Comparative biochemical analysis of HIV-1 subtype B and C integrase enzymes

Background Integrase inhibitors are currently being incorporated into highly active antiretroviral therapy (HAART). Due to high HIV variability, integrase inhibitor efficacy must be evaluated against a range of integrase enzymes from different subtypes. Methods This study compares the enzymatic activities of HIV-1 integrase from subtypes B and C as well as susceptibility to various integrase inhibitors in vitro. The catalytic activities of both enzymes were analyzed in regard to each of 3' processing and strand transfer activities both in the presence and absence of the integrase inhibitors raltegravir (RAL), elvitegravir (EVG), and MK-2048. Results Our results show that integrase function is similar with enzymes of either subtype and that the various integrase strand transfer inhibitors (INSTIs) that were employed possessed similar inhibitory activity against both enzymes. Conclusion This suggests that the use of integrase inhibitors against HIV-1 subtype C will result in comparable outcomes to those obtained against subtype B infections.


Background
Integration of viral cDNA into the host genome is one of the definitive features of retroviral replication. Integration is mediated by the HIV pol-encoded integrase enzyme. Recently, integrase inhibitors have been added to the arsenal of antiviral drugs used in therapy. RAL (Merck) was the first integrase inhibitor to be approved by the US Food and Drug Administration (FDA) after clinical trials demonstrated that this drug promoted a rapid and sustained antiretroviral effect [1]. EVG (GS-9137, Gilead), another integrase inhibitor, is currently in phase III clinical trials [2]. Other integrase inhibitors, such as MK-2048 (Merck), are still in pre-clinical development.
Integrase inhibitors are active against both B-and non-B subtypes in therapy [3,4]. Subtype C variants are responsible for approximately 50% of global infections, mostly in Sub-Saharan Africa and India [5]. It is therefore important to determine whether the integrase enzymes of different HIV-1 subtypes behave in a parallel manner to one another and whether they respond similarly to the use of integrase inhibitors of HIV-1 replication.
After viral entry and reverse transcription, reverse-transcribed double-stranded blunt-ended DNA is incorporated into the host cell genome through two catalytic activities mediated by integrase: 3' end processing and strand transfer [6,7]. During 3' end processing, a dinucleotide adjacent to the conserved 3' terminal CA is excised from the 3' end of the recently reverse transcribed HIV-1 DNA genome, generating 3' hydroxyl ends. During the strand transfer reaction, both newly generated 3' ends are covalently linked to target DNA in a concerted fashion via a one-step transesterification reaction [8]. In vitro, integrase can also catalyze two additional reactions: disintegration and specific internal endonucleolytic cleavage [9,10].
Variability between different HIV-1 integrases at an amino acid level is low, ≈ 8-12%. However, sites of amino acid differences between subtypes are often close to resistancerelated amino acids. We were therefore interested in analyzing whether such minor differences might be important in differential acquisition of INSTI resistance mutations in a subtype-specific manner [11]. Furthermore, natural polymorphisms in non-B integrase proteins might alter INSTI binding or activity [12,13]. An in silico comparison of subtype B and CRF A/G integrase predicted that polymorphisms within subtypes might affect structure and substrate binding characteristics of IN enzymes [13]. In this study, we compared the enzymatic activities of subtype B and C recombinant integrases in the context of inhibition by RAL, EVG, and the novel INSTI MK-2048.

Purification of active subtype C integrase
Subtype C integrase was PCR amplified from the pINDIE-C1 molecular clone and introduced into the expression vector pET-15B, replacing the ORF of subtype B integrase previously cloned by Bushman et al. [14]. To increase the solubility of subtype C recombinant proteins, two amino acid changes were introduced: a phenylalanine at codon 185 was changed to a histidine, and a cysteine at codon 280 was changed to a serine. These changes mimic those previously introduced into subtype B integrase to increase solubility and are known to not affect catalytic activity [15,16]. Expression and purification of the subtype B and C integrase enzymes were performed simultaneously as previously described for subtype B integrase [15] with minor modifications. Subtype B and C integrases were successfully purified to > 95% homogeneity ( Figure 1). The N-terminal His tag was removed from recombinant integrase enzymes by thrombin cleavage (Figure 1). When the enzymatic activities of both subtype B and C purified recombinant proteins in the presence or absence of the Nterminal His tag were compared, no difference was detected (data not shown). Therefore, all further experiments were orchestrated using recombinant integrase that did not undergo His tag removal.

