Enzymatic properties and inhibition tolerance analysis of key enzymes in β-phenylethanol anabolic pathway of Saccharomyces cerevisiae HJ

Abstract

Huangjiu is known for its unique aroma, primarily attributed to its high concentration of β-phenylethanol (ranging from 40 to 130 mg/L). Phenylalanine aminotransferase Aro9p and phenylpyruvate decarboxylase Aro10p are key enzymes in the β-phenylethanol synthetic pathway of Saccharomyces cerevisiae HJ. This study examined the enzymatic properties of these two enzymes derived from S. cerevisiae HJ and S288C. After substrate docking, Aro9p^HJ (−24.05 kJ/mol) and Aro10p^HJ (−14.33 kJ/mol) exhibited lower binding free energies compared to Aro9p^S288C (−21.93 kJ/mol) and Aro10p^S288C (−12.84 kJ/mol). ARO9 and ARO10 genes were heterologously expressed in E. coli BL21. Aro9p, which was purified via affinity chromatography, showed inhibition by l-phenylalanine (L-PHE), but the reaction rate V_max(Aro9p^HJ: 23.89 μmol·(min∙g)⁻¹) > Aro9p^S288C: 21.3 μmol·(min∙g)⁻¹) and inhibition constant K_i values (Aro9p^HJ: 0.28 mol L⁻¹>Aro9p^S288C 0.26 mol L⁻¹) indicated that Aro9p from S. cerevisiae HJ was more tolerant to substrate stress during Huangjiu fermentation. In the presence of the same substrate phenylpyruvate (PPY), Aro10p^HJ exhibited a stronger affinity than Aro10p^S288C. Furthermore, Aro9p^HJ and Aro10p^HJ were slightly more tolerant to the final metabolites β-phenylethanol and ethanol, respectively, compared to those from S288C. The study suggests that the mutations in Aro9p^HJ and Aro10p^HJ may contribute to the increased β-phenylethanol concentration in Huangjiu. This is the first study investigating enzyme tolerance mechanisms in terms of substrate and product, providing a theoretical basis for the regulation of the β-phenylethanol metabolic pathway.

Graphical abstract

Highlights

• •Aro9p^HJ and Aro10p^HJ had lower free energies for docking with the substrate than Aro9p^S288C and Aro10p^S288C.
• •Aro9p^HJ was more tolerant to inhibition when the substrate L-PHE concentration exceeded 2 mM.
• •Aro10p^HJ had a stronger affinity for Phenylpyruvate than Aro10p^S288C.
• •Aro9p^HJ and Aro10p^HJ showed higher tolerance to β-phenylethanol and ethanol than Aro9p^S288C and Aro10p^S288C.

1. Introduction

Huangjiu is fermented by adding wheat qu and some enzyme preparation, yeast, and other saccharifying starters, such as alcohol, aldehyde, ester, acid, and other substances, to enrich the body of Huangjiu mellow and typica [1,2]. As one of the aromatic alcohols, β-phenylethanol is one of the main by-products produced by Saccharomyces cerevisiae during the brewing process of Huangjiu [3,4]. The content of β-phenylethanol in Huangjiu is 40 mg/L-130 mg/L β-phenylethanol to a certain extent enhances the flavor of seasonings such as Huangjiu and other fermented foods, thereby positively impacting their sensory characteristics [5]. However, when its concentration is excessively high, it may potentially pose negative effects on the human body. Therefore, understanding the metabolic regulation mechanisms of β-phenylethanol is crucial for healthy consumption of such products.

The aromatic alcohols in Huangjiu are mainly produced by S. cerevisiae, and their concentration is highly positively correlated with the concentration of ethanol. Previous studies on promoter strength and two-dimensional gene expression analysis show that aromantic alcohol biosynthesis of S. cerevisiae is primarily based on the Ehrlich pathway. The concentration of total aromatic alcohols produced by S. cerevisiae HJ01 was higher than that of the model strain [6].

