HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 25, 22345-22352, June 21, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/25/22345    most recent M200282200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Fisher, T. S. Articles by Prasad, V. R. Search for Related Content PubMed PubMed Citation Articles by Fisher, T. S. Articles by Prasad, V. R. Social Bookmarking                      What's this?

Substitutions of Phe⁶¹ Located in the Vicinity of Template 5'-Overhang Influence Polymerase Fidelity and Nucleoside Analog Sensitivity of HIV-1 Reverse Transcriptase^*

Timothy S. Fisher and Vinayaka R. Prasad^§

From the Department of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, New York 10461

Received for publication, January 10, 2002, and in revised form, March 25, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Human immunodeficiency virus type 1 reverse transcriptase (RT) is an error-prone DNA polymerase. Structural determinants of its fidelity are incompletely understood. RT/template primer contacts have been shown to influence its fidelity and sensitivity to nucleoside analog inhibitors. The Phe⁶¹ residue, located within the 3 sheet of the finger subdomain, is highly conserved among retroviral RTs. The crystal structure of a ternary complex revealed that Phe⁶¹ contacts the first and second bases of the 5'-template overhang. To determine whether such contacts influence the dNTP-binding pocket, we performed a limited vertical scanning mutagenesis (Phe  Ala, Leu, Trp, or Tyr) at Phe⁶¹. The F61A mutant displayed the highest increase in fidelity, followed by the F61L and F61W variants, which had intermediate phenotypes. F61Y RT had a minimal effect. The increase in fidelity of the F61A mutant was corroborated by a 12-fold decrease in its forward mutation rate. The Phe⁶¹ mutant RTs also displayed large reductions in sensitivity to 2',3'-dideoxythymidine triphosphate and 2',3'-dideoxy,2'3'-didehydrothymidine triphosphate. Mutants displaying the largest increase in fidelity (F61A and F61L) were also the most resistant. These results suggest that contacts between the finger subdomain of human immunodeficiency virus type 1 RT and the template 5'-overhang are important determinants of the geometry of the dNTP-binding pocket.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To date, the lack of success in the development of an effective vaccine against human immunodeficiency virus (HIV)¹ infections has led to the practice of long term administration of cocktails of antiretroviral drugs for the efficient suppression of viremia. However, the inherent genomic hypervariability of HIV-1 (1, 2) in combination with a high rate of replication (3, 4) allows for the generation of drug-resistant variants during antiretroviral therapy and evasion of the host immune system. As a result, treatment consisting of the currently used drugs for treatment of AIDS inevitably leads to the emergence of resistant viruses (reviewed in Ref. 5). Therefore, determining viral factors that influence either the rate of viral replication or mutation continues to be of utmost importance in the further development of better antiretroviral therapies.

One of the major sources of HIV-1 sequence variation is its reverse transcriptase (RT). HIV-1 RT is required to convert the HIV-1 single-stranded RNA genome into double-stranded proviral DNA. The complete synthesis of proviral DNA requires a multitude of activities inherent to HIV-1 RT, including RNA- and DNA-dependent DNA polymerase ribonuclease H, strand transfer, and strand displacement synthesis (reviewed in Ref. 6). HIV-1 RT is thought to be an error-prone DNA polymerase (7, 8). In addition to its poor polymerase fidelity, HIV-1 RT lacks a 3' to 5' proofreading activity, which precludes the removal of misinserted dNTPs (8). Even when compared with other retroviral RTs also lacking proofreading, such as avian myeloblastosis virus and Moloney murine leukemia virus RTs, HIV-1 RT is significantly more error prone (8, 9). Although the importance of the poor polymerase fidelity of HIV-1 RT in HIV-1 variation has been well known for some time, knowledge of the molecular determinants contributing to its error-prone nature is only beginning to emerge.

Studies on the polymerase fidelity of HIV-1 RT have focused on two major categories of RT variants. In one class are the mutants that emerge in patients receiving therapy with NRTIs or in cell culture systems under drug pressure, and in a second are those selected via structure-based mutagenesis. Work from our laboratory and that of others have shown that a subset of mutations conferring resistance to NRTIs enhances HIV-1 RT polymerase fidelity relative to the wild type enzyme (10-16). Mutations that confer resistance to NRTIs, with the exception of 2',3'-dideoxy-3'-azidothymidine resistance mutations, enable RT to discriminate between a nucleoside analog inhibitor and its normal dNTP substrate. Such mutations may also allow RT to better discriminate between a correct versus an incorrect dNTP based on the geometry of the dNTP-binding pocket. This was demonstrated initially for the ()2',3'-dideoxy-3'-thiacytidine resistance mutation M184V by work previously reported from our laboratory as well as that of others (11-13). Similarly, the multi-2',3'-dideoxynucleoside analog-resistant variant K65R displayed a 8-fold increase in overall fidelity (16). The Lys⁶⁵ residue directly contacts the -phosphate of incoming dNTP, whereas the Met¹⁸⁴ side chain contacts the sugar and the base of the 3'-nucleotide in the primer, but the substitution with a -branched side chain such as Leu or Ile creates contact with the dNTP sugar ring as well (17). Thus, the influence of these mutations on the dNTP insertion fidelity or the overall forward mutation rate can be readily appreciated.

