Characterization of the Coronavirus Mouse Hepatitis Virus Strain A59 Small Membrane Protein E

Cells and viruses.

Mouse L cells, OST7-1 cells (9), and RK13 cells were maintained in Dulbecco modified Eagle medium (DMEM) containing 10% heat-inactivated fetal calf serum (FCS)–100 IU of penicillin per ml–100 μg of streptomycin per ml (DMEM–10% FCS), supplemented in the case of OST7-1 cells with 400 μg of G-418 (Geneticin; Gibco BRL) per ml. BHK-21 cells and CEF cells were maintained in Glasgow minimal essential medium (MEM) containing the same additional ingredients.

MHV-A59 was propagated in Sac(−) cells as described previously (42). The recombinant vaccinia viruses vTF7-3 and MVA-T7 expressing bacteriophage T7 RNA polymerase were propagated in RK13 and CEF cells, respectively, as described previously (11, 43).

Expression vectors.

The expression construct pT7Ts-E carrying the MHV-A59 E gene behind a T7 promoter was made by using the vector pT7Ts (a kind gift of P. Krieg). The E gene was obtained as a PCR fragment amplified from cDNA clone pRG86 (2) by using primer 471 (5′-dCACGCAGCTCGAAACATATGTTTA-3′) and primer 496 (5′-dGGATTAGATATCATCCACCTCTA-3′). It was cloned into pNoTa (5 Prime→3 Prime, Inc.). From the resulting plasmid pNoTa/T7-5b, the E gene was subsequently retrieved with NdeI and BamHI, and this fragment was treated with Klenow DNA polymerase. pT7Ts was first digested with EcoRI and XbaI and then treated with Klenow enzyme and religated. The plasmid was then cut with SpeI and treated again with Klenow DNA polymerase, and finally the E gene fragment was ligated into it to yield pT7Ts-E.

The expression plasmid pTUM-M containing the MHV-A59 M gene behind a T7 promoter has been described elsewhere (28), as has pAVI02, a pBluescript KS(+) construct carrying the equine arteritis virus (EAV) G_s gene (6).

Antibodies.

To obtain a serum against the MHV-A59 E protein, we constructed a plasmid encoding an N-terminally His-tagged fusion protein of the major part of the EAV G_s protein ectodomain linked to the C-terminal 34-residues sequence of the E protein (Fig. 1B). To this end, we amplified the relevant region of the plasmid pNoTA/T7-5b by PCR using the primers 485 (5′-dGGAATTCTTTGGTGCTGTCCCCTTC-3′; nucleotides 15554 to 15571 of the MHV-A59 RNA sequence) and the M13 reverse sequence primer (Pharmacia). This PCR product was digested with EcoRI and HindIII and cloned into pBluescript SK(−) to generate pBM.cE. To obtain the fragment encoding the 161-residues G_s polypeptide, we performed a PCR with the primers 483 (5′-dCGGGATCCATCGCCCGCAGCTTGG-3′, nucleotides 9897 to 9914 of the EAV RNA sequence) and primer 484 (5′-dGGAATTCAACACTTCCACAGATGAAGC-3′; genomic positions 10362 to 10382). This fragment was cloned as a BamHI/EcoRI fragment into pBluescript SK(−) to generate pBE.GS-ecto. Then the EcoRI/HindIII fragment from pBM.cE was cloned into pBE.GS-ecto behind the EAV fragment which resulted in the unintentional insertion of two codons at the G_s-E junction site adding two amino acids (Asn and Ser). The chimeric G_s-E fragment was cloned as a BamHI fragment into the His-tag vector pQE10 (Qiagen) yielding pGSectoEendo. The protein was expressed in Escherichia coli, isolated, and purified on a Ni-nitriloacetic acid column (Invitrogen) according to the manufacturer's instructions. Antibodies were elicited by subcutaneous injection of the fusion protein (75 μg in Freund complete adjuvant) into a rabbit from which blood had been taken the day before to obtain a preserum (designated preE). The rabbit was boosted similarly after 3, 7, 11, and 15 weeks using 250 μg of protein in Freund incomplete adjuvant each time. Animals were bled 17 weeks after the first immunization.

