Role of the Reverse Transcriptase, Nucleocapsid Protein, andTemplate Structure in the Two-step Transfer Mechanism inRetroviral Recombination*

Received for publication, May 2, 2003, and in revised form, June 4, 2003Published, JBC Papers in Press, June 11, 2003, DOI 10.1074/jbc.M304608200

Ricardo H. Roda‡, Mini Balakrishnan‡, Mark N. Hanson‡, Birgitta M. Wohrl¶,Stuart F. J. Le Grice�, Bernard P. Roques**, Robert J. Gorelick‡‡, and Robert A. Bambara‡ §§

From the ‡Department of Biochemistry and Biophysics, University of Rochester, Rochester, New York 14642, ¶AbteilungPhysikalische Biochemie, Max-Planck-Institut fur Molekulare Physiologie, Otto-Hahn-Strasse 11, D-44227 Dortmund,Germany, �HIV Drug Resistance Program and ‡‡AIDS Vaccine Program, Science Applications International Corporation,NCI Frederick, Frederick, Maryland 21702, and the **Department de Pharmacochemie Moleculaire et Structurale,INSERM U266, CNRS UMR 8600, 4 Ave. de l’Observatoire, 75270 Paris Cedex 06, France

Template switching during reverse transcription pro-motes recombination in retroviruses. Efficient switcheshave been measured in vitro on hairpin-containing RNAtemplates by a two-step mechanism. Pausing of the re-verse transcriptase (RT) at the hairpin base allowedenhanced cleavage of the initial donor RNA template,exposing regions of the cDNA and allowing the acceptorto base pair with the cDNA. This defines the first ordocking step. The primer continued synthesis on thedonor, transferring or locking in a second step. Here wedetermine the enzyme-dependent factors that influencetemplate switching by comparing the RTs from humanimmunodeficiency virus, type 1 (HIV-1), and equine in-fectious anemia virus (EIAV). HIV-1 RT promoted trans-fers with higher efficiency than EIAV RT. We found thatboth RTs paused strongly at the base of the hairpin.While stalled, HIV-1 RT made closely spaced cuts,whereas EIAV RT made only a single cut. Docking oc-curred efficiently at the multiply cut but not at the sin-gly cut site. HIV-1 nucleocapsid (NC) protein stimulatedstrand transfers. It improved RNase H activity of bothRTs. It allowed the EIAV RT to make a distribution ofcuts, greatly stimulating docking at the base of the hair-pin. Most likely, it also promoted strand exchange, al-lowing transfers to be initiated from sites throughoutthe hairpin. Minor pause sites beyond the base of thehairpin correlated with the locking sites. The strandexchange properties of NC likely promote this step. Wepresent a model that explains the roles of RNase H spec-ificity, template structure, and properties of NC in thetwo-step transfer reaction.

Recombination occurs frequently in retroviruses (1–3). For asimple retrovirus such as spleen necrosis virus, recombination

happens once every three replication cycles (4, 5), whereas inHIV-11 as many as three recombination events take place in asingle replication cycle (6, 7). More importantly, recombinationappears to be an essential mechanism for retroviral survival. Itallows viruses to repair genomic errors (8–11) and to combineresistance markers against anti-retroviral drugs (12–14), andit promotes genomic shuffling that frustrates immune surveil-lance and allows for the emergence of new mosaic viral strains(15, 16).

Recombination takes place during reverse transcriptionwhen the virus-encoded reverse transcriptase (RT) copies thesingle-stranded viral RNA into double-stranded viral DNA.Synthesis of the integration-competent double-stranded viralDNA involves two obligatory strand transfers from the ends ofthe genome (17–19). In addition to these end transfers, tem-plate switching can also occur within internal sequences of thegenome (4, 6, 20–23) and is driven by template homology(24–30). Strand transfer or template switching results when aprimer being extended on one copy of the template (donor) istransferred to a homologous site on a second copy of the tem-plate (acceptor). Although it has been proposed to occur duringplus (31) and minus strand synthesis (32), most templateswitching occurs during RNA-templated synthesis (7, 20, 23,33, 34). As a result, most mechanistic studies have focused onrecombination from RNA templates. Intra-molecular transferswithin the same molecule lead to deletions (17, 25, 35). How-ever, because retroviruses package two complete copies of theirRNA genome, template switching can also take place inter-molecularly. When two nonidentical (heterogeneous) genomesare co-packaged, such inter-molecular transfers lead to recom-binant progeny (4).

Pausing of RT during synthesis was found to correlate withthe locations of template switching (36–38). Because RT hasDNA polymerase and RNase H activities within the same mol-ecule (39, 40), it was suggested that pausing would temporarilyinhibit nucleotide addition, but it would not affect the nucleasefunction. This would result in degradation of the RNA donortemplate in the region of pausing and subsequent exposure ofthe cDNA. Base pairing between a homologous acceptor mole-cule and the cDNA would then result in a transfer event (36–38) and polymerization would now resume on the acceptortemplate. In support of this model, RNase H activity has beenfound to be necessary for strand transfers. Reducing or deleting

* This work was supported by National Institutes of Health GrantsGM49573 and T32-CA09363 (to M. N. H.) and also in part by NIH/NCIContract N01-CO-12400 with SAIC-Frederick, Inc. The costs of publi-cation of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§ Trainee in the Medical Scientist Training Program funded by Na-tional Institutes of Health Grant T32-GM07356 and is also supportedby National Institutes of Health Predoctoral Fellowship 5F31 GM20911-02.

§§ To whom correspondence should be addressed: Dept. of Biochem-istry & Biophysics, University of Rochester Medical Center, 601 Elm-wood Ave., Box 712, Rochester, NY 14642. Tel.: 585-275-2764; Fax:585-271-2683; E-mail: robert_bambara@urmc.rochester.edu.

1 The abbreviations used are: HIV-1, human immunodeficiency virus,type 1; RT, reverse transcriptase; NC, nucleocapsid; FL, full length;EIAV, equine infectious anemia virus.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 34, Issue of August 22, pp. 31536–31546, 2003Printed in U.S.A.

This paper is available on line at http://www.jbc.org31536

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

the RNase H activity of RT leads to a significant reduction inits ability to switch between RNA templates in vitro (37, 41–44) and a reduction in direct repeat deletion in virions (26, 27,29, 30, 45, 46). The transfer process has been described as a“dynamic copy choice,” meaning that an appropriate balancebetween the RNase H and polymerase activities is required toobtain efficient direct repeat deletion (27, 46).

The role of RNA template structure in triggering transfershas also been studied (47–54). RNA hairpins facilitate trans-fers in a number of ways. The kissing hairpin of HIV-1, whichcontains a palindromic loop sequence that induces templatedimerization, has been shown to facilitate strand transfer invitro (55, 56), and to provide a hot spot for recombination rescueof defective murine leukemia virus genomes in vivo (11). Sim-ilarly, kissing interactions promoted between the TAR loopsequences of the cDNA intermediate and acceptor RNA areproposed to facilitate the minus strand transfer (47).

RNA hairpins may also promote transfer by creating a tran-sient pause site (36, 47, 48, 54, 57, 58). Although our initialexpectation was that most transfers would take place at thesite of pausing at the hairpin base (48), sequence analysis of thetransfer products revealed that the transfer of the primer ter-minus took place at a distance, after synthesis through thepause site (48, 54). Detailed analysis revealed that templateswitching occurred through a two-step process (54). Pausingincreased RNase H activity at the base of the hairpin, allowingthe acceptor to “dock” to the cDNA. This initial step was fol-lowed by continued synthesis through the structure of thehairpin on the donor. The transfer event is not finished or“locked” until the cDNA primer terminus had annealed to theacceptor beyond the pause site. These results presented clearevidence for a previously suggested two-step transfer mecha-nism (27, 55, 59). More recently, transfers taking place in the 5�end of the HIV-1 genome have also been shown to take placethrough such a “dock and lock” process (55).

