key: cord-0008552-27fd59qe
authors: Meulenberg, Janneke J.M.; Petersen-Den Besten, Annelien; De Kluyver, Eric P.; Moormann, Rob J.M.; Schaaper, Wim M.M.; Wensvoort, Gert
title: Characterization of proteins encoded by ORFs 2 to 7 of Lelystad virus
date: 1995-01-10
journal: Virology
DOI: 10.1016/s0042-6822(95)80030-1
sha: 8ec8e4151b2b8d6a61816f179b510df439a19de9
doc_id: 8552
cord_uid: 27fd59qe

The genome of Lelystad virus (LV), a positive-strand RNA virus, is 15 kb in length and contains 8 open reading frames (ORFs) that encode putative viral proteins. ORFs 2 to 7 were cloned in plasmids downstream of the Sp6 RNA polymerase promoter, and the translation of transcripts generated in vitro yielded proteins that could be immunoprecipitated with porcine anti-LV serum. Synthetic polypeptides of 15 to 17 amino acids were selected from the amino acid sequences of ORFs 2 to 7 and antipeptide sera were raised in rabbits. Antisera that immunoprecipitated the in vitro translation products of ORFs 2 to 5 and 7 were obtained. Sera containing antibodies directed against peptides from ORFs 3 to 7 reacted positively with LV-infected alveolar lung macrophages in the immunoperoxidase monolayer assay. Using these antipeptide sera and porcine anti-LV serum, we identified three structural proteins and assigned their corresponding genes. Virions were found to contain a nucleocapsid protein of 15 kDa (N), an unglycosylated membrane protein of 18 kDa (M), and a glycosylated membrane protein of 25 kDa (E). The N protein is encoded by ORF7, the M protein is encoded by ORF6, and the E protein is encoded by ORF5. The E protein in virus particles contains one or two N-glycans that are resistant to endo-β-N-acetyl-d-glucosaminidase H. This finding indicates that the high-mannose glycans are processed into complex glycans in the Golgi compartment. The protein composition of the LV virions further confirms that LV is evolutionarily related to equine arteritis virus, simian hemorrhagic fever virus, and lactate dehydrogenase-elevating virus.

Lelystad virus (LV) is a small enveloped virus containing a positive-strand RNA genome. It was first identified in 1991 in The Netherlands by Wensvoort eta/. (1991) and in the United States by Collins eta/. (1992) as the causative agent of porcine reproductive respiratory syndrome (PRRS). PRRS is mainly characterized by reproductive failure in sows and respiratory problems in pigs 0fall ages (Loula, 1990; Wensvoort, 1993) .

The genome of LV is a polyadenylated RNA molecule 0fab0ut 15 kb, which contains eight open reading frames (0RFs) that probably encode the replicase genes (ORFs la and lb), the envelope proteins (ORFs 2 to 6) and the ,%cle0capsid protein (ORF7; Meulenberg et al., 1993a; 00nzelmann et al., 1993) . ORFs 2 to 7 are most likely expressed from six subgenomic RNAs, which are synthesized during replication (Meulenberg et al., 1993a) . These subgenomic RNAs form a 3' coterminal nested set and are composed of a common leader, derived from the 5' end of the viral genome which is joined to the body of the subgenomic RNAs at the consensus junction sequence UOAAOC (Meulenberg et aL, 1993b) .

LV(also named PRRS virus) resembles equine arteritis virus (EAV), lactate dehydrogenase-elevating virus (LDV), 'T0 whom correspondence and reprint requests should be addressed. Fax= 31-320042804. 155 and simian hemorrhagic fever virus (SHFV) in genome organization, replication strategy, amino acid sequence of the proteins, and preference for infection of macrophages, both in vivo and in vitro (Conzelmann et aL, 1993; Meulenberg et al., 1993a) . Because of these similarities, proposals have been made to classify LV, EAV, SHFV, and LDV into a new virus family, tentatively named the Arteriviridae (Conzelmann et al., 1993; Meulenberg et al., 1993a; Plagemann and Moennig, 1991) .

When isolates of LV from Europe and the United States were tested in an immunoperoxidase monolayer assay using potyclonal sera containing antibodies against these isolates, they appeared to differ antigenically, but to share common antigens (Wensvoort eta/., 1992) . Murtaugh eta/. (1993) also observed genomic heterogeneity between the nucleotide sequence of a United States isolate and LV. Apparently, these viruses have diverged from each other by point mutation and possibly through recombination.

