Efficient conversion of aromatic and phenylpropanoid alcohols to acids by the cascade biocatalysis of alcohol and aldehyde dehydrogenases

Abstract

Benzyl and phenylpropanoid acids are widely used in organic synthesis of fine chemicals, such as pharmaceuticals and condiments. However, biocatalysis of these acids has received less attention than chemical synthesis. One of the main challenges for biological production is the limited availability of alcohol dehydrogenases and aldehyde dehydrogenases. Environmental microorganisms are potential sources of these enzymes. In this study, 129 alcohol dehydrogenases and 42 aldehyde dehydrogenases from Corynebacterium glutamicum, Pseudomonas aeruginosa, and Bacillus subtilis were identified and explored with various benzyl and phenylpropanoid alcohol and aldehyde substrates, among which four alcohol dehydrogenases and four aldehyde dehydrogenases with broad substrate specificity and high catalytic activity were obtained. Moreover, a cascade whole-cell catalytic system including ADH-90, ALDH-40, and the NAD(P)H oxidase LreNox was established, which showed high efficiency in converting cinnamyl alcohol and p-methylbenzyl alcohol into the respective carboxylic acids. Remarkably, this biocatalytic system can be easily scaled up to gram-level production, facilitating preparation purposes.

Graphical abstract

1. Introduction

Benzyl compounds and phenylpropanoids serve as crucial intermediates in the production of fine chemicals and active metabolites that have broad applications and high industrial value [1,2]. Their aldehydes, such as vanillin, benzaldehyde, and cinnamaldehyde, are prized for their aromatic properties and account for the fragrance industries, which have an estimated global market value of approximately $ 1 billion [3]. Their carboxylic acids, including salicylic acid, ferulic acid, gallic acid, and p-coumaric acid, are commercially important compounds that have diverse uses in the food, pharmaceutical, cosmetic, and chemical industries [4]. Derivatives of benzyl and phenylpropanoid acids, such as flavonoids and resveratrol, are noteworthy for their remarkable health benefits and medicinal activity [5]. The market share of flavonoids is projected to reach $ 3.5 billion by 2025 [6], while resveratrol is anticipated to attain a market share of $ 99.4 million [7].

Carbonyl compounds, including benzyl and phenylpropanoid aldehydes and carboxylic acids, can be produced by the oxidation of alcohols, and the catalytic synthesis process has attracted considerable attention from organic chemistry chemists and biochemists. Selective oxidation of benzyl alcohols and phenylpropanoid alcohols can afford their benzyl and phenylpropanoid aldehyde analogues, which can be further oxidized to yield the corresponding benzyl and phenylpropanoid carboxylic acids [8,9]. Direct selective oxidation of alcohols and aldehydes in a chemical process is well-established [[10], [11]], such as chromium and manganese oxide catalysis [12,13], Dess-Martin oxidation [14], Swern oxidation [15], and noble metal catalyst systems [16]. However, there are fewer reports on biocatalysis, and achieving high-efficient and green biocatalysis remains a significant challenge.

Alcohol dehydrogenase (ADH) is a member of the dehydrogenase family found in various organisms. It catalyzes the oxidation of primary and secondary alcohols into the corresponding aldehydes or ketones [17]. Aldehyde dehydrogenase (ALDH) belongs to the nicotinamide adenine dinucleotide (phosphate) (NAD(P))-dependent enzyme family. It oxidizes a wide range of aliphatic and aromatic aldehydes, both endogenous and exogenous, to form the corresponding carboxylic acids [18]. The biocatalytic potential of ADH and ALDH is particularly attractive for the selective cascade oxidation of benzyl alcohols and phenylpropanoid alcohols to produce corresponding carboxylic acids [19]. Environmental microorganisms, such as Corynebacterium [20], Pseudomonas [21], and Bacillus [22], exhibit remarkable abilities to degrade and oxidize phenolic compounds through catalysis of alcohols, aldehydes and acids, highlighting their potential as candidate dehydrogenases [[23], [24], [25], [26], [27], [28]]. Exploring potential dehydrogenases in environmental microorganisms opens up opportunities for the efficient biocatalysis of these compounds [29].

