INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Equine arteritis virus (EAV) is the type member of the family Arteriviridae (38), which was recently grouped together with the coronaviruses and the toroviruses in the newly established order of the Nidovirales (4, 13). Other members of the Arteriviridae are Porcine reproductive and respiratory syndrome virus, Lactate dehydrogenase-elevating virus, and Simian hemorrhagic fever virus.

EAV is a spherical, enveloped virus with a diameter of 50 to 60 nm (18, 30) and a positive-stranded RNA genome of about 12,700 nucleotides (nt) (8). The virion envelope is derived from intracellular host cell membranes and contains two major and three or four minor structural proteins (12, 38, 39). The envelope surrounds an isometric nucleocapsid of about 35 nm (18), which is composed of the genomic RNA and multiple copies of the nucleocapsid protein (N).

The EAV replicase is produced in the form of two large polyproteins: the open reading frame 1a (ORF1a) protein and the ORF1ab protein, the C-terminal part of which is expressed by ORF1a/1b ribosomal frameshifting (8). An apparently complete proteolytic processing scheme for the ORF1a and ORF1ab nonstructural polyproteins, which results in the generation of 12 end products (nsp1 to nsp12) and a large number of processing intermediates, was recently obtained (41, 46, 47, 50).

The EAV structural proteins are translated from a 3'-coterminal nested set of subgenomic (sg) mRNAs, which also carry a common 5' leader sequence derived from the 5' end of the genome (11). The mechanism of sg mRNA transcription, which resembles that of the coronaviruses in many aspects, involves a discontinuous step of which many details remain to be elucidated (4, 24, 38, 48).

Little is known about the genome replication of arteriviruses at the molecular level. The genomic 3' end is polyadenylated and contains a nontranslated region (NTR) of 59 to 117 nt (38). A short, conserved sequence is present just upstream of the poly(A) tail, but its function remains to be elucidated (16). The 5' end of the genomic RNA is capped (38) and contains an NTR of 156 to 224 nt. It was shown by Hwang and Brinton (19) that at least four host proteins from MA-104 cells interact with in vitro-transcribed RNA representing the 3' end of the genomic negative strand of simian hemorrhagic fever virus, lactate dehydrogenase-elevating virus, and EAV. This region is the complement of the genomic leader sequence and is assumed to be involved in the initiation of plus-strand RNA synthesis.

Encapsidation of EAV sg mRNAs in virus particles has never been observed (R. Molenkamp, B. C. D. Rozier, and E. J. Snijder, unpublished data), which suggests that only genomic RNA is encapsidated and that a specific encapsidation signal that is lacking from the sg mRNAs could be located in the ORF1ab region. It has been shown that the same region in defective genomes of mouse hepatitis coronavirus (MHV) contains an encapsidation signal that is required for the specific encapsidation of MHV DI RNAs (2, 15, 35, 43). However, the infectious bronchitis coronavirus is able to encapsidate small amounts of sg mRNAs (54), suggesting that the determinant for specific encapsidation is not exclusively located in the ORF1ab region of this virus.

Recently, a full-length EAV cDNA clone from which infectious transcripts can be produced has been generated (45). Although many aspects of the viral life cycle can be investigated by using this tool, the study of replication and encapsidation signals would strongly benefit from the development of EAV replicons that can be replicated in trans. Defective interfering (DI) RNAs have been widely used for the analysis of cis-acting replication signals (6, 15, 20, 22, 26, 28, 36, 43, 52). DI RNAs are truncated, and in some cases rearranged, genomes that have usually lost the potential to replicate autonomously due to deletions in the viral replicase gene(s). They can be generated by serial undiluted passaging of viruses (3, 10, 31, 33, 34) during which their replication depends on the replicative enzymes encoded by the helper virus. DI RNAs have retained all replication signals and frequently also the sequences required for RNA encapsidation (15, 26, 36, 43, 52). Because DI RNAs are generally much smaller than the helper virus genome, they are replicated more efficiently. Together with competition for viral proteins, the replicative advantage of DI RNAs explains the interference with helper virus replication. Although arterivirus DI genomes would be very useful, the detection or isolation of arterivirus DI RNAs has not been reported.