Reverse transcription takes place in the presence of theaccessory protein nucleocapsid (NC) that binds to the genome.NC is packaged within the virion, is processed from the Gagprecursor poly-protein, and it has RNA chaperone activity, aswell as a modulating activity on RT (60). NC has been shown toincrease the efficiency of the strand transfer process, either bypromoting nucleic acid annealing or by increasing the RNase Hactivity of RT (38, 48, 51, 54, 56, 59, 61–71).

Here we seek to explore the enzymatic factors that contrib-ute to the pause-driven two-step transfer mechanism. We havedone so by carrying out assays with the RTs from humanimmunodeficiency virus, type I (HIV-1), and equine infectiousanemia virus (EIAV) on hairpin containing RNA substrates.EIAV and HIV-1 belong to the lentivirus subfamily of retrovi-ruses. Both replicate in cells of the immune system and havehigh mutation rates (72, 73). The RTs share 40% of amino acidhomology, are heterodimers of a catalytically active p66 sub-unit and an inactive p51 subunit, and appear to be structurallysimilar (74–80). The present study demonstrates that despitesuch similarities in their overall properties, HIV-1 RT wasmore efficient in carrying out strand transfers from hairpinstructures than EIAV RT. An important contributor to thiswere differences in the RNase H cleavages carried out when theRTs are paused at the base of the hairpin. We have found thatalthough both RTs paused at the base of the hairpin, HIV-1 RTdegraded the template into smaller fragments that could easilydissociate from the cDNA and could therefore allow more effi-cient docking. HIV-1 NC, but not EIAV NC, caused a robustincrease in transfers with both RTs. We believe this enhance-ment to be a result of HIV-1 NC’s enhancement of RNase H

activity, and by making previously inaccessible regions of theacceptor more amenable to the invasion or docking step.

EXPERIMENTAL PROCEDURES

Materials

Expression plasmids, pKK-p66(6his) and pKK-p51(6his) (81), foroverexpression of the RT p66 and p51 subdomains were kindly providedby Dr. Neerja Kaushik. EIAV RT was generated and purified as de-scribed previously (76, 79). Nucleocapsid protein-(1–72)-NCp7 waschemically synthesized (82) and was stored at �80 °C in a buffer of 50mM Tris-HCl (pH 7.5) and 5 mM dithiothreitol. DH5�-competent cells,XbaI, and EcoRI were from Invitrogen. Radiolabeled compounds werefrom PerkinElmer Life Sciences, and Micro Bio-Spin columns were fromBio-Rad. Plasmid purification kits were from Qiagen. All DNA primerswere synthesized by Integrated DNA Technologies, Inc. All other prod-ucts were purchased from Roche Applied Science.

Methods

HIV-1 RT Purification—Overexpression and purification of the RTheterodimer was done as described previously (83), with a slight mod-ification. Briefly, JM109 cells harboring the expression plasmids weregrown in 2� YT media containing ampicillin (100 �g/ml). Cell pellets ofthe p66 and p51 clones were mixed together at the appropriate ratio(84), lysed, sonicated, and centrifuged at 28,000 � g for 45 min. Thesupernatant was diluted to achieve 0.5 M NaCl, treated with strepto-mycin sulfate at a final concentration of 4%, and centrifuged again at28,000 � g for 30 min. The supernatant was applied to a nickel affinitycolumn (Amersham Biosciences), and the heterodimer was purified asdescribed earlier (83). Eluted factions containing equal amounts of p66and p51, as judged from SDS-PAGE, were pooled together and dialyzed.

FIG. 1. Sequence and structure of EIAV primer binding site(PBS)-derived hairpin substrates. Schematic of the DI donor, AI-2acceptor, and ssAcceptor templates. Donor and acceptor RNAs share ahomology of 93 and 75 nucleotides, respectively. Synthesis is initiatedfrom a DNA primer (dashed arrow) annealed to the donor 3� end. Theacceptors are longer at their 5� end, than the donor. Base substitutionmarkers are shown as filled circles on the acceptors with the markernumber and sequence indicated on top. Structural regions of the tem-plate with respect to the marker are also highlighted. The sequence andstructure of DI donor RNA shown. Donor and acceptor RNAs shared thesame structure. The dark line represents regions present only on thedonor, and the sequence represents the region of homology betweendonor and acceptors. Nucleotide positions on the donor RNA are indi-cated with a � prefix, starting with the 5� end as �1. Marker numbersare indicated with a # prefix, with the first marker closer to the 3� endof the RNA.

Retrovirus Recombination Mechanism 31537

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

Purified protein was more than 95% pure as judged by CoomassieBlue-stained gels and was stored at �20 °C. Protein concentration wasdetermined by Bio-Rad assay.

Generation of EIAV-derived Donor and Acceptor Templates—Thedonor construct pEIAV-Donor and the acceptor construct pEIAV-A2have been described previously (54). RNA templates were generated byrun-off transcription in vitro from linearized plasmids using T7 RNApolymerase, as per the manufacturer’s protocol. The pEIAV-Donor was

linearized with NotI to generate DI donor templates, and pEIAV-A2was linearized with HindIII and HaeIII to generate the AI-2 and ASacceptor templates, respectively. RNA substrates were gel-purified andtested for integrity.

Labeling and Annealing of Substrates—DNA primers or RNA tem-plates (RNA templates first calf intestine phosphatase-treated) werelabeled at the 5� end as described previously (54). For annealing, donorRNA and DNA primer at a ratio of 1:2 were heated for 2 min at 95 °C

FIG. 2. HIV-1- and EIAV RT-catalyzed transfer reactions. Transfer reactions catalyzed by HIV-1 (A) and EIAV RT (B). The 5� end-labeledSP6 primer was annealed to donor DI, and extension reactions were carried out in the presence of acceptor AI-2 and in either the absence (lanes1–6) or presence of HIV-1 NC (lanes 8–13). Enough NC for 200% coating of the templates (100% � enough NC to cover all nucleotides in thereaction once, assuming each NC molecule covers 7 nucleotides) was added to the reaction and incubated for 5 min prior to the addition of RT. Thereaction was sampled at the times indicated on top of the lanes, and products were separated in an 8% denaturing polyacrylamide gel. Lane L wasloaded with a 10-nucleotide DNA ladder. Products observed over the time of the assay correspond to products resulting from full-length synthesisto the end of the donor (FL), transfer product (TP) resulting from switching and synthesis to the end of the acceptor, and a fold-back (FB) resultingfrom self-priming of the full-length product. Locations of the markers are indicated on the left of the panel and product sizes on the right. Theschematic on the right indicates the region of primer binding as well as the corresponding regions on the template for the pause sites. C, transferefficiencies supported by the HIV-1 and EIAV RTs. Transfer efficiency was determined using the formula %TE � TP/(FL � FB � TP) � 100.Transfer efficiencies were determined using donor to acceptor ratios of 1:2 (black bars) and 1:8 (gray bars) for HIV-1 and EIAV RTs in the presenceand absence of nucleocapsid proteins at 200% coating. Results represent the average of at least three independent experiments.

