Oxidoreductases have broad applications in industrial synthesis of various chemicals, encompassing enzymatic preparation of sugar alcohols [1], reduction of amines to produce α-amino acids by amino acid dehydrogenases [2], etc. Compared with chemical methods, enzyme-catalyzed synthesis provides benefits such as gentle reaction conditions, high catalytic efficiency, and precise stereo-/regio- specificity. Despite aforementioned merits, enzymatic redox reactions come with a significant challenge in applications. Nicotinamide cofactors NAD(P)H/NAD(P)+ which consumed in enzymatic redox reactions are in high demand across the majority of industrial processes especially cell-free bioprocesses. However, substantial cost is associated with the production of unstable natural cofactors. Thus, imperative pursuit of cofactor regeneration is crucial for the practical application of oxidoreductases. Various means for cofactor recycling have been developed, including homogeneous catalysis [3], heterogeneous catalysis [4], electrochemical approaches [5], photochemical techniques [6], and enzymatic methods [7].

Enzymatic cofactor recycling remains a preferable choice in large-scale applications, due to its mild reaction conditions and high catalytic efficiency. However, addition of extra enzymes and co-substrates for cofactor regeneration also leads to production of byproducts, which complicates subsequent product recovery [8]. In photochemical methods, artificially synthesized chemical catalysts are adopted to absorb light energy for cofactor regeneration, eliminating the necessity for additional enzyme reactions. Additionally, the photochemical catalysts used in the reaction can be easily separated and recycled, representing a cost-effective and environmentally friendly option compared with other methods [9]. Various types of photocatalysts have been reported for cofactor regeneration. For example, an artificial photosynthetic system was constructed based on porphyrin. Porphyrin acts as a photosensitizer, attached to SiO2 microspheres coated with polydopamine and polyethyleneimine containing thiol groups. Under optimal conditions, 81.5 % cofactor regeneration yield was achieved using this artificial photosynthetic system after 60 min reaction [10]. The photocatalytic antenna-reactor system, utilizing gold-rhodium nanoflowers (Au@Rh NFs), efficiency by introducing favorable electrostatic interactions in enhance electron transfer. Using this system, the photocatalytic regeneration yield of NADH reached 30 % [11]. Another approach involves preparation of carbon nanodot-SiO2 hybrid photocatalyst through a reverse microemulsion method using non-ionic surfactants. The cofactor regeneration yield of 74 % was reached in a photocatalyst-biocatalyst coupled system [12]. Additionally, non-metallic polymer semiconductor materials such as g-C3N4 have received considerable attentions in photocatalysis due to their excellent stability and favorable band edge positions. Meng et al. used g-C3N4 to regenerate NADH, the yield of NADH reached 27.8 % in 30 min [13]. Although the efficiency of NADH regeneration in the g-C3N4 system could be comparatively lower than that of other photocatalytic systems, its intrinsic characteristics, such as non-toxicity, cost-effective synthesis, high stability in acidic/alkaline environments, and responsiveness to visible light, position it as a promising photocatalyst.

The progress of photocatalytic reactions requires not only photocatalysts, but also electron mediators (EM) that transfer photoexcited electrons and electron sacrificial agents to fill photoexcited holes. The strong functionality of organometallic complex Cp*Rh(bpy)H2O2+ makes it an excellent EM [14]. It exhibits high stability and activity under different reaction conditions (wide pH and temperature range), and is highly active to nicotinamide cofactors. Besides this, it exhibits high universality toward a variety of reducing equivalents, which enables it to capture photo-excited electrons from various photocatalysts. In addition, there are many options for electronic sacrificial agents, among which triethanolamine (TEOA) is widely used due to its economic and efficient characteristics. TEOA can produce photoinduced homeostasis of RO-COOH followed by a hydrogen atom transfer (HAT) process under UV light irradiation, which can interfere with the photocatalytic reaction. It is however a preferable choice in photocatalytic reactions without the use of UV light [15].

Compared with mild enzymatic regeneration reaction conditions, photocatalytic regeneration may pose challenges to the stability of cofactors. Synthetic nicotinamide cofactors have substantial potential for economical and biorthogonal cofactor regeneration compared with costly and delicate natural cofactors, owing to their robustness and easy preparation [16]. In recent years, NCBs have been demonstrated as promising substitutes for natural cofactors in redox catalysis. Seiber et al. synthesized a simple nicotinamide cofactor mimic P2NAH, retaining the nicotinamide moiety but replacing the remaining part with a single phenylpropyl. Enoate reductase (TsER, from Thermus scotoductus) and glucose dehydrogenase (SsGDH, from Sulfolobus solfataricus) could utilize the biomimetic P2NAH as a cofactor to produce 2-methylbutanal, demonstrating it is applicable in enzymatic reactions [17]. Additionally, NCBs have been demonstrated to serve as redox intermediate in reactions catalyzed by OYE [18]. By utilizing NCBs, OYE can catalyze the reduction of α, β-unsaturated carbonyl compounds.

Here, a photocatalytic system mediated by g-C3N4 was constructed for the regeneration of NAD+ and four distinct NCBsox. In this cofactor regeneration system, the semiconductor g-C3N4 served as a photocatalyst, being photoexcited to generate photoelectrons. TEOA acted as the sacrificial electron donor, supplying electrons to the entire reaction system. Cp*Rh(bpy)H2O2+ functioned as EM, absorbing electrons from the photocatalyst and selectively catalyzing the reduction of NAD+ and NCBsox. The regeneration efficiency among NCBs and natural cofactor were evaluated in terms of yield, total turnover number (TTN) and turnover frequency (TOF), etc. Furthermore, the feasibility of photocatalytic cofactor regeneration was validated by coupling with OYE (XenA from Pseudomonas putida) catalyzed reduction of 4-Ketoisophorone, and was also evaluated in terms of yield, TTNXenA, and TOFXenA.