A series of enzymes in the Shikimic acid and Ehrlich pathways catalyze the synthesis of β-phenylethanol by S. cerevisiae [7,8]. In the Ehrlich pathway, S. cerevisiae transaminates l-phenylalanine (L-PHE) in the presence of phenylalanine aminotransferase to form phenylpyruvate (PPY), and decarboxylase decarboxylates the decarboxylation of carboxyl to form phenylacetaldehyde. Finally, β-phenylethanol is reduced to β-phenylethanol by dehydrogenase [9,10], the other is the Shikimic acid pathway. However, little is known about the biosynthetic pathway of aromatic alcohols in Huangjiu fermentation and the key enzymes involved in this process. In our previous studies, two key enzymes have been isolated, purified, and characterized from S. cerevisiae HJ and S. cerevisiae S288C. Mutations in PDT (I161K, L239P, and Q250H) were identified in S. cerevisiae HJ01. Pha2p^HJ exhibited higher substrate affinity (K_m, 11.1 mM–15.9 mM), catalytic efficiency, and thermal stability but was less sensitive to feedback inhibition. Therefore, the PDT mutation of S. cerevisiae HJ01 leads to a higher concentration of β-phenylethanol in Huangjiu. There were four amino acid differences (58(V-T), 127(Q-E), 147(Q-E), and 151(I–V)) in the active catalytic region of alcohol dehydrogenase Adh1p [11], which led to a higher tolerance to β-phenylethanol and ethanol of yeast strain Adh1p^HJ than Adh1p^S288C.

Transaminization is the rate-limiting step in the Ehrlich pathway, and the ARO10 gene mainly determines phenylpyruvate decarboxylase activity and that PPY will be preferentially catalyzed by the ARO10-encoded enzyme during reaction [12,13]. Many studies have shown a close link between enzyme expression and aromatic alcohol production in beer and wine [14]. By overexpressing ARO80, ARO9, and ARO10 genes simultaneously, the concentration of β-phenylethanol produced by S. cerevisiae was increased threefold. ARO80 is a crucial transcription factor in the aromatic alcohol synthesis pathway [15]. Therefore, it is essential to explore the phenylalanine aminotransferase and phenylpyruvate decarboxylase encoded by ARO9 and ARO10 genes in huangjiu(Fig. 1).

Fig. 1.

Phenylalanine aminotransferase and phenylpyruvate decarboxylase of the Ehrlich pathway of S. cerevisiae

Aromatic amino acid aminotransferase I and aromatic amino acid aminotransferase II, both of which belong to the first family of aminotransferases, are encoded and synthesized by genes ARO8 and ARO9, respectively. Through the research and comparison of genes, it is found that the amino acid difference between Aro8p^HJ and Aro8p^S288C is located at the edge of the protein structure, so it will not be studied for the time being (data not shown). According to previous studies, PDC1, PDC5, PDC6, and THI3 mainly encode pyruvate decarboxylase, and the activity of phenylpyruvate decarboxylase is mainly determined by the ARO10 gene. Therefore, this study focused on ARO9 and ARO10.

In this study, our aim was to enhance the understanding of the enzyme regulatory mechanism involved in the key biosynthetic pathway of β-phenylethanol - the Ehrlich pathway, during Huangjiu fermentation. We primarily investigated the structure, enzymatic properties and kinetics of Aro9p and Aro10p from S. cerevisiae HJ and model strain S288C, with a particular focus on the mechanisms of substrate and end product resistance [16].

2. Materials and methods

2.1. Reagents

Sodium dodecyl sulfate (SDS), Histrap™ HP column, and molecular mass markers were obtained from GE Healthcare Life Sciences (Sweden). β-mercaptoethanol, kanamycin, and IPTG were obtained from Sigma-Aldrich. All chemicals were of analytical grade and commercially available.

2.2. Strains, plasmids, and materials

Strain S. cerevisiae HJ was isolated from the traditional Huangjiu fermentation process at Zhejiang Guyuelongshan Shaoxing Wine Co., Ltd., the largest Huangjiu producer in the world, as a β-phenylethanol producing yeast strain. S. cerevisiae S288C was used as the comparator strain (Table 1). The genomic DNA of S. cerevisiae and ARO9 and ARO10 were prepared according to our previous reports [17]. The primers for ARO9 amplification were GGGAATTCCATATGATGACTGCTGGTTCTGTCCC, which contained a Nde I site (underlined), and CGAGCTCTCAACTTTTATAGTTGTCAA, which contained a Sac I site (underlined). And the primers for ARO10 amplification were GGGAATTCCATATGATGGCACCTGTTACAATTGA, which ontained a Nde I site (underlined), and GGATCCCTATTTTTTATTTCTTTTAA, which ontained a BamH I site (underlined). The PCR products were ligated with vector pESI-T and then digested by Nde I and Sac I(BamH I). The resulting fragment was ligated with similarly restricted pET-28a (+) to construct pET-28a-ARO9^S288C and pET-28a-ARO9^HJ, pET-28a-ARO10^S288C and pET-28a-ARO10^HJ, respectively. These two plasmids were used individually to transform E. coli BL21(DE3), and Aro9p production was induced in the resulting strains with IPTG. TB medium (12 g/L tryptone, 24 g/L yeast extract, 5 g/L glycerol, 2.31 g/L KH₂PO₄ and 12.54 g/L K₂HPO₄) supplemented with kanamycin (30 mg/L) was used for protein production [18].

