Simian immunodeficiency virus from the sooty mangabey and rhesus macaque is modified with O-linked carbohydrate

Abstract

Although stretches of serine and threonine are sometimes sites for O-linked carbohydrate attachment, specific sequence and structural determinants for O-linked attachment remain ill defined. The gp120 envelope protein of SIVmac239 contains a serine-threonine-rich stretch of amino acids at positions 128 to 139. Here we show that lectin protein from jackfruit seed (jacalin), which binds to non- and monosialylated core 1 O-linked carbohydrate, potently inhibited the replication of SIVmac239. Selection of a jacalin-resistant SIVmac239 variant population resulted in virus with specific substitutions within amino acids 128 to 139. Cloned simian immunodeficiency virus (SIV) variants with substitutions in the 128-to-139 region had infectivities equivalent to, or within 1 log unit of, that of SIVmac239 and were resistant to the inhibitory effects of jacalin. Characterization of the SIVmac239 gp120 O-linked glycome showed the presence of core 1 and core 2 O-linked carbohydrate; a 128-to-139-substituted variant gp120 from jacalin-resistant SIV lacked O-linked carbohydrate. Unlike that of SIVmac239, the replication of HIV-1 strain NL4-3 was resistant to inhibition by jacalin. Purified gp120s from four SIVmac and SIVsm strains bound jacalin strongly in an enzyme-linked immunosorbent assay, while nine different HIV-1 gp120s, two SIVcpz gp120s, and 128-to-139-substituted SIVmac239 gp120 did not bind jacalin. The ability or inability to bind jacalin thus correlated with the presence of the serine-threonine-rich stretch in the SIVmac and SIVsm gp120s and the absence of such stretches in the SIVcpz and HIV-1 gp120s. Consistent with sequence predictions, two HIV-2 gp120s bound jacalin, while one did not. These data demonstrate the presence of non- and monosialylated core 1 O-linked carbohydrate on the gp120s of SIVmac and SIVsm and the lack of these modifications on HIV-1 and SIVcpz gp120s.

FIG. 1.

Common mucin-type O-linked carbohydrates. The Tn antigen, sialylated Tn antigen, nonsialylated and sialylated forms of core 1, immature core 2, and nonsialylated and sialylated forms of core 2 are the most common mucin-type carbohydrates. The main synthesis pathway of the common core types is indicated by black arrows from top to bottom. Initially, N-acetylgalactosamine is attached to serine or threonine to form the Tn antigen. Then attachment of galactose in a β-1,3 linkage forms the nonsialylated core 1 structure. An additional attachment of N-acetylglucosamine yields an immature core 2, which is not commonly modified with N-acetylneuraminic acid (sialic acid). Galactose is added to the immature core 2 to form nonsialylated core 2. N-Acetylneuraminic acid is attached to the Tn antigen, nonsialylated core 1, and nonsialylated core 2 structures. The sialylated forms of the common core structures are shown.

FIG. 2.

Selection of jacalin-resistant SIVmac239. (A) Representative replication of SIVmac239 in the absence of jacalin (squares) and in the presence of jacalin at 1 μg/ml (circles), 10 μg/ml (triangles), or 100 μg/ml (diamonds). (B) Replication of SIVmac239 and the jacalin-selected virus population in the absence of jacalin. (C) Replication of SIVmac239 and the jacalin-selected population in the presence of 250 μg/ml jacalin. CEMx174 cells were infected with SIVmac239 containing 5 ng of p27 or the jacalin-selected virus containing 5 ng p27. The cultures were divided in half 24 h later, and the medium was replaced with a medium that contained either no lectin or the indicated concentration of jacalin. SIV capsid protein (p27) that was released into the culture medium was measured on the days indicated in order to follow virus production.

FIG. 3.

Substitutions of amino acids in the jacalin-selected SIV population cluster in the potential O-linked carbohydrate attachment site. Envelope sequences were amplified from viral RNA, and 20 clones were sequenced. Five clones representative of the population are shown. The predicted O-linked glycosylation site (amino acids 128 to 139) in gp120 is boxed. The gp120 variable regions (V1 to V5), the junction between gp120 and gp41, and the transmembrane domain (TM) are also indicated. Amino acids in the jacalin-selected virus population that are the same as those in the parental SIVmac239 strain are shown as periods. Dashes indicate deletions. Asterisks represent stop codons.

