Molecular Impact of the M184V Mutation in Human Immunodeficiency Virus Type 1 Reverse Transcriptase

The generation of drug-resistant variants of human immunodeficiency virus (HIV) type 1 (HIV-1) in vivo as well as in vitro is a consequence of the error proneness of the HIV-1 reverse transcriptase (RT) enzyme. During viral replication, RT copies the single-stranded RNA genome into double-stranded DNA. Due to the lack of 3′ exonuclease proofreading activity, the incorporation of missense nucleotides occurs at a relatively high frequency. The misincorporation rate of the RT enzyme ranges from 10⁻⁴ to 10⁻⁵, depending on the nature of the template and the source of the RT (4, 7, 31, 39, 54). During in vivo and in vitro selections, drug-resistant variants emerge; competition for replication then results in outgrowth of the fittest variant.

The M184V mutation in HIV-1 RT is associated with high-level resistance to the antiviral drug 2′,3′-dideoxy-3′-thiacytidine (3TC) and has been extensively studied from the clinical, biological, and enzymatic perspectives. In fact, the first report of a change from a methionine to valine at residue 184 (M184V) in HIV was in regard to selection for resistance in tissue culture to didanosine (ddI) (23). Later, this mutation was shown to be responsible for both high-level resistance to 3TC (6, 17, 18, 22, 57, 63, 69) and low-level resistance to almost all the molecules that act as nucleoside analog RT inhibitors (NRTIs) (Table 1) (25, 53, 64). The appearance of the M184V substitution was found to be transiently preceded by another mutation, M184I, that also confers high-level resistance (about 1,000-fold) to 3TC (6) and occurs in patients treated with 3TC (58). Longitudinal sampling revealed a transient appearance of the M184I variant, which subsequently disappeared from the viral population due to the outgrowth of the M184V variant (3, 37). Eventual outgrowth of the M184V variant at the expense of M184I during therapy is consistent with superior RT polymerase function (3, 8, 11) and a higher viral replication rate of the M184V variant in primary cells (3). Inspection of the nucleotide sequences of both 3TC-resistant variants indicates that M184V (GTG) originates from wild-type (WT) Met (ATG) and not from the initial M184I variant (ATA) (34). Both variants are generated from the WT ATG sequence by transitional substitutions (G to A for 184I and A to G for 184V). The reason that M184I appears before M184V is that the G→A substitution is the type of mutation that most commonly occurs during HIV-1 replication (13, 24).

TABLE 1.

Resistance profile of the M184V mutation in RT

Drug	Selection in:		Fold increase in IC50
	Culture	Patients
ddI	Yes	Rarely	2-5
ddC	Yes	Rarely	2-5
FTC	Yes	Yes	>100
ABC	Yes	Yes	2-5
3TC	Yes	Yes	>100

CHARACTERISTICS OF THE M184V MUTATION

The M184V substitution changes the geometry of the YMDD motif, which is located at positions 183 to 186 in RT. This motif is, in general, conserved among all retroviruses. Attempts to deliberately mutagenize the sequence at the M184 position, with substitution of other amino acids in place of methionine, led to viral attenuation or nonviability as well as greatly diminished enzymatic function (70). The M184V mutation in HIV-1 RT has been characterized in the following ways. (i) M184V is the only single mutation known to encode as much as 1,000-fold increased resistance to an NRTI, i.e., 3TC, as well as low-level resistance to other drugs. The increase in the 50% inhibitory concentration (IC₅₀) of 3TC can approach 1,000-fold in tissue culture evaluations. (ii) The mutation is observed very rapidly both in cell culture and in patients receiving treatment with 3TC, but not in those receiving ddI or zalcitabine (ddC) (6, 32, 58, 63). This is because the high rate of HIV mutagenesis means that all single point mutations in HIV-1 that are compatible with viral survival, in fact, preexist before patients are ever treated with antiviral drugs. Thus, the use of drugs such as 3TC in monotherapy can rapidly select mutated forms. (iii) The substitution is located in the catalytic domain of the RT enzyme that, as stated above, is conserved among almost all retroviral RTs (70). (iv) Recombinant RT containing M184V exhibits a 50-fold diminished sensitivity to 3TC 5′-triphosphate (3TC-TP) compared to that of WT RT but has K_m and V_max values for ddC-TP similar to those of the WT RT (15, 35, 49).

