Identification and Characterization of Viral Structural Proteins of Infectious Salmon Anemia Virus

Abstract

Infectious salmon anemia virus (ISAV) is an unclassified Orthomyxovirus that has been shown to contain a segmented genome with eight single-stranded RNA species coding for 10 viral proteins. Four major structural proteins were characterized in the present study: two glycosylated proteins with estimated molecular masses of 42 and 50 kDa, one 66-kDa phosphoprotein, and one 22-kDa protein. Examination of lysed virions revealed the two glycoproteins and the 22-kDa protein in the soluble fraction, while the 66-kDa phosphoprotein and a minor part of the 22-kDa protein were found in the pelleted fraction. Immunofluorescence staining of infected cells demonstrated that the 22-kDa protein was a late protein accumulating in the nucleus. We conclude that the 66-kDa protein is the nucleoprotein, the 22-kDa protein is the matrix protein, and the 42- and 50-kDa proteins are the surface proteins. Radioimmunoprecipitation analysis of the 42-kDa glycoprotein, which was previously shown to represent the ISAV hemagglutinin, indicated that this protein exists at least as dimers. Further, by labeling of purified ISAV with [1,3-³H]diisopropyl fluorophosphate, it was also demonstrated that the viral esterase is located with the hemagglutinin. This finding was confirmed by demonstration of acetylesterase activity in affinity-purified hemagglutinin preparations. Finally, the active-site serine residue could be tentatively identified at position 32 within the amino acid sequence of the hemagglutinin of ISAV strain Glesvaer/2/90. It is proposed that the ISAV vp66 protein be termed nucleoprotein, the gp42 protein be termed HE protein, and the vp22 protein be termed matrix protein.

Infectious salmon anemia (ISA) is one of the most important viral diseases in farmed Atlantic salmon (Salmo salar L). The disease was first described for salmon parr in a hatchery on the southwest coast of Norway in 1984 (34), and later the disease was also observed in Canada, Scotland, the Faroe Islands, and the United States (2, 4, 26, 30). In addition, the ISA virus has been observed in Pacific coho salmon (Oncorhyncus kisutch) in Chile (1, 13) and in rainbow trout (Oncorhyncus mykiss) in Ireland.

The ISA virus has been shown to share several morphological, physiochemical, and biochemical characteristics with the influenza viruses, suggesting that it belongs to the Orthomyxoviridae (12, 18, 24). These characteristics include a single-stranded RNA genome that consists of eight segments with a tentative negative polarity (24), conservation of partially self-complementary terminal ends of the genomic segments (32), a replication strategy that both includes RNA production and protein accumulation in the nucleus (12, 32), morphological similarities, and the ability to hemagglutinate red blood cells (RBC) (12). The nucleotide sequences of all ISAV genome segments have now been published (3, 6, 18, 19, 24, 29, 33). In common with the influenza A and B viruses, segments 7 and 8 reveal two open reading frames (ORFs), while the others contain one ORF each. No significant sequence homology with the influenza viruses was found, although some sequence similarities suggesting coding assignment of the different genes have been reported.

Influenza A and B viruses contain four major structural proteins and six minor proteins. The major proteins include the nucleoprotein (NP), which interacts with the viral RNA together with the polymerases (PB1, PB2, and PA) to form the viral ribonucleoprotein (vRNP) complex. The matrix protein (M1) is the most abundant structural protein and is believed to form a bridge between the vRNPs and the viral membrane. The two other major structural proteins of the virion are integral to the viral envelope. The hemagglutinin (HA) is a homotrimeric glycoprotein with receptor-binding and acid-activated membrane fusion activity, and the neuraminidase is a homotetramer with receptor-destroying enzyme (RDE) activity. Influenza C virus has only seven gene segments and has one surface glycoprotein, the HA-esterase (HE) protein. Further, while the RDEs of influenza A and B viruses are neuraminidases, the RDE of influenza C virus is an esterase (21).

ISAV has been shown to possess hemagglutinating, RDE, and fusion activities (9, 12). The RDE has been suggested to be an acetylesterase with specificity different from that of influenza C virus (12). Recently, it was found that 4-O-acetylated sialic acids serve both as a substrate for the RDE and as a receptor determinant for virus binding (12a). Four major structural proteins with estimated molecular masses of 71, 53, 43, and 24 kDa have been detected in purified ISAV particles (12, 14). Further, by immunoprecipitation of lysates from radiolabeled infected cells, Kibenge et al. (14) detected a total of 12 proteins. It is currently unknown whether these additional proteins represent minor structural proteins, precursor proteins, or breakdown products. Based on functional analysis and protein expression, the 43-kDa protein has been suggested to represent the HA and to be coded by genome segment 6 (6, 17, 28). The 71-kDa protein has been suggested to be encoded by segment 3 and to represent the NP, though no experimental data supporting this assumption have been presented (6, 29). Recently the 24-kDa structural protein was linked to one of the two ORFs of segment 8 (3), and the 53-kDa was linked to segment 5 (6); however, no function or characterization of these two proteins has so far been presented.

