|Home | About | Journals | Submit | Contact Us | Français|
Severe acute respiratory syndrome coronavirus (SARS-CoV) encodes 3 major envelope proteins: spike (S), membrane (M), and envelope (E). Previous work identified a dibasic endoplasmic reticulum retrieval signal in the cytoplasmic tail of SARS-CoV S that promotes efficient interaction with SARS-CoV M. The dibasic signal was shown to be important for concentrating S near the virus assembly site rather than for direct interaction with M. Here, we investigated the sequence requirements of the SARS-CoV M protein that are necessary for interaction with SARS-CoV S. The SARS-CoV M tail was shown to be necessary for S localization in the Golgi region when the proteins were exogenously coexpressed in cells. This was specific, since SARS-CoV M did not retain an unrelated glycoprotein in the Golgi. Importantly, we found that an essential tyrosine residue in the SARS-CoV M cytoplasmic tail, Y195, was important for S-M interaction. When Y195 was mutated to alanine, MY195A no longer retained S intracellularly at the Golgi. Unlike wild-type M, MY195A did not reduce the amount of SARS-CoV S carbohydrate processing or surface levels when the two proteins were coexpressed. Mutating Y195 also disrupted SARS-CoV S-M interaction in vitro. These results suggest that Y195 is necessary for efficient SARS-CoV S-M interaction and, thus, has a significant involvement in assembly of infectious virus.
Coronaviruses are enveloped positive-strand RNA viruses that infect a wide variety of mammalian and avian species. These viruses generally cause mild disease in humans and are one major cause of the common cold (34). However, severe acute respiratory syndrome coronavirus (SARS-CoV), a novel human coronavirus which emerged in the Guangdong province in China in 2002 (30, 48), caused a widespread pandemic. SARS-CoV caused severe disease with a mortality rate of approximately 10%, the highest for any human coronavirus to date (62). The phylogeny and group classification of SARS-CoV remain controversial (17), but it is widely accepted to be a distant member of group 2. While SARS-CoV is no longer a major health threat, understanding the basic biology of this human pathogen remains important.
Coronaviruses encode three major envelope proteins in addition to various nonstructural and accessory proteins. The envelope protein (E) is the least abundant structural protein in the virion envelope, although it is expressed at robust levels during infection (21). E plays an essential role in assembly for some but not all coronaviruses (31-33, 45) and may also be a viroporin (reviewed in reference 21). The spike glycoprotein (S) is the second most abundant protein in the envelope. S determines host cell tropism, binds the host receptor, and is responsible for virus-cell, as well as cell-cell, fusion (15). The S protein is a type I membrane protein with a large, heavily glycosylated luminal domain and a short cytoplasmic tail that has been shown to be palmitoylated in some coronaviruses (47, 58). The membrane protein (M) is the most abundant protein in the virion envelope and acts as a scaffold for virus assembly. M has three transmembrane domains, a long cytoplasmic tail, and a short glycosylated luminal domain (reviewed in reference 21). Unlike many enveloped viruses, coronaviruses assemble at and bud into the lumen of the endoplasmic reticulum (ER)-Golgi intermediate compartment (ERGIC) and exit the infected cell by exocytosis (29). In order to accomplish this, the envelope proteins must be targeted to the budding compartment for assembly.
For most coronaviruses, the E and M proteins localize in the Golgi region near the budding site independently of other viral structural proteins. We have previously shown for infectious bronchitis virus (IBV) E protein that the cytoplasmic tail contains Golgi targeting information (9). IBV M contains Golgi targeting information in its first transmembrane domain (57), while the transmembrane domains and cytoplasmic tail of mouse hepatitis virus (MHV) M appear to play a role in Golgi targeting (1, 36). Some coronavirus S proteins contain targeting information in their cytoplasmic tails; however, some do not (38, 39, 52, 63). Since S proteins can escape to the cell surface when highly expressed, S may rely on lateral interactions with other viral envelope proteins to localize to the budding site and be incorporated into newly assembling virions.
In line with its role in virus assembly, M is necessary for virus-like particle (VLP) formation (3, 10, 26, 40, 55, 59). M has been shown to interact with itself to form homo-oligomers (12). In addition, M interacts with E, S, and the viral nucleocapsid and is essential for virion assembly (reviewed in reference 21). Lateral interactions between the coronavirus envelope proteins are critical for efficient virus assembly. The interaction of S and M has been studied for MHV, and the cytoplasmic tail of each protein is important for interaction (16, 44). Specifically, deletion of an amphipathic region in the MHV M cytoplasmic tail abrogates efficient interaction with MHV S (11). The S and M proteins of IBV, bovine coronavirus, feline infectious peritonitis virus, and SARS-CoV have been shown to interact; however, information about the specific regions that are important for interaction remains elusive (16, 22, 26, 42, 64). Due to the presence of several accessory proteins in the virion envelope (23-25, 28, 51, 53), it is possible that the requirements for SARS-CoV S and M interaction could be different from those of previously studied coronaviruses.