Retrovirus Recombination Mechanism31538

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

and slowly cooled to room temperature in 50 mM Tris-HCl (pH 8.0), 50mM KCl, 1 mM dithiothreitol, and 1 mM EDTA.

RT Assays—Donor DI was primed for reverse transcription with theDNA oligonucleotide SP6 (5�-TACGATTTAGGTGACACTATAG-3�). Do-nor extension and strand transfer assays were carried out in a finalvolume of 12 �l. Two units of HIV-1 or EIAV reverse transcriptase (50ng) were incubated for 5 min at room temperature with 25:50 fmol(donor/primer) of hybridized substrate. For reactions with NC, theprotein was added to the reactions prior to the addition of RT andincubated at room temperature for 5 min. By assuming that each NCmolecule would cover 7 nucleotides of RNA, the amount of NC requiredfor 200% coating was determined based on the length and amount of theprimers and RNA templates in the reaction. Reactions were startedwith the addition of dNTPs and MgCl2 at a final concentration of 50 �M

and 6 mM, respectively, incubated at 37 °C, and stopped at the appro-priate time by adding 1 volume of termination buffer (90% formamide,10 mM EDTA (pH 8.0), and 0.1% each of xylene cyanol, and bromphenolblue). Strand transfer reactions were carried out under the same con-ditions as extension assays, except that acceptor was included in theannealing reactions. For blocking oligonucleotide (EIAV-3, 5�-CAGAC-CATACCTGAAGCTTAC-P-3�) experiments the oligomer was added tothe reaction mixture at the same time as the start mix. The oligomerhas a 3�-phosphate modification that prevents priming of synthesis.

RNase H Assays—RNase H assays were performed with 5� end-labeled donor RNA and unlabeled SP6, under the same reaction condi-tions as the extension assays but in a final volume of 80 �l. For the timecourse analysis, 10 �l of reaction was sampled at each time point andmixed to 10 �l of termination buffer. Reaction products were resolvedon 8% polyacrylamide-urea gels, and visualized and quantitated usinga PhosphorImager and ImageQuant software (Amersham Biosciences).

Analysis of Transfer Products—To determine the transfer distribu-tion, the transfer products were isolated, amplified, cloned, and se-quenced as described previously (54). Each clone represented an indi-vidual transfer event, and its sequence was used to determine the pointof crossover between donor and acceptor.

RESULTS

We propose that the characteristics of the two-step transferprocess are determined by a combination of qualities of thereverse transcriptase, the nucleocapsid protein, the structureof the substrates, and the transfer intermediates these threecomponents form. To distinguish specifically the roles of theproteins from those of the substrate, we examined each step inthe transfer mechanism in the presence and absence of NC,using HIV-1 and EIAV RTs. The system we used to modelstrand transfers in vitro consists of a hairpin containing donorRNA molecule, on which reverse transcription is initiated, anda homologous acceptor RNA molecule onto which the primercan transfer. Single base markers on the acceptor molecule areused to determine the location of the crossover (38, 48, 54, 55).

HIV-1 RT and EIAV RT Show Similar Pattern of Pausing—Donor RNA DI contains a strong hairpin, and when used inconjunction with acceptor AI-2 most transfers are completedbefore the last nucleotide on the donor is copied (54). Thesequence of DI and the location of the markers on acceptor AI-2used to locate the locking site are shown in Fig. 1. Pausingduring reverse transcription was proposed to affect both dock-ing and locking during strand transfer (54). To compare thepausing profiles of HIV-1 and EIAV RTs extension reactionswere carried out using labeled primer on donor DI. Previousassays carried out with these substrates and HIV-1 RT weredone at 80 mM KCl. However, to get efficient synthesis withEIAV RT, a lower salt concentration (50 mM KCl) was neces-sary. As a result, assays using the HIV-1 RT have been re-peated at the lower salt condition and incorporated in thepresent report for appropriate comparison.

The sites of pausing for both RTs were similar but not iden-tical (Fig. 2, A and B). Full-length extension products were notseen until 5 min with EIAV RT, as opposed to 3 min with HIV-1RT. EIAV RT paused more prominently at �78 and �79, cor-responding to stalling at the base of the hairpin between mark-ers 1 and 2. We proposed previously (54) that pausing and

associated RNase H activity at this point is responsible forcreating a major acceptor invasion site. HIV-1 RT had a uniquepause site at location �68 (between markers 3 and 4). The nexttwo pause sites, �44 and �57, both of which fall between theregion delineated by markers 4 and 5 on the acceptor, were

FIG. 3. RNase H cleavage profiles of HIV-1 and EIAV RT duringsynthesis. Reaction conditions for HIV-1 (A) and EIAV RT (B) were thesame as those used for extension reactions in the absence (lanes 1–6) orpresence of HIV-1 NC (lanes 7–12). The radiolabel (shown as a star) wason the 5� end of the donor RNA instead of the primer, and no acceptorwas present. The schematic on the right indicates the region of primerbinding as well as the corresponding regions on the template for thecleavage sites. Enough NC for 200% coating of the templates was addedto the reaction and incubated for 5 min prior to the addition of RT. Thereactions were sampled at the times indicated above each lane, andproducts were separated through an 8% polyacrylamide gel run underdenaturing conditions. Sizes of the cleavage products are indicated. Ctrllane, reactions done in absence of RT.

Retrovirus Recombination Mechanism 31539

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

present in extensions by both RTs. EIAV RT presented aunique pause site at �31 (between markers 5 and 6), whereasboth polymerases paused at position �16 albeit with differentintensities. EIAV paused more intensely at position �16 thanat any site upstream of the base of the hairpin. HIV-1 RT, onthe other hand, paused approximately the same at �16, �44,and �57. The overall similarity in pausing profiles betweenboth enzymes, especially at the base of the hairpin, led us tohypothesize that strand transfer occurs through an analogousmechanism for both RTs.

EIAV RT Has Low Strand Transfer Activity in Hairpin Con-taining Substrates—Despite similar pausing profiles, a lowertransfer efficiency was observed for EIAV RT as compared withHIV-1 RT. The transfer efficiency is a measure of how manytransfer products are produced in relation to the total full-length products. It follows the formula TE � TP/(TP � FL �FB) � 100%. Here the transfer efficiency (TE) is related to theamount of transfer product (TP), full-length extension productson the donor template (FL), and the product resulting fromfoldback and additional extension of the donor extension prod-uct (FB). At a 1:2 donor to acceptor ratio (Fig. 2C) HIV-1 RTand EIAV RT supported a TE of 5 and 3%, respectively. Theinitial “docking” step in the proposed two-step transfer involvesan interaction between the acceptor and the nascent cDNA. Wetherefore reasoned that higher acceptor concentrations wouldincrease the chances for such an intermolecular event andtherefore enhance TE. When the concentration of acceptor wasquadrupled the TE increased 4-fold to 21% for HIV-1 RT butonly doubled to 6% for EIAV RT. The lower TE for EIAV RT andits lesser responsiveness to increases in the acceptor concen-tration suggested that it was less efficient in the docking stepthat is dependent on RNase H activity. This lead us to examineand compare the events leading up docking for both RTs.

Differences in RNase H Cleavages by the RTs—To examineRNase H cleavages during synthesis, extension reactions werecarried out using unlabeled primer and 5� end-labeled donor.Both RTs made numerous cuts in two regions on the donorRNA as follows: between positions 80 and 100 (Fig. 3, A and B),corresponding to RTs stalled at the base of the stem, and in thedescending limb of the hairpin, corresponding to cleavages atthe end of the template, which produces fragments smaller

than 20 nucleotides in length. The characteristics of those cuts,however, were different for the two RTs.

