Specificity of Initiation of Plus-Strand DNA by Rous Sarcoma Virus JULIE K

Home , Restriction enzyme

JOURNAL OF VIROLOGY, Nov. 1984, p. 314-319 Vol. 52, No. 2 0022-538X/84/110314-06$02.00/0 Copyright C 1984, American Society for Microbiology Specificity of Initiation of Plus-Strand DNA by Rous Sarcoma Virus JULIE K. SMITH, ANITA CYWINSKI, AND JOHN M. TAYLOR* The Institute for Cancer Research, Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111 Received 3 May 1984/Accepted 19 July 1984

We previously reported that in the endogenous reaction of Rous sarcoma virus disrupted by melittin, plus- strand DNA initiates on a small oligonucleotide primer and that this initiation can be reconstructed in vitro in reactions containing purified minus-strand DNA as template, viral RNA as a source of primer, and reverse transcriptase (Smith et al., J. Virol. 49:200-204, 1984). Further studies on the specificity of initiation in the endogenous reaction have shown the following. (i) The primer was 12 nucleotides in length. Its sequence began with a 5' pyrimidine, followed by 11 purines, ending with rGrA-3'. This sequence was in agreement with the known plus-strand RNA sequence immediately upstream from the initiation site. Thus, the primer began one nucleotide 5' to the so-called polypurine tract that has been found on all retrovirus genomes. (ii) The transition point between RNA primer and DNA product was precisely located. It was before the end of the polypurine tract. Thus the polypurine tract, although essential for virus replication and probably a flag for the priming event, did not define the limits of the RNA primer. After primer removal, the DNA had a 5' phosphate, consistent with generation by the viral RNase H activity. The priming specificity in reconstructed reactions was also examined further, with the following observations. (i) When the source of RNA primer was prehybridized to the template viral DNA, the generation, utilization, and subsequent removal of primer were essentially the same as those observed in the endogenous reaction. In the absence of deliberate prehybridization, some specificity was lost. There were then additional locations for the 5' end of the primer as well as the transition point between RNA primer and DNA. (ii) Purine-rich oligoribonucleotides created by RNase A digestion of viral RNA could prime strong-stop plus DNA, but again with the loss of specificity relative to that in the endogenous reaction. (iii) The 5' end of the minus-strand DNA template was not required for initiation of strong-stop plus DNA. Therefore, the specificity of initiation did not depend upon an intramolecular interaction requiring the two inverted repeat sequences that flank the long terminal repeat.

During the replication of retroviruses a viral enzyme, junctions have been detected in the total plus-strand DNA reverse transcriptase, copies the viral RNA genome into a synthesized in a reconstructed reaction (17). Our data on the double-stranded DNA intermediate (reviewed in reference specificity of initiation, as presented here, are consistent 27). Minus-strand DNA (complementary to the viral RNA with such a role for RNase H. genome) is synthesized first, using a cellular tRNA as the In all retroviruses examined, the initiation site of strong- primer (5). The growing minus-strand DNA is the template stop plus DNA is adjacent to a series of purine nucleotides in for the synthesis of plus-strand DNA. The first plus-strand the viral RNA (28). This polypurine tract (PPT) is required in DNA seen is a discrete species called strong-stop plus DNA cis for retroviral replication (22). Thus the circumstantial (14, 29). Strong-stop plus DNA is initiated at a precise evidence is that the PPT plays a role in the priming of plus- location (13). For avian sarcoma virus, this species is ca. 340 strand DNA synthesis. Our studies have determined the nucleotides in length (29). actual relationship between this PPT and the RNA species We previously demonstrated that in the endogenous reac- used as the primer. tion of avian sarcoma virus virions permeabilized with melittin, the strong-stop plus DNA is initiated on an