Directed evolution of enzymes at the crossroads of tradition and innovation.

Abstract

Directed evolution has become a central methodology for engineering proteins with improved or entirely new functions, enabling applications across biotechnology, medicine, and synthetic chemistry. By iteratively coupling genetic diversification with screening or selection, directed evolution allows functional optimization even when detailed structural or mechanistic knowledge is unavailable. While display-based selection platforms have enabled the efficient evolution of binders from extremely large libraries, enzyme evolution relies primarily on quantitative screening strategies that preserve genotype-phenotype linkage, often through compartmentalization. This review focuses primarily on enzyme directed evolution, using binder evolution as a comparative reference point to highlight key methodological differences and parallel advances. Major technological advances-including in vitro emulsions, droplet microfluidics, ultrahigh-throughput sorting, genetically encoded biosensors, and alternative detection modalities-have dramatically expanded screening capacity and analytical resolution. We also discuss why stability remains a central constraint on evolvability, why assay design continues to limit translational relevance, and how failures such as surrogate-substrate bias, droplet leakage, tracking errors, and overfitted machine-learning models can misdirect campaigns. By integrating classical strategies with emerging continuous and data-driven approaches, enzyme directed evolution is moving toward more predictive, automated, and industrially translatable workflows.