(6e) Computationally Engineering SARS Antibodies for COVID-19 Therapeutics

The emergence of Corona Virus Disease 2019 (COVID-19) has resulted in a pandemic with nearly 1,000,000 global cases as of April 1, 2020. There is a dire need for possible therapeutic agents or vaccines. The causative agent of this outbreak is the Betacoronavirus SARS-CoV-2, which has high genomic and structural homology with the SARS-CoV virus that caused the Severe Acute Respiratory Syndrome (SARS) pandemic in 2003-2004. Given the similarity between the two viruses, previous research on SARS immune responses can be leveraged to develop therapeutic treatments for COVID-19.

Antibodies play an important role in the protection against and recovery from viral infections. Neutralizing antibodies have been looked into as potential strategy for treating patients with SARS and other viral infections, because of their ability to neutralize the biological effects of the viral particles. SARS-CoV-2 uses the spike (S) protein to interact with the ACE2 cell receptor to enter host cells, similar to SARS-CoV. Thus, SARS-CoV specific antibodies are a promising potential therapeutic approach against the current disease. However, three antibodies (M396, S230, and 80R) which bind the S protein of SARS-CoV and neutralize the virus failed to show measurable binding to SARS-CoV-2 in early experiments.

We have employed computational methods to understand why those three antibodies have lost binding to SARS-CoV-2 and predict mutations to recover it. Our results show that the loss of binding is due to SARS-CoV-2 mutations that disrupt antibody “hotspot” interactions. Interestingly, almost all of the lost interactions in the three antibodies were salt bridges with the remainder being disruptions of hydrophobic patches. Based on this finding, a comprehensive interface mutational analysis was conducted to identify antibody mutations that can introduce novel hotspots with SARS-CoV-2. For all three antibodies, salt bridge and hydrophobic patch mutations were identified that are predicted to recover binding to SARS-CoV-2. Initial experiments were being conducted at the time this abstract was prepared. This presentation will describe the computational findings for why binding was disrupted and how it can be recovered, with experimental results hopefully included pending progress in the coming weeks and months.