The monoclonal antibody J1.3 (a kind gift of J. Flemming) recognizes the N-terminal domain of the MHV-A59 M protein (designated αM_N), as has been described (5). Also described previously was a rabbit serum against a peptide corresponding to the M protein's C terminus (αM_C [19rsqb;), as well as a polyclonal rabbit serum against MHV virions (αMHV [32rsqb;).

Infection, transfection, and metabolic labeling.

Subconfluent monolayers of OST7-1 cells grown in 10-cm² dishes were washed with phosphate-buffered saline (PBS) containing 50 μg of DEAE-dextran per ml (PBS-DEAE) and 1% of FCS and inoculated using this medium with MHV-A59 at a multiplicity of infection (MOI) of 10 to 50 for 60 min at 37°C.

For expression of the MHV-A59 M and E genes and of the EAV G_s gene, subconfluent monolayers of OST7-1 or BHK-21 cells in 10-cm² dishes were washed with DMEM and inoculated with vTF7-3 or MVA-T7 at an MOI of 10 in DMEM for 45 min at 37°C. Cells were then washed with DMEM and transfected with the plasmid of choice. For this purpose, 200 μl of DMEM containing 10 μl of Lipofectin reagent (Life Technologies, Inc.) was mixed with 5 μg of plasmid DNA and added to the cells. After a 10-min incubation at room temperature, 800 μl of DMEM was added, and the cells were incubated further initially at 37°C for 1 h and subsequently at 32°C for the indicated times; in the case of MVA-T7-driven expression the cells were kept at 37°C at all times.

For radioactive labeling of proteins, MHV-A59-infected cells and cells expressing genes from plasmids were starved for 30 min by incubation with MEM lacking methionine and cysteine (GIBCO) and supplemented with 10 mM HEPES (pH 7.4). Cells were labeled with 80 μCi of ³⁵S In Vitro Labeling Mix (Amersham) in the same medium for the times indicated. For pulse-chase experiments, the cells were washed twice after the labeling with prewarmed chase medium (DMEM–10% FCS, 2 mM methionine and cysteine, and 10 mM HEPES [pH 7.4]) and incubated further in the same medium.

Immunoprecipitation, immunoisolation, and protein analysis.

After labeling, cells were washed once with PBS containing 50 mM Ca²⁺ and 50 mM Mg²⁺. They were then lysed on ice in 1 ml of lysis buffer (20 mM Tris, pH 7.6; 150 mM NaCl; 1% Nonidet P-40 [NP-40]; 0.5% deoxycholate; and 0.1% sodium dodecyl sulfate [SDS]) per 10-cm² dish. The lysates were cleared by centrifugation for 15 min at 13,000 rpm at 4°C in an Eppendorf centrifuge. To analyze proteins present in the culture media, cell supernatants were taken off, cleared by centrifugation for 15 min at 4°C and 4,000 rpm, and then mixed with a one-fifth volume of a 5× concentrated lysis buffer: 100 mM Tris (pH 7.6), 150 mM NaCl, 5% (vol/vol) NP-40, 2.5% (wt/vol) deoxycholate (DOC), 0.5% (wt/vol) SDS. To analyze proteins present in cells and culture media together, as was done in the experiments of Fig. 2 and 3, total lysates were prepared by adding a one-fifth volume of 5× concentrated lysis buffer to the culture media. The combined lysates of cells and media were cleared by centrifugation as described above for conventional cell lysates.

For immunoprecipitation of proteins, 100- or 150-μl aliquots of lysates were diluted to 1 ml with immunoprecipitation buffer (IP buffer; 20 mM Tris [pH 7.6], 150 mM NaCl, 5 mM EDTA, 0.5% [vol/vol] NP-40, 0.1% [wt/vol] DOC, 0.1% [wt/vol] SDS), and antibodies (3 μl of rabbit sera and 100 μl of monoclonal antibody tissue culture supernatant) were added. The solutions were incubated at 4°C for at least 3 h, after which 30 μl of Pansorbin (Calbiochem) was added, and the incubation continued for at least 1 h. Immune complexes were then collected by centrifugation, washed three times with wash buffer I (20 mM Tris [pH 7.6], 150 mM NaCl, 5 mM EDTA, 0.1% NP-40) and once with wash buffer II (20 mM Tris [pH 7.6], 0.1% NP-40).

