Substrate variations that affect the nucleic acid clamp activity of reverse transcriptases

Transcription

1 Substrate variations that affect the nucleic acid clamp activity of reverse transcriptases Iris Oz-Gleenberg, Eytan Herzig, Nickolay Voronin and Amnon Hizi* Department of Cell and Developmental Biology, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel Keywords nucleic acids clamp activity; reverse transcriptases; reverse transcription; strand transfer; substrate variations Correspondence A. Hizi, Department of Cell and Developmental Biology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel Fax: Tel: *A. Hizi is an incumbent of the Gregorio and Dora Shapira Chair for the Research of Malignancies (Received 29 January 2012, revised 29 February 2012, accepted 14 March 2012) doi: /j x We have recently shown that reverse transcriptases (RTs) perform template switches when there is a very short (two-nucleotide) complementarity between the 3 ends of the primer (donor) strand and the DNA or RNA template acceptor strands [Oz-Gleenberg et al. (2011) Nucleic Acids Res 39, ]. These dinucleotide pairs are stabilized by RTs that are capable of clamping together the otherwise unstable duplexes. This RT-driven stabilization of the micro-homology sequence promotes efficient DNA synthesis. In the present study, we have examined several factors associated with the sequence and structure of the DNA substrate that are critical for the clamp activity of RTs from human immunodeficiency virus type 1 (HIV-1), murine leukemia virus (MLV), bovine immunodeficiency virus (BIV) and the long terminal repeat retrotransposon Tf1. The parameters studied were the minimal complementarity length between the primer and functional template termini that sustains stable clamps, the effects of gaps between the two template strands on the clamp activity of the tested RTs, the effects of template end phosphorylations on the RT-associated clamp activities, and clamp activity with a long hairpin double-stranded primer comprising both the primer and the complementary non-functional template strands. The results show that the substrate conditions for clamp activity of HIV-1 and MLV RTs are more stringent, while Tf1 and BIV RTs show clamp activity under less rigorous substrate conditions. These differences shed light on the dissimilarities in catalytic activities of RTs, and suggest that clamp activity may be a potential new target for anti-retroviral drugs. Introduction Reverse transcription is a vital step in the life cycle of retroviruses and related long terminal repeat retrotransposons. This process is performed by retroviral reverse transcriptases (RTs), which convert (+)-sense single-stranded viral RNA into integration-competent double-stranded viral DNA [1 3]. To perform this multi-faceted step, RTs possess DNA polymerase activity, which is both DNA- and RNA-dependent, in addition to ribonuclease H activity that, concomitantly with DNA synthesis, cleaves the genomic RNA template within the generated RNA DNA duplex [4]. DNA synthesis activity produces both ()) and (+) DNA strands, while ribonuclease H removes the RNA primers and the viral genomic RNA template [1,3 5]. Two strand transfers are known to take place during reverse transcription, whereby the 3 end of the growing DNA strand switches to a second template [1 3,5]. In the first transfer, designated ())-strand transfer, the DNA copied from the 5 end of the viral genomic RNA is translocated onto the 3 end of this RNA. Abbreviations BIV, bovine immunodeficiency virus; HIV-1, human immunodeficiency virus type 1; MLV, murine leukemia virus; NTA, non-templated addition; RT, reverse transcriptase; T, template primer FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

