key: cord-0843224-vnlsjhrv authors: Shajahan, Asif; Archer-Hartmann, Stephanie; Supekar, Nitin T; Gleinich, Anne S; Heiss, Christian; Azadi, Parastoo title: Comprehensive characterization of N- and O- glycosylation of SARS-CoV-2 human receptor angiotensin converting enzyme 2 date: 2020-10-29 journal: Glycobiology DOI: 10.1093/glycob/cwaa101 sha: e7714f0a5573b248eb8471ab1de0a85eb8aae664 doc_id: 843224 cord_uid: vnlsjhrv The emergence of the COVID-19 pandemic caused by SARS-CoV-2 has created the need for development of new therapeutic strategies. Understanding the mode of viral attachment, entry and replication has become a key aspect of such interventions. The coronavirus surface features a trimeric spike (S) protein that is essential for viral attachment, entry and membrane fusion. The S protein of SARS-CoV-2 binds to human angiotensin converting enzyme 2 (hACE2) for entry. Herein, we describe glycomic and glycoproteomic analysis of hACE2 expressed in HEK293 cells. We observed high glycan occupancy (73.2 to 100%) at all seven possible N-glycosylation sites and surprisingly detected one novel O-glycosylation site. To deduce the detailed structure of glycan epitopes on hACE2 that may be involved in viral binding, we have characterized the terminal sialic acid linkages, the presence of bisecting GlcNAc, and the pattern of N-glycan fucosylation. We have conducted extensive manual interpretation of each glycopeptide and glycan spectrum, in addition to using bioinformatics tools to validate the hACE2 glycosylation. Our elucidation of the site-specific glycosylation and its terminal orientations on the hACE2 receptor, along with the modeling of hACE2 glycosylation sites can aid in understanding the intriguing virus-receptor interactions and assist in the development of novel therapeutics to prevent viral entry. The relevance of studying the role of ACE2 is further increased due to some recent reports about the varying ACE2 dependent complications with regard to age, sex, race, and pre-existing conditions of COVID-19 patients. The emergence of the COVID-19 pandemic caused by SARS-CoV-2 has created the need for development of new therapeutic strategies. Understanding the mode of viral attachment, entry and replication has become a key aspect of such interventions. The coronavirus surface features a trimeric spike (S) protein that is essential for viral attachment, entry and membrane fusion. The S protein of SARS-CoV-2 binds to human angiotensin converting enzyme 2 (hACE2) for entry. Herein, we describe glycomic and glycoproteomic analysis of hACE2 expressed in HEK293 cells. We observed high glycan occupancy (73.2 to 100%) at all seven possible N-glycosylation sites and surprisingly detected one novel O-glycosylation site. To deduce the detailed structure of glycan epitopes on hACE2 that may be involved in viral binding, we have characterized the terminal sialic acid linkages, the presence of bisecting GlcNAc, and the pattern of N-glycan fucosylation. We have conducted extensive manual interpretation of each glycopeptide and glycan spectrum, in addition to using bioinformatics tools to validate the hACE2 glycosylation. Our elucidation of the site-specific glycosylation and its terminal orientations on the hACE2 receptor, along with the modeling of hACE2 glycosylation sites can aid in understanding the intriguing virus-receptor interactions and assist in the development of novel therapeutics to prevent viral entry. The relevance of studying the role of ACE2 is further increased due to some recent reports about the varying ACE2 dependent complications with regard to age, sex, race, and pre-existing conditions of COVID-19 patients. In late 2019, the emergence of the highly transmissible coronavirus disease led to a global health crisis within weeks and was soon declared a pandemic. The new underlying pathogen belongs to the family of the coronaviridae and has been named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), initially termed 2019 novel coronavirus (2019-nCoV) (Gorbalenya, A.E., Baker, S.C., et al. 2020) . More than ten months into the pandemic, no promising vaccine candidates are currently available for COVID-19 and a recent WHO solidarity trial results indicated that the currently available drugs have little or no effect on overall mortality, initiation of ventilation and duration of hospital stay (Li, G.D. and De Clercq, E. 2020 , Pan, H., Peto, R., et al. 2020 , World Health Organization (WHO) 2020 . The elucidation of key structures involved in the transmission of SARS-CoV-2 will provide insights towards design and development of suitable vaccines and drugs against COVID-19 to curb the global public health crisis (Wang, N., Shang, J., et al. 2020) . Located on the viral surface is the spike (S) protein, which attaches the SARS-CoV-2 pathogen to target cells in the human body. The trimeric spike protein belongs to the class I fusion proteins. Its two subunits S1 and S2 orchestrate its entry into the cell: The S1 subunit facilitates the attachment of the virus via its receptor binding domain (RBD) to the host cell receptor, while the S2 subunit mediates the fusion of the viral and human cellular membranes (Hoffmann, M., Kleine-Weber, H., et al. 2020 , Shang, J., Ye, G., et al. 2020 . The glycosylation pattern of the spike protein, which