Biochemical properties of subtype C integrase
Integrase mediates the insertion of viral cDNA into host chromatin through two unique enzymatic activities: 3' processing and strand transfer [6,17]. Oligonucleotides that mimic the viral LTR ends can be utilized to analyze these two catalytic activities in vitro. First, subtype B and C integrases were tested for their ability to perform 3' processing ( Figure 2) and strand transfer ( Figure 3). Time course experiments show similar results for both enzymes. Disintegration was also analyzed and subtype C recombinant protein catalyzed this activity to a similar extent as did subtype B recombinant protein ( Figure 4). These experiments confirm the activity of our subtype C purified recombinant protein.
Purification of recombinant subtype B and C integrase enzymes Figure 1 Purification of recombinant subtype B and C integrase enzymes. N-terminal His tags of the enzymes were removed from purified subtype B and C recombinant proteins by thrombin cleavage. Lane 1, protein ladder (10-250 kDa) (New England Biolabs); INB, subtype B integrase; INC, subtype C integrase.

Subtype B and C enzymes are inhibited to a similar extent by RAL, MK-2048 and EVG
RAL and EVG are INSTIs with high specific activity against strand transfer [18,19]. MK-2048 is a prototype secondgeneration INSTI with a resistance profile that is distinct from RAL and EVG [20,21]. These three drugs have been reported to be approximately 100-fold less specific for the inhibition of 3' processing activity compared to strand transfer [18,22,23].
Purified recombinant subtype B and C integrase enzymes were incubated with increasing concentrations of integrase inhibitors and corresponding templates. The results of Table 1 and Figures 5, 6 and 7 show that 3' processing mediated by recombinant enzymes of both subtypes was inhibited to a similar extent (p > 0.05) by all three drugs in the presence of MnCl 2 . The inhibition of 3' processing required much higher concentrations of integrase inhibitors than those needed to block strand transfer for both subtype enzymes (Table 1), consistent with previously reported data for subtype B integrase [18].
The strand transfer activity of subtype B and C recombinant proteins was inhibited by all three inhibitors. The IC 50 values of RAL for subtype B and C integrase strand transfer were 0.37 μM and 0.15 μM, respectively, in assays that employed Mn 2+ as the cation ( Figure 8, Table 1). The IC 50 values for EVG inhibition of strand transfer in Mn 2+based assays were 0.014 μM and 0.018 μM for the subtype B and C enzymes, respectively ( Figure 9, Table 1). The IC 50 values for MK-2048 against subtype B and C enzymes were 0.075 μM and 0.08 μM, respectively ( Figure 10, Table 1).
Disintegration was inhibited by high concentrations of MK-2048 to a comparable extent with both subtype B and C enzymes ( Figures 11,12,13). In contrast, neither RAL nor EVG had much effect on this process, which is a discovery that is consistent with work by others [22]. We also evaluated strand transfer in the presence of MgCl 2 rather than MnCl 2 and obtained similar IC 50 values (p > 0.05) Figure 2 3' Processing assay. One representative reaction (out of five reactions) is illustrated. Recombinant enzyme was incubated at 37°C with templates (radiolabeled double stranded oligonucleotide INT1/2) for the indicated times up to 120 minutes. The 21-mer substrate and 19-mer 3' processing products are indicated. Figure 3 Strand transfer assay. One representative reaction (out of five reactions) is depicted. Recombinant integrase enzyme was incubated at 37°C for 3 minutes for the initial 3' processing reaction. T35/SK70, double stranded oligonucleotide substrate, was added and reaction tubes were incubated at 37°C for the indicated times up to 120 minutes. The 21-mer substrate, 19-mer 3' processing, and strand transfer products are indicated.

Strand transfer assay
for each of subtype B versus C enzymes with each of RAL, EVG and MK-2048 in a microtiter plate system [24] (Table  2). Consistent with previous observations, IC 50 values were lower when these reactions were performed with MgCl 2 than with MnCl 2 [2,24].
Inhibition of replication by integrase inhibitors was also evaluated in cell culture based assays using cord blood mononuclear cells (Table 3). Subtype B and C clinical isolates were inhibited to a similar extent by each of RAL, EVG and MK-2048.

