A conserved Gly436-Trp-Leu-Ala-Gly-Leu-Phe-Tyr motif in hepatitis C virus glycoprotein E2 is a determinant of CD81 binding and viral entry

Abstract

The hepatitis C virus (HCV) glycoproteins E1 and E2 form a heterodimer that mediates CD81 receptor binding and viral entry. In this study, we used site-directed mutagenesis to examine the functional role of a conserved G436WLAGLFY motif of E2. The mutants could be placed into two groups based on the ability of mature virion-incorporated E1E2 to bind the large extracellular loop (LEL) of CD81 versus the ability to mediate cellular entry of pseudotyped retroviral particles. Group 1 comprised E2 mutants where LEL binding ability largely correlated with viral entry ability, with conservative and nonconservative substitutions (W437 L/A, L438A, L441V/F, and F442A) inhibiting both functions. These data suggest that Trp-437, Leu-438, Leu-441, and Phe-442 directly interact with the LEL. Group 2 comprised E2 glycoproteins with more conservative substitutions that lacked LEL binding but retained between 20% and 60% of wild-type viral entry competence. The viral entry competence displayed by group 2 mutants was explained by residual binding by the E2 receptor binding domain to cellular full-length CD81. A subset of mutants maintained LEL binding ability in the context of intracellular E1E2 forms, but this function was largely lost in virion-incorporated glycoproteins. These data suggest that the CD81 binding site undergoes a conformational transition during glycoprotein maturation through the secretory pathway. The G436P mutant was an outlier, retaining near-wild-type levels of CD81 binding but lacking significant viral entry ability. These findings indicate that the G436WLAGLFY motif of E2 functions in CD81 binding and in pre- or post-CD81-dependent stages of viral entry.

FIG. 1.

(A) ClustalW alignment of prototype strains representative of each HCV genotype: 1a, H (GenBank accession no. M62631) and H77c (AF011751); 1b, J (D90208) and Con-1 (AJ238799); 1c, HC-G9 (D14853); 2a, JFH1 (AB047639); 2b, HC-J8 (D10988); 2c, BEBE1 (D50409); 2k, VAT96 (AB031663); 3a, NZL1 (D17763); 3b, HCV-TR (D49374); 3k, JK049 (D63821); 4a, ED43 (Y11604); 5a, SA13 (AF064490); 6a, EUHK2 (Y12083); 6b, TH580 (D84262); 6d, VN235 (D84263); 6g, JK046 (D63822); 6h, VN004 (D84265); and 6k, VN405 (D84264). Highlighted are the hypervariable regions located in E2 (black), the region under examination in this study (gray), the E2 stem region (crosshatched) (12), predicted locations of the transmembrane domains/signal peptide sequences (vertical stripes), and regions implicated in CD81 binding (horizontal stripes) (23, 31, 36, 43). Asterisk, identical; colon, highly conserved; period, partially conserved. (B) Fusion peptide sequences (underlined) of class II fusion proteins (tick-borne encephalitis virus Neudoerfl [U27495] and dengue virus type 2 [P12823]) are shown and compared to the region encompassing the G⁴³⁶WLAGLFY sequence (underlined) in HCV E2.

FIG. 2.

Heterodimerization of E1E2 mutants in cell lysates and pseudotyped particles. (A) Metabolically labeled lysates of 293T cells, transfected with 2 μg wild type (wt), mutated pE1E2H77c, or empty pcDNA4 vector (no E1E2), were immunoprecipitated with the conformation-dependent anti-E2 MAb H53. The immunoprecipitants were examined by nonreducing SDS-PAGE on 10 to 15% polyacrylamide gradient gels. E1 and E2 were visualized by phosphorimage analysis. (B and C) Heterodimerization of E1E2 mutants incorporated into HIV-1 pseudotypes. Metabolically labeled viral particles were pelleted from the tissue culture fluid of pE1E2H77c plus pNL4-3.LUC.R⁻E⁻-transfected cells prior to lysis in RIP buffer. E1E2 heterodimers were immunoprecipitated with MAb H53 (B) or polyclonal IgG obtained from an HCV-infected individual (C) prior to nonreducing SDS-PAGE on 10 to 15% gradient gels and phosphorimage analysis.

