Polymorphic substitution E157Q in HIV-1 integrase increases R263K-mediated dolutegravir resistance and decreases DNA binding activity

Abstract

Objectives

The E157Q substitution in HIV-1 integrase (IN) is a relatively common natural polymorphism associated with HIV resistance to IN strand transfer inhibitors (INSTIs). Although R263K is the most common resistance substitution for the INSTI dolutegravir, an INSTI treatment-experienced individual recently failed dolutegravir-based therapy, with E157Q being the only resistance-associated change reported. Given that different resistance pathways can sometimes synergize to confer high levels of resistance to antiretroviral drugs, we studied the effects of E157Q in association with R263K. Because Glu157 is thought to lie within the binding site of HIV IN DNA binding inhibitors such as FZ41, we also evaluated DNA binding activity and resistance to IN inhibitors in the presence of E157Q.

Methods

Purified recombinant IN proteins were assessed in cell-free assays for their strand transfer and DNA binding activities. NL4.3 viral stocks harbouring IN mutations were generated and characterized in the presence and absence of IN inhibitors in tissue culture.

Results

E157Q alone had little if any effect on the biochemical activity of IN, and partially restored the activity of R263K-containing IN. The E157Q/R263K double viral mutant displayed infectiousness in culture equivalent to WT, while increasing resistance to dolutegravir by 10-fold compared with lower-level resistance associated with R263K alone. None of the mutations tested showed significant resistance to either raltegravir or FZ41.

Conclusions

This study shows that E157Q may act as a compensatory mutation for R263K. Since E157Q is a natural polymorphism present in 1%–10% of HIV-positive individuals, it may be of particular importance for patients receiving INSTI therapy.

Introduction

HIV-1 is a highly heterogeneous virus, existing as a quasispecies with different variants both within a single patient and at a population level.¹ This is due to the error-prone reverse transcriptase (RT) enzyme of HIV, which frequently inserts incorrect nucleotides during reverse transcription, leading to the generation of multiple mutations, some with the potential to confer resistance to antiretroviral drugs (ARVs).²

Although resistance has emerged in patients for every currently available ARV class,³ the integrase (IN) strand transfer inhibitor (INSTI) dolutegravir has yet to be shown to select for resistance mutations in treatment-naive individuals, unlike the earlier drugs of this class, raltegravir and elvitegravir.^4–6 These latter drugs can initially select for primary resistance mutations that cause drug resistance at the expense of viral replicative capacity and later for compensatory mutations that can restore replication capacity while further increasing the level of drug resistance.⁷ When used as an initial INSTI in a highly treatment-experienced patient population, dolutegravir did not select for classical INSTI primary resistance mutations; instead, the R263K substitution has been observed in treatment-experienced individuals who failed therapy with this drug.⁸ We have also observed this substitution during tissue culture selection studies with HIV-infected cells passaged in increasing concentrations of dolutegravir.⁹ R263K decreases both viral replicative capacity and enzymatic strand transfer activity by 20%–30% while conferring 2- to 5-fold resistance to dolutegravir.¹⁰

It was recently reported that a patient failed therapy with raltegravir, and subsequently dolutegravir, with only one known resistance-associated substitution present in IN: E157Q.¹¹ Although considered a polymorphic substitution, E157Q has long been implicated in clinical INSTI resistance.¹² We previously characterized E157Q in the context of the N155H/R263K HIV-1 subtype B double mutant and showed that E157Q did not have a restorative effect in this background.¹³

The present study was designed to determine what effect, if any, E157Q might have on dolutegravir resistance, viral infectivity and IN strand transfer activity. We also assessed the effect of E157Q together with R263K. The results show that E157Q acts as a compensatory mutation, increasing enzymatic activity, infectivity and dolutegravir resistance in a R263K-containing background, even though this substitution has little effect on its own on IN catalytic activity and infectivity in the absence of dolutegravir while also rendering the virus hypersusceptible to this inhibitor.