Discussion
Most HIV-1 patients are infected with non-B subtypes, most commonly subtype C [5], and subtype-specific differences in the development of drug resistance have been reported [25]. Therefore, it is important to understand the activity of enzymes of different subtypes. In our study, subtype B and C integrase enzymes were evaluated; and the activity of integrase inhibitors against them were compared, since a role for polymorphisms and structure-function differences between subtypes in drug resistance has been demonstrated [11,12,23].
Strand transfer inhibitors have been shown as efficient inhibitors of integration amongst a wide range of retroviruses [26]. In silico observations suggest that subtype-specific differences in regard to key amino acids in integrase, including those close to the catalytic site, may pose an effect on the binding of RAL [13,27,28]. Therefore, subtype-specific variations in DNA-binding domains could also affect the affinity of RAL for integrase. In vitro, subtype B and C recombinant proteins retain similar enzymatic capacities in the absence of drug (Figures 2, 3, 4), with comparable strand transfer, 3' processing and disintegration activities, as measured by time course experiments. We also show that RAL and EVG had similar effects against both subtype B and C integrase enzymes, regardless of whether Mg 2+ or Mn 2+ was used as a cation (Tables  1 and 2). In addition to the foregoing, we have evaluated the IC 50 values of RAL, EVG and MK-2048 in cell-based assays using clinical isolates of viruses of either subtype B or subtype C origin (Table 3). No significant differences were observed between subtypes in regard to drug susceptibility. These findings are consistent with recent results on similarities vis-à-vis biochemical activity and susceptibility to antiretroviral drugs of reverse transcriptase enzymes derived from HIV-1 subtypes B and C [29].

Conclusion
Our results provide biochemical and tissue culture evidence that integrase enzymes from HIV-1 subtypes B and C are inhibited by each of RAL, EVG and MK-2048 to a similar extent. These findings are supportive of the use of these inhibitors in patients infected with subtype C virus.  Inhibition of 3' processing as a function of increasing RAL concentration       Inhibition of disintegration as a function of increasing RAL concentration Figure 11 Inhibition of disintegration as a function of increasing RAL concentration. Subtype B and C disintegration activity (presented as relative percentage) in reaction to increasing RAL concentration. This graph was prepared with GraphPad Prism 4.0, the combined result of quantification and analyses of 3 independent experiments.
RAL and MK-2048 were obtained from Merck Pharmaceuticals, Inc. EVG was obtained from Gilead Biosciences.

Cloning and Mutagenesis
Subtype C integrase was PCR amplified from the molecular clone pINDIE-C1 (accession number: AB023804) and subcloned into the bacterial expression vector pET15B, replacing the subtype B integrase ORF kindly obtained from Dr. Robert Craigie, NIH.
The QuickChange II (Stratagene) site-directed mutagenesis kit was utilized for the introduction of the solubility mutations F185H and C280S into subtype C integrase.

Protein purification
Wild type and mutant His-tagged integrase proteins were expressed in Escherichia coli BL21(DE3) and purified under non-denaturing conditions. Bacterial cultures were grown at 37°C. When cultures of BL21 achieved an optical density of 0.5 at 600 nm, protein expression was induced by the addition of isopropyl-β-D-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. The cultures were incubated for 3 hours at 37°C, centrifuged (5000 rpm for 12 min), and frozen at -80°C. The pellets were resuspended in lysis buffer (20 mM Hepes pH 7.5, 100 mM NaCl, 2 mM β-ME, protease inhibitors) and lysed by sonication. The lysates were centrifuged (12 500 rpm for 30 min), the supernatants discarded, and the pellets resuspended in binding buffer (1 M NaCl, 20 mM imidazole, 20 mM Hepes pH 7.5, 2 mM β-ME, 100 μM ZnCl 2 , protease inhibitors). Following centrifugation at 12500 rpm for 30 min, the supernatants were incubated with nickel-nitrioltriacetic acid (Ni-NTA) agarose beads (Qiagen) for 1 hour at 4°C with mild agitation. Proteins were purified utilizing propylene columns (Qiagen). Histagged integrase protein was then eluted, utilizing a gradient of increasing imidazole concentration (0-2 M) in elution buffer (1 M NaCl, 20 mM Hepes pH7.5, 10% glycerol, 2 mM β-ME, 100 μM ZnCl 2 ). The eluates were Inhibition of disintegration as a function of increasing MK-2048 concentration Figure 12 Inhibition of disintegration as a function of increasing MK-2048 concentration. Subtype B and C disintegration activity (presented as relative percentage) in reaction to increasing MK-2048 concentration. This graph was prepared with GraphPad Prism 4.0, the combined result of quantification and analyses of 3 independent experiments.  analyzed on 10% SDS-polyacrylamide gels with Coomassie staining (Sigma-Aldrich). Proteins were dialyzed overnight at 4°C against 1 M NaCl, 200 mM Hepes pH 7.5, 100 μM ZnCl 2 , 10% glycerol and 2 mM DTT in dialysis cassettes (10,000 MWCO, ThermoScientific). The samples were aliquoted and fast frozen at -80°C. Protein concentration was measured by Bradford assay utilizing the Bradford Protein Assay kit (Bio-Rad Laboratories).

Thrombin His tag exclusion
The His tags were removed from purified recombinant subtype B and C integrases using the Thrombin Clean-Cleave kit (Sigma) according to manufacturers' instructions. Proteins were dialyzed overnight at 4°C against 1 M NaCl, 200 mM Hepes pH 7.5, 100 μM ZnCl 2 , 10% glycerol and 2 mM DTT in dialysis cassettes (10,000 MWCO, ThermoScientific).

