ANALYTICALBIOCHEMISTRY

Analytical Biochemistry 331 (2004) 296–302

www.elsevier.com/locate/yabio

A capillary electrophoretic assay for ribonuclease H activity

King C. Chan,a Scott R. Budihas,b Stuart F.J. Le Grice,b Michael A. Parniak,c

Robert J. Crouch,d Sergei A. Gaidamakov,d Haleem J. Isaaq,a Antony Wamiru,e James B. McMahon,e and John A. Beutlere,¤

a Laboratory of Proteomics and Analytical Technologies, SAIC-Frederick, Inc., Center for Cancer Research, NCI at Frederick, Frederick, MD 21702, USA

b HIV Drug Resistance Program, Center for Cancer Research, NCI at Frederick, Frederick, MD 21702, USAc Division of Infectious Diseases, School of Medicine, University of Pittsburgh, Pittsburgh, PA 15213, USA

d Laboratory of Molecular Genetics, National Institute of Child Health and Human Development, Bethesda, MD 20892, USAe Molecular Targets Development Program, Center for Cancer Research, NCI at Frederick, Frederick, MD 21702, USA

Received 20 February 2004Available online 17 June 2004

Abstract

A capillary electrophoretic assay was developed to measure the ribonuclease (RNase) H activity of human immunodeWciencyvirus (HIV) type 1 reverse transcriptase. Cleavage of a Xuorescein-labeled RNA–DNA heteroduplex was monitored by capillaryelectrophoresis. This new assay was used as a secondary assay to conWrm hits from a high-throughput screening program. SinceautoXuorescent compounds in samples migrated diVerently from both substrate and product in most cases, the assay was extremelyrobust for assaying enzymatic inhibition of such samples, in contrast to a simple well-based approach. The assay was broadly appli-cable to other RNases H, speciWcally those from human, Escherichia coli, and HIV-2, although product proWles varied for eachenzyme.Published by Elsevier Inc.

Keywords: Capillary electrophoresis; Ribonuclease H; HIV-1; Reverse transcriptase; Fluorescence

Human immunodeWciency virus-type 1 (HIV-1)1

reverse transcriptase (RT) is a dual function enzyme cat-alyzing both DNA polymerization and ribonuclease(RNase) H activities. The RNase H domain catalyzescleavage of the RNA strand contained in an RNA/DNAduplex; this is an important step in the process of reversetranscription. While both enzymatic activities are criticalfor infectivity [1], only the polymerase activity has previ-ously been successfully exploited as a target for commer-cial antiviral drugs. The investigation of RNase H as atherapeutic target has been reviewed previously [1–3].

¤ Corresponding author. Fax: +1-301-846-6177.E-mail address: jb123w@nih.gov (J.A. Beutler).1 Abbreviations used: HIV-1, human immunodeWciency virus-type

1; RT, reverse transcriptase; RNase, ribonuclease; FRET, Xuorescenceresonance energy transfer; CE, capillary electrophoresis; MALDI-TDF, matrix-assisted laser desorption ionization-time of Xight; ACN,acetonitrile.

0003-2697/$ - see front matter. Published by Elsevier Inc.doi:10.1016/j.ab.2004.05.017

Reasons that this target has not been eVectivelyexploited include the facts that previous assays forRNase H have been slow, cumbersome, or expensive toconduct.

Our laboratories are involved in the identiWcation ofcompounds that inhibit the nonspeciWc HIV-1 RNase Hactivity of HIV-1 RT. We recently described a homoge-neous Xuorescence resonance energy transfer (FRET)quenching assay designed for high-throughput screeningof inhibitors of HIV-1 RNase H and have used this forlarge-scale screening of small-molecule libraries [4].BrieXy, the FRET assay uses an 18-nucleotide-3�-FAM-labeled heteropolymeric RNA annealed to a comple-mentary 5�-dabcyl-labeled DNA. RT-associated RNaseH cleaves the RNA strand four nucleotides from the 3�end. The labeled tetranucleotide dissociates from theDNA and a Xuorescent signal is detected by a conven-tional Xuorescence plate reader.