In this study, the genomes of representative environmental microorganisms C. glutamicum ATCC 13032, P. aeruginosa PAO1, and B. subtilis subsp. 168 were analyzed, and 129 alcohol dehydrogenases and 42 aldehyde dehydrogenases were identified. Representative benzyl and phenylpropanoid alcohols and aldehydes were tested as substrates, and 4 ADHs and 4 ALDHs with broad substrate specificity and high catalytic activity were obtained. Furthermore, a whole-cell cascade biocatalytic system including the optimal dehydrogenases ADH-90 and ALDH-40, as well as the NAD(P)H oxidase LreNox, which was introduced to regenerate NAD⁺ and NADP⁺, was established. Efficient production of benzyl and phenylpropanoid acids was achieved and successfully scaled up to the gram level. Our results provide new insights into the design and construction of other biological redox reactions for natural product synthesis.

2. Materials and methods

2.1. Strains, plasmids and culture conditions

The strains and plasmids used in this study are listed in Table S1 and Table S2. E. coli DH5α served as the host for gene cloning and plasmid construction, and was grown on Luria-Bertani (LB) agar plates or LB broth. For protein expression, purification, and biotransformation, E. coli BL21 (DE3) was used and grown in LB or Terrific Broth (TB) medium (1.2% tryptone, 2.4% yeast extract, 0.94% K₂HPO₃, 0.22% KH₂PO₃, 4% glycerol). Unless otherwise stated, all E. coli strains were incubated at 37 °C, with appropriate antibiotics added to the broth as needed.

2.2. Molecular manipulation, chemicals and reagents

Standard molecular cloning techniques including gene cloning, plasmid transformation, and agarose gel electrophoresis were performed following standard protocols. Primers were listed in Table S2, which were synthesized by Sangon Biotech (Shanghai, China) The primary screened ADHs and ALDHs were amplified from the genome by PCR and cloned between the EcoRI and NdeI sites of pET30a using restriction enzyme ligation. The LreNox from Lactobacillus reuteri was amplified using primers P347 and P348. The pET-28b plasmid and LreNox gene were digested with EcoRI (NEB, Frankfurt, Germany) and NdeI (NEB, Frankfurt, Germany) at 37 °C for 1 h for linearization, and subsequently constructed into pET28a-LreNOX vector using T4 DNA ligase. The construction method of pETDuet-1-ADH-90-2-ALDH-40 is the same as above. Firstly ADH90 was cloned between HindIII and BamHI of pETDuet-1 to construct pETDuet-1-ADH-90. Then, ALDH40 was cloned between NdeI and XhoI of pETDuet-1-ADH-90 to construct pETDuet-1-ADH-90-2-ALDH-40. The recombinant plasmid was transformed into E. coli BL21 (DE3) to generate the recombinant strain. E.Z.N.A.™ Plasmid Miniprep Kit (Omega Biotek, Norcross, GA) was used for plasmid isolation. E.Z.N.A.™ Gel Extraction Kit (Omega Biotek, Norcross, GA) was used for DNA fragment purification. T3 Super PCR Mix (Tsingke Biotechnology, Beijing, China) was used for colony PCR. PrimeSTAR (Takara Bio) or Phanta Max Super-Fidelity DNA polymerase (Vazyme Biotech, Nanjing, China) was used for other conventional PCR amplifications. ClonExpress Ultra One Step Cloning Kit (Vazyme Biotech, Nanjing, China) was used for plasmid construction. All substrates and chemicals were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd., Shanghai Macklin Biochemical Technology Co., Ltd. or Sigma Aldrich (Shanghai) Trading Co., Ltd.

2.3. Protein expression, purification and concentration determination

A single colony of E. coli BL21 (DE3) harboring the specific expression plasmid was grown in 5 mL LB broth with 50 μg/mL kanamycin and incubated at 37 °C, 220 rpm, overnight. Then, the overnight culture (500 μL) was inoculated into 500 mL of fresh TB broth containing kanamycin and incubated at 37 °C, 220 rpm. When the cell density reached an OD₆₀₀ of 0.6–0.8, IPTG was added to a final concentration of 0.2 mM. The cells were continued to grow at 18 °C for 24 h and then harvested by centrifugation at 6000 g, 4 °C. The cell pellets were stored at −80 °C for later use.