Although the replication strategy of LV has been studied, and the complete nucleotide sequence of the viral genome has been determined, little is still known about the structural proteins of LV. Using hyperimmune pig antisera and monoclonal antibodies, Nelson eta/. (1993) have identified three viral proteins of 15, 19, and 26.kDa in lysates of cells infected with a United States isolate of LV. They considered the 15-kDa protein to be the nucleocapsid protein. In the work described in this paper the viral proteins of LV were studied in more detail. The proteins encoded by ORFs 2 to 7 were characterized by in vitro transcription and translation experiments; proteins encoded by ORFs 5 to 7 were shown to be structural proteins of LV.

LV was grown either on porcine alveolar macrophages or on CL2621 cells (courtesy of Boehringer-lngelheim, St. Joseph, MO). Macrophages were maintained as described before (Wensvoort et al., 1991) . CL2621 cells were maintained in Eagle's basal medium supplemented with 5% fetal bovine serum, 100 IU/ml penicillin, and 100 /~g/ml streptomycin. To prepare concentrated and purified virions, confluent monolayers of CL2621 cells were infected at a multiplicity of infection (m.o.i.) of 0.1. At the beginning of cytopathic changes (48-56 hr after infection), the medium was harvested and centrifuged for 20 min at 1200 g. The virus in the medium was concentrated by precipitating it with 6% polyethylene glycol 20,000 overnight at 4 ° and was then centrifuged at 10,000 g for 45 min. The pellet was resuspended in TNE buffer (0.01 MTris-HCI, pH 7.2, 0.1 M NaCI, 1 mM EDTA) and layered either on a 30-0% glycerol, 0-50% di-K-tartrate gradient (Purchio et aL, 1984) or on a 20-50% sucrose gradient (van Berlo et aL, 1982) .

To identify specific characteristics in the amino acid sequences of ORFs 2 to 7, the computer analysis program PCGENE (Intelligenetics, Inc., Mountain View, CA) was used. Prosite was used to detect sites in the protein sequence that might have a putative specific biological function (Bairoch et al., 1991) . Hydrophobic membranespanning sequences and signal sequences were identified by the methods of Kyte and Doolittle (1982) and von Heijne (1986), respectively.

Specific oligonucleotides located upstream and downstream of QRFs 2 to 7 were synthesized. These were 5' CTGCCGCCCGGGCAAGTGCC 3' and 5' AGAACAGGG-CATGCATATGG 3' for ORF2, 5' CACGCCAGGTACCAG-GCCCA 3' and 5' AAAGCATCTGCAGGTCCGCG 3' for ORF3, 5' GGCAATTGGATCCAq-I-i-GGA 3' and 5' AGA-AGCAAGCTTGCGGAGTC 3' for ORF4, 5' GAGGTGGGA-TOCAACCATTG 3' and 5' CTAGCACAAGC-FFI-FGTGCG 3' for ORF5, and 5' CCCTTGAOGAGCTCTTCGGC 3' and 5' CCATCGGATCCGTACT-i-FFC 3' for ORF6, and 5' GGT-TAACCTCGTCGACTATG 3' and 5' CCTGA-I-I-AAGCT-TGACCCCC 3' for ORF7. The underlined nucleotides are mutated with respect to the genome sequence of LV, in order to create restriction sites in the primers. These primers were used to amplify ORFs 2 to 7 by the polymerase chain reaction (PCR) from cDNA clones described earlier (Meulenberg et aL, 1993a) . The ORFs were amplified with AmpliTaq DNA polymerase (Perkin-Elmer-Cetus) under the following conditions: Double-stranded DNA was denatured for 1 min at 94 ° , the primer was annealed for 1 min at 55 ° and extended for 1 min at 72 ° . The number of cycles was 30. The restriction sites incorporated in the primers were used to clone the ORFs downstream of the Sp6 RNA polymerase promoter of pGEM-4Z (Promega). Recombinant DNA techniques were performed essentially as described by Sambrook et al. (1989) .

Plasmids containing ORFs 2 to 7 were linearized and transcribed in vitro by use of Sp6 RNA polymerase (Promega), according to the protocol of Promega. The synthesized RNA was translated in rabbit reticulocyte lysate (Promega) in the presence of L-[3~S]methionine (>1000 Ci/mmol, Amersham), as recommended by the manufacturer, in either the presence or the absence of microsomal membranes.

In vitro translation products, purified LV preparations, or immunoprecipitates were resuspended in 25 #1 endoglycosidase buffer (1% NP-40, 1 mM phenylmethylsulfonyl fluoride, 1 #g/ml aprotinin, 1 #g/ml pepstatin A, and 1 #g/ml leupeptin in phosphate-buffered saline (PBS)). Then 800 mU of peptide N-glycosidase F (PNGaseF; Boehringer-Mannheim) was added and the reaction mixture was incubated overnight at 37 ° . Controls were treated similarly, except that the PNGaseF was omitted. Endo-/~-N-acetyi-D-glucosaminidase H (EndoH) digestion was performed similarly, except that the protein samples were resuspended in 100 mM sodium acetate (pH 5.5), 1% NP-40, 1 mM phenylmethylsulfonyl fluoride, 1 #g/ml aprotinin, 1 #g/ml pepstatin A, 1 #g/ml leupeptin, and 20 mU EndoH (Boehringer-Mannheim). Endo-/~-gaiactosidase digestions were performed as described by de Vries et aL (1992) .