K.C. Chan et al. / Analytical Biochemistry 331 (2004) 296–302 297

Due to the nature of the primary screening assay itwas not clear that we would be able to distinguish truehits from small molecules that did not inhibit RT butwould quench the FAM signal. This was of particularconcern for the screening of natural product extracts.Our solution was to develop a capillary-electrophoresis(CE)-based assay that allowed for quantitative and qual-itative analyses in a moderate-throughput format, anextension of the well-based assay in which the substratelacks the dabcyl quencher and its cleavage product isseparated by CE. This CE approach has been validatedand used as a moderate-throughput secondary screeningassay. We also present its use for RNases H from otherdiverse sources. The assay was adopted as a “gate-keeper” assay for conWrmation of hits identiWed in theFRET assay, since it has the ability to separate Xuores-cent or Xuorescence quenching sample compounds fromthe assay components, thereby yielding a more reliablemeasurement of enzyme inhibition.

Capillary electrophoretic techniques have been previ-ously applied to enzymatic assays of proteases [5] andkinases [6,7]. With currently available automation it ispossible to substantially increase throughput in this typeof assay by using a 96-capillary-array electrophoreticinstrument [7]. The sensitivity, speed, and potential forminiaturization have made CE an attractive approachfor conducting enzymatic assays [8].

Methods

Reagents

The oligonucleotides 5�-GAU CUG AGC CUGGGA GCU-Xuorescein-3� and 5�-AGC TCC CAG GCTCAG ATC-3� were synthesized and provided as theannealed RNA/DNA hybrid by TriLink Biotechnolo-gies (San Diego, CA). The RNA oligo fragmentsAGCU-Xuorescein, GCU-F, CU-F, and U-F were alsoobtained from TriLink and were used without furtherpuriWcation. Recombinant wild-type p66/p51 HIV-1 RTwas expressed and puriWed as described [9]. HIV-2 RTwas prepared using the same Escherichia coli expressionsystem [9]. E. coli RNase HI was prepared as previouslydescribed [10], as was recombinant human RNase H1[11].

Samples for assay were prepared in 96-well blackpolystyrene plates in a total volume of 100 �l. Stock solu-tions of the substrate and HIV-1 RT were diluted to theappropriate concentration immediately before use, sinceextended dilution reduced enzymatic activity. Fiftymicroliters of a 0.4 �M solution of RNA/DNA hybrid in50 mM Tris, pH 8.0, containing 60 mM KCl and 10 mMMgCl2 was added to individual wells of the microplateusing a Beckman BioMek FX. Reactions were initiatedby addition of 50 �l of 13.6 nM HIV-1 RT in 50 mM Tris,

pH 8.0, containing 60 mM KCl and 10 mM MgCl2 andallowed to proceed at room temperature for 30 min.Reactions were quenched by the addition of 10 �l of0.5 M EDTA, pH 8.0. To assess the eVect of inhibitors,6 �l of inhibitor in dimethyl sulfoxide was added to themicroplate well after addition of substrate but prior toRT addition. Completed assay plates were stored frozenat ¡20 °C before CE analysis. For HIV-2, E. coli, andhuman RNases H, the reactions were carried out usingWnal enzyme concentrations of 22, 0.9, and 1.6 nM,respectively. These concentrations were empiricallydetermined to give equivalent levels of substrate cleav-age using the FRET quenching assay.

Instrumentation

A Beckman MDQ CE system equipped with a 488nm laser-induced Xuorescence module was used to ana-lyze the samples. Separations were performed with eithera 30- or 50-�m i.d. £ 30-cm capillary (Polymicro) at20 °C. A new capillary was treated by Wlling it with 1 Nsodium hydroxide overnight. The capillary was rinsedwith 1 N NaOH and running buVer for 1 min eachbetween runs. All samples were diluted Wvefold with dis-tilled water before injection. Samples were injected byvacuum (0.5 psi at 5 s) or voltage (5 kV at 5 s) and typi-cally separated by applying a voltage of 15 kV.

