INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Classical swine fever virus (CSFV) is the causative agent of a highly contagious disease in pigs. Virulent strains cause characteristic disease symptoms such as fever, neurological disorders, hemorrhages, and high mortality rates, whereas infection with avirulent strains remains clinically inapparent but induces a protective immunity. Furthermore, prenatal infection of fetuses can lead to persistently infected animals which shed virus over a long period, as do pigs which exhibit the chronic form of disease (3, 4, 29, 31).

CSFV, bovine viral diarrhea virus (BVDV), and border disease virus form the pestivirus genus within the family Flaviviridae (33). CSFV is a small enveloped virus containing a positive-strand RNA genome of 12.3 kb, which consists of a 5' untranslated region (5'UTR), a large open reading frame (ORF) encoding a single polyprotein, and a 3' untranslated region (3'UTR) (16, 29). The 5'UTR contains an internal ribosomal entry site for cap-independent translation initiation of the viral polyprotein which is co- and posttranslationally processed by host cell and viral proteases (reviewed in reference 16). The cleavage sites of the polyprotein have been determined for BVDV, except for p7-NS2 (5, 16, 23, 26). C, E^rns, E1, and E2 represent the structural proteins found in mature virions. As demonstrated for other flaviviruses (2), RNA replication of pestiviruses is thought to occur on cytoplasmatic membranes via the synthesis of a negative-stranded full-length genome (6, 7, 23), while the components of the viral replication complex have not been identified. It has been shown recently that the viral protein NS5B of BVDV has RNA-dependent RNA polymerase activity (37). The NS3 protease is responsible for the cleavage of the viral polyprotein downstream of NS3 and requires NS4A as a cofactor (28, 35). Furthermore, it possesses both helicase and NTPase activity (27, 32), suggesting a role in RNA replication.

According to their behavior in tissue culture pestiviruses can be divided into two biotypes, noncytopathogenic (NCP) and cytopathogenic (CP). The mechanism responsible for the cytopathic effect (CPE) is poorly understood, but the overexpression of NS3 is a common feature of all CP pestiviruses (16). Furthermore, it has been shown that cells infected with CP BVDV undergo apoptosis (10, 36). CP pestiviruses represent mutants which arise from NCP viruses, presumably by RNA recombination during replication (16). Several mutants have been described for BVDV. Most of them contain rearrangements of viral sequences and/or insertions of host cell sequences (reviewed in reference 16). However, some CP BVDV and all CP CSFV isolates described so far are composed of defective interfering (DI) particles and NCP helper virus. CSFV DI genomes, with the identical deletion of 4,764 nucleotides (nt) corresponding to the genes encoding N^pro through NS2, have been isolated from different sources (12, 14, 17). A genome with a similar deletion of 4,746 nt has been described recently (12). Furthermore, a defective genome was reported which lacks the 4,263 nt encoding C through NS2 but retains the foreign murine ubiquitin gene which had been used to replace N^pro at the 5' terminus of the ORF in the parental genome (30). After transfection into susceptible cells, such defective CSFV genomes, obtained either by extraction from infected cells or by transcription in vitro from the respective cloned cDNA, replicate and are packaged efficiently in the presence of helper virus RNA, thereby causing a CPE (14, 15).

An autonomously replicating defective BVDV genome which lacks the genes encoding C, E^rns, E1, E2, p7, and NS2 has been described recently by Behrens et al. (1). It demonstrates that none of these proteins is essential for RNA replication. In the same study it was suggested that the 5' terminal region of the ORF contains a cis signal required for RNA replication, since the N^pro gene could neither be deleted completely nor be replaced by a ubiquitin gene. However, the replacement of the N^pro gene by ubiquitin in the CSFV Alfort/187 genome yielded infectious virus (30).

