Double-Stranded RNA-Dependent Rnase Activity Associated

Proc. Nadl. Acad. Sci. USA Vol. 89, pp. 927-931, February 1992 Biochemistry Double-stranded RNA-dependent RNase activity associated with human immunodeficiency virus type 1 reverse transcriptase (retrovirus ribonudease) HANA BEN-ARTZI, ELISHA ZEELON, MARIAN GoRECKI, AND AMOS PANET BioTechnology General (Israel) Ltd., Kiryat Weizmann, Rehovot, 76326, Israel Communicated by Ephraim Katchalski-Katzir, October 24, 1991

ABSTRACT Early events in the retroviral replication cycle of the RNA template and, to complete synthesis of the include the conversion of viral genomic RNA into linear double-stranded proviral DNA, the enzyme switches tem- double-stranded DNA. This process is mediated by the reverse plates at the region of the PBS. While the general scheme of transcriptase (RT), a multifunctional enzyme that possesses reverse transcription is known, details of specific steps, such RNA-dependent DNA polymerase, DNA-dependent DNA poly- as template switching, are still missing (5, 8). In an endoge- merase, and RNase H activities. In the course of studies of a nous reaction mixture containing detergent-disrupted viri- recombinant RT of human immunodeficiency virus type 1 ons, synthesis of a full-length proviral DNA with two long (HIV-1), we observed an additional, unexpected activity of the terminal repeats can be detected. It is not clear, however, enzyme. The purified RT catalyzes a specifiC cleavage in HIV-1 whether auxiliary viral proteins other than RT are involved in RNA hybridized to tRNALYS, the primer for HIV-1 reverse the various steps of the reverse transcription reaction. To transcription. The cleavage at the primer binding site (PBS) of study the initiation and synthesis of the proviral DNA, we HIV RNA is dependent on the double-stranded structure ofthe have constructed a system with recombinant RT and HIV-1 HIV RNA-tRNALys complex. This RNase activity appears to be RNA-tRNALYs template-primer complex, which resembles distinct from the RNase H activity of HIV-1 RT, as the the initiation complex utilized by the RT. During these substrate specificity and the products of the two activities are studies, we noticed that HIV-1 RT cleaves specifically at the different. Moreover, Escherichia coli RNase H and avian viral RNA PBS sequence. We propose that this newly myeloblastosis virus RT are unable to cleave the IIV RNA- described activity may be important in the reverse transcrip- tRNALYs complex. We refer to this unusual activity as RNase tion process. D. Two lines of evidence indicate that the specific RNase D activity is an integral part ofrecombinant Hil RT. The specific MATERIALS AND METHODS RNase D activity comigrates with the other RT activities, DNA polymerase, and RNase H upon filtration on a Superose 6 gel Construction of Plasmids. The plasmid pBENN7 (derived column or chromatography on a phosphocellulose column. from HIV-1 lymphadenopathy-associated virus DNA) (11) Moreover, three recombinant HIV-1 RT preparations ex- was used for the isolation of a HindIII/Cla I restriction pressed and purified in different laboratories by various pro- fragment of 2% base pairs (nucleotides 76-371; ref. 11). This cedures exhibit RNase D activity. Sequence analysis indicated DNA fragment includes part of R, U5, PBS, and part of gag. that RNase D activity cleaves the substrate HIV-1 RNA- To restore the entire R region, this fragment was ligated with tRNALYS at two distinct sites within the PBS sequence 5'- a synthetic DNA into the plasmid pSP64 (pHIV). The tran- UGGCGCCCGA | ACAG GGAC-3'. The sequence specific- scription of the HIV-1 gene starts with the first base of the R. I The plasmid pGEM-1 (Promega) was modified and a new ity of RNase D