The sequence of human ACE2 is suboptimal for binding the S spike protein of SARS coronavirus 2

The rapid and escalating spread of SARS coronavirus 2 (SARS-CoV-2) poses an immediate public health emergency, and no approved therapeutics or vaccines are currently available. The viral spike protein S binds membrane-tethered ACE2 on host cells in the lungs to initiate molecular events that ultimately release the viral genome intracellularly. The extracellular protease domain of ACE2 inhibits cell entry of both SARS and SARS-2 coronaviruses by acting as a soluble decoy for receptor binding sites on S, and is a promising candidate for therapeutic and prophylactic development. Using deep mutagenesis, variants of ACE2 are identified with increased binding to the receptor binding domain of S at a cell surface. Mutations are found across the protein-protein interface and also at buried sites where they are predicted to enhance folding and presentation of the interaction epitope. The mutational landscape offers a preliminary blueprint for engineering high affinity ACE2 receptors to meet this unprecedented challenge. Overall design: The cDNA encoding full-length human ACE2 (GenBank NM_021804.1) with an N-terminal c-myc tag was diversified by site-saturation mutagenesis at 117 positions. This was accomplished by introducing degenerate NNK codons at the respective sites, which fully spanned the known binding epitope for SARS-CoV as well as additional sites in the ACE2 substrate-binding cleft. The ACE2 library was expressed in human Expi293F cells under conditions where most cells express no more than a single coding variant. The cell culture was then incubated with the receptor binding domain (RBD; a.a. 333-529) of the S1 subunit (GenBank YP_009724390.1) from SARS-CoV-2. The RBD was fused at its C-terminus to sfGFP for fluorescence detection in an oxidizing extracellular environment. The cell culture was simultaneously stained with Alexa 647-conjugated anti-myc to measure surface ACE2 expression. After washing, bound RBD-sfGFP was found to be proportional to ACE2 levels. There were unambiguous populations of cells expressing ACE2 variants with high and low RBD-sfGFP binding; these cell populations were collected by FACS and are called nCoV-S-High and nCoV-S-Low, respectively. Transcripts in the collected cell populations were Illumina sequenced and their frequencies compared to the naive plasmid library. The enrichment or depletion of all 2,340 single amino acid variants in the library was calculated, which acts as a proxy for relative “binding” to the RBD of SARS-CoV-2. This “binding” signal is a product both of S/ACE2 affinity and surface display of properly folded ACE2.