In the present study we confirm the four major structural proteins and also document that the virus has one surface glycoprotein in addition to the earlier-described HA. We have also demonstrated, for the first time, that the HA is an HE protein. Furthermore, we provide both experimental data showing that the 22-kDa major structural protein is the matrix (M) protein and data indicating that the 66-kDa protein represents the NP.

MATERIALS AND METHODS

Virus and cells.

The Norwegian ISA virus isolate Glesvaer/2/90 (7) was used throughout this study. Cultures of SHK cells (7) or ASK cells (8) were used for virus propagation. The cells were grown at 20°C in Leibovitz L-15 medium (L-15) supplemented with 5% (SHK cells) or 10% (ASK cells) fetal bovine serum, glutamine (4 mM), and gentamicin (50 μg ml⁻¹). Cells were incubated at 15°C after inoculation with virus. Unless otherwise stated, cells with the fluid overlay removed were inoculated with virus at a multiplicity of infection of 0.1 in serum free L-15. All assays were performed with the fourth or fifth passage of the virus. Infectivity titrations were done by end point titration in 96-well culture plates as previously described (11).

For preparation of radiolabeled whole virus, 25 μCi of [³⁵S]methionine (ProMix; Amersham Biosciences) ml⁻¹ was added to SHK cell cultures in 175-cm² tissue culture flasks at 1 day postinfection (p.i.), and incubation was continued for 3 days before harvest of the virus-containing cell culture medium. Radiolabeled cell lysates were prepared by infecting confluent monolayers of ASK cells in 25-cm² tissue culture flasks. Following a 4-h adsorption period, 350 μCi of [³⁵S]methionine, 350 μCi of [³H]mannose (Amersham Biosciences), or 350 μCi pf [³²P]phosphate (Amersham Biosciences) ml⁻¹ in 2.5 ml of methionine-, mannose-, or phosphate-free medium was added, and incubation was continued for 24 h. The cell cultures were then washed twice with PBS and lysed on ice for 15 min in lysis buffer (50 mM Tris [pH 7.6], 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate), and insoluble debris were removed by centrifugation at 14,000 rpm (20,000 × g) for 20 min at 4°C. Lysates were stored at −80°C until used.

Antibodies.

Antibodies used in the present study included a monoclonal antibody (MAb) directed at the ISAV HA (anti-HA MAb) (11), a rabbit antibody against purified whole virus (anti-ISAV polyclonal antibody [PAb]) (12), and a rabbit antibody against a recombinant protein based on the first ORF in ISAV genome segment 8 (anti-rVp22 PAb) (3). All polyclonal antisera were adsorbed by incubation on monolayers of acetone-fixed SHK or ASK cells prior to use.

Virus purification.

Purified whole virus was prepared by sucrose gradient centrifugation of polyethylene glycol 8000-precipitated cell culture supernatant as described previously (12). Alternatively, virus was affinity purified directly from cell culture supernatants by using anti-mouse immunoglobulin G (IgG)-conjugated immunomagnetic beads (Dynabeads M-280; Dynal, Oslo, Norway) and the anti-HA MAb. Briefly, the beads were washed twice in PBS (10 mM phosphate [pH 7.4], 150 mM NaCl) with 0.1% bovine serum albumin (BSA) (PBS-BSA), incubated with the MAb in PBS for 1 h at room temperature (RT), and washed four times in PBS-BSA. Culture supernatant with [³⁵S]methionine-labeled virus was then incubated with the coated beads at 4°C for 3 h (0.1 mg of beads per 1 ml of supernatant). Finally, the beads were washed four times in PBS-BSA, resuspended in dissociation buffer, and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by autoradiography (see below).

RIPA of virus-specific proteins.

Immune precipitation of radiolabeled viral antigen was performed on crude cytosol extracts. Antibody-coated magnetic beads (see above) were used for the precipitation. Radiolabeled cell lysates were first absorbed with uncoated beads for 1 h at RT, and then 250 μg of coated beads was incubated with 10 μl of cell lysate in 100 μl of PBS for 1 to 2 h at RT. Antibody-antigen complexes were washed two times with radioimmunoprecipitation assay (RIPA) buffer (PBS containing 1% Triton X-100, 0.1% BSA, 1% sodium deoxycholate, 0.5 M lithium chloride, and 0.1% SDS), once with PBS-BSA, and once with Tris-HCl (50 mM; pH 6.8). Finally, the beads were resuspended in dissociation buffer and analyzed by SDS-PAGE followed by autoradiography (see below). Noninfected, radiolabeled cell lysates were used as controls.

SDS-PAGE and Western blotting.

Samples were treated with dissociation buffer (50 mM Tris-HCl [pH 6.8], 1% SDS, 50 mM dithiothreitol, 8 mM EDTA, 0.01% bromophenol blue) and heated for 5 min at 95°C. In the nonreduced samples, dithiothreitol was omitted and incubation was for 30 min at RT. The proteins were then separated by SDS-PAGE with the discontinuous system devised by Laemmli (20) and 0.5-mm-thick precast 12.5% polyacrylamide gels (ExcelGel SDS; Amersham Biosciences). ¹⁴C-methylated marker proteins (Amersham Biosciences; range, 14.3 to 220 kDa) were run in adjacent lanes. For autoradiography, the gel was fixed in 10% acetic acid-40% ethanol in double-distilled water for 30 min, treated with Amplify (Amersham Biosciences) containing 3% glycerol for 30 min, dried, and exposed to X-ray film at −80°C.