activity suggests that it might be involved in two BspMI site was introduced adjacent to the T7 promoter. Four stages during the reverse transcription process: displacement synthetic oligodeoxynucleotides spanning the sequence of of the PBS to enable copying of tRNALS sequences into tRNALYS and including an additional BspMI site at the 3' end plus-strand DNA or to facilitate the second template switch, of the tRNALYS gene were cloned into BspMI/HindIII- which was postulated to occur at the PBS sequence. cleaved modified pGEM-I. The ptRNALYs cleaved with BspMI endonuclease was used as a template to yield com- The reverse transcriptase (RT) of human immunodeficiency plete tRNALYS molecules 76 bases long. virus (HIV), like the enzyme of other retroviruses, exhibits Preparation of [32PJRNA and Synthetic tRNALYS Complexes. several activities: RNA- and DNA-directed DNA synthesis pHIV DNA was linearized with BssHII endonuclease at (1-3), RNase H (4, 5), and tRNALYS binding (6, 7). These nucleotide 256 (12) and transcribed with SP6 RNA polymer- well-characterized activities participate in various stages of ase (13). [a-32P]CTP (105 cpm/pmol) was used for labeling the double-stranded proviral DNA synthesis. The synthesis of HIV RNA. The reaction mixture was treated with DNase I the first strand, minus-strand DNA, is primed by a tRNALYS and phenol/chloroform extracted, and RNA was ethanol of cellular origin hybridized to the viral genomic RNA at a precipitated and purified by electrophoresis on 7 M urea/6% complementary sequence of 18 nucleotides, the primer bind- polyacrylamide denaturing gel (14). ing site (PBS) (8). Synthesis ofthe plus-strand DNA is primed ptRNALYS was transcribed with T7 RNA polymerase (15). by an oligoribonucleotide of virus origin hybridized to the The reaction mixture was treated with DNase I, and RNA nascent minus-strand DNA, which now serves as a template was extracted with phenol/chloroform followed by ethanol (9, 10). To produce a proviral DNA with two long terminal precipitation and subjected to electrophoresis on a denatur- repeats, the enzyme has to switch templates at two stages of ing gel (14). the reverse transcription process: during synthesis of minus- For complex preparation, purified HIV [32P]RNA and strand DNA the enzyme jumps from the 5' end to the 3' end tRNALYS were hybridized at a 1:10 molar ratio in a solution

The publication costs of this article were defrayed in part by page charge Abbreviations: RT, reverse transcriptase; HIV, human immunode- payment. This article must therefore be hereby marked "advertisement" ficiency virus; PBS, primer binding site; AMV, avian myeloblastosis in accordance with 18 U.S.C. §1734 solely to indicate this fact. virus.

927 Downloaded by guest on October 2, 2021 928 Biochemistry: Ben-Artzi et al. Proc. Nati. Acad Sci. USA 89 (1992) containing 30%o formamide, 40 mM Tris-HCI (pH 7.4), 1 mM of the HIV-1 genome (R-U5-PBS) hybridized to a synthetic EDTA, and 0.3 M NaCi. Hybridization was for 10 min primer tRNALYS (Fig. 1). without salt at 650C and proceeded with salt at 420C for 2 hr. It should be noted that the synthetic tRNALYS is unmodi- The RNA complex was purified on a nondenaturing 8% fied; however, it does not differ from the authentic tRNALYS polyacrylamide gel. Elution ofthe RNA complex from the gel in its ability to bind HIV RT and to initiate minus-strand DNA was done as described (14). A hybrid of 18-mer synthetic synthesis (18). oligodeoxynucleotide, complementary to the PBS region, The recombinant HIV-1 RT was produced in E. coli and and HIV [32P]RNA was prepared under similar conditions for purified by four consecutive column steps to apparent ho- the RNase H reaction. mogeneity. The HIV RT consists of two subunits, 