Identification of a Simian Immunodeficiency Virus Reverse Transcriptase Variant with Enhanced Replicational Fidelity in the Late Stage of Viral Infection*

Genomic hypermutation of human and simian immunodeficiency viruses (HIV and SIV) enables these viruses to adapt and escape from various types of anti-viral selection by altering the molecular properties of viral gene products. In this study, we examined whether the biochemical and catalytic properties of SIV DNA polymerases (reverse transcriptases; RT) can change during the course of viral infection. For this test, we analyzed RTs obtained from two SIV clones, SIVMNE CL8 and SIVMNE 170. SIVMNE 170 was isolated during the late symptomatic phase of infection with the parental strain, SIVMNE CL8. We found these two RTs have identical DNA polymerase specific activities and kinetics with three different DNA and RNA templates. In addition, the processivity of these two SIV RT proteins were also similar. However, as demonstrated by a misincorporation assay, the SIVMNE 170 RT showed much higher fidelity than SIVMNE CL8. The fidelity difference between these two SIV RTs was also confirmed by a steady state kinetic fidelity assay. These findings suggest that the fidelity of lentiviral RTs may change during the course of viral infection, possibly in response to alterations of host anti-viral immune capability. In addition, our sequence analysis of these two RT genes proposes possible structural strategies that the virus may employ to alter RT fidelity.

Evolution of infectious parasitic organisms is driven by host defense mechanisms that generate various types of anti-parasite selective pressures during the course of infection. Microorganisms that undergo escape from host selection systems are equipped with their own defense devices that can antagonize the anti-parasite mechanisms of the infected hosts. In the case of lentiviruses such as human and simian immunodeficiency viruses (HIV and SIV), genomic hypervariability has been considered to be an evolutionary tool that allows the viruses to escape from highly developed host anti-viral immune responses (1)(2)(3).
Alterations in the biochemical and structural properties of virally encoded proteins have consistently been observed during the course of lentiviral infection (4,5). Genomic hypermutation of lentiviruses underlies these molecular and functional shifts of viral gene products. For example, hypermutation at the variable (V) regions of lentiviral envelope proteins enables the virus to alter Env-related virological properties (i.e. cell tropism, infectivity, and cytopathicity) (6,7). These molecular alterations of env genes allow for adaptation and selection of the viral population, leading to viral escape from the host anti-viral immune system. However, even with constant molecular diversification of viral envelope protein, it is obvious that fundamental biological functions of the envelope protein must remain intact (8). Sequence variation has also been observed in the pol gene, which encodes the viral DNA polymerase (reverse transcriptase; RT) 1 (9). It is also obvious that the basic enzymatic activities of RT, such as DNA polymerase and RNase H activities, should remain active enough to support viral replication. However, it is not clear whether other biochemical properties of lentiviral RTs such as replicational fidelity remain constant in the face of sequence variations during the course of viral infection. Lentiviral RTs are the most error prone DNA polymerases known to be involved in DNA replication (10,11), and RT infidelity is thought to be the major driving force behind the generation of new viral variants, which are essential for viral escape from host immune pressure (10,12). In fact the level of host immune capability gradually but progressively decreases during the course of lentiviral infection, ultimately declining to a level that results in clinically apparent immune deficiency (i.e. AIDS) and in overwhelming opportunistic infections. Therefore, it is a logical assumption that the viral capability to mutate, evolve, and escape may also vary, particularly as host anti-viral selective pressure becomes diminished during the late stages of viral infection. Therefore, if RT infidelity is a source of viral hypermutagenesis, it is possible that the level of RT fidelity may change when the virus grows under conditions of reduced or no anti-viral selective pressure (i.e. the symptomatic phase of infection). Even though several RT mutants with slightly increased fidelity have been isolated following the anti-viral chemotherapy (13,14), RT mutants with greatly enhanced fidelity have not been isolated from the natural lentiviral population. Therefore, whether RT fidelity changes during the course of viral infection remains unanswered.
In this study, we examined the enzymatic properties of SIV RTs obtained from two representative viruses (SIVMNE CL8 and SIVMNE 170) that were isolated from the same animal at two different stages of infection (15)(16)(17). SIVMNE 170 is a molecular clone of SIVMNE that was isolated from a pig-tailed macaque during the late symptomatic phase of infection with the parental SIVMNE CL8 virus (17,18). These two clones have been used as representative clones to investigate alterations in various viral phenotypes (i.e. co-receptor use, cytopathicity, and pathogenicity) during lentiviral infection (15)(16)(17)(18)(19).
