Cysteine-S-conjugate β-lyase: a green catalyst for the production of flavor-active mercaptoketones

Abstract

Abstract

Sulfur-containing aroma compounds, which are key determinants of the sensory profile and consumer acceptance of many foods, have received limited research attention regarding the liberation of these thiol aromas via exogenous enzymes. In this study, we characterized and successfully overexpressed an active cysteine-S-conjugate β-lyase (MetC) from the marine bacterium Shewanella putrefaciens, and we also established its optimal catalytic conditions. Kinetic analysis demonstrated effective activity of the enzyme toward natural substrates and aroma mercaptoketone precursors. Size-exclusion chromatography indicated a tetrameric structure for MetC. For most substrates, the apparent catalytic efficiency ($V_{\text{max}}$/$K_{\text{m}}$)_app was calculated from the initial slopes of the velocity-concentration curves, which was used to evaluate catalytic efficiency due to solubility limits. Notably, only Cys-4MP-one reached saturation and conformed to sigmoidal model, indicating a unique cooperative behavior. These findings provide a quantitative basis for comparing the substrate specificity and catalytic effectiveness of MetC toward diverse precursors. The practical application of the enzyme was validated through preparative enzymatic conversion of cysteine-4-mercapto-4-methylpentan-2-one to 4-mercapto-4-methylpentan-2-one, with a 48.5% conversion achieved from its cysteine-conjugated precursor. Our findings demonstrate that S. putrefaciens MetC functions effectively as a biocatalyst for aroma thiol mercaptoketone production. Additionally, this research provides perspectives on the enzymatic production of sulfur-containing aroma compounds for applications in the fragrance and food industries.

Key points

• Eco-friendly enzymatic bioconversion produced aroma mercaptoketone.

• Cysteine-S-conjugate β-lyase from Shewanella putrefaciens exhibited sigmoidal kinetics toward Cys-4MP-one, while other substrates followed linear kinetics.

• A conversion of 48.5% was achieved for 4MMP during the preparative-scale biotransformation.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00253-026-13927-6.

Introduction

Sulfur-containing aroma compounds play an essential role in determining the sensory characteristics and consumer acceptance of numerous food products, and these compounds, characterized by their potent aroma, significantly influence flavor profiles even at extremely low concentrations because of their low odor thresholds (Zhang et al. 2023a). Within the food industry, their controlled synthesis and application are of substantial commercial interest, particularly in products such as wine, cheese, coffee, beer, and meat where highly valued complex flavor profiles are required (Ben Alzacha and Gargouri 2015; Singh et al. 2024).

Volatile thiols such as 4-mercapto-4-methylpentan-2-one (4MMP) and 3-mercapto-1-hexanol (3MH) are crucial in the beverage industry; they impart distinctive fruity and tropical notes. Reminiscent of blackcurrant, grapefruit, and passion fruit that significantly influence the typicity and marketability of wines like Sauvignon Blanc (Tominaga et al. 1998; Dubourdieu et al. 2006). Another compound, dimethyl trisulfide (DMTS), known for its cabbage-like and sulfurous aroma, is essential in defining flavor profiles of cooked meats and various fermented foods (Pardon et al. 2008). Given the importance of these aroma compounds, elucidating the biochemical mechanisms underlying their formation from nonvolatile precursors has become a growing area of interest. A particularly crucial group of precursors involves cysteine conjugates, which are nonvolatile and odorless yet can be enzymatically cleaved by cysteine-S-conjugate β-lyase (MetC) to release volatile aromatic thiols (Bonnaffoux et al. 2021). The enzyme-mediated cleavage and subsequent release of aromatic thiols provide a highly promising approach for flavor enhancement strategies in the food industry, particularly in products such as fruits, alcoholic beverages, and tea. Those approaches enable the modulation and intensification of desirable aromas across various food products (Liang et al. 2022; Zhang et al. 2023b; Alises et al. 2024; Schwartz et al. 2024).

However, despite this potential, significant gaps remain regarding our understanding and optimization of enzymatic processes for the industrial production of sulfur-containing aroma compounds. Current industrial production of sulfur-containing aroma ingredients primarily relies on natural extraction or chemical synthesis. However, these methods are often limited by low natural abundance and the formation of undesirable off-flavors during chemical processing. Recent studies have demonstrated that optimizing enzymatic pathways in Saccharomyces cerevisiae can significantly enhance the yield of key compounds, such as 3-methylthio-1-propanol, under mild conditions (Sun et al. 2024). Such biocatalytic routes offer superior chemo-specificity and sensory quality compared to traditional methods. Thus, developing efficient biotechnological processes that utilize microbial enzymes presents an attractive alternative, one that promises better control, higher yields, and improved consumer acceptance due to the perceived naturalness of biocatalytic methods (Dickey et al. 2024). In this context, the enzyme MetC from S. putrefaciens is of particular interest, as MetC has demonstrated the potential to efficiently catalyze cleavage of cysteine conjugates, thereby releasing volatile thiols such as 3MH and 4MMP (Holt et al. 2011). Li et al. (2021) screened the crude cell extract from bacteria and yeast strains for β-lyase activities, and the S. putrefaciens cell extract exhibited the highest β-lyase activity for all tested substrates.

