Structural basis of SARS-Cov-2 spike recognition by engineered synthetic multivalent VHH antibodies

High-throughput technologies such as next-generation sequencing (NGS), microarray-based gene synthesis, and phage display have empowered the discovery and engineering of precisely defined, synthetic antibodies with high avidity and drug-like features. Here, we describe a scalable process for engineering homo- and hetero-hexavalent variable domains of camelid heavy-chain (VHH)-Fc antibodies against the severe acute respiratory coronavirus 2 (SARS-CoV-2) spike (S) protein. Overall, we demonstrate that VHH trimerization is an effective and modular approach for increasing the affinity of anti-S1 VHH-Fc antibodies for the highly mutated S proteins of SARS-CoV-2 variants. We show that one specific nanobody (named TB201-1) binds spike trimer protein at the interface of two neighboring RBDs, recognizing one distinct epitope on one RBD but making a set of secondary interactions with the neighboring RBD. From this structure, we determine the epitope-paratope residues responsible for spike-nanobody interaction and how mutations found in the SARS-CoV-2 variants contribute to oblate TB201-1 binding. This approach could be leveraged to improve existing antibody-based diagnostics and therapeutics targeting SARS-CoV-2 as the virus evolves.