Coronavirus structural proteins and ER stress

The accumulation of viral proteins at the ER during an infection can induce ER stress as outlined before via three main pathways, PERK, ATF6, and IRE1, which are commonly refereed to as the unfolded Protein Response (UPR) or ER stress response. This response results in the transcriptional activation of genes encoding for ER chaperones, increasing autophagy, or regulating apoptosis.

Coronaviruses are no exception. Both structural and non-structural proteins localise to the ER - followed by the formation of Replication Transcription Centers in the case of non-structural proteins or the co-localisation in ER-Golgi Intermediate Compartments (ERGIC), the sites of viral genome replication and assembly respectively.

Prototype Coronavirus particle

                                                    S (Spike) protein

The Coronavirus S protein plays a pivotal role in the entry of the Coronavirus particle, by mediating receptor binding, membrane fusion as well as eliciting the formation of neutralising antibodies. newly synthesised S protein co-localises with the other structural proteins in the ERGIC where viral assembly takes place. Although the viral assembly site is distinct from the ER, newly synthesised S protein localises to the ER where the protein is slowly folded and glycosylated prior release into the cytoplasm of the cell where it is retained in the ERGIC by a dilysine signal (in the case of IBV S) or a dibasic signal (KXHXX) (SARS-CoV S, TGEV S); interestingly, the S proteins of BCV and MHV-A59 both do not contain similar localisation signals. These signals not only localise S to the ERGIC, but also retain S protein in the cytoplasm instead of being localised to the plasma membrane.

Since the S is glycosylated and localises to the ER, the viral S protein modulates the ER stress response. As it the case with the viral E, HCOV-Nl63 orf3, and SARS-CoV orf3a proteins, S activates PERK, but not ATF6 and IRE1. Similar to these proteins, the expression of S results in PERK and phosphorylated eIF2α induced transcriptional activation of intraluminal ER chaperones GRP94/78, thus facilitating the correct folding and processing of viral proteins inside the ER. Indeed, the expression of S does not induce CHOP, nor ATF4 or splicing of XBP1, although in SARS-CoV slightly elevated levels of sXBP1 can be detected (independent of S). If the latter is required for viral replication since sXBP1 has also chaperone activity or not is not clear.

                                              M (Membrane protein)

The Coronavirus M is the main structural component of the virion, its function associated not only with incorporating the nascent nucleocapsid via interaction with the viral N protein into the viral particle but also recruiting other viral  structural components to the site of viral assembly. Although the precise topology of the M protein is still disputed, the consensus is that the three putative transmembrane domains predominantly localise in the Golgi, as well as in trafficking vesicles, with the N and C termini are either in a N_exo//C_exo  or N_exo/C_endo orientation (exo referring to ectodomain, endo to endodomain). In the case of SARS-CoV M, the protein is N-Glycosylated on the N-terminal ectodomain prior to insertion into the ERGIC at the ER, with the glycosylation site facing the ER lumen and thus favoring a N_exo-C_endo orientation. Consequently, the highly immunogenic C terminus is located inside the mature virus particle with the less immunogenic N terminus protruding from the viral surface. This is different from Transmissible Gastroenteritis Virus (TGEV) and the feline Coronavirus (FCoV),  where the M protein adopts not only a N_exo-C_endo orientation but also a N_exo - C_exo configuration with the N- and C-terminus both protruding the viral particle and thus exposing the C-terminus; indeed, TGEV and FCoV infection induces the formation of highly specific antibodies targeting the viral M protein. The N_exo-C_endo orientation not only prevents the formation of antibodies against M but also is also required for the interaction with the N protein and subsequent recruitment of the viral positive sense ssRNA to sites of viral assembly. Indeed, in TGEV virions we find both conformations of M, with the N_exo-C_endo conformation being predominant. 

