N-glycoproteomic Proling Revealing New Coronavirus Therapeutic Targets That Maybe Involved in Cepharanthine’s Intervention

N-glycosylation is an important post-translational modication involved in protein folding, signal transduction, extracellular matrix organization and immune response. Evidence showed that glycosylated SARS-CoV-2 Spike protein may be a potential target in viral pathogenesis and drug/vaccine design. To investigate the mechanism of coronavirus infestation and drug targets from glycosylation perspective, we constructed a SARS-CoV-2 cellular model using GX_P2V-infected VeroE6 cells to study the effects of GX_P2V on glycoproteins in presence or absence of Cepharanthine (CEP) through N-glycoproteomics proling. The results showed that coronavirus GX_P2V could cause aberrant protein glycosylation, whereas CEP can partially maintain GX_P2V-induced aberrant glycoproteins at homeostasis. Further study revealed that proteins LAMB1 and FN1 were pivotal in counteracting coronavirus-induced aberrant protein glycosylation by CEP. Furthermore, CEP can dramatically regulate the glycosylation of viral proteins S, M and N. Our results suggest that despite the strong anti-coronavirus effects of CEP, drug combinations need be considered to achieve optimal therapeutic strategies.


Introduction
The current global pandemic of COVID-19 caused by the highly contagious coronavirus SARS-CoV-2 has had a huge impact on international community. Currently, no anti-SARS-CoV-2 drugs or vaccines have been o cially and clinically approved due to the absence of adequate evidence for exact mechanism of SARS-COV-2 infestation. Therefore, it is of great urgent need to deeply understand its mechanism to screen potential drug targets 1 .
Glycosylation is the most abundant and complex post-translational modi cation which can have profound structural and functional effects on the conjugate 2 . The oligosaccharide fraction is considered to be related with multiple biological processes and affects the physical properties of proteins. It has been known that protein glycosylation is de nitely stable among the normal people, although it is different from person to person. Only when a person's homeostasis changes due to lifestyle or pathological conditions, will glycosylation changes signi cantly 3,4 . The glycosylation has been labeled an important quality attribute of biopharmaceuticals 5 . The S1 N-terminal domain of the spike protein (S1-NTD) may "hide" sites for recognition and binding of sialic acid-containing glycans, which may provide possibilities for ACE2 alternative receptors 6 . Therefore, it is reasonable to take the glycosylation in consideration in the treatment of coronavirus infestation.
GX_P2V is a SARS-CoV-2 related coronavirus isolated from pangolin. Its spike protein shares 92.2% amino acid identity with the spike protein of SARS-CoV-2 7 . Besides, GX_P2V also has the shared receptor with SARS-CoV-2 but has no pathogenicity to human. Based on the above characteristics, it has been reported that GX_P2X can be used as an accessible in vitro model for developing therapies or for mechanism investigation against SARS-CoV-2 7 . Here we adopted this well-established coronavirus SARS-CoV-2 cellular model by using GX_P2V infestation with Vero E6 to investigate the mechanisms of coronavirus SARS-CoV-2 infestation with respect to the dysregulation of glycoprotein and screening for potential targets for therapeutic intervention.
We have reported that cepharanthine (CEP), a clinically approved drug to increase immune function after chemotherapy, has anti-coronavirus properties that prevent the infected host cells from further damage 7 .
But the mechanism of CEP on glycosylation targets remains unclear. In this paper, we carried out Nglycosylation modi cation proteomics analysis to study the mechanisms of SARS-CoV-2 infection and the anti-coronavirus effects of CEP. Using the advanced technique 4D label-free quanti cation technology, we constructed the N-glycoproteomic pro ling, which revealed new therapeutic targets against coronavirus infestation and CEP's intervention. Our results provide important insights into the mechanism of coronaviruses infection and the identi cation of drug targets for anti-coronavirus of multi-target drug combination usage.

Results And Discussion
N-glycoproteomic pro ling highlights the differentially expressed glycoproteins in SARS-CoV-2 infection cellular model and CEP intervention.
N-Glycosylation is one of the most important post-translational modi cations of proteins 8 . Although there are only a few hundreds of glycoproteins in total, their highly complex glycosylation increases the number of theoretical protein morphology by several orders of magnitude 9 . Many cell surface and extracellular matrix-related proteins become glycosylated through secretory pathways, thereby regulating their adhesion interactions, physical and chemical properties, and diversifying their functions 10 .