In this paper, we describe the generation of the first documented EAV DI genomes by serial undiluted passaging of a virus stock. We characterized the smallest of the DI RNAs that were generated, the 5.6-kb DI-b, and constructed a DI-b cDNA clone. Transcripts derived from this construct (pEDI) were replicated and packaged in helper virus-infected cells. Sequence analysis revealed that in comparison with the genome, DI-b contains two large in-frame deletions in the replicase gene. Remarkably, the entire region encoding the structural proteins was retained in DI-b. Density gradients showed that virus particles containing EDI RNA are of the same density as virus particles containing only full-length EAV RNA. Finally, introduction of the chloramphenicol acetyltransferase (CAT) gene in the DI genome allowed us to identify sequences required for replication/transcription and encapsidation of EAV DI RNAs and to reduce the maximal size of a replication-competent EAV DI RNA to 3.0 kb.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and virus. Baby hamster kidney cells (BHK-21 cells) were grown in BHK-21 medium (Life Technologies) supplemented with 5% fetal calf serum, 10% tryptose phosphate broth, and 10 mM HEPES. The EAV Bucyrus strain (14) was used in all experiments. All infections with EAV were carried out at 39.5°C.

Recombinant DNA techniques. Standard recombinant DNA procedures were used (37). Restriction enzymes, T4 DNA ligase, and T7 RNA polymerase were obtained from Life Technologies. All enzyme incubations and biochemical reactions were performed according to the instructions of the manufacturers. Sequencing reactions were performed with the Big Dye Terminator kit (Perkin-Elmer) and analyzed using an ABI PRISM 310 genetic analyzer (Perkin-Elmer).

Serial undiluted passages. BHK-21 cells were seeded in 10-cm² dishes and infected with the EAV Bucyrus strain at a multiplicity of infection (MOI) of 50. Medium was harvested after incubation for 24 h at 39.5°C, at which time point complete cytopathogenic effect (CPE) was observed. For each virus passage, a similar dish with fresh BHK-21 cells was infected with one third (600 µl) of the culture medium from the previous passage. To preserve high helper virus titers during passaging, extra helper virus was added to the inoculum at an MOI of 10 whenever a delay of CPE was observed.

Isolation and analysis of viral RNAs. Intracellular RNA was isolated at 12 h postinfection (p.i.) by using Trizol (Life Technologies) followed by isopropanol precipitation. Denaturing RNA electrophoresis was carried out using 1 to 1.5% agarose gels containing 10 mM MOPS (morpholinepropanesulfonic acid) and 2.2 M formaldehyde. Gels were dried and hybridized with appropriate 5'-end-labeled oligonucleotides (Table 1) as described by Meinkoth and Wahl (32).

RT-PCR and cDNA cloning. Intracellular poly(A)-containing RNA was purified using oligo(dT)-coupled magnetic beads (Dynal) according to the instructions of the manufacturer. cDNA synthesis was performed using Superscript II reverse transcriptase (Life Technologies), and subsequently cDNA was amplified by using the GeneAmp XL-PCR system (Perkin-Elmer). Oligonucleotides used as primers during reverse transcription (RT) and PCR are described in Table 2. The DI RNA-specific RT-PCR product was digested with MluI and BamHI and cloned between the MluI and BamHI sites in pEAV030 (EMBL/GenBank accession no. Y07862) (45), which are located at nt 589 and 9149, respectively. The resulting plasmid was named pEDI.