Polyvalent antiserum 21 directed against LV was obtained from a specific-pathogen-free (SPF) pig infected intranasally with 105 TCID~o of a fifth cell culture passage of LV (CDI-NL-91; Institute Pasteur 1-1102). Blood samples were taken 42 days after infection. Gene-specific rabbit sera directed against ORFs 2 to 7 were obtained by use of synthetic peptides of 15 to 17 residues containing an amino acid sequence specific for each ORF. The amino acid sequences were selected on the basis of hydrophilicity, which was determined by the algorithms of Hopp and Woods (1981) and the antigen prediction program Asciflex (Janin, 1979; Parker et aL, 1986) . The peptides were conjugated to keyhole limpet hemocyanin. SPF rabbits were immunized intramuscularly and subcutaneously with 1 mg peptide conjugated to keyhole limpet hemocyanin in complete Freund's adjuvant. After 1 month, the rabbits were given a booster injection of the same amount of conjugated peptide in incomplete Freund's adjuvant. The rabbits were bled at 12 weeks after the first immunization. Sera were tested for their reactivity with the various peptides in an enzyme-linked immunosorbent assay (ELISA), using pep'~ides coated to M96 plates (Suter, 1982) . Sera were also tested for their reactivity with viral antigen in an immunoperoxidase m0n01ayer assay (IPMA), essentially as described by Wensvoort et al. (1991) . Porcine alveolar lung macrophages were seeded in microtiter plates as described and grown for 16 hr. Then 50 #1 medium containing 10 3-104 TCID~o LV was added, and the macrophages were grown for 24 hr. The plates were then fixed and stained with the rabbit antipeptide sera and goat anti-rabbit IgG horseradish peroxidase, as described. As previously established, these incubation times resulted in 30% of the macrophages being infected and showing good cytoplasmic staining in IPMA. Sera 690, 694, 698, 704, 710, and 714 directed against ORFs 2, 3, 4, 5, 6, and 7, respectively, were selected for further use in immunoprecipita-ti0n and Western blot analysis of viral proteins produced in vitro or in vivo. Monoclonal antibodies raised against a United States isolate (SDOW12 and SDOW17) and an English isolate of LV (WBE1, WBE 4 to 6) were obtained from D. Benfield and T. Drew, respectively.

Viral protein samples were suspended in Laemmli sample buffer (Laemmli, 1970) , heated for 2 min at 100 °, and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on gels containing 12.5 or 15% polyacrylamide. The separated proteins were transferred to nitrocellulose paper by elec-tr0bl0tting (Towbin et aL, 1979) . Polyclonal antiserum 21 and antipeptide sera 690, 694, 698, 704, 710, and 714 were diluted 1,50 in PBS containing 2% NaCl, 0.05% Tween 80, and 5% horse serum. Nitrocellulose strips were incubated with these diluted antisera for 1 hr at 37 °. The strips were washed three times with PBS containing 2% NaCI and 0.05% Tween 80. They were then incubated with rabbit anti-swine IgG horseradish peroxidase (1,500) or goat anti-rabbit IgG horseradish peroxidase (1,1000) diluted in PBS containing 2% NaCl, 0.05% Tween 80, and 5% horse serum for 1 hr at 37 ° . Finally, the strips were washed three times in PBS and stained in a solution of 0.6 mg/ml 4-chloro-l-naphthol, 20% (v/v) methanol, and 0.3 #l/ml HzO2 (30%).

00nfluent monolayers of CL2621 were infected with LVas described above, at an m.o.i, of 0.1. At 24 hr after infection, the medium was replaced by methionine-and cysteine-free medium containing 10% fetal calf serum, 90 #Ci/ml L-[3~S]methionine, and 125 #Ci/ml h-[3~S]cysteine (>1000 Ci/mmol, Amersham). Cells were harvested 40 hr after infection, rinsed in PBS, and lysed in PBS-TDS (1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, and 1 mM phenylmethylsulfonyl fluoride in PBS). Lysates were sonified and clarified for 10 min at 120,000 g. The supernatant was used for immunoprecipitation.