For Western blotting, separated proteins were transferred to a nitrocellulose membrane in a semidry electroblotter (NovaBlot; Amersham Biosciences). The protein blot was then treated with blocking solution (PBS containing 1% Tween 20 and 5% nonfat dry milk) overnight at 4°C, followed by incubation for 2 h at RT with the primary rabbit immune serum diluted in blocking solution. The membrane was washed three times in blocking solution and reacted with a goat anti-rabbit-horseradish peroxidase (HRP) conjugate for 1 h at RT. Diaminobenzidine was used for detection of bound HRP conjugate.

For staining of glycosylated proteins with lectins, the protein blot was first treated with PBS with 0.1% Tween 20 and 5% BSA overnight at 4°C. The protein blot was then reacted with 15 mg of biotinylated concanavalin A (Sigma) ml⁻¹ diluted in PBS with 0.1% Tween 20 and 1% BSA. The membrane was washed three times in PBS with 0.1% Tween 20 and reacted with HRP-conjugated streptavidin for 1 h at RT. Bound HRP-conjugate was detected as described above.

Labeling of ISAV protein with [1,3-³H]diisopropyl fluorophosphate (DFP) (New England Nuclear; specific activity, 310.8 GBq/mmol) was done by incubating purified ISAV, containing 2 mU of esterase activity, with 74,000 Bq of [³H]DFP for 1 h at RT, followed by an overnight incubation at 4°C. Reduced and nonreduced viral proteins were analyzed by electrophoresis on SDS-12 and 8% polyacrylamide gels, respectively. The gels were then incubated with Amplify, dried, and subjected to fluorography.

The autoradiographs and blots were scanned in a desktop scanner (Image Scanner; Amersham Biosciences) and subsequently analyzed and printed by using Gel-Pro gel scanning software (Media Cybernetics, Silver Spring, Md.).

Deglycosylation of viral glycoprotein.

Immunoprecipitated [³⁵S]methionine-labeled viral proteins were heated at 95°C for 5 min in TNE buffer (10 mM Tris-HCl [pH 7.2], 100 mM NaCl, 1 mM EDTA) with 0.05% SDS. N-Octylglycoside (Roche) was added to 0.5%, and then the mixture was briefly heated again to 95°C and incubated overnight at 37°C with or without 2 U of N-glycosidase-F (Roche) ml⁻¹. Samples were subsequently prepared in dissociation buffer, followed by SDS-PAGE and autoradiography.

Trypsin treatment of virus-specific proteins.

Immunoprecipitated [³⁵S]methionine-labeled viral proteins were treated with 20 μg of trypsin (Promega) ml⁻¹ in TBS (50 mM Tris [pH 7.2], 100 mM NaCl) for 30 min at 37°C and immediately prepared in dissociation buffer, followed by SDS-PAGE and autoradiography.

Dissociation of virions.

Virus metabolically labeled with [³⁵S]methionine was purified by sucrose gradient centrifugation. Purified virus was then diluted in lysis buffer (50 mM Tris [pH 7.6], 100 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, Complete Mini protease inhibition mix [Roche]) and incubated on ice for 15 min. The lysate was then pelleted through a cushion of 20% sucrose in lysis buffer at 50,000 rpm for 2 h at 4°C in a Beckman SW-60 rotor. The supernatant was collected and precipitated with trichloroacetic acid at a final concentration of 10% on ice for 30 min. The precipitate was washed three times with ethanol-ether (1:1), and both this precipitate and the pelleted fraction were resuspended in dissociation buffer and analyzed by SDS-PAGE, followed by autoradiography.

Hemagglutination and HI assays.

Hemagglutination in microtiter plates was carried out as previously described (12). The hemagglutination inhibition (HI) test with the serine esterase inhibitors DFP (Sigma) and 3,4-dichloroisocoumarin (DCIC) (Sigma) was performed by first mixing 25 μl of inhibitor, diluted in PBS, with 25 μl of virus suspension containing 4 HA units, followed by incubation for 1 h. Fifty microliters of 0.5% Atlantic salmon RBC, 0.75% rabbit RBC, or 0.75% horse RBC was then added, and the agglutination end points were read following 1 and 6 h of incubation. All incubations were performed at RT. The hemagglutination activity (HA units) is expressed as the reciprocal value of the highest dilution showing complete agglutination of RBC.

Affinity purification of HA and detection of acetylesterase activity.

The viral HA was affinity purified by using anti-HA MAb-coated immunomagnetic beads as described above for purification of whole virus. Briefly, 500 μg of coated beads was added to 300 μl of ISAV cell lysate and incubated for 30 min at RT. The beads were then removed and washed, and both beads and cell lysate were assayed for esterase activity.

