Vanillin, a plant secondary metabolite, is widely used as a flavoring agent in beverages, foods, and pharmaceuticals (Olatunde et al., 2022). The market size of vanillin is expected to reach $724.5 million by 2025 (Martău et al., 2021). Traditionally, vanillin has been primarily produced via phytoextraction or chemical synthesis to satisfy its diverse market demands (Arya et al., 2021). However, producing vanillin from phytoextraction is labor-intensive and low-yield (Liaqat et al., 2023). Furthermore, chemical synthesis of vanillin presents several challenges, such as a high environmental impact and a low-quality, with consequences on the potential restrictions imposed by food safety control authorities (Banerjee and Chattopadhyay, 2018). Thus, there is a growing interest in the sustainable and environmentally friendly biosynthesis of vanillin. In this scenario, the use of microbial cells is anticipated for the conversion or transformation of renewable biomass-based substrates into vanillin (Xu et al., 2024).

As such, by means of metabolic engineering and synthetic biology, the reconstruction of heterologous vanillin synthesis pathways into microbes is an exciting area of research, with significant progress being made in recent years (Martău et al., 2021). Inspired by the naturally occurring vanilla pathways in plants, the artificial vanillin biosynthetic system was initially developed in a recombinant Escherichia coli strain (Li and Frost, 1998). A heterogeneous carboxylic acid reductase (CAR, also referred to aldehyde dehydrogenase, or aldehyde oxidoreductase) from Neurospora crassa, was utilized for the reduction of vanillic acid to vanillin, led to a 66% overall yield (mol/mol) (Li and Frost, 1998). Afterwards, Rosazza et al. purified a CAR from Nocardia sp. strain NRRL 5646, which was found to catalyze 34 mg of vanillic acid to generate 8 mg of vanillin (Li and Rosazza, 2000). Later on, Venkitasubramanian et al. recombined a CAR, a phosphopantetheinyl tranferase and a glucose dehydrogenase (GDH) for NADPH regeneration in the whole cell reaction, achieved a production of 2 g L-1 of vanillin in 5 h (Venkitasubramanian et al., 2008). In 2009, Hansen et al. introduced a CAR derived from Nocardia and a phosphopantethelylmercaptoethyl aminotransferase of Corynebacterium glutamicum in yeasts, generated 0.045 g L-1 vanillin (Hansen et al., 2009). Moreover, CARs together with terephthalate 1, 2-dioxygenase (TPADO), dihydroxy-3, 5-cyclohexadiene-1, 4-dicarboxylate dehydrogenase (DCDDH) and catechol-o-methyltransferase (COMT) were incorporated in E. coli to directly convert plastic derivative monomer terephthalic acid into vanillin, yielding a maximum synthetic titer of 119 mg L-1 (Sadler et al., 2021). Recently, Qiu et al. successfully conducted a de novo biosynthesis of vanillin in engineered Saccharomyces cerevisiae, resulting in a vanillin yield of 8.54 mg L-1 (Qiu et al., 2022). Overall, the conversion of vanillic acid to vanillin catalyzed by CAR represents a key step in the aforementioned studies.

CARs have gained significance for its capacity to catalyze the direct reduction of various acids to the corresponding aldehydes under mild conditions, using ATP and NADPH as cofactors (Winkler, 2018, Qu et al., 2018, Derrington et al., 2019, Butler and Kunjapur, 2020). The first crystal structures of the CAR from Segniliparus rugosus (SrCAR) have been reported recently, unveiled the multidomain architectures and intricate allosteric control (Gahloth et al., 2017). Based on the X-ray structures, our group documented the conformational dynamics and catalytic mechanism of SrCAR by computational calculations (Qu et al., 2019a), and exploited the hotspot residues that may influence activity by protein-ligand interaction analysis (Qu et al., 2019b). Nevertheless, natural CARs have limitations in terms of both catalytic efficiency and stability, which require enzyme engineering to meet industrial demands (Thomas et al., 2019, Basri et al., 2022, Qin et al., 2022).

Rational design has proved to be a powerful technique to endow enzymes with new catalytic repertoires (Buller et al., 2023, Reetz et al., 2023a). It typically modifies protein structures by using saturation mutagenesis at residues proximal to the substrate binding pocket, which has been formalized as the Combinatorial Active-site Saturation Test (CAST) and Iterative Saturation Mutagenesis (ISM) (Qu et al., 2020, Reetz et al., 2023b). The combination of CAST/ISM and rational design has led to the development of a relatively new technique called Focused Rational Iterative Site-specific Mutagenesis (FRISM), which has been extensively summarized in recent reviews (Bao et al., 2023, Reetz et al., 2024, Paul et al., 2024).

In this study, we sought to elevate the activity and thermostability of SrCAR in the process of reducing vanillic acid (1a) to vanillin (1b) (Scheme 1). After one round of CAST and two additional rounds of site-specific mutagenesis in a FRISM manner, the simultaneous improvement of the dual parameters of activity and thermostability of SrCAR was achieved. The engineered variants also demonstrated efficacy in reducing other vanillic acid derivatives, followed by a scaled-up reaction.