In regard to other drugs, one study of 255 M184V-containing clinical isolates showed that this mutation did not result in broad cross-resistance to NRTIs other than 3TC (43). However, tissue culture findings have shown that M184V confers low-level resistance to abacavir (ABC), ddI, and ddC (17, 23, 42, 61) (Table 1). In order to attain higher levels of resistance to the aforementioned drugs, other substitutions in RT are required, e.g., K65R, L74V, Y115F, and Q151M for resistance to ABC (42, 64). Higher-level resistance to ddI and ddC is often associated with M184V, L74V, and K65R substitutions in tissue culture; and interactions between M184V and mutations that confer resistance to ddI and ddC have recently been investigated (K. L. White, N. A. Margot, T. Wrin, C. J. Petropoulos, L. K. Naeger, and M. D. Miller, 2002 XIV Int. AIDS Conf., poster 3033, 2002).

The M184V mutation has also been reported to result in diminished RT processivity relative to that of the WT enzyme both in infected cells (3, 8) and in virions (59). Processivity is defined as the number of nucleoside incorporation events before the enzyme dissociates from its nucleic acid substrate. Recombinant RT containing M184V also possesses diminished processivity (3, 37). Structural studies have shown that processive synthesis by HIV RT involves interactions between the minor groove of the template-primer and the minor-groove binding track, which is a significant structural element comprising five amino acids, four of which (Q258, G262, W266 and Q269) are located in α helix H and one of which (I94) is located in β sheet 5b (5). This provides insight into the altered processivity of M184V RT, since this mutation results in an altered interaction with the DNA duplex in the minor groove and may indirectly affect translocation due to the minor-groove binding track, hence causing reduced processivity (59).

MOLECULAR MECHANISMS RESPONSIBLE FOR ALTERED INTERACTION OF M184V-CONTAINING RT WITH 3TC

3TC-TP is about 50 times less effective against RT with the M184V mutation than WT RT, yet little difference exists between these enzymes in regard to recognition of deoxynucleoside triphosphate (dNTP) substrates (15, 49). A 200-fold increased level of resistance to 3TC-TP has also been seen in an endogenous RT assay (49, 50). Although the level of resistance to 3TC-TP in cell-free assays is much lower than the 1,000-fold increase in IC₅₀s measured in tissue culture, the biochemical data clearly indicate diminished rates of incorporation of 3TC monophosphate (3TC-MP). However, the probability that 3TC-MP is incorporated by RT with the M184V mutation in vivo is not negligible. Biochemical studies have shown that the incorporation efficiency of 3TC-MP is as high as the rate of formation of frequently occurring G-T mismatches by the same mutant enzyme (16).

Even though patients harbored HIV strains with the M184V mutation, patients treated with 3TC maintained a relatively low plasma RNA copy number over protracted periods (58). In this context, it is probably important that the mutant enzyme may not be capable of removing incorporated 3TC-MP via pyrophosphorolysis or nucleotide-dependent primer unblocking reactions (20). Indeed, recent results have shown that 3TC can still retain moderate inhibitory effects against M184V-containing viruses in primary CD4⁺ cells, despite the high level of resistance to 3TC conferred by this substitution (51). This may be related to the fact that resting lymphocytes in vivo maintain relatively low dNTP pools, which can potentiate the effect of 3TC, since the ratio of inhibitor to substrate may be higher under these conditions than might be the case in dividing cells (18). Thus, the in vivo benefit of 3TC in the aftermath of the M184V mutation might be greater in nondividing lymphocytes than in activated lymphocytes. Even if the intracellular levels of 3TC-TP in the peripheral blood mononuclear cells of treated patients are lower than the levels of either the diphosphates or the monophosphates (44), this may not rule out an effect of 3TC-TP as a chain terminator (51). Chain termination experiments with 3TC, both in tissue culture systems and in cell-free systems, showed that 3TC continues to function as a chain terminator, albeit at a greatly reduced efficiency, since a molecule of 3TC-TP, once incorporated, might be displaced only with relative difficulty (51).

