Mutations in HIV-1 gag cleavage sites and their association with protease mutations.

Abstract

Protease inhibitors severely inhibit the replication of HIV-1 wild-type virus, but may select drug-resistant variants. The initial mutations that confer resistance are observed in the protease gene, but there is a continued evolution both in the protease gene and in secondary loci, especially in the cleavage sites of gag and gag–pol polypeptides, which are substrates for the protease [1]. Mutations in the cleavage sites were first described in 1996 [2], and chiefly concerned the p7/p1 and p1/p6 cleavage sites [2,3]. In vitro, these mutations appeared simultaneously or after resistance mutations in the protease gene, and both seemed to follow a common evolutionary pathway [2,4,5]. In vivo, the correlation between cleavage sites and protease mutations was not observed so clearly [6–9]. In order to clarify the relationship between these two types of mutations, we analysed 62 samples from 57 patients in Marseille (France) hospitals, either not treated (22 samples) or treated (40 samples) with protease inhibitors at the time of the analysis. The data in Table 1 concerned those viruses for which a mutation in the cleavage sites or in the protease was observed.Table 1: Association of mutation in the cleavage sites with major resistance mutations in the protease. To sequence the cleavage sites coding region and the protease gene, viral RNA was purified from plasma, reverse transcribed and amplified by nested polymerase chain reaction. The polymerase chain reaction products were sequenced with the ABI dye terminator technology as previously described [10], and the sequences were aligned on the HXB2 reference using Sequence Navigator software (Applied Biosystems, Courtaboeuf, France). In six cases (patients 7, 18, 24, 34, 40, and 41), the sequences of the protease gene clustered with non-B subtypes, as demonstrated by phylogenetic analysis (Table 1). The analysis of major protease and cleavage site mutations of the 62 samples showed that 17 samples (29.8%) have an A–V mutation at codon 431 in the p7/p1 cleavage site. In 16 of these 17 samples (94%), mutation A431V was associated with mutation M46I/L in the protease gene. Moreover, only two protease sequences with the M46IL mutation did not display A431V, confirming the close association between both mutations. In one case (patient 16), the A431V mutation disappeared after switching the therapeutic regimen from stavudine, lamivudine and saquinavir to zidovudine, lamivudine and ritonavir. These data showing a close association between M46IL and A431V mutations are consistent with a previous characterization of the protease/cleavage sites co-evolution upon antiprotease treatment [3]. However, our data differed significantly from those of Bally et al.[9], who reported a weaker M46IL/A431V association in a recent case–control study (65% of A431V viruses and 12% of A431 viruses were M46IL; versus 94 and 4.4%, respectively, in our study). Moreover, Bally et al.[9] found a better correlation between mutation V82AFT and A431V than we did in our study (71% of A431V viruses and 22% of A431 viruses were V82AFT; versus 58.8 and 26.6%, respectively, in our study). For the P1/P6 cleavage site, the mutation pattern was more heterogeneous, without significant associations. Most of the mutations observed in this site have been described previously in naive patients [7,11,12]. The most frequent mutation in the P6/P1 cleavage site was P453L (11 cases). However, we could not confirm the preferential association between this mutation and I84A and L90M protease mutations, nor the mutual exclusion between P453L and V82AFT reported by Bally et al.[9]. Overall, the discrepancies observed between the available data in the literature and this study underscore the complexity of the compensatory mechanisms that may occur in the cleavage sites in response to the emergence of drug-resistance mutations in HIV-1 protease [1]. Except for the significant association between M46IL in the protease and A431V in the P7/P1 cleavage site, no general model could be drawn from the comparative study of protease and gag cleavage site sequences. This may be because of specific physico-chemical features of the protease–gag polypeptide complex (i.e. the enzyme–substrate complex), which may accomodate a wide variety of structural changes both in the enzyme and the substrate. In this respect, one should consider the possibility of long-range conformational changes that could propagate from a given cleavage site to the others as the result of single mutations. Such effects have previously been described for drug-resistant mutant HIV-1 reverse transcriptase [13]. Nathalie Kocha Nouara Yahia Jacques Fantinib Catherine Tamaleta