Table 1.

Microbial strains and plasmids used in this study.

Strain and plasmids	Relevant characteristics	Source or reference
E. coli BL21(DE3)	F−ompT hsdSB (rB−mB−) gal dcm (DE3)	Novagen
S. cerevisiae S288C	MATα (leu2-3, can1-100,112 ura3-1 trp1-1, his3-11, ade2-1)	ATCC 27325
S. cerevisiae HJ	huangjiu brewing yeast	huangjiu fermentation broth
pESI-T Vector	ampr	TaKaRa, Japan
pET-28a(+)	KanrlacI	Novagen, USA
pET-28a-ARO9S288C/pET-28a-ARO9HJ	KanrlacI ARO9S288C/ARO9HJ	This work
pET-28a-ARO10S288C/pET-28a-ARO10HJ	KanrlacI ARO10S288C/ARO10HJ	This work

pESI-T Vecto is r as a cloning vector.

pET-28a (+) is used as a subcloning vector.

2.3. Sequence analysis and 3D structure analysis based on homology modeling

The ARO9^HJand ARO10^HJ gene in plasmid pET-28a-ARO9^HJ and pET-28a-ARO10^HJ were sequenced by Sangon Biotech (Shanghai) Co., Ltd. The ProtParam tool (https://web.expasy.org/protparam/) was used to predict the molecular weight of genes through the amino acid sequence. The online tool TMHMM 2.0 (tool website: http://www.cbs.dtu.dk/services/TMHMM/) was used to predict transmembrane helices. The predicted three-dimensional (3D) structure of Aro9p^HJ was obtained using the Automated Mode in the SWISS-MODEL server (http://swissmodel.expasy.org/). The template sequence for the enzyme was obtained using the PDB (http://www1.rcsb.org/) database. A multiple sequence alignment of the deduced amino acid sequences of Aro9p^S288C and Aro9p^HJ the Aro9p from Candida albicans SC5314 was prepared using DNAMAN software. The structural models were viewed using PyMOL software [19].

Interactions between Aro9p (Aro10p) and L-PHE (PPY) were evaluated by a molecular docking study. 3D structures of ligands were retrieved from the PubChem database in SDF format. Molecular docking was carried out using AutoDock 4.2, and a grid box covering the 3D coordinates in X, Y, and Z-dimensions of active sites was created using AutoGrid. After a successful docking run, the conformation with the smallest binding energy was selected from the 100 results [20].

2.4. Protein expression, enzyme preparation, and purification

Pure enzymes were performed by metal-affinity chromatography. Firstly, When the optical density (OD) at 600 nm reached 0.6–0.8 (about 3–4 h), the cultures were cooled on ice, and gene expression was induced by the addition of 0.5 mM (0.4 mM) IPTG, and the mixture was cultured at 25 °C for an additional 12 h [5]. The biomass was harvested by centrifugation for 10 min at 12000 rpm and 4 °C. The resulting cell pellets were washed, and resuspended in PBS buffer (300 mM NaCl and 50 mM phosphate buffer at pH 8.0). Then the resuspended cells were incubated on ice for 20 min and lysed by sonication (Sonics Vibra - Cell) (300 W, 2 s, 1 s, Crushing time: 10 min). The crude extract was centrifuged at 12000 rpm for 30 min at 4 °C to remove cellular debris. As a control, E. coli BL21(DE3) cells harboring plasmid pET-28a (+) were treated similarly [21]. The crude extracts were used for activity determination and purification. Then the soluble portion was loaded on a Histrap™ HP chelated column connected to an AKTA avant 25 (GE Healthcare). The bound enzyme was eluted with 250 mM imidazole in 20 mM phosphate buffer (pH 7.0). Enzyme solutions were desalted and stored at 4 °C until further use [22].