Whereas HIV-1 RT is paused at the first base of the stem ofthe hairpin, it creates closely spaced cuts at positions �101, 98,94, 91, and 83, with the most intense cut at �94. These corre-spond to cuts in the donor 23, 20, 16, 13, and 5 nucleotidesdownstream of the �78 pause site. These cleavages could bethe result of either single cuts on different RNA molecules orthe result of several cuts on a single molecule. If the formerwere true, then over time one would observe a net accumula-tion of all cleavage products, and only the full-length RNA bandwould be reduced. However, products at �101, 98, and 94appear within 1 min of the reaction, but then their intensitiesdecrease, while other bands become more intense. We interpretthis as signifying that these early products are further cleavedinto smaller products that appear at later time points. Appar-ently in our reactions a significant portion of the RNA mole-cules are cut in multiple locations. We also note that the per-sistence of some uncut RNA in reactions without NC mostlikely reflects incomplete priming of the RNA, rather than alow level of cutting.

The cleavage pattern suggests that at least some of the RNAsare fragmented into pieces either small enough to dissociateinto solution or to be readily displaced by a docking acceptorRNA molecule. The RNase H cleavage profile of EIAV RTproduced a major fragment 98 nucleotides in length and minorcuts at 101 and 94. The resulting segments of RNA may be toolarge to dissociate, inhibiting docking by an acceptor moleculeleading to poor transfer efficiency.

Throughout the template, the cleavage pattern shows thatHIV-1 RT cuts the RNA with much more diversity of location,e.g. at positions �63, 50, 27, and 17, than EIAV RT. Thisprovides the opportunity for nearby cuts on the same template,which could allow gap formation that would promote primer-donor interaction. The positive correlation between diverse po-sitions of cleavage and high transfer efficiency suggests thatclosely spaced cuts are an important prerequisite for efficientdocking. The inability of the EIAV RT to make diverse-positioncuts suggests the basis for its low TE relative to the HIV RT.

Transfers Carried Out by EIAV RT Take Place by an Ineffi-cient Two-step Mechanism—In a previous study we showed

FIG. 4. Effect of deleting the base ofhairpin on transfer efficiency. ThessAcceptor (54), lacking the region 3� tothe base of the hairpin, was used insteadof AI-2 in standard transfer assays. Aschematic of the donor-acceptor substratepair in indicated above the panel. Trans-fer efficiencies and NC coating levels werecalculated as in Fig. 2. Transfer efficien-cies were determined using donor to ac-ceptor ratios of 1:2 (black bars) and 1:8(gray bars) for HIV-1 and EIAV RTs inthe presence and absence of nucleocapsidprotein at 200% coating. Results repre-sent the average of at least three inde-pendent experiments.

Retrovirus Recombination Mechanism31540

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

that deleting the docking region of an acceptor leads to a verylarge reduction in transfers supported by HIV-1 RT (54). Usingsuch an acceptor (ssAcceptor) leads to a 10-fold reduction intransfer efficiency for HIV-1 RT (Fig. 4). Unlike the acceptorthat includes the docking region, increasing the acceptor con-centration 4-fold did not increase the TE with the shortenedacceptor. Similar to HIV-1 RT, EIAV RT also displayed a re-duction in transfers in the absence of a docking region. At thelower acceptor ratio the transfer efficiency was reduced from3% with the original acceptor to 1% with ssAcceptor (Fig. 4).For both RTs, the drop in transfer efficiency with ssAcceptor isgreater at the higher acceptor concentration (Fig. 4). This sug-gests that transfers promoted by EIAV RT were also initiatedat the base of the hairpin, although very inefficiently due topoor docking. The reduction in TE observed with the ssAcceptoralso implies that the base of the hairpin is the primary dockingsite for transfers promoted by both RTs. Otherwise the transfer

efficiency would have been maintained despite the deletion.Distribution of Transfer Products—The second step in strand

transfer, or “lock” step, involves the translocation of the 3� endof the cDNA from the donor to the acceptor RNA. This producesthe genetic shift. The position of locking was obtained by ana-lyzing cloned transfer products to determine which markersfrom the acceptor they incorporated. The data in Fig. 5, A andB, were generated by dividing the number of colonies showinga transfer between two markers by the total number of coloniessequenced. Fig. 5A compares the distribution of crossovers atthe lower acceptor concentration for both HIV-1 and EIAV RT.The distributions show some similarities. Both RTs transferredmostly within the hairpin, downstream of the base of the hair-pin. They also both had a peak of transfer between markers 4and 5, with 52 and 40% for HIV-1 and EIAV RT, respectively.There were few transfers on the ascending limb of the hairpinbetween markers 2–4 or after the last nucleotide had been

FIG. 5. Distribution of primer terminus transfers for HIV-1 and EIAV RTs. Transfer reactions using HIV-1 and EIAV RT were carriedout using a donor to acceptor ratio of 1:2 (A) or 1:8 (B). Products were gel-purified, amplified, and cloned, and individual colonies were sequenced.To obtain the distribution, the number of clones containing transfers within each individual segment were divided by the total number of clonessequenced and multiplied by 100. Two independent experiments were carried out for each set. The graph shows the average values for each setand the range of values for each experiment within that set. For reactions performed using HIV-1 RT (black bars), a total of 55 clones weresequenced; for reactions using EIAV RT (gray bars), 60 clones were sequenced. A schematic of the donor-acceptor substrate pair is indicated abovethe panel.

Retrovirus Recombination Mechanism 31541

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

copied from 5� to 8. Differences arose in the second-most pre-ferred site of transfer; for HIV-1 RT it was the base of thehairpin between markers 1 and 2 (25%) and the 5� half of theloop, between markers 5 and 6 (12.5%), whereas for EIAV RT itwas between markers 5 and 8 where segments 5–6, 6–7, and7–8 supported 19, 21, and 14% of total transfers.

Increasing the acceptor concentration 4-fold had an interest-ing effect on the transfer distributions. Fig. 5B shows that thegeneral pattern of transfers for HIV-1 RT maintained a bi-modal nature. However the number of transfers at the basewas significantly increased (43%), to the point that they be-come the most likely inter-marker site of primer terminustransfer. The higher acceptor concentration also triggered amajor redistribution of transfers for EIAV RT. The increasefrom 2 to 6% in the transfer efficiency translated mostly into anincrease from 19 to 39% in the fraction of transfers betweenmarkers 5 and 6 and from 21 to 33% of those between 6 and 7,whereas the number of transfers between 4 and 5 were reducedfrom 40 to 21%. The overall data showed that at a loweracceptor concentration the distributions of transfers as well asthe transfer efficiencies were similar for both enzymes. How-ever, increasing the concentration of acceptor affected theamount and location of transfers differently for HIV-1 andEIAV RTs.

Nucleocapsid Protein Increases Transfer Efficiency—BecauseHIV-1 NC has been shown to promote transfers, we tested theeffects of both HIV-1 and EIAV NC on the template switchingreactions carried out by their respective RTs. We observed thatat the low acceptor concentration, HIV-1 NC increased trans-fers promoted by HIV-1 RT from 5 to 22% (Fig. 2C). Interest-ingly, at the higher acceptor concentration, NC did not havemuch of a significant effect on the TE. This suggests that theincrease in transfers mediated by HIV-1 NC on HIV-1 RT canbe saturated at high acceptor concentrations.