66 and 51 RT Activities. Reaction mixtures (10 pl) contained 50 mM kDa, as reported before (16, 19). The DNA polymerase- Tris-HCl (pH 8.0), 50 mM KCI, 2 mM dithiothreitol, 8 mM specific activity of this enzyme is similar to that of other MgCl2, and the appropriate substrate. For standard DNA purified HIV RT preparations (16, 19). The RT preparation is polymerase reactions, poly(A)/oligo(dT) and [a-32PJdTTP free of nonspecific RNase activities as incubation with HIV were used (16). One unit ofenzyme activity is defined as nmol [32P]RNA single strand does not result in detectable degra- ofdTMP incorporated per 10 min under the standard reaction dation (Fig. 2, lanes 1 and 2). Surprisingly, we noticed that conditions (16). The substrate for RNase H reactions, HIV-1 incubation ofHIV RT with the template RNA annealed to the [32P]RNA-DNA hybrid (2500 cpm), and the substrate for primer tRNALYS resulted in a unique cleavage (lanes 3-5). complex The size of the two RNA products, -200 and -60 nucleo- RNase D reactions, HIV-1 [32P]RNA-tRNALYs tides, suggests that cleavage occurs at the PBS sequence (2500 cpm), were incubated at 37°C. Samples (4 ,ul) were hybridized to tRNALYS and that the degradation products of withdrawn, quenched with sample buffer (80%o formamide/50 200 and 60 bases are derived from the 5' end and the 3' end mM EDTA), and heated at 95°C for 2 min. RNA was of HIV RNA, respectively (see Fig. 1A). subjected to electrophoresis on a 6% polyacrylamide/7 M To investigate the origin of this unique RNase activity, we urea denaturing gel. have obtained purified HIV RT preparations from two groups Recombinant HIV RTs. Details of the expression and (16, 19). These two recombinant RTs were expressed and purification ofHIV RT are described elsewhere (unpublished purified to apparent homogeneity by different procedures. data). Briefly, Escherichia coli containing plasmid pHIV RT The enzyme expressed by Mizrahi et al. (19) was processed that expresses BH10 HIV RT (12) was the source ofenzyme. in the bacterium by the HIV-1 coded protease. Thus, the RT was purified to apparent homogeneity by ammonium purified enzyme is a heterodimer p66/pSl identical to RT sulfate fractionation followed by chromatographies on four purified from virions (3). The purified RT produced by Hizi resins-phenyl-Sepharose, Q-Sepharose, phosphocellulose, et al. (16) is a dimerof66 kDa. The three enzyme preparations and Superose gel. Specific activity of the purified RT (5000 have similar DNA polymerase specific activity, and they units/mg) was determined by a standard RT assay using exhibit identical cleavage activity on the HIV RNA-tRNALYs poly(A)/oligo(dT) as template/primer (16). complex (Fig. 2, lanes 6-9). The sequence of RNA fragments after RNase D reaction It may be argued that the HIV RNA-tRNALYs substrate was determined by DNA primer extension (14) with a 17-mer contains contaminating DNA and that the observed cleavage 32P-labeled 5'-GCCGAGTCCTGCGTCGA-3', complemen- simply represents the RNase H activity of HIV RT. Several tary to the 3' end of the HIV-1 RNA (Fig. 1) and avian measures were taken to ensure removal of all DNA contam- myeloblastosis virus (AMV) RT (IBI). DNA products were inations from the substrate HIV RNA-tRNALYs. After syn- analyzed by electrophoresis on a 6%o polyacrylamide sequencing gel (14). A DNA sequence ladder of HIV was generated with the same 32P 5' primer and dideoxynucle- RT( RT( ) RT(I1) RT0111 otides (14) on DNA plasmid (pHIV). 0 10' 0 5 15 0101 ) 10 RESULTS For the studies of HIV reverse transcription we have pro- - 256 duced in vitro a template RNA that corresponds to the 5' end - ~ 1~~~~~~~~~~~~~~~~~~~~~~~~~~~ 200 Go -j0 z n: B <- A z z tRNAIL.ys cc cc DNA synthesis ,Ct m I 956 a 'l I a - 5.