Glycosylation of gp41 of simian immunodeficiency virus shields epitopes that can be targets for neutralizing antibodies

Abstract

Human immunodeficiency virus type 1 and simian immunodeficiency virus possess three closely spaced, highly conserved sites for N-linked carbohydrate attachment in the extracellular domain of the transmembrane protein gp41. We infected rhesus monkeys with a variant of cloned SIVmac239 lacking the second and third sites or with a variant strain lacking all three of SIVmac239's glycosylation sites in gp41. For each mutation, asparagine (N) in the canonical N-X-S/T recognition sequence for carbohydrate attachment was changed to the structurally similar glutamine such that two nucleotide changes would be required for a reversion of the mutated codon. By 16 weeks, experimentally infected monkeys made antibodies that neutralized the mutant viruses to high titers. Such antibodies were not observed in monkeys infected with the parental virus. Thus, new specificities were revealed as a result of the carbohydrate attachment mutations, and antibodies of these specificities had neutralizing activity. Unlike monkeys infected with the parental virus, monkeys infected with the mutant viruses made antibodies that reacted with peptides corresponding to the sequences in this region. Furthermore, there was strong selective pressure for the emergence of variant sequences in this region during the course of infection. By analyzing the neutralization profiles of sequence variants, we were able to define three mutations (Q625R, K631N, and Q634H) in the region of the glycosylation site mutations that conferred resistance to neutralization by plasma from the monkeys infected with mutant virus. Based on the reactivity of antibodies to peptides in this region and the colocalization of neutralization escape mutations, we conclude that N-linked carbohydrates in the ectodomain of the transmembrane protein shield underlying epitopes that would otherwise be the direct targets of neutralizing antibodies.

FIG. 1.

Antibody reactivity to linear peptides. (A) Locations of peptides in the region of the HIV gp41 protein proximal to the transmembrane domain that interact with antibodies (17). The beginning of the membrane-spanning domain is indicated. The conserved sites of N-linked glycosylation in the ectodomain of gp41 are also indicated by a gray box. (B) Antibody reactivities to overlapping peptides spanning the entire envelope protein by ELISA using a pool of SIV-positive plasmas (AE625) from monkeys infected with SIVmac239. The location of the variable loops, the beginning of the gp41 protein, and the predicted membrane-spanning domain are indicated. The conserved sites of N-linked glycosylation in the ectodomain of gp41 are also indicated by a gray box. (C) Same as B, except that plasma from an individual monkey (monkey 18-01) 16 weeks after infection with SIVmac239 was used as the source of antibody. (D) Same as B, except that plasma from an individual monkey (monkey 414-98) 22 weeks after infection with SIVmac239 was used as the source of antibody. Plasma from monkeys 18-01 and 414-98 was not present in the pool used for B.

FIG. 1.

Antibody reactivity to linear peptides. (A) Locations of peptides in the region of the HIV gp41 protein proximal to the transmembrane domain that interact with antibodies (17). The beginning of the membrane-spanning domain is indicated. The conserved sites of N-linked glycosylation in the ectodomain of gp41 are also indicated by a gray box. (B) Antibody reactivities to overlapping peptides spanning the entire envelope protein by ELISA using a pool of SIV-positive plasmas (AE625) from monkeys infected with SIVmac239. The location of the variable loops, the beginning of the gp41 protein, and the predicted membrane-spanning domain are indicated. The conserved sites of N-linked glycosylation in the ectodomain of gp41 are also indicated by a gray box. (C) Same as B, except that plasma from an individual monkey (monkey 18-01) 16 weeks after infection with SIVmac239 was used as the source of antibody. (D) Same as B, except that plasma from an individual monkey (monkey 414-98) 22 weeks after infection with SIVmac239 was used as the source of antibody. Plasma from monkeys 18-01 and 414-98 was not present in the pool used for B.

FIG. 2.

Comparative infectivities of SIVmac239 and mutants with different N-linked glycosylation sites in the transmembrane protein mutated. (A) Amino acid sequence of the glycosylated region of SIVmac239 gp41. (B and C) Relative infectivities of virus stocks. Virus stocks obtained from the transfection of HEK-293T cells were normalized for the amount of p27 and used to infect C8166-45 SIV-SEAP cells. SEAP activity was measured by use of a Phosphalight kit according to the manufacturer's recommendations at 3 days postinfection.

