key: cord-0898899-eugpes2b
authors: Minasov, George; Rosas-Lemus, Monica; Shuvalova, Ludmilla; Inniss, Nicole L.; Brunzelle, Joseph S.; Daczkowski, Courtney M.; Hoover, Paul; Mesecar, Andrew D.; Satchell, Karla J. F.
title: Mn2+ coordinates Cap-0-RNA to align substrates for efficient 2′-O-methyl transfer by SARS-CoV-2 nsp16
date: 2021-02-01
journal: bioRxiv
DOI: 10.1101/2021.01.31.429023
sha: b9933122af9718deb7861179eae514ef5b37a5b9
doc_id: 898899
cord_uid: eugpes2b

Capping viral messenger RNAs is essential for efficient translation and prevents their detection by host innate immune responses. For SARS-CoV-2, RNA capping includes 2′-O-methylation of the first ribonucleotide by methyltransferase nsp16 in complex with activator nsp10. The reaction requires substrates, a short RNA and SAM, and is catalyzed by divalent cations, with preference for Mn2+. Crystal structures of nsp16-nsp10 with capped RNAs revealed a critical role of metal ions in stabilizing interactions between ribonucleotides and nsp16, resulting in precise alignment of the substrates for methyl transfer. An aspartate residue that is highly conserved among coronaviruses alters the backbone conformation of the capped RNA in the binding groove. This aspartate is absent in mammalian methyltransferases and is a promising site for designing coronavirus-specific inhibitors.

The recently emerged human pathogenic SARS-CoV-2 is a positive-stranded RNA virus responsible for the on-going pandemic of highly transmissible fatal respiratory coronavirus infectious disease , which has caused over two million deaths worldwide (1) . SARS-CoV-2 proteins are translated from nine canonical subgenomic mRNAs, generated by a discontinuous transcription process that results in all mRNAs having identical 5¢-ends (2) . To promote translation and to protect viral RNA from host surveillance by the innate immune system (3) , coronaviral mRNAs are capped with guanosine monophosphate. Next, the guanosine cap is methylated by the nsp14-nsp10 heterodimer to generate Cap-0-RNA. Finally, a methyl group is transferred from S-adenosylmethionine (SAM) to the 2¢-OH of the first adenosine residue to form Cap-1-RNA. For coronaviruses, this last reaction is catalyzed by the 2¢-O-methyltransferase (MTase), a heterodimeric complex of nsp16 with the activator nsp10 ( Fig. 1A) (3) (4, 5) . Inhibitors of nsp16 reduce viral titer and delay the interferon response in mice, validating nsp16 as a target for the development of anti-viral small molecules (6, 7) . To support drug discovery efforts, researchers around the globe have determined crystal structures of the SARS-CoV-2 nsp16-nsp10 in complex with ligands (8) (9) (10) (11) (12) , complementing prior structural biology observations for this enzyme from SARS-CoV and MERS-CoV (13) (14) (15) . This structural information has facilitated a detailed examination of the SARS-CoV-2 nsp16-nsp10 substrate binding sites (10, 11) . It is well established that the 2¢-O-MTase in SARS-CoV-2 and other RNA viruses can be activated by divalent metal ions (13, 14, 16) . Yet, despite this extensive structural information, a major gap exists in understanding of the role of metal ions in the 2¢-O-methyl transfer reaction and the position of ribonucleotides in the RNA binding groove since structures of these complexes have not been reported.

