INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Canine distemper virus (CDV) is an enveloped virus with a monopartite negative-stranded RNA genome. Together with Measles virus (MV) and Rinderpest virus, it belongs to the morbilliviruses which form a serologically closely related genus in the family Paramyxoviridae. CDV primarily affects dogs, but infections of other terrestrial carnivores, in both captivity and the wild, have been reported (1, 2, 18, 21, 22, 25, 26, 29, 33, 36). The mortality rates associated with CDV infection vary among susceptible species and range from 0% in domestic cats to 50% in dogs and 100% in ferrets. One of the currently available vaccines (Onderstepoort strain) efficiently protects dogs, but it is insufficiently attenuated for other species, and high levels of mortality can occur due to its remaining virulence (10). Thus, there is a need to develop more attenuated vaccines for CDV to fully protect susceptible animals.

A rescue system for CDV would provide the means to study attenuating effects of defined mutations and might subsequently facilitate generation and examination of new vaccines. Mutations that cause persistence of the virus could also be determined. This system would be useful in gaining a better understanding of CDV infection which can more easily be studied in cell culture and animal models by coexpression of additional reporter genes (e.g., the enhanced green fluorescent protein) from the recombinant viral genome (19, 20).

The CDV genome is 15,690 nucleotides (nt) in length and consists of a short 3' leader region and six genes encoding the N nucleocapsid (N), phospho- (P), matrix (M), fusion (F), hemagglutinin (H) and large (L) proteins (4, 5, 7, 16, 32, 37, 42). They are separated by intergenic regions of 3 nt and are followed by a short 5' trailer region. The nonstructural proteins V and C are encoded within the P gene. V is expressed by cotranscriptional RNA editing, and C is expressed from an overlapping reading frame (8, 12). The genome of paramyxoviruses does not consist of a naked RNA molecule. In many members of the subfamily Paramyxovirinae, 6 nt are likely to be tightly associated with N protein (9, 30). The exact structure of these ribonucleoprotein complexes is not known but it has recently been suggested for vesicular stomatitis virus that the N protein binds to the sugar-phosphate backbone of the RNA, exposing the bases to the outside. As a consequence, the RNA of the Mononegavirales may be transcribed without dissociation from the nucleoprotein (27). The phosphoprotein and the RNA-dependent RNA polymerase are also associated with the RNP. Paramyxovirus genomic or antigenomic RNA as for all Mononegavirales cannot function as mRNA. Thus, to initiate an infectious cycle by introducing genome analogues into the cell, all viral proteins involved in transcription and replication have to be provided in trans. Since the generation of infectious rabies virus from a cDNA clone in 1994 (40), several other negative-stranded RNA viruses have been rescued using a number of different approaches. To date two morbilliviruses, namely MV (35) and Rinderpest virus (3), have been recovered from cDNA. These viruses are closely related and distinct from CDV and phocine distemper virus (PDV) in terms of genome length and sequence homology (6, 17, 24, 34, 41).

In this study, we have established a rescue system for CDV. We generated a full-length cDNA clone of the CDV strain Onderstepoort [large plaque-forming variant (OND-LP)] and the helper plasmids encoding N, P, and L proteins. Recombinant virus (rCDV) was recovered from cell cultures transfected with all four plasmids. Immunofluorescence and a genetic tag identified rCDV. The growth characteristics of rCDV were compared with the original CDV strain.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and viruses. Vero cells were maintained in BHK medium supplemented with 8% newborn calf serum. HeLa cells were grown in RPMI 1640 medium supplemented with 10% fetal calf serum. All media and sera were obtained from Life Technologies/Gibco BRL. For transfections the cells were seeded into six-well trays and grown to approximately 80% confluence. For immunofluorescence Vero cells were grown on glass coverslips (diameter, 13 mm) to 100% confluence and infected with rCDV at a multiplicity of infection (MOI) of 0.1. For the growth analysis Vero cells were grown in 25-cm² flasks to 100% confluence and infected with rCDV or CDV Onderstepoort (OND-LP) at an MOI of 0.1. Virus was removed after an incubation of 2 h, and new medium was added. The samples of cell-associated and cell-free virus at time point 0 were collected immediately after addition of new medium and, after that, every 4 h. For a 50% tissue culture-infective dose, which was performed using standard methods, Vero cells were grown in 96-well trays until confluent. For phase-contrast microscopy, a confluent monolayer of Vero cells was infected with rCDV at an MOI of 0.5. When cytopathic effects (CPEs) were visible, cells were formamide fixed and stained with methylene blue. Both viruses, OND-LP and rCDV, were propagated in Vero cells. Cells and viruses were grown at 37°C under 5% CO₂.

