Preferred Sequences within a Defined Cleavage Window Specify DNA 3′ End-directed Cleavages by Retroviral RNases H*

The RNase H activity of reverse transcriptase carries out three types of cleavage termed internal, RNA 5′ end-directed, and DNA 3′ end-directed. Given the strong association between the polymerase domain of reverse transcriptase and a DNA 3′ primer terminus, we asked whether the distance from the primer terminus is paramount for positioning DNA 3′ end-directed cleavages or whether preferred sequences and/or a cleavage window are important as they are for RNA 5′ end-directed cleavages. Using the reverse transcriptases of human immunodeficiency virus, type 1 (HIV-1) and Moloney murine leukemia virus (M-MuLV), we determined the effects of sequence, distance, and substrate end structure on DNA 3′ end-directed cleavages. Utilizing sequence-matched substrates, our analyses showed that DNA 3′ end-directed cleavages share the same sequence preferences as RNA 5′ end-directed cleavages, but the sites must fall in a narrow window between the 15th and 20th nucleotides from the recessed end for HIV-1 reverse transcriptase and between the 17th and 20th nucleotides for M-MuLV. Substrates with an RNA 5′ end recessed by 1 (HIV-1) or 2–3 (M-MuLV) bases on a longer DNA could accommodate both types of end-directed cleavage, but further recession of the RNA 5′ end excluded DNA 3′ end-directed cleavages. For HIV-1 RNase H, the inclusion of the cognate dNTP enhanced DNA 3′ end-directed cleavages at the 17th and 18th nucleotides. These data demonstrate that all three modes of retroviral RNase H cleavage share sequence determinants that may be useful in designing assays to identify inhibitors of retroviral RNases H.

During reverse transcription, a retrovirus produces a doublestranded terminally redundant DNA from a single-stranded plus-sense RNA genome (for reviews, see Refs. 1 and 2). Minusstrand DNA synthesis is initiated with a host cell-derived tRNA, whereas plus-strand DNA synthesis is initiated with a primer generated from a polypurine tract (PPT) 2 in the viral RNA genome. This replication process is carried out by a virally encoded protein termed reverse transcriptase that contains two enzymatic activities. The amino-terminal two-thirds of reverse transcriptase has a DNA polymerase activity that utilizes RNA or DNA as a template, whereas the carboxyl-terminal one-third has an RNase H activity that degrades the RNA strand of RNA/ DNA hybrids. Both activities are required for viral replication (3)(4)(5)(6).
RNase H has several roles in reverse transcription (for reviews, see Refs. 7 and 8). First, RNase H extensively degrades the RNA genome, which assists plus-strand synthesis, strand transfers, and recombination. Second, RNase H specifically cleaves the viral genome to generate the PPT primer required for plusstrand synthesis. Third, RNase H removes the tRNA and PPT primers after minus-strand and plus-strand DNAs are initiated. Because of these multiple functions, RNase H is considered a potential target of antivirals in the treatment of patients infected with human immunodeficiency virus, type 1 (HIV-1) (for a review, see Ref. 9).
Both the heterodimeric reverse transcriptase of HIV-1 and the monomeric reverse transcriptase of Moloney murine leukemia virus (M-MuLV) represent excellent model systems to study the enzymatic mechanism and properties of retroviral RNase H. Crystallography studies have shown that the DNA polymerase domains of the human and murine enzymes have similar nucleic acid binding clefts for the double-stranded primer-template and that their RNase H domains share very comparable tertiary folds (10 -14). In addition, co-crystal structures have shown that the 3Ј end of a DNA primer makes multiple contacts with the polymerase domain of HIV-1 reverse transcriptase and that the active site of the RNase H domain is 17 or 18 nucleotides away, depending upon whether the substrate is a DNA/DNA or an RNA/DNA duplex (12,13,15). However, the human and murine enzymes each display distinctive structural features that might uniquely influence substrate interactions and consequently RNase H activity, such as a longer connection domain in the M-MuLV reverse transcriptase or the absence of the C-helix in the HIV-1 RNase H domain.
Depending upon how reverse transcriptase associates with an RNA/DNA hybrid, the RNase H activity carries out three distinct types of cleavage: internal, RNA 5Ј end-directed, and DNA 3Ј end-directed (for reviews, see Refs. 8 and 16). Internal cleavage can occur when reverse transcriptase binds a hybrid without associating with the end of a recessed strand (17)(18)(19)(20)(21). Instead of occurring at random sites, our recent studies have shown that sequence features both upstream and downstream of a cleavage site represent important determinants for the positioning of internal cleavages by the RNases H of HIV-1 and M-MuLV (22). However, the polymerase domain of reverse transcriptase preferentially associates with a recessed strand end in an RNA/DNA hybrid, and such interactions direct more extensive RNase H cleavages that are termed RNA 5Ј end-di-rected or DNA 3Ј end-directed, depending upon whether the recessed end is RNA or DNA, respectively (23)(24)(25)(26).
Many studies have examined the RNA 5Ј end-directed or the DNA 3Ј end-directed mode of cleavage separately. RNA 5Ј enddirected cleavages have previously been observed to occur as close as 7 nucleotides and as far as 21 nucleotides from the recessed RNA 5Ј end (for example, see Refs. 23 and 26 -31).
Recently, we showed that, like internal cleavages, a site recognized by RNA 5Ј end-directed cleavage is characterized by preferred nucleotides flanking the cleavage site, and in addition, cleavage preferentially occurs within a "cleavage window" that ranges between 13 and 19 nucleotides from the RNA 5Ј end (31). A possible explanation for the existence of this cleavage window is that the enzyme associates only loosely with the recessed RNA 5Ј end and then prior to cleavage slides in either direction to a site corresponding to the nucleotide sequence preference of the retroviral RNase H.
Unlike RNA 5Ј end-directed cleavages where the structural basis for the association of the enzyme with the recessed end is unknown, the polymerase domain of reverse transcriptase is expected to strongly engage a recessed DNA 3Ј end for potential use as a primer in DNA synthesis (12,13,15). This consideration suggested the possibility that DNA 3Ј end-directed cleavages by RNase H might occur at a fixed distance from the recessed end and without any preference for nucleotide sequence. The alternative hypothesis is that DNA 3Ј end-directed cleavages occur at specific sites within a window similar to what we have described previously for RNA 5Ј end-directed cleavage. In other reports where DNA 3Ј end-directed cleavage has been studied, cleavage sites have been reported in the range from 15 to 20 nucleotides from the recessed DNA 3Ј terminus (for example, see Refs. 27 and 32-34), but no attempt was made to systematically determine whether the cleavage sites must fall within a cleavage window or simply reflected differences in the structure of the RNA/DNA hybrid substrate. To distinguish between these possibilities and to further explore the basis for the interaction of reverse transcriptase with substrates containing different types of recessed ends, in this study we directly compared the cleavage sites preferred by the DNA 3Ј end-directed versus RNA 5Ј end-directed types of RNase H cleavage on matched hybrid substrates with the same nucleotide sequence. The results indicate that, similar to RNA 5Ј enddirected cleavage, DNA 3Ј end-directed cleavage occurs at specific sites in the substrate, but the allowable window for cleavage is distinct from that described for RNA 5Ј end-directed cleavage.
