Skip to main content
NCBI home page
Acta Biochimica et Biophysica Sinica logoLink to Acta Biochimica et Biophysica Sinica
. 2024 Jul 26;56(8):1118–1129. doi: 10.3724/abbs.2024127

O-glycosylation of SARS-CoV-2 spike protein by host O-glycosyltransferase strengthens its trimeric structure

O-glycosylation strengthens the trimer of the SARS-CoV-2 S protein

Zhijue Xu 1,2, Han Zhang 1, Jiaqi Tian 1,3, Xin Ku 1, Rumeng Wei 1, Jingli Hou 4, Can Zhang 1, Fang Yang 1, Xia Zou 1, Yang Li 1, Hiroyuki Kaji 2, Sheng-Ce Tao 1, Atsushi Kuno 2,5, Wei Yan 1, Lin-Tai Da 1, Yan Zhang 1,2,*
PMCID: PMC11399440  PMID: 39066577

Abstract

Protein O-glycosylation, also known as mucin-type O-glycosylation, is one of the most abundant glycosylation in mammalian cells. It is initially catalyzed by a family of polypeptide GalNAc transferases (ppGalNAc-Ts). The trimeric spike protein (S) of SARS-CoV-2 is highly glycosylated and facilitates the virus’s entry into host cells and membrane fusion of the virus. However, the functions and relationship between host ppGalNAc-Ts and O-glycosylation on the S protein remain unclear. Herein, we identify 15 O-glycosites and 10 distinct O-glycan structures on the S protein using an HCD-product-dependent triggered ETD mass spectrometric analysis. We observe that the isoenzyme T6 of ppGalNAc-Ts (ppGalNAc-T6) exhibits high O-glycosylation activity for the S protein, as demonstrated by an on-chip catalytic assay. Overexpression of ppGalNAc-T6 in HEK293 cells significantly enhances the O-glycosylation level of the S protein, not only by adding new O-glycosites but also by increasing O-glycan heterogeneity. Molecular dynamics simulations reveal that O-glycosylation on the protomer-interface regions, modified by ppGalNAc-T6, potentially stabilizes the trimeric S protein structure by establishing hydrogen bonds and non-polar interactions between adjacent protomers. Furthermore, mutation frequency analysis indicates that most O-glycosites of the S protein are conserved during the evolution of SARS-CoV-2 variants. Taken together, our finding demonstrate that host O-glycosyltransferases dynamically regulate the O-glycosylation of the S protein, which may influence the trimeric structural stability of the protein. This work provides structural insights into the functional role of specific host O-glycosyltransferases in regulating the O-glycosylation of viral envelope proteins.

Keywords: O-glycosylation, SARS-CoV-2, spike protein, glycosyltransferase, ppGalNAc-T

Introduction

Protein O-linked glycosylation, also known as mucin-type O-glycosylation, is one of the most abundant glycosylation in mammalian cells. More than 80% of secreted and cell membrane proteins undergo O-glycosylation [1]. This type of glycosylation plays crucial roles in modulating various functions, including ligand-receptor interactions, proprotein processing, and subcellular localization sorting [ 2, 3]. The initiation of protein O-glycosylation is facilitated by a glycosyltransferase family known as polypeptide N-acetylgalactosaminyltransferases (ppGalNAc-Ts), which transfer GalNAc from UDP-GalNAc to the Thr/Ser residues of substrate proteins or peptides [4]. In human cells, the ppGalNAc-T family contains up to 20 isoenzymes, exhibiting distinct spatial and temporal expression patterns in tissues and cells. Furthermore, these ppGalNAc-T isoenzymes share both redundant and partially specific preferences for protein substrates [ 5, 6]. Therefore, protein O-glycosylation is a kind of systematic modification in human cells.

Viruses, as simple life forms, exploit the host’s transcription, translation, and modification systems to synthesize and package their required proteins and nucleic acids [7]. Consequently, viral proteins carry post-translational modifications that are homologous to those of the host cells, such as glycosylation. Many envelope proteins of pathogenic viruses, including HIV, MERS, and SARS, are heavily glycosylated [8]. Viruses utilize host glycosylations to evade host immunity, bind to host receptor cells, and regulate hydrolysis by host proteases [9]. Therefore, glycosylation plays a crucial role in interactions between viruses and their hosts.

Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), which is responsible for the novel coronavirus disease 2019 (COVID-19), posed a pandemic threat to global public health during 2019–2023 [10]. Currently, both long COVID-19 and sporadic COVID-19 cases continue to burden global health systems [11]. The spike protein (S protein) of SARS-CoV-2 facilitates membrane fusion between the virus and host cells [12]. This type I membrane protein is a homotrimer [13]. It has been reported that two trimeric S proteins can simultaneously bind to one angiotensin converting enzyme 2 (ACE2) homodimer from host cells, and then mediate the membrane fusion process [14]. Each monomer of the S protein is coated with 22 N-glycan sites and numerous O-glycan sites in diverse expression systems [15]. N-glycosylation of the S protein has been extensively studied and found to directly affect interactions with ACE2 [16]. For example, deletions of the N-glycosites at N331 and N343 significantly reduce viral infectivity [17]. Additionally, O-glycosylation is also essential for SARS-CoV-2 S protein function. Blocking O-glycan elongation on the S protein can reduce viral entry by 35%‒75% [18]. The O-glycans at T676 and T678 near the furin cleavage site of the S protein inhibit the formation of virus-infected syncytia [ 19, 20]. Although a large number of O-glycosites on the S protein have been characterized using mass spectrometry (MS) analysis [ 21‒ 23], the relationship between host ppGalNAc-Ts and S protein O-glycosylation remains unclear.

In this study, we employed a higher-energy collisional dissociation (HCD) product dependent triggered electron-transfer dissociation (ETD), i.e., HCD-pd-ETD MS approach to characterize the O-glycan sites and structures across the complete S protein. A total of 15 O-glycosites were identified, with the glycan structures for each site are determined or inferred. Using a peptide array of the S protein as the acceptor substrate, we detected the catalytic activities of several ppGalNAc-T isoenzymes, including T1, T2, T3, and T6, for the S protein. Moreover, the upregulation of ppGaNAc-T6 in cells resulted in pronounced alterations in the O-glycosylation patterns of the S protein. Molecular dynamics (MD) simulations provided insights into the potential functional roles of O-glycosylation on the S protein induced by ppGalNAc-T6 in cells.