In this report a series of in vitro-constructed defective CSFV genomes derived from strain Alfort/187 were analyzed with respect to autonomous RNA replication in SK-6 cells. The question of whether these defective genomes can be packaged in the presence of a helper virus and whether they are cytopathogenic was also addressed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and viruses. The swine kidney cell line SK-6 (11), kindly provided by M. Pensaert (Faculty of Veterinary Medicine, Ghent, Belgium), was propagated in Dulbecco's modified Eagle medium supplemented with 5% horse serum. CP CSFV vA187-1 used as a control was derived from a persistently infected SK-6 cell culture in which DI particles had been spontaneously generated (17).

Plasmid constructs. Plasmids were amplified in Escherichia coli XL-1 Blue cells (Stratagene). Restriction enzymes were purchased from New England Biolabs except for SrfI (Stratagene). Plasmid DNA was purified with the Wizard Mini- or Maxiprep kit (Promega). Primers used for reverse transcription (RT) and PCR are shown in Table 1. Mutant CSFV genomes shown in Fig. 1 were constructed on the basis of the full-length cDNA clones pA187-1 (25) and pA187-CAT (20). CSFV-specific restriction sites are indicated by superscript numbers corresponding to their positions on the genome of vA187-1 (25).

View larger version (31K): [in this window] [in a new window]   FIG. 1.   Schematic display of plasmid-encoded full-length and defective genomes of CSFV. (A) Full-length genomes of A187-1 and A187-CAT. The 5'- and 3'UTRs are depicted as black lines, and the open reading frame is shown as a box composed of the individual viral genes encoding structural (hatched) and nonstructural (white) proteins as indicated. Vertical lines indicate the SrfI12298 and the NruI11301 sites which were used to linearize the respective plasmids prior to in vitro RNA transcription. The CAT marker gene inserted in A187-CAT at nt 386 of the genome is shown as a separate box (black). (B) Defective genomes. The position and extent of the individual deletions (dashed lines) are given in parentheses for each construct. Numbers refer to the nucleotide sequence of CSFV Alfort/187 (25). The ubiquitin gene is shown as a black box.

(i) pA187-p7, pA187-3390, and pA187-4263. The ExSite PCR-based site-directed mutagenesis kit (Stratagene) and primer pairs NS2-LMfe and E2-RMfe (p7), NTR-R and NS2-L (3390), and N^pro-R and NS3-L (4263) were used to construct these deletion mutants. In the following description, CSFV-specific restriction endonuclease sites are given with their positions (indicated by superscript numbers) in the viral genome. Since the cleavage site between p7 and NS2 and, therefore, the exact 5' terminus of the NS2 gene is not known, the 3' borders of the deletions in the constructs A187-3390 and -p7 were chosen within the genome region suggested by Elbers et al. (5) to contain the respective cleavage site. Two pBluescript-derived subclones of pA187-1 were used as templates for mutagenesis PCR: pBS-E/K-A187, containing the EagI⁸² to KpnI⁴⁴⁵³ fragment of pA187-1, and pBS-T7G1P (25), containing the EagI⁸² to BamHI⁶⁴³⁷ fragment of pA187-1. Deletions and flanking regions were confirmed by DNA sequencing. Fragments AflII³³⁹⁶ to KpnI⁴⁴⁵³, EagI⁸² to KpnI⁴⁴⁵³, and BspDI⁷⁷⁸ to NcoI⁵⁵³³, each carrying the appropriate deletion, were isolated and used to replace the corresponding fragment in pA187-1 to obtain pA187-p7, -3390, and -4263, respectively.

(ii) pA187-E2. The SpeI²⁴²⁹ to AflIII³⁰⁰⁰ fragment of pA871-1 was subcloned into pBluescript to obtain pBS-E2. This plasmid was cleaved with NheI²⁴⁴⁴ and EcoNI²⁹⁰⁶, treated with Klenow polymerase, and religated. A clone with an in-frame deletion ranging from nt 2443 to 2907 was identified by DNA sequencing, and the SpeI²⁴²⁹ to AflIII³⁰⁰⁰ fragment of this clone was used to replace the corresponding fragment in pA187-1 and pA187-CAT to produce pA187-E2 and pA187-E2-CAT, respectively.

