Plant-based production of virus-like particles from hepatitis C virus

TL;DR

Scientists produced hepatitis C virus-like particles in plants for the first time, which could help in developing safe and scalable vaccines.

Contribution

This is the first successful production of complete hepatitis C virus-like particles in plants using a polycistron-like expression system.

Findings

HCV structural proteins Core, E1, and E2 were expressed in Nicotiana benthamiana plants.

Purified proteins assembled into spherical virus-like particles confirmed by electron microscopy.

The binary vector system achieved higher protein yields compared to the deconstructed system.

Abstract

Virus-like particles (VLP) from hepatitis C virus were successfully produced in Nicotiana benthamiana plants for the first time, by co-expressing three viral proteins (Core, E1 and E2) in a polycistron-like arrangement.

Hepatitis C virus (HCV) remains a global health challenge, underscoring the need for a preventive vaccine. Virus-like particles (VLP) offer a safe alternative, as they resemble native virions without infectious genomes. We expressed the HCV structural proteins Core, E1, and E2 in Nicotiana benthamiana using binary and deconstructed viral vector systems. Western blot confirmed expression, with the binary system achieving higher yields. Purified proteins assembled into spherical VLP (40–60 nm) were confirmed by electron microscopy. These findings demonstrate for the first time the feasibility of producing complete HCV-VLP in plants, supporting their potential as a scalable platform for vaccine development.

The online version contains supplementary material available at 10.1007/s00425-026-04956-6.

Figures

Diagram of the constructed vector LJMS-HCV used for transient transformation. The HCV/UBQ sequence contains the coding regions for the Core, E1, and E2 proteins, which are separated by ubiquitin-encoding sequences

HCV proteins detected in HCV-VLP and HCV-VLP TEM micrographs. a–c The three HCV proteins E1 (a), E2 (b) and Core (c) were detected by western blot (arrows) from HCV-VLP purified by iodixanol gradient. d-e Micrographs obtained by TEM showed the presence of HCV-VLP. Scale bar = 150 nm

Full text

Introduction

Hepatitis C continues to be a prevalent disease all over the world. It is caused by hepatitis C virus (HCV), a bloodborne pathogen that mainly affects the liver, which leads to both acute and chronic infections. The World Health Organization (WHO) estimates that approximately 58 million people live with chronic HCV infection, with about 1 million new infections occurring each year (Sallam et al. 2024; WHO 2025).

The development of direct-acting antivirals has revolutionized HCV treatment. This oral therapy can cure more than 95% of cases within 8–12 weeks with minimal side effects, making it the standard of care for HCV infection (https://stateofhepc.org/). The cost of HCV treatment has been high, with a 12-week course ranging from $24,000 to over$74,000. However, in some low- and middle-income countries, generic versions of the compounds are available at substantially lower prices, sometimes as low as $60 for a full course (Stateofhepc, 2025). The WHO aims to eliminate HCV as a public health threat by 2030, but achieving this goal requires increased screening, affordable treatment options, and comprehensive public health strategies.

HCV is an enveloped virus with a positive sense, single-stranded RNA genome of approximately 9.6 kb and a member of the Flaviviridae (Horsley-Silva et al. 2017). In spite of the availability of effective treatments, the development of a prophylactic vaccine still remains a necessity. Nevertheless, the generation of a vaccine has been challenging mainly due to the high variability of the virus, the limited availability of adequate animal models, and the complex immune response to HCV. Different strategies have been tested towards the development of a vaccine, including recombinant proteins, synthetic peptides, DNA, live vector-based vaccines, and combinations of these approaches, which have been tested with variable efficacy (Hartlage et al. 2021; Costa et al. 2025). Possible explanations for the limited efficacy of many of these approaches may include few protective viral epitopes included in the vaccine, the high variability of the virus, and the use of inadequate adjuvants. In this respect, virus-like particles (VLP) might represent an attractive alternative.

