Emerging variants of SARS-CoV-2 NSP10 highlight strong functional conservation of its binding to two non-structural proteins, NSP14 and NSP16

Reviewer #1 (Public Review):

Huan Wang et al. analyzed more than 10 million sequences and find that T12I, T102I and A104V were the top 3 frequently occurring mutations. They verified whether these mutations affect the stability and binding ability of NSP10, and whether there are structural changes. They find that three mutations destabilize the NSP10 by NMA prediction and determine their prediction by TSA. In addition, the Kd values shows that variants have similar binding ability or slightly improved affinity to NSP14 and NSP16 than native NSP10. Even though crystallization of the two variants is missing, the comparison of the crystallization of the T102I crystalline protein with the native shows that there is no structural change. Simultaneously, the dihedral angles in the variants do not explore any additional minima than that observed in wild-type NSP10, and there is no major conformational change.

Reviewer #2 (Public Review):

The authors of this study levered large-scale genomics data on SARS-CoV2, and extracted non-synonymous mutations of NSP10. The overall frequency was little, compared to other significantly mutating Spike protein. Further they performed stability and binding analysis to report changes in three variants and found modest differences. However, crystallography and simulations study reported almost no changes.

The strength of the work clearly is merging genomics data and reporting quantitative frequencies with high-resolution structural data. Some open ended questions remain. For instance, The DynaMut2 and thermal shift assays point towards less stable variants than wild type, with Tm values slightly lower. On the other hand, the Kd value of variants reported stronger binding of NSP10 with NSP16. How do authors explain this, as the change due to point mutation may not fall within error range?

The crystal structures and the simulations have been under-analysed. For instance, the conformational ensemble could be utilized for docking with NSP16 and NSP14 . There could be a potential alternative pathway for explaining the above changes in Kd. This should be attempted for understanding the role in its functional activity.

Previous extensive EM work on Spike protein variants also displayed subtle differences locally. However, allosteric pathways with D614G have been reported. Therefore, more quantitative analysis is required to explain structural changes. The free energy landscape reported in the paper may not capture rare transition events or slight rearrangements in side chain dynamics, both these could offer better understanding of mutations.