(iii) pA187-Apa and pA187-Acc. Plasmid pBS-E/K-A187 was digested with ApaI^660,3260 or with AccI^445,4407 and religated. From the resulting plasmids, the respective EagI⁸² to KpnI⁴⁴⁵³ fragment, comprising either a 2,601-nt or 3,963-nt in-frame deletion, was replaced in pA187-1, to obtain pA187-Apa and pA187-Acc, respectively. Similarly, pA187-Apa-CAT and pA187-Acc-CAT were derived from pA187-CAT via the subclone pBS-E/K-A187-CAT.

(iv) pA187-4764 and pA187-4764Ubi. Defective CSFV genomes carrying these deletions were spontaneously generated in SK-6 cells persistently infected either with biologically cloned strain Alfort/187 (17) or vA187-Ubi (30). To generate molecular clones of these genomes, viral RNA was extracted from the respective cell culture medium and subjected to RT-PCR with the primers CSF-L001 and PR5 as described before (20). The resulting PCR products of 832 and 1,060 bp, respectively, were digested with EagI⁸² and NcoI⁵⁵³³ and used to replace the corresponding fragment in pA187-1, to obtain pA187-4764 and pA187-4764Ubi, respectively.

(v) pA187-4764BN. pA187-4764 was digested with BspEI⁵⁴⁶⁰ and NcoI⁵⁵³³, treated with Klenow polymerase, and religated. The presence of the predicted in-frame deletion of 69 nt within the protease domain of NS3 was confirmed by DNA sequencing.

In vitro transcription and electroporation. In vitro transcription from SrfI¹²²⁹⁸- or NruI¹¹³⁰¹-linearized plasmids was performed by using the T7 Megascript kit (Ambion) as described before (20, 25, 30). After DNase I digestion, transcripts were purified through MicroSpin S-400 HR columns (Pharmacia) and quantified with a GeneQuant II photometer (Pharmacia). SK-6 cells were washed twice and resuspended in ice-cold phosphate-buffered saline (PBS). A total of 2 × 10⁷ cells in a volume of 0.8 ml were mixed with 15 µg of RNA, transferred to a 0.4-cm cuvette (Bio-Rad), and electroporated immediately by using a Gene Pulser (Bio-Rad) set at 450 V and 500 µF. Alternatively, 0.4 ml of cell suspension at a density of 10⁷ cells/ml was mixed with 5 µg of RNA, transferred into a 0.2-cm cuvette, and electroporated twice at 200 V and 500 µF. After electroporation the cell suspension was kept for 10 min at room temperature, then diluted in Dulbecco's modified Eagle medium containing 5% horse serum, seeded, and harvested for analysis at different times after electroporation.

Cell staining. After electroporation 4 × 10⁵ cells per well were seeded in 24-well plates and heat fixed at the indicated time as described before (17). For the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labelling (TUNEL) assay performed with the in situ cell death detection kit (Boehringer Mannheim) cells were seeded in 96-well plates at a density of 4 × 10⁴ cells per well. Fixation and staining were performed as suggested by the manufacturer. After heat fixation or TUNEL assay the cells were incubated overnight at 4°C with primary antiviral antibody diluted in PBS containing 0.01% Tween 20. Monoclonal antibodies (MAb) C16 (9, 21) and HC/TC 26 (8) were used for the detection of NS3 and envelope protein E2, respectively. Bound primary antibody was labelled with biotinylated anti-mouse antibody (LSAB kit; DAKO) and streptavidin-horseradish peroxidase conjugate (LSAB kit; DAKO) and visualized by adding chromogen solution (17). The cells were examined by fluorescence and bright-field microscopy to identify TUNEL-positive and immunostained cells, respectively.