Membrane topology of M proteins derived from SARS-CoV and TGEV

Although the N-glycosylation of SARS-CoV M and the O-glycosylation of MHV M are dispensable for viral assembly -glycosylation mutants do recruit both the viral S and N proteins to the ERGIC-, budding, and infectivity, the N-glycosylation of M might be required to induce antiviral signaling, specifically the induction of Interferon-α and -β.

Since the Coronaviral M protein is localised to the ER prior to be transported to the ERGIC and glycosylated by ER resident enzymes, it might be conceivable to assume that the expression of M induces the ER stress response by lipid depletion. So far, however the induction of the UPR has not been reported.

                                              E (Envelope)

In addition to the M protein, the coronaviral E protein is also involved in the formation of viral and virus-like particles and as such localises to the ERGIC, the viral assembly site located in close proximity to the ER and Golgi. Indeed if transfected into cells,  the SARS-CoV and PEDV E protein E can be detected both in the ER and the Golgi, although not in a reticulate pattern but in membrane clusters. In contrast to the viral M or the SARS-CoV orf3a protein, both the N- and C-Termini of SARS-CoV E are located on the cytoplasmic side with no lumenal domain, although tin the case for IBV E a different topology has been reported. In the case of SARS-CoV E protein, this transmembrane helical hairpin domain might not only be responsible for the membrane curvature of the viral particle but is also palmitoylated. In contrast to the M protein, the E protein from both SARS-CoV and PEDV have been shown to interfere with the ER stress response, albeit the results are contradictory. In the case of SARS-CoV, the expression of the E protein inhibits the IRE1 pathway but PERK or ATF6 mediated signalling and thus inhibits apoptosis induced by IRE1 activation in cells infected with a recombinant SARS-CoVΔE virus or RSV, as well as in cells treated with thapsigargin or tunicamycin.  On the other hand, both the expression of SARS-CoV and PDEV E activates PERK and PERK mediated expression of pro inflammatory cytokines such as Interleukin-8 by activating NF-κB as well as inducing the expression of Bcl2 and thus preventing apoptosis. Indeed cells infected with rSARS-CoVΔE not only have an increased stress response but also exhibit a decreased inflammatory response - yet rSARS-CoVΔE is attenuated.

Membrane topology of Coronavirus E proteins derived from SARS-CoV, MHV, and IBV

                          SARS-CoV orf3a/HCoV-NL63 orf3

The SARS-CoV open reading frame (orf) 3a protein is encoded by one of the so called “group specific genes” and has no known structural nor sequence homology to any of the known proteins of other coronaviruses, although it bears some similarities to the Coronavirus M protein. Similar to the M protein, orf3a has three transmembrane domains with the same topology as SARS-CoV M, it localizes predominately to the Golgi and ERGIC, both are structural proteins, both are glycosylated, and both interact with the viral S and E proteins. In contrast to M, the three transmembrane domains have been postulated to form an ion channel with the domain 2 and 3 forming the pore via a cysteine rich domain (AA 81-60) and the central region (AA 125-200) be required for binding the 5’-UTR of the SARS-CoV genome. Although both orf3a and M are glycosylated, orf3a is O-glycosylated postranslationally similar to MHV M whereas SARS-CoV M is N-Glycosylated cotranslationally.

The C-terminus contains both the Yxx𝛟 (with x representing any AA and 𝚽 is a hydrophobic amino acid residue) and the diacidic motifs (ExD, where x represents any AA). Whilst the Yxx𝛟 domain is required for the localisation of the viral particle to the endosome during viral entry, the diacidic motif is required for the export from the ER to the ERGIC/Golgi as well as preventing the retrograde Rab6GTPase dependent transport from the Golgi to the ER, thus facilitating the accumulation of orf3a at the plasma membrane. As mentioned in a previous post, the expression of orf3a has been associated with the induction of the ER stress response through the activation of the PERK pathway but does not trigger the activation of IRE1 or ATF6 nor the Endoplasmic Reticulum Associated Degradation (ERAD) pathway. Constitutive activation of PERK by orf3a induces apoptosis through the expression of ATF4 and CHOP as described before as well as activating p38 MAPK mediated release of mitochondrial Cytochrome c.