To initiate this study, we first extended prior studies on SARS-CoV and confirmed that the SARS-CoV-2 2¢-O-MTase also requires divalent cations. We used a custom-synthesized Cap-0-RNA substrate comprised of the N 7 -methylated guanosine (m 7 G) attached via a triphosphate bridge to a short RNA (AUUAAA), which matches the naturally occurring ribonucleotides at the 5¢-end of SARS-CoV-2 mRNAs (17) . At a concentration of 3 mM, both Mg 2+ and Mn 2+ significantly increased MTase activity (Fig. 1B) . In contrast, 3 mM Ca 2+ yielded only 50% of the activity observed with Mg 2+ and Na + did not stimulate activity. These data are consistent with observations for SARS-CoV nsp16 (4, 13, 14) . The activity of SARS-CoV-2 nsp16 with the Cap-0-RNA substrate (m 7 GpppAUUAAA) is over 10-times higher than that of the Cap-0 analog (m 7 GpppA) ( Fig. 1B , blue bars). Isothermal calorimetry (ITC) measurements were utilized to show that while it is a poor substrate for catalysis, m 7 GpppA does bind nsp16-nsp10 (Kd = 6.6 ± 0.3 µM) with 3fold higher affinity than to nsp16 alone (Kd = 28.0 ± 5.5µM) (Table S1 ). In contrast, m 7 GpppG bound only to the nsp16-nsp10 complex (Kd = 20.0 ± 2.7 µM). We also determined the binding affinities for the methyl donor SAM and the product SAH for nsp16 and the nsp16-nsp10 heterodimer. Neither SAM nor SAH bound to nsp16 alone, but both SAM and SAH bound to the nsp16-nsp10 heterodimer with Kd values of 6.9 ± 1.3 µM and 13.0 ± 1.2 µM, respectively. These results from biochemical and ITC studies together, as well as work by others (4, 9, (11) (12) (13) , indicated that the Cap-0 analog is capable of binding to nsp16 alone and that nsp10 greatly enhanced the binding affinity. However, a short capped RNA and the presence of Mg 2+ or Mn 2+ dramatically increased the rates of catalysis.

To gain further insight into how metal ions stimulate catalysis, we took a structural biology approach. Crystals of nsp16-nsp10 in complex with SAM from different crystallization conditions were soaked with the custom-synthesized m 7 GpppAUUAAA substrate in the presence of Mg 2+ or Mn 2+ (see methods). Multiple datasets were collected and, ultimately, three with the highest resolution and the best data statistics were selected for further analysis. Crystal #1 (PDB 7JYY) grew from a high NaCl concentration condition and was soaked with substrates and low MgCl2 concentration for 1.5 hours. In this crystal, we observed Cap-0-RNA, SAM, and Mg 2+ (Fig. 1C , D). Crystal #2 (PDB 7L6R) grew from a high (NH4)2SO4 concentration condition and was soaked with substrates and MnCl2 for 6 hours. In this crystal, we observed Mn 2+ and the products of the reaction, Cap-1-RNA and SAH, indicating that the methyl transfer reaction occurred in the crystal (Fig. 1E ). Crystal #3 (PDB 7L6T) grew from a high magnesium formate concentration condition and was soaked with substrates for 6 hours. In this crystal we observed two Mg 2+ ions and products of the reaction, Cap-1-RNA and SAH. The first Mg 2+ occupied the same metal binding site as in Crystals #1 and #2 and the second Mg 2+ directly interacted with phosphate groups of the capped RNA (Fig. S1 ). The number of ribonucleotides in Crystal #1 and #3 included the cap and first three ribonucleotides (m 7 GpppAUU) and the phosphate group of the fourth nucleotide. Crystal #2 contained the whole A4 ribonucleotide and the phosphate group of A5.

An overlay of the previously reported Cap-0 analog (PDB 6WRZ (10)) and Cap-0-RNA (PDB 7JYY, this study) structures showed they are very similar with a root-mean-square-deviation of 0.33 Å ( Fig. 2A) . We previously showed that the cap binding site, also called the High Affinity Binding Site (HBS), is bordered by flexible loops that adopt an open conformation upon Cap-0 analog binding (10) . Interactions of the nsp16 residues Tyr6828, Tyr6930, Lys6935, Thr6970, Ser6999 and Ser7000 with the Cap-0 analog and Cap-0-RNA are similar in both structures. The m 7 GpppA in the HBS is stabilized by stacking of the m 7 G and A1 bases with Tyr6828 and Tyr6930 residues, respectively ( Fig. 2A) . The O2¢ of the A1 ribose interacts with the conserved nsp16 catalytic residues (13, 14) , corresponding to Lys6839-Asp6928-Lys6968-Glu7001 in our structures (Fig. 2B) , as well as with the conserved water molecule we previously identified (10) . The interactions between Asn6841, SAM and O2¢ from the A1 are also consistent between these structures.