RNA analysis. Total RNA was extracted from 5 × 10⁶ electroporated cells with Trizol reagent (Gibco). Irrespective of the RNA concentration, one-third of each sample was used for Northern blotting in order to standardize the samples based on the number of electroporated cells. Northern blotting and hybridization were performed as described before (17) with ³²P-labelled riboprobes: either JL1, which is complementary to the 3'-terminal 204 nt, for the detection of positive-stranded viral RNA of the CSFV Alfort/187 genome, or GM, corresponding to nt 327 to 582 of the CSFV Alfort/187 genome for the detection of negative-stranded RNA.

Protein analysis. SK-6 cells were lysed in a hypotonic buffer (20 mM MOPS [morpholinepropanesulfonic acid], 10 mM NaCl, 1.5 mM MgCl₂, 1% Triton X-100 [pH 6.5]), and the extracts were used either for chloramphenicol acetyltransferase-enzyme-linked immunosorbent assay (CAT-ELISA) (Boehringer Mannheim) or for Western blotting as described before (17, 20). Porcine antipestivirus hyperimmune serum N8T12 (19) and MAb 49DE directed against pestivirus NS3 (kindly provided by E. Peterhans, Institute of Veterinary Virology, University of Bern, Bern, Switzerland) served for the detection of NS3.

Packaging of replicon RNA. After electroporation of defective genomes together with full-length helper A187-CAT RNA, 5 × 10⁶ SK-6 cells were incubated for 48 h before the virus was harvested by freezing and thawing of the cultures. After low-speed centrifugation, 1 ml of undiluted supernatant was used to infect 2 × 10⁶ SK-6 cells seeded the day before and the inoculum was replaced after 1 h. An aliquot of the cell culture medium was collected 48 h after infection for RNA extraction and RT-PCR, which was performed as described before (20).

CPE assay. CPE induced by replicon RNA was monitored on SK-6 cells seeded at a density of 1.5 × 10⁵ (for subsequent infection) or 4 × 10⁵ (after electroporation) per well in 24-well plates. For infection, virus obtained either from the cultures in which the packaging had been performed or from cultures in which the respective packaged replicons had been passaged once or twice was used. Undiluted supernatant obtained after centrifugation of the respective frozen and thawed cultures served as the inoculum, which was replaced with fresh medium 1 h after infection. The cells were examined twice a day by light microscopy for the development of a CPE. After 48 or 72 h the medium was removed, and the cells were fixed and stained by the addition of a crystal violet solution (0.4 mg of crystal violet per ml and 0.37% formaldehyde in PBS) for 1 h to visualize the CPE macroscopically.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Replication competence of defective genomes. The defective genomes shown in Fig. 1 were transcribed in vitro from the respective plasmids, and 15 µg of each RNA was transfected into 2.5 × 10⁷ SK-6 cells. Subsequently, the cells were split in order to perform immunostaining (Fig. 2), to analyze viral RNA (Fig. 3 and 4) and viral protein expression (Fig. 5) at different times after electroporation.

View larger version (71K): [in this window] [in a new window]   FIG. 2.   Expression of NS3 from CSFV replicons. SK-6 cells were stained for NS3 expression 16 and 44 h after electroporation of in vitro transcripts obtained from the respective SrfI-linearized plasmid DNA. A, A187-p7; B, A187-E2; C, A187-Apa; D, A187-3390; E, A187-Acc; F, A187-Apa transcript from NruI-linearized plasmid; G, A187-4263; H, A187-4764Ubi; I, A187-4764; K, A187-4764BN; L, mock.

View larger version (89K): [in this window] [in a new window]   FIG. 3.   Northern blot analysis of total RNA extracted from SK-6 cells 1 h (A), 20 h (B and D), and 48 h (C) after electroporation with the indicated in vitro transcripts. Total RNA from 1.7 × 106 cells, determined before electroporation, was analyzed in each lane. Two riboprobes were used, probe JL1 (17) specific for positive-stranded viral RNA (A, B, and C) and probe GM specific for viral negative strands (D). The specificity of the latter probe is illustrated by hybridizing the probe to 100 ng of in vitro transcripts of A187-1, of either positive (+) or negative () polarity (D).