Materials and Methods

Materials

The recombinant SARS-CoV-2 S ectodomain (FLAG tag) was constructed in-house and expressed in human embryonic kidney 293T (HEK-293T) cells. The SARS-CoV-2 S protein subunit 1 (His tag) expressed by HEK-293 cells was purchased from Sanyoubio (Shanghai, China). The essential reagents including dithiothreitol (DTT), iodoacetamide (IAA), formic acid (FA), trifluoroacetic acid (TFA), acetonitrile (ACN), ammonium bicarbonate (NH 4HCO 3), anti-FALG antibody, anti-FLAG M2 affinity gel, and neuraminidase were purchased from Sigma-Aldrich (St Louis, USA). Sequencing-grade trypsin was purchased from Enzyme & Spectrum (Beijing, China). C18 Tips were purchased from Thermo Fisher Scientific (Waltham, USA). Enzymes such as PNGase F and O-glycosidase were purchased from NEB Labs (Ipswich, USA).

Cell culture

Human embryonic kidney 293T (HEK-293T) cells were obtained from the Chinese National Human Genome Center (Shanghai, China). HEK-293T cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (Cytiva, Marlborough, USA), supplemented with 10% heat-inactivated fetal bovine serum (Excell, Shanghai, China) and maintained in a 5% CO 2 atmosphere at 37°C.

Protein expression and purification

To express the recombinant ectodomain of the S protein (S-ECD), a mammalian-codon-optimized gene encoding SRAS-CoV-2 S (GenBank: QHD43416.1) residues 1–1208, with proline substitutions at residues 986 and 987 [24], as well as a “GAAG” substitution at the furin cleavage site (residues 682–685), was synthesized and cloned into the mammalian expression vector pCMV-3×FLAG-14. This expression vector was then used to transiently transfect HEK-293T cells using polyethylenimine. The C-terminal 3×FLAG-tagged S-ECD protein (S-ECD-3×FLAG) was subsequently purified from the cell medium or whole-cell lysate using an anti-FLAG M2 affinity gel, followed by competitive elution with 3×FLAG peptide.

Glycosidase treatment

The purified S-ECD-3×FLAG proteins (50 ng) were subjected to treatment with PNGase F, neuraminidase and/or O-glycosidase according to the manufacturer’s protocol. Briefly, the proteins were denatured in 1× Glycoprotein Denaturing Buffer at 98°C for 10 min, followed by the addition of GlycoBuffer 2 and NP-40. The proteins were then incubated with PNGase F, neuraminidase, and/or O-glycosidase at 37°C for 2 h. Subsequently, the mixture was analyzed via western blotting using anti-FLAG antibodies.

O-glycoprotease treatment

The purified S-ECD-3×FLAG proteins (50 ng) were treated with neuraminidase according to the manufacturer’s protocol as described above. Subsequently, 200 ng of home-purified O-glycoproteases, including BT4244, StcE, OgpA, IMPa and AM0627, was added individually and then incubated at 37°C for 2 h. An equal volume of PBS was added to serve as the vehicle control. The mixture was then analyzed via western blotting using anti-FLAG antibodies.

Protein digestion

The N-glycans of the S1 and S-ECD-3×FLAG proteins were initially removed by PNGase F. Subsequently, the proteins were separated via SDS-PAGE, and the gel was stained with Coomassie blue. The bands corresponding to S1 (70 to 110 kDa) and S-ECD-3×FLAG (130 to 170 kDa) were then excised into approximately 1 mm 3 cubes. The gel pieces were destained with 50% ACN in 50 mM NH 4HCO 3 buffer at room temperature (RT) for 30 min, followed by dehydration with 100% ACN. The proteins in the gel were reduced with 10 mM DTT at 56°C for 45 min, followed by dehydrating with 100% ACN. The proteins in the gel were then alkylated with 60 mM IAA at RT for 45 min in the dark. Next, the proteins were digested by trypsin in 50 mM NH 4HCO 3 for 18 h at 37°C. The peptides were extracted from the gel by the addition of 50% ACN containing 5% formic acid, and the released peptides were desalted using C18 Tips and then speed-dried.

Mass spectrometry analysis

LC-MS/MS analysis was conducted by coupling a nanoLC (Dionex Ultimate 3000; Thermo Fisher Scientific) to an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific). Each sample was dissolved in 0.1% formic acid and injected onto an analytical column (inspire C18, 3 μm; Dikma, Toronto, Canada, 150 mm×75 μm, self-packed). Separation was achieved using a gradient at a flow rate of 0.3 μL/min: 0‒3 min, 1%‒4.8% solvent B; 3‒38 min, 4.8%‒16% solvent B; 37‒47 min, 16%‒28% solvent B; 47‒48 min, 28%‒72% solvent B; 48‒60 min, 72% solvent B; 60‒70 min, 90% solvent B; and 70‒80 min, equilibration with 1% solvent B [solvent A, 100% H 2O; solvent B, 100% acetonitrile; both containing 0.1% (v/v) formic acid].

For tandem MS analysis, an “HCD triggered subsequent ETD scan” strategy was employed for most runs. Initially, a precursor MS1 scan (m/z 300–1800) for intact peptides was acquired in the Orbitrap at 60,000 resolution, followed by MS2 scans of the 20 most abundant multiply charged precursors in the MS1 spectrum (using the “top 20” method) with HCD fragmentation (m/z 100–2000, resolution 30,000) in the Orbitrap mass analyzer. Subsequent ETD-based MS2 acquisition for the same precursor was triggered (m/z 150–2000, resolution 30,000) and acquired in the Orbitrap when a HexNAc fragment at either m/z 204.0867, 186.0766 or 138.0545 was detected in the previous HCD-MS2 spectrum. The normalized collision energy for HCD was set to 30%, and ETD fragmentation was performed with an automated calibrated charge-dependent reaction time supplemented by 15% HCD activation. The automatic gain control targets were set to 400,000 ions for MS1 and 50,000 ions for MS2 scans. Dynamic exclusion of 8 s was enabled to prevent repeated acquisition of the same precursor.

Mass spectrometric data analysis

The LC-MS/MS raw data were manually analyzed for each glycopeptide using Xcalibur 3.0.63 software (Thermo Fisher Scientific). The spectra were classified as O-glycopeptides only when the saccharide oxonium ions of HexNAc containing the pair of ions at m/z 204.087 and 186.076 were detected. The identification of O-glycosites and O-glycan structures was performed manually. The mass tolerances for the precursors and fragment ions were set to 10 ppm and 0.02 Da, respectively. The glycosites and glycoforms of the S-ECD protein purified from HEK-293T cells were validated based on the mass spectrometry results of commercial S1 using manual “match between runs” analysis. Specifically, precursor ions (±10 ppm) with a similar retention time (±5 min) in the mass spectrometry results of commercial S1 or S protein purified from HEK-293T cells were considered as identical glycopeptides with the same O-glycosites and O-glycans (refer to Supplementary Table S1). Quantification of each O-glycan at each site was determined by using the top intensity of the corresponding MS1 peaks. Normalization across different samples was performed using the total amount of S protein in each sample, which was achieved by using the protein label-free quantification module in MaxQuant v2.2.0.