carries N-linked, as well as O-linked glycosylation, has been the subject of several recent scientific studies (Shajahan, A., Supekar, N.T., et al. 2020b , Watanabe, Y., Allen, J.D., et al. 2020 , Zhang, Y., Zhao, W., et al. 2020 . Insight into the glycosylation pattern expands the understanding of the viral binding to receptors, the fusion event, host cell entry, replication, as well as the design of suitable antigens for vaccine development. Our recent study on the site-specific quantitation of N-linked and O- Glycosylation of human ACE2 5 linked glycosylation of the subunits S1 and S2 of the SARS-CoV-2 spike protein revealed the presence of an O-glycosylation site at the RBD of subunit S1 (Shajahan, A., Supekar, N.T., et al. 2020b ), although Watanabe et al. (Watanabe, Y., Allen, J.D., et al. 2020) reported lower abundances of O-linked glycans on the full-length spike protein. A number of amino acids within the RBD have been shown to be crucial determinants for binding in general and the binding affinity in particular of the virus to the host cell receptors (Andersen, K.G., Rambaut, A., et al. 2020 . Several studies have identified the human angiotensinconverting enzyme 2 (hACE2, ACEH) as the key virus receptor of SARS-CoV-1 in the early 2000s (Kuba, K., Imai, Y., et al. 2005 , Li, W., Moore, M.J., et al. 2003 . Moreover, hACE2 was identified recently as the receptor for cell entry of SARS-CoV-2 (Hoffmann, M., Kleine-Weber, H., et al. 2020 , Zhou, P., Yang, X.L., et al. 2020 . However, the SARS-CoV-1 and SARS-CoV-2 pathogens exploit the hACE2 receptor for cell entry only, which is unrelated to its physiological function (Li, F. 2013) . Also a recent study hypothesizes that N165 and N234 glycans of SARS-CoV-2 S protein stabilize the RBD "up" conformation and thus facilitate hACE2 binding, with N234 playing a more important role than N165 (Casalino, L., Gaieb, Z., et al. 2020) . hACE2 is a type-I transmembrane protein and includes an extracellular, a transmembrane, and a cytosolic domain within a total of 805 amino acids (Jiang, F., Yang, J., et al. 2014 ). Interestingly, it has been found that the transmembrane form can be cleaved to a soluble form of hACE2 (sACE2) that lacks the transmembrane and cytosolic domains but is enzymatically active, since the catalytic site, as well as the zinc-binding motif lie within the extracellular region (Tipnis, S.R., Hooper, N.M., et al. 2000) . hACE2 is secreted by endothelial cells and is involved in the renin-angiotensin system (RAS) (Donoghue, M., Hsieh, F., et al. 2000) . The presence of N-glycosylation was experimentally verified by treatment with the endoglycosidase PNGase F and monitoring the shift in the gel migration of hACE2 to the predicted molecular mass of ∼85 kDa. (Li, W., Moore, M.J., et al. 2003, Tipnis, S.R., Hooper, , et al. 2000 ). An earlier analysis on hACE2 N-linked glycans via sequential exoglycosidase digestion and HPLC after 2-aminobenzamide labeling identified mainly bi-antennary N-linked glycans with sialylation and core fucosylation, along with sialylated tri-and tetra-antennary Nglycans. A recent study and a patent have described glycoproteomics of hACE2 on non-human cells. However, glycosylation on such systems is not expected to reflect what is found in human ACE2 (Schuster, M., Loibner, H., et al. 2013 , Sun, Z., Ren, K., et al. 2020 . Our study complements a recent report on the N-glycosylation mapping on hACE2 (Zhao, P., Praissman, J.L., et al. 2020) expressed in HEK293 cells, particularly because only six out of seven potential N-glycosylation sites and no O-glycosylation on hACE2 was reported in that study. Moreover, we could not find any reports with detailed structure elucidation of both hACE2 N-and Oglycans including sialic acid linkages through glycomics, even on multiplatform glycosylation resource glygen.org (Chen, R., Jiang, X., et al. 2009 , Kristiansen, T.Z., Bunkenborg, J., et al. 2004 , Towler, P., Staker, B., et al. 2004 , York, W.S., Mazumder, R., et al. 2020 . The study of glycosylation can help in understanding the key roles glycans play during the physiological function of hACE2 and more importantly, its involvement in facilitating viral binding. We examined the cryo-electron microscopy (cryo-EM) structures of the SARS-CoV-1 S protein trimer in complex with hACE2 (PDB ID -6ACG) and SARS-CoV-2 RBD complexed with hACE2 dimer (PDB ID -6M17) to understand the binding location of S protein with hACE2 and the proximity of glycosylation sites (Song, W., Gui, M., et al. 2018 , Yan, R., Zhang, Y., et al. 2020 to the RBD of S protein. We concluded that N-glycosylation, particularly at sites N90 and N322, is close to the RBD of S protein and thus could play a critical role in the binding between the two proteins. A recent MD simulation study confirmed the glycan-mediated interactions between the S trimer and glycans at N90 and N322 of ACE2. Zhao et al.'s MD simulation, which used actual glycan structures identified by MS, furthermore indicated that the glycan at N546 of ACE2 is also involved in the interaction with the RBD of S protein (Zhao, P., Praissman, J.L., et al. 2020) . The binding location of individual monomers of the S protein with each monomer of the hACE2 dimer is shown in Figure 1 . Recently, it was reported that a dimeric ACE2 can accommodate