The acetylesterase activity was determined by incubating 25 μl of sample with 1 ml of 1 mM p-nitrophenyl acetate (pNPA) (Sigma) in double-distilled water. The release of acetate was monitored by determining the optical density at 400 nm. One unit of viral esterase was defined as the amount of enzymatic activity resulting in cleavage of 1 μmol of pNPA per min (16). Inhibition of acetylesterase activity was tested by incubating virus samples and inhibitors for 30 min at RT before testing for acetylesterase activity.

Immunofluorescence.

Indirect immunofluorescence staining was performed on ISAV-infected ASK cells grown on coverslips. Following incubation, the cells were washed once in PBS, fixed in 10% buffered formalin for 10 min, rinsed in PBS, and subsequently permeabilized with 2% Triton X-100 in PBS for 30 min. After blocking with 5% nonfat dry milk in PBS for 30 min, the cells were incubated for 1 h with the primary antibody diluted in PBS with 2.5% nonfat dry milk. Alexa Fluor 546 goat anti-mouse IgG (Molecular Probes) and Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes) in PBS were used for the detection of bound antibodies. PBS was used for washing, and all incubations were performed at RT. The cell preparations were mounted with the SlowFade antifade kit (Molecular Probes) and examined in a confocal microscope (Leica TCS-SP).

Sequence alignments.

For analysis we used the amino acid sequences of the European ISAV strain Glesvaer (accession number AAF32361) and the antigenically different Canadian isolate ISAV Bay of Fundy/97 (AAK30064) (15, 19). Accession numbers for the other HE sequences used for alignment are AAA43785 for influenza C virus, AAA46442 for mouse hepatitis virus strain JHM, P15776 for bovine coronavirus strain Mebus, and CAA71819 for Breda virus. The sequences were aligned by using the CLUSTAL method as described previously (38).

RESULTS

ISA virus structural proteins.

Radioactively labeled virus preparations purified either by sucrose gradient centrifugation or by affinity purification with immunomagnetic beads coated with anti-HA MAb were analyzed by SDS-PAGE (Fig. 1a, lanes 1 and 2). Both purification methods gave essentially the same results, revealing four major structural proteins, as observed earlier by Falk et al. (12). The molecular masses were calculated based on the average of 10 different lanes, and the three most distinct bands were estimated to be approximately 22, 50, and 66 kDa, respectively, while the wide, less distinct band had a molecular mass peaking at 42 kDa (range, 40 to 43 kDa). By scanning the autoradiograms of the gels and performing a densiometric analysis, integrating the area under each peak, the relative amounts of the 22-, 42-, 50-, and 66-kDa bands were estimated to be approximately 40, 37, 11, and 12%, respectively.

FIG. 1.

(a) Autoradiograms of [³⁵S]methionine-labeled ISAV proteins resolved in SDS-12.5% polyacrylamide gels. Lanes: 1, whole virus purified by sucrose gradient centrifugation; 2, whole virus purified by affinity purification with immunomagnetic beads and an anti-HA MAb; 3, immunoprecipitation of cell lysates from virus-infected cell cultures with an anti-whole virus antibody (30- and 82-kDa proteins are indicated by asterisks); 4, control lysates of [³⁵S]methionine-labeled cells. (b) Detection of glycoproteins. Lanes: 1, Western blot of purified virus stained with biotinylated concanavalin A; 2, immunoprecipitation of cell lysates from virus-infected cells labeled with [³H]mannose; 3, control lysates of [³H]mannose-labeled cells; 4, [³⁵S]methionine-labeled ISAV proteins were immunoprecipitated, followed by deglycosylation; 5, nondeglycosylated control preparation; 6, control lysates of [³⁵S]methionine-labeled cells. (c) Detection of phosphoproteins. Lanes: 1, immunoprecipitation of cell lysates from virus-infected cells labeled with ³²P by using an anti-whole virus antibody; 2, control lysates of ³²P-labeled cells.

By RIPA of [³⁵S]methionine-labeled cell lysates of infected cells with anti-ISAV PAb, we also found two minor proteins migrating at 30 and 82 kDa (Fig. 1a, lane 3), which probably represented precursor proteins, breakdown products, dimeric forms of one of the major proteins, or minor viral proteins. Comparing cell lysate RIPA protein profiles with those obtained with purified virus preparations, it was also evident that the relative amount of the 66-kDa protein had increased while the amount of the 22-kDa protein had decreased, probably reflecting differences in the immune response to these proteins in the rabbit. A relatively weak reaction to the 22-kDa protein was also seen on Western blots (see Fig. 3b, lane 3) supporting this observation.

FIG. 3.

Dissociation of ISAV with NP-40. [³⁵S]methionine-labeled virus was purified by sucrose gradient centrifugation, treated with 1% NP-40, loaded onto a 20% sucrose cushion, and centrifuged. The supernatant (lanes 1) and pellet (lanes 2) were then examined by SDS-12.5% PAGE. (a) Autoradiograms of the two fractions. (b) Western blot analysis of the two fractions stained with a PAb to recombinant vp22. Lane 3 is included for reference and contains purified virus stained with a polyclonal anti-whole virus antibody.