RNA transcription and transfection. pEAV030, pEDI, and derivatives were linearized with XhoI. DNA templates were extracted with phenol-chloroform and ethanol precipitated. RNA was synthesized in vitro by using T7 RNA polymerase for 2 h at 37°C. Reactions contained 4 µg of linearized template DNA, 1 mM each GTP, CTP, ATP, and UTP, 5 mM dithiothreitol, 0.1 µg of bovine serum albumin per µl, 125 U of T7 RNA polymerase, 63.5 U of unmethylated capping analog [G(5')ppp(5')G; New England Biolabs], and 40 U of RnaseOut (Life Technologies) in a final volume of 50 µl of T7 transcription buffer (Life Technologies). For transfections with pEDI transcripts, BHK-21 cells were infected with EAV at an MOI of 10. After 1 h, the inoculum was removed and the cells were transfected with EDI RNA as previously described (45). Alternatively, BHK-21 cells were double transfected with EAV030 RNA and EDI RNA.

Sucrose gradient purification of EAV and EDI-containing particles. BHK-21 cells were infected with EDI-containing EAV stocks at high MOI, and viral RNAs were labeled with [³H]uridine (100 µCi/ml) at 10 h p.i. in the presence of 10 µg of actinomycin D per ml. Medium from 10⁶ infected cells was harvested at 24 h p.i., cleared by low-speed centrifugation, and loaded on a linear 20 to 50% (wt/wt) sucrose gradient in 10 mM Tris-HCl (pH 7.5)-1 mM EDTA-100 mM NaCl. The gradient was centrifuged at 4°C for 16 h at 40,000 rpm in a Beckman SW41 rotor. Subsequently, 600-µl fractions were collected from bottom to top and assayed for incorporation of label by trichloroacetic acid precipitation and liquid scintillation counting. RNA from sucrose gradient fractions was isolated as described by Spaan et al. (42).

Construction and analysis of pEDIC2, pEDIC7, and derivatives. pEDI-based CAT expression constructs were created by introducing a BamHI-XhoI (nt 9149 to 12704) restriction fragment from the previously described constructs pEAVCAT2 and pEAVCAT7 (45) into BamHI-XhoI-digested pEDI. The resulting plasmids, pEDIC2 and pEDIC7, contained the CAT reporter gene downstream of the RNA2 and RNA7 sg mRNA promoters, respectively. An additional MluI restriction site was engineered at nt 435 to 440 by PCR mutagenesis (TGGCTT to ACGCGT). Truncated derivatives of pEDIC2 and pEDIC7 were generated by fusion of the restriction sites indicated in Table 3. With the exception of pEDIC2-1820, all deletions in the ORF1ab region of pEDIC2 were made in frame with respect to the replicase reading frame. In vitro-transcribed RNA from pEDIC2, pEDIC7, and derivatives was transfected into helper virus-infected cells as described above. As controls, pEDIC2 and pEDIC7 RNA was transfected into noninfected BHK-21 cells. At 12 h p.i., cell lysates were made by using the CAT enzyme-linked immunosorbent assay (ELISA) lysis buffer supplemented with the CAT ELISA kit (Boehringer Mannheim). At 16 h p.i., virus was harvested (P0 [passage 0] virus), mixed with helpervirus (MOI of 10), and used to infect a fresh monolayer of BHK-21 cells. Cell lysates were made at 12 h p.i., and P0 and P1 CAT expression was determined by CAT-ELISA (Boehringer Mannheim) by using an ELISA reader set at the recommended wavelength of 405 nm.