two S protein trimers, each through a monomer of ACE2 (Yan, R., Zhang, Y., et al. 2020) . Our own recent study on the glycosylation of individual S protein subunits (Shajahan, A., Supekar, N.T., et al. 2020b) showed the prevalence of high mannose type glycans at the RBD of SARS-CoV-2 spike protein (N331 and N343), while the results of Watanabe et al. and Zhao et al. on trimeric S protein (Watanabe, Y., Allen, J.D., et al. 2020, Zhao, P., Praissman, J.L., et al. 2020 ) indicated a predominance of complex glycans at these sites. The relevance of the different types of glycans and their implication on the receptor binding needs further investigation. ACE2 enzyme inhibitors were demonstrated as not promising to prevent viral binding as the binding location of physiological ACE2 inhibitors and SARS-CoV viruses are different (Rico-Mesa, J.S., White, A., et al. 2020 , Towler, P., Staker, B., et al. 2004 ). Nevertheless, in order to understand the effect of glycosylation on the enzymatic activity of ACE2, we evaluated the proximity of the N-glycans to the peptidase activity and inhibitor binding sites of hACE2 ( Figure 1 ). Inspection of the co-crystallization structure of hACE2 (PDB ID -1R4L) with its inhibitor shows that the inhibitor is bound inside the cleft formed by the two monomers of ACE2, which is also the site of peptidase enzymatic activity (Towler, P., Staker, B., et al. 2004) . The viral RBDbinding region with its N-glycans, on the other hand, is located outside the cleft, nearly on the opposite face of each ACE2 monomer, as shown in Figure 1 . Within this study, we report the quantitative site-specific N-linked and O-linked glycan characterization of hACE2 by a glycomic and glycoproteomic approach. We identified glycosylation at all seven potential N-glycosylation sites on hACE2 and also report one Oglycosylation site. The N-and O-glycans were released from hACE2 and structurally 8 characterized by MALDI-MS and ESI-MS n . Particular focus was placed upon the location of fucosylation, the presence of bisecting GlcNAc residues, and the linkages of sialic acids. For the amino acid numbering we have followed the numbering used in UniProt [Q9BYF1] and the literature, although the numbering differs from that of the recombinant hACE2 we used for the study, which is missing the 1-17 signal peptide. Initially, we conducted a combination of trypsin and chymotrypsin digests to map the N-glycosylation on hACE2. Even though all the sites were covered by this digestion, the glycoform assignment at site N53 was ambiguous, as the peptide + HexNAc fragment (Y1 ion) was missing from the spectra. We conducted LC-MS/MS on de-N-glycosylated hACE2 and confirmed that the amino acid sequence of the peptide with site N53 is as expected. Nevertheless, we conducted a combination of GluC and chymotrypsin digestion to characterize the N-glycosylation at site N53. To ensure whether the detected glycans were indeed derived from ACE2, we searched the LC-MS/MS data of the ACE2 digest against the complete human UniProt database (downloaded Aug. 2018) and validated that hACE2 was the most abundant protein in the digest with a log probability value of 511.82, while the second abundant protein had a log probability value of 83.49 (Table SII) . Beside ACE2, no other detected proteins showed presence of glycopeptide spectra with acceptable fragment ions. We observed substantial (73.2% to 100% total site occupancy rate) glycan occupancy at all seven predicted N-glycosylation sites of hACE2 (Figures 2, 3 , 4, and S1-S7). While sites N53, N90, N103, N322, N546 showed no unglycosylated peptides, sites N432 and N690 were 26.8% and 1.1% unoccupied, respectively. Complex-type, bi-antennary glycans were much more abundant than high-mannose and hybrid type glycans across all Nglycosylation sites. We discovered highly processed sialylated, complex-type bi-antennary, tri- antennary and tetra-antennary glycans on all sites (Figures 2 and 4) . Different sialylation and glycan extension patterns were observed among the N-glycosylation sites. Site N53 featured predominantly sialylated and extended complex type glycans, whereas sites N90, N103, N322, and N546 carried abundant non-sialylated, bi-anntennary structures. Sites N432 and N690 were occupied with sialylated and non-sialylated N-glycans of almost equal abundance. All sites showed high levels of mono-fucosylated glycans. The N-glycan structures at each site were confirmed by the presence of characteristic peptide fragments, glycan oxonium ions and neutral losses, in addition to high resolution precursor mass determination (Figures 5A, S1-S7, Table SI ). dissociates glycans while generating few peptide backbone fragments, which complicates the identification of peptide sequences. Moreover, low m/z signature ions of glycopeptides, such as glycan oxonium ions may not be detected in the CID MS/MS spectra. Conversely, HCD enables detection of small diagnostic oxonium ions of monosaccharides and the peptide backbone while preserving labile modifications at the glycan cores (Cao, L., Qu, Y., et al. 2016) . Thus, by employing tandem HCD and CID fragmentation we obtained a wide spectrum of characteristic fragments of glycopeptides such as oxonium ions, peptide fragments, core glycan-peptide fragments (Y1, Y2 ions) and glycan neutral