2 I. Oz-Gleenberg et al. Nucleic acid clamp activity of various RTs In the second switch, the (+)-strand transfer, the 3 end of the (+) DNA strand with the primer binding site sequence switches to a complementary sequence in the ()) DNA strand. These transfers depend on stable sequence complementarities between the ends of the growing (donor) DNA and the acceptor template RNA or DNA strands. These complementary sequences are relatively long. Thus, the primer binding site sequence is usually 18 nucleotides long in most retroviruses, and the sequence promoting ())-strand transfer may even reach 100 nucleotides in length [1,6]. We have recently presented in vitro evidence that RTs perform template switches even with very short complementarities (two nucleotides) between the 3 ends of the primer donor strand and the DNA or RNA template acceptor strands [7]. These dinucleotide duplexes are stabilized by RTs, which clamp together these otherwise unstable duplexes. Consequently, RT-driven stabilization of this sequence micro-homology efficiently promotes DNA synthesis. This apparently newly discovered clamp function is performed by all RTs, but not by cellular DNA polymerases. Thus, this activity potentially affects the efficiency of reverse transcription, allowing the RT-associated DNA synthesis to bridge over nicks in the copied DNA or RNA templates while synthesizing DNA. The combined clamp activity and the capacity of some RTs to perform template-independent blunt-end DNA synthesis also support strand switches during DNA synthesis onto compatible acceptor strands [7,8]. In the present study, we examined the influence of several factors, such as nucleotide sequences, DNA folding and 5 and 3 end phosphorylations, on the clamp activities of a variety of RTs. The studied RTs included those from human immunodeficiency virus type 1 (HIV-1) and murine leukemia virus (MLV), which serve as prototype retroviruses of the lentivirus and gammaretrovirus groups, respectively [1,7]. In addition, we analyzed the RTs from lentivirus, bovine immunodeficiency virus (BIV) [9] and the long terminal repeat retrotransposon Tf1 [10]. Results and Discussion In our previous study, we characterized the clamp activity associated with RTs [7]. Such activity was observed with HIV-1 RT when the primer s 3 end overhang, which is annealed to the 3 end of the functional template (designated ), was at least two nucleotides long. Efficient clamp activity was also shown for the RTs of HIV-2, MLV, Tf1, BIV and avian myeloblastosis virus [7]. Another property of the clamp activity associated with HIV-1 RT was that it is substantial only when the functional template strand (designated ) adjoins a second DNA (or RNA) strand () that is annealed to most of the primer and is located upstream of [7]. We examined here whether these substrate restrictions also apply to the other RTs tested (from BIV, MLV and Tf1). We also determined whether phosphorylating the 3 end of the functional template () and or the 5 end of the adjacent strand () interferes with the correct alignment of the two template strands across the 3 end tail of the primer. Finally, we tested whether a long hairpin substrate, which includes both the primer and its complementary template () sequences, serves as an efficient clamp substrate after adding the strand. All assays used in the present study and previously [7,8] measure the combined two-step clamp DNA polymerase activity, where the initial clamp is followed by DNA synthesis. We have used similar DNA polymerase activities for all tested RTs (data not shown). Thus, the variations in the apparent combined clamp polymerase activities of the tested RTs result solely from differences in clamp activity per se, as the only rate-limiting step in the RT-driven reaction. In the experiments described, we have used the clampdependent DNA synthesis assay, as described previously in detail [7,8]. Minimal length of complementarity between the termini of the primer and functional template strands that supports a stable clamp Each RT was tested for its clamp activity with two sets of primer template substrates. In the first one (designated set #2 and already tested previously [7]), most of the 5 end-labeled 21 nucleotide primer (designated ) is annealed to the non-functional template (), but its last two 3 end nucleotides (with the GT sequence) form an overhang that is complementary to the last two nucleotides (AC) at the 3 end of the functional template (). The second substrate set (#52) is similar to set #2, except that the primer is shorter at its 3 end by a single nucleotide. Accordingly, in this case, only a single-nucleotide duplex (G C) is formed between the 3 end tails of and (see top of Fig. 1). The resulting combined clamp DNA synthesis showed that all tested RTs have efficient activities with two-nucleotide clamps, but to different extents. However, HIV-1 and MLV RTs lost all apparent activity when tested with the single-nucleotide clamp substrate, in contrast to BIV and Tf1 RTs, which were still quite active with this substrate although their activity was substantially lower relative to the two-nucleotide FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS 1895

