Mechanism and fidelity of HIV reverse transcriptase.

THEJOURNAL OF BIOLOGICAL CHEMISTRY Q 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 36,Issue of December 25, pp. 25988-25997.1992 Printed in U.S.A.

Mechanism andFidelity of HIV Reverse Transcriptase* (Received for publication, May 18, 1992)

Warren M. Kati andKenneth A. Johnson$ From the Department of Mokculur and Cell Biology, The Pennsylvania State University, University Park, Pennsylvania 16802

L. Fred Jerva and KarenS. Anderson$ From the Department of Pharmacology, Y a k University School of Medicine, New Haven, Connecticut 06510

We have examined the RNA-dependent and DNAdependent polymerase and ribonuclease H catalytic activities of human immunodeficiency virus reverse transcriptase using rapid transient kinetic methods with defined synthetic25/45-mer DNA/RNA and DNA/DNA primer/templates. The &value for interaction of the enzyme with duplex DNA was 4.7 nM, and the value for RNA/DNA heteroduplex was of similar magnitude. A pre-steady state burst of nucleoside triphosphate incorporation was observed for both DNA and RNA templates. Analysis of the dATP concentration dependence of the burst rate provided Kd values for dATP of 4 and14 I.IM and maximum rates of single of 33 and 74 s-l, for nucleotide incorporation, b,, DNA and RNA templates, respectively. Subsequent turnovers were limited by the rate of dissociation of the primer/template from the enzyme at rates of 0.18 and 0.06 s-’ for duplex DNA and RNA/DNA heteroduplex, respectively. Analysis of rates of DNA polymerization and RNA cleavage using the RNA template revealed that thetwo activities are independent of one another. The polymerization rate (4-70 s-’) was dependent on dATP concentration, whereas the RNA cleavage occurred at a constant rate of 10 s” over the 100-fold dATP concentration range (2-200 p ~ ) . Examination of the RNA cleavage products resulting from a single turnover indicates that the polymerase and ribonuclease domains of the enzyme are separated b y a distance corresponding to 19 bases of RNA/DNA heteroduplex, consistent with the recently published crystal structure (Kohlstaedt, L. A., Wang, J., Friedman, J., Rice, P. A., and Steitz, T. A. (1992) Science 266, 1783-1790). Analysis of the kinetics of processive synthesis suggested that the initial binding of dNTP leads to a faster rateof dissociation of DNAfrom the enzyme. Further investigation supported a twostep dNTP binding mechanism with the formation of an initial E*DNA*dNTPcomplex followed by a more stable E’*DNA*dNTPcomplex. The Kd values for incorporation of incorrect nucleoside triphosphates opposite a DNA template thymidine were 1010 pM for dGTP, 1240 I.IM for dCTP, and 840 PM for dTTP. The 1 rates were 4.8 s-’ for corresponding maximum b

* This work was supported by National Institutes of Health (NIH) Grant GM44613 (to K. A. J.) andNIH Postdoctoral Fellowship GM14469 (to W. M. K.). 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 solelyto indicate this fact. $ To whom correspondence should be addressed. K. A. J.: Dept. of Molecular and Cell Biology, 106 Althouse Laboratory, The Pennsylvania State University, University Park, Pennsylvania 16802; K. S. A.: Dept. of Pharmacology, 333 Cedar St., Yale University School of Medicine, New Haven, CT 06510.

dGTP, 0.52 s-l for dCTP, and 0.41 s” for dTTP.These values provide fidelity estimates of 1740 for discrimination against dGTP, 19,700 for dCTP, and 16,900 for dTTP misincorporations at this site.