MALDI-TOF mass spectroscopy was performed onan Axima-CFR MALDI mass spectrometer (Shimadzu)in reXectron mode. Substrate and product spectra wereobtained by drying the reaction mixtures describedabove after incubation with or without enzyme. Theenzyme concentration was doubled to increase the extentof substrate cleavage. The resulting samples were pro-cessed through C-18 ZipTips (Millipore) according tothe manufacturer’s speciWcations, and their spectra wereobtained, without additional puriWcation, using 3-hydroxypicolinic acid (Fluka) or 6-aza-2-thiothymine(Aldrich) as matrix.

Results

Development of the CE separation

Because the charge-to-mass ratio of short oligonucle-otides such as the substrate and product diVer, they canbe readily separated by CE using either acidic or alkalibuVers without the inclusion of sieving media[12–15]. Analkali buVer was used in this study because acidic buVersquench the Xuorescein Xuorescence, lowering the sensi-tivity of the assay. In addition, by using an alkali buVer,separations could be carried out with an uncoated capil-lary, which would be expected to be more robust in ahigh-throughput screening system employing a 96-capil-lary array. Using a 20 mM sodium tetraborate, pH 9.2,

298 K.C. Chan et al. / Analytical Biochemistry 331 (2004) 296–302

buVer, the RNase H reaction mixtures were separatedinto two major peaks of substrate and product (Fig. 1A).This buVer system was useful for rapid screening, but ahigher resolution was desired to ascertain that there wasno background interference and to better understand thenature of the reaction product(s). When the borate con-centration was increased, the viscosity of the buVer alsoincreased, resulting in enhancement of separation due tothe lower electroosmotic Xow (Fig. 1B). When the boratewas increased to 60 mM or higher, one major and threeminor product peaks were resolved. Although the reso-lution was enhanced by increasing the buVer concentra-tion, the Joule heating also increased. Alternatively, theelectroosmotic Xow could be reduced by using a buVercontaining an organic modiWer and resolution enhancedwithout increasing Joule heating. As exempliWed inFig. 1C, the separations were performed with a 20 mMtetraborate buVer containing 30% acetonitrile (ACN) inwhich the four products were baseline separated fromthe substrate, as was observed with the high-ionic-strength borate buVer. However, the selectivity of thepeaks changed; the substrate peak migrated faster thanthe product peaks, the reverse of that seen with theborate buVer. Since a 20 mM borate buVer containing20% ACN allowed good resolution of the substrate andproducts within 7 min, these were chosen as optimumconditions for this study.

Since there is no discrimination of the origin ofXuorescence signals in a plate assay, any background

Fig. 1. EVect of increasing sodium tetraborate (pH 9.2) concentrationon separation of RNase H substrate (s) and products (p); (A) 20 mMtetraborate; (B) 80 mM tetraborate; and (C) eVect of 30% acetonitrilein 20 mM tetraborate A 30-mm £ 30-cm capillary was used in all cases.Separation voltage in (A) and (B) was 12 kV; in (C) it was 15 kV.

Xuorescence signal can potentially contribute to a falsenegative result. Likewise, a compound that quenchesXuorescence could lead to a false positive result. The useof CE provides the capability to separate interferingXuorophores or quenchers from the substrate and prod-uct peaks. For example, in Fig. 2 the two intense sample-derived peaks are well resolved from the substrate andproduct peaks, leading to a more reliable result.

Voltage eVect

Occasionally, a small ssRNA that migrated fasterthan the substrate peak was observed (peak r in Fig. 3).The identity of the ssRNA was conWrmed by spikingwith a ssRNA standard. This ssRNA peak was usuallysmall, most likely arising from strand separation of theDNA–RNA duplex before or during CE. One variablefor the strand separation was the separation voltage. Incontrast to the oligonucleotide product peaks, the rela-tive magnitude of the substrate and RNA peaks wereobserved to be dependent on the separation voltage.While a higher voltage permits faster separations, thepeak height of the substrate (s) decreased and the ssRNApeak (r) increased when the voltage reached 18 kV (Figs.3A and B). This conversion (denaturation) may be due toa heating eVect inside the column, since at higher temper-atures, the stability of DNA–RNA duplexes willdecrease. The substrate peak completely disappeared atan applied voltage greater than 18 kV. Nevertheless, thissubstrate–RNA peak conversion was not observed witha 30-�m i.d. capillary even at very high voltage (Figs. 3Cand D); presumably a smaller-diameter capillary is moreeYcient in dissipating Joule heating. Therefore, the sumof the peak height (or volume) of the substrate and thepresumed RNA peaks can be used instead of the sub-strate peak alone to accurately measure the extent ofenzymatic activity. Alternatively, by using a higher volt-age, one can observe only the ssRNA substrate peak.This condition would be desirable if higher-speed analy-ses were required using the higher voltage.

