Journal list menu

Volume 582, Issue 18 p. 2799-2805
Short communication
Free Access

Mutagenesis of Gln294 of the reverse transcriptase of human immunodeficiency virus type-2 and its effects on the ribonuclease H activity

R. Bochner

R. Bochner

Department of Cell and Developmental Biology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 62263, Israel

Molecular Genetics and Biochemistry at The Sackler School of Medicine, Tel Aviv University. Search for more papers by this author
A. Duvshani

A. Duvshani

Department of Cell and Developmental Biology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 62263, Israel

Present address: Columbia University School of Medicine, New York, USA. Search for more papers by this author
N. Adir

N. Adir

Faculty of Chemistry, Institute of Catalysis, Science and Technology, Technion, Israel Institute of Technology, Haifa 32000, Israel

Search for more papers by this author
A. Hizi

A. Hizi

Department of Cell and Developmental Biology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 62263, Israel

Incumbent of the Gregorio and Dora Shapira Chair for the Research of Malignancies at Tel-Aviv University. Search for more papers by this author
First published: 14 July 2008
Citations: 5
Corresponding author. Fax: +972 3 6407432.

Abstract

Despite the high homology between human immunodeficiency virus type-1 (HIV-1) and human immunodeficiency virus type-2 (HIV-2) reverse transcriptases (RTs), the ribonuclease H (RNase H) level of HIV-2 RT is lower than that of HIV-1 RT, while the DNA polymerase of both RTs is similar. We conducted mutagenesis of HIV-2 RT Gln294 (shown to control the RNase H activity level when modified to a Pro in the smaller p54 subunit and not in the larger p68 subunit) to various residues, and assayed the activities of all mutants. All exhibited an RNase H that is higher than the wild-type (WT) HIV-2 RT level, although the DNA polymerase of all mutants equals WT HIV-2 RT level. These results represent a unique case, where every mutation induces an increase rather than a decrease in an enzyme's activity.

1 Introduction

Reverse transcriptases (RTs) of retroviruses have two enzymatic activities, a DNA polymerase (that copies both RNA and DNA templates) and the ribonuclease H (RNase H) activity that hydrolyzes the RNA in DNA–RNA hybrids [1, 2]. The fundamental role in the replication of retroviruses makes RT a major target for the drug treatment against human immunodeficiency virus type-1 (HIV-1) and human immunodeficiency virus type-2 (HIV-2), the lentiviruses that cause human acquired immunodeficiency syndrome (AIDS). Both HIV RTs are heterodimeric proteins, each composed of two polypeptides, the p66/p51 for HIV-1 RT or p68/p54 for HIV-2 RT [3-7]. The larger subunit in both RTs is composed of two domains, DNA polymerase and RNase H. The polymerase domain, located in the amino terminal portion, resembles a right hand with fingers, palm and thumb subdomains. It is joined to the RNase H domain, located at the carboxyl terminus of the larger subunit, by a connection subdomain [3, 5-7]. The DNA polymerase active site is located within the palm subdomain of the p66 or p68 subunit of either HIV-1 or HIV-2 RT, respectively. The smaller RT subunit is a cleavage product of the larger one (where the entire carboxyl terminal segment is removed); hence, the smaller subunit lacks the RNase H domain. Even though the amino acid sequence of this subunit is identical to the larger subunit, the fingers, palm and thumb fold differently in the heterodimeric protein. Other polymerases also show an overall resemblance to a right hand structure with the active site located within the palm subdomain. These include the DNA-dependent DNA polymerases of both polymerase β and the Klenow large fragment of Escherichia coli polymerase I [8, 9], the DNA-dependent RNA polymerase of T7, the RNA-dependent RNA polymerase of poliovirus [10, 11] and the RT of murine leukemia retrovirus [12, 13].