-0 R I U5 I PBS 12 3 4 5 67 89 1OOb 85b 18b [3-P] RNA 256b * 0 - 256 FIG. 2. Cleavage of HIV RNA-tRNALYs complex by HIV RT. 5,1I Recombinant RTs (300 ng per 3 units) were incubated with -OOb [32P]RNA-tRNALYS, and samples were electrophoresed on 6% poly- M urea Lanes: 1 and incubation of 1 3 acrylamide/7 gel. 2, single- 60b 1 2 stranded HIV [32P]RNA (see Fig. 1B, lane 1) with RT I (RT I was purified in our laboratory as described); 3-5, incubation of FIG. 1. Preparation of HIV RNA-tRNALYs template primer [32P]RNA-tRNALYs with RT I; 6 and 7, incubation of [32P]RNA- complex. (A) Scheme of the 5' region of HIV RNA and degradation tRNALYS with RT II [RT II was prepared by Hizi et al. (16)]; 8 and products of the double-stranded RNA-dependent RNase activity. b, 9, incubation of [32P]RNA-tRNALYs with RT III (300 ng per 3 units). Bases. (B) Construction of HIV [32P]RNA-tRNALYs complex. RT III was purified by Mizrahi et al. (19). DNA polymerase activities Purified HIV [32P]RNA (100,000 cpm) (lane 1) was hybridized with (units) were determined on the three RT preparations with poly(A)/ 0.5 ,ug of purified tRNALYS. The complex (lane 2) was subjected to oligo(dT) (16). Sizes of RNA fragments (bases) were estimated electrophoresis in 8% polyacrylamide nondenaturing gel. according to DNA size markers under denaturing conditions. Downloaded by guest on October 2, 2021 Biochemistry: Ben-Artzi et al. Proc. Natl. Acad. Sci. USA 89 (1992) 929

thesis, the RNA preparations were treated with DNase I and A D purified by electrophoresis on denaturing polyacrylamide gels. The two RNAs were hybridized and the complex was '700 l 0.5 600 i ? 0.4 separated from unhybridized nucleic acid by electrophoresis 0 I on nondenaturing polyacrylamide gel (Fig. 1B). To further exclude the possibility of RNase H activity, the substrate gL400 0.3 -T 3- ' .2300200 ;\ coJ HIV RNA/tRNALYS was incubated with AMV RT or with E. 0.2< 2 < coli RNase H. It is evident from the results presented in Fig. .-200 0.1 3A that while HIV RT cleaves the HIV RNA-tRNALYs 0100 complex (lanes 1 and 2), neither AMV RT nor E. coli RNase H possessed the specific RNase activity (lanes 3-6). In a 10 15 20 control experiment, an RNADNA complex was prepared by hybridizing the HIV RNA to a 18-base oligodeoxynucle- RNaseH otides, complementary to the PBS sequence. The three B E RNaseH enzymes, HIV RT, AMV RT, and E. coli RNase H, exhibited extensive RNase H activity on the RNA-DNA complex as expected (Fig. 3B). It is important to note that the smaller 200- 200- *e.*.* RNA product of -60 nucleotides obtained after cleavage of Czar the RNA-tRNALYs substrate (Fig. 3A, lane 2) differs from the degradation product of70 nucleotides obtained by the RNase H activity on the RNA-DNA substrate (Fig. 3B, lane 2). The difference between the 3-end-derived RNA products is a 60_- * 60- u*r* reflection of RT enzyme specificities toward the two substrates. All of these results indicate that the unique RNase l l I l l II I II cleavage of HIV RNA-tRNALYs complex differs from the 10 12 14 20 22 24 well characterized RNase H activity of HIV RT. RNaseD RNaseD Next, we investigated whether the specific RNase activity C is an integral part of the RT heterodimer. This was done by F analyzing the chromatographic profiles of RT protein and the 200 - specific RNase on resins that resolve proteins on the basis of 200 - molecular mass or charge. As shown in Fig. 4, purified RT migrates on a Superose 6 gel filtration column as a native __ protein of -100 kDa. The three activities-DNA polymerase (Fig. 4A), RNase H (Fig. 4B), and RNA-tRNALYs cleavage activity, termed RNase D (Fig. 4C)-comigrate on this column; the most active fractions are fractions 12-14. To 60- 1F 9 further analyze the