Plasmid constructions. All cloning procedures were performed following standard protocols. PCR amplifications were carried out using Pwo (Boehringer) or Taq DNA polymerases (Life Technologies). PCR products were first cloned into pGEM-T (Promega) and then subcloned into pEMC vectors or pBS SK II (Stratagene). The vector backbone pEMC, used for cloning of coding sequences of CDV N, P, and L proteins, has been described elsewhere (35). The plasmids pEMC-Na, pEMC-Pa, and pEMC-La, which code for MV N, P, and L proteins, pCDV():CAT, and p107MV():CAT, were a kind gift from M. A. Billeter, University of Zürich.

View larger version (33K): [in this window] [in a new window]   FIG. 1.   Schematic representation of p(+)CDV: cloning procedure and full-length cDNA clone. (A) The multiple cloning sites of pBS SK II and adjacent T3/T7 promoters were excised from the plasmid with BssHII. The remaining plasmid then served as backbone for construction of p(+)CDV. The first cDNA coding for the T7 promoter, viral leader sequence, and mainly the N protein was ligated into the vector. The restriction enzyme MluI was used to block one BssHII recognition site during the cloning procedure (see panel C). (B) BssHII was used throughout the construction of p(+)CDV to cut at the 3' ends. (C) The full-length clone of 18,743 nt was generated in 12 separate cloning steps. The restriction sites used for cloning are indicated. As the genetic tag, a Csp45I site was introduced into the coding sequence of the L protein. The  ribozyme and T7 terminator sequences were required for generation of RNAs of exact length. The representation is not to scale.

Chloramphenicol acetyl transferase (CAT) assays. Transfections were performed following the same protocol as for rescue (below). Plasmids were used in the following amounts: 1.2 µg of pEMC-N, 1.2 µg of pEMC-P, 0.4 µg of pEMC-L, and 1.0 µg of pCDV():CAT or p107MV():CAT. At 20 h posttransfection (p.t.), crude cell extracts were generated and tested for CAT activity using standard protocols (Promega Protocols and Applications Guide). Incubation time for the enzyme assay was set to 2 h. The substrate [C¹⁴]chloramphenicol (50 mCi/mmol) was obtained from Amersham. Radioactivity of the butyrylated chloramphenicol products was measured in a liquid scintillation counter (LKB).

Virus rescue. Upon reaching a confluence of approximately 80% in six-well trays, HeLa cells were incubated in 2 ml of Optimem (Life Technologies) for 30 min. The cells were infected with the host range mutant of vaccinia virus Ankara expressing T7 polymerase (MVA-T7) at an MOI of 0.5 for 30 min. Prior to this infection, 8 µl of Lipofectin was mixed and incubated with 92 µl of Optimem following the supplier's instructions (Life Technologies). Plasmids [1.5 µg of pEMC-N, 1.5 µg of pEMC-P, 0.5 µg of pEMC-L, and 5.0 µg of p(+)CDV] were mixed into 100 µl of Optimem. The plasmid mixtures were then carefully pipetted into the Lipofectin. MVA-T7 was removed from the cells, and plasmids in Lipofectin were added to the HeLa cells. The cells were incubated with an additional 2 ml of Optimem at 37°C for 16 h to recover from the procedure before the transfection medium was exchanged with normal growth medium. At 48 h p.t., 5 × 10⁴ Vero cells per well were added. At 3 days p.t., 1 ml of the supernatant was removed and added to a confluent monolayer of Vero cells in a 75-cm² flask. At 6 days p.t., when the infection had progressed well, cells and supernatants were harvested and stored at 70°C.

