Logo of jbcAbout JBCASBMBSubmissionsSubscriptionsContactJBCThis Article
J Biol Chem. 2008 Aug 29; 283(35): 23599–23609.
PMCID: PMC3259799

Biochemical Analysis of HIV-1 Integrase Variants Resistant to Strand Transfer Inhibitors*

Abstract

In this study, eight different HIV-1 integrase proteins containing mutations observed in strand transfer inhibitor-resistant viruses were expressed, purified, and used for detailed enzymatic analyses. All the variants examined were impaired for strand transfer activity compared with the wild type enzyme, with relative catalytic efficiencies (kp/Km) ranging from 0.6 to 50% of wild type. The origin of the reduced strand transfer efficiencies of the variant enzymes was predominantly because of poorer catalytic turnover (kp) values. However, smaller second-order effects were caused by up to 4-fold increases in Km values for target DNA utilization in some of the variants. All the variants were less efficient than the wild type enzyme in assembling on the viral long terminal repeat, as each variant required more protein than wild type to attain maximal activity. In addition, the variant integrases displayed up to 8-fold reductions in their catalytic efficiencies for 3′-processing. The Q148R variant was the most defective enzyme. The molecular basis for resistance of these enzymes was shown to be due to lower affinity binding of the strand transfer inhibitor to the integrase complex, a consequence of faster dissociation rates. In the case of the Q148R variant, the origin of reduced compound affinity lies in alterations to the active site that reduce the binding of a catalytically essential magnesium ion. Finally, except for T66I, variant viruses harboring the resistance-inducing substitutions were defective for viral integration.

HIV-12 is a retrovirus with a single-stranded RNA genome (1). After infection, the viral reverse transcriptase synthesizes a DNA copy (vDNA) of the viral RNA. HIV integrase is responsible for inserting the viral DNA into the host genome, an event that includes a number of biochemically discrete steps (25). Integrase acts in a multimeric form (6), binding to ends of the long terminal repeats (LTR) of viral DNA and, although still in the cytoplasm, cleaving two bases from each 3′ terminus (4, 7). In the nucleus, HIV integrase mediates an internal cleavage of the host genome followed by a so-called strand transfer reaction, resulting in a joining of the 3′ ends of viral DNA to the 5′ ends of host DNA (8). This generates a 5-bp duplication of the host DNA at the site of integration (9, 10), which is thought to be filled in by the host DNA repair system. Integration of the vDNA allows the viral genes to be expressed, leading to the assembly of new virions that are processed into infectious particles.

HIV integrase strand transfer inhibitors (STIs), which selectively target the strand transfer activity of integrase, possess potent anti-HIV activity in cell culture (1114), and the first one, raltegravir (15), has been approved for use in treatment experienced patients. Replication of HIV in the presence of increasing concentrations of STIs selects for resistant viruses, with amino acid changes within the integrase coding region (11, 16). In this study we report detailed in vitro biochemical analyses of purified integrase enzymes containing substitutions resulting in decreased sensitivity to STIs. These include the reported previously substitutions T66I, V75I, E92Q, Q148R, and N155H and M154I (1722). We have also examined combinations of these changes (T66I/E92Q, T66I/N155H, and V75I/M154I). Compared with the wild type (WT) enzyme, all variant enzymes were impaired with respect to assembly and multimerization into active complexes and their subsequent in vitro strand transfer abilities. Reduced susceptibility to one STI, STI 1, (23), was shown to be a result of decreased affinity of the inhibitor for the integrase enzymes containing the selected mutations. In addition, isogenic viruses harboring these resistance mutations showed defects in integration in cells.

EXPERIMENTAL PROCEDURES

Compounds—STI 1 was described previously (23). STI 2 is a diketo acid prepared at Bristol-Myers Squibb.

Purification of HIV-1 Integrase Proteins—Wild type and variant integrase enzymes were expressed in Escherichia coli and purified for biochemical analyses. The variant enzymes used in this study were T66I, V75I, E92Q, Q148R, N155H, T66I/E92Q, T66I/N155H, and V75I/M154I, reflective of their amino acid substitutions compared with the WT NL4-3 enzyme. Variant enzymes were constructed through site-specific mutagenesis of the NL4-3 wild type integrase enzyme using the QuickChange® site-directed mutagenesis kit, according to manufacturer's recommendations (Stratagene, City). The complete sequence of the integrase enzyme was obtained to confirm the construction of each variant enzyme. HIV-1 integrases were purified as reported previously (23). Purities were greater than 95%, as indicated by SDS-polyacrylamide gels.

3-Processing Activity of HIV Integrase—Radiolabeled double-stranded HIV LTR substrate 5′-pACCCTTTTAGTCAGTGTGGAAAATCTCTAGCA[32P]GT and 3′-GAAAATCAGTCACACCTTTTAGAGATCGTCA was synthesized by a fill-in reaction of 5′-pACCCTTTTAGTCAGTGTGGAAAATCTCTAGCA and 3′-GAAAATCAGTCACACCTTTTAGAGATCGTCA using Klenow DNA polymerase and [32P]dGTP and dTTP. Unlabeled carrier DNA substrate was also generated by a fill-in reaction with Klenow using unlabeled dGTP and dTTP. Initial cleavage assays were performed by titrating integrase concentrations from 100 to 800 nm, to determine the proper concentration and time course for kinetic studies. The cleavage reactions were performed by preincubating integrase variants (100–400 nm) with 5–320 nm [32P]DNA at room temperature in 20 mm HEPES, pH 7.5, 30 mm NaCl, and 10 mm dithiothreitol for 15 min. This preincubation step was performed because variable extents of reaction were observed (especially at short time periods), as noted in a previous kinetics analysis (24). Reactions were initiated by adding MgCl2 to a final concentration of 7.5 mm and incubated at 37 °C over 30 min for wild type integrase and 30–90 min for variant integrases. An equal volume of formamide stop solution (U. S. Biochemical Corp.) was added, and samples were heated to 95 °C for 90 s and quick-chilled in ice water before loading onto 20% acrylamide/urea gels. After electrophoresis, gels were exposed to phosphorescent screens (GE Healthcare) and scanned in a PhosphorImager (GE Healthcare). Under the reaction conditions used, minimal strand transfer occurred (less than 5% of the cleaved substrate). Scans were quantitated for substrate and the –2 nucleotide cleavage product with ImageQuant software (GE Healthcare) and exported to Microsoft Excel for calculation of reaction rates. Curve fitting was performed with GraphPad Prism version 4.03 to generate Km and maximal velocity values (Vmax).