3' processing, strand transfer and disintegration assays evaluated in urea gels
Oligonucleotide-based assays were performed to measure integrase enzymatic activities. All oligonucleotide probes were gel purified and phenol-chloroform extracted. INT1 oligonucleotide was radiolabeled using the T4 Polynucleotide Kinase kit (Ambion) with [γ-32 P] ATP (Perkin Elmer). Unincorporated nucleotides were discarded utilizing 'NucAway columns' (Ambion). A double stranded oligonucleotide substrate was obtained by mixing equal concentrations of INT1 and INT2, heating to 95°C and stepwise cooling to 37°C in 100 mM NaCl. The oligonucleotides T35 and SK70 were annealed to form a second double stranded oligonucleotide. INT1/2 mimicked the HIV-1 U5 long terminal repeat (LTR) and acted as a substrate for 3' processing. T35/SK70 acted as a site for inte-gration of the processed INT1/2 fragment and mimicked host DNA.
Concentrations of enzymes used in the following reactions were optimized in a series of preliminary experiments. Integrase reactions were performed in a buffer containing 20 mM Hepes (pH 7.5), 30 mM NaCl, 1 mM dithiothreitol, 125 μM ZnCl 2 and 0.125 pmol dsDNA substrate (INT1/2) in a final volume of 10 μl. Recombinant integrase (3.1 μM), and varying concentrations of integrase inhibitors or water/DMSO as control, were mixed and preincubated at 37°C for 15 min. 3' processing reactions were initiated by addition of 7.5 mM of MnCl 2 and incubated at 37°C for 3 min unless otherwise indicated. For strand transfer reactions, 1.25 pmol of dsDNA template (T35/SK70) were added and the reaction mixture was further incubated at 37°C for 1 hour unless otherwise indicated. Reactions were stopped by adding 5 times the volume of gel loading dye (formamide containing 1% SDS, 0.25% bromophenol blue, and 0.25% xylene cyanol) and heating to 95°C. Reaction products were separated on 6% acrylamide, 7 M urea sequencing gels. Gels were dried and exposed utilizing phosphorimager screens (GE Healthcare) and scanned in a Molecular Dynamics Typhoon Phosphorimager (GE Healthcare). Product analysis and quantification were conducted using Image-Quant and GraphPad Prism 4.0 software. Quantification of standard error of the mean (SEM) was performed with GraphPad Prism 4.0 software.

The use of microtiter plates for strand transfer analysis
A microtiter plate assay was utilized to evaluate MgCl 2mediated strand transfer as previously described [24,31]. Briefly, biotinylated oligonucleotides mimicking LTR donor DNA (5'-biotin-ACCCTTTTAGTCAGTGT-GGAAAATCTCTAGCAGT and 5'-ACTGCTAGAGATTTTC-CACACTGACTAAAAG) were immobilized onto blackcolour Reacti-Bind Streptavidin-coated plates (Ther-moFisher). 313 nM recombinant enzyme was bound to donor DNA on the plates in the presence of 25 mM MnCl 2 and the plates were then washed to remove excess unbound enzyme. 3'-FITC labelled dsDNA (5'-TGAC-CAAGGGCTAATTCACT-FITC-3' and 3'-FITC-ACTGGTTC-CCGATTAAGTGA-5'), used as a reaction target, was added to the wells in 25 mM Hepes (pH 7.8), 25 mM NaCl, 2.5 mM MgCl 2 , and 50 μg/mL BSA [24,32] and the plate was incubated at 37°C for 1 hour. Covalently linked target DNA was detected through use of an anti-FITC antibody conjugated to alkaline phosphatase (Roche) and a chemiluminescence substrate (CSPD Sapphire II, Applied Biosystems). Integrase inhibitors were added at increasing concentrations shortly before the addition of target DNA. Strand transfer was evaluated by chemiluminescence.

Statistical Analysis
Unpaired two-tailed t-tests were used to examine statistical significance in subtype B versus subtype C integrase enzymatic assays using GraphPad Prism 4.0 software.

Determination of activity of integrase inhibitors in cell culture
Recombinant viruses (subtype B) (pNL4-3 and IIIb) and viruses obtained from either our primary HIV infection cohort or from long-term infected patients (subtype B and C) were amplified as previously described [33,34]. Drug susceptibility was measured in cell culture-based phenotypic assays using cord blood mononuclear cells to determine the extent of in vivo HIV replication blockage by integrase inhibitors. 50% drug effective concentrations (EC50s) were determined for each of RAL, EVG and MK-2048 by monitoring the production of p24 antigen, as previously elaborated [33].