A Hepatitis C Virus Envelope Polymorphism Confers Resistance to Neutralization by Polyclonal Sera and Broadly Neutralizing Monoclonal Antibodies

CBH-5 (23), HC84.22 and HC84.26 (18), and HC33.4 (25) were gifts from Steven Foung (Stanford University School of Medicine, Stanford, CA, USA). AR3A (22) and AR4A (21) were gifts from Mansun Law (The Scripps Research Institute, La Jolla, CA, USA).

Plasma samples were obtained from the Baltimore Before-and-After Acute Study of Hepatitis (BBAASH) (42) cohort (Andrea Cox, Johns Hopkins University School of Medicine). Samples from each of the 18 HCV-infected subjects who had previously shown at least 50% neutralization by 1:100 plasma dilution of at least 2 HCV pseudoparticles (HCVpp) in the 19 HCVpp panel previously described (33) were selected.

HCVpp were generated by cotransfection of pNL4-3.Luc.R-E- plasmid and an expression plasmid containing the gene encoding Bole1a HCV E1E2 as described elsewhere (35, 43, 44). Virus-containing medium was collected at 48 and 72 h, pooled, and stored at −4°C. For infectivity and neutralization testing of HCVpp, 8,000 Hep3B cells (American Type Culture Collection) per well were plated in flat-bottom 96-well tissue culture plates and incubated overnight at 37°C. The following day, HCVpp were mixed with either MAb (10 μg/ml) or heat-inactivated plasma (1:100 dilution) and then incubated at 37°C for 1 h. The medium was removed from the cells and replaced with 50 μl of HCVpp mixture. The plates were placed in a CO₂ incubator at 37°C for 5 h, after which the HCVpp were removed and replaced with 100 μl of phenol-free Hep3B media and incubated for 72 h at 37°C. The medium was removed from the cells, 50 μl of 1× cell culture lysis reagent (Promega) was added and left to incubate for >5 min, and then 45 μl from each well was transferred to a white, low-luminescence 96-well plate (Berthold) and luciferase activity measured in relative light units (RLUs) in a Berthold Luminometer (Berthold Technologies Centro LB960). Infection by pseudoparticles was measured in the presence of MAb/test plasma (HCV_ppRLU_test) or nonspecific IgG/HCV-negative normal human plasma (HCV_ppRLU_control) at the same dilution. The percentage of neutralization was calculated as follows: 100% × [1 − HCV_ppRLU_test/HCV_ppRLU_control)]. Each sample was tested in duplicate. A mock pseudoparticle (no envelope) was used as a negative control. Neutralization was tested only for HCVpp with infectivity at least 10× greater than typical mock pseudoparticle values.

E1E2 site-directed mutants were generated using a QuikChange Lightning site-directed mutagenesis kit (Agilent Technologies).

Human hepatoma Huh7.5.1 cells (a gift from Jake Liang, NIH, Bethesda, MD, USA) were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 1% sodium pyruvate, and 1% l-glutamate. At 48 h after infection, cells were fixed with 4% formaldehyde for 20 min and then stained for HCV NS5A using a mixture containing primary anti-NS5A 9E10 antibody (a gift from Charles Rice, The Rockefeller University, New York City, NY, USA) at a 1:10,000 dilution in phosphate-buffered saline (PBS), 3% bovine serum albumin, and 0.3% Triton X-100 for 1 h at room temperature. Cells were washed twice with PBS and stained using a mixture of Alexa Dylight 488-conjugated goat anti-mouse IgG secondary antibody (Life Technologies) at a 1:500 dilution in PBS, 3% bovine serum albumin, and 0.3% Triton X-100 for 1 h at room temperature. Cells were washed twice in PBS and then stored covered in 100 μl PBS at 4°C.

