Mice were handled according to the regulations of Spanish and European legislation. Protocols were approved by the Ethics Committee of “Instituto de Salud Carlos III” (Ref.: CBA 10_2019-v3).

Huh7.5 cells derived from human hepatoma were obtained from Apath LLC (Brooklyn, NY, United States), and Huh7.5.1 clone 2 was kindly provided by Dr. Francis V. Chisari (The Scripps Research Institute, La Jolla, CA, United States). Cell lines were maintained in Dulbecco’s Modified Eagle Medium (DMEM; Lonza, Basel, Switzerland), supplemented with 10% fetal bovine serum (FBS; Biological Industries, Beit Haemek, Israel), 4 mM L-glutamine (Lonza), and antibiotics (100 U/mL penicillin, 100 U/mL streptomycin; Lonza) at 37°C, 5% CO₂.

The plasmid encoding the JFH1 genome (Gt2a) was obtained from Apath LLC (Wakita et al., 2005). The plasmids encoding JFH1-based chimeric viruses containing the Core-NS2 region of Gt1a (H77/JFH1) (Scheel et al., 2008), Gt1b (J4/JFH1) (Gottwein et al., 2009), Gt3a (S52/JFH1) (Gottwein et al., 2007), and Gt4a (ED43/JFH1) (Scheel et al., 2008) were kindly provided by Jens Bukh (Copenhagen University Hospital, Copenhagen, Denmark). Cell culture infectious HCVs (HCVcc) were produced from plasmid-transcribed RNAs and titrated as previously reported (Sepulveda-Crespo et al., 2022).

Huh7.5 clone 2 cells were infected at a multiplicity of infection (moi) of 0.01 focus-forming units per cell. The cells were then passed every 2–3 days until the virus titers reached maximum levels. At this point, cells were scraped and pelleted by low-speed centrifugation. Culture supernatants were stored at −80°C, and cell pellets were washed with PBS and resuspended in extraction buffer (10 mM Tris-HCl, pH 7.6, 5 mM EDTA, 140 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS).

Two sets of truncated forms of the HCVcAg (amino acid residues 1-125 and 1-169) were produced (Figures 1A–C). DNA sequences encoding the HCVcAg of five different genotypes, Gt1a (H77; GenBank accession no: EU363761), Gt1b (J4; accession no: FJ230881), Gt2a (JFH1; accession no: AB047639), Gt3a (S52; accession no: EU204645), and Gt4a (ED43; accession no: EU36376 0) with the addition of two six-histidine tags (His tag) at the 5′and 3′ ends (HCVcAg 1-169) or one His tag at the 3′ end (HCVcAg 1-125) were amplified by PCR and cloned in the E. coli expression vector pET21d (Novagen) to obtain the recombinant HCVcAg[1-125] and HCVcAg[1-169] plasmids. Two His tag additions to the 1-169 constructions were necessary to stabilize the expressed recombinant proteins against degradation.

Escherichia coli strain C41(DE3) cells were transformed with the corresponding recombinant plasmids. A single colony was selected and used to inoculate 150–800 mL of Luria-Bertani media supplemented with 100 μg/mL ampicillin. This culture was grown to an optical density at 600 nm of 0.6, and then IPTG (Sigma) was added to a final concentration of 0.5 mM and incubated at 37°C for 5 h to induce protein expression. Cells were harvested by centrifugation at 7,400 g for 10 min, and cell pellets were resuspended in denaturing buffer, 20 mM Tris, 50 mM NaCl, 6 M Guanidinium Chloride (GdmCl) pH 8.0, and lysed by tip sonication. The lysate was clarified by centrifugation at 27,000 g for 60 min.

Recombinant proteins from the supernatant were purified under denaturing conditions using a single affinity chromatography step in a 6BCL-QHNi High-Density Nickel column (ABT) equilibrated with denaturing buffer. Once the protein solution had entered the column, it was washed with a denaturing buffer containing 10 mM imidazole and later with the same buffer containing 30 mM imidazole. The proteins were eluted with denaturing buffer containing 200 mM imidazole. After elution, imidazole and GdmCl were readily removed from the protein solution using a PD-10 desalting column (GE Healthcare) equilibrated in 20 mM Tris, 50 mM NaCl, pH 8.0 buffer. The purified and renatured recombinant proteins were aliquoted and stored at −20°C. The presence of the HCVcAg was monitored throughout the purification by SDS-PAGE.

