key: cord-0904851-5fh477c4
authors: Le, Tra M.; Wong, Hui H.; Tay, Felicia P. L.; Fang, Shouguo; Keng, Choong‐Tat; Tan, Yee J.; Liu, Ding X.
title: Expression, post‐translational modification and biochemical characterization of proteins encoded by subgenomic mRNA8 of the severe acute respiratory syndrome coronavirus
date: 2007-07-20
journal: FEBS J
DOI: 10.1111/j.1742-4658.2007.05947.x
sha: 7e90b235cee44affd8f1f18e0178075196e78438
doc_id: 904851
cord_uid: 5fh477c4

The most striking difference between the subgenomic mRNA8 of severe acute respiratory syndrome coronavirus isolated from human and some animal species is the deletion of 29 nucleotides, resulting in splitting of a single ORF (ORF8) into two ORFs (ORF8a and ORF8b). ORF8a and ORF8b are predicted to encode two small proteins, 8a and 8b, and ORF8 a single protein, 8ab (a fusion form of 8a and 8b). To understand the functions of these proteins, we cloned cDNA fragments covering these ORFs into expression plasmids, and expressed the constructs in both in vitro and in vivo systems. Expression of a construct containing ORF8a and ORF8b generated only a single protein, 8a; no 8b protein expression was obtained. Expression of a construct containing ORF8 generated the 8ab fusion protein. Site‐directed mutagenesis and enzymatic treatment revealed that protein 8ab is modified by N‐linked glycosylation on the N81 residue and by ubiquitination. In the absence of the 8a region, protein 8b undergoes rapid degradation by proteasomes, and addition of proteasome inhibitors inhibits the degradation of protein 8b as well as the protein 8b‐induced rapid degradation of the severe acute respiratory syndrome coronavirus E protein. Glycosylation could also stabilize protein 8ab. More interestingly, the two proteins could bind to monoubiquitin and polyubiquitin, suggesting the potential involvement of these proteins in the pathogenesis of severe acute respiratory syndrome coronavirus.

The most striking difference between the subgenomic mRNA8 of severe acute respiratory syndrome coronavirus isolated from human and some animal species is the deletion of 29 nucleotides, resulting in splitting of a single ORF (ORF8) into two ORFs (ORF8a and ORF8b). ORF8a and ORF8b are predicted to encode two small proteins, 8a and 8b, and ORF8 a single protein, 8ab (a fusion form of 8a and 8b). To understand the functions of these proteins, we cloned cDNA fragments covering these ORFs into expression plasmids, and expressed the constructs in both in vitro and in vivo systems. Expression of a construct containing ORF8a and ORF8b generated only a single protein, 8a; no 8b protein expression was obtained. Expression of a construct containing ORF8 generated the 8ab fusion protein. Site-directed mutagenesis and enzymatic treatment revealed that protein 8ab is modified by N-linked glycosylation on the N81 residue and by ubiquitination. In the absence of the 8a region, protein 8b undergoes rapid degradation by proteasomes, and addition of proteasome inhibitors inhibits the degradation of protein 8b as well as the protein 8b-induced rapid degradation of the severe acute respiratory syndrome coronavirus E protein.

Glycosylation could also stabilize protein 8ab. More interestingly, the two proteins could bind to monoubiquitin and polyubiquitin, suggesting the potential involvement of these proteins in the pathogenesis of severe acute respiratory syndrome coronavirus. accessory proteins are dispensable for viral replication in cultured cells [6] , their exact roles in the pathogenesis and virulence of SARS-CoV in its natural hosts are yet to be established.

Two such accessory proteins, proteins 8a and 8b, are encoded by two overlapping ORFs (8a ⁄ b) present in subgenomic mRNA8 (Fig. 1 ). Subgenomic mRNA8 from most SARS-CoV isolates obtained from animals and patients at early stages of the SARS epidemic contains a single ORF8 with potential to encode a single protein, 8ab [7] . A 29 nucleotide deletion between T27867 and A27868 for strain SG2774 (accession number AY283798) was found in strains collected from human isolates at late stages of the outbreak, resulting in the splitting of the single ORF8 into two overlapping ORFs, ORF8a and ORF8b ( Fig. 1 ) [1, 8] . The deletion of the 29 nucleotides and the formation of two overlapping ORFs were predicted to be the consequence of adaptation of SARS-CoV from animals to humans [7, 9] . Experimental infection of civets with SARS-CoV isolates BJ01 (without the 29 nucleotides) and GZ01 (with the 29 nucleotides) showed that animals infected with strain BJ01 tend to have a higher average temperature and slightly stronger antibody responses [10] . Two recent studies have shown that the putative protein 8b can downregulate the SARS-CoV E protein in virus-infected cells and induce DNA synthesis in cells expressing the protein [11, 12] . However, more detailed characterization of the expression, biochemical properties and functions of the products encoded by these two mRNA isoforms has yet to be done.