The specific DNA polymerase activities of purified recombinant RTs of HIV-1 and HIV-2 are similar. However, the specific RNase H activity of HIV-2 RT from both Rod and D194 strains was shown in several studies to be up to about ten-fold lower than that of HIV-1 RT [14-19]. Unexpectedly, part of the polymerase domain, rather than the RNase H domain per se was found by us to have a significant effect on the level of the RNase H activity [17, 18]. The observed difference in the RNase H activity between the HIV-1 and HIV-2 RTs is quantitative rather than qualitative, as the pattern of RNA cleavage by the two RTs is indistinguishable. We have found that the thumb subdomain of the smaller RT subunit has a major role in determining the level of the RNase H activity, either through interactions with the RNase H catalytic domain of the larger subunit or with the substrate molecules themselves [17]. Interestingly, previous observations have shown that the isolated thumb subdomain of HIV-1 RT has only a marginal stimulatory effect on the RNase H activity of the RT [20]. Later on, a single residue, Gln294, located in the thumb of the small HIV-2 RT subunit, was shown to be solely responsible for the outlined difference in specific RNase H activity. Thus, a mutant HIV-2 RT, where Gln294 was replaced by a proline (that is the comparable residue located at this position in HIV-1 RT), has an activity as high as that of HIV-1 RT [16]. Moreover, it was shown that while the Gln294 in the small RT subunit has a major impact on the RNase H activity, the identical residue on the larger RT subunit has hardly any effect. Steady state kinetics revealed that residue 294 in the small RT subunit affects the K m rather than the k cat value; thus influencing the affinity of the enzyme towards the RNA–DNA substrate. As could be expected, we also found that the reciprocal mutation P294Q mutation of HIV-1 RT leads to a substantial reduction in the enzyme's specific RNase H activity [16].

In the present investigation we have addressed the question as to how various amino acids, other that Gln and Pro at position 294 of HIV-2 RT affect the enzyme's RNase H activity. This was done by a systematic mutagenesis of Gln294 to 10 different residues (in addition to the Pro mutant already studied) that represent all different chemical types of amino acid side chains. The results show that all modifications of HIV-2 RT Gln294 lead to an increase of the RNase H activity of the enzyme relative to the wild-type (WT) HIV-2 RT and also support our previous observations that residue 294 in the p54 subunit of the RT is the major contributor to the level of this activity in HIV-2 RT. These results indicate an exceptional case where every mutation of a particular residue brings about an increase rather than a decrease in the specific catalytic activity of the studied enzyme.

2 Materials and methods

2.1 Plasmid construction for the expression of RTs

All recombinant versions of both HIV-1 and HIV-2 RT used in this study were expressed in E. coli with six-histidine tags [16]. The plasmids, designated pHIV-1 RT and pHIV-2 RT, induce the expression of the WT heterodimeric RTs of the HIV-1 BH-10 isolate and the HIV-2 Rod isolate [14, 21]. In these plasmids, the genes that encode the large subunit of the RT (p66 or p68 for HIV-1 RT or HIV-2 RT, respectively) were introduced in the pT5m plasmid along with the genes encoding the relevant HIV-1 or HIV-2 protease [17]. The co-expression of the RT and protease proteins was induced by isopropyl β-thiogalactopyranoside (IPTG) in the BL21(DE3)pLysS E. coli strain, leading to the simultaneous expression of both the large RT subunit and the protease. This causes the homodimeric RT to be proteolytically cleaved in the bacteria, leading to the accumulation of the heterodimeric p66/p51 and p68/p54 for HIV-1 RT or HIV-2 RT, respectively.

In order to generate the new Gln294 mutants of HIV-2 RT used in this study, we have employed a cassette mutagenesis method described for the mutagenesis of HIV-1 RT and HIV-2 RT [4, 16]. We have constructed a BspMI cassette by introducing the BspMI restriction site into the HIV-2 RT coding region in the pUC12N HIV-2 RT expressing plasmid (in which the HindIII site downstream to the RT-coding insert was replaced by a NotI restriction site). The cassette differs from the parental HIV-2 RT expressing plasmid in that 100 bp of the RT sequence were deleted. This deleted segment, which codes for the residues from glycine 285 to tyrosine 318 in the HIV-2 RT protein, was replaced by DNA segment containing a HindIII recognition site flanked by two BspMI sites, orientated at opposite directions. The synthetic DNA fragments inserted in the BspMI cassette contained the RT coding segments with the specific mutations. The phosphorylated oligomers were ligated into the linearized BspMI cassette – containing plasmid. After transforming E. coli DH5α, the DNA of the selected plasmids was sequenced for verification. The mutant RTs-expressing genes were used to replaced the gene for the WT HIV-2 RT in the vector HIV-2 RT (Rod) [16]; thus generating the vectors for expressing the heterodimeric mutant HIV-2 RT versions with identical modifications in both subunits. The subunit-selective mutant of HIV-2 RT was also expressed from the plasmid constructed to co-express separately the two HIV-2 RT subunits [16, 17], the large one as a WT protein and the small subunit as the Q294R mutant.