BALB/c female mice between 6 and 8 weeks of age were inoculated intramuscularly with 10 μg of purified recombinant HCVcAg mixed with adjuvant (CpG Magic Mouse Adjuvant, Creative Diagnostics). Successive doses of HCVcAg from different genotypes were inoculated at 2–3 weeks intervals (Figure 1D). Four days after the last dose was administered, the mice were euthanized by isoflurane inhalation, and the spleen cells fused with Sp2-0 myeloma cells using polyethylene glycol (Stemcell Technologies). Hybridomas were seeded in 96 well plates and selected in Clonacell-HY selection medium (Stemcell Technologies) supplemented with HAT (hypoxanthine, aminopterin, and thymidine). After 7 days, the hybridomas were transferred to Clonacell-HY growth medium E (Stemcell Technologies). The presence of antibodies in hybridoma culture supernatants was analyzed by ELISA against purified HCV H77 HCVcAg and cell extracts from cells infected with H77 (Gt1a) recombinant HCV. Positive cultures were recloned at least twice by limiting dilution, expanded, and stored in liquid nitrogen.

The antibodies were purified from the hybridoma culture supernatant using CaptureSelect LC-kappa (mur) affinity resin (Catalog # 191315005, Thermo Scientific), concentrated (Clean-Up kit, ab102778, Abcam), and HRP-labeled (Lightning-link, ab102890, Abcam), following the manufacturer’s instructions.

The antibody isotype was determined using the Pro-Detect™ Rapid Antibody Isotyping Assay Kit, mouse (Catalog # A38550, Thermo Scientific).

Two hundred ng of purified recombinant proteins were separated in SDS-15% PAGE gels and transferred to an immobilon-P membrane (Millipore). The membrane was incubated for 1 hour with primary antibodies from selected hybridoma clones (1C, 2C, 4C, and 8C). Horseradish peroxidase-linked anti-mouse Ig (ab97040, Abcam) was used as a secondary antibody. Proteins were visualized by chemiluminescence using Pierce ECL Western Blot Substrate (Catalog # 32209, Thermo Fisher Scientific) on a ChemiDoc MP Imaging System instrument (Bio-Rad).

The amounts of proteins loaded on the gels were previously normalized in cell extracts according to their HCVcAg content using a pool of 1C, 2C, and 4C antibodies.

For dot blot assays, 3 µL of PBS-0.05% Tween20 containing 200 ng of purified recombinant proteins or normalized amounts of cell extracts were spotted on immobilon-P membranes (Millipore). The membranes were treated as described for Western blotting.

All the assays were run in a Biacore T200 instrument (Cytiva) using the recombinant HCVcAg as captured ligands and the antibodies as analytes. The Cytiva His-Capture kit (#28-9950-56) was used for covalently binding an anti-6xHis monoclonal antibody (11,000 response units) to the sample and the reference cells of a CM5 chip (Cytiva, #BR100530). Approximately 70 RUs of the recombinant HCVcAg proteins were captured in the sample cell by the anti-6xHis mAb. To reduce ligand heterogenicity, the captured HCVcAg proteins contained the first N-terminal 125 aa and just one histidine tag at the C-terminal end. In binding experiments, the antibodies were injected in a multicycle format using nine different antibody concentrations (0.39–100 nM). Calculating affinity or kinetic parameters was impossible since binding data was unsuccessfully fitted to any affinity or kinetic model. Therefore, the binding properties for each interaction were evaluated by early (20 s after the end of antibody injection) and late (280 s after the end of antibody injection) stability points achieved at the highest antibody concentration used (100 nM). For competition assays, the binding points (5 s before the end of antibody injection) achieved at the 12.5 nM antibody concentration were evaluated after a previous competitor injection. The injections of antibodies as competitors were performed at different concentrations (6.25–200 nM) and were prolonged for 20s without a dissociation step. All antibody injections were carried out using a flow rate of 50 μL/min. Regeneration between cycles was performed by 60s injections of 10 mM Glycine pH 1.5 solution at a 30 μL/min flow rate.