3 Nucleic acid clamp activity of various RTs I. Oz-Gleenberg et al. T/ #2 (2nts clamp) T/ #52 (1nt clamp) * 5 ca cg 5 * G 3 5 c cg 5 HIV-1 RT Tf1 RT BIV RT MLV RT rimer 21 rimer Overhang (nts): Clamp (%): ± ± Fig. 1. Urea AGE analysis of combined clamp polymerase activity of the various RTs performed using substrates with oneand two-nucleotide clamp complementarities. Reactions were performed using the indicated T sets and RTs, as described in Experimental procedures. Asterisks indicate the 5 end 32 label. The ends of the primer () and the first template () are indicated by capital letters, and those of the functional template () are indicated by lower-case letters. The percentage clamp formation is the mean of two independent experiments for each RT with each T set. The results were normalized to the value for clamp formation obtained with T set #2 (two-nucleotide overhang). clamp. Thus, Tf1 RT retains approximately 24% of its initial clamp activity (single-nucleotide clamp versus double-nucleotide clamp), while BIV RT retains approximately 36% of its initial activity. Taken together, the data presented in Fig. 1 strongly suggest that ability to form a stable clamp structure between the tails of the primer, the functional template and the RT depends not only on the stability of the nucleic acid duplexes themselves, but also on the specific RT involved in stabilizing these complexes. The RTs from Tf1 and BIV are able to stabilize even a single-nucleotide complementarity between the duplexed, but the HIV-1 and MLV RTs are unable to do so. In the case of Tf1 RT, products significantly longer than the expected 59 nucleotides were also generated for both substrates (Fig. 1, lanes 3 and 4). This results from the high non-templated addition (NTA) (or terminal deoxynucleotide transferase) activity that is typical of Tf1 RT [10,11]. Thus, the 59-nucleotide blunt-ended products of the first combined clamp polymerase activity cycle are extended by a few nucleotides due to this DNA synthesis activity of Tf1 RT. Only these NTA products (which are slightly longer than the initial 59-nucleotide products and have GT or G tails) are further extended by a second round of clamp polymerase activity after annealing to and copying another strand (as molecules are in excess over the duplex). This process generates products that are 99 nucleotides long, i.e. the length of the initial product (59 nucleotides) plus the length of (oligonucleotide #I8, +40 nucleotides; Tables 1 and 2). This phenomenon is apparent in the results obtained with Tf1 RT (Figs 2A and 3). BIV RT also has significant NTA activity, but lower than that of Tf1 RT (data not shown). HIV-1 RT also shows NTA activity, but considerably lower than that for Tf1 and BIV RTs, while MLV RT shows hardly any NTA activity [10 12]. Therefore, some second-cycle products of clamp synthesis were also observed with BIV RT (Figs 2A and 3), but no significant activity was detected with HIV-1 and MLV RTs in any of the experiments (Figs 1 3) FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

4 I. Oz-Gleenberg et al. Nucleic acid clamp activity of various RTs Table 1. rimers templates sets used in this study. The designations of the labeled primers are shown for each set, together with the compatible templates. The 5 and 3 ends of the primers and first templates () are indicated by capital letters, and those of the functional template () are indicated by lower-case letters. Nicks or gaps between two adjacent templates (bottom strands) are indicated by arrowheads. Asterisks indicate the [ 32 ] 5 ends of primers. T set 5 end-labeled primer Template Scheme #2 I1 I3, I8 5 * CAGT 3 3 CA GT ca cg 5 #3 I1 I8 5 * CAGT 3 3 ca cg 5 #6 I1 I4, I8 5 * CAGT 3 3 CA G