Of particular note was the space between the Tyr6930 and Asp6873 side chains, which is occupied by the A1 base in the Cap-0 analog structure. In the Cap-0-RNA structure, this space accommodates the stacked bases of A1 and U2, with the U2 base forcing repositioning of the A1 ribonucleotide ( Fig. 2A) . Superposition of the Cap-0 and Cap-0-RNA structures revealed that this repositioning involves: i) a 0.6 Å shift of the A1 base towards the side chain of Tyr6930 without notable changes in the positions of Asn6873 and Tyr6930, and ii) a 0.4 Å decrease in the distance between O2¢ of A1 and the SAM methyl group (Fig. 3A) , which occurs without significant changes in position of catalytic residues (Fig. 2B) . The movement and alignment of the A1 O2¢ atom toward the methyl group of SAM may explain why the additional ribonucleotides increase the efficiency of the methyl transfer reaction (18) (19) (20) .

Although the superposition of Cap-0 and capped RNA structures revealed differences in the position of the first adenosine, no significant deviations were observed between Cap-0-RNA and Cap-1-RNA conformations ( Fig. 2A,B) , indicating that the A1 base is not repositioned after the methyl transfer. The structures are essentially identical with the only difference being the methyl group, which moves from SAM to the A1 ribose hydroxyl group during methyl transfer (Fig.  2C ,D).

The structures in complex with capped RNAs also revealed the importance of the low affinity binding site (LBS) residues for the conformation of the mRNA in the catalytic site. The best resolved and most complete electron density for capped RNA was observed in Crystal #2. The position of U2 in the active site is "locked" by multiple interactions (Fig. 3A) . The phosphate group of U2 interacts with waters from the hydration sphere of the metal ion and the side chain nitrogen of Lys6844; the O2 atom of the U2 base interacts with the main chain nitrogen of Asp6873 and the O2¢ of the U2 ribose makes direct interactions with one of the oxygens of the side chain of Asp6873 (Fig. 3B ). The phosphate group of U3 interacts directly with the side chain nitrogen of Lys6874 and forms water mediated interactions with residues Asp6873, Lys6874, Met6840 and Asn6841. The base of U3 interacts directly with the main chain oxygen and nitrogen atoms of Ala6832 and forms a water bridge interaction with the nitrogen of the main chain of Leu6834. The whole nucleotide A4 and phosphate group of A5, the last ordered part of the capped RNA in the Crystal #2 structure, are solvent exposed and connected with protein via a hydrogen-bond between O4¢ of the A4 and the side chain oxygen of Ser6831 (Fig. 3B ). The stacking interactions between bases of U3 and A4 define the position of the A4 nucleotide. It is unknown if the conformation of A4 reflects the natural interaction of nucleotides, or if a longer mRNA would form different contacts with the nsp16-nsp10 heterodimer. However, the position of m 7 G and first three nucleotides of the capped RNA closely match in all three structures and likely represent the accurate binding mode for this part of the capped RNA.

The primary metal binding site is located near the HBS and loop1 ( Fig. 1C ) with either Mg 2+ or Mn 2+ occupying the same site with similar interactions (Fig. 3B,C) . The metal ions make both direct and water-mediated interactions with sidechains of nsp16 residues and the backbone of the capped RNA. The best electron density maps were observed for Crystal #3 with two magnesium ions, both of which have near-ideal octahedral geometry (Fig. 3C) . The Mg 2+ in the primary metal binding site is coordinated in part by interactions with phosphate groups of the triphosphate bridge linking the cap to A1, the phosphate group of U2, and the ribose of U3. The second Mg 2+ directly interacts with the U3 and A4 phosphate group oxygens and through waters with the side chain oxygens of the Asp6873 and the side chain nitrogen of Lys6874 (Fig. 3C) . Thus, the metal ion that occupies the primary metal binding site of nsp16 properly aligns capped RNA with SAM for an efficient methyl transfer reaction. The role of metal ions in facilitating orbital alignments for efficient catalysis has been demonstrated structurally for hydride transfer reactions (19) . Structural evidence for the requisite orbital alignment of substrates in RNA 2'-O-MTases (18) and ribozymes (20) has also been observed.

Although Mn 2+ and Mg 2+ are catalytic co-factors in solution (Fig. 1B) and were observed in our crystal structures, biochemical assay showed that Mn 2+ best stimulated the methyl transfer reaction (Fig. 1B) . More importantly, Mn 2+ is present at higher concentration than Mg 2+ in the endoplasmic reticulum and Golgi, where the SARS-CoV-2 replication vesicles are formed (21, 22) . These findings suggest that Mn 2+ may be the natural co-factor for nsp16 from SARS-CoV-2 and other coronaviruses.