Structural studies have also shown that M184V can influence both the dNTP and the primer terminus (28). The methionine side chain contacts the sugar ring and base of the 3′ nucleotide in the primer, but introduction of a β-branched side chain (isoleucine in the case of M184I or valine in the case of M184V) also creates a contact with the dNTP sugar ring (56). Modeling of the correct (−) enantiomer of 3TC into the structure shows that interference with an isoleucine or valine at position 184 is enhanced (with respect to that with a conventional dNTP) by the configuration of the oxathiolane ring, accounting for the strong effects of these mutations on 3TC inhibition (28). Taken together, the biochemical and structural data provide insight into the ability of the M184V mutation to confer high-level resistance to 3TC.

INTERACTION OF THE M184V SUBSTITUTION IN RT WITH OTHER DRUGS AND RESISTANCE- CONFERRING MUTATIONS

Sometimes, the mutations that are responsible for resistance to one drug can cause hypersusceptibility to a different compound or resensitize strains that are resistant to that drug. Such is the case for the M184V mutation and 3′-azido-3′-deoxythymidine (ZDV) (6, 26, 37, 60, 62, 63, 69). The presence of M184V during highly active antiretroviral therapy (HAART) with various combinations of drugs or in vitro during drug resistance selection experiments in cell culture is associated with the reversal of resistance to certain drugs, e.g., ZDV, d4T, and tenofovir (TDF). Moreover, M184V is known to have antagonistic and suppressor effects for a variety of NRTI mutations, particularly those that are responsible for ZDV resistance (40, 47). The mechanism of resistance to ZDV is increased rates of removal of the chain terminator in the case of ZDV-resistant mutant enzymes (2, 41), while that of 3TC-resistant mutant enzymes is likely to be discrimination between the natural substrate and the chain terminator (35). These observations suggest that patterns of resistance to ZDV and 3TC show a certain degree of incompatibility (19), which helps to explain why the M184V mutation can resensitize viruses containing ZDV resistance-associated mutations to the latter drug, at least on a transient basis (37).

In vivo hypersusceptibility to NNRTIs, specifically efavirenz, has also been associated with the M184V substitution in a background of ZDV resistance mutations (60). However, due to the small size of the cohort, further studies are needed in vivo as well as in vitro to determine the contribution of individual RT mutations to this observed hypersusceptibility.

The M184V mutation is also known to affect the phenotypic expression of other drug resistance mutations (21, 37, 42, 69) and render virus more susceptible to inhibition by the nucleotide analogs adefovir (phosphonylmethoxyethyl adenine) (42; J. M. Cherrington, A. S. Mulato, P. D. Lamy, N. A. Margot, K. E. Anton, and M. D. Miller, Abstr. 2nd Int. Workshop HIV Drug Resist. Treatment Strategies, abstr. 4, 1998) and TDF (phosphonylmethoxypropyl adenine) (69; G. Skowron, J. Nadler, M. Thompson, and M. D. Miller, 8th Eur. Conf. Clin. Aspects Treatment HIV Infect., poster 226, 2001; White et al., XIV Int. AIDS Conf., 2002). For example, although the K65R mutation in RT can confer resistance to TDF (69), the combination of K65R and M184V increased the level of susceptibility to this drug over the diminished levels that resulted from K65R alone (White et al., XIV Int. AIDS Conf., 2002). RT with the K65R substitution also decreased the level of binding of TDF diphosphate as well as decreased the level of processivity, which adds to the defect associated with M184V when both mutations are combined in the same RT enzyme.