We also investigated the effects of E157Q on IN DNA binding activity and susceptibility to a DNA binding inhibitor in tissue culture, in part because we recently identified Glu157 as a putative interaction domain with the HIV IN DNA binding inhibitor FZ41.¹⁴ However, E157Q did not have any significant effect on either DNA binding or resistance to FZ41, alone or in combination with R263K.

These results suggest that the natural polymorphism E157Q may partially compensate for impairment of the IN protein conferred by the dolutegravir resistance mutation R263K, while increasing levels of drug resistance, and this may have implications for the use of dolutegravir in some HIV-infected individuals.

Materials and methods

Experimental design

Our objectives were to evaluate the effects of the E157Q polymorphism on the biochemical activities of purified recombinant IN in cell-free assays as well as on viral replication in the presence or absence of IN inhibitors in tissue culture. We studied the effect of E157Q both alone and in combination with the R263K dolutegravir mutation, compared with WT and R263K alone.

Cells and reagents

TZM-bl and 293T cells were cultured as reported previously.⁹ Merck & Co., Inc., ViiV Healthcare Ltd and LBPA, ENS Cachan, CNRS supplied raltegravir, dolutegravir and FZ41 (CID 5481653), respectively.

IN strand transfer activity assay

Site-directed mutagenesis was performed to create both pET15b_E157Q and pET15b_E157Q/R263K plasmids as previously described for the pET15b_R263K IN subtype B expression vector using the following primers: sense, 5′-GTCAAGGAGTAATAGAATCTATGAATAACAGTTAAAGAAAATTATAGGACAGGATAGAG-3′; and antisense, 5′-CTCTTACCTGTCCTATAATTTTCTTTAACTGTTTATTCATAGATTCTATTACTCCTTGAC-3′.¹⁰ The expression and purification of recombinant IN proteins has also been described previously.⁹ Strand transfer assays were performed as previously published.¹⁰ Briefly, DNA-Bind 96-well plates (Corning) were coated with HIV-1 pre-processed long terminal repeat (LTR) DNA at 4°C for 48 h and then incubated with 600 nM purified proteins. Increasing concentrations of biotinylated target DNA were then added to the plates and the reaction was allowed to occur for 1 h at 37°C. Europium (Eu)-labelled streptavidin, which stably bound the biotin tag of the target DNA molecules, was added and strand transfer activity was quantified via Eu fluorescence.

IN DNA binding activity assay

DNA binding activity assays were performed as described previously using purified recombinant proteins.^15,16 Briefly, High Bind black 96-well plates (Corning) were coated with 600 nM IN proteins at 4°C for 16 h. Fluorescently labelled Rhodamine Red (RhoR)–LTR HIV-1 DNA duplexes were then added and binding was allowed for 1 h at room temperature in the dark. Unbound DNA was washed away and DNA binding activity was measured via RhoR fluorescence.

Generation of NL4.3 HIV-1 viral clones

The generation of the pNL4.3_IN(R263K) plasmid has been reported previously⁹ and similar methods were used to generate the pNL4.3_IN(E157Q) and pNL4.3_{IN(E157Q/R263K)} plasmids using the primers described above. Viral stocks were produced as described previously.⁹ 293T cells were transfected with plasmid DNA using Lipofectamine 2000. At 48 h after transfection, cell culture fluids were collected and filtered at 0.45 μm to remove cell debris. Viral stocks were then aliquotted and stored at −80°C. Quantification of infectious viral particles was on the basis of measuring cell-free RT activity.

HIV inhibition assay

HIV susceptibilities to dolutegravir, raltegravir and FZ41 were measured by infection of 30 000 TZM-bl cells with 100 000 RT units of each virus in the presence of serial drug dilutions. After 48 h, cells were lysed and luciferase production was measured using the Luciferase Assay System (Promega, Madison, WI, Canada).

HIV infectivity assay

HIV-1 infectivity was measured through the infection of 30 000 TZM-bl cells per well using serial 1:4 dilutions of each virus. Levels of infection were measured as described above.