All protein purification procedures were performed at 4 °C. Briefly, the cell pellet was resuspended by vortexing in 50 mL of lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10% glycerol, and 10 mM imidazole, pH 8.0). After sonication (5 s on, 5 s off, 30 min total), the crude cell lysate was centrifuged at 12,000×g for 30 min. The supernatant fraction was collected, mixed with 1 mL of Ni-NTA resin (Qiagen, Hilden, Germany), and incubated at 4 °C for 30 min on a gentle rotator. The slurry was then loaded onto an empty column for protein purification. The resin was washed with 100 mL of wash buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10% glycerol, and 20 mM imidazole, pH 8.0) until no protein is detectable in the flow-through by Coomassie Brilliant Blue G250 assay. The target protein bound to Ni-NTA resin was eluted with 10 mL of elution buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10% glycerol and 250 mM imidazole, pH 8.0). The eluate was concentrated at 5000 g for 30–60 min using an Amicon Ultra centrifugal filter (Merck KGaA, Darmstadt, Germany). Next, the protein solution was loaded onto a pre-equilibrated PD-10 column (GE Healthcare, Buckinghamshire, UK) for buffer exchange with 5 mL of desalting buffer (50 mM NaH₂PO₄, 10% glycerol, pH 7.5). Finally, the desalted protein fraction was aliquoted, snap-frozen with liquid nitrogen, and stored at -80 °C. Protein concentrations were determined by the Bradford method using BSA as a standard.

2.4. Preparation of a crude enzyme extract

Sterilized 96-well deep well plates were filled with 300 μL of LB medium containing 50 μg/mL kanamycin. Single colonies were inoculated and grown at 37 °C and 220 rpm for 12 h. Then, 10% of the overnight culture was transferred to a new 96-well deep well plate with 400 μL of TB medium containing 50 μg/mL kanamycin per well. The plate was incubated at 37 °C and 220 rpm for 4 h. After that, IPTG was added to a final concentration of 0.2 mmol/L and the plate was incubated at 20 °C and 220 rpm for 20 h. The cells were harvested by centrifugation at 3700 g for 10 min and resuspended in 200 μL of potassium phosphate buffer (50 mmol/L, pH 7.4) containing 100 mg/L lysozyme, 300 U/mL DNase I and 10% Triton X-100. The plate was centrifuged, and the supernatant was collected as the crude enzyme extract.

2.5. Enzymatic assays and kinetics determination

To screen ADHs and ALDHs initially, the in vitro enzymatic reaction system consisted of 1 mM substrate, 1 mM NAD⁺, 1 mM NADP⁺, and 150 μL crude enzyme. To test the substrate promiscuity of pure enzymes, the in vitro enzymatic reaction system contained 10 μM ADH or ALDH, 10 μM LreNox, 1 mM substrate, 1 mM NAD⁺, and 1 mM NADP⁺ for 3 h.

To test the thermostability and kinetic parameters, the in vitro enzymatic reaction system contained 5 μM ADH-90 or 0.5 μM ALDH-40, 1 mM substrate, and 2 mM NAD⁺ at seven temperatures of 0, 25, 30, 35, 40, 45, and 50 °C for 30 min. To test the cofactor preference, the in vitro enzymatic reaction system contained 0.5 μM ADH or ALDH, 1 mM substrate, and 2 mM NAD⁺ or NADP⁺ for 10 min. All screening enzymatic assays were performed in 100 μL of 50 mM Tris·HCl buffer (pH 8.0) at 30 °C. The reaction was quenched with two volumes of methanol. The enzyme concentrations of ADH-90 and ALDH-40 in kinetics assays were 5 μM and 0.5 μM, respectively. The triplicated data were fitted to the Michaelis-Menten equation for calculating the k_cat and Km values of each substrate using Origin 8.5.

2.6. Whole-cell transformation

Recombinant E. coli were incubated overnight at 37 °C and 220 rpm. Then, the overnight culture (500 μL) was inoculated into fresh TB broth containing the corresponding antibiotic and incubated at 37 °C and 220 rpm. When the cell density reached an OD₆₀₀ of 0.6–0.8, IPTG was added to a final concentration of 0.2 mM, and the cells were allowed to grow at 18 °C for 24 h. Afterward, cells were harvested by centrifugation at 6000 g and 4 °C, and washed twice with 0.9% NaCl solution. The bacterial cells were resuspended in a solution of OD₆₀₀ = 27 for following whole-cell catalysis. A 10 mL volume of the cell suspension was used in a 50 mL Erlenmeyer flask at 30 °C and 220 rpm for the catalytic test. During sampling, 0.5 mL of the transformation solution was mixed with 0.5 mL of methanol and shaken for 1 h, followed by centrifugation at 12,000 rpm for 10 min. The supernatant was then analyzed using HPLC (three parallel experiments were performed).