In this report, we present evidence that SIVMNE 170 RT has much higher fidelity than the RT of initial SIVMNE CL8. This finding suggests that RT fidelity can change during the course of viral infection. Presumably, this fidelity change could be due to a viral response to alterations in host immune capability. Possible molecular and structural strategies that the virus may employ to alter its RT fidelity are also discussed.
Cloning of RT Genes from SIVMNE CL8 and 170 Viral DNAs-For the expression plasmids for SIVMNE CL8 and 170 RTs, RT genes were amplified by Pfu DNA polymerase (CLONTECH, San Francisco, CA) from pSIV MNECL8 and pSIV MNE170 by using SRT SD F forward (5Ј-TTTTTTAAGCTTGAAGGAGATATACATATGCCCATAGCTAAGGTA-GAGCC-3Ј; NdeI site in underline, start codon in bold) and SRT C-T R reverse primers (5Ј-TTTTTTGGATCCGTCGACTTAAACTTGTCTAAT-CCCTTG-3Ј; BamHI in underline, termination codon in bold). Amplified 1.6-kilobase products were cloned to the pHis/NdeI plasmid after digestion with NdeI and BamHI. Two independent clones of each SIV RT were analyzed for both sequencing and biochemical analysis.
Purification of SIVMNE RT Proteins-SIV RT proteins were purified using our purification protocol for HIV-1 RT as previously described (20). Expression of SIVMNE RTs was induced by addition of 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside to 1000 ml of log phase E. coli BL21 with the RT expression plasmids (pBK90 and pBK125), grown in 2 ϫ YT to an A 600 of 0.5. After incubation for 2 h, cells were harvested by centrifugation and the pellets were resuspended and frozen (Ϫ70°C) with 20 ml of 1 ϫ binding buffer and lysozyme (200 g/ml). All buffers and binding resin used in this work were purchased from Novagen. Frozen cells were thawed and lysed on ice for ϳ2 h. The lysed cells were centrifuged (27,000 ϫ g) and the supernatant solution was applied to a charged 10-ml His-Bind column (1 ϫ 10 cm). The resin was prepared by successive washes with deionized water (30 ml), 1 ϫ charge buffer (30 ml) and 1 ϫ binding buffer (45 ml). All chromatographic steps were carried out at 4°C at a flow rate of 20 ml/h. Following application of the crude supernatant solution, the column was washed with 1 ϫ binding buffer (40 ml) and mixture of 1 ϫ binding buffer and 1 ϫ wash buffer (7:3, 10 ml). RT proteins were eluted with 1 ϫ elute buffer (20 ml); 90% of the recovered RT protein was released from the resin in the first 8 ml. Fractions containing purified SIVMNE RT proteins were analyzed by electrophoresis in a 12% SDS-polyacrylamide gel. Fractions containing RT were dialyzed against 1 ϫ dialysis buffer (50 mM Tris-Cl, pH 7.5, 1 mM EDTA, 200 mM NaCl, 10% glycerol) for 16 h and 1 ϫ dialysis buffer with 1 mM dithiothreitol for 6 h. In this protocol the purity of SIVMNE was typically greater than 95%, estimated by visual inspection of Coomassie Blue-stained gels. RT proteins were purified from two independent clones of pBK90 or pBK125. A p66/p66 homodimer of SIVRT proteins was used in this study.
DNA Polymerase Assay-Three templates were used to measure both DNA dependent DNA polymerase and RNA dependent DNA polymerase activities of RT proteins. The heteropolymeric DNA template is a gapped salmon sperm DNA prepared as described (21). The homopolymeric RNA template-primer is poly(rA) (286 -428 nucleotides long, Amersham Pharmacia Biotech, Piscataway, NJ) annealed to a 20-base pair long oligo(dT) (Amersham Pharmacia Biotech; T:P, 1:4). The heteropolymeric RNA template-primer is a 3,569-nucleotide long MS2 phage (Roche Molecular Biochemicals) annealed to 20-mer DNA primer The protocol for this assay was previously described (22).