To systematically study the differentially expressed glycoproteins during SARS-CoV-2 infestation, we adopted our well-established SARS-CoV-2 cellular model 7 with GX_P2V-treated Vero E6 cells, and performed the N-glycoproteomic pro ling via the 4D label-free LC-MS/MS analysis in presence or absence of GX_P2V and/or CEP. The work ow for the experiment was outlined in supplementary Figure S1. First, the cell samples were harvested and enzymatically digested and then analyzed by 4D label-free LC-MS/MS and bioinformatics. A total of 737,202 secondary spectrograms were obtained by mass spectrometry (supplementary Figure S1B). A total of 1770.0 N-glycosylation modi cation sites on 828.0 proteins were identi ed with database searching, of which 1298.0 sites on 624.0 proteins have quantitative information. The detected N-glycosylation tends to obey N-X-T, N-C-S, N-G-S, N-X-S, N-X-C, N-N rules (supplementary Figure S1C).
We then performed the analyses of global proteomics and respective of glycosylation occupancy ( Figure  1A), and took the fold changes (Fc= expression levels of glycosylated peptide in Group2 (or 3 or 4)/Group 1 ratios) at Fc ≥ 2 folds (up-regulated) or Fc≤0.5 folds (down-regulated) as differentially expressed protein, where group 1 was the mock control (Vero), group 2 was treated with CEP only (Vero_C) , group 3 was treated by coronavirus GX_P2V only (Vero_P) and group 4 was treated by CEP and GX_P2V (Vero_C_P). In comparison with the mock control group, GX_P2V treatment can up-regulate 26 glycosylation sites in 23 proteins and down-regulate 58 glycosylation sites in 47 proteins, whereas in presence of CEP, GX_P2V treatment can induce 26 glycosylation sites up-regulated in 22 proteins and 110 glycosylation sites down-regulated in 75 proteins in host Vero E6 cells ( Figure 1B).
To investigate the protein-protein interaction (PPI) networks and the interaction modules of these differentially expressed proteins in Vero E6 cells in response to GX_P2V treatment, we conducted STRING and Cytoscape analysis ( Figure 1C). We found that the key interaction module obtained from PPI network with MCODE plugin involved in 9 aberrant glycoproteins including 4 up-regulated glycoproteins namely Integrin Subunit Alpha 1(ITGA1), Integrin Subunit Beta 3(ITGB3), Laminin Subunit Gamma 1 (LAMC1) and Heparan Sulfate Proteoglycan 2(HSPG2), and 4 down-regulated glycoproteins namely Integrin Subunit Beta 1 (ITGB1), Laminin Subunit Beta 1(LAMB1), Integrin Subunit Alpha 5(ITGA5) and Fibronectin 1 (FN1). The protein (ITGA3) contains both up-and down-regulated glycosylation sites ( Figure   1D). LAMC1 was reported to be involved in cell proliferation, angiogenesis, growth, migration and invasion 11,12 . Remarkably, majority of the 9 aberrant glycoproteins in the key interaction module were enriched in ECM-receptor interaction pathways ( Figure 1D  Coronavirus GX_P2V-induced alteration in protein glycosylation are prevalent in infected cells N-linked glycosylation was involved in multiple biological processes, such as receptor interactions, immune responses, protein secretion and transport, cell adhesion, signal transduction, etc. As described above, we detected 84 sites of aberrant glycosylation sites present in 69 glycoproteins from coronavirus GX_P2V-treated Vero E6 cells. To study these sites holistically, we performed functional enrichment clustering analysis of these differential glycoproteins using Metascape (Figure 2A), and the results were graphically presented using the Sangerbox mapping toolbox ( Figures 2B and 2C).
Upon the treatment of GX_P2V, the involved functions alteration related to differentially expressed proteins were present in Figure 2A and 2B. The most involved top 20 functions were listed in Figure 2A, and corresponding glycosylation counts of each function presented in Figure 2B. The circle diagram of protein function enrichment analysis demonstrated that ITGB1 was involved in the most relevant function and the extracellular matrix organization was enriched for the most genes. Besides, in response to coronavirus GX_P2V infection, the top 5 aberrant glycoproteins most involved relevant functions include ITGB1, TGFB2, ITGB3, ITGB5 and FN1, and the top 5 most aberrant functions involved most genes were those related to extracellular matrix organization, response to wounding, regulation of cell adhesion, signaling pathway regulated exocytosis and transmembrane receptor protein tyrosine kinase. Other important functions such as protein maturation, viral entry into host cell and a platelet degranulation are also affected.
To have an overview about the effects of coronavirus GX_P2V treatment on host cells, we summarized the subcellular localization of differentially expressed glycosylated proteins and PPI networks by UniProt database 13 and STRING software ( Figure 3).