Fig. 2. Separation of of RNase H substrate and product peaks in pres-ence of Xuorescent sample components; 20 mM borate buVer, pH 9.2,20% ACN; b, peaks attributed to sample components of extract ofStelletta crater; r, ssRNA peak derived from substrate; s, substratepeak; p, cleavage product peak.

K.C. Chan et al. / Analytical Biochemistry 331 (2004) 296–302 299

Nature of the cleavage product

We sought to conWrm the position of substrate cleav-age by MALDI-TOF mass spectrometry. The uncleavedsubstrate gave the expected mass (6322.4 Da) for theRNA strand. The primary cleavage product observedwith the HIV-1 enzyme (1843.8 Da) was consistent witha 4-nt RNA oligonucleotide when the mass of theXuorophore was taken into account (Fig. 4). The minorproduct peaks corresponded to 3-, 2- and 1-nt RNA spe-cies (1513.3, 1167.2, and 862.2 Da, respectively). It shouldbe noted that the relative peak heights of the major andminor peaks remained constant as the enzymaticreaction progressed, indicating that they were likely tobe produced in tandem by the action of the enzyme [1].

Reproducibility

We found that the reproducibility of the CE assaywas highly dependent on the condition of the capillary.

Fig. 3. Separation of of RNase H substrate and product at diVerentapplied voltages; 20 mM borate buVer, pH 9.2, 20% ACN, 50-�m-diameter capillary; r, ssRNA peak derived from substrate; s, duplexsubstrate; p, cleavage product. (A) 12 kV with 50-�m-diameter capil-lary; (B) 18 kV with 50-�m-diameter capillary; (C) 24 kV with 30-�m-diameter capillary; (D) 30 kV with 30-�m-diameter capillary.

Good peak proWles could be obtained only when thecapillary was adequately cleaned before use. In addition,the reproducibility also depended on the injection mode.Since a 96-well plate was used to present samples in theCE assay, only vacuum and voltage injection modeswere available to us. With vacuum injection, thepercentage RSDs of migration time and peak height(n D 6) were 0.3 and 20%, respectively. With voltageinjection, the percentage RSDs of migration time andpeak height (n D 6) were 0.2 and 5%, respectively. Judg-ing by the relative peak heights, bias injection of thereaction mixture was insigniWcant. Thus, voltage injec-tion was chosen as the preferred injection mode.

Kinetics

To be able to use the CE assay for measuring RNaseH activity in a quantitative manner, it is necessary to

Fig. 4. Product identiWcation by MALDI mass spectrometry. (A)RNA/DNA duplex substrate without enzyme. (B) HIV-1 RNase Hcleavage products. (C) Control oligo. The synthesized oligo is not 5�

phosphorylated while the RNase H products possess a 3� hydroxyland a 5� phosphate, accounting for the mass diVerence of 53 Da.Large- and small-product RNAs are m/z 6322.4 and m/z 1843.8 Da,respectively.

300 K.C. Chan et al. / Analytical Biochemistry 331 (2004) 296–302

know that the assay can measure the reaction rate. Datapresented in Fig. 5 show a plot of the ratio of peakheights of the largest (4-nt) product peak to that of thesubstrate as a function of time. RNase H activity isquenched at the diVerent time points by addition ofEDTA. The CE assay appears suitable for kineticanalysis. In practice, shorter time points (010 min)would be preferred. Our studies used longer time pointsfor direct comparison with the well-based assay, whichwas optimized for signal-noise ratio and compatabilitywith robotic liquid handling.