ca cg 5 #9 I1 I5, I8 5 * AGCAGT 3 3 CA TC ca cg 5 #52 I11 I3, I8 5 * CAG 3 3 CA GTca cg 5 #53 I1 I6, I8 5 * TAGCAGT 3 3 CA AT ca gt 5 #54 I1 I85, I8 5 * CA GT 3 3 CA GT-ca cg 5 #55 I1 I3, I86 5 * CA GT 3 3 CA GT -ca ct 5 #56 I1 I85, I86 5 * CA GT 3 3 CA GT--ca cg 5 #64 I94 I8 CCGTGTGGAAAATCTCTAGCAGT-3 C CCCACACCTTTTAGAGATCGT*ca gt-5 #65 I94 Not labeled I8 CCGTGTGGAAAATCTCTAGCAGT-3 C CCCACACCTTTTAGAGATCGT ca gt-5 Effects of gaps between the two adjacent template DNA strands ( and ) on the clamp activity of the tested RTs The nucleic acid clamp model implies that the clamp activity associated with HIV-1 RT depends on tandem alignment of the template strands [7]. Therefore, in order to achieve an efficient clamp, the 3 end of the functional template () must be positioned immediately next to the 5 end of the second template strand () that is annealed to most of the length of the primer (top of Fig. 1) [7]. To further investigate how gaps between the and strands affect efficient clamp activity, we tested the four RTs using a substrate containing no gap (set #2) and similar substrates containing gaps ranging from 1 to 3 nucleotides. These gaps were created by shortening the 5 end of the strand. The data in Fig. 2 show that MLV RT is the most sensitive to gaps introduced between and, as no FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS 1897

5 Nucleic acid clamp activity of various RTs I. Oz-Gleenberg et al. Table 2. The synthetic oligonucleotides used in this study. Oligonucleotide designations Sequence 5 - to 3 I1 (21-mer) I3 (19-mer) I4 (18-mer) I5 (17-mer) I6 (16-mer) I8 (40-mer) I11 (20-mer) I40 (20-mer) I85 (19-mer) I86 (25-mer) I94 (45-mer) GTGTGGAAAATCTCTAGCAGT TGCTAGAGATTTTCCACAC GCTAGAGATTTTCCACAC CTAGAGATTTTCCACAC TAGAGATTTTCCACAC GCCGGCCCATGGTCTTCCTAGAAAATATCCCCTCAGCCAC GTGTGGAAAATCTCTAGCAG CGTATGCGGCAAGCTTTACC -TGCTAGAGATTTTCCACAC TCCTAGAAAATATCCCCTCAGCCAC- TGCTAGAGATTTTCCACACCCCCCGTGTGGAAAATCTCTAGCAGT A HIV-1 RT Tf1 RT BIV RT MLV RT 59 rimer T/ set: #3 #2 #6 #9 #53 #3 #2 #6 #9 #53 #3 #2 #6 #9 #53 #3 #2 #6 #9 #53 Gap (nts): no no no no B Clamp (%) Gap (nts) 3 HIV-1 RT Tf1 RT BIV RT MLV RT 4 Fig. 2. Effect of gaps between the two template strands on formation of a stable clamp by the tested RTs. Reactions were performed using the indicated T sets and RTs, as described in Experimental procedures. For the description of each T set, see Tables 1 and 2. (A) Urea AGE analysis of the clamp polymerase activity. (B) Results obtained in two independent experiments similar to the one described in (A). The results were normalized to the value for clamp formation obtained with T set #2 (containing no gaps) FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

6 I. Oz-Gleenberg et al. Nucleic acid clamp activity of various RTs * T/ #2 ca cg 5 T/ #54 * p-5 ca cg 5 T/ #55 T/ #56 * 3 -p-ca ct 5 * p-5 3 -p-ca ct 5 HIV-1 RT Tf1 RT BIV RT MLV RT Fig. 3. Urea AGE analysis of the clamp polymerase activity of the RTs obtained with phosphorylated templates. Reactions were performed using the indicated T sets and RTs, as described in Experimental procedures. Each T set is described at the top. Lower-case p indicates phosphate groups at the 5 and or 3 ends of the template strands. rimer T/ set: #2 #54 #55 #56 #2 #54 #55 #56 #2 #54 #55 #56 #2 #54 #55 #56 hosphate: no /3 no /3 no /3 no /3 location / / / / rimer 21 clamp activity was detected even with a single-nucleotide gap (Fig. 2A, lane 18). HIV-1 RT is also highly gap-sensitive, with a sharp reduction of approximately 95% in the initial clamp activity after a single-nucleotide gap was introduced (Fig. 2A, lane 3, and Fig. 2B). Longer gaps totally abrogate clamp formation by HIV-1 RT. Similar to the experiment studying the minimal complementarity sequence required for a stable clamp (Fig. 1), the RTs of BIV and Tf1 differ from those of MLV and HIV-1 with regard to gap length. Thus, the clamp activity of BIV RT is reduced by only approximately 44% when a single-nucleotide break was introduced (Fig. 2A, lane 13, and Fig. 2B), but longer gaps totally abolish any clamp formation. The clamp activity of Tf1 RT even tolerates longer gaps. There is only approximately 30% reduction in activity with a single-nucleotide break (Fig. 2A, lane 