FIG. 4.

Schematic representation of cloned SIV variants. Combinations of amino acid substitutions in the potential O-linked site in gp120 were introduced into the SIVmac239 proviral backbone. SIV variants Jacalin1 to Jacalin8 were based on changes present in the jacalin-selected population. Amino acid substitutions introduced in SIV01 to SIV07 were designed to disrupt the serine-threonine-rich stretch (amino acids 128 to 139).

FIG. 5.

Comparative infectivities of SIV variants that lack the O-linked carbohydrate attachment site. (A) SIV variants with V1 sequences that mirror the jacalin-selected virus population were within 0.8 log unit as infectious as the parental SIVmac239 strain. (B) SIV variants that were noninfectious or had an infectivity within 1 log unit of that of SIVmac239. Virus stocks for each cloned SIV variant and the cloned SIVmac239 were produced by transient transfection of HEK293T cells. Virus containing normalized amounts of p27 was used to infect C8166-secreted alkaline phosphatase (SEAP) cells, which contain a stably integrated, Tat-inducible SEAP reporter gene. Viral infectivity is directly correlated to the amount of SEAP released into the culture supernatant. At 72 h after infection, SEAP activity was measured in triplicate with a Tropix Phospha-Light kit. The mean SEAP activity (± standard deviation) is shown for each viral input (ng p27) indicated.

FIG. 6.

Jacalin resistance is conferred by sequence variations in the V1 domain of SIV gp120. (A) SIV variants with V1 sequences that mirror the jacalin-selected virus population were replication competent in the absence of jacalin. (B) SIV V1 variants replicated similarly to SIVmac239. (C) SIV variants with V1 sequences that mirror the jacalin-selected virus population were more resistant to jacalin than the sensitive SIVmac239 strain. (D) SIV V1 variants were more resistant to jacalin than the sensitive SIVmac239 strain. Virus stocks for each cloned SIV variant and cloned SIVmac239 were produced by transient transfection of HEK293T cells. CEMx174 cells were infected with SIV variant virus or SIVmac239 virus, each containing 5 ng p27. One day following infection, the culture was split, and the medium was replaced with a medium that did not contain lectin or with a medium that contained jacalin (250 μg/ml). SIV capsid protein (p27) released into the culture medium was measured on the days indicated in order to follow virus production.

FIG. 7.

Jacalin binds O-linked carbohydrate attached to the V1 domain of SIVmac239. (A) SIV04 and Jacalin4 gp120 proteins have markedly lower jacalin binding capacities than that of SIVmac239. HEK293T cells were transfected with no DNA (mock) or with the gp120 eukaryotic expression vector for SIVmac239, SIV04, or Jacalin4. Protein released into serum-free medium was purified with the lectin from Galanthus nivalis. Equivalent amounts of protein from each sample were loaded into six wells of a high-protein-binding plate. Jacalin-HRP was used to probe each sample in quadruplicate. Filled bars represent mean jacalin signals; error bars, standard deviations. Sera from SIV-positive rhesus macaques were used to probe each sample in duplicate. Shaded bars represent signals from sera binding gp120. (B) Normalized jacalin binding to gp120 from SIV04 or Jacalin4. The mean jacalin signal was normalized to the mean signal from sera. Filled bars represent normalized jacalin signals; error bars, standard deviations.

FIG. 8.