Glycosylation of the major proteins was examined both by staining of Western blots with biotinylated concanavalin A (Fig. 1b, lane 1) and by RIPA of cell lysates labeled with [³H]mannose (Fig. 1b, lane 2). The results revealed three bands comigrating with the 42-, 50-, and 82-kDa proteins, indicating the glycosylated nature of these proteins. This was further confirmed by deglycosylation experiments performed with [³⁵S]methionine-labeled viral proteins precipitated with the anti-ISAV PAb. The results of these experiments are shown in Fig. 1b, lane 4, and indicate that the deglycosylated gp50 has a molecular mass of 45 kDa while deglycosylated gp42 resulted in a wide band peaking at 35 kDa (range, 34 to 37 kDa). The protein band just below the 45-kDa band in Fig. 1b, lane 4, comigrated with a corresponding band in the control (Fig. 1b, lane 6) and was considered nonspecific.

Analysis of cell lysates labeled with inorganic ³²P revealed one major band comigrating with the 66-kDa major protein. (Fig. 1c, lane 1). Two minor bands were also found in the control preparations (Fig. 1c, lane 2), and thus the 66-kDa protein is the only major phosphorylated protein of the virion.

Trypsin treatment of proteins obtained by precipitation with the anti-HA MAb revealed that the gp42, as well as the gp82 also found by precipitation with this antibody, was almost completely digested, leaving one band at 34 kDa (Fig. 2a, lane 1), indicating that the HA could be subjected to proteolytic activation similar to that seen with influenza virus HA. However, when proteolytic activation of the influenza virus HA occurs, the two resulting proteins are bound together by S-S binding. To test whether this had happened during the trypsin digestion of the ISAV HA, nonreduced preparations were examined. The results revealed a digestion profile similar to that with reduced samples, indicating no S-S-binding (Fig. 2a, lane 3) and thus making the possibility of proteolytic activation of the HA less likely. The control in this experiment (Fig. 2a, lane 2) demonstrates, for the first time, that the anti-HA MAb binds to gp42.

FIG. 2.

(a) Effect of trypsin treatment of immunoprecipitated [³⁵S]methionine-labeled ISAV HA. Following treatment, proteins were resolved in SDS-12.5% polyacrylamide gels. Lanes: 1, trypsin treatment of proteins immunoprecipitated with an anti-HA MAb; 2, nontreated control of proteins in lane 1; 3, nonreduced sample of trypsin-treated proteins in lane 1; 4, nontreated control of proteins in lane 3. (b) SDS-8 to 18% polyacrylamide gels of nonreduced, non-heat-treated samples of the ISAV HA protein immunoprecipitated from [³⁵S]methionine-labeled ISAV cell lysate, demonstrating that the HA protein exists at least as dimers and probably as trimers (lane 1). Lane 2 is included for reference and contains proteins immunoprecipitated with a polyclonal anti-whole virus antibody.

Furthermore, in RIPA experiments with the anti-HA MAb examined in SDS-12% polyacrylamide gels, we also observed an 82-kDa glycoprotein. This protein was also seen, to a greater or less degree, in RIPA experiments with the anti-ISAV PAb. In addition, we also frequently observed a protein band in the interphase between the 5% stacking and 12% resolving gels. To find out whether these observations could be explained as a result of dimeric and trimeric HA, too tightly bound to be completely resolved by SDS-PAGE, we examined nonreduced, non-heat-treated samples of precipitated HA in an 8 to 18% gradient gel. The results are shown in Fig. 2b, lane 1, revealing three protein bands and thus suggesting that the ISAV HA exists at least as dimers and probably as trimers. It was also seen that the mass of gp82 relative to gp42 in these nonreduced samples had increased as measured by densitometry of scanned films, from approximately 1:5 to 1:2, further supporting this assumption.

Location of the proteins in the virion.

Purified [³⁵S]methionine-labeled ISA virions were solubilized with 1% NP-40. Following ultracentrifugation of the detergent-treated virions through a sucrose layer, the supernatant and the pelleted fraction were recovered and analyzed by SDS-PAGE. The results are presented in Fig. 3a and show that this treatment solubilized gp42, gp50, and a major part of vp22 (Fig. 3a, lane 1), leaving a minor part of vp22 together with all of vp66 in the pelleted fraction (Fig. 3a, lane 2).

To demonstrate that the vp22 found in the pelleted fraction was similar to the vp22 found in the soluble fraction, the experiment was repeated and the two fractions were analyzed by Western blotting with rabbit antibodies to a recombinant protein encoded by the first ORF of genome segment 8 (anti-rVp22 PAb), which was recently shown to be associated with vp22 (3). The results of this experiment are presented in Fig. 3b and show that this antibody detects vp22 both in the solubilized and in the pelleted fraction. An additional protein was detected in the pelleted fraction (Fig. 3b, lane 2); however, this protein did not colocalize with any of the ISAV proteins, as shown in the control preparation (Fig. 3b, lane 3). The most reasonable interpretation of these results is that the two glycoproteins in the soluble fraction are associated with the viral envelope; that the nonglycosylated vp22 found in both soluble and pelleted fractions is analogous to the M protein of influenza A, B, and C viruses; and that the vp66 found in the pelleted fraction is the NP.