In earlier work, we reported that SARS-CoV M retains SARS-CoV S intracellularly at the Golgi region when both proteins are expressed exogenously (39). We also demonstrated that the SARS-CoV S cytoplasmic tail interacts with in vitro-transcribed and -translated SARS-CoV M (39). Here, we show that the SARS-CoV M cytoplasmic tail is necessary for specific retention of SARS-CoV S at the Golgi region. We found a critical tyrosine residue at position 195 to be important for retaining SARS-CoV S Golgi membranes when coexpressed with M. When Y195 was mutated to alanine, the mutant protein, MY195A, did not reduce the amount of SARS-CoV S at the plasma membrane or reduce the extent of S carbohydrate processing as well as wild-type SARS-CoV M does. Additionally, mutation of Y195 in SARS-CoV M disrupted the S-M interaction in vitro. Thus, Y195 is likely to play a critical role in the assembly of infectious SARS-CoV.
pCAGGS/SARS-CoV M and pCAGGS/SARS-CoV S were previously described (39). QuikChange (Stratagene, La Jolla, CA) site-directed mutagenesis was used to introduce deletions in pcDNA3.1/SARS-CoV M (39), using primers that flanked the deleted regions, according to the manufacturer's protocol. The amino acids indicated were deleted or mutated for the following constructs: MΔ1 (Y177-T207), MΔ2 (L137-T207), MΔ1a (L205-D214), MΔ1b (Y195-K204), MΔ1c (R185-A194), MΔbc (S191-R199), MS190A, MG191A, MF192A, MA193V, MY195A, MR197A, MY198A, and MY203. pcDNA3.1/SARS-CoV MY195A was created using primers that changed the nucleotides coding for SARS-CoV M Y195 to A195 according to the manufacturer's protocol. The sequenced deletions and mutations were then subcloned into the mammalian expression vector pCAGGS-MCS (43) (a kind gift from P. Bates) to increase expression levels using KpnI and XhoI sites. cDNA encoding vesicular stomatitis virus (VSV) G was excised from pBS/G (9) using HindIII and XbaI sites and subcloned into pcDNA3.1 (Invitrogen, Grand Island, NY) to create pcDNA3.1/VSV G. To create pCAGGS/VSV G, cDNA encoding the VSV G luminal and transmembrane domains was excised from pcDNA3.1/VSV G using EcoRI and BamHI and cDNA encoding the VSV G cytoplasmic tail was excised from pBS/GCT (9) using BamHI and XhoI; these fragments were then subcloned into pCAGGS-MCS. To create pCAGGS/SARS-CoV M-Gtail, the cDNA sequence encoding M1-R100 of SARS-CoV M was amplified from pcDNA3.1/SARS-CoV M by PCR flanked by KpnI and BamHI sites and the cDNA sequence encoding the cytoplasmic tail of VSV G (200 bp) was excised from pBS/GCT using BamHI and XhoI; these fragments were then subcloned into pCAGGS-MCS using KpnI and XhoI.
HeLa cells and HEK293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen/Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS) (Atlanta Biologicals, Lawrenceville, GA) and 0.1 mg/ml Normocin (InvivoGen, San Diego, CA) at 37°C and 5% CO2.
Fugene6 (Roche, Indianapolis, IN) was used to transfect each cDNA into cells according to the manufacturer's instructions. One day after plating of cells, 50% confluent HeLa or HEK293T cells were transfected with a total of 2 μg of DNA. When SARS-CoV S was coexpressed with SARS-CoV M or any M mutant, there was a decrease in the SARS-CoV S expression level. To counteract this decrease in S expression, we transfected S and M at a 3:1 ratio (1.5 μg pCAGGS/SARS-CoV S and 0.5 μg pCAGGS/SARS-CoV M, M mutant, or empty vector). In cases where SARS-CoV S was expressed alone, the remaining concentration of DNA was composed of empty pCAGGS vector. Fugene6 was first diluted in Opti-MEM (Invitrogen/Gibco, Grand Island, NY) before plasmid DNA was added at a 3:1 Fugene6:DNA ratio according to the manufacturer's instructions. The localization of SARS-CoV S and M was similar in HeLa and HEK293T cells; however, HeLa cells are flatter and easier to image and were thus used for indirect immunofluorescence microscopy, while 293T cells were used for the biochemical experiments because the transfection efficiency was higher. All mutants were tested by indirect immunofluorescence in HEK293T cells and behaved identically as in HeLa cells.
Rabbit anti-SARS-CoV S and rabbit anti-SARS-CoV M polyclonal antibodies were previously described (39). Mouse monoclonal SARS-CoV S antibodies were from Biodefense and Emerging Infections Research (BEI) Resources (Manassas, VA). The hybridoma line for mouse anti-VSV G monoclonal antibody I1 was grown as described previously (35). Rabbit anti-VSV G polyclonal antibody (61) and rabbit anti-golgin-160 antibody (19) were previously described. Mouse anti-GM130 antibody was from BD Biosciences (San Diego, CA). Alexa 488-conjugated donkey anti-mouse IgG was from Invitrogen/Molecular Probes (Eugene, OR). Texas red-conjugated donkey anti-rabbit IgG was from Jackson ImmunoResearch (Westgrove, PA). Horseradish peroxidase (HRP)-conjugated anti-rabbit IgG was from Amersham/GE Healthcare (Piscataway, NJ).