loss fragments ( Figure 5A ). Through a glycoproteomics approach, we have identified one O-glycosylation site on hACE2 by searching the LC-MS/MS data for common O-glycosylation modifications. We have observed very strong evidence for the presence of O-glycosylation at site Thr730 on peptide LGIQPTLGPPNQP, as the precursor masses, oxonium ions, neutral losses, and peptides fragments (b and y ions) were detected with high mass accuracy (Figures 2, 4B , 5B, and S8, (Figures 4, 5B , and S8). Even though we observed minor peak of O-glycan corresponding to GalNAcGalNeuAc in the chromatogram, we did not consider that for the calculation since such structure can be generated by in-source fragmentation ( Figure S8 ) (Grunwald-Gruber, C., Thader, A., et al. 2017 ). We aligned the amino acid sequence of human (Homo sapiens), bat (Rhinolophus sinicus), pig (Sus scrofa), cat (Felis catus), mouse (Mus musculus), rabbit (Oryctolagus cuniculus) , Malayan pangolin (Manis javanica), and chicken (Gallus gallus) ACE2 and compared the Nglycosylation sites among these species. This is shown schematically in Figure 6 . Whilst displaying overall sequence similarities of about 53 percent, human ACE2 possesses 7, bat (Chinese rufous horseshoe bat; Rhinolophus sinicus) ACE2 also 7, porcine ACE2 8, cat ACE2 a total of 9, murine ACE2 only 6, rabbit ACE2 8, pangolin ACE2 7, and chicken ACE2 10 potential N-glycosylation sites. Our sequence alignment studies showed that in human ACE2, five Nglycosylation sites share similarities with bat ACE2, but only three N-glycosylation sites share similarities with mouse ACE2, which is not susceptible to SARS-CoV-1 binding. Pig ACE2 showed four similar sites with human, whereas cat ACE2 showed five sites that are similar ( Figure 6 ). Based on recent reports, human, bat, cat, rabbit, and pangolin ACE2 exhibit binding affinity to SARS-CoV-2 viral RBD, while pig, mouse, and chicken ACE2 do not bind (Shi, J., Wen, Z., et al. 2020 , Zhao, X., Chen, D., et al. 2020 . Overall alignment of N-glycosylation sites among all these 8 species indicate that site N90 is the site that stands out among the susceptible and non-susceptible species. The susceptible species have an N90 glycosylation site, whereas the non-susceptible species do not. Interestingly, site N90 is closer to the SARS- CoV-2 RBD binding domain of ACE2 with respect to other sites, as there are reports indicating that abrogation of N90 glycosylation enhances the RBD binding with ACE2 (Procko, E. 2020 , Stawiski, E.W., Diwanji, D., et al. 2020 . Site N322 which is proximal to the SARS-CoV-2 RBD binding domain is aligned in human, bat, cat, pig and chicken but pig and chicken among them are non-susceptible. N-and O-glycomic studies of the hACE2 receptor were performed by methods described previously . Briefly, N-glycans were released by treating the Table I ). The glycans were predominantly of the complex type and were primarily (~60%) bi-antennary, with tri-antennary and quaternary structures also present in significant amounts. More than 85% of the structures were fucosylated, and roughly half of the structures were sialylated. The glycomics results corroborate with the glycoproteomics results except that few minor glycoforms present on the glycomics data were not annotated to the Nglycosylation sites of ACE2 because of the lack of their MS 2 scans on the LC-MS/MS data. The glycan structures were annotated based on high resolution precursor mass, ESI-MS n fragmentation and the common biosynthetic pathways of mammalian N-glycans ( Figure S11 ). MS/MS sequencing of the observed N-glycans was conducted with an automated top-down program, collecting MS 2 spectra of the highest intensity peaks with CID. This helped to confirm overall structure (for example hybrid forms versus complex type) and determine placement of the fucose. We observed that, while most of the fucosylated structures were primarily corefucose, small amounts of the same glycoform appeared to be fucosylated on the antennae ( Figure 8A ). This can be determined by a diagnostic terminal GalGlcNAcFuc fragment of m/z 660 [M+Na + ], which corresponds to the fragmentation of the antenna ( Figure 8A ) appearing as a b-ion. The corresponding y-ion m/z 1402 [M+Na + ] is also found, confirming this structure. Additional glycoforms displaying two fucoses (core and antennal) were also observed among these structures ( Figure 7D , Table I ). A separate MS n procedure was used to determine the linkages of sialic acids in the complextype N-glycans in hACE2. The method, which uses direct infusion of the permethylated glycans in a lithium carbonate and methanol solution, causes sequential fragmentation of the sialylated arms of a complex-type N-glycan, down to the galactose monosaccharide residue, which carries different methyl substitution, depending on the position through which it was linked to sialic acid in the glycan. The methyl substitution is then determined by analysis of the cross-ring fragments produced from the galactose monosaccharide ion. The cross-ring fragments obtained from the To determine presence of bisecting GlcNAc, an MS n strategy to trim down the permethylated N-glycan to the trimannose core was utilized ( and higher-order complex-type glycans (Zhao, X., Guo, F., et al. 2015) . The released O-glycans from hACE2 