oligori- MATERIALS AND METHODS bonucleotide primer (21). The specific initiation is the result Virus, cells, and reagents. The Prague A strain of Rous of three precise nucleolytic events: two to generate the sarcoma virus was derived from the transformed quail clone discrete ends of the primer and a third to remove the primer Q-PrA-4 (19). The virus was propagated in quail embryo after use. fibroblasts as previously described (7). Restriction enzymes We also showed that strong-stop plus initiation can be were used according to the instructions of the manufacturer reconstructed with purified components: minus-strand DNA (Bethesda Research Laboratories), with the addition of 0.5 as template, high-molecular-weight viral RNA as a source of U of Ribonucleasin (Bolton Biologicals) per ,ll. Reverse primers, and reverse transcriptase (21). Champoux et al. (3) transcriptase, purified from avian myeloblastosis virus, was have used similar reconstructed reactions to show that generously provided by the Office of Program Resources and RNase A-digested viral RNA can provide primers for the Logistics, Virus Cancer Program, Beltsville, Md. initiation of strong-stop plus DNA by Moloney murine Plasmid pSRA2, which contains Rous sacroma virus DNA leukemia virus. (Schmidt-Ruppin A strain) inserted into pBR322, was ob- It has been suggested that degradation of the viral RNA by tained from W. DeLorbe and co-workers (4). RNase H, an activity of the reverse transcriptase, could DNA synthesis and analysis. Conditions for the synthesis provide primers for plus-strand DNA synthesis (27, 29). and analysis of strong-stop plus DNA from endogenous Such a mechanism has been demonstrated in the reverse reactions and from reconstructed reactions were as previ- transcription of homopolymers (31). In addition, RNA-DNA ously described (21) and are presented briefly here. In the endogenous reactions, pelleted virions were per- * Corresponding author. meabilized with melittin and incubated with deoxyribonucle- 314 VOL. 52, 1984 SPECIFIC INITIATION OF RETROVIRAL PLUS-STRAND DNA 315

Hpa Sph Eco Hr larger than band D represents strong-stop plus DNA in H_-/ < DNA- which only a single base has been copied from the tRNA " A I | I§ - DNA + primer. Species A and B correspond to the 5' EcoRI fragments with or without the oligoribonucleotide primer. I /f I ~I II Relative to the single-nucleotide calibration provided by the adjacent sequencing ladder, the primer was reproducibly -570 0 200 400 measured to be 12 bases in length (Fig. 2; unpublished data). To confirm this size assignment and to characterize the NUCLEOT IDES sequence of the primer, samples were digested to completion FIG. 1. Features of the viral DNA around the origin of strong- with base-specific RNases. RNase A, which cuts 3' to stop plus DNA. DNA- represents the minus-strand template, and the one nucleotide DNA+ represents the two sizes of strong-stop plus DNA. Positions pyrimidines, reduced primer length by of restriction enzyme sites used in this study are based on the (Fig. 2a, lane 1). RNase T1, which cuts 3' to rG, reduced the nucleotide sequence data of Schwartz et al. (20). Hpa, HpaII; Sph, primer to one nucleotide in length (Fig. 2a, lane 3). RNase SphI; Eco, EcoRI. Also shown are the inverted repeat sequences U2, which cuts 3' to rA, and alkali, which hydrolyzes all (>ol) and the RNA primers used to initiate each DNA strand ribonucleotides, both removed the entire primer (Fig. 2a, (w)- lanes 4 and 5, respectively). From these digestions the sequence of the RNA primer was deduced to be 5'-Y(R)9GA- 3' where Y = T or C and R = A or G. This sequence is otide triphosphates. Labeled DNA products were made by consistent with the known RNA sequence of the viral RNA including [a-32P]TTP in the reaction mixture. After digestion (20) immediately upstream from the initiation site. with either EcoRI or SphI, the DNA product was denatured and hybridized in solution to an excess of DNA from a (a) (b) recombinant M13 bacteriophage which contained the minus Y R l 2 3 4 5 2 3 strand of the long terminal repeat (LTR) (25). The hybrid b)1 AI molecules were isolated by rate zonal sedimentation. This .N.A provided a partial purification of those labeled molecules 0 containing sequences related to strong-stop plus DNA. This *4s_" -A 0 DNA was then analyzed on an 80-cm gel of 6% acrylamide- _: s..