key: cord-0737646-iih307fm
authors: Raghuvamsi, Palur; Tulsian, Nikhil; Samsudin, Firdaus; Qian, Xinlei; Purushotorman, Kiren; Yue, Gu; Kozma, Mary; Lescar, Julien; Bond, Peter; MacAry, Paul; Anand, Ganesh
title: SARS-CoV-2 S protein ACE2 interaction reveals novel allosteric targets
date: 2020-10-13
journal: bioRxiv
DOI: 10.1101/2020.10.13.337212
sha: 78d3a16f75ed2f08d1adf1755150ae92ccbe83c5
doc_id: 737646
cord_uid: iih307fm

The Spike (S) protein is the main handle for SARS-CoV-2 to enter host cells through surface ACE2 receptors. How ACE2 binding activates proteolysis of S protein is unknown. Here, we have mapped the S:ACE2 interface and uncovered long-range allosteric propagation of ACE2 binding to sites critical for viral host entry. Unexpectedly, ACE2 binding enhances dynamics at a distal S1/S2 cleavage site and flanking protease docking site ~27 Å away while dampening dynamics of the stalk hinge (central helix and heptad repeat) regions ~ 130 Å away. This highlights that the stalk and proteolysis sites of the S protein are dynamic hotspots in the pre-fusion state. Our findings provide a mechanistic basis for S:ACE2 complex formation, critical for proteolytic processing and viral-host membrane fusion and highlight protease docking sites flanking the S1/S2 cleavage site, fusion peptide and heptad repeat 1 (HR1) as allosterically exposed cryptic hotspots for potential therapeutic development. One Sentence Summary SARS-CoV-2 spike protein binding to receptor ACE2 allosterically enhances furin proteolysis at distal S1/S2 cleavage sites

The COVID-19 pandemic caused by the SARS-CoV-2 virus has sparked extensive efforts 44 to map molecular details of its life cycle to drive vaccine and therapeutic discovery.(1) SARS- class I viral fusion protein that has distinctive 'head' and 'stalk' regions ( Fig. 1A) . 54 A characteristic feature of SARS-CoV-2 is proteolysis of Pre-fusion S protein by host 55 proteases into S1 and S2 subunits. The S1 subunit comprises an N-terminal domain (NTD) and a 56 receptor binding domain (RBD) that interacts with the host receptor Angiotensin converting 57 enzyme-2 (ACE2)(10, 11) to initiate viral entry into the host.(12) Cryo-electron tomography (cryo-58 ET) has been used to capture the distribution and organization of trimeric S protein on the intact 59 virion,(9) revealing that 25 ± 9 S protein trimers decorate a single virion with a small percentage 60 (3%) of embedded S proteins in a post-fusion state adopting an extended helical conformation. 61 The first virus-host interaction is mediated by the viral S protein with the host ACE2 receptor.(10) 62 Binding to ACE2 primes the S protein for proteolysis at S1/S2 cleavage site into individual S1 and 

Localizing subunit specific dynamics and domain motions of S protein trimer 81 Structural snapshots of the ACE2 binding to the SARS-CoV-2 S protein interface have 82 been obtained with the RBD alone.(10, 16, 21-23) In this study, we have mapped this interface for 83 the S protein construct (1-1208) with mutations at the S1/S2 cleavage site (PRRAS to PGSAS) 84 and proline substitution at 986-987,(16) to block proteolysis during expression and purification 85 (Fig. S1A ). The S protein and isolated RBD constructs showed high affinity binding to ACE2 (Fig.   86 S1B). We measured dynamics of a trimer of this near-full length S protein by amide hydrogen-87 deuterium exchange mass spectrometry (HDXMS). Pepsin proteolysis generated 321 peptides 88 with high signal to noise ratio, accounting for ~87% of the entire S protein (Fig. S2) . Glycosylation of at least 22 sites have been predicted on S protein.(24) Average deuterium exchange at these 90 reporter peptides was monitored for comparative deuterium exchange analysis of S protein, ACE2 91 receptor and S:ACE2 complex, along with a specific ACE2 complex with the isolated RBD. While 92 glycosylation is an important posttranslational modification, our HDXMS study has measured 93 deuterium exchange of non-glycosylated segments of S protein alone. Deuterium exchange (t = 1 94 and 10 min) across all peptides of the free S protein trimer are shown in (Fig. 2) . We built an around the HR2 domain, compared to the rest of the protein (Fig. S3, Fig. S4 ).

