key: cord-0847252-a0uop959 authors: Lee, Yun-Bin; Jung, Minkyo; Kim, Jeesoo; Kang, Myeong-Gyun; Kwak, Chulhwan; Kim, Jong-Seo; Mun, Ji-Young; Rhee, Hyun-Woo title: Endomembrane systems are reorganized by ORF3a and Membrane (M) of SARS-CoV-2 date: 2021-06-01 journal: bioRxiv DOI: 10.1101/2021.06.01.446555 sha: 60309c8682f7019588792d0ed093068bb75c5c89 doc_id: 847252 cord_uid: a0uop959 The endomembrane reticulum (ER) is largely reorganized by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). SARS-CoV-2 ORF3a and membrane (M) protein expression affects ER-derived structures including cubic membrane and double membrane vesicles in coronavirus-infected cells; however, the molecular mechanisms underlying ER remodeling remain unclear. We introduced a “plug and playable” proximity labeling tool (TurboID-GBP) for interactome mapping of GFP-tagged SARS-CoV-2 ORF3a and M proteins. Through mass spectrometric identification of the biotinylated lysine residue (K+226 Da) on the viral proteins using Spot-TurboID workflow, 117 and 191 proteins were robustly determined as ORF3a and M interactomes, respectively, and many, including RNF5 (E3 ubiquitin ligase), overlap with the mitochondrial-associated membrane (MAM) proteome. RNF5 expression was correlated to ORF3a ubiquitination. MAM formation and secreted proteome profiles were largely affected by ORF3a expression. Thus, SARS-CoV-2 may utilize MAM as a viral assembly site, suggesting novel anti-viral treatment strategies for blocking viral replication in host cells. Highlights SARS-CoV-2 proteins ORF3a and M alter endoplasmic reticulum proteome profile ORF3a affects mitochondrial-associated membrane formation SARS-CoV-2 may utilize mitochondrial-associated membrane as viral assembly site ORF3a and M interactome proteins may serve as targets for COVID-19 treatment eTOC Blurb ER remodelling by SARS-CoV-2 ORF3a and M protein The endomembrane reticulum (ER) is largely reorganized by severe acute respiratory syndrome coronavirus 24 2 (SARS-CoV-2). SARS-CoV-2 ORF3a and membrane (M) protein expression affects ER-derived structures 25 including cubic membrane and double membrane vesicles in coronavirus-infected cells; however, the 26 molecular mechanisms underlying ER remodeling remain unclear. We introduced a "plug and playable" to possess pro-apoptotic activity (Ren et al., 2020) . Similarly, M of SARS-CoV-2 shows 96.4% sequence 154 similarity with M of SARS-CoV (Gordon et al., 2020) , which was reported to target the ER, ER-Golgi 155 intermediate compartment, and Golgi apparatus, and is associated with apoptosis (Chan et al., 2007) and 156 nuclear factor-(NF) κB signaling (Fang et al., 2007) . We hypothesized that these proteins may perturb the 157 ultrastructure of endomembrane systems, as it has been shown that many viral proteins target the 158 endomembrane systems of host cells (Inoue and Tsai, 2013) . 159 To test this hypothesis, we prepared ORF3a-APEX2 and M-APEX2 constructs, with APEX2 tagged to the 160 C-terminus of the vPOIs (Figure 3A) , and used them for EM imaging in HeLa cells following APEX-mediated Similarly, ER was clearly disrupted in M-expressing cells ( Figure 3D ) and appeared to curl into whorl 175 patterns (also termed as aggresomes). These patterns were also observed in herpes simplex virus-infected cells 176 (Nii et al., 1968 ) and ER stress-induced cells (Schuck et al., 2014; Snapp et al., 2003) . In CLEM imaging of 177 M-GFP (Figure S5D) , we observed the formation of multilamellar bodies (MLB) at the ER membrane ( Figure 178 3D, F) and electron-dense autophagic vesicles (green arrows in Figure S5D ). The biogenesis of MLB 179 regulated is by autophagy (Hariri et al., 2000) and MLBs are utilized during extracellular vesicle-type viral 180 8 egress (Bunyavirus) (Sanz-Sánchez and Risco, 2013) . In flavivirus replication (such as Zika virus), autophagy 181 is usually activated and autophagic vesicles accumulate inside the cells in in cellulo (Hamel et al., 2015) and 182 in vivo (Cao et al., 2017) experiments. In our study, these structures were not detected in non-transfected HeLa 183 ( Figure S5B ) and A549 cell controls ( Figure S6A) . Moreover, we confirmed that the structure of ER, Golgi, 184 and mitochondria did not change in SEC61B-APEX2 or APEX2-expressed cells (Figure S6B-C ORF3a and M interactomes showed perturbed proteomic landscape at the ER membrane. 