Surprisingly, EIAV NC did not significantly stimulate trans-fers with EIAV RT at the low or high acceptor concentrations(Fig. 2C). An additional surprise was that HIV-1 NC had arobust effect on transfers by EIAV RT. A 4.5-fold increase wasregistered at low and high acceptor ratios. This suggested thatat least some of the enhancements produced by retroviral NCsare not dependent on specific interactions between autologousRTs and their nucleocapsid proteins. Also, further raising the

concentration of acceptor 4-fold increased transfers for EIAVRT in the presence of HIV-1 NC (Fig. 2C). This suggests thatHIV-1 NC alleviates but does not eliminate an inefficient in-termediate step in the EIAV RT-promoted transfer reaction.Because of the much greater stimulatory effects of HIV-1 NCon both reactions, we carried out additional NC assays for bothRTs using HIV-1 NC.

HIV-1 NC Effects on Synthesis and RNase H Activity—In anattempt to define the manner in which NC promotes transfers,we examined both extensions and RNase H cleavages in thepresence of HIV-1 NC (Figs. 2 and 3). Fig. 2, A and B, showsextension reactions in the presence of NC. The protein did notsignificantly alter the extension profile or the pausing profile.However, quantification of band intensities indicated an in-crease (2–3-fold) of those products making it to full-length (FL)versus those paused at the hairpin base (�78 and �79) for bothRTs. This showed that NC enhances synthesis through themajor pause site.

The effect of NC on the RNase H profile was more striking(Fig. 3, A and B). For HIV-1 RT, NC creates a new set of cuts atpositions �86, 24, and 20, whereas the intensities of bands atposition �101 are increased (Fig. 3A). The product at position94 does not accumulate at the early time points. The overalleffect is an increase in the number of cuts between positions�83 and 101, the region adjacent to the base of the hairpin.One could deduce then that this would certainly increase theexposure and availability of the “docking” site on the cDNAprimer. Also, the intensity of products 17 nucleotides or smallerin length is significantly increased, consistent with an overallincrease in RNase H cleavage.

A similar effect is also observed for EIAV RT (Fig. 3B). In thepresence of NC new cleavage sites were observed in the regionadjacent to the base of the hairpin, in particular cuts appearingat positions �86 and 83. Intensities of cleavages at �101, 94,and those corresponding to the smaller products were alsoincreased. Thus NC increases cleavages in the expected dock-ing site of EIAV RT as well. This result suggests a mechanismby which NC could substantially increase transfers byEIAV RT.

HIV-1 NC Improves the Docking Step of EIAV RT—The effectof NC on RNase H cleavages suggested that at least part of itsenhancement of transfers by EIAV RT could derive from more

FIG. 6. Effect of using blocking oli-gonucleotide on transfer efficiency.Transfer assays were carried out withHIV-1 RT and EIAV RT in the absence(filled bars) and presence of a DNA oli-gomer (gray bars) (see “Experimental Pro-cedures” for details). Assays were per-formed using DI donor and AI-2 acceptorat a 1:5 ratio, in the absence and presenceof NC at 200% coating level. Descriptionand details are same as in Fig. 2. Sche-matic of the RNA templates showing ac-tion of the blocking oligomer duringtransfer is indicated above the panel.

Retrovirus Recombination Mechanism31542

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

efficient cuts followed by more docking. To test this hypothesiswe measured strand transfers with the shortened ssAcceptor inthe presence of NC. In the presence of NC between an 8- and10-fold increase in transfers was observed with the short ac-ceptor (Fig. 4), regardless of acceptor concentration. This in-crease in transfers could not have arisen from enhanced dock-ing at the base of the hairpin, because that region of thesubstrate was deleted. Even if more cleavages were madethroughout the donor RNA, the highly annealed structure ofthe acceptor should have suppressed docking. The robust in-crease in transfers seen with NC could have resulted from acombination of an increase in cleavages throughout the donor,possibly creating new invasion sites, and a reduction in thestability of the acceptor structure allowing previously inacces-sible regions to dock with the cDNA.

We were able to distinguish docking sites by using blockingoligonucleotides. These are short DNA molecules complemen-tary only to a specific region of the cDNA and not to sequencesof any other strands in the assay (54, 56). For the originalacceptor a blocking oligonucleotide complementary to thecDNA at the base of the hairpin was used (54). Adding thisoligonucleotide to the HIV-1 RT reactions reduced transfers by60%, but it did not have a measurable effect on transfers byEIAV RT (Fig. 6). We attributed this result to the poor hairpin-base docking carried out by EIAV RT. A paradoxical effect wasobserved when oligonucleotide interference was carried out inthe presence of NC. HIV-1 RT transfers were less efficientlyblocked in the presence of NC (Fig. 6), a result previouslyobserved (54). However, interference was now observed for theEIAV RT, and as much as a third of all transfers could beblocked. In the case of HIV we determined that NC destabilizedthe blocking oligonucleotide-cDNA complex, thus reducing itsefficiency (data not shown). However, the result with EIAV RTsuggested that an NC-induced improvement in invasions at thebase of the hairpin was more effective than the destabilizationof any blocking oligonucleotide-cDNA complexes made.

HIV-1 NC Significantly Affects the Distribution of Transfers

by EIAV RT—We had previously reported that the presence ofNC protein did not significantly change the distribution oftransfers by HIV-1 RT at high salt concentration (80 mM KCl)(54). We re-checked the effects of NC on the distributions at thecurrently used 50 mM KCl. A bi-modal distribution of transferswas still observed for HIV-1 RT (Fig. 7); however, 40% fewertransfers took place at the base of the hairpin and 20% fewerbetween markers 4 and 5. An increase in transfers occurredbetween markers 5 and 7 (14.3–26%). Despite these slightdifferences, the larger features of the distribution were pre-served. A substantial alteration in the distribution of EIAVtransfer products occurred in the presence of NC. Without NC,most transfers took place between markers 4 and 5, whereas inits presence they occurred mostly beyond marker 5 (Fig. 7).This constituted a shift in transfers toward the 5� end ofthe template.

DISCUSSION

We recently proposed that strand transfer in retrovirusesoccurs through a two-step mechanism (54). An initial dockingstep starts the process on the donor RNA template, whereas afinal “locking” step completes it on the acceptor template. Weobserved these two steps separately through the use of highlystructured RNA templates. In the present work we address themanner in which enzyme-dependent factors, such as pausingand RNase H activity of the RT, as well as template structureand NC protein affect each of the two steps and therefore theoverall transfer process.

Pausing during synthesis has been proposed to improvetransfer, because the RT can concentrate RNase H cuts whilestalled (36–38, 54, 56). We observed an inhibition of transferwith both the ssAcceptor, which lacks sequences at the base ofthe hairpin (docking site), and a blocking oligonucleotide,which interferes with the docking site. This suggests that dock-ing occurred predominantly at the base of the hairpin, a pre-ferred RNase H cleavage region on the donor template. Wepropose that docking, which is the first step in the transfer,

FIG. 7. Effect of NC on distribution of primer terminus transfers for HIV-1 and EIAV RT. Transfer reactions were carried out usingHIV-1 and EIAV RTs at a 1:2 donor to acceptor ratio with 200% NC coating. Transfer products were gel-purified, amplified, and cloned, andindividual colonies were sequenced. Two independent experiments were carried out for each set. The distributions were calculated as in Fig. 5legend. For reactions performed using HIV-1 RT (black bars) a total of 60 clones were sequenced; for reactions using EIAV RT (gray bars) a totalof 57 clones were sequenced.