The deuterium exchange heat map showed the highest relative exchange in the S2 subunit 105 (Fig. S3 ) and helical segments, while peptides spanning the fusion peptide showed relatively lower 106 deuterium exchange. Individually, S1 and S2 subunits showed different intrinsic deuterium 107 exchange kinetics, where the average relative fractional deuterium uptake (RFU) of S1 subunit 108 (~0.25) was lower than the average RFU (~0.35) of S2 subunit (Fig. S3 , Table S1 ). Moreover, 109 peptides connecting the RBD to the remainder of the S protein showed greater deuterium the 'down'-conformation relative to the hinge region (Fig. 2D, Fig. S4A ). Interestingly, a part of 114 the receptor binding motif, specifically residues 476-486, exhibited a higher degree of flexibility 115 based on its average atomic fluctuations ( Fig. 2A, 4C) , suggesting that binding to ACE2 receptor 116 would be required to stabilize its motion.

The NTD of the S protein showed low overall RFU (~0.2), consistent with its well-118 structured arrangement of β-sheets connected by loops (Fig. 1B) . Importantly, certain regions 

Peptides spanning residues 351-375 and 432-452 showed significantly increased deuterium 127 uptake, and these correspond to the NTD interdomain interaction sites. Interestingly, loci of the 128 RBD implicated in the interface (453-467, 491-510) with ACE2 showed relative higher exchange.

Overall, the S2 subunit showed relatively higher RFU than the S1 subunit, with each 130 domain exhibiting specific conformational changes (Fig. 1E, Fig. S4 ). Peptides spanning the 131 region immediately downstream of the S1/S2 cleavage site showed the highest deuterium uptake 132 (0.6), reflecting the rapid dynamics it undergoes for facilitating cleavage of S protein into two 133 subunits. Congruently, our MD simulations revealed the unstructured loop housing the S1/S2 134 cleavage site (residues 677-689) to be highly dynamic ( Figure S4C ), with RMSFs reaching >1.0 135 nm. It is important to note that the S1/S2 cleavage site has been abrogated in the construct of the 136 S protein used in this study to block proteolytic processing into S1 and S2 subunits during 137 expression in host cells. We thus observed lower deuterium uptake (and lower RMSF values) at 138 peptides in the central helix and connector domain, suggesting that these act as the central core of critically at the S1/S2 cleavage sites (Fig. 4D ). Even though the construct used in this study has 187 the proteolysis site mutated, it still resulted in increased dynamics at this S1/S2 locus. Furthermore, 188 this region exhibited high RMSF values during simulations. (Fig. S4B ). These results clearly 189 indicate that ACE2 binding induces allosteric enhancement of dynamics at this locus, providing 190 mechanistic insights into the conformational switch from the pre-fusion to fusogenic intermediate.

