The increasing demand for efficient and sustainable chemical processes necessitates the development of scalable biocatalytic systems for fine chemical production. Continuous flow biocatalysis offers advantages like improved reaction control, waste minimization, and energy efficiency. This approach is particularly suitable for biocatalysis, leveraging the stereo- and regioselective power of enzymes. Asymmetric reduction of activated C=C bonds (ene-reduction), catalyzed by Old Yellow Enzymes (OYEs), is a valuable reaction for generating stereogenic centers. OYEs utilize a flavin mononucleotide (FMN) cofactor and typically require NAD(P)H for FMNH₂ regeneration. However, OYEs exhibit promiscuity towards reductants, including reduced free flavins. Alternative regeneration methods like photochemical and electrochemical approaches have limitations in conversion rates and stability. *In vivo* systems show high conversion rates but present downstream processing challenges. This study aims to develop a scalable *in vitro* flow system for flavin-based biocatalysis, utilizing molecular hydrogen (H₂) as a sustainable and waste-free reductant produced by water electrolysis. Molecular hydrogen (H₂) is an ideal reductant due to its non-wasteful nature and production via water electrolysis using renewable energy. Soluble hydrogenase from *Cupriavidus necator* (SH) can reversibly oxidize H₂ to reduce NAD⁺ to NADH, exhibiting high atomic efficiency and oxygen tolerance. SH's NAD⁺-reducing ability can be extended to NADPH via mutagenesis. Coupling SH with thermostable OYE from *Thermus scotoductus* (TSOYE) allows FMNH₂ regeneration for C=C double bond reduction. *In situ* H₂ production via water electrolysis minimizes H₂ handling risks. While photoelectrochemical water oxidation has been explored, its limited efficiency poses upscaling challenges. Previous work demonstrated electrically driven NADH-dependent biocatalytic processes in flow using H₂ as a mediator, but these methods involved less efficient gas transfer. This research addresses these limitations by developing a closed-loop flow system with efficient H₂ transfer for electro-driven flavin-dependent biocatalysis.

The literature extensively covers continuous flow biocatalysis as a sustainable manufacturing method in pharmaceuticals and fine chemicals. Advancements in flow chemistry and enzyme immobilization techniques have made biocatalysis in continuous flow increasingly feasible. Studies highlight the use of OYEs for asymmetric alkene reduction, but traditional cofactor regeneration methods using NAD(P)H are costly and may generate waste. Alternative methods for regenerating flavin cofactors include photochemical and electrochemical approaches, but these methods face challenges in scalability and efficiency. The use of hydrogenases for cofactor regeneration has gained traction due to the atom-efficient nature of H2 and its sustainable production via water electrolysis. Several studies demonstrate the successful use of soluble hydrogenases for regenerating NAD(P)H and reduced flavin cofactors in batch reactions. The use of PEM electrolyzers for H2 production in flow systems offers advantages in terms of high current densities, instant current response and coupling to renewable energy sources compared to other methods. However, efficient integration of H2 produced by PEM electrolyzers into biocatalytic flow systems remains a challenge, and this paper aims to address this gap.

A closed-loop flow system was designed to facilitate electro-driven enzymatic reactions. The system incorporates a commercially available PEM electrolyzer (IrRuOx anode, Pts/Ptc cathode) for H₂ production. The generated H₂ is introduced into the flow system via a gas addition module containing gas-permeable tubing (PVMS or PTFE). PVMS tubing exhibited higher H₂ transfer efficiency than PTFE. Thermostable TsOYE was immobilized via coordination bonds, and SH was immobilized by affinity binding. Catalase was included to mitigate ROS formation. Online monitoring of H₂, O₂, and FMN was performed to assess system functionality. Different substrates (cyclohexenone, ketoisophorone, (S)- and (R)-carvone) were used to evaluate the system's adaptability. Immobilization of SH and TsOYE was achieved using Strep-Tactin XT 4flow resin and EziG beads respectively for simplified product isolation and enzyme reusability. The system's reusability and scalability were investigated by performing multiple reaction cycles and by scaling up the reaction volume from 17 mL to 185 mL. Analytical techniques employed included GC-FID for conversion rate determination, GC analysis for enantiomeric excess, GC-MS for side product identification, and NMR for product confirmation. Faradaic efficiency was calculated to assess the system’s energy efficiency. The E factor was calculated to quantify waste production.

The developed closed-loop flow system consistently achieved >99% conversion of ketoisophorone to levodione over seven cycles, demonstrating high stability and reusability. High turnover numbers (TTN) were observed for both SH-Tactin and TsOYE-EziG, indicating robust biocatalyst immobilization. The system showed versatility in reducing various cyclic enones, though the optical purity of chiral products was lower than anticipated, potentially due to non-enzymatic racemization in water. Upscaling the reaction to 185 mL maintained high conversion, yielding 471 mg of levodione with 96% purity. The Faradaic efficiency was 0.15%, limited by the H₂ gas outflow requirement and the high voltage needed to overcome internal resistance in the commercial PEM electrolyzer. The E factor was calculated to be 6.0 and 5.7 for the 17 mL and 185 mL reactions respectively, indicating low waste production. The system demonstrated a high TTN for both enzymes, especially when compared to the performance of previous studies. Strep-Tactin resin and EziG beads are highlighted as being robust immobilization carriers for SH and TsOYE, respectively, compared to other methods like covalent irreversible immobilization. Online monitoring provided valuable insights into the reaction progress, allowing for assessment of enzyme activities and identification of potential bottlenecks.

The results demonstrate the feasibility of a sustainable, scalable platform for flavin-dependent biocatalysis using H₂ generated from water electrolysis. The high conversion rates, reusability of the immobilized enzymes, and successful upscaling suggest the system's potential for industrial applications in fine chemical synthesis. The lower-than-expected optical purity of chiral products highlights the need for further optimization, potentially involving the incorporation of immiscible organic solvents to enhance substrate solubility and mitigate racemization. The relatively low Faradaic efficiency indicates areas for improvement in the system's energy efficiency, perhaps by utilizing zero-gap cells to reduce energy consumption. The observed side reactions, especially with cyclohexenone, highlight the importance of further investigation into the optimization of reaction conditions. The high TTN values obtained for the enzymes demonstrate a considerable improvement over previous batch processes.

This study successfully developed a closed-loop flow system for electro-driven flavin-dependent biocatalysis, utilizing H₂ from water electrolysis. The system demonstrates high conversion, reusability, and scalability. Future work should focus on optimizing reaction conditions to improve optical purity and Faradaic efficiency, and exploring the adaptability of the system to other gas-dependent enzymes.

The main limitations include the relatively low Faradaic efficiency, which can be improved by implementing zero-gap cells, and the lower-than-expected optical purity of chiral products, which may require modifications to reaction conditions or solvent systems. The system's current design uses aqueous solutions, which can lead to side reactions and reduce optical purity for some substrates. Further research is needed to fully understand the differences in enzyme behavior between flow and batch reactions to further optimize conditions.