In this study, ORF8a was expressed in vitro in both wheat germ extracts and rabbit reticulocyte lysates as a 5.3 kDa protein. Expression of the protein was also observed in Cos-7 cells, as a Flag-tagged protein. Similarly, translation of ORF8ab was observed in all the translation systems used. In contrast, protein 8b was expressed only when ORF8b was separately cloned. When constructs containing the two overlapping ORFs ORF8a ⁄ ORF8b and the single ORF8ab, respectively, were expressed, expression of proteins 8a and 8ab were observed. Protein 8ab was shown to be an N-linked glycosylated protein, and the glycosylation site was identified to be the N81 residue. Proteins 8b and 8ab could be modified by ubiquitination, and in the absence of the 8a region, protein 8b undergoes rapid degradation by proteasome. Addition of the proteasome inhibitors inhibits the degradation of protein 8b as well as the protein 8b-induced rapid degradation of the SARS-CoV E protein. In addition, glycosylation could also stabilize protein 8ab. Furthermore, proteins 8b and 8ab could bind covalently and noncovalently to monoubiquitin and polyubiquitin. As no homology with any known ubiquitin-binding domains (UBDs) was found, they may represent a novel group of ubiquitin-binding proteins.

Cloning, expression and post-translational modification of proteins encoded by the SARS-CoV mRNA8

In some animal and early human isolates, the subgenomic mRNA8 of SARS-CoV was predicted to encode a single ORF (ORF8). Owing to the deletion of 29 nucleotides [between T27867 and A27868 for strain SG2774 (accession number AY283798)], two overlapping ORFs (ORF8a ⁄ ORF8b) were found in most human isolates (Fig. 1 ). ORF8a and ORF8b are predicted to encode two small proteins, 8a and 8b, whereas ORF8 encodes a single protein, 8ab, representing a fused form of proteins 8a and 8b. To understand the functions of these proteins, cDNA fragments covering these ORFs were cloned into pFlag, giving rise to five constructs either with or without a Flag-tag at the N-terminus (Fig. 1) . These constructs were then expressed in both in vitro expression systems and in intact cells. When constructs p8a ⁄ b, pF-8a ⁄ b, pF-8b and pF-8ab were expressed in TnT transcription-coupled translation wheat germ extracts in the presence of [ 35 S]methionine, single protein bands of approximately 5.3, 6.5, 10.2 and 14.4 kDa, representing untagged protein 8a, Flag-tagged protein 8a, Flag-tagged protein 8b and Flag-tagged protein 8ab, respectively, were detected ( Fig. 2A, lanes 1-4) . Expression of the same four constructs in rabbit reticulocyte lysates in the presence of [ 35 S]methionine gave rise to the same four products (Fig. 2B, lanes 1-4) . In addition, a series of bands (ladder bands) with increases of approximately 10 kDa were detected when pF-8b and pF-8ab were expressed in the system (Fig. 2B, lanes 3 and 4) . The patterns of these bands suggest that they may represent ubiquitinated forms of proteins 8b and 8ab. To determine whether these bands are related to the 8b region, immunoprecipitation was carried out using rabbit polyclonal antibodies to protein 8b. As shown in Fig. 2B , the 10.2 kDa protein 8b and the 14.4 kDa protein 8ab, together with their corresponding ladder bands, were precipitated with the antibodies (lanes 7 and 8). The fact that these bands were efficiently immunoprecipitated by the antibodies to protein 8b suggests that proteins 8b and 8ab may be modified by ubiquitination.

Expression of these constructs was then carried out in HeLa cells using the vaccinia ⁄ T7 expression system. As shown in Fig. 2C , Flag-tagged proteins 8a, 8b and 8ab were detected in cells transfected with the corresponding constructs (lanes 2-4). Interestingly, a prominent band (8ab*), migrating more slowly than the 14.4 kDa protein 8ab, was detected in cells expressing pF-8ab (Fig. 2C, lane 4) . This may represent a posttranslationally modified protein 8ab; characterization of this modification was conducted, and the results are presented in a later section. In addition, multiple species of more slowly migrating bands were also detected from cells expressing F-8b and F-8ab (Fig. 2C , lanes 3 and 4).