Aberrant alteration of protein glycosylation can be seen in cytoplasm, endoplasmic reticulum, Golgi apparatus, lysosome, nucleus, cell membrane and extracellular matrix. The PPI networks of differentially expressed glycoproteins in each subcellular distribution were also highlighted ( Figure 3). These results suggest that coronavirus-induced alterations in glycosylation of host cell proteins are widespread in infected cells and have multiple subcellular localizations. The most involved glycoproteins locate in cell membrane, extracellular matrix, lysosomes, and endoplasmic reticulum system and cytosol.
Effects of CEP on GX_P2V-induced aberrant glycosylation of proteins in Vero E6 cells.
We have reported that CEP could effectively inhibit coronavirus infestation 7 . To identify the potential targets of CEP in coronavirus-infected cells in response to GX_P2V treatment, we compared the aberrant glycosylation sites and glycoproteins with and without coronavirus treatment in presence or absence of CEP and perform Reactome Pathway Analysis ( Figure 4).
We then carried out Reactome Pathway Analysis on the CEP-affected 12 glycoproteins to understand intervention mechanism of CEP ( Figure 4B). The results showed that CEP can exert its antiviral and cellprotective effects by targeting multiple pathways such as ECM proteoglycans, extracellular matrix organization, platelet degranulation, response to elevated platelet, neutrophil degranulation, and etc.
Among these 12 proteins, LAMB1 and FN1 are also the pivotal proteins in the key interaction module which consists of 9 aberrant glycoproteins in response to coronavirus GX_P2V infestation ( Figure 1D), indicating that proteins LAMB1 and FN1 might be the key targets of CEP against GX_P2V infestation ( Figure 3).
In addition, several proteins can be over-regulated by CEP against GX_P2V (Supplementary Table S1). There are two aberrant glycosylation sites on MRC2 induced by GX_P2V infection. In presence of CEP, the site N69 on MRC2 protein can remain normally glycosylation state, while the other site N1134 is shifted from virus-induced up-regulation to down-regulation. MRC2 may contribute to cellular uptake, remodeling and degradation of extracellular collagen matrices. Like the glycosylation regulation at N1134 site on MRC2 regulated by CEP against GX_P2V, the glycosylation of N83 site on plasminogen activator urokinase receptor (PLAUR) can be over-regulated by CEP, but it is shifted from down-regulation to upregulation, by which CEP might regulate plasminogen activation system controlling the extracellular level of PLAUR/PLAU, and thus may regulate protease activity at the cell surface 14,15 . Over-regulated glycosylation by CEP can also be seen at the site N277 of granulin precursor (GRN), which is a secreted glycoprotein that acts as a key regulator of lysosomal function and as a growth factor involved in in ammation, wound healing and cell proliferation 16 . It involves in regulating protein tra cking to lysosomes and, also the activity of lysosomal enzymes 16,17 . Alteration of glycosylation GRN may contribute to lysosome-related functions. Therefore, CEP might also perform its anti-viral effect via regulation of glycosylation of MRC2, PLAUR, and GRN.
Effect of CEP on the N-glycosylation of GX_P2V viral proteins.
Coronavirus viral proteins are normally extensively glycosylated, especially coronavirus spike proteins where it encodes around 66-87 N-linked glycosylation sites per trimeric spike. It has been known that the extensive glycosylation of viral protein plays a role in protein folding and shielding immunogenic epitopes, resulting in immune evasion 18,19,20 . Remarkably, SARS-CoV and SARS-CoV-2 recognize the human angiotensin converting enzyme-2 (ACE-2) receptor via their glycosylated spike proteins 21 .
From the dataset for N-glycoproteomic pro ling, we observed that the identi ed viral proteins include M, N and Spike were highly glycosylated ( Figure 5A), which might offer the viral proteins a glycan shield 18,19 . Interestingly, we found that the glycosylation of viral proteins M, N and S was affected by CEP treatment. In particular, CEP can up-regulate the glycosylation of the N5 site of M protein, the N-terminal sites of the viral S protein (including the N30, N122, N331 and N343 sites) and the sites N1070 and N1094 sites in the S2 region (N1070 and N1094, the corresponding positions on SARS-CoV-2 spike protein are N1074 and N1098, respectively). However, CEP can dramatically decrease N25-site glycosylation of viral N protein, but no obvious effect can be seen in the glycosylation of the C-terminal sites N1154, N1169, and N1190 of S proteins under CEP treatment.