At 24 h posttransfection, HeLa cells were washed in phosphate-buffered saline (PBS) and fixed for 10 min in 3% paraformaldehyde in PBS. Fixative was quenched in PBS containing 10 mM glycine (PBS-gly). Cells were permeabilized for 3 min in 0.5% Triton X-100 in PBS-gly and then stained for 20 min with primary antibodies diluted in 1% bovine serum albumin (BSA) in PBS-gly. The following combinations were used: rabbit anti-SARS-CoV M (1:400) and mouse anti-GM130 (1:200), rabbit anti-SARS-CoV M (1:400) and mouse anti-SARS-CoV S (1:25), affinity-purified rabbit anti-VSV G (1:25) and mouse anti-GM130 (1:200), affinity-purified rabbit anti-VSV G (1:25) and mouse anti-SARS-CoV S (1:25), mouse anti-VSV G hybridoma tissue culture supernatant (undiluted) and rabbit anti-golgin-160 (1:200), and mouse anti-VSV G hybridoma tissue culture supernatant (undiluted) and rabbit anti-SARS-CoV M (1:400). Cells were then stained for 20 min with secondary antibodies in the following combination: Alexa 488-conjugated donkey anti-mouse IgG (1:500) and Texas red-conjugated donkey anti-rabbit IgG (1:400). Cells were washed twice with PBS-gly after each step for a total of 5 min and mounted in glycerol containing 0.1 M N-propylgallate. Images were obtained with a Axioscop microscope (Zeiss, Thornwood, NJ) equipped for epifluorescence using a Sensys charge-coupled-device camera (Photometrics, Tucson, AZ) and IP Lab software (Scanalytics, Vienna, VA).
For biotinylation experiments, HEK293T cells were seeded on poly-l-lysine-treated (1 mg/ml) tissue culture dishes to increase cell adherence. At 24 h posttransfection, HEK293T cells were washed twice in PBS. Cells were incubated for 30 min on ice with 1 mg/ml biotin (EZ-Link Sulfo-NHS-LC-Biotin; Pierce/ThermoScientific, Rockford, IL) in PBS. Cells were washed twice and then incubated for 3 min in PBS containing 50 mM glycine to quench the biotin. Cells were lysed in biotinylation lysis buffer (10 mM HEPES [pH 7.2], 0.2% NP-40, 150 mM NaCl) containing protease inhibitor cocktail (Sigma, St. Louis, MO) at 0°C for 10 min. Lysates were clarified at 16,000 × g for 10 min at 4°C. Ten percent of the lysate was removed for quantifying total S. The remainder of the lysate was added to 75 μl of washed streptavidin agarose resin (Pierce/ThermoScientific, Rockford, IL) to bind biotinylated proteins overnight at 4°C with rotation. Streptavidin beads were washed 3 times in lysis buffer, and biotinylated proteins were eluted at 100°C in 1× sample buffer (50 mM Tris-HCl [pH 6.8], 2% sodium dodecyl sulfate [SDS], 20% glycerol, 0.025% bromophenol blue) and 5% 2-mercaptoethanol. Samples were subjected to 8% SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene difluoride membrane (PVDF) (Millipore, Bedford, MA) for Western blotting. Membranes were blocked for 30 min in 5% nonfat dry milk made in Tris-buffered saline containing Tween (TBST) (150 mM NaCl, 10 mM Tris-HCl [pH 7.4], 0.05% Tween 20) and then washed three times in TBST. Membranes were incubated overnight at 4°C in primary rabbit anti-SARS-CoV S polyclonal antibody diluted 1:5,000 in 3% BSA and 0.02% NaN3 in TBST. Membranes were washed three times in TBST and then incubated for 1 h in HRP-conjugated anti-rabbit IgG diluted 1:10,000 in 5% nonfat dry milk made in TBST. Membranes were washed three times in TBST and then treated with HyGlo chemiluminescence reagent (Denville Scientific, Metuchen, NJ) according to the manufacturer's instructions. Membranes were analyzed using a Versa Doc imaging station (Bio-Rad, Hercules, CA) and quantified using Quantity One software (Bio-Rad).
At 24 h posttransfection, HEK293T cells were pulse-labeled and chased as described previously (39). Briefly, HEK293T cells were starved in methionine- and cysteine-free DMEM for 20 min, labeled for 20 min with 50 μCi of Expre35S35S (Perkin Elmer, Waltham, MA) [35S]methionine-cysteine per dish in methionine- and cysteine-free DMEM, and then chased for various times in DMEM/10% FBS. Cells were washed in PBS and then lysed in detergent solution (50 mM Tris-HCl [pH 8.0], 1% NP-40, 0.4% deoxycholate [DOC], 62.5 mM EDTA) containing protease inhibitor cocktail. Lysates were clarified, SDS was added to a final concentration of 0.2%, and SARS-CoV S proteins were immunoprecipitated with rabbit-anti-SARS-CoV S polyclonal antibodies overnight at 4°C. Immune complexes were collected with washed Staphylococcus aureus Pansorbin cells (Calbiochem, San Diego, CA) and washed 3 times in radioimmunoprecipitation assay (RIPA) buffer (0.1% SDS, 50 mM Tris-HCl [pH 8.0], 1% DOC, 150 mM NaCl, 1% Triton X-100). Samples were eluted in 1% SDS in 50 mM Tris (pH 6.8) at 100°C and digested with 0.1 mU/μl endoglycosidase H (endo H) (New England Biolabs, Beverly, MA) in 150 mM sodium citrate [pH 5.5] overnight at 37°C. Concentrated sample buffer was added to 1×, and samples were subjected to 8% SDS-PAGE. Labeled proteins were visualized by using a Molecular Imager FX phosphorimager (Bio-Rad) and quantified using Quantity One software.