Immunofluorescence. Vero cells were grown on glass coverslips and transfected as described above, with 1.2 µg of pEMC-N or pEMC-P to detect expression of either protein, or infected with rCDV at an MOI of 0.1. At 24 h postinfection (p.i.) or p.t., cells were fixed in ice-cold acetone and stored at 20°C. The primary serum of a patient with subacute sclerosing panencephalitis (SSPE) or monoclonal antibodies (MAbs) anti-N 2.78 and anti-P 2.98, obtained from N. Duffy, and anti-M 2.47, anti-F IC5, and anti-H 2.80, obtained from A. Trudget (both from The Queen's University of Belfast), were added in appropriate dilutions. Secondary fluorescein isothiocyanate (FITC)-labeled rabbit anti-mouse or rabbit anti-human antibodies (DAKO) were diluted 1:40 in phosphate-buffered saline. The procedure was performed using standard protocols. The nuclei were counterstained with propidium iodide (120 ng/µl), and the coverslips were mounted on slides with Citifluor (Amersham). The indirect fluorescence of recombinant viral proteins was examined under a confocal microscope equipped with a krypton-argon laser (Leica).

RNA isolation, cDNA synthesis, PCR, and automated sequencing. For viral RNA isolation, Vero cells were infected with OND-LP or rCDV at an MOI of 0.1. The cells were harvested in guadinium-isothiocyanate solution, and RNA was isolated using CsCl-gradient centrifugation (38). The total RNA was stored in 70% ethanol at 70°C. For first-strand cDNA synthesis, 1 µg of total RNA was used following the protocol supplied with AMV reverse transcriptase (Promega). The cDNA synthesis was primed with an oligo-dT_(12-18) primer or a primer complementary to base pairs 15,657 through 15,690, for cloning of the full-length cDNA p(+)CDV or CDV-FL-Avr II rev (Table 1), respectively, for amplification of the area carrying the genetic tag. PCR reactions were carried out following protocols supplied with Taq (Life Technologies) or Pwo (Boehringer) DNA polymerases. Cycling conditions were adjusted to templates and primers by varying the standard set-up by prolonging elongation time up to 2.5 min or increasing annealing temperature from 50 to 55°C.

Nucleotide sequence accession number. The CDV insert sequence of plasmid p(+)CDV is accessible under GenBank accession no. AF 305419.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of plasmids expressing recombinant CDV N, P, and L proteins. For successful rescue of most negative-stranded RNA viruses in the Paramyxoviridae and Rhabdoviridae, the N, P, and L proteins have to be provided and expressed from different plasmids to the vector coding for the full-length genome (13, 40). We generated three plasmids, pEMC-P, pEMC-N, and pEMC-L, to code for the respective proteins of CDV. The successful cloning of the coding sequences was verified by restriction endonuclease analysis and determination of nucleotide sequences of the respective plasmids (data not shown).

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Recombinant N and P proteins visualized by immunofluorescence. Vero cells (105) were transfected with 1.2 µg of pEMC-N or pEMC-P, acetone fixed at 24 h p.i. and subjected to immunofluorescence. (A) N protein was indirectly visualized with SSPE antiserum and FITC-labeled mouse anti-human secondary antibody. (B) Phosphoprotein expression was examined using an anti-PDV P primary and FITC-labeled rabbit anti-mouse secondary antibody. Nuclei were counterstained with propidium iodide (red). Photographs were taken with a confocal laser microscope (magnification, ×160).

Rescue of minigenome constructs. We tested the functionality of N, P, and L proteins by assessing their ability to rescue minigenome constructs (14) which contain the CAT reporter gene in antisense. The only viral sequences present are the leader and trailer regions of the genome. We cotransfected the plasmids encoding the N, P, and L proteins with such a minireplicon. Rescue events were monitored by assaying of CAT protein expression in crude cell extracts. Unfortunately, we found that active protein was produced from pCDV():CAT in transfected HeLa and Vero cells in the absence of pEMC-L. This meant that pCDV():CAT could not be used to test activity of N, P, and L proteins. To overcome this problem, we tested for functionality of the proteins using p107MV():CAT as a minireplicon. Unlike pCDV():CAT, the MV version of the plasmid did not give rise to CAT activity by itself (300 cpm) or when one of the helper plasmids was omitted. CAT activity was equally high when cells were transfected with a combination of CDV N, P, and L plasmids (22,500 cpm) as with MV-derived plasmids (23,100 cpm). Recombinant N, P, and L proteins were deemed to be functional and could be used in CDV rescue.