Immunoisolation of virus particles was performed by diluting 450 μl of cleared cell culture supernatant with a similar volume of TNE (20 mM Tris [pH 7.6], 50 mM NaCl, 1 mM EDTA) containing antibodies (3 μl of rabbit serum or 100 μl of monoclonal antibodies). The solution was incubated overnight at 4°C, after which 30 μl of Pansorbin was added to each sample. Incubation was continued for 1 h, after which immune complexes were collected by centrifugation and washed three times with TNE.

For analysis of the proteins, the washed immune complexes were resuspended in 30 μl of Laemmli sample buffer containing 50 mM dithiothreitol (DTT) and analyzed by electrophoresis in SDS-polyacrylamide gels (PAG). The gels were fixed in 10% acetic acid–50% methanol for at least 30 min and incubated for another 30 min in 1 M sodium salicylate. Finally, the gels were dried and subjected to fluorography at −80°C. Radioactivity in protein bands in PAG was quantitated by using a PhosphorImager and ImageQuant (Molecular Dynamics) analysis.

Proteinase K treatment of MHV virions.

After labeling of MHV-A59-infected OST7-1 cells, the culture medium was collected, cleared, and adjusted to 0.01 M Tris (pH 7.8) and 1 mM CaCl₂ (proteinase K reaction buffer) by using a 100× concentrated stock solution. Proteinase K was then added from a stock solution to some samples to a concentration of 150 μg/ml, and all samples were incubated for 1 h on ice. The enzymatic treatment was stopped by adding phenylmethylsulfonyl fluoride (PMSF) to a concentration of 7.5 mM while keeping the samples on ice. A one-fifth volume of concentrated lysis buffer was then added to each sample, and immunoprecipitations were performed as detailed above but in the presence of 5 mM PMSF in all solutions.

In vitro transcription and translation.

In vitro transcription reactions were carried out by using T7 RNA polymerase (Boehringer Mannheim) according to the manufacturer's instructions in 50-μl volumes containing 4 μg of PstI-linearized pT7Ts-E. After a 1-h incubation at 37°C, the RNA was purified by chromatography in a Sephadex G50 column and precipitated with ethanol, and the dried pellet was dissolved in 50 μl of TNE.

Translations of the mRNAs (0.8 μl of RNA transcript in a 10-μl reaction) were done for 1 h at 30°C in the Promega rabbit reticulocyte lysate system in the presence or absence of canine microsomal membranes (Promega). The samples were then split and diluted to 1 ml with IP buffer. Antisera (3 μl of the rabbit preE serum or the anti-E serum) were added, and the normal immunoprecipitation procedure was followed.

Immunofluorescence analysis of protein membrane topology.

BHK-21 cells expressing the MHV-A59 E or M protein were surface permeabilized by using Streptolysin O (SLO; purchased from S. Bhakdi, Johannes Gutenberg-Universität, Mainz, Germany). At 7 h postinfection, culture media were taken off and cells were rinsed with SLO buffer (25 mM HEPES [pH 7.4], 115 mM potassium acetate, 2.5 mM Mg₂Cl). They were then incubated for 15 min on ice with SLO buffer containing 1 mM DTT and 1 μg of SLO per ml and subsequently washed twice with SLO buffer containing 1 mM DTT. To activate the SLO, the cells were incubated for 30 min at 37°C with prewarmed SLO buffer containing 1 mM DTT. The cells were put on ice and directly rinsed with SLO wash buffer (50 mM HEPES [pH 7.4], 5 mM Mg₂Cl, 2 mM EGTA, 50 mM KCl). This washing step was repeated once, and the cells were then washed with PBS and subjected to fixation with 3% paraformaldehyde (PFA) for 30 min. Proteinase K treatment of SLO-permeabilized cells was carried out by preincubation of the cells after the PBS washing step for 5 min with proteinase K buffer, followed by incubation for 30 min on ice with proteinase K buffer containing 50 μg of proteinase K per ml. Cells were then rinsed twice with proteinase K wash buffer (2 mM EGTA, 5 mM EDTA, 1.15 mM PMSF [pH 7.4]) and once with PBS containing 50 mM glycine and 1.15 mM PMSF, after which they were fixed with 3% PFA.