A sequence alignment of M cytoplasmic tail amino acids was generated using MultAlin software (8; http://bioinfo.genotoul.fr/multalin/). The GenBank nucleotide sequence accession numbers of the M proteins aligned are as follows: IBV M, NP_040835.1; MHV M, NP_045301.1; BCV M, AAK29779.2; HCV OC43 M, NP_937953.1; HCV HKU1 M, YP_173241.1; SARS-CoV M, NP_828855.1; HCV NL63 M, YP_003770.1; HCV 229E M, NP_073555.1; FIPV M, YP_239357.1; and TGEV M, ABF72147.1. The full-length M sequences were used; however, only the relevant portion of the alignment is shown (see Fig. Fig.77).
Recombinant C-terminally His-tagged full-length SARS-CoV S expressed with baculovirus and purified from Sf9 insect cells was obtained from BEI Resources. Full-length radiolabeled SARS-CoV M and MY195A were translated in vitro using the TNT quick coupled transcription/translation system (Promega Corporation, Madison, WI). Briefly, pcDNA3.1/SARS-CoV M or pcDNA3.1/SARS-CoV MY195A was incubated with TNT master mix in the presence of Easy Tag [35S]methionine (Perkin Elmer) and canine pancreatic microsomal membranes (Promega Corporation, Madison, WI) for 90 min at 30°C. In vitro-transcribed and -translated proteins were solubilized in in vitro transcription and translation (IVTT) binding buffer (50 mM HEPES [pH 7.1], 100 mM NaCl, 10 mM EDTA, 5 mM MgCl2, 1 mM dithiothreitol [DTT], 0.1% NP-40). Equal amounts of the IVTT reaction mixtures (with SARS-CoV M or MY195A) were added to full-length His-tagged SARS-CoV S tail prebound to Ni-nitrilotriacetic acid (NTA)-agarose or Ni-NTA-agarose alone and incubated with rotation at room temperature for 1 h. Bound proteins were washed in IVTT binding buffer, eluted in sample buffer, and subjected to 15% SDS-PAGE. Gels were stained with Coomassie blue to ensure equal S protein load, and labeled proteins were visualized by using a Molecular Imager FX phosphorimager and quantified using Quantity One software.
We have shown that SARS-CoV S localizes to the plasma membrane when exogenously expressed alone in cells. However, when SARS-CoV S is coexpressed with SARS-CoV M, S is retained intracellularly at the Golgi complex near the virus assembly site (39). This phenomenon has also been demonstrated for the S and M proteins of several other coronaviruses (42, 44). To determine if the ability of M to retain S was specific, we coexpressed SARS-CoV M with the VSV glycoprotein (VSV G). VSV G is a well-studied viral glycoprotein that traffics quickly through the secretory pathway, localizes to the plasma membrane at steady state (38), and has no homology with SARS-CoV M. When coexpressed with VSV G, SARS-CoV M did not retain VSV G intracellularly at the Golgi complex (Fig. (Fig.1B).1B). This suggests that the ability of the SARS-CoV M tail to retain SARS-CoV S is specific.
It has been shown that SARS-CoV S and M can interact in vitro and in vivo (22, 26, 39); however, the specific requirements for S-M interaction have not been determined. Published data for MHV implicate the cytoplasmic tail of each protein in the S and M interaction, although the specific region important for SARS-CoV S-M interaction has not been demonstrated (16, 44). To determine if the region of SARS-CoV M necessary for interaction with S was limited to the cytoplasmic tail, we created a chimeric protein (SARS-CoV M-Gtail) consisting of the SARS-CoV M luminal and transmembrane domains (M1-R100) fused to the cytoplasmic tail of VSV G. The chimeric protein SARS-CoV M-Gtail was localized to the Golgi complex when exogenously expressed in HeLa cells (Fig. (Fig.1C).1C). This was expected since the SARS-CoV M transmembrane domains are important for Golgi localization (unpublished data) (60). However, unlike full-length M, M-Gtail did not retain SARS-CoV S intracellularly at the Golgi complex when the two proteins were coexpressed, suggesting that the SARS-CoV M cytoplasmic tail was necessary to retain SARS-CoV S in the Golgi region of transfected cells (Fig. (Fig.1C).1C). A reciprocal chimera was also made, consisting of the VSV G luminal and transmembrane domains fused to the cytoplasmic tail of SARS-CoV M. The resulting chimera, VSV G-Mtail, was found to be mislocalized with both ER and Golgi staining (data not shown). However, when coexpressed with SARS-CoV S, VSV G-Mtail was able to retain S intracellularly, again indicating that the SARS-CoV M cytoplasmic tail was necessary to retain SARS-CoV S in the Golgi region (data not shown).
To map the interaction between the SARS-CoV S and M proteins, we made deletions of 120, 70, or 30 amino acids in the SARS-CoV M cytoplasmic tail near the C terminus. The M protein lacking 120 amino acids was unstable and was therefore not included in the subsequent analysis. The mutant protein SARS-CoV MΔ1 lacks 30 amino acids (Y177-T207) of the M cytoplasmic tail. The mutant protein SARS-CoV MΔ2 lacks 70 amino acids (L137-T207) of the M cytoplasmic tail (Fig. (Fig.2).2). Our rabbit anti-SARS-CoV M antibody recognizes the extreme 15 C-terminal amino acids, so these residues were not deleted in either mutant.