were sequenced by both MALDI-MS and ESI-MS n after permethylation ( Figures 8D and S10 ). Mostly the disialylated core-1 O-glycan GalNAcGalNeuAc 2 was observed along with a minor peak of monosialylated GalNAcGalNeuAc glycan. This is supported by our O-glycoproteomics findings, which showed that disialylated Core-1 was the major glycoform. The MS 2 fragmentation of the O-glycan confirms its structure as a sialylated Core-1 O-glycan ( Figure 8D ). According to recent cryo-EM studies on SARS-CoV-2, the binding of S protein to the hACE2 receptor primarily involves extensive polar residue interactions between RBD and hACE2 (Hoffmann, M., Kleine-Weber, H., et al. 2020 . The RBD located in the S1 subunit of SARS-CoV-2 S protein undergoes a hinge-like dynamic movement that enhances the capture of the spike protein RBD with hACE2 (Wrapp, D., Wang, N., et al. 2020) . This enhanced affinity for the human ACE2 receptor is predicted to be 10-20-fold higher for SARS-CoV-2 than SARS-CoV-1, which may be responsible for the increased transmissibility of the new virus (Wrapp, D., Wang, N., et al. 2020 , Yan, R., Zhang, Y., et al. 2020 . The protease domain of ACE2 interacts with the RBD of coronaviruses, and thus soluble ACE2 (sACE2), which is devoid of neck and transmembrane domains are capable of binding with RBD, neutralizing infection (Hofmann, H., Geier, M., et al. 2004 , Yan, R., Zhang, Y., et al. 2020 . ACE2 is expressed in most vertebrates. The ACE2 variants from human, bat, domestic pig, domestic cat, mouse, rabbit, and pangolin are all composed of 805 amino acid residues ( Figure 6 ), whereas chicken ACE2 contains 808 amino acids. SARS-CoV-1 and SARS-CoV-2 can infect a wide variety of organisms, including but not limited to humans, palm civets, cats and bats. In vitro virus infectivity studies conducted by Zhou el al indicated that SARS-CoV-2 is able to exploit the ACE2 proteins from humans, bats, pigs and civets to infect the cell cultures expressing the respective receptor. However, a cell culture expressing murine ACE2 was not infected (Zhou, P., Yang, X.L., et al. 2020) . This is based on varying degrees of receptor recognition. Receptor recognition is largely determined by two factors, (i) the binding specificity and (ii) the binding affinity of the RBD of the virus's spike protein to the cell entry receptor ACE2. This attachment step has been identified as a crucial limiting step for infection, as well as cross-species infection (Hou, Y., Peng, C., et al. 2010 , Li, F. 2013 ). Our comparison of Nglycosylation sites across species indicated that human ACE shares several glycosylation sites with other species. Studying the correlation between glycosylation sites on ACE2 of different species and their susceptibility to viral binding can help in understanding the key sites involved ( Figure 6 ). According to the data published by Qiu et al, the overall sequence identity of the ACE2 between different organisms and the hACE2 cannot be directly translated into prediction of transmissibility (Qiu, Y., Zhao, Y.B., et al. 2020) . For example, mouse ACE2 matches human with a higher sequence similarity than bat and pig, but the murine ACE2 cannot be exploited by the coronavirus, the study found (Qiu, Y., Zhao, Y.B., et al. 2020) . This indicates the possibility of involvement of glycosylation sites, glycans and glycan terminal epitopes in dictating the binding affinity of coronavirus RBD with ACE2. Our sequence analysis of ACE2 among different species draws attention to the correlation between glycosylation at site N90 and susceptibility, N90 being a site proximal to viral RBD binding (Figures 1, 3, and 6) . A recent in silico study by Stawiski et al (Stawiski, E.W., Diwanji, D., et al. 2020) predicted that glycosylation at N90 is an important modification that partly disrupts the interaction of the coronavirus with the ACE2 receptor (Figures 1 and 3) . Therefore, mutation at either N90 or T92 removes the glycosylation motif and makes the unglycosylated variant prone to interactions with SARS-CoV-2. The deleterious effect of N90 glycosylation upon binding to S protein has been experimentally confirmed by Chan et al. (Chan, K.K., Dorosky, D., et al. 2020) . A recent combined atomic force microscopy (AFM) and steered molecular dynamics (SMD) simulation study showed that the glycan on N90 can have, on the one hand, hindering effects on the association but, on the other hand, a retaining effect on the interaction and hinders dissociation of the complex formed between RBD of SARS-CoV-2 and ACE2 (Cao, W., Dong, C., et al. 2020) . Also, the elucidation of the structure of murine ACE2 would be beneficial for the investigation of insusceptible organisms. It might, in turn, provide valuable insight into the molecular structures that make humans prone to infection with SARS-CoV-2 via hACE2. A previous co-crystallization study of hACE2 with its inhibitor indicates that the spike protein RBD binding region of hACE2 is distal to the surface that binds with the ACE2 inhibitors, suggesting that the use of ACE2 inhibitors may not be beneficial in preventing viral binding ( Figure reason for the lack of 9-O-acetyl-sialic acid on the glycans from hACE2 we used for this study, as it is produced from HEK293 cells. We have detected both core fucosylation and antennal