For classical indirect immunofluorescence, permeabilization of all cellular membranes was performed after fixation with 3% PFA by treatment for 5 min with 1% Triton X-100–5% FCS in PBS–50 mM glycine. The cells were washed three times with PBS containing 50 mM glycine and 5% FCS.

Fluorescence labeling of the cells with antibodies was done by incubating them with primary antibodies, followed by fluorescein isothiocyanate (FITC)- or tetramethyl rhodamine isocyanate (TRITC)-conjugated secondary antibodies. The primary antibodies used were rabbit anti-E (1:400), mouse anti-M_N (undiluted), and rabbit anti-M_c (1:300) sera. The secondary antibodies used were FITC- or TRITC-conjugated goat anti-mouse and goat anti-rabbit immunoglobulin G (Cappel; 1:100).

Electron microscopy.

Mouse L cells infected with MHV-A59 and BHK-21 cells expressing the E protein were fixed at 5.30 and 4 h postinfection (hpi), respectively, and prepared for cryosectioning as described earlier (21). Thawed cryosections were double labeled with the E antibody at a dilution of 1:100 and with Rab-1 antibodies (36) diluted 1:40. The double labeling was carried out according to the sequential protocol of Slot et al. (39). Embedding of cells in Epon was done as described elsewhere (13).

Kinetics of appearance of the E protein in MHV-A59-infected cells.

Sera raised against purified MHV-A59 are very poor in detecting the E protein. Thus, a fusion protein was prepared in order to obtain protein E-specific antibodies for immunological studies. It consisted of the 34-amino-acid carboxy-terminal domain of MHV-A59 E protein preceded by the amino-terminal 161-residue ectodomain of the EAV G_S protein (Fig. 1B). By choosing G_S as a fusion partner to enhance immune responses, we aimed to also acquire a polyclonal serum to this EAV protein. For purification purposes, the fusion protein was extended amino terminally with a His-tag. The protein was produced in E. coli and injected into rabbits. The immune serum obtained was found to immunoprecipitate a protein with an apparent molecular mass of 9.8 kDa from lysates of radiolabeled cells infected with MHV-A59 or expressing the viral E gene by using the vaccinia virus T7 system (data not shown). This molecular mass closely corresponds to the predicted size of the E protein (9.6 kDa).

We used the antiserum to evaluate the synthesis of the E protein during MHV infection, comparing it with that of the other viral structural proteins. Cultures of infected cells were therefore labeled for successive 1-h periods with ³⁵S-labeled amino acids. Immunoprecipitates were prepared from the cell lysates by using a polyclonal rabbit serum against MHV-A59 that recognizes predominantly the N, M, and S proteins and by using the rabbit serum against the E protein. Precipitates were analyzed in 15 and 20% PAG, respectively, as shown in Fig. 2A and B. For each of the viral proteins the time course of synthesis was determined by quantitation of the radioactivity in the respective gel segments, the results of which, normalized to allow comparison, are compiled in Fig. 2C. It is clear that all four proteins were synthesized with very similar kinetics. They were all first detected at between 3 and 4 hpi, after which their production rapidly and simultaneously increased, reaching a maximal rate around 6 hpi, which was maintained for several hours. Since the proteins were probably all immunoprecipitated with different efficiencies, these results do not allow an estimation of the relative molar ratios of their synthesis.

Stability of the E protein during infection.

To further study the fate of the E protein in infection, we performed a pulse-chase analysis, during which we again monitored the radioactivity in the E protein in comparison with that in the other structural proteins. In order to obtain an overall picture, we included in our analysis the fraction of the proteins that is released from the cells with virions. For this purpose we prepared total lysates in which cells and culture fluids were combined. Immunoprecipitates of M, S, and N proteins were prepared with the anti-MHV serum and analyzed in a 15% gel (Fig. 3A); E protein precipitated with the E-specific serum was electrophoresed in a 20% gel (Fig. 3B). Quantitations of the radioactivity in the different proteins are graphically represented in Fig. 3C. The data show that the structural proteins synthesized at around 6 hpi turned over with quite similar kinetics. It was deduced from this and other experiments that the proteins had half-lives of approximately 4 h. A similar pattern of turnover was obtained when an otherwise identical experiment was performed in which protein synthesis was blocked after the pulse-labeling by including cycloheximide (0.5 mM) in all chase media (data not shown). This result ruled out the possibility that the decline in the amount of radioactive proteins was caused by a gradual shortage of antibodies during the immunoprecipitation as a result of the continued synthesis of (unlabeled) viral proteins.