Statistical analysis

Each experiment was performed with at least two biological replicates in triplicate (n = 6). Biochemical estimates of strand transfer activity were normalized to WT activity at 1600 nM protein (Figure 1a) or 128 nM target DNA (Figure 1b). DNA binding activity was normalized to WT at 6 μM LTR DNA (Figure 2a). Relative EC₅₀ was normalized to WT EC₅₀ (Table 1). Fold change was normalized to WT IC₅₀ (Table 2). The kinetic constant K_M, V_max, EC₅₀, IC₅₀, standard error of the mean (SEM) and 95% CI were calculated using Prism 5.0 software. All figures were visualized by Prism 5.0 software. Student's t-tests were performed using the OpenEpi toolkit, accessible free online at www.openepi.com.

Figure 1.

Strand transfer activities of purified recombinant IN proteins. Biotinylated target DNA incorporation was measured by Eu–streptavidin fluorescence as a function of (a) protein concentration or (b) target DNA concentration, displayed as relative fluorescence units (RLU). (c) K_M values for each IN mutation were obtained from the curves in (b). (d) Enzyme proficiency was quantified by dividing V_max [taken from plateau in (b)] by K_M. Error bars display SEM. *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001 by Student's t-test.

Figure 2.

Cell-free DNA binding activities of purified recombinant IN proteins. (a) Enzymatic activity as a function of increasing HIV LTR DNA concentration, displayed as relative fluorescence units (RLU). (b) K_M values for each mutation were obtained from the curves in (a). (c) Enzyme proficiency was quantified by dividing V_max [taken from plateau in (a)] by K_M. Error bars display SEM. *P ≤ 0.05 by Student's t-test.

Table 1.

Relative EC₅₀ values for NL4.3 viral stocks in TZM-bl cells

Genotype	EC50 relative to WT	95% CI
WT	1.00	0.64–1.56
E157Q	1.23	1.02–1.50
R263K	2.67*	1.99–3.58
E157Q/R263K	0.66	0.53–0.82

*Significantly different compared with WT by Student's t-test (P ≤ 0.01).

Table 2.

IC₅₀ values for NL4.3 viral stocks in TZM-bl cells for dolutegravir, raltegravir and FZ41

Genotype	Dolutegravir			Raltegravir			FZ41
	IC50 (nM)	95% CI	fold change relative to WT	IC50 (nM)	95% CI	fold change relative to WT	IC50 (nM)	95% CI	fold change relative to WT
WT	52.10	16.83–161.2	1.00	23.59	15.28–36.41	1.00	9029	8804–9259	1.00
E157Q	0.4277	0.304–0.602	0.01*†	10.70	7.356–15.55	0.454*†	7605	6419–9011	0.84*
R263K	106.1	20.73–543.4	2.04	4.131	3.124–5.463	0.175*	8114	7590–8674	0.90*
E157Q/R263K	1035	407.3–2632	19.9*†	38.56	18.14–81.96	1.63†	7068	6147–8126	0.78*

*Significantly different compared with WT by Student's t-test (P ≤ 0.05).

†Significantly different compared with R263K by Student's t-test (P ≤ 0.05).

Results

E157Q partially restored the strand transfer activity of R263K-containing IN

Figure 1 displays the strand transfer activity of WT, E157Q-containing, R263K-containing and E157Q/R263K-containing HIV-1 IN subtype B proteins as a function of protein (Figure 1a) or target DNA substrate (Figure 1b) concentrations. Consistent with previous studies,^13,15 all proteins were maximally active at 400 nM. It is evident from both panels that the polymorphic E157Q substitution had no effect on the observed enzyme activity since the curves generated are similar to that of the WT enzyme. As we have previously reported, the R263K-containing IN was impaired in strand transfer activity at every concentration of protein or substrate tested;^10,17–19 however, the double E157Q/R263K mutant displayed an activity that was intermediate between WT or E157Q and R263K. Figure 1(c) displays the affinity of the IN enzymes for the DNA substrate as K_M, where a lower value indicates a higher affinity.²⁰ We have shown previously that IN strand transfer K_M strongly correlates with viral infectivity and replicative capacity and is thus a good indicator of the possible phenotypic effects of various substitutions.^9,17 The presence of R263K resulted in a sharp increase in the observed K_M, which was partially restored in the case of the double mutant. When the measure of enzyme proficiency, V_max/K_M, where V_max represents the maximal strand transfer activity of the protein, was taken into consideration (Figure 1d), the same trend was observed. Thus, the polymorphic substitution E157Q had little effect on its own, but improved the strand transfer activity of IN enzymes containing R263K.