Mercaptoketones, including 4MMP, characterized by passion fruit and black currant aromas, and 4-mercaptopentan-2-one (4MP-one), associated with leek and potato aromas, play a fundamental role in developing high-impact aromatic profiles in food and beverage products. The enzymatic biosynthesis of these compounds, particularly via mechanisms that facilitate release or formation of thiol groups, represents a significant area of scientific inquiry with promising applications in sustainable manufacturing processes. This research aims to evaluate the kinetic parameters of mercaptoketones aroma compounds 4MMP and 4MP-one through S. putrefaciens MetC transformation from their cysteine-conjugated precursors. The cysteine-conjugated precursors of 4MMP and 4MP-one are synthesized, followed by overexpression, purification, and optimization of the MetC enzyme activity. Additionally, the scalability of this enzymatic production process is assessed, providing new opportunities for industrial applications.

Materials and methods

Preparation of cysteine-conjugated precursors

Commercially sourced reagents and solvents of appropriate grade for their intended application were used without further purification unless otherwise specified. Flash column chromatography was performed using SiliCycle silica gel (70–230 mesh, product number: R10040B). Thin-layer chromatography (TLC) was performed on pre-coated silica plates (Merck, cat. no. 1.05554) and visualized with 0.5% ninhydrin in acetone. Reaction progress and chromatographic fractions were routinely monitored by TLC. Nuclear magnetic resonance (NMR) spectra were recorded on Bruker Avance™ III and Bruker NEO 600 MHz spectrometers. Spectral data were processed using TopSpin 3.6.5 (Bruker).

Preparation of Cys-4MMP

l-Cysteine (40 mmol) was reacted with di-tert-butyl dicarbonate (44 mmol) in a CH₃CN/H₂O mixture (220:80 mL, v/v) containing triethylamine (60 mmol) to synthesize N-(tert-butoxycarbonyl)-l-cysteine (N-Boc-Cys), following a modified literature procedure (Bissolino et al. 1993). After stirring at room temperature for 4 h, the reaction mixture was worked up as previously described, affording N-Boc-Cys (~ 9 g) as a pale oil, used without further purification.

Freshly prepared N-Boc-Cys (~ 9 g) was then dissolved in CH₃CN (400 mL) and reacted with 4-methyl-3-penten-2-one (7.85 g, 80 mmol, also known as mesityl oxide) in the presence of Cs₂CO₃ (20.85 g, 64 mmol). The mixture was stirred at room temperature for 2 h, concentrated under reduced pressure, acidified with 0.4 M HCl_(aq) (320 mL), and extracted with ethyl acetate. The combined organic layers were washed with brine, dried over anhydrous Na₂SO₄, and concentrated. The resulting crude intermediate was purified by flash chromatography using a mixture of ethyl acetate, hexane, and acetic acid (5:15:1, v/v/v) to afford the corresponding intermediate (8.74 g, 68% over two steps).

N-Boc-protected intermediates (8.74 g, 27.3 mmol) were treated with trifluoroacetic acid (93.4 g, 819 mmol) in CH₂Cl₂ (273 mL) at room temperature for 2 h. After concentration under reduced pressure, the residue was directly purified by flash chromatography without aqueous workup (acetone/ethyl acetate/water = 5:4:1, v/v/v) to yield the target Cys-4MMP (65%). The NMR spectra were identical to those reported in our previous work (Li et al. 2021).

Preparation of Cys-4MP-one

Starting from 40 mmol of l-cysteine, Cys-4MP-one was prepared under analogous conditions by replacing mesityl oxide with 3-penten-2-one, affording the product in 60% overall yield. The compound was obtained as an inseparable mixture of diastereomers: ¹H NMR (600 MHz, D₂O) δ = 4.00 (1H, dd, J = 4.1 and 7.6 Hz), 3.30 (1H, app sext, J ≈ 6.8 Hz), 3.20 (1H, dd, J = 4.1 and 14.7 Hz), 3.07 (1H, dd, J = 7.7 and 14.8 Hz), 2.91 (1H, dd, J = 7.4 and 17.6 Hz), 2.83 (1H, dd, J = 6.4 and 17.6 Hz), 2.25 (3H, s), and 1.33 (3H, d, J = 6.8 Hz). Significant overlap between the two diastereomeric sets of signals is observed; the reported chemical shifts and coupling constants therefore correspond to one well-resolved set, while integrations reflect the combined signals of both diastereomers (Fig. S1): ¹³C NMR (150 MHz, D₂O) δ = 213.5, 213.4, 172.7, 172.6, 54.0, 53.8, 50.0, 49.7, 35.4, 35.1, 30.9, 30.8, 29.8, 29.7, 20.7, and 20.6. Each carbon gives rise to a pair of resonances corresponding to the two diastereomers (Fig. S2). Cys-4MP-one, as prepared in this study, was obtained as a diastereomeric mixture. All mentions of Cys-4MP-one throughout the manuscript refer to this mixture.

Strains and plasmids

The Escherichia coli strain DH5α (genotype: Δ(argF-lac)169, φ80dlacZ58(M15), ΔphoA8, glnX44(AS), deoR481, rfbC1, gyrA96(NalR), recA1, endA1, thiE1, and hsdR17) (Song et al. 2015) was chosen as the competent cells for their high efficiency and their ability to facilitate intracellular recombination of homologous DNA fragments (Thomason et al. 2023). Moreover, the recA1 and endA1 genotypes of this strain ensure stable plasmid propagation and the maintenance of high-quality DNA for subsequent downstream applications. E. coli strain C43 (DE3) was used to overexpress MetC protein, and the purified MetC protein was analyzed to determine the protein identification and the kinetic parameters. The metC gene was cloned into the pET21a vector at the NdeI and XhoI restriction sites. The construct was designed to keep the C-terminal His-tag for subsequent purification by nickel affinity chromatography. The metC gene sequence from S. putrefaciens (NCBI accession number: WP_011919183.1) was synthesized and optimized, and it was cloned onto pET21a for expression in E. coli (MDBio, Inc., Taiwan) (Fig. S3).