Huh7.5.1 cells were plated at 10,000 cells per well onto 96-well plates. On the following day, antibodies were serially diluted 2.5-fold in growth medium starting from 50 μg/ml MAb. Controls were performed with a similar dilution series of nonspecific IgG. Viral supernatants diluted to fall in a linear range of infectivity were added to antibody dilutions in duplicate for 1 h at 37°C. The medium was removed from cells, and the antibody-virus mix was added onto cells and incubated overnight. The antibody-virus mix was replaced with 100 μl fresh medium the following day. After 48 h, the medium was removed and cells were fixed and stained. Images were acquired, and spot-forming units (SFU) were counted in the presence of MAb (HCV_ccSpots_test) or nonspecific IgG (HCV_ccSpots_control) using AID iSpot Reader Spectrum 219 operating AID ELISpot Reader version 7.0. Percent neutralization was calculated as 100% × [1 − (HCV_ccSpots_test/HCV_ccSpots_control)]. HCVcc neutralization assays using human plasma were performed similarly using 1:100 dilutions of test plasma and HCV-negative normal human plasma.

MAb or plasma binding to E1E2 was quantitated using an enzyme-linked immunosorbent assay (ELISA) as previously described (11). 293T cells were transfected with E1E2 expression constructs. At 48 h posttransfection, cell lysates were harvested. Plates were coated with 500 ng Galanthus nivalis lectin (Sigma-Aldrich) and blocked with phosphate-buffered saline containing 0.5% Tween 20, 1% nonfat dry milk, and 1% goat serum. E1E2 cell lysates were added. MAb or plasma was assayed in duplicate 2.5-fold serial dilutions, starting at 10 μg/ml or a 1:100 dilution, respectively. Binding was detected using horseradish peroxidase (HRP)-conjugated anti-human IgG secondary antibody (BD Pharmingen catalog no. 555788). For ELISA performed under denaturing conditions, lysates of cells were diluted in denaturing buffer (Tris-buffered saline with 10% fetal bovine serum, 1.0% sodium dodecyl sulfate, and 50 mM dithiothreitol) and then boiled for 5 min. Lysates were cooled on ice and then added to G. nivalis lectin-coated plates. The assay was completed as described above.

HCV RNA levels in infection supernatants were quantified using a process of RNA extraction and utilization of commercial real-time reagents (Abbot HCV Real Time assay) migrated onto a research-based real-time PCR platform (Roche 480 LightCycler).

To introduce an AfeI restriction site, HCVcc chimera H77/JFH1 (45), a gift from Jens Bukh (Copenhagen University Hospital, Copenhagen, Denmark), was amplified in three sections, using primers insert_R_new (CACCAGCTGATATAGCGTTTGTAATATGGCGACAGAGTC), insert_F_new (GGATTCCGATCTACCAGCGCTTTGGAGAACCTCGTAATACTCAATGCAGCATCCCTGGCC), back_mid_F (CGGAATATGACCTGGAGCTAATAACATCC), backbone_outer_R (GGTGACATGGTAAAGCCCCG), Back-mid_R (CCAGGTCATATTCCGGTCTGG), and Backbone_outer_F (CTGTCGCCATATTACAAACGC).

This omitted nucleotides 916 to 2579. Amplified sections were reassembled using In-Fusion cloning (Clontech). This backbone was digested with enzyme AfeI (New England BioLabs). E1E2 insertions were amplified from library plasmids using primers HCVcc_E1E2_1R (GAGGTTCTCCAAAGCCGCCTCCGC) and HCVcc E1E2_1F (TGTGCCCGCTTCAGCCTACCAAG). The insertions were cloned into the digested HCVcc backbone using In-Fusion cloning.

To make HCVcc RNA, 2 μg plasmid DNA was linearized using XbaI (New England BioLabs). Linear DNA was used for in vitro RNA transcription with a T7 MEGAscript kit (Ambion). RNA cleanup was performed using a RNeasy minikit (Qiagen) and quantified using NanoDrop. RNA products were then stored at −80°C.