SARS-CoV orf3a and HCoV-NL63 orf3

In a similar way to SARS-CoV orf3a, the HCoV-NL63 orf 3 protein co-localises with the viral S, M, E, and N proteins within the ERGIC compartment in Huh7 cells transfected with the respective expression plasmids. In addition, HCoV-NL63 is N-glycosylated at the ER luminal side. If the expression of HCoV-NL63 induces the ER stress response however is not known, but the author of these lines would predict that akin to SARS-CoV orf 3a, HCoV-NL63 orf3 induces the PERK pathway.

                                        N (Nucleocapsid)

In contrast to the aforementioned proteins, the Nucleocapsid protein does not integrate into the membrane of either the virion, the ER, or the ERGIC due to the absence of a transmembrane domain but binds both the M and E protein (in addition to nsp3) and  in cells infected with CoV or co-transfected with either M and E respectively, although N is not required for the formation of virus like particles. Mass spectrometric characterisation of the SARS-CoV revealed 12 potential glycosylation sites were identified, suggesting that N is N-glycosylated within the ER co- or postranslationally. Based on these findings is seems possible that the expression of N induces a ER stress response although the N protein itself does not predominantly localise to the ER but both within the cytoplasm and as well as the nucleolus in infected cells and cells transfected with IBV, MHV, SARS-CoV, and TGEV N. An exception might be however the PEDV (Porcine Epidemic Diarrhoea Virus) N protein, which seems to localise to the ER (according to the authors at least) -I should note that I however  profoundly disagree with the authors of the paper in question about since from the published figures seems to be absent from the region labeled with ER tracker. In any case, the published data suggest that PEDV N might activate NF-κB signalling in a similar way to HCoV-OC43 via binding to microRNA9 instead. In my opinion, it is the glycosylation of N, which might induce the ER stress pathway instead as indicated by unregulated expression of GRP78, a ER chaperone - a hypothesis that could and should have been tested by mutating the putative glycosylation sites. In infected cells, the viral nsp-3 protein binds N and thus localises N to the replication complex, as does binding of N to non-glycosylated M. In these cases however, it is not N itself which would cause ER stress, but the respective binding partner, although it might be possible that N excaberates ER stress. The SARS-CoV N protein itself however localises to stress granules upon arsenite treatment which can be inhibited by phosphorylation of a Serine residue via SR protein kinase 1in the C-terminal domain, maybe similar to PEDV N. Indeed, mutating this Serine residue to Alanine stabilises the formation of N containing stress granules. So in the end, it is not clear if the glycosylation of N induces the ER stress response  but there is strong evidence that the C-terminal domain is involved in inducing the ER stress response via the formation of protein aggregates in HeLa cells treated with arsenite.

Interestingly, the author of this post did some research in the past where preliminary data suggested that B23 remains nucleolar in the presence of IBV N in contrast to a nucleolar and perinuclear distribution in mock-transfected cells. If this is the case indeed, this could explain how IBV N might interfere with the ER stress response.

B23 in Vero cells transfected with backbone vector (top) or IBV-N (bottom)

So what is final conclusion? The expression of coronaviral structural proteins certainly interferes with the ER stress, but the general pattern is, that this interference is limited to the PERK pathway and does not involve the ATF6 or IRE1 response, thus preventing CHOP mediated apoptosis. If genes related to autophagy are induced remains to be seen; if so, then this might provide a mechanism to evade apoptosis.