VLP are a powerful tool for eliciting broadly reactive immune responses that may be superior to individual antigens. They mimic the authentic conformation of viral capsid proteins but without genetic material and they may be more cost-efficient than co-inoculation of multiple antigens (Wijesundara et al. 2022; Zahmanova et al. 2023). A very convenient VLP feature is the fact that they may self-assemble in vivo when the proper proteins are expressed in heterologous systems. The resultant particles are structurally similar to authentic virus and able to elicit strong humoral and cytotoxic T lymphocyte response (Venkataraman et al. 2021).

In the case of the HCV, the genome translates into a single polyprotein of about 3000 amino acids that contains the viral proteins core, E1 and E2. These proteins have been co-expressed in bacteria, yeast, insect and mammalian cells resulting in VLP formation, albeit yields have been low (Gupta et al. 2023; Ali et al. 2025). According to Keyvani et al. (2012) and Gupta et al. (2023), formation of VLP from HCV (HCV-VLP) relies on the N-terminal region of the core protein, which interacts with an untranslated region that contains structural elements of the HCV RNA, which are crucial for VLP assembly. Several studies have confirmed that HCV-VLP can be divided into two types (Ali et al. 2025). The first type is made up only of the core protein, which is able to self-assemble into a capsid of approximately 30–50 nm in diameter; the second type comprises fully enveloped VLP, with an inner capsid surrounded by an outer lipid bilayer envelope containing the viral proteins E1 and E2, averaging 60–65 nm in size. Both forms of VLP had been investigated as potential vaccine candidates, and studies have confirmed that they are able to induce broad neutralizing antibodies and strong HCV-specific CD8 + T-cell response in preclinical studies (Chua et al. 2012; Gupta et al. 2023; Ali et al. 2025). Consequently, several systems have been tested for their production, such as yeast, mammalian and, as of late, plant cells, which can be an ideal platform.

Over the last 30 years, plants have demonstrated an enormous potential for production of recombinant proteins, including candidate antigens. As a production platform, the system is scalable, cost-effective and free of mammalian pathogens. The use of “humanized” plants has helped circumvent the problem of post-translational modifications (Schoberer et al. 2018). Both transgenic as well as transient approaches have been employed to successfully produce enveloped and non-enveloped VLP for a number of animal viruses (Uribe-Campero et al. 2015; Gonzalez-Castro et al. 2018; Mardanova et al. 2024). Using plant viral vectors, expression levels have been enhanced, enabling VLP production at suitable scales for vaccine development (Shi et al. 2025).

The use of plants for expression of HCV antigens has been explored in several instances. Han et al. (2000) expressed the HCV core protein in tobacco plants using a tobacco mosaic virus-based vector. Immunization with a crude extract from tobacco expressing HCV E2 hyper variable region 1 epitope induced specific neutralizing antibodies in mice (Nemchinov et al. 2000). Denis et al. (2007) used the papaya mosaic virus capsid protein as a carrier fused to a C-terminal HCV E2 epitope which elicited a long-lasting humoral response (> 120 days). Mohammadzadeh et al. (2014) also expressed the HCV core protein in tobacco plants employing a non-replicative, binary vector (pBI121) and a potato virus X-based viral replicative vector. The potato virus X-based vector provides higher expression levels which were enhanced by expression of the gene silencing suppressor P19 up to 0.022% of total soluble protein (TSP). Mohammadzadeh et al. (2015) employing a potato virus X-based viral replicative vector expressed a polytope from HCV fused to the Hep B surface antigen. An HCV E1-E2 heterodimer was transiently expressed in lettuce and intramuscular immunization of BALB/c female mice with E1-E2 dimers followed by two oral boosts induced systemic and mucosal immune responses as demonstrated by the presence of anti-HCV secretory immunoglobulin A in feces (Clarke et al. 2017).