Extension of Mismatched Primers-To measure RTs capability to extend mismatched primer, we used two different mismatched primers annealed to either the RNA or DNA template used in the misincorporation assay. 16-mer G/T mismatched primer (5Ј-CGCGCCGAATTC-CCGT-3Ј: mismatch with underline) or 19-mer C/A mismatched primer (5Ј-CGCGCCGAATTCCCGCTAA-3Ј) was annealed to the 40-mer RNA and DNA templates (see above), respectively. For control matched primer, we used the 15-mer C primer (5Ј-CGCGCCGAATTCCCG-3Ј). Reaction conditions for these mismatched primers were identical to those used in the misincorporation assay described above. The same RT activities of SIVMNE RT proteins were used in this assay, as determined by the extension reaction with the matched C primer. Two RT concentrations (4ϫ and 1ϫ) of RT proteins and all dNTPs were used in this assay. The reactions were also analyzed by 14% denaturing gel.
Kinetic Measurements for Misinsertion Fidelity of Two SIVMNE RT Proteins-A modification of the steady-state kinetic assay of Boosalis et al. (23), was used to determine misinsertion frequencies. Reaction conditions were the same as those described in the misincorporation assay except for the primers, the concentrations of dNTPs. For RNA template reaction (1 st DNA strand synthesis), 32 P-labeled C primer (5Ј-CGCGC-CGAATTCCCG-3Ј) annealed to 40-mer RNA template (see above) was extended by the two SIVMNE RTs with increasing concentrations of correct (dCTP) or incorrect (TTP) nucleotide at 37°C (Fig. 3). The amounts of RTs and incubation time were adjusted to yield extension of 30 -50% of the total labeled primer at the highest dNTP concentrations. Products were resolved in 14% polyacrylamide-urea gel and quantitated by phosphor-imaging analysis (Cyclone, Packard, Meriden, CT). For the DNA template reaction (2 nd DNA strand synthesis), G primer (5Ј-CGCGCCGAATTCCC-3Ј) annealed to 40-mer DNA template (see above) was extended with correct (dGTP) or incorrect (dATP). The k cat and K m values were determined from Hanes-Woolf plots (24). The misinsertion fidelity, f ins ((k cat /K m for incorrect dNTP)/(k cat /K m for correct dNTP)) was calculated from the k cat and K m values.
Construction and Purification of SIVmac239 RT-A plasmid expressing SIVmac239 RT was constructed using SIVmac239 RT genes amplified from pSIVmac239 -5Ј using the SRT SD F and SRT C-T R primers that were used for construction of pBK90 and pBK125 (see above). SIVmac239 RT protein was purified by using the same protocol as one used for SIVMNE RT proteins (see above).
Processivity Assay-Reaction condition for measuring processivity requires a single round of primer extension (25,26). RT proteins were preincubated with 20 nM poly(rA) (average size ϭ 260 nucleotides, Amersham Pharmacia Biotech, Piscataway, NJ) annealed to the 32 Plabeled 20-mer oligo(dT) (Amersham Pharmacia Biotech, Piscataway, NJ) for 3 min at 20°C. The concentrations of wild type RTs and M184I RTs of SIVMNE CL8 and 170 were 10 and 25 nM, respectively. The extension reactions were initiated by adding the trap mixture containing dNTPs (0.5 mM with final concentration 5 mM MgCl 2 ), molar excess of cold poly(rA)/oligo(dT) (20 M) and heparin (10 g/20 l). The extension reactions were terminated with 4 l of stop solution after 3 min incubation at 37°C. Two control reactions were performed as described (25,26), confirming the single round of primer extension under this reaction condition. In the ϩTrap control reaction, the trap mixture was added during preincubation of RT and T/P, which prevents RT from extending the primer. In the ϪTrap reaction, the T/P preincubated with RT was extended by only dNTP in the absence of the trap molecules, which allows multiple round of primer extension. The terminated processivity reaction and control reactions were analyzed by 8% polyacrylamide-urea gel after 3 min heat inactivation.