2.7. Analytical methods

Samples were analyzed on an Agilent 1260 system equipped with a photodiode array detector. Separation was achieved within 15 min using a linear mobile phase gradient on a YMC Triart-C18 column (150 × 4.6 mm, 5 μm). The solvent system consisted of water with 0.1% TFA (solvent A) and acetonitrile (solvent B), and the following linear gradient program was used: 5% buffer B (0–2 min), 5–90% linear gradient of buffer B (2–12 min), 90 to 5% linear gradient of buffer B (12–13 min), and 5% buffer B (13–15 min). The flow rate was set at 1 mL/min, and the injection volume was 10 μL. Full-wavelength scans (210–400 nm) were recorded. Substrate consumption and product formation were quantified by HPLC peak area integration using corresponding authentic compounds as standards.

3. Results and discussions

3.1. Genome mining of environmental microorganisms for alcohol dehydrogenases and aldehyde dehydrogenases

Environmental microorganisms are crucial for ecosystem functioning, as they participate in biogeochemical cycles of various elements, such as carbon, nitrogen, and sulfur [30]. Moreover, they have strong redox capabilities, which enable them to perform functions such as bioremediation, organic matter decomposition, pollutant degradation, and restoration of degraded land [31]. C. glutamicum, P. aeruginosa, and B. subtilis, isolated from sewage and other environments, can efficiently degrade phenol [32,33], formaldehyde [34], p-nitrophenol [35], p-cresol [36], and diverse polyphenolic compounds [33]. These processes often involve dehydrogenase enzymes that catalyze the oxidation of hydroxyl groups to carbonyl groups or further to carboxyl groups. Some of these dehydrogenases, such as ADH CreG (for oxidation of benzyl alcohols) and ALDH CreC (for oxidation of benzyl aldehyde) from C. glutamicum [36], and ADH AdhA (for biotransformation a wide variety of aromatic alcohols) from R. opacus [37], have been studied for their roles and mechanisms in the metabolism of phenolic compounds. These enzyme resources in the genomic data of environmental microorganisms provide a material basis for mining efficient dehydrogenases.

The ADHs include two main families differing in their structural features: the “short-chain” enzymes have a single domain of enoyl-(acyl carrier protein) reductase, adh_short_C2 (PF13561), while the “medium-chain” zinc enzymes have two domains, namely a GroES-like alcohol dehydrogenase domain, ADH_N (PF08240), and a zinc-binding dehydrogenase domain, ADH_zinc_N (PF00107) [38]. On the other hand, the ALDHs share one common domain of aldehyde dehydrogenase family, Aldedh (PF00171), which is responsible for the oxidation of aldehydes to carboxylic acids [18]. To identify ADH genes in the genomes of P. aeruginosa PAO1, C. glutamicum ATCC 13032, and B. subtilis subsp. 168, we conducted Hidden Markov Model (HMM) searches using the ADH Pfam models of adh_short_C2 (PF13561), ADH_N (PF08240), and ADH_zinc_N (PF00107) as queries. This resulted in a total of 129 candidate ADHs (Tables S3 and S4), comprising 89 “short-chain” type and 40 “medium-chain” type enzymes (Fig. 1). Similarly, we identified 42 candidate ALDHs by querying the genomes with the ALDH Pfam model of Aldedh (PF00171) (Fig. 1, Table S5). Then, all candidate ADHs and ALDHs were amplified from the genomes, cloned into pET30a and transformed into E. coli BL21(DE3) to obtain heterologous expression strains (namely BADH1-BADH129 and BALDH1-BALDH42, respectively; Table S1).

Fig. 1.

Genome mining of P. aeruginosa PAO1, C. glutamicum ATCC 13032, and B. subtilis subsp. 168 for ADHs and ALDHs.