Authentic VLP from HCV have been produced by expression of one or more HCV proteins in different expression systems, except plants (Ali et al. 2025). The produced VLP showed clear potential in stimulating specific humoral and T cell responses against HCV. Considering the advantages described above for VLP, the aim of this work was to obtain HCV-VLP in plants. Since the work of Dahari et al. (2010) has shown that inclusion of core, E1, and E2 proteins in vaccines is more significantly associated with protective immunity compared with vaccines based on other non-structural proteins, we decided to express simultaneously the three proteins in a polycistron-like arrangement, which allowed recovery of well-defined HCV-VLP.

Material and methods

Design of the constructs

A synthetic gene sequence was designed to include the tobacco etch virus enhancer, a stabilizing peptide (Harbury et al. 1993), and the three structural proteins of HCV (Uniprot accession P27958). Sequences were codon-optimized (see Suppl. Data S) for plant expression and linked via monomeric ubiquitin units (Walker et al. 2007). The sequence was synthesized by Gene Script (Piscataway, NJ, USA). Two plant expression vectors were constructed: LJMS-HCV and pICH11599-HCV (Suppl. Data S). For LJMS-HCV, the synthetic sequence was cloned into the KpnI and PstI sites of a binary vector constructed in our lab (Fig. 1). For pICH11599-HCV, the fragment was inserted into pICH11599. Vector maps are shown in Supplementary Data S. These constructs were used to transform Agrobacterium tumefaciens via electroporation. LJMS-HCV was introduced into strain LBA4404, maintained on rifampicin (50 mg/L) and kanamycin (50 mg/L); pICH11599-HCV was introduced into strain PGV3101, maintained on rifampicin (50 mg/L) and carbenicillin (100 mg/L).Fig. 1. Diagram of the constructed vector LJMS-HCV used for transient transformation. The HCV/UBQ sequence contains the coding regions for the Core, E1, and E2 proteins, which are separated by ubiquitin-encoding sequences

Nicotiana benthamiana agroinfiltration

Two transient expression systems were employed: a binary system using LJMS-HCV and pPSP19 carrying the Tomato Bushy Stunt Virus p19 suppressor gene (Huang et al. 2009), and a deconstructed virus system involving pICH11599-HCV, pICH14011, and pICH15879, generously provided by NOMAD Bioscience GmbH (Halle, Germany). Agrobacterium infiltration was performed as described by Uribe-Campero et al. (2015). Plants were maintained at 25 °C with a 16/8 h light/dark cycle post-infiltration.

Total soluble protein extraction and Western-blot analysis

Leaves were harvested and processed as described by Uribe-Campero et al. (2015). Soluble protein concentration was determined according to Bradford (1976). Protein integrity was verified by 12% SDS-PAGE with Coomassie staining. Western-blot analysis was performed on 12% SDS-PAGE gels using 80 µg of sample, 50 ng of core antigen, and 37.5 ng of E2 antigen (Austral Biologicals, San Ramon, CA, USA). Membranes were stained with Ponceau red, blocked with 3% skim milk in TBS and probed overnight at 4 °C with primary polyclonal antibodies at 1:1000 dilution for anti-core (Affinity BioReagents, Golden, CO, USA), anti-E1 (1:500; US Biological, Salem, MA, USA) and anti-E2 (1:1000, Austral Biologicals). Following TBS washes, membranes were incubated with secondary antibody diluted 1:3000 (goat anti-mouse IgG-AP; Zymed), washed, and developed using Sigma Fast BCIP/NBT.

VLP purification and analysis

VLP were purified by iodixanol gradient ultracentrifugation (Bertolotti-Ciarlet et al. 2002). Extracts were diluted 1:1 with iodixanol, filtered sequentially (0.8, 0.45, 0.22 µm), and centrifuged at 350,000 g for 3.5 h at 4 °C. Fractions (200 µL) were collected, and densities measured via refractometry. Protein presence was confirmed by Western blotting.