2.2 Protein purifications

All recombinant six-histidines-tagged RT versions were purified as described in detail previously [17]. Briefly, the bacterial extracts underwent first an affinity chromatography using Ni-Agarose to enrich for the histidines-tagged proteins and the RT-containing fractions were further purified by carboxyl-methyl Sepharose ion exchange chromatography. All RT versions were highly pure as judged from polyacrylamide gel analysis and were devoid of any detectable non-specific nuclease activities (data not shown).

2.3 RNase H assay

This assay has been previously described in detail [16, 17]. In brief, it monitors the cleavage of [32P]5′-end labeled 267 nt RNA, synthesized in vitro from the pBLRA30 plasmid. This RNA, derived from the U5 and R regions of HIV-1 genome, was annealed to a 20 nt-long synthetic oligonucleotide DNA with the sequence 5′-AGTTAGCCAGAGAGCTCCCA-3′ (see scheme in Fig. 1 A). The RNase H cleavage products were electrophoretically resolved on a denaturating sequencing polyacrylamide gel, and the dried gels underwent autoradiography.

figure image
RNase H activity of the Gln294 HIV-2 RT mutants studied. (A) Schematic description of the enzymatic assay. The 267 nt RNA, synthesized in vitro from the pBLRA30 plasmid, as was described in Section 2. This RNA was annealed to a synthetic 20 nt complementary oligomeric DNA. The primary RNase H cleavages are near position 47, approximately 17 bases downstream of the 5′ end of the RNA and the secondary cleavages are near position 38 (8 nt upstream from the primary cleavage). The RNA and DNA substrates and the cleavages are shown schematically. The drawings are not to scale. (B) 20 ng of each heterodimeric RT were incubated with the [32P]-end labeled RNase H substrate for 15 s or 1 min at 37 °C (marked as 1 and 2, respectively). The reaction products were resolved by urea-PAGE and the gel underwent autoradiography. The two controls RTs of this assay, wild-type HIV-1and HIV-2 RTs (designated HIV-1 and HIV-2, respectively) were assayed along with the 11 HIV-2 RT mutants tested in this study. Each RT mutant is designated by the single letter code of the substituting amino acid. The subunit mutated HIV-2 RT (with a wild-type p68 and a Q294R p54) is designated Q/R, and is shown next to the Q294R HIV-2 RT (in which both subunits possess the Arg294 mutation).

2.4 Strand-transfer assay

This assay was also described in detail previously [4, 16, 17]. A [32P]5′-end labeled 20nt synthetic DNA (5′-CACACAACAGACGGGCACAC-3′) was annealed to the same 267 nt RNA used in the RNase H assay (and serves in the strand-transfer assay as the primary RNA template). A second 144 nt RNA (synthesized from the pG89 plasmid) was used as the secondary or the acceptor RNA template. (See scheme in Fig. 2 A). The products of the combined DNA polymerase/RNase H reactions (performed in the presence of RTs and all four dNTPs) were resolved on denaturating polyacrylamide gels and the pattern of the 5′-end labeled DNA products was analyzed after autoradiography.

figure image
The strand-transfer activity of the Gln294 HIV-2 RT mutants. (A) A schematic description of the strand-transfer assay of RT. The 32P end-labeled 20nt synthetic DNA (thick line) was annealed to the 267 nt RNA transcript (thin line) that served as the primary RNA template. This hybrid was mixed with 144 nt PG89 RNA transcript (dotted line), which serves as the acceptor RNA template. The lengths of the lines are not drawn to scale. The four steps in the strand transfer process were described in detail previously [17]. (B and C) Each reaction consists of 0.1 μg heterodimeric HIV-2 RT mutants that were incubated for 1 h at 37 °C. Reaction products were analyzed by urea-PAGE followed by autoradiography. We show here the autoradiograms of two gels of two independent experiments, conducted with the specified RT versions. The HIV-2 RT mutants were designated as described in Fig. 1. The strong-stop 139 nt DNA transcript that is the product of the template primary RNA is marked by S and the full-length 238 nt long transfer DNA product, generated after strand-transfer, is marked by T. The left lanes of each panel (marked as con) show the products of the control reactions with HIV-1 RT in the absence of the secondary (acceptor) RNA transcript, generating only strong-stop 139 nt products with no strand-transfer.

