Understanding SARS-CoV-2 antigenic evolution using complete genotype-to-phenotype maps of the effects of mutations on antibody binding

Abstract

Since the COVID-19 pandemic began, global sequencing efforts have allowed scientists to follow the evolution of the SARS-CoV-2 virus in real time. The consequences of that evolution, however, are less readily apparent. Traditional methods to test the functional or antigenic effects of a new mutation can take weeks, resulting in a lag between surveillance and the corresponding experimental data. Thus, we developed a prospective approach to comprehensively measure the effects of mutations and generate complete genotype-to-phenotype maps for the SARS-CoV-2 spike receptor-binding domain (RBD). The RBD binds to the angiotensin converting enzyme 2 (ACE2) receptor on host cells, mediating viral entry. We developed a yeast-display deep mutational scanning system to measure the effects of all possible single amino-acid mutations to the RBD on expression (a correlate of protein folding and stability) and ACE2 binding. We identified parts of the RBD that are mutationally constrained, as well as mutations that enhance expression or ACE2 binding. We created an interactive visualization to quickly identify the effects of any mutation to the RBD. This "lookup table" has been useful in interpreting the evolution of SARS-CoV-2 as new SARS-CoV-2 variants have emerged. The RBD is also a major target of neutralizing antibodies that can block the virus' ability to enter cells. We extended the deep mutational scanning system to comprehensively measure the effects of mutations to the RBD on antibody binding. We found that the resulting antibody-escape maps predict which mutations will be selected during viral growth in the presence of antibody. Additionally, we identified mutations that can escape binding and neutralization of antibody therapeutics used to treat COVID-19. As new SARS-CoV-2 variants emerged, these antibody-escape maps could predict which variants would be resistant to antibody treatment. As greater fractions of the population gain immunity to SARS-CoV-2 through infection and/or vaccination, immune selection is likely to become a major driving force of the virus' evolution. I applied our deep mutational scanning system to measure the effects of RBD mutations on the binding of polyclonal antibodies from convalescent or vaccine-elicited plasmas and sera. We found that mutations to site 484 within one epitope (the "class 2" epitope) had some of the largest effects on antibody binding. For convalescent plasmas, mutations to site 484 often reduced neutralization to the same degree as removing all RBD-binding antibodies--indicating that the infection-elicited neutralizing response was highly focused on this one site. In late 2020, multiple SARS-CoV-2 variant lineages with mutations to site 484 began to emerge, suggesting that immune selection may already be shaping SARS-CoV-2 evolution. In addition to evading preexisting immunity, new SARS-CoV-2 variants may also alter the specificity of the antibody response. We compared the antibody response elicited by infection with early 2020 viruses to that elicited by the B.1.351 (Beta) variant. We found that the B.1.351 variant induces an antibody response with a shifted immunodominance hierarchy that is more focused on a different epitope (the "class 3" epitope) spanning sites 443 to 452 in the RBD. While this epitope is conserved between early 2020 and B.1.351 viruses, it is mutated in the Delta lineage, which contains an L452R mutation, and has risen to near-fixation in many countries. Thus, as SARS-CoV-2 continues to evolve, it will be necessary to reevaluate which mutations might evade immunity elicited by new variants. Overall, my thesis work has contributed to our understanding of how mutations affect the function and antigenicity of the SARS-CoV-2 RBD, and improve our ability to anticipate the consequences of the virus’ evolution.