correlation of the RNase D to the other 60 activities of the enzyme, purified RT was chromatographed ii liii 1 11 11 1 on a phosphocellulose column and eluted by a salt concen- 10 12 14 20 22 24 tration gradient. The activity profiles ofthe DNA polymerase FIG. 4. Cochromatography of the double-stranded RNA- dependent RNase activity (RNase D) with HIV RT on Superose 6 A B (A-C) and phosphocellulose (D-F) columns. (A and D) RT protein profile A280 (--) and DNA polymerase activity (-). (B and E) HIV AMV RNASE HIV AMV RNASE RNase H activity with the hybrid substrate [32P]RNA-18-base DNA. RT RT H RT RT H Numbers on left are bases. (C and F) Double-stranded RNA- dependent RNase (RNase D) activity with the substrate [32P]RNA- 0 10' 0 10'0 10' 0 10' 0 10' 0 10 tRNALYS. Numbers on left are bases. (A-C) Purified RT (10 mg per 0.5 ml) was injected to a Superose 6 gel column (10 x 0.5 cm) (Pharmacia; HR 10/30) equilibrated with 20 mM Tris-HCl, pH 7.8/150 mM NaCl. One-milliliter fractions (fractions 5-21) were h i i s > - ~~~~256 collected. (D-F) Purified RT (10 mg) was subjected to phosphocellulose column chromatography. The column (0.5 x 2 cm) was equilibrated with 10 mM potassium phosphate (pH 6.7) and RT was eluted by a linear gradient ofpotassium phosphate (10-500 mM) (see diagonal line in D); 0.5-ml fractions (fractions 17-30) were collected. Aliquots (1 IAl; 1:100 dilution) were taken for DNA polymerase * * ~70 activity with poly(A)/oligo(dT) (16). Incorporation of [a-32P]dTMP into acid-insoluble material was monitored. Aliquots of 1 1ul (1:10 -60 dilution) were taken for RNase H activity with the RNADNA substrate. Undiluted aliquots of 1 /Al were taken for RNase D activity with [32P]RNA-tRNALYs. 2 3 4 5 6 1 2 3 4 5 6 1 (Fig. 4D), the RNase H (Fig. 4E), and the specific RNase D FIG. 3. Comparison of the double-stranded RNA-dependent (Fig. 4F) are very similar. The highest activities are exhibited RNase and RNase H activities of RT. Reaction conditions and the in fractions 21-24. In repeated experiments, we observed that analysis of RNA degradation products are as described. (A) RNase the three activities coeluted in the same fractions. However, reactions with HIV [32P]RNA-tRNALYS substrate. (B) RNase H while the DNA polymerase reaction is quantitative and reactions with the hybrid HIV[32P]RNA (2500 cpm) annealed to D is 18-base oligodeoxynucleotide complementary to the PBS sequence. linear, the analysis of both RNase H and RNase only Lanes: 1 and 2, HIV RT (3 units); 3 and 4, AMV RT (2 units) (IBI); semiquantitative. Therefore, the apparent peak activities of 5 and 6, E. coli RNase H (2 units) (Promega). Numbers on right are RNase H and RNase D appear to be less distinct as compared bases. with the DNA polymerase. Comigration of RT and RNase D Downloaded by guest on October 2, 2021 930 Biochemistry: Ben-Artzi et al. Proc. Natl. Acad Sci. USA 89 (1992) activity on two resins, which separate proteins by different RNase D activity could be detected (see Fig. 2, lane 1). As principles, further supports the notion that the double- expected, addition of the oligodeoxynucleotide primer and stranded RNA-dependent RNase is yet another function of AMV RT to this reaction mixture resulted in synthesis of the HIV-1 RT molecule. DNA 256 bases long (data not shown). A control experiment carried out in parallel involved the To learn more about the RNase D activity and its corre- assay of proteins expressed in a culture of E. coli harboring lation with other RT activities, DNA polymerase and RNase a mock RT plasmid; i.e., a plasmid lacking a promoter H, an experiment was designed that allowed reverse tran- upstream of the RT coding sequence. The mock proteins scription and subsequent analysis of the RNA template