To determine if SARS-CoV MΔ1 and SARS-CoV MΔ2 were correctly targeted to the Golgi compartment, the mutant proteins were transiently expressed and analyzed by indirect immunofluorescence microscopy. At 24 h posttransfection, cells were fixed, permeabilized, and stained with antibodies to SARS-CoV M and the Golgi marker GM130. Similar to wild-type SARS-CoV M, both MΔ1 and MΔ2 localized to the Golgi region (Fig. (Fig.3A3A).
Since the results of our immunofluorescence experiments suggested that SARS-CoV S and M interact in transfected cells, we used this assay to determine if SARS-CoV MΔ1 and SARS-CoV MΔ2 could interact with and retain SARS-CoV S intracellularly at the Golgi complex. HeLa cells transiently expressing SARS-CoV S with SARS-CoV M, SARS-CoV MΔ1, or SARS-CoV MΔ2 were fixed, permeabilized, and stained with antibodies to SARS-CoV S and SARS-CoV M. As shown earlier, when SARS-CoV S was expressed alone, it was present at the cell surface. Unlike full-length SARS-CoV M, neither SARS-CoV MΔ1 nor SARS-CoV MΔ2 retained SARS-CoV S in Golgi membranes, and S accumulated at the plasma membrane similarly to its accumulation when it was expressed alone (Fig. (Fig.3B3B).
To quantify the steady-state level of SARS-CoV S at the plasma membrane, we performed a cell surface biotinylation assay. We used HEK293T cells for this and other biochemical assays due to their higher transfection efficiency. All mutants were tested by indirect immunofluorescence in HEK293T cells and behaved identically to HeLa cells. HEK293T cells expressing SARS-CoV S with SARS-CoV M, SARS-CoV MΔ1, SARS-CoV MΔ2, or empty vector were biotinylated at 0°C with a non-membrane-permeating biotinylating reagent. Biotinylated S proteins were isolated with streptavidin agarose resin and analyzed by SDS-PAGE and Western blotting with antibodies to SARS-CoV S (Fig. (Fig.3C).3C). The total amount of S protein in the sample was determined by running 10% of the lysate. Total S protein, as well as biotinylated surface S protein, was quantified and normalized to the amount of SARS-CoV S coexpressed with empty vector. When SARS-CoV S was coexpressed with wild-type SARS-CoV M for 24 h, we observed a 60% reduction in the amount of SARS-CoV S present at the plasma membrane (Fig. (Fig.3D).3D). However, coexpression with either SARS-CoV MΔ1 or SARS-CoV MΔ2 did not reduce the amount of SARS-CoV S present on the cell surface (Fig. (Fig.3D3D).
SARS-CoV S is trafficked through the secretory pathway when expressed alone and accumulates at the plasma membrane (4, 20, 39, 52, 54, 56). Since SARS-CoV M but not MΔ1 or MΔ2 retained S intracellularly at the Golgi compartment and reduced the amount of S present at the plasma membrane (Fig. (Fig.3),3), we next analyzed S protein trafficking through the secretory pathway when coexpressed with the M mutants. As a glycoprotein moves through the secretory pathway, its sugars are modified (18); in the medial Golgi complex, N-linked sugars become resistant to digestion with endoglycosidase H (endo H). The rate and extent of glycoprotein trafficking through the secretory pathway en route to the plasma membrane can be inferred from the amount of endo H-resistant protein over time after pulse-labeling. HEK293T cells exogenously expressing SARS-CoV S and SARS-CoV M or empty vector were pulse-labeled with [35S]methionine-cysteine and chased for various times. S protein was immunoprecipitated, denatured, and digested with endo H. When SARS-CoV S was expressed alone, approximately 70% of labeled SARS-CoV S was resistant to digestion with endo H after 40 min of chase. However, when SARS-CoV S was coexpressed with SARS-CoV M, only 25% of SARS-CoV S was resistant to endo H digestion after 40 min of chase, indicating that the majority of labeled S had not reached the medial Golgi compartment (Fig. (Fig.4).4). Coexpression with SARS-CoV MΔ1 or MΔ2 had little effect on the amount of SARS-CoV S carbohydrate processing. This suggests that wild-type M but not MΔ1 or MΔ2 retains S predominantly in a pre-medial Golgi compartment. Given that SARS-CoV MΔ1 and SARS-CoV MΔ2 had similar phenotypes when coexpressed with SARS-CoV S, we concluded that the 30 amino acids near the C terminus were important for S and M interaction, since this region was missing from both SARS-CoV MΔ1 and MΔ2.
To further map the SARS-CoV S and M interaction, we made smaller deletions in and surrounding the region deleted from SARS-CoV MΔ1 (Fig. (Fig.2).2). SARS-CoV MΔ1a, SARS-CoV MΔ1b, and SARS-CoV MΔ1c were missing amino acids L205-D214, Y195-K204, or R185-A194, respectively. When exogenously expressed for 24 h in HeLa cells, SARS-CoV MΔ1a, MΔ1b, and MΔ1c all localized to the Golgi region (Fig. (Fig.5A).5A). However, when coexpressed with SARS-CoV S, only SARS-CoV MΔ1a retained S intracellularly at the Golgi region (Fig. (Fig.5B).5B). When SARS-CoV MΔ1b and MΔ1c were coexpressed with SARS-CoV S, S was present on the plasma membrane at steady state (Fig. (Fig.5B5B).
