key: cord-0296198-a2g7fm1q
authors: Ren, Wenlin; Lan, Jun; Ju, Xiaohui; Gong, Mingli; Long, Quanxin; Zhu, Zihui; Yu, Yanying; Wu, Jianping; Zhong, Jin; Zhang, Rong; Fan, Shilong; Zhong, Guocai; Huang, Ailong; Wang, Xinquan; Ding, Qiang
title: Mutation Y453F in the spike protein of SARS-CoV-2 enhances interaction with the mink ACE2 receptor for host adaption
date: 2021-08-24
journal: bioRxiv
DOI: 10.1101/2021.08.24.457448
sha: f57bda57c34ff94b6ddebc41f0bc93eb88d4c13c
doc_id: 296198
cord_uid: a2g7fm1q

COVID-19 patients transmitted SARS-CoV-2 to minks in the Netherlands in April 2020. Subsequently, the mink-associated virus (miSARS-CoV-2) spilled back over into humans. Genetic sequences of the miSARS-CoV-2 identified a new genetic variant known as “Cluster 5” that contained mutations in the spike protein. However, the functional properties of these “Cluster 5” mutations have not been well established. In this study, we found that the Y453F mutation located in the RBD domain of miSARS-CoV-2 is an adaptive mutation that enhances binding to mink ACE2 and other orthologs of Mustela species without compromising, and even enhancing, its ability to utilize human ACE2 as a receptor for entry. Structural analysis suggested that despite the similarity in the overall binding mode of SARS-CoV-2 RBD to human and mink ACE2, Y34 of mink ACE2 was better suited to interact with a Phe rather than a Tyr at position 453 of the viral RBD due to less steric clash and tighter hydrophobic-driven interaction. Additionally, the Y453F spike exhibited resistance to convalescent serum, posing a risk for vaccine development. Thus, our study suggests that since the initial transmission from humans, SARS-CoV-2 evolved to adapt to the mink host, leading to widespread circulation among minks while still retaining its ability to efficiently utilize human ACE2 for entry, thus allowing for transmission of the miSARS-CoV-2 back into humans. These findings underscore the importance of active surveillance of SARS-CoV-2 evolution in Mustela species and other susceptible hosts in order to prevent future outbreaks.

Coronaviruses are enveloped, positive-stranded RNA viruses that circulate broadly 53 among humans, other mammals, and birds and can cause respiratory, enteric, and hepatic 54 disease 1 . In the last two decades, coronaviruses have caused three major outbreaks: 55 severe acute respiratory syndrome (SARS), Middle East respiratory syndrome (MERS) 56 and the recent Coronavirus Disease 2019 (COVID-19) 2,3 . COVID-19, caused by severe 57 acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is a major global health threat. 58

The receptor-binding domain (RBD) of the SARS-CoV-2 spike (S) protein binds its 59 cellular receptor angiotensin-converting enzyme 2 (ACE2), thus mediating viral entry 4,5 . It 60 has been demonstrated that the interaction of a virus with (a) species-specific receptor(s) 61 is a primary determinant of host tropism and constitutes a major interspecies barrier at the 62 level of viral entry 6 . Our previous study found that numerous mammalian ACE2 orthologs 63 could function as receptors to mediate virus entry in vitro 7 , and other studies have 64 demonstrated that rhesus macaques, dogs, cats, cattle, hamsters, ferrets, minks and other 65 animals are susceptible hosts [8] [9] [10] [11] [12] [13] . Together, these findings suggest that SARS-CoV-2 has 66 a broad host range, with many species that could serve as potential reservoirs and thus 67 pose a risk for spillover to humans in the future 7 . 68

Recently, the first animal-to-human transmission of SARS-CoV-2 was reported 14,15 . In 69

April 2020, SARS-CoV-2 was transmitted to minks at two farms in the Netherlands by 70 infected employees, and the virus subsequently circulated among the minks 16,17 . On 5 71

November 2020, the Danish public health authorities reported 214 cases of humans 72 infected with mink-associated SARS-CoV-2 variants (miSARS-CoV-2) containing a 73 combination of mutations not previously observed 15 . Genetic analysis grouped the 74 miSARS-CoV-2 variants into 5 clusters with seven mutations 18 . Of note, the "Cluster 5" 75 variant with four amino acid changes in the spike protein was identified in mink and isolated 76 from 12 human cases in North Jutland 14, 18 . The implications of the mutations in this variant 77

are not yet well characterized. Preliminary results suggested that the "Cluster 5" miSARS-78