~~ *T urea. A cloned fragment of DNA which contained the origin of strong-stop plus DNA (an EcoRI-PvuII fragment of 0 pSRA2) was submitted to base-specific cleavages described . S by Maxam and Gilbert (11) and served as a size standard on : a , B~~~ *_. so the high-resolution acrylamide gels. For the reconstructed reactions we used as a template minus-strand DNA synthesized in an endogenous reaction. The higher-molecular-weight DNA species were isolated by two cycles of rate zonal sedimentation through alkaline sucrose. This DNA was incubated with viral RNA and reverse transcriptase. Labeled DNA products were made by the addition of [(x-32P]TTP to the reaction. Labeled species related to strong-stop plus DNA were partially purified and FIG. 2. Acrylamide gel electrophoresis of strong-stop plus DNA then analyzed by gel electrophoresis as described above. from endogenous reactions. Radiolabeled DNA synthesized by Unlabeled reaction products were glyoxalated (12) and sub- melittin-disrupted virions was digested with EcoRI (a) or SphI (b) to on slab 2% agarose. and then hybridized in solution to an excess of recombinant M13 jected electrophoresis gels of After DNA containing inserted LTR minus-strand sequences. Hybrids transfer to nitrocellulose (23), sequences related to strong- were isolated by sedimentation on a 5 to 20% sucrose gradient. The stop plus or its complement were detected with probes made fractions enriched for strong-stop plus DNA were subjected to the from recombinant M13 DNAs containing inserted LTR treatments indicated and then resolved on 80-cm gels of 6% acryl- sequences (9, 25). amide-urea. The dots indicate single nucleotide spacings, and labeled bands are as described in the text. (a) EcoRI digestion. At RESULTS the left is a sequencing ladder generated from cloned viral DNA Size and sequence of the RNA primer in endogenous reac- containing the origin of strong-stop plus DNA which was 3'-end tions. In our previous studies, the RNA primer on plus- labeled at the EcoRI site. Sequencing reactions were performed by DNA to be uniform in size The the method of Maxam and Gilbert (11). Lanes: Y, (T + C)-specific strand appeared (21). reaction; R, (A + G)-specific reaction; 1, product digested with following studies were undertaken to determine whether the RNase A (10 mM Tris [pH 7.6], 1 mM EDTA, 10 mM NaCl; total primers were predominantly of a single sequence and, if so, reaction volume, 100 ,ul including 5 ,ug of tRNA and 10 p.g of RNase to compare that sequence to the published sequence of the A [Sigma Chemical Co.]; 30 min at 37°C); 2, untreated product; 3, viral DNA (20). To do this we purified those labeled species product digested with RNase T, (same reaction buffer as RNase A, from an endogenous reaction that were related to strong-stop 10 U of T1 [Calbiochem-Behring]; 30 min at 37°C); 4, product plus DNA. A map of the DNA around the origin of strong- digested with RNAse U2 (50 mM sodium acetate [pH 4.7], 2 mM stop plus and the locations of the restriction sites used in this EDTA; total reaction volume, 100 p.1 including 5 ,ug of tRNA and 5 study are shown in Fig. 1. This DNA was digested with U of U2 [Sankyo]; 20 min at 50°C); 5, product hydrolyzed with alkali and on an lane (0.3 N NaOH; 10 min at 100°C). Bands A to D refer to lane 2. (b) EcoRI analyzed acrylamide gel (Fig. 2a, 2). SphI digestion. Lanes: 1, untreated product; 2, product treated with As described previously (21), species C and D correspond to phosphatase (10 mM Tris [pH 8.0], 120 mM NaCl; total reaction the 3' EcoRI fragments of strong-stop plus DNA with or volume, 100 ,u1 with 125 U of bacterial alkaline phosphatase ([Bethesda without the 18 nucleotides copied from the minus-strand Research Laboratories]; 60 min at 65°C); 3, product hydrolyzed with tRNA primer. The labeled band which is one nucleotide alkali (0.3 N NaOH; 10 min at 100°C). A' and B' refer to lane 1. 316 SMITH, CYWINSKI, AND TAYLOR J. VIROL.