Proteins were immunoprecipitated essentially according to.the method of Hulst eta/. (1993) . In vitro translation products were diluted in PBS-TDS and incubated overnight at 4 ° with 10 #1 polyvalent serum 21 or antipeptide serum. The immunocomplexes were then precipitated with protein A-Sepharose CL-4B beads (Pharmacia) for 2 hr at 4 ° . The protein A-Sepharose beads were pelleted (2 min at 1000 g) and were washed four times with PBS-TDS. Pellets were then resuspended in Laemmli sample buffer and heated for 2 min at 100 ° before analysis by SDS-PAGE on gels containing 12.5 or 15% polyac~lamide. Gels were dried and immunoprecipitated proteins were visualized by autoradiography. The same procedure was used to precipitate proteins from cell lysate or purified virus, except that the samples were first immunoprecipitated with negative pig serum or rabbit preserum to reduce nonspecific precipitation of proteins.

The positive-strand RNA genome of LV is composed of 15,088 bases and contains two large ORFs (ORFla and ORFlb) encoding the RNA-dependent RNA polymerase and six smaller ORFs (ORFs 2 to 7) located at the 3' part of the viral genome. ORFs 2 to 7 are 128 to 265 amino acids in length and possibly encode structural proteins of LV (Table 1) . These ORFs contain putative Nglycosylation sites, which may mean that they encode gtycoproteins. ORFs 2, 3, and 5 contain a hydrophobic amino terminus followed by a putative signal cleavage site as predicted by the method of Von Heijne (1986). These sequences may be involved in the transport of ORFs 2, 3, and 5 to the endoplasmic reticulum. ORF4 also contains an N-terminal hydrophobic sequence, but no signal sequence cleavage site was predicted. Hydrophobic sequences encoded at the carboxy termini of ORFs 2 and 4 and in the middle of the coding region of ORF5 may function as membrane anchors. ORF6 contains three putative membrane-spanning fragments at the N-terminus. The hydrophobicity profile of this protein is comparable to that of the membrane protein M of coronaviruses and toroviruses and the proteins encoded by ORF6 of LDV and EAV (Meulenberg eta/., 1993a) . This suggests a similar membrane topology and membraneassociated function. In contrast to the amino acid sequence of ORFs 2 to 6, the amino acid sequence of ORF7 is basic, it does not contain hydrophobic segments. Because the amino acid sequence of ORF7 closely resembles that of other viral nucleocapsid proteins, it is thought to be the nucleocapsid protein of LV.

In vitro transcription and translation of ORFs 2 to 7

ORFs 2 to 7 were cloned in plasmid vector pGEM-4Z, and transcripts from these ORFs were translated in nuclease-treated rabbit reticulocyte lysate. Figure 1 shows the translation products analyzed by SDS-PAGE. The apparent sizes of the products, translated in the absence of microsomal membranes were 26 kDa for ORF2, 27 kDa for ORF3, 19 kDa for ORF4, 20 kDa for ORF5, 18 kDa for ORF6, and 15 kDa for ORF7. Although the sizes of these proteins, determined by SDS-PAGE, do not necessarily correlate with true molecular weights, they are in fair agreement with the sizes that were estimated from the amino acid sequence (Table 1) . Translation of the transcripts of ORFs 2 to 5 in the presence of microsomal membranes resulted in larger products, the major protein bands being 30, 45, 31, and 25 kDa, respectively (Fig. 1) . The presence of microsomal membranes during translation did not affect the sizes of the gene products of ORFs 6 and 7. The expression of the ORF6 gene product was somewhat higher when microsomal membranes were added during in vitro translation. Maybe these membranes enhance translation or increase the stability of the membrane protein encoded by ORF6. The results indicate that the gene products of ORFs 2 to 5 were glycosylated during in vitro translation, whereas the gene products of ORFs 6 and 7 were not. The shift in size of the glycosylated gene products of ORFs 3 and 4 with 7 and 4 putative glycosylation sites, respectively (Table 1) , was greater than the shift in size of the gene products of ORFs 2 and 5 with both 2 putative N-glycosylation sites ( Table 1 ). The glycosylation of the gene products of ORFs 2 to 5 that were synthesized in the presence of microsomal membranes was further confirmed by digesting the products with PNGaseF, an enzyme which cleaves off N-glycans. This procedure reduced the size of the gene products of ORFs 2 to 5 to the size of the unglycosylated gene products (data not shown). For ORFs 2, 4, and 5 also a product of smaller size than the unglycosyiated products was observed, when translated in the presence of microsomal membranes (Fig. 1) , and increased in intensity after PNGase treatment (data not shown). These products may represent unglycosylated gene products from which the Nterminal signal sequence is cleaved off during membrane insertion. For ORF2 a second protein of smaller size than the unglycosytated protein was observed upon Fla. 2. RIP and SDS-PAGE of radiolabele.d proteins encoded by 0RFs 2 to 7 synthesized in vitro. The transcripts of the ORFs were translated in the absence (-) or presence (+) of microsemal membranes (M) and were immunoprecipitated with convalescent serum 21 collected from a pig infected with LV (A). As a control, the in vitro translation products of Brome mosaic virus (BMV) RNA (supplemented with the in vitro translation kit of Pharmacia) were immunoprecipitated with serum 21 (indicated as ORFC). The in vitro translation products of0RFs 6 and 7 and BMV RNA (indicated with C) were also immunoprecipitated with a monoclonal antibody SDOW17 raised against a United States isolate of LV (B). Identical precipitation reactions were obtained with monoclonel antibodies SDOW12, WBE1, and WBE4 to WBE6. translation in the presence of microsomal membranes (Fig, 1) . The nature of this protein is not understood. The translation products of ORFs 6 and 7, when generated in the presence of microsomal membranes, did not change in size after treatment with PNGaseF (data not shown). This finding indicates that the putative N-glycosylation sites of ORFs 6 and 7 are not in fact glycosylation sites, at least in in vitro experiments.