Immunofluorescence assays. Immunofluorescence assays were essentially performed as described before (47). A CAT-specific rabbit antiserum was obtained from 5 Prime  3 Prime Inc. and used at a 1:500 dilution. A Cy3-conjugated donkey anti-rabbit immunoglobulin G (Jackson ImmunoResearch Laboratories) was used as the secondary antibody at a 1:1,000 dilution.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of EAV DI RNAs during serial undiluted virus passaging. For many viruses, serial undiluted passaging in tissue culture has resulted in the generation of DI viruses (3, 10, 31, 33, 34). Likewise, in this study an EAV stock was passaged up to 40 times in BHK-21 cells. At each passage, intracellular RNA was isolated, subjected to denaturing gel electrophoresis, and hybridized to a radiolabeled oligonucleotide complementary to the 3' end of the EAV genome and all sg mRNAs. At P6 we observed, the first new EAV RNA species, which migrated between the 12.7-kb genome (RNA1) and sg mRNA2 (3.2 kb) in denaturing agarose gels (Fig. 1A) and which appeared to be passaged from dish to dish. The size of this new RNA (DI-a) was estimated to be approximately 8 kb. Serial passaging was continued and at P27 a second new RNA species was observed (Fig. 1B). This RNA was approximately 6 kb long and was named DI-b. Both DI-a and DI-b accumulated in subsequent passages and remained present up to P40. During the passaging experiments, a decrease in CPE was regularly observed, suggesting that that at least one of the new RNA species interfered with helper virus replication.

View larger version (98K): [in this window] [in a new window]   FIG. 1.   Generation of natural EAV DI RNAs by serial undiluted passaging of a wt virus stock. Intracellular RNA was isolated from infected BHK-21 cells at P3 to P12 (A) and P27 to P37 (B) and subjected to gel electrophoresis and hybridization with an oligonucleotide recognizing the 3' end of all viral mRNAs. The positions of the new RNA species DI-a and DI-b are indicated.

cDNA cloning and composition of an EAV DI RNA. To determine which regions of the EAV genome were represented in DI-a and DI-b, intracellular RNA was isolated at P35 and subjected to gel electrophoresis and hybridization with a number of antisense oligonucleotides. The results of this analysis are summarized in Table 1. Apparently, DI-a and DI-b both lacked large portions of the 9.5-kb EAV replicase gene. Therefore, we attempted to generate and clone a cDNA copy of this region by using an RT-PCR approach based on a primer set derived from the terminal sequences of the replicase gene. cDNA was synthesized with oligonucleotide E275 (nt 9192 to 9209; Table 2) and a PCR product was generated using oligonucleotide E275 and E261 (nt 228 to 247; Table 2). An RT-PCR product (Fig. 2A) specific for DI RNA-containing intracellular RNA was obtained, and this approximately 1.9-kb fragment was cloned between the 5'- and 3'-terminal sequences of the infectious EAV cDNA clone, pEAV030 (45), using internal MluI (nt 589) and BamHI (nt 9149) restriction sites. This resulted in a plasmid containing a central replicase gene region that was presumably derived from a DI RNA and pEAV030-derived sequences representing nt 1 to 589 and 9149 to 12704 of the EAV genome. A schematic representation of this construct, which was named pEDI, is shown in Fig. 2B. The MluI-BamHI restriction fragment of pEDI was sequenced and found to consist of three segments of the EAV replicase gene (Fig. 2B). Two DI RNA-specific fusion sites, designated FsA (nt 1057 fused to 1388 of the EAV genome) and FsB (nt 1684 fused to 8530 of the EAV genome), were identified. These fusions were in frame with respect to the replicase gene and are also shown in Fig. 2B.

View larger version (6721K): [in this window] [in a new window]   FIG. 2.   (A) RT and PCR on RNA isolated from P35 or wt-infected cells (EAV) using a primer set derived from the terminal sequences of the replicase gene. An approximately 1.9-kb DI RNA-specific PCR product was detected. Molecular weight markers are indicated. (B) Schematic representation of the EAV genome and EDI RNA. The replicase gene region is indicated by shaded boxes. The 296-nt middle segment of EDI is indicated as a black box. The internal MluI and BamHI restriction sites used for cloning of the DI replicase gene region and the sequences and locations of the two fusion sites FsA and FsB are also depicted. (C) RT-PCR approach to determine the sequence of the 5'- and 3'-terminal sequences of EDI RNA. An EDI RNA-specific RT-PCR was carried out on P35 RNA using a primer specific for FsA (primer E299) or FsB (primer E298). The specific RT-PCR products representing the 3' (4 kb; lane 3)- and 5' (1.1 kb; lane 5)-terminal sequences of the DI RNA were sequenced and found to contain sequences collinear with the wt EAV genome only. As a negative control, the same RT-PCR was performed on RNA isolated from wt EAV-infected cells (lanes 2 and 4).