DNA Polymerase Activities of Purified SIVMNE RT Proteins-
To test whether the DNA polymerase activity of lentiviral RT can change during the course of viral infection, we purified RT proteins from two SIV representative clones, SIVMNE CL8 (early asymptomatic phase) and SIVMNE 170 (late symptomatic phase) using a bacterial overexpression system. We determined the DNA polymerase specific activities of these two purified SIVMNE RT proteins with a heteropolymeric DNA template and two homo-and heteropolymeric RNA templates. As seen in Table I, these two SIV RT proteins showed very similar levels of DNA-and RNA-dependent DNA polymerase activities with both heterogeneous and homogeneous DNA and RNA templates. The similarity of DNA polymerase activity between these two SIV RT proteins was also confirmed by the RT assay, with a gel-based primer extension reaction (see below and Fig. 1, lanes containing all dNTPs), and steady state kinetic measurement (K m and k cat for correct dNTPs, dCTP, and dGTP; Fig. 3 and Table II). We purified RTs from two independent clones of each SIVMNE strain. Purified RT proteins from two independent clones of each SIVMNE also showed similar specific activities with both DNA and RNA templates. This result shows that the DNA polymerase catalytic activities of these two SIVMNE RTs, isolated in different stages of viral infection, are largely identical.
Misincorporation Assay with Matched Primer-Next, we examined the fidelity of SIVMNE CL8 and 170 RT proteins using the misincorporation assay. This is a primer extension assay that monitors both misinsertion and mismatch extension of a 17-mer primer annealed to a 40-mer DNA or RNA template with biased dNTP pools containing only three kinds of dNTPs. In this assay with only three kinds of dNTPs, the primer extension stops at one base before the template nucleotide for which the complementary dNTP was deleted (stop sites; see the sites with "*" in Fig. 1). Therefore, the higher efficiency of primer extension beyond the stop sites will reflect higher utilization of incorrect dNTPs and the lower fidelity of the RT protein assayed. We used concentrations of SIVMNE CL8 and 170 RTs that showed identical levels of primer extension with all dNTPs present. As shown in Fig. 1, A (RNA template) and b Gapped salmon sperm DNA prepared as described (20). c Poly(rA)/oligo(dT) 20 . d MS2 phage RNA annealed to 20-mer primer.
FIG. 1. Misincorporation assay with SIVMNE CL8 and 170 RT proteins with RNA and DNA templates. The 32 P-labeled 17-mer primer ("S": A primer) annealed to 40-mer RNA (A) or DNA (B) template was extended by the same concentrations (10 nM for RNA template reactions; 40 nM for DNA template reactions) of SIVMNE CL8 and 170 RT proteins at 37°C for 5 min. The extension reactions were performed in the presence of all 4, or only 3 complementary dNTPs (minus dATP and minus dCTP for RNA template reaction; minus dATP and minus TTP for DNA template reaction). The reactions were analyzed by 14% polyacrylamide-urea denaturing gel. The sequence of the extended part of the primer is shown. F, fully extended products; S, un-extended primer; "*", stop sites. B (DNA template), in reactions containing all 4 dNTPs, equivalent concentrations of SIVMNE CL8 and 170 RT proteins catalyzed approximately the same amounts of DNA synthesis and fully extended primers (50% in RNA template; 80% in DNA template). However, when incubated with mixtures of only 3 dNTPs (Fig. 1, A, ϪdATP and ϪTTP; B, ϪdGTP and -TTP), SIVMNE CL8 RT catalyzed substantial extension past the stop sites (*), indicating low fidelity (11,25,26). On the other hand, the SIVMNE 170 RT protein catalyzed much less synthesis and shorter extension beyond the stop sites than SIVMNE CL8 RT protein in reactions with 3 dNTPs, indicative of higher replicational fidelity on both RNA and DNA templates.
Extension of Mismatched Primers-Creation of one mutation during DNA synthesis consists of two sequential steps, 1) incorporation of the incorrect dNTP (misinsertion) and 2) extension of the mismatched primer generated by the misinsertion.
In the misincorporation assay shown in Fig. 1, extension of primer beyond the stop sites measured the capability of the RT proteins to misinsert and then to extend the mismatched primer. Next, we examined the capability of SIVMNE 170 and CL8 RT proteins only to extend the mismatched primer (2 nd step of mutation synthesis). In this study, we used two different mismatched primers (MP) annealed to either RNA or DNA template. One T/P contains a G/T mismatch at the 3Ј end of the primer annealed to the RNA template. The other T/P contains a C/A mismatch at the 3Ј end of the primer annealed to the DNA template. These C/T (1 st DNA strand synthesis) and C/A (2 nd DNA strand synthesis) mismatched T/Ps represent the replication intermediates that can be synthesized during the process of G to A mutation, which is the predominant lentiviral mutation (27). We tested whether SIVMNE 170 RT (high fidelity) has a reduced capability to extend these two mismatched  Table II). Arrow, un-extended primer. T/Ps for creating the G to A mutation. The primer extension reactions were performed with all dNTPs, which allowed the extension of the matched and mismatched primers to the end of the templates. As seen in Fig. 2A, in reactions with the matched primer, both SIVMNE 170 and CL8 RT proteins showed the same level of total primer extension (or remaining unextended primer), indicating that the total RT activities of these RT proteins added in this assay were the same. However, SIVMNE 170 RT protein showed a much lower level of primer extension from both the G/T (Fig. 2B) and C/A (Fig. 2C) mismatched primers than SIVMNE CL8 RT protein. These results suggest that the SIVMNE 170 RT has a reduced capability to extend the mismatched primer, compared with SIVMNE CL8 RT.