3.2. Activity screening of alcohol dehydrogenases and aldehyde dehydrogenases

Single clones of strains BADH1-BADH129 and BALDH1-BALDH42 were picked and cultured overnight to make seed liquid, which was then inoculated into TB medium. The strains were grown in shake flasks at 37 °C until the mid-logarithmic growth phase, and the enzyme expression was induced at 18 °C for 24 h. Cells were then harvested, and the crude enzyme solutions were obtained by cell lysis and used for the screening assays. p-Hydroxybenzyl alcohol and cinnamyl alcohol were chosen as representative substrates of benzyl alcohols and phenylpropanoid alcohols for the activity screening of ADHs, respectively. Among the tested ADH enzymes, ADH-49, ADH-64, ADH-88, ADH-90, and ADH-96 showed high activity in oxidizing p-hydroxybenzyl alcohol to p-hydroxybenzaldehyde, with over 30% conversion rate (Fig. 2a). Moreover, ADH-49, ADH-88, ADH-90, and ADH-96 also exhibited remarkable activity in oxidizing cinnamyl alcohol to cinnamaldehyde (Fig. 2b). Consequently, these 4 ADHs were selected as potential candidates for further characterization. Similarly, p-hydroxybenzaldehyde and cinnamaldehyde were used as test substrates for ALDHs, resulting in the identification of four potential ALDHs (ALDH-11, ALDH-25, ALDH-26, and ALDH-40) with significant catalytic activity among 42 ALDH candidates. The four ALDHs were able to oxidize p-hydroxybenzaldehyde and cinnamaldehyde into corresponding acid products with >70% and 100% conversion ratios, respectively (Fig. 2c and d). We then investigated the cofactor preferences of the top enzymes ADH-49, ADH-88, ADH-90, ADH-96, ALDH-11, ALDH-25, ALDH-26 and ALDH-40 using either NADP⁺ or NAD⁺. ADH-88, ADH-90 and the four ALDH could use either NADP⁺ or NAD⁺ (Fig. S6). Both ADH-49 and ADH-96 could only use NAD⁺ as cofactor. ALDH-11, ALDH-26, ALDH-40 and ADH-90 showed a strong preference for NAD⁺ over NADP⁺, while ADH-88 exhibited increased activity with NADP⁺ compared to NAD⁺. These results demonstrate the feasibility of screening and identifying novel ADHs and ALDHs from genomic library of environmental microorganisms for the biosynthesis of benzyl and phenylpropanoid compounds.

Fig. 2.

Activity screening of ADHs and ALDHs. (a) p-hydroxybenzyl alcohol, (b) cinnamyl alcohol, (c) p-hydroxybenzyl aldehyde, and (d) cinnamaldehyde were used as substrates.

3.3. Exploration of the substrate promiscuity of alcohol dehydrogenases and aldehyde dehydrogenases

To fully exploit the potential of dehydrogenases for the biosynthesis of aromatic compounds, the substrate promiscuity of the selected ADHs and ALDHs was investigated in detail. A major challenge for the application of dehydrogenases is the requirement of costly cofactors (NAD(P)⁺) as co-substrates for their catalytic activity [39]. To overcome this limitation, a NAD(P)H oxidase (LreNox) [40] from L. reuteri was co-expressed with the dehydrogenases to achieve cofactor recycling. The activity and kinetic parameters of NOX had been determined using NADH or NADPH in previous report, showing that LreNox exhibits high efficiency for oxidizing both NADH and NADPH. The kinetic parameters for LreNox were determined under increasing concentrations of NADH and NADPH, revealing that the K_m and K_cat values for both NADH and NADPH were similar [40]. The effect of LreNox on the conversion rate of m-hydroxybenzyl alcohol (4) to m-hydroxybenzaldehyde (4a) by ADH-49 was tested as an example. The results showed that co-expression of LreNox with ADH-49 increased the conversion rate by as high as 7.2-fold within 270 min of reaction (Fig. S1), demonstrating the effectiveness of this strategy. Following above activity screening, the 4 ADH and 4 ALDH candidate enzymes were purified (Fig. S2), and the substrate promiscuity was further evaluated under the LreNox-mediated cofactor recycling conditions. Twenty-four types of benzyl alcohols and five types of phenylpropanoid alcohols were used as substrates to test the 4 ADHs (ADH-49, ADH-88, ADH-90, and ADH-96). The reaction mixtures containing 10 μM ADH and 10 μM LreNox were incubated at 30 °C for 3 h, and the conversion rates were measured.

ADH-49 demonstrated remarkable efficiency and selectivity in oxidizing most of the meta- or para-substituted benzyl alcohols, achieving over 90% conversion of m-hydroxybenzyl alcohol (4), p-aminobenzyl alcohol (6), p-methylbenzyl alcohol (11), m-methylbenzyl alcohol (12), p-bromobenzyl alcohol (14), m-chlorobenzyl alcohol (15), p-methylbenzyl alcohol (17), and m-methylbenzyl alcohol (18) into the corresponding aldehydes (Fig. 3), while exhibited poor or no oxidation for ortho-substituted benzyl alcohol substrates (5, 7, 10, 13, and 16). Interestingly, the oxidation of p-nitrobenzyl alcohol (8) and m-nitrobenzyl alcohol (9) was relatively low, with no more than 50% conversion observed. Meanwhile, ADH-49 was able to oxidize over 90% of phenylpropanoid alcohols (27, 29, 30, and 31) without nitro substituents into corresponding phenylpropanoid aldehydes. ADH-88 exhibited efficient oxidation of some benzyl alcohols (5, 6, 11, 17, 20, and 22) and phenylpropanoid alcohols (29, 30, and 31) with more than 60% conversion, but it showed a narrower substrate range than ADH-49. Furthermore, when both ADH-49 and ADH-88 could catalyze the same benzyl alcohols (6, 11, 17, 20, and 22), ADH-49 showed higher catalytic efficiency than ADH-88.