E. coli transformed with pET21a-metC

Competent E. coli DH5α cells were prepared using the heat shock method. The pET21a-metC plasmid was transformed into E. coli DH5α for amplification and subsequently into E. coli C43 (DE3) for protein expression. Both transformation processes were performed using the calcium chloride-mediated heat-shock method as previously described (Sambrook and Russell 2001). Recombinant colonies were screened on LB agar plates containing ampicillin (100 µg/mL). The identity of the positive transformants was rigorously confirmed by colony PCR (targeting the 1200 bp metC) and plasmid sequencing. Validated plasmids were then extracted using a commercial miniprep kit to ensure high-quality DNA for the generation of the expression strain.

Overexpression and purification of MetC

E. coli C43 harboring the pET21a plasmid encoding the metC gene was cultivated at 37 °C. When the culture reached an OD₆₀₀ of 0.6–0.8, it was supplemented with 0.2 mM IPTG in 2 × YT broth and further incubated at 25 °C for 12 h for protein overexpression. After centrifugation (8000 × g, 4 °C, 10 min), the supernatant was removed, and the cells were resuspended in 20 mM Tris buffer (pH = 8) containing 250 mM NaCl and 10% (v/v) glycerol as A buffer. For the extraction of intracellular proteins, cells were resuspended in a lysis buffer supplemented with 1000 U/L DNase I and 13 mM benzamidine. Cell lysis was subsequently performed using a High-Pressure Cell Disrupter (Constant Systems CF1) at 20 kpsi. The objective was to extract intracellular proteins. The cell lysate was then centrifuged (10,000 × g, 4 °C, 15 min), and the supernatant was collected. The MetC proteins solubilized in the buffer were purified using a nickel affinity column (Hitrap chelating column, GE). The column was initially washed with 5–15 column volumes of de-ionized water and buffer. The protein was loaded onto the nickel affinity column using a peristaltic pump at a rate of 1 mL/min. A buffer was supplemented with 500 mM imidazole to prepare B buffer, which served as the elution buffer for protein purification. The split ratio of A/B buffer was adjusted with imidazole at varying concentrations for purification using Fast Protein Liquid Chromatography (ÄKTA purifier). The overexpression and purification process yielded concentrated recombinant MetC protein from the nickel affinity column, and a desalting step was immediately performed to remove imidazole. The final purified MetC protein reached a concentration of 4.59 mg/mL, as determined by BCA assay. The homogeneity and molecular weight of the purified protein were verified by SDS-PAGE.

MetC protein identification by using LC-MS/MS

The band displayed on the SDS gel was excised and cut into small pieces (approximately 1 mm²). These pieces were then transferred to new Eppendorf tubes. The gel pieces were covered with wash buffer (50% acetonitrile/50% 25 mM triethylammonium bicarbonate buffer), vortexed for 1 min, and this process was repeated several times until the dye was completely washed away. Subsequently, the gel pieces were transferred to microtubes and dried using a bench-top microcentrifuge. Next, the gel pieces were incubated with 20 mM dithiothreitol (DTT) solution, covered, and allowed to absorb the solution for 10 min at room temperature. After incubation at 55 °C for 30 min, the DTT solution was removed. Subsequently, 55 mM iodoacetamide (IAA) solution was added, and the mixture was allowed to react for 60 min at room temperature in the dark before being removed. Wash buffer (50% acetonitrile/50% 25 mM triethylammonium bicarbonate buffer) was added, and the mixture was vortexed for 1 min twice. The gel pieces were then shrunk by centrifugation, and all liquid was removed. A sufficient amount of trypsin solution (20 µg/mL) was added to cover the gel pieces, which were rehydrated at 4 °C for 10 min. Excess trypsin solution (20 µg/mL) was removed using a gel-loading tip. Subsequently, the gel pieces were incubated either at 37 °C for 4 h or at 25 °C overnight to complete the digestion. The digestion was terminated, and peptides were extracted by vortexing and sonicating for 15 min with extraction solution A (50% acetonitrile/5% formic acid), repeated twice. The extracted solution was collected, and extraction solution B (80% acetonitrile/5% formic acid) was added three times. The combined extraction solution was then transferred to a new Eppendorf tube and subjected to vacuum centrifugation. LC-MS analyses were conducted using a Waters Acquity nano-UPLC coupled with a Waters G2 Q-TOF mass spectrometer. Protein samples (10 µg/mL) were infused directly onto the mass spectrometer at a flow rate of 1 µL/min using a syringe pump. The G2 Q-TOF mass spectrometer operated in positive ion, high-resolution mode with detection in the range of 600 to 2300 m/z. Source parameters were set as follows: capillary voltage, 2.50 kV; source temperature, 90 °C; desolation temperature, 200 °C; cone gas flow, 20 L/h; and desolation gas flow, 500 L/h. The protein peak was deconvoluted using the MassLynx MaxEnt1 function with the following parameters: output resolution, 1.0 Da/channel; uniform Gaussian width at half height, 0.75 Da; minimum intensity ratios, 30% for left and right; and iteration to convergence for completion.

