A single-amino-acid polymorphism in Chikungunya virus E2 glycoprotein influences glycosaminoglycan utilization

Abstract

Chikungunya virus (CHIKV) is a reemerging arbovirus responsible for outbreaks of infection throughout Asia and Africa, causing an acute illness characterized by fever, rash, and polyarthralgia. Although CHIKV infects a broad range of host cells, little is known about how CHIKV binds and gains access to the target cell interior. In this study, we tested whether glycosaminoglycan (GAG) binding is required for efficient CHIKV replication using CHIKV vaccine strain 181/25 and clinical isolate SL15649. Preincubation of strain 181/25, but not SL15649, with soluble GAGs resulted in dose-dependent inhibition of infection. While parental Chinese hamster ovary (CHO) cells are permissive for both strains, neither strain efficiently bound to or infected mutant CHO cells devoid of GAG expression. Although GAGs appear to be required for efficient binding of both strains, they exhibit differential requirements for GAGs, as SL15649 readily infected cells that express excess chondroitin sulfate but that are devoid of heparan sulfate, whereas 181/25 did not. We generated a panel of 181/25 and SL15649 variants containing reciprocal amino acid substitutions at positions 82 and 318 in the E2 glycoprotein. Reciprocal exchange at residue 82 resulted in a phenotype switch; Gly(82) results in efficient infection of mutant CHO cells but a decrease in heparin binding, whereas Arg(82) results in reduced infectivity of mutant cells and an increase in heparin binding. These results suggest that E2 residue 82 is a primary determinant of GAG utilization, which likely mediates attenuation of vaccine strain 181/25.

Importance: Chikungunya virus (CHIKV) infection causes a debilitating rheumatic disease that can persist for months to years, and yet there are no licensed vaccines or antiviral therapies. Like other alphaviruses, CHIKV displays broad tissue tropism, which is thought to be influenced by virus-receptor interactions. In this study, we determined that cell-surface glycosaminoglycans are utilized by both a vaccine strain and a clinical isolate of CHIKV to mediate virus binding. We also identified an amino acid polymorphism in the viral E2 attachment protein that influences utilization of glycosaminoglycans. These data enhance an understanding of the viral and host determinants of CHIKV cell entry, which may foster development of new antivirals that act by blocking this key step in viral infection.

FIG 1

Soluble GAGs inhibit 181/25 infectivity and binding. (A and B) Purified virions of strain 181/25 (A) or SL15649 (B) (MOI of ∼2.5 PFU/cell) were incubated with buffer alone or buffer containing heparin, heparan sulfate, chondroitin sulfate A, dermatan sulfate, or bovine serum albumin (BSA) at the concentrations shown at 4°C for 30 min prior to adsorption to BHK-21 cells. At 2 hpi, the inoculum was replaced with medium containing 20 mM NH₄Cl. At 18 hpi, infected cells were detected by indirect immunofluorescence. Results are expressed as the mean percentage of infected cells normalized to untreated controls for three (181/25) or two (SL15649) independent experiments performed in triplicate. Error bars indicate SEM. *, P < 0.05; **, P < 0.01 (in comparison to untreated controls as determined by one-way ANOVA followed by Dunnett's post hoc test). (C) Strain 181/25 (MOI of ∼5 PFU/cell) was incubated with buffer alone or buffer containing heparin, heparan sulfate, chondroitin sulfate A, or dermatan sulfate at the concentrations shown at 4°C for 30 min. Virus-GAG mixtures were adsorbed to BHK-21 cells at 4°C for 30 min and stained with a CHIKV-specific antibody. The MFI of each sample was determined by flow cytometry. The data were normalized to the MFI of untreated virus controls for three independent experiments performed in triplicate. Error bars indicate SEM. *, P < 0.01 (in comparison to untreated controls as determined by one-way ANOVA followed by Dunnett's post hoc test).

FIG 2

Kinetics of inhibition of 181/25 by heparan sulfate and ammonium chloride. (A) 181/25 virions (MOI of ∼2.5 PFU/cell) were adsorbed to BHK-21 cells at 37°C. At the times shown prior to or during adsorption, heparan sulfate (250 μg/ml) was added to the virus inoculum. After 2 h adsorption, unbound virus was removed, and cells were incubated with medium containing 20 mM NH₄Cl. At 18 hpi, infected cells were detected by indirect immunofluorescence. Results are expressed as the mean percentage of infected cells normalized to BSA-treated controls from two independent experiments performed in triplicate. Error bars indicate SEM. *, P < 0.01 (in comparison to BSA-treated controls as determined by one-way ANOVA followed by Dunnett's post hoc test). (B) 181/25 virions (MOI of ∼2.5 PFU/cell) were adsorbed to BHK-21 cells at 37°C. At the times shown following adsorption, the virus inoculum was removed, and cells were incubated with medium containing 20 mM NH₄Cl. At 18 hpi, infected cells were detected by indirect immunofluorescence using a CHIKV-specific polyclonal antibody. Results are expressed as the mean percentage of infected cells normalized to the percentage of infected cells when NH₄Cl was added at 120 min from three independent experiments performed in triplicate. Error bars indicate SEM. *, P < 0.01 (in comparison to untreated controls as determined by one-way ANOVA followed by Dunnett's post hoc test).