Inspection of Fig. 3A reveals the appearance during the chase of a polypeptide (marked by an asterisk) running slightly faster in the gel than the normal set of M proteins. This polypeptide, which could also be observed in the experiment depicted in Fig. 2, appeared concomitantly with the conversion of the unglycosylated M protein (M₀) synthesized during the pulse into the different, slower-migrating O-glycosylated forms. Using domain-specific antibodies, we have demonstrated that the polypeptide represents an M protein that lacks its normal amino terminus (data not shown). Its discrete size, which is about 2.4 kDa smaller than that of the M₀ form, indicates that it has completely lost the domain that is exposed at the luminal side of intracellular membranes, i.e., at the outside of virions.

Stability of individually expressed E protein.

We also studied the stability of the independently expressed E protein in a pulse-chase experiment. Because no E protein appeared to be released into the culture medium (data not shown) only the cell-bound protein was analyzed. For comparison, we similarly expressed in parallel the M protein (known to be stable [34rsqb;) and the EAV G_S protein, which we found earlier to be subject to degradation both in EAV-infected BHK-21 cells and when expressed individually in these cells (7). As the results compiled in Fig. 4 show, the M and G_S proteins behaved as expected, although the latter protein, immunoprecipitated by the antiserum directed against G_S-E fusion protein, appeared to be less prone to turnover in the OST7-1 cells than in the BHK-21 cells used before. The stability of the E protein was comparable with that of the G_S protein under these conditions; its level remained constant initially but started to decrease thereafter.

Membrane integration of E occurs without signal sequence cleavage.

The MHV-A59 E protein contains a rather long hydrophobic domain at its amino terminus which is likely to mediate its membrane integration. To find out whether its membrane insertion involves the functioning of a classical, cleavable signal sequence, we compared the size of the primary translation product with that of the membrane-integrated form. Capped RNA was transcribed in vitro from an E gene plasmid construct and translated in a reticulocyte lysate system in the absence and in the presence of rough microsomal membranes. The labeled products were immunoprecipitated by using the E antiserum, and their electrophoretic mobilities were analyzed in parallel with that of E protein synthesized in transfected cells by using the vaccinia virus expression system. Figure 5A shows that the E proteins synthesized in vitro and in cells comigrated. Apparently, no signal sequence is cleaved during or after membrane integration of the protein unless the size reduction that would accompany such cleavage is compensated for by a secondary modification, for which we have no indications. The MHV-A59 E protein sequence does not contain an N-glycosylation motif. Moreover, as we show in Fig. 5B, we were unable to confirm the conclusion made by Yu et al. (49), on the basis of hydroxylamine treatment, that the E protein is acylated: in our experiment the protein remained unaffected. Attempts to label the protein in infected cells with ³H-labeled palmitic acid were equally unsuccessful (not shown). These observations are consistent with those of others who were similarly unable to label the TGEV E protein in infected cells with ³H-labeled palmitic acid (12).

Topology of the membrane-assembled E protein.

Besides the lack of a cleaved signal sequence at its amino terminus the E protein also lacks the typical anchor sequence occurring in the carboxy-terminal region of so many other membrane proteins. Rather, its hydropathy plot suggests that the polypeptide is buried within the lipid bilayer for most of its amino-terminal 60 residues. To determine the disposition of the residual more-hydrophilic carboxy terminus, we took an immunofluorescence approach based on the availability of an antiserum specifically recognizing this part of the protein. In this assay cells were permeabilized selectively in their plasma membrane by using SLO to allow intracellular access of antibodies. The assay was first tested with cells expressing the M protein whose membrane topology has been well established (33, 35): its amino- and carboxy-terminal domains are exposed luminally and cytoplasmically, respectively, and antibodies directed to these domains are available. Accordingly, surface-permeabilized cells could not be stained with the anti-M_N antibodies unless the intracellular membranes were also permeabilized by using Triton X-100 (Fig. 6). In contrast, the anti-M_C antibodies readily stained the surface-permeabilized cells, thus confirming the cytoplasmic exposure of the carboxy terminus. Consistently, staining was fully abrogated by prior treatment of the permeabilized cells with proteinase K, a treatment that only affected the cytoplasmically exposed protein domains since the amino terminus of the M protein remained intact and could still be visualized by the anti-M_N antibodies after additional Triton X-100 permeabilization. This latter treatment did not enable staining with the anti-M_C antibodies, indicating that the M protein assumes one defined membrane topology, in contrast to the TGEV M protein, which was reported to take two alternative orientations (31).