Removal of the RNA primer in endogenous reactions. To 1 2 3 4 5 6 examine the 5' terminus more closely, the strong-stop plus .,.17 DNA was digested with SphI. This enzyme cuts ca. 100 bases from the 5' end (Fig. 1), releasing fragments A' and B', -A which are smaller and hence better resolved than the corresponding fragments (A and B) released by EcoRI (Fig. 2b, lane 1). The migration of the fragments was measured by B scanning the autoradiograms with a densitometer and then measuring the distance between the various peaks relative to the single nucleotide spacing. When the sample was first treated with alkali, band A' disappeared, as expected, and a novel species appeared which migrated slightly slower than band B' (Fig. 2b, lane 2). The difference in migration relative to B' was equivalent to one third of a single nucleotide increment. The identity of the novel species became apparent when the sample was treated with alkaline phosphatase before electrophoresis (Fig. 2b, lane 3). Band B' then disappeared, and only the novel species remained. We interpreted these results to mean that the natural removal of the primer in the endogenous reaction leaves a 5' phosphate FIG. 3. Acrylamide gel electrophoresis of strong-stop plus DNA on strong-stop plus DNA. In contrast, removal of the primer from reconstructed reactions. Minus-strand viral DNA containing with alkali leaves a 5' OH, resulting in a species with a 0.6 ng of LTR minus-strand sequences was incubated in a reaction relatively slower electrophoretic mobility. The ability of a mixture containing reverse transcriptase and viral RNA as indicated. Labeled products related to strong-stop plus DNA were en- terminal phosphate to increase mobility has been previously riched for and analyzed on acrylamide gels as described in the described (24). RNase H is known to produce oligoribonu- legend to Fig. 2. Lanes: 2, 180 ng of viral RNA (containing 6.6 ng of cleotides with a 5'-terminal phosphate (2, 30). This experi- LTR sequences) added; 1, sample as in lane 2 except treated with ment showed that the strong-stop plus DNA also has a 5' alkali (0.3 N NaOH; 100°C; 10 min); 3, 180 ng of viral RNA added, phosphate after primer removal. prehybridized (to a Crt of 1 mol - s/liter) to DNA template; 4, sample Specificity of initiation in reconstructed reactions. Studies as in lane 3 except treated with alkali; 5, 180 ng of viral RNA added, were undertaken to compare the specificity of priming and predigested with RNase A; 6, sample as in lane 5 except treated with plus-strand initiation observed in endogenous reactions to alkali. Bands labeled A and B are as described in the text, and the that dots represent single nucleotide spacings from a sequencing ladder obtained in reconstructed reactions with purified com- run in an adjacent as in ponents: reverse transcriptase, minus-strand DNA as tem- lane, Fig. 2. plate, and viral RNA as a source of primers. The products from such reconstructed reactions, with or without alkali initiated in such a reconstructed reaction (using RNase A- treatment to remove the primer, are shown in Fig. 3, lanes 1 digested viral RNA as a source of primers) relative to an and 2. Similar to the products of endogenous reactions, we endogenous reaction. As before, labeled strong-stop plus- observed the bands previously designated A, B, and D. Band containing DNA products were isolated, digested with C was absent. This was expected since in the endogenous EcoRI, and subjected to electrophoresis on polyacrylamide reaction band C corresponds to the 3' end of strong-stop plus gels without or with a previous alkali treatment (Fig. 3, lanes DNA with an additional 18 nucleotides copied from the 5 and 6). The resulting bands were compared to an adjacent minus-strand tRNA primer. The minus-strand template used sequencing ladder of cloned DNA, as in Fig. 2A. The in the reconstructed reactions had been treated with both indicated band B in lane 6 corresponds to the single DNA RNase A and alkali so species C could not be made. For the initiation site, after primer removal, as seen in the endoge- products of an endogenous reaction, bands A and B were nous reactions (Fig. 2a). Clearly in this reconstructed reac- homogeneous (Fig. 2). However, for the reconstructed reaction DNA synthesis initiated at two additional sites, appar- tions a size heterogeneity was observed in both bands A and ently located one nucleotide on either side of that site. The B (Fig. 3, lanes 1 and 2). There were a different number of indicated band A in lane 5, with RNA primer still associated, extra species around