Scintillation Proximity Assay (SPA) for Determination of Kinetics of Strand Transfer and Binding of Radiolabeled Inhibitor—An SPA format was used to measure the strand transfer kinetics of the integrase proteins. The details of the assay have been reported previously (23). Briefly, viral LTR DNA is attached, via a biotin linker, to streptavidin-coated SPA PVT beads, after which the enzyme is added to form active complexes. For kinetic studies, 33P-labeled target DNA was added, which gets integrated into the LTR DNA. The amount of integrase enzymes used to prepare each of the different complexes was determined by titrating variant enzymes onto the SPA/DNA beads and noting the minimal concentration of enzyme required to produce the maximal amount of product (saturation), using 0.92 nm 33P-labeled target DNA. An amount equal to 1.1 times this concentration was used for subsequent kinetic studies. For binding studies [3H]STI 1, rather than target DNA, was added. The extent of radioligand binding was measured on a topcount instrument (PerkinElmer Life Sciences), and affinity constants toward the wild type and variant integrase enzymes were calculated as described previously (23). For compound sensitivity studies, variant enzyme concentrations were adjusted so that total product was approximately the same as that derived from using WT enzyme. The extent of integration is measured though SPA analysis, and the degree of inhibition is determined compared with non-compound containing controls. For STI 1 binding kinetics, data points were acquired every 30 s and the data fit to a one-site binding model (GraphPad Prism 4.01).

Determination of kcat, kp, and Km from Strand Transfer Kinetics—Reactions were carried out in 96-well white microtiter dishes (catalog number 3600, Corning Glass) containing 40 μl of complexes. Reactions were initiated by rapid addition, with shaking, of 10 μl of 33P-labeled target DNA (final concentration 0.23–29.4 nm) in SPA buffer using dilutions of 33P-labeled target DNA (original specific activity between 0.8 to 1.9 × 106/cpm per μl at a concentration of 600 nm) mixed with unlabeled DNA to achieve the final concentrations. Incubations proceeded for various times at 24 °C and were stopped by addition of 200 μl of phosphate-buffered saline, 50 mm EDTA. Products were allowed to sit at 24 °C for at least 1 h prior to centrifugation at 2000 rpm or overnight without centrifugation. The reaction products were quantitated using a Topcount scintillation counter, and the counts/min readout were used to calculate molar concentrations of product.

Kinetic curves (product versus time) were prepared, and the data in the linear range were used to calculate slopes for given concentrations of 33P-labeled target DNA. Slopes versus target concentrations were used to calculate Vmax and Km values by fitting to the Michaelis-Menten equation using GraphPad Prism, version 4.01 (GraphPad Software, Inc.). The kcat values (kcat = Vmax/integrase]) were calculated from Vmax and ¼ the concentration of integrases used to prepare the complexes (i.e. integrase was treated as a tetramer) (6, 2527). The kp values (Vmax/Eactive]) were calculated from the actual concentration of active complexes (Eactive values), which were determined experimentally. Eactive is defined as the maximal amount of product that could be formed under conditions of saturating amounts of target DNA and represents the conversion of all active complexes to product complexes, assuming a single turnover of enzyme (23, 24). All data points in individual experiments were obtained in duplicate, and the kinetic constants from three independent experiments were averaged, and standard deviations were calculated.

Inhibitor Sensitivity Studies—Variant and WT integrases were evaluated for STI 1 sensitivities using the SPA assay (23), with minor modifications as follows. The amount of integrase used to prepare the complexes was adjusted so that approximately equal amounts of products were formed after 2 h at 37 °C; integrase complexes were not diluted with ⅓ volume of 25% DMSO prior to use (30 μl of undiluted complexes were used rather than 40 μl of diluted complexes); 5-fold serial dilutions of STI 1 (in 25% v/v DMSO/H2O) were added to integrase complexes. The inhibitor-enzyme complexes were formed by incubating the complexes at 37 °C for 10 min prior to initiating strand transfer reactions by adding 1 × 106 cpm 33P-labeled target DNA in SPA buffer (the target DNA was added at room temperature). The plates were then returned to 37 °C. After 2 h, strand transfer reactions were stopped by adding 200 μl of phosphate-buffered saline, 50 mm EDTA. Plates were allowed to settle for at least 2 h and then read on a Topcount instrument.

PCR Analysis of Integrated and 2-LTR Viral DNA—In the DNA analysis experiments, MT-2 cells were infected with recombinant replicating viruses through spinoculation (28). Briefly, cell pellets were resuspended in a volume of virus corresponding to a multiplicity of infection of 0.1 TCID50/cell, and media were added to a final volume of 2 ml. The cells and virus were centrifuged at 1200 × g for 1 h. The supernatant was removed, and the cells were resuspended at 1 × 105 cells/ml. At different time points post-infection, 1 × 105 cells were removed and lysed at 65 °C in 100 μl of buffer (100 mm KCl, 10 mm Tris-HCl, pH 8.3, and 2.5 mm MgCl2) containing 1% Tween 20, 0.4 mg/ml proteinase K, and 1% Nonidet P-40. Proteinase K was subsequently inactivated by heating the samples to 95 °C for 15 min, and the various forms of total, circular, and integrated viral DNA were quantified, as described below (29). First, the total copy number of all the forms of HIV DNA (integrated, linear and circular) was determined using the primer/probe set (forward primer, 5′-TGTGTGCCCGTCTGTTGTGT, reverse primer, 5′-GAGTCCTGCGTCGAGAGAGC, and probe, 5′-FAM-CAGTGGCGCCCGAACAGGGA-TAMRA). These primers are able to amplify a fragment present within each LTR region. NL4-3 proviral DNA of known concentration was serially diluted over a million-fold to generate a standard curve for determining the total HIV DNA copy number. Second, the amount of integrated HIV DNA was quantitated by first pre-amplifying (in a PCR) isolated genomic DNA with one primer homologous to genomic Alu sequences and a second primer homologous to an HIV Gag sequence (forward primer, GCCTCCCAAAGTGCTGGGATTACAG, and reverse primer, GCTCTCGCACCCATCTCTCTCC). Next, the Alu-Gag products were re-amplified using a real time PCR on a Taqman instrument using LTR primers (forward, GCCTCAATAAAGCTTGCCTTGA, reverse, TCCACACTGACTAAAAGGGTCTGA, and probe, 5′-FAM-GCGAGTGCCCGTCTGTTGTGTGACTCTGGTAACTAGCTCGC-BHQ-3′). These primers amplify a fragment within the LTR region. Amplified Alu-Gag amplicons were converted to DNA copy numbers by reference to a standard curve generated on the same plate using the same primers acting on serial dilutions of NL4-3 proviral DNA of known concentrations. The fractional amounts of Alu-Gag copy numbers were then obtained by dividing them by the total DNA copy numbers. The fractional ratios were then normalized to the maximal ratio = 100, to enable comparison between samples. The 2-LTR circular DNA was quantitated as described elsewhere (30). Briefly, to generate the standard DNA for 2-LTR analysis, a 324-bp PCR product representing the junction of the 2-LTR circles was cloned. Standard curves for the amount of 2-LTR circles present in the virus infections were prepared based upon this DNA construct. The copy number for the circular DNA present in the proteinase K-digested cell lysates was then determined by a real time PCR amplification (forward primer, 5′-AACTAGGGAACCCACTGCTTAAG, reverse primer, 5′-TCCACAGATCAAGGATATCTTGTC, and probe, 5′-FAM-ACACTACTTGAAGCACTCAAGGCAAGCTTT-TAMRA). Copy numbers were then normalized using the standard curve.