FIG. 3.

Replication kinetics of viruses with different N-linked glycosylation sites in the transmembrane mutated in PBMC cultures obtained from four different rhesus macaques. (A) Animal 206-03; (B) animal 330-03; (C) animal 333-03; (D) animal 467-03.

FIG. 4.

Viral RNA loads in plasma after intravenous inoculation of two macaques with an SIVmac239 strain lacking two gp41 N-linked glycosylation sites (animals 323-92 and 293-92) and two macaques lacking three sites (animals 394-91 and 103-94). Each animal was inoculated with virus equivalent to 20 ng of p27. Viral RNA loads in plasma were determined using a quantitative RT-PCR assay. d.p.i., days postinfection.

FIG. 5.

Neutralizing antibody responses. For each monkey, neutralizing activity in plasma was tested against the virus used for the inoculation. (A and B) SIVmac239-gp41/g23 virus against plasmas from animals 323-92 (A) and 293-92 (B). (C and D) SIVmac239-gp41/g123 virus against plasmas from animals 394-91 (C) and 103-94 (D).

FIG. 6.

Neutralization sensitivity of SIVmac239 to plasma samples from monkey 103-94 at weeks 16, 24, and 28 after infection with SIVmac239-gp41/g123.

FIG. 7.

Antibody reactivity to overlapping peptides spanning the entire envelope protein by ELISA using plasma from monkeys infected with SIVmac239-gp41/g23 (A and B) and SIVmac239-gp41/g123 (C and D) at week 16 postinoculation. The location of the variable loops, the beginning of the gp41 protein, and the predicted membrane-spanning domain are indicated. The conserved sites of N-linked glycosylation in the ectodomain of gp41 are also indicated by a gray box.

FIG. 8.

Antibody reactivity to overlapping peptides of the region spanning the sites for the attachment of N-linked carbohydrates in gp41. Plasma obtained at week 16 postinoculation of monkeys infected with the mutant viruses (monkeys 394-92, 293-92, 323-92, and 103-94) and weeks 16, 17, and 22 postinoculation of monkeys infected with SIVmac239 (monkeys 187-93, 414-98, 189-01, and 18-01) were tested for antibody reactivity to peptides with the parental and mutant sequence. (A) Sequences of the peptides that were used. (B) Pepscan reactivity. O.D., optical density.

FIG. 9.

Alignments of SIV envelope sequences from plasma RNA obtained from monkeys infected with the mutant viruses (five clones each) at weeks 16 (animals 293-92 and 394-91) and 22 (animals 323-92 and 103-94) postinfection. The location of the variable loops, the beginning of the gp41 protein, and the predicted membrane-spanning domain (TM) are indicated. Periods indicate conservation with the SIVmac239 sequence. The conserved sites for N-linked glycosylation in the ectodomain of gp41 are highlighted in gray. New sites of N-linked glycosylation (designated by a box) appeared during viral replication in monkeys 394-91 and 103-94.

FIG. 10.

Characterization of the chimeric viruses used to define mutations that confer neutralization resistance. (A) Sequence alignments in the region spanning the sites for the attachment of N-linked carbohydrates in gp41 of the chimeras and mutants used. The conserved sites of N-linked glycosylation in the ectodomain of gp41 are highlighted in gray. The new sites of N-linked glycosylation are designated by a box. (B and C) Relative infectivity. Virus stocks were obtained from the transfection of HEK-293T cells. Stocks were normalized for the amount of p27 and used to infect C8166-45 SIV-SEAP cells. SEAP activity was measured by use of a Phosphalight kit according to the manufacturer's recommendations at 3 days postinfection.

FIG. 11.

Neutralization sensitivities of SIVmac239-gp41/g123 mutant virus and two chimeric viruses (RNNH and RNH) to plasma samples from monkeys infected with the mutant viruses at week 16 postinoculation. Chimeric virus RNNH has a new site of N-linked glycosylation that has been removed in chimeric virus RNH. Sources of week 16 plasma were animals 323-92 (A), 293-92 (B), 394-91 (C), and 103-94 (D).