fucosylation on hACE2 N-glycans ( Figure 8A ). Moreover, we found evidence of bisecting GlcNAc on the hACE2 N-glycans ( Figure 8C ). The discovery of such glycan epitopes of hACE2 provides a better understanding of viral binding preferences and can guide the research for the development of suitable therapeutics. Our N-and O-glycosylation characterization of hACE2 expressed in a human cell system through both glycoproteomics and glycomics can help future studies in understanding the roles glycans play in the function and pathogen binding of hACE2. We have conducted extensive manual interpretation for the assignment of each glycopeptide, glycan and linkage structures in order to eliminate false detections. We are currently exploring the protein polymorphism on hACE2 and how the glycosylation profile varies in the variants which alter the glycosylation sites as it is reported that natural ACE2 variants are predicted to alter the virus-host interaction and thereby potentially alter host susceptibility (Stawiski, E.W., Diwanji, D., et al. 2020) . Detailed glycan analysis is important for the development of hACE2 or sACE2-based therapeutics which are suggested as a therapeutic measure to neutralize the viral pathogens (Hofmann, H., Geier, M., et al. 2004 , Yan, R., Zhang, Y., et al. 2020 . Evaluation of glycosylation on glycoprotein therapeutics produced from various human and non-human expression systems is critical from the point of view of immunogenicity, stability, as well as therapeutic efficacy (Beck, A., Cochet, O., et al. 2010, Sola, R.J. and Griebenow, K. 2010) . Studies indicate that although children generally express more ACE2 than adults, they tend to experience milder symptoms of COVID-19, raising the question whether the glycosylation profile of hACE2 changes with age or not (Ciaglia, E., Vecchione, C., et al. 2020) . Recent reports on COVID-19 infection of children, the triggering of severe cardiac problems, and protection observed in asthma patients necessitates the study on disease etiology and the contribution of hACE2 receptors (Hofmann, H., Simmons, G., et al. 2006 , Jackson, D.J., Busse, W.W., et al. 2020 . The understanding of complex sialylated N-glycans and sialylated mucin type O-glycans, on hACE2, along with their linkage and structural isomerism provides the basic structural knowledge that is useful for elucidating their interaction with viral surface protein and can aid in future therapeutic possibilities. We have procured recombinant hACE2 receptor expressed on human HEK293 cells and conducted the experiment with two biological replicates. The protein was expressed with a Cterminal His tag and comprises residues Gln18 to Ser740. We performed a combination of trypsin (1:20 enzyme protein ratio) and chymotrypsin (1:20 enzyme protein ratio) digestion on reduced and alkylated hACE2 and generated glycopeptides, each containing a single N-linked glycan site (based on in silico digestion). In order to map the site N53 unambiguously we conducted sequential digest of reduced/alkylated hACE2 with GluC (1:20 enzyme protein ratio) and chymotrypsin (1:20 enzyme protein ratio). The glycopeptides were directly subjected to The purified hACE2 (20 µg) expressed on HEK293 cells in 50 mM ammonium bicarbonate solution was reduced by adding 25 mM DTT and incubating at 60 °C for 30 min. The protein was further alkylated by the addition of 90 mM IAA and incubating at RT for 30 min in the dark. Subsequently, the protein was desalted using a 10-kDa centrifuge filter and digested by sequential treatment with trypsin and chymotrypsin or GluC and chymotrypsin by incubating for 18 h at 37 °C during each digestion step. The digest was filtered through a 0.2-µm filter and directly analyzed by LC-MS/MS. N-and O-linked glycans were released by following the methods described previously . N-linked glycans were released from about 80 µg of reduced and alkylated (as mentioned previously) hACE2 sample by treatment with PNGase F at 37˚C for 16 h. The released N-glycans were isolated by passing the digest through a C18 SPE cartridge with a 5% acetic acid solution (3 mL) and dried by lyophilization. The remaining de-N-glycosylated hACE2 protein, still containing O-glycans was then eluted from the column using 80% aqueous acetonitrile with 0.1% formic acid (3 mL The released N-and O-linked glycans were then permethylated using methods described elsewhere ). The glycoprotein digests were analyzed on an Orbitrap Fusion Tribrid mass spectrometer equipped with a nanospray ion source and connected to an Ultimate 3000 RSLCnano nano The permethylated N-and O-glycans were dissolved in 20 µL of methanol. A 0.5-µL portion of sample was mixed with an equal volume of DHB matrix solution (10 mg/mL in 1:1 methanolwater) and spotted on to a MALDI plate. MALDI-MS spectra were acquired in positive ion and reflector mode using an AB Sciex 5800 MALDI-TOF-TOF mass spectrometer. Original glycoform assignments were made based on full-mass molecular weight. Additional structural details were determined by MS n and modeling with the GlycoWorkbench software. Glycoform quantitation was conducted by calculating the relative intensity of glycan peaks after the deconvolution of ESI-MS spectrum by FreeStyle 1.4 software (Thermo Fisher). A solution of the permethylated N-glycans from hACE2 