To quantify the amount of SARS-CoV S at the plasma membrane, we performed a cell surface biotinylation assay as described above. Coexpression with SARS-CoV MΔ1a reduced the amount of SARS-CoV S present at the plasma membrane by approximately 60%, similar to coexpression with wild-type SARS-CoV M (Fig. (Fig.5D5D and and3D).3D). However, coexpression with SARS-CoV MΔ1b or SARS-CoV MΔ1c did not reduce the amount of SARS-CoV S present at the cell surface.
To determine if SARS-CoV MΔ1a, MΔ1b, or MΔ1c could reduce the extent of SARS-CoV S carbohydrate processing, we performed a pulse-chase endo H-trafficking assay as described above. After 40 min of chase, approximately 70% of radiolabeled SARS-CoV S had moved to the medial Golgi compartment and was resistant to endo H digestion. When SARS-CoV S was coexpressed with SARS-CoV MΔ1a, only 25% was resistant to digestion with endo H after 40 min of chase (Fig. (Fig.6),6), which was similar to the result when S was coexpressed with wild-type M (Fig. (Fig.4).4). This suggests that MΔ1a was able to retain S in a pre-medial Golgi compartment. However, when SARS-CoV MΔ1b or MΔ1c was coexpressed with SARS-CoV S, the extent of S carbohydrate processing was not reduced (Fig. (Fig.6)6) and was similar to the amount of carbohydrate processing when S was coexpressed with SARS-CoV MΔ1 (Fig. (Fig.4).4). Thus, neither MΔ1b nor MΔ1c was able to retain S in a pre-medial Golgi compartment. When SARS-CoV S was coexpressed with SARS-CoV MΔ1, SARS-CoV MΔ1b, or SARS-CoV MΔ1c, the amount of SARS-CoV S that was resistant to endo H digestion after 40 min of chase was less than the amount of endo H-resistant S after 40 min of chase when S was expressed alone (Fig. (Fig.4).4). Thus, it is possible that some residual interaction of S with the mutant M proteins still exists. The implications of this observation will be discussed below.
Since neither SARS-CoV MΔ1b nor MΔ1c retained SARS-CoV S intracellularly at the Golgi and they do not contain overlapping deleted amino acids, we concluded that the S interaction site was near the shared border between the two deleted regions. To test this, we created another deletion mutant, SARS-CoV MΔbc, which is missing 5 amino acids from both MΔ1b and MΔ1c (S191-R199), spanning the MΔ1b/MΔ1c junction. SARS-CoV MΔbc was unable to retain SARS-CoV S intracellularly as assayed by immunofluorescence microscopy, surface biotinylation, or pulse-chase (data not shown), and it behaved similarly to MΔ1, MΔ2, MΔ1b, and MΔ1c. Sequence alignment of various coronavirus M protein cytoplasmic tails revealed the presence of several conserved residues spanning the SARS-CoV MΔ1b and MΔ1c junction (Fig. (Fig.7).7). To determine if the conserved amino acids were important for SARS-CoV S and M interaction, we mutated them individually to alanine or valine (MS190A, MG191A, MF192A, MA193V, MY195A, MR197A, MY198A, or MY203) (data not shown). Only one mutant, SARS-CoV MY195A, was unable to retain SARS-CoV S intracellularly. When analyzed by indirect immunofluorescence microscopy, SARS-CoV MY195A was localized to the Golgi region but did not retain SARS-CoV S intracellularly (Fig. (Fig.8A).8A). Unlike wild-type M, coexpression of SARS-CoV MY195A did not reduce the amount of SARS-CoV S present on the plasma membrane as measured by surface biotinylation (Fig. (Fig.8B).8B). In addition, SARS-CoV MY195A behaved similarly to MΔ1 when SARS-CoV S carbohydrate processing was measured by endo H resistance after pulse-chase (Fig. (Fig.8C).8C). Thus, we conclude that disruption of Y195 was responsible for the inability of SARS-CoV MΔ1, MΔ2, MΔ1b, and MΔ1c to interact with SARS-CoV S. Although Y195 is present in MΔ1c, the context in which it is presented is probably changed, since it is the first amino acid after the deletion.
Our previously published results suggest that SARS-CoV S and M interact directly (39). To determine if the Y195A mutation disrupts SARS-CoV S and M interaction in vitro, we performed a binding assay using in vitro-transcribed and -translated SARS-CoV M or MY195A incubated with full-length His-tagged SARS-CoV S. The solubilized in vitro-transcribed and -translated M proteins were incubated with His-tagged SARS-CoV S prebound to nickel beads or with nickel beads alone. Samples were analyzed by SDS-PAGE and phosphorimaging. In vitro-transcribed and -translated SARS-CoV M bound specifically to His-tagged SARS-CoV S (Fig. (Fig.9).9). However, when Y195 was mutated to alanine, the amount of SARS-CoV MY195A binding to His-tagged SARS-CoV S was reduced to background levels (Fig. (Fig.9).9). We were unable to detect in vitro binding when only the soluble, recombinant tail of each protein was examined, suggesting that transmembrane regions of each may be important for orienting the tails (data not shown).