Kinetic Measurements for Misinsertion Fidelity of Two SIVMNE RTs-Next, we employed a steady state kinetic assay for measuring misinsertion fidelity (f ins ) SIVMNE CL8 and 170 RT proteins. The f ins value is calculated by two kinetic parameters, k cat and K m , which are indicative of the capability of DNA polymerases to incorporate dNTPs (correct or incorrect). Since the G to A mutation is the most dominant mutation in lentiviral mutagenesis (27), we measured misinsertion fidelity for the C (correct) to T (incorrect) mutation using an RNA template (1 st DNA strand synthesis: Fig. 2, Table I) and the G (correct) to A (incorrect) mutation using a DNA template (2 nd DNA strand synthesis: Table I). The k cat and K m values of two SIV RT proteins with correct versus incorrect nucleotides were measured by a gel-based analysis and Hanes-Woolf equation (24). As seen in Fig. 3 and Table II, both SIVMNE RT proteins showed very similar K m and k cat values with correct dCTP and dGTP, indicating kinetic similarity between these two RTs. However, SIVMNE 170 RT protein showed a much lower level of extension (k cat ) with incorrect TTP and dATP than SIVMNE CL8 RT. As shown in Table II, and consistent with the misincorporation assay data (Figs. 1 and 2), SIVMNE 170 RT shows 8 and 11 times higher misinsertion fidelity than SIVMNE CL8 RT in the G to A viral mutation events.
Effect of Met-184 Mutations on Fidelity of SIVMNE RT Proteins-3TC-resistant RT mutations, M184V and M184I, have been reported to be high fidelity RTs in HIV, even though the M184V mutant showed only a very slight increase (Ͻ2 times) (28) of fidelity, compared with M184I (ϳ4 times) (28). Treatment of SIV with 3TC also results in the development of M184 mutations (29). We tested whether M184I high fidelity mutations affect the fidelity of SIVMNE CL8 and 170 wild type RTs by using the misincorporation assay. As seen in Fig. 4B (minus dATP), the M184I mutant of SIVMNE CL8 showed a reduced level of misincorporation, compared with wild type SIVMNE CL8 RT, indicating that M184I enhances the fidelity of both SIV RT proteins. However, the M184V mutant of SIVMNE CL8 showed a wild type (high) level of misincorporation. The M184I mutation also exerted similar effects on both SIVMNE RTs in misincorporation assays with different deleted dNTPs (ϪdGTP, ϪTTP, and ϪdCTP; data not shown). These findings suggest that the fidelity of both SIVMNE CL8 and 170 RTs can be enhanced by the M184I mutation.
Misincorporation Assay with SIVMNE170 and SIVmac239 RT Proteins-We compared the fidelity of SIVMNE 170 RT and SIVmac239 RT by using the misincorporation assay. Two concentrations (4ϫ and 1ϫ) of RT proteins were used in this assay. As shown in Fig. 5, SIVMNE 170 RT showed less extension of the primer beyond the stop sites (Fig. 5B for ϪT), indicating that SIVMNE 170 RT has higher fidelity than SIVmac239. Similar differences between SIVMNE 170 and SIVmac239 RTs in primer extension assays were also observed in reactions with other kinds of biased dNTP pools (data not shown). This result shows that SIVmac239 RT, like SIVMNE CL8 RT, has low replicational fidelity.
