Search

Several functions have been assigned to the extensive glycosylation of HIV-1 envelope glycoprotein gp120, especially immune escape mechanisms, but the intramolecular interactions between gp120 and its carbohydrate complement are not well understood. To analyse this phenomenon we established a new microwell deglycosylation assay for determining N-linked glycan accessibility after binding of gp120-specific agents. Orientation-specific exposition of gp120 in ELISA microplates was achieved by catching with either anti-C5 antibody D7324 or anti-V3 antibody NEA-9205. We found that soluble CD4 inhibited the deglycosylation of gp120 only when gp120 was caught by D7324 and not by NEA9205. In contrast, antibodies from HIV-infected individuals inhibited the deglycosylation best when gp120 was caught by NEA9205. These results demonstrated that both the CD4-binding site and the epitopes recognised by antibodies from HIV-infected individuals have N-glycans in the close vicinity. However, the difference in gp120 orientation indicates that antibodies in HIV-infected individuals, at least partly, bind to epitopes different from the CD4-binding site. Finally, we determined the structural class of the glycan of one V1 glycosylation site of prototype HIV-1 LAI gp120, which remained unsolved from previous studies, and found that it belonged to the complex type of glycans.

DNA encoding HIV-1 env is a poorly efficient B-cell immunogen and one probable explanation is that the numerous gp120 N-linked glycans gp120 may interfere with B-cell epitope presentation. The N306 glycan in gp120 shields HIV-1 from neutralizing antibodies. A DNA immunogen lacking the N306 glycosylation signal (T308A) was constructed to determine whether this glycan affected the immune response. Mice were immunized intranasally twice with DNA containing either the wild type or the mutant env. Two additional groups were primed with wild type or mutant env and boosted with rgp160 protein, containing the complete set of N-linked glycans. Immunization with DNA alone resulted in priming of B-cell clones but was not sufficient to induce a complete antibody response. Animals primed with the N306 mutant and subsequently boosted with rgp160 protein displayed higher serum IgG-binding titers to gp120 than animals primed with wild type env DNA. The manipulation of the glycosylation sites of the env DNA strongly primes antibody responses (but non-neutralizing) as well as T-cell responses to the wild type strain gp160. However, priming with mutant plasmid did not result in higher neutralization titers to wild type or T308A-mutated virus than did the wild type plasmid. With the N306 mutant DNA we thus immunized a non-neutralization epitope, but obtained strong env-binding IgG after rgp160 boosting.

We have previously shown that an N-glycosylation site of N306 of HIV-1 gp120 is not necessary for the HIV-1 infectivity but protects HIV-1 from neutralising antibodies. In contrast Nakayama et al. [FEBS Lett. (1998) 426, 367-372], using a virus with an identical V3 region, suggested that elimination of this particular glycan reduced the ability of T-tropic HIV to bind to CXCR4 and hence its ability to infect T cell lines. We therefore re-examined the ability of a mutant virus, lacking the N306 glycan, to replicate in various types of cells and found no change in co-receptor usage for mutant virus. The ability of mutant virus to replicate or to induce syncytia in infected cells was similar to that of wild type virus. These results corroborate our original observation, confirming that the induced mutation in the N306 glycosylation site neither impairs nor improves the ability of mutant virus to replicate in permissive cells.

HIV-1 gp120 is heavily glycosylated containing 24 N-glycosylation sites, and this makes elucidation of the significance of glycans at individual glycosylation sites a difficult task. A procedure is described where a complex mixture of biologically radiolabelled glycans of gp120, derived from a relatively small number of virus-infected cells may be characterized by a combination of N-glycanase release, single lectin separation, and normal phase HPLC (NP-HPLC). The method was applied in analysis of three N-linked glycosylation sites essential for the in vivo priming of T-cells, specific for an epitope in their vicinity (Sjölander, S., Bolmstedt, A., Akerblom, 1996. Virology 215, 124-133.). The carbohydrate compositions of wild type gp120 and of mutant variants gp120 lacking one, two, or all of these three active N-linked glycans were analysed. Cells were infected with r-vaccinia virus expressing wild-type gp120 or mutated gp120, or were infected with HIV-1BRU (wild type) or mutant virus variants. HIV-1 glycoproteins were purified by immunosorbent affinity chromatography and released glycans were separated on lectins, then analysed with NP-HPLC. Our data showed that the structural composition of glycans occupying two of the three glycosylation sites was heterogeneous but the site located adjacent to the T-cell epitope was equipped with one large, high mannose-type structure (> 11 units) with the capacity to cover a substantial part of the gp120 surface.

Fucosylated N-linked glycans are important constituents of membrane glycoproteins, owing to their significance as biologically active ligands for several selectins and their role in modulating protein conformation of viral glycoproteins. The human immunodeficiency virus type 1 (HIV-1) glycoprotein contains more than 30 different glycan structures but so far fucose was found associated solely with the innermost GlcNAc of N-linked glycans. In the present report we determined whether fucose units also were linked to the distal GlcNAc via alpha(1-3) or alpha(1-4) linkages in N-linked glycans of gp 120. [3H]-fucose labelled gp 120 was subjected to endoglycosidase F digestion, releasing diantennary complex type N-linked glycans, but leaving the inner polypeptide-bound carbohydrates, GlcNAc and possibly associated fucose units, intact. Gel filtration of the digested material revealed that [3H]-fucose label was released from gp 120 by this treatment, indicating presence of peripheral fucose units. Furthermore, [3H]-focuse label was also released by treatment of the labelled gp 120 with an alpha-L-fucosidase specifically removing fucose in alpha(1-3) and alpha(1-4) linkages. Altogether the results indicated presence of fucose units linked to peripheral GlcNAc of gp 120 N-linked glycans. We have earlier shown that other peripheral carbohydrate determinants, i.e. beta(1-4)-galactose on N-linked glycans, maintain a correct antigenic conformation of gp 120. Using a coupled ELISA system, where changes in antigenic behaviour of a viral glycoprotein were correlated to stepwise elimination of peripheral monosaccharides from N-linked glycans, we found that treatment of gp 120 with a pan-specific alpha-fucosidase as well as an enzyme specific for alpha(1-3)- or alpha(1-4)-linked fucose disclosed a hidden linear epitope situated in the gp 120 C2 region. The effects of the general fucosidase on epitope exposure was more prominent than those obtained with the enzyme with narrow specificity, suggesting that peripheral and inner fucose units co-operate in the maintenance of gp 120 conformation.