To compare the relative extents of integration by STI-resistant variant viruses, the relative amounts of integrated HIV DNA were first plotted as a function of time for samples harvested between 0 and 36 h post-infection. Next, the area under the curve was calculated for each variant, and the areas were normalized as a percent of WT control. The multiplicity of infection in these experiments was 0.1. In addition, 5 μm ritonavir and amprenavir were added after infection, so the assumption in the calculation is that the extent of replication represents ∼1 replication cycle.

HIV Assays—MT-2 cells and 293T cells were obtained from the National Institutes of Health AIDS Research and Reference Reagent Program and propagated in MT-2 cells. MT-2 cells were propagated in RPMI 1640 media supplemented with 10% heat-inactivated fetal bovine serum, 10 mm HEPES buffer, pH 7.55, 2 mm l-glutamine, 100 units/ml penicillin G, 100 units/ml streptomycin, and 0.25 μg/ml amphotericin B. The 293T cells were propagated in Dulbecco's modified Eagle's media supplemented with 10% heat-inactivated fetal bovine serum, 10 mm HEPES buffer, pH 7.55, 2 mm l-glutamine, 100 units/ml penicillin G, 100 units/ml streptomycin, and 0.25 μg/ml amphotericin B. The proviral DNA clone of NL4-3 was obtained from the National Institutes of Health AIDS Research and Reference Reagent Program. Recombinant NL-Rluc virus, in which a section of the nef gene from NL4-3 was replaced with the Renilla luciferase gene, was used as a reference virus (31). Briefly, the NL-RLuc virus was prepared by co-transfection of two plasmids, pNLRLuc, containing the NL-RLuc DNA, and pVSVenv, a plasmid containing the vesicular stomatitis virus envelope gene expressed via an LTR promoter (32). Transfections were performed at a 1:3 ratio of pNLRLuc to pVSVenv in 293T cells using Lipofectamine PLUS from Invitrogen, according to the manufacturer's instruction. The pseudotype virus was titered in MT-2 cells using luciferase enzyme activity as a marker. Luciferase was quantitated using the dual luciferase kit from Promega (Madison, WI). Antiviral activities of STI 1 toward the recombinant viruses were quantified by measuring luciferase activity in cells infected with NLRluc recombinants in the presence of the inhibitor. STIs 1 and 2 (23) were prepared by the Medicinal Chemistry group at Bristol-Myers Squibb.

RESULTS

Inhibition of Strand Transfer Activity by the Integrase Variants Correlates with STI 1 Binding—[3H]STI 1 (23) was used in binding assays with the eight variant enzymes (Fig. 1, A and B). The data were fit to a simple one-site binding model. As summarized in Table 1 and Fig. 7, the variant enzymes generally had binding affinities for [3H]STI 1 which correlated with their IC50 values for inhibition of integrase activity by nonradiolabeled STI 1. For instance, the Kd for V75I is 68 ± 13 nm, whereas the IC50 for STI 1 against the enzyme is 85 nm, and the Kd for N155H is 289 ± 76 nm, whereas the IC50 for STI 1 against the enzyme is 334 nm, etc. This correlation holds for 6 of the 8 enzymes examined. However, although the Kd values for E92Q and T66I were elevated versus WT (46 and 61 versus 21 nm), the magnitude of these increased Kd values was not as large as the increases in the IC50 values for inhibition of strand transfer versus WT (145 and 183 versus 20 nm, respectively). These small differences continue to be analyzed. Overall, the binding affinity of [3H]STI 1 to the most resistant integrase enzymes was also the poorest (Q148R, N155H, and the double variants). Additionally, isogenic recombinant viruses containing these resistance substitutions were also less sensitive to inhibition by STI 1 (Table 1), and these reduced sensitivities parallel the data for both inhibition of strand transfer in vitro and for the [3H]STI 1 binding affinities. Together, the data indicate that resistance of these various enzymes to inhibition by STI 1 is directly related to the binding affinities of these variant enzymes for the inhibitor. One consequence of the reduced inhibitor sensitivities by the mutants is that inhibitor concentrations that are sufficient to abolish 90% of WT activity (a concentration of STI 1 of ∼120 nm) have little effect on the in vitro integrase activity of such highly resistant enzymes as Q148R (whose in vitro IC50 for STI 1 inhibition is ∼10-fold higher (1.1 μm)).

TABLE 1
Inhibition of ST and of isogenic viruses by STI 1 compared with binding constants for [3H]STI 1 toward variant integrases
FIGURE 1.
Representative binding curves for [3H]STI 1 to WT and variant integrases. Inhibitor-bound complexes (WT and variant enzymes) were formed by adding increasing amounts of radiolabeled STI 1, and the amount of bound compound was measured. A, data for ...
FIGURE 7.
Binding affinities (Kd) for STI 1 toward the variant integrases compared with the kinetically determined Km values for target DNA in strand transfer reactions.

Reduced Inhibitor Affinity for Resistant Integrases Is Because of Faster Dissociation—Off rates for binding of [3H]STI 1 at 20 °C were measured by the loss of binding of [3H]STI 1 upon the addition of a large excess of unlabeled STI 1. Examination of the off rates to the resistant integrases (Table 1, 6th column) shows that decreases in binding affinity are mainly because of increases in the off rates. The off rates of [3H]STI 1 to the Q148R, N155H, and T66I enzymes increased 14-, 12-, and 9-fold, respectively.

Variant Enzymes Display a Defect in Assembly—Studies were performed to assess whether assembly of the active multimer was affected. Assembly efficiencies of the integrases were determined by adding increasing amounts of integrase proteins to a viral LTR sequence that was immobilized on scintillation proximity beads. Radiolabeled target DNA was added, and the amounts of products generated were measured after 2 h at 37 °C. The results are presented in Fig. 2, A and B, and Tables Tables22 and and7.7. The concentrations of enzymes required to give 50% of the maximal signal (Max50) varied from 0.21 (WT) to 0.72 (T66I). Hill slopes, an indication of the cooperativity in assembly (33), varied from 6.1 (WT) to 2.4 (E92Q/T66I), suggesting improper assembly of the latter enzyme complex. All eight variant enzymes required higher enzyme concentrations to maximally carry out the strand transfer reaction.

TABLE 2
Amount of integrases used in the drug sensitivity and kinetic studies
TABLE 7
Summary of relative fold increases in defects in in vitro and cell culture characteristics of the variant integrases versus WT control
FIGURE 2.
Strand transfer reaction profiles for WT and variant integrases. The product of in vitro strand transfer reactions of WT and variant integrases were determined as a function of the concentration of purified integrase enzyme. The product is displayed ...

Resistant Integrase Enzymes Have Reduced Strand Transfer Activities—The eight variant enzymes were next examined for their catalytic activities in strand transfer reactions. Enzymes were compared by standardization of an SPA format using defined amounts of each component. A fixed amount of SPA beads containing a fixed amount of pre-bound viral (donor) DNA was used in all experiments. The various integrase variants were initially titrated onto the complexes prior to the addition of the 33P-labeled target DNA, and the minimal amount of enzyme required to produce the maximal amount of product (performed at 37 °C for 2 h) was determined. An amount equal to 1.1 times this saturating amount of enzyme was used in all subsequent experiments. Once this enzyme concentration was determined, experiments were performed using 2-fold serial dilutions of target DNA, beginning at 29 nm and diluting down to 0.23 nm.

