Evolutionary dynamics of the glycan shield of the human immunodeficiency virus envelope during natural infection and implications for exposure of the 2G12 epitope

Abstract

Elucidation of the kinetics of exposure of neutralizing epitopes on the envelope of human immunodeficiency virus type 1 (HIV-1) during the course of infection may provide key information about how HIV escapes the immune system or why its envelope is such a poor immunogen to induce broadly efficient neutralizing antibodies. We analyzed the kinetics of exposure of the epitopes corresponding to the broadly neutralizing human monoclonal antibodies immunoglobulin G1b12 (IgG1b12), 2G12, and 2F5 at the quasispecies level during infection. We studied the antigenicity and sequences of 94 full-length envelope clones present during primary infection and at least 4 years later in four HIV-1 clade B-infected patients. No or only minor exposure differences were observed for the 2F5 and IgG1b12 epitopes between the early and late clones. Conversely, the envelope glycoproteins of the HIV-1 quasispecies present during primary infection did not expose the 2G12 neutralizing epitope, unlike those present after several years in three of the four patients. Sequence analysis revealed major differences at potential N-linked glycosylation sites between early and late clones, particularly at positions known to be important for 2G12 binding. Our study, in natural mutants, confirms that the glycosylation sites N295, N332, and N392 are essential for 2G12 binding. This study demonstrates the relationship between the evolving "glycan shield " of HIV and the kinetics of exposure of the 2G12 epitope during the course of natural infection.

FIG. 1.

Antigenic profiles and sCD4 binding abilities of the recombinant envelope glycoproteins (gp120 and gp160) derived from the early and late clones for each patient. The four patients are presented from left to right. The binding indices for IgG1b12, 2G12, 2F5, and sCD4 are presented from top to bottom. Each frame shows the distribution of the binding index for both gp120 and gp160 of early clones (left panels) and late clones (right panels). For each distribution, the horizontal lines represent the 10th, 25th, 50th (median), 75th, and 90th percentiles.

FIG. 2.

Phylogenetic analysis of early and late full-length HIV-1 envelope gene sequences from the four patients. A distance scale is given for each neighbor-joining tree. Bootstrap values are expressed as percentages per 1,000 replicates, and only values above 80% are indicated on nodes. (A) Neighbor-joining tree including 94 full-length env sequences, 11 other clade B sequences obtained previously from proviral DNA isolated at the time of primary infection from our four patients (PIH133, PIH153, PIH159, and PIH309) and seven other patients (PIH120, PIH146, PIH155, PIH160, PIH373, PIH374, and PIH384) from the same cohort (accession numbers AF041125 to AF041135 [1]), and the two reference sequences HxB2 and MN (accession numbers AF033819 and M17449, respectively). The sequence of the env gene from a clade D virus (DGOB) was used as an outlier. Triangles correspond to clusters of early or late clones. (B to E) Details for patients 133 (B), 153 (C), 159 (D), and 309 (E). The different groups of late clones are indicated by circled numbers.

FIG. 2.

Phylogenetic analysis of early and late full-length HIV-1 envelope gene sequences from the four patients. A distance scale is given for each neighbor-joining tree. Bootstrap values are expressed as percentages per 1,000 replicates, and only values above 80% are indicated on nodes. (A) Neighbor-joining tree including 94 full-length env sequences, 11 other clade B sequences obtained previously from proviral DNA isolated at the time of primary infection from our four patients (PIH133, PIH153, PIH159, and PIH309) and seven other patients (PIH120, PIH146, PIH155, PIH160, PIH373, PIH374, and PIH384) from the same cohort (accession numbers AF041125 to AF041135 [1]), and the two reference sequences HxB2 and MN (accession numbers AF033819 and M17449, respectively). The sequence of the env gene from a clade D virus (DGOB) was used as an outlier. Triangles correspond to clusters of early or late clones. (B to E) Details for patients 133 (B), 153 (C), 159 (D), and 309 (E). The different groups of late clones are indicated by circled numbers.

FIG. 3.

Location of the modifications affecting potential N-linked glycosylation sites within the envelope glycoproteins of early and late clones from each patient and correlation with 2G12 binding. For clarity, regions of the envelope that contain N-linked glycosylation sites that were conserved between early and late clones are not shown. Potential N-linked glycosylation sites are highlighted in green. N-linked glycosylation sites implicated in the 2G12 epitope are indicated in red, with amino acid numbers corresponding to the HxB2 sequence (18). Clones that did not bind 2G12 are underlined, and those that weakly bound 2G12 are in italics and underlined.

FIG. 3.

Location of the modifications affecting potential N-linked glycosylation sites within the envelope glycoproteins of early and late clones from each patient and correlation with 2G12 binding. For clarity, regions of the envelope that contain N-linked glycosylation sites that were conserved between early and late clones are not shown. Potential N-linked glycosylation sites are highlighted in green. N-linked glycosylation sites implicated in the 2G12 epitope are indicated in red, with amino acid numbers corresponding to the HxB2 sequence (18). Clones that did not bind 2G12 are underlined, and those that weakly bound 2G12 are in italics and underlined.

FIG. 4.

Amino acid sequences including the linear 2F5 epitope among all clones isolated from the four patients and correlation with 2F5 binding. The core epitope is indicated in dark green. Amino acid numbers correspond to the HxB2 sequence (18). Clones that did not bind 2F5 are underlined. The asterisk corresponds to a stop codon.