FIG. 3.

Ability of E1E2-pseudotyped HIV-1 luciferase reporter viruses to enter Huh7 cells. Data represent the means ± standard deviations of an experiment performed in quadruplicate and are representative of five independent assays. wt, wild-type E1E2-pp; empty, HIV-1 particles lacking the E1E2 envelope glycoproteins.

FIG. 4.

Ability of E1E2 precursors in cell lysates containing mutations in the G⁴³⁶WLAGLFY motif of E2 to bind CD81. (A) Detection of noncovalently associated E2 glycoprotein in cell lysates. Proteins were separated in 12% SDS-polyacrylamide gels under nonreducing conditions and then Western blotted with the E2-specific MAb All and goat anti-mouse Alexa 680 antibody. The amount of E2 in each lane was quantitated using Odyssey software and used to normalize the assay shown in panel B. Numbers at left of each panel are molecular masses in kilodaltons. (B) Ability of E2 in cell lysates to bind CD81. Dimeric MBP-LEL^113-201 was applied as a coating to enzyme immunoassay plates at 5 μg/ml. The cell lysates analyzed in panel A were titrated, and bound E2 was detected with MAb H53 and rabbit anti-mouse horseradish peroxidase in an enzyme immunoassay (13). The data were normalized against the amount of nonreduced E2 glycoproteins present and are representative of two independent experiments. Solid diamonds, wild-type pE1E2H77c-transfected cell lysates; open circles, empty vector control cell lysates.

FIG. 5.

Ability of E1E2 pseudotyped particles containing mutations in the E2 G⁴³⁶WLAGLFY motif to bind CD81. (A) Dimeric MBP-LEL^113-201 was applied as a coating to enzyme immunoassay plates at 5 μg/ml. Wild-type (WT) and mutated E1E2-pp were titrated twofold, and bound E2 was detected with MAb H53 and rabbit anti-mouse horseradish peroxidase. Data were calculated as (optical density of each dilution of mutant E1E2-pp/optical density of undiluted wild-type E1E2-pp) × 100. Maximal binding values for the individual assays were 0.40, 0.334, 0.525, and 0.39 optical density units (450 nm). Data represent the means ± standard errors of four independent experiments. P values were determined for wild type versus G436P, wild type versus A439S, wild type versus A439G, wild type versus G440P, and A439G versus G440P using a two-sample Student t test assuming unequal variances at 75% of maximal binding by wild-type E2 (1/4 dilution). (B) Immunoprecipitation of E2 and HIV-1 core proteins. E1E2-pps were produced in 293T cells and metabolically labeled with Tran³⁵S-label for 18 h. E1E2 glycoproteins were immunoprecipitated with MAb H53 and protein G Sepharose, analyzed under nonreducing conditions in 10 to 15% SDS-polyacrylamide gels, and subjected to phosphorimage analysis (upper panel). Bands corresponding to E2 are shown. The HIV core proteins were immunoprecipitated with IgG from an HIV-positive individual, analyzed under reducing conditions in 7.5 to 15% SDS-polyacrylamide gels, and subjected to phosphorimage analysis (lower panel). RT, reverse transcriptase. Numbers at right are molecular masses in kilodaltons.

FIG. 6.

Ability of E2 glycoproteins containing mutations in the G⁴³⁶WLAGLFY motif to bind full-length CD81. (A) Recombinant forms of E2 were expressed from the pcE2⁶⁶¹myc vector in 293T cells. The secreted glycoproteins were concentrated from the tissue culture fluid of transfected cells and examined by SDS-PAGE and Western blotting with MAb 9E10 and goat anti-mouse Alexa 680 antibody. Blots were scanned in an Odyssey detector, and the amount of E2 in each lane was quantitated using Odyssey software. Numbers at left are molecular masses in kilodaltons. (B) Equivalent amounts of E2⁶⁶¹myc (derived from panel A) were added to CHO-K1 cells transfected with a vector encoding human CD81. Bound E2⁶⁶¹myc was detected using ¹²⁵I-MAb 9E10 and counted in a Packard gamma counter. Shown are the averages of two independent assays ± standard deviations. wt, wild type.