Fig. 3 shows representative initial progress curves for the wild type, T66I, Q148R, and E92Q enzymes as a function of time for each target DNA concentration (in nm). Except for T66I, which exhibited a short lag phase, the other enzymes produced products linearly. The cause of the lag phase is unknown but is reproducible. As expected, the reaction velocities are a function of target DNA concentrations, and the wild type enzyme exhibits the greatest catalytic activity of all the integrases. This was confirmed by determining the maximal velocities, Vmax, for each enzyme, by constructing Michaelis-Menten plots (Fig. 4, A and B), using the slopes of the activity curves shown in Fig. 3. Fig. 4A illustrates the slower enzymes (N155H, Q148R, T66I, E92Q, T66I/N155H, and T66I/E92Q), whereas the plots of WT and the less defective enzymes (V75I and V75I/M154I) are shown in Fig. 4B.

FIGURE 3.
Progress versus time curves for strand transfer catalyzed by WT and variant integrases. Strand transfer reactions using NL4-3 WT and variant enzymes were initiated on active integrase complexes by adding increasing concentrations of 33P-labeled target ...
FIGURE 4.
Representative Michaelis-Menten plots (velocity of reactionversus concentration of target DNA) for strand transfer catalyzed by WT and variant integrases. A, enzymes with reaction velocities significantly slower than that of WT enzyme: Q148R, ♦; ...

Values for the kcat values for strand transfer (Vmax, divided by the enzyme concentration) are shown in Table 3 (4th column). For easier comparison, relative kcat values (normalized to WT = 100) are also shown (Table 3, 6th column). Most of the resistance variants are substantially impaired for catalytic activity (kcat) in vitro, with relative catalytic activity ranging from 52% (V75I) to as little as 0.1% (T66I/E92Q). The Michaelis-Menten analysis also provided Km values, the concentrations of target which produced half-maximal velocities. The absolute values of these Km values varied from a low of 7.1 to a high of 58 nm, with the WT enzyme exhibiting a Km of 14 nm (Table 3, 8th column).

TABLE 3
Calculated kcat, Km, and Vmax values for strand transfer activities of variant integrase enzymes

The preceding analysis of the reaction rates of the various enzymes incorporates two main variables in the in vitro reaction: complex assembly and catalytic activity. As suggested by the data described in Fig. 2, A and B, and Tables Tables22 and and7,7, all eight integrase variants exhibited various degrees of defects in complex assembly compared with the wild type enzyme, requiring that more enzyme be added (relative to WT) to achieve maximal product formation. To remove this “assembly factor” from the analysis, kcat values were recalculated by dividing Vmax values by the actual in situ concentrations of active integrase complexes that formed (Eactive), rather than the total amount of integrase added into the reaction (Etotal) (Table 3, 3rd and 2nd columns, respectively). These re-calculated kcat values are referred to as kp values (p = product). The Eactive values are the maximal concentrations of strand transfer products that could be formed under saturating concentrations of target, and are assumed to be the amounts of active integrase complexes present, because integrase operates only once, i.e. is a single turnover enzyme, as was recently concluded from a detailed study of the 3′-processing (24) and strand transfer (23) functions. This methodology ignores any multimerization or assembly defects, because only the pre-formed active complexes are being evaluated (which is why this method is often used in the analysis of single turnover enzyme systems) (34). The Eactive concentrations (Table 3, 3rd column) for most of the enzymes are <2.5 nm, which are 20–440-fold lower than Etotal values used to prepare the complexes (Table 3, 2nd column).

A comparison of the kp values relative to WT (kp relative, Table 3, 7th column) shows that all the variants are catalytically less active than WT (0.8–71%). The T66I double variants, T66I/N155H and T66I/E92Q, are the least catalytically active enzymes, with relative kp values of 6.7 and 0.8, respectively. The least defective variant was V75I, which exhibited a relative kp of 71% and exhibited only 1.8-fold resistance to STI 1 in the strand transfer assay (Table 1). Another interesting enzyme is the Q148R variant, which demonstrated relatively high levels of catalytic activity, with a relative kp value of 63% of wild type. Interestingly, the Q148R mutation induces one of the highest levels of resistance to STI 1 in the strand transfer assay (87-fold) and in cell culture (48-fold) (Table 1). Thus, besides displaying varying defects in assembly, the variants are all intrinsically impaired for catalytic activity.

To draw overall conclusions from the kinetic data, the Km values of the integrases can be compared. As noted, the Km value is equal to the concentration of the target, which gives 50% of the maximal activity, and is roughly akin to the efficiency with which the enzyme uses the target (lower Km = more efficient, higher Km = less efficient). The Km value for the NL4-3 wild type integrase was 14.2 ± 1.3 nm (Table 3, 8th column). Interestingly, several variants exhibited a lower Km value than the wild type enzyme. Examples of this are T66I (8.4 ± 3.0 nm), E92Q (7.1 ± 4.0 nm), and the V75I/M154I double variant (9.4 ± 0.2 nm). The V75I variant had a Km value similar to the wild type enzyme (20 ± 7), whereas several variants have Km values higher than the wild type enzyme. These are the N155H (29 ± 12 nm), T66I/N155H (36 ± 13 nm), and Q148R (58 ± 5nm) enzymes.

3-Processing Activities of Variant Integrases—Five of the integrase variants were analyzed for the effect of these amino acid changes on their 3′-processing activities. T66I integrase showed a 2-fold higher Km value than wild type, indicating that the LTR DNA substrate still bound reasonably well to this enzyme (Table 4). The Km values for the V75I, E92Q, Q148R, and N155H were elevated less than 2-fold, suggesting that binding of the LTR substrate was not significantly altered by these substitutions. The reduced cleavage observed for these altered integrase enzymes was found to be due to a reduction in kcat (kcat values were calculated from Vmax values, which had been divided by one-quarter of the total integrase concentrations, using the assumption of a tetramer as the active integrase species). Except for V75I, all the variant integrases displayed 2–6-fold reductions in their kcat values for 3′-processing. The Q148R variant had the greatest defect, with the catalytic turnover reduced 6-fold versus wild type. The V75I integrase was the most cleavage competent enzyme, with only a 15% reduction in catalytic turnover compared with the wild type enzyme.

