key: cord-0933729-6h6dmbl1
authors: Guo, Liang; Bi, Wenwen; Wang, Xinling; Xu, Wei; Yan, Renhong; Zhang, Yuanyuan; Zhao, Kai; Li, Yaning; Zhang, Mingfeng; Bao, Xingyue; Cai, Xia; Li, Yutang; Qu, Di; Jiang, Shibo; Xie, Youhua; Zhou, Qiang; Lu, Lu; Dang, Bobo
title: Engineered Trimeric ACE2 Binds and Locks “Three-up” Spike Protein to Potently Inhibit SARS-CoVs and Mutants
date: 2020-09-01
journal: bioRxiv
DOI: 10.1101/2020.08.31.274704
sha: c6603118730f5f22fc2267352883becc075e75e1
doc_id: 933729
cord_uid: 6h6dmbl1

SARS-CoV-2 enters cells via ACE-2, which binds the spike protein with moderate affinity. Despite a constant background mutational rate, the virus must retain binding with ACE2 for infectivity, providing a conserved constraint for SARS-CoV-2 inhibitors. To prevent mutational escape of SARS-CoV-2 and to prepare for future related coronavirus outbreaks, we engineered a de novo trimeric ACE2 (T-ACE2) protein scaffold that binds the trimeric spike protein with extremely high affinity (KD < 1 pM), while retaining ACE2 native sequence. T-ACE2 potently inhibits all tested pseudotyped viruses including SARS-CoV-2, SARS-CoV, eight naturally occurring SARS-CoV-2 mutants, two SARSr-CoVs as well as authentic SARS-CoV-2. The cryo-EM structure reveals that T-ACE2 can induce the transit of spike protein to “three-up” RBD conformation upon binding. T-ACE2 thus represents a promising class of broadly neutralizing proteins against SARS-CoVs and mutants.

motif and ACE2 would be around 60 Å, which corresponds to the length of the (GGGGS)3 154 linker. Thus, the (GGGGS)5 flexible linker in some of our designed proteins is long enough for 155 three ACE2s to bind, but is not optimal. The more rigid (EAAAK)5 linker is shorter than 156 (GGGGS)5 and can effectively separate different functional domains of fusion proteins 46 . We 157 think the effective length of the (EAAAK)5 linker is probably around 60 Å, making it an optimal 158 linker for T-ACE2. The rigid nature of this (EAAAK)5 linker probably helps to orient ACE2 159 right around RBD for immediate rebinding even if one ACE2 monomer from T-ACE2 160 dissociates from the spike protein. we engineered trimeric ACE2 proteins based on wild-type ACE2 and showed that T-ACE2 could 176 bind spike protein with extremely high affinity to potently inhibit all tested pseudotyped viruses 177 including SARS-CoV-2, SARS-CoV, eight naturally occurring SARS-CoV-2 mutants, two 178

SARSr-CoVs as well as authentic SARS-CoV-2. The rigid linker employed in T-ACE2 was 179 previously injected into mice and didn't seem to show strong immunogenicity 50 . The 3HB and 180 foldon trimerization motifs have been observed to cause immunogenicity, but the introduction of 181 glycans could silence the immunogenicity without disrupting the trimer formation 51 . Carrying 182 these advancements a few steps beyond, the modular design of T-ACE2 demonstrates that other 183 oligomerization motifs and linkers could be further explored to improve properties of T-ACE2 or 184 higher oligomeric ACE2s. 185