fate. were purified according to the same procedure used for the The 32P-labeled HIV RNA-tRNALYS complex was incubated recombinant RT. The mock fraction contains small amounts with HIV RT in the presence or absence of four deoxyribo- ofbacterial proteins that may copurify with RT. This fraction nucleotide triphosphates and the RNA products were sub- was concentrated 5:1 and an amount of protein (300 ng) jected to electrophoresis on denaturing gels (Fig. 6). Two equivalent to that used in standard RNase D assays was RNA degradation products of -200 bases and 60-64 bases incubated with the InPIRNA-tRNALYs substrate. No RNase were detected in the absence of DNA synthesis (Fig. 6A). D activity was detected in this protein fraction, giving sup- Addition of deoxynucleotide triphosphates resulted in DNA port to the suggestion that this activity is intrinsic to the synthesis and simultaneous degradation ofthe RNA template recombinant RT. Ideally, one would like to show a correla- (Fig. 6B). While the small RNA fragments (60 and 64 bases) tion between RNase D activity and the RT by using enzyme were unchanged in the presence ofDNA synthesis, the larger isolated from HIV virions. However, due to the difficulty in RNA fragment (=200 bases) was diminished. We interpreted propagating quantities ofvirions, the small amount ofenzyme these results as follows: the large fragment derived from the purified from virions is generally contaminated with nonspe- 5' end of HIV RNA served as a template for DNA synthesis cific nucleases. Consequently, the template HIV [32P]RNA and, as expected, was degraded by the RNase H activity of as well as the complex HIV RNA-tRNALYs were nonspecif- RT. In fact, in a similar reaction mixture in which ically degraded by HIV RT isolated from detergent-treated [a-32P]dCTP was included, a radiolabeled DNA product of virions. 200 nucleotides was obtained, as was described by Barat et To determine more precisely the sequences at which al. (18) (data not shown). Since the RNADNA hybrid RNase D cleaves the HIV-1 RNA, standard reaction prod- produced during DNA synthesis spans only 200 bases of the ucts were subjected to sequence analysis (Fig. 5). The RNA HIV RNA, cleavage by RNase H at the 5' end of the PBS, products were annealed to oligodeoxynucleotide comple- should have left an RNA fragment >70 bases (see Fig. 1A). mentary to the 3' end ofthe 256-base HIV RNA (Fig. 1). This Nonetheless, the production of identical RNA molecules 60 DNA primer was elongated with AMV RT and the resulting and 64 bases long in the presence and absence of DNA [32P]DNA was resolved on a sequencing gel, alongside a synthesis indicates that cleavage occurs closer to the 3' end ladder sequence derived from the same region of HIV-1 of the PBS site, as expected of the RNase D activity. DNA. Two main DNA products of 60 and 64 nucleotides are detected (lane 1). These DNAs correspond in size to the small RNA degradation products (=60 bases) observed after the DISCUSSION RNase D reaction. In fact, the small products of RNase D We describe in this work an enzymatic activity associated reactions are usually resolved by gel electrophoresis into two following this unexpected RNA species (see, for example, Fig. 2). We conclude from with the HIV-1 RT. Our objective PBS finding has been to analyze whether this RNase activity is these data that RNase D cleaves twice within the indeed part of the HIV RT molecule. Several independent sequence 5'-UGGCGCCCGA l ACAG I GGAC-3'. In a con- RNase D is trol reaction with single-stranded HIV RNA (256 bases), no lines of evidence support the notion that yet A 1 A C G T B A (-dNTP's) (--dNTP'so 0 10 20 0 10 20 A. _

T ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~,*.J r !i .jd S.