FIG 3

CHIKV 181/25 and SL15649 infectivity of and binding to parental and mutant CHO cells. (A) Parental CHO-K1, pgsA745, pgsB761, and pgsD677 cells were adsorbed with an MOI of ∼10 PFU/cell of either 181/25 or SL15649. At 2 hpi, the inocula were replaced with medium containing 20 mM NH₄Cl. At 18 hpi, infected cells were detected by indirect immunofluorescence. Results are expressed as the mean percentage of infected cells normalized to the percentage of infected parental CHO-K1 cells for three (181/25) or two (SL15649) independent experiments performed in triplicate. Error bars indicate SEM. *, P < 0.001 (in comparison to the infectivity of the appropriate parental virus in CHO-K1 cells as determined by one-way ANOVA followed by Tukey's multiple-comparison test). (B) Parental CHO-K1, pgsA745, pgsB761, and pgsD677 cells were adsorbed with an MOI of ∼5 PFU/cell of either 181/25 or SL15649 at 4°C for 1 h and stained with a CHIKV-specific antibody. The mean fluorescence intensity (MFI) of each sample was determined by flow cytometry. The data were normalized to the MFI of virus bound to parental CHO-K1 cells for three (181/25) or two (SL15649) independent experiments performed in triplicate. Error bars indicate SEM. *, P < 0.001 (in comparison to the binding of the appropriate parental virus to CHO-K1 cells as determined by one-way ANOVA followed by Tukey's multiple-comparison test). GAG, glycosaminoglycan; HS, heparan sulfate; CS, chondroitin sulfate.

FIG 4

E2 residue 82 is a primary determinant of GAG utilization. Parental CHO-K1 and pgsB761 cells were adsorbed with an MOI of ∼10 PFU/cell of each parental virus (181/25 or SL15649) or the E2 mutants shown. At 2 hpi, the inoculum was replaced with medium containing 20 mM NH₄Cl. At 18 hpi, infected cells were detected by indirect immunofluorescence. Results are expressed as the mean percentage of infected pgsB71 cells normalized to parental CHO-K1 cells for each virus for two independent experiments performed in triplicate. Error bars indicate SEM. *, P < 0.001 (in comparison to infectivity of the appropriate parental virus in CHO-K1 cells as determined by one-way ANOVA followed by Tukey's multiple-comparison test).

FIG 5

CHIKV E2 R82G mediates a direct interaction with heparin. Approximately 5 × 10⁷ genome equivalents of purified 181/25, 181/25-E2 R82G, SL15649, and SL15649-E2 G82R virions were incubated with 75 μl of washed heparin-agarose beads (A) or unconjugated beads (B) at 4°C for 30 min. Beads were washed three times, and the bound material as well as 12.5% of the input virus was resolved by SDS-PAGE, transferred to a PVDF membrane, and immunoblotted for E2 as a marker for captured virus using a CHIKV-specific monoclonal antibody. The results of an experiment representative of three performed are shown. P, parental virus. (C) Densitometric analysis of virus bound to heparin-agarose beads. Data are expressed as the mean percent bound virus calculated from the densitometric analysis of captured virus divided by the estimated total input virus for three independent experiments. Error bars indicate SEM.

FIG 6

Electrostatic potentials of SL15649 and 181/25 E1/E2 trimers. (A) Top view of the electrostatic potential map displayed on the molecular surface of E1/E2 trimers of SL15649 (left panel) compared with a model of 181/25 (right panel) based on the crystal structure of CHIKV E1/E2 (PDB accession no. 2XFB). Positive potential is depicted in blue, and negative potential is depicted in red. (B) Enlarged view of the boxed areas from panel A highlighting the central cavity of the E1/E2 trimer. A white arrow indicates the position of Gly⁸² in SL15649 (left panel) or Arg⁸² in 181/25 (right panel) in one of the E2 monomers. (C) Enlarged view of the inner cavity rotated by 90° around the horizontal axis from the top view in panel B. A ribbon tracing of E2 is shown with a semitransparent view of the electrostatic surface and amino acids Gly⁸² (SL15649, left panel) or Arg⁸² (181/25, right panel) and Lys¹²⁰ shown in stick representations.