A Virus-Like Particle System Identifies the Endonuclease Domain of Crimean-Congo Hemorrhagic Fever Virus

ABSTRACT

Crimean-Congo hemorrhagic fever virus (CCHFV; genus Nairovirus) is an extremely pathogenic member of the Bunyaviridae family. Since handling of the virus requires a biosafety level 4 (BSL-4) facility, little is known about pathomechanisms and host interactions. Here, we describe the establishment of a transcriptionally competent virus-like particle (tc-VLP) system for CCHFV. Recombinant polymerase (L), nucleocapsid protein (N) and a reporter minigenome expressed in human HuH-7 cells resulted in formation of transcriptionally active nucleocapsids that could be packaged by coexpressed CCHFV glycoproteins into tc-VLPs. The tc-VLPs resembled authentic virus particles in their protein composition and neutralization sensitivity to anti-CCHFV antibodies and could recapitulate all steps of the viral replication cycle. Particle attachment, entry, and primary transcription were modeled by infection of naive cells. The subsequent steps of genome replication, secondary transcription, and particle assembly and release can be obtained upon passaging the tc-VLPs on cells expressing CCHFV structural proteins. The utility of the VLP system was demonstrated by showing that the endonuclease domain of L is located around amino acid D693, as was predicted in silico by B. Morin et al. (PLoS Pathog 6:e1001038, 2010, http://dx.doi.org/10.1371/journal.ppat.1001038). The tc-VLP system will greatly facilitate studies and diagnostics of CCHFV under non-BSL-4 conditions.

IMPORTANCE Crimean-Congo hemorrhagic fever virus (CCHFV) is an extremely virulent pathogen of humans. Since the virus can be handled only at the highest biosafety level, research is restricted to a few specialized laboratories. We developed a plasmid-based system to produce virus-like particles with the ability to infect cells and transcribe a reporter genome. Due to the absence of viral genes, the virus-like particles are unable to spread or cause disease, thus allowing study of aspects of CCHFV biology under relaxed biosafety conditions.

INTRODUCTION

Crimean-Congo hemorrhagic fever (CCHF) is a severe viral disease in Eastern Europe, the Middle East, Asia, and Africa, reported to have a case/fatality rate of approximately 30% (1). Infections of humans are associated with an acute febrile disease that can lead to hemorrhages, hypovolemic shock, and death, while in infected animals no clinical signs can be detected. The CCHF virus (CCHFV) is transmitted via tick bites (principally from the Hyalomma genus) or by direct contact with blood or tissues from infected persons or animals (2). The efficiency of ribavirin as an antiviral in humans is still under debate (3), and there are currently no established prophylaxis, no specific treatment, and no FDA-approved vaccine. The pathogenesis as well as immune responses are poorly characterized, mostly due to the restriction to high-biosafety-level facilities (biosafety level 4 [BSL-4]) for handling of virus and biological samples. The only animal models available so far are based on mice lacking antiviral interferon (IFN) responses, thus hampering studies on innate immune system interactions (4, 5). Clearly, there is paucity in tools and methods to better study the virus and its host cell interactions.

CCHFV belongs to the Nairovirus genus of the family Bunyaviridae. This enveloped virus has a single-stranded negative-sense RNA genome (viral RNA [vRNA]) of an unusually large total size of 19,146 nucleotides (nt). The genome is divided into 3 segments and encodes 4 structural proteins: the RNA-dependent RNA polymerase (RdRp) L on the large (L) segment (12,108 nt), the two glycoproteins (GPs) Gc and Gn on the medium (M) segment (5,366 nt), and the nucleoprotein N on the small (S) segment (1,672 nt). The nucleoprotein N binds each RNA segment to form ribonucleoprotein complexes (RNPs), the functional template of the viral polymerase (pol) L for transcription and replication. Shortly after infection, CCHFV performs the so-called primary transcription leading to initial mRNA synthesis. The protein products are necessary for subsequent replication of the viral genome, resulting in copy RNA (cRNA) (positive sense) and progeny genomic vRNA (negative sense). The untranslated regions (UTRs) at both the 5′ and 3′ ends of each segment serve as a promoter and are required for the virion assembly and the budding into the Golgi apparatus (6, 7).

Reverse genetics systems are useful to dissect and study the different steps of the viral life cycle, the implication of viral and cellular proteins, and the function of viral protein domains (8). The most elementary setup, the minireplicon system, is based on the expression of an artificial, genome-like reporter RNA flanked by viral UTRs, which is encapsidated by recombinant N and L proteins to be transcribed and replicated. Minireplicon systems have been developed for several bunyaviruses, namely, Bunyamwera virus (9), Uukuniemi virus (10), Hantaan virus (HTNV) (11), Rift Valley fever virus (RVFV) (12), La Crosse virus (LACV) (13), and, recently, CCHFV (14). Minireplicon systems have helped the understanding of specific aspects of the viral cycle, such as transcription and encapsidation of the viral genome (15–17). Additional information, however, can be obtained by expanding to a system for transcriptionally competent virus-like particles (tc-VLPs) (8). Hereby, expression of viral glycoproteins (GPs) in addition to the minireplicon components leads to self-assembly into tc-VLPs, which structurally and functionally mimic the parental virus. Since VLPs are devoid of viral coding sequences, they cannot propagate and hence represent a safe tool to study highly pathogenic viruses under less-stringent-biosafety-level conditions (8, 18–20). Here, we describe the establishment of a tc-VLP system for CCHFV and demonstrate its potential to reconcile and investigate viral entry, primary transcription, genome replication, particle propagation, and antibody neutralization. Moreover, the system allowed us to confirm the previously hypothesized location (21) of the CCHFV endonuclease domain within the L protein.