To transfect RNA, Nucelofector kit T was used (Amaxa). A 5-μg volume of RNA was transfected into 1.8e6 Huh7.5.1 cells and plated in a 6-cm-diameter plate. Transfection supernatants were collected 4 to 6 days later, once cells had reached confluence, and stored at −80°C for titer determinations by HCV NS5A immunostaining. High-titer transfection supernatants were used to infect Huh7.5.1 cells, and infection supernatants were collected; then, the titers of the infection supernatants were measured, and the infection supernatants were stored at −80°C for use in infectivity and neutralization experiments.

HCVpp supernatants were concentrated and purified by ultracentrifugation through a 20% sucrose cushion (123,000 × g, 2 h, 4°C) with pellets resuspended in TNE buffer (50 mm Tris-HCl [pH 7.4], 100 mm NaCl, 0.1 mm EDTA). HCVpp and E1E2 lysates used in ELISAs were diluted and then denatured and run on a 4% to 12% Bis-Tris gel. After transfer, blots were probed with HC33.1.53 (25) (a gift from Steven Foung, Stanford University School of Medicine, Stanford, CA) and binding was detected with HRP-conjugated anti-human IgG secondary antibody (BD Pharmingen catalog no. 555788). HCVpp Western blots were also probed with mouse anti-HIV1 p24 (ab9044; Abcam) followed by HRP-conjugated anti-mouse IgG secondary antibody (ab97265; Abcam).

Percent neutralization values were grouped by individual mutations, and a one-way analysis of variance (ANOVA) with a Tukey test was performed to compare the levels of significance of changes in neutralization mediated by different amino acid mutations at the same position (see Fig. 2). For each pair of R424S ELISAs, a two-way ANOVA was performed to determine the significance of the change in binding between E1E2 lysates.

Through the use of Bayesian phylogenetics, ancestral sequence reconstruction, and covariance analysis, the most likely ancestral genotype 1a amino acid was predicted at each position across the HCV polyprotein, generating the Bole1a genome (39). For the majority of polyprotein positions, the most likely ancestral amino acid could be determined with very high posterior probability ranging from 0.99 to 1, but for 94 of the 3,012 codons analyzed, the posterior probability of the most likely ancestral codon was less than 0.99 (Fig. 1A), suggesting complex evolution at these positions deep in the phylogenetic tree. We selected three of these positions for further analysis. At position 242, which falls in HCV E1, leucine (L) was the most likely ancestral amino acid. At position 424, which falls near the CD81 binding site of E2, arginine (R) was the most likely ancestral amino acid. At position 570, which falls in the intergenotypic variable region (igVR) (46) near the carboxy terminus of E2, asparagine (N) was most likely. We also determined the most common amino acids at these positions in circulating viral strains (Fig. 1B). Circulating strains showed significant variability at the 242 and 570 positions, with 4 and 9 different amino acids observed at each position, respectively, in a reference alignment of 390 genotype 1a sequences. In contrast, the 424 codon is highly constrained, with either R or S observed in 387 of 390 circulating strains in the reference alignment.

HCV pseudoparticles (HCVpp) with Bole1a E1E2_L242, E1E2_R424, or E1E2_N570 or Bole1a E1E2 with other naturally occurring amino acids at these positions were generated by cotransfection of E1E2 expression plasmids with an HIV pNL4-3.Luc.R-E- reporter construct. The amino acid polymorphisms to be tested were chosen based on the predicted ancestral sequence, the most common amino acids in circulating strains, and some additional less frequently observed amino acids. To efficiently measure the effect of mutations at each of the three positions in multiple experiments, 13 HCVpp were generated with each mutation alone or in combination with mutations at one or both of the other two positions. These HCVpp were tested for sensitivity to neutralization by HCV-infected plasma samples, and all variants with the same amino acid at position 242, 424, or 570 were grouped for analysis (Fig. 2). Mutation of L242 to valine (V) or methionine (M) and mutation of N570 to valine (V) or aspartic acid (D) did not have a significant effect on neutralization sensitivity, but mutation of arginine (R) 424 to serine (S) conferred a significant increase in resistance to neutralization by plasma. Since variation at positions 242 and 570 did not significantly affect Bola1a neutralization sensitivity, subsequent experiments focused on variation at the 424 position, with L242 and N570 held constant.