To test further whether proteins 8b and 8ab are posttranslationally modified by ubiquitination, coexpression of these proteins with a Myc-tagged ubiquitin in HeLa cells was carried out, and the expression and interaction of the proteins were determined by western blot and coimmunoprecipitation assays. As shown in Fig. 3A , western blot analysis using antibody to Myc resulted in the detection of monoubiquitin and . These results strongly suggest that proteins 8b and 8ab are modified by ubiquitination.

In a similar way, western blot analysis using antibody to Flag resulted in the detection of Flag-tagged protein 8b (Fig. 3B , lane 2) and protein 8ab (Fig. 3B , lanes 3 and 4), either expressed alone or coexpressed with the Myc-tagged ubiquitin. The same cell lysates were immunoprecipitated with antibody to Myc and subsequently analyzed by western blot with antibody to Flag. In addition to the detection of several more slowly migrating bands, the Flag-tagged protein 8b (lane 6) and protein 8ab (lane 7) were efficiently detected when they were coexpressed with the Myctagged ubiquitin. Protein 8ab was not detected when it was expressed on its own (Fig. 3B, lane 8) . The reason for preferential precipitation of the nonubiquitinated forms of proteins 8b and 8ab by antibody to Myc (Fig. 3B, lanes 9-12) is not clear. However, it may indicate that the unmodified proteins 8b and 8ab might have higher binding affinity to the monoubiquitin and polyubiquitin molecules than their ubiquitinated counterparts.

The coimmunoprecipitation results shown above demonstrated that proteins 8b and 8ab could interact with monoubiquitin and polyubiquitin. To further confirm these results, proteins 8ab and 8b were expressed in Escherichia coli as glutathione S-transferase (GST) fusion proteins and purified. Relatively pure GST, GST-8b and GST-8ab were obtained (Fig. 3C,  lanes 1-3) . Equal amounts of GST, GST-8b and GST-8ab bound to GST beads were then used in a pull-down assay by mixing them with [ 35 S]methioninelabeled lysates prepared from cells overexpressing the Myc-tagged ubiquitin. Efficient pull-down of monoubiquitin and polyubiquitin by GST-8b and GST-8ab was observed (Fig. 3C, lanes 6 and 7) . However, no monoubiquitin or polyubiquitin bands were detected when GST protein was used in the same experiment ( Fig. 3C, lane 5) . These results confirm that proteins 8b and 8ab can efficiently bind to the monoubiquitin and polyubiquitin molecules.

To study the effect of ubiquitin binding and ubiquitination on the degradation and accumulation of proteins 8b and 8ab, a system that mimics the expression level of the proteins in SARS-CoV-infected cells was developed on the basis of an infectious clone system derived from the coronavirus infectious bronchitis virus (IBV) [13] . As shown in Fig. 4A , the 8b and 8ab genes were cloned into the IBV genome between the N gene and the 3¢-UTR under the control of a transcriptional regulatory sequence for the subgenomic mRNA5 of IBV. Electroporation of the in vitro-synthesized full-length RNAs containing the 8b and 8ab genes, respectively, into Vero cells resulted in the recovery of infectious viruses, which were stable during propagation in cells (data not shown).

Northern blot analysis was then carried out to determine whether an additional mRNA species deriving from the inserted 8b and 8ab genes was produced in cells infected with the recombinant viruses. The results showed that additional mRNA species with sizes corresponding to the inserted 8b and 8ab genes, respectively, were produced in cells infected with the rIBV ⁄ 8b and rIBV ⁄ 8ab viruses (Fig. 4B, lanes 2 and  3) . Analysis of cells infected with the recombinant viruses by western blot with antibodies to protein 8b resulted in efficient detection of proteins 8ab and 8ab* in cells infected with the rIBV ⁄ 8ab virus (Fig. 4C , lane 3). However, no protein 8b was detected in cells infected with the rIBV ⁄ 8b virus (Fig. 4C , lane 2). Considering that similar amounts of mRNAs were detected in cells infected with the two recombinant viruses, and that the two proteins could bind to ubiquitin and be modified by ubiquitination, the failure to detect 8b expression at the protein level in this system suggests that protein 8b may undergo rapid degradation via the proteasome pathway.