Interestingly, several non-nucleoside reverse transcriptase inhibitors that inhibit the polymerase activity of HIV-1 reverse transcriptase have been reported to increase DNA 3Ј end-directed cleavages while partially suppressing RNA 5Ј end-directed cleavages (35). These results indicate that antiviral drugs can differentially influence what type of RNase H cleavage occurs. Consequently, identifying and characterizing all of the determinants that direct the internal, DNA 3Ј end-directed, and RNA 5Ј end-directed types of RNase H cleavage are essential both for a general understanding of retroviral replication and for developing assays to identify new inhibitors of HIV-1 reverse transcriptase.

EXPERIMENTAL PROCEDURES
Reagents and Enzymes-Recombinant M-MuLV reverse transcriptase and ddNTPs were purchased from Amersham Biosciences. Recombinant HIV-1 reverse transcriptase was obtained from Worthington Biochemicals, and dNTPs were from Invitrogen. T7 DNA polymerase and calf intestinal alkaline phosphatase were obtained from USB/Affymetrix, whereas all other DNA-modifying enzymes and restriction enzymes were purchased from New England Biolabs. DNA oligonucleotides were obtained from Eurofins MWG Operon and Bioneer, Inc. In vitro transcripts were prepared using the MEGAshortscript TM kit from Ambion.
RNA and DNA Strands Used in Hybrid Substrates-Preparation of RNAs Md1, Md4, Md7, Md9, and R46 was described previously (22,31). The 5Ј end of 46-mer R46 begins with 8 nucleotides of non-viral sequence derived from the template used for in vitro transcription (numbered T1 to T8 from 5Ј to 3Ј; see Fig. 1) followed by the first 38 nucleotides of sequence (numbered ϩ1 to ϩ38) downstream of the cleavage site in the M-MuLV genome that creates the PPT primer (defined as the ϩ1/Ϫ1 site). To prepare RϪ63/Ϫ23, a 49-mer DNA oligonucleotide (5Ј-CGCCTATAGAGTACGAGCCATAGATAA-AATAAAAGATTTTATTGAGACG-3Ј) was annealed to a 57-mer DNA oligonucleotide (5Ј-AATTCGTCTCAATAAA-ATCTTTTATTTTATCTATGGCTCGTACTCTATAGGC-GAGCT-3Ј) to form a duplex that was placed into EcoRI-and SacI-linearized pGEM9Zf(Ϫ) (Promega), and the resulting plasmid was linearized with BsmBI and transcribed in vitro as described previously (20). The resulting 49-mer RNA, RϪ63/ Ϫ23, begins with the same 8 nucleotides as R46 (T1 to T8 from 5Ј to 3Ј; see Fig. 1) followed by 41 nucleotides of sequence (numbered as Ϫ63 to Ϫ23 from 5Ј to 3Ј) upstream of the M-MuLV PPT primer cleavage site.
The sequences of DNAs Cd1, Cd4, Cd7, and Cd9 are complementary to RNAs Md1, Md4, Md7, and Md9, respectively, and are shown in Fig. 1. DNA D46 is identical in length and complementary to RNA R46. RNAs and oligonucleotides were gel-purified in denaturing polyacrylamide gels as described previously (20,22).
Preparation of Hybrid Substrates-RNAs were labeled at the 5Ј end or the 3Ј end (described previously (20,22)) as specified in the figure legends and annealed to the indicated DNAs at an RNA:DNA molar ratio of 1:2 in 10 mM Tris-HCl (pH 8.0) and 200 mM KCl at 90°C for 3 min followed by cooling to room temperature.
For the hybrid substrates used to analyze the transition between RNA 5Ј end-directed cleavage and DNA 3Ј enddirected cleavage, the substrate with a recessed DNA 3Ј end was generated by annealing Cd9 to R46. The blunt end substrate contained RNA Md9 annealed to a DNA strand with the sequence 5Ј-GCTAGCTTGCCAAACCTACAGGTGG-GGT-3Ј such that the DNA 3Ј end was co-terminal with the 5Ј end of Md9 and the DNA 5Ј end was recessed by 1 base from the 3Ј end of Md9. Substrates in which the RNA 5Ј end of Md9 was progressively recessed were generated by annealing Md9 to DNAs containing the preceding underlined sequence and additional bases at the DNA 3Ј end as follows: recessed 1 base, 5Ј-…C-3Ј; recessed 2 bases, 5Ј- …CT-3Ј; recessed 3 bases,  5Ј-…CTT-3Ј; recessed 8 bases, 5Ј-…CTTTCATT-3Ј; recessed  11 bases, 5Ј-…CTTTCATTGAG-3Ј; and recessed 38 bases,  5Ј-…CTTTCATTGAGCTCCCCCCCCCTTTTTCTGGAG-ACTAA-3Ј. The size of the cleavage window for DNA 3Ј end-directed cleavages was determined with hybrids containing R46 or RϪ63/Ϫ23 (Fig. 1). For hybrid substrates containing R46, the complementary DNAs had 5Ј ends beginning at position ϩ36 and 3Ј ends located from position T6 to position ϩ13 in R46. For hybrid substrates containing RϪ63/Ϫ23, the complementary 30-mer DNAs had 5Ј ends that were recessed by at least 2 bases from the 3Ј end of RϪ63/Ϫ23 and 3Ј ends that were positioned at 13 successive nucleotides, beginning at position Ϫ62 and ending at position Ϫ50 in RϪ63/Ϫ23.
To analyze the effects of dNTPs on RNase H cleavage, hybrid substrates were generated by annealing 31-mer DNA strands with 3Ј ends located 15-20 nucleotides upstream of the site Ϫ41/Ϫ40 to RϪ63/Ϫ23 (Fig. 1B). To prepare these DNA strands, the complementary 30-mer DNAs with 3Ј ends at every position from Ϫ59 to Ϫ54 in RϪ63/Ϫ23 (described above) were annealed to a DNA version of RϪ63/Ϫ23 and then extended by 1 nucleotide with the appropriate ddNTP and T7 DNA polymerase. Reactions were carried out in 40 mM Tris-HCl (pH 7.5), 20 mM MgCl 2 , 50 mM NaCl, 0.1 mM dithiothreitol, and 300 M ddNTP for 60 min at 37°C followed by gel purification of the dideoxy-extended DNA. The resulting primers were named Primers 15,16,17,18,19, and 20, referring to the distance between the 3Ј end of each primer and the nucleotide 5Ј of site Ϫ41/Ϫ40.