In vitro-transcribed EDI RNA can be replicated and packaged by EAV helper virus. To show that pEDI transcripts are replication competent and behave similarly to the DI RNAs generated in our passaging experiments, we transfected in vitro-transcribed EDI RNA into EAV-infected cells and performed passaging experiments. BHK-21 cells were infected with helper virus at an MOI of 10. Half of the cells were transfected with EDI RNA, and the other half were mock transfected. Cells were plated and incubated for 16 to 20 h at 39.5°C. Culture supernatant (P0 virus) was harvested, and at the same time intracellular RNA was isolated (P0 RNA). One-third of the P0 virus was used to infect a fresh monolayer of BHK-21 cells, and at 16 to 20 h p.i. the medium was harvested (P1 virus). This procedure was repeated two more times. P0 to P2 virus was used to infect BHK-21 cells, and at 12 h p.i. intracellular RNA was isolated and analyzed. Figure 3 shows the results of such a transfection/passaging experiment. In lane 1 (EDI-P0 RNA), a small amount of EDI and a relatively large amount of helper virus genome could be observed. EDI accumulated in P2 and P3 (lane 3 and lane 4) at the expense of genomic RNA, which strongly suggested that EDI indeed interfered with the replication of the latter. The control experiment without EDI transcript (lanes 5 to 8) did not show the appearance of a DI RNA. Throughout this EDI passaging experiment, the amount of sg mRNAs produced in infected cells seemed to remain constant, while the amount of helper virus genomic RNA decreased drastically. This suggested that EDI was used in trans as a template for the synthesis of sg mRNAs. EDI comigrated with DI-b from P35 (lane 9), which again suggested that EDI is identical to DI-b. Hybridization of intracellular RNA from P35 with the EDI FsB-specific oligonucleotide (Table 2) supported this assumption.

View larger version (94K): [in this window] [in a new window]   FIG. 3.   Transfection and passaging of EDI RNA. In vitro-transcribed EDI RNA was transfected into EAV-infected BHK-21 cells. At 16 to 20 h p.i., medium was harvested and used to infect fresh BHK-21 cells. This procedure was repeated two times. RNA isolated from each passage (EDI-P0 to EDI-P3) was subjected to gel electrophoresis and hybridization with an oligonucleotide recognizing the 3' end of all viral RNAs. As a control, BHK-21 cells were infected with wt EAV and passaged in the same manner (EAV-P0 to EAV-P3). RNA isolated from P35 of the initial passaging experiment was put on the same gel and hybridized with an oligonucleotide recognizing all viral mRNAs (P35/E154) or hybridized with an FsB-specific oligonucleotide (P35/E288).

View larger version (51K): [in this window] [in a new window]   FIG. 4.   In vitro-transcribed EDI RNA was cotransfected with EAV030 RNA into BHK-21 cells. At 16 to 20 h p.i., medium was harvested and used to infect fresh BHK-21 cells. This procedure was repeated one more time. RNA isolated from each passage (P0 to P2; lanes 4 to 6) was subjected to gel electrophoresis and hybridization with an oligonucleotide recognizing the 3' end of all viral mRNAs. As a control, BHK-21 cells were transfected with EAV030 RNA only, and virus was passaged in the same manner (lanes 1 to 3).