We demonstrated that T-ACE2 could induce the transit of spike protein to a unique "three-186 up" RBD conformation and bind all three RBDs simultaneously. Whether this T-ACE2-induced 187 spike protein conformation change represents a transition state during virus infection cannot be 188 definitively answered here. Full-length ACE2 protein functions as a dimer 42 , the two monomers 189 from this ACE2 dimer are related by two-fold symmetry. They are also situated close in space, 190 with the distance between D615 being about 53 Å. Thus, the native dimeric ACE2 is unlikely to 191 engage more than one RBD from the same spike protein without substantial conformational 192 changes. It is however possible that ACE2 dimers on the cell surface might further cluster to 193 induce more RBDs to adopt up conformation and help virus to transit from the prefusion state to though further studies are still needed to confer this advantage. Thus, proteins engineered based 202 on wild-type ACE2, such as T-ACE2, can potently and broadly inhibit virus infections, they also 203 have the added benefits of regulating RAS and alleviating ARDS. These potential beneficial 204 effects distinguish proteins like T-ACE2 from neutralizing antibodies. We believe T-ACE2 205 represents a promising class of proteins to broadly inhibit SARS-CoVs and to treat viruses 206 infected patients. Finally, the extremely high binding affinity between T-ACE2 and spike protein 207 (KD < 1pM) suggests that T-ACE2 could also be useful for virus detection. The fact that T-ACE2 208 was engineered based on native ACE2 sequence also makes such detection methods widely 209 useful for all SARS-CoVs and related viruses. 

To construct trimeric ACE2s, we inserted the linker sequences (GGGGS)5, (EAAAK)5 or 290 GGGS after ACE2(1-615), followed by trimerization motifs, an HRV3C cleavage sequence, an 291 eGFP tag and a His8 tag. Monomeric ACE2 was constructed as ACE2(1-615)-(GGGGS)5-292 HRV3C-eGFP-His8 for direct comparison. For purification of ACE2 proteins, the cell supernatants were harvested by centrifugation at 306 1000×g for 5 minutes. Then the supernatants were loaded on Ni-NTA beads (Smart-Lifesciences, 307

Cat. SA004100) and washed with washing buffer (5 mM imidazole, 1 × PBS). Proteins were 308 then eluted with elution buffer (50 mM imidazole, 1 × PBS). 309

The eluted proteins were concentrated and subjected to size-exclusion chromatography 310 (Superose 6 Increase 10/300 GL, GE Healthcare) in the PBS buffer. The peak fractions were 311 collected and concentrated. The proteins were then analyzed by size exclusion chromatography 312 (AdvanceBio SEC 300Å) in PBS buffer pH 7.4. The standard proteins were purchased from GE 313 S2) . 314

To remove C-terminal tags of ACE2 proteins, 16 µg HRV3C protease (expressed and 315 purified in house) was add to 1mg ACE2 protein and incubated at 4 °C overnight, followed by 

For kinetics analyses, S-ECD was captured on streptavidin biosensors. Biotinylated S-ECD 361 was diluted to 20 µg/mL in dilution buffer (PBS with 0.02% Tween 20 and 0.1% BSA). Then 362 sensor baselines were equilibrated in the dilution buffer for 90 seconds. Next, the S-ECD was 363 loaded until the thickness signal was 0.6 nm or 0.3 nm (low loading). After loading, the sensors 364 were washed for 60 seconds in the dilution buffer. The sensors were then immersed into wells 365 containing ACE2 proteins for 100 seconds (association phase), followed by immersion in 366 dilution buffers for an additional 300 seconds (dissociation phase). The background signal was 367 measured using a reference sensor with S-ECD loading but no ACE2 protein binding and was 368 subtracted from the corresponding ACE2 binding sensor. Curve fitting was performed using a 

The live SARS-CoV-2 inhibition assay was performed as previously described 56 . Briefly, 412

Vero-E6 cells were seeded into the 96-well cell culture plate at 3×10 4 per well and cultured for 413 12 hours. Recombinant proteins were diluted with FBS-free DMEM, mixed with 100 TCID50 of 414 SARS-CoV-2, and incubated at 37 ℃ for 1 hour. Then, the protein-virus mixtures were added to 415

Vero-E6 cells and incubated at 37 ℃ for 1 hour. After removing the mixtures, cells were 416 cultured with fresh DMEM containing 2% FBS for another 48 hours. Then, the supernatants 417 were collected to detect viral RNA titer. The cells were fixed to perform immunofluorescence 418 analysis. After fixing with 4% paraformaldehyde, the cells were permeabilized by 0.2% Triton 419 X-100 and blocked with 3% BSA for 1 hour. Then the SARS-CoV-2 Nucleocapsid Antibody 420