band A as compared with band B, and showed the same degree of heterogeneity as band B (Fig. 3). we interpreted this to mean that the heterogeneity was due to We interpreted this as evidence that the primers were differences in both the size of the RNA primer and the predominantly of a single size, 11 bases, and that the position of the RNA-DNA junction. These heterogeneities heterogeneity around bands A and B was due to two new were significantly reduced in reconstructed reactions in positions for the location of the RNA-DNA junction. A which the template DNA and source of RNA primers were comparison of lanes 5 and 6 also suggests that RNase H hybridized before synthesis (Fig. 3, lanes 3 and 4). removes the RNA primers more efficiently from incorrectly We have previously shown that an RNase A digest of viral initiated molecules than from those initiated at the correct RNA is an efficient source of primers for the synthesis of location (Fig. 3). The significance of this finding is being strong-stop plus DNA in reconstructed reactions (21). This further investigated. was expected since RNase A only cleaves at pyrimidines and Role of the template DNA in specific plus-strand initiation. since the sequence upstream from the initiation site of In the above studies with endogenous reactions, and even strong-stop plus DNA is made up of purines, the PPT. Figure with reconstructed reactions, the size of the RNA primers 4d shows the RNA oligomer that would be generated from an and the location of the RNA-DNA junction were remarkably RNase A digestion of viral sequences around the strong-stop specific. The following experiment was undertaken to test an plus DNA origin. Further experiments were undertaken to hypothesis of how the viral minus-strand DNA template characterize the RNA species that were actually used as might contribute to this specificity. primers and to determine where the plus-strand DNA was The viral LTR is essentially a double-stranded copy of VOL. 52, 1984 SPECIFIC INITIATION OF RETROVIRAL PLUS-STRAND DNA 317

5'-UpGpCpAipUpApGpGpGpAp GpGpGpGpGpAlpApApUpGpUpA-3' (a) viral RNA I PPT II I I IpNpNpNpNpNpN- (b) endogenous I I DNA + IpYp RpRpRpRpRpRpRpRpRpGpA'pNpNpNpNpNpN- (c) DNA+ with primer A pRG p p p p p p p p G p Alp NAp p p I ApGpG pGpApG pGpG pG pG pApAp ApU p (d) RNAse A fragment

40 1 (e) reconstructed DNA + FIG. 4. Features of the nucleotide sequence around the origin of strong-stop plus DNA. N, Nucleotide; Y, pyrimidine; R, purine. (a) Published sequence of viral RNA (20); (b) deduced 5' end of strong-stop plus DNA, as synthesized in the endogenous reaction; (c) sequence of the RNA primer attached to strong-stop plus DNA from an endogenous reaction; (d) RNase A fragment that would be released from viral RNA in the region including strong-stop plus DNA; (e) interpretation of the three initiation sites observed in reconstructed reactions that used RNase A-digested viral RNA as a source of primers with indicated relative amounts, as deduced by densitometry (as from Fig. 3, lane 6).

strong-stop plus DNA. It is flanked by a short inverted virions several remarkably specific events are involved (Fig. repeat sequence (IR) which for Rous sarcoma virus is 15 4b and c). (i) Only one size of primer was seen, indicating bases in length with three mismatches (20). Thus on the precise 5' and 3' ends. (ii) The site of initiation of DNA was minus-strand DNA template, a copy of the IR is located always at the same nucleotide. (iii) The primer was apparent- precisely at the initiation site of strong-stop plus DNA. A ly removed in a single step, since intermediates containing a copy complementary to this is present at the 5' end of the partial primer were not detected. Furthermore, the sequence template (Fig. 1). This suggested that a possible mechanism of the primer was both uniform and consistent with the RNA for the specific primer generation and utilization might sequence (20) expected upstream from the origin of strong- involve an interaction between these two IR elements. To stop plus DNA. test this hypothesis we digested a sample of endogenously These specificities of the endogenous reaction were also synthesized DNA with HpaII before purifying the fragment achieved in reconstructed reactions with purified compo- containing the origin of strong-stop plus DNA on an alkaline nents. The necessary specificities thus could be provided by sucrose gradient. Such a digestion should remove the 5' the template DNA, the source of RNA primers, and reverse terminus of the minus-strand template, including one copy of transcriptase. Apparently