To confirm the resistance phenotype of S424, Bole1a_R424 and Bole1a_S424 HCVpp were tested for neutralization by an additional panel of 19 HCV-infected plasma samples (Fig. 3A). As in prior experiments, S424 was strongly associated with resistance to neutralization (median neutralization, 91% for Bole1a_R424 and 44% for Bole1a_S424, P = 2E−10). Strikingly, for every plasma sample tested, Bole1a_S424 was more neutralization resistant than Bole1a_R424, despite the fact that each plasma sample was from a different donor and each sample presumably contained polyclonal NAbs. This result was confirmed with replication-competent cell culture virus (HCVcc) expressing either Bole1a_R424 or Bole1a_S424 E1E2 (Fig. 3B). In addition, mutation of R424 to S in two natural E1E2 variants isolated from HCV-infected donors (1a53 and 1a79) also conferred resistance to neutralization by heterologous HCV-infected plasma, and mutation of S424 to R in a third naturally occurring E1E2 variant (H77) conferred increased sensitivity to neutralization (Fig. 3C). To investigate whether changes in the neutralization sensitivity observed in Bole1a_R424 and Bole1a_S424 might be influenced by incorporation of different amounts of E2 per particle, we performed a Western blot analysis using purified HCVpp (Fig. 3D). Both E2 and the p24 loading control of Bole1a_S424 HCVpp showed approximately 2-fold less protein than Bole1a_R424 HCVpp, indicating that the two HCVpp incorporate approximately equal amounts of E2 protein per particle.

To elucidate the mechanism of variable neutralization sensitivity between Bole1a_R424 and Bole1a_S424 variants, the same panel of 19 HCV-infected plasma samples was used in E1E2 binding assays. Lysates of cells transfected with Bole1a_R424 or Bole1a_S424 E1E2 expression constructs were used as the target antigen in enzyme-linked immunosorbent assays (ELISAs) performed with HCV-infected plasma (Fig. 4A). Binding of plasma antibodies to the S424 E1E2 protein was significantly lower than binding to the R424 variant (median optical density [OD] values of 1.44 for Bole1a_R424 and 0.0009 for Bole1a_S424, P = 1E−05), suggesting that mutation of R424 to S confers neutralization resistance by reducing binding of NAbs. To confirm that the reduced binding to Bole1a_S424 E1E2 relative to Bole1a_R424 was not due to a difference in levels of expression of the two E1E2 proteins, we performed a Western blot analysis with serial dilutions of both E1E2 lysate preparations, confirming that the quantities of the proteins used in the ELISAs were similar (Fig. 4B).

