Abstract
Synthetic signal processing in cells can be achieved through complex networks of molecules or simpler, single-molecule-based mechanisms. The latter approach stands out for its reduced genetic load and ease of delivery. Multivalent protein binders are such single-molecule signal processors that exhibit distinct behaviors when interacting with cells presenting various antigen profiles. However, a quantitative and precise method to design multivalent interactions in the context of cell surface binding is still lacking. Here, we developed Multivalent Antigen Sensing Simulator (MASS) to guide the design of such multivalent proteins. We verified the model through in vitro and cell binding experiments, showing the predictive accuracy of our model in designing multivalent binders to sense cell surface antigen identities and elicit selective killing. Our model also accurately captured, which was verified via experiments, the nonmonotonic interplay between valency and monovalent affinity in designing multivalent binders that sense cell surface antigen densities. Our work suggests practical guidelines in designing multivalent proteins that could simultaneously sense both antigen identities and densities and enables fast and precise in silico prescreening of multivalent protein binders.