C~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~40.:

110 -

90 - .a. * 76 -- 67 --a

FIG. 5. Determination of the cleavage site in HIV [32P]RNA- tRNALYS complex. Lane 1, after the RNase D reaction, a 17-mer [32P]oligodeoxynucleotide was hybridized to the HIV RNA frag- FIG. 6. Double-stranded RNA-dependent RNase activity during ments for the primer-extension reaction (14). The DNA products DNA synthesis. RT (300 ng per 3 units) was incubated with were resolved on a 6% polyacrylamide sequencing gel. Lanes A, C, [32P]RNA-tRNALYS complex (2000 cpm) in the absence (A) or in the G, and T, products of dideoxynucleotide reaction with the same presence (B) offour deoxyribonucleotide triphosphates (1 mM each). 5'-end-phosphorylated 17-mer primer and a plasmid harboring the Samples were subjected to electrophoresis in a 6% polyacrylamide/7 corresponding HIV sequences. M urea denaturing gel. Numbers on left are bases. Downloaded by guest on October 2, 2021 Biochemistry: Ben-Artzi et al. Proc. Natl. Acad. Sci. USA 89 (1992) 931 another enzymatic activity of HIV RT. (i) The activity is RNA structure at the PBS may be facilitated by the cleavages present in several preparations of highly purified HIV RT introduced at this sequence by RNase D (Fig. 7). Thus, prepared in three laboratories by different expression and RNase D activity may be involved in positioning the two purification procedures. (ii) The RNase D activity cochro- complementary nascent DNA strands, plus and minus, for matographed with RT on two different resins-phosphocel- the second template switch (Fig. 7). While the second tem- lulose and Superose 6 gel. The molecular masses, under plate switch at the PBS site has not been defined yet, it is of native conditions, of the heterodimer RT and of the RNase D interest to note that nicks in the template RNAs induce are similar ('100 kDa). In contrast, the known E. coli template switch and recombination events during proviral nucleases, including RNase III, a double-stranded RNA- DNA synthesis (17). dependent enzyme, have molecular masses in the range of We thank N. Sarver for valuable discussions and for useful 10-50 kDa (20). (iii) The bacterial mock protein fraction comments on the manuscript, R. Lifshitz and B. Amit for help in the purified by the same procedure as RT did not cleave the work, A. Hizi and C. Debouck for the gift of HIV RT, and L. Nir for substrate HIV RNA-tRNALYs. typing the manuscript. Part of this work was supported by U.S. Our experiments indicate that RNase D activity is distinct Public Health Service Contract NO1-A182696 of the National Insti- from the RNase H of RT in terms of substrate specificity and tutes of Health (National Institute of Allergy and Infectious Dis- the sequence at the cleavage site. The two small RNA eases). pBENN7 was obtained through the AIDS Research and products (60 and 64 bases) appear to represent two unique Reference Reagent Program, AIDS Program. endonucleolytic cuts by the RNase D activity as, on a 1. Baltimore, D. (1970) Nature (London) 226, 1209-1211. high-resolution sequencing gel, two large RNA products (200 2. Temin, H. M. & Mizutani, S. (1970) Nature (London) 226, and 196 bases) could also be detected. RNase H activity, on 1211-1213. the other hand, functions as both endo- and exonuclease 3. Di Marzo-Veronese, F., Copeland, T. D., DeVico, A. L., Rahman, R., Oroszlan, S., Gallo, R. C. & Sarngadharan, M. G. (21-23). (1986) Science 231, 1289-1291. The specificity of RNase D activity is indicated by ana- 4. Molling, K., Bolognesi, D. P., Bauer, H., Busen, W., Plass- lyzing different substrates. tRNALYS, which represents in the mann, H. W. & Hausen, P. (1971) Nature (London) 234, cloverleaf structure a double-stranded RNA region (12 base 241-243. pairs of the 18 nucleotides complementary to the PBS), is 5. Goff, S. P. (1990) J. AIDS 3, 817-831. resistant to RNase D cleavage. However, a duplex of 6. Panet, A., Haseltine, W. A., Baltimore, B., Peters, G., Harada, tRNALYS hybridized to antisense to produce a F. & Dahlberg, J. E. (1975) Proc. Natl. Acad. Sci. USA 72, [32P]tRNALYS, 2535-2539. 