193 For mass spectrometric analysis, we used the aforementioned Spot-TurboID method (Figure 2A ORF3a-GFP and M-GFP co-expressed samples versus the control GFP-expressed sample were selected as 208 interactome proteins of ORF3a and M, respectively (log2(fold change, FC) ≥ 2.3 and -log10 p-value ≥ 2.3, 209 Figure 4B , S7E). The major population of the interactomes for ORF3a and M comprises endomembrane 210 system proteins, suggesting that SARS-CoV-2 ORF3a and M are primarily localized at the ER. Among the 117 filtered proteins for the ORF3a interactome, 40.2% were located at the plasma membrane 212 and 37.6% were located at the endomembrane (Figure 4C, S7F) . Among the 191 filtered proteins of the M 213 interactome, 49.7% were located at the plasma membrane and 25.1% were located at the endomembrane 214 ( Figure 4C, S7F) . Thus, a large proportion of plasma membrane and endomembrane proteins were present in 215 both, the ORF3a and M interactomes. According to gene ontology analysis, proteins associated with regulation 216 of localization and ubiquitin-dependent ER-associated protein degradation (ERAD) pathway were highly 217 enriched among the 117 proteins of the ORF3a interactome ( Figure 4D, S7G) . Proteins related to localization 218 and nitrogen compound transport were highly enriched among the 191 proteins comprising the M interactome 219 (Figure 4D, S7G) . From these analyses, we could see that the molecular functions of ORF3a and M are highly 220 related to localization and transport of other substances within the endomembrane of host cells. To further our understanding of these interactions, we then filtered the top 30 most strongly biotinylated 222 proteins (red or blue-colored dots in Figure 4B ) among the ORF3a and M interactomes, respectively. The 30 223 filtered proteins of the ORF3a interactome were classified according to localization and function ( Figure 4E ). 224 Surprisingly, 13 of these 30 proteins (43.3%) overlapped with the recently identified mitochondrial-associated membrane. In our work, membrane topological information of the biotin-labeled membrane proteins was 260 easily extracted. Using our Spot-TurboID method, we obtained mass spectrometry data that included the 11 biotinylation sites of each digested peptides. All digested peptides isolated following SA-bead enrichment had 262 at least one biotin-modified site at a lysine (K) residue. Since the biotinylation by TurboID-GBP occurred via 263 cytosolic regions of the protein, biotinylated sites on proximal interacting proteins of vPOI-GFP logically 264 faced the cytosol. We then obtained membrane topological information for proteins harboring a TM domain 265 using mass spectrometry data. It is noteworthy that using similar approaches, we successfully revealed the Figure 5 ). 274 It is noteworthy that this detailed topological information can be only obtained by direct detection of 275 biotinylated peptides, however, conventional methods that overlook these biotin-PTM signatures cannot report 276 it. We believe that our data provide detailed information on the domains heading toward cytosolic domain of 277 ORF3a and M proteins at the membrane, which lays foundation for further studies on host membrane proteins. Notably, among the biotinylated membrane proteins, several lysosome-associated proteins (i.e., VPS16, In regard to the ORF3a interaction with membrane proteins at the autophagosome and lysosome, RNF5, which 12 is strongly biotinylated by Spot-TurboID (Figure 6A ), interests us because it is an ER-anchored E3 ubiquitin- Since the molecular function of RNF5 with SARS-CoV-2 is not fully characterized yet, we conducted 295 further studies to reveal its functional relationship with SARS-CoV-2 proteins. From confocal microscopy 296 experiments, we observed that cytoplasmic dispersed