The influence of a second class of mutations on polymerase fidelity, however, appears to be indirectly mediated via contacts with the template primer duplex. One of the earliest reported examples of such mutations is the E89G substitution in the template grip 5a in the palm region. The multi-2',3'-dideoxynucleoside analog-resistant E89G mutant RT displayed an increase in dNTP insertion fidelity (10, 11). The Glu⁸⁹ residue is known to contact the template strand at the penultimate base pair (2 bases toward the 3' of the templating nucleotide in the duplex region) Therefore, a change in the conformation of the template primer duplex could lead to alterations in the geometry of the dNTP-binding pocket. Similarly, mutation at the residue Pro¹⁵⁷, which contacts the same template nucleotide, conferred a sequence-specific increase in resistance to nucleoside analogs (18). Additional evidence that contacts made between HIV-1 RT and the template primer affect fidelity include mutations in the fingers (R78A and D76V) and thumb (G262A and W266A) subdomains (19-21). D76V and R78A variants were selected via an Escherichia coli complementation screen and structure-based mutagenesis, and both displayed an overall fidelity 9-fold greater than wild type RT (19, 20). Asp⁷⁶ and Arg⁷⁸ are H-bonded to each other, and Asp⁷⁶ contacts the template nucleotide base paired to the incoming dNTP. Residues Gly²⁶² and Trp²⁶⁶ constitute part of the minor groove binding track, a structural element found primarily within the H helix of the thumb subdomain (22). Alterations at these two positions were found to dramatically affect polymerase processivity, template primer affinity, and frameshift fidelity (21, 23). Thus, several mutations shown to affect HIV-1 RT polymerase fidelity presumably do so by "repositioning" the template primer, ultimately leading to a more or less stringent dNTP-binding site and/or polymerase active site. In either case, repositioning of the template primer affects the ability of HIV-1 RT to discriminate between correct versus incorrect dNTP substrates.

Most of the template- or primer-contacting residues that influence fidelity (Glu⁸⁹, Met¹⁸⁴, Gly²⁶², and Trp²⁶⁶) contact the nucleic acid within the template primer duplex with the exception of residues such as Asp⁷⁶ and Arg⁷⁸, which contact the sugar-phosphate backbone of the templating nucleotide. Thus, mutations that affect misincorporation fidelity or the overall mutation rates of HIV-1 RT appear to contact the incoming dNTP (K65R and M184V), the primer terminus (M184V), the templating base (L74V), or the template strand within the duplex region (E89G, G262A, and W266A). None of the residues that contact the template overhang beyond the templating base have been reported to affect fidelity. The recent crystal structure of HIV-1 RT complexed with a DNA-DNA template primer and dNTP shows that two highly conserved amino acid residues in the finger subdomain, Phe⁶¹ and Trp²⁴, interact with the extended single-stranded template (17). Earlier crystallographic data suggested that the extended template would continue in a helical path, entering the polymerase active site through the cleft formed by the fingers, palm, and thumb subdomains of HIV-1 RT (24, 25). However, as shown in the crystal structure of the ternary complex, the template does not continue in a helical path but instead bends away from the duplex, making contact along the face of the finger subdomain. Trp²⁴ is proposed to interact with the phosphate group between the second and third nucleotide, whereas Phe⁶¹ contacts both the first and second nucleotides of the extended template. As mentioned above, several mutations within the finger subdomain of HIV-1 RT affect the fidelity of the enzyme and its sensitivity to NRTIs. Therefore, it is possible that these two conserved contacts with the extended template may contribute to the observed poor fidelity of HIV-1 RT during DNA synthesis.

In this report, we have evaluated the influence of the Phe⁶¹ residue on polymerase fidelity. Substitutions created at Phe⁶¹ were found to increase the fidelity of DNA synthesis by RT. Interestingly, the changes in fidelity of Phe⁶¹ mutant RTs correlated with their sensitivity to NRTIs. We conclude that contacts between the finger subdomain of HIV-1 RT and the extended single-stranded template are functionally important in determining the overall polymerase fidelity and dNTP analog sensitivity of the enzyme.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Poly(rA) and oligo(dT)_12-18 were purchased from Amersham Biosciences. 16 S rRNA, dNTPs, and ddTTP were purchased from Roche Molecular Biochemicals. Radiolabeled nucleotides were from PerkinElmer Life Sciences, and all of the oligonucleotides used in these studies were purchased from Gene Link, Inc. d4TTP was purchased from Sierra Bioresearch.

Phage and Bacterial Strains-- The bacteriophage M13mp2 was used to prepare the gapped duplex DNA substrate. M13 phage was grown in the E. coli strain NR9099 ((pro-lac), thi, ara, recA56/F' (proAB, lacIqZM15)) to make single-stranded and replicative form DNAs. Electrocompetent E. coli strain MC1061 (hsdR, hsdM+, araD, (ara, leu), (lacIPOZY), galU, galK, strA) was used to generate the mutant phage. The -complementation strain of E. coli, CSH50 ((pro-lac), thi, ara, strA/F' (proAB, lacIqZM15, traD36)), was used to visualize the phenotype of the mutant phage.

Enzymes-- Recombinant heterodimeric HIV-1_HXB2 RT and mutant RTs carrying Phe⁶¹ substitutions only in the p66 subunit were generated as described (26). All of the enzymes were found to be fee of contaminating DNase and RNase activities under conditions that were more stringent (10 times the RT amount and 10 times the reaction time) than that used in our polymerase assays (data not shown).

Determination of RT Activities-- The polymerase activity of wild type and mutant RTs on various RNA and DNA templates was measured in standard RT reactions. The following template primer pairs were used: poly(rA) annealed to oligo(dT), 16 S rRNA annealed to a 22-nt DNA primer VP₂₀₀ (5'-TAACCTTGCGGCCGTACTCCCC-3') (nucleotides 885-906 of 16 S rRNA), and an oligodeoxynucleotide template primer pair consisting of a 29-nt DNA PBS_A primer annealed to a 55-nt DNA VP229 template: VP229, 3'-GCGAAAGTCCAGGGACAAGCCCGCGGTGACGATCTCTAAAAGGTGTGACTGATTT-5', and PBS_A, 5'-CGCTTTCAGGTCCCTGTTCGGGCGCCAC-3'.