Cleavage Analysis of Hybrid Substrates-Typically, 20-l reactions were preincubated with 10 nM hybrid substrate in 50 mM Tris-HCl (pH 8.0), 50 mM KCl, 6 mM MgCl 2 , and 5 mM dithiothreitol at 37°C for 3 min, and then 0.2 pmol of HIV-1 or ϳ0.6 -0.8 pmol of M-MuLV reverse transcriptase was added to start the reaction (31). At the indicated times, aliquots were added to formamide stop mixture (95% (v/v) formamide and 20 mM EDTA) and analyzed in denaturing 20% polyacrylamide gels that were subsequently dried. Products were visualized by PhosphorImager analysis and analyzed using ImageQuant software (GE Healthcare).
In the experiments comparing the effects of cognate versus non-cognate dNTPs on RNase H cleavage, reactions additionally contained a 50, 200, or 1000 M concentration of a noncognate dNTP or the cognate dNTP as determined by the template strand sequence. For Primers 15,16,17,18,19, and 20, the cognate dNTPs used were dCTP, dTTP, dATP, dTTP, dATP, and dGTP, respectively, and the non-cognate dNTPs used were dGTP, dATP, dTTP, dATP, dTTP, and dCTP, respectively. Products were analyzed in denaturing 15% polyacrylamide gels and visualized as described above.

RESULTS
Comparison of DNA 3Ј End-and RNA 5Ј End-directed Cleavages Using Matched Hybrid Substrates-Earlier studies with HIV-1 and M-MuLV reverse transcriptases have shown that the sequence containing the PPT region of M-MuLV has several RNase H cleavage sites (previously named A-I) that are located downstream of the specific cleavage site generating the 3Ј end of the PPT primer (defined as occurring between positions Ϫ1 and ϩ1 in the M-MuLV sequence) (18,20). Depending upon substrate structure, the sites that are recognized by each enzyme can be cleaved in either the RNA 5Ј end-directed or internal mode of RNase H cleavage. Recognition of a particular site sometimes differs between the two enzymes and is explained by the unique nucleotide preferences of the HIV-1 or M-MuLV reverse transcriptase for the sequence surrounding a given RNase H cleavage site (22). The sequence downstream of the M-MuLV PPT containing these sites is found in RNA R46, and the positions of the relevant sites E-I are indicated in Fig. 1A.
Using the HIV-1 and M-MuLV reverse transcriptases, we examined how these sites were recognized by the DNA 3Ј enddirected type of RNase H cleavage. Hybrids that contained recessed DNA 3Ј ends at varied locations relative to sites E-I were generated by annealing DNAs Cd1, Cd4, Cd7, and Cd9 to RNA R46 (Fig. 1A). To directly compare the relative positions of DNA 3Ј end-directed cleavages with RNA 5Ј end-directed cleavages, matching substrates containing recessed RNA 5Ј ends with identical hybrid sequences were generated as well. As an example in Fig. 1A, the matched hybrid substrates with end positions designated as Pair 1 consist of one substrate containing Cd1 annealed to R46 (recessed DNA 3Ј end) and one substrate containing Md1 annealed to the DNA equivalent of R46 (recessed RNA 5Ј end; not shown). A total of four pairs of matched hybrid substrates containing the sequences for Cd1, Cd4, Cd7, and Cd9 and their RNA counterparts were generated ( Fig. 2A). Importantly, these hybrids all contained cleavage sites E-I, but these sites were found at different distances from the recessed ends in each hybrid pair.

Characterizing End-directed Cleavages by Retroviral RNases H
The pairs of matched hybrids containing 5Ј end-labeled RNAs were treated with HIV-1 reverse transcriptase in time course assays, and the resulting products generated by RNase H cleavage at the same sites in each pair are indicated by dashed lines in Fig. 2, B-E. To consider the cleavage sites initially recognized by RNA 5Ј end-or DNA 3Ј end-directed cleavage, only the major products generated in the 0.25-and 1-min time points were evaluated and are indicated by filled circles at the left of each panel. In substrate Pair 1, sites E, F, and G are found 13, 16, and 19 nucleotides, respectively, from the recessed ends in both hybrids (Fig. 2B, lanes 1-10). In the substrate with a recessed RNA 5Ј end, HIV-1 RNase H primarily cleaved at sites from E to F and at site G (Fig. 2B, lanes 2 and 3). In contrast, HIV-1 RNase H cleaved the equivalent substrate with a recessed DNA 3Ј end primarily at sites F and G but not at sites closer to the 5Ј end, including site E (Fig. 2B, lanes 7 and 8). In Pair 4, sites F, G, and H are found 13, 16, and 19 nucleotides from the recessed ends, respectively (Fig. 2C, lanes 1-10).
When the substrate has a recessed RNA 5Ј end, site F was cleaved, but when the substrate has a recessed DNA 3Ј end, sites G and H were cleaved efficiently (Fig. 2C, compare lanes 2 and 3 and lanes 7 and 8). In Pair 7, sites F, G, and H are 10, 13, and 16 nucleotides, respectively, from the recessed hybrid ends, and RNA 5Ј end-directed cleavages occurred at each of these sites, but DNA 3Ј end-directed cleavage strongly favored site H (Fig.  2D, lanes 1-10). In a last example, the substrates of Pair 9 have sites H and I located 14 and 19 nucleotides from the recessed end ( Fig. 2E, lanes 1-10), and HIV-1 RNase H cleaved at site H when the recessed end was RNA and at site I when the recessed end was DNA (Fig. 2E, compare lanes 2 and 3 and lanes 7 and 8).
These data reveal that HIV-1 RNase H did not cleave at the same distance from a recessed DNA 3Ј end in the hybrid substrates as would be predicted if rigid positioning of the 3Ј end in the polymerase active site strictly determined the cleavage specificity. Instead, HIV-1 RNase H recognized the same sites by DNA 3Ј end-directed and RNA 5Ј end-directed cleavages, but site selection was different in each pair of substrates. For example, the most extensive cleavage at site H occurred by DNA 3Ј end-directed cleavage in Pair 7 and by RNA 5Ј enddirected cleavage in Pair 9 ( Fig. 2, D, lane 7 and E, lane 2, respectively). Also, a very similar cleavage pattern was observed for Pair 4 substrate with a DNA 3Ј end and Pair 7 substrate with an RNA 5Ј end (Fig. 2, compare C, lane 7 with D, lane 2, respectively). These observations indicated that DNA 3Ј end-directed cleavage occurs only at sites also recognized by RNA 5Ј enddirected cleavage and that the type of recessed end determines the locations and extent of cleavage at each site.