Retrovirus Recombination Mechanism 31543

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

occurs at locations where the RNase H activity of the RT clearsaway a segment of the donor RNA so that the cDNA caninteract with structurally accessible sections of the acceptor.

Comparisons of the transfers supported by the EIAV andHIV-1 RTs reveal that efficient docking is required for templateswitching and that this step is strongly influenced by the en-zymatic properties of the RTs. We found the pausing profiles ofthe EIAV and HIV-1 RTs to be very similar. In particular, theyboth paused strongly once synthesis reached the base of thehairpin. However, HIV-1 RT was much better at promotingstrand transfers than EIAV RT. We believe that this differencederives from the nature of the cleavages executed by the RTswhile stalled at this major pause site. HIV-1 RT catalyzed anumber of closely spaced cuts, which efficiently exposed thecDNA for docking. EIAV RT made a single prominent cut18–20 nucleotides away from the pause site. This limited cleav-age is unlikely to create an efficient docking site for invasion,thereby lowering its efficiency of transfers.

Increasing the acceptor concentration leads to an overallincrease in template switching by HIV-1 RT, most likely byincreasing the frequency of docking interactions (Fig. 9C). Highacceptor concentration did not substantially increase the trans-fer efficiency of EIAV, presumably because it could not improvethe initially poor docking efficiency.

Studies have shown that HIV-1 RT makes a primary RNaseH cut �18 nucleotides behind the primer terminus on the RNAstrand, followed by secondary cuts 8–9 nucleotides behind theprimer terminus (85, 86). To make a secondary cut, HIV-1 RThas to translocate forward to a new position after making theprimary cut (86). Once paused at the base of the hairpin, EIAVRT made a cut �18 nucleotides away, but unlike HIV-1 RT it isnot able to make cuts closer to the hairpin. In addition, weobserved that EIAV RT had low strand displacement activitywhen paused at the base of the hairpin. Possibly, structuralfeatures of the EIAV RT that restrict strand displacementsynthesis also limit the movement necessary to make second-ary RNase H cleavages. As described above, this would restrictdocking and lower the transfer efficiency. An analogous mech-anism has been described in the case of an RT inhibitor whichspecifically restricts strand transfer by inhibiting RNase Hsecondary cuts (87).

HIV-1 NC is an RNA chaperone protein that promotes an-nealing of complementary strands and also influences the fold-

ing and stability of nucleic acid structures (64, 88–92). More-over, it has been shown previously to enhance overall RNase Hactivity and secondary RNase H cuts by HIV-1 RT (86). HIV-1

FIG. 8. Distribution of primer ter-minus transfers normalized to thetransfer efficiency. Distribution ofprimer terminus transfers for HIV-1 andEIAV RTs at 1:2 and 1:8 donor to acceptorratio in the absence of NC (Fig. 5, A andB) and at 1:2 donor:acceptor ratio in thepresence of NC (Fig. 7) were normalizedfor the transfer efficiency measured un-der that specific condition (Fig. 2C) foreach RT. This was done by multiplyingthe fraction of transfers taking place ineach segment for a given set of assay con-ditions by the transfer efficiency meas-ured for each RT under those conditions.Conditions are as follows: donor to accep-tor ratio of 1–2 (black bars), donor to ac-ceptor ratio of 1–8 (gray bars), donor toacceptor ratio of 1–2, and 200% NC coat-ing (white bars).

FIG. 9. Dock and lock model of strand transfer from hairpinstructures. Schematic depicts mechanisms in the overall transferprocess. Once synthesis on the donor (thick black lines) reaches the baseof the hairpin (A), it pauses and creates a set of adjacent cuts (B) thatallows the acceptor (thick gray lines) to dock (base pair) with thenascent cDNA (thin black line) at the exposed site (C). This could leadeither to continued synthesis on the donor while the acceptor is docked(D), or to immediate transfer of the cDNA to the acceptor (E). In theformer case, the invasion can lead to branch migration (F) which even-tually locks the cDNA onto the acceptor (H) and transfer is completed.In the latter case, locking occurs in the vicinity of the initial pause site,completing the transfer and synthesis continues on the acceptor (G).

Retrovirus Recombination Mechanism31544

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

NC increased transfers by both RTs. We observed that NCwidened the distribution of cuts at the proposed invasion sitefor both HIV-1 RT and EIAV RT. For EIAV RT, this very likelypromoted docking, as evidenced by increased inhibition by thecorresponding blocking oligonucleotide, and resulted in moreefficient transfers. Both with and without NC, high transferefficiency correlated with the production of a distribution ofcuts. Furthermore, the stimulatory effects of HIV-1 NC onHIV-1 RT transfer efficiency is partly attributable to its chap-erone properties, because HIV-1 RT is capable of efficientlycleaving the donor RNA and facilitating docking in its absence.Interestingly, only in the presence of HIV-1 NC were both RTsefficient in promoting transfers with the ssAcceptor. We believethat this results from the creation of new docking sites throughincreased cutting of the donor RNA and to the relief of struc-tural constraints on the acceptor RNA. This would make pre-viously unavailable regions accessible for base pairing with thecDNA and hence for invasion.

Once docked, what triggers the primer terminus transfer orlock step? We had observed previously that although a fractionof the primers locked close to their docking site (Fig. 9, C andE), for the majority of primers, RT continued synthesis past thepause site on the donor (54). Whereas the site of dockingremained unchanged, this synthesis allowed for the locking tooccur further downstream (Fig. 9, C and D) (54). The location ofgenetic shift occurs at the point where the primer terminus istransferred. To compare the locking distributions, the transferefficiency for each of the RTs at all conditions tested wasmultiplied by the fraction of transfers in each template seg-ment and re-plotted as shown in Fig. 8.

We noted a correlation between the intensity and location ofpause sites beyond the base of the hairpin and the distributionof locking. For example, EIAV RT paused more strongly be-tween markers 6 and 7 (pause �16) than HIV-1 RT, and at lowacceptor concentration more transfers mapped in this regionfor EIAV RT than for HIV-1 RT (Fig. 8). Possibly, pause sitesobserved later in synthesis slow down the polymerizing RT,allowing the invasion initiated at the dock site to catch up tothe primer terminus through branch migration, completing thetransfer event (Fig. 9, F and H). Another possibility is that thelatter pause sites create a second interaction site between theprimer terminus and acceptor, in addition to the docking. Wehave already shown that transfers by HIV-1 RT with ssAccep-tor, although very inefficient, result in a similar distribution astransfers from the original length template (54). This suggeststhat some sequences may initiate locking even before branchmigration from the docking site catches up to the primer ter-minus. DeStefano and colleagues (67) have shown that theprocess of branch migration is necessary to complete transfersafter docking by HIV-1 RT but that it is not efficient in theabsence of NC. This second interaction is likely to promotebranch migration, which ultimately spans the distance be-tween docking and locking sites.