Reaction mixtures (50 µl) contained 1 µM template primer (primer 3'-OH ends), 4.2 nM (25 ng) RT, 80 mM KCl, 50 mM Tris-Cl (pH 8.0), 6 mM MgCl₂, 1 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, 10 µM [-³²P]dTTP, 25 µM each of the remaining three dNTPs. The reactions were incubated at 37 °C for 15 min before being stopped by spotting 40 µl of the reaction mixture onto DE81 filter squares. The filters were washed with 2× SSC (30 mM sodium citrate, 300 mM sodium chloride, pH. 7.0) for 20 min three times to remove unincorporated dNTPs. The filters were dried and counted using a scintillation counter (1218 Rack Beta; LKB-Wallac, Stockholm, Sweden).

Steady-state Kinetic Studies-- The kinetic constants K_m and V_max were determined for RNA-dependent DNA polymerase activity on both homopolymeric (poly(rA)·oligo(dT)) and heteropolymeric (16 S rRNA·VP₂₀₀ DNA primer) template primer pairs. The reaction mixtures (25 µl) carried out with the poly(rA) template annealed to a oligo(dT) primer contained 80 mM KCl, 50 mM Tris-Cl (pH 8.0), 6 mM MgCl₂, 10 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, 8.4 nM (2.5 ng) RT, and 1.0 µM template primer. Similar reactions were carried out with a 16 S rRNA template annealed to a 22-nt DNA primer, VP₂₀₀, and contained 0.84 nM (25 ng) RT and 50 µM each of dATP, dCTP, and dGTP. For both sets of reactions the dTTP substrate concentration was varied from 0.1 to 50 µM, and the amount of DNA synthesis was determined by measuring [-³²P]dTMP incorporation. The reactions were carried out for 10 min at 37 °C and were stopped by spotting 20 µl of the reaction mixture onto DE81 filter squares. The reactions were further processed as mentioned above, and the dNMP incorporation was quantitated as before. The kinetic constants (K_m and V_max) were determined by fitting results from at least three independent experiments to a Michaelis-Menten curve using nonlinear regression (GraphPad Software Inc.).

Single dNTP Exclusion Assay-- Primer extension reactions were performed using a 5'-³²P-labeled DNA 28-nt primer, VP₂₂₉, annealed to a 55-nt DNA template, PBS_A at a molar ratio of 5:1 template to primer. The reaction mixtures (20 µl) contained 80 mM KCl, 50 mM Tris-Cl (pH 8.0), 6 mM MgCl₂, 10 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, 3 or 4 dNTPs (250 µM each), and 10 nM template primer. Two different concentrations (2.0 and 4.0 nM) of the wild type and Phe⁶¹-substituted RTs were employed in the reactions. The reactions were carried out for 5 min at 37 °C before being terminated by the addition of 40 µl of stop solution (95% formamide, 10 mM EDTA, and 0.1% each of xylene cyanol and bromphenol blue). The reaction products were separated by 10% denaturing PAGE. The gels were dried, and the radiolabeled products were analyzed using a PhosphorImager and ImageQuant software. Under these conditions, the reactions not containing dTTP resulted in minimal DNA synthesis by all RTs tested and thus were not included (data not shown).

Mispair Extension Assay-- The reactions were similar to those for the single dNTP exclusion assay. The mispair extension reactions were performed using two different 5'-³²P-labeled 31-nt DNA primers (G-C and G-T) annealed to a 55-nt DNA template (VP229). The primary sequences of primers used were as follows: G-C primer (5'-CGCTTTCAGGTCCCTGTTCGGGCGCCACTGC-3') and G-T primer (5'-CGCTTTCAGGTCCCTGTTCGGGCGCCACTGT-3'). The reaction mixtures (20 µl) contained 25 µM each dATP, dGTP, and dTTP and 10 nM template primer. Two different concentrations (2.0 and 4.0 nM) of the wild type and Phe⁶¹-substituted RTs were employed in the reactions. The reactions were carried out for 6 min at 37 °C before being terminated by the addition of 20 µl of stop solution. The reaction products were separated by 10% denaturing PAGE. The gels were dried, and the radiolabeled products were analyzed as above.

Determination of Forward Mutation Frequency-- The mutation frequencies of wild type and F61A mutant RT were measured essentially as described previously (27, 28). M13mp2 DNA containing a single-stranded gap of 361 nucleotides was prepared as specified (27) and used as a template-primer for fill-in DNA synthesis by wild type and F61A mutant RTs. Fill-in reactions (25 µl) contained 80 mM KCl, 75 mM Tris-Cl (pH 8.0), 10 mM dithiothreitol, 6 mM MgCl₂, 0.5 mM each dATP, dCTP, dGTP, and dTTP, 75 ng of gapped duplex DNA, and 170 nM (500 ng) of purified RT. The reactions were incubated for 1 h at 37 °C before stopping with the addition of 1 µl of 0.5 M EDTA. Complete synthesis across the gapped region was confirmed by agarose gel electrophoresis.

