Alternative sequence–function optima in rodent ancestor ACE2 serendipitously confers resistance to SARS-CoV-2 in extant mice.

(A) Phylogeny of jawed vertebrates used in the ancestral reconstruction of ACE2 from the last common ancestor of rodents (RodAnc). (B) Evolution of high ACE2 activity in the rodent ancestor. The carboxyl terminus of mouse (M) or human (H) was used to determine hydrolysis rates of a fluorometric peptide. N = 4 to 5 biological replicates. Standard error is shown. (C) RodAnc ACE2 displays none of the rare sequence variants unique to the RBD binding domain of mice. (D) Flow cytometry analysis of SARS-CoV-2 S RBD binding to RodAnc-M and human ACE2. Sextuple mutants of each contain the 6 mouse variants. N = 3 biological replicates. Standard deviation is shown. (E) Hydrolysis rates of WT and mutant ACE2. N = 5 biological replicates. Standard error is shown. (F) Residues conferring mice with resistance to SARS-CoV-2 are unlikely to evolve in humans due to antagonistic pleiotropy with human ACE2 hydrolysis rates. This antagonistic pleiotropy is absent from dogs and the rodent ancestor. All data points relative to human WT, except dog, for which RBD binding is shown relative to dog WT. (G) Systolic blood pressure divergence between rodents (red) and primates (blue) is correlated with differences in body size (p = 0.001; phylogenetically independent least squares linear regression). All data are available in S1 Data. ACE2, angiotensin converting enzyme 2; PIC, phylogenetic independent contrast; RBD, receptor-binding domain; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2; WT, wild-type.