Better docking at high acceptor concentrations shifted moreof the primer terminus transfers closer to the docking site (Fig.8), probably by forming the docking intermediate before the RTcould move very far forward on the donor, favoring step E overstep D in Fig. 9. The effect of NC on the location of primerterminus transfer may derive from its chaperone properties.NC has been recently shown to enhance fraying (partial unan-nealing) at the ends of nucleic acids duplexes (93–96). A reduc-tion of the stability of the structure of the acceptor and in-creased fraying at the ends of paused primers should makelocking events more likely to occur within previously inacces-sible regions of the hairpin. Also, we observed that for HIV-1RT, NC permitted less transfers at the base of the hairpin than

are observed at high acceptor concentration (Fig. 8). NC in-creased primer extension past the base of the hairpin, thusincreasing the separation between the site of docking and thatof locking (Fig. 9, C and D). Clearly NC and higher acceptorconcentration both promote transfer but by at least partiallydifferent mechanisms.

Overall, our results suggest that efficient strand transferrequires promotion of both the dock and lock steps. Docking ismost efficient when the RT pauses in a way that makes a seriesof adjacent RNase H cleavages. Invasion at the site is promotedat high acceptor concentration and by the presence of NC. Thelocking step is most proficient in regions of weak pausing thatlack stable structure and is promoted by the chaperone prop-erties of NC.

Acknowledgments—We thank Roslyn Chang for help in purifyingHIV-1 RT. We also thank Yan Chen for helpful discussions.

REFERENCES

1. Blair, D. G. (1977) Virology 77, 534–5442. Kawai, S., and Hanafusa, H. (1972) Virology 49, 37–443. Vogt, P. K. (1971) Virology 46, 947–9524. Hu, W. S., and Temin, H. M. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,

1556–15605. Hu, W. S., and Temin, H. M. (1990) Science 250, 1227–12336. Jetzt, A. E., Yu, H., Klarmann, G. J., Ron, Y., Preston, B. D., and Dougherty,

J. P. (2000) J. Virol. 74, 1234–12407. Yu, H., Jetzt, A. E., Ron, Y., Preston, B. D., and Dougherty, J. P. (1998) J. Biol.

Chem. 273, 28384–283918. Schwartzberg, P., Colicelli, J., and Goff, S. P. (1985) J. Virol. 53, 719–7269. Colicelli, J., and Goff, S. P. (1987) Virology 160, 518–522

10. Murphy, J. E., and Goff, S. P. (1994) Virology 204, 458–46111. Mikkelsen, J. G., Lund, A. H., Kristensen, K. D., Duch, M., Sorensen, M. S.,

Jorgensen, P., and Pedersen, F. S. (1996) J. Virol. 70, 1439–144712. Gu, Z., Salomon, H., Cherrington, J. M., Mulato, A. S., Chen, M. S., Yarchoan,

R., Foli, A., Sogocio, K. M., and Wainberg, M. A. (1995) Antimicrob. AgentsChemother. 39, 1888–1891

13. Gunthard, H. F., Leigh-Brown, A. J., D’Aquila, R. T., Johnson, V. A.,Kuritzkes, D. R., Richman, D. D., Wong, J. K., and D’Aquila, R. T. (1999)Virology 259, 154–165

14. Kellam, P., and Larder, B. A. (1995) J. Virol. 69, 669–67415. Robertson, D. L., Sharp, P. M., McCutchan, F. E., and Hahn, B. H. (1995)

Nature 374, 124–12616. Peeters, M., and Sharp, P. M. (2000) AIDS 14, S129–S14017. Temin, H. M. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 6900–690318. Gilboa, E., Mitra, S. W., Goff, S., and Baltimore, D. (1979) Cell 18, 93–10019. Goff, S. P., and Telesnitsky, A. (1993) in Reverse Transcriptase (Goff, S. P., and

Skalka A. M., ed) pp. 49–83, Cold Spring Harbor Laboratory Press, ColdSpring Harbor, NY

20. Anderson, J. A., Teufel, R. J., II, Yin, P. D., and Hu, W. S. (1998) J. Virol. 72,1186–1194

21. Clavel, F., Hoggan, M. D., Willey, R. L., Strebel, K., Martin, M. A., andRepaske, R. (1989) J. Virol. 63, 1455–1459

22. Goodrich, D. W., and Duesberg, P. H. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,2052–2056

23. Zhang, J., Tang, L. Y., Li, T., Ma, Y., and Sapp, C. M. (2000) J. Virol. 74,2313–2322

24. Pathak, V. K., and Temin, H. M. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,6019–6023

25. Pathak, V. K., and Temin, H. M. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,6024–6028

26. Pfeiffer, J. K., Georgiadis, M. M., and Telesnitsky, A. (2000) J. Virol. 74,9629–9636

27. Svarovskaia, E. S., Delviks, K. A., Hwang, C. K., and Pathak, V. K. (2000)J. Virol. 74, 7171–7178

28. Zhang, J., and Temin, H. M. (1993) Science 259, 234–23829. Delviks, K. A., and Pathak, V. K. (1999) J. Virol. 73, 7923–793230. Halvas, E. K., Svarovskaia, E. S., and Pathak, V. K. (2000) J. Virol. 74,

10349–1035831. Skalka, A. M., Boone, L., Junghans, R., and Luk, D. (1982) J. Cell. Biochem.

19, 293–30432. Coffin, J. M. (1979) J. Gen. Virol. 42, 1–2633. Hu, W. S., and Temin, H. M. (1992) J. Virol. 66, 4457–446334. Stuhlmann, H., and Berg, P. (1992) J. Virol. 66, 2378–238835. Hu, W. S., Bowman, E. H., Delviks, K. A., and Pathak, V. K. (1997) J. Virol. 71,

6028–603636. DeStefano, J. J., Buiser, R. G., Mallaber, L. M., Fay, P. J., and Bambara, R. A.

(1992) Biochim. Biophys. Acta 1131, 270–28037. DeStefano, J. J., Bambara, R. A., and Fay, P. J. (1994) J. Biol. Chem. 269,

161–16838. Wu, T., Guo, J., Bess, J., Henderson, L. E., and Levin, J. G. (1999) J. Virol. 73,

4794–480539. Baltimore, D. (1970) Nature 226, 1209–121140. Temin, H. M., and Mizutani, S. (1970) Nature 226, 1211–121341. Luo, G. X., and Taylor, J. (1990) J. Virol. 64, 4321–432842. Peliska, J. A., and Benkovic, S. J. (1992) Science 258, 1112–111843. DeStefano, J. J., Mallaber, L. M., Rodriguez-Rodriguez, L., Fay, P. J., and

Retrovirus Recombination Mechanism 31545

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

Bambara, R. A. (1992) J. Virol. 66, 6370–637844. DeStefano, J. J., Roberts, B., and Shriner, D. (1997) Arch. Virol. 142,

1797–181245. Brincat, J. L., Pfeiffer, J. K., and Telesnitsky, A. (2002) J. Virol. 76, 88–9546. Hwang, C. K., Svarovskaia, E. S., and Pathak, V. K. (2001) Proc. Natl. Acad.

Sci. U. S. A. 98, 12209–1221447. Berkhout, B., Vastenhouw, N. L., Klasens, B. I., and Huthoff, H. (2001) RNA

(New York) 7, 1097–111448. Kim, J. K., Palaniappan, C., Wu, W., Fay, P. J., and Bambara, R. A. (1997)

J. Biol. Chem. 272, 16769–1677749. Moumen, A., Polomack, L., Roques, B., Buc, H., and Negroni, M. (2001) Nucleic

Acids Res. 29, 3814–382150. Moumen, A., Polomack, L., Unge, T., Veron, M., Buc, H., and Negroni, M.