2.5 The DNA polymerase activity

This RT-associated activity was determined using poly(rA) n oligo(dT)12-18 as the template-primer and [3H]dTTP as the incorporated deoxynucleotide triphosphate, as was described in detail previously [15, 16].

3 Results

The research presented herein was designed to address the question how various residues, other than Gln and Pro at position 294 of HIV-2 RT can affect the RNase H activity of the enzyme. This was done by a “vertical” mutagenesis of Gln294 to various residues that represent the different amino acids groups, and by observing the effects of these mutations on the activities of the various RT mutants generated. In these HIV-2 RT mutants, Gln294 was modified in both the p68/p54 subunits to either of the follow residues: Asn, Ser, Cys, Ala, Glu, Arg, His, Met, Trp and Tyr. These residues were chosen, since they represent the various amino acid groups, as described in Table 1 . Additionally, we have tested the Q294P mutant, already studied by us previously, that represents the conversion of the Gln in WT HIV-2 RT to the Pro of WT HIV-1 RT [16]. All recombinant RT versions were expressed in bacteria as heterodimeric p68/p54 RT isoforms and the proteins underwent extensive purifications, yielding highly purified enzymes (data not shown).

Table Table 1. The relative enzymatic activities of the various Q294 HIV-2 RT mutants studied
HIV RT version Type of side-chain DNA polymerase activity (%) RNase H activity (fold) Strand-transfer activity (fold)
WT HIV-1 (Pro) Non-polar 100 6.5 ± 0.7 5 ± 0.5
WT HIV-2 (Gln) Uncharged polar with amide group 100 1 1
Pro Non-polar 113 ± 10 5.1 ± 0.4 3.9 ± 0.5
Asn Uncharged polar with amide group 136 ± 13 4.9 ± 1.4 4.4 ± 0.1
Ser Uncharged polar with hydroxyl group 119 ± 2 6.9 ± 0.3 4.2 ± 0.8
Cys Thiol group 134 ± 9 9.2 ± 1 4.4 ± 0.5
Ala Small non-polar 112 ± 8 6.6 ± 0.2 2.9 ± 0.2
Glu Acidic 111 ± 3 4.0 ±0.2 4.1 ± 0.3
Arg Basic 135 ± 12 4.8 ± 0.6 3.4 ± 0.5
Q (p68)/R (p54) Basic (only in only p54) 106 ± 18 4.2 ± 0.3 3.7 ± 0.1
His Imidazole group 116 ± 1 4.2 ± 1.2 3.5 ± 0.4
Met Non-polar aliphatic with thiol-ether group 104 ± 5 4.3 ± 0.2 3.5 ± 0.6
Trp Aromatic with indole ring 113 ± 2 5.1 ± 0.5 4.5 ± 0.2
Tyr Aromatic with hydroxyl group 103 ± 4 4.8 ± 0.1 4.0 ± 0.1

The specific RNase H and resulting strand-transfer activities were calculated relative to the activity of wild-type (WT) HIV-2 RT and are expressed as the factor by which the activity was increased. The initial RNase H activities were based on the quantitative evaluation of the reaction products, generated after 15-second incubations, as detected by the PAGE analyses (shown in Fig. 1B). After background subtraction, the activity of every RT version was calculated as follows: (A + B)/(A + B + C), where A, is the amount of 47 nt RNA products, B is the amount of the 38 nt products and C is the amount of the uncleaved 267 nt RNA substrate. Then, the activity of each RT version was calculated relative to that of WT HIV-2 RT. Each presented figure is the average (+/− range) of the data shown in Fig. 1B and those obtained from two other independent experiments (data not shown). The strand-transfer activities were calculated from the generated full length 238 nt DNA product, as shown in Fig. 2B and C. After background subtraction, the strand-transfer activity of every RT version was calculated as follows: (T)/(T + S), where T is the 238 nt strand-transfer product and S is the strong stop 139 nt product (See Fig. 2B and C). Here again, the activity of each RT version was calculated relative to that of WT HIV-2 RT, and each presented figure is the average (+/− range) of the data shown in Fig. 2B and C and two additional fully-independent experiments (data not shown). In both RNase H and strand-transfer assays the autoradiography films were scanned and the activity levels were calculated using the densitometric software TINA (version 2.07d; Raytest Isotopenmessgeraete GmbH). The DNA polymerase activity of the all RT versions studied was also calculated relative to wild-type HIV-2 RT and is expressed as percentages. In this assay, we have used 5 ng of each purified RT variant. All values presented in the table are the averages (+/− range) of three independent experiments.