Flag-conjugated RNF5 was largely altered by ORF3a Interestingly, we also observed that many of the ORF3a and M interactome proteins overlapped with our 313 recently identified MAM proteins. We have identified 115 local resident proteins using Contact-ID, which 314 utilized the reconstituted biotinylating activity of split BioID fragments at the MAM, which is an ER and 13 mitochondrial contact site (Kwak et al., 2020). Twenty-eight of the proteins that filtered out from the ORF3a 316 interactome (23.9%) and the 24 from the M interactome (12.6%), overlapped with the 115 proteins previously 317 identified to be a part of the MAM proteome. A total of 18 proteins overlapped between ORF3a, M, and the 318 MAM proteome ( Figure 7A ). The complete list of proteins in the ORF3a and M proteomes that overlap with 319 the MAM proteome, classified by function, are given in Figures 7B and 7C , respectively. Since many of the 320 ORF3a interactome proteins overlapped with those from the MAM proteome rather than the M interactome, 321 we questioned whether ORF3a could more strongly hinder MAM formation. To address this, we used the Contact-ID tool to monitor changes in the proteomic composition at the MAM 323 in response to ORF3a expression ( Figure 7D ). (Contact-ID generated more biotinylated proteins in the 324 ORF3a-expressing sample compared to that in the control sample (blue arrows in Figure 7E ; Figure S11A -325 D), which implies that ORF3a may lead to a considerable change of the proteomic landscape at the MAM. We 326 also found that ORF3a was biotinylated by Contact-ID (red arrows in Figure 7E ). Thus, we are currently 327 conducting further investigations to analyze which proteins are recruited to the MAM in response to ORF3a 328 expression. Similar to ORF3a, we observed that M was also biotinylated by Contact-ID ( Figure S12 ). 329 We speculate that the expression of vPOI may change the secretory proteome because our EM results TurboID-expressing cells can be detected in the culture media or mouse bloodstream. In this study, we utilized 337 the iSLET method to assess whether the secretome profiles of the host cells changed in response to ORF3a 338 and M expression ( Figure 7F) . 339 In the SA-HRP western blot results, the amounts of biotinylated proteins in the culture media of cells co-340 expressing SEC61B-TurboID and ORF3a were substantially higher compared to that of the control sample 341 lacking ORF3a expression ( Figure 7G , S13A-B). In addition, we found ORF3a in the culture media of the 342 14 cells (red arrows in Figure 7H , S13C). In whole cell lysates, ORF3a was also biotinylated by SEC61B-TurboID ( Figure S14 ). Similar to ORF3a, we observed that the amounts of biotinylated proteins increased in 344 the secretome of M-expressed cells (Figure S15A) , and M protein was also secreted from cells ( Figure S15B ) 345 and was biotinylated by SEC61B-TurboID ( Figure S15C ). 346 These results imply that ORF3a and M may upregulate cargo secretion from the ER and there may be a 347 specific sorting pathway for the secretion of ORF3a in the ER lumen, as has been shown for other secreted In addition, we also found that many of the proteins in the ORF3a and M interactomes are targeted by 398 approved drug molecules, which may be useful for drug repurposing therapeutic strategies. The druggable 399 proteins targeted by previously approved drug molecules are listed in Supplementary Table 2-3. 400 In this study, we also provide evidence that our Spot-TurboID workflow using the GFP/GBP binding system No. 21126, 1:10,000 dilution) were incubated for 1 h at room temperature. After washing three times with 517 TBST for 5 min at room temperature, results of immunoblotting assay were obtained using ECL solution using 518 GENESYS program. Line scan analysis was performed using image J software program (NIH). After without streptavidin beads was transferred to new tubes. This elution step was repeated at least 4 times. These 578 total elution fractions were dried for 5 h using Speed-vac (Eppendorf). Before samples were injected to mass 579 spectrometry, they