(1:500) (Sino Biological) was added to cells and reacted at 4℃ overnight. Finally, the cells were Particles were automatically picked using Relion 3.0.6 62-65 from manually selected 462 micrographs. After 2D classification with Relion, good particles were selected and subjected to 463 two cycles of heterogeneous refinement without symmetry using cryoSPARC 66 .The good 464 particles were selected and subjected to Non-uniform Refinement (beta) with C1 symmetry, 465 resulting in the 3D reconstruction for the whole structures that were further subjected to 3D 466 classification, 3D auto-refinement and post-processing with Relion. For interface between RBD 467 and ACE2, the datasets were subjected to focused refinement with adapted mask on each RBD 468 and ACE2 subcomplex to improve the map quality. Then the dataset of three RBD and ACE2 469 sub-complexes were combined and subjected to focused refinement with Relion, resulting in the 470 3D reconstruction of better quality on the interface between S-ECD and ACE2. 471

The resolution was estimated with the gold-standard Fourier shell correlation 0.143 criterion 472 67 with high-resolution noise substitution 68 . Refer to (fig S6-8) and Supplemental Table S1 for 473 details of data collection and processing. geometry restraints to prevent overfitting. To monitor the potential overfitting, the model was 485 refined against one of the two independent half maps from the gold-standard 3D refinement 486 approach. Then, the refined model was tested against the other map. Statistics associated with 487 data collection, 3D reconstruction and model building were summarized in Table S1 . 

A pneumonia outbreak associated with a new coronavirus of probable bat 588 origin

Structure of SARS coronavirus spike 590 receptor-binding domain complexed with receptor

Cryo-EM structure of the 2019-nCoV spike in the prefusion 592 conformation

Architecture of the SARS 594 coronavirus prefusion spike

Angiotensin-converting enzyme 2 is a functional receptor for the SARS 596 coronavirus

Human neutralizing antibodies elicited by SARS-CoV-2 infection

Potent Neutralizing Antibodies against SARS-CoV-2 Identified by High-600

Throughput Single-Cell Sequencing of Convalescent Patients' B Cells

A noncompeting pair of human neutralizing antibodies block COVID-19 603 virus binding to its receptor ACE2

Antibody cocktail to SARS-CoV-2 spike protein prevents rapid 605 mutational escape seen with individual antibodies

Studies in humanized mice and convalescent humans yield a SARS-607

CoV-2 antibody cocktail

Mutation rates among RNA viruses

Viral mutation rates

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G 613 Increases Infectivity of the COVID-19 Virus

The Impact of Mutations in SARS-CoV-2 Spike on Viral Infectivity and 615 Antigenicity

Escape from neutralizing antibodies by SARS-CoV-2 spike protein 617 variants

Emergence of RBD mutations in circulating SARS-CoV-2 strains enhancing 619 the structural stability and human ACE2 receptor affinity of the spike protein

A crucial role of angiotensin converting enzyme 2 (ACE2) in SARS 622 coronavirus-induced lung injury

Differential Downregulation of ACE2 by the Spike Proteins of Severe 624

Acute Respiratory Syndrome Coronavirus and Human Coronavirus NL63

Clinical and biochemical indexes from 2019-nCoV infected patients linked 627 to viral loads and lung injury

Angiotensin II plasma levels are linked to disease severity and predict 629 fatal outcomes in H7N9-infected patients

Angiotensin-converting enzyme 2 protects from severe acute lung failure. 631

Angiotensin-converting enzyme 2 protects from lethal avian influenza A 633 H5N1 infections

A pilot clinical trial of recombinant human angiotensin-converting 635 enzyme 2 in acute respiratory distress syndrome

Susceptibility to SARS coronavirus S protein-driven infection 637 correlates with expression of angiotensin converting enzyme 2 and infection can be 638 blocked by soluble receptor

Neutralization of SARS-CoV-2 spike pseudotyped virus by recombinant 640 ACE2-Ig

Inhibition of SARS-CoV-2 Infections in Engineered Human Tissues 642 Using Clinical-Grade Soluble Human ACE2

Trimeric SARS-CoV-2 Spike interacts with dimeric ACE2 with limited intra-644 Spike avidity

Engineering human ACE2 to optimize binding to the spike protein of 646 SARS coronavirus 2