Immunoprecipitation of in vitro translation products of 0RFs 2 to 7

The gene products of ORFs 2 to 7 synthesized in the presence or absence of microsomal membranes were immunoprecipitated with convalescent serum, serum 21, 0fa pig infected with LV. Serum 21 recognized the gene products of all these ORFs ( Fig. 2A) . The gene product of 0RF7 was immunoprecipitated best, whereas the gene products of ©RFs 2 and 6 were immunoprecipitated only weakly (and were observed clearly only after longer exposure; data not shown). The 12-and 13-kDa bands, which were coprecipitated with the major 15-kDa product of ORF7, are probably minor products formed as a result of proteolytic degradation, premature translational stop, or internal initiation of translation. Serum 21 did not immunoprecipitate in vitro translation products of Brome mosaic virus RNA, which served as negative control (Fig.  2A) . These results indicate that the proteins encoded by O R F s 2 to 7 are made during virus replication in pigs and that antibodies are raised against all these proteins. Monoclonal antibodies raised against a United States isolate (SDOW12 and SDOWl 7; Nelson et al., 1993) and an English isolate of LV (WBE1, WBE4 to WBE6; Drew et al., 1994) all recognized the 15-kDa gene product of QRF7. Figure 2B shows a representative radio immunoprecipitation (RIP) assay using SDOW17.

Gene-specific antisera, containing antibodies directed against peptides of ORFs 2 to 7 were raised in rabbits. Five s e r a -694 (anti-ORF3), 698 (anti-ORF4), 704 (anti-ORF5), 710 (anti-ORF6), and 714 (anti-ORF7) --were selected that reacted positively with the corresponding peptide in an ELISA and that reacted positively in an IPMA with alveolar macrophages infected with LV ( Table 2) . The latter finding indicates that these sera recognize LVspecific proteins. The antipeptide sera obtained for ORF2 reacted negatively in the IPMA. However, antFORF2 serum 690 immunoprecipitated the gene product of ORF2 synthesized in vitro in the presence of microsomal membranes, whereas the preserum did not recognize this product ( Fig. 3 ; Table 2 ). Furthermore, no cross-reactivity of serum 690 with other in vitro translation products was observed, proving that the anti-ORF2 serum 690 is specific for the ORF2 product (data not shown). Similarly, antipeptide sera 694 (anti-ORF3), 698 (anti-ORF4), 704 . 3 . RIP of in vitro translation products of ORFs 2 to 5 and 7, with gene-specific antipeptide sera raised in rabbits. The translation of transcripts was performed in the presence of microsomal membranes. The gene products of ORFs 2, 3, 4, 5, and 7 were immuneprecipitated with sera 690 (anti-ORF2), 694 (anti-ORF3), 698 (anti-ORF4), 704 (anti-ORF5), and 714 (anti-ORF7), respectively, and their corresponding preserum, which is indicated with the extension -P.

(anti-ORF5), and 714 (anti-ORF7) immunoprecipitated specifically the in vitro translation products of ORFs 3, 4, 5, and 7, respectively ( Fig. 3 ; Table 2 ). The glycosylation of the in vitro translation products synthesized for these RIP experiments was less complete than that in previous experiments. Therefore, also partially glycosylated products were observed in the immunoprecipitation reactions with ORFs 2 to 5. Although serum 710 (anti-ORF6) reacted positively in the IPMA, it did not recognize the ORF6 gene product synthesized by in vitro translation in the RIP assay. We hypothesize that the conformation of ORF6 in this RIP assay is unfavorable for exposure of the linear peptide epitope recognized by serum 710.