MATERIALS AND METHODS

Cells and viruses.

HuH-7 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), 1% l-glutamine, and 1% penicillin-streptomycin. SW-13 cells were grown in Rega 3 minimum essential medium (MEM) (Life Technologies), 1% l-glutamine, 1% penicillin-streptomycin, and 0.00075% sodium bicarbonate. CCHFV strain IbAr10200 was propagated on SW-13 cells under BSL-4 conditions.

Plasmids.

The constructs encoding wild-type (wt) and RdRp-inactive CCHFV L polymerase (pCAGGS_V5_L_wt and pCAGGS_V5_L_ΔDD), CCHFV N nucleoprotein (pCAGGS_N), CCHFV-specific Gaussia luciferase minigenome (T7-vS-Gluc and T7-vL-Gluc), T7 polymerase (pCAGGS_T7), and negative-control protein (pcDNA3.1_3×Flag_ΔMx) were described previously (14, 22). The plasmids pGL3-luc and pRL-SV40, constitutively expressing firefly luciferase (FF-Luc) or Renilla luciferase (REN-Luc), respectively, were purchased from Promega.

All other plasmids were generated using standard molecular cloning techniques and confirmed by DNA sequencing. PCR was carried out with the Phusion HotStartII enzyme (FinnZymes). Plasmid pCAGGS_GP was constructed by subcloning the CCHFV major open reading frame (M-ORF) (23) into pCAGGS. The two genomic T7 polymerase (pol)-driven constructs pT7riboSM2_vS_Ren and pT7riboSM2_vL_Ren contain the REN-Luc gene in antisense orientation, flanked by the 3′ and 5′ genomic untranslated regions (promoter) of the CCHFV S and L segments, respectively. Those plasmids were obtained by a two-step method. In the first step, the CCHFV minigenome sequences vS_Gluc and vL_Gluc, encoding Gaussia luciferase in antisense orientation, were amplified by PCR from plasmids T7-vS-Gluc and T7-vL-Gluc (14), respectively. Restriction sites for Esp3I were engineered into the forward and reverse primers, to generate ends that are compatible with plasmid pT7riboSM2 (24) cut by the same enzyme. After ligation, the plasmids pT7riboSM2_vS_Gluc and pT7riboSM2_vL_Gluc were obtained. In a second step, the Gaussia reporter gene was replaced by the REN-Luc gene. Plasmid pRL-SV40 served as the PCR template for the REN-Luc sequence, using forward and reverse primers that contained restriction sites for BglII and KpnI, respectively. Both the PCR product and the recipient plasmids pT7riboSM2_vS_Gluc and pT7RiboSM2_vL_Gluc were cleaved with these enzymes. The insert containing the Gaussia luciferase gene was discarded, and the PCR fragment with the REN-Luc sequence was inserted into the vector backbone. The resulting plasmids pT7riboSM2_vS_Ren and pT7riboSM2_vL_Ren contained the REN-Luc gene in antisense orientation flanked by promoter sequences of CCHFV S and L segments, respectively.

Plasmid constructs pCAGGS_V5_L_D693A, pCAGGS_V5_L_D718A, and pCAGGS_V5_L_K734A, expressing mutated CCHFV polymerase sequences, were obtained by site-directed mutagenesis and cDNA cloning as follows. Three overlapping PCR products were separately amplified using the pCAGGS_V5_L_wt plasmid as the template. The overlapping PCR products covered the region between the BstBI and PacI sites of the L sequence and contained either mutation D693A, D718A, or K734A, introduced by the respective primers (sequences available upon request). The three PCR fragments were combined and PCR reconstituted to a full-length product by using the flanking primers only. The DNA was then digested with BstBI and PacI and cloned into pCAGGS_V5_L_wt which was cut with the same enzymes. Sequencing of the full-length L sequences was carried out to confirm the individual point mutations and to rule out any other unspecific mutations.

Production and amplification of tc-VLPs expressing Renilla luciferase.

Subconfluent monolayers of HuH-7 cells (“donor cells”) seeded in 6-well plates were transfected with 600 ng of pCAGGS_V5_L (_wt or _ΔDD), 200 ng of pCAGGS_N, 200 ng of pT7riboSM2_vS_Ren or pT7riboSM2_vL_Ren, 500 ng of pCAGGS_GP, 500 ng of pCAGGS_T7, and 100 ng of pGL3-Luc control, using Nanofectin transfection reagent (PAA Laboratories). The pGL3-Luc was cotransfected as a marker for potential carryover transfection of indicator cells. Depending on the specific experimental setup, some plasmids were omitted from the transfection mix. Four hours after transfection, the transfection medium was replaced with fresh medium. Cell supernatants were collected 72 h posttransfection, and REN-Luc and FF-Luc activities were measured in cell lysates using the dual-luciferase reporter assay system (Promega) and a Centro LB 960 microplate luminometer (Berthold Technologies). The supernatants were treated with 25 U/ml Benzonase (Novagen) at 37°C for 3 h and centrifuged at 12,000 × g for 5 min to remove cellular debris. The presence of tc-VLPs in donor cell supernatant was validated by infecting HuH-7 cells (“indicator cells”). Briefly, HuH-7 cells grown in 6-well plates were washed with phosphate-buffered saline (PBS) and incubated with donor cell supernatants for 1 h at 37°C before fresh medium containing FCS was added. Luciferase activities were measured 24 h later as described above. In some experiments, indicator cells were pretransfected 20 h before supernatant transfer with 600 ng pCAGGS_V5_L_wt and 200 ng pCAGGS_N.