Fig. 3.

Exploration of the substrate promiscuity of potential ADHs and ALDHs. Upper panel, structures of substrates and products. Lower panel, conversion rates.

ADH-90 exhibited excellent oxidation of para-monosubstituted benzyl alcohols (6, 8, 11, 14, and 17), but had limited activity for other mono-substituted ones (4, 5, 7, 9, 10, 12, 13, 16, and 18), as well as some di- or multi-substituted benzyl alcohol substrates (19, 22, 24, 25, and 26). Intriguingly, ADH-90 showed good catalytic activity (with the conversion ratio of 100%) for p,m-difluorobenzyl alcohol 23, likely due to the minimal steric hindrance of the fluorine atoms. Furthermore, compared with the low conversion ratio (15.4%) of ADH-49, ADH-90 converted 61.5% of p-nitrobenzyl alcohol (8) into p-nitrobenzaldehyde (8a). The above results showed that although different from ADH-49 in having broader substrate specificity, ADH-90 is more efficient at catalyzing para substituted substrates. Additionally, ADH-90 oxidized the most tested phenylpropanoid alcohols (except for 31) into aldehydes completely, indicating its advantage in oxidizing small hindered phenylpropanoid alcohols, consistent with its oxidation pattern for benzyl alcohols. Moreover, ADH-96 exhibited limited catalytic activity on benzyl alcohols but demonstrated excellent preference for all tested phenylpropanoid alcohols.

Following the same approach as for the alcohol dehydrogenases, the substrate promiscuity of the 4 ALDHs (ALDH-11, ALDH-25, ALDH-26, and ALDH-40) was assessed using 24 benzyl aldehydes and 5 phenylpropanoid aldehydes. The 4 ALDHs completely oxidized almost all of the tested aldehydes. ALDH-40 showed oxidation activity for all tested substrates except 24a (conversion ratio, 84%), with remarkable conversion rates above 90% for the remaining substrates. On the other hand, the catalytic efficiencies of ALDH-11 and ALDH-25 for the tested tri-substituted benzyl aldehyde (26a) were considerably lower than that of ALDH-40. In addition, ALDH-26 exhibited a slightly lower oxidation efficacy (<50%) for o-aminobenzaldehyde (7a), o-nitrobenzaldehyde (10a), o-tolualdehyde (13a), and other ortho-monosubstituted benzyl aldehydes, and its catalytic efficiency for 25a was also inferior to the other three ALDHs.

Taken together, the results indicated that the four ADHs and ALDHs had different substrate preferences and catalytic efficiencies for benzyl alcohols and phenylpropanoid alcohols. ADH-49 was effective for meta-/para-substituted benzyl alcohols and phenylpropanoid alcohols without nitro substituents, but less so than ADH-90 for the para-substituted benzyl alcohols and ADH-96/ADH-90 for the phenylpropanoid alcohols. Therefore, ADH-49 was suitable for meta-substituted benzyl alcohol and para- and meta-disubstituted benzyl alcohol, while ADH-90 was the best enzyme for para-substituted benzyl alcohols. Both ADH-90 and ADH-96 showed high catalytic efficiency for phenylpropanoid alcohols with small steric hindrance, while ADH-96 was the most efficient enzyme for phenylpropanoids with large steric hindrance. In contrast, ADH-88 had a narrow substrate range and low catalytic efficiency, making it an unlikely choice as the preferred candidate ADH under normal conditions. Among the 4 ALDHs, ALDH-40 exhibited superior ability for oxidizing benzyl and phenylpropanoid aldehydes, surpassing the other three enzymes and thus being the most preferred ALDH.