3.1 RNase H activity of all Gln294 mutants of HIV-2 RT

All highly purified heterodimeric RT versions were assayed for their RNase H activities, using the method described in detail previously [4, 16, 17] and briefly in Section 2 (see also Fig. 1A). Fig. 1B shows the pattern of the RNA cleavage in the RNA–DNA hybrids by the various mutant HIV-2 RTs, in comparison to the WT HIV-1 and HIV-2 RTs. We have already shown that the patterns generated by both WT HIV-1 and HIV-2 RTs, as well as by all previous chimeric and mutant RTs tested, were similar [4, 16, 17]. Hence, it is not surprising that the same is also true for all newly generated mutants of HIV-2 RT. Since the first cleavage generates a 47 nt long product, the RNA is cut 17 nt away from the 3′-end of the DNA primer. Consequently, in thereby-employed polymerase-dependent RNase H assay the molecular distance between the DNA polymerase and the RNase H active sites of the RT equals to about 17 nt. The secondary cleavage generates 38 nt-long RNA fragments (the −8 nt cleavage). The extent of RNA cleavage was quantified and the initial velocity of RNA cleavage (calculated after 15 s of incubation) is summarized in Table 1. The presented values are the averages of three independently-performed experiments. All mutations incorporated into the HIV-2 RTs have led to a surprising increase in the RNase H specific activity rather than losing any activity. Generally, all mutations induced the RNase H specific activity to a level comparable to that typical to WT HIV-1 RT and the related Q294P mutant HIV-2 RT. Interestingly, even the substitution of Gln294 to Asn294, a modification that is considered the most conservative one (since both residues are uncharged derivatives of acidic residues, containing a terminal carboxamide in place of a carboxylic acid), has led to an increase in the specific RNase H activity. Therefore, it seems that the only residue at position 294 that suppresses the RNase H activity of HIV-2 RT is indeed the Gln found in the WT enzyme. Nonetheless, none of the 11 mutations incorporated at position 294 of HIV-2 RT has any significant effect on the noticeable level of the DNA polymerase activity of the enzyme. These results strongly suggest that, despite its significant effect on the RT-associated RNase H activity, Q294 appears to be dispensable for the second RT-associated activity, the DNA polymerase.

3.2 The strand-transfer activity of the mutant RTs

During the process of reverse transcription of the retroviral RNA genome into double stranded DNA, there are two template switches or strand-transfers [1]. Both the DNA polymerase and RNase H functions of RT are simultaneously involved in this course of action. The data presented here show (as shown before for WT HIV-1 and HIV-2 RTs, chimeric HIV-1/HIV-2 RT versions and the Q294P mutant of HIV-2 RT [16, 17]) that all new HIV-2 RT mutants possess similar DNA polymerase activities (Table 1), as we believe that the minor differences observed reflect insignificant experimental fluctuations. This confirms that Gln294 is involved primarily in the RNase H RT function. With equal DNA polymerase activities, differences in strand-transfer would be determined for the most part by the differences in the RNase H activity. Therefore, we have assessed the effects of the specific RNase H activity of the mutants of HIV-2 RT on the levels of the RT-associated strand-transfer function.

In the specific in vitro strand-transfer assay system employed, the synthesized full-length 238 nt DNA molecules represent the products of strand-transfer activity (see Fig. 2A). The results indicate that the strand-transfer activity of all HIV-2 RT Gln294 mutants tested was indeed higher than that of the WT RT (Fig. 2B and C). These results were also quantified, using three independently-performed experiments, and the results are shown in Table 1, and as expected, there is a good correlation between the strand-transfer activity and the RNase H activity of all RT versions tested (Table 1).

3.3 Subunit-directed mutagenesis of HIV-2 RT Gln294

We have already shown that the involvement of Gln294 in affecting the RNase H activity of HIV-2 RT is primarily through the p54 small RT subunit [16]. As we have shown in the present study that all modifications performed in both RT subunits have led to a substantial increase in the RNase H activity, it was of interest to test whether the mutations affect the activity when located in the small RT subunit. To this aim, we have generated by subunit-directed mutagenesis an RT version with the WT sequence in the p68 large subunit and the Q294R mutation in the p54 smaller subunit (designated Q/R), and the DNA polymerase, RNase H and strand-transfer functions of the mutant were assayed (1, 2). As could be expected, the data show that mutating just the p54 subunit was enough to increase the specific RNase H activity by about four-fold relative to the WT HIV-2 RT a factor that is quite similar to the one obtained with the Q294R HIV-2 RT mutant, in which both subunits carry the same mutation.