On-chip ppGalNAc-Ts assay

The peptide microarray of the S protein was constructed in-house as described previously [25]. The on-chip ppGalNAc-Ts assay was modified from previously reported methods [25]. Briefly, the microarrays were incubated with a 5% w/v BSA/TBS solution (50 mM Tris and 150 mM NaCl, pH 7.4) for 1 h to block nonspecific binding. Subsequently, the blocked microarrays were incubated overnight at 37°C with a ppGalNAc-Ts reaction mixture containing 10 ng/μL ppGalNAc-Ts, 25 mM Tris-HCl (pH 7.4), 5 mM MnCl 2, 0.2% v/v Triton X-100, and 0.5 mM UDP-GalNAc in 5% w/v BSA. Following the incubation, the microarrays were washed three times (10 min each) with wash buffer (TBST containing 200 mM NaCl, 0.1% v/v Tween 20, and 0.2% w/v SDS). Subsequently, the microarrays were incubated with 1 μg/mL biotinylated Vicia villosa agglutinin (VVA) lectin (Vector Labs, Newark, USA) for 1 h at room temperature, followed by another round of washing (three times, 10 min each) with wash buffer. Next, the microarrays were incubated with 1 μg/mL Cy5-labeled streptavidin, followed by a final round of washing (three times, 10 min each) with wash buffer. The microarrays were then scanned at Ex 532 nm using a GenePix 4200A slide scanner (Molecular Devices, San Jose, USA). The foreground and background intensities were extracted from the microarray images using GenePix Pro 6.0 software. The signal-to-noise (S/N) ratio of each protein was calculated as the average of triplicate spots.

3D modeling and molecular dynamics simulations

For each of the six protomer-interface O-glycosites, including T1027, T912, T768, S316, T315, and T572, we constructed the Tn-modified S-protein based on the one-crystal structure of the S-protein trimer (pdb ID: 6xr8) [13]. Additionally, we constructed a T structure modified S-protein at the T912 residue. Specifically, we synthesized the GalNAc saccharide and Gal-GalNAc disaccharide utilizing the resources available on the GLYCAM-Web site ( www.glycam.org). Subsequently, the hydroxyl group at the C1 position of GalNAc was removed in PyMOL, and then the GalNAc was linked to the corresponding Thr/Ser residues using the ‘bound’ command. We then solvated each O-glycosylated S-trimer in a cubic box filled with TIP3P water. Three Na + ions were added to neutralize the whole system. The final system was then subjected to energy minimization to relieve local steric clashes, followed by a heating simulation from 0 to 310 K within 50 ps. Next, we performed 500 ps equilibrium MD simulations at 310 K by constraining all solute heavy atoms. Finally, we conducted 20-ns molecular dynamics (MD) simulations for each system by keeping the temperature at 310 K, controlled by the Langevin thermostat [26]. All the MD simulations were performed using the Amber14 package [27]. The ff14SB and GLYCAM06j-1 force fields were employed to describe the protein and GalNAc, respectively [ 28, 29]. The SHAKE algorithm was used to constrain the bond lengths involving hydrogen atoms [30]. The non-bonded cutoff distance was set as 10 Å, and the long-range electrostatic interaction was calculated using the particle mesh Ewald (PME) method [31]. As a control, we also conducted a 20-ns MD simulations for the un-glycosylated S-protein trimer.

Mutation frequency analysis

The mutation data for SARS-CoV-2 variants were obtained from the Global Initiative on Sharing All Influenza Data (GISAID, https://www.gisaid.org), including 16,662,958 sequences (updated to 24/04/20). Mutations at each O-glycosite on the S protein were searched and collected. The frequency of mutations was calculated by comparing them to the total number of 16,662,958 variants.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism 5.0 (GraphPad software, San Diego, USA). The significance of differences between two groups was determined using Student’s t-tests (two-tailed). A P value less than 0.05 was considered statistically significant.

Results

SARS-CoV-2 S protein is highly O-glycosylated in human HEK-293T cells

The recombinant S-ECD was expressed in human embryonic kidney 293T (HEK-293T) cells ( Figure 1A). To prevent the furin cleavage of the S protein at the 682RRAR 685 site, we introduced a “GAAG” substitution at the above cleavage site and additionally designed two proline mutations at Lys986 and Val987 to enhance the stability of the trimeric S protein, as suggested by previous studies [24]. Purification of the S protein from culture medium and cell lysates was achieved via immunoprecipitation with an anti-FLAG antibody, and the protein purity was confirmed by silver staining ( Figure 1B). The S protein purified from culture medium exhibited smeared bands, while the protein derived from cell lysate displayed a sharp band, indicating high glycosylation of the secreted S protein. To characterize the nature of the glycans, we treated the S protein with either the N-glycosidase PNGase F or neuraminidase. Both treatments resulted in a significant decrease in protein molecular weight ( Figure 1C), confirming that the S protein was coated with N-glycans. Notably, further treatment with O-glycosidase, together with PNGase F and neuraminidase, led to an additional decrease in the molecular weight of the S protein compared to treatment with PNGase F or neuraminidase alone ( Figure 1C). These findings indicate that the S protein expressed in HEK-293T cells undergoes both N- and O-glycosylation.

Figure 1 .

Figure 1

Open in a new tab

Expression and purification of the SARS-CoV-2 S protein

(A) Schematic diagram of the SARS-CoV-2 S protein and recombinant S-ECD-3×FLAG protein. The protein domains are illustrated: SS, signal sequence; NTD, N-terminal domain; RBD, receptor-binding domain; FP, fusion peptide; HR1, heptad repeat 1; CH, central helix; CD, connector domain; HR2, heptad repeat 2; and TM, transmembrane domain. In the recombinant S-ECD-3×FLAG protein, “GAAG” was introduced to replace the cleavage site “RRAR”, and the K986-V987 motif was substituted with “PP”. (B) Silver staining of purified S-ECD-3×FLAG protein. The S-ECD-3×FLAG protein was expressed in human embryonic kidney 293T cells and purified from culture medium or whole cell lysate (WCL) using anti-FLAG beads. (C) Glycosidase treatment of purified S-ECD-3×FLAG protein. Purified S-ECD-3×FLAG proteins (50 ng) were treated with PNGase F, neuraminidase and/or O-glycosidase. After treatment, the proteins were separated by SDS-PAGE and immunoblotted with an anti-FLAG antibody. (D) O-glycoprotease treatments of the purified S-ECD-3×FLAG protein. All the samples were first subjected to neuraminidase treatment. Data are shown as the mean±SEM (n=3). *P<0.05, n.s., not significant.