The generated gene-specific antipeptide sera were used to identify the proteins incorporated in virus parti-cles. Lelystad virus was purified on a glycerol-di-K-tartrate gradient or sucrose gradient, and infectious peak fractions, found at densities of 1.16-1.17 g/cm 3, were subjected to Western blot analyses using convalescent serum 21 and the gene-specific antipeptide sera. Serum 21 recognized three structural proteins with an apparent molecular weight (MW) of 25, 18, and 15 kDa (Fig. 4A ). in addition to these three proteins, two faint bands of 28 and 42 kDa were observed. These were not detected on the control strip incubated with a negative pig serum. Antipeptide serum 704 (specific for ORF5), recognized the 25-kDa protein (E) and a faint band of 42 kDa. We concluded that the E protein is encoded by ORF5. The much fainter protein band observed at 42 kDa might be a home-.or heterodimer of the ORFS-encoded protein, still present to a limited extent under denaturing conditions. Antipeptide serum 710 stained an 18-kDa protein (M), indicating that this structural protein is expressed from ORF6. Antipeptide serum 714 reacted with a 15-kDa protein (major band) and a 28-kDa protein (minor band). This finding proves that the 15-kDa protein (N) is encoded by ORF7. The 28-kDa protein is probably a dimeric form of the N protein. No staining was observed when these Western blot strips were incubated with 704, 710, or 714 presera or with antipeptide sera specific for ORFs 2, 3, and 4.

The results of the in vitro translation experiments indicated that the gene product of ORF5 was N-glycosylated, whereas the gene products of ORFs 6 and 7 were net. To establish whether these proteins are similarly glycosylated in vivo, we treated purified LV with PNGaseF. The treated and untreated virus preparations were analyzed on Western blots stained with convalescent serum 21 and the antipeptide sera specific for ORFs 5, 6, and 7. As is shown in Fig. 4B , after treatment with PNGaseF the apparent MW of the E protein was reduced to approxirnately 17 to 18 kDa. The apparent MW of the dimer of the E protein was reduced from 42 to 34 kDa. The size 0fthe M and N proteins remained the same before and after treatment with PNGaseF. These results show that E is an N-glycosylated structural protein, whereas M and N are not. Furthermore, these results are in line with our endoglycosidase studies on in vitro translation products, mentioned above. If the apparent size difference (_+7 kDa) between the unglycosylated and glycosylated E protein correlates with true molecular weight, this would suggest that both putative N-glycosylation sites are functional in vivo. The 17-to 18-kDa backbone of the unglycosylated E protein (which was also observed after PNG-aseF treatment of the ORF5-encoded gene product translated in the presence of microsomal membranes)was smaller than the ORF5-encoded gene product (20 kDa) translated in the absence of microsomal membranes (Fig. 1) . This finding suggests that the signal sequence of ORF5 is removed during membrane insertion. The E protein was also treated with EndoH (an endoglycosidase that digests N-linked high-mannose N-glycans but not complex N-glycans) and was shown to be resistant, although not completely (Fig. 4C) . A shift in apparent molecular weight of +1 kDa was observed. We cannot explain this slight reduction in size after EndoH treatment. The EndoH digestion of other substrates was complete, under identical conditions. Although proteins containing both EndoH-resistant and EndoH-sensitive oligosaccharides have been observed by others (Vennema et al., 1990; Doyle eta/., 1986) , a shift in apparent molecular weight of +1 kDa, if correlating with true molecular weight, is too low to account for one EndoH-sensitive oligosaccharide. The E protein was completely resistant to endo-/~-galactosidase (Fig. 4C ). This indicates that E, in contrast to G~, the counterpart of E in EAV (de Vries etaL, 1992) , does not contain N-acetyllactosamine residues.

In order to identify additional (non)structural LV-specific proteins, porcine anti-LV serum 21 and the genespecific sera were also used to immunoprecipitate pro-35 reins from LV-infected L-[ S]methionine-and L-[3~S]cysteine-labeled cell lysates. Convalescent serum 21 was used to immunoprecipitate the E, M, and N proteins from LV-infected cell lysates (Fig. 5) . The N protein was also immunoprecipitated with specific antipeptide serum for 0RF7 (Table 2) though the size of the M protein seemed to be somewhat larger in cell lysates (Fig. 5 ) than in the virion (Fig. 4) , this size difference is most likely due to the use of a different percentage of acrylamide and a different gel electrophoresis system (mini-gel) for the gel of Fig. 5 , compared to the gels prepared for the Western blots, shown in Fig. 4 . The gene products of ORFs 2, 3, and 4 could not be detected with serum 21, nor with specific antipeptide sera, which were otherwise able to immuneprecipitate the in vitro translation products of ©RFs 2 t o 4 .