Linear response of CE RNase H assay to drug inhibitorconcentration

The main purpose of the CE assay was to serve as asecondary assay to evaluate putative RNase H inhibitors.Therefore, the CE assay had to be sensitive enough toobtain inhibitor dosing information. Fig. 6 shows the CEseparation of the HIV-1 RNase H reaction mixture at sev-eral diVerent concentrations of a standard hydrazoneinhibitor, N-(3,4,5-trihydroxybenzoyl)-2-methoxy-1-naphthaldehyde hydrazone [16]. Dose–response resultsfrom the CE assay (IC50 0.15�M) are comparable to thosedetermined with the FRET-based assay (IC50 0.18�M).

Other enzymes

We explored the utility of the CE format with RNasesH from other sources. HIV-2 RT gave a product peakproWle which diVered substantially from that found forthe HIV-1 enzyme. There is high sequence homologyand virtually identical structural homology between thereverse transcriptases from HIV-1 and HIV-2 [17,18], yet

Fig. 5. Kinetic analysis of HIV-1 RNase H enzymatic activity. Pointsrepresent triplicate determinations.

signiWcant diVerences were noted both in the amountand in the type of products generated upon hydrolysis ofthe hybrid duplex substrate by HIV-1 and HIV-2 RT-RNase H (Figs. 7A and B). The amount of substrate

Fig. 6. EVect of four concentrations of the standard RNase H inhibitorN-(3,4,5-trihydroxybenzoyl)-2-methoxy-1-naphthaldehyde hydrazone(KMMP), a standard inhibitor [16]. (A) 10 �M KMMP; (B) 1 �MKMMP; (C) 0.5 �M KMMP; (D) 0.1 �M KMMP; s, duplex substrate;p, cleavage product.

Fig. 7. (A) Separation of substrate and products after cleavage byHIV-1 RNase H; (B) HIV-2 RNase H; (C) E. coli RNase HI; (D)human RNase H.

K.C. Chan et al. / Analytical Biochemistry 331 (2004) 296–302 301

hydrolysis by HIV-2 RT RNase H was much reducedcompared to that by HIV-1 RT RNase H. This is consis-tent with several previous studies that showed theRNase H speciWc activity of HIV-2 RT to be about 10%that of HIV-1 RT [19–22]. Although crystallographicdata [23] indicate that the DNA polymerase and RNaseH catalytic centers of HIV-1 RT are separated by»17 bp, our studies with the RTs of HIV-1 [24], HIV-2[22], and equine infectious anemia virus [25] indicate thateach enzyme displays a slightly relaxed hydrolysis pro-Wle, cleaving the RNA strand with varying eYciencybetween positions ¡16 and ¡19.

More interesting is the diVerence in product distribu-tion formed by the two enzymes. The predominantcleavage during hydrolysis of the 18-bp hybrid duplex byHIV-1 RT RNase H occurs 4 nt from the 3� end of theRNA strand, suggesting that this enzyme has a preferredbinding mode for the blunt end hybrid substrate corre-sponding to a position 16 nt from the 3� end of the DNAstrand. However, three additional minor products arealso noted, implying that there is some Xexibility in thepositioning of the substrate in the RNase H domainHIV-1 RT. In contrast, HIV-2 RT RNase H does notform the 4-nt cleavage product but rather produces onlythose products equivalent to the minor products seenwith HIV-1 RT, in essentially the same ratio, suggestingthat HIV-2 RT does not possess a preferred bindingmode for the 18-bp hybrid duplex substrate. Recentstudies have shown that diVerences in the thumb domainof the small subunit of HIV-2 RT, notably Q294 (theequivalent residue in HIV-1 RT is proline), result inreduced aYnity of HIV-2 RT for RNase H substrates[17]. It is possible that this reduced aYnity impacts onthe ability of the HIV-2 enzyme to precisely position therelatively short RNase H substrate in the RNase Hactive site to enable the ¡4 cut in the RNA strand seenwith HIV-1 RT. Similar diVerences in substrate speciWc-ity have been noted with RT from equine infectious ane-mia virus [25].