Many previous studies reported that the S protein was O-glycosylated with low occupancy [ 16, 23, 32]. To further determine the percentage of O-glycosylated S-ECD in our expression system, we utilized O-glycoproteases to digest the O-glycosylated S-ECD and detected the remaining un-glycosylated S-ECD protein via anti-FLAG immunoblotting ( Figure 1D). The mucin-selective proteases BT4244, StcE, OgpA, IMPa and AM0627 were used ( Supplementary Figure S1). Following treatment, less than 82%, 70%, 78%, 73%, and 94% of S-ECD remained in the BT4244, StcE, OgpA, IMPa and AM0627 groups, respectively, with significant reductions observed in the BT4244, StcE, and OgpA groups compared to the vehicle group ( Figure 1D). These findings suggested that more than 20%‒30% of the S-ECD protein was O-glycosylated in our expression systems. Taken together, these results indicate that the S-ECD expressed in HEK-293T cells is O-glycosylated with more than 20%‒30% occupancy.

Site-specific analysis of O-glycosylation of the SARS-CoV-2 S protein

To determine the O-glycosylation pattern on the S protein, we analyzed the O-glycosites and the corresponding glycan structures using a sequential fragmentation MS strategy (HCD triggered ETD). After removing the N-glycans ( Figure 2A), 7 high-confidence O-glycosites and 10 O-glycan structures were first identified from a commercialized recombinant S1 subunit used as a standard sample ( Supplementary Figures S2‒S4). Using this analysis method, we identified a total of nine O-glycosites from the purified S protein expressed in HEK-293T cells, including five Thr-sites (T315, T323, T638, T912, and T1027), three Ser-sitse (S31, S316, and S640), and one ambiguous site within the S939‒S943 motif ( Figure 2B). Based on the MS/MS analysis, we detected five O-glycan structures. Among these, GalNAc O-glycan (Tn) and GalNAcGal O-glycan (T) structures were the predominant O-glycan types on the S protein, while sialylated T structure (ST), core-2 type GalNAcGal (GlcNAc) O-glycan (C2), or core-2 type GalNAcGal(GlcNAcGal) O-glycan (C2-2) were mostly detected at T323 ( Figure 2B,C and Supplementary Figure S5).

Figure 2 .

Figure 2

Open in a new tab

Site-specific O-glycosylation analysis of the S protein

(A) Scheme of O-glycosylation analysis of the SARS-CoV-2 S protein. (B) Identified O-glycosites and the corresponding O-glycan structures on the S-ECD-3×FLAG protein. Monosaccharide symbols follow the Symbol Nomenclature for Glycans (SNFG) system. a, an uncertain site at S939, S940, T941, or S943. (C) Representative HCD and ETD MS/MS spectra of the O-glycopeptide with a Tn structure at the T323 residue.

An on-chip ppGalNAc-Ts assay on an S protein peptide microarray

To explore the unique role of glycosyltransferases in the O-glycosylation of the S protein, a 12-aa length peptides construct array including 211 peptides covering the full-length S protein was used to investigate the glycosyltransferase activities against the S protein ( Figure 3A). A previous study showed that ppGalNAc-T1, -T2, -T3, -T6, -T10, -T13, and -T16 could glycosylate the S protein in human cells [19]. For this study, we selected ppGalNAc-T1, -T2, -T3, and -T6 isoenzymes for analysis on a chip. ppGalNAc-T10, which exhibits specific activity toward GalNAc-glycopeptides and low activity toward unmodified peptides, was excluded [33]. Similarly, ppGalNAc-T13 and -T1 belong to the same subfamily and exhibit similar substrate activities [34]. ppGalNAc-T16 and -T2 also belong to the same subfamily, but ppGalNAc-T2 exhibits higher extent of glycosylation and wider substrate selection than ppGalNAc-T16 [35], leading to their exclusion from analysis. The peptides modified with O-GalNAc glycosylation produced by each ppGalNAc-Ts were detected using the lectin of VVA. Peptides with a signal-to-noise ratio greater than 5.0 were considered positive substrates for each ppGalNAc-T ( Figure 3B). In total, 22 peptides could be catalyzed by these ppGalNAc-Ts, and the most 15 peptides of them were glycosylated by ppGalNAc-T6 ( Figure 3C). These results clearly show that ppGalNAc-T6 exhibits higher enzymatic activity against a broader set of S peptides compared to the other three isoenzymes. Furthermore, ppGalNAc-T1 and ppGalNAc-T2 prefer to catalyze the flexible regions connecting functional domains, whereas ppGalNAc-T6 can also target structured units such as the N-terminal domain (NTD) and connector domain (CD). This microarray assay suggests that ppGalNAc-T6 is the major ppGalNAc-T isoenzyme involved in the O-glycosylation of the S protein.

Figure 3 .

Figure 3

Open in a new tab

On-chip ppGalNAc-Ts assay on the S protein peptide microarray(A) Scheme of the on-chip ppGalNAc-Ts assay. The peptide microarray contains short peptides covering the full length of the S proteins. ppGalNAc-T transfers GalNAc from UDP-GalNAc to the targeted Ser/Thr residues, and the resulting O-GalNAc glycan is immobilized with biotin-labeled VVA lectin and Cy5-labeled streptavidin. The signal was analyzed using a microarray scanner. (B) Overview of S protein peptide glycosylation modified by T1/T2/T3/T6 and (C) heatmap of glycopeptides with a signal-to-noise (S/N) intensity greater than 5.0.

Overexpression of ppGalNAc-T6 dramatically alters the O-glycan patterns of the S protein

To further investigate the potential impact of ppGalNAc-T6 on the S protein, we overexpressed ppGalNAc-T6 in HEK-293T cells and purified the S protein to analyze the resultant O-glycan features. As shown in Figure 4A,B, overexpression of ppGalNAc-T6 led to a profound alteration in the O-glycosylation pattern of the S protein compared to that observed in WT HEK-293T cells by increasing O-glycosites, diversities of O-glycans and elevated levels of O-glycosylation. Specifically, five additional O-glycosites were observed upon ppGalNAc-T6 overexpression, including three Thr-sites (T29, T307 and T768), one Ser site (S459), and one uncertain site at either T572 or T573. In addition, more diversified O-glycan structures were observed at specific sites, including additional T structures at S31 and T912, ST structures at T315, S316, and T638, and C2 structures at T638 residue ( Figure 4A and Supplementary Figure S6). More intriguingly, we noted that the O-glycosylation levels at the S31, T315, T638, T912, and T1027 residues were increased significantly following ppGalNAc-T6 overexpression, while the levels at S316, T323, S640, and one uncertain site within the S939-S943 motif remained unchanged ( Figure 4C). This finding aligns with the results of our on-chip assays showing that ppGalNA-T6 preferred to catalyze peptides containing T29, S31, T307, and T912 ( Figure 3C). Collectively, these findings indicate that upregulation of ppGalNAc-T6 significantly alters the O-glycosylation patterns of the S protein.