Construction of p(+)CDV. The cDNA clone p(+)CDV was generated using genomic RNA from the vaccine strain CDV OND-LP. A strategy was chosen using unique endonuclease recognition sites within the genomic sequence of CDV. This reduced the risk of introducing changes, which might impair growth of any recombinant virus. A plasmid of 18,743 kb was generated with a T7 promoter directly adjacent to the viral leader sequence, the antigenome of CDV followed by -ribozyme and T7 terminator sequences for correct termination of RNA molecules (Fig. 1C). After each cloning step, the resulting plasmids were screened for correct uptake of the new fragment and sequenced across the cloning site. The complete p(+)CDV was sequenced because Taq DNA polymerase was used for the majority of PCR amplifications. We confirmed that no major sequence changes had taken place. In addition, sequencing confirmed the correct construction of the plasmid and presence of the genetic tag within the coding region of L. Two more mutations were detected after comparison with published CDV Onderstepoort sequences (4, 5, 7, 16, 32, 37). One nt exchange was found in the M-F intergenic at nt position 4,724 (T to A), and one was detected within the coding region of the L at position 9,067 (A = T, L13 E = V).

Rescue of CDV. The CDV rescue system was based on the MVA-T7-mediated rescue established for MV by Schneider et al. (39). After the functionality of N, P, and L proteins had been determined, the protocol for rescue was optimized. HeLa cells were chosen for the rescue experiments, which were set up in triplicate. Negative controls without pEMC-L were included. The effect of MVA-T7 on the cells was observed in control experiments without plasmids or Lipofectin. At 2 days p.t., the HeLa cells started to detach, and in order not to lose any of the transfected cells, Vero cells were added. The obvious effect of this supplement was that the HeLa cells reattached to the substrate, and both cell types settled to form a monolayer. At 3 days p.t., 1 ml of each supernatant was added to confluent monolayers of Vero cells. The next day syncytia were clearly visible in the six-well trays. Due to the different potentials of HeLa and Vero cells for fusion, the syncytia did not appear as defined round structures but were rather diffuse in nature. Syncytia also formed in the flask which contained the 3-day p.t. supernatant, and these showed normal progression over the next 2 days. Rescued virus (rCDV) was then propagated in Vero cells. Preliminary observations revealed no impairment in the progression of infection, and CPEs looked like normal syncytia caused by OND-LP infection (Fig. 3A). Control transfections without pEMC-L never gave rise to infectious virus, and CPE did not appear. In those HeLa cells which were infected only with MVA-T7, no CPE could be observed, other than that cells also started to detach at 2 days p.i. This indicated that the detachment of the HeLa cells at that time was a result of the MVA-T7 infection rather than of the transfection procedure. The virus has been rescued successfully in every one of five attempts.

View larger version (97K): [in this window] [in a new window]   FIG. 3.   Immunofluorescence of rCDV. Vero cells were infected with rCDV at an MOI of 0.1 and acetone fixed at 24 h p.i. when CPEs of different stages of progression were visible. Expression of viral proteins was detected with specific MAbs, and nuclei were counterstained with propidium iodide. Staining for the N (B), P (C), M (D), F (E), and H (F) proteins is shown by the green fluorescence of FITC-labeled rabbit anti-mouse antibodies. (A) Light-microscopic photograph of a syncytium in an infected native monolayer of Vero cells stained with methylene blue (magnification, ×160).

Immunofluorescence of rCDV. Indirect fluorescence with antibodies specific to all viral proteins except the L polymerase was used to characterize rCDV and to confirm that the infectious virus expressed all major structural proteins. Two MAbs (anti-F IC5 and anti-M 2.47) specifically recognized CDV proteins and did not cross-react with PDV or MV structures, while the N, P, and H MAbs cross-reacted with PDV and MV. Vero cells grown on coverslips were infected with passage-four rCDV at an MOI of 0.1 and fixed at 24 h p.i., when syncytia of different stages were visible. The green fluorescence of indirectly stained viral proteins was observed by confocal laser microscopy (Fig. 3). We demonstrated that N, P, M, F, and H proteins were expressed in infected cells.