Previous studies demonstrated that WT viruses and viruses with the M184V mutation that had been passaged in the presence of monoclonal antibodies (MAbs) (termed MAb 447-52D) yielded neutralization-resistant viruses after times ranging between 15 and 22 days and 25 and 32 days, respectively; thus, a delay occurred in the case of the M184V variant. For virus with the M184V mutation, the escape mutation was located outside of the known GPGR epitope, which is known to confer resistance to MAb 447-52D (29). A slower emergence of epitope variants has also been observed in patients treated with 3TC (67). The slower emergence of such escape variants (29) may be due to increased RT fidelity, as assessed by deoxynucleoside misincorporation and misinsertion experiments (27, 48, 68), as well as to the low level of replication fitness associated with M184V.

Other studies have shown that M184V rapidly disappears under conditions of 3TC discontinuation in the clinic (55) and ZDV pressure (5 to 10 weeks) and that M184V may disappear within 9 to 20 weeks in the absence of 3TC or other drug pressure (14). Indeed, it has been shown that deselection of the M184V variant to the WT occurred in the presence of ZDV alone or ZDV plus low levels of 3TC, but not in the presence of ZDV and a higher yet physiologically attainable concentration of 3TC (0.25 μM or greater) (14). Overall, these data show the nonstability of the M184V mutation under ZDV pressure and in the absence of 3TC. Taken together, these data show that 3TC pressure is necessary to maintain the M184V mutation. Furthermore, relatively high concentrations of 3TC may be needed to maintain the M184V mutation in the presence of ZDV.

PRIMER NUCLEOSIDE UNBLOCKING AND RESENSITIZATION TO ZDV

As stated earlier, the M184V mutation increases viral sensitivity to ZDV by 5- to 10-fold (6, 37, 62, 63). This effect is seen both in the presence and in the absence of mutations that confer resistance to ZDV, which suggests the possibility that M184V might interfere with ZDV 5′-monophosphate (ZDV-MP) excision from the end of the primer. Moreover, the effect of M184V on ZDV sensitivity is seen in the absence of any specific ZDV resistance mutations; this suggests that WT RT of HIV-1 might be able to excise ZDV-MP in vivo, a concept supported by in vitro data (9). This excision mechanism involves pyrophosphorolysis; however, there has been controversy about the nature of the pyrophosphate (PP_i) donor (2). Although it was originally suggested that the in vivo PP_i donor might be PP_i itself, more recent work has suggested that ATP is likely to be the PP_i donor (9, 41).

In addition, the M184V mutation may reduce the rate of excision of ZDV-MP (7, 20), although one group reported that M184V does not impair ZDV-MP excision (46). It has also been shown that M184V reduces the rate of ZDV-MP excision when M184V is combined with the ZDV resistance mutations M41L, D67N, K70R, and T215Y but not with the M41L or T215Y mutation (38). More recently, use of a long template for the excision reaction suggested that M184V may alter the position of the template-primer and the RT active site (10); this repositioning is thought to move the end of the primer away from the ATP PP_i donor and render excision less likely to occur (10); however, experimental data may be needed in order to confirm these observations, which may support the concept that diminution of ZDV-MP excision may reduce the level of resistance to ZDV in the presence of the M184V mutation.

To assess the effect of M184V on the incidence of thymidine analog mutations, comparative studies involving 3TC-naive versus 3TC-experienced patients infected with viruses that possessed M184V were undertaken. By multivariate analysis, it was determined that M184V was associated with a lower incidence of thymidine analog mutations as well as lower levels of resistance to both ZDV and d4T (1). As the majority of the patients in that study were receiving dual regimens, it would be premature to generalize these findings to patients receiving first-line HAART. However, these results suggest that if patients receiving HAART are maintained on a failing regimen, they might have lower levels of resistance to thymidine analogs with regimens that include 3TC than with regimens that exclude 3TC. Prospective studies are needed to clarify whether there is benefit to maintaining selective pressure on M184V after the patient is switched to a new regimen. It has also been shown that viruses harboring both K65R and M184V have decreased processivities and low replication capacities (White et al., XIV Int. AIDS Conf., 2002). Taken together, these data suggest a possible fitness barrier for viruses containing K65R in vivo. This is consistent with the low frequency of occurrence of this mutation in viruses from antiretroviral agent-experienced patients (White et al., XIV Int. AIDS Conf., 2002).