2.5. Enzyme assays

2.5.1. Enzyme assays of phenylalanine aminotransferase Aro9p

The reaction mixture (1 mL) consisted of potassium dihydrogen phosphate buffer (0.1 mmol L⁻¹, pH 8.0), pyridoxine phosphate (0.1 mmol L⁻¹), l-phenylalanine (1 mmol L⁻¹), pyruvate (10 mmol L⁻¹) as amino receptor, and enzyme (0.2 mg protein-0.6 mg protein). The reaction started with the addition of l-phenylalanine. When 1 mL of 1 mol L⁻¹ sodium hydroxide was added to the ice surface and reacted at 30 °C for 2 min and 10 min, the two parallel reactions of transaminase activity stopped. PPY, the reaction product of phenylalanine aminotransferase, was determined at 320 nm. The assay was carried out at least in triplicate [23].

$$\mathrm{U}/\left(\mathrm{\mu }\text{mol}/\min \right)=\frac{V\bullet {△A}_{320}}{t\bullet \varepsilon \bullet \mathrm{b}}\times 100$$

In the formula, V was the volume of the reaction system (mL); △A₃₂₀ was the absorbance difference; t was reaction time (min); The $\varepsilon$ is in mole extinction coefficient (17.5 × 10³ L (mol·min)⁻¹); b was the optical path (cm); The enzyme activity unit (U) of one phenylalanine aminotransferase was defined as the amount of enzyme required to produce 1 μmo1 product PPY within 1min. Specific enzyme activity (U∙g⁻¹) was defined as the unit of enzyme activity per g of phenylalanine aminotransferase Aro9p [24].

2.5.2. Enzyme assays of phenylpyruvate decarboxylase Aro10p

The activity of phenylpyruvate decarboxylase was determined by coupled optical analysis assisted by alcohol dehydrogenase, and the activity of phenylpyruvate decarboxylase was determined by consuming NADH at 340 nm. The assay system was as follows: 200 μL reaction system: containing sodium phosphate (20 mmol L⁻¹, pH 7.0), ThDP (0.2 mmol L⁻¹), Mg²⁺ (0.1 mmol l⁻¹), NADH (0.35 mmol L⁻¹), yeast alcohol dehydrogenase(1 U∙mL⁻¹). The reaction started with the addition of pure enzyme (40 μL, 50 μg mL⁻¹) and reacted at 37 °C for 15 min. The activity of phenylpyruvate decarboxylase was determined by monitoring the reduction of NAD⁺ at 340 nm. The calculation formula is as follows [25]:

$$\mathrm{U}/\left(\mathrm{\mu }\text{mol}/\min \right)=\frac{V\bullet {△A}_{340}}{t\bullet \varepsilon \bullet \mathrm{b}}\times 100$$

A unit of phenylpyruvate decarboxylase activity (U) is defined as the amount of enzyme required to reduce 1 μmo1 of coenzyme NADH in 1 min. The specific enzyme activity(U∙g⁻¹) was defined as the enzyme activity unit per g of phenylpyruvate decarboxylase Aro10p [26].

2.6. Enzyme characterization and kinetics

2.6.1. Effect of temperature on activity and stability of the purified enzyme

The optimal temperature of purified Aro9p (Aro10p) was determined at temperatures ranging from 15 °C to 70 °C. For thermal stability, residual activity was determined after incubation by the value of the temperature range employed for 1 h using L-PHE(PPY) as substrate at 40 °C and pH 7.0. The experiment was performed in triplicate [27].

2.6.2. Effect of pH on activity and stability of the purified enzyme

The optimal pH of purified Aro9p (Aro10p) was determined using L-PHE (PPY) as substrate in 20 mM sodium phosphate-sodium citrate (pH 3.0–8.0) and 20 mM sodium carbonate-sodium bicarbonate (pH 9.0–10.0). For pH stability, the residual activity was measured at 40 °C and pH 6.0 using L-PHE(PPY) as substrate after incubation at various pH values for 1 h. The experiment was performed in triplicate [18].

