An Autoantigen Profile of Human A549 Lung Cells Reveals Viral and Host Etiologic Molecular Attributes of Autoimmunity in COVID-19

We aim to establish a comprehensive COVID-19 autoantigen atlas in order to understand autoimmune diseases caused by SARS-CoV-2 infection. Based on the unique affinity between dermatan sulfate and autoantigens, we identified 348 proteins from human lung A549 cells, of which 198 are known targets of autoantibodies. Comparison with current COVID data identified 291 proteins that are altered at protein or transcript level in SARS-CoV-2 infection, with 191 being known autoantigens. These known and putative autoantigens are significantly associated with viral replication and trafficking processes, including gene expression, ribonucleoprotein biogenesis, mRNA metabolism, translation, vesicle and vesicle-mediated transport, and apoptosis. They are also associated with cytoskeleton, platelet degranulation, IL-12 signaling, and smooth muscle contraction. Host proteins that interact with and that are perturbed by viral proteins are a major source of autoantigens. Orf3 induces the largest number of protein alterations, Orf9 affects the mitochondrial ribosome, and they and E, M, N, and Nsp proteins affect protein localization to membrane, immune responses, and apoptosis. Phosphorylation and ubiquitination alterations by viral infection define major molecular changes in autoantigen origination. This study provides a large list of autoantigens as well as new targets for future investigation, e.g., UBA1, UCHL1, USP7, CDK11A, PRKDC, PLD3, PSAT1, RAB1A, SLC2A1, platelet activating factor acetylhydrolase, and mitochondrial ribosomal proteins. This study illustrates how viral infection can modify host cellular proteins extensively, yield diverse autoantigens, and trigger a myriad of autoimmune sequelae.


Introduction
In total, Orf3 affected 71 DS-affinity proteins identified from A549 cells, which includes those directly interacting with Orf3 and those perturbed by Orf3 protein expression in A549 cells. The large number of Orf3-affected host proteins implicates important roles of Orf3 in SARS-CoV-2 infection. Network analysis reveals these proteins to be mostly associated with gene expression regulation, cytoplasmic vesicles, apoptosis, response to stress, monosaccharide biosynthesis, or hydrolase activity (Fig. 5). Several of these are classical nuclear autoAgs, e.g., PNCA, SSB (Lupus La), XRCC5 (Lupus Ku80, thyroid-lupus autoAg), XRCC6 (Ku70), and SNRPB (SmB/B'). A few are unknown autoAgs but with important relevance to COVID, e.g., PAFAH1B2 and PAFAH1B3 (the alpha catalytic subunits of the cytosolic type I platelet-activating factor (PAF) acetylhydrolase). PAF is produced by a variety of cells involved in host defense, and PAF signaling can trigger inflammatory and thrombotic cascades. The modulation of PAF by SARS-CoV-2 Orf3 may partially explain the frequently occurring thrombotic complications and coagulopathy in COVID -19 patients. PAF also induces apoptosis in a PAF receptor independent pathway that can be inhibited by PAFAH1B2 and PAFAH1B3 (38).
SARS-CoV-2 E protein affects a number of ribonucleoproteins that are related to translation initiation and mRNA splicing, e.g., hnRNP (U and UL1) and ribosomal protein (L7, L8 L11, L12, L35A). E-affected proteins are associated with establishment of protein localization to membrane, regulation of autophagy, and post-translational protein modification (Fig. 5). SARS-CoV-2 M, Nsp1, and N proteins also affect various ribonucleoproteins, whereas Nsp13 appears to affect proteins associated with the cytoskeleton. Overall, the majority of DS-affinity proteins found affected by individual SARS-CoV-2 proteins are known autoAgs ( Fig. 5 and Table 1), which indicates that host proteins perturbed by viral components are an important source of autoAgs.
Orf9b of SARS-CoV has been shown to localize to mitochondria, trigger ubiquitination and proteasomal degradation of dynamin-like protein 1, limit host cell interferon signaling by targeting mitochondrial associated adaptor molecule MAVS signalosome, and manipulate the mitochondrial function to help evade host innate immunity (39). Orf9b of SARS-CoV-2 has been reported to suppress the type I interferon response by targeting TOM70 (40). In COVID-19 pneumonia patients, monocytes show altered bioenergetics and mitochondrial dysfunction with depolarized and abnormal ultrastructure (41).
Currently, little is known about the involvement of mitochondrial ribosomes or mitochondrial translation in SARS-CoV-2 infection. Expression of mitochondrial ribosomal proteins associated with protein synthesis has been found to be the most striking transcriptional difference among dengue virus-infected children, as revealed by a genome-wide microarray analysis of whole blood RNA from 34 infected children collected on days 3-6 of illness (42). In human cytomegalovirus infection, proteins involved in biogenesis of the mitochondrial ribosome changed early during the viral replication cycle (43). Mitochondria are vital to cell survival and apoptosis as they produce the majority of adenosine triphosphate (ATP) that provide chemical energy to cells. Especially for cells such as muscles that require much ATP, mitochondrial dysfunction will certainly lead to problems such as muscle weakness and fatigue. The roles of mitochondrial ribosomal proteins play in COVID and long-term sequelae merit further investigation.