Figure 4 .

Figure 4

Open in a new tab

Site-specific O-glycosylation of the S protein with overexpressed ppGalNAc-T6

(A) Identified O-glycosites and O-glycans on S-ECD-3×FLAG from the ppGalNAc-T6-overexpressing HEK-293T cells. Newly discovered O-glycosites in ppGalNAc-T6-overexpressing cells are highlighted in orange. Representative MS/MS spectra are shown in Supplementary Figure S6. Twenty-two known N-glycosites were also labeled. (B) Heatmap of the intensities of the total O-glycopeptides at each O-glycosite in the S-ECD-3×FLAG from the WT and ppGalNAc-T6-overexpressing cells. (C) Quantitative comparisons of the O-glycosylation levels of the O-glycosites detected on S-ECD-3×FLAG in both WT and ppGalNAc-T6-overexpressing cells. Data are shown as the mean±SEM (n= 3). *P<0.05, n.s., not significant, n.a., not available.

The enhanced O-glycosylation of the S protein induced by ppGalNAc-T6 stabilizes its trimeric conformation

As described above, we identified a total of 15 O-glycosites on the S protein, including 14 sites on the S-ECD ( Figure4) and an additional site at S325 on the S1 subunit ( Supplementary Figure S2). By locating these sites onto the trimeric structure of the S-protein (pdb ID: 6xr8), nine of them were found to be located on the S protein surface, and the other six are located at the interfaces between the protomers ( Figure 5A,B). For the six interface-glycosites, the Tn structure is the most abundant O-glycan. To examine how the interface O-glycosites affect the structure of the S protein trimer, we conducted MD simulations for each of the six interface-glycosites, both with and without Tn modification. Intriguingly, we found that O-glycosylation at T1027, T912, T768, T572 and S316 profoundly promotes the formation of the S protein trimer by strengthening the inter-protomer interactions ( Figure 5C). Specifically, for the T1027-site, the Tn groups from the three protomers can form stable non-polar contacts through their -CH3 groups and can also establish one hydrogen bond (HB) with the R1039 sidechain from the neighboring protomer ( Figure 5D and Supplementary Figure S7). These interactions contribute to stabilizing the S-trimer conformation, reflected by the shortened distance between the T1027 Cα atoms from each pair of adjacent protomers ( Figure 5D). Similarly, the GalNAc group at the T912 site can directly contact F1121 from the adjacent protomer by stacking interactions, which also strengthens the S-trimer conformation ( Figure 5E).

Figure 5 .

Figure 5

Open in a new tab

3D modeling and molecular dynamics simulations of O-glycosylation on the S protein

(A) A schematic illustration of the S-protein, with 15 identified O-glycosites labeled on top. In particular, the O-glycosites located on the S protein surface and at the protomer interfaces are colored green and orange, respectively. (B) Structural representation of the S protein trimer from two different views. Protomers A, B, and C are shown in blue, gray, and violet, respectively. The modeled Tn structure on protomer A is highlighted as spheres, which are colored differently according to the site. (C) A zoomed-in view of six O-glycosites located at the protomer interfaces. (D) Structural highlight of the T1027-Tn from three protomers, with the CH3 group of Tn shown in spheres. Inter-Tn:CH3 distance (middle panel) and T1027 Cα atom distance (right panel) between each pair of S-protomers. (E) Structural highlights of T912-Tn from three protomers. F1121 is shown as colored sticks. Distances between the center of mass (COM) of GalNAc and the COM of the F1121 sidechain from the neighboring protomer (middle panel). The T912 Cα atom distance (right panel) between each pair of S-protomers. (F) Structural highlight of the T768-Tn in protomer A; Q314 from protomer C is shown as violet sticks. Comparisons of the inter-protomer contact numbers with and without Tn modifications (middle panel), where the contact number is calculated by the sum of each two protomer contacts surrounding the Tn structure with a distance cutoff of 6 Å; HB occupancy between Tn and Q314 from the neighboring protomer (right panel). (G) Structural highlight of T572-Tn in protomer A; N856 from protomer B is shown as gray sticks. Comparisons of the inter-protomer contact numbers with and without Tn modifications (middle panel). HB occupancy between T572-Tn and N856 from the neighboring protomer (right panel). Each HB is highlighted by a black dashed line.

In addition, for the T768, T572, and S316 sites, the presence of the Tn structure significantly increases the contact number between two adjacent protomers compared to the non-glycosylated conformations ( Figure 5F,G and Supplementary Figure S7A). Moreover, T768-Tn and T572-Tn can interact with Q314 and N856 from different protomers, respectively, via HB interactions ( Figure 5F,G). The GalNAc-group on S316-Tn, on the other hand, can form a water-mediated interaction with T761 from neighboring protomers ( Supplementary Figure S7A). In contrast, the O-glycosylation of T315 imposes no apparent structural perturbations on the S protein trimer ( Supplementary Figure S7B). Overall, most of the interface O-glycosites play critical roles in stabilizing the S protein trimer structure by forming hydrophobic and/or HB interactions with other protomers. Notably, the O-glycosylation levels of four interface O-glycosites, T1027, T912, T768, and T572, are greatly enhanced by the upregulation of ppGalNAc-T6 ( Figure 4C), again highlighting the significant role of ppGalNAc-T6 in modulating the trimeric structure of the S protein.

Analysis of the mutation frequency of O-glycosites on the S protein

To gain further insight into the mutation frequency at O-glycosites on the S protein, we summarized data on unambiguous O-glycosites identified via MS analysis from various workflows using recombinant full-length or truncation constructs of the S protein expressed in human cells or insect cells. More than 27% of the S/T residues on the S protein were detected with O-glycosylation. Specifically, 37 O-glycosites are located within the S1 subunit, and 17 are located within the S2 subunit ( Figure 6A). We then searched for mutations at these 54 O-glycosites in public available SARS-CoV-2 variants from the Global Initiative on Sharing All Influenza Data (GISAID, https://www.gisaid.org). We investigated the virus details of each variant of concern (VOC) containing mutations in both the N- and O-glycosylated sites and calculated the corresponding mutation frequencies ( Figure 6B). Notably, N-glycosites appear to be more conserved in the evolution of viruses than O-glycosites. Of the 54 glycosylation sites, 50 exhibit a mutation rate of less than 1% of the total variants, including all 15 O-glycosites identified in our study ( Figure 6C). These findings suggest that the majority of O-glycosites on the S protein have remained conserved during SARS-CoV-2 evolution.

Figure 6 .