M184V DECREASES VIRAL FITNESS AND INITIATION OF REVERSE TRANSCRIPTION

M184V is also associated with decreased viral fitness in culture in comparison with the fitness of WT virus (3, 15, 37, 49, 71; K. Diallo et al., unpublished data). Despite the rapid selection of highly 3TC-resistant M184V-containing HIV-1 variants, sustained reductions in plasma viral RNA levels have been observed in 3TC-treated individuals (58, 65, 67). Until recently, it was thought that these reductions in viral RNA load were most likely explained by a reduction in viral replication capacity that was due to a low level of processivity (3, 59). However, the experimental conditions under which processivity is measured may not adequately reflect physiological conditions. Reverse transcription takes place within the viral capsid that contains RT in large excess over the copackaged RNA genome. As a consequence, dissociated enzymes are likely to be rapidly replaced by others that continue the polymerization process. Thus, the effects of a moderate diminution in enzymatic processivity may not contribute as much to an overall reduction in DNA synthesis as has been reported for intrinsic rate-limiting steps during reverse transcription. The events associated with initiation of RNA-primed synthesis of minus- and plus-strand viral DNA have recently been characterized as rate limiting; moreover, these steps can reduce rates of DNA synthesis by 2 to 3 orders of magnitude in the context of RT with the M184V mutation in comparison with that of WT RT. The initiation of minus-strand DNA synthesis is accompanied by frequent pausing and is, in general, a distributive process that involves frequent dissociation of the enzyme from the template (30, 36).

Using well-established methodologies, others have shown a delayed and reduced initiation of reverse transcription by RT with the M184V mutation, specifically in the context of viruses containing deletions in an A-rich loop located upstream of the primer binding site (71). In this system, deletion of the A-rich loop, when present in WT HIV-1, caused reduced replication kinetics and was associated with the emergence of revertants in long-term culture. These reversions were attributable to a broad variety of genetic alterations in the vicinity of the deletion, which helps to explain the structural and functional requirements for efficient tRNA^Lys3-primed initiation of reverse transcription. However, dually mutated viruses that contained both the A-rich loop deletion as well as the M184V mutation in RT were unable to revert even after 20 passages. Cell-free assays with an RNA template-primer system showed a reduced affinity of the system for WT RT, which was further compromised in the presence of the M184V mutation, thus explaining the low level of replication fitness observed in cell culture (71). Taken together, these data show that the joint presence of the A-rich loop deletion and the M184V mutation affected the initiation of reverse transcription in a complementary and possibly a synergistic fashion, resulting in a level of impairment of minus-strand DNA synthesis greater than that seen with either alteration on its own. Overall, the presence of the M184V mutation under conditions of stress can lead to strong impairment of the replication capacities of viruses harboring this substitution.

Using a similar approach, our laboratories have shown that the diminished release of pausing sites during initiation of both minus- and plus-strand DNA synthesis on the part of RTs containing either the M184V or the L74V substitution is an important factor that delays and decreases the overall process of reverse transcription (unpublished data). These data help to explain the low levels of replication fitness of viruses containing the L74V or the M184V mutation. Thus, the low levels of replication fitness of viruses harboring M184V are due not only to low levels of processivity but also to impaired initiation of reverse transcription.

RT enzymes that contained the L74V, Y115F, and M184V mutations did not show any selectivity advantage over those containing M184V alone and were severely impaired in their ability to remove a chain terminator; thus, there does not appear to be a kinetic basis for the increased levels of resistance of such viruses to antiviral drugs in cellular systems (52). The mechanism of resistance to ABC by the combination of the L74V, Y115F, and M184V mutations remains unclear; however, similarities may exist in regard to the mechanism seen with L74V and M184V, which involves low levels of viral replication capacity, due to the low levels of processivity, and delayed and decreased levels of synthesis of both minus- and plus-strand viral DNA. An HIV-1 double mutant with the K65R and M184V mutations also showed a reduced replication capacity in vitro, consistent with a potential fitness defect in vivo and the low prevalence of the K65R mutation among isolates from antiretroviral agent-experienced patients (White et al., XIV Int. AIDS Conf., 2002).