To confirm this possibility, high concentrations of two proteasome inhibitors, lactacystin (8 lm) and trileucine vinyl sulfone (NLVS; 32 lm), were added to cells infected with rIBV ⁄ 8b virus at a multiplicity of infection of approximately 1. Western blot analysis with antibodies to IBV N protein resulted in efficient detection of the N protein from cells infected with the virus in the presence of dimethylsulfoxide (Fig. 4D,  upper panel, lane 1) . In rIBV ⁄ 8b-infected cells treated with NLVS, approximately the same amounts of N protein were detected (Fig. 4D, upper panel, lane 3) . However, in cells treated with lactacystin, much less N protein was detected (Fig. 4D, upper panel, lane 2) , suggesting that this inhibitor may have some inhibitory effect on viral replication. Analysis of the same cell lysates by western blot with antibodies to protein 8b resulted in no detection of the protein in cells treated with dimethylsulfoxide alone (Fig. 4D, lower panel,  lane 1) . However, in rIBV ⁄ 8b-infected cells treated with the two inhibitors, protein 8b was clearly detectable (Fig. 4D, lower panel, lanes 2 and 3) , confirming that the two proteasome inhibitors can indeed stabilize protein 8b and increase the accumulation of the protein in infected cells.

In a recent report, protein 8b was shown to mediate rapid degradation of the SARS-CoV E protein [11] . To determine whether the two proteasome inhibitors could also inhibit the protein 8b-mediated rapid degradation of the SARS-CoV E protein, coexpression of protein 8b with E protein in the presence or absence of the inhibitors was carried out. Analysis of the E protein expression by western blot with antibodies to E protein showed that addition of the two inhibitors blocked the rapid degradation of the SARS-CoV E protein (Fig. 5, top panel) . Similar results were obtained when the transfected cells were incubated in the presence of another proteasome inhibitor, MG132 (50 lm) (Fig. 5, second panel) . However, addition of 50 lm MG132 did not significantly affect the accumulation of protein 8b when the protein was overexpressed in cells (Fig. 5, third panel) .

As mentioned, detection of protein 8ab* and its migration properties on SDS ⁄ PAGE suggested that the protein may undergo some post-translational modification in addition to ubiquitination. ORF8b was predicted to encode a potential N-linked glycosylation site on asparagine 43 (N43, equivalent to N81 in protein 8ab). To test the possibility that proteins 8b and 8ab may be modified by N-linked glycosylation, total cell lysates prepared from cells expressing the Flag-tagged proteins 8b and 8ab were treated with peptide N-glycosidase F (PNG-aseF). The protein 8ab* band disappeared after treatment with the enzyme (Fig. 6A, lanes 2 and 4) , but the treatment did not affect the migration of protein 8b (Fig. 6A, lanes 1 and 3) . This suggests that protein 8b is not modified by N-linked glycosylation. During the course of this study, it was also noted that variable amounts of the N-linked glycosylated protein 8ab* were detected when pF-8ab was expressed in different cells.

As can be seen in Fig. 6A , similar amounts of the modified and unmodified protein 8ab were detected when the construct was expressed in Cos-7 cells (lanes 5 and 6). In HeLa cells, however, much more glycosylated protein 8ab was detected under the same expression and detection conditions (Fig. 6A, lanes 5 and 6) .

To further determine whether N81 is the site for the N-linked glycosylation of protein 8ab, mutation of the residue to an aspartic acid was carried out. Expression of the mutant construct resulted in detection of the 14.4 kDa protein 8ab but not the more slowly migrating glycosylated form (Fig. 6B, lanes 1 and 2) , confirming that N81 is the site for N-linked glycosylation of protein 8ab. In addition to the 14.4 kDa protein 8ab, several more rapidly migrating bands were observed in cells expressing pF-8abM (Fig. 6B, lane 1) . Western blot analysis using antibody to Flag resulted in the detection of two bands of approximately 10 and 5 kDa (Fig. 6B , lane 1). They represent the N-terminal portions of protein 8ab (Fig. 6B, upper diagram) . Western blot analysis of the same lysates using polyclonal antibodies to protein 8b resulted in the detection of two bands of approximately 7 and 3 kDa (Fig. 6B,  lane 3) , representing the C-terminal portions of protein 8ab (Fig. 6B, upper diagram) . The detection of multiple degraded products in cells expressing the glycosylation-defective mutant protein 8ab suggests that, like protein 8b, the unglycosylated protein 8ab is unstable.