Amplification of tc-VLPs was conducted by transfer of cell supernatants (named passage n) onto new HuH-7 cells that were pretransfected 20 h earlier with 600 ng pCAGGS_V5_L_wt, 200 ng pCAGGS_N, and 500 ng pCAGGS_GP. Supernatants from passage n + 1 were collected 72 h later and processed as described above.

Freshly produced supernatants were used, unless specified otherwise in the text.

Immunoblotting of tc-VLP content.

The protein content of tc-VLPs was assessed by Western blotting, using CCHFV strain IbAr10200 as a control. CCHF tc-VLPs were produced in HuH-7 cells in T75 flasks by proportionally increasing the number of cells and plasmid amounts according to the flask surface. Corresponding control supernatants, from mock- and minireplicon (L+N)-transfected HuH-7 cells, were also produced. CCHFV was amplified on SW-13 cells infected at a multiplicity of infection (MOI) of 0.01, and supernatant was collected 3 days postinfection (p.i.) (T75 flasks). A corresponding SW-13 mock control was also produced. Supernatants were prepared as follows: clarification from cell debris by centrifugation for 5 min at 500 × g and concentration by ultracentrifugation through a 20% sucrose cushion at 28,000 rpm, for 2 h at 4°C, using an SW-32 rotor (Beckman). Pellets were dried for 10 min at room temperature before resuspension in Opti-MEM (Life Technologies). All lysates were prepared as follows: cells were scraped off in 10 ml PBS and centrifuged for 5 min at 500 × g, and the pellet was resuspended in 150 μl radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris HCl, pH 7.5, 15 mM NaCl, 1% NP-40) containing protease inhibitors (Roche). Supernatants and cell lysate samples were mixed with 4× sample buffer (143 mM Tris-HCl, pH 6.8, 4.7% SDS, 28.6% glycerol, 20% β-mercaptoethanol, 4.3 mM bromophenol blue) and incubated for 10 min at 105°C. Proteins were analyzed by Western blotting (10% SDS-PAGE gel), using rabbit polyclonal antibodies against CCHFV N (1:5,000), CCHFV Gn (1:400), or CCHFV Gc (1:400) (25). Staining of the blot with 0.1% Ponceau S in 5% acetic acid served as a loading control. A horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody (Thermo Fisher) (1:20,000) was used to detect the primary antibodies. Quantification of the signals was performed using the Image Lab 4.0 software.

VLP neutralization assay.

Human serum neutralizing CCHFV was kindly donated by G. Korukluoglu (Refik Saydam National Public Health Agency, Ankara, Turkey). Sheep and mouse antisera raised against RVFV strain MP12 were kindly provided by A. Brun (INIA, Spain). The human control serum was kindly provided by A. Kaufmann (Philipps-University Marburg, Institute for Immunology, Germany). CCHF tc-VLPs were incubated for 1 h at 37°C with antisera diluted 1:200. HuH-7 cells, grown in 6-well plates and pretransfected with 600 ng of pCAGGS_V5_L_wt and 200 ng of pCAGGS_N, were infected with the treated tc-VLP-containing supernatants. Luciferase activity was measured 24 h postinfection as described above.

Titration of CCHF tc-VLPs by immunofluorescence staining.

Subconfluent monolayers of HuH-7 cells grown on coverslips were transfected with 600 ng of pCAGGS_V5_L_wt and 200 ng of pCAGGS_N, as described above. After overnight incubation at 37°C, the cells were infected with 10-fold dilutions of tc-VLP-containing supernatants. Twenty hours later, cells were fixed in 4% paraformaldehyde for 10 min, permeabilized with 0.2% Triton X-100 for 5 min, and blocked with PBS containing 2.5% FCS for 30 min. Rabbit polyclonal antibody against Renilla luciferase (MBL), diluted 1:200 in 2.5% FCS-PBS, was incubated for 30 min at room temperature, washed twice with PBS, and followed by an incubation with Alexa Fluor 488 donkey anti-rabbit IgG (Molecular Probes), diluted 1:400 in 2.5% FCS-PBS, for 30 min at room temperature (RT). Cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (1 μg/ml) for 5 min at room temperature. Images were obtained with a Zeiss Axiovert 200M microscope. Cell nuclei and Renilla luciferase-positive cells were counted in 4 random fields per condition, and titers were calculated accordingly.

Purification, stability, and cryoprotection assays.

Freshly produced CCHFV tc-VLPs were treated with 25 U/ml Benzonase (Novagen) at 37°C for 3 h, centrifuged at 12,000 × g for 5 min to remove cellular debris, and aliquoted.