Filled in M13mp2 DNA was electroporated into E. coli MC1061 cells, and transformants were plated to M9 plates containing 5-bromo-4-chloro-3-indolyl -D-galactopyranoside (X-gal; Labscientific Inc.) and isopropyl-1-thio--D-galactopyranoside (Sigma) with CSH50 lawn cells. The plates were incubated at 37 °C for ~15 h and then screened for plaques that did not display the dark blue wild type phenotype. Mutant phenotypes were confirmed by plating equal inputs of wild type and potential mutant phage on an indicator lawn as described above. The mutation frequency was determined as the ratio of confirmed mutant (pale blue and clear) plaques to total plaques as described (27). The background mutation frequency was determined by electroporating unfilled gapped duplex and scoring for mutant plaques as described above.

Determination of Sensitivity to dTTP and d4TTP-- The sensitivity of the wild type and mutant RTs to the NRTIs, ddTTP, and d4TTP was measured in reactions similar to those detailed above. The reaction mixtures (50 µl) contained 1 µM 16 S rRNA template annealed to a slight excess of the primer, VP₂₀₀ (sequence described above under "Determination of RT Activities") and 25 ng of RT. The concentrations of dATP, dCTP, and dGTP were each at 25 µM, whereas dTTP was at 5 µM and was -³²P-labeled. The reactions were carried out in the absence or presence of increasing concentrations of ddTTP or d4TTP (50 nM to 400 µM) for 15 min at 37 °C before being terminated by spotting 40 µl of the reaction onto DE81 paper. The reactions were further processed as mentioned above, and the dNMP incorporation was quantitated as before. IC₅₀ values of each NRTI for a given RT variant were determined by fitting results from at least three independent experiments to a dose-response curve using nonlinear regression (GraphPad Software Inc.).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Enzymatic Activities of Wild Type and Phe⁶¹ Mutant HIV-1 RTs-- The effects of various substitutions at Phe⁶¹ on the polymerase activity of HIV-1 RT are shown in Fig. 1. Replacing Phe⁶¹ with Leu, Trp, or Tyr resulted in variant RTs with a polymerase activity that was the same as wild type RT or higher on all substrates tested. On both heteropolymeric RNA (16 S rRNA·VP₂₀₀) and DNA (PBS_A·VP200) templates, the F61Y mutant displayed at least a 2-fold increase in polymerase activity and was even more active (>4-fold) when a poly(rA)·oligo(dT) template primer was used. The F61L mutant also displayed higher activity (~2-fold) than wild type RT on either a homopolymeric RNA or heteropolymeric DNA template. In contrast, replacement of Phe⁶¹ with Ala resulted in an RT variant with polymerase activity either similar to or lower than (~2-fold) wild type RT depending on whether an RNA or DNA template was used.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Polymerase activity of wild type and Phe61 mutant RTs. RNA-dependent DNA polymerase or DNA-dependent DNA polymerase activities determined on homopolymeric (poly(rA)·oligo(dT)12-18) or heteropolymeric RNA-DNA (16 S rRNA·VP200) or DNA-DNA (PBSA·VP229) template primer pairs are shown. The values were normalized to the activity obtained with wild type RT on each template primer. The reactions were carried out as detailed under "Experimental Procedures." The results shown are the averages of at least three independent experiments, and the deviation between experiments was below 10% in all cases.

We also examined the steady-state kinetic parameters (K_m and V_max) for both wild type and mutant RTs. The kinetic parameters were determined using both homopolymeric (poly(rA)·oligo(dT)) and heteropolymeric (16 S rRNA·VP₂₀₀) RNA-DNA template primer pairs and the results shown in Table I. On the homopolymeric template, each Phe⁶¹ substitution resulted in a 2-4-fold increase in V_max. For F61L, F61W, and F61Y mutants, there was a corresponding 2-4-fold increase in the catalytic efficiency (V_max/K_m). However, for F61A, in addition to the increase in V_max, a 6-fold increase in K_m dTTP was also observed, leading to an overall decrease of 2-fold in catalytic efficiency. No significant difference in the K_m dTTP of other Phe⁶¹ mutants was observed when using the homopolymeric template primer. A similar trend in the K_m dTTP of wild type and Phe⁶¹ mutant RTs was seen on a heteropolymeric template, in that the F61A mutant displayed the highest K_m dTTP of all the substitutions (3-fold increase from wild type) with F61L showing a 2-fold increase from wild type RT. The mutants F61W and F61Y showed values similar to whose of the wild type. The 3-fold increase in K_m dTTP for F61A, combined with a 2.5-fold decrease in V_max, resulted in an overall 8-fold decrease in catalytic efficiency, suggesting that the F61A mutation had a larger effect on polymerization on heteropolymeric substrates. The effect of the F61L mutation was also more severe on the heteropolymeric template, resulting in a nearly 2-fold loss of catalytic efficiency, in contrast to a 2-fold increase on a homopolymeric template. Both F61W and F61Y had similar effects on V_max using a heteropolymeric template, resulting in modest increases in catalytic efficiency (Table I). These data suggested that the observed increases in polymerase activity of F61W and F61Y was likely due to an increase in V_max, with little effect on K_m of dNTP. In addition, the loss of activity observed with F61A on a heteropolymeric RNA template was likely due to both an increase in K_m and a decrease in V_max. Importantly, the effects of substitutions on both the overall polymerase activity and kinetic constants were due to the presence of the mutation only in the p66 subunit of HIV-1 RT.