These matched pairs of hybrid substrates were also used in RNase H cleavage assays with M-MuLV reverse transcriptase (Fig. 3). In Pair 1, the substrate with a recessed RNA 5Ј end was primarily cleaved at sites E and F, whereas the substrate with a recessed DNA 3Ј end was primarily cleaved at site G (Fig. 3A, lanes 1-10). With Pair 4, the substrate with a recessed RNA 5Ј end was mostly cleaved at site F and the site just 3Ј of F, whereas a substrate with a recessed DNA 3Ј end was primarily cleaved at site H (Fig. 3B, lanes 1-10). In Pair 7, a more subtle difference in cleavage was observed with RNA 5Ј end-directed cleavages occurring at sites H and HЈ and DNA 3Ј end-directed cleavages occurring at site HЈ and beyond (Fig. 3C, lanes 1-10). Finally, using Pair 9 substrates, cleavage primarily occurred at site HЈ with a recessed RNA 5Ј end and at site I with a recessed DNA 3Ј end (Fig. 3D, lanes 1-10). Thus, similar to the observations for HIV-1 reverse transcriptase, DNA 3Ј end-directed cleavages and RNA 5Ј end-directed cleavages occurred at different locations within a given pair of matched hybrid substrates.
Together, these data showed that DNA 3Ј end-directed cleavages are not positioned strictly by distance and that the M-MuLV and HIV-1 reverse transcriptases can recognize the same sites by the DNA 3Ј end-directed or RNA 5Ј end-directed type of RNase H cleavage, depending upon the substrate structure and sequence. In a matched hybrid, there is a clear tendency for more distal sites to be recognized by DNA 3Ј enddirected cleavages and more proximal sites to be preferred by RNA 5Ј end-directed cleavages.
Nature of Recessed End Responsible for RNA 5Ј Versus DNA 3Ј End-directed Cleavages-Because matched substrates containing RNA 5Ј or DNA 3Ј ends at the same positions demonstrated distinct differences in which RNase H cleavage sites are recognized, we evaluated how the structure of a recessed end might promote an RNA 5Ј end-directed or DNA 3Ј end-directed type of cleavage. To distinguish the two types of end-directed RNase H cleavages, substrates containing the hybrid sequence of Pair 9 were used for RNA 5Ј end-directed cleavages at site H for HIV-1 RNase H or site HЈ for M-MuLV RNase H and for DNA 3Ј end-directed cleavage at site I for both enzymes (Figs. 2E and 3D).
For these experiments, we used hybrid substrates with an RNA 5Ј overhang (DNA 3Ј end-recessed), with a blunt end, and with a recessed RNA 5Ј end. A hybrid substrate with an RNA 5Ј overhang was used in the experiments described above ( Fig. 2A, DNA 3Ј end-recessed). To generate substrates with a blunt end or an RNA 5Ј end recessed at different positions from the DNA 3Ј end, Md9 was annealed to complementary DNAs that increased successively in length at their 3Ј ends (Fig. 4A, blunt and recessed). Initial assays with HIV-1 reverse transcriptase comparing cleavage of the RNA 5Ј overhang substrate containing R46 with the blunt end substrate containing Md9 showed that site I was the primary cleavage product (compare Fig. 2E, lane 8 with Fig. 4B, lane 2). Consequently, only data using the blunt end substrate containing Md9 are shown in Fig. 4. These data indicated that blunt ends promoted DNA 3Ј end-directed cleavages and that a recessed RNA 5Ј end was required to direct cleavage at site H.
The cleavage pattern of HIV-1 reverse transcriptase was analyzed using hybrid substrates containing a blunt end or recessed RNA 5Ј ends of 1, 2, 3, 8, 11, or 38 bases. When the RNA 5Ј end was recessed by 1 base, equal amounts of cleavage occurred at sites I and H (Fig. 4B, lane 3 and graph). As the RNA 5Ј end was progressively recessed, cleavage at site H increased, and cleavage at site I decreased (Fig. 4B, lanes 4 -8 and graph).
The transition between RNA 5Ј end-directed and DNA 3Ј end-directed types of cleavage was also analyzed for M-MuLV reverse transcriptase using the same substrates and monitoring cleavage at sites HЈ and I for RNA 5Ј end-directed and DNA 3Ј end-directed cleavages, respectively (Fig. 4C). As was observed for HIV-1 reverse transcriptase, M-MuLV reverse transcriptase cleaved at site I with the substrate containing a recessed DNA 3Ј end (RNA 5Ј overhang) and with the substrate containing a blunt end (Fig. 3D, lane 8 and Fig. 4C, lane 2, respectively). Similarly, a substrate with an RNA 5Ј end recessed by 1 base was also cleaved predominately at site I (Fig.  4C, lane 3 and graph). Sites HЈ and I were equally cleaved in substrates containing an RNA 5Ј end recessed by 2 or 3 bases (Fig. 4C, lanes 4 and 5  and graph). When the substrates contained RNA ends recessed by 8 bases or more, cleavages were primarily at site HЈ (Fig. 4C, lanes 6 -8  and graph).
In other experiments with HIV-1 and M-MuLV reverse transcriptases, we compared cleavage of substrates with RNA 5Ј ends recessed between 1 and 6 bases to find that RNA 5Ј ends recessed by 1-3 bases had similar levels of RNA 5Ј endand DNA 3Ј end-directed cleavage and that ends recessed by 4, 5, or 6 bases predominantly showed RNA 5Ј end-directed cleavages (data not shown). Also, similar results were obtained using substrates with hybrids containing Md1 ( Fig. 1A; data not shown). Together, these observations indicate that the shift in generating DNA 3Ј end-directed versus RNA 5Ј end-directed cleavage was determined by the distance that the RNA 5Ј end was recessed. Substrates that appeared to allow both DNA 3Ј end-directed and RNA 5Ј enddirected types of RNase H cleavage had a 1-base RNA 5Ј end recession for HIV-1 reverse transcriptase and a 2-3-base recession for M-MuLV reverse transcriptase.
Size of DNA 3Ј End-directed Cleavage Window-For recognition by RNA 5Ј end-directed cleavage, we have previously shown that, for both M-MuLV and HIV-1 reverse transcriptases, a site in an RNA/DNA hybrid must be located within a cleavage window of 13-19 nucleotides from the recessed RNA 5Ј end (31). Using our pairs of matched hybrid substrates, we observed that DNA 3Ј end-directed cleavages similarly occurred at sites with preferred sequences but were generally found further from the recessed end than RNA 5Ј end-directed cleavages (Figs. 2 and 3). Thus, we wished to establish whether these two distinct types of retroviral RNase H cleavage might have unique cleavage windows.
To define the optimal distance for DNA 3Ј end-directed cleavages using two different sequences, RNase H assays were designed to monitor cleavage at sites F, G, H, and I in R46 hybrids where the DNA 3Ј ends ranged in increments of 1 nucleotide from positions T6 to ϩ13 over a total distance of 16 nucleotides (Fig.  5A). As a second sequence, DNA 3Ј ends were similarly varied over a total distance of 13 nucleotides relative to site Ϫ41/Ϫ40 in RϪ63/Ϫ23 hybrids (position Ϫ62 to position Ϫ50; Fig. 5B).