was diluted into a solution of 1 mM lithium carbonate/50% MeOH and directly infused (0.5 µL/min) into an Orbitrap Fusion Tribrid mass spectrometer equipped with a nanospray ion source. The sialic acid-containing N-glycans (determined by MALDI-TOF-MS and ESI-MS n experiments) were probed with MS n analysis described previously (Anthony, R.M., Nimmerjahn, F., et al. 2008 , Lin, N., Mascarenhas, J., et al. 2015 , Shajahan, A., Supekar, N.T., et al. 2017 . Isolation was conducted in the quadrupole, while detection was conducted in the IT. The isolation width of each fragmentation was 2 mass units, and the maximum injection time was 100 ms. More than 300 spectra were collected for each glycoform, which were then spectrally averaged. A solution of the permethylated N-glycans from hACE2 was diluted into a solution of 1 mM sodium hydroxide/50% MeOH and directly infused (0.5 µL/min) into an Orbitrap Fusion Tribrid mass spectrometer equipped with a nanospray ion source. The neutral complex-type N-glycans (determined by MALDI-TOF-MS and ESI-MS n experiments) were probed with MS n analysis described previously (Ashline, D.J., Duk, M., et al. 2015) . Because of the extra fragmentation steps, sialylated N-glycans were not probed since they yielded too low of a signal. Isolation was conducted in the quadrupole while detection was conducted in the IT. The isolation width of each fragmentation was 2 mass units, and the maximum injection time was 100 ms. More than The raw data files and search results can be accessed from glycopost repositoryhttps://glycopost.glycosmos.org/preview/18398096625f7659b46e0ba; pin: 7697. Financial support from the US National Institutes of Health (S10OD018530 and R24GM137782) is gratefully acknowledged. This work was also supported in part by the U. Glycosylation of human ACE2 27 The authors certify that they have no competing interests. Supplemental material: Annotated glycopeptide MS/MS spectra (S1 -S8), N-and O-glycan MALDI-MS data (S9, S10), and ESI-MS/MS data (S11) are incorporated as a separate supplemental material file. Supplemental Tables SI and SII are also included as separate files. Figure 1 : Cryo-electron microscopy (cryo-EM) structure of complexes of A. SARS-CoV-1 trimer with hACE2 (PDB ID -6ACG) and B. SARS-CoV-2 RBD with hACE2 (PDB ID -6M17) shows the binding of individual monomer of spike (S) protein with peptidase domain (PD) of each monomer of hACE2 dimer. Inset -Crystallography data of binding of hACE2 with inhibitor (PDB ID -1R4L). N-glycosylation at sites N90, N322 and N546 could a play critical role in the binding of hACE2 with RBD of S protein but these three glycans are likely too far from the enzyme active and inhibitor binding sites to have direct influence on inhibitor binding or hACE2 activity (location of glycans based on cryo-EM / crystallography data, representative glycan structures are core GlcNAc). Chem3D 18.1 was used for the mapping of the glycosylation sites and representation of PDB files. S C R I P T Figure 2 : Glycosylation profile on hACE2 characterized by high-resolution LC-MS/MS. All the seven potential Nglycosylation sites were found occupied along with one O-glycosylation site bearing core-1 type O-glycans. Mostly complex type glycans were observed in all N-glycosylation sites. Some N-glycosylation sites were partially glycosylated. Monosaccharide symbols follow the SNFG (Symbol Nomenclature for Glycans) system (Varki, A., Cummings, R.D., et al. 2015) . . Visual inspection of the potential binding location of SARS-CoV-2 RBD with hACE2 receptor suggests that the N-glycans at N90 and N322 could influence the receptor binding of coronavirus S protein RBD. However, recent MD simulation studies using full glycan structures (Zhao, P., Praissman, J.L., et al. 2020) indicate that site N546 of hACE2 may also be involved in intermolecular glycan-glycan interaction with S protein. ). The number of N-linked glycosylation motifs, -NXS/T-(XP), accounts for 7 potential sites in the human ACE2, 7 potential N-linked glycosylation sites in the bat ACE2, 8 potential N-glycan sites in the porcine ACE2, 9 potential sites in the cat ACE2, 6 potential N-linked glycosylation sites in the murine ACE2, 8 sites in rabbit ACE2, 7 sites in pangolin ACE2, and 10 sites in chicken ACE2. Blue colored are SARS-CoV-2 susceptible species and orange are nonsusceptible species. Glycosylation of human ACE2 38 Figure 7 : N-glycomics profiling of hACE2: N-glycans were released, purified, and permethylated prior to analysis with ESI-MS n . A. Deconvoluted profile of permethylated N-glycans from hACE2 (Only major glycoforms are shown). Masses are shown as deconvoluted, molecular masses. The N-glycan structure were confirmed by ESI-MS n ; B. The proximal origin of SARS-CoV-2 Recapitulation of IVIG anti-inflammatory activity with a recombinant IgG Fc The structures of glycophorin C N-glycans, a putative component of the GPC receptor site for Plasmodium falciparum EBA-140 ligand Expression of 9 GlycoFi's technology to control the glycosylation of recombinant therapeutic proteins Essentials of Glycobiology. Cold Spring Harbor (NY) Intact glycopeptide characterization using mass spectrometry Biomechanical Characterization of SARS-CoV-2 Spike RBD and Human ACE2 Protein-Protein Interaction Beyond Shielding: The Roles of Glycans