were stored at -20 °C. Figures S1-S15. is given in Figure S8 and Dataset S1. Identification of proteins interacting with vPOI using GFP-GBP system A Protein transport and chaperone Secretion of the human T 723 cell leukemia virus type I transactivator protein tax RNA granules Upregulated function of 727 mitochondria-associated ER membranes in Alzheimer disease Modular Detection of GFP-Labeled Proteins for Rapid Screening by Electron 730 Microscopy in Cells and Organisms Efficient proximity labeling in living cells and organisms with TurboID Inhibition of autophagy limits vertical 735 transmission of Zika virus in pregnant mice Ubiquitination of SARS-CoV-2 ORF7a promotes antagonism of interferon response The SARS-Coronavirus Membrane protein 739 induces apoptosis through modulating the Akt survival pathway The ion channel activity of the SARS-coronavirus 3a protein is linked to its pro-apoptotic function Integrative Imaging Reveals SARS-CoV-2-Induced Reshaping of 746 Comparative multiplexed interactomics of SARS-CoV-2 and homologous coronavirus non-structural proteins identifies unique and shared host-cell 749 dependencies. bioRxiv Wolfram syndrome: MAMs' connection? Coronavirus-Induced Host Cubic Membranes and Lipid-Related Antiviral 753 Therapies: A Focus on Bioactive Plasmalogens Alternatively spliced soluble 755 coxsackie-adenovirus receptors inhibit coxsackievirus infection The membrane 757 protein of SARS-CoV suppresses NF-kappaB activation Reovirus forms neo-organelles for progeny particle assembly within reorganized cell 760 membranes Coronavirus-induced ER stress response and its involvement in 762 regulation of coronavirus-host interactions A SARS-CoV-2 protein interaction map reveals targets for drug 765 repurposing Human VAP-B is involved in hepatitis C virus replication through interaction 768 with NS5A and NS5B Autophagosomes form at ER-mitochondria contact sites The coxsackie-adenovirus receptor (CAR) is used by 828 reference strains and clinical isolates representing all six serotypes of coxsackievirus group B and by swine 829 vesicular disease virus COVID-19 virus SARS-CoV-2 blocks HOPS complex-mediated assembly of the SNARE complex required 832 for autolysosome formation Zika virus induces massive cytoplasmic vacuolization and paraptosis-835 like death in infected cells Electron microscopy of herpes simplex virus. II. Sequence of 837 development A Global Map of Lipid-Binding Proteins and Their Ligandability in 840 Cells Formation of cytoplasmic heat shock granules in tomato 842 cell cultures and leaves Links between recognition and degradation of cytoplasmic 844 viral RNA in innate immune response Downregulation of the small GTPase 846 SAR1A: a key event underlying alcohol-induced Golgi fragmentation in hepatocytes Structure and function of ER membrane contact sites with other 848 organelles A deficient public health system as a contributing cause 850 of Severe Acute Respiratory Syndrome (SARS) epidemic in mainland China Neurobiochemical Cross-talk Between 853 COVID-19 and Alzheimer's Disease The ORF3a protein of SARS-CoV-2 induces apoptosis in cells Proteomic 857 mapping of mitochondria in living cells via spatially restricted enzymatic tagging A promiscuous biotin ligase fusion protein identifies 859 proximal and interacting proteins in mammalian cells A SARS-CoV-2 -host proximity interactome. bioRxiv Multilamellar structures and filament bundles are found on the cell 864 surface during bunyavirus egress Mitochondria: the next (neurode)generation ER-phagy mediates selective degradation of endoplasmic 867 reticulum independently of the core autophagy machinery SARS-CoV-2 protein ORF3a is pathogenic in Drosophila and 869 causes phenotypes associated with COVID-19 post-viral syndrome Severe acute respiratory syndrome coronavirus ORF3a protein activates the 872 NLRP3 inflammasome by promoting TRAF3-dependent ubiquitination of ASC Formation of stacked ER cisternae by low affinity protein interactions A unifying structural and functional model of the coronavirus replication 878 organelle: Tracking down RNA synthesis