Three methods were tested for tc-VLP purification and concentration: ultracentrifugation, ultracentrifugation through a sucrose cushion, and polyethylene glycol (PEG) precipitation. Ultracentrifugation (with or without a sucrose cushion) was conducted as described above for CCHFV particles. CCHFV tc-VLPs were also concentrated by polyethylene glycol precipitation. Briefly, a solution of 30% PEG 8000 (Sigma-Aldrich) was made in NTE buffer (100 mM NaCl, 10 mM Tris, pH 6.5, 1 mM EDTA). Cleared supernatants containing tc-VLPs were mixed with the PEG 8000 solution at a 10% final concentration and NaCl at a 230 mM final concentration. After a 30-min incubation with rocking at 4°C, samples were centrifuged for 1 h at 6,500 × g and 4°C. In the 3 methods, the pellets containing tc-VLPs were resuspended in Opti-MEM (Life Technologies), to be 100-fold concentrated. For luciferase assays, concentrated tc-VLPs were diluted 1:100 in Opti-MEM before infection of indicator cells.

For stability assays, the freshly produced tc-VLP samples were either boiled for 10 min at 95°C or stored for 3 days at 37°C, room temperature (RT), 4°C, −20°C, or −80°C. Samples were also frozen (−80°C) and thawed (RT) three times or kept at −80°C for 1 year. For cryoprotection assays, aliquots were frozen either crude or in the presence of various additives (0.002% DEAE-dextran [Sigma Aldrich], 0.3 M l-lysine [Sigma Aldrich], 20% sucrose [Serva], 0.05% Tween 20 [Sigma Aldrich], 10% glycerol [Roth], or 20% d-sorbitol [anhydrous] [ICN Biomedicals Inc.]) and kept for 3 days at −20°C or −80°C or frozen-thawed 3 times. All additive solutions were previously sterilized with 0.22-μm polyethersulfone filters (Sarstedt). Preservation of the different tc-VLP preparations was assessed by infection of HuH-7 cells pretransfected with 600 ng of pCAGGS_V5_L_wt and 200 ng of pCAGGS_N. Luciferase activity was measured 24 h later as described above and compared to freshly prepared, crude tc-VLP activity.

tc-VLP assay with endonuclease-inactive mutant polymerases.

CCHFV tc-VLPs were produced as described above, using different polymerase constructs (pCAGGS_V5_L_ΔDD, _wt, _D693A, _D718A, or _K734A). HuH-7 cells, grown in 6-well plates and pretransfected with 600 ng of pCAGGS_V5_L_wt and 200 ng of pCAGGS_N, were infected with Benzonase-treated tc-VLP-containing supernatants. Luciferase activity was measured 24 h postinfection as described above.

Real-time RT-PCR.

Protocols and primers for real-time reverse transcription-PCR (RT-PCR) were previously described (22).

Statistics.

The two-tailed, paired Student t test was used to compare the different conditions. The P value was considered significant if below 0.05.

RESULTS

Generation of CCHF virus-like particles expressing a reporter gene.

For our attempts to generate tc-VLPs for CCHFV, we built on the previously established minireplicon system (14). The original minireplicon system consists of BSR-T7/5 cells transiently transfected with constructs for CCHFV L and N and an in vitro-transcribed RNA representing a CCHFV minigenome harboring a Gaussia luciferase gene. However, Gaussia luciferase is a secreted reporter and hence not suitable to monitor VLP activity in supernatants of transfected cells. Therefore, we replaced it with the cell-bound Renilla luciferase (REN-Luc) reporter. We created two minigenome constructs, vS-Ren and vL-Ren, which contain the REN-Luc gene in negative-sense orientation, flanked by viral promoters of the CCHFV S and L segments, respectively. Coexpression of the viral glycoproteins (GPs) should enable packaging of the minireplicon RNPs into VLPs (Fig. 1A), as was shown elsewhere for several other viruses (8). Initial attempts with a series of cell lines (hamster BSR-T7/5, human HEK 293 and SW-13, and primate CV-1 and Vero) failed to yield detectable activity indicative of tc-VLPs, but the human hepatoma cell line HuH-7 turned out to be a reliable system. Curiously, in these cells the minireplicon system (i.e., expression of minigenome, L, and N) had very little activity over the background set by the transcriptionally inactive L mutant (LΔDD), but addition of the GP construct led to strong reporter activity (Fig. 1B and D). Experiments employing a neutralizing antibody showed that the GP-mediated boost (which was also observed for LACV and RVFV systems, though in a less extreme manner [18, 22]) is due to reinfection of transfected cells by the generated VLPs (data not shown). Differently from the previously published minireplicon system (14), we obtained very little background activity by the minigenome plasmid if transfected without L polymerase (data not shown). The supernatants of the transfected HuH-7 cells (“donor cells”) were harvested after 72 h of incubation and transferred onto fresh cells expressing CCHFV L and N (“indicator cells”). Measurement of REN-Luc activity 24 h later suggested formation of tc-VLPs derived from the full set of VLP plasmids (Fig. 1C and E). As expected, supernatants from the negative controls (inactive L mutant and minireplicon system only) did not yield substantial activities in indicator cells. Of note, usage of the L segment promoter for the REN-Luc minireplicon resulted in approximately 50-fold-higher activity than that with the S segment promoter (compare Fig. 1C and E). It was therefore decided to use the L segment-derived minireplicon for subsequent experiments.

FIG 1.