When the immunofluorescence assay was applied to cells expressing the E protein, it became clear that its carboxy terminus protrudes into the cytoplasm. The domain was accessible to the anti-E antibodies after surface permeabilization (Fig. 6). Proteinase K treatment completely abolished the staining, and subsequent permeabilization with Triton X-100 did not reveal any “hidden” carboxy-terminal domains, a result consistent with a unique membrane topology.

A cytoplasmic exposure of the E protein's carboxy terminus would topologically correlate with a disposition of this domain on the inside of MHV particles. Accordingly, we were unable to label VLPs, obtained by expressing the MHV M and E proteins, in an immunogold labeling with the anti-E antibodies, as judged under the electron microscope (data not shown). To further corroborate this point, we also used a biochemical approach. Because of its great sensitivity, we performed an immunoisolation of radioactively labeled virus particles from the culture medium of MHV-A59-infected cells by using various antibodies. As the protein pattern of Fig. 7A shows, the S-specific antibodies (Fig. 7A, αS) not only isolated S protein but in addition all other structural proteins M, N, and E, the latter one being visible only after prolonged exposure (not shown). Similarly, viral particles could be isolated with the anti-M_N antibodies but not with the anti-M_C antibodies, a finding consistent with the known exterior and interior disposition of the corresponding M domains, respectively. As expected, no trace of viral particles was detected when the anti-E serum was used in the immunoisolation, confirming the interior disposition of the E protein's carboxy terminus in virions.

Due to its very low abundance, the E protein is difficult to detected in MHV-A59 particles. We nevertheless analyzed the effect of proteolytically removing the exposed domains of the viral membrane proteins. Figure 7B shows that treatment of ³⁵S-labeled virions eliminated the epitope present in the amino terminus of the M protein recognized by the monoclonal J1.3 (αM_N). It also reduced the size of the polypeptide by about 2.5 kDa as shown previously (33). The E protein seemed unaffected. Its appearance as a diffuse band, visible only after long exposure of the gel, did not change as a result of the treatment. The mere fact that the protein remained precipitable by the anti-E antibody implies that the carboxy terminus was indeed protected. In addition, it shows that no other part(s) of the protein is significantly exposed on the outside of virions.

Electron microscopic observations.

One of the marked observations made during the immunofluorescence studies described above was the peculiar punctate staining pattern of the expressed E protein. To study this aspect in more detail, BHK-21 cells expressing the protein were prepared for and then analyzed by electron microscopy. It appeared that the E protein induced the formation of characteristic electron-dense structures (Fig. 8A) that were not seen in control cells, for instance, when the M protein was expressed similarly. They were, however, also observed in MHV-A59-infected L cells (as shown after cryosectioning in Fig. 8B) and are actually very reminiscent of the tubular structures described in MHV-A59-infected cells many years ago (4). They seem to consist of masses of tubular, smooth membranes with much curvature that form complicated networks. The membrane clusters are relatively homogeneous in that other organelles are excluded. Continuities with the endoplasmic reticulum (ER) are, however, often observed. The structures were obviously induced by the E protein since they labeled heavily with the anti-E antibody (Fig. 8B to D). In fact, in highly expressing cells the E protein often appeared to localize almost exclusively to them. Interestingly, the tubular structures are part of the ER-intermediate compartment (IC) network, as was demonstrated by their colabeling on cryosections with a marker for these compartments, Rab-1 (Fig. 8B to D; see also references 14 and 36).