Differences in deuterium exchange between free S protein and S:ACE2 complex shows 192 stabilization at ACE2 interacting site and local destabilization at peptides juxtaposed to S1/S2 193 cleavage site (residues 931-938). This suggests that ACE2 binding potentiates peptide of residues 194 931-938 and other high exchanging regions flanking the S1/S2 cleavage site for enhanced furin 195 protease binding and cleavage. Importantly, these results suggest that the S1/S2 cleavage site is a 196 critical hotspot for S protein dynamic transitions for viral entry into the host, and therefore 197 represents a new target for inhibitory therapeutics against the virus. 199 Considering the indispensable role of ACE2 binding in SARS-CoV-2 infection, it is crucial to 200 assess the effects of S protein and RBD binding on ACE2 dynamics. We therefore mapped the 201 corresponding binding sites of RBD, both isolated and within the Spike, on ACE2. The S:ACE2 202 complex represents the prefusion pre-cleavage state wherein full-length S protein is bound to the 203 ACE2 receptor (Fig. 1B ii) , while the RBDisolated:ACE2 complex represents the post-furin cleavage 204 product formed by the S1 subunit and ACE2 (Fig. 1B iii) . Previous studies have shown that 14 key 205 amino acids of RBD interact with ACE2, wherein mutations at 6 amino acids resulted in higher (Fig. S7, Fig. S8 ). We observed a reduction in conformational changes promoting Furin (red) proteolysis at the S1/S2 cleavage site (red arrows, 337 leading to dissociation of S1 and S2 subunits, mechanism of which is unknown. (iii) The residual 338 ACE2-bound S1 subunit stably bound to ACE2 and S2 subunits dissociate (iv) Conformational (RBDS) mapped on to the structure of RBD extracted from S protein model (see Table S2 ). High 

Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation

Specificity of immobilized porcine 142 pepsin in H/D exchange compatible conditions

Site-specific glycan analysis of the 145 SARS-CoV-2 spike

Protein analysis by hydrogen exchange mass 147 spectrometry

Comparative protein modelling by satisfaction of spatial restraints

Cryo-EM structures of MERS-CoV and SARS-CoV spike glycoproteins reveal the 151 dynamic receptor binding domains

Solution structure of the severe acute 153 respiratory syndrome-coronavirus heptad repeat 2 domain in the prefusion state

Structural basis for membrane anchoring of HIV-1 envelope spike

A composite score for predicting errors in protein structure models

Stereochemistry of polypeptide chain 160 configurations

Palmitoylation of the cysteine-rich endodomain of the SARS-coronavirus spike 162 glycoprotein is important for spike-mediated cell fusion

Lipids of the Golgi membrane

Characterization of the budding 165 compartment of mouse hepatitis virus: evidence that transport from the RER to the Golgi 166 complex requires only one vesicular transport step

Coronavirus M proteins accumulate in the Golgi complex beyond the site 168 of virion budding

CHARMM-GUI Membrane Builder for Complex Biological Membrane Simulations 170 with Glycolipids and Lipoglycans

GROMACS: High performance molecular simulations through multi-level 172 parallelism from laptops to supercomputers

CHARMM36 all-atom additive protein force field: validation based 174 on comparison to NMR data

CHARMM-GUI Input Generator for NAMD, GROMACS, AMBER, OpenMM, and 176 CHARMM/OpenMM Simulations Using the CHARMM36 Additive Force Field

A molecular dynamics method for simulations in the canonical ensemble

Canonical dynamics: Equilibrium phase-space distributions

Polymorphic transitions in single crystals: A new molecular dynamics 183 method

A smooth particle mesh Ewald method

LINCS: A linear constraint solver for 187 molecular simulations

Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2

ACE2 receptor expression and severe acute respiratory syndrome coronavirus 191 infection depend on differentiation of human airway epithelia

Membrane lipidome of an epithelial cell line

VMD: visual molecular dynamics

A) Images of denaturing polyacrylamide gel electrophoresis 203 of purified proteins of the S protein (mutant), isolated RBD and ACE2 are shown, and their 204 molecular sizes are highlighted with red arrow, alongside protein standards

Deuterium uptake profile for ACE2 receptor and all-atom MD simulation of the 246 ACE2-B0AT1 complex

A) RFU values of pepsin proteolysed peptides listed in N-to C-terminus of ACE2 (peptide 18-248 615) for deuterium labelling times are shown. (B) Differences in deuterium exchange (Y-axis) of 249 ACE2 peptides listed from N-to C-terminus

Deuterium exchange significance threshold of ±0.3 D is indicated 251 in red and standard errors in gray. (C) The first principal motion of the all backbone atoms of the 252 ACE2 monomer as determined by PCA. (D) The RMSF values of the ACE2 receptor