Taken together, our results suggest that the SARS-CoV S and M interaction is mediated by the cytoplasmic tail of SARS-CoV M and that a single residue, Y195, is necessary for efficient S and M interaction.
Unlike many well-studied enveloped viruses, coronaviruses assemble intracellularly at the ERGIC (reviewed in reference 21). Thus, the proteins in the virion envelope must be targeted to the ERGIC for assembly. As for most other enveloped viruses, lateral interactions between coronavirus envelope proteins are important for the proper assembly of infectious virus. The coronavirus M protein was originally thought to determine the location of virus assembly, since it can interact with the other viral structural proteins. However, since the M proteins of several coronaviruses localize past the budding site at steady state (29), it is thought that other factors, such as lateral interactions with other structural proteins, influence the location of the assembly site.
Previously, we determined that a dibasic ER retrieval signal found in the cytoplasmic tail of SARS-CoV S promoted interaction with the SARS-CoV M protein, presumably by increasing the possibilities for interaction during cycling (39). Here, we identified a critical tyrosine residue in the SARS-CoV M cytoplasmic tail that is necessary for efficient interaction with SARS-CoV S. When SARS-CoV S is expressed alone, it localizes to the plasma membrane, past the budding compartment. However, when SARS-CoV S is coexpressed with SARS-CoV M (39), M mediates intracellular retention of S in a pre-medial Golgi compartment near the virus assembly site. When coexpressed with SARS-CoV S, SARS-CoV M also reduced the amount of S present at the plasma membrane. Importantly, VSV G was not retained intracellularly by SARS-CoV M. This was expected, since coronavirus lateral interactions exclude host proteins from the assembling virion (12). Additionally, a chimeric protein consisting of the cytoplasmic tail of VSV G fused to the luminal and transmembrane domains of SARS-CoV M, M-Gtail, did not retain SARS-CoV S intracellularly. However, the reciprocal chimera consisting of the cytoplasmic tail of SARS-CoV M fused to the luminal and transmembrane domains of VSV G was able to retain SARS-CoV S intracellularly. These results suggest that the cytoplasmic tail of SARS-CoV M is necessary to mediate the SARS-CoV S-M interaction. This lateral S-M interaction is presumably how S is localized to the budding site and incorporated into newly assembled virions in infected cells.
We found that a critical tyrosine residue, Y195, in the cytoplasmic tail of SARS-CoV M was important for retaining S at the Golgi complex. When Y195 was mutated to alanine, the mutant protein, SARS-CoV MY195A, did not retain S in the Golgi region when the two proteins were coexpressed. The expression of SARS-CoV MY195A did not reduce the amount of SARS-CoV S present at the plasma membrane or reduce S carbohydrate processing due to retention in a pre-medial Golgi compartment. Also, MY195A was unable to interact with SARS-CoV S in vitro. Importantly, MY195A was properly localized to the Golgi complex (Fig. (Fig.8A)8A) and had a half-life similar to that of wild-type SARS-CoV M (data not shown). The importance of SARS-CoV MY195 in S-M interaction was confirmed by using an in vitro binding assay. We do not believe that phosphorylation of Y195 is the basis for its importance in S-M interaction, since we have found no evidence that M is phosphorylated (unpublished data). Other tyrosine-based posttranslation modifications, such as tyrosine nitration, are possible (27), but more work needs to be done to implicate or rule out tyrosine modifications as the basis for SARS-CoV S and M interaction. It is also possible that the critical tyrosine residue could be structurally important for S-M interaction. A well-known tyrosine-based protein-protein interaction motif is YXXΦ, where X is any amino acid and Φ is a bulky, hydrophobic residue (5). This tyrosine-based motif is recognized by the clathrin adaptor proteins and mediates transmembrane protein endocytosis from the plasma membrane (5). The structure of the adaptor protein subunit responsible for binding the YXXΦ motif has been solved and highlights the importance of the tyrosine hydroxyl in hydrogen bonding, as well as its insertion in a hydrophobic binding pocket (46).
Although Y195 clearly plays a role in S-M interaction, it is unlikely to be the only region of SARS-CoV M that is important for interaction with S. When SARS-CoV S was expressed alone, 75% of S was resistant to endo H digestion after 40 min of chase. However, when MY195A was coexpressed with S, 60% of S was resistant to endo H digestion after 40 min (Fig. (Fig.44 and and8C).8C). Although the ability of S and MY195A to interact was clearly and significantly reduced, S was not trafficked to the same extent as when it was expressed alone. Thus, it is likely that there is a region outside of the regions tested in this study that is also important for S-M interaction. Extra information could be either in the region of the tail not deleted in MΔ2 or in the luminal or transmembrane domains. In support of this idea, Voss et al. recently implicated the N-terminal 134 amino acids of SARS-CoV M, which includes the transmembrane domains and a small portion of the cytoplasmic tail, in S-M interaction (60). The N-terminal 134 amino acids of M were able to retain S at the Golgi compartment by indirect immunofluorescence microscopy. The ability to interact with S was only measured by qualitative microscopy in that study. However, we did not observe Golgi retention of S with M containing residues M1-R100 (M-Gtail, Fig. Fig.1C).1C). Since the SARS-CoV M cytoplasmic tail alone was able to retain S intracellularly (Fig. (Fig.1),1), the role of the transmembrane domains in S-M interaction remains unclear, and they may function primarily to present the cytoplasmic tail in the proper orientation.