THE M184V MUTATION IN SIV RT

M184V is also rapidly selected in the case of simian immunodeficiency virus (SIV) and results in high-level resistance to 3TC (12, 66). In SIV, M184V also confers modest levels of resistance to ddI but not to ddC, while it may also confer increased sensitivity to d4T (12). In SIVmac239, which contains a deletion within a 97-nucleotide region of the untranslated or leader sequence, the presence of M184V led to decreased viral fitness and an impaired potential for compensatory mutagenesis, reversion, and replication competence both in cell lines and in primary cells (72).

In vivo studies with rhesus macaques have shown that M184V-containing SIVs were initially less fit (slower replication during the first week of infection and reversion in the absence of drug) but maintained the full potential for virulence (66). Although in vitro studies often provide useful information on viral replicative fitness, they cannot model complex interactions with the immune system and predict in vivo virulence (33). Thus, animal models are the best tools that give insights into fitness and virulence; clearly, reductions in the replicative fitness of SIV caused by M184V are not high enough either to alter virulence or to prevent compensation by other mechanisms (66).

More recently, in vitro studies with SIVmac239 have shown that maintenance of the M184V substitution in culture diminished the effect of K65R on phenotypic resistance to TDF (as is also observed with HIV [69]) (45). An in vivo study compared juvenile macaques infected with either SIVmac239 or SIVmac239-184V and showed that treatment of the former animals with TDF or the latter animals with both TDF and emtricitabine (FTC) resulted in viral loads lower than those observed in the absence of drug. This suggested that the M184V mutation may confer hypersusceptibility to TDF in vivo. In that study, FTC, which is an analog of 3TC, was used at concentrations below those that are active in vivo in order to maintain the M184V mutation. This series of experiments also showed both in vitro and in vivo that the M184V mutation in the RT of SIV may have been deselected in favor of the WT RT in the presence of both TDF and FTC, while M184V was stable when TDF was not administered. These findings are similar to earlier results that showed that the simultaneous presence of both ZDV and 3TC resulted in the loss of the M184V mutation in tissue culture (14). These findings may suggest therapeutic strategies that may select for the loss of a resistance-conferring mutation through the use of pressure by other drugs.

CONCLUSION

The emergence of drug-resistant variants of HIV remains of prime interest in regard to both HIV pathogenesis and chemotherapy. The existence of such variants has limited the number of options available for the successful treatment of AIDS. Furthermore, the spread of drug-resistant variants of HIV among populations is a serious public health problem. Interestingly, several of the mutations associated with HIV drug resistance also result in diminished replicative fitness. Of these mutations, the M184V substitution in RT has been the most extensively studied. This characteristic may result from the location of M184V within the conserved YMMD motif, which is close to the enzyme's catalytic site. This results in a negative impact on RT enzyme activity, hence impairing virus replication capacity.

This decreased replication capacity of M184V-carrying viruses may be of clinical benefit and has prompted some scientists and clinicians to suggest that maintenance of this mutation following its appearance may represent a good strategy. Indeed, the low replication fitness of viruses with M184V and the biochemical mechanisms associated therewith may also have consequences on antiviral immune responsiveness and in vivo virulence. For example, viruses that are less fit may be much slower than WT virus to evade antiviral immune responses. Clinical trials, however, are the best way to monitor any effect that is associated with the M184V mutation, and studies that directly test the clinical benefit of the M184V benefit hypothesis should be started. For example, one can envisage a simple strategy in which 3TC would be maintained as part of a new treatment regimen, as opposed to the use only of new drugs that will be directly active against viral replication after initial treatment failure. In contrast, 3TC and other drugs should never be deliberately used to select for this or other mutations, and all antiviral compounds should be used initially for their intended purpose, which is suppression of viral replication.

(The work performed by K. Diallo was in partial fulfillment of the Ph.D. degree, Faculty of Graduate Studies and Research, McGill University, Montreal, Quebec, Canada.)