View larger version (39K): [in this window] [in a new window]   FIG. 4.   RT-PCR of total RNA extracted from SK-6 cells 48 h after electroporation with the indicated in vitro transcripts. Prior to RT-PCR with primers HL3 and HR3, the RNA was digested with DNase I. The CSFV-specific PCR product of 241 bp is located within the NS3 gene.

View larger version (53K): [in this window] [in a new window]   FIG. 5.   Western blot analysis of SK-6 cell lysates. Analysis was performed 20 h after electroporation with the indicated in vitro transcripts. Each lane represents the extract from 5 × 105 cells. An extract from the same number of SK-6 cells obtained 20 h after infection with CP vA187-1 served as a positive control. Pig hyperimmune serum N8T12 (A) and MAb 49DE directed against pestivirus NS3 (B) were used for the detection of NS3.

View larger version (26K): [in this window] [in a new window]   FIG. 6.   CAT expression from CSFV replicons. CAT protein measured by ELISA was standardized to 106 cells, determined before electroporation. (A) SK-6 cells were lysed 20 h after electroporation with the indicated in vitro transcripts, and the extracts were analyzed by ELISA. Results of immunoperoxidase (IPA) staining for NS3 carried out in parallel cultures are indicated at the bottom (+ and ). (B) Kinetics of CAT expression. After electroporation of the respective transcripts, the cells were split into six aliquots, which were seeded in separate culture flasks. Protein extraction and subsequent CAT-ELISA was carried out 5, 20, 30, 45, 58, and 68 h after electroporation.

Packaging of replicons in virus particles. To show the packaging of viral RNA, SK-6 cells were coelectroporated with helper A187-CAT RNA and the respective defective RNAs. At 48 h after transfection the cell-free supernatants were collected and used to infect SK-6 cells. The medium of these cultures was collected 48 h postinfection. The RNA was extracted and used to perform RT-PCR with the CSFV-specific primers Pest1 and PR5. A 5.8-kb product representing the A187-CAT helper genome was amplified from all samples except for those containing either A187-4764Ubi or A187-4764 RNA, which both had caused an extensive CPE. Shorter PCR products corresponding in length to the input replicon RNAs were detected for A187-Apa, -4263, -4764Ubi, and -4764 (Fig. 7A). An additional PCR specific for the marker gene of the helper virus confirmed the presence of vA187-CAT in all samples derived from coelectroporations with the respective in vitro transcript (Fig. 7B).

View larger version (83K): [in this window] [in a new window]   FIG. 7.   Packaging of defective RNA into CSFV particles. The indicated defective genomes were electroporated together with helper A187-CAT RNA into SK-6 cells. Virus was harvested after 48 h and used to infect SK-6 cells. At 48 h postinfection, viral RNA was extracted from the clarified cell culture medium and assayed by RT-PCR. Cytopathogenic vA187-1, which is composed of standard CSFV particles and DI particles carrying the 4764-type defective RNA (17), served as a positive control. The predicted lengths of the amplified products are indicated for the defective RNA (left) and for the helper virus RNA (right). (A) PCR with primers flanking the deletions of defective genomes. (B) PCR specific for the CAT marker gene of the helper virus.