All proposed in silico drugs that could target the residues of the LBS (9, (11) (12) (13) have relied on the crystal structures of 2¢-O-MTases with capped RNA for DENV NS5 (PDB 5DTO (23)), VACV VP39 (PDB 1AV6 (24)) and hCMTr1 (PDB 4N48 (25) ). Comparison of the RNA binding site and the capped RNA conformation from these structures with SARS-CoV-2 nsp16-nsp10 (PDB 7L6R) revealed that the conformation of the m 7 G and the location of the cap binding pockets relative to the active sites are dramatically different (Fig. S2A-D) . In the nsp16-nsp10 and VP39 structures, the N 7 -methyl groups are nestled in the HBS pocket (Fig. S2A,C) . In contrast, in NS5 and hCMTr1, for which methylation at the N 7 position of the G0 is not required for the 2¢-O-MTase activity (25) , the N 7 -methyl groups are pointed toward the solvent (Fig. S2B,D) . Although m 7 G positions do not overlap, nucleotides N1 and N2 in all these structures are closely matched, which is consistent with the conserved mechanism of action and the structure of the catalytic site (Fig. 4A in stereo view) . In all but the nsp16-nsp10 structure, the N2 nucleotide is sandwiched between the N1 and N3 by base stacking interactions and these three nucleotides have limited interactions with the 2¢-O-MTase residues of RNA binding grove. In the nsp16-nsp10 structure, all functional groups of the nucleotides N2 and N3 are involved in an integrated and complex network of hydrogen bond interactions with residues of the LBS. The U2 base and the ribose directly interact with main chain and side chain atoms of Asp6873. The side chain of Asp6873 occupies the space that is filled by N3 in all other structures. The presence of Asp6873 essentially forces the U3 nucleotide to move to the opposite side of the RNA binding groove, where it is involved in direct and water-mediated interactions with Ala6832 and Leu6834 (Fig. 4A,B) . The superposition of SARS-CoV-2 nsp16-nsp10 (PDB 7JYY) with MERS-CoV (PDB 5YNM (26)) revealed almost identical conformations of the proteins and the promontory Asp6873 residue, as well as the position of the Cap moiety (Fig. 4B) . Alignment of the amino acid sequences of nsp16 from representative coronaviruses ( Fig.  4C and Fig. S3 ) showed that Asp6873 is conserved across the coronaviruses, except for feline coronavirus (F-CoV). Structural and sequence alignments of SARS-CoV-2 nsp16 with the other MTases revealed that the Asp6873 is unique to coronaviruses and located in a four-residue insertion in the loop between b1 and aA (Fig. 4D,E) . Thus, this aspartate-containing loop and the redirection of the RNA, which is also coordinated by metals, is a unique feature that is shared across coronaviruses. Importantly, its absence in mammalian methyltransferases makes the region surrounding this residue a promising site for selective coronavirus-specific inhibitors.

Protein expression, purification, and crystallization. Recombinant nsp16 and nsp10 proteins with 6xHis-tags removed were purified from Escherichia coli and crystallized as previously described (10) .

MTase activity. The Cap-0 analog (m 7 GpppA) was obtained from New England Biolabs and the Cap-0-RNA (m 7 GpppAUUAAA) was custom-synthesized. The MTase activity was measured using the MTase-Glo Methyltransferase bioluminescence assay (Promega) (27) in buffer conditions suitable for use with SARS-CoV-2 nsp16. Luminescence was measured using a TECAN Safire2 microplate reader in arbitrary units and normalized assigning 100% to the activity in presence of Mg 2+ and Cap-0-RNA. The plot was created using GraphPad Prism V9 and shows the average and the standard deviation of three measurements using two different protein purifications.