there was no requirement for the IR (Fig. 1). We compared the activity of this DNA cellular or additional viral factors, such as specific nucleic template relative to undigested minus-strand template in acid binding proteins. This does not exclude the possibility reconstructed reactions. Unlabeled reaction products were that in vivo other factors could play an auxiliary role. The analyzed on agarose gels. After Southern transfer (23), LTR- observed in vitro use of RNA primers, in the absence of specific probes were used to detect template and product ribonucleotide precursors, has also ruled out de novo syn- DNA (Fig. 5). The three lanes labeled (-) are from recon- thesis of primers, for example, by a primase. structed reactions with the undigested minus-strand tem- The combined evidence indicates that the RNA primer is plate, and those labeled (+) are from reactions with the generated from the viral RNA by the action of the RNase H HpaII-digested template. An M13 probe specific for LTR activity present in the viral reverse transcriptase. The main minus-strand sequences was used to show the size of the argument is that it was the only nuclease available in the minus-strand DNA template. The undigested DNA template reconstructed reaction. RNase H degrades the RNA portion was large and heterogeneous, but after HpaII digestion was of DNA-RNA hybrids (15). The avian retrovirus RNase H reduced to a discrete band of 850 bases, the predicted size of generally releases oligoribonucleotides of 10 to 12 bases with the HpaII fragment-containing LTR sequences (Fig. 5, lanes a 5'-terminal phosphate (2, 30), in agreement with our 1). Lanes 2 and 3 show what was detected with an M13 probe observations. However, the documented properties of the specific for LTR plus-strand sequences in reconstructed viral RNase H are not sufficient to explain the specificities reactions with no added primers (lanes 2) or in reactions with we have observed. It may be that RNase H is more specific viral RNA added as a source of primers (lanes 3) (Fig. 5). than has been previously described, or additional factors The main plus-strand product from the undigested template may be involved. One such factor would be that we assay the was strong-stop plus DNA, whereas the major product from functional combination of primer generation and primer the Hpa II-digested template was 260 bases (Fig. 5, lanes 3). utilization; RNase H must generate many other oligonucleo- This was the size predicted for correct initiation of strong- tides that fail to act as primers. Another factor could be the stop plus DNA and for elongation to the HpaII site. In primary structure of the template. As others have noted, summary, this experiment did not detect any role for an every retroviral genome contains a PPT in this region (28), interaction between the IR sequences in the specific initia- and it is essential for infectivity (22). For another "retroid tion of strong-stop plus DNA. virus", cauliflower mosaic virus, there are two specific sites for plus-strand DNA initiation, and both have a PPT in the DISCUSSION immediate vicinity (6). Nevertheless, it is important to stress These studies have shown that in the synthesis of strong- that the PPT may have other roles, and as far as the initiation stop plus DNA in the endogenous reaction of disrupted of plus-strand DNA synthesis, the PPT is at most a flag. It is 318 SMITH, CYWINSKI, AND TAYLOR J. VIROL.

(-) (+) expected from this region of the viral RNA is 14 bases long (Fig. 4d). Before it could be used as a primer the 3' Kb M 1 23 123 ___, phosphate would need to be removed. Also, to be consistent with our data, an additional two to four nucleotides would 3.1 have to be removed. In a similar study with murine leukemia 2.4 virus, Champoux et al. detected two rather than three DNA initiation sites (3). The three initiations we detected began with deoxyadenosine (Fig. 4e), and this was also true for the two murine leukemia virus initiations. It is remarkable that of the many oligoribonucleotides released from viral RNA by RNase A, rather than RNase H, only a narrow-size class 0.3 l F SS+ were utilized as primers, and furthermore, the initiation site was almost coincident with that observed in endogenous FIG. 5. Agarose gel electrophoresis of reconstructed reactions reactions. with undigested or HpaIl-digested templates. Reconstructed reac- In tions contained reverse transcriptase with or without added viral summary, a variety of factors seem to influence the RNA. Reactions indicated by (-) had high-molecular-weight minus- specific initiation of