8). Moreover, even a two-nucleotides gap is accepted by Tf1 RT (with only approximately 72% reduction in activity, Fig. 2A, lane 9, and Fig. 2B); however, a three-nucleotide gap results in almost complete abolishment of the clamp activity (Fig. 2A, lane 10, and Fig. 2B). The apparent disparities between the tested RTs may be explained by the possibility that the 3 end of is extended by the RT towards, thus filling the gaps initially introduced between these two bottom strands. RT variations in gap-filling activity, in which serves as the template, could possibly explain the results observed in Fig. 2. However, in an experiment designed to test this issue, we did not observe differences in the gap-filling activities of the various RTs, suggesting that this is not the reason for the observed differences in the effects of gaps between and on the clamp activity of the RTs (data not shown). Effects of phosphorylating the 5 end of and the 3 end of DNA strands on the clamp activity of the RTs The results presented so far emphasize the importance of the sequence microenvironment near the clamped region in the substrates used. All clamp studies above were performed using a strand with a 5 -OH moiety and a functional template with a free 3 -OH moiety [7,8]. Another factor that may potentially affect the clamp activity may be phosphorylation of either the 5 end of or the 3 end of. To study this issue, we have used custom-synthesized and phosphorylated DNA oligonucleotides that are similar in sequence to those used in set #2, except for their lengths and the presence of a phosphate at either the 5 end of or the 3 end of. FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS 1899

7 Nucleic acid clamp activity of various RTs I. Oz-Gleenberg et al. The results presented in Fig. 3 and Table 3 show that a phosphate moiety on the 5 end of (T set #54) causes a substantial reduction in the clamp activity of HIV-1 and MLV RTs. Thus, their activity is suppressed by approximately 90% and 52%, respectively, relative to their control activity with non-phosphorylated templates (Fig. 3, lanes 2 and 14, respectively, and Table 3). In contrast, the RTs of Tf1 and BIV were hardly affected (Fig. 3, lanes 6 and 10, respectively, and Table 3). Similarly, the clamp activity obtained with the 3 end-phosphorylated strand (T set #55) is suppressed by approximately 60% and 80%, respectively, for HIV-1 and MLV RTs (Fig. 3, lanes 3 and 15, respectively, and Table 3). However, the Tf1 and BIV RTs were barely affected by 3 end phosphorylation of the 25-nucleotide template (Fig. 3, lanes 7 and 11, respectively). The expected length of the products is 44 nucleotides ( ) 2). When and are both phosphorylated (T set #56), the clamp activity of the Tf1 and BIV RTs was reduced by approximately 65% (Fig. 3, lanes 8 and 12, and Table 3). As expected, the clamp activity of HIV-1 and MLV RTs was fully lost (Fig. 3, lanes 4 and 16, respectively, and Table 3). These results indicate that, in the case of HIV-1 and MLV RTs, substitution of the OH group at either the 5 end of or the 3 end of by the bulkier O 4 group (and or the negative charge contribution of the O 4 group) significantly interferes with stable clamp formation. As the reduction in clamp formation is less pronounced with Tf1 and BIV RTs, these RTs are able to achieve a stable clamp with relatively little interference from a single O 4 moiety or even two O 4 groups. It should be also noted that Tf1 RT pauses quite frequently when tested with 3 endphosphorylated and when both and are phosphorylated (Fig. 3, lanes 7 and 8). In addition, under the conditions used, the Tf1 RT-associated NTA activity yielding products longer than the expected 44 nucleotides is also very pronounced. Table 3. Clamp polymerase activity of RTs obtained with phosphorylated templates in three independent experiments similar to the one described in Fig. 3. The results were normalized to the value for clamp formation obtained with T set #2 (i.e. without template phosphorylation and with regular 5 end labeling of the primer). All values are percentages. RT T #2 T #54 T #55 T #56 HIV ± ± ± 0.3 BIV ± ± ± 2 Tf ± ± 9 35 ± 6 MLV ± ± 4 0 Clamp activity with a long hairpin substrate containing