When HIV-1 reverse transcriptase was analyzed using hybrid substrates containing R46, it was evident that cleavage at sites F, G, H, and I depended upon the location of the DNA 3Ј end relative to each site (Fig. 5C, lanes 1-17). Most cleavage was observed at site F in substrates T6 to ϩ2 (lanes 2-6), at site G in substrates ϩ1 to ϩ5 (lanes [5][6][7][8][9], and at site H in substrates ϩ4 to ϩ8 (lanes 8 -12). Site I only showed strong cleavage in substrates ϩ9 and ϩ10 (lanes 13-14) where the DNA 3Ј ends were located 18 or 19 nucleotides away, respectively. When no cleavage site was present within the opti-mal window for cleavage, no particular cleavage sites were favored (for example, lane 15), or cleavage shifted back to a closer site (for example, lane 16, site H). Notably, for a very strong cleavage site such as F, significantly lower but detectable levels of cleavage occurred outside of the optimal window, most likely as weaker secondary cleavages (Fig. 5C, lanes 7-13).
To determine the size of the DNA 3Ј end-directed cleavage window with R46 hybrids, the amount of cleavage products generated by cleavage at sites F-I in each substrate was plotted as a function of distance from the DNA 3Ј end (Fig.  5E). Although there was some variation in the amount of cleavage products produced, in all cases optimal cleavage fell between 15 and 20 nucleotides from the DNA 3Ј end. This distance was slightly narrower and slightly further from the recessed end than the RNA 5Ј enddirected cleavage window of 13-19 nucleotides determined previously for HIV-1 reverse transcriptase (31). Notably, the amount of products resulting from cleavage at site I was significantly less than those observed for the other sites when site I was located 15-17 nucleotides from the recessed DNA 3Ј end (Fig. 5, C and E). This likely indicated that site I is not recognized as strongly as sites F, G, and H and explains the smaller cleavage window surrounding site I.
To measure the cleavage window using an alternative site and sequence, similar analyses were carried out using the site Ϫ41/Ϫ40 in a series of hybrids containing RϪ63/Ϫ23 (Fig. 5B). Site Ϫ41/ Ϫ40 is an exceptionally strong and relatively isolated cleavage site as the closest flanking cleavage sites are at least 3 bases upstream or downstream. In time course assays using hybrids containing RϪ63/Ϫ23 and HIV-1 reverse transcriptase, optimal cleavage at site Ϫ41/Ϫ40 occurred in substrates Ϫ60 to Ϫ55 (Fig. 5D, lanes  3-8). When present at an optimal distance in the cleavage window, products resulting from cleavage at sites 3-5 bases upstream of site Ϫ41/Ϫ40 were also observed (Fig. 5D, lanes  1-4), but very little cleavage was observed in the sequence found 3Ј of the Ϫ41/Ϫ40 site even when site Ϫ41/Ϫ40 cleavage decreased substantially and was not in the cleavage window (Fig. 5D, lanes 9 -13). The amount of product generated by cleavage at site Ϫ41/Ϫ40 in each substrate was plotted as a function of the distance from the DNA 3Ј end (Fig. 5F). As  NOVEMBER 20, 2009 • VOLUME 284 • NUMBER 47 JOURNAL OF BIOLOGICAL CHEMISTRY 32231 described above for sites F-I in the R46 hybrids, the window for significant amounts of cleavage at site Ϫ41/Ϫ40 was located between 15 and 20 nucleotides from the DNA 3Ј end.

Characterizing End-directed Cleavages by Retroviral RNases H
The R46 and RϪ63/Ϫ23 hybrid series (shown in Fig. 5, A and  B) were also used to determine the DNA 3Ј end-directed cleav-age window of M-MuLV reverse transcriptase (Fig. 6). As seen for HIV-1 RNase H, optimal cleavage of sites F, G, H, and I occurred in different substrates and depended upon the distance of the DNA 3Ј end from each cleavage site. Site F was cleaved in substrates T6 through T8 (Fig. 6A, lanes 2-4); site G To assay cleavage at sites F, G, H, and I, 5Ј end-labeled (indicated by an asterisk) R46 (gray) was annealed to 16 different complementary DNA strands that shared the same DNA 5Ј end but successively had 1 nucleotide removed from the DNA 3Ј end. The span of the 3Ј end positions is indicated from T6 to ϩ13 on R46 with dashed lines, but only two DNA strands (black), with 3Ј ends at T6 and ϩ13, are shown. Note that the relative positions of sites F, G, H, and I change according to the 3Ј end positions of the 16 different hybrids (indicated by dashed lines). B, schematic of hybrid substrates containing RϪ63/Ϫ23. To assay cleavage at site Ϫ41/Ϫ40, 5Ј end-labeled RϪ63/Ϫ23 (gray) was annealed to 13 different complementary DNA strands of mostly 30 nucleotides in length that successively changed their DNA 3Ј and 5Ј ends by 1 nucleotide. The span of the 3Ј end positions is indicated from Ϫ62 to Ϫ50 on RϪ63/Ϫ23 with dashed lines, but only two DNA strands (black), with 3Ј ends at Ϫ62 and Ϫ50, are shown. DNA strands Ϫ52, Ϫ51, and Ϫ50 are slightly less than 30 nucleotides in length and share the same 5Ј end as strand Ϫ53, which is recessed by 2 nucleotides from the 3Ј end of RϪ63/Ϫ23. C and D, HIV-1 reverse transcriptase was incubated with hybrid substrates containing R46 (C) or RϪ63/Ϫ23 (D), and 1-min samples from each reaction were analyzed as described in Fig. 2. In C, the positions of R46 and products resulting from cleavage at sites F, G, H, and I are indicated (lanes 1-17). In D, the positions of RϪ63/Ϫ23 and only the products resulting from cleavage at site Ϫ41/Ϫ40 are indicated (lanes 1-14). E and F, the amount of product generated by cleavage (percentage of total) at sites F, G, H, and I or at site Ϫ41/Ϫ40 is plotted as a function of distance of the site in nucleotides (nt) from the DNA 3Ј end of each substrate with the distance measured between the 3Ј end (nucleotide 1) and the 5Ј nucleotide bordering the cleavage site. For the graph in E, data from six independent experiments were averaged. For the graph in F, data from a single representative experiment was used.
was cleaved in substrates ϩ1 and ϩ2 (lanes 5 and 6); site H was cleaved in substrates ϩ3, ϩ4, and ϩ5 (lanes 7-9); and site I was cleaved in substrates ϩ9 and ϩ10 (lanes 13 and 14). Notably, site HЈ, which produces a fragment 1 nucleotide longer than site H, was cleaved optimally in substrates ϩ5 and ϩ6 (lanes 9 and 10). When sites F-I were not optimally positioned in the cleavage window, fewer products resulting from weaker cleavages occurred (Fig. 6A, see lanes 11, 12, and 16). As was observed for HIV-1, for very strong cleavage sites such as F, significantly lower but detectable levels of cleavage occurred outside of the optimal window, most likely from weaker secondary cleavages (Fig. 6A, lanes 5-13). In the RϪ63/Ϫ23 hybrid series, cleavage at site Ϫ41/Ϫ40 was optimal in substrates with 3Ј ends at positions Ϫ60, Ϫ59, and Ϫ58 (Fig. 6B, lanes 3-5), and only weaker cleavages were seen in the other substrates (Fig. 6B, lanes 1, 2,  and 6 -13). Similar to the data shown for HIV-1 reverse transcriptase, cleavage at site Ϫ41/Ϫ40 was clearly both strong and not flanked by other strong sites, and the weaker sites that were observed followed the same general trend regarding the optimal distance from the DNA 3Ј end for cleavage.