Phe⁶¹ RT Mutants Display Increase in Misincorporation Fidelity-- To investigate the effects of mutations at Phe⁶¹ of the finger subdomain of HIV-1 RT on the fidelity of DNA synthesis, we first employed a gel-based assay that measures primer extension in the presence of only three of four dNTPs complementary to template nucleotides, referred to here as a "single dNTP exclusion" assay (7). For these studies, a 22-nt 5'-end labeled DNA primer was annealed to a 55-nt DNA template consisting of an HIV-1 PBS sequence. As a result of omitting a single dNTP from the reactions, a barrier to primer elongation is created at the template position at which the enzyme would normally incorporate the missing dNTP. Therefore, any elongation past this barrier site requires that the enzyme both insert an incorrect nucleotide (misinsertion) and then extend the resulting mismatched primer terminus (mispair extension). By measuring the amounts of products formed past the barrier, one can obtain an overall estimate of misincorporation fidelity, which is a combined sum of changes in the efficiencies of misinsertion and mispair extension for polymerases lacking an exonucleolytic proofreading activity.

In reactions in which all four dNTP were included, wild type and mutant RTs synthesized similar amounts of full-length and intermediately sized products (Fig. 2A, panel labeled All dNTPs). When increasing amounts of enzyme were included in these reactions, the amount of full-length product subsequently increased with all RTs tested, indicating that equal amounts of enzyme, based on DNA-dependent DNA polymerase activity, were used in all reactions. Wild type HIV-1 RT efficiently extended the primer past the first barrier site when each of three dNTPs was missing (Fig. 2A, panel labeled Minus dNTPs). In each case of excluding a single dNTP, wild type RT was capable of passing through several barrier sites and, in reactions without dATP or without dGTP, synthesized a limited amount of full-length product (Fig. 2A, panels labeled Minus dATP and Minus dGTP). Under identical conditions, heterodimeric RTs carrying substitutions at Phe⁶¹ in the p66 subunit extended significantly fewer products past the first barrier site and synthesized fewer products past this site compared with wild type RT (Fig. 2A). To investigate trends in the overall fidelity of Phe⁶¹ mutant RTs, we quantitated the amount of products formed past the first barrier site for all RTs in relation to the total amount of products formed both at and above this site. The percentage of products formed past each site for all Phe⁶¹ mutants was then compared with the percentage of products formed past each site by wild type RT. Shown in Fig. 2B are the relative product formation of both wild type and mutant RT past the barrier site for each set of reactions lacking a single dNTP. As mentioned above, wild type RT extended most efficiently past each barrier site compared with Phe⁶¹ mutant RTs. In general, the relative product formation past the barrier site by Phe⁶¹ mutants was correlated with the degree of conservation between the original and substituted amino acid, i.e. the less conserved substitution led to the least amount of extension past the barrier site (Phe⁶¹ > F61Y or F61W > F61L > F61A). These data indicate that each substitution at Phe⁶¹ confers greater accuracy of DNA synthesis by increasing the misinsertion and/or mispair extension fidelities of mutant enzymes. The effect of such mutations on fidelity was observed to be greatest with the least conserved F61A substitution at this amino acid position.

View larger version (97K): [in this window] [in a new window]   Fig. 2.   Single dNTP exclusion assay of WT and Phe61 mutant RTs. A, analysis of products resulting from synthesis in the absence of one dNTP. The DNA sequence of the first 13 nucleotides synthesized is indicated on the left side of the first panel, with asterisks indicating the first position where the excluded dNTP would be incorporated in reactions containing only three dNTPs. In each of the three panels where a dNTP was excluded (Minus dATP, Minus dCTP, and Minus dGTP), the band representing termination caused by the first barrier site would be immediately prior to this site. Each set of reactions contained two different concentrations (2 and 4 nM) of WT or Phe61-substituted RT. In the first panel, all four dNTPs are present (All dNTPs), whereas in subsequent panels a single dNTP was missing (Minus dATP, Minus dCTP, and Minus dGTP). The reactions in which TTP was excluded resulted in minimal activity and are not shown. B, relative fidelities of wild type and Phe61-substituted RTs. The amount of extension products formed past the first barrier site was quantitated as described under "Experimental Procedures." The total intensity was calculated as the band immediately prior to the first barrier site plus all extension products above, whereas extension was calculated as the sum of all products above the first barrier site. The percentage of extension past the barrier site as a fraction of the total intensity is shown for wild type and Phe61 mutant RTs. All of the results represent the means ± S.E. of three independent experiments.

Phe⁶¹ Substitutions in HIV-1 RT Also Affect Mispair Extension Efficiency-- Primer extension past the barrier sites in reactions missing a single dNTP was the result of both misinsertion of an incorrect nucleotide and mispair extension from the subsequent mispaired primer terminus. Therefore, we were interested in determining the contribution of mispair extension alone relative to the observed increases in fidelity of Phe⁶¹ mutants compared with wild type RT. To investigate the effects of Phe⁶¹ substitutions on the relative fidelity of mispair extension, we used a gel-based mispair extension assay. This assay measured the relative efficiency at which both wild type and Phe⁶¹ mutant RTs extended either a properly base-paired template primer or a mispaired template primer terminus. In both sets of reactions, a 31-nt 5'-end labeled DNA primer was annealed to a 55-nt DNA template consisting of an HIV-1 PBS sequence. To ensure that reaction products were the result of mispair extension, dCTP was not included in reactions, thus preventing exonucleolytic removal of the mispaired primer terminus followed by a possible insertion of the correct dCTP substrate.