Figure 6

Open in a new tab

Mutations of the S protein O-glycosites among SARS-CoV-2 strains

(A) Summary of O-glycosites identified on the S protein in reported studies. (B) Comparison of the mutation ratios of the O-glycosites and N-glycosites in SARS-CoV-2 alpha, beta, gamma, delta, and omicron strains. Data were sourced from the Global Initiative on Sharing All Influenza Data (GISAID, https://www.gisaid.org). (C) Mutation ratios of all 54 O-glycosites on the S protein in all SARS-CoV-2 strains in GISAID. The O-glycosites identified in our study are labeled in red.

Discussion

In this study, we confirmed that at least 20%–30% of the S protein expressed in HEK-293T cells is O-glycosylated via O-glycoprotease treatment. By employing MS analysis in HCD-pd-ETD mode, we identified a total of 15 O-glycosites and 10 distinct O-glycan structures on the S protein from HEK-293T cells. Among ppGalNAc-T1, -T2, -T3, and -T6, ppGalNAc-T6 exhibited significant activity toward the S protein in a peptide array. Overexpression of ppGalNAc-T6 in cells dramatically altered the O-glycosylation patterns of the S protein, affecting glycosites, O-glycan structures, and O-glycosylation levels. Further MD simulations suggested that the interface O-glycosylation between adjacent protomers could strengthen the S protein trimer structure by establishing nonpolar contacts or HB networks. Moreover, the mutation rates of the identified O-glycosites were conserved during the evolution of SARS-CoV-2.

O-glycosylation in human is controlled by up to 20 ppGalNAc-Ts. These ppGalNAc-T isoenzymes exhibit shear redundancy but a partially specific substrate sequence preference [25]. Our on-chip ppGalNAc-T assays revealed that ppGalNAc-T1, -T2, -T3, and -T6 exhibit distinct preferences for different regions of the S protein. For instance, the peptide R319-to-P330 on the array, representing the most abundant O-glycosite at T323, is favored by ppGalNAc-T1 and -T2 but not by -T3 and -T6. This finding is consistent with the lack of changes in the O-glycosylation of T323 in cells overexpressing ppGalNAc-T6. Furthermore, ppGalNAc-T6 overexpression significantly enhanced O-glycosylation at sites such as T29, S31, T307, and T912, all of which were preferred by ppGalNAc-T6 but not by -T1 and -T2 on the peptide array. Thus, alterations in the ppGalNAc-Ts expression pattern in host cells could profoundly impact O-glycosylation on the S protein of viruses. ppGalNAc-T6 is selectively expressed in human tissues, with high expression levels in the salivary gland, stomach, and breast and low expression levels in the liver, testis and muscle according to the Human Protein Atlas database ( Supplementary Figure S8) [36]. pGalNAc-T6 exhibits low expression in normal lung tissues, and its expression increased significantly, by more than 30-fold, when mouse lungs were infected with SARS-CoV-2 [37]. This finding suggested that SARS-CoV-2 infection induces the expression of ppGalNAc-T6 in lung cells. Proteomic and single-cell transcriptome sequencing studies of lung tissues from SARS-CoV-2-infected patients revealed alterations in the expression patterns of ppGalNAc-Ts in response to SARS-CoV-2 infection [ 38, 39]. SARS-CoV-2 infection has been shown to remodel the host cellular RNA-bound proteome to facilitate the virus life cycle [40]. These findings suggest that viruses such as SARS-CoV-2 may also modulate the protein O-glycosylation process in host cells to obtain an optimal O-glycosylation pattern for their envelope proteins.

Moreover, the ppGalNAc-T6 overexpression gives rise to increase the Tn structure O-glycosylation levels at T1027, T912, T768, and T572, which are located at protomer interfaces, thereby strengthening the trimerization of the S protein through stable interactions with adjacent protomers, as revealed by MD simulations. In addition to the Tn structure, the T structure at the T912 residue exhibits stronger steric hindrance effects compared to E1098-Q1100 of the same protomer ( Supplementary Figure S7D), indicating that the Tn structure, rather than the T structure, of the O-glycan stabilizes the trimerization of the S protein. Notably, a well-folded trimeric S protein plays crucial roles in SARS-CoV-2 infection [41]. Hence, SARS-CoV-2 infection might exploit host T6 to dynamically upregulate the O-glycosylation level of the S protein, thereby stabilizing the trimeric S protein structure and affecting viral infection. Another MD study indicated that longer O-glycan on S494 located in the RBD domain facilitates RBD‒ACE2 interaction and demonstrated that the truncated O-glycans have no impact on the affinity of the virus for ACE2 but reduce viral entry [42]. Additionally, O-glycans regulate the dissociation of the S1 and S2 subunits by inhibiting furin protease hydrolysis to decrease host membrane fusion [19]. Combined with our results, it is speculated that O-glycans on the S protein could reduce viral infection by stabilizing trimers, affecting ACE2 interactions, and protecting the S protein from dissociation or hydrolysis.

Numerous MS-based studies have elucidated the O-glycosylation of the S protein [ 21– 23]. Interestingly, the O-glycosites identified in these studies exhibit a low repeatability rate. Only the most abundant O-glycosite at T323 was identified by all studies, including our own work ( Supplementary Figure S9A). The O-glycan structures identified in these studies were also different, as only the Tn and T structure O-glycans were detected ( Supplementary Table S2). This may be attributed to variations in the expression systems used, such as insect cells, Drosophila cells, and human 293T cells ( Supplementary Figure S9B). The expression patterns and levels of ppGalNAc-T isoenzymes differ among these cells and significantly influence S protein O-glycosylation. Additionally, MS-based studies have revealed a low proportion of O-glycosylation on the S protein [15]. However, using O-glycoproteases such as StcE, we discovered that at least 20%–30% of the S protein in our expression system was O-glycosylated. In addition to the different expression systems potentially causing variations in O-glycosylation proportions, differences in the ionization efficiency of glycopeptides versus that of naked peptides in mass spectrometry might also lead to quantitative biases [43]. The O-glycoprotease of StcE has a specific S/T*–X–S/T cleavage motif, where cleavage occurs before the second serine or threonine, and X represents any amino acid [44]. The amino acid flanking the T323 residue of the S protein is QPT*ES, which fits the cleavage motif of StcE, rendering the S protein suitable for cleavage. Of course, O-glycoproteases are recently discovered glycosylation tools, and the hydrolysis specificities for O-glycosylation require further exploration.

In summary, our study revealed that the O-glycosylation sites of the S protein remain relatively conserved during SARS-CoV-2 evolution. Dynamic changes in the expression levels of host ppGalNAc-Ts can significantly alter the O-glycosylation sites, structural diversity, and glycosylation levels of the S protein. O-glycosylation at the interface of the S protein trimer promotes trimer stability, indicating that O-glycosylation potentially plays a role in the virus invasion process into host cells.

Supplementary Data

Supplementary data is available at Acta Biochimica et Biophysica Sinica online.

COMPETING INTERESTS

The authors declare that they have no conflict of interest.