For both the R46 and RϪ63/Ϫ23 series of substrates using M-MuLV reverse transcriptase, the amount of cleavage product generated at each cleavage site, plotted as a function of the distance from the DNA 3Ј end, revealed that the optimal cleavage window was between 17 and 20 nucleotides from the recessed end (Fig. 6, C and D). This cleavage window is 2 nucleotides narrower than that observed for HIV-1 reverse transcriptase (compare Fig. 6, C and D with Fig. 5, E and F, respectively). Within the cleavage window for M-MuLV reverse transcriptase, DNA 3Ј end-directed cleavages appeared most optimal when located at 18 or 19 nucleotides from the DNA 3Ј end.
Effects of dNTP Binding on DNA 3Ј End-directed Cleavage by RNase H-Unlike internal or RNA 5Ј end-directed cleavages, DNA 3Ј end-directed cleavages can occur during DNA synthesis. RNase H assays are often carried out in the absence of dNTPs to simplify analyses, including the previous experiments in this study. However, given that a free DNA 3Ј end is a substrate for polymerase addition and that the polymerase domain of reverse transcriptase can accommodate the RNA template strand, the 3Ј primer terminus of the DNA strand, and the incoming dNTPs, we asked whether the binding of cognate versus non-cognate dNTPs might affect the positioning of DNA 3Ј end-directed cleavages.
To assess whether nucleotide binding affects DNA 3Ј enddirected RNase H cleavage by M-MuLV or HIV-1 reverse tran- scriptase, we used the strong and relatively isolated cleavage site Ϫ41/Ϫ40 in RϪ63/Ϫ23 hybrids (Fig. 1B). The hybrids used for this analysis contained DNA primers that positioned site Ϫ41/ Ϫ40 over a distance of 15-20 nucleotides from a DNA 3Ј end and that were terminated with a ddNTP to prevent primer extension (Fig. 7A). When RNase H assays were carried out comparing hybrids with primers that had a 3Ј-hydroxyl terminus with hybrids with primers containing a dideoxy 3Ј end, cleavage products generated by HIV-1 or M-MuLV reverse transcriptase were identical in all cases (data not shown). This indicated that primers with a dideoxy 3Ј terminus did not affect DNA 3Ј end-directed cleavage for either enzyme.
The RϪ63/Ϫ23 hybrid series containing 3Ј ddNTP substitutions was used in RNase H assays in the absence of dNTPs or in the presence of either a non-cognate dNTP or the next cognate dNTP to test the effects of dNTP binding on DNA 3Ј enddirected RNase H cleavages. When HIV-1 reverse transcriptase was added to reactions containing a 200 M concentration of a non-cognate dNTP, cleavage at site Ϫ41/Ϫ40 was unaffected and identical to that observed in the absence of a dNTP for all primers (Fig. 7B, compare lanes 3, 7, 11, 15, 19, and 23 with  lanes 2, 6, 10, 14, 18, and 22, respectively). For Primers 17 and 18, the cleavage pattern remained unchanged when the next cognate dNTP was included in reactions (Fig. 7B, compare  lanes 10 -12 and 14 -16, respectively).
However, with primers positioning site Ϫ41/Ϫ40 at 15, 16, 19, or 20 nucleotides from the DNA 3Ј terminus, the presence of the cognate dNTP changed the cleavage pattern. Using substrate containing Primer 15, the presence of cognate dNTP reduced cleavage at site Ϫ41/Ϫ40 and eliminated cleavage at a site 19 nucleotides from the DNA 3Ј end, and a new cleavage site was observed 18 nucleotides from the DNA 3Ј end (Fig. 7B,  lanes 3 and 4). Similarly, for substrate 16, the cognate dNTP decreased cleavage at site Ϫ41/Ϫ40 and significantly increased cleavage at a site present 18 nucleotides from the DNA 3Ј end (Fig. 7B,  lanes 7 and 8). When site Ϫ41/Ϫ40 was positioned 19 nucleotides from the recessed end, the presence of the cognate dNTP slightly increased cleavages at sites found 17 and 18 nucleotides from the DNA 3Ј primer terminus while decreasing cleavages at sites 16 and 15 nucleotides from the primer terminus but did not seem to reduce cleavage of site Ϫ41/Ϫ40 significantly (Fig. 7B,  lanes 19 and 20). Similarly, when site Ϫ41/Ϫ40 was positioned 20 nucleotides from the DNA 3Ј end, cleavage at site Ϫ41/Ϫ40 appeared unchanged, but an increase in the 17-and 18-nucleotide cleavage products and a decrease in the 16-nucleotide cleavage product were observed (Fig. 7B, lanes 23  and 24). When compared with substrates 17 and 18, cleavage of site Ϫ41/Ϫ40 in substrates 19 and 20 is less, which correlates with the position of this site in the cleavage window as shown in the Ϫ59 and Ϫ60 substrates in Fig. 5D, lanes 4 and 3, respectively.
Thus with DNA primer 3Ј ends located 15, 16, 19, or 20 nucleotides away, the extent of cleavage at the Ϫ41/Ϫ40 site decreased, and the extent of cleavage at positions that were 17 or 18 nucleotides from the DNA 3Ј end increased. Substrates containing Primers 17 and 18, which already had DNA 3Ј ends directing cleavage at these positions, were unaffected. In other experiments, a 50 or 1000 M concentration of cognate and non-cognate deoxynucleotides had similar effects on RNase H cleavage by HIV-1 reverse transcriptase (data not shown).
The effects of cognate versus non-cognate deoxynucleotides on the RNase H activity of M-MuLV reverse transcriptase were also tested using these substrates. Unlike the observations presented above for HIV-1 reverse transcriptase, no differences in DNA 3Ј end-directed cleavages were observed for M-MuLV reverse transcriptase in reactions containing a 50, 200, or 1000 M concentration of cognate versus non-cognate dNTPs (data not shown).