AutoAgs related to ubiquitination alteration in SARS-CoV-2 infection
Ubiquitination provides a universal signal for protein degradation. By comparing our data with the ubiquitinome of SARS-CoV-2 infected cells, we identified 102 DS-affinity proteins that are altered by ubiquitination during viral infection (Supplemental Table 1). These ubiquitination-altered proteins are significantly associated with gene expression, catabolic process, regulation of apoptotic process, cytoplasmic vesicles, and cytoskeleton ( Fig. 6). They include 15 ribosomal proteins, 8 heat shock proteins, 5 hnRNP proteins, 5 histones, 4 translation elongation factors, and 3 translation initiation factors, and a majority of them are known autoAgs (Table 1). Three ubiquitination/de-ubiquitination enzymes (UBA1, UCHL1, and USP7) are COVID-altered and possess DS-affinity, with UBA1 and UCHL1 being known autoAgs. UBA1 catalyzes the first step in ubiquitin conjugation to mark proteins for degradation through the ubiquitin-proteasome system. USP7 is a hydrolase that deubiquitinates target proteins. UCHL1 is a thiol protease that recognizes and hydrolyzes a peptide bond at the C-terminal glycine of ubiquitin, and is involved in the processing of ubiquitin precursors and of ubiquitinated proteins. UBA1 is found down-regulated by Orf3 expression. UCHL1 is found in the Orf3 interactome, up-regulated by SARS-CoV-2 E protein expression, and down-regulated by Nsp12, Nsp8, Orf8, and Orf9b (Supplemental Table 1).
PLPs, along with other proteases, are responsible for processing replicase proteins that are required from viral replication. PLP of SARS-CoV-2 is able to reverse host ubiquitination and remove interferonstimulated gene product 15 (ISG15), and its substrate activity mirrors closely that of PLP of MERS (45).
Ubiquitin modifications can regulate innate immune response and apoptosis, and ISG15 is a ubiquitin-like modifier typically expressed during host cell immune response. Overall, various components of SARS-CoV-2 appear to be able to alter uniquitination of host proteins. The large pool of ubiquitin-altered proteins in SARS-CoV-2 infection indicates that ubiquitin modification, such as differential abundance and dynamic ubiquitination pattern change, may be a major origin of autoAgs.