Further confirming that the unglycosylated protein 8ab is less stable than the glycosylated protein 8ab*, addition of high concentrations of the two proteasome inhibitors lactacystin (8 lm) and NLVS (32 lm) to cells expressing F-8ab showed that the two inhibitors significantly enhanced the accumulation of the unglycosylated protein 8ab (Fig. 6C ). However, a much lower enhancement effect was observed for accumulation of the glycosylated form (Fig. 6C ).

The subgenomic mRNA8 of SARS-CoV isolated from humans at late stages of the epidemic contains a deletion of 29 nucleotides, resulting in splitting of a single ORF, ORF8, into ORF8a and ORF8b [7] . In this study, ORF8a was expressed in vitro in both wheat germ extracts and rabbit reticulocyte lysates as a 5.3 kDa protein. Owing to the lack of a specific antibody, the protein was not detected in cells. In the genomic RNA of SARS-CoV, ORF8a is located immediately downstream of a strong body transcriptional regulatory sequence (AGUCUAAACGAAAUG) [14] , and is the 5¢-most ORF of the subgenomic mRNA.

With the recent detection of proteins 3a and 7a from subgenomic mRNA3 and mRNA7, respectively, the majority of the 5¢-most ORFs of genomic and subgenomic mRNAs of SARS-CoV are known to encode proteins [13, 15] . In patients infected with SARS-CoV strains carrying the 29 nucleotide deletion, antibodies against the 8a peptide were present, confirming the expression of the protein [16, 17] . ORF8ab was expressed as a 14.4 kDa Flag-tagged protein in this study. Although expression of ORF8ab in virusinfected cells and production of antibodies against this protein in patients are yet to be established, protein 8ab was likely to be expressed by a mechanism similar to ORF8a, as it is the only ORF in the subgenomic mRNA8 encoded by SARS-CoV strains without the 29 nucleotide deletion. ORF8b was expressed as a 

However, translation of the protein was not detected from expression of the overlapping ORF8a ⁄ ORF8b. This ORF was previously speculated to be translationally silent [14] , and no antibodies against the protein have been found in patients [16] . Although ORF8b was located downstream of a weak transcriptional regulatory sequence (CUAAUAA ACUCAUG) [1] , no additional subgenomic mRNA starting from this position was identified. If protein 8b was synthesized from ORF8a ⁄ ORF8b, it would be unlikely to be translated by the conventional ribosomal scanning mechanism. Protein 8ab was found to be post-translationally modified by N-linked glycosylation on the N81 residue. Although the same residue is present in protein 8b, glycosylation of protein 8b was not detected. Glycosylation plays a fundamental role in the oligomerization, sorting, transport and folding of proteins [18] . During glycosylation, glycan residues are transferred to Asn of the target sequence motif Asn-X-Ser ⁄ Thr on a nascent peptide by oligosaccharyltransferase in eukaryotic cells [19] . The glycan is the 'admission ticket' for proteins to undergo folding or oligomerization, sorting and transport [18] . Substitution mutation of the N81 to an Asp in protein 8ab made the protein more susceptible to degradation. In the absence of the 8a region, protein 8b was also rapidly degraded, as the protein was not glycosylated, although it contains the potential target sequence motif for glycosylation. It is therefore speculated that glycosylation may assist in folding of protein 8b. In general, the folding process happens in the endoplasmic reticulum (ER) and the Golgi apparatus [20, 21] . In the ER, the calnexin-calreticulin cycle is involved in the folding process of glycosylated proteins [22] . In this process, calnexin and calreticulin bind transiently to glycosylated proteins, and subsequently form a complex with cochaperone ERp57 [23] . The three-protein complex folds glycosylated proteins by forming the appropriate disulfide bonds [24] . Protein 8ab is rich in cysteine residues in the 8a region, and may attain the correct conformation by forming disulfide bonds in this region.