TABLE 4
Kinetics of 3′-cleavage by integrase enzymes

Q148R Requires Higher Concentrations of Magnesium to Efficiently Bind Inhibitors—There is a loosely bound Mg2+ ion in the active site of integrase that is required for inhibitor binding (23). We previously determined that Q148R resistance to STI 1 was due, at least in part, to a requirement for an elevated Mg2+ concentration compared with WT (23). Here, we extend the data to T66I. As shown in Table 5, T66I and N155H have similar Kd values (4.8–5.2 mm) for Mg2+ in the presence of [3H]STI 1. This implies that the environment of the Mg2+ in the active sites of these enzymes, with respect to STI 1 binding, is not significantly altered versus the WT enzyme. However, the Kd value for Mg2+ in the binding of this compound to the Q148R variant was shifted to a >20-fold higher concentration. This implies that the enzyme binds this Mg2+ less well than WT, and might at least partly explain its resistance to STIs. To confirm this, the IC50 for inhibition by STI 2, a diketo acid, was examined as a function of Mg2+ concentration, using WT and T66I enzymes as controls. As shown in Fig. 5, the relative potency of STI 2 increased against all three enzymes as the Mg2+ concentration was raised from 3.1 to 100 mm Mg2+, but the increase was 7-fold greater for Q148R. The minor increase in potency at 100 mm magnesium for T66I was similar to that of the WT enzyme. This result shows that compound binding to Q148R is much more sensitive to changes in magnesium ion concentration versus the WT and T66I enzymes, and it may help explain why the Q148R-containing enzyme does not bind inhibitors as avidly as WT enzyme.

TABLE 5
Kd values for Mg+2 in the binding of [3H]STI 1 to WT and variant integrases
FIGURE 5.
Comparison of the fold decrease in the IC50 for inhibition of strand transfer as a function of the Mg2+ concentration (3.1–100 mm) by STI 2. Complexes pre-formed in the presence of Mg2+ were pelleted and resuspended in buffer lacking Mg2+ ...

DNA Analysis of STI-resistant Variants in Infected Cells—To probe the fitness of the STI-resistant viruses in infected cells, QPCR was used to measure the extent of viral DNA integration into chromosomal DNA (29). Different QPCR assays were designed to evaluate total levels of HIV-1 DNA present within a cell, the amount of HIV-1 DNA integrated into the chromosomal DNA, and the amount present as 2-LTR circles. It has been reported that integrase inhibitors cause a decrease in integrated DNA, with a concomitant increase in the levels of circular forms of viral DNA in the nucleus (known as 2-LTR circles) (11, 35). These 2-LTR circles are thought to be a consequence of abortive integration.

The kinetics of the appearance of the three kinds of viral DNA in WT, T66I, and Q148R viruses were analyzed. MT-2 cells were infected by spinoculation (28), and samples were analyzed for integration and 2-LTR circles. Maximal integration (Fig. 6A) was reached around 18 h post-infection for all three viruses, whereas 2-LTR circle formation (Fig. 6B) began to level off after ∼30 h. The T66I virus behaved similarly to WT, indicating no defect in the kinetics of integration for this virus. However, by 18 h post-infection, Q148R peak integration was reduced by 2.5-fold, while producing 3-fold more LTR circles. By 36 h, both integrated and total DNA decreased, most likely due to some cellular degradation. These results demonstrate that the Q148R substitution directly reduces the rate and extent of viral integration by ∼3-fold. The greater defect in Q148R versus T66I might be due to its lower overall catalytic strand transfer efficiency (15 versus 36% of WT, respectively), because 3′-processing activities were similarly reduced (12 versus 13% of WT, respectively). The same QPCR method was used to semiquantitatively compare the integration capacity of other resistant recombinant viruses (Table 6). Several of the variants display a reduction in integration by factors of 0.7–0.5-fold per cycle (E92Q, N155H, and T66I/E92Q), whereas Q148R shows the greatest defect (0.3-fold of WT). These variants also exhibit the greatest increase in 2-LTR circle formation (1.2 to 3.5-fold WT; Table 6). Overall, the cellular integration data indicate that most of these integrase variants have reduced integrative capacity during infection. This reduced integrative capacity translates into poorer replicative capacity (data not shown; studies to be reported elsewhere), as has also been reported for these and other IN-resistant viruses (36, 37).

TABLE 6
Extent of integration and 2-LTR formation in variant-infected cells
FIGURE 6.
Kinetics of viral integration (A) and 2-LTR formation (B) in WT (filled circles), T66I (squares), and Q148R (+) virus-infected cells. DNA was isolated at the indicated times post-infection, and QPCR was used to determine the proportion of each type ...

DISCUSSION

Strand transfer inhibitors are an important new class of anti-HIV agents that specifically target integrase function (11). Selection for resistance to the action of strand transfer inhibitors has been shown to map to the integrase gene and to encode changes near the catalytic triad responsible for catalytic activity (16, 38). Such changes have also been observed in the clinic to raltegravir, the first HIV integrase approved for treatment of HIV-infected patients (15). We performed biochemical experiments to characterize eight variants and also compared the results using isogenic viruses containing these resistance substitutions. The results of these studies are summarized in Table 7.

The kinetic strand transfer assays require that the strand transfer reactions be staged by first pre-forming the integraseviral LTR complex, prior to addition of the target DNA. This staging allows an accurate measurement of the kinetics also providing the Km for target utilization. The Km value (14.2 ± 1.3 nm) for the WT enzyme for target DNA utilization in the strand transfer reaction is similar to the reported dissociation constant (Kd) for double-stranded DNA binding to integrase using BIA-core measurements (23 ± 2 nm) (39) and slightly higher than that reported for the association of a fluorescein-labeled oligoduplex by fluorescence anisotropy changes (35 nm) (40) or the inhibition constant (Ki) for the binding of a nonspecific oligoduplex DNA to integrase measured by inhibition of 3′-processing (40 nm) (41). This is evidence that the approach taken here of using a two-phase system to measure the kinetics is valid.

A number of conclusions can be drawn from these studies concerning the properties of these integrase variants. First, it is clear that most of the single and double amino acid mutants are substantially catalytically impaired for in vitro strand transfer efficiencies. Predominantly, the origins of the reduced overall catalytic efficiencies (kp/Km) were reductions in catalytic activities (kp). However, smaller, second-order effects were caused by up to 4-fold increases in the Km value for target DNA utilization in some of the variants (e.g. Q148R). Together, the kinetic data indicate that STI-resistant integrases have alterations in the microenvironment of the active sites responsible for strand transfer and inhibitor binding.

A comparison of the relative catalytic efficiencies (Table 3, 10th column, relative kp/Km) shows that they range from as little as 0.6% for the T66I/E92Q variant to as high as 50% of WT for the V75I enzyme. Besides V75I, the next most efficient enzymes were V75I/M154I and T66I, with 44 and 36% of wild type efficiency, respectively. The N155H, Q148R, and E92Q enzymes were all between 4 and 15% as efficient as WT, with N155H being the weakest single variant (4%). The N155S virus has been reported to be less fit than WT (42, 43), and the N155E, N155K, and N155L variants were reported to have 6% of WT strand transfer activity (44), similar to the N155H results reported here. T66I had a relative efficiency of 36%, indicating that this enzyme is not much altered catalytically, as compared with WT, similar to what has been reported previously (38). Certain double variants were particularly defective. For example, the relative efficiencies of the single T66I (36%) and E92Q (15%) enzymes are reduced by ∼40-fold to nearly undetectable levels in the T66I/E92Q double mutant (relative efficiency 0.6% of WT). The main reason the double variant is so debilitated is that the catalytic activity, kp, of this enzyme is crippled, with only 0.8% of the wild type activity. This effect of a second amino acid change on integrase activity was similar to what was reported for the addition of L74M to T66I (38). The double mutant was more defective for 3′-processing and strand transfer activity as compared with either single variant. On the other hand, the activity of certain double mutants was dominated by one of the amino acid changes. For example, the N155H mutation dominates the behavior of the T66I/N155H double mutant. The relative efficiency of this double mutant (3%) was the same as for N155H alone (4%) and was significantly lower than that for T66I alone (36%).