Virally encoded reverse transcriptase (RT)’ is required for replication of the human immunodeficiency virus (HIV), an etiological agent for acquired immunodeficiency syndrome (AIDS) (Barre-Sinnoussi et Q L , 1984; Popovic et Q L , 1984). The RT enzyme has three distinct catalytic activities which enable itto convert the viral RNAgenome into doublestranded DNA 1) RNA template-dependent DNA polymerization, 2) RNase H degradation of the RNA, and 3) DNA template-dependent DNA polymerization (Goff, 1990). The native enzyme exists as a heterodimer which contains poly, peptides of 66 and 51 kDa (DiMarzo-Veronese et ~ l . 1986; Lightfoote et Q L , 1986). The polymerase domain has been shown to be associated with the amino terminus of one or both subunits, while the RNase H activity has been shown to be associated with the carboxyl-terminal domain of only the 66-kDa polypeptide (Hansen et Q L , 1988). A recent crystal structure indicates that theheterodimer possess only one open polymerase site residing on the 66-kDa polypeptide adjacent to the RNase H site; the putative polymerase site on the 55kDa polypeptide is closed by a conformational collapse of the structure and movement of a domain into the polymerase site (Kohlstaedt et ~ l . 1992). , RT is considered to be the molecular target of several therapeutic agents that interfere with HIV replication, including the 5’-O-triphosphates of 3’-azido-3’-deoxythymidine (AZT, Furman et ~ l . 1986) , and 2’,3’-dideoxyadenosine and 2’,3’-dideoxycytidine (Mitsuya et ~ l . 1987; , Chen and Oshasa, 1987). The rational design of moreeffective and specific inhibitors of RT aspotentialanti-AIDS drugs should be facilitated by a complete kinetic and thermodynamic description of the RT enzymatic reaction mechanism. Transient state kinetic methods are required for investigating a complex enzymatic mechanism, because these methods permit the direct measurement of the individual reactions occurring within the enzyme’s active site. In particular, rapid chemicalquench-flow techniques have successfully solvedthe complete kinetic mechanism for T7 DNA polymerase (Pate1 et ~ l .1991; , Wong et ~ l . 1991) , and theKlenow fragment from Escherichi~ coli DNA polymerase I (Kuchta et Q L , 1987). In the present study we have examined all three catalytic The abbreviations used are: RT, reverse transcriptase; dNTP, deoxynucleoside 5“triphosphate; dATPaS, deoxyadenosine 5’-0-(3a-thi0)triphosphate; HIV-1, human immunodeficiency virus type 1; AIDS, acquired immunodeficiency syndrome; HPLC, high-performance liquid chromatography.

25988

Reverse Pathway Kinetic Transcriptase

25989

activities of HIV-1 RT using rapid transient kinetic methods. Biospin 30 columns were from Bio-Rad, and the DE81 filters (2.5 The results provide a quantitative descriptionof the interac- cm) were from Whatman. Synthetic Oligonucleotides-TheDNA oligonucleotides (25- and tions of D N A / R N A and D N A / D N A primer/templates as well 45-mer) shown in TableI were synthesized on an Applied Biosystems as with nucleotide triphosphates. We also begin to address 380A DNA synthesizer (DNA synthesis facility, Yale University) and the question of how the polymerase and ribonuclease catalytic purified using denaturing polyacrylamide gel electrophoresis (20% activities are coordinated. acrylamide, 8 M urea). MATERIALS AND METHODS