Potential host range of multiple SARS-like coronaviruses and an improved 648 ACE2-Fc variant that is potent against both SARS-CoV-2 and SARS-CoV-1

Highly stable trimers formed by human immunodeficiency virus type 1 651 envelope glycoproteins fused with the trimeric motif of T4 bacteriophage fibritin

Foldon, the natural trimerization 654 domain of T4 fibritin, dissociates into a monomeric A-state form containing a stable beta-655 hairpin: atomic details of trimer dissociation and local beta-hairpin stability from residual 656 dipolar couplings

Thermodynamic analysis 658 of a designed three-stranded coiled coil

A basis set of de novo coiled-coil peptide oligomers for rational 660 protein design and synthetic biology

Stabilized coronavirus spikes are resistant to conformational 662 changes induced by receptor recognition or proteolysis

Cryo-EM structure of the SARS coronavirus 664 spike glycoprotein in complex with its host cell receptor ACE2

Distinct conformational states of SARS-CoV-2 spike protein

Structure, Function, and Antigenicity of the SARS-CoV-2 Spike 669 Glycoprotein

Cryo-electron microscopy structures of the SARS-CoV spike glycoprotein 671 reveal a prerequisite conformational state for receptor binding

Probing the structure of the SARS coronavirus using scanning electron 674 microscopy

Supramolecular architecture of severe acute respiratory syndrome 676 coronavirus revealed by electron cryomicroscopy

Fusion protein linkers: property, design and 678 functionality

Structural basis for the recognition of SARS-CoV-2 by full-length human 680 ACE2

Structural and Functional Basis of SARS-CoV-2 Entry by Using 682 Human ACE2

A pH-dependent switch mediates conformational masking of SARS-CoV-684 2 spike

Conformational dynamics of SARS-CoV-2 trimeric spike glycoprotein in 686 complex with receptor ACE2 revealed by cryo-EM

Fusion Proteins for Half-Life Extension of Biologics as a Strategy to Make 689 Biobetters

Engineered ACE2 receptor traps potently neutralize SARS-CoV-2

Molecular Mechanism for Antibody-Dependent Enhancement of 693 Coronavirus Entry

A perspective on potential antibody-dependent enhancement of SARS-695 CoV-2

Improving the oral efficacy of recombinant granulocyte colony-697 stimulating factor and transferrin fusion protein by spacer optimization

Immunosilencing a 700 highly immunogenic protein trimerization domain

Angiotensin-converting enzyme 2 in lung diseases

Much More Than Just a Receptor for SARS-COV-2

A pan-coronavirus fusion inhibitor targeting the HR1 domain of human 706 coronavirus spike

Inhibition of SARS-CoV-2 (previously 2019-nCoV) infection by a highly 708 potent pan-coronavirus fusion inhibitor targeting its spike protein that harbors a high 709 capacity to mediate membrane fusion

Potent antiviral effect of protoporphyrin IX and verteporfin on SARS-CoV-2 711 infection

A neutralizing human antibody binds to the N-terminal domain of the Spike 713 protein of SARS-CoV-2

Automated acquisition of cryo-electron micrographs for single particle 715 reconstruction on an FEI Tecnai electron microscope

MotionCor2: anisotropic correction of beam-induced motion for 717 improved cryo-electron microscopy

Measuring the optimal exposure for single particle cryo-EM 719 using a 2.6 A reconstruction of rotavirus VP6

Real-time CTF determination and correction

New tools for automated high-resolution cryo-EM structure 723 determination in RELION-3

Accelerated cryo-EM 725 structure determination with parallelisation using GPUs in RELION-2

RELION: implementation of a Bayesian approach to cryo-EM structure 728 determination

A Bayesian view on cryo-EM structure determination

cryoSPARC: algorithms for 732 rapid unsupervised cryo-EM structure determination

Optimal determination of particle orientation, absolute 734 hand, and contrast loss in single-particle electron cryomicroscopy

High-resolution noise substitution to measure overfitting and validate 737 resolution in 3D structure determination by single particle electron cryomicroscopy

Flexible fitting of atomic 740 structures into electron microscopy maps using molecular dynamics

PHENIX: a comprehensive Python-based system for macromolecular 745 structure solution