To observe and analyze the location of these affected glycosylation sites on S proteins and their possible roles, we compared the sequences of spike protein of SARS-CoV-2 and GX_P2V viruses. Due to the structure of GX_P2V viral S protein is not available by far, and considering the high sequence similarity between SARS-CoV-2 and GX_P2V, we adopted the 3D structure of SARS-CoV-2 (6VXX, closed state; 6VYB, closed state) to label the affected glycosylation sites (Fig. 5B-G). Among the sites affected by CEP on Spike protein, N30 and N122 are located at NTD region, N331 and N343 are in CTD/RBD region, and N1070 and N1094 are in S2 region 22 . As can be seen from the cartoon image ( Figure 5G), the CEPaffected Spike protein glycosylation sites are located at both ends of the S protein. Four of the sites (including those located in the NTD and CTD/RBD regions) are at one end away from the virion, while the sites affected by the two sites affected locate in the S2 region at one end close to the virion. This suggests that one of the effects of CEP on viral S proteins might be to interfere with viral assembly and viral recognition or binding to the S protein receptors to exert its antiviral effects.

Multi-target drug combination is essential for the treatment of coronavirus infestation
Through the analysis of coronavirus GX_P2V-induced host cell glycoproteomics, we found that the viral effect on infected host cells is multifaceted. It involved in dysfunction of multiple proteins and pathways, including ECM-receptor interaction, extracellular matrix organization, response to wounding, platelet degranulation, and etc. (Figure 1 and Figure 2). Our data uncovered coronavirus GX_P2V-induced aberrant glycosylation of a variety of proteins on membrane ( Figure 3). There are more than 75% of aberrant glycosylation of proteins after GX_P2V infection in correspondence to membrane-related subcellular structures, involving cell membrane, endoplasmic reticulum membrane, lysosomal membrane, Golgi apparatus membrane and nuclear membrane (Figure 3). These abnormal changes in glycoproteins caused by coronaviruses may play an important role in viral invasion of host cells, resulting in cellular dysfunction.
We found that the antiviral mechanism of CEP is also diverse, as it exerts its antiviral function by regulating the glycosylation of about 12 proteins which locate in endoplasmic reticulum, lysosomal, nucleus, cell membrane and ECM ( Figure 3). However, at least 57 aberrant glycoproteins caused by coronavirus were not signi cantly improved by CEP treatment. Therefore, multi-target drug use in combination with CEP is essential for the treatment of coronavirus infection.
Taken together, the results highlight the differentially expressed glycoproteins targets and related pathways in coronavirus infection and CEP intervention. Generally, coronavirus GX_P2V-induced aberrant N-glycosylation alteration in infected cells are prevalent. CEP can reduce the N-glycosylation of proteins located on cell membrane surface, indicating that CEP may achieve the protective effect against virus invasion by interfering potential N-glycoprotein targets in the affected cells. Besides, CEP can regulate the glycosylation of viral proteins S, M, and N, revealing targets or pathways relevant for viral pathogenicity. The enhancement of glycan shield on the viral proteins S and M might weaken coronavirus' binding to its recognition targets. Nevertheless, the dramatic down-regulation of N-glycosylation witnessed at the N25 site in viral N protein caused need further study.
This study provides a landscape of N-glycoproteomic pro ling using the established SARS-CoV-2 cellular model, suggesting that it is of great importance in therapeutic target screening for drug discovery. The results indicated that coronavirus can cause alterations in the glycosylation of proteins at multiple levels in infected cells, whereas CEP can partially maintain GX_P2V-induced aberrant N-glycoprotein targets as well as partially regulate the glycosylation of viral proteins, suggesting that multi-target drug combination is essential for the treatment of coronavirus infestation so as to achieve optimal therapeutic results. These aberrant N-glycosylation of GX_P2V infestation might be potential targets for combination therapy.

Protein extraction and trypsin digestion
Samples stored in -80°C refrigerator were thawed on ice and centrifuged at 12,000 g at 4°C for 10 min to remove cell debris. The supernatant was transferred to a new centrifuge tube for protein concentration determination using BCA kit.
Equal amount of each protein sample was enzymatically lysed with same volume of lysis buffer containing the appropriate amount of reference protein. The protein samples were mixed with ve times of the sample volume and precipitated with pre-cooled acetone at -20℃ for 2 h, followed by centrifugation at 4500 g for 5 min, the precipitate was collected and washed twice with pre-cooled acetone. The protein precipitate from each sample was dried, and then sonicated in buffer containing 200 mM Triethylammonium bicarbonate (TEAB). The digestion was performed by incubation with trypsin at a ratio of 1: 50 (enzyme: protein, m/m) with protein samples for overnight. Finally, the samples were desalted according to the C18 ZipTips instructions, and vacuum freeze-dried for HPLC analysis.