AutoAgs related to phosphorylation alteration in SARS-CoV-2 infection
Comparing our data with currently available phosphoproteome data of COVID-19 (26,34), 97 phosphoproteins are found with both DS-affinity and COVID-induced alteration, with a majority (52/97) related to gene expression (Fig. 7). Notably, they include 8 heterogeneous nuclear ribonucleoproteins affected by phosphorylation (HNRNPA2B1, HNRNPC, HNRNPH1, HNRNPK, HNRNPM, HNRNPU, HNRNPUL1, HNRNPUL2), with 4 being known autoAgs (Table 1). HNRNPs are involved in many cellular . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made There are 25 phosphorylation-altered proteins that are related to vesicle-mediated transport, most of which are known autoAgs, including ACLY, ACTA2, ACTB, ALB, ALDO, ANXA2, FLNA, COPA, SPTAN1, SPTBN1, TLN1, TUBB4, and VCL ( Fig. 7 and Table 1). The coatomer, to which COPA (coatomer subunit alpha) belongs, is a cytosolic protein complex that associates with Golgi non-clathrin-coated vesicles and is required for budding from the Golgi membrane. COPA is associated with autoimmune interstitial lung, joint, and kidney disease (47).
There are 18 phosphorylation-altered potential autoAgs with ATP binding activity, and 12 with kinase binding activity (Fig. 7). In particular, PRKDC (DNA-dependent protein kinase catalytic subunit) is identified with strong DS-affinity and is a known autoAg. It is a serine/threonine-protein kinase that acts as a molecular sensor for DNA damage, with involvement in numerous biological processes such as DNA damage and repair, immunity, innate immunity, ribosome biogenesis, and apoptosis. PRDKC is found in the interactomes of M and Nsp4 proteins of SARS-CoV-2 and up-regulated by expression of Nsp10, Nsp9, Orf7a, or Orf7b protein in A549 cells (19,34). PRKDC is also found up-regulated at 0 h and 4 h in SARS-CoV-2 infected Vero E6 cells (26) and up-regulated at 24 h in SARS-CoV-2 infected Caco-2 cells (20). These findings suggest that phosphorylation by PRKDC plays extensive and important roles in COVID.
Proteins phosphorylated during apoptosis are common targets of autoantibodies. For example, the U1-70 snRNP autoAg undergoes specific changes in the phosphorylation/dephosphorylation balance and cellular localization during apoptosis (48), and phosphorylated U1-snRNP complex induced by apoptosis is recognized by autoantibodies in patients with systemic lupus erythematosus (49). A high degree of phosphorylation of SSB (lupus La autoAg) substantially diminished its poly(U) binding capacity, but its binding to human autoantibodies increased 2-fold with increased phosphorylation (50). On the other hand, SSB autoAg has also been reported to be dephosphorylated and cleaved during early apoptosis (51).
During apoptosis, ribosomal protein P1 and P2 autoAgs are completely dephosphorylated while P0 autoAg is partially dephosphorylated (52). Therefore, alterations in phosphorylation, either hyper-or hypophosphorylation, may lead to changes in self-molecules and render them autoantigenic.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Conclusion
In our quest for a comprehensive autoantigen atlas for COVID-19, we report an autoantigen profile of 191 confirmed autoAgs and 150 putative autoAgs in SARS-CoV-2 infection. These proteins are initially identified from human lung epithelial A549 cells using a unique DS-affinity autoAg enrichment strategy, and then compared with currently available COVID-omics data. Our study reveals that cellular processes and components integral to viral infection are major origins of autoAgs, including gene expression, ribonucleoprotein biogenesis, translation and mitochondrial translation, vesicle and vesicle-mediated transport, and cytoskeleton. Ubiquitination and phosphorylation are particular post-translational modifications that cause changes in self-molecules and render them autoantigenic. Impaired clearance of apoptotic and dead cell material is considered a major pathogenic attribute to autoimmune disease. We have previously shown that DS possesses unique affinity to apoptotic cells and their released autoAgs, and our current study further demonstrates that ubiquitination and phosphorylation associated with apoptosis are possibly major sources of molecular alterations in self-molecule to autoantigen transformation. Overall, our study demonstrates that SARS-CoV-2 causes extensive alterations of host cellular proteins and produces a large number of potential autoAgs, indicating that there may be an intimate relationship between COVID infection and autoimmunity.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

A549 cell culture
The A549 cell line was obtained from the ATCC (Manassas, VA, USA) and cultured in complete F-12K medium at 37 °C in 75 cm 2 flasks to 80% confluency. The growth medium was supplemented with 10% fetal bovine serum and a penicillin-streptomycin-glutamine mixture (Thermo Fisher).