in the SARS-CoV-2 Spike Protein Engineering human ACE2 to optimize binding to the spike protein of SARS coronavirus 2 Glycan topology determines human adaptation of avian H5N1 virus hemagglutinin Glycoproteomics analysis of human liver tissue by combination of multiple enzyme digestion and hydrazide chemistry COVID-19 Infection and Circulating ACE2 Levels: Protective Role in Women and Children A novel angiotensin-converting enzyme-related carboxypeptidase (ACE2) converts angiotensin I to angiotensin 1-9 Severe acute respiratory syndrome-related coronavirus: The species and its viruses -a statement of the Coronavirus Study Group Determination of true ratios of different N-glycan structures in electrospray ionization mass spectrometry Modulation of TNF-alpha-converting enzyme by the spike protein of SARS-CoV and ACE2 induces TNF-alpha production and facilitates viral entry SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor Susceptibility to SARS coronavirus S protein-driven infection correlates with expression of angiotensin converting enzyme 2 and infection can be blocked by soluble receptor Human coronavirus NL63 employs the severe acute respiratory syndrome coronavirus receptor for cellular entry Highly conserved regions within the spike proteins of human coronaviruses 229E and NL63 determine recognition of their respective cellular receptors Angiotensin-converting enzyme 2 (ACE2) proteins of different bat species confer variable susceptibility to SARS-CoV entry Association of respiratory allergy, asthma, and expression of the SARS-CoV-2 receptor ACE2 Ectodomain shedding of angiotensin converting enzyme 2 in human airway epithelia Angiotensin-converting enzyme 2 and angiotensin 1-7: novel therapeutic targets A proteomic analysis of human bile A crucial role of angiotensin converting enzyme 2 (ACE2) in SARS coronavirus-induced lung injury Angiotensin-converting enzyme 2 ectodomain shedding cleavage-site identification: determinants and constraints Receptor recognition and cross-species infections of SARS coronavirus Therapeutic options for the 2019 novel coronavirus (2019-nCoV) Identification of sialic acid-binding function for the Middle East respiratory syndrome coronavirus spike glycoprotein Angiotensin-converting enzyme 2 is a functional receptor for the SARS coronavirus Chinese hamster ovary (CHO) host cell engineering to increase sialylation of recombinant therapeutic proteins by modulating sialyltransferase expression Unravelling the Role of O-glycans in Influenza A ADAM17 inhibition may exert a protective effect on COVID Structures of MERS-CoV spike glycoprotein in complex with sialoside attachment receptors The sequence of human ACE2 is suboptimal for binding the S spike protein of SARS coronavirus 2 Predicting the angiotensin converting enzyme 2 (ACE2) utilizing capability as the receptor of SARS-CoV-2. Microbes Infect Outcomes in Patients with COVID-19 Infection Taking ACEI/ARB Recombinant ACE2 polypeptide dimer Glycomic and glycoproteomic analysis of glycoproteins-a tutorial Simplifying Glycan Profiling through a High-Throughput Micropermethylation Strategy Deducing the N-and O-glycosylation profile of the spike protein of novel coronavirus SARS-CoV-2 Tool for Rapid Analysis of Glycopeptide by Permethylation via One-Pot Site Mapping and Glycan Analysis Structural basis of receptor recognition by SARS-CoV-2 Susceptibility of ferrets, cats, dogs, and other domesticated animals to SARS-coronavirus 2 Glycosylation of therapeutic proteins: an effective strategy to optimize efficacy Cryo-EM structure of the SARS coronavirus spike glycoprotein in complex with its host cell receptor ACE2 Mass Spectrometry Analysis of Newly Emerging Coronavirus HCoV-19 Spike Protein and Human ACE2 Reveals Camouflaging Glycans and Unique Post-Translational Modifications Database analysis of O-glycosylation sites in proteins A human homolog of angiotensin-converting enzyme. Cloning and functional expression as a captopril-insensitive carboxypeptidase Structural basis for human coronavirus attachment to sialic acid receptors ACE2 X-ray structures reveal a large hinge-bending motion important for inhibitor binding and catalysis Concepts and principles of Olinked glycosylation Receptor Recognition by the Novel Coronavirus from Wuhan: an Analysis Based on Decade-Long Structural Studies of SARS Coronavirus Subunit Vaccines Against Emerging Pathogenic Human Coronaviruses. Front Microbiol Site-specific glycan analysis of the SARS-CoV-2 spike World Health Organization (WHO). 2020. Coronavirus, Licence: CC BY-NC-SA 3.0 IGO Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation Mass spectrometric glycan rearrangements Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2 Site-specific N-glycosylation Characterization of Recombinant SARS-CoV-2 Spike Proteins using High-Resolution Mass Spectrometry Virus-Receptor Interactions of Glycosylated SARS-CoV-2 Spike and Human ACE2 Receptor Broad and differential animal ACE2 receptor usage by SARS-CoV-2 Inhibition of endoplasmic reticulum-resident glucosidases impairs severe acute respiratory syndrome coronavirus and human coronavirus NL63 spike protein-mediated entry by altering the glycan processing of angiotensin I-converting enzyme 2 A pneumonia outbreak associated with a new coronavirus of probable bat origin