Size-exclusion chromatography analysis of the oligomeric state of MetC

The experimental procedure was adapted from Beattie (1998) with minor modifications. Protein oligomeric states were analyzed using size-exclusion chromatography (SEC) on a HiLoad 16/600 Superdex 200 PreGrade column (GE Healthcare) integrated with an FPLC system (ÄKTA purifier). The column was pre-equilibrated with two column volumes (CVs) of degassed deionized water, followed by an equivalent volume of degassed buffer (20 mM Tris-base, pH 8.0, and 250 mM NaCl). Samples and standards (300 µL) were injected into the loading loop and eluted at a constant flow rate of 0.8 mL/min. The standards included ovalbumin, conalbumin, aldolase, ferritin, and blue dextran; the latter of which was used to determine the column void volume (V₀) due to its complete exclusion from the stationary phase. To ensure reproducibility, the column was washed and re-equilibrated after each run. Molecular weights were estimated by correlating sample elution volume (V_e) with a standard calibration curve. Finally, the oligomeric state of each protein was subsequently determined by comparing the SEC-derived molecular weight with the subunit size determined by SDS-PAGE.

MetC for β-lyase activity assays

The β-lyase activity of the MetC protein was assessed by measuring the absorbance of the product formed through the spontaneous disulfide exchange reaction between free thiol groups and 5,5′-dithiobis (2-nitrobenzoic acid) (DTNB) (Irmler et al. 2008). Various substrates, such as l-cystine and l-cystathionine, were individually added to a Bis-Tris propane buffer at pH 9 containing 50 mM Bis-Tris propane, 1 mM EDTA, 20 µM pyridoxal-5′-phosphate, and 0.01 µM of purified MetC enzyme, resulting in a final substrate concentration of 0.2 mM. Subsequently, 2 µL of 50 mM DTNB was added to achieve a final DTNB concentration of 1 mM. The reaction mixture was incubated at room temperature for 15 min, after which the absorbance was measured at 412 nm using a spectrophotometer. The rate of β-lyase activity was calculated using a l-cysteine standard curve. To evaluate the effects of pH and temperature on β-lyase activity, assays were performed across a pH range of 6.0–10.0 and temperatures spanning 4–70 °C (4,15, 25, 30, 37, 42, 50, 60, and 70 °C). The pH profiles were determined using a series of 50 mM buffer systems, each supplemented with 20 µM pyridoxal 5′-phosphate (PLP) and 1 mM EDTA: citrate (pH 6.0), phosphate (pH 7.0), Bis-Tris propane (pH 8.0 and 9.0), and carbonate (pH 10.0). To establish the pH optimum, assays were initially performed at 37 °C. Based on these preliminary results, pH 9.0 was identified as the optimal pH and was subsequently maintained as the constant pH for temperature optimization studies (ranging from 4 to 70 °C). Enzymatic activity was expressed as relative activity (%), with the maximum rate observed within each experimental group defined as 100%. Under these optimized conditions, the 100% relative activity corresponded to an absolute reaction rate of 84.7 µmol/mg/min.

Enzyme MetC kinetic assays

Based on the method described by Lu et al. (2024) with modifications, enzymatic assays were conducted using 0.01 µM of purified enzyme and various concentrations of substrates, including l-cystine, l-cystathionine, Cys-4MMP, and Cys-4MP-one. The enzyme concentration was calculated based on the molecular weight of the MetC monomer (43 kDa) and achieved precisely by stepwise serial dilutions in the Bis-Tris propane buffer immediately prior to the reaction. The reactions were carried out in 50 mM Bis-Tris propane buffer (pH 9.0) containing 20 µM pyridoxal-5′-phosphate (PLP), 1 mM EDTA, and 1 mM DTNB. All reactions were performed in 96-well microplates and initiated by immediately adding the enzyme to the substrate solution. The reaction mixtures were incubated at 37 °C for 15 min, and the absorbance at 412 nm was recorded at 1-min intervals. Simultaneously, a control group without enzymes was established. The net increase in absorbance after subtracting the background signal was used to quantify the concentration of thiols released during enzyme-catalyzed substrate degradation using a standard curve. The concentration of thiol group released as a result of substrate degradation catalyzed by the enzyme was quantified by referencing a standard curve. Rates of thiol release (µmol/mg/min) were plotted versus substract concentrations, and the resulting data were fitted to the Linear equation or Hill equation models with SigmaPlot to obtain the kinetic parameters.

$$\mathrm{Hill}\mathrm{equation}:v=\frac{V_{\mathit{max}}{\left[S\right]}^h}{K_{0.5}^h+{\left[S\right]}^h}$$ 1

Preparative scale of enzymatic conversion of Cys-4MMP to 4MMP

To further assess the bioconversion capability of the enzyme under preparative conditions relevant to potential industrial applications, a 100-mL scale enzymatic conversion of Cys-4MMP to 4MMP was carried out in a Florence flask equipped with continuous magnetic stirring. Purified MetC protein solution was added to give a final concentration of 1 µM, and the substrate Cys-4MMP was supplied at 68 mM. The reaction mixture was adjusted to 100 mL with 50 mM Bis-Tris propane buffer (pH 9.0) containing 1 mM EDTA and 20 µM pyridoxal-5′-phosphate. The suspension was maintained at 30 °C and stirred without interruption for 6 h. At the end of the incubation, the volatile products were extracted with diethyl ether, concentrated, and quantified by GC analysis, from which the product yield was calculated.