With respect to the other catalytic activity of integrase, 3′-processing, all five of the variants that were evaluated were found to be impaired. The most defective in terms of catalytic efficiency were Q148R and T66I (both 13% of WT). Both enzymes were also attenuated for strand transfer efficiency (15 and 36% of WT, respectively). However, the attenuation reported here for T66I is greater than that reported by others. In a combined 3′-processing/strand transfer reaction in the presence of Mn2+ the specific activity of T66I was 43% of wild type integrase (45). Others have reported that the specific activity of T66I in 3′-cleavage was 110% of wild type, also using Mg2+ as the metal cofactor (38). However, those reaction conditions differed from that used in this study (1 μm enzyme, 20 nm DNA (50:1 enzyme to substrate, 15% PEG 8000, and 15% DMSO)), which may have contributed to the observed differences. We found that V75I was the least defective integrase (85% of WT), similar to the low level defect observed for strand transfer (50% of WT). The other integrases were impaired for 3′-processing by 3-fold (E92Q), 5-fold (N155H), and 8-fold each (T66I and Q148R). This is to be compared with strand transfer defects of 7-, 27-, 3-, and 7-fold for the same enzymes (Table 3). Thus, resistance substitutions, in general, give rise to defects in both catalytic activities of integrase, presumably because of the overlapping spatial arrangements and shared geometry of the active sites. For example, resistance to the integrase inhibitor l-chicoric acid generated E92K and Q148A, which were found to have greatly attenuated catalytic activities (4 and 14% of wild type, respectively) in a combined 3′-processing/strand transfer reaction (45).

These strand transfer and 3′-processing kinetic data appear relevant to the situation in the replicating virus, because several isogenic viruses harboring these resistance mutations were also defective for integration in cells in a single cycle infection assay (Tables (Tables66 and and7).7). In this assay, integration of E92Q, T66I/E92Q, and N155H was reduced 1.4–2-fold (per unit of viral DNA produced during infection) versus the WT virus. A similar reduction in integration efficiency measured by this technique was observed for several other variants selected by diketo acid inhibitors (45). The virus with the greatest integration defect (3-fold) was Q148R. In vitro, Q148R has an altered dependence on magnesium (23), requiring 10-fold higher magnesium concentrations to effect maximal levels of strand transfer and inhibitor binding, and the overall strand transfer efficiency (kp/ Km) of Q148R was 15% of WT, similar to that reported for Q148A (46). However, although the kp for Q148R is only modestly reduced (63%) compared with the WT enzyme, Q148R has the largest elevation in the Km value for target DNA utilization of all the enzymes studied (4-fold). Together, the data argue that the poorer magnesium binding is responsible for the elevated Km value and may be the underlying reason for the Q148R integration defect within the viral replication cycle. Interestingly, the Q148K mutation was also reported to attenuate virus growth (47).

Another observation from these studies is that several of the integrase enzymes have strand transfer Km values that differ more than 2-fold from that of the WT enzyme. The T66I, E92Q, and V75I/M154I integrases have ∼2-fold lower Km values than the wild type enzyme, although the N155H, T66I/N155H, and Q148R enzymes have Km values 2–4-fold higher than the wild type enzyme. A lower Km value than WT implies that the binding of the target DNA within the active site is more avid to the variant enzyme than to the wild type enzyme. Conversely, higher Km values than the wild type imply that the target DNA binds less avidly. Thus, besides the primary effects that the resistance substitutions have on assembly and catalytic activities, these mutations also influence overall catalytic efficiency (defined here as kp/Km) by altering the efficiency with which the target is recognized/bound in the active site. It appears that these mutations alter the binding of the target DNA to the complex in the transition state for the reaction, but the effects can either increase or decrease the target binding (Km) in the transition state. This indicates that the environment in the active site that binds the STI, and that is altered in the variant enzymes, is also an environment in which the binding of the target DNA is altered. The simplest explanation for these results is that STIs bind, at least in part, to certain residues that are also involved in the binding to the target DNA in the transition state of the strand transfer reaction. This conclusion is supported by the observation that the dissociation constants (Kd values) for [3H]STI 1 for the resistant enzymes correlated with the Km values for target DNA utilization (Fig. 7) in the strand transfer reaction. This also suggests a common or overlapping binding site for both inhibitor and target DNA at the active site. This commonality is further supported by the observation that inhibition of strand transfer by STI 1 was competitive with target DNA (13, 23).

As noted previously, the Kd values for E92Q and T66I, although elevated versus WT, were not elevated to the same extent as were the increases in the relative IC50 values for inhibition of these enzymes by unlabeled STI 1. Both of these Kd values are toward the lower end of Fig. 7, bottom left. Were the Kd values to have been in strict accord with the inhibition values, there would not be a material change to the correlation of Fig. 7.

Interestingly, there are examples of how opposing effects of catalytic activity (kp) and target DNA utilization (Km) can influence the efficiency of certain mutant enzymes. For instance, even though Q148R has a nearly 8-fold higher relative kp than E92Q (Table 3, 7th column), the overall relative efficiency (kp/ Km) of Q148R (Table 3, 10th column) is no different from that of E92Q (15%), due in part because the Km value for E92Q (7.7 nm) is ∼8-fold lower than the Km value for Q148R (58 nm). It is not clear which parameter, kp or Km, is more important in driving resistance development in cells. Integration by the Q148R containing recombinant virus in cells was reduced 3-fold versus WT, an indication that this variant is also defective in the context of a viral infection.

Integrase needs to assemble as a multimer onto a viral LTR-containing DNA sequence to carry out integration of the LTR ends into a target sequence (6, 26, 27, 4851), although the exact structure of the active integrase multimer is not known with certainty. A recent study indicates that the tetramer, arranged as a dimer of dimers, is required for the enzyme to acquire strand transfer activity (25). In an effort to determine whether resistance-inducing mutations might multimerize differently from the WT, we compared the concentrations of the variant integrases required to achieve 50% of the maximal product (Max50). It was observed that the eight variants required from 1.7 to 3.5 times more enzyme than the wild type enzyme to achieve this maximum (Table 1). Interestingly, the T66I variants, either as single or double variants, exhibited the greatest defects, as T66I/N155H, T66I/E92Q, or T66I required 2.4-, 2.5-, and 3.5-fold more enzyme than wild type to achieve Max50 (Fig. 2B). These differences could be the result of an increase in misfolding of the variant enzymes during expression in bacteria, or it could indicate that the T66I-containing enzymes do not assemble into active multimers as efficiently as the wild type enzyme. Analysis of the Hill slopes of the product curves in Fig. 2, A and B, provides another measure of the ability of the enzymes to cooperatively multimerize to form catalytically active complexes. The Hill slope for WT enzyme was steep (6.1) (Table 2), indicating a highly cooperative binding event. By comparison, several of the variant enzymes exhibited shallower slopes, indicative of an assembly defect. The most defective enzymes were the T66I/E92Q and T66I/N155H double variants (Hill slope 2.4 ± 1.2 and 3.2 ± 0.8, respectively), whereas T66I had a slope close to wild type (Hill slope 5.3 ± 0.4) but, as indicated, had the highest Max50 values.