Funding Statement

This work was supported by the grants from Shanghai Pilot Program for Basic Research-Shanghai Jiao Tong University (No. 21TQ1400210), the National Natural Science Foundation of China (Nos. 32371332, 32071271, and 22007065), and the Fundamental Research Funds for the Central Universities (No. KLSB2023KF-04).

References

  • 1.Yang W, Tian E, Chernish A, McCluggage P, Dalal K, Lara A, Ten Hagen KG, et al. Quantitative mapping of the in vivo O-GalNAc glycoproteome in mouse tissues identifies GalNAc-T2 O-glycosites in metabolic disorder . Proc Natl Acad Sci USA. . 2023;120:e2303703120. doi: 10.1073/pnas.2303703120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Schjoldager KT, Narimatsu Y, Joshi HJ, Clausen H. Global view of human protein glycosylation pathways and functions. Nat Rev Mol Cell Biol. . 2020;21:729–749. doi: 10.1038/s41580-020-00294-x. [DOI] [PubMed] [Google Scholar]
  • 3.Wandall HH, Nielsen MAI, King‐Smith S, de Haan N, Bagdonaite I. Global functions of O‐glycosylation: promises and challenges in O‐glycobiology. FEBS J. . 2021;288:7183–7212. doi: 10.1111/febs.16148. [DOI] [PubMed] [Google Scholar]
  • 4.de las Rivas M, Lira-Navarrete E, Gerken TA, Hurtado-Guerrero R. Polypeptide GalNAc-Ts: from redundancy to specificity. Curr Opin Struct Biol. . 2019;56:87–96. doi: 10.1016/j.sbi.2018.12.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.de las Rivas M, Paul Daniel EJ, Narimatsu Y, Compañón I, Kato K, Hermosilla P, Thureau A, et al. Molecular basis for fibroblast growth factor 23 O-glycosylation by GalNAc-T3. Nat Chem Biol. . 2020;16:351–360. doi: 10.1038/s41589-019-0444-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Schjoldager KT, Joshi HJ, Kong Y, Goth CK, King SL, Wandall HH, Bennett EP, et al. Deconstruction of O‐glycosylation—GalNA c‐T isoforms direct distinct subsets of the O‐glycoproteome. EMBO Rep. . 2015;16:1713–1722. doi: 10.15252/embr.201540796. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Koonin EV, Dolja VV. Virus world as an evolutionary network of viruses and capsidless selfish elements. Microbiol Mol Biol Rev. . 2014;78:278–303. doi: 10.1128/MMBR.00049-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Feng T, Zhang J, Chen Z, Pan W, Chen Z, Yan Y, Dai J. Glycosylation of viral proteins: Implication in virus-host interaction and virulence. Virulence. . 2022;13:670–683. doi: 10.1080/21505594.2022.2060464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Vigerust DJ, Shepherd VL. Virus glycosylation: role in virulence and immune interactions. Trends Microbiol. . 2007;15:211–218. doi: 10.1016/j.tim.2007.03.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Msemburi W, Karlinsky A, Knutson V, Aleshin-Guendel S, Chatterji S, Wakefield J. The WHO estimates of excess mortality associated with the COVID-19 pandemic. Nature. . 2023;613:130–137. doi: 10.1038/s41586-022-05522-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Brodin P, Casari G, Townsend L, O′Farrelly C, Tancevski I, Löffler-Ragg J, Mogensen TH, et al. Studying severe long COVID to understand post-infectious disorders beyond COVID-19. Nat Med. . 2022;28:879–882. doi: 10.1038/s41591-022-01766-7. [DOI] [PubMed] [Google Scholar]
  • 12.Lan J, Ge J, Yu J, Shan S, Zhou H, Fan S, Zhang Q, et al. Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor. Nature. . 2020;581:215–220. doi: 10.1038/s41586-020-2180-5. [DOI] [PubMed] [Google Scholar]
  • 13.Cai Y, Zhang J, Xiao T, Peng H, Sterling SM, Walsh Jr. RM, Rawson S, et al. Distinct conformational states of SARS-CoV-2 spike protein. Science. . 2020;369:1586–1592. doi: 10.1126/science.abd4251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Yan R, Zhang Y, Li Y, Xia L, Guo Y, Zhou Q. Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2. Science. . 2020;367:1444–1448. doi: 10.1126/science.abb2762. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Watanabe Y, Allen JD, Wrapp D, McLellan JS, Crispin M. Site-specific glycan analysis of the SARS-CoV-2 spike. Science. . 2020;369:330–333. doi: 10.1126/science.abb9983. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Zhao P, Praissman JL, Grant OC, Cai Y, Xiao T, Rosenbalm KE, Aoki K, et al. Virus-receptor interactions of glycosylated SARS-CoV-2 spike and human ACE2 receptor. Cell Host Microbe. . 2020;28:586–601.e6. doi: 10.1016/j.chom.2020.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Yang J, Wang W, Chen Z, Lu S, Yang F, Bi Z, Bao L, et al. A vaccine targeting the RBD of the S protein of SARS-CoV-2 induces protective immunity. Nature. . 2020;586:572–577. doi: 10.1038/s41586-020-2599-8. [DOI] [PubMed] [Google Scholar]
  • 18.Yang Q, Hughes TA, Kelkar A, Yu X, Cheng K, Park S, Huang WC, et al. Inhibition of SARS-CoV-2 viral entry upon blocking N- and O-glycan elaboration. eLife. . 2020;9:e61552. doi: 10.7554/eLife.61552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Zhang L, Mann M, Syed ZA, Reynolds HM, Tian E, Samara NL, Zeldin DC, et al. Furin cleavage of the SARS-CoV-2 spike is modulated by O-glycosylation. Proc Natl Acad Sci USA. . 2021;118:e2109905118. doi: 10.1073/pnas.2109905118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Wang S, Ran W, Sun L, Fan Q, Zhao Y, Wang B, Yang J, et al. Sequential glycosylations at the multibasic cleavage site of SARS-CoV-2 spike protein regulate viral activity. Nat Commun. . 2024;15:4162. doi: 10.1038/s41467-024-48503-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Dong X, Chen C, Yan J, Zhang X, Li X, Liang X. Comprehensive O-glycosylation analysis of the SARS-CoV-2 spike protein with biomimetic Trp-Arg materials. Anal Chem. . 2021;93:10444–10452. doi: 10.1021/acs.analchem.0c04634. [DOI] [PubMed] [Google Scholar]