sequences derived from both the primer and the strand The observation of clamp formation by RTs in the presence of the appropriate substrate raises the question of whether nucleic acid structures, which are different from the classical three-oligonucleotide format studied so far [7], also support RT-dependent clamp formation. To assess this, we designed a new substrate: T sets #64 and #65 (Fig. 4 and Tables 1 and 2). Here, the functional primer is extended at its 5 end by a sequence derived from the original strand used in the previous substrate set (oligonucleotide #I3 in T set #2, Tables 1 and 2). Consequently, this 45-nucleotide oligonucleotide (designated #I94) folds back on itself (due to internal sequence pairing), leaving a 3 end GT overhang that is free to form a duplex with the AC tail of the strand (Tables 1 and 2). First, we tested this substrate with 32 5 end-labeled oligonucleotide #I94 and (T set #64). HIV-1 RT showed hardly any combined clamp polymerase activity (Fig. 4A, lane 1), but the other tested RTs were effective to varying extents in using this substrate, with the strongest activity obtained using Tf1 RT (Fig. 4A, lane 2). However, the poor utilization of this unique substrate by HIV-1 RT may not be due to the substrate itself, but instead may result from the fact that the 5 end of the folded oligonucleotide #I94 is phosphorylated by 32 (as this substrate is analogous to substrate set #54 shown above in Fig. 3). As 5 end phosphorylation of substantially reduced the stability of the clamp formed by HIV-1 RT (Fig. 3, lane 2, and Table 3), it is possible that the poor activity observed in Fig. 4A, lane 1, results directly from interference by the 5 end phosphorylation. To test this issue, we repeated this experiment using a non-5 end-labeled oligonucleotide #I94 (Fig. 4B). Synthesis was followed by incorporation of [a 32 ]- labeled dtt. Thus, direct utilization of the hairpin substrate with no interference from 5 end phosphorylation of the 45-nucleotide folded primer was tested. The results show that HIV-1 RT utilizes the long hairpin substrate quite well; hence, the low activity observed with the 5 end-labeled substrate reflects the hindrance imposed by the 5 end O 4 moiety (Fig. 4B, lane 1). The other tested RTs also utilized this substrate quite efficiently (Fig. 4B). Overall, it appears that the order of utilization for this substrate is: Tf1 RT BIV RT > HIV-1 RT > MLV RT. Interestingly, under the assay conditions used, DNA synthesis by Tf1 RT pauses quite frequently, leading to premature terminations. This finding, in addition 1900 FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

8 I. Oz-Gleenberg et al. Nucleic acid clamp activity of various RTs A T/ #64 B T/ #65 *5 ca gt 5 5 ca gt 5 HIV-1 RT Tf1 RT BIV RT MLV RT HIV-1 RT Tf1 RT BIV RT MLV RT Fig. 4. Clamp polymerase activity of the tested RTs utilizing a long hairpin substrate. Reactions were performed using the indicated T sets (described at the top) and RTs, as described in Experimental procedures. (A) Urea AGE analysis of the clamp polymerase activity obtained with T set #64 (5 end-labeled with [c 32 ]AT) and the indicated RTs. (B) Urea AGE analysis of the clamp polymerase activity obtained with T set #65 (not labeled at its 5 end) and the indicated RTs. Here, synthesis was followed by adding [a 32 ]labeled dtt (at a final concentration of 5 lm) to the three other non-labeled dnts (each at a final concentration of 50 lm) rimer to the synthesis pattern shown above in Fig. 3 (lanes 7 and 8), suggests that Tf1 RT has marked low processivity of DNA synthesis with some of the substrates used. Conclusions In summary, the tested RTs were shown to differ in their clamp activity when tested with the various substrates. HIV-1 and MLV RTs have more stringent requirements with respect to substrates, while Tf1 and BIV RTs perform clamping under more relaxed conditions. Thus, unlike HIV-1 and MLV RTs, BIV and Tf1 RTs clamp substrates with a single-nucleotide complementarity between the primer and the functional temple strand, and also tolerate short gaps between the two template strands. We have previously suggested that a biological explanation for the clamp activity associated with RTs is the possibility