In addition, previous studies have shown that ADHs and ALDHs employ different catalytic mechanisms in the alcohol and aldehyde oxidation processes \[18,41]. The “short-chain” type of ADHs uses a universal base (B) as the acceptor of the hydroxyl proton, and the alcohol substrate is oxidized to the aldehyde by hydride transfer (Fig. S3a) [42]. The “medium-chain” type of ADHs interact with the coenzyme NAD(P)⁺ and coordinate with zinc(II) ions to bind alcohol substrates. Through a series of deprotonation steps, the hydrides are transferred from the alkoxide ions to NAD(P)⁺ to form zinc-bound aldehydes or ketones and NAD(P)H, and then the aldehydes are then released (Fig. S3b) [19]. ALDHs also depend on the cofactor NAD(P)⁺. Aldehydes enter the enzyme's active pocket through surface channels. The sulfur group of the active cysteine residue undergoes nucleophilic attack on the aldehydes' carbonyl carbon and the hydrides then attack NAD(P)⁺ to generate NAD(P)H. Subsequently, a conformational change occurs in the active pocket, leading to NAD(P)H displacement and water molecule entry. Under the mediation of the active glutamate residue, the water molecule attacks the carbonyl carbon, resulting in the formation of carboxylic acid (Fig. S4) [18]. Interestingly, in this study, some of the identified alcohol dehydrogenases (ADH-49, ADH-88 and ADH-90) have been found to have the ability to continuously oxidize alcohol substrates into carboxylic acids (Fig. S5). The enzymes' dual functionality disrupts the isolation between alcohol dehydrogenases and aldehyde dehydrogenases. It can be speculated that these enzymes may have evolved different catalytic mechanisms to achieve the oxidation of both alcohols and aldehydes. A thorough understanding of the mechanism underlying these bifunctional enzymes can help future efforts to rationally engineer ADHs and ALDHs and identify more potential dehydrogenases.

3.4. Scaled-up production of benzyl and phenylpropanoid acids via a whole-cell catalytic system

Many benzyl and phenylpropanoid alcohols are environmental hazards due to their cytotoxicity [43]. Nevertheless, these compounds can be transformed into synthetic drug intermediates or products by oxidizing them to carboxylic acid derivatives [44]. Hence, we aimed to develop an enzymatic cascade whole-cell system, employing dehydrogenase oxidation and cofactor regeneration processes to convert benzyl and phenylpropanoid alcohols directly into carboxylic acids.

Cinnamic acid (2b), renowned for its honey-like aroma, plays a pivotal role in the synthesis of methyl cinnamate, ethyl cinnamate, and benzyl cinnamate in the perfume industry [45]. Furthermore, it serves as a key intermediate in the biosynthesis of various natural compounds such as lignin, flavonoids, isoflavones, coumarin, aurone, stilbenes, catechins, and phenylpropenes [46]. Additionally, 2b acts as a precursor for the sweetener aspartame [47]. Likewise, p-toluic acid (11b) has a crucial role as an intermediate in polymer stabilizers, photosensitizing compounds, animal feed supplements, and other organic chemicals including pharmaceuticals, pigments, and dyes [48]. The production of 2b and 11b has significant medical and economic value. In view of the aforementioned applications, we aimed to establish a comprehensive whole-cell catalytic system for the conversion of 2 and 11 into 2b and 11b, respectively. For substrate 2, belonging to para-substituted benzyl alcohols, ADH-90 emerges as the optimal choice (K_m = 1.7 ± 0.3 mM, k_cat = 14.3 ± 1.2 min⁻¹, Fig. S8a). Substrate 11, a phenylpropanoid alcohol with limited steric hindrance, suggests the potential use of ADH-90 and ADH-96 as preferred ADHs. Meanwhile, ALDH-40 exhibits the broadest substrate range and promising catalytic activity for aldehyde substrates, making it the most suitable ALDH (The K_m and k_cat values towards 2a were 10.0 ± 2.7 mM and (3.3 ± 0.7) × 10² min⁻¹, and the K_m and k_cat values towards 11a were 10.0 ± 4.1 mM and (1.8 ± 0.6) × 10² min⁻¹. Figs. S8c and d). Taking these factors into account, we selected ADH-90 and ALDH-40 as biocatalysis enzymes, together with the NAD(P)H oxidase LreNox to construct the whole-cell catalytic system. The K_m and k_cat values of ADH-90 towards 11 were 1.2 ± 0.1 mM and 6.7 ± 0.3 min⁻¹ (Fig. S8b).