More intriguingly, proteins 8b and 8ab could covalently and noncovalently bind to monoubiquitin and polyubiquitin, and facilitate the binding of the two proteins to cellular ubiquitinated proteins. The noncovalent binding of proteins 8b and 8ab to ubiquitin suggests that the two proteins may contain a ubiquitin-binding domain (UBD). UBDs are able to bind to ubiquitin noncovalently and covalently [25] . So far, nine UBDs, such as the ubiquitin-binding motif and ubiquitin-associated domain, have been identified [26] [27] [28] . Many UBDs were initially discovered by homol-ogy search with the identified UBDs [25, 29, 30] . However, proteins 8b and 8ab do not share homology with any known UBD, suggesting that they might contain a novel UBD. As proteins 8b and 8ab but not protein 8a are conjugated with polyubiquitin and form complexes with monoubiquitin and polyubiquitin, the novel UBD is probably located in the 8b region. Ubiquitin conjugation of UBDs could facilitate the formation of complexes between a UBD and free mono ⁄ polyubiquitin. It is generally believed that UBDs are able to bind to other ubiquitinated proteins [25] , disrupting the function or transportation of polyubiquitinated proteins [27, 28] . As ubiquitination and degradation of cellular proteins by proteasomes may play an important role in maintaining a constant level of these proteins in cells [30, 31] , a wider effect of proteins 8b and 8ab on cellular ubiquitinated proteins would be expected.

Another intriguing and essential issue is whether splitting of the single ORF8 into ORF8a and ORF8b by the deletion of 29 nucleotides is evolutionarily advantageous. The following considerations and clues provided by this study would suggest that it is beneficial to the virus. First, as splitting of the single ORF8 results in the formation of a bicistronic mRNA species, expression of the downstream ORF8b would be regulated by various translational control mechanisms. Evidence obtained from in vitro and in vivo expression of constructs containing the two separate ORFs suggests that the downstream ORF is either silent or expressed at a low level. It is unclear at the moment whether this ORF is expressed in virus-infected cells. One recent study found positive staining in some of the SARS-CoV-infected cells with polyclonal antibodies to protein 8b, suggesting that the protein may be expressed [11] . Second, the data presented in this study show that the 8b region may contain multiple functional domains, such as for ubiquitin binding, ubiquitination and glycosylation, involved in the regulation of the stability, host-virus interaction and biochemical properties of protein 8ab. However, no functional domain has so far identified in the 8a region of the protein. Although protein 8a would certainly be expressed from the bicistronic mRNA8, the absence of the 8b region would result in the loss of the functions assigned to the 8b region of protein 8ab. Finally, as ubiquitination and the binding to mono ⁄ polyubiqutins, as well as ubiquitinated cellular proteins, may regulate cellular signaling pathways involved in the host antiviral response, silencing the expression of protein 8ab may constitute a viral strategy to evade the host antiviral defense mechanisms. More systematic studies of the functions of these proteins are required to address this issue further.

Plasmid DNA was transcribed and translated in vitro by using the TNT T7 coupled wheat germ extracts or reticulocyte lysates (Promega, Madison, WI, USA) in a 50 lL reaction mixture at 30°C for 1.5 h. [ 35 S]Methionine was used to label the synthesized proteins. 

The expression of GST, GST-8b and GST-8ab was induced in E. coli BL21 cells with 1 mm isopropyl thio-b-d-galactoside for 4 h. Cells were washed three times with cold NaCl ⁄ P i , and broken by sonication. Pellets from cells expressing GST-8b and GST-8ab, respectively, were lysed with 8 m urea, and purified with gluthionine-Sepharose beads (Amersham, Little Chalfont, UK) after dialysis.

Radiolabeled lysates prepared from cells overexpressing ubiquitin were precleared by mixing with GST-prebound beads for 2 h at 4°C to remove nonspecifically binding proteins. Beads prebound with GST, GST-8b, and GST-8ab, respectively, were mixed with the precleared [ 35 S]methionine-labeled or unlabeled cell lysates at 4°C for 4 h, washed three times with RIPA buffer (50 mm Tris ⁄ HCl, pH 7.5, 150 mm NaCl, 0.5% sodium deoxycholate, 0.5% Nonidet P-40, 0.05% SDS), and resolved by SDS ⁄ PAGE. The labeled samples were visualized by autoradiography, and the unlabeled samples were analyzed by western blot with antibody to Myc.

Cells transfected or cotransfected with plasmids were washed once with NaCl ⁄ P i , harvested, and lysed with 500 lL of RIPA buffer in the presence of proteinase inhibitors (Roche Diagnostics AG, Rotkreuz, Switzerland). Cell debris was removed by centrifugation at 15 000 g at 4°C for 30 min using an Eppendorf 5415-R, and the supernatants were immunoprecipitated with 1 lg of appropriate antibodies at room temperature for 2 h. The mixtures were incubated with 50 lL of protein A agarose beads for 2 h, and washed three times with RIPA buffer. The immunoprecipitated proteins were separated on SDS ⁄ PAGE and analyzed by western blot using appropriate antibodies.