Binding affinity was determined using a MicroCal PEAQ-ITC system (Malvern, Worcestershire, UK) at 25°C. The sample cell volume was 200 µL and the total syringe volume was 40 µL. For each titration, the first injection was performed using 0.4 µl which was then followed by 18 additional injections at 2 µl per injection. The first injection was considered a void and was automatically removed from data analysis. Each injection was spaced by 120 s after a 60 s initial delay. SARS-CoV-2 samples of nsp10, nsp16 and nsp16-nsp10 were individually loaded into the sample cell and then titrated with either SAH, SAM, m 7 GpppA Cap analog or the m 7 GpppG Cap analog (New England Biolabs). All samples were dialyzed overnight in ITC buffer (200 mM NaCl, 50 mM HEPES (pH 8.0), 0.1 mM ZnCl2, and 1 mM DTT). The concentrations of SARS-CoV-2 nsp10 or nsp16 used in experiments were 200 µM and 40 µM, respectively, and the concentrations of all substrates used were 500 µM. For titration experiments of the SARS-CoV-2 nsp16-nsp10, the two proteins were mixed to the final concentrations 25µM and 200 µM, respectively, and incubated for 30 mins at room temperature. A titration of SARS-CoV-2 nsp10 into nsp16 was performed at the 8:1 ratio to ensure no enthalpy was detected for the complex formation alone. The SARS-CoV-2 nsp16-nsp10 was titrated with substrates SAH and SAM at 375 µM and each Cap analog at 280 µM. Individual titration data were analyzed with MicroCal PEAQ-ITC Analysis Software using a single-site binding model and non-linear curve fitting. Each experiment was performed in triplicate and the resulting values and standard error in the fitted parameters for n, Kd, DH, DTS, and DG were obtained and are summarized in Table S1 . Data collection, processing, structure solution and refinement. Data sets were collected at the beam lines 21ID-D and 21ID-F of the Life Sciences-Collaborative Access Team (LS-CAT) at the Advanced Photon Source (APS), Argonne National Laboratory. Images were indexed, integrated and scaled using HKL-3000 (28) . Data quality and structure refinement statistics are shown in Table S2 . All structures were determined by Molecular Replacement with Phaser (29) from the CCP4 Suite (30) using the crystal structure of the nsp16-nsp10 heterodimer from SARS-CoV-2 as a search model (PDB ID 6W4H). The initial solutions went through several rounds of refinement in REFMAC v. 5.8.0266 (31) and manual model corrections using Coot (32) . The water molecules were generated using ARP/wARP (33) followed by an additional rounds of refinement in REFMAC. All structures were carefully examined, and three data sets were selected for further structural studies. For all structures, the Cap-0-RNA, SAM and Mg 2+ were fit into electron density maps and further refined. Inspection of anomalous and Fourier difference electron density maps revealed that, for Crystal #1 the nsp16-nsp10 heterodimer formed the complex with the Cap-0-RNA, SAM and Mg 2+ , for Crystal #2 the complex was formed with the Cap-1-RNA, SAH and Mn 2+ , and for Crystal #3 the complex was formed with the Cap-1-RNA, SAH and two Mg 2+ . In Crystals #1 and #3, no additional electron density was detected beyond phosphate group of A4. In Crystal #2, the presence of well-defined electron density near phosphate group of A4 allowed unambiguously extending of the RNA model by adding sugar and base for A4 and phosphate group of A5. All structures were further refined with the Translation-Libration-Screw (TLS) group corrections, which were created by the TLSMD server (34) . The quality control of the models during refinement and for the final validation of the structures were done using MolProbity (35) (http://molprobity.biochem.duke.edu/). All structures were deposited to Validated SARS-CoV-2 related structural models of potential drug targets (https://covid19.bioreproducibility.org/) and to the Protein Data Bank (https://www.rcsb.org/) with the assigned PDB codes 7JYY (Crystal #1), 7L6R (Crystal #2) and 7L6T (Crystal #3). A fourth structure (PDB 7JZ0) was also determined as part of this study and deposited but was not analyzed further as the metal was modeled as Na + , which does not catalyze the reaction, although data on this structure are included in Table S2 for reference. All models of the structures were created in PyMOL open source V 2.1 (36), diagram of interactions was created in LigPlot+ (37) .

Structural, sequence alignment and phylogenic tree. The PDB coordinates of SARS-CoV-2 nsp16 and nsp10 were analyzed using the FATCAT (38) , POSA (39) and DALI (40) ) on the right. The catalytic site residues, SAM and capped RNAs are labeled and shown as stick models with atoms colored in wheat, green, and grey for carbons of nsp16, SAM and capped RNA, respectively, with red for oxygens, blue for nitrogens, yellow for sulfurs. Conserved catalytic waters are shown as cyan spheres, hydrogen bond interactions as black, dashed lines, and the omit |Fo-Fc| electron density maps contoured at the 3s level as blue mesh. The methyl group of SAM and Cap-1-RNA are marked with black triangles. (43) . c Validation was done using MolProbity (35) .