The RNA oligonucleotide (45-mer, Table I), also synthesized at Yale University, was obtained protected as the t-butyldimethyl silyl ether at the 2"ribose position. Deprotection and purification were carried out according to Webster and Spicer (1990). This involved deprotection with tetrabutylammonium fluoride (Aldrich Chemical Co.), chromatography on a Vydac C4-RP HPLC column (4.6 mm X 25 cm) to remove partially deprotected oligonucleotide and purification using denaturing polyacrylamide get electrophoresis (20% acrylamide, 8 M urea). The duplex 25/45-mer primer/templates of DNA/RNA and DNA/ DNA were formed by annealing approximately equimolar pure 25and 45-mer at 80 "C for 4 min and 50 "C for 30 min. To ensure that proper annealing had taken place the duplex mixtures were analyzed by nondenaturing polyacrylamide gel electrophoresis (20%). Concentrations of the oligonucleotides were estimated by UV absorbance at 260 nm using the following calculated extinction coefficient: DNA 25-mer, e = 249,040; DNA 45-mer, t = 491,960; and RNA 45-mer, t = 507,960. Buffers-All experiments using HIV RT were carried out in50 mM Tris-C1,50 mM NaCl buffer a t pH 7.5 and ata temperature of 37 "C. All experimental procedures were carried out using sterile buffers, reagents, and glassware where feasible. 5'-32P-Labelingof 25/45-mers-Before annealing both primer and template strands of the DNA/RNA and DNA/DNA 25/45-mer were 5'-radiolabeled with T4 polynucleotide kinase (New England Biolabs) using the procedure described in Maniatis et al. (1982). The reaction was quenched with 0.1 M EDTA followedby purification using a Biospin-30 column whichremoved the contaminating nucleotides from the labeled primer/template. Rapid Quench Experiments-Rapid quench experiments were carried out in an apparatus designed by Johnson (1986) and built by KinTek Instruments (University Park, PA). The apparatus contained a computer-controlled stepping motor and was modified for using small reaction volumes (15 p l ) . Typically, the experiments were carried out by loading the enzyme in one loop (15 p l ) and substrate in the second loop (15 pl) of tubing. The reactions were started by rapidly mixing the two reactants and then quenched with 0.3 M EDTA (final concentration) after time intervals ranging from 3 ms to several seconds. All concentrations reported are final concentrations after mixing in the rapid quench apparatus. Product Analysis-The products were analyzed by sequencing gel electrophoresis (16% polyacrylamide gel) under denaturing conditions with 8 M urea. The product and thesubstrate DNAs and RNAs were quantitated by either scanning the dried gel using a Betascope and/ or by visualizing the radioactive bands by autoradiography followed by cutting the bands from the gel and counting in a scintillation counter. Data Analysis-Data were fitted by nonlinear regression using the program RS 1(BBN Software Products Corp., Cambridge, MA). Data from the active site titration (see Fig. 2) was fitted to the quadratic Eo + Do)equation: [E.DNA] = 0.5 (Kd Eo Do) - 0.5((& 4EoDo)". The kinetic data for incorporation of correct nucleotides into duplex DNA were modeledusing the KINSIM kinetic simulation program provided by Carl Frieden and Bruce Barshop, Washington University, St Louis, MO (Barshop et al., 1983) as modified by Anderson et al. (1988) to accept x-y data pairs. Final fitting of the data was accomplished by nonlinear regression based upon kinetic simulation using a modification of the program FITSIM (Zimmerlie and Frieden, 1989).

Overexpression and Purification of Recombinant HIV RT-Two E. coli cell lines containing plasmids for the expression of the 66- and 51-kDa subunits of HIV-1 R T (Hizi et al., 1988) were obtained from Dr. Stephen Hughes a t the National Cancer Institute-Frederick Cancer Research Center. The 51-kDa subunit contains a modified carboxyl-terminal sequence that improves overexpression and stability. Thetwo cell lines were grown separately in Superbroth (Clark et al., 1990) containing ampicillin (50 pg/ml) a t 37 "C with shaking for 12 h to an optical density of 1.5. After harvesting by centrifugation, the two cell lines were resuspended separately in Buffer A (50 mM Tris-C1, pH 8.0,2 mM EDTA, 0.1% (3-mercaptoethanol,1mM phenylmethylsulfonyl fluoride, and 10% glycerol) which also contained 0.5 M NaCl. The cell lines were combined proportionally based on an sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis. The combined cells were lysed by treatment with lysozyme (0.3 mg/ ml) at room temperature for 15 min, followed by the addition of Triton X-100 to 0.1% and mild sonication. The purification procedure was performed on ice or a t 4 "C and is outlined as follows: 1) precipitation of nucleic acids by addition of 5% poly(ethy1enimine) t o 0.3% and centrifugation; 2) ammonium sulfate fractionation (060%); 3) phosphocellulose chromatography using a linear gradient of 0-1 M NaCl in Buffer A; 4) single-stranded DNA cellulosechromatography (Sigma) using a linear gradient of 50-1000 mM NaCl in Buffer A, 5) Q-Sepharose chromatography using a linear gradient from 10 t o 250 mM NaCl in Buffer A. The final pooled fractions were concentrated and then diluted with one volume of Buffer A containing50% glycerol and stored at -20 "C (Mizrahi et al., 1989). The ratio of the Coomassie Blue staining intensities of the two bands suggested a finalsubunit composition of1.0:1.24 (66/51 kDa). The enzyme concentration was determined by using extinction coefficients of 136,270 and 124,180 for the 66- and 51-kDa subunits, respectively, a t 280 nm (calculated from the amino acid composition). Enzyme concentrations were corrected for the fraction of active protein based upon active sitetitrations, as described under "Results." Several preparations gave a reaction amplitude of approximately 0.5 sites per 66/51 heterodimer. We also examined the single turnover kinetics using a preparation of RT prepared by Muller et al. (1989). Because the active site titration based upon single turnover kinetics (see below) indicated that only 50% of the enzyme in our preparation was active, it was important toexamine enzyme prepared by other methods. A sample of enzyme was generously provided by Barbara Muller and Roger Goody (Max-Planck Institute,Heidelberg), prepared from their clone which provides the expression of both the large and small subunit within the same cell (Muller et al., 1989). Active site titrationsof two preparations made by this method and shipped on dry ice from Germany gave amplitudes lower than our preparations. A fresh preparation made in our laboratory using a clone provided by Roger Goody gave a burst amplitude of approximately 45%. All of the experiments reported inthis paper were performed with protein purified using the clones obtained from Stephen Hughes as outlined above. While this paper was under review, a report of the HIV R T presteady state kinetics appeared (Reardon, 1992)examining the kinetics of incorporation of dTTP into a DNA/RNA heteroduplex (d21:r44mer). An apparent burst amplitude of 1 per RT dimer was observed. Reardon used a significantly higher extinction coefficient for RT (334,000 for the dimer) than was used in our study, but thisdifference alone can only account for 20%. We have seen no significant change RESULTS in extinction coefficient in the presence of urea and our concentration measurements agree with those by Muller et al. (1989) based upon a of gravimetric standard. Further experiments are underway to examine Steady State and Pre-steady State Polymerization Kinetics d A T P Incorporation changes in the methods of purification and other variables that might affect the active site titration amplitude in our preparations. In the D N A as a Substrate-The steady state kinetics of elongapresent work, we have used the active site titration to establish the tion of the 25/45-mer duplex D N A to a 26/45-mer product concentration of active enzyme reported for each experiment. Nucleotide Triphosphates and Other Materials-All four dNTPs were measured using a large excess of 25/45-mer (200 nM) were purchased from Pharmacia LKB Biotechnology Inc. [ T - ~ ~ P ] A T Prelative to RT (0.8 nM) and including a saturating concentrawas purchased from Du Pont-New England Nuclear Labs or ICN. tion (see below) of d A T P (25 PM) in the reactionmixture.