2.6.3. Kinetic parameter analysis and effectiveness factor

The kinetic parameters of Aro9p (Aro10p), such as the maximum velocity (V_max), Michaelis-Menten constant (K_m), turnover number (k_cat), and catalytic efficiency (k_cat/K_m), were determined after calculating the activity in the presence of different concentrations (0 mmol L⁻¹ - 5 mmol L⁻¹ substrate) at the optimum temperature and pH. A Lineweaver-Burk plot of 1/v versus 1/s was employed for the V_max and K_m calculations. The data were fitted to the linear model using a double-reciprocal plot. K_m is the value of the Michaelis–Menten constant without the presence of the inhibitor. Inhibition Constant K_i was calculated with reference to previous studies [28].

2.7. Statistical analysis

Origin Pro 9.0 software (OriginLab Corp., Northampton, MA, USA) was used for data sorting and calculation and atlas production, and Prism7.0 software (GraphPad Software, San Diego, CA) software was used for significance analysis. Except for special instructions, three parallels were set for each sample, and data were expressed as mean ± standard deviation.

3. Results

3.1. Sequence and structural analysis

3.1.1. Sequence and structural analysis

Physicochemical characterizations of proteins are closely related to their structures, which can lay the foundation for studying their functions and interactions with small molecules. Physicochemical parameters such as pI, amino acid composition, instability index, amino acid hydrophilicity, hydrophobicity, etc., were analyzed for the Aro9p and Aro10p. This could provide a reference for predicting the advanced structure of the protein [29]. The Prot Param program, based on a sequence of enzyme bioinformatics prediction, found that the molecular formula of Aro9p^HJ and Aro10p^HJ were C₂₆₄₅H₄₀₇₂N₆₇₆O₇₈₆S₁₉ and C₃₁₇₆H₄₉₃₁N₈₃₃O₉₄₈S₂₆, respectively, and the molecular weight were 58.53 kDa and 70.79 kDa. The total number of atoms were 8198 and 9914, and the computed pI value were 5.27 and 5.76 (pI < 7). This suggests that the protease is least soluble when the pH of the solution is at its isoelectric point [30]. The total number of basic amino acid residues (Arg + Lys) were 55 and 59, less than those of acidic amino acid residues (Asp + Glu), which were 66 and 70, indicating that phenylalanine aminotransferase Aro9p^HJ and phenylpyruvate decarboxylase Aro10p are acidic protein. The Grand Average Hydropathy (GRAVY) values were found to be −0.322 and −0.183. These negative GRAVY values suggest that these proteins could exhibit better interaction with water (hydrophilic in nature) [31].

3.1.2. Homology modeling analysis and substrate molecular docking study

The open reading frames of Aro9p^HJ and Aro9p^S288C both encompass 1542 base pairs, encoding protein sequences of 513 amino acids each. According to a BLAST analysis against the UniProt database and homology modeling using the SWISS-MODEL, these two proteins exhibit the highest amino acid identities of 37.97 % and 38.35 % respectively, to the aromatic aminotransferase Aro9 from Candida albicans(PDB ID:6HND) [32]. The tertiary structure of these proteins provides important insights into their functional molecular mechanisms. Homology-based protein modeling is an effective computational tool for predicting the structure of unknown proteins based on the determined 3D structures of other proteins in the same family that have similar folds and/or functions [33]. The homology modeling prototypes of Aro9p^HJ and Aro9p^S288C were determined to be aromatic-amino-acid: 2-oxoglutarate transaminase from Candida albicans SC5314. As shown in Fig. 2 (A and C), the structure of Aro9p includes a dimer of 59 kDa protein subunits, each consisting of 513 amino acids. This is similar to the structure of CaAro9p [32]. The N-terminal region of phenylalanine aminotransferase Aro9p contains two antiparallel β-folded sheets. These are distinct characteristics of Aro9p that differ from other transaminases and are related to the formation of the Aro9p dimer [34].

Fig. 2.

The sequences and structures of phenylalanine aminotransferase Aro9p An L-PHE residue is bound covalently (forming an external aldimine) to the PLP cofactor in the ligand-binding cavity of Aro9p. Residues of subunits I and II surround the L-PHE ligand. Aro9p^S288C -dimer (A); Aro9p^HJ -dimer (B); L-PHE bound to Aro9p^S288C (C); L-PHE bound to Aro9p^HJ (D); Multiple alignments of amino acid sequences of phenylalanine aminotransferase Aro9p from Candida albicans SC5314. (PDB ID: 6HND), S. cerevisiae HJ(ARO9-HJ), and S. cerevisiae S288C(ARO9-S288C) (E). Aro9p^S288C: NP_012005.1; ARO9^S288C: GENE ID:856539.