Though E. coli RNase H markedly diVers from theHIV enzyme in lacking a DNA polymerase domain, thepattern of cleavage of the substrate in the CE assay(Fig. 7C) was quite comparable to that of HIV-1. Thecleavage pattern observed using human RNase H(Fig. 7D) was more complex than that for either HIV-1,HIV-2, or E. coli enzymes, with additional peaksobserved at diVerent migration times.

Conclusions

A capillary electrophoretic approach to the measure-ment of RNase H activity that provides an attractivealternative to existing assays was developed. While itsspeed in a single-capillary format is not comparable tothe well-based assay that we previously described [4], it

provides a more meaningful endpoint for conWrmationof activity with samples which may otherwise be diYcultto properly analyze. It is suitable for implementation in a96-channel format using a CE array, which would makethis assay competitive with regard to speed andthroughput. Several hundred pure compounds and natu-ral product extracts have been tested using the capillaryelectrophoresis assay for HIV-1 RNase H to qualifythem for further pharmacologic evaluation and leaddevelopment.

Acknowledgments

This project was funded in part with Federal fundsfrom the National Cancer Institute, National Institutesof Health, under Contract No. NO1-CO-12400.

References

[1] J.W. Rausch, S.F.J. Le Grice, Reverse transcriptase-associatedribonuclease H activity as a target for antiviral chemotherapy,Antiviral Chem. Chemother. 8 (1997) 173–185.

[2] G.J. Klarmann, M.E. Hawkins, S.F.J. Le Grice, Uncovering thecomplexities of retroviral ribonuclease H reveals its potential as atherapeutic target, AIDS Rev. 4 (2002) 183–194.

[3] M.L. Andreola, V.R. De Soultrait, M. Fournier, V. Parissi, C. Des-jobert, S. Litvak, HIV-1 integrase and RNase H activities as thera-peutic targets, Expert Opin. Ther. Targets 6 (2002) 433–446.

[4] M.A. Parniak, K.L. Min, S.R. Budihas, S.F.J. Le Grice, J.A. Beu-tler, A Xuorescence-based high-throughput screening assay forinhibitors of HIV-1 reverse transcriptase associated ribonucleaseH activity, Anal. Biochem. 322 (2003) 33–39.

[5] F. Welder, E.M. McCorquodale, C.L. Colyer, Proteinase assay bycapillary electrophoresis employing Xuorescence-quenched pro-tein-dye conjugates, Electrophoresis 23 (2002) 1585–1590.

[6] F. Sano, K. Ueno, H. Kamimori, Enzyme assay for protein kinaseusing micellar electrokinetic chromatography with laser-inducedXuorescence detection, J. Chromatogr. B 794 (2003) 149–156.

[7] Y. He, E.S. Yeung, High-throughput screening of kinase inhibitorsby multiplex capillary electrophoresis with UV absorption detec-tion, Electrophoresis 24 (2003) 101–108.

[8] J.J. Bao, J.M. Fujima, N.D. Danielson, Determination of minuteenzymatic activities by means of capillary electrophoretic tech-niques, J. Chromatogr. B 699 (1997) 481–497.

[9] J.T. Miller, J.W. Rausch, S.F.J. Le Grice, Evaluation of retroviralribonuclease H activity, Methods Mol. Biol. 160 (2001) 335–354.

[10] W.P. Ma, S.E. Hamilton, J.G. Stowell, S.R. Byrn, V.J. Davisson,Sequence speciWc cleavage of messenger RNA by a modiWed ribo-nuclease H, Bioorg. Med. Chem. 2 (1994) 169–179.

[11] F. Pileur, M.L. Andreola, E. Dausse, J. Michel, S. Moreau, H.Yamada, S.A. Gaidamakov, R.J. Crouch, J.J. Toulme, C. Cazen-ave, Selective inhibitory DNA aptamers of the human RNase H1,Nucleic Acids Res. 31 (2003) 5776–5788.

[12] E.F. Fabriziom, A. Nadim, J.D. Sterling, Resolution of multiplessDNA structures in free solution electrophoresis, Anal. Chem.75(2003) 5012–5021.