Polypeptides were separated by SDS ⁄ PAGE, and transferred to poly(vinylidene difluoride) membrane (Bio-Rad, Hercules, CA, USA). The membrane was blocked with 10% nonfat dry milk for 1 h, and this was followed by 1 h of incubation with an appropriate primary antibody. The membrane was then washed five times with PBST (NaCl ⁄ P i containing 1% Tween-20), and this was followed by incubation with an appropriate horseradish peroxidase-conjugated secondary antibody (DAKO) at room temperature for 1 h; detection was with an ECL plus system (Amersham Biosciences, Piscataway, NJ, USA). When necessary, the membrane can be reincubated with another antibody after stripping off the signals with the stripping buffer (2% SDS, 100 mm b-mercaptoethanol, 62.5 mm Tris ⁄ HCl, pH 6.8) at 55°C for 30 min, and washed extensively with PBST.

Vero cells were infected with rIBV, rIBV ⁄ 8b and rIBV ⁄ 8ab, respectively, at a multiplicity of approximately one plaqueforming unit per cell. Total RNA was extracted from cells infected with wild-type or mutant viruses. Ten micrograms of RNA was added to a mixture of 1 · Mops, 37% formaldehyde and formamide, and incubated at 65°C for 20 min before being subjected to gel electrophoresis. The segregated RNA bands were transferred onto a Hybond N+ membrane (Amersham Biosciences) via capillary action overnight, and fixed by UV crosslinking (Stratalinker tm 2400, Stratagene, La Jolla, CA, USA). Hybridization of Dig-labeled DNA probes was carried out at 50°C in hybridization oven overnight. Membranes were washed three times for 15 min each with the probe buffer, before proceeding to detection with CDP-Star (Roche) according to the manufacturer's instructions.

Construction of a full-length cDNA clone for IBV, insertion of the 8b and 8ab genes into the clones, and rescue of recombinant viruses Construction of a full-length cDNA clone for IBV was carried out essentially as described previously [13, 32] . Briefly, five fragments spanning the entire IBV genome were obtained by RT-PCR from Vero cells infected with the Vero cell-adapted IBV p65. The PCR products were purified from agarose gels and cloned into pCR-XL-TOPO (Invitrogen, Carlsbad, CA, USA) or pGEM-T Easy (Promega) vectors. Subsequently, fragment A was removed from pCR-XL-TOPO by digestion with NheI and EcoRI, and subcloned into the pKT0 vector. Plasmids were digested with either BsmBI (fragment A) or BsaI (fragments B, C, D and E). The digested plasmids were separated on 0.8% agarose gels containing crystal violet. Bands corresponding to each of the fragments were cut from the gels and purified with a QIAquick gel extraction kit (Qiagen). Fragments A and B, and fragments C, D and E, were first ligated with T4 DNA ligase at 4°C overnight. The two reaction mixtures were then mixed and further ligated at 4°C overnight. The final ligation products were extracted with phenol ⁄ chloroform ⁄ isoamyl alcohol (25 : 24 : 1), precipitated with ethanol, and detected by electrophoresis on 0.4% agarose gels.

Full-length transcripts were generated in vitro using the mMessage mMachine T7 kit (Ambion, Austin, TX). The N transcripts were generated by using a linearized pKT0-IBVN containing the IBV N gene and the 3¢-UTR region as templates. The in vitro-synthesized full-length and N transcripts were treated with DNaseI and purified with phenol ⁄ chloroform. Vero cells were grown to 90% confluence, trypsinized, washed twice with cold NaCl ⁄ P i , and resuspended in NaCl ⁄ P i . RNA transcripts were added to 400 lL of Vero cell suspension in an electroporation cuvette, and electroporated with one pulse at 450 V and 50 lF with a Bio-Rad Gene Pulser II electroporator. The transfected Vero cells were cultured overnight in 1% fetal bovine serum-containing MEM in a 60 mm dish or a six-well plate, and further cultured in MEM without fetal bovine serum.

The 8b and 8ab genes were inserted into the IBV genome between the N gene and the 3¢-UTR under the control of the TRS for the subgenomic mRNA5 of IBV.

Total lysates (30 lL) prepared from Cos-7 cells expressing protein 8ab were directly treated with 4 lL of PGNaseF (New England Biolabs, Beverly, MA, USA) at 37°C for 2 h. The samples were then separated on SDS ⁄ PAGE and analyzed by western blot with antibody to Flag.