EDI RNA and EAV genomic RNA are packaged in virus particles of equal density. Since EDI RNA is much smaller than the EAV genome, we investigated whether EDI RNA is packaged into virus particles with a density different from that of regular EAV particles. BHK-21 cells were infected with EDI- and helper virus-containing P3 virus from the passaging experiment shown in Fig. 3. Viral RNAs were metabolically labeled by using [³H]uridine. The culture supernatant was loaded on a 20 to 50% (wt/wt) linear sucrose gradient, which was centrifuged to equilibrium. Subsequently, 600-µl fractions were collected from bottom to top and assayed for the presence of label by trichloroacetic acid precipitation and liquid scintillation counting.

View larger version (25K): [in this window] [in a new window]   FIG. 5.   Sucrose gradient analysis of an EDI-containing EAV stock. BHK-21 cells were infected at high MOI, and viral RNAs were labeled with [3H]uridine. Medium was loaded on a 20 to 50% (wt/wt) sucrose gradient which was centrifuged to equilibrium; 600-µl fractions were collected from bottom to top and analyzed for the presence of label. (A) Sucrose gradient profile with the peak of radioactivity in fraction 10. (B) RNA isolated from fractions 8 to 11 was subjected to gel electrophoresis and hybridization with an oligonucleotide recognizing all viral mRNAs. Both EAV genomic RNA and EDI RNA were found almost exclusively in fraction 10.

EDI replication and packaging monitored by expression of the CAT reporter gene. The analysis of the replication and packaging of an RNA replicon is, obviously, most straightforward by direct metabolic labeling of the replicon in transfected cells. However, in the case of EDI, such an approach was hampered by low transfection and replication efficiencies of the RNA replicon. We therefore engineered the expression of a reporter gene to monitor replication and packaging of EDI-based replicons in a sensitive and convenient manner. A similar approach has been used successfully during molecular studies on the replication, packaging, and transcription of, e.g., coronavirus and poliovirus subgenomic replicons (1, 20, 27, 29, 53). Clearly, this kind of analysis is most straightforward when the assay is based directly on translation of the replicating RNA itself. However, in the case of EDI, the presence of a truncated replicase ORF in the 5' end of the replicon posed a potential problem, in particular because the importance of a similar ORF in certain coronavirus DI RNAs has been well established in our laboratory (7, 44). To avoid the risk of impairing EDI replication by inserting a reporter gene in its 5'-terminal region, we chose an indirect approach to study replication and packaging. The CAT reporter gene was inserted at the position normally occupied by ORF2b or ORF7 (Fig. 2B). Consequently, the CAT gene should be expressed from sg mRNA 2 or 7, both of which are normally used to translate EAV structural proteins. We have previously shown that in the EAV full-length cDNA clone pEAV030 efficient CAT expression from these positions is possible (45). Therefore, we engineered pEDI derivatives carrying the CAT gene at the same positions (pEDIC2 and pEDIC7) (Fig. 6A and 7), thereby allowing its expression from EDI-derived sg mRNAs.

View larger version (34K): [in this window] [in a new window]   FIG. 6.   (A) Schematic representation of EDIC2 and EDIC7. Subgenomic mRNAs from which the CAT reporter gene is expressed are indicated. (B) CAT protein expression from EDIC2 RNA transfected into EAV-infected BHK-21 cells during a 12-h time period. OD, optical density. (C) Immunofluorescence staining with a CAT-specific antiserum on EAV-infected BHK-21 cells, transfected with EDIC2 RNA. The transfection efficiency in this particular experiment was estimated to be approximately 10%.

View larger version (18K): [in this window] [in a new window]   FIG. 7.   Replication and packaging of EDIC2 and EDIC7 RNA deletion mutants. EDIC2, EDIC7, and restriction sites used for deletion mutagenesis are schematically depicted. The position of the CAT reporter gene behind the RNA2 sg mRNA promoter (EDIC2) or RNA7 sg mRNA promoter (EDIC7) is indicated. Deleted regions are indicated by dashed lines. CAT expression levels in P0 and P1 comparable to those observed for EDIC2 or EDIC7 are indicated by +; background CAT levels are indicated by .

