3.5. The conformational changes of mink S glycoprotein with mvACE2 receptor 
Herein, we performed a simultaneous structural comparison of mvACE2-unbound and mvACE2-bound to spike at various RBD angles to investigate the different conformational changes induced by binding. Previous studies have reported that spike glycoprotein could only accommodate ACE2 binding when its RBD is up at an angle of at least 50° with respect to the horizontal plane of the S glycoprotein14,33,39. However, this data shows that about half of the mink S glycoprotein bound to one mvACE2 accommodated the mvACE2 binding with the RBD at a lower angle than previously reported13,14,33 (Fig. 2 and 6A).
Further structural comparison of our density map shows that the mvACE2-free conformation had its RBD at 27.1° in relation to the horizontal plane of mink S glycoprotein (Fig 2A). In comparison, one mvACE2 and two mvACE2 bound spikes had their RBD rotate outward with an up position at 62.8° and 69.2°, respectively (Fig 2C and 2D). The angle at which mvACE2 is bound to the upRBD in our study is comparable to the angles previously reported14,33,39 (Fig 6B and 6C). The intermediate stage of the mvACE2-bound mink S glycoprotein with downRBD captured was at an angle of 28.9°, similar to the angle of mvACE2-free conformation and approximately 34°- 40° lower than when it had up RBD (Fig. 6A). These findings challenge the previously reported notion that ACE2 binding may only occur when the RBD is up at a high angle as well as provide insight into the conformational changes that are induced upon ACE2 binding.
4. Discussion:
Our in vitro binding studies demonstrate a weaker binding affinity of the SD614G protein to the mvACE2 receptor compared to the hACE2 receptor, and this suboptimal fitness in a new host may be overcome by selecting virus with host-adaptive mutations. Indeed, mutations were frequently observed in SARS-CoV-2 isolates derived from infected mink26,41. We show that the addition of the mink-associated S glycoprotein mutations Δ69-70 and Y453F enhance binding to mvACE2 while having little impact on binding to the hACE2 receptor, suggesting they are host-adaptive mutations to American mink. The similarity in binding affinity between SD614G+Y453F and SD614G+Δ69-70+Y453Fwith the mvACE2 receptor suggests that the Y453F change is the main driver of the observed enhanced binding to mvACE2. Y453F has been shown by others to enhance mustelid ACE2 usage8,9,42. Residue 453 is located in the S glycoprotein RBD and interacts with hACE2 receptor residue H34, which is a Y34 in mvACE2. To elucidate the enhanced mvACE2 receptor binding mechanism, we solved the structure of the American mink S glycoprotein trimer bound to the mvACE2 receptor.
As previously reported, the RBD was at a range of tilts and angles regarding the horizontal plane of spike33. The general consensus was that the RBD of the spike protein undergoes a conformational shift in its RBD from an inactive ”down” state to an active ”up” state at an angle of at least 50° to access the ACE2 of the host cell14,33,39. Our study captured a novel intermediate step in which the mvACE2 binds to mink S glycoprotein with the RBD at a relatively lower angle than previously reported (Fig 4 and 6A). The simultaneous structural comparison suggests that the binding of the mvACE2 receptor facilitates the further opening of the CTD1 of the S glycoprotein. This would expose the fusion cleavage site of S2 in the spike, triggering the release of the S1 subunit from the S glycoprotein trimer43.
Overall, our structural analysis of the full trimeric mink S-mvACE2 complex is mainly in agreement with previously published structures5-14 in its architecture. Additionally, comparing the molecular interface between previously published mink S RBD-mink ACE2 complex and our full trimeric mink S-mvACE2 complex when RBD is up at >50° revealed comparable interacting residues8,9 (Fig 6). This includes the enhanced interaction between S F453 with mvACE2 Y34 via π-π stacking from the Y453F substitution mutation in the mink S glycoprotein.
Further analysis of the mink S-mvACE2 complex revealed that the interacting residues in the interface of the mink S and mvACE2 molecules differ depending on the angle at which the RBD is bound to mvACE2 (Fig 5C and 6C). These results provide insights into the residues in the spike trimer that are important for initiating the binding of the host ACE2 receptor. For example, while we observed that S R403, K417, and F486 may play a role in initiating the binding of the ACE2 receptor, we did not note any interaction in these residues when the ACE2 is bound to S at a higher angle. Similarly, we note that S Y505 plays a role in both initiating and maintaining the binding of host ACE2 receptors. This is congruent with previous reports that Y505 is a critical viral determinant for specific recognition of ACE2 by SARS-CoV-2 RBD and why many potent neutralizing antibodies interact with this specific residue42,44,45. Overall, our sequence alignment indicates that these interacting residues of the spike in downRBD conformation are well-conserved across different variants except the Y505H substitution in the Omicron BA.1 variant (Fig 3). It is reported that while the Y505H mutation in Omicron BA.1 significantly reduced ACE2 binding, and other mutations in the RBD compensated for its decreased binding affinity13,46.
The interactions between SARS-CoV-2 S glycoproteins and ACE2 receptors are of prime interest due to the essential role it plays in species specificity, transmission and pathogenesis. SARS-CoV-2 is now endemic in humans and will give rise to periodic epidemics similar to that of influenza A and B viruses and respiratory syncytial virus. Mutations will continue to emerge in the S glycoprotein to escape host immunity and/or optimize interactions with hACE2, and performing similar structural studies on new variants is necessary for understanding the disease and updating vaccines. Monitoring S glycoprotein variants for expanded or altered species specificity will also help assess the risk of zoonosis and reverse zoonosis.