Production of CCHF virus-like particles containing a reporter gene (tc-VLPs). (A) General outline of the procedure to generate tc-VLPs. HuH-7 donor cells were transfected with an expression plasmid encoding CCHFV polymerase L, nucleoprotein N, and glycoproteins (GP), along with a CCHFV-specific REN-Luc minigenome. A plasmid constitutively expressing FF-Luc served as a transfection control. Seventy-two hours after transfection, supernatants were collected and treated with Benzonase to remove free nucleic acids. The presence of tc-VLPs in donor cell supernatants was detected by transfer of supernatants on HuH-7 indicator cells expressing L and N and measurement of REN-Luc and FF-Luc activities 24 h postinfection. (B to E) REN-Luc activities in donor and indicator cells, using S segment promoter minigenome (vS-Ren) (B and C) or L segment promoter minigenome (vL-Ren) (D and E). Donor cells were transfected as depicted in panel A. The CCHFV polymerase plasmid encoded either an RdRp-inactive (LΔDD) or a wild-type (L) polymerase. For minireplicon conditions, the plasmid encoding the glycoproteins was omitted. Luciferase activities were detected in donor cells after harvesting of supernatants and in indicator cells 24 h postinfection. Luciferase counts were normalized on the inactive L polymerase (LΔDD) condition. Mean values and the standard deviations of 3 independent experiments are shown. p.t., posttransfection; p.i., postinfection.

CCHF VLPs have characteristics of the parental virus.

So far, we have failed to visualize our tc-VLPs by immunoelectron microscopy, probably because of their inherent physical instability (see below) and the lack of a suitable antiserum. To obtain other, independent evidence of their composition, we analyzed the viral proteins present in tc-VLPs. Supernatants from HuH-7 cells transfected with different plasmid combinations were concentrated and purified by ultracentrifugation through a sucrose cushion and analyzed by Western blotting. As a control, we used supernatants of mock- and CCHFV-infected SW-13 cells, which were processed in the same way (Fig. 2A, lanes 1 and 2). As expected, no viral protein can be detected in mock- and untransfected cells (Fig. 2A, lanes 1 and 3). The CCHFV glycoproteins Gn and Gc, as well as nucleoprotein N, can be detected approximately at their expected size (37, 75, and 54 kDa, respectively) in both virus and tc-VLP supernatants (Fig. 2A, lanes 2 and 6). Interestingly, all these viral proteins are also found in concentrated supernatants from the control cells expressing an RdRp-inactive polymerase (LΔDD) (Fig. 2A, lane 4). Moreover, transfection of the GP plasmid alone is sufficient to secrete Gn and Gc in the supernatant (data not shown). Thus, VLPs can be produced in the absence of viral replication. However, as shown in Fig. 1C and E, LΔDD VLPs seem to not have packaged authentic nucleocapsids, as they cannot be transcomplemented in indicator cells. Of note, supernatants from minireplicon-transfected cells contained spurious amounts of N, as demonstrated by the faint signal in the immunoblot (Fig. 2A, lane 5). This confirms previous observations for CCHFV and RVFV that N can be secreted independently of GP (26, 27). Nonetheless, the signals for N and the GPs are stronger in the supernatants containing tc-VLPs. A ratio among N, Gc, and Gn representation in concentrated CCHFV particles, empty VLPs, and tc-VLPs was made based on the band signal intensities of 3 independent immunoblotting experiments (Fig. 2B). This semiquantitative method did not reveal significant differences in protein ratios between the different particle types, suggesting that tc-VLPs have a protein composition similar to that of authentic virions.

FIG 2.

CCHF tc-VLPs are similar to CCHFV particles. (A) Western blot analysis of CCHF tc-VLPs and virus. Lysate and ultracentrifuged supernatant from mock- (lane 1) or CCHFV IbAr10200-infected (lane 2) SW-13 cells and from HuH-7 cells either left untransfected (lane 3) or transfected with different plasmid combinations (lanes 4 to 6) were analyzed by Western blotting using polyclonal antibodies specific for CCHFV glycoprotein Gn (top panels), glycoprotein Gc (second panels), or nucleoprotein N (third panels). One representative blot of 3 is shown. (B) Ratios of N, Gc, and Gn expression in CCHFV, LΔDD VLPs, and L VLPs, based on band intensities in 3 independent Western blotting experiments performed with ultracentrifuged supernatants. The two-tailed, paired t test did not show any significant differences among virus, empty VLPs, and L VLPs. (C) Neutralization of tc-VLPs by CCHFV-specific antisera. tc-VLPs were generated as described for Fig. 1, using the L segment promoter minigenome (vL-Ren). The crude tc-VLP supernatants were incubated for 1 h at 37°C with human CCHFV-specific antiserum at a final dilution of 1:200. As controls, normal human serum was used, as well as sheep and mouse sera specific for RVFV. HuH-7 indicator cells were pretransfected with CCHFV L and N plasmids and infected with the different tc-VLP preparations. REN-Luc activity was measured 24 h postinfection, and the activity was normalized on the inactive L polymerase (LΔDD) supernatant condition. Luciferase fold activation is shown only for tc-VLP-containing supernatants. Mean values and the standard deviations of 3 independent experiments are shown. A two-tailed, paired t test was used to compare the treated supernatants with the untreated one (no serum). *, P < 0.05; n.s., not significant.

We also tested neutralization by specific antisera. The tc-VLPs were incubated either with two different antisera reactive against RVFV, with a human control antiserum, or with a human antiserum derived from a CCHFV-infected patient. Indicator cells expressing L and N were infected with the antiserum-pretreated tc-VLPs, and reporter activities were measured as outlined above. Neutralization of tc-VLP activity was specific for the CCHFV antiserum (Fig. 2C). Thus, our VLPs have antigenic properties of the parental CCHFV particles.

