Biophysical fitness landscape design traps viral evolution

Abstract Evolutionary adaptation is often visualized as a population’s stochastic climb toward the top of a fitness landscape. While there exist approaches to design or synthetically evolve proteins into desired structures, there is a lack of methodology for designing, tuning, and quantitatively reshaping the fitness landscapes themselves on which protein evolution takes place. Here, we introduce foundational principles of fitness landscape design (FLD) to customize the structural peaks and valleys of biophysical fitness landscapes with quantitative accuracy, offering robust control of long-term evolutionary outcomes. Our FLD algorithms use stochastic optimization of a chemically derived biophysical fitness model to consistently discover optimal antibody ensembles which force a target protein to evolve according to a user-specified target fitness landscape. We then apply FLD to suppress the fitnesses of two SARS-CoV-2 genotype neutral networks and to discover proactive vaccines that preemptively restrict escape variant fitness trajectories before they arise. Significance Statement Rapidly evolving viruses mutate to escape antibodies generated by the human immune system, leading to periodic waves of infection and death. Modern vaccine design approaches that focus on currently prevalent strains can be vulnerable to emerging escape mutations. An ideal strategy for proactive vaccine design requires not only immediate effectiveness, but also control over the viral fitness landscape to ensure optimal suppression of escape variants. Here, we introduce biophysics-based computational algorithms to discover optimal antibody ensembles that quantitatively reshape the viral fitness landscape to induce long-term suppression of fitness trajectories. These protocols, called biophysical fitness landscape design, open the door to improved pandemic preparedness via proactive vaccine, antibody, and peptide design, thinking several steps ahead of pathogen evolution. Competing Interest Statement The authors have declared no competing interest. Footnotes In vivo model, experimental and epidemiological data. New figures and text.