Activation of PERK during Coronavirus infection by the structural proteins

Further reading

Fung TS, & Liu DX (2014). Coronavirus infection, ER stress, apoptosis and innate immunity. Frontiers in microbiology, 5 PMID: 24987391

Liu DX, Fung TS, Chong KK, Shukla A, & Hilgenfeld R (2014). Accessory proteins of SARS-CoV and other coronaviruses. Antiviral research, 109C, 97-109 PMID: 24995382

Nal B, Chan C, Kien F, Siu L, Tse J, Chu K, Kam J, Staropoli I, Crescenzo-Chaigne B, Escriou N, van der Werf S, Yuen KY, & Altmeyer R (2005). Differential maturation and subcellular localization of severe acute respiratory syndrome coronavirus surface proteins S, M and E. The Journal of general virology, 86 (Pt 5), 1423-34 PMID: 15831954

Lontok E, Corse E, & Machamer CE (2004). Intracellular targeting signals contribute to localization of coronavirus spike proteins near the virus assembly site. Journal of virology, 78 (11), 5913-22 PMID: 15140989

Chan CP, Siu KL, Chin KT, Yuen KY, Zheng B, & Jin DY (2006). Modulation of the unfolded protein response by the severe acute respiratory syndrome coronavirus spike protein. Journal of virology, 80 (18), 9279-87 PMID: 16940539

Versteeg GA, van de Nes PS, Bredenbeek PJ, & Spaan WJ (2007). The coronavirus spike protein induces endoplasmic reticulum stress and upregulation of intracellular chemokine mRNA concentrations. Journal of virology, 81 (20), 10981-90 PMID: 17670839

Chang YJ, Liu CY, Chiang BL, Chao YC, & Chen CC (2004). Induction of IL-8 release in lung cells via activator protein-1 by recombinant baculovirus displaying severe acute respiratory syndrome-coronavirus spike proteins: identification of two functional regions. Journal of immunology (Baltimore, Md. : 1950), 173 (12), 7602-14 PMID: 15585888

de Haan CA, de Wit M, Kuo L, Montalto-Morrison C, Haagmans BL, Weiss SR, Masters PS, & Rottier PJ (2003). The glycosylation status of the murine hepatitis coronavirus M protein affects the interferogenic capacity of the virus in vitro and its ability to replicate in the liver but not the brain. Virology, 312 (2), 395-406 PMID: 12919744

 Risco C, Antón IM, Suñé C, Pedregosa AM, Martín-Alonso JM, Parra F, Carrascosa JL, & Enjuanes L (1995). Membrane protein molecules of transmissible gastroenteritis coronavirus also expose the carboxy-terminal region on the external surface of the virion. Journal of virology, 69 (9), 5269-77 PMID: 7636969

Voss D, Pfefferle S, Drosten C, Stevermann L, Traggiai E, Lanzavecchia A, & Becker S (2009). Studies on membrane topology, N-glycosylation and functionality of SARS-CoV membrane protein. Virology journal, 6 PMID: 19534833

Escors D, Camafeita E, Ortego J, Laude H, & Enjuanes L (2001). Organization of two transmissible gastroenteritis coronavirus membrane protein topologies within the virion and core. Journal of virology, 75 (24), 12228-40 PMID: 11711614

Oostra M, de Haan CA, de Groot RJ, & Rottier PJ (2006). Glycosylation of the severe acute respiratory syndrome coronavirus triple-spanning membrane proteins 3a and M. Journal of virology, 80 (5), 2326-36 PMID: 16474139

Padhan K, Minakshi R, Towheed MA, & Jameel S (2008). Severe acute respiratory syndrome coronavirus 3a protein activates the mitochondrial death pathway through p38 MAP kinase activation. The Journal of general virology, 89 (Pt 8), 1960-9 PMID: 18632968

Minakshi R, Padhan K, Rani M, Khan N, Ahmad F, & Jameel S (2009). The SARS Coronavirus 3a protein causes endoplasmic reticulum stress and induces ligand-independent downregulation of the type 1 interferon receptor. PloS one, 4 (12) PMID: 20020050

 Fossati M, Colombo SF, & Borgese N (2014). A positive signal prevents secretory membrane cargo from recycling between the Golgi and the ER. The EMBO journal PMID: 25063674

Nishimura N, & Balch WE (1997). A di-acidic signal required for selective export from the endoplasmic reticulum. Science (New York, N.Y.), 277 (5325), 556-8 PMID: 9228004