Electron microscopy

Samples were examined by transmission electron microscopy (TEM) using negative staining. Samples in phosphate buffered saline (PBS, 1:50) were placed on nickel grids (200 mesh) coated with formvar/carbon, incubated for 15 min, stained with 1% uranyl acetate, and imaged with a JEOL 2000 TEM at 80 kV.

Quantification of proteins by ELISA

Structural protein concentrations were determined by indirect ELISA using iodixanol gradient-purified extracts. A standard curve was generated with recombinant core antigen (Austral Biologicals). Polystyrene plates were coated overnight (4 °C) with 100 µL/well of standards or samples. Plates were blocked (1% BSA in PBS, 30 min), incubated with anti-core antibody (1:5000, Affinity BioReagents) for 2 h, washed, then incubated with secondary antibody (goat anti-mouse IgG-HRP 1:50,000, Millipore) for 1 h. Plates were washed and developed with SIGMAFAST OPD (Sigma), and OD450 was measured (Multiskan, Thermo Fisher Scientific).

Results

Plant transformation and protein detection

Expression of core, E1, and E2 HCV proteins should be sufficient for self-assembly into VLP. We expressed in planta simultaneously the full-length sequences of the three proteins intervened by ubiquitin monomers, employing a binary and a viral deconstructed vector (Fig. 2a–g). The integrity of the protein extracts was verified by SDS-PAGE, and the presence of each of the three structural proteins of the hepatitis C virus was determined by western blotting for both expression systems. E1 protein (30 kDa), E2 protein (70 kDa), and Core protein (21 kDa) were clearly detectable in crude extracts from both transformation systems. Nevertheless, all three proteins were detected in higher concentration in plants corresponding to 8 days post-infiltration in the binary system against the 12-day disassembled virus system (Fig. 2h–j). An ELISA was performed to determine the quantity of reactive protein in the samples. In the semi-purified extracts from the deconstructed virus system, a concentration of 1.01 mg/mL of reactive Core protein was detected, equivalent to 113.5 μg of reactive protein/g of fresh tissue and 3.5% of total soluble protein (TSP). In contrast, for the binary system, 5.5 mg of reactive protein/mL of sample was obtained in the semi-purified extracts, corresponding to 219 μg of reactive protein/g of fresh tissue and 6.7% of TSP.Fig. 2Nicotiana benthamiana plants expressing HCV proteins and Western blot of co-expressed proteins. a–d Visualization of 12-day post-infiltration plants transformed with the disassembled virus system producing HCV (c) and GFP (d) proteins. Wild type plants with no treatment were used as a negative control (a and** b**). g N. benthamiana plant visualized 8 days post-infiltrated with the binary system. h–j Verification by Western blot of the production of the E1 (h), E2 (i), and Core (j) proteins of the hepatitis C virus in N. benthamiana plants transformed by disassembled virus and binary systems. Black arrows show the detected proteins

VLP detection

Based on our results, it was decided to focus on the binary expression system as it yields higher levels of VLP. Once VLP were produced and purified as described before, detection was carried out by Western blotting using the three antibodies against the HCV capsid proteins: anti-core, anti-E1, and anti-E2 (Fig. 3a–c). These results confirmed that the VLP produced in plants were made up from the three co-expressed viral proteins as detectable by specific antibodies. To determine the presence of VLP-like particles with morphology and size consistent with the reported size of native HCV virions (Gastaminza et al. 2010), TEM was employed. The results clearly showed the presence of assembled HCV-VLP with the correct size and morphology (Fig. 3d-e) in samples derived from plants transformed using the binary vector system. TEM only provides ultrastructural evidence; additional biochemical/functional characterization will be addressed in future work.Fig. 3HCV proteins detected in HCV-VLP and HCV-VLP TEM micrographs. a–c The three HCV proteins E1 (a), E2 (b) and Core (c) were detected by western blot (arrows) from HCV-VLP purified by iodixanol gradient. d-e Micrographs obtained by TEM showed the presence of HCV-VLP. Scale bar = 150 nm