that it allows the RT-associated DNA synthesis to bridge over nicks in the copied DNA or RNA templates while synthesizing DNA [7]. The data presented here show that the RTs of Tf1 and BIV are even able to bridge over gaps in the copied template of one nucleotide, or even two nucleotides, in the case of Tf1 RT. Overall, the four tested RTs may be divided, according to their clamp parameters, into two groups: the first comprising Tf1 and BIV RTs and the second comprising HIV-1 and MLV RTs. Based on sequence alignments [calculated using CLUSTALW pairwise alignment ( clustalw2/)], this overall RT grouping is unexpected because the homology between full-length HIV-1 and BIV RTs (both lentiviruses) is significantly higher (homology score 37) than that between MLV and HIV-1 RTs (homology score 16) or between Tf1 and BIV RTs (homology score 14). Another feature that distinguishes the lentiviral RTs from the RTs of MLV and Tf1 is that the former are heterodimeric in structure, while the latter two are monomeric [3,13]. Based on structure modeling of HIV-1 RT, we previously proposed that a number of residues in HIV-1 RT are associated with clamp activity [7]. These residues are Glu89, Val90, Gln91, Leu92, Lys154, ro157 and Ala158, which are located in the palm sub-domain of the large subunit of the p66 p51 heterodimeric HIV-1 RT [14]. Multiple sequence alignment of the amino acids sequences within the fingers and palm subdomains of the four RTs studied here showed that, out these suggested residues, only ro157 is fully conserved among all tested RTs (Fig. 5). Thus, this alignment does not explain the suggested grouping of RTs according to their clamp activities. Therefore, mutational analyses are being performed to further elucidate these findings. For example, we are testing whether replacing several of the residues in HIV-1 RT (e.g. residues 89 92) by the comparable ones in Tf1 RT converts the clamp activity of HIV-1 RT into activity typical of Tf1 RT. Furthermore, other RTs that belong to additional retroviral groups will be tested to see how general the apparent variations in clamp activity are. The observed differences between the clamp activities of the tested RTs shed light on the disparities between the catalytic functions of these RTs. This FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS 1901

9 Nucleic acid clamp activity of various RTs I. Oz-Gleenberg et al. RT: Glu 89 Val 90 Gln 91 Leu 92 HIV-1 77 FRELNKRTQDFW-EVQLGIHAGLK-KKKSVTVLDVGDAYFSVLDEDFRKYTAFTIS 134 BIV 76 FRELNKITVKGQ-EFSTGLYGIK-ECEHLTAIDIKDAYFTILHEDFRFTAFSVV 133 MLV 115 LREVNKRVEDIHTVNYNLLSGLSHQWYTVLDLKDAFFCLRLHTSQLFAFEWRD 174 Tf1 117 YKLNKYVKNIYLLIEQLLAKIQ-GSTIFTKLDLKSAYHLIRVRKGDEHKLAFR : :**... : * :*:.*:. : :. ** HIV INNETGIRYQYNVLQGWKGSAIFQSSMTKILEFKKQNDIVIYQYMDDLYVGSDLE 194 BIV 134 VNREGIERFQWNVLQGWVCSAIYQTTTQKIIENIKKSHDVMLYQYMDDLLIGSNRD 193 MLV 175 E-MGISGQLTWTRLQGFKNSTLFDEALHRDLADFRIQHDLILLQYVDDLLLAATSE 233 Tf CRGVFEYLVMYGISTAAHFQYFINTILGEAKESH----VVCYMDDILIHSKSE 224 : :* * :*: :: : :.: : *:**: : : : Lys 154 ro 157 Ala 158 Fig. 5. Multiple sequence alignment of the tested RTs. Alignment was performed using the ClustalW2 multiple sequence alignment tool ( Asterisks indicate positions at which there is a single, fully conserved residue; colons indicate residues with strongly similar properties; full points indicate residues with weakly similar properties. Only the relevant part of the alignment performed with the full protein sequence of each RT is shown. Residues that were suggested in our previous study to be associated with the clamp activity of HIV-1 RT [7] are indicated in bold; their positions (referring to the HIV-1 RT sequence) and amino acid abbreviation are also shown. study may also potentially assist in defining the possible involvement of clamp activity in retroviral reverse transcription in various retroviruses and retroelements. Moreover, such clamp activity may be a new potential target for the development of anti-retroviral drugs. Experimental procedures Reverse transcriptases All purified recombinant RTs carry six-histidine tags, as previously described by us. These RTs were from the BH-10 strain of HIV-1 [15], Moloney MLV [16], the long terminal repeat retrotransposon Tf1 [10] and BIV [9]. The DNA polymerase activities of all tested RTs were normalized as described in detail previously [7]. Oligonucleotides All oligonucleotides were custom-synthesized and HLCpurified by Metabion (Martinsried, Germany) (Tables 1 and 2). Combined clamp polymerase assay reactions All reactions were performed as described previously [7]. In brief, primers were 5 end-labeled using T4 polynucleotide kinase and [c 32 ]AT, and annealed to the appropriate templates (Tables 1 and 2). The molar ratio of templates to primers was 4:1. Reactions were performed using the indicated template primer (T ) sets and 128 nm HIV-1 RT, 122 nm BIV RT, 171 nm MLV RT or 108 nm Tf1 RT, calibrated and assayed as described previously [7]. Reactions were initiated by adding all four dnts, each at a final concentration of 50 lm (unless otherwise indicated) for 30 min at 37 C, and then stopped by adding formamide loading buffer. The reactions were analyzed by urea polyacrylamide gel electrophoresis (urea AGE), followed by autoradiography, as previously described [16,17]. Acknowledgements This research was supported in part by a grant from the Israeli Science Foundation (number ). References 1 Coffin JM, Hughes SH & Varmus HE (1997) Retroviruses. Cold Spring Harbor Laboratory ress, Cold Spring Harbor, NY. 2 Menendez-Arias L & Berkhout B (2008) Special issue on: retroviral reverse transcription. Virus Res 134, Herschhorn A & Hizi A (2010) Retroviral reverse transcriptases. Cell Mol Life Sci 67, Schultz SJ & Champoux JJ (2008) RNase H activity: structure, specificity, and function in reverse transcription. Virus Res 134, Basu V, Song M, Gao L, Rigby ST, Hanson MN & Bambara RA (2008) Strand transfer events during HIV- 1 reverse transcription. Virus Res 134, Le Grice SF (2003) In the beginning : initiation of minus strand DNA synthesis in retroviruses and LTRcontaining retrotransposons. Biochemistry 42, FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

10 I. Oz-Gleenberg et al. Nucleic acid clamp activity of various RTs 7 Oz-Gleenberg I, Herschhorn A & Hizi A (2011) Reverse transcriptases can clamp together nucleic acids strands with two complementary bases at their 3 -termini for initiating DNA synthesis. Nucleic Acids Res 39, Oz-Gleenberg I & Hizi A (2011) Strand selections resulting from the combined template-independent DNA synthesis and clamp activities of HIV-1 reverse transcriptase. Biochem Biophys Res Commun 408, Avidan O, Bochner R & Hizi A (2006) The catalytic properties of the recombinant reverse transcriptase of bovine immunodeficiency virus. Virology 351, Kirshenboim N, Hayouka Z, Friedler A & Hizi A (2007) Expression and characterization of a novel reverse transcriptase of the LTR retrotransposon Tf1. Virology, 366, Oz-Gleenberg I, Herzig E & Hizi A (2012) The template-independent DNA synthesis activity associated with the reverse transcriptase of the LTR retrotransposon Tf1. FEBS J 279, Golinelli M & Hughes SH (2002) Nontemplated nucleotide addition by HIV-1 reverse transcriptase. Biochemistry 41, Hizi A & Herschhorn A (2008) Retroviral reverse transcriptases (other than those of HIV-1 and murine leukemia virus): a comparison of their molecular and biochemical properties. Virus Res 134, Jacobo-Molina A, Ding J, Nanni RG, Clark AD Jr, Lu X, Tantillo C, Williams RL, Kamer G, Ferris AL & Clark et al. (1993) Crystal structure of human immunodeficiency virus type 1 reverse transcriptase complexed with double-stranded DNA at 3.0 A resolution shows bent DNA. roc Natl Acad Sci USA 90, Sevilya Z, Loya S, Hughes SH & Hizi A (2001) The ribonuclease H activity of the reverse transcriptases of human immunodeficiency viruses type 1 and type 2 is affected by the thumb subdomain of the small protein subunits. J Mol Biol 311, Avidan O, Loya S, Tonjes RR, Sevilya Z & Hizi A (2003) Expression and characterization of a recombinant novel reverse transcriptase of a porcine endogenous retrovirus. Virology 307, Oz-Gleenberg I, Avidan O, Goldgur Y, Herschhorn A & Hizi A (2005) eptides derived from the reverse transcriptase of human immunodeficiency virus type 1 as novel inhibitors of the viral integrase. J Biol Chem 280, FEBS Journal 279 (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS 1903