  • 22.Tian W, Li D, Zhang N, Bai G, Yuan K, Xiao H, Gao F, et al. O-glycosylation pattern of the SARS-CoV-2 spike protein reveals an “O-Follow-N” rule. Cell Res. . 2021;31:1123–1125. doi: 10.1038/s41422-021-00545-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Bagdonaite I, Thompson AJ, Wang X, Søgaard M, Fougeroux C, Frank M, Diedrich JK, et al. Site-specific O-glycosylation analysis of SARS-CoV-2 Spike protein produced in insect and human cells. Viruses. . 2021;13:551. doi: 10.3390/v13040551. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Pallesen J, Wang N, Corbett KS, Wrapp D, Kirchdoerfer RN, Turner HL, Cottrell CA, et al. Immunogenicity and structures of a rationally designed prefusion MERS-CoV spike antigen. Proc Natl Acad Sci USA. . 2017;114:E7348. doi: 10.1073/pnas.1707304114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Xu Z, Li X, Zhou S, Xie W, Wang J, Cheng L, Wang S, et al. Systematic identification of the protein substrates of UDP‐GalNAc:polypeptide N‐acetylgalactosaminyltransferase‐T1/T2/T3 using a human proteome microarray. Proteomic. . 2017;17:1600485. doi: 10.1002/pmic.201600485. [DOI] [PubMed] [Google Scholar]
  • 26.Wu X, Brooks BR. Self-guided Langevin dynamics simulation method. Chem Phys Lett. . 2003;381:512–518. doi: 10.1016/j.cplett.2003.10.013. [DOI] [Google Scholar]
  • 27.Case DA, V Babin, J Berryman, RM Betz, Q Cai, DS Cerutti, TE Cheatham III, et al. AMBER 14, 2014, University of California, San Francisco, http://ambermd.org/
  • 28.Kirschner KN, Yongye AB, Tschampel SM, González‐Outeiriño J, Daniels CR, Foley BL, Woods RJ. GLYCAM06: a generalizable biomolecular force field. Carbohydrates. J Comput Chem. . 2008;29:622–655. doi: 10.1002/jcc.20820. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Maier JA, Martinez C, Kasavajhala K, Wickstrom L, Hauser KE, Simmerling C. ff14SB: improving the accuracy of protein side chain and backbone parameters from ff99SB. J Chem Theor Comput. . 2015;11:3696–3713. doi: 10.1021/acs.jctc.5b00255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ryckaert JP, Ciccotti G, Berendsen HJC. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J Comput Phys. . 1977;23:327–341. doi: 10.1016/0021-9991(77)90098-5. [DOI] [Google Scholar]
  • 31.Essmann U, Perera L, Berkowitz ML, Darden T, Lee H, Pedersen LG. A smooth particle mesh Ewald method. J Chem Phys. . 1995;103:8577–8593. doi: 10.1063/1.470117. [DOI] [Google Scholar]
  • 32.Tian Y, Parsons LM, Jankowska E, Cipollo JF. Site-specific glycosylation patterns of the SARS-CoV-2 spike protein derived from recombinant protein and viral WA1 and D614G strains. Front Chem. . 2021;9:767448. doi: 10.3389/fchem.2021.767448. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Bennett EP, Mandel U, Clausen H, Gerken TA, Fritz TA, Tabak LA. Control of mucin-type O-glycosylation: a classification of the polypeptide GalNAc-transferase gene family. Glycobiology. . 2012;22:736–756. doi: 10.1093/glycob/cwr182. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Zhang Y, Iwasaki H, Wang H, Kudo T, Kalka TB, Hennet T, Kubota T, et al. Cloning and characterization of a new human UDP-N-acetyl-α-d-galactosamine: polypeptideN-acetylgalactosaminyltransferase, designated pp-GalNAc-T13, that is specifically expressed in neurons and synthesizes GalNAc α-Serine/Threonine antigen. J Biol Chem. . 2003;278:573–584. doi: 10.1074/jbc.M203094200. [DOI] [PubMed] [Google Scholar]
  • 35.Kong Y, Joshi HJ, Schjoldager KTBG, Madsen TD, Gerken TA, Vester-Christensen MB, Wandall HH, et al. Probing polypeptide GalNAc-transferase isoform substrate specificities by in vitro analysis. Glycobiology. . 2015;25:55–65. doi: 10.1093/glycob/cwu089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Uhlén M, L Fagerberg, BM Hallström, C Lindskog, P Oksvold, A Mardinoglu, Å Sivertsson, et al. Tissue-based map of the human proteome, Science. 2015, 347: 1260419 . [DOI] [PubMed]
  • 37.Tian E, ZA Syed, ML Edin, DC Zeldin, KG Ten Hagen. Dynamic expression of mucins and the genes controlling mucin-type O-glycosylation within the mouse respiratory system, Glycobiology. 2023, 33: 476-489 . [DOI] [PMC free article] [PubMed]
  • 38.Nie X, Qian L, Sun R, Huang B, Dong X, Xiao Q, Zhang Q, et al. Multi-organ proteomic landscape of COVID-19 autopsies. Cell. . 2021;184:775–791.e14. doi: 10.1016/j.cell.2021.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Wang S, Yao X, Ma S, Ping Y, Fan Y, Sun S, He Z, et al. A single-cell transcriptomic landscape of the lungs of patients with COVID-19. Nat Cell Biol. . 2021;23:1314–1328. doi: 10.1038/s41556-021-00796-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Kamel W, Noerenberg M, Cerikan B, Chen H, Järvelin AI, Kammoun M, Lee JY, et al. Global analysis of protein-RNA interactions in SARS-CoV-2-infected cells reveals key regulators of infection. Mol Cell. . 2021;81:2851–2867.e7. doi: 10.1016/j.molcel.2021.05.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Jackson CB, Farzan M, Chen B, Choe H. Mechanisms of SARS-CoV-2 entry into cells. Nat Rev Mol Cell Biol. . 2022;23:3–20. doi: 10.1038/s41580-021-00418-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Rahnama S, Azimzadeh Irani M, Amininasab M, Ejtehadi MR. S494 O-glycosylation site on the SARS-CoV-2 RBD affects the virus affinity to ACE2 and its infectivity; a molecular dynamics study. Sci Rep. . 2021;11:15162. doi: 10.1038/s41598-021-94602-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Cao L, Tolić N, Qu Y, Meng D, Zhao R, Zhang Q, Moore RJ, et al. Characterization of intact N- and O-linked glycopeptides using higher energy collisional dissociation. Anal Biochem. . 2014;452:96–102. doi: 10.1016/j.ab.2014.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Malaker SA, Pedram K, Ferracane MJ, Bensing BA, Krishnan V, Pett C, Yu J, et al. The mucin-selective protease StcE enables molecular and functional analysis of human cancer-associated mucins. Proc Natl Acad Sci USA. . 2019;116:7278–7287. doi: 10.1073/pnas.1813020116. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Acta Biochimica et Biophysica Sinica are provided here courtesy of Acta Biochimica et Biophysica Sinica Editorial Office

RESOURCES