In reactions with a properly base-paired G-C template primer terminus, wild type and Phe⁶¹ mutant RTs synthesized comparable amounts of the +10 product in addition to intermediately sized products (Fig. 3A). Of note, both the F61A and F61L mutant RTs produce more intermediately sized products than wild type RT, which is likely due to a defect in processive DNA synthesis through runs of nucleotides.² Regardless of the size of extension products, the total number of primers extended by wild type and F61A and F61L RTs are similar, indicating that equal amounts of enzyme, based on DNA-dependent DNA polymerase activity, were used in all reactions. Wild type HIV-1 RT efficiently extended the mispaired G-T template primer, resulting in most primers being extended by 10 nucleotides (Fig. 3B). F61A mutant RT was clearly inefficient at extending the G-T mispaired primer terminus compared with wild type RT (Fig. 3B). F61L formed more mispair extension products than F61A but was still considerably less efficient than wild type RT. To compare the relative mispair extension carried out by wild type and mutant RTs, we quantitated the amount of products formed on G-T template primers by each enzyme. As shown in Fig. 3C, a similar trend to that of the dNTP exclusion assay was observed when comparing the efficiency of Phe⁶¹ mutant RTs in extending the G-T mispair (WT > F61Y > F61W > F61L > F61A). Together, these data suggest that the observed increase in fidelity shown in the dNTP exclusion assay was at least in part due to an increase in the fidelity of mispair extension of Phe⁶¹ mutant RTs.

View larger version (41K): [in this window] [in a new window]   Fig. 3.   Mispair extension assay of WT and Phe61 mutant RTs. The reactions were carried out with both a correctly base-paired G-C (A) and an incorrectly base-paired G-T template primer (B) as described under "Experimental Procedures." Each set of reactions contained two different concentrations (2 and 4 nM) of WT or Phe61-substituted RT. The first lanes of each panel are reactions without enzyme. Both the position of unextended primer (P) and full-length extension product (+10) are indicated on the left side of each panel. C, relative mispair extension fidelities of wild type and Phe61-substituted RTs. The amount of mispair extension products formed on a G-T template primer terminus was quantitated as described under "Experimental Procedures." The total intensity was calculated as the unextended primer band plus all extension products above, whereas extension was calculated as the sum of all products above the primer band. The percentage of mispair extension past as a fraction of the total intensity is shown for wild type and Phe61 mutant RTs. All of the results represent the means ± S.E. of three independent experiments.

F61A Substitution Leads to a Large Increase in Overall Fidelity-- We have previously shown that mutations known to increase the fidelity of dNTP misinsertion or mispair extension often fail to affect the overall fidelity (29). Therefore, we wished to determine whether increases in misincorporation and mispair extension fidelities observed for the Phe⁶¹ mutants translate into increases in overall fidelity. From the single dNTP exclusion and the mispair extension assays, it appeared that the F61A substitution led to the highest increase in dNTP insertion and mispair extension fidelities of HIV-1 RT. Therefore, we tested the overall mutation rate of the F61A mutant in an M13-based forward mutation assay with lacZ as a reporter (27). Two independent gap-filling DNA synthesis reactions were performed with purified wild type and F61A HIV-1_HXB2 RTs. The ability of both RTs to synthesize across the entire gap was confirmed by agarose gel electrophoresis (data not shown). Circular double-stranded M13 DNA was electroporated into host E. coli and plated along with the -complementation strain of E. coli and the resulting plaques screened for -galactosidase activity. The mutations generated during DNA synthesis across the gap were phenotypically scored by counting the number of plaques with an altered color phenotype (light blue to clear). The mutation frequencies were calculated as the ratio of mutant plaques scored by the total number of plaques screened, and then the background mutation frequency was subtracted. The background-adjusted overall mutation frequencies for the wild type and F61A RTs were determined to be 9.7 × 10³ and 0.83 × 10³, respectively (Table II). Therefore, the overall polymerase fidelity of the F61A mutant RT was ~12-fold higher than wild type RT. These data corroborate the results on misincorporation and mispair extension fidelities, showing a significant increase in polymerase fidelity of Phe⁶¹ mutant RTs. We note that the background values measured for the two preparations of the gapped duplex used for wild type (1.7 × 10³) and the mutant (2.5 × 10³), respectively, were both higher than the background-corrected frequency of the mutant (0.83 × 10³). We believe that these values are significant, because the background mutation rates we measure are consistently within the above range, and the mutant frequencies were derived from a rather large denominator (~40,000 plaques).