Processivity Assay with SIVMNE RT Proteins-We examined the processivity of the two SIVMNE RTs by using poly(rA) (template)/oligo(dT) (primer). The processivity assay requires a single round of primer extension, which can be established by using trap molecules (heparin and molar excess of cold T/P). As shown in two trap controls (Fig. 6), primer extension was inhibited by trap molecules which were added at the beginning of the reaction (ϩTrap control), whereas primer extension was made in multiple rounds in the absence of the trap molecule, as shown in synthesis of long products (ϪTrap control). As shown in the processivity reaction (Proc) where the RT molecules pre-bound to T/P can extend the primer only once, both SIVMNE CL8 and 170 RTs showed identical processivity. It has been shown that M184I mutation reduces processivity of HIV-1 RT (30). As shown in Fig. 6, M184I mutants of both SIVMNE CL8 and 170 RTs also showed reduced processivity, suggesting that the processivity of both SIV RTs was affected by the M184I mutation.
Sequence Comparison of SIVMNE CL8 and 170 RT-We determined the sequence of the SIVMNE 170 RT gene between residues 1 and 396 (this corresponds to the DNA polymerase domain of RT). By sequence comparison of SIVMNE CL8 and 170 RT genes, we found that six residues in this region have sequence variations between these two SIVMNE RT genes (Fig. 7). Residues affected are 73, 148, 173, 211, 303, and 332 (Fig. 7). We also compared the sequences of these two SIVMNE RT genes with that of the SIVmac239 RT gene (31). As shown above, both SIVmac239 and SIVMNE CL8 RTs have low fidel- ity, compared with SIVMNE 170 RT. Among the 6 residues showing variations between SIVMNE CL8 and 170 RTs, four residues (148, 173, 211, and 303) of SIVMNE 170 RT also show variations from SIVmac239 RT. The location of these residues of SIVMNE RT was examined by using a structural model of HIV-1 RT solved by Huang et al. (32). Interestingly, none of the six residues, which differ between SIVMNE 170 RT and CL8 RT (or SIVmac239 RT), directly interact with RT substrate molecules (template, primer and incoming dNTPs). Furthermore, none of the six putative SIVMNE 170 fidelity mutations overlap with previously defined HIV-1 RT fidelity mutations (Fig. 7). DISCUSSION Genomic hypervariability of lentiviruses constantly shifts molecular and functional properties of viral gene products during the course of viral infection. For example, functional modifications of the lentiviral envelope gene result in temporal, but essential, alterations in viral infectivity, cellular tropism, cytopathicity, and in vivo pathogenesis (escape from host immune). In this study, we examined whether biochemical properties of lentiviral RTs, DNA polymerase activities, fidelity, and processivity, can also be altered during the course of viral infection. For this test, we have analyzed SIV RT proteins derived from two cloned SIV strains, SIVMNE CL8 and SIVMNE 170, that are representative clones for asymptomatic (early) and symptomatic (late) phases of viral infection, respectively (15)(16)(17)(18). SIVMNE 170 clone was obtained from the symptomatic (late) phase of infection of a macaque initially inoculated with SIVMNE CL8 (the parent strain). These two representative clones have been used previously to understand alterations in various viral phenotypes (i.e. co-receptor use, replication kinetics, cytopathicity, and pathogenicity) during lentiviral pathogenesis (15)(16)(17)(18)(19).
For biochemical analysis of these two SIV RTs, we purified these RT proteins from a bacterial overexpression system. DNA polymerase assays with three different DNA and RNA templates revealed that these two SIVMNE RTs contain almost identical DNA-and RNA-dependent DNA polymerase specific activities. These two proteins also showed identical processivity (Fig. 6). These data indicate that the DNA polymerase activities and processivity of RT may be kept unchanged over time, presumably in order to maintain a constant level of viral replication (33).