For antibody screening, 96-well microtiter plates were coated for indirect ELISA with 20 ng per well of purified recombinant proteins (169 aa versions) or protein extracts from cells infected with JFH1 recombinant HCV. After blocking with PBS-TW20 0.05% supplemented with 5% porcine serum (26250-084, Thermo Fisher Scientific), primary antibodies (supernatants from hybridoma or purified antibodies 1C, 2C, 4C, and 8C) were added, and plates were incubated for 1 h at RT. Horseradish peroxidase-linked anti-mouse Ig (ab97040, Abcam) was used as the secondary antibody. Lastly, the substrate (OPD, ref. P8287, Sigma-Aldrich) was added, and the optical density was read at 493 nm. Extensive washing with water was done after each step.

For DAS-ELISA, 96-well microtiter plates were coated with 500 ng per well of purified capture antibodies. After blocking with PBS-5% BSA, antigens (40 ng of purified HCVcAg, concentrated supernatants from infected cells) were added, and plates were incubated for 30 min at RT. Subsequently, an HRP-conjugated detection antibody was added and incubated for 1 hour at RT, followed by substrate (OPD, ref. P8287, Sigma-Aldrich). Extensive washing with water was done after each step. The optical density was read at 493 nm.

Infected and non-infected control cell supernatants, obtained as described in the “Chimeric HCV viruses and cell infections” section, were concentrated 15 times using Vivaspin^® 20, 100,000 MWCO PES ultrafiltration concentrator units (VS 2041, SARTORIUS). Virus titers in concentrated supernatants were determined as focus-forming units per milliliter (FFU/mL) after staining with the anti-NS5A 9E10 antibody (a gift from Charles Rice; The Rockefeller University, New York, NY, United States), goat anti-mouse IgG-horseradish peroxidase (HRP) secondary antibody (Abcam, Cambridge, United Kingdom) and 3-amino-9-ethylcarbazole (AEC; Alfa Aesar, Heysham, United Kingdom) (Sepulveda-Crespo et al., 2022).

Before being tested, culture supernatants were solubilized by adding Triton X-100 (ref. T8532, Sigma-Aldrich) to a final concentration of 1%.

Commercial mouse anti-His tag (MCA1396) and anti-HCVcAg C7-50 (ab2740) antibodies were purchased from Bio-Rad and Abcam, respectively.

Five mice were immunized following the protocol outlined in Figure 1, and the hybridoma selection process is detailed in Supplementary Table S1. On average, each mouse yielded approximately 1300 hybridoma clones. These clones were screened using an ELISA assay, wherein they were simultaneously tested against purified recombinant HCVcAg H77 (Gt1a) and HBV Pre-S (utilized as a negative control). In some cases, complementary ELISA screening was performed by simultaneously testing the clones against protein cell extracts obtained from cells infected with the recombinant HCV (H77) and mock-infected cells. A selection of 10–28 hybridomas per mouse exhibiting high reactivity towards HCVcAg and no reactivity towards HBV Pre-S were chosen for further cloning.

Ultimately, a total of eight clones were selected for further analysis (Supplementary Table S1). Clones 5C, 6C, and 7C are not described in this manuscript, as they recognized the recombinant HCVcAg but not the HCVcAg from HCV-infected cells.

It is important to emphasize that despite initial screenings yielding a substantial number of positive clones, most of them lost their ability to produce antibodies during subsequent cloning rounds. This unexpected result may be due to genomic instability in hybridoma clones caused by their hyperploidy (Andreeff et al., 1985).

Four monoclonal antibodies were selected that specifically recognize HCVcAg of different genotypes: 1C (IgG2b), 2C (IgG2a), 4C (IgG2b), and 8C (IgG2b). Western blot (WB) and dot-blot (DB) assays showed that antibodies 1C, 2C, and 4C specifically reacted with purified HCVcAg from HCV genotypes Gt1a, Gt1b, Gt2a, Gt3a, and Gt4a (Figure 2A). However, antibody 8C did not react with HCVcAg of genotype Gt2a and barely recognized the protein of genotype Gt3a. The four monoclonal antibodies were directed against epitopes placed in the first 125 aa of HCVcAg since they reacted with a truncated version of the protein comprising only those amino acids (Figure 2A).