CoV-2 strain has moderately decreased sensitivity to neutralizing antibodies 14,19,20 . Further 79 As expected, the binding of S1-Fc or S1 (Y453F)-Fc to HeLa cells expressing mouse 118 ACE2 was very low and comparable to that of the empty vector control while S1-Fc 119 efficiently bound to HeLa cells expressing human ACE2, which is consistent with previous 120 reports 5 . S1 (Y453F)-Fc bound human ACE2 more efficiently than WT S1-Fc (99.1% vs 121 86.5%). Notably, WT S1-Fc bound mink ACE2 with limited efficiency, but in contrast, S1 122 (Y453F)-Fc bound mink ACE2 with 77% efficiency, demonstrating that the miSARS-CoV-2 123 mutation enhances binding to mink ACE2. Moreover, after replacing the amino acid residue 124 at position 34 of human ACE2 with its mink counterpart to generate human ACE2 (H34Y), 125 binding to S1-Fc was reduced (86.5% [WT] vs 42.0%) but increased to S1 (Y453F)-Fc 126 (92.25% [WT] vs 98.38%). Performing the converse by substituting the amino acid residue 127 at position 34 of mink ACE2 with its human counterpart to generate mink ACE2 (Y34H) 128 only slightly increased binding to S1-Fc (0.04% vs 0.92%) and decreased binding to 129 S1(Y453F)-Fc (45.05% vs 42.17%) (Fig. 2B) . 130

As we used cell-surface staining of ACE2 to sort the cells used for these experiments 131 so that they had comparable ACE2 expression, the limited or undetectable binding of 132 certain ACE2 variants with the S1 variants was not due to low expression of ACE2 or 133 alteration of its cell surface localization (Fig.S1A) . The expression level of the ACE2 134 variants was also assessed by immunoblotting using an anti-Flag antibody and cell surface 135 localization by immunofluorescent microscopy ( Fig. 2C and D) . Together, these results 136 showed that all the ACE2 orthologs were expressed and localized at the cell surface at 137 comparable levels, excluding the possibility that the limited binding efficiencies of ACE2 138 orthologs with S1-Fc variants was attributable to varied cell surface localization. 139

To further quantify the binding of ACE2 variants with the spike protein variants, we 140 expressed and purified recombinant WT and Y453F SARS-CoV-2 RBD as well as ACE2 141 variants to assay binding in vitro by surface plasmon resonance (SPR) analysis ( Fig. S3A-142 B and Fig. 2E ). Mink ACE2 bound Y453F SARS-CoV-2 with a KD of 78.22nM but binding 143 to WT RBD was not detectable. Human ACE2 bound SARS-CoV-2 RBD with a KD of 6.5nM, 144

and Y453F RBD with a KD of 0.98nM. These SPR results are thus consistent with the 145 findings of our cell-based assay. 146

Collectively, our results demonstrate that the SARS-CoV-2 spike binds mink ACE2 147 with limited affinity. However, the Y453F mutation dramatically increased the binding affinity 148 with mink ACE2 without compromising binding to human ACE2, which suggests that Y453F 149 is an adaptive mutation to improve its fitness in a new host-mink. 150 151 Structural basis for the enhanced binding of mink ACE2 with Y453F RBD 152 To elucidate at the atomic level the molecular basis for the enhanced binding, we 153 determined the crystal structure of SARS-CoV-2 Y453F RBD bound to mink ACE2 at 3.01 154 Å resolution ( Table 1) . The overall binding mode of Y453F RBD to mink ACE2 was very 155 similar to that of WT RBD to human ACE2 21 ( Fig. 3A and Table S1 ) as evidenced by the 156 low root mean square deviation (RMSD) value of 0.9 Å for the aligned 597 Cα atoms. A 157 surface of 1668 Å 2 is buried at the binding interface of Y453F RBD/mink ACE2 and is 158 comprised of 18 residues from Y453 RBD and 18 residues from mink ACE2. The WT 159 RBD/human ACE2 binding interface has a similar buried surface of 1687 Å 2 , made of 17 160 WT RBD residues and 20 human ACE2 residues. 161