To further investigate the surprising result that an R424S mutation conferred resistance to polyclonal sera from numerous donors, we measured the sensitivity of replication-competent cell culture virus (HCVcc) expressing Bole1a_R424 or Bole1a_S424 E1E2 to neutralization by a panel of some of the most broadly neutralizing human monoclonal antibodies (MAbs) described to date, including CBH-5 (23), HC84.22, HC84.26 (18), HC33.4 (25), AR3A (22) and AR4A (21). These MAbs were selected because they are broadly neutralizing and also because they bind to distinct epitopes across E2. CBH-5, AR3A, HC84.22, and HC84.26 bind to distinct epitopes at or near the CD81 binding site of E2, while HC33.4 binds near the amino terminus of E2, and AR4A binds to E1 and the carboxy terminus of E2 (Fig. 5A). Amino acid 424 is a known binding residue for AR3A as determined by alanine scanning mutagenesis (22), but it is not a known binding residue for any other MAbs in the panel. Despite the differences in E2 binding sites of these bNAbs, mutation of R424 to S conferred resistance to each antibody (Fig. 5B). As shown in Fig. 5C, the 50% inhibitory concentration (IC₅₀) of each MAb against Bole1a_R424 HCVcc was less than 2 μg/ml, while the IC₅₀ of the majority of MAbs against Bole1a_S424 HCVcc was 50 μg/ml or greater. Similar results were observed with HCVpp expressing Bole1a_R424 or S424 E1E2 and with HCVpp expressing R424 and S424 variants of a naturally occurring HCV strain, 1a53 (Fig. 5D and E). Interestingly, no difference in MAb sensitivity was observed between R424 and S424 variants of two other naturally occurring HCV strains, 1a79 and H77. Taken together, these results indicate that mutation of Bole1a_R424 to S confers resistance to bNAb neutralization, regardless of the antibody binding epitope.

Binding of MAbs to Bole1a_R424 and Bole1a_S424 E1E2 was measured using an E1E2 ELISA. MAbs HC84.22, CBH-5, AR3A, and HC33.4 showed greater binding to the R424 E1E2 lysate than the S424 E1E2 lysate (Fig. 6A and B). There was no detectable difference in binding of AR4A to either variant, suggesting either that the Bole1a_S424 resistance to neutralization by AR4A is not mediated by a reduction in binding of the MAb or that the binding ELISA is not sensitive enough to detect changes in binding affinity sufficient to influence neutralization. Taken together, these results show a small but consistent reduction in binding of MAbs to S424 E1E2 relative to R424 E1E2, in agreement with the observed reduction in binding of plasma antibodies to the S424 E1E2 protein, suggesting that resistance to bNAb binding is a likely mechanism of the observed neutralization resistance of Bole1a_S424.

Since mutation of R424 to S conferred resistance to binding of MAbs for which 424 is not a known binding residue, we tested whether the effect is due to an induced structural change in the E1E2 proteins. We were able to test this directly for one MAb, HC33.4, since it binds to both native and denatured protein. Binding of HC33.4 to Bole1a_R424 and Bole1a_S424 E1E2 protein lysates was measured with the E1E2 protein in either native or denatured states (Fig. 6B). Surprisingly, the difference in MAb HC33.4 binding to Bole1a_R424 and Bole1a_S424 E1E2 was significantly more pronounced with the protein in a denatured state, suggesting that induction of a structural shift may not be the mechanism by which mutation of R424 to S reduces binding of HC33.4. Instead, R424 could be a previously undetected HC33.4 binding residue.

We also measured binding of four MAbs to R424 and S424 variants of 1a53, 1a79, and H77 E1E2 proteins (Fig. 6C). Introduction of S424 into 1a53 E1E2 reduced binding of all four MAbs, which is consistent with the observed resistance to these MAbs conferred to 1a53 HCVpp by S424. For 1a79 and H77 E1E2, S424 conferred resistance to binding by some MAbs but not others, which is also consistent with minimal resistance to MAb neutralization conferred to 1a79 and H77 HCVpp by S424.

Despite the broad neutralization resistance conferred by S424, both R424 and S424 variants are abundant in circulating virus populations, suggesting selection pressure favoring R424 acting in opposition to the bNAb selection pressure favoring S424. We examined the relative levels of fitness of Bole1a_R424 HCVcc and Bole1a_S424 HCVcc in the absence of antibody (Fig. 7). Bole1a_R424 HCVcc showed higher specific infectivity than Bole1a_S424 HCVcc (18.9 SFU/10⁵ international units [IU] for Bole1a_R424 versus 5.4 SFU/10⁵ IU for Bole1a_S424). Taken together, these results suggest that selection by bNAbs favors persistence of S424 variants, while greater replicative fitness favors persistence of R424 variants.