This may suggest that T66I multimerizes well, but whether the enzyme preparation itself contains a higher percentage of misfolded enzyme (unable to enter into forming complexes) or the contacts that drive multimerization are weaker, and therefore require higher concentrations to form the final multimerized form, is not clear.

The other variants had intermediate Hill slopes between 3.3 and 5.0. Q148R was the least affected, with a Hill slope of 6.7. Thus, although misfolding in bacteria cannot be ruled out, the data suggest that all eight integrase variants do not assemble into active complexes as efficiently as the wild type enzyme. It is interesting that mutations that cluster near the active site for carrying out the strand transfer reaction also play a role in the efficiency of the multimerization reaction, but the relevance of this in vitro characteristic to that of the replicating virus in cell is not known. We define here the parameter enzyme/Hill slope relative to WT = 1 to give an idea of the relative efficiency of complex formation (Table 2, 5th column).

The binding and inhibition studies reveal that the mechanistic origin of resistance is a decrease in the affinity of the inhibitor for the integrase complexes, as shown by the poorer affinity of the radiolabeled inhibitor, [3H]STI 1, to resistant integrases, and the higher IC50 values for inhibition of strand transfer by STI 1. Dissociation kinetics were used to understand the origin for reduced inhibitor affinities toward these variants. It was found that STI 1 dissociates slowly (t½ = 27 min) from WT integrase complexes at 20 °C. The dissociation of STI 1 from the Q148R, N155H, and T66I integrases were 14-, 12-, and 9-fold faster, respectively, than that from the WT enzyme. Because these differences mirror the differences in affinity for STI 1, these results indicate that the dissociation rate is the controlling factor for reduced inhibitor affinities. It was also observed that the relative in vitro resistance of STI 1 toward the strand transfer activities of the variant enzymes correlated with the relative resistance of isogenic recombinant viruses carrying only these mutations in cellular assays (Tables (Tables11 and and7).7). Taken together, all these data provide strong evidence that the origin of viral resistance is entirely due to reduced STI affinity for the integrase enzyme.

Notes

*The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Footnotes

2The abbreviations used are: HIV-1, human immunodeficiency virus, type 1; WT, wild type; LTR, long terminal repeat; STI, strand transfer inhibitor; QPCR, quantitative PCR; SPA, scintillation proximity assay.

