Abstract
ABSTRACT
Although viral antigenic evolution is shaped by the competing demands of immune evasion and receptor-binding fitness, the physical principles underlying divergent viral evolutionary dynamics remain poorly understood. Here, by integrating structural biology, deep mutational scanning, and genomic surveillance data, we show that receptor-binding geometry constrains the spatial coupling between antibody escape and intrinsic viral fitness. In SARS-CoV-2, the exposed convex Spike-ACE2 interface spatially overlaps dominant neutralizing epitopes with receptor-contact residues, producing a stringent fitness-evasion trade-off. Consequently, viable escape mutations are largely restricted to a narrow peripheral region surrounding the receptor-binding core. As population immunity progressively saturates this limited escape space, continued adaptation increasingly requires coordinated saltational remodeling of the receptor-binding interface. In contrast, the influenza H3N2 hemagglutinin binds the small sialic acid ligand through a recessed pocket architecture that partially separates dominant antibody pressure from receptor-binding fitness constraints. This topological decoupling permits continuous stepwise antigenic drift without requiring frequent remodeling of the core binding site. Building on these findings, we develop a lightweight geometry-based framework that accurately identifies receptor-competitive antigenic hotspots across divergent viruses. Together, our results support a general model in which receptor geometry and protein surface topology strongly bias accessible antigenic evolutionary trajectories.
IMPORTANCE
Why some viruses evolve through gradual antigenic drift whereas others undergo recurrent evolutionary leaps remains a fundamental question in virology. By integrating protein structures, mutational scanning data, and genomic surveillance, we show that the geometry of the receptor-binding interface strongly influences how immune escape mutations affect viral fitness. In SARS-CoV-2, receptor-binding residues are highly exposed and overlap with dominant antibody targets, restricting immune escape to a narrow set of mutations and promoting periodic evolutionary leaps when this space becomes exhausted. In contrast, the receptor-binding pocket of influenza A virus H3N2 is partially shielded from antibody recognition, allowing continuous antigenic drift with fewer fitness constraints. We further demonstrate that these structural principles can be used to identify antigenic hotspots directly from virus-receptor complexes. These findings provide a mechanistic explanation for divergent viral evolutionary trajectories and offer a structural framework for studying and anticipating antigenic evolution.