To further confirm the identity of the suggested M protein, infected cell cultures on coverslips were fixed every 4 h and immunostained with the anti-rVp22 PAb. Parallel sections were stained with the anti-whole virus antibody (anti-ISAV PAb) in order to determine the relative time of appearance of the vp22. For reference purposes, the preparations were double stained with antibodies to the HA (anti-HA MAb). The results are presented in Fig. 4 and show that vp22 first appears in the nucleus as well as in the cytoplasm together with the HA at 24 h p.i. (Fig. 4b). In contrast, staining with the anti-whole virus antibody revealed an early protein in the nucleus, presumably the NP, at 12 h p.i. (Fig. 4c) The NP in other influenza viruses is known to be an early protein which initially accumulate in the nucleus (21). This demonstrates both that the vp22 accumulates in the nucleus during the infectious cycle and that vp22 and HA are late proteins.

FIG. 4.

Immunofluorescent staining of ISAV-infected ASK cells with PAbs to recombinant vp22 (green) (a and b) and whole virus (green) (c and d). All preparations were double stained with the anti-HA MAb (red). Neither vp22 nor HA was detected at 12 h p.i. (a and c), while virus protein was detected in the nucleus (c). At 24 h p.i., vp22 was detected in the nucleus (b), while HA was detected in the cytoplasm (b and d). The nuclear staining in panel d probably represents both the nucleoprotein and vp22.

Biological activities associated with the virus particle.

To characterize hemagglutination and esterase activity, the two serine hydrolase inhibitors DFP and DCIC were tested both by HI, which showed an eventual lack of elution of the hemagglutination reaction, and by the pNPA test. The results of the HI experiments are presented in Fig. 5 and show that these inhibitors did not inhibit hemagglutination, while both DFP and DCIC inhibited elution from rabbit erythrocytes by the esterase at concentrations of as low as 0.1 mM and 12.5 μM, respectively. Similar results were obtained with horse and rainbow trout RBC. The inhibition experiments were further confirmed by testing the esterase inhibitors on purified virus and measuring esterase activity in solution by using the pNPA esterase assay. The results revealed a similar inhibition of the esterase activity (data not shown).

FIG. 5.

Inhibition of virus elution by DFP and DCIC. Four HA units of virus was incubated with either 0.1 to 5 mM DFP or 12.5 to 100 μM DCIC, followed by addition of 0.75% rabbit RBC. Controls: 0, no inhibitor; C, no virus. (a) Incubation for 1 h, showing complete agglutination in all wells. (b) After 6 h of incubation, the agglutination had eluted in the control wells but not in wells containing esterase inhibitors.

To determine the location of the esterase activity, purified virus was incubated with [^1,3H]DFP, followed by SDS-PAGE analysis of reduced and nonreduced samples, followed by autoradiography. This approach was used earlier to identify the HE proteins of influenza C virus (25, 37) and of bovine coronavirus (36). When we incubated purified ISAV Glesvaer/2/90 with [1,3-³H]DFP, we identified a single tritiated protein after SDS-PAGE and autoradiography. This protein comigrated with the 42-kDa HA protein, as shown by Western blot analysis (Fig. 6a). In nonreduced samples, an additional weak band was seen at 82 kDa, comigrating with the suggested dimeric HA (Fig. 6b). Since the viral esterase activity is inhibited by DFP, it appears likely that the labeled proteins represents the HA and the esterase.

FIG. 6.

Analysis of [1,3-3H]DFP-labeled ISAV protein. Purified ISAV was incubated with [1,3-³H]DFP for 1 h at room temperature, followed by an overnight incubation at 4°C. Viral proteins were analyzed by electrophoresis on a reducing SDS-12% polyacrylamide gel (a) or a nonreducing SDS-8% polyacrylamide gel (b). The gels were then cut into two pieces, and one part of each gel was blotted onto nitrocellulose and probed with ISAV-specific antiserum. The rest of the gel was incubated with Amplify (Amersham-Pharmacia), dried, and subjected to fluorography. Lanes: 1, Western blot; 2, [1,3-³H]DFP.

In order to confirm these results, the ISAV HA protein was depleted from a Triton X-100 lysate of infected cells by using immunomagnetic beads coated with the anti-HA MAb. Following the first round of depletion, new antibody-coated beads were added to the original virus lysate in order to pick up any remaining HA protein. Altogether, three rounds were performed, and the lysate and beads were tested for esterase activity after each round by using the pNPA esterase assay. The results are presented in Table 1 and show that the esterase activity in the lysate decreased after each round, leaving the activity on the immunomagnetic beads, which confirms that the esterase activity is indeed located on the viral HA protein. Radioimmunoprecipitation of [³⁵S]methionine-labeled Triton X-100 cell lysates with the same coated beads was used to confirm that gp42, but not gp50, was precipitated by the beads, and the results were similar to those presented in Fig. 2a, lane 2.

TABLE 1.