References

1. Chiu, T. K., and Davies, D. R. (2004) Curr. Top. Med. Chem. 4 965–977 [PubMed]
2. Sherman, P. A., and Fyfe, J. A. (1990) Proc. Natl. Acad. Sci. U. S. A. 87 5119–5123 [PMC free article] [PubMed]
3. Bushman, F. D., and Craigie, R. (1991) Proc. Natl. Acad. Sci. U. S. A. 88 1339–1343 [PMC free article] [PubMed]
4. Craigie, R., Fujiwara, T., and Bushman, F. D. (1990) Cell 62 829–837 [PubMed]
5. Vink, C., Van Gent, D. C., Elgersma, Y., and Plasterk, R. H. A. (1991) J. Virol. 65 4636–4644 [PMC free article] [PubMed]
6. Cherepanov, P., Maertens, G., Proost, P., Devreese, B., Van Beeumen, J., Engelborghs, Y., De Clercq, E., and Debyser, Z. (2003) J. Biol. Chem. 278 372–381 [PubMed]
7. Katz, R. A., Merkel, G., Kulkosky, J., Leis, J., and Skalka, A. M. (1990) Cell 63 87–95 [PubMed]
8. Engelman, A., Mizuuchi, K., and Craigie, R. (1991) Cell 67 1211–1221 [PubMed]
9. Vincent, K. A., York-Higgins, D., Quiroga, M., and Brown, P. O. (1990) Nucleic Acids Res. 18 6045–6047 [PMC free article] [PubMed]
10. Vink, C., and Plasterk, R. H. A. (1993) Trends Genet. 9 433–437 [PubMed]
11. Hazuda, D. J., Felock, P., Witmer, M., Wolfe, A., Stillmock, K., Grobler, J. A., Espeseth, A., Gabryelski, L., Schleif, W., Blau, C., and Miller, M. D. (2000) Science 287 646–650 [PubMed]
12. Assante-Appiah, E., and Skalka, A. M. (1999) Adv. Virus Res. 52 351–369 [PubMed]
13. Espeseth, A. S., Felock, P., Wolfe, A., Witmer, M., Grobler, J., Anthony, N., Egbertson, M., Melamed, J. Y., Young, S., Hamill, T., Cole, J. L., and Hazuda, D. J. (2000) Proc. Natl. Acad. Sci. U. S. A. 97 11244–11249 [PMC free article] [PubMed]
14. Pommier, Y., Johnson, A. A., and Marchand, C. (2005) Nat. Rev. Drug Discov. 4 236–248 [PubMed]
15. Cooper, D. A., Gatell, J., Rockstroh, J., Katlama, C., Yeni, P., Lazzarin, A., Chen, J., Isaacs, R., Teppler, H., Nguyen, B.-Y., and Group, B.-S. (2007) 14th Conference on Retroviruses and Opportunistic Infections, Los Angeles, CA, February 25–28, 2007
16. Fikkert, V., Hombrouck, A., Van Remoortel, B., De Maeyer, M., Pannecouque, C., De Clercq, E., Debyser, Z., and Witvrouw, M. (2004) AIDS 18 2019–2020 [PubMed]
17. Ren, C., May, S., Miletti, T., and Bedard, J. (2007) 16th International HIV Resistance Workshop, Bridgetown, Barbados, June 12–16, 2007, Abstr. 1
18. Hazuda, D. J., Miller, M. D., Nguyen, B. Y., and Zhao, J. (2007) 16th International HIV Resistance Workshop, Bridgetown, Barbados, June 12–16, 2007, Abstr. 8
19. Jones, G., Ledford, R. M., Yu, F., Chen, X., Miller, M. D., Tsiang, M., and McColl, D. J. (2007) 14th Conference on Retroviruses and Opportunistic Infections, Los Angeles, CA, February 25–28, 2007, Abstr. 627
20. McColl, D. J., Fransen, S., Gupta, S., Parkin, N., Margot, N., Ledford, R., Chen, J., Chuck, S., Cheng, A. K., and Miller, M. D. (2007) 16th International HIV Resistance Workshop, Bridgetown, Barbados, June 12–16, 2007, Abstr. 9
21. Shimura, K., Kodama, E., Sakagami, Y., Matsuzaki, Y., Watanabe, W., Yamataka, K., Watanabe, Y., Ohata, Y., Doi, S., Sato, M., Kano, M., Ikeda, S., and Matsuoka, M. (2007) J. Virol. 82 764–774 [PMC free article] [PubMed]
22. Yoshinaga, T., Nakahara, K., Kobayashi, M., Miki, S., Sato, A., Garvey, E., Foster, S., Underwood, M., Johns, B., and Fujiwara, T. (2008) 15th Conference on Retroviruses and Opportunistic Infections, Boston, MA, February 3–5, 2008, Abstr. 860
23. Dicker, I. B., Samanta, H. K., Li, Z., Hong, Y., Tian, Y., Banville, J., Remillard, R. R., Walker, M. A., Langley, D. R., and Krystal, M. (2007) J. Biol. Chem. 282 31186–31196 [PubMed]
24. Smolov, M., Gottikh, M., Tashlitskii, V., Korolev, S., Demidyuk, I., Brochon, J. C., Mouscadet, J. F., and Deprez, E. (2006) FEBS J. 273 1137–1151 [PubMed]
25. Li, M., Mizuuchi, M., Burke, T. R., and Craigie, R. (2006) EMBO J. 25 1295–1304 [PMC free article] [PubMed]
26. Faure, A., Litvak, S., and Parissi, V. (2005) Nucleic Acids Res. 33 977–986 [PMC free article] [PubMed]
27. Wang, L. D., Liu, C. L., Chen, W. Z., and Wang, C. X. (2005) Biochem. Biophys. Res. Commun. 337 313–319 [PubMed]
28. O'Doherty, U., Swiggard, W. J., and Malim, M. H. (2000) J. Virol. 74 10074–10080 [PMC free article] [PubMed]
29. O'Doherty, U., Swiggard, W. J., Jeyakumar, D., McGain, D., and Malim, M. H. (2002) J. Virol. 76 10942–10950 [PMC free article] [PubMed]
30. Butler, S. L., Johnson, E. P., and Bushman, F. D. (2002) J. Virol. 76 3739–3747 [PMC free article] [PubMed]
31. Blair, W., and Spicer, T. P. (January 16, 2003) U.S. Patent Application 2003001378 A1
32. Bartz, S. R., Rogel, M. E., and Emerman, M. (1996) J. Virol. 70 2324–2331 [PMC free article] [PubMed]
33. Cheng, H. C. (2004) Pharmacol. Res. 50 21–40 [PubMed]
34. Segel, I. H. (1975) Enzyme Kinetics, p. 19, Wiley-Interscience, New York
35. Wiskerchen, M., and Muesing, M. A. (1995) J. Virol. 69 376–386 [PMC free article] [PubMed]
36. Fransen, S., Gupta, S., Danovich, R., Hazuda, D., Miller, M., Witmer, M., Petropoulos, C. J., Parkin, N. T., and Huang, W. (2008) XVII HIV Drug Resistance Workshop, Sitges, Spain, June 10–14, 2008, Abstr. 7
37. Miller, M. D., Danovich, R. M., Ke, Y., Witmer, M., Zhao, J., Harvey, C., Nguyen, B.-Y., and Hazuda, D. (2008) XVII International HIV Drug Resistance Workshop, Sitges, Spain, June 10–14, 2008, Abstr. 6
38. Fikkert, V., Van Maele, B., Vercammen, J., Hantson, A., Van Remoortel, B., Michiels, M., Gurnari, C., Pannecouque, C., De Maeyer, M., Engelborghs, Y., De Clercq, E., Debyser, Z., and Witvrouw, M. (2003) J. Virol. 77 11459–11470 [PMC free article] [PubMed]
39. Yi, J., Asante-Appiah, E., and Skalka, A. M. (1999) Biochemistry 38 8458–8468 [PubMed]
40. Deprez, E., Barbe, S., Kolaski, M., Leh, H., Zouhiri, F., Auclair, C., Brochon, J. C., Le Bret, M., and Mouscadet, J. F. (2004) Mol. Pharmacol. 65 85–98 [PubMed]
41. Bugreev, D. V., Baranova, S., Zakharova, O. D., Parissi, V., Desjobert, C., Sottofattori, E., Balbi, A., Litvak, S., Tarrago-Litvak, L., and Nevinsky, G. A. (2003) Biochemistry 42 9235–9247 [PubMed]
42. Hazuda, D. J., Anthony, N. J., Gomez, R. P., Jolly, S. M., Wai, J. S., Zhuang, L., Fisher, T. E., Embrey, M., Guare, J. P., Jr., Egbertson, M. S., Vacca, J. P., Huff, J. R., Felock, P. J., Witmer, M. V., Stillmock, K. A., Danovich, R., Grobler, J., Miller, M. D., Espeseth, A. S., Jin, L., Chen, I. W., Lin, J. H., Kassahun, K., Ellis, J. D., Wong, B. K., Xu, W., Pearson, P. G., Schleif, W. A., Cortese, R., Emini, E., Summa, V., Holloway, M. K., and Young, S. D. (2004) Proc. Natl. Acad. Sci. U. S. A. 101 11233–11238 [PMC free article] [PubMed]
43. Kehlenbeck, S., Betz, U., Birkmann, A., Fast, B., Goller, A. H., Henninger, K., Lowinger, T., Marrero, D., Paessens, A., Paulsen, D., Pevzner, V., Schohe-Loop, R., Tsujishita, H., Welker, R., Kreuter, J., Rubsamen-Waigmann, H., and Dittmer, F. (2006) J. Virol. 80 6883–6894 [PMC free article] [PubMed]
44. Gerton, J. L., Ohgi, S., Olsen, M., DeRisi, J., and Brown, P. O. (1998) J. Virol. 72 5046–5055 [PMC free article] [PubMed]
45. Lee, D. J., and Robinson, W. E., Jr. (2004) J. Virol. 78 5835–5847 [PMC free article] [PubMed]
46. Lee, D. J., and Robinson, W. E., Jr. (2006) Antimicrob. Agents Chemother. 50 134–142 [PMC free article] [PubMed]
47. Neamati, N. (2002) Expert Opin. Ther. Patents 12 709–724
48. Deprez, E., Tauc, P., Leh, H., Mouscadet, J. F., Auclair, C., and Brochon, J. C. (2000) Biochemistry 39 9275–9284 [PubMed]
49. Deprez, E., Tauc, P., Leh, H., Mouscadet, J. F., Auclair, C., Hawkins, M. E., and Brochon, J. C. (2001) Proc. Natl. Acad. Sci. U. S. A. 98 10090–10095 [PMC free article] [PubMed]
50. Gao, K., Butler, S. L., and Bushman, F. (2001) EMBO J. 20 3565–3576 [PMC free article] [PubMed]
51. Jones, K. S., Coleman, J., Merkel, G. W., Laue, T. M., and Skalka, A. M. (1992) J. Biol. Chem. 267 16037–16040 [PubMed]

Articles from The Journal of Biological Chemistry are provided here courtesy of American Society for Biochemistry and Molecular Biology

Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...