The V loop contains a site for N glycosylation

The V3 loop contains a site for N-glycosylation at the 6th position which is conserved in most HIV-1 strains but the significance of this high degree of conservation has been poorly addressed. To our knowledge, it has only been shown that the N-glycan in the V3 of some T-tropic strains (HXB2, HXB10, LAI and BRU) may be dispensable for virus infectivity 4, 5, 6, 7, 8. But removal of this N-glycan enhanced the virus sensitivity to neutralizing antibodies, suggesting its role in masking the potential antigenicity of V3 peptide 5, 6, 7, 8. Here, we re-evaluated, by site-specific mutagenesis, the significance of this N-glycan of HIV-1 strains with various cell tropisms. Contrary to the above reports, our data show that removal of the N-glycan from V3 impaired efficient virus replication. This impairment appeared to be associated with a reduced level of CXCR-4-dependent fusion and virus entry. Therefore, our data suggest the importance of the highly conserved V3 N-glycan not only for the antigenicity of HIV-1, but also for viral infectivity and cellular host range. However, no such requirement of the glycan has been found for CCR-5-dependent fusion and virus entry characteristic of M-tropic strains. Thus, there appear to be considerable strain-specific differences in the structure-functional relationships in HIV-1-co-receptor interactions.
Materials and methods
Results
Discussion In the present study, we generated N-glycosylation mutants of HIV-1 strains which differ in their cellular host range and demonstrated that the removal of the N-glycan from V3 impaired replication capability of a T-tropic HIV-1. The same KU55933 caused no such impediment for an M-tropic strain. The specificity of co-receptor usage of each mutant Env was confirmed by a Vac-based fusion assay, which demonstrated that the mutation affected CXCR-4-dependent fusion but not CCR-5-dependent fusion. Co-receptor-specific requirement of the N-glycan was further supported by similar mutagenesis of a dual tropic strain, as removal of the glycan affected the use of CXCR-4 but not CCR-5. Previously, the N-glycan in V3 of HXB2, HXB10, LAI and BRU strains was reported to be dispensable for their infectivity to T cell lines, although detailed growth kinetics and experimental conditions were not described 4, 5, 6, 7, 8. The experiments appear to have been done by infection initiated with only a single virus dose that was not specified clearly. We have to note that differences between the wild-type and mutant viruses were not observed with higher input virus doses, suggesting that the effect of glycan removal would essentially be small in a single step replication, and become appreciable only after numerous rounds of replication. Even though small to a single replication cycle, the contribution of the glycan would be great enough to in vivo pathogenesis which involves persistence of highly active viral replications for a long period. It was previously noted that the N-glycosylation site in V3 is less conserved among T-tropic isolates than among M-tropic isolates [5]. This is apparently paradoxical if the N-glycan is important for CXCR-4- but not CCR-5-dependent entry as proposed in this study. However, inspection of V3 sequences revealed that the loss of the N-glycan is closely associated with a basic amino acid substitution at the 11th position. Furthermore, the basic amino acid substitution at the 11th position in NL43HNT partially but significantly restored the infectious capacity. Therefore, the N-glycan in V3 seems to be dispensable only in the particular context of the peptide backbone, and plays an important role for CXCR-4-dependent membrane fusion when the 11th position of V3 is occupied by a non-basic residue.