The enrichment of glycosylation modi cation of peptides The peptide fragments were dissolved in 40 μL of enrichment buffer (80% acetonitrile/1% tri uoroacetic acid) and transferred to a hydrophilic micro-column. The enrichment was completed by centrifugation at 4000 g for approximately 15 min in HILIC. The hydrophilic micro-column was washed for 3 times with enrichment buffer. The glycopeptides were eluted with 10% acetonitrile, then the eluate was collected and vacuum dried. After drying, the eluate was reconstituted in 50 μL of 50 mM ammonium bicarbonate buffer dissolved in 50 μL of hydrogen peroxide with 2 μL of water. The glycopeptides were incubated with PNGase F glycosidase overnight at 37 °C. Finally, salt was removed according to the C18 ZipTips instructions and vacuum freeze-dried for liquid-liquid analysis.

Liquid Chromatography-Mass Spectrometry (LC-MS)
The peptides were dissolved in liquid chromatography mobile phase A (0.1% (v/v) formic acid aqueous solution) and then separated using NanoElute ultra-e cient liquid phase system. Mobile phase A is an aqueous solution containing 0.1% formic acid; mobile phase B is an acetonitrile solution containing 0.1% formic acid. Liquid phase gradient setting: 0-50 min, 2%~22%B; 50-52 min, 22%~35%B; 52-55 min, 35%~90%B; 55-60 min, 90%B, the ow rate is maintained at 450 nL/min. The peptides were separated by an ultra-high-performance liquid phase system and injected into a capillary ion source. Then the peptide segments are analyzed by a TIMS-TOF Pro mass spectrometer. The ion source voltage was set to 1.6 kV and both the peptide parent ion and its secondary fragments were detected and analyzed using TOF. The secondary mass spectrometry scan range was set to 100-1700 m/z. The data acquisition mode used parallel cumulative serial fragmentation (PASEF) mode. One primary mass spectrometry acquisition followed by 10 PASEF mode acquisitions of secondary spectra with parent ion charge numbers in the range 0-5. The dynamic exclusion time of the tandem mass spectrometry scan was set to 30 seconds to avoid repeated scans of the parent ion.

Database Search
Retrieval parameter settings: the database was FA105LPNg_ Chlorocebus_sabaeus_60711_Ensembl_GX_P2V_protein_TX_combine_20200413 (19267 sequences), an anti-library was added to calculate the false positive rate (FDR) caused by random matching, and a common pollution library was added to the database to eliminate the contamination protein in the identi cation results Impact; the digestion method was set to Trypsin/P; the number of missed cleavage sites was set to 2; the minimum length of the peptide was set to 7 amino acid residues; the maximum modi cation number of the peptide was set to 5.
The mass error tolerance of the primary precursor ions in First search and Main search was respectively set to 20 ppm and 20 ppm, and the mass error tolerance of the secondary fragment ions was 0.02 Da. The cysteine alkylation was set as a xed modi cation, and the variable modi cation was the oxidation of methionine, the acetylation of the N-terminus of the protein, and the deamination of asparagine (18O). The FDR for protein identi cation and PSM identi cation was set to 1%.

Bioinformatics analyses
Glycoproteomic data of GX_P2V infection group and 14 glycosylation sites in 12 glycoprotein targets of CEP treatment in presence of GX_P2V l was analyzed by Metascape web-based platform 24 . Pathway and process enrichment analysis were carried out with the following ontology sources: KEGG Pathway, GO Biological Processes, Reactome Gene Sets, Canonical Pathways and CORUM. PPI networks were constructed by string (STRING; http://string-db.org. version 11.0) 25      Schematic diagram of subcellular location and PPI of differentially expressed glycoproteins in GX_P2V treated Vero E6 cells as well as the intervention targets of CEP. The subcellular location of aberrant glycoproteins was determined using UniProt. The PPI network maps of these dysregulated glycoproteins were generated using STRING. The red color highlights CEP-regulated proteins. Upon GX_P2V treatment in presence of CEP, 12 out of 69 glycoproteins dysregulated by coronavirus GX_P2V remains in normal glycosylation state, suggesting that these proteins and glycosylation sites might be the targets of CEP.   Note: Since the 3D structure of GX_P2V viral S proteins is not available, considering the high sequence