Identification and quantitation of scale-up enzymatic conversion products using gas chromatography–mass spectrometry analysis (GC-MS)

Analyses were performed using an Agilent 8890 gas chromatograph with a flame ionization detector (FID) and a parallel 5977 C mass selective detector (MSD). Separation occurred on a VF-5 ms capillary column (60 m × 0.32 mm i.d., 0.25 µm film thickness; Agilent J&W Scientific, USA). The injector operated at 250 °C, with helium as the carrier gas at a constant pressure of 19.6 psi and a split ratio of 10:1. The oven temperature program began at 50 °C, ramped to 90 °C at 10 °C/min, then to 110 °C at 5 °C/min, and finally to 275 °C at 30 °C/min, held for 1.5 min. The FID was operated at 300 °C, with hydrogen and air flow rates of 30 and 400 mL/min, respectively, and a nitrogen makeup gas flow rate of 25 mL/min. For MS detection, the transfer line, ion source, and quadrupole were set at 280, 230, and 150 °C, respectively. Electron impact ionization was performed at 70 eV, and mass spectra were acquired in scan mode over a range of 20–400 m/z. The product and co-product were identified by comparison of their mass spectra with standard spectra in NIST 23 library and commercially available standards.

Quantification of the product and co-product

To quantify the concentrations of the enzymatic product 4MMP and the co-product mesityl oxide by GC analysis, a calibration method was established using furfuryl mercaptan as the internal standard. Prior to use, furfuryl mercaptan was purified by distillation to ensure analytical accuracy. Multi-point calibration curves were generated using ethanol-based standard solutions of 4MMP and mesityl oxide, each spiked with furfuryl mercaptan at a fixed concentration of 0.01% (w/w). The standard concentrations ranged from 0.001% to 0.2% (w/w). Calibration curves were constructed by linear regression of the GC-FID peak area ratios (4MMP/furfuryl mercaptan and mesityl oxide/furfuryl mercaptan) versus their respective concentrations.

The established calibration curves were subsequently applied to determine the concentrations of the product and co-product in the extracted reaction samples based on GC-FID measurements. To address the potential solubility limitations of the analytes, the concentrated extracts were diluted tenfold in ethanol prior to analysis. As a result, the established calibration curves enabled quantification over an effective concentration range of 0.01% to 2% (w/w).

Products extraction from enzymatic reaction solution

The products generated during the enzymatic transformation were recovered via solvent extraction. The reaction mixture was extracted twice with diethyl ether using a total solvent volume equal to the reaction volume, and the combined organic layers were diluted with ethanol. Diethyl ether was then removed under reduced pressure, yielding an ethanol solution containing the enzymatic products. Recovery efficiency was validated using commercial 4MMP at a known concentration; experimental details and representative raw data are provided in the Supplementary Materials.

Statistics analysis

Data were analyzed statistically using SPSS Version 12.0 (SPSS Inc., Chicago, IL, USA). One-way analysis of variance (ANOVA) was used to determine statistical differences between sample means, followed by Tukey’s HSD post hoc test for multiple comparisons. The level of significance was set at p < 0.05. Each experiment was performed in three independent biological replicates (n = 3). Data are presented as means, with error bars representing the standard deviation. Different capital letters above the bars denote statistically significant differences (p < 0.05).

Results

Cys-4MMP and Cys-4MP-one synthesis

In this study, previously established protocols by Lu et al. (2024) were adapted to prepare Cys-4MMP and Cys-4MP-one. The synthetic route (Fig. 1), beginning with freshly prepared N-Boc-cysteine, involves N-terminal protection, thia-Michael addition, and subsequent deprotection of the amino group.

Fig. 1.

Synthesis of mercaptoketone precursors Cys-4MMP and Cys-4MP-one

The full ¹H and ¹³C NMR spectra of Cys-4MP-one are shown in the Supplementary materials (Figs. S1 and S2).

Edited-HSQC spectra (Fig. 2) confirmed the structures of Cys-4MMP and Cys-4MP-one, displaying CH₂ signals in red and CH/CH₃ signals in blue with opposite phases. The spectra clearly distinguish the structural differences between the two compounds. Regarding purity, both NMR spectra showed clean baselines without detectable impurity signals, indicating high purity and suitability for subsequent experimental applications.

Fig. 2.

The multiplicity-edited ¹H-¹³C HSQC spectra of Cys-4MMP and Cys-4MP-one

Overexpression and purification of MetC

The synthesized metC gene was cloned onto the pET21a vector and transformed into E. coli C43 (DE3). Expression of MetC was induced with 0.2 mM isopropyl β-D-1-thiogalactopyranoside, and purification was carried out using a nickel affinity column.

Purity and identity of the purified MetC protein were confirmed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). The MetC protein consists of 400 amino acid residues with a predicted molecular weight of approximately 44 kDa. In SDS-PAGE analysis (Fig. S4), a distinct protein band was observed in lanes 4–8, with an estimated molecular weight of approximately 41 kDa. Following in-gel digestion, the MetC protein treated with trypsin was analyzed by LC-MS/MS, and the results revealed 79% sequence coverage and Mascot score was 6363, identifying the protein as the cystathionine β-lyase from S. putrefaciens (NCBI accession number: WP_011919183.1; Fig. S5).