The most extensive information on coronavirus S-M interaction comes from studies of MHV. The key regions for MHV S-M interaction have been mapped on both the S and M proteins (7, 11). Deletion of 70 amino acids in the amphipathic region of the MHV M cytoplasmic tail (residues 121 to 195) disrupts S-M interaction, as determined by indirect immunofluorescence microscopy and coimmunoprecipitation (11). Specific detergent combinations are needed to preserve MHV S-M interactions for coimmunoprecipitation (44). Here, we have used more sensitive assays to characterize the SARS-CoV S and M interaction. By using cell surface biotinylation and pulse-chase assays, we examined both steady-state localization and trafficking of SARS-CoV S to examine S-M interaction without the complication of maintaining S-M complexes in detergent lysates. Assays using virus-like particle (VLP) formation to assess contributions to virus assembly are often used to characterize coronavirus protein-protein interaction, although the exact requirements for VLP formation are controversial (3, 6, 10, 16, 26, 40, 44, 55, 59). However, in our hands, SARS-CoV S and M proteins were released in membrane vesicles from HEK293T and HeLa cells when expressed individually (data not shown), so we could not use this assay to study assembly.
We identified a different region of SARS-CoV M as involved in S interaction than was identified for MHV (11). Coronavirus M proteins are similar in size, ranging from approximately 220 to 260 amino acids. When the SARS-CoV M and MHV M protein sequences are aligned, Y195 is approximately 10 amino acids downstream in comparison to the 70 amino acids deleted in MHV studies. It is likely that the distance of the tyrosine from the membrane affects interaction with S. The coronavirus M protein cytoplasmic tail is considerably longer than that of the S cytoplasmic tail, but a large region of the M tail was shown to be tightly folded (2, 49, 50). Nevertheless, deletion of 70 amino acids from the MHV M tail could have altered the orientation and/or distance of a motif outside the 70-amino-acid stretch required for interaction with S. Importantly, the point mutations used here to dissect SARS-CoV S-M interaction are unlikely to disrupt overall features of the SARS-CoV M cytoplasmic tail.
Interestingly, in addition to the 70 amino acids deleted, a cytoplasmic tyrosine residue, Y211, was shown to be important for MHV S and M interaction (11). Although S-M interaction was not fully abolished when Y211 was mutated to glycine in MHV M, MHV S-M interaction was severely reduced (11). MHV M Y211 is downstream in comparison to SARS-CoV M Y195; however, when the analogous tyrosine in SARS-CoV M, Y205, was mutated to alanine, it had no affect on S-M interaction as determined by the assays used here (data not shown). Although MHV M Y211 is outside of the 70 amino acids implicated in MHV S-M interaction, it clearly plays a role. This suggests that the spacing of the tail from the membrane may be important for S-M interaction and virus assembly.
Coronavirus S proteins form trimers prior to export from the ER, while M proteins form larger oligomers in the Golgi (13, 37). It is thought that coronavirus S-M interaction occurs in a pre-Golgi compartment, since endo H-sensitive proteins can be isolated by coimmunoprecipitation (42). It is possible that one S trimer could interact with multiple M oligomers, which could overcome a weak affinity through avidity. Indeed, if there are 16 to 25 M proteins for each S trimer, as estimated in recent cryoelectron microscopy (cryo-EM) studies of the SARS-CoV virion (41), only a fraction of M would be interacting with S. Although cryo-EM analyses of SARS-CoV and MHV virions did not detect a difference in the amounts of M proteins next to S spikes, it is not clear if a difference in packing would be detectable at the resolution achieved in these studies. In addition, the C-terminal region of M interacts with the nucleocapsid protein N (14). Since the interaction site on M for S is also near the C terminus, it is possible that only M that is not associated with S is available for interaction with the viral nucleocapsid during assembly. Interestingly, the analogous region of the SARS-CoV M tail that was predicted to be tightly folded in MHV lacks the interaction sites for both S and N. This region may also be tightly folded in SARS-CoV M and could function to regulate the spacing of the C-terminal region from the membrane.
In conclusion, our results suggest that the SARS-CoV M cytoplasmic tail specifically retains the SARS-CoV S protein near the virus assembly site, and Y195 in the M cytoplasmic tail appears to be important for this interaction. Ongoing work is directed toward determining the region of SARS-CoV S that is recognized by this region of the M tail.
This work was supported by National Institutes of Health grant R21 A1072312 and a Ford Foundation Predoctoral Diversity Fellowship.
The following reagents were obtained through the NIH Biodefense and Emerging Infections Research Resources Repository, NIAID, NIH: C-terminally histidine-tagged SARS-CoV spike (S) protein NR-686, a recombinant expressed from baculovirus, and monoclonal antibody anti-SARS-CoV S protein (similar to 341C) NR-617.
We thank Jie Li for constructing pCAGGS/VSV G and pCAGGS/SARS-CoV M-Gtail. We also thank members of the Machamer laboratory for helpful discussions and comments on the manuscript.
Published ahead of print on 9 December 2009.