Müller MA, van der Hoek L, Voss D, Bader O, Lehmann D, Schulz AR, Kallies S, Suliman T, Fielding BC, Drosten C, & Niedrig M (2010). Human coronavirus NL63 open reading frame 3 encodes a virion-incorporated N-glycosylated membrane protein. Virology journal, 7 PMID: 20078868

Yang Y, Zhang L, Geng H, Deng Y, Huang B, Guo Y, Zhao Z, & Tan W (2013). The structural and accessory proteins M, ORF 4a, ORF 4b, and ORF 5 of Middle East respiratory syndrome coronavirus (MERS-CoV) are potent interferon antagonists. Protein & cell, 4 (12), 951-61 PMID: 24318862

Yuan Q, Liao Y, Torres J, Tam JP, & Liu DX (2006). Biochemical evidence for the presence of mixed membrane topologies of the severe acute respiratory syndrome coronavirus envelope protein expressed in mammalian cells. FEBS letters, 580 (13), 3192-200 PMID: 16684538

Chen SC, Lo SY, Ma HC, & Li HC (2009). Expression and membrane integration of SARS-CoV E protein and its interaction with M protein. Virus genes, 38 (3), 365-71 PMID: 19322648

Arbely E, Khattari Z, Brotons G, Akkawi M, Salditt T, & Arkin IT (2004). A highly unusual palindromic transmembrane helical hairpin formed by SARS coronavirus E protein. Journal of molecular biology, 341 (3), 769-79 PMID: 15288785

DeDiego ML, Nieto-Torres JL, Jiménez-Guardeño JM, Regla-Nava JA, Alvarez E, Oliveros JC, Zhao J, Fett C, Perlman S, & Enjuanes L (2011). Severe acute respiratory syndrome coronavirus envelope protein regulates cell stress response and apoptosis. PLoS pathogens, 7 (10) PMID: 22028656

Xu X, Zhang H, Zhang Q, Dong J, Liang Y, Huang Y, Liu HJ, & Tong D (2013). Porcine epidemic diarrhea virus E protein causes endoplasmic reticulum stress and up-regulates interleukin-8 expression. Virology journal, 10 PMID: 23332027

Ruch TR, & Machamer CE (2012). A single polar residue and distinct membrane topologies impact the function of the infectious bronchitis coronavirus E protein. PLoS pathogens, 8 (5) PMID: 22570613 

Hurst KR, Koetzner CA, & Masters PS (2013). Characterization of a critical interaction between the coronavirus nucleocapsid protein and nonstructural protein 3 of the viral replicase-transcriptase complex. Journal of virology, 87 (16), 9159-72 PMID: 23760243 

Xu X, Zhang H, Zhang Q, Huang Y, Dong J, Liang Y, Liu HJ, & Tong D (2013). Porcine epidemic diarrhea virus N protein prolongs S-phase cell cycle, induces endoplasmic reticulum stress, and up-regulates interleukin-8 expression. Veterinary microbiology, 164 (3-4), 212-21 PMID: 23562137

Wurm T, Chen H, Hodgson T, Britton P, Brooks G, & Hiscox JA (2001). Localization to the nucleolus is a common feature of coronavirus nucleoproteins, and the protein may disrupt host cell division. Journal of virology, 75 (19), 9345-56 PMID: 11533198 

Lai FW, Stephenson KB, Mahony J, & Lichty BD (2014). Human coronavirus OC43 nucleocapsid protein binds microRNA 9 and potentiates NF-κB activation. Journal of virology, 88 (1), 54-65 PMID: 24109243

Peng TY, Lee KR, & Tarn WY (2008). Phosphorylation of the arginine/serine dipeptide-rich motif of the severe acute respiratory syndrome coronavirus nucleocapsid protein modulates its multimerization, translation inhibitory activity and cellular localization. The FEBS journal, 275 (16), 4152-63 PMID: 18631359