A lentiviral vector deep mutational scanning system for studying virus evolution and escape from neutralization by antibodies and polyclonal serum

Abstract

Viral entry proteins are critical for viral replication of enveloped viruses. These proteins allow viruses to bind receptors on cells and accomplish membrane fusion. Since viral entry proteins are located on the viral surface and the surface of infected cells, they are also the targets of host immune responses. For all of these reasons, mutations to viral entry protein can have a wide range of effects. These effects include the ability to better infect new host species, changes to receptor tropism and affinity, and evasion of host immune responses. By studying the effects of mutations to viral entry proteins, we can inform monitoring of emerging viral pathogens as well as the design of vaccines and other therapeutics.Studying the effects of mutations on viral entry proteins is not straightforward. Traditional methods of measuring mutation effects on viral entry proteins have drawbacks. Often, mutations are cloned individually or in a few combinations and tested in viral replication or neutralization assays, which can be time and resource expensive. These methods may rely on using replicative viruses with full-length genomes, raising the biosafety level required to perform these experiments. Mutations also often have different effects in different strains of a virus, making it difficult to gain a full picture of the mutation’s potential effects unless these already intensive experiments are extended to multiple strains. Some of these challenges have been alleviated by high throughput techniques like phage display, yeast display, deep mutational scanning of full- length replicative viruses, and cell surface display. However, these methods each have drawbacks of their own, such as not displaying the full viral entry proteins, only being able to measure ligand or antibody binding rather than viral entry protein function, not being able to measure effects of combinations of mutations, or being difficult to perform using certain viruses due to technical reasons or biosafety concerns. We have developed a lentiviral vector-based system for deep mutational scanning of viral entry proteins that overcomes many limitations of previous studies. Non-replicative lentiviral vectors can be pseudotyped by displaying viral entry proteins from viruses of interest on their surface. We developed a method to generate a genotype-phenotype link between a lentivirus genome and a viral entry protein mutant displayed on the surfaces of the virion for large mutant libraries of viral entry proteins. These mutant lentivirus libraries can then be used to measure mutation effects on function and immune escape of full viral entry proteins. By using a nucleotide barcoding method, this system is also able to measure the effects of combinations of mutations in viral entry proteins. Measurements of effects of combinations of mutations allows us to investigate epistasis in mutation effects and map virus escape from immunity targeting multiple epitopes simultaneously. We used the lentiviral vector system to perform deep mutational scanning of the SARS- CoV-2 spike protein. This allowed us to safely measure the effects of mutations to the full SARS-CoV-2 Spike protein in a biosafety level 2 setting. Using mutant libraries of Omicron BA.1 and Delta strain spike proteins, we mapped escape mutations to neutralizing antibodies targeting various domains of the spike protein. We also compared our measurements of the functional effects of mutations to the spike protein to previous studies and natural sequence data, and found our results were more correlated with enrichment of mutations in natural sequences. This approach can be used to rapidly characterize mutation effects on function and escape from neutralization by sera or therapeutics for emerging pathogens. We also tailored the lentiviral vector system to map the neutralizing specificity of human anti-HIV sera. We designed mutant libraries of HIV Envelope mutants with combinations of mutations based on previous deep mutational scanning studies and natural sequence data. We used these libraries to map escape from neutralizing antibodies and human sera targeting CD4 binding site of HIV Envelope. Individual mutation effects and antibody epitopes were inferred using a biophysical model to deconvolute mutation effects from the combinations of mutations in the mutant libraries. Most sera mapped had neutralizing specificities similar to individually characterized monoclonal antibodies, but the neutralizing specificity of one serum was best explained by two epitopes within the CD4 binding site. Thus, this approach allows us to map multi-epitope targeting polyclonal immunity and can be used to characterize and evaluate infection or vaccination elicited polyclonal antibody responses to viral entry proteins.