Sensitivity of Wild Type and Phe⁶¹-substituted HIV-1 RTs to NRTIs-- The dNTP analog inhibitors of HIV-1 RT compete with the natural dNTP substrate of RT, blocking primer extension following their incorporation. Our results show that mutations at Phe⁶¹ affect the dNTP-binding pocket in a manner that renders RT selective against the misincorporation of incorrect dNTPs (Figs. 2 and 3 and Table II). Therefore, we were interested in determining whether Phe⁶¹ mutations would affect the ability of HIV-1 RT to discriminate against another class of "incorrect" dNTP substrates, the nucleoside analog RT inhibitors. As shown in Table III, RTs carrying mutations at Phe⁶¹ were less sensitive to inhibition by both ddTTP and d4TTP compared with wild type RT. The F61A mutant was highly resistant to both drugs, displaying a greater than 70-fold increase in IC₅₀ to ddTTP and nearly a 60-fold increase in the IC₅₀ to d4TTP. The F61L mutant was somewhat less resistant than F61A to inhibition by the NRTIs tested; however, it displayed a significant 40-fold increase in the IC₅₀ to ddTTP. Both F61Y and F61W mutant RTs were also somewhat resistant to inhibition by both NRTIs, displaying nearly a 2-5-fold increase in their IC₅₀ values. These data indicate that similar to their effect on the polymerase fidelity of HIV-1 RT, substitutions at Phe⁶¹ lead to an increased ability of the enzyme to discriminate between dNTP substrates versus NRTIs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In this study, we carried out a structure-based mutagenesis of Phe⁶¹ in HIV-1 RT to determine whether proposed contacts between this residue and the extended single-stranded template play a functional role in polymerase fidelity and sensitivity to nucleoside analog drugs. A limited vertical scanning mutagenesis (Phe  Ala, Leu, Trp, or Tyr) at position 61 of HIV-1 RT was used to evaluate any influence on enzyme function. First, most substitutions at Phe⁶¹ had a minimal effect on the RNA-dependent DNA polymerase and DNA-dependent DNA polymerase activities of HIV-1 RT, although the F61A variant was considerably less efficient when using a heteropolymeric RNA template (Fig. 1 and Table I). Second, measurements of misincorporation and mispair extension fidelity of wild type and mutant RTs indicated that all substitutions at Phe⁶¹ led to a higher fidelity, with less conserved substitutions, e.g. F61A and F61L, displaying the greatest increase (Figs. 2 and 3). The increase in fidelity of F61A was further corroborated by a nearly 12-fold higher fidelity observed in the forward mutation assay (Table II). Third, along with an increase in fidelity, Phe⁶¹ variants were significantly less sensitive to inhibition by NRTIs tested (Table III). Therefore, we conclude that Phe⁶¹, located within the 3 strand of the finger subdomain of HIV-1 RT, plays an important role in determining the fidelity of DNA synthesis and sensitivity to NRTIs.

The effect of substitutions at Phe⁶¹ on polymerase fidelity was determined using two gel-based assays measuring misincorporation and mispair extension efficiencies in addition to an M13-based forward mutation assay used for F61A mutant. Does the dNTP exclusion assay reflect the true fidelity of RT? Two issues need to be addressed. First, it may be argued that the lowered synthesis in the absence of a single dNTP is due to a reduced enzymatic activity. However, as seen in Fig. 1, the relative activities of F61L, F61W, and F61Y in DNA-dependent DNA polymerase assays are similar despite wide variations in their relative fidelities in assays lacking dATP, dCTP, and dGTP, respectively (Fig. 2B). Even though F61A is the only mutation that led to reduced DNA-dependent DNA polymerase activity (Fig. 1), all substitutions at Phe⁶¹ appear to reduce the ability of HIV-1 RT for misincorporation and mispair extension (Figs. 2 and 3). Therefore, the lack of extension products past the barrier site was likely not due to a decrease in catalytic efficiency. Furthermore, the control reactions in the dNTP exclusion (Fig. 2A, panel labeled All dNTPs) and mispair extension (Fig. 3; properly base-paired G-C template primer) assays each demonstrate that all enzymes were used at input levels capable of extending similar amounts of primer under the reaction conditions used. Thus, variations in the specific activities of the mutant RTs did not contribute to differences in misincorporation. Second, it is possible that the products resulting from synthesis beyond the barrier site represent those that are extended by the missing dNTP, which is often present at contaminating levels in the remaining three dNTP preparations used in the reaction. It is known that the addition of different dNTPs to a 5'-end labeled primer will result in different degrees of electrophoretic mobility shifts and thus can be easily verified by denaturing PAGE (10, 13). As shown in those earlier studies by our laboratory, the purity of the dNTPs used in our reactions is high enough that contaminating dNTPs are not detected in primer extension assays.

The first nucleoside analog resistance mutations shown to increase dNTP misinsertion fidelity, E89G and M184V, both displayed no significant difference from wild type in their overall mutation frequency (10, 13, 29). However, variants of HIV-1 RT containing several other RT mutations, M184I, K65R, L74V, D76V, and R78A, displayed a positive correlation between changes in fidelity based on the dNTP misincorporation assays (14, 15) and in the overall fidelity of RT (16, 19, 20, 30, 31). It is important to point out that four of the five mutations for which such positive correlation was previously demonstrated map to the finger domain, as does the F61A mutation, which also shows increased fidelity both in misincorporation assays and in forward mutation assays.

The crystal structure of HIV-1 RT in a covalently trapped RT double-stranded DNA template primer-dNTP ternary complex shows a movement of the finger subdomain toward the polymerase active site (17). As a result, a "closed" complex is formed, with several residues previously implicated in NRTI resistance brought into direct contact with the incoming dNTP. Similar to other polymerase structures solved (32-34), the single-stranded template 5'-overhang is bent away from the helical path of the duplex, making contact with several amino acid residues on the outer surface of the finger subdomain of HIV-1 RT. Two such residues that make contact with the first nucleotide of the extended template (the templating nucleotide), Asp⁷⁶ and Arg⁷⁸, were previously shown to be involved in polymerase fidelity and processivity (19, 20). In addition to crystallographic and mutagenesis data, biochemical evidence has indicated that the finger subdomain makes contacts with the extended template ahead of the polymerase active site. For example, footprinting studies of HIV-1 RT-template primer complexes using DNase I and hydroxyl radicals have shown areas of protection on the template strand from +7 to 23 and +3 to 15, respectively (35, 36). The affinity of HIV-1 RT for the template-primer was greatly increased with templates extended at least four nucleotides 5' to the site of dNTP incorporation (37, 38). Cross-linking experiments have also demonstrated that several residues within the finger subdomain contact the extended template (39).

Earlier work has documented the influence of residues contacting the template primer duplex upstream of the primer 3'-hydroxyl terminus on the geometry of the dNTP-binding pocket (10, 18, 40). The results reported here show for the first time that mutations at residues that may be contacting the template overhang also influence the geometry of the dNTP-binding pocket. In fact, the level of resistance observed here (60-72-fold) is one of the highest reported for substitutions at template-contacting residues.