Repeated depletion of ISAV esterase from a cell lysate of ISAV-infected cells by using immunomagnetic beads coated with an anti-ISAV HA MAb

Precipitation	Esterase activity measured by pNPA assay in:
	Lysate (mU)	Beads (absorbancea)
Control	6.9	0
First round of depletion	2.3	0.52
Second round of depletion	1.2	0.33
Third round of depletion	0.3	0.03

^aOnly absorbance values were measured, as it was not possible to determine units in these samples.

These findings were also corroborated by a comparison of the amino acid sequence of the ISAV protein with the HEs of influenza C virus, group 2 coronaviruses, and Breda torovirus. As shown in Fig. 7A, substantial sequence similarities were observed with the E1 and E′ regions (31) of the influenza C virus esterase. Specifically the region around the active-site serine residue is conserved among all viral esterases. Amino acid sequences around the other residues of the catalytic triad are less well conserved. The putative active-site residues were detected at positions 32 (serine) (Fig. 7A) and 261 (aspartic acid) and 264 (histidine) (Fig. 7B) of the ISAV Glesvaer sequence. We found no substantial sequence similarities to the HAs of influenza A and B viruses. Our finding is consistent with the lack of esterase activity in those viruses. As demonstrated earlier by the structural analysis of the influenza C virus surface glycoprotein by Rosenthal et al. (31), the RDE domain is absent in the HA proteins of influenza A and B viruses. Instead of an esterase, those viruses possess a sialidase activity, which is encoded by the neuraminidase gene (21).

FIG. 7.

Alignment of amino acid sequences around the active-site residues. The HA sequences of ISAV Glesvaer/2/90 and ISAV Bay of Fundy/97 were aligned by using the CLUSTAL algorithm with the sequences around the active-site serine residue (A) and Asp-352/His-355 (B) of the HE proteins of influenza C virus, mouse hepatitis virus strain JHM (MHV JHM), bovine coronavirus (BCoV) strain Mebus, and Breda torovirus. Amino acid residues identical to the consensus sequence are in boldface, and gaps are indicated by hyphens. Amino acid residues are numbered according to the GenBank entries. Active-site residues are marked by asterisks.

DISCUSSION

Virus purified by two independent methods revealed the same four major structural proteins that have been described earlier (12, 14). However, the estimated molecular masses, i.e., 22, 42, 50, and 66 kDa, differed slightly, probably due to the use of different molecular mass standards. Further, these new estimates are closer to the predicted molecular masses, which are 22.0, 42.4, 48.7, and 68.0 kDa, respectively (3, 6, 17). Metabolic labeling with tritiated mannose or with ³²P established that the 42- and 50-kDa proteins were glycosylated, while the 66-kDa protein was the only phosphorylated structural protein. The glycosylated nature of the gp42 and gp50 was further substantiated by lectin binding and deglycosylation experiments.

When purified virions were solubilized by NP-40 treatment followed by ultracentrifugation, the two glycoproteins and the nonglycosylated vp22 were recovered in the solubilized fraction. These results confirm that the two glycoproteins are surface proteins and also suggest that the nonglycosylated vp22 represents the ISAV M protein, analogous to the influenza virus M1 protein. The major protein in the pelleted fraction was the phosphorylated vp66, and there was also a minor fraction of vp22. Thus, vp66 represents the viral NP, which is the major part of the RNP complex, together with the viral RNA and the viral polymerases, in influenza viruses. Also, in influenza viruses this complex can be pelleted following solubilization. The finding of M protein in the pelleted fraction was not surprising, as complete separation of M1 from the RNP complex by using nonionic detergents is difficult to achieve and is dependent on NaCl concentration and pH in both orthomyxo- and paramyxoviruses (39). Attempts to demonstrate the effect of NaCl concentration and pH on solubilization were not successful, although indications that higher NaCl concentration promoted the separation of these proteins were observed (data not shown). The finding that vp66 represents the NP is supported by computer simulation studies of gene segment 3, which is thought to code for the vp66 (6, 29).

By immunofluorescence and confocal microscopy with an anti-vp22 immune serum, we were able to demonstrate that the putative ISAV M protein both accumulated in the nucleus and was a late protein. The M1 protein of the influenza viruses is the most abundant protein and has an important structural role in the influenza virus virion (21). It is located on the interior side of the lipid bilayer, forming a shell which surrounds the RNP complex. After entry of the virus through endosomal vesicles, the M1 protein undergoes a pH-dependent conformational change, resulting in release of the vRNPs into the cytoplasm. This is followed by vRNP import into the nucleus, where the viral genome is transcribed and replicated (21). In the course of the infection cycle, the M1 accumulates in the nucleus like the NP does, but in contrast to the NP, the M1 protein is a late protein, and it has been shown to inhibit viral transcription and may contribute to the shift of viral RNA synthesis towards replication in the late phase of infection (5, 21, 22). Furthermore, the nuclear accumulation of the M1 protein is a prerequisite for the nuclear export of newly synthesized vRNPs in the late phase of infection (23). Thus, based on the facts that the ISAV vp22 is the most abundant protein in the virion, that it is the only nonglycosylated soluble protein, and that it is a late protein that accumulates in the nucleus during the infectious cycle, we conclude that the vp22 is the ISAV M protein, analogous to the influenza virus M1 protein.