Increasing VLP production.

Compared to systems for other bunyaviruses like LACV and RVFV (18, 22), the reporter activity of the CCHFV tc-VLPs is rather low (data not shown). One measure to obtain a robust VLP system was to employ the minireplicon with the stronger L promoter (Fig. 1). Moreover, we devised a strategy to increase particle yields. Repeated passaging on cells expressing the virus structure proteins, i.e., L, N, and the GPs, is expected to enable the replication and encapsidation of the VLP-borne minireplicon RNA and the packing into second-generation progeny particles (Fig. 3A). Indeed, in indicator cells the addition of a GP construct to the mix of support plasmids leads to a slight increase in reporter activity, suggesting formation and spreading of daughter particles (Fig. 3B, passage 0). This difference between nucleocapsid amplification (N and L expression) and VLP amplification (N, L, and GP expression) was even more apparent when indicator cells were incubated for longer times. At 72 h postinfection, VLP activity was decreased if only N and L were provided in trans, whereas it increased when GPs were provided in addition (compare Fig. 3B, passage 0, with Fig. 3C, passage 0). When VLP supernatants were further passaged on indicator cells, each step further increased reporter activities, suggesting efficient VLP amplification (Fig. 3B). The increased VLP yields are also reflected by a weak but measurable primary transcription, i.e., VLP reporter activity in the absence of coexpressed L and N proteins (Fig. 3B, passages 1 and 2, left columns).

FIG 3.

Amplification of CCHF tc-VLPs. (A) General outline of the procedure to amplify tc-VLPs. HuH-7 donor cells were transfected with expression plasmids encoding CCHFV L, N, and GP, along with the L segment promoter minigenome (vL-Ren). Crude supernatants called tc-VLPs passage 0 were collected 72 h posttransfection, treated with Benzonase, and used to infect indicator cells pretransfected with L and N plasmids, combined with the GP or not. tc-VLP supernatants from passage 1 were collected 72 h posttransfection, and the whole procedure was repeated to generate tc-VLPs passage 2 and tc-VLPs passage 3. At each VLP passage step, REN-Luc activity was measured 24 h and 72 h postinfection in pretransfected indicator cells. (B) Reporter activity in donor cells (72 h posttransfection) and indicator cells (24 h after infection) with tc-VLPs from passage 0, 1, or 2. (C) Reporter activity in the indicator cells at 72 h after infection with tc-VLPs from passage 0, 1, or 2. Luciferase counts were normalized on the inactive L polymerase (LΔDD) condition. Luciferase fold activation is shown for the tc-VLP-containing supernatants. Values from one representative experiment out of 3 are shown. p.t., posttransfection; p.i., postinfection. (D) CCHF tc-VLP titers after amplification on cells expressing CCHFV structural proteins. Experiment 1 corresponds to results shown in panels B and C.

To determine the actual VLP titers after each passage, we employed immunofluorescence. Indicator cells expressing L and N were incubated with 10-fold dilutions of tc-VLP-containing supernatants collected 72 h postinfection and analyzed 24 h later by immunostaining against REN-Luc. Similar to the VLP reporter activity, the titers, which were calculated from the number of REN-Luc-positive cells (Fig. 3D), demonstrate that repeated passaging is a useful method to optimize VLP production. Despite repeated attempts and further passaging, however, we were unable to produce VLP yields above approximately 10E5 particles per ml, possibly an underestimate caused by the comparatively low REN-Luc production by the CCHFV promoter.

Importantly, the passaging strategy allows us to recapitulate all steps of the viral replication cycle under non-BSL-4 conditions. The preparations with increased titer that were obtained after several passages can detectably infect naive cells, thereby enabling the study of particle attachment, entry, and primary transcription. The passaging of tc-VLPs on cells expressing L, N, and GP extends the infection cycle to the subsequent steps of genome replication, secondary transcription, and particle assembly and release.

Stability of CCHF tc-VLPs.

To obtain concentrations higher than the 10E5 particles per ml from the repeated passaging, we precipitated VLPs with polyethylene glycol (PEG) or ultracentrifuged them through a sucrose cushion. This allowed a further 10- to 100-fold concentration of VLP stocks (Fig. 4A). Ultracentrifugation without a sucrose cushion, by contrast, resulted in a significant loss of activity (Fig. 4A). Other important aspects for downstream VLP applications are environmental stability and shelf life. To determine the stability, we incubated VLPs for 3 days at either 37°C, room temperature, +4°C, −20°C, or −80°C. Moreover, we freeze-thawed particle preparations 3 times from −80°C to room temperature or incubated them for 10 min at 95°C as a negative control. As shown in Fig. 4B, only storage at −20°C or below was able to preserve VLP quality, whereas any prolonged storage at higher temperatures, as well as repeated freeze-thawing, was detrimental. On the other hand, when kept at −80°C, even after over a year of storage the VLPs were still active to about 40% (Fig. 4C). Similar results were obtained with VLPs containing the minigenome with the weak S segment promoter (data not shown). We also tested the influence of various additives known to enhance VLP stability (28, 29) (Fig. 4D). The presence of DEAE-dextran (to dissolve particle aggregates) did not change VLP activity, whereas sucrose (to prevent conformational changes of viral envelope proteins) decreased REN-Luc activity similarly to lysine or the detergent Tween 20, either of which is known to disrupt VLP structure. Apparently, for unknown reasons the presence of glycerol or sorbitol (reducer of protein aggregation and stabilizer of VLP tertiary structure, respectively) not only preserved but slightly increased REN-Luc activity, although this was not statistically significant (Fig. 4D). Thus, the CCHF tc-VLPs can be stored for a long time at −80°C without any major degradation, either crude or in the presence of glycerol or sorbitol solutions.