Based on both previous work and our current results, we propose that upon closing the finger subdomain with bound dNTP into a closed complex, contacts are formed between Phe⁶¹ and both the first and second nucleotide of the extended template. Presumably, these contacts between the wild type Phe⁶¹ residue and the extended template contribute to an error-prone polymerase active site capable of accepting incorrect dNTPs and NRTIs. If Phe⁶¹ plays a role in the observed bending of the template out of its helical path, this may provide a more flexible dNTP-binding pocket. In the case of the F61A mutant, contacts with the +1 and +2 template position may be lost; bending of the template may be reduced, resulting in an altered template path and ultimately a more stringent dNTP-binding pocket. Recent work suggests that the size and shape complementarity, rather than the ability of dNTPs to hydrogen bond to the template nucleotide, are responsible for the high fidelity of DNA replication (41, 42). Therefore, based on our results, we propose that Phe⁶¹-template contacts contribute significantly to a "loose" dNTP-binding pocket that enables HIV-1 RT to both misinsert incorrect dNTPs and extend mispaired primer termini, leading to an overall low polymerase fidelity and sensitivity to NRTIs. Previously solved structures of HIV-1 RT in complex with a template-primer contained a template with only a single nucleotide overhang, and therefore it was not possible to determine whether Phe⁶¹ interacts with the extended template in this conformation. However, contacts between HIV-1 RT and the template primer would likely affect the ability of RT to discriminate between correct and incorrect dNTPs upon dNTP binding or after misinsertion takes place. Therefore, contacts between Phe⁶¹ of HIV-1 RT and the extended template in the closed RT-template primer-dNTP ternary complex are likely to be functionally important in determining polymerase fidelity. Whether the F61A mutation results in a complete loss of this contact with the template, a change in the geometry of the dNTP-binding pocket, or an altered trajectory of the extended template requires further crystallographic analysis of this mutant RT-template primer-dNTP complex.

The Phe⁶¹ residue of HIV-1 RT is highly conserved among several groups of retroviral RTs (43). Phe⁶¹ is likely to have been retained because it supports both catalytic activity and error-prone DNA synthesis. The error-prone nature of HIV-1 RT, in combination with a robust rate of virus replication enables the virus to evade both drug and immune selection in vivo. The F61A and F61L mutations, both of which confer significant resistance to NRTIs, have not been found in drug-resistant clinical isolates of HIV-1. As described in this report, both of these mutants displayed a significant decrease in the catalytic efficiency on a heteropolymeric RNA template (Table I) and may contribute to a reduction in the rate of replication and/or viral fitness.

Furthermore, both F61A and F61L mutant RTs also displayed significantly lower polymerase processivity compared with the wild type enzyme (data not shown). It must be cautioned that a decreased sensitivity of these mutants to NRTIs may also be due to increased dissociation of RT from chain terminated primer termini. Measurements of processivity of Phe⁶¹ mutants³ revealed that the dissociation from template primer correlates with decreased susceptibility to NRTIs. Whether the mutations have dual effect on both the geometry of dNTP-binding pocket as well as an indirect effect because of RT dissociation needs to be further examined. A loss in RT processivity has been previously proposed to be the primary reason for the lower fitness of some of the NRTI resistance viruses (44, 45). Similarly, the D76V and R78A mutations have not been isolated in vivo. The only mutations observed in vivo that have been shown to increase HIV-1 RT overall polymerase fidelity are K65R and M184I (16, 30). K65R was shown to not affect viral replication (46), whereas M184I appears only transiently during treatment with ()2',3'-dideoxy-3'-thiacytidine, likely because of a lower viral fitness (44). Therefore, both F61A and F61L mutations are likely not found in HIV-1 isolates because of a loss in enzyme function and viral fitness. However, using complementary approaches, such as structure-based mutagenesis and E. coli complementation, mutants of HIV-1 RT have been isolated with altered fidelity and drug sensitivity. Such mutants should prove useful in determining the molecular determinants of HIV-1 RT fidelity and NRTI resistance, thereby helping to prevent resistance to currently used drugs and in the development of novel antiretroviral therapies.

	ACKNOWLEDGEMENTS

We thank William Drosopoulos for reading the manuscript and Roopa Narasimhaiah and Kenneth Curr in help with the preparation of the gapped duplex DNA and in generating the wild type forward mutation data.

	FOOTNOTES

^* This work was supported by Public Service Grant RO1 AI30861 (to V. R. P.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Supported by National Institutes of Health Predoctoral Training Grant T32-GM07491.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom correspondence should be addressed: Dept. of Microbiology and Immunology, Albert Einstein College of Medicine, 1300 Morris Park Ave., Rm. GB 401, Bronx, NY 10461. Tel.: 718-430-2517; Fax: 718-430-8976; E-mail: prasad@aecom.yu.edu.

Published, JBC Papers in Press, April 10, 2002, DOI 10.1074/jbc.M200282200

² T. S. Fisher and V. R. Prasad, manuscript in preparation.

³ T. S. Fisher and V. R. Prasad, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: HIV, human immunodeficiency virus; HIV-1, HIV type 1; RT, reverse transcriptase; WT, wild type; ddTTP, 2',3'-dideoxythymidine triphosphate; NRTI, nucleoside analog RT inhibitor; d4TTP, 2',3'-dideoxy,2'3'-didehydrothymidine triphosphate; nt, nucleotide; PBS, primer-binding site.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us