Optimum pH and temperature of MetC for β-lyase activity

The MetC from S. putrefaciens, expressed and purified in E. coli C43 (DE3), exhibited highest activity at pH 9 (100% relative activity, 84.7 µmol/mg/min), retaining about 57.1% activity at pH 10 (Fig. 3A). Activity at pH 5–8 and 11 was markedly lower. The enzyme also exhibited highest activity at 37 °C and maintained approximately 80–90% relative activity at 15–30 °C (Fig. 3B). When the temperature increased to 42 °C, 50 °C, 60 °C, and 70 °C, only about 60–70% of its original activity remained.

Fig. 3.

Effect of A pH and B temperature on β-lyase activity of Shewanella putrefaciens MetC toward l-cystathionine. Enzymatic activity was normalized to the optimal conditions (pH 9.0, 37 °C), where the absolute reaction rate was 84.7 µmol/mg/min in the reaction mixture. The lines represent a non-linear Gaussian fit, illustrating the bell-shaped trend of the enzyme activity. Data were presented as means, with error bars representing the standard deviation. Different capital letters above each bar denote statistically significant differences (p < 0.05)

Kinetic parameters of MetC for β-lyase activity against sulfur-containing amino acids and mercaptoketones

Except for Cys-4MP-one, the enzymatic kinetic curves could not achieve the saturation because of substrate solubility limits. Under the condition of [S] <  < $K_{\text{m}}$, the initial velocity ($V_0$) maintains a linear proportionality to the substrate concentration, as the Michaelis-Menten equation simplifies to $V_0$ ≈ $\frac{V_{\text{max}}}{K_{\text{m}}}$ [S]. In this regime, the reaction rate is limited by the enzyme-substrate encounter. Therefore, these substrates exhibited a linear relationship between reaction velocity and concentration, and their catalytic efficiencies were determined from the initial slopes of the linear regression, representing the apparent second-order rate constants [($V_{\text{max}}$/$K_{\text{m}}$)]_app. This value serves as an apparent second-order rate constant, allowing for a rigorous comparison of the substrate specificity and catalytic effectiveness toward different substrates, regardless of their individual saturation levels.

As illustrated in Fig. 4A and B, L-cystine (10–500 µM) and l-cystathionine (10–800 µM) degradation by S. putrefaciens MetC increased linearly within the tested concentration range. At their maximum soluble concentrations, both substrates exhibited strong linear relationships (R² > 0.99). The apparent maximum velocities ($V_{\text{max},\text{app}}$) were 90.5 and 150 µmol/mg/min, with the apparent catalytic efficiencies [($V_{\text{max}}$/$K_{\text{m}}$)]_app of 0.177 and 0.183 L/mg/min, respectively (Table S1). These results indicate that MetC efficiently catalyzes the degradation of natural sulfur-containing amino acids, even under sub-saturating conditions.

Fig. 4.

Catalytic rate profiles illustrating the degradation of sulfur-containing amino acids A l-cystine and B l-cystathionine by Shewanella putrefaciens MetC. Measurements at each substrate concentration were conducted in triplicate. Data were presented as means, with error bars representing the standard deviation

In addition to its activity toward canonical substrates, the kinetic parameters of S. putrefaciens MetC were further evaluated using the mercaptoketone precursors Cys-4MMP and Cys-4MP-one (Fig. 5 and Table S1). For Cys-4MMP, MetC also exhibited a strong linear relationship (R² > 0.99), with an apparent maximum velocity ($V_{\text{max},\text{app}}$) of 37.7 µmol/mg/min. The apparent catalytic efficiencies [($V_{\text{max}}$/$K_{\text{m}}$)]_app was 0.0150 L/mg/min (Table S1). Notably, for Cys-4MP-one, which was analyzed using sigmoidal curve fitting, the enzyme showed a $V_{\text{max}}$ of 38.7 µmol/mg/min, $K_{0.5}$ of 887 mM, $k_{\text{cat}}$ of 64.5 s⁻¹, and $k_{\text{cat}}/K_{0.5}$ of 7.2 × 10⁴ M⁻¹ s⁻¹, accompanied by a Hill coefficient (h) of 2.21.

Fig. 5.

Catalytic rate profiles illustrating the degradation of mercaptoketone precursors A Cys-4MMP and B Cys-4MP-one by Shewanella putrefaciens MetC. Measurements at each substrate concentration were conducted in triplicate. Data were presented as means, with error bars representing the standard deviation

To examine the quaternary structure of MetC in solution, size-exclusion chromatography was performed. The analysis revealed an apparent molecular weight of approximately 177 kDa (Fig. 6 and Fig. S6). The SDS-PAGE results showed that molecular weight of MetC monomer is 44 kDa (Fig. S4). Together, these results indicate that S. putrefaciens MetC exists predominantly as a tetramer under the tested conditions.

Fig. 6.