Mass spectra of SIVmac239, SIV04, and Jacalin4 gp120 permethylated O-glycans released by reductive elimination. The spectra were acquired in positive-ion mode using the reflectron. Molecular ions of m/z 534, 895, and 1,256 correspond to non-, mono-, and disialylated core 1 structures, respectively, whereas molecular ions of m/z 779, 983, 1,344, and 1,705 were defined as core 2 O-linked glycans. Signals from hexose oligomers are indicated by X's. All molecular ions are [M + Na]⁺, and each panel is normalized to the most abundant signal (100% intensity) in the raw data. The abundance reported in the text for each O-glycan corresponds to its percentage of the total observed O-glycome (based on peak heights). The O-glycome value was obtained by summing the peak heights of the molecular ions attributable to O-glycans. These are noted in each of the panels. Structural assignments are based on monosaccharide composition, fragmentation analyses, and knowledge of the glycan biosynthetic pathways. The sugar symbols used are those employed by the Consortium for Functional Glycomics (www.functionalglycomics.org). (A) SIVmac239; (B) SIV04; (C) Jacalin4; (D) mock transfection.

FIG. 9.

HIV-1 NL4-3 is resistant to jacalin inhibition. (A) Replication of HIV-1 NL4-3 and SIVmac239 in the absence of lectin. (B) Replication of HIV-1 NL4-3 and SIVmac239 in the presence of 250 μg/ml jacalin. CEMx174 cells were infected with SIVmac239 containing 5 ng p27 or with HIV-1 NL4-3 containing 5 ng of p24. One day following infection, the culture was split, and the medium was replaced either with a medium that did not contain lectin or with a medium that contained jacalin (250 μg/ml). The amount of SIV or HIV capsid protein that was released into the culture medium was measured on the days indicated in order to follow virus production.

FIG. 10.

The jacalin and PNA binding capacities of SIVmac/sm, HIV-2 Ben, and HIV-2 UC1 gp120s differ from those of HIV-2 NIH-Z, HIV-1, and SIVcpz gp120 proteins. (A) Jacalin binding capacities of SIVmac, SIVsm, HIV-2, HIV-1, and SIVcpz gp120 proteins. Equivalent amounts of histidine-tagged gp120 proteins were each loaded into six wells of a high-protein-binding plate. Jacalin-HRP was used to probe each sample in quadruplicate. A histidine-HRP antibody was used to probe each sample in duplicate. The mean jacalin signal was normalized to the mean signal from the histidine tag. The normalized jacalin signal (± standard deviation) is shown. (B) Same as panel A but without normalization. (C) Comparison of the jacalin binding capacities of SIVmac, SIVsm, and HIV-1 gp120s. HEK293T cells were transiently transfected either with no DNA (mock) or with a gp120 eukaryotic expression vector for SIVmac, SIVsm, or HIV-1. Env was purified from the cell culture supernatant with lectin from Galanthus nivalis. Equivalent amounts of total protein were loaded into the wells of a high-protein-binding plate. Jacalin-HRP was used to probe each sample in quadruplicate. The mean jacalin signal (± standard deviation) is shown. (D) Lectin from peanut (PNA) binds to the gp120 of SIVmac/sm but not to the gp120 of SIVcpz or HIV-1. Equivalent amounts of histidine-tagged gp120 proteins were each loaded into six wells of a high-protein-binding plate. PNA-HRP was used to probe each sample in quadruplicate. A histidine-HRP antibody was used to probe each sample in duplicate. The mean PNA signal was normalized to the mean signal from the histidine tag. The normalized PNA signal (± standard deviation) is shown.

FIG. 11.

Comparative neutralization of SIVmac239 and jacalin-resistant SIV V1 variants. (A and B) Neutralization of infectivity by pooled SIV-positive sera. (C and D) Neutralization of infectivity with soluble CD4 (sCD4). (E and F) Representative neutralization of infectivity by rhesus monoclonal antibody 1.9C. HEK293T-produced virus stocks for each cloned SIV variant and the parental SIVmac239 strain were incubated for 1 h with either pooled sera from SIVmac239-positive rhesus macaques, sCD4, or a rhesus monoclonal antibody. C8166-SEAP cells, which contain a stably integrated, Tat-inducible secreted alkaline phosphatase (SEAP) reporter gene, were then added. SEAP activity was measured 72 h later with a Tropix Phospha-Light kit. A lower percentage of SEAP activity is indicative of neutralization, while 100% SEAP activity indicates the lack of neutralization of virus infectivity.