Cytopathogenicity of replicons. Cell morphology and immunostaining after electroporation suggested that the replicons A187-4263, -4764Ubi, and -4764 were cytopathogenic (Fig. 2G, H, and I). To further characterize the cytopathogenic nature of replicons, SK-6 cells were electroporated with the respective RNAs and examined 20 h later by TUNEL assay for fragmentation of genomic DNA as a marker for cell death as well as for the expression of NS3. A high percentage of the cells scored positive in the TUNEL assay in all cultures including mock-electroporated cells (Table 2). This was likely due to the electroporation procedure, which causes extensive damage of the cells. Thus, the defective genomes were packaged into virions by coelectroporation of SK-6 cells with helper A187-CAT RNA as described above. As expected, CPE was observed in cells coelectroporated with helper RNA and, more dramatically, after passage of the supernatants for the transcripts A187-4263, -4764Ubi, and -4764 but not for any of the other defective RNAs or for helper RNA alone (Table 2).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In order to define the genes required for RNA replication of CSFV, a series of defective genomes carrying deletions within the 5' half of the ORF was constructed, based on the infectious full-length clones pA187-1 and pA187-CAT. The genes encoding N^pro through NS2 proved to be dispensable for RNA replication, since A187-4764 RNA replicated autonomously. As replicon RNA with partially or completely deleted structural genes is not expected to spread from cell to cell, a highly efficient transfection procedure was required to detect low levels of RNA replication. In fact, for the defective genomes A187-Apa, -4764Ubi, and -4764, more than 50% of the cells stained positive for NS3 16 h after electroporation (Fig. 2C, H, and I), and direct evidence for replication was obtained by the detection of negative-stranded RNA of these genomes (Fig. 3D). In contrast, no NS3-positive cells were observed after electroporation of a 3' truncated transcript from pA187-Apa. Lower proportions of cells staining positive for NS3 were observed for A187-E2, -3390, and -4263 RNA. We conclude that these mutant genomes are replicons as well and that negative-stranded RNA could not be detected due to a less efficient transfection and/or replication of these genomes. Furthermore, in cells transfected with A187-E2-CAT or A187-Apa-CAT the marker protein CAT was expressed but not after transfection of the respective 3' truncated RNAs or A187-Acc-CAT (Fig. 6). This confirms the result obtained with immunostaining for NS3 and shows that translation from input RNA does not yield detectable amounts of protein. The concentration of CAT for the replicons A187-E2-CAT and A187-Apa-CAT was only slightly above the detection limit of the ELISA and about 100-fold lower than for infection of SK-6 cells with vA187-CAT (19). Nevertheless, this shows that CAT is a useful marker for the detection and quantification of replication of mutant CSFV genomes.

The RNAs A187-p7, -Acc, and -4764BN did not replicate at detectable levels (Fig. 2). This was expected for the latter construct, since the serine protease domain is essential for replication (13, 35). In contrast, the constructs A187-p7 and A187-Acc contain all genes required for replication (Fig. 1). We speculate that these RNAs might have an altered secondary structure which affects their ability to serve as a template for RNA replication or translation, since sequencing the template plasmids revealed no mutations which could account for a failure to replicate. Alternatively, correct processing of the viral polyprotein might be affected by the deletions in the E2-p7-NS2 region, yielding nonfunctional proteins. Although the amount of input RNA was standardized, the transfection efficiency as reflected by the proportion of NS3-positive cells 16 h after transfection varied from below 1% to nearly 100%, depending on the construct (Fig. 2). Furthermore, the amount of RNA detected in cells 1 h after transfection varied strongly, indicating that the stability of the respective transcripts varies. We exclude the possibility that the RNA detected at this time by Northern blot analysis represents newly synthesized RNA. Even after infection of SK-6 cells with wild-type CSFV at a multiplicity of infection of 10 and by using the same detection procedure, we have never observed viral RNA earlier than 6 h after infection (18). Differences in RNA stability might also be responsible for the strong variation in the transfection efficiency between individual constructs (Fig. 2). In addition, it has to be considered that in vitro transcripts of different constructs do not necessarily yield equal proportions of RNA molecules with a correct, replication-competent secondary structure. Thus, we conclude that the number of input RNA molecules required to start replication in a single cell varies according to the nature of the particular construct.