Plasmid p8a ⁄ b was constructed by cloning a PCR fragment covering the SARS-CoV sequence from nucleotides 27 763 to 28 102. The PCR fragment was digested with EcoRV and XbaI, and cloned into EcoRV-and XbaI-digested pKT0. The sequences of the two primers used were 5¢-GCG AATTCAACATGAAATTCTTCTCAT-3¢ and 5¢-GCTCT AGATTAATTTGTTCGTTTATT-3¢. Plasmid pF-8a ⁄ b was created by cloning an EcoRV-XbaI-digested PCR fragment into EcoRV-and XbaI-digested pFlag. The two primers used were 5¢-GGGATATCCATGAAACTTCTC AT-3¢ and the same reverse primer as used for the construction of p8a ⁄ b. Plasmid pF-8a was made by cloning a PCR fragment covering the ORF8a region into pFlag.

Plasmid pF-8ab was cloned by adding the 29 nucleotide insertion into pF-8a ⁄ b using two rounds of PCR. The two primers used to introduce the 29 nucleotide insertion were 5¢-ATTCAGGTTGGTAACCAGTAGGACAAGGATCTT CAA-3¢ and 5¢-TACCAACCTGAATGGAATATAAGGT ACAACACTAGG-3¢.

All constructs were confirmed by automated nucleotide sequencing, and diagrams of these plasmids are shown in Fig. 1A .

The genome sequence of the SARS-associated coronavirus

Severe acute respiratory syndrome coronavirus 3a protein is a viral structural protein

Antibody responses to individual proteins of SARS coronavirus and their neutralization activities

Overexpression of 7a, a protein specifically encoded by the severe acute respiratory syndrome coronavirus, induces apoptosis via a caspasedependent pathway

Characterization of the 3a protein of SARS-associated coronavirus in infected Vero E6 cells and SARS patients

Severe acute respiratory syndrome coronavirus group-specific open reading frames encode nonessential functions for replication in cell cultures and mice

Isolation and characterization of viruses related to the SARS coronavirus from animals in southern China

Clinical progression and viral load in a community outbreak of coronavirus-associated SARS pneumonia: a prospective study

Bats are natural reservoirs of SARS-like coronaviruses

Civets are equally susceptible to experimental infection by two different severe acute respiratory syndrome coronavirus isolates

The human severe acute respiratory syndrome coronavirus (SARS-CoV) 8b protein is distinct from its counterpart in animal SARS-CoV and down-regulates the expression of the envelope protein in infected cells

Expression and functional characterization of the putative protein 8b of the severe acute respiratory syndrome-associated coronavirus

An arginine-to-proline mutation in a domain with undefined function within the RNA helicase protein (NSP13) is lethal to the coronavirus infectious bronchitis virus in cultured cells

Unique and conserved features of genome and proteome of SARS-coronavirus, an early split-off from the coronavirus group 2 lineage

Identification of a novel protein 3a from severe acute respiratory syndrome coronavirus

SARS corona virus peptides recognized by antibodies in the sera of convalescent cases

Profile of antibody responses against SARS-coronavirus recombinant proteins and their potential use as diagnostic markers

Intracellular functions of N-linked glycans

Structural requirements of N-glycosylation of proteins. Studies with proline peptides as conformational probes

Role of N-linked oligosaccharide recognition, glucose trimming, and calnexin in glycoprotein folding and quality control

Association of folding intermediates of glycoproteins with calnexin during protein maturation

Cellular functions of endoplasmic reticulum chaperones calreticulin, calnexin, and ERp57

Identification and characterization of structural domains of human ERp57: association with calreticulin requires several domains

The eS-sence of -SH in the ER

Ubiquitin-binding domains

A ubiquitin-interacting motif conserved in components of the proteasomal and lysosomal protein degradation systems

GAT (GGA Tom1) domain responsible for ubiquitin binding and ubiquitination

Eap45 in mammalian ESCRT-II. binds ubiquitin via a phosphoinositide-interacting GLUE domain

Ubiquitin-mediated sequestration of normal cellular proteins into polyglutamine aggregates

The p53 pathway

The NF-jB and IjB proteins. New discoveries and insights

Amino acid residues critical for RNA-binding in the N-terminal domain of the nucleocapsid protein Are essential determinants for the replication and infectivity of coronavirus in cultured cells

This work was supported by the Agency for Science Technology and Research, Singapore.