(2003) J. Biol. Chem. 278, 15973–1598251. Negroni, M., and Buc, H. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 6385–639052. Negroni, M., and Buc, H. (2001) Annu. Rev. Genet. 35, 275–30253. Negroni, M., and Buc, H. (2001) Nat. Rev. Mol. Cell. Biol. 2, 151–15554. Roda, R. H., Balakrishnan, M., Kim, J. K., Roques, B. P., Fay, P. J., and

Bambara, R. A. (2002) J. Biol. Chem. 277, 46900–4691155. Balakrishnan, M., Fay, P. J., and Bambara, R. A. (2001) J. Biol. Chem. 276,

36482–3649256. Balakrishnan, M., Roques, B. P., Fay, P. J., and Bambara, R. A. (2003) J. Virol.

77, 4710–472157. Harrison, G. P., Mayo, M. S., Hunter, E., and Lever, A. M. (1998) Nucleic Acids

Res. 26, 3433–344258. Klarmann, G. J., Schauber, C. A., and Preston, B. D. (1993) J. Biol. Chem. 268,

9793–980259. Negroni, M., and Buc, H. (1999) J. Mol. Biol. 286, 15–3160. Darlix, J. L., Lapadat-Tapolsky, M., de Rocquigny, H., and Roques, B. P. (1995)

J. Mol. Biol. 254, 523–53761. Darlix, J. L., Vincent, A., Gabus, C., de Rocquigny, H., and Roques, B. (1993)

C. R. Acad. Sci. III (Paris) 316, 763–77162. Allain, B., Lapadat-Tapolsky, M., Berlioz, C., and Darlix, J. L. (1994) EMBO J.

13, 973–98163. Peliska, J. A., Balasubramanian, S., Giedroc, D. P., and Benkovic, S. J. (1994)

Biochemistry 33, 13817–1382364. You, J. C., and McHenry, C. S. (1994) J. Biol. Chem. 269, 31491–3149565. DeStefano, J. J. (1995) Arch. Virol. 140, 1775–178966. Rodriguez-Rodriguez, L., Tsuchihashi, Z., Fuentes, G. M., Bambara, R. A., and

Fay, P. J. (1995) J. Biol. Chem. 270, 15005–1501167. DeStefano, J. J. (1996) J. Biol. Chem. 271, 16350–1635668. Guo, J., Henderson, L. E., Bess, J., Kane, B., and Levin, J. G. (1997) J. Virol.

71, 5178–518869. Davis, W. R., Gabbara, S., Hupe, D., and Peliska, J. A. (1998) Biochemistry 37,

14213–1422170. Raja, A., and DeStefano, J. J. (1999) Biochemistry 38, 5178–518471. Johnson, P. E., Turner, R. B., Wu, Z. R., Hairston, L., Guo, J., Levin, J. G., and

Summers, M. F. (2000) Biochemistry 39, 9084–909172. Montelaro, R., Ball, J. M., and Rushlow, K. (1992) in Viruses (Levy, J. A., ed)

Vol. 4, pp. 257–360, Plenum Publishing Corp., New York73. Coffin, J. M., Hughes, S. H., and Varmus, H. (1997) Retroviruses, pp. 88–89,

Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY74. Stephens, R. M., Casey, J. W., and Rice, N. R. (1986) Science 231, 589–59475. DeVico, A., Montelaro, R. C., Gallo, R. C., and Sarngadharan, M. G. (1991)

Virology 185, 387–39476. Le Grice, S. F., Panin, M., Kalayjian, R. C., Richter, N. J., Keith, G., Darlix,

J. L., and Payne, S. L. (1991) J. Virol. 65, 7004–700777. Shaharabany, M., Rice, N. R., and Hizi, A. (1993) Biochem. Biophys. Res.

Commun. 196, 914–92078. Rubinek, T., Loya, S., Shaharabany, M., Hughes, S. H., Clark, P. K., and Hizi,

A. (1994) Eur. J. Biochem. 219, 977–98379. Wohrl, B. M., Howard, K. J., Jacques, P. S., and Le Grice, S. F. (1994) J. Biol.

Chem. 269, 8541–854880. Rausch, J. W., Arts, E. J., Wohrl, B. M., and Le Grice, S. F. (1996) J. Mol. Biol.

257, 500–51181. Lee, R., Kaushik, N., Modak, M. J., Vinayak, R., and Pandey, V. N. (1998)

Biochemistry 37, 900–91082. de Rocquigny, H., Ficheux, D., Gabus, C., Fournie-Zaluski, M. C., Darlix, J. L.,

and Roques, B. P. (1991) Biochem. Biophys. Res. Commun. 180, 1010–101883. Pandey, V. N., Kaushik, N., Rege, N., Sarafianos, S. G., Yadav, P. N., and

Modak, M. J. (1996) Biochemistry 35, 2168–217984. Le Grice, S. F., Naas, T., Wohlgensinger, B., and Schatz, O. (1991) EMBO J.

10, 3905–391185. Wisniewski, M., Balakrishnan, M., Palaniappan, C., Fay, P. J., and Bambara,

R. A. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 11978–1198386. Wisniewski, M., Chen, Y., Balakrishnan, M., Palaniappan, C., Roques, B. P.,

Fay, P. J., and Bambara, R. A. (2002) J. Biol. Chem. 277, 28400–2841087. Davis, W. R., Tomsho, J., Nikam, S., Cook, E. M., Somand, D., and Peliska,

J. A. (2000) Biochemistry 39, 14279–1429188. De Rocquigny, H., Gabus, C., Vincent, A., Fournie-Zaluski, M. C., Roques, B.,

and Darlix, J. L. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 6472–647689. Dib-Hajj, F., Khan, R., and Giedroc, D. P. (1993) Protein Sci. 2, 231–24390. Herschlag, D. (1995) J. Biol. Chem. 270, 20871–2087491. Prats, A. C., Housset, V., de Billy, G., Cornille, F., Prats, H., Roques, B., and

Darlix, J. L. (1991) Nucleic Acids Res. 19, 3533–354192. Tsuchihashi, Z., and Brown, P. O. (1994) J. Virol. 68, 5863–587093. Urbaneja, M. A., Wu, M., Casas-Finet, J. R., and Karpel, R. L. (2002) J. Mol.

Biol. 318, 749–76494. Hong, M. K., Harbron, E. J., O’Connor, D. B., Guo, J., Barbara, P. F., Levin,

J. G., and Musier-Forsyth, K. (2003) J. Mol. Biol. 325, 1–1095. Azoulay, J., Clamme, J. P., Darlix, J. L., Roques, B. P., and Mely, Y. (2003) J.

Mol. Biol. 326, 691–70096. Bernacchi, S., Stoylov, S., Piemont, E., Ficheux, D., Roques, B. P., Darlix, J. L.,

and Mely, Y. (2002) J. Mol. Biol. 317, 385–399

Retrovirus Recombination Mechanism31546

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from

Grice, Bernard P. Roques, Robert J. Gorelick and Robert A. BambaraRicardo H. Roda, Mini Balakrishnan, Mark N. Hanson, Birgitta M. Wöhrl, Stuart F. J. Le

the Two-step Transfer Mechanism in Retroviral RecombinationRole of the Reverse Transcriptase, Nucleocapsid Protein, and Template Structure in

doi: 10.1074/jbc.M304608200 originally published online June 11, 20032003, 278:31536-31546.J. Biol. Chem. 

  10.1074/jbc.M304608200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/278/34/31536.full.html#ref-list-1

This article cites 94 references, 50 of which can be accessed free at

by guest on June 16, 2020http://w

ww

.jbc.org/D

ownloaded from