Timed batch inputs unlock substantially higher yields for enzymatic cascades

Cell-free enzymatic reaction networks (ERNs) enable the production of value-added compounds in a single reaction. However, allosteric interactions, product inhibition, reversibility and competition for shared cofactors only become apparent once the ERN is assembled. These emergent dynamics create kinetic barriers that limit the overall yield. Here we introduce a model-guided optimal design strategy to generate time-dependent ‘recipes’ for batch reactions, in which every component can be added repeatedly at specified amounts, at any time. We apply the method to two ERNs: the pentose phosphate pathway and a branched nucleotide salvage pathway. In the pentose phosphate pathway, optimized inputs increased AMP production up to 5.7-fold and raised glucose-to-product conversion from ~12% in the control to ~48% using a time-dependent input. In the salvage pathway, time-dependent dosing balanced competing branches and increased UTP yield ~21-fold relative to composition-matched all-at-once dosing. Timed batch inputs provide a generally applicable route to optimizing a complex reaction sequence. Optimizing the performance of complex enzymatic reaction networks remains challenging as hidden interactions and unfavourable kinetic barriers often arise when enzymes operate together in networks. Now a model-guided design strategy for generating time-dependent protocols for batch reactions has been developed. This strategy offers a general approach for optimizing multi-step reaction sequences.