Between mink ACE2 and human ACE2, there are differences in five of the amino acids 162 involved in RBD binding: L24 (mink ACE2)/Q24 (human ACE2), E30/D30, Y34/H34, 163 E38/D38 and H354/G354 (Table S1) . We especially focused on the residue at ACE2 164 position 34 as it directly interacts with RBD residues including that at position 453. 165

Structural analysis indicated that the Y453F substitution in the spike RBD is a species-166 specific adaptive mutation increasing the binding to mink ACE2 ( Fig. 3B and Table S1 ). At 167 the WT RBD/human ACE2 interface, RBD Y453 interacts with H34 of human ACE2 and 168 the change to Tyr, as seen in mink ACE2, results in a bulkier side chain that may increase 169 steric clashes and disfavor the binding of WT RBD to mink ACE2 (Fig. 3B) . However, with 170 the mutation to Phe at position 453 in the RBD, the Tyr at position 34 in mink ACE2 does 171 not face the same steric clashes (Fig. 3B ). In addition, while the human ACE2 H34 also 172 has interactions with WT RBD L455 and Q493, mink ACE2 Y34 interacts with residue R403 173 in addition to L455 and Q493 ( Fig. 3B and Table S1 ). These differing interactions may 174 account for the enhanced binding of Y453F RBD to mink ACE2. Finally, it is also predicted 175 that the Y453F mutation would not bring steric clashes with human ACE2 H34 and thus 176 would not significantly change the surrounding interactions, which may explain the retained 177 binding of Y453F RBD to human ACE2. 178

The Y453F mutation in miSARS-CoV-2 spike increases its interaction with other 180

Mustela ACE2 orthologs 181

Mink is a member of the Mustela genus which also includes stoats and ferrets, the 182 latter of which have been used as animal models owing to their susceptibility to SARS-183

CoV-2 25 . Due to the high similarity of ACE2 proteins in the Mustela genus (Fig. S4A) , we 184 performed the binding experiments of the S1 variants (WT or Y453F) with ferret and stoat 185 ACE2. Consistent with the results of mink ACE2, ferret and stoat ACE2 exhibited limited 186 binding capability with WT S1-Fc but increased binding ability with S1 (Y453F)-Fc (Fig.  187 4A). As before, we confirmed that the ACE2 orthologs expressed and localized at the cell 188 surface at comparable levels to exclude the possibility that differences in binding capability 189 were due to variation in ACE2 expression and/or mislocalization ( Fig.4B and Fig. S4B ). In To further demonstrate the biological consequences of the enhanced binding affinity 198 of S1 (Y453F) with Mustela ACE2, we generated MLV retroviral particles (Fluc as the 199 reporter) pseduotyped with SARS-CoV-2 WT spike (SARS-CoV-2pp) or variant Y453F 200 (SARS-CoV-2pp Y453F) (Fig. 5A) . A549 cells transduced with ACE2 orthologs were 201 inoculated with the pseudotyped virus, and luciferase (Luc) activity assayed 48h later to 202 determine entry efficiency. Mouse ACE2 was included as a negative control. As expected, 203

Luc activities were limited in A549-mouse ACE2 cells infected with WT or Y453F SARS-204

CoV-2pp. However, Luc activity increased by 350 and 400 fold in A549-human ACE2 cells 205 infected with WT or Y453F SARS-CoV-2pp, respectively, compared to that of A549-mouse 206 ACE2. In mink or ferret ACE2-transduced cells infected with SARS-CoV-2pp, the Luc 207 activity increased by 9 or 12 fold relative to that of mouse ACE2, respectively. The signal 208 was further enhanced in A549-mink ACE2 or A549-ferret ACE2 cells inoculated with SARS-209

CoV-2pp Y453F (110-or 120-fold, respectively, compared to A549-mouse ACE2). These Minks can act as a reservoir of SARS-CoV-2, circulating the virus among them and 258 thus becoming a source for virus spill-over into humans as observed in the Netherlands 14,18 . 259

The miSARS-CoV-2 variants harbor mutations in the spike protein, which potentially could 260 alter its transmission, pathogenicity, and/or sensitivity to vaccine-elicited immunity. In this 261 study, we demonstrated that the Y453F substitution in the miSARS-CoV-2 spike protein is 262 an adaptive mutation that significantly enhances interaction with mink ACE2 and promotes 263 infection of minks. Moreover, the Y453F substitution does not compromise the ability of the 264 spike protein to utilize human ACE2, explaining at least in part the circulation of the mink-265 associated virus observed in humans. The miSARS-CoV-2 variant exhibited partial 266 resistance to neutralization by convalescent sera, suggesting that the variant has the 267 potential to also escape protection induced by infection or vaccines. 268