However, as shown in the assays that we undertook to analyze the fidelity of SIVMNE CL8 and 170 RT proteins, the RT of SIVMNE 170 has much higher fidelity than that of SIVMNE CL8. Both the steady state kinetic assay for misinsertion frequency and the mismatched primer extension assay revealed a fidelity difference between these two SIVMNE RT proteins. Thus, SIVMNE 170 seems to have enhanced fidelity in both steps (misinsertion and extension of mismatched primer) of mutation synthesis. Furthermore, as seen in the steady state fidelity kinetic assay, SIVMNE 170 RT has greatly altered Poly(rA) annealed to 32 P-labeled 20-mer oligo(dT) was extended by wild type (W) and M184I (M) proteins of SIVMNE CL8 and 170 RTs. In the processivity assay, T/P was first preincubated with RT proteins, and then the extension reactions were initiated by adding trap mixture containing dNTPs, poly(rA)/oligo(dT), and heparin. T/P was also extended by the preincubated RT and dNTP in the absence of the trap mixture, allowing multiple rounds of extension (ϪTrap). T/P preincubated with both RT and the trap mixture was extended by dNTP, showing a complete block of RT activity by the trap mixture (ϩTrap). These two controls show that there was only single round of primer extension in the processivity assay. All reactions were analyzed by 8% polyacrylamide denaturing gel electrophoresis. L, 100 base pair ladder. fidelity with respect to the generation of the predominant G to A mutation (27). Since the insertion fidelities of SIVMNE 170 RT for both RNA-and DNA-dependent DNA polymerase activities were higher than those of SIVMNE CL8 RT, it is likely that SIVMNE 170 may produce less mutations in both 1 st and 2 nd strand syntheses of viral replication.
3TC resistant RT mutations, M184V and M184I, were initially reported to be high fidelity HIV-1 RTs, even though later biochemical studies showed that the overall mutation rate of M184V is very similar to that of wild type. A recent in vivo study also showed that M184V mutant HIV-1 did not show an altered mutation rate, compared with wild type virus (34). In contrast to M184V, the overall mutation rate of the M184I mutation is slightly higher (ϳ3 times) than wild type. As shown in our misincorporation assay with SIVMNE CL8 wild type and its M184V mutant (Fig. 4), the SIVMNE CL8 M184V mutant also showed a wild type level of misincorporation. However, the M184I mutation was able to enhance fidelity of both SIVMNE CL8 and 170 RTs. Indeed, the M184I mutation can further elevate the already enhanced fidelity of SIVMNE 170 RT.
Our sequencing analysis revealed six residue sequence differences in the DNA polymerase domain (residues 1 to 396) between the SIVMNE 170 and CL8 RT. Four of these SIVMNE 170 RT residues (148, 173, 211, and 303) were also different from the corresponding sequences of SIVmac239 RT which is also a low fidelity RT. As examined by using HIV-1 RT structure, the four residues showing sequence variations between low fidelity RTs (SIVMNE CL8 and SIVmac239) and SIVMNE 170 RT do not provide direct interactions with RT substrates (i.e. incoming dNTPs and T/P). Residues 211 and 302 lie near the bottom of the palm domain and surface side of the thumb domain, respectively, which are distant from the DNA polymerase active site of RT. Furthermore, none of the six residues of SIVMNE 170 RT with sequence variations overlap with known HIV-1 RT fidelity mutations. Even though they do not interact with RT substrates directly, two residues of SIVMNE 170 RT, 148 and 173, are adjacent to several key RT residues that are known to be involved in not only RT activity, but also fidelity and drug resistance. Residue 148 is near the ␣E-␤5 loop region (Fig. 7) that has been found to be important to both fidelity and drug resistance of HIV-1 RT (26,35). The residue 173 locates at the outer portion of the ␤9-␤10 palm region containing the conserved YMDD catalytic site. Since these residues do not interact directly with RT substrates, it is possible that the mutations found in SIVMNE 170 RT may enhance enzyme fidelity by affecting the adjacent key residues involved in interaction with incoming dNTP and template.
The identification of a SIVMNE isolate with enhanced RT fidelity from the late, symptomatic phase of viral infection suggests that RT fidelity may respond to changes in host antiviral immune capability. It is not clear, however, whether there might be a selective advantage for high fidelity RT in the viral populations generated from the late-stage of viral infection. It is possible that viral population with high fidelity RT might be generated simply because RT fidelity and viral mutagenesis are not significant factors for viral survival any longer in this late-stage of infection with depleted host immune capability. Further studies will be necessary to explore this issue, and to determine the relationship between viral RT fidelity and host immune selective pressure.
Overall, the data presented in this report support the notion that RT fidelity can change during the course of viral infection. However, it still remains unclear whether RT infidelity, which can initiate error-prone viral replication, is a key element for viral escape from host immune selection. It is highly likely that the interplay between error-prone DNA synthesis and other virological factors, such as constant and massive viral replication and production (33), high degree of viral recombination (36), and infection of cell types with limited dNTP pools (37), may enhance the viral genomic complexity that is essential to viral evolution and escape from host antiviral selective pressures.