Almost identical results were obtained when antibodies were tested against protein extracts from Huh7.5.1 cells infected with chimeric HCV expressing HCVcAg of the same genotypes (Figure 2B). However, antibody 4C reacted poorly with HCVcAg of the Gt3a genotype in this case (Figure 2B). One possible explanation for the differences observed between purified proteins and proteins from cell extracts is the occurrence of post-translational modifications in the HCVcAg of Gt3a during infection. Such modification could impact the accessibility of 4C to the epitope. It is worth noting that these modifications are not anticipated to take place in E. coli, where recombinant proteins are expressed. Various post-translational modifications have been reported for HCVcAg in human cells, including phosphorylation, ubiquitination, and palmitoylation (Hundt et al., 2013).

The binding of monoclonal antibodies to purified HCVcAg was measured by SPR (The raw sensorgrams are shown in Supplementary Figure S1). This technique confirmed that the 8C antibody did not bind Gt2a and poorly Gt3a HCVcAg (Supplementary Figure S1).

Since binding data at the antibody concentration employed did not fit any affinity or kinetic binding model, probably due to the bivalency of antibodies and/or the presence of different oligomerization/conformational states in HCVcAg preparations, the binding properties for each antibody/protein interaction were determined by response values corresponding to early and late stability points achieved with the highest antibody concentration used (100 nM, Figure 3).

The strongest interaction was observed with the 2C antibody for all proteins, followed by 8C > 4C > 1C > C7-50 (Figure 3). As stated above, 8C did not interact with Gt2a and poorly with Gt3a HCVcAg. Additionally, the early and late stability points of the 1C and 4C antibodies with Gt3a HCVcAg were lower than those with the other proteins (Figure 3). It is important to remark that in all the cases where antibody/protein interaction was detected, the complexes formed were stable, which shows the suitability of these antibodies for developing DAS-ELISA or lateral flow immunochromatography tests.

The percentage of a particular antibody binding to HCVcAg Gt1b (125 aa) was measured at increasing concentrations of a competitor antibody. Figure 4 shows that antibodies 1C, 2C, and 4C competed with each other but not with 8C. The strongest competitor antibody was 2C, probably due to its stronger binding properties. Sensorgrams of the competition assays are shown in Supplementary Figure S2. We also performed a competition assay with the anti-HCVcAg commercial antibody C7-50, which competed with all other antibodies (1C, 2C, 4C, and 8C), indicating that it is directed against an epitope that overlaps with those recognized by them (Supplementary Figure S3).

The ability of the antibodies to work in a DAS-ELISA was tested with two different samples: 1) purified recombinant proteins, 2) culture supernatants from in vitro HCV-infected cells.

Most antibody combinations worked similarly with recombinant proteins (Figure 5). However, the 8C antibody did not detect HCVcAg Gt2a, and Gt3a was only slightly detected (Figure 5). Furthermore, when the best binder 2C antibody was used as the capture antibody, the signals were lower than with the other capture antibodies (Figure 5), probably due to its stronger binding properties that hinder the binding of detection antibodies. The DAS-ELISA detected HCVcAg at concentrations below 0.1 ng/μL (Supplementary Figure S4).

Culture supernatants of HCV-infected cells were concentrated and assayed in the DAS-ELISA to confirm that the monoclonal antibodies also detect HCVcAg proteins produced during an in vitro infection. HCVcAg Gt1a, Gt1b, Gt3a, and Gt4a were detected in the supernatants of infected cells when antibodies 1C, 4C, and 8C were used as capture antibodies (Figure 6). Despite the highest virus titer, the HCVcAg Gt2a signal was slightly above that of the supernatant of mock-infected cells only when 1C and 4C were used (Figure 6). However, these differences were not statistically significant. No significant differences were observed when the 2C antibody was used as a capture antibody (Figure 6).