The interaction of spike with ACE2 is a major genetic determinant of SARS-CoV-2 host 269 range 4,7,30,31 , and thus changes in the spike protein can adapt the virus to new hosts. Our 270 findings revealed that SARS-CoV-2 was genetically flexible in its ability to adapt to new 271 hosts. As we have shown, SARS-CoV-2 spike has limited binding capability to mink ACE2 272 ( Fig. 2B and E) , and our previous work demonstrated that mink ACE2 supported SARS-273

CoV-2 entry at 20% the efficiency of human ACE2 7 . It is conceivable that after transmission 274 from humans to minks, the virus evolved further in mink to adapt to the new host, with 275 efficient binding of SARS-CoV-2 S protein to mink ACE2 a prerequisite for ready 276 transmission among minks. Of interest, a recent study reported the absence of natural 277 SARS-CoV-2 human-to ferret transmission in a high-exposure setting, and genetic analysis 278

suggested that infection of ferrets may require viral adaptation 32 . Here we demonstrate that 279 the emergence of Y453F in SARS-CoV-2 spike significantly enhanced its interaction with 280

other Mustela ACE2 orthologs-namely ferret and stoat-conferring a potential fitness 281 advantage in Mustela species that could subsequently promote virus transmission and 282 possible risk of animal-adapted virus with genetic alterations to spilling over into humans. 283

Since the spike protein is a major target for prophylactic vaccines and antibody-based 284 therapeutics, such mutations could have implications for treatment, diagnostic tests and 285 viral antigenicity. It has been shown that the D614G substitution in the SARS-CoV-2 spike 286 can promote virus entry into host cells and enhance infectivity as well as make the mutant 287 virus resistant to neutralizing antibody 33,34 . The Y453F mutation was shown to be an 288 escape mutation for the monoclonal antibody RGN10933 35 . Our results showed that 289 miSARS-CoV-2 Y453F spike pseudotyped virions exhibited 3.5-fold resistance to 290 neutralization by convalescent serum (Fig. 6A) , suggesting the antibody response induced 291 by infection or vaccine might not offer sufficient protection against mink-associated SARS-292

CoV-2 variant infection. However, the miSARS-CoV-2 was still sensitive to ACE2-lg, which 293

suggests that ACE2-based therapeutics may represent an effective antiviral strategy 294 against the continued emergence of new variants. 295

In summary, our study suggests that Y453F is an adaptive mutation in SARS-CoV-2 296 that results in a virus more competent for infection and transmission among mink. In 297 addition, the miSARS-CoV-2 Y453F spike mutant maintained its ability to interact with 298 human ACE2 and exhibited partial resistance to neutralizing antibodies, potentially 299 explaining the ability of miSARS-CoV-2 to transmit back to humans and subsequent Luminescence was recorded on a GloMax® Discover System (Promega). 366

Protein expression and purification. The SARS-CoV-2 RBD (residues Arg319-Phe541) 367 and the N-terminal peptidase domain of human ACE2 (residues Ser19-Asp615) were 368 expressed using the Bac-to-Bac baculovirus system (Invitrogen) as described previously 21 . 369 ACE2-lg, a recombinant Fc fusion protein of soluble human ACE2 (residues Gln18-Ser740), 370 was expressed in 293F cells and purified using protein A affinity chromatography as 371 described in our previous study 36 . 372

Surface plasmon resonance anlysis. ACE2 was immobilized on a CM5 chip (GE 373

Healthcare) to a level of around 500 response units using a Biacore T200 (GE Healthcare) 374 and a running buffer (10 mM HEPES pH 7.2, 150 mM NaCl and 0.05% Tween-20). Serial 375 dilutions of the SARS-CoV-2 RBD were flowed through with a concentration ranging from 376 400 to 12.5 nM. The resulting data were fit to a 1:1 binding model using Biacore Evaluation 377