[13] M.F. Mangano, C. Battaglia, G. Salani, L.R. Bernardi, G. De Bel-lis, Composition dependent separation of oligonucleotides by cap-illary electrophoresis in acidic buVers with application to thequality control of synthetic oligonucleotides, J. Chromatogr. A.848 (1999) 435–442.

302 K.C. Chan et al. / Analytical Biochemistry 331 (2004) 296–302

[14] P.G. Righetti, C. GelW, M. Perego, A.V. Stoyanov, A. Bossi, Capil-lary zone electrophoresis of oligonucleotides and peptides in iso-electric buVers: theory and methodology, Electrophoresis 18(1997) 2145–2153.

[15] G. De Bellis, G. Salani, Oligonucleotide analysis by capillary zoneelectrophoresis at low pH, Anal. Chim. Acta 345 (1997) 1–4.

[16] G. Borkow, R.S. Fletcher, J. Barnard, D. Arion, D. Motakis, G.I.Dmitrienko, M.A. Parniak, Inhibition of the ribonuclease H andDNA polymerase activities of HIV-1 reverse transcriptase by N-(4-tert-butylbenzoyl)-2-hydroxy-1-naphthaldehyde hydrazone,Biochemistry 36 (1997) 3179–3185.

[17] H. Kuhnel, H. von Briesen, U. Dietrich, M. Adamski, D. Mix, L.Biesert, R. Kreutz, A. Immelmann, K. Henco, C. Meichner, Molec-ular cloning of two west African human immunodeWciency virustype 2 isolates that replicate well in macrophages: a Gambian iso-late, from a patient with neurologic acquired immunodeWciencysyndrome, and a highly divergent Ghanian isolate, Proc. Natl.Acad. Sci. USA 86 (1989) 2383–2387.

[18] J. Ren, L.E. Bird, P.P. Chamberlain, G.B. Stewart-Jones, D.I.Stuart, D.K. Stammers, Structure of HIV-2 reverse transcriptaseat 2.35-A resolution and the mechanism of resistance to non-nucleoside inhibitors, Proc. Natl. Acad. Sci. USA 99 (2002)14410–14415.

[19] N. Fan, K.B. Rank, S.M. Poppe, W.G. Tarpley, S.K. Sharma,Characterization of the p68/p58 heterodimer of human immuno-

deWciency virus type 2 reverse transcriptase, Biochemistry 35(1996) 1911–1917.

[20] Z. Sevilya, S. Loya, S.H. Hughes, A. Hizi, The ribonuclease Hactivity of the reverse transcriptases of human immunodeWciencyviruses type 1 and type 2 is aVected by the thumb subdomain ofthe small protein subunits, J. Mol. Biol. 311 (2001) 957–971.

[21] Z. Sevilya, S. Loya, N. Adir, A. Hizi, The ribonuclease H activityof the reverse transcriptases of human immunodeWciency virusestype 1 and type 2 is modulated by residue 294 of the small subunit,Nucleic Acids Res. 31 (2003) 1481–1487.

[22] K. Post, J. Guo, K.J. Howard, M.D. Powell, J.T. Miller, A. Hizi,S.F. Le Grice, J.G. Levin, Human immunodeWciency virus type 2reverse transcriptase activity in model systems that mimic steps inreverse transcription, J. Virol. 77 (2003) 7623–7634.

[23] S.G. SaraWanos, K. Das, C. Tantillo, A.D. Clark Jr., J. Ding, J.M.Whitcomb, P.L. Boyer, S.H. Hughes, E. Arnold, Crystal structureof HIV-1 reverse transcriptase in complex with a polypurine tractRNA:DNA hybrid, EMBO J. 20 (2001) 1449–1461.

[24] C.E. Cameron, M. Ghosh, S.F. Le Grice, S.J. Benkovic, Mutations inHIV reverse transcriptase which alter RNase H activity and decreasestrand transfer activity are suppressed by HIV nucleocapsid protein,Proc. Natl. Acad. Sci. USA 94 (1997) 6700–6705.

[25] J.W. Rausch, S.F.J. Le Grice, Substituting a conserved residue ofthe ribonuclease H domain alters substrate hydrolysis by retrovi-ral reverse transcriptase, J. Biol. Chem. 272 (1997) 8602–8610.