DNA binding activity of IN is restored to WT levels upon the addition of E157Q to R263K

We recently identified position Glu157 as a putative interactive residue with the IN DNA binding inhibitor FZ41;¹⁴ therefore, we evaluated the effect of the E157Q substitution on the in vitro DNA binding activity of IN. Figure 2(a) shows the DNA binding activity of each protein studied as a function of fluorescently labelled HIV LTR DNA concentration. All the enzymes studied appeared to have similar activities. However, K_M calculations showed that R263K alone led to an improved DNA binding activity compared with the other proteins (Figure 2b), and this increase was statistically significant, as shown in enzyme efficiency studies (Figure 2c). Thus, the polymorphic substitution E157Q acts to decrease DNA binding levels of the R263K-containing mutant.

E157Q-containing NL4.3 viruses are infectious

Table 1 displays the infectivity of subtype B NL4.3 viral stocks in TZM-bl reporter cells as a function of increasing RT activity. Although the E157Q/R263K double mutant displayed a lower EC₅₀ value than WT, this trend did not reach statistical significance. As has been shown in the past, the R263K substitution resulted in a ∼30% decrease in HIV-1 infectivity relative to WT.^9,10,17–19 This decrease was completely restored in the case of the E157Q/R263K double mutant, a finding that is consistent with the biochemical results presented here and further showing that E157Q can compensate for the R263K mutation.

E157Q is hypersusceptible to dolutegravir, but enhances R263K-mediated resistance by 10-fold

Table 2 shows the IC₅₀ values of dolutegravir for the viruses studied. The E157Q virus was hypersusceptible to dolutegravir, displaying an IC₅₀ of 0.43 nM, which was 100-fold lower than WT (52.1 nM). The R263K virus displayed low levels of dolutegravir resistance, consistent with previous reports.^9,10,17–19 However, the combination of E157Q and R263K increased dolutegravir resistance by 20- and 10-fold compared with the WT and R263K enzymes, respectively, with the IC₅₀ for the double mutant being elevated to ∼1 μM.

Neither E157Q nor R263K alone or in combination decreased susceptibility to raltegravir or FZ41

Also summarized in Table 2 are IC₅₀ values derived for each virus for raltegravir. Despite previous reports that identified E157Q as a raltegravir resistance-associated substitution in patients failing treatment with this drug,^11,21 we did not observe a significant increase in IC₅₀ compared with WT. The presence of E157Q also had little effect when added to a R263K-containing background; however, it did significantly increase levels of raltegravir resistance compared with R263K alone.

Table 2 also displays the infectivity of viruses in response to the IN DNA binding inhibitor FZ41. Although we had previously identified position 157 in IN as a potential interactor with this inhibitor,¹⁴ the E157Q mutation, alone or in combination with R263K, did not display any increase in IC₅₀ compared with the WT virus.

Discussion

In this study, we addressed the potential role of the E157Q polymorphic HIV-1 substitution in individuals treated with dolutegravir. A previous report had suggested that a patient failed raltegravir, and subsequently dolutegravir, with only E157Q in IN being present. There was a strong indication that this patient was adherent to his ART regimen due to the high plasma concentration of medication and the lack of resistance to the background regimen.¹¹ The authors concluded that HIV bearing this substitution was more fit than WT and also highly resistant to both of the above-mentioned INSTIs. In contrast, our results show that E157Q on its own has only minimal impact on IN, both in terms of enzyme activity and viral infectivity. We have also shown that E157Q on its own confers hypersusceptibility to dolutegravir without significantly altering susceptibility to raltegravir. Discrepancies between our and previous studies can be explained in part by differences in experimental protocols, since we uniquely evaluated the contribution of each mutation by site-directed mutagenesis using the NL4.3 backbone while the other studies utilized the entire patient-derived IN sequence. Considerable viral sequence evolution can occur during treatment failure and this may have contributed to the observed phenotype in the previous report.²²