Delimitation of sequences required for replication and packaging of EDI. Several deletions were made in pEDIC2 and pEDIC7 (Fig. 7), and in vitro-transcribed RNA from these derivatives was transfected into helper virus-infected BHK-21 cells. With the exception of EDIC2-1820, deletions in ORF1ab were in frame with respect to the replicase reading frame. As controls, EDIC2 and EDIC7 RNA were transfected into infected and mock-infected cells. Transfection efficiencies were determined by immunofluorescence assays (data not shown) and were judged to be similar for all CAT-expressing constructs. CAT protein expression in P0 and P1 was quantitated by using a sensitive CAT ELISA. The results of this analysis are summarized in Fig. 7.

Further characterization of EDIC2-DD1 and EDIC7-DD2. To further characterize the EDIC2-DD1 and EDIC7-DD2 replicons, CAT expression levels during P0 were monitored for EDIC2, EDIC7, EDIC2-DD1, and EDIC7-DD2. Transfection efficiencies were determined by immunofluorescence assay, and CAT values were corrected for transfection efficiency. For all constructs, significant CAT protein expression levels were detected at 4 h posttransfection (Fig. 8A) and CAT levels increased between 8 and 12 h posttransfection, when EAV sg mRNA synthesis reaches its maximum (9). No significant differences in replication efficiencies between the various EDI derivatives were found. To investigate whether the EDIC2, EDIC7, EDIC2-DD1, and EDIC7-DD2 replicons were packaged in virus particles, and to exclude the possibility that the RNA replicons could be passaged otherwise, a gradient analysis was performed. Medium harvested from transfected cells was passaged once, and the resulting P1 virus was loaded on a 20 to 50% linear sucrose gradient, which was centrifuged to equilibrium. Virus was isolated from the fraction that was earlier shown to contain the virus peak (Fig. 5, fraction 10) and fractions 9 and 11. Virus isolated from these fractions was mixed with helper virus (MOI of 10) and used to infect a fresh monolayer of BHK-21 cells. At 12 h postinfection, lysates were made and CAT expression was determined. Alternatively, immunofluorescence assays with a CAT-specific antibody were performed. For gradient-purified EDIC2, EDIC7, and EDIC2-DD1 virus, significant CAT expression (Fig. 8B) and a positive immunofluorescence signal were observed in a low percentage of cells (data not shown). In contrast, gradient-purified EDIC7-DD2 virus did not show CAT expression or a positive signal in immunofluorescence. From these data we conclude that the EDIC2, EDIC7, and EDIC2-DD1 replicons are indeed packaged into virus particles, while EDIC7-DD2 is not.

View larger version (31K): [in this window] [in a new window]   FIG. 8.   (A) Replication of EDIC2, EDIC7, EDIC2-DD1, and EDIC7-DD2 replicons during the first 12 h posttransfection (pt) as measured by the expression of the CAT reporter gene (optical density at 405 nm [OD 405]) in the presence () or absence () of helper virus. (B) CAT expression in BHK-21 cells infected with sucrose gradient-purified virus from EDIC2, EDIC7, EDIC2-DD1, and EDIC7-DD2 replicons.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of the first arterivirus DI RNA. To our knowledge, the work presented in this paper constitutes the first characterization of a defective arterivirus genome. A DI RNA of approximately 8 kb (DI-a) was readily observed in P6 of our serial passaging experiment, but the 6-kb DI-b RNA used in this study surfaced only after 21 subsequent passages. No other new RNA species were observed up to P40, at which passage the experiment was terminated. DI-b cDNA was generated, and a presumably full-length DI-b cDNA clone (pEDI) was constructed and subjected to sequence analysis. Our preliminary hybridization analysis (Table 1) suggests that DI-a contains an additional part of ORF1b and may therefore be a precursor of DI-b. However, since the nature of DI-a was not further investigated, it remains unclear whether DI-b has evolved independently or is derived from DI-a. Synthetic pEDI transcripts were replicated and packaged in EAV-infected cells and also in cells transfected with transcripts derived from the full-length EAV cDNA clone. Density gradient centrifugation showed that virus particles containing EDI RNA were of the same density as virus particles containing only EAV genomic RNA. Finally, the insertion of the CAT reporter gene downstream of the RNA2 or RNA7 sg mRNA promoter allowed us to delimit the sequences required for replication/EDI-based transcription and packaging of EDI RNA. Based on the results obtained with deletion mutagenesis, these sequences are now reduced to 589 nt at the 5' end, 1,066 nt at the 3' end, and an internal sequence of 583 nt derived from ORF1b. However, the deletion mutant EDIC7-DD2, which contained these EAV sequences only, did replicate but was not packaged into virions.