In ISAV we found only the NP protein to be phosphorylated. This was surprising, as previous studies have indicated major similarities between the ISAV and influenza virus structures and reproduction cycles (12, 32). In influenza A, B, and C viruses, both NP and M1 are phosphorylated, a property that it is thought to be associated with the active transport of these proteins in and out of the nucleus during the infection cycle (5, 27).

Previously, gp42 has been shown to hold the hemagglutinating activity and to be encoded by ISAV gene segment 6 (17, 28). Our findings further support this conclusion, showing the glycosylated nature of this protein, which it is solubilized by NP-40, and, by using RIPA, that the anti-ISAV HA MAb actually binds to gp42 (Fig. 2, lane 4). Both the RIPA using the anti-ISAV HA MAb with either [³⁵S]methionine-labeled or [3H]mannose-labeled cell lysates and the [1,3-³H]DFP-labeling experiment revealed a minor 82-kDa glycosylated protein. This 82-kDa glycoprotein had approximately twice the molecular mass of gp42, suggesting that it might be a dimer of gp42. When nonreduced, non-heat-treated samples were examined in an 8 to 18% gradient gel, the relative proportion of this protein increased and a third protein also was detected (Fig. 2b, lane 1). However, this possible trimeric form of gp42 was not detected in the [1,3-³H]DFP-labeling experiment, which could be due to low sensitivity in this assay. Thus, the existence of a trimeric form of gp42 remains to be further documented.

In addition to the hemagglutinating activity, the surface glycoproteins of the influenza viruses exhibit RDE activity and fusion activity. Both of these activities have been detected in ISAV (9, 12); however, the localization of these activities has so far not been determined. Also, the RDE of ISAV has been suggested to be an acetylesterase different from that found in influenza C viruses (12). We demonstrated here that the ISAV RDE, but not hemagglutination, is inhibited by the serine hydrolase inhibitors DFP and DCIC, further supporting the idea that the ISAV RDE indeed is an acetylesterase. Both of these inhibitors have been shown to be potent inhibitors of the influenza C virus esterase (37).

In addition to influenza C viruses, several coronaviruses, including human and bovine coronavirus, hemagglutinating encephalomyelitis virus, mouse hepatitis virus, puffinosis coronavirus and sialodacryoadenitis virus, have also been shown to possess an esterase as the RDE (10, 35). All these viruses have a surface HE glycoprotein; i.e., both receptor binding and RDE activities are located on the same protein and they bind to cellular receptors, which contain O-acetylated sialic acids as the major receptor determinant. However, it should be noted that in coronaviruses the second surface glycoprotein, termed spike protein, is the major receptor-binding protein, whereas the coronavirus HE is a minor HA. On the other hand, both influenza A and B viruses, which also possess two surface structural glycoproteins, have hemagglutinating and fusion activities on one protein and RDE (a neuraminidase) activity on the other (21). We now present data showing that we have identified another HE protein in ISAV. By labeling purified ISAV with the serine hydrolase inhibitor [1,3-³H]DFP, we detected one tritiated protein band on SDS-PAGE that comigrated with the viral HA. This finding was further confirmed by the demonstration of acetylesterase activity in preparations of affinity-purified viral HA. Finally, by comparing the amino acid sequence of the ISAV HA protein with those of the HEs of influenza C virus and several coronaviruses, substantial sequence similarities were observed. Studies of the three-dimensional structure of influenza C virus have revealed two different binding sites for O-acetylated sialic acids: the receptor-binding domain at the top of the trimeric molecule and the catalytic site, which is present at a position closer to the viral membrane. The active site is built by a catalytic triad composed of a serine residue, a histidine residue, and an aspartic acid residue, located at positions 57, 352, and 355, respectively (30). In ISAV strain Glesvaer/2/90, these residues were identified at positions 32, 261, and 264, respectively. Thus, in analogy to the HEs of influenza C virus and coronaviruses, it is proposed that the ISAV protein be renamed HE protein.

Data suggesting proteolytic activation of ISAV by trypsin treatment during replication in cell cultures have previously been presented (12). In the present work, trypsin treatment of immunoprecipitated gp42 followed by SDS-PAGE analysis revealed a 35-kDa breakdown product, suggesting that this protein could be subjected to proteolytic activation. However, examination of nonreduced samples revealed the same results, indicating no S-S binding of the cleaved products, and therefore it is not likely that gp42 is subjected to proteolytic activation. In the influenza viruses, HA or HE is synthesized as a precursor, which upon cleavage activation generates the two subunits, the HA₁ (or HE₁) and the N terminus of HA₂ (or HE₂) bound together by an S-S bridge. This cleavage primes the HA or HE for activation of its fusion activity and is required for the viruses to be infectious (21). The function of the gp50 surface glycoprotein is currently not known. Based on the fact that the hemagglutinating and esterase activities are associated with gp42, it is tempting to suggest that this protein holds the fusion activity. However, no documentation or indications supporting such an assumption have so far been presented, and hence this question remains to be resolved. In summary, we provide experimental evidence that the ISAV vp66 represents the phosphorylated NP, the gp42 is the viral HE, and the vp22 represents the analogue to the influenza virus M protein.