Software (GE Healthcare). 378

Protein crystallization. The gene encoding 333-527aa of SARS-CoV-2 RBD (WT or 379 Y453F mutant) was cloned into the pFastbac-dual vector using the BamH1 and Hind3 380 restriction enzyme. A GP67 signal peptide was added to the N-terminus for protein 381 secretion and a 6×His tag was added to the C-terminus for protein purification as described 382 previously 21 . The gene encoding 1-615aa of ACE2 from human or mink was constructed 383 in the same way as SARS-CoV-2 RBD except using its own peptide for protein secretion. Facility. The diffraction data were then processed using the HKL2000. The two structures 402 were determined using the molecular replacement method with Molrep in the CCP4 suite 37 , 403 using the model of the human ACE2 and SARS-CoV-2 complex 21 . There were three 404 complex molecules in one asymmetric unit. Subsequent model building and refinement 405 were performed using COOT and PHENIX, respectively 38,39 . The data-processing statistics 406 and structure refinement statistics are listed in Table 1 

All infections were performed in triplicate, and the data are representative of two independent 508 experiments (mean ± SD). *, P < 0.05, **, P < 0.01. Significance assessed by one-way ANOVA. 

Coronaviruses post-SARS: update on replication and 613 pathogenesis

A decade after SARS: strategies for controlling 615 emerging coronaviruses

Transcriptomic characteristics of bronchoalveolar lavage fluid and 617 peripheral blood mononuclear cells in COVID-19 patients. Emerging Microbes & 618 Infections 9

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is 620 Blocked by a Clinically Proven Protease Inhibitor

A pneumonia outbreak associated with a new coronavirus of probable bat 623 origin

Genetic Dissection of the Host Tropism of Human-Tropic Pathogens

Functional and genetic analysis of viral receptor ACE2 orthologs reveals a 627 broad potential host range of SARS-CoV-2

Respiratory disease in rhesus macaques inoculated with SARS-CoV-2. 630

From People to Panthera: Natural SARS-CoV-2 Infection in Tigers and 632 Lions at the Bronx Zoo

Detection of SARS-CoV-2 in a cat owned by a COVID-19−affected 634 patient in Spain

Susceptibility of ferrets, cats, dogs, and other domesticated animals to SARS-637 coronavirus 2

Pathogenesis and transmission of SARS-CoV-2 in golden hamsters

Infection of dogs with SARS-CoV-2

Transmission of SARS-CoV-2 on mink farms between humans 643 and mink and back to humans

SARS-CoV-2 Transmission between Mink (Neovison vison)

SARS-CoV-2 infection in farmed minks, the Netherlands

Recurrent mutations in SARS-CoV-2 genomes isolated from mink point 650 to rapid host-adaptation

SARS-CoV-2 and the human-animal interface: outbreaks on mink farms

Effect of RBD mutation (Y453F) in spike glycoprotein 654 of SARS-CoV-2 on neutralizing antibody affinity

SARS-CoV-2 mutations acquired in mink reduce antibody-mediated 656

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the 658 ACE2 receptor

Structural basis of receptor recognition by SARS-CoV-2

Structural and Functional Basis of SARS-CoV-2 Entry by Using Human 662

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

Infection and Rapid Transmission of SARS-CoV-2 in Ferrets

A novel cell culture system modeling the SARS-CoV-2 life cycle

Multiplex genome engineering using CRISPR/Cas systems

Human neutralizing antibodies elicited by SARS-CoV-2 infection

Inhibition of SARS-CoV-2 Infections in Engineered Human Tissues Using 674

Adaptation of SARS-CoV-2 in BALB/c mice for testing vaccine efficacy

A Mouse-adapted SARS-CoV-2 induces Acute Lung Injury (ALI) and 679 mortality in Standard Laboratory Mice

Host barriers to SARS-CoV-2 demonstrated by ferrets in a high-681 exposure domestic setting

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases 684

Infectivity of the COVID-19 Virus

Spike mutation D614G alters SARS-CoV-2 fitness

Antibody cocktail to SARS-CoV-2 spike protein prevents rapid mutational 689 escape seen with individual antibodies

SARS-CoV-2 and Three Related Coronaviruses Utilize Multiple ACE2 Orthologs 692

Are Potently Blocked by an Improved ACE2-Ig

Molecular replacement with MOLREP

Coot: model-building tools for molecular graphics

PHENIX: building new software for automated crystallographic structure 699 determination

PyMod 2.0: improvements in protein 702 sequence-structure analysis and homology modeling within PyMOL