Chemoenzymatic synthesis planning by evaluating the synthetic potential in biocatalysis and chemocatalysis

Chemoenzymatic synthesis integrates the advantages of chemocatalysis and biocatalysis to design efficient synthesis routes. However, current computer-assisted chemoenzymatic synthesis planning tools lack a heuristic method to unify step-by-step chemoenzymatic synthesis planning and molecule-by-molecule identification of chemo-/biocatalysis opportunities in synthesis routes. Here we develop an asynchronous chemoenzymatic retrosynthesis planning algorithm (ACERetro) which employs a search strategy that prioritizes the exploration of a molecule's promising catalytic methods. The suitability of a molecule to be synthesized via chemo- or biocatalysis is quantitatively evaluated by a data-driven Synthetic Potential Score (SPScore) using a neural network model. Additionally, the SPScore can be used to heuristically identify chemo-/biocatalysis opportunities in synthesis routes. For a given synthesis route, this algorithm uses SPScore to identify the molecules that offer optimization potential when synthesized by an alternative catalytic method, and then ACERetro is used to search synthesis routes. Case studies on synthesis planning for ethambutol and epidiolex demonstrate that our strategy can design concise chemoenzymatic synthesis routes by applying enzymatic steps to introduce stereochemistry and find shortcuts. Moreover, case studies on synthesis route optimization for rivastigmine and (R,R)-formoterol demonstrate how our strategy finds bypasses to form alternative, shorter chemoenzymatic synthesis routes. Our findings demonstrate that ACERetro with evaluating the synthetic potential of molecules represents a versatile and effective search framework for chemoenzymatic synthesis planning.