4 Discussion

The results presented here further illuminate the critical role of Gln294 in determining the level of the HIV-2 RT-associated RNase H activity. Elucidation of the mechanism by which Gln294 affects the specific RNase H activity (and, consequently, strand-transfer) requires structural insights. However, no experimentally-derived structures of both HIV-1 and HIV-2 RTs with full-length substrate molecules (extending past the RNase H active site towards the Gln294 position) exist. In contrast, in the 1HYS structure [22] of HIV-1 RT, the bound RNA–DNA hybrid appears to veer away from the position of Pro294 (the HIV-1 RT homologous residue), which would seem to preclude any involvement in the RNA degradation mechanism. It has been proposed that HIV-RT can undergo structural changes upon substrate binding and during the different stages of its reactions [23]. Indeed, superposition of the apo-HIV-1 RT structure (1JKH) with the 1HYS structure shows that the position of Pro294 moves about 5 Å closer to the RNase H domain (data not shown). However, whether this would bring either the substrate or the RNase H active site into contact with the Gln294 cannot be determined. We can, nonetheless, use structural insights, obtained from other crystallographic studies of the RT, RNase H and other enzymes, to model such possible interactions.

The human RNase H1 catalytic domain was recently crystallized with different lengths of the RNA–DNA heroduplex substrate molecules [24]. These authors utilized this structure to explore the function of the HIV-1 RT by superimposing the human enzyme structure onto the RNase H domain of the 1HYS HIV-1 RT structure. Using conserved structural elements, it could be seen that six nucleotides straddle the active site, as the substrate passes through a charged cleft between the RNase H domain and the p51 thumb domain. Slight modifications in the substrate are then required to avoid clashes. These modifications force the RNA–DNA hybrid to pass directly over the position of Pro294 (in HIV-1 RT), which is identical to the position of the Gln294 residue in HIV-2 RT. In light of this proposal, the presence of a polar residue (such as Gln) instead of the Pro of HIV-1 RT could interact with the RNA strand and, consequently, inhibit the progress of the enzyme.

The above argument does not, however, explain why a Gln at position 294 would behave in a different manner than all other residues. If we look closely at the values in Table 1, we can separate the different residues into two subclasses. The three smallest residues (Ala, Cys and Ser) have RNase H activities equal or better than that of HIV-1 RT. The second group is broader with respect to their chemical identity and the effects on the RNase H activity. This group includes the slightly larger hydrophobic (Pro and Met), aromatic (Trp and Tyr) or polar/charged (Glu, Arg, His and Asn) residues, which have all RNase H activity levels that are reduced by between 20% and 40% compared to that of WT HIV-1 RT. Both groups of mutations have only a small reduction (60% and 90%) in strand-transfer rates compared to HIV-1 RT. From analyzing these mutated enzymes, we can conclude that the inhibitory effect of Gln (which has an ∼80% reduction in both RNase H activity and strand-transfer activity when compared to HIV-1 RT) must be due primarily to the combination of its length and polarity (but lack of charge) with its potential to form hydrogen bonds with the substrate (see below).