DISCUSSION
In this study, we tested whether DNA 3Ј end-directed cleavages by the RNase H activity of reverse transcriptase are strictly dictated by the interaction of the DNA primer terminus engaged in the active site of the polymerase domain of reverse transcriptase or whether DNA 3Ј end-directed cleavages are determined by sequence preferences within a cleavage window. Our results clearly demonstrate that DNA 3Ј end-directed cleavages have sequence preferences that are identical to those described for internal and RNA 5Ј end-directed cleavages and that preferred sequences must fall within an acceptable distance or cleavage window from the DNA 3Ј end to be cleaved. Thus, the tight association of the polymerase domain with a potential 3Ј primer terminus does not restrict cleavage to a fixed distance from the recessed 3Ј end. If the DNA 3Ј end-directed cleavage window should happen to lack preferred sequences, then little or no DNA 3Ј end-directed cleavage occurs, despite the positioning afforded by the DNA 3Ј end on a hybrid substrate.
Previous studies with the HIV-1 and M-MuLV reverse transcriptases have shown that the internal and RNA 5Ј end-directed RNase H cleavage sites are flanked by preferred nucleotides at specific positions and that the preferences surrounding a site are different for the human and murine enzymes (for reviews, see Refs. It is likely that the observed sequence preferences reflect features of the substrate structure that interact well with reverse transcriptase and thereby facilitate RNase H cleavage (16).
In a number of earlier studies characterizing DNA 3Ј enddirected cleavages by the human or murine reverse transcriptase, sequence was not examined as a determinant of cleavage, and the precise locations of the sites were not mapped. However, in reports where the sites of DNA 3Ј end-directed cleavages were precisely determined, the nucleotides found at positions Ϫ4, Ϫ2, and ϩ1 for HIV-1 reverse transcriptase or at positions Ϫ2 and ϩ1 for M-MuLV reverse transcriptase agree well with the nucleotide preferences that we identified (23, 27, 29, 32, 36 -38). Notably, a few of these sites do have a nonpreferred nucleotide at a single position and still allow some cleavage, but we have found that changes at one or even two positions are typically insufficient to completely prohibit DNA 3Ј end-directed cleavage and instead reduce cleavage from the optimal level. 3 These data argue that sequence is an important determinant for both types of end-directed RNase H cleavage by the retroviral enzymes and that an eligible site, as defined by the nucleotide preferences identified for internal cleavage, is also recognized in the DNA 3Ј end-directed or RNA 5Ј enddirected mode of RNase H cleavage provided the site is positioned appropriately in a hybrid (see below).
Using matched hybrids, we found that DNA 3Ј end-directed cleavages consistently occur further from the recessed terminus than RNA 5Ј end-directed cleavages. For HIV-1 or M-MuLV reverse transcriptase, DNA 3Ј end-directed RNase H cleavages are located 15-20 or 17-20 nucleotides away from the recessed end, respectively, whereas for both enzymes, optimal RNA 5Ј end-directed cleavages fall as close as 13 nucleotides from the recessed end (31). Within the DNA 3Ј end-directed cleavage window, a site is cleaved distinctly better by M-MuLV reverse transcriptase when located 18 or 19 nucleotides from the DNA 3Ј end and by HIV-1 reverse transcriptase when located at all distances but 17 nucleotides from the DNA 3Ј end. By contrast, no significant distance preferences were observed within the window for sites in the RNA 5Ј end-directed cleavage mode (31).
In a noteworthy study examining the substrate requirements of secondary cleavage by HIV-1 reverse transcriptase, hybrid substrates containing primarily one RNA sequence were used to examine both DNA 3Ј end-directed and RNA 5Ј end-directed cleavages (38). Although these substrates were not designed to compare end-directed cleavages on matched hybrid sequences, the observed primary cleavages were at 18 or 19 nucleotides from the recessed DNA 3Ј end (or a blunt end) and at 15 nucleotides from the recessed RNA 5Ј end, which is in agreement with our data.
The binding interactions between reverse transcriptase and its substrate determine which types of RNase H cleavage are possible (23,24). Reverse transcriptase primarily interacts with a substrate near the RNase H and polymerase active sites (11)(12)(13) and binds RNA/DNA hybrids more tightly than duplexes of DNA or RNA (24, 39 -41). There are several additional contacts predicted between reverse transcriptase and RNA/DNA hybrids, including 11 interactions with RNA 2Ј-hydroxyls (13), and these contacts involve as few as 5 nucleotides of the RNA/ DNA hybrid interacting with the polymerase domain near the active site (42). When presented with a hybrid substrate, reverse transcriptase preferentially binds to a recessed DNA 3Ј or RNA 5Ј end (23)(24)(25)(26), and this affinity favors end-directed RNase H cleavages over internal RNase H cleavages (26). Rather than dissociating from a substrate to assume different binding interactions, recent experiments using single molecule techniques have revealed that HIV-1 reverse transcriptase can both flip orientations and slide between the termini on long DNA/ DNA and RNA/DNA substrates (43,44). Because the data in our biochemical assays were generated from a population of molecules, the importance of determinants such as sequence and distance from a recessed nucleic acid end reflects preferences in that population rather than absolute requirements in directing RNase H cleavages.
Several observations suggest that the contacts between the polymerase active site and the hybrid substrate coordinate the different types of RNase H cleavage. Although both the poly-merase and RNase H domains interact with substrate, most protein-nucleic acid contacts are proximal to the polymerase active site (12,13). Also, the distance between the polymerase and RNase H active sites is 17-18 nucleotides on the substrate as measured in co-crystal structures or enzymatic assays (12,13,15,27,32,37), and the window for DNA 3Ј end-directed cleavage begins at least 15 nucleotides from the recessed end. These considerations suggest that the preferred positions and nucleotides shared by all three types of cleavage (ϩ1, Ϫ2, and Ϫ4 for HIV-1 and ϩ1 and Ϫ2 for M-MuLV) are recognized near the scissile phosphate in the active site of the RNase H domain and that the sequence preferences at the more distal positions observed only for internal cleavage (Ϫ6 and Ϫ11 for M-MuLV and Ϫ7, Ϫ12, and Ϫ14 for HIV-1) and the recognition of a recessed end itself are related to substrate interactions near the active site of the polymerase domain (13,22).