Although the replication strategy of LV has been studied, and the complete nucleotide sequence of the viral genome has been determined, little is known about the structural proteins of LV. The purpose of our study was to analyze these proteins in more detail. Three protein components of LV virions were identified. These proteins have apparent molecular weights of 15, 18, and 26 kDa and are designated N, M, and E, respectively. The 15-kDa protein (N) is encoded by ORF7 and is not N-glycosylated. Its amino acid sequence is extremely basic and has 41 and 20% amino acid identity with the nucleocapsid protein of LDV (Chert etaL, 1993; Godeny eta/., 1990) and EAV (den Boon etaL, 1991a), respectively. The identity of the nucleocapsid proteins of EAV and LDV was established by virus fractionation experiments. After the virus particles were treated with detergent, only the N proteins of 15 to 16 kDa cosedimented with the viral genome in the bottom fractions of a sucrose gradient (Brinton-Darnell and Plagemann, 1975; Michaelides and Schlesinger, 1973; Zeegers et a/., 1976) . In a sucrose gradient layered with NP-40-treated LV virions, the N protein was found in the bottom fraction of the gradient at a density of 1.18 g/cm 3, whereas the M and E proteins cosedimented in the middle of the sucrose gradient at a density of 1.10 g/cm 3. This supplies further evidence supporting the assumption that the 15-kDa N protein of LV is the nucleocapsid protein.

The 18-kDa protein (M) is a non-N-glycosylated envelope protein encoded by ORF6, which has the same hydrophobicity profile as the M proteins of MHV (Boursnell, 1984) and IBV (Rottier et al., 1986) , the E protein of BEV (den Boon et al., 1991b) , the M protein of EAV (de Vries eta/., 1992) , and the gene product of ORF6 of LDV Godeny et al., 1993) . These proteins are characterized by the presence of three hydrophobic segments at the N-terminus. Protease protection experiments have shown that the M proteins of MHV-A59 and IBV are type Ill integral membrane proteins (Rottier eta/., 1984; Cavanagh et al., 1986) . They are anchored in the membrane by the three successive hydrophobic domains, whereas the C-terminal part is thought to be associated with the membrane surface. We assume that the M protein of LV has a similar membrane-associated topology.

The structural protein of 25 kDa, designated E, is the gene product of ORF5. We showed that E is N-glycosylated, probably at two different sites. The E protein incorporated in virus particles was sensitive to PNGaseF, but partially resistant to EndoH. These findings indicate that during virus maturation, E is transported through the Golgi apparatus and its N-linked oligosaccharides undergo Golgi-specific modifications. The apparent size of E decreased about 1 kDa when virions were treated with EndoH, indicating that it contains an EndoH-sensitive Nglycan moiety. However, if this reduction correlates with true molecular weight, it is too small to account for one N-glycan moiety being removed by EndoH. We cannot explain this reduction.

E is the counterpart of G~, a structural envelope protein encoded by ORF5 of EAV (de Vries et al., 1992) . G~ migrated as a heterogeneous protein of 30 to 42 kDa on SDS-PAGE, because a variable number of lactosamine repeats were added to the N-linked core oligosaccharide. The E protein of LV, however, was not susceptible to endo-/3-galactosidase. Therefore the maturation of the N-linked oligosaccharide side chains of the E and G~ protein is probably different. The topology of the E protein is still unknown. Both the E and G~ proteins contain a large internal hydrophobic region, which might anchor these proteins in the membrane.

The N, M, and E proteins were detected both in Western blot analysis of purified virus and in RIP assays of cell lysates. These proteins were also detected by Nelson et a/. (1993) in cell lysates of 0L2621 cells infected with LV or with a United States isolate of LV. Although we have generated gene-specific antipeptide sera, which recognized the N-glycosylated in vitro translation products of ORFs 2, 3, and 4, we were not able to detect the proteins encoded by these ORFs in cell lysates or purified LV. Perhaps these proteins were expressed only at very low levels in CL2621 cells and only small amounts of these proteins were incorporated in virus particles. Furthermore, the affinity of the polyvalent sera and monospecific peptide sera might not be high enough to detect such low amounts of protein. Alternatively, the ORF2, 3, and 4 gene products may be present in heterocomplexes that are not denatured under the conditions used. We do not yet know whether the gene products of ORFs 2 to 4 are structural proteins. The hydrophobicity profile of ORF2 of LV is similar to that of ORF2 of EAV and LDV. The gene product of ORF2, a 25-kDa glycoprotein designated Gs, was detected in a purified virus preparation of EAV (de Vries et aL, 1992) . It constituted only 1-2% of the virion protein and was not recognized by a polyvalent antivirion serum. Therefore it is not clear whether this protein is only incidentally incorporated into virions or is present in the virion preparation because it is contaminated with cellular proteins.