viral plus-strand DNA, but in vitro at strand DNA as a template. For reactions indicated by (+), the least, all the factors are contributed by the template, primer, template was the HpaII restriction fragment of viral DNA which and enzyme. This is in marked contrast to the multicompon- contained the origin of strong-stop plus DNA. Lanes: 1, sequences ent systems known to be required for correct initiation of detected with an M13 probe specific for the LTR minus strand; 2, procaryotic genomes and bacteriophages (8). sequences detected with an M13 probe specific for the LTR plus strand; reaction run without added RNA; 3, sequences detected ACKNOWLEDGMENTS with an M13 probe specific for LTR plus-strand sequences; reaction run with the addition of viral RNA. At the left is cloned viral DNA This work was supported by Public Health Service grants CC- (4) digested with EcoRI as an LTR size marker. 22651, CA-06927-20, and RR-05539 from the National Institutes of Health; grant MV-7 from the American Cancer Society; and by an appropriation from the Commonwealth of Pennsylvania. J.K.S. was not the primer. The primer begins one nucleotide outside the supported by Public Health Service Predoctoral Training Grant GM- 5' end of the PPT and ends three nucleotides within the 3' 07071-08 to the University of Pennsylvania. We thank W. Mason for discussions. end (Fig. 4). Secondary structure of the template may yet be helpful involved in the of with other specificity initiation, by analogy LITERATURE replication origins (8), in particular for the bacterial plasmid CITED Col El, in which both secondary structure and an RNase H 1. Anderson, S., and M. DePamphilis. 1979. Metabolism of Okaza- determine the initiation site (10, 26). Our limited study shows ki fragments during simian virus 40 DNA replication. J. Biol. Chem. that there is no for secondary structure 254:11495-11504. only requirement 2. Baltimore, D., and D. Smoler. 1972. Association of an endori- mediated by the two IR sequences on the template DNA. bonuclease with the avian myeloblastosis virus deoxyribonucle- The RNase H activity of the reverse transcriptase must ic acid polymerase. J. Biol. Chem. 247:7282-7287. also be responsible for the efficient removal of RNA primer. 3. Champoux, J. J., E. Gilboa, and D. Baltimore. 1984. Mechanism In reconstructed reactions, RNase H has been shown to of RNA primer removal by the RNase H activity of avian remove the avian tRNATrP primer for minus-strand DNA myeloblastosis virus reverse transcriptase. J. Virol. 49:686-691. synthesis (18) and to remove RNase A-derived oligoribonu- 4. DeLorbe, W. J., P. A. Luciw, H. M. Goodman, H. E. Varmus, cleotides used as primers for reverse transcription of murine and J. M. Bishop. 1980. Molecular cloning and characterization leukemia virus sequences (3). We observed that after primer of avian sarcoma virus circular DNA molecules. J. Virol. 36:50- 61. removal a 5' remained on the DNA, analogous to phosphate 5. Faras, A. J., J. M. Taylor, W. E. Levinson, H. M. Goodman, the 5' phosphate left by RNase H on its RNA degradation and J. M. Bishop. 1973. RNA-directed DNA polymerase of products (2). Rous sarcoma virus: initiation of synthesis with 70S viral RNA Approximately 60% of the strong-stop plus DNA mole- as template. J. Mol. Biol. 79:163-183. cules completely lacked 5'-terminal ribonucleotides, where- 6. Hohn, T., M. Pietrzak, L. Dixon, I. Koenig, J. Penswick, B. as the remainder contained the full 12-base primer. This all- Hohn, and P. Pfeiffer. 1983. Involvement of reverse transcrip- or-none property of the primer removal is in contrast to what tase in cauliflower mosaic virus replication, p. 28-33. In H. D. has been described for RNA primer removal during the Robertson, S. H. Howell, M. Zaitlin, and R. L. Malmberg (ed.), of Escherichia coli (16) and simian virus 40 (1). Plant infectious agents. Cold Spring Harbor Laboratory, Cold replication N.Y. are 40% of the RNA primer molecules apparently Spring Harbor, Why 7. Hsu, T. W., and J. M. Taylor. 1982. Single-stranded regions on spared from RNase H action under our reaction condition? unintegrated avian retrovirus DNA. J. Virol. 44:47-53. Perhaps, as Champoux et al. have suggested, dissociation of 8. Kornberg, A. 1980. DNA replication, p. 347-413. W. H. Free- the primer-product complex from the template DNA occurs man & Co., San Francisco. so that the RNA primer is no longer a substrate for RNase H 9. Mason, W. S., C. Aldrich, J. Summers, and J. M. Taylor. 