyon
,
L.
,
Thibeault
,
D.
,
Mckercher
,
G.
,
Pilote
,
L.
,
Croteau
,
G.
et al.  (
1997
)
Impaired fitness of human immunodeficiency virus type 1 variants with high-level resistance to protease inhibitors
.
J. Virol.
71
,
1089
1096
PMID:
[PubMed]
24
Myint
,
L.
,
Matsuda
,
M.
,
Matsuda
,
Z.
,
Yokomaku
,
Y.
,
Chiba
,
T.
,
Okano
,
A.
et al.  (
2004
)
Gag non-cleavage site mutations contribute to full recovery of viral fitness in protease inhibitor-resistant human immunodeficiency virus type 1
.
Antimicrob. Agents Chemother.
48
,
444
452
25
Clavel
,
F.
and
Mammano
,
F.
(
2010
)
Role of gag in HIV resistance to protease inhibitors
.
Viruses
2
,
1411
1426
26
Gatanaga
,
H.
,
Suzuki
,
Y.
,
Tsang
,
H.
,
Yoshimura
,
K.
,
Kavlick
,
M.F.
,
Nagashima
,
K.
et al.  (
2002
)
Amino acid substitutions in gag protein at non-cleavage sites are indispensable for the development of a high multitude of HIV-1 resistance against protease inhibitors
.
J. Biol. Chem.
277
,
5952
5961
27
Nijhuis
,
M.
,
Van Maarseveen
,
N.M.
,
Lastere
,
S.
,
Schipper
,
P.
,
Coakley
,
E.
,
Glass
,
B.
et al.  (
2007
)
A novel substrate-based HIV-1 protease inhibitor drug resistance mechanism
.
PLoS Med.
4
,
e36
28
Marlowe
,
N.
,
Flys
,
T.
,
Hackett
,
J.
,
Schumaker
,
M.
,
Jackson
,
J.B.
and
Eshleman
,
S.H.
(
2004
)
Analysis of insertions and deletions in the gag p6 region of diverse HIV type 1 strains
.
AIDS Res. Hum. Retroviruses
20
,
1119
1125
29
Martins
,
A.N.
,
Arruda
,
M.B.
,
Pires
,
A.F.
,
Tanuri
,
A.
and
Brindeiro
,
R.M.
(
2011
)
Accumulation of P(T/S)AP late domain duplications in HIV type 1 subtypes B, C, and F derived from individuals failing ARV therapy and ARV drug-naive patients
.
AIDS Res. Hum. Retroviruses
27
,
687
692
30
Neogi
,
U.
,
Rao
,
S.D.
,
Bontell
,
I.
,
Verheyen
,
J.
,
Rao
,
V.R.
,
Gore
,
S.C.
et al.  (
2014
)
Novel tetra-peptide insertion in Gag-p6 ALIX-binding motif in HIV-1 subtype C associated with protease inhibitor failure
.
AIDS
28
,
2319
2322
31
Fun
,
A.
,
Wensing
,
A.M.J.
,
Verheyen
,
J.
and
Nijhuis
,
M.
(
2012
)
Human immunodeficiency virus gag and protease: partners in resistance
.
Retrovirology
9
,
1
14
32
Lastere
,
S.
,
Dalban
,
C.
,
Collin
,
G.
,
Descamps
,
D.
,
Girard
,
P.M.
,
Clavel
,
F.
et al.  (
2004
)
Impact of insertions in the HIV-1 p6 PTAPP region on the virological response to amprenavir
.
Antivir. Ther.
9
,
221
227
PMID:
[PubMed]
33
Mammano
,
F.
,
Petit
,
C.
and
Clavel
,
F.
(
1998
)
Resistance-associated loss of viral fitness in human immunodeficiency virus type 1: phenotypic analysis of protease and gag coevolution in protease inhibitor-treated patients
.
J. Virol.
72
,
7632
7637
PMID:
[PubMed]
34
Dauber
,
D.S.
,
Ziermann
,
R.
,
Parkin
,
N.
,
Maly
,
D.J.
,
Mahrus
,
S.
,
Harris
,
J.L.
et al.  (
2002
)
Altered substrate specificity of drug-resistant human immunodeficiency virus type 1 protease
.
J. Virol.
76
,
1359
1368