Here, we show that the E157Q polymorphism has no significant effect on either IN strand transfer activity or DNA binding affinity of HIV-1 subtype B IN protein. However, as shown here, E157Q is able to partially restore deficits in IN enzymatic activity caused by the R263K substitution, thereby acting as a secondary, compensatory mutation. Again, this is consistent with the fact that E157Q can function as a compensatory mutation for the primary raltegravir/elvitegravir resistance mutation N155H.^13,23 Some reports have stated that E157Q is deleterious for viral replication without conferring much drug resistance, while others claim that the reverse is true.^11,21,24 We recently reported that E157Q had beneficial effects on some enzymatic activities of the IN protein, while also conferring moderate dolutegravir resistance when combined with N155H and R263K in biochemical assays.¹³ It is also worth noting that the Stanford Drug Resistance Database does not consider E157Q to be a major INSTI resistance determinant.⁷

There has yet to be a report of E157Q in combination with R263K in the clinic, despite high levels of dolutegravir resistance and viral replication capacity, as shown in the current study. Interestingly, the E157Q substitution in isolation rendered the virus hypersusceptible to dolutegravir. This may be because this residue is located near the active site of the protein, and while we have shown that it does not have a significant effect on enzymatic activity, E157Q may affect the binding of dolutegravir to this region. As we have shown E157Q to be hypersusceptible to dolutegravir, it follows that this substitution would not arise first in response to drug pressure, but this does not explain its absence as a secondary mutation after R263K. Furthermore, >5 years of dolutegravir selection in our laboratory with the R263K-containing virus has not yet led to any compensatory mutation for R263K (ongoing). This supports the hypothesis that R263K may represent an evolutionary dead-end pathway for the virus, from which it may be unable to escape. This may be due to deleterious effects of R263K on HIV RT activity, which may, in turn, influence the mutational capability of R263K-containing virus; investigations are currently under way to address this question. The fate of HIV-1 bearing E157Q in dolutegravir selections is also currently under evaluation. Of course, first-line dolutegravir resistance has so far been a rare occurrence and hence further monitoring of patients on dolutegravir therapy is necessary to indicate whether E157Q will ultimately be shown to have clinical relevance or not.²⁵

E157Q also fully compensated for the infectious deficit of R263K when the two substitutions were combined in NL4.3, while having no significant effect on its own. The E157Q/R263K double mutant also displayed enhanced dolutegravir resistance, with a calculated IC₅₀ of ∼1 μM compared with R263K alone, with an IC₅₀ of 106.1 nM, which represents a 20-fold decrease in susceptibility compared with WT. This finding may have implications for the treatment of HIV-positive individuals with dolutegravir, given that E157Q is naturally present in 1%–10% of untreated individuals, depending on subtype.²⁶ Thus, the presence of E157Q at baseline could potentially have a negative connotation. This is dependent on whether the hypersusceptibility of this mutant to dolutegravir would be recapitulated in vivo.

Glu157 was also recently identified as a putative interactor with a new class of HIV IN inhibitors that specifically target the DNA binding activity of the enzyme.¹⁴ Therefore, we also wanted to evaluate whether this polymorphic substitution could have effects on resistance to this class of inhibitors in vitro, as this would have implications for the further clinical development of compounds of this class. However, we found that E157Q had little effect on the DNA binding activity of purified recombinant IN, and this translated to a lack of resistance to FZ41 in tissue culture experiments.

In conclusion, this study highlights the importance of genetic background in viral evolution under drug pressure. We have shown the compatibility of the dolutegravir R263K mutation with the polymorphic substitution E157Q. This combination appeared replication competent and yielded a higher level of resistance to dolutegravir than either mutation on its own. The combination of the E157Q and R263K mutations should be monitored in clinical practice to determine whether they could provide a mechanism through which HIV might be able to escape dolutegravir pressure in the clinic.

Funding

This work was funded by grants from the Canadian Institutes of Health Research (CIHR). K. A. is the recipient of a doctoral research award from the Faculty of Medicine, McGill University.

Transparency declarations

None to declare.