It is interesting to consider why the cleavage windows for DNA 3Ј end-directed versus RNA 5Ј end-directed cleavages are different. There are multiple interactions between residues near the polymerase active site in the fingers and thumb subdomains of HIV-1 reverse transcriptase and a substrate with a recessed DNA 3Ј primer terminus (12). Because the cleavage window for DNA 3Ј end-directed cleavages is narrower and further from the recessed terminus, perhaps the numerous contacts with a recessed DNA 3Ј end on a longer RNA strand allow the substrate to fit more securely in the active site of the polymerase domain. These interactions might reduce the flexibility of the RNase H domain, restrict access of the RNase H domain to the RNA strand, and/or place the RNase H active site at a further distance from the polymerase domain. Conversely, the closer but also broader window observed for RNA 5Ј enddirected cleavages may reflect that, although there are multiple contacts between the recessed end and the polymerase active site that account for the higher level of RNA 5Ј end-directed as compared with internal cleavage, these interactions may allow more lateral movement of the substrate in the binding cleft. Some possible features of a substrate with a recessed RNA 5Ј end that might relax positioning in the polymerase active site are the absence of an RNA template strand that continues out of the active site, the presence of a long DNA 3Ј overhang, or the RNA 5Ј end itself. For example, perhaps the RNA 5Ј end acts as a "notch" that catches some region near to the polymerase active site and consequently enhances cleavage by the RNase H domain.
Based on the co-crystal structure of HIV-1 reverse transcriptase with a DNA duplex (12), there are multiple residues (Trp-24, Pro-25, Phe-61, Ile-63, Leu-74, Asp-76, and Gly-152) that contact the template strand downstream of the primer 3Ј terminus, and any of these residues might be candidates that facilitate RNA 5Ј end recognition. Notably, substitutions at Phe-61 in HIV-1 reverse transcriptase interfere with RNA 5Ј end-directed cleavages as well as with strand displacement synthesis (45)(46)(47). However, substitutions at Try-64 in M-MuLV reverse transcriptase (equivalent to Trp-24 in HIV-1) do not affect any type of RNase H cleavage but do interfere with strand displacement synthesis (48). Unfortunately, the co-crystal structures of HIV-1 and M-MuLV reverse transcriptases to date only reveal one orientation of enzyme-substrate binding in which the polymerase domain is bound to the 3Ј end of a DNA primer (11)(12)(13)(14), and thus, structures that reveal the interactions between reverse transcriptase and an RNA 5Ј end remain to be determined.
An interesting and related issue is how reverse transcriptase transitions between the two types of end-directed cleavage. In this work, we asked whether the cleavages directed by a recessed DNA 3Ј or RNA 5Ј end are mutually exclusive or whether a substrate could equally favor both types of end-directed cleavage. For the HIV-1 enzyme, DNA 3Ј and RNA 5Ј end-directed RNase H cleavages appeared to occur at approximately equal levels on hybrids containing an RNA 5Ј end recessed by 1 base, whereas for M-MuLV equal cleavage occurred with ends recessed by 2-3 bases. This intermediate cleavage pattern indicates that the positioning of reverse transcriptase to carry out RNA 5Ј versus DNA 3Ј end-directed cleavages is dynamic and that a slightly recessed RNA 5Ј end in a substrate can allow positioning within the polymerase domain to permit either mode of cleavage. Because the structural basis for the binding of a recessed RNA 5Ј end remains unknown, precisely how this positioning shift might occur within the polymerase domain is unclear. Again, insights will come from cocrystal structures that describe the interactions between reverse transcriptase and a recessed RNA 5Ј end in a hybrid substrate.
Our RNase H cleavage assays showed that the presence of cognate dNTP had subtle effects on DNA 3Ј end-directed RNase H cleavages for HIV-1 reverse transcriptase. The binding of primer-template by reverse transcriptase causes a conformational change in the position of the thumb subdomain, and when dNTP is bound to form the ternary complex, a second conformational change clamps the thumb subdomain more tightly to the substrate and moves the fingers subdomain closer to the active site of reverse transcriptase (12). Our data suggest that the formation of the ternary complex alters the relationship between the RNase H and DNA polymerase active sites in the HIV-1 reverse transcriptase such that the binding of cognate dNTP changes the positioning of RNase H cleavages to favor cleavage at 17 or 18 nucleotides from the primer 3Ј end over cleavages either closer or further from this distance. Consistent with our results, a recent study has shown that a DNA primer terminated with ddTTP in the absence of a cognate dNTP was cleaved between the 19th and 20th nucleotides from the 3Ј end, and in the presence of the next cognate dNTP, cleavage was shifted back to between the 18th and 19th nucleotides from the 3Ј end (49). The preference for cleavages at 17-18 nucleotides from the DNA 3Ј end in the ternary complex is consistent with co-crystal structures of HIV-1 reverse transcriptase showing a distance of 17-18 nucleotides between the polymerase and RNase H active sites (10,11,15).
Curiously, the formation of the ternary complex with M-MuLV reverse transcriptase does not affect the positioning of DNA 3Ј end-directed RNase H cleavages and is an interesting difference between the two enzymes. There are 32 additional amino acids found between the connection domain and the RNase H domain of M-MuLV reverse transcriptase as compared with HIV-1 that are predicted to make the murine RNase H domain more flexible (14). In addition, the M-MuLV reverse transcriptase has 11 amino acids that form the C-helix and loop that are absent in the HIV-1 enzyme (14,50). It is possible that either of these features of the M-MuLV reverse transcriptase minimizes any structural consequences of forming the ternary complex on DNA 3Ј end-directed RNase H cleavages.
Because RNase H is essential for viral replication (3)(4)(5)(6), this activity remains a promising target in developing successful drug therapies for HIV-1. Reverse transcription most likely requires all three types of RNase H cleavage, although the relative contributions by internal, DNA 3Ј end-directed, and RNA 5Ј end-directed cleavages remain unknown (8,16,51). At least four different categories of RNase H inhibitors might be considered possible. First are inhibitors that block substrate binding in the active site of the retroviral RNase H. Second are inhibitors that interfere with overall substrate binding such that RNase H cleavages are increased, decreased, or changed in a manner deleterious to reverse transcription. Third are inhibitors that chelate the metal ions required for RNase H activity. And fourth are compounds that target the DNA polymerase activity of reverse transcriptase but also influence RNase H activity. RNase H inhibitors representing several of these categories have been identified and show antiviral activity, but most display limited cellular uptake and/or high cell toxicity (for reviews, see Refs. 9 and 52), and none have been examined in clinical trials.
Although the active sites of the human RNase H1 and HIV-1 RNase H are highly conserved, a comparison of co-crystal structures of these enzymes has revealed intriguing structural differences that may influence substrate binding or enzyme activity in a manner that would allow for the design of inhibitors specifically targeting the viral RNase H (53). Notably, a recently identified hydroxylated tropolone displays a 30-fold selectivity for the HIV-1 RNase H over the human enzyme (54). To screen for drugs that preferentially inhibit HIV-1 RNase H, assays that discriminate between internal, RNA 5Ј end-directed, and DNA 3Ј end-directed cleavages should be considered. For example, an inhibitor that specifically blocks the end-directed modes of retroviral RNase H cleavage but does not alter internal RNase H cleavages might selectively interfere with the retroviral RNase H and reverse transcription without significantly impacting the endogenous human RNase H1 activity.