Although LV is able to grow in CL2621 cells, we were not able to perform a single step growth curve in these cells, since not all cells are infected after 1 hr absorption at an m.o.i, of 10. The virus growth and cell labeling conditions must be improved in order to increase the amount of label incorporated in proteins of LV, which may allow us to perform pulse-chase experiments in labeled cells and RIP assays on purified labeled LV particles. In addition, the generation of more monospecific or monoclonal antibodies recognizing the gene products of ORFs 2 to 6 will be useful for further elucidation of the function and topology of these proteins. Until now, only monoclonal antibodies recognizing the N protein of LV have been isolated (Nelson et aL, 1993; Drew eta/., 1994) .

Prosite, a dictionary of sites and patterns in proteins

Sequence of the membrane protein gene from avian coronavirus IBV

Structure and chemical-physical characteristics of lactate-dehydrogenase-elevating virus and its RNA

Coronavirus IBV glycopeptides: Locational studies using proteases and saponin, a membrane permeabalizer

Sequences of 3' end of genome and of 5' end of open reading frame la of Lactate dehydrogenase-elevating virus and common junction motifs between 5' leader and bodies of seven sugenomic mRNAs

Isolation of swine infertility and respiratory syndrome virus (Isolate ATCC-VR

North America and experimental reproduction of the disease in gnotobiotic pigs..L Vet. Diagn. Invest

Molecular characterization of porcine reproductive and respiratory syndrome virus, a member of the Arterivirus group

Another triple-spanning envelope protein among intracellularly budding RNA viruses= The tor0virus E protein

Structural proteins of equine arferitis virus

Analysis of progressive deletions of the transmembrane and cytoplasmic domains of influenza hemagglutinin

Production, characterisation and reactivity of monoclonal antibodies to porcine reproductive and respiratory syndrome (PRRSV)

Map location of lactate dehydrogenase-elevating virus (LDV) capsid protein (Vpl) gene

Complete genomic sequence and phylogenetic analysis of the lactate-dehydrogenase-elevating virus (LDV)

Formalin inactivation of the lactate dehydrogenase-elevating virus reveals a major neutralizing epitope not recognized during natural infection

Prediction of protein antigenic determinants from amino acid sequences

Glycoprotein E1 of Hog Cholera virus expressed in insect cells protects swine from hog cholera

Surface and inside volumes of globular proteins

A simple method for displaying the hydrophatic character of a protein

Cleavage of structural proteins during the assembly of the head of bacteriophage T4

Clinical presentation of mystery pig disease in the breeding herd and suckling pig}ets

Lelystad virus, the causative agent of porcine epidemic abortion and respiratory syndrome (PEARS) is related to LDV and EAV

Subgenomic RNAs of Lelystad virus contain a conserved junction sequence

Structural proteins of lactic dehydrogenase-elevating virus

Comparison of the structural protein coding sequences of Lelystad virus and VR2332 strains of the PRRS virus

New hydrophilicity scale derived from high performance liquid chromatography peptide retention data, Correlation of predicted surface residues with antigenicity and X-ray-derived accessible sites

Lactate dehydrogenaseelevating virus, equine arteritis virus, and simian hemorrhagic fever virus-A new group of positive-strand RNA viruses

Characterization of bovine viral diarrhea proteins

Assembly in vitro of a spanning membrane protein of the endoplasmic reticulum, The E1 gtycoprotein of coronavirus mouse hepatitis virus A59

Predicted membrane topology of the coronavirus protein El

Molecular Cloning, A Laboratory Manual

Modified ELISA techniques for anti-hapten antibodies

Electrophoretic transfer of proteins from polyacrylamide gels to nitro cellulose sheets= Procedure and some applications

Intracellular transport of recombinant coronavirus spike proteins= Implications for virus assembly

Equine arteritis virus-infected cells contain six potyadenylated virusspecific RNAs

A new method for predicting signal sequence cleavage sites

Lelystad virus and the porcine epidemic aborion and respiratory syndrome

Antigenic comparison of Lelystad virus and swine infertility and respiratory symdrome (SIRS) virus..L Vet

Mystery swine disease in the Netherlands= The isolation of Lelystad virus

The structural proteins of equine arteritis virus

We thank J. Langeveld for the synthesis and analysis of the synthetic peptides and P. Rottier for fruitful discussions. We thank D. Benfield and T. Drew for the supply of monoclonal antibodies. Par[ of this work was supported by Boehringer-lngelheim, Germany, and the Produktschap voor Vee en Vlees (PW), The Netherlands.