Detection of the genetic tag of rCDV. The rescued virus was characterized to prove its distinctiveness from the parental Onderstepoort vaccine strain. During the cloning procedure, a genetic tag was introduced into the coding region of the L protein. We determined the presence of this artificial Csp45I recognition site by two different methods. The region containing the tag was amplified by reverse transcription-polymerase chain reaction (RT-PCR) from RNA of passage-four virus with primers generating a 1,733-bp product. A PCR product of this size was separated by agarose gel electrophoresis, and following treatment with endonuclease Csp45I, two fragments 354- and 1,379-bp in size were visualized. RNA obtained from CDV Onderstepoort was amplified by RT-PCR and treated with Csp45I as a control. As expected the PCR product could not be cut using this enzyme. To ensure that no plasmid contamination was carried over from the transfections, rCDV RNA was directly used for PCR amplification without the initial RT step. No PCR product was obtained (Fig. 4A). Second, we sequenced the region containing the genetic tag directly from first-strand cDNA, derived from genomic RNA of the same virus stock. The results clearly showed that the two nucleotide exchanges were present in the genomic RNA of rCDV (Fig. 4B).

View larger version (46K): [in this window] [in a new window]   FIG. 4.   Characterization of rCDV. Vero cells were infected with passage-four rCDV and harvested for RNA isolation when the monolayer was fused. First-strand cDNA was synthesized with AMV-RT and used either for sequencing or for PCR amplification of the area containing the genetic tag. (A) Shown is a photograph of a 1.5% agarose gel with separated PCR products either cut with endonuclease Csp45I or undigested. In lanes 1 and 8, molecular weight markers have been separated ( EcoRI/HindIII). The cDNA of rCDV was digested with Csp45I (lane 2) or untreated (lane 3). RNA of CDV Onderstepoort was used for PCR amplification and separated either after treatment with Csp45I (lane 4) or after no treatment (lane 5). For lanes 6 and 7, RNA of rCDV was subjected to PCR without the RT step as a control for plasmid contaminations. Shown is a chromatogram of the sequence analysis of the area containing the genetic tag (B). Arrows indicate the two nt changes.

Growth kinetics of rCDV. In order to confirm the general observation that rCDV did not show any alterations in growth compared to the Onderstepoort vaccine strain, we analyzed a growth curve. Vero cells were infected at an MOI of 0.1, and supernatant and cell-associated virus was collected every 4 h. At 20 h p.i., CPE was clearly visible, and by 36 h p.i. the whole-cell monolayer was fused. After 52 h the cell monolayer started to detach, and it was no longer possible to separate virus in supernatant and cell-associated virus. Virus titers were estimated for each time point in triplicate (Fig. 5). The rescued virus grew to equivalent titers, and the growth kinetics showed no significant difference.

View larger version (25K): [in this window] [in a new window]   FIG. 5.   Growth analysis of rCDV and CDV Onderstepoort separated into cell-associated virus and virus in supernatant. Vero cells were infected at an MOI of 0.1, and virus was collected every 4 h over a period of 52 h. Shown are the mean values of three separate 50% tissue culture-infective doses for CDV Onderstepoort and rCDV as virus in supernatant(s) and cell-associated virus (ca).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have established a rescue system that allows the generation of infectious recombinant CDV from a full-length cDNA clone p(+)CDV. Plasmids coding for the viral proteins N, P, and L, which are indispensable in replication and transcription, were constructed. Expression of N and P proteins was shown by means of immunofluorescence. The functionality of N, P, and L proteins was demonstrated in minirescue of CAT protein as a reporter. When the plasmid pCDV():CAT was used, it gave rise to CAT activity without the normally required L protein, indicating the presence of promoter-like elements. The same phenomenon was observed when using the plasmid pCDV():CAT-3, which lacks an extra stop codon at the end of the CAT open reading frame and as such does not conform to the rule of six (9). This indicated that RNA replication was probably not required to generate this enzyme activity. Hence, detailed studies of the CDV promoter would be difficult using this system. Because p107MV():CAT could be rescued with CDV helper proteins, we did not pursue the problem. Successful minirescue with p107MV():CAT confirmed that CDV N, P, and L were expressed as functional proteins and interestingly demonstrated that the helper proteins of one morbillivirus could be used to rescue the minigenome constructs of another. It has been shown for other members of the Paramyxoviridae, namely Sendai virus and bovine respiratory syncytial virus, that minigenome analogues can be rescued by closely related viruses (15, 46).