Size-exclusion chromatography analysis of Shewanella putrefaciens MetC. The black arrow indicates tetrameric MetC (K_av = 0.256), corresponding to a calculated molecular weight of 177 kDa

Preparative enzymatic conversion and product recovery

As a preparative-scale study to evaluate the potential of S. putrefaciens MetC, a larger-scale enzymatic transformation of Cys-4MMP was conducted. To enable accurate quantification of MetC-converted products by GC-FID, a calibration protocol was first established. As shown in Fig. S7A, the calibration curve derived from the GC peak area ratio of 4MMP to furfuryl mercaptan as an internal standard exhibited excellent linearity (R² = 0.999). Mesityl oxide concentrations were determined using the same internal standard approach, and the corresponding calibration curve also demonstrated high linearity (R² = 0.999) (Fig. S7B).

A preparative-scale biotransformation was then carried out using 1 µM S. putrefaciens MetC in a 100 mL reaction volume containing 68 mM Cys-4MMP. By operating at 30 °C, we tended to provide a safety margin to prevent the reaction temperature from exceeding the enzyme’s optimal range, which would otherwise lead to thermal inactivation. Following the enzymatic reaction, the concentrations of reaction products were determined by GC-FID. The reaction produced 33.0 mM 4MMP, while 8.1 mM mesityl oxide was also detected in the reaction mixture (Fig. S8), corresponding to a conversion of 48.5% of 4MMP from its cysteine-conjugated precursor. No other notable peaks were observed in the chromatograms, indicating a high degree of reaction specificity. The identity of the enzymatic reaction products was further confirmed by GC-MS. The product 4MMP and the co-product mesityl oxide were identified by comparison of their mass spectra with reference spectra from the NIST 23 library and commercially available standards. For 4MMP, the chromatographic peak at 8.18 min was subjected to library search and showed an excellent match with the NIST reference spectrum (Fig. 7A–C). The electron ionization (EI) mass spectrum of the bioconversion product displayed a base peak at m/z 43 and characteristic fragment ions at m/z 41, 55, 59, 75, 83, 89, 99, and 132, including the expected molecular ion at m/z 132. The relative fragment ion intensities closely matched those of the NIST reference spectrum, yielding high match factor (MF) and reverse match factor (RMF) values. Mesityl oxide was similarly confirmed by GC-MS through spectral comparison with the NIST library (Fig. 8A–C).

Fig. 7.

Mass spectrum of 4MMP obtained from A NIST 23 library, B the enzymatic reaction solution, and C the overlaid spectrum of (A) and (B). In the overlaid spectrum: red, enzymatic reaction; blue, NIST database

Fig. 8.

Mass spectrum of mesityl oxide obtained from A NIST 23 library, B the enzymatic reaction solution, and C the overlaid spectrum of (A) and (B). In the overlaid spectrum: red, enzymatic reaction; blue, NIST database

To evaluate downstream processing, the recovery of 4MMP from the enzymatic conversion broth was investigated. Solvent extraction followed by solvent-removal condensation afforded recoveries in excess of 90%. Prolonged solvent removal under reduced pressure resulted in a slight decrease in recovery yield, likely due to evaporation losses of the volatile thiol during concentration. Using the optimized extraction procedure, the final concentration of bioconverted 4MMP in ethanol exceeded 1% (w/w), a level suitable for industrial flavor applications (Iwabuchi et al. 2011).

Discussion

Previously established protocols by Lu et al. (2024) were adapted to prepare the precursors Cys-4MMP and Cys-4MP-one. Although the direct thia-Michael addition of l-cysteine to mesityl oxide has been reported previously (Li et al. 2021), inconsistent yields often limit its practical application. In contrast, the N-terminal protection strategy employed in this study improved yield stability and scalability. Moreover, by selecting appropriate α, β-unsaturated ketone precursors, this route provides a generalizable framework for preparing other mercaptoketone derivatives. To the best of our knowledge, NMR characterization of Cys-4MP-one is presented here for the first time.

The alkaline pH optimum observed for overexpressed S. putrefaciens MetC is consistent with previous studies on MetC, including enzymes from Lactobacillus delbrueckii subsp. bulgaricus ATCC BAA-365 (Allegrini et al. 2017). Human cystathionine γ-lyase, which shares over 30% sequence similarity with Shewanella MetC, has also been reported to exhibit optimal activity at pH 9 (Miyamoto et al. 2024). This pH dependence can be attributed to the PLP-dependent catalytic mechanism, in which Schiff base formation between PLP and the ε-amino group of an active-site lysine residue is favored under mildly alkaline conditions (Mehta and Christen 2000; Liang et al. 2019). The optimal temperature range of 30–37 °C reflects the mesophilic nature of S. putrefaciens (Moser and Nealson 1996). In contrast, β-lyase homologs derived from hyperthermophilic organisms such as Thermotoga maritima exhibit substantially higher temperature optimum, which is typically associated with enhanced structural rigidity and dense intramolecular interactions (Priyakumar 2012; Miyamoto et al. 2022).