FIG 4.

Purification, stability, and conservation of CCHF tc-VLPs. A large stock of tc-VLPs was generated as described for Fig. 1, using the vL-Ren minigenome. Differently prepared or treated tc-VLP preparations were used to infect HuH-7 indicator cells expressing CCHFV L and N. REN-Luc was measured 24 h postinfection and normalized on the inactive L polymerase (LΔDD) supernatant condition, and the activity was expressed as a percentage of the fresh and crude REN-VLP supernatant condition. Mean values and the standard deviations of 3 independent experiments are shown. A two-tailed, paired t test was used to compare the treated supernatants with the untreated one. (A) REN-Luc activity in cells infected with either crude, ultracentrifuged, ultracentrifuged through a sucrose cushion, or PEG-precipitated tc-VLPs. The 100-fold-concentrated tc-VLPs were diluted 100-fold for the luciferase assay. (B) REN-Luc activity in cells infected with crude tc-VLP supernatants used immediately after collection (fresh), stored for 3 days at different temperatures, or incubated for 10 min at 95°C. One tc-VLP sample was frozen (−80°C) and thawed (RT) 3 times (Fr/th). (C) REN-Luc activity in cells infected with crude tc-VLP supernatants that were stored for 1 year at −80°C. (D) REN-Luc activity in cells infected with crude tc-VLP supernatants frozen at −20°C or −80°C for 3 days in the presence of various additives. Some samples were also frozen (−80°C) and thawed (RT) 3 times (Fr/th). *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant.

Endonuclease domain of CCHFV.

Bunyaviruses share with arenaviruses and orthomyxoviruses the ability to cleave host cell mRNAs and use the resulting 5′ capped oligonucleotides as primers for their own transcription (“cap-snatching”). The viral endonuclease domains had been identified and the atomic structures had been solved for arenaviruses (21), orthomyxoviruses (30, 31), and the bunyavirus LACV (32). The endonuclease domain of LACV (genus Orthobunyavirus) is situated at the N terminus of the L protein. Sequence alignments indicated a similar N-terminal endonuclease domain for three other Bunyaviridae genera, Phlebovirus, Hantavirus, and Tospovirus (32), and for phleboviruses, this was confirmed by functional data (21, 22). For the genus Nairovirus (to which CCHFV belongs), however, the N terminus of the L protein harbors the so-called OTU domain, which has immune escape function (33). Morin et al. proposed that the endonuclease domain of nairoviruses is located further downstream (Fig. 5A), since the amino acid sequence and predicted structure around position 700 exhibit similarities to the N terminus of the other bunyaviruses (21) (Fig. 5B). We employed our tc-VLP production/infection system to test this hypothesis. The predicted endonuclease-signature amino acids 693, 718, and 734 of CCHFV L were individually exchanged against alanine. In donor cells producing VLPs, this resulted in a substantial impact on reporter activity, with mutants D693A and K734A not displaying any activity and D718A being down to one-fourth of wild-type (wt) activity (Fig. 5C). However, when supernatants of the donor cells were transferred on indicator cells that express wt L (together with N), all mutants could be rescued to some extent (Fig. 5D). VLPs of mutant D693A were even transcomplemented to an activity level that slightly (though not significantly) exceeded that of wt VLPs. The results indicate that transcription-inactive nucleocapsids of mutant D693A had been assembled, replicated, and packaged into VLPs. Indeed, direct analysis of RNA synthesis by strand-specific real-time RT-PCR of donor cells revealed that the D693A mutant, which did not generate translatable reporter mRNA (Fig. 5C), produced substantial amounts of positive-strand RNA (Fig. 5E). This indicates unhindered production of cRNA, i.e., minigenome replication. The other two mutants exhibited a similar correlation between positive-strand RNA synthesis and the ability to be transcomplemented (Fig. 1D) but had a slightly (D718A) or significantly (K734A) lower rate of RNA synthesis, making it difficult to draw firm conclusions. Thus, mutant D693A displays the most striking phenotype, with a complete lack of mRNA transcription but an uninhibited RNA replication activity. These results demonstrate that D693 in the CCHFV L sequence is needed for mRNA transcription but not for genome replication, indicating that the surrounding domain is indeed harboring the predicted endonuclease domain (21).

FIG 5.