However, there may be an additional complexity in the comparison between the two HIV RTs. Using the coordinates of the HIV-2 RT crystal structure (1MU2), the electrostatic potential of the surface of the RNase H domain and the p54 thumb domain was calculated using the PyMOL algorithm [DeLano, W.L. The PyMOL Molecular Graphics System (2002) on World Wide Web http://www.pymol.org]. The calculated potential is depicted in Fig. 3 , overlaying the protein structure. In order to identify the position of the bound RNA–DNA hybrid, the 1HYS structure was superimposed onto the 1MU2 structure (for clarity the 1HYS protein subunits are not shown). The p54 Gln294 residue (orange spheres) is situated between a very positively charged cleft and a negatively charged patch. A yellow arrow shows the probable direction of the RNA–DNA hybrid as it exits the RNase H active site [24], and it is certainly feasible that this residue interacts with the RNA–DNA hybrid. In HIV-1 RT, the protein surface behind Pro294 is positive and not negative; suggesting that the direction of the movement of this hybrid in the two related RTs may be different. While most reduction in RNase H activity of HIV-2 RT can be associated with the presence of Gln (as opposed to the Pro residue in HIV-1 RT), the addition of a change in the surface electrostatic potential in proximity to Gln294 may also affect the enzymes’ activity. It is also possible that if a potentially-charged residue is introduced at the HIV-2 RT 294 site (Glu, Asp, Lys or Arg), the existence of the negative potential (behind) and positive potential (in front) may affect the ability to form hydrogen bonds with the substrate. Thus, for chemical and spatial reasons, it is likely that only the relatively long and uncharged Gln can inhibit movement of the substrate. A specific interaction between a Gln residues and nucleic acids has been identified in the cases of a U1hpII small RNA [25]. In addition, in the case of elongation factor Tu [26], it was shown that a Gln can decrease the rate of polymerization of poly U, with respect to the same enzyme with a smaller hydrophobic residue (Leu).

figure image
Molecular characteristics of p54 Gln294 in the structure of HIV-2 reverse transcriptase. The crystal structure of HIV-2 RT (PDB code 1MU2) was superimposed over the HIV-1 RT crystal structure that includes the RNA–DNA hybrid (PDB code 1HYS). For clarity, only, the two HIV-2 RT protein subunits (from 1MU2, p68 in green and p54 in black) and the hybrid (from 1HYS, RNA in brown and DNA in magenta) are depicted. The p54 Gln294 residue is shown in orange spheres. RH, denotes the RNase H domain. The protein molecules are overlaid with semi-transparent calculated electrostatic potential (blue-positive, red-negative). The positively-charged cleft, which may serve as the pathway by which the nucleic acids hybrid exits the enzyme, is identified by the yellow arrow. A negatively-charged patch seen behind Gln294 is positive in HIV-1 (see text for further discussion).

In light of the results presented here, it was of interest to examine in all available published sequences of HIV-2 RT variants how conserved is Gln294 (Table 2 derived from http://www.hiv.lanl.gov). All 21 sequenced isolates of HIV-2 RT have a Gln294; hence, this unique residue is highly conserved. A further support to this high conservation level emerges from the fact that SIVsmm, which is the SIV strain mostly related to HIV-2, harbors also Gln294 at all the sequenced isolates analyzed. This information strongly suggests that Gln294 is absolutely required for the optimal viability of HIV-2 and SIVsmm. This may also imply that the higher level of RNase H activity, similar to that exhibited by HIV-1 RT, may be disadvantageous in the context of the life cycle of HIV-2, and the lower RNase H activity is sufficient for its proliferation. As to HIV-1 RT, an inspection of 914 isolates of HIV-1 pol gene reveals that Pro294 is conserved in an overwhelming majority of 745 isolates (Table 2); thus, implying the importance of this residue. However, in a relatively large number of isolates, 130, there is a Thr294 and 14 isolates possess a Ser294. Since both Thr and Ser are uncharged polar residues with an OH group, the relatively high incidence of both these mutations (altogether 15.7%) is exceptional. Additionally, in 0.7% of the HIV-1 isolates, there is an Ala294, in 0.4% an Ile, and in lower incidences there are- Leu, Asn, Val, or a Glu at position 294. Interestingly, in 0.9% of HIV-1 RT variants there is a Gln294. We have found that there is substantial reduction in the RNase H activity of the recombinantly expressed P294Q mutant of HIV-1 RT [16]. Therefore, it is likely that a combination of other RT mutations in these eight HIV-1 isolates may compensate for the potentially-low RNase H activity of these P294Q mutants of HIV-1 RT (by increasing the enzyme's activity). The relatively high incidence of the P294T and P294S natural mutants should also indicate potentially interesting RT properties. All in all, we plan to study soon these issues and related ones.

Table Table 2. The frequencies of the residues at position 294 in the RTs of various variants of HIV-2, HIV-1 and SIVsmm
RT Q294 (%) No. of isolates
HIV-2 Q 100 21
SIVsmm Q 100 45
HIV-1 P 81.5 745
T 14.2 130
S 1.5 14
Q 0.9 8
A 0.7 6
I 0.4 4
L 0.2 2
N 0.2 2
V 0.2 2
E 0.1 1
Total 100 914

The presented figures were calculated from the data available at http://www.hiv.lanl.gov.