The primary challenge in kinetic studies was the inherent solubility limits of the substrates in the reaction buffer, which prevented us from reaching the concentrations ([S] > 10 × $K_{\text{m}(0.5)}$) required to define a saturation plateau (l-cystine and l-cystathionine, and Cys-4MMP). Thus, we reported an apparent maximum velocity ($V_{\text{max},\text{app}}$) and apparent catalytic efficiencies [($V_{\text{max}}$/$K_{\text{m}}$)]_app for each substrate at their tested concentrations (Table S1). Although less efficient than with canonical substrates, MetC from S. putrefaciens exhibits significant activity toward the cysteine-conjugated precursors of both mercaptoketones. The non-Michaelis–Menten kinetics and substrate-dependent Hill coefficients observed in this study suggest the presence of cooperative or allosteric effects during catalysis. MetC functions as a PLP-dependent homotetramer, a structural arrangement often associated with allosteric regulation via transitions between low-affinity (T) and high-affinity (R) conformational states. Our data suggests that this transition is highly sensitive to the chemical nature of the substrate. The MetC from S. putrefaciens exhibited broad catalytic versatility across various structural precursors, demonstrating a multifunctional nature over narrow substrate specificity. Size-exclusion chromatography analysis demonstrated that S. putrefaciens MetC exists predominantly as a tetramer in solution. This quaternary organization is consistent with previous reports on MetC homologs, as the native Escherichia coli MetC enzyme has been shown to form a homotetramer composed of four identical subunits assembled into the functional enzymatic complex (Clausen et al. 1996). The conserved tetrameric assembly of MetC suggests its physiological importance. This study suggests that inter-subunit interactions may contribute to the observed cooperative kinetic behavior, which is directly linked to enzymatic activity. Previous research has pointed out that similar kinetic complexities have been reported for other PLP-dependent enzymes, including human mercaptopyruvate sulfurtransferase and cystathionine β-synthase (Sen and Banerjee 2007; Yadav et al. 2020). These findings further support the functional significance of the oligomeric arrangement.

A preparative-scale biotransformation was then carried out using 1 µM S. putrefaciens MetC in a 100-mL reaction volume containing 68 mM Cys-4MMP, with a conversion of 48.5% of 4MMP from its cysteine-conjugated precursor. For comparison, Holt et al. (2011) reported that Saccharomyces cerevisiae released 12.3 µM 4MMP from 2 mM Cys-4MMP, corresponding to a conversion below 1%. Similarly, the high-thiol-producing yeast strain S. cerevisiae Esperide® achieved a conversion efficiency of 0.84% (Bonnaffoux et al. 2018, 2021). The preparative-scale conversion of Cys-4MMP demonstrated a substantially higher conversion than previously reported systems. Notably, the substrate concentration used in this study (68 mM) is higher than the 2 mM used in previous studies of S. cerevisiae (Holt et al. 2011; Bonnaffoux et al. 2021), which may influence the observed conversion efficiency.

Despite this significantly higher substrate loading, S. putrefaciens MetC maintained high conversion, whereas prior yeast-based systems reported conversions below 1%. This performance highlights the enzyme’s robustness and a superior observed catalytic rate, which remains highly competitive even under the non-saturating conditions tested in this study, thereby suggesting its potential for industrial-scale applications where high space-time yields and tolerance to high substrate concentrations are required (Bonnaffoux et al. 2018). The formation of 4MMP confirmed successful cleavage of the cysteinyl moiety by S. putrefaciens MetC, releasing a free thiol known for its potent tropical aroma in wine and fermented foods (Swiegers and Pretorius 2007). To better understand the origin of mesityl oxide formation, the chemical stability of both the precursor and product was evaluated under the reaction buffer conditions. Cys-4MMP and 4MMP were incubated separately in buffer in the absence of enzyme, and mesityl oxide formation increased over time in both cases, indicating spontaneous formation under mildly alkaline conditions. The decomposition of Cys-4MMP is consistent with previous reports on the instability of Michael addition adducts under alkaline conditions and likely represents a major pathway contributing to mesityl oxide formation. In contrast, 4MMP degradation appears to play a minor role, potentially via β-elimination. Together, these observations suggest that the intrinsic chemical instability of the precursor and product plays a prominent role in mesityl oxide formation during the biotransformation process (Escobar et al. 2016).

Finally, the MetC from S. putrefaciens displayed broad catalytic activity toward structurally diverse precursors, indicating a multifunctional rather than strictly substrate-specific profile. The high recovery yield and ability to concentrate 4MMP to levels well above its sensory threshold underscore the potential of S. putrefaciens MetC as a promising biocatalyst for the industrial production of aroma-active thiols.

Conclusion

This study successfully characterized a MetC from the marine bacterium S. putrefaciens, demonstrating its significant catalytic potential. Enzyme kinetics analysis revealed efficient activity toward mercaptoketone precursors, specifically Cys-4MMP and Cys-4MP-one. A conversion of 48.5% was achieved in our preparative-scale enzymatic production of 4MMP, a compound known for its desirable fruity and tropical organoleptic properties. While promising, minimizing the formation of mesityl oxide as a co-product will require further optimization of catalysis conditions. Overall, these findings provide perspectives on the production of valuable thiol aroma compounds for sustainable industrial manufacture of aromas and fragrances.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

H.-T.V.L. conceived and designed the experiments; Y.-C.H., H.-K.L., Y.-L.L. H.H., T.-C.L. performed the experiments; Y.-C.H., H.-K.L., and T.-C.L. analyzed the data; Y.-C.H., H.-K.L., T.-C.L., and H.-T.V.L. wrote the paper. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by National Science and Technology Council (NSTC 113-2221-E-019-009) and the Center of Excellence for the Oceans, National Taiwan Ocean University, from The Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan (NTOU-RD-AA-2021-1-02018). HFNMRC is funded by Academia Sinica Core Facility and Innovative Instrument Project (AS-CFII-111-214).

Availability of data and materials

Data will be made available on request.

Declarations

Ethical approval

Not applicable.

Competing interests

The authors declare no competing interests.