76-base double-stranded region, is cleaved by RNase D (data 7. Barat, C., Lullien, V., Schatz, O., Keith, G., Nugeyre, M. T., not shown). The substrate for RNase D described here Gnuninger-Leitch, F., Barre-Sinoussi, F., LeGrice, S. F. J. & consists of a synthetic tRNALYS, which differs from the Darlix, J. L. (1989) EMBO J. 8, 3279-3285. authentic tRNALYS (8) in 6 modified bases. Similar cleavage 8. Weiss, R., Teich, N., Varmus, H. & Coffin, J., eds. (1985) RNA activity was, however, also observed when HIV-1 RNA was Tumor Viruses (Cold Spring Harbor Lab., Cold Spring Harbor, complexed to tRNALYS derived from total bovine tRNA (data NY), pp. 369-513. not shown). The RNase D activity appears to be specific, as 9. Luo, G., Sharmeem, L. & Taylor, J. (1990) J. Virol. 64, AMV RT does not cleave the substrate HIV-1 RNA- 592-597. tRNALYS. Specific retrovirus substrates, AMV RNA- 10. Wohre, B. M. & Molling, K. (1990) Biochemistry 29, 10141- 10147. tRNA~rP and murine leukemia virus RNA-tRNAPro, are 11. Gendelman, H. E., Phelps, W., Feigenbaum, L., Ostrove, being prepared now to further analyze the substrate speci- J. M., Adachi, A., Howley, P. M., Khoury, G., Ginsberg, ficity of the RNase D activity. It is interesting to note in this H. S. & Martin, M. A. (1986) Proc. Natl. Acad. Sci. USA 83, respect that binding of tRNALYS and tRNATmr with their 9759-9763. respective HIV RT and AMV RT is species specific (6, 18). 12. Ratner, L., Haseltine, W., Patarca, R., Livak, K. J., Starcich, What is the possible function of RNase D activity in the B., Josephs, S. F., Doran, E. R., Rafalski, J. A., Whitehorn, virus replication cycle? During reverse transcription, after E. A., Baumeister, K., Ivanoff, L., Petteway, S. R., Pearson, the first template switch, the DNA minus-strand is extended M. L., Lautenberger, J. A., Papas, T. S., Ghrayeb, J., Chang, by RT on the genomic RNA template to the 3' end ofthe PBS N. T., Gallo, R. C. & Wong-Staal, F. (1985) Nature (London) 313, 277-284. sequence (Fig. 7). The point at which synthesis stalls and the 13. Melton, D. A., Krieg, P. A., Rebagliati, M. R., Maniatis, T., second template switch occurs has not been precisely deter- Zinn, K. & Green, M. R. (1984) Nucleic Acids Res. 12, 7035- mined (8). We propose that cleavage at the PBS sequence 7056. introduced by RNase D determines cessation ofminus-strand 14. Sambrook, J., Fritsch, E. F. & Maniatis, T., eds. (1989) DNA elongation and perhaps facilitates the template switch. Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Simultaneously with minus-strand DNA elongation, the plus- Lab., Cold Spring Harbor, NY). strand DNA is initiated. This strand is extended by copying 15. Davanloo, P., Rosenberg, A. H., Dunn, J. J. & Studier, W. F. the minus-strand DNA and the first 18 nucleotides of the (1984) Proc. Natl. Acad. Sci. USA 81, 2035-2039. tRNALYS primer (8). Utilization of tRNALYS as a template for 16. Hizi, A., McGill, C. & Hughes, S. (1988) Proc. Natl. Acad. Sci. RT be the PBS USA 85, 1218-1222. should dependent on its displacement from 17. Luo, G. & Taylor, J. (1990) J. Virol. 64, 4321-4328. sequence ofviral RNA. The dissociation oftRNALYS from the 18. Barat, C., Le Grice, S. F. J. & Darlix, J. (1991) Nucleic Acids Res. 19, 751-757. tRNALys 19. Mizrahi, V., Lazarus, G. M., Miles, L. M., Meyers, C. A. & Debouck, C. (1989) Arch. Biochem. Biophys. 273, 347-358. +)Dl\A 5' 5 5 5 b b t p p p | 20. March, P. E. & Gonzalez, M. A. (1990) Nucleic Acids Res. 18, (-)DNA 3' A C C GC G ( C TT G T C CC T G/ (-)DNA 3293-3298. 21. Krug, M. S. & Berger, S. L. (1989) Proc. NatI. Acad. Sci. USA Geneomic +RNA 5' 11 C C G C CCG AtA C A GtGG A C 3' R U5 86,3539-3543. R'\ase[) cleavaue site 22. Oyama, F., Kikucki, R., Crouch, R. J. & Uchida, T. (1989) J. Biol. Chem. 264, 18808-18817. FIG. 7. Proposed model for RNase D activity and second tem- 23. Schatz, O., Mous, J. & Le Grice, S. J. F. (1990) EMBO J. 9, plate switch. 1171-1176. Downloaded by guest on October 2, 2021