An interactive web-based dashboard to track COVID-19 in real time

The Architecture of SARS-CoV-2 Transcriptome

mRNA capping: biological functions and applications

In vitro reconstitution of SARScoronavirus mRNA cap methylation

Molecular mechanisms of coronavirus RNA capping and methylation

Short peptides derived from the interaction domain of SARS coronavirus nonstructural protein nsp10 can suppress the 2'-O-methyltransferase activity of nsp10/nsp16 complex

Coronavirus nsp10/nsp16 Methyltransferase Can Be Targeted by nsp10-Derived Peptide In Vitro and In Vivo To Reduce Replication and Pathogenesis

Crystal Structure of the methyltransferase-stimulatory factor complex of NSP16 and NSP10 from SARS CoV-2. RCSB Protein Data Bank

Structural analysis of the SARS-CoV-2 methyltransferase complex involved in RNA cap creation bound to sinefungin

High-resolution structures of the SARS-CoV-2 2'-O-methyltransferase reveal strategies for structure-based inhibitor design

Structural basis of RNA cap modification by SARS-CoV-2

Methylation of RNA Cap in SARS-CoV-2 captured by serial crystallography

Biochemical and structural insights into the mechanisms of SARS coronavirus RNA ribose 2'-O-methylation by nsp16/nsp10 protein complex

Crystal structure and functional analysis of the SARS-coronavirus RNA cap 2'-O-methyltransferase nsp10/nsp16 complex

Crystal structure of MERS-CoV nsp10/nsp16 complex. RCSB Protein Data Bank

RNA 3'-end mismatch excision by the severe acute respiratory syndrome coronavirus nonstructural protein nsp10/nsp14 exoribonuclease complex

RNA genome conservation and secondary structure in SARS-CoV-2 and SARS-related viruses: a first look

Mechanism of RNA 2'-O-methylation: evidence that the catalytic lysine acts to steer rather than deprotonate the target nucleophile

Orbital steering in the catalytic power of enzymes: small structural changes with large catalytic consequences

Ribozyme catalysis via orbital steering

Ion and Redox Homeostasis: How Much Do They Really Matter? Front

The Ca2+/Mn2+ pumps in the Golgi apparatus

Molecular basis for specific viral RNA recognition and 2'-O-ribose methylation by the dengue virus nonstructural protein 5 (NS5)

Structural basis for sequence-nonspecific recognition of 5'-capped mRNA by a cap-modifying enzyme

Structural analysis of human 2'-O-ribose methyltransferases involved in mRNA cap structure formation

Crystal structure of MERS-CoV nsp16/nsp10 complex bound to SAM and m7GpppA. RCSB Protein Data Bank

An enzyme-coupled ultrasensitive luminescence assay for protein methyltransferases

HKL-3000: the integration of data reduction and structure solution--from diffraction images to an initial model in minutes

Phaser crystallographic software

Overview of the CCP4 suite and current developments

REFMAC5 for the refinement of macromolecular crystal structures

Coot: model-building tools for molecular graphics

ARP/wARP and automatic interpretation of protein electron density maps

Optimal description of a protein structure in terms of multiple groups undergoing TLS motion

MolProbity: all-atom structure validation for macromolecular crystallography

Pymol: An open-source molecular graphics tool. CCP4 Newsletter On Protein Crystallog

LigPlot+: multiple ligand-protein interaction diagrams for drug discovery

FATCAT: a web server for flexible structure comparison and structure similarity searching

POSA: a user-driven, interactive multiple protein structure alignment server

Dali server: conservation mapping in 3D

Deciphering key features in protein structures with the new ENDscript server

Global indicators of X-ray data quality

Linking crystallographic model and data quality

Protein chains are shown as ribbons in orange for SARS-CoV-2 (PDB 7JYY) and green for MERS-CoV (PDB 5YNM) overlayed on the semitransparent solvent-exposed surface of nsp16 (wheat) and nsp10 (teal). Nucleotides are labeled, conserved Asp is marked with black triangle and direct hydrogen bond interactions between U2 and conserved Asp are shown as black, dashed lines. (C) Multiple sequence alignment of the loop region between 1 and A from different coronaviruses Severe Acute respiratory Syndrome (SARS-CoV)

We thank Grant Wiersum, Olga Kiryukhina and Ivgeniia Dubrovska for technical assistance in protein expression, purification and crystallization. We also thank Masoud Vedadi (University of Toronto) for helpful advice on obtaining the substrate.