1982. (3). Assymetric replication of duck hepatitis B virus DNA in liver We have reconstructed the initiation of plus-strand DNA cells: free minus-strand DNA. Proc. Natl. Acad. Sci. U.S.A. synthesis with an RNase A digest of viral RNA. Three sites 79:3997-4001. of initiation of DNA synthesis were found: one the same as 10. Masukata, H., and J.-I. Tomizawa. 1984. Effects of point mutations on formation and structure of the RNA primer for Col occurs in the endogenous reaction and the other two adja- El DNA replication. Cell 36:513-522. cent and one nucleotide away (Fig. 4e). We also detected 11. Maxam, A. M., and W. Gilbert. 1980. Sequencing end-labeled three alkali-sensitive fragments that were 11 bases longer DNA with base-specific chemical cleavages. Methods Enzymol. than the three bands seen after alkali. An obvious interpreta- 65:499-560. tion was that an 11-base RNA primer was used at each of the 12. McMaster, G. K., and G. G. Carmichael. 1977. Analysis of three initiation sites. The RNase A fragment that would be single- and double-stranded nucleic acids on polyacrylamide VOL. 52, 1984 SPECIFIC INITIATION OF RETROVIRAL PLUS-STRAND DNA 319

and agarose gels by using glyoxal and acridine orange. Proc. acting function. J. Virol. 43:482-488. Natl. Acad. Sci. U.S.A. 74:4835-4838. 23. Southern, E. M. 1975. Detection of specific sequences among 13. Mitra, S. W., M. Chow, J. Champoux, and D. Baltimore. 1982. DNA fragments separated by gel electrophoresis. J. Mol. Biol. Synthesis of murine leukemia virus strong stop DNA initiates at 98:503-517. a unique site. J. Biol. Chem. 257:5983-5986. 24. Tapper, D. P., and D. A. Clayton. 1981. Altered mobility of 14. Mitra, S. W., S. Goff, E. Gilboa, and D. Baltimore. 1979. polydeoxyribonucleotides in high resolution polyacrylamide Synthesis of a 600 nucleotide-long plus-strand DNA by virions gels due to removal of terminal phosphates. Nucleic Acids Res. of Moloney murine leukemia virus. Proc. Natl. Acad. Sci. 9:6787-6794. U.S.A. 76:4355-4359. 25. Taylor, J. M., A. Cywinski, and J. K. Smith. 1983. Discontinui- 15. Moelling, K., D. P. Bolognesi, H. Bauer, W. Busen, H. W. ties in the DNA synthesized by an avian retrovirus. J. Virol. Plassman, and P. Hausen. 1971. Association of viral reverse 48:654-659. transcriptase with an enzyme degrading the RNA moiety of 26. Tomizawa, J.-I., and T. Itoh. 1982. The importance of RNA RNA-DNA hybrids. Nat. New Biol. 234:240-243. secondary structure in Col El primer formation. Cell 31:575- 16. Ogawa, T., and T. Okazaki. 1984. Function of RNAse H in 583. DNA replication revealed by RNAse H defective mutants of 27. Varmus, H., and R. Swanstrom. 1982. Replication of retrovi- Escherichia coli. Mol. Gen. Genet. 193:231-237. ruses, p. 369-512. In R. Weiss, N. Teich, H. Varmus, and J. 17. Olsen, J. C., and K. F. Watson. 1980. Avian retrovirus RNA- Coffin (ed.), RNA tumor viruses, 2nd ed., Cold Spring Harbor directed DNA synthesis by purified reverse transcriptase. Co- Laboratory, Cold Spring Harbor, N.Y. valent linkage of RNA to plus strand DNA. Biochem. Biophys. 28. Varmus, H. E. 1983. Retroviruses, p. 411-503. In J. A. Shapiro Res. Commun. 97:1376-1383. (ed.), Mobile genetic elements. Academic Press, Inc., New 18. Omer, C. A., and A. J. Faras. 1982. Mechanism of release of the York. avian retrovirus tRNATrP primer molecule from viral DNA by 29. Varmus, H. E., S. Heasley, H.-J. Kung, H. Opperman, V. C. ribonuclease H during reverse transcription. Cell 30:797-805. Smith, J. M. Bishop, and P. R. Shank. 1978. Kinetics of synthe- 19. Sabran, J. L., T. W. Hsu, C. Yeater, A. Kaji, W. S. Mason, and sis, structure, and purification of avian sarcoma virus-specific J. M. Taylor. 1979. Analysis of integrated avian RNA tumor DNA made in the cytoplasm of acutely infected cells. J. Mol. virus DNA in transformed chicken, duck, and quail fibroblasts. Biol. 120:55-82. J. Virol. 29:170-178. 30. Verma, I. M. 1975. Studies on reverse transcriptase of RNA 20. Schwartz, D. E., R. Tizard, and W. Gilbert. 1983. Nucleotide tumor viruses. III. Properties of purified Moloney murine sequence of Rous sarcoma virus. Cell 32:853-869. leukemia virus DNA polymerase and associated RNase H. J. 21. Smith, J. K., A. Cywinski, and J. M. Taylor. 1984. Initiation of Virol. 15:843-854. plus-strand DNA synthesis during reverse transcription of an 31. Watson, K. F., P. L. Schendel, M. J. Rosok, and L. R. Ramsey. avian retrovirus genome. J. Virol. 49:200-204. 1979. Model RNA-directed DNA synthesis by avian myeloblas- 22. Sorge, J., and S. H. Hughes. 1982. Polypurine tract adjacent to tosis virus DNA polymerase and its associated RNase H. the U3 region of the Rous sarcoma virus genome provides a cis- Biochemistry 18:3210-3219.