The effect of thermal stability on ADH-90 and ALDH-40 were monitored at seven temperatures of 0, 25, 30, 35, 40, 45, and 50 °C for 30 min. Reactions were carried out in triplicates with 1 mM substrate, 0.5 μM ADH-90 or ALDH-40, and 2 mM NADP+. For ADH-90, the maximum conversion of substrate 2 (8.0%) was achieved at temperatures of 40 °C and 45 °C (Fig. S7a). The conversion rate variance between 30 and 40/45 °C is relatively small, with a difference of less than 1.5%. The maximum conversion of substrate 2a by ALDH-40 was attained at 30 °C, resulting in a yield of 61.0 % (Fig. S7b). A decrease in yield was observed at elevated temperatures of 35 °C, 40 °C, 45 °C, and 50 °C. Compared to ALDH-40, ADH-90 has better thermal stability. Among the temperatures tested, 30 °C and 40 °C demonstrated enhanced catalytic efficiency for both ADH-90 and ALDH-40. While 40 °C resulted in a relatively higher conversion rate for ADH-90, it was noted that this temperature would dramatically decrease the activity of ALDH-40 by 16 %. Therefore, 30 °C was selected as the optimal temperature for driving the cascade biocatalysis.

To obtain the whole-cell transformant WCS1 (Fig. 4a), we co-transformed E. coli BL21(DE3) cells with two plasmids: one expressing the LreNox gene (pET28a-LreNOX) and another expressing both the ADH-90 and ALDH-40 genes (pETDuet-1-ADH-90-2-ALDH-40). After enzyme expression, the bacterial cells were collected and resuspended in a solution of OD₆₀₀ = 27 for following whole-cell catalysis. We tested different concentrations of 11 (0.5 g/L, 1.0 g/L, and 1.5 g/L) as substrates and collected samples at 2 h, 4 h, and 6 h of reaction to evaluate the production of 11b (Fig. 4b). Within 2 h of the whole-cell catalytic reaction, substrate 11 at concentrations of 0.5, 1.0 and 1.5 g/L were converted into 11b with conversion efficiencies of 100%, 92.0%, and 88.7%, respectively. After 4 h of reaction, complete oxidation of 1.0 g/L substrate was achieved, while 97.0% of 1.5 g/L substrate was oxidized. For 6 h of reaction, 1.5 g/L of 11 was fully converted into product.

Fig. 4.

Whole-cell cascade catalytic system for production of benzyl and phenylpropanoid acids. (a) The whole-cell catalysis scheme. (b) Whole-cell transformation of 11. (c) Whole-cell transformation of 2.

Likewise, we used various concentrations of 2 (1 g/L, 2 g/L, 6 g/L, and 10 g/L) as substrates and took samples were taken at 2 h and 4 h of reaction to measure production (Fig. 4c). After 2 h of reaction, the whole-cell system completely oxidized 1 g/L, 2 g/L, and 6 g/L of substrate 2 into product 2b. For 10 g/L of the substrate, 8.85 g/L of the product was produced, corresponding to a conversion rate of 80.1%. After 4 h of reaction, complete oxidation of 10 g/L substrate was achieved, indicating the high efficiency of the whole-cell system for phenylpropanoid alcohol conversion.

These results demonstrate that the whole-cell cascade catalytic system developed can effectively biocatalyze 11 and 2 into corresponding carboxylic acids, which can be easily scaled up to gram level, providing a method for the preparation of pharmaceutical intermediates and products. It is anticipated that, by adjusting the candidate enzymes according to the substrate, the whole-cell system can be easily extended to the production of other benzyl and phenylpropanoid alcohols, offering substantial economic and production benefits.

4. Conclusions

Multiple ADHs and ALDHs were identified and characterized from genomic data of environmental microorganisms that can oxidize benzyl/phenylpropanoid alcohols and aldehydes. The optimal ADH-90 and ALDH-40 with higher catalytic activity and broader substrate preferences were selected, together with the NAD(P)H oxidase LreNox, to construct a whole-cell cascade catalytic system. This system can oxidize many benzyl alcohols and phenylpropanoid alcohols in a green and efficient manner. It effectively transformed gram-scale quantities of p-methylbenzyl alcohol and cinnamyl alcohol to the respective carboxylic acids. The mining-test-analyze-build strategy of this study also provides new insights for realizing the green biosynthesis of other valuable aromatic products.

CRediT authorship contribution statement

Zetian Qiu: performed the experiment and prepared original draft. Xiaohui Liu and Yushuo Zhao: performed the experiment. Jie Yu and Kun Liu: analyzed the data. Guang-Rong Zhao: prepared the manuscript. Shengying Li: prepared the manuscript. Lei Du: conceived, and designed the experiment, prepared the manuscript. Li Ma: cconceived, and designed the experiment, prepared the manuscript. All authors read and approved the final manuscript.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix B. Supplementary data

The following is the Supplementary data to this article.