Molecular characterization of DI-b. Sequence analysis of the cloned DI-b RNA revealed some intriguing features. DI-b is composed of three genome segments joined in frame with respect to the replicase gene. The first segment (nt 1 to 1057) consists of the 5' NTR (leader) sequence and part of replicase ORF1a. The second segment is a small (296-nt) region corresponding to nt 1388 to 1684 of the EAV genome, which encodes part of nsp2. The third segment represents the region from nt 8530 up to the 3' end of the EAV genome. As a result, EDI contains a reading frame of 783 amino acids, starting at the ORF1a translation initiation codon and terminating at the ORF1b stop codon. Since the sequence upstream of the translation initiation codon of this fusion ORF is identical to the 5' NTR of the EAV genome, we assume that this truncated replicase gene is indeed translated. The translation product consists of the entire nsp1 sequence, two segments of nsp2, the C-terminal domain of nsp10, nsp11, and nsp12. In EAV- and EDI-infected cells the proteolytic processing of the EDI 1ab fusion protein will probably be similar to that of the corresponding parts of the full-length EAV 1ab polyprotein (41, 46, 47, 50). The nsp1 autoprotease can be expected to cleave itself at the nsp1/2 site (40). The nsp10/11 and nsp11/12 sites will probably be processed in trans by the nsp4 protease encoded by the helper virus (46). The remainder of the EDI translation product would then be a small (185-amino-acid) fusion protein containing two segments of nsp2 and the C-terminal part of nsp10. It is unclear whether any of these EDI translation products has a function in DI RNA replication. It was proposed that for MHV DI RNAs, translation of a fusion reading frame might promote the stability and enhance the replication of DI RNAs (7). Studies to investigate whether the truncated EDI ORF is indeed translated and whether translation per se is required for efficient DI RNA replication are in progress.

Delimitation of sequences required for replication and packaging of EDI RNA. On the basis of the results obtained with constructs EDIC2-0406 and EDIC2-0613, the 5'-terminal sequence required for efficient replication or transcription and packaging of EAV DI RNAs has been reduced to a maximum of 589 nt. Likewise, constructs EDIC2-3457 and EDIC2-3450 demonstrated that the 3'-terminal genomic sequence required for replication and packaging of EAV DI RNAs is at most 1,068 nt long. In addition to these 5'- and 3'-terminal sequences, a 583-nt sequence from the central part of replicase ORF1b (nt 8566 to 9149 in the EAV genome) was found to be essential for EDI-driven CAT expression (EDIC2-1318 and EDIC2-1820). Since our assay did not differentiate between replication and sg mRNA transcription defects, we cannot exclude that the minimal sequences that are required are in fact smaller. With the exception of EDIC2-1820, all deletions in the EDI ORF1ab region were made in frame with respect to the replicase reading frame. Thus, if the inability of EDIC2-0406 and EDIC2-1318 to express the CAT protein is due to an effect at the level of replication, this defect is not caused by the disruption of the truncated EDI ORF1ab.