Protein extraction
About 100 million A549 cells were suspended in 10 ml of 50 mM phosphate buffer (pH 7.4) containing the Roche Complete Mini protease inhibitor cocktail. Cells were homogenized on ice with a microprobe sonicator until the turbid mixture turned nearly clear with no visible cells left. The homogenate was centrifuged at 10,000 g at 4 °C for 20 min, and the total protein extract in the supernatant was collected.
Protein concentration was measured by absorbance at 280 nm using a NanoDrop UV-Vis spectrometer (ThermoFisher).

DS-Sepharose resin preparation
The DS-affinity resins were prepared as previously described (3,5). In brief, 2 ml of EAH Sepharose 4B

DS-affinity fractionation
The total proteins extracted from A549 cells were fractionated in a DS-Sepharose column with a BioLogic . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Mass spectrometry sequencing
Protein sequencing was performed at the Taplin Biological Mass Spectrometry Facility at Harvard Medical School. Proteins in gels were digested with sequencing-grade trypsin (Promega) at 4 °C for 45 min. Tryptic peptides were separated on a nano-scale C 18 HPLC capillary column and analyzed in an LTQ linear ion-trap mass spectrometer (Thermo Fisher). Peptide sequences and protein identities were assigned by matching the measured fragmentation pattern with proteins or translated nucleotide databases using Sequest. All data were manually inspected. Only proteins with ≥2 peptide matches were considered positively identified.

COVID data comparison
DS-affinity proteins were compared with currently available proteomic and transcriptomic data from SARS-CoV-2 infection compiled in the Coronascape database (as of 12/14/2020) . These data had been obtained with proteomics, phosphoproteomics, interactome, ubiquitome, and RNA-seq techniques.
Up-and down-regulated proteins or genes were identified by comparing COVID-19 patients vs. healthy controls and cells infected vs. uninfected by SARS-CoV-2. Similarity searches were conducted between our data and the Coronascape database to identify DS-affinity proteins (or their corresponding genes) that are up-and/or down-regulated in the viral infection.

Protein-protein interaction network analysis
Protein-protein interactions were analyzed by STRING (16). Interactions include both direct physical interaction and indirect functional associations, which are derived from genomic context predictions, high-throughput lab experiments, co-expression, automated text mining, and previous knowledge in databases. Each interaction is annotated with a confidence score from 0 to 1, with 1 being the highest, indicating the likelihood of an interaction to be true. Only interactions with high confidence (a minimum score of 0.7) are shown in the figures.

Pathway and process enrichment analysis
Pathways and processes enrichment were analyzed with Metascape (17), which utilize various ontology sources such as KEGG Pathway, GO Biological Process, Reactome Gene Sets, Canonical Pathways, CORUM, TRRUST, and DiGenBase. All genes in the genome were used as the enrichment background. Terms with a p-value <0.01, a minimum count of 3, and an enrichment factor (ratio between the observed counts and . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 22, 2021. ; https://doi.org/10.1101/2021.02.21.432171 doi: bioRxiv preprint the counts expected by chance) >1.5 were collected and grouped into clusters based on their membership similarities. The most statistically significant term within a cluster was chosen to represent the cluster.

Autoantigen confirmation literature text mining
Literature searches in Pubmed were performed for every DS-affinity protein identified in this study. Search keywords included the protein name, its gene symbol, alternative names and symbols, and the MeSH keyword "autoantibodies". Only proteins with their specific autoantibodies reported in PubMed-listed journal articles were considered "confirmed" autoAgs in this study.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Authors' contributions
JYW directed the study, analyzed data, and wrote the manuscript. WZ performed some experiments and reviewed the manuscript. MWR and VBR assisted with data analysis and manuscript preparation. MHR consulted on the study and data analysis and edited the manuscript. All authors have approved the manuscript.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 22, 2021. ; https://doi.org/10.1101/2021.02.21.432171 doi: bioRxiv preprint