Localization of the CCHFV endonuclease domain. (A) Domain organization of bunyavirus polymerases: comparison between the L proteins of LACV, RVFV, HTNV, and CCHFV. (B) Amino acid sequence alignment of confirmed (LACV and RVFV) and predicted (HTNV and CCHFV) endonuclease domains, as proposed by Morin et al. (21). (C to E) Endonuclease mutants of CCHFV L. (C) tc-VLP production by donor cells. CCHF tc-VLPs were generated using constructs encoding either wild-type polymerase (L), various point mutants (L_D693A, L_D718A, and L_K734A), or the RdRp-inactive mutant (L_ΔDD). REN-Luc activity was measured 72 h posttransfection in donor cells. (D) Transcomplementation assays. Indicator cells were transfected with plasmids encoding CCHFV N and wt L at 20 h prior to the infection with crude tc-VLPs, and REN-Luc activity was measured 24 h postinfection. The activity of the tc-VLPs with L was set to 100%. Mean values and standard deviations of 4 independent experiments are shown. A two-tailed, paired t test was used to compare the mutant polymerase activity to the wild-type polymerase activity. (E) Positive-strand RNA synthesis by the CCHFV L mutants. Total RNAs from 72-h-transfected HuH-7 donor cells were tested by strand-specific real-time RT-PCR for the presence of positive-strand REN-Luc RNA. Untransfected cells (UT), cells transfected with a plasmid constitutively expressing REN-Luc (pRL-SV40), and the RdRp-inactive mutant L_ΔDD were used as controls. Note that the amount of the negative-sense vL_Ren minigenome plasmid was lowered to 10 ng per well to reduce the RT-PCR background signal. rel., relative. Mean values and the standard deviations of 4 independent experiments are shown. A two-tailed, paired t test was used to compare the mutant polymerase activity to the wild-type polymerase activity. **, P < 0.01; ***, P < 0.001; n.s., not significant.

Taken together, the results show that our tc-VLP system is a suitable tool to investigate aspects of CCHFV biology such as neutralizing antibody responses, cell infection, mRNA transcription, and genome replication under convenient BSL-2 conditions.

DISCUSSION

tc-VLP systems and minireplicons are useful tools for studying and dissecting individual steps of the viral multiplication cycle. In the case of CCHFV as well as of other highly pathogenic viruses (18, 20, 22, 34, 35), they confer the additional advantage of abrogating the need for high-containment facilities, since they are nonmultiplying and lack authentic virus genes (8). CCHFV tc-VLPs were capable of primary transcription, which is detectable when the REN-Luc reporter is expressed via the strong L segment promoter. This demonstrates that the L polymerase packaged into the VLPs is active and confirms previous observations on the high activity of the bunyaviral L promoter (14, 16, 36, 37). Using a tc-VLP system for RVFV, we have previously shown that primary transcription is sufficient to trigger innate immune responses mediated by the pathogen recognition receptor RIG-I (38) and that the antiviral host cell protein MxA is targeting RVFV primary transcription (18). Given that the CCHFV genome is lacking the RIG-I-activating 5′ triphosphate group (39) and that CCHFV is also inhibited by MxA (40), the tc-VLPs presented here will complement the set of available systems to study and compare bunyavirus biology with respect to antiviral responses, requirements for RIG-I activation, and transcription activity.

Expression of the nucleocapsid proteins L and N in trans strongly increased reporter activity by tc-VLPs, indicating a switch to genome replication mode. Moreover, the additional supply of GPs further increased reporter activity and led to formation of second-generation VLPs that reinfected the cells. Consequently, repeated passaging of tc-VLPs allowed us to substantially increase VLP yields. To eliminate the need for repeated plasmid transfections, we are aiming toward the generation of a packaging cell line expressing all structural proteins, similar to what was reported for RVFV (27).

VLPs are used for numerous downstream applications, e.g., diagnostics and vaccines. For this, it is important to purify, concentrate, and store them without losing stability. Purification by polyethylene glycol (PEG) precipitation or by ultracentrifugation through a sucrose cushion allows concentration of the VLPs without substantial loss of activity. Biological activity was also maintained after freezing the crude VLP supernatants at −20°C or −80°C for 3 days, whereas positive temperatures were destructive. At −80°C, the crude VLPs could be frozen for longer periods without losing much activity.

Direct diagnosis of CCHFV infection is based either on virus isolation or on molecular detection of the genome (41–45). However, the period in which the virus can be detected in patients is rather short. Therefore, most of the current diagnostic assays are based on serology. Indirect immunofluorescence assays (IFAs) and enzyme-linked immunosorbent assays (ELISAs) are the most reliable and useful methods, since they can detect and differentiate IgM and IgG antibodies (46). Inactivated CCHFV particles are expected to be the best material for such assays, but they require a BSL-4 facility to produce. Our structurally similar tc-VLPs provide a useful alternative, as attested by the presented assay to determine neutralizing antisera.

The tc-VLP production/infection system allowed us to verify the prediction by Morin et al. that the endonuclease domain of nairoviruses is located around amino acid D693 in the L sequence (21). The L alanine-substitution mutant at this position has entirely lost the ability to transcribe mRNA but has retained the ability to replicate the minigenome and assemble nucleocapsids that could be packaged into virions. D693 coaligns with D79 of LACV, D111 of RVFV, and D89 of arenaviruses. Mutants of these corresponding amino acids were shown to display a selective loss of transcription activity similar to that of CCHFV L D693A (21, 22, 32). It is hence likely that D693 inhabits the same key position in the active center and complexes the two Mn²⁺ ions that are necessary for endonucleolytic cleavage of host mRNAs. The D693A mutant will now enable us to study the RNA replication activity of CCHFV independently of mRNA transcription.

In summary, we have established a tc-VLP system which enables studies of the CCHFV infection cycle under BSL-2 conditions and may help to develop vaccines and antivirals. We are confident that this system will accelerate future research on the highly pathogenic CCHFV.