+ +

+

Pathway

Kinetic Transcriptase Reverse

25990

TABLEI 25/45-mer

DNA

25/45-mer

RNA

Oligonucleotides GCCTCGCAGCCGTCCAACCAACTCA CGGAGCGTCGGCAGGTTGGTTGAGTTGGAGCTAGGTTACGGCAGG GCCTCGCAGCCGTCCAACCAACTCA CGGAGCGUCGGCAGGUUGGUUGAGUUGGAGCUAGGUUACGGCAGG

-3

n

FIG. 1. Steady state and presteady state kinetics of incorporation of dATP. A , the steady state kinetics of 25/45-mer duplex DNA (Table I) elongating to 26/45-mer were measured by preincubating25/45-mer (200 nM) with R T (0.8 nM active site concentration) and then starting the reactions bytheaddition of Mg+' (10 mM) and dATP (25 WM). Reactions were terminated by theaddition of 0.3 M EDTA (final concentrations) and the product 26/45-mer DNA was quantitated by sequencing gel analysis (16%acrylamide, constant rate 8 M urea). state The steady was calculatedfrom the slope of the graph divided by the enzyme concentra(established tion by active titration, site see Fig. 2) tovalue give a of 0.18 s-'. B, the pre-steady state kinetics for incorporation into theduplex DNA substrate were measured by mixing apreincubated solution of R T (44 nM active site concentration) and DNA/DNA25/45-mer (200 nM) with Mg+' (10 mM) and dATP (10 PM) under rapid quench conditions. The reactions were quenched and quantitated as above. The data ( 0 )were fittedto a burst equation with rate constants equal to 20 and 0.18 s-', for the exponential and linear phases,respectively. The solid line was calculated from a computer simulationtothemechanism shown in Scheme I. C, pre-steadystateburstexperimentin whicha preincubated s o h tion of R T (100 nM active siteconcentration) and 25/45-mer DNA/RNA heteroduplex (300 nM) was mixed with Mg+' (10 mM) and dATP (100 p ~ ) The . rate for the exponential phasewas 66 s-' and the slower linear phase was 0.06 s-'.

5

60

I-

= * I w 0 30

z

8

2o

TIME (secs)

1

0

L

2

3

5

4

TIME (secs)

c

1 4 -120

x

-

C

'100

z