Discussion

Here we show for the first time the successful production of complete VLP from HCV in N. benthamiana plants. The expression of the structural proteins of HCV was achieved using both a deconstructed viral vector system and a binary vector system. Western-blot analysis of extracts from infiltrated plants confirmed the presence of proteins with the expected molecular weights: approximately 20 kDa for Core, 31 kDa for E1, and 70 kDa for E2. The analysis using antibodies against E2 yielded a weak signal. We repeated the experiment using higher antibody titers, loading more protein on the gels and other modifications, but it did not improve the signal. We do not know the reason, but possible explanations may include low affinity of the commercial antibody, incomplete protein transfer, masked epitopes, or another unknown reason. We are working to resolve this issue. In addition, in the case of core and E1 we detected multiple reactive bands both above and below the target protein. We hypothesize that the low molecular weight bands, which are also detectable in E2, may be degradation products. It is also possible that the bands above the target proteins may be oligomerized products, as it is known that core and E1 proteins can form diverse complexes (Strosberg et al. 2010; Pierce et al. 2024). A final explanation might be that these bands are contaminating proteins which happened to be recognized by the antibodies. The detection of bands at the predicted sizes demonstrated successful expression of the recombinant proteins in plants with both systems, although a higher yield was observed in samples obtained from the binary vector system. These findings suggest that the binary system enables higher protein accumulation compared to the deconstructed viral system, likely due to differences in vector design and expression regulation.

Once protein expression was verified, TEM analyses were performed on crude extracts from plants transformed with the binary system, to evaluate the capacity of the structural proteins to assemble into VLP. The micrographs revealed spherical particles ranging from 40 to 60 nm in diameter, which was consistent with the reported size of native HCV virions (Gastaminza et al. 2010). These results indicate that the co-expression of core, E1, and E2 in plants is necessary and sufficient for self-assembly into VLP that closely resemble authentic viral particles in morphology. Negative-stain TEM is a standard first-line technique to verify that VLP are assembled and display the expected morphology before employing biochemical or functional assays, and it has been employed to confirm morphology consistent with VLP assembly for a number of viruses (Uribe-Campero et al. 2015; Gonzalez-Castro et al. 2018; Mardanova et al. 2024). In the particular case of HCV-VLP, a number of publications have employed negative-stain TEM to confirm the morphology of VLP from structural proteins (Baumert et al. 1999; Acosta-Rivero et al. 2001; Blanchard et al. 2002; Gastaminza et al. 2010), and for that reason this was our method of choice to verify the morphology of our HCV-VLP.

The successful formation of VLP in a plant-based system represents an important advancement, as these structures are considered highly immunogenic due to their ability to mimic viral surface protein conformations without containing infectious genetic material (Bachmann et al. 2025). Thus, the VLP obtained in this study represents promising candidates for the development of safe and effective HCV vaccines.

The use of plants as bio-factories for recombinant protein and VLP production offers several advantages over conventional expression systems such as yeast, insect, or mammalian cells. Plant-based systems are cost-effective, scalable, and devoid of risks associated with animal pathogen contamination (Shi et al. 2025). Furthermore, they support the correct folding and post-translational modifications of complex glycoproteins, which are essential for the biological function and immunogenicity of viral proteins such as E1 and E2. As Ali and Tabil (2025) clearly establish, VLP can elicit strong humoral and cellular immune responses, which makes them a promising tool for development of HCV vaccines. In this sense, plants represent an ideal platform for the development of next-generation HCV vaccines, providing a safe, scalable, and economically viable alternative to traditional production systems.

Future work will involve additional biophysical and functional assays to confirm the presence and functionality of HCV-VLP. In addition, we will be employing murine models to perform immunogenicity studies and determine humoral and cellular specific responses of our HCV-VLP.

Supplementary Information

Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 126 KB)