A novel cloning strategy was used to assemble the plasmid p(+)CDV, which may find wider applications. Instead of cloning entire genes and having to manipulate the nt sequences in order to generate the plasmid, unique restriction endonuclease recognition sites within the viral genome were used such that the plasmid did not contain any sequence changes except for the genetic tag. Additionally, the endonuclease recognition site BssHII was introduced into reverse primer sequences used to amplify cDNAs to allow sequential assembly of p(+)CDV. Sequence analysis of p(+)CDV confirmed that the tag, 2 nt changes which do not cause any amino acid changes, had been inserted as designed and revealed two additional mutations. The first is situated in the noncoding region between M and F genes. The second mutation, which causes an amino acid exchange, was detected at the start of the L-coding sequence. This region is highly conserved among Paramyxoviridae (32). These mutations did not appear to have a detrimental effect on the growth of rCDV. The recombinant virus replicated normally in tissue culture and showed no differences when compared to the parental virus, as was confirmed by analysis of the growth kinetics of rCDV and OND-LP. The Onderstepoort vaccine strain causes mortality of up to 100% in ferrets (10, 43). It will be interesting to see whether rCDV has the same potential.

The T7 polymerase in our rescue system for CDV was delivered by MVA-T7, a derivative of the modified vaccinia virus Ankara. This host cell-restricted, highly attenuated mutant does not produce progeny in most mammalian cell lines and therefore does not interfere with recovery of rCDV after rescue (44). The system has previously been used to rescue MV to the same efficiency (39) as other systems (23, 31, 40, 45) with the vaccinia virus vTF7-3, which requires discrimination between the rescued viruses and the vaccinia virus. The efficiency of recovery of rCDV is similar to that of the other morbillivirus systems, with about 1 to 2 rescue events per 10⁶ transfected cells. One limiting factor in rescue systems using T7 polymerase for expressing the antigenome transcripts can be overcome by introducing G residues at transcription start sites. This approach has been used in rescue of, e.g., rabies virus, vesicular stomatitis virus, and Sendai virus because T7 polymerase initiates more efficiently at multiple G residues (23, 40, 45). We used the normal viral sequence starting with ACCA, as was done before with MV and Rinderpest virus (3, 35).

One potential problem in rescue experiments involving a poxvirus is the homologous recombination catalyzed by vaccinia virus proteins between plasmids coding for the genome and the helper plasmids (23). In the reported work, the helper plasmids N and P were constructed using the same RNA of OND-LP, and hence recombination would not have had any impact. The tag within the coding region of the L protein of p(+)CDV and the mutations present in the helper plasmid pEMC-L could have been exchanged by homologous recombination. However, this was not detected in this region in the rescued virus. In future work involving defined mutations in N, P, or L genes, which confer a replicative disadvantage on the virus, it has to be considered that recombination revertants might be recovered preferentially.

For verification that the rescued virus was a morbillivirus and specifically rCDV, we performed immunofluorescence studies with specific antisera. The MAbs that we used were cross-reactive with PDV, MV, and CDV proteins except for the anti-M and anti-F antibodies, which exclusively bind to the respective CDV proteins (N. Duffy, 1995. M.Sc. thesis, The Queen's University of Belfast, Belfast, Northern Ireland; A. Trudget, personal communication). The staining for N, P, M, F, and H proteins was positive. As M-protein-deficient MV (11) as well as respiratory syncytial virus, which does not express SH and G proteins (28), can be propagated in cell culture, it was important to confirm that rCDV expressed all of the major viral structural proteins.

CDV is now the third morbillivirus available for detailed analysis applying reverse genetics. We plan to introduce the green fluorescent protein into the cDNA clone to study the infection of the virus in cell culture and in an animal model as this has already been proven to be very useful in MV (19, 20). The rescue system may also facilitate studies on determinants of persistence in cells and animals and attenuating mutations in a natural host such as the ferret.