We cannot exclude that in the case of constructs with low transfection efficiency, such as A187-3390, the few cells which stain positive for NS3 (Fig. 2D) contain RNA with an altered sequence compared to its parental template plasmid. In this case, the replicating RNA might have acquired mutations either during in vitro transcription by T7 RNA polymerase or during initial low-level replication in cells. To obtain sufficient material for sequencing and identification of such mutations, RT-PCR amplification of these genomes would be required. However, virus-specific RT-PCR yielded strong signals 48 h after electroporation (Fig. 4), even for genomes unable to replicate such as A187-4764BN or A187-Apa RNA lacking its 3'UTR. This demonstrates that input RNA remains detectable over a long time and, therefore, that sequencing of RT-PCR products would be biased.

Interestingly, two types of RNA replicons were found with respect to their replication kinetics and their effect on host cells. The defective genomes A187-E2, -Apa, and -3390 represent the noncytopathogenic replicon type, whereas A187-4263, -4764Ubi, and -4764 are cytopathogenic replicons. NCP replicons encode NS2-NS3, replicate at low levels, and persist in transfected cells for at least 10 cell passages (data not shown). NS3 expressed by NCP A187-Apa was detectable by immunostaining in the majority of the electroporated cells and barely on Western blots, indicating that the expression level is significantly lower than that for CP replicons or virus infection (Fig. 5). On the other hand, CP replicons which do not contain the NS2 gene replicate very efficiently and express large amounts of NS3 (Fig. 2 and 5), the latter being the common feature of CP pestiviruses (16). No functions of pestivirus NS2 have been identified so far (16), but in the closely related hepatitis C virus it was suggested to be involved in the processing of NS2-NS3 (34). Our findings indicate that CSFV NS2 has a regulatory function in RNA replication; nonetheless, it is not essential. Possibly, NS2 prevents an uncontrolled accumulation of viral RNA and proteins leading to host cell death. It is not clear whether mature NS2, a precursor protein, e.g., NS2-NS3, or both could account for this function. Alternatively, the NS2 gene might act at the RNA level, either by altering the secondary structure or by binding modulating factors.

The CP replicons described here correspond in their overall genome structure to the naturally occurring defective CSFV RNAs (12, 14, 17, 30), which carry a large deletion comprising most of or the complete N^pro gene, and the coding sequences for all structural proteins, p7, and NS2. We report an additional CP CSFV RNA (A187-4263), which by analogy to the BVDV replicon CP9 (1, 16), contains the complete N^pro gene. Remarkably, three different genes with no sequence homology, namely, N^pro, murine ubiquitin, and NS3, are located directly downstream of the translation start signal of the respective CP replicons A187-4263, A187-4764Ubi, and A187-4764. This supports our earlier suggestion that the internal ribosome entry site of CSFV Alfort/187 does not overlap with the ORF (30) but is in contrast to what has been reported for CSFV strain C (24) and for BVDV strains NADL (22) and CP9 (1). Upon coelectroporation with a full-length marker genome, the CP replicons were packaged efficiently by the structural proteins provided in trans from a helper genome, giving rise to CP CSFV (Table 2). Analysis of the NCP replicons revealed detectable amounts of packaged RNA for A187-Apa but not for A187-E2 and A187-3390. This is not surprising since the transfection efficiencies of the latter two RNAs are very low when compared to A187-Apa (Fig. 2). The RT-PCR data (Fig. 7) further indicate that the replicons propagated in the presence of helper virus retained their overall genome structure upon passage, although minor mutations cannot be ruled out. We consider that the transfection efficiency rather than a specific signal on the RNA is critical for the packaging of the replicons described here.

Future studies will aim at a more detailed characterization of NS2 and its role in viral replication. Also, CSFV replicons packaged in virions provided in trans from a heterologous expression system will allow helper virus-free experiments. Finally, CSFV replicons may serve as a model for the investigation of hepatitis C virus replication in cell culture.