Researchers have combined generative protein design with directed evolution to create artificial metalloenzymes driven by metal polypyridine photocatalysts, detailed in a 2026 study published in Nature by S. Mishra, Payal, D. Evans, K. Bortey, H. Bootwala, G. Gonzalez-Gutierrez, R. J. Vázquez, D. Baker, and J. C.
Scientists trying to harness synthetic metal complexes for visible-light chemistry hit a structural wall. Linking metal polypyridine photocatalysts to natural protein scaffolds to achieve high enantioselectivity proved difficult.
Computational Protein Engineering Solves Structural Walls
A research team turned to computational methods to engineer a way around the problem. By applying generative protein design, the investigators created entirely new scaffolds designed from scratch to selectively bind metal polypyridine complexes. These custom-built protein environments act as more than passive containers for the metal complex.
Two-Step Engineering Pipeline Drives the Breakthrough
The breakthrough relies on a two-step engineering pipeline described in the 2026 study titled De novo design of enantioselective artificial metalloenzyme photocatalysts.
First, generative protein design creates the de novo scaffolds required to accommodate synthetic metal photocatalysts that lack natural protein partners. Because these initial computational models only provide a starting point for binding, the researchers then deploy directed evolution to refine the structures.
Computer-Generated Scaffolding Bridges Catalysis Gap
This combination of computer-generated scaffolding and laboratory evolution bridges a major gap in asymmetric catalysis, allowing the resulting artificial metalloenzymes to support diverse light-driven reactions.
Laboratory Durability Confirms Proof-of-Concept
Beyond achieving high enantioselectivity, the newly developed artificial metalloenzymes demonstrate practical durability in the laboratory.
The study authors emphasize that this work serves as a proof-of-concept for a combined protein-design and directed-evolution strategy. The reported mechanism is specific to the designed systems and does not imply that every protein scaffold can control natural or engineered metal cofactors in the same way, nor is it evidence that arbitrary proteins or metal complexes can be converted into efficient artificial metalloenzymes.