The enzymes Aro9p catalyzes the interconversion of amino acids and oxoacids through a transfer of amino groups. Aminotransferases utilized pyridoxal-5′-phosphate (PLP) as a coenzyme to transit the amino groups. Sequence analysis revealed noticeable differences between amino acids 12 (T-A), 79 (N-D), 454 (L-V), and 460 (V-L) in Aro9p^HJ and Aro9p^S288C, which are located in the unconserved region. The Center Grid box coordinates were set as (X: 65.506, Y: 39.431, Z: 13.620), and the small molecule compound (L-PHE) connected to the binding site of Aro9p, as shown in Fig. 2B and D. In general, the configuration with the lowest docking energy was chosen as the most likely binding conformation. In this case, the minimum Aro9p^HJ docking score was −24.05 kJ/mol, with docking in Aro9p^HJ - PHE compounds found in the ASN144, PHE168, ASN232, LYS317, ARG324, ARG481, GLY50(from the other subunit), TYR102(from the other subunit), and ARG347(also from the other subunit) residues as the essential residues. The minimum Aro9p^S288C docking score was −21.93 kJ/mol, with docking in Aro9p^S288C - PHE compounds found in the ASN 144, PHE168, PRO170, ASN232, LYS317, ARG324, PHE426, ARG481, LYS32 (from the other subunit), GLY50 (also from the other subunit), TYR102 (also from the other subunit), and ARG347 (also from the other subunit) residues as the critical residues. These results could assist in understanding the enzyme's conformational specificity and illustrate the superior spontaneous binding of the substrate (L-PHE) to Aro9p^HJ.

Simultaneously, both Aro10p^HJ and Aro10p^S288C are coded by open reading frames of 1908 base pairs, which encode amino acid sequences of 635 residues each. The BLAST analysis compared to the UniProt database and homology modeling against the SWISS-MODEL revealed that these two sequences have the highest amino acid identity of 34.92 % and 34.95 % respectively, to the pyruvate decarboxylase from S. cerevisiae (PDB ID:2W93) [35]. This suggests that the homologous modeling prototypes for Aro10p^HJ and Aro10p^S288C are derived from the pyruvate decarboxylase of S. cerevisiae. The enzymatic function of Aro10p relies on the thiamine diphosphate (ThDP) cofactor, which primarily binds to the protein component at the interface via a divalent metal ion, typically magnesium. Pyruvate decarboxylases are enzymes composed of multiple subunits, with a typical subunit having a molecular weight between 59 kDa and 61 kDa (Fig. 3A and C). The tetramer is the catalytically active form of most PDCs [26,36]. Based on the crystal structure of ScPDC, it is hypothesized that only in the tetrameric state can the intermediate (R-) domain, which is very flexible and contains C221, interact with other domains in the other subunits [35,36]. Sequence analysis revealed that the notable difference in amino acids 377 (A-V) and 527 (K–R) between Aro10p^HJ and Aro10p^S288C was located in the unconserved region (Fig. 3F).

Fig. 3.

The sequences and structures of pyruvate decarboxylase Aro10p

PPY residue is bound covalently (forming an external aldimine) to the PPY cofactor in the ligand-binding cavity of Aro10p. Aro10p^S288C - monomer (A); Aro10p^HJ - monomer (B); PPY bound to Aro10p^S288C (C); PPY bound to Aro10p^HJ (D); Multiple alignments of amino acid sequences of pyruvate decarboxylase Aro9p from S. cerevisiae HJ(ARO10-HJ), and S. cerevisiae S288C(ARO10-S288C) (E). Mapping of amino acid differences (Yellow: 2W93; Grey: Aro10p^HJ; Magentic red: structural loop 288–304 and the enzyme regulatory site C221; Red: amino acids of Aro10p^HJ V377 and R527) (F); Aro10p^S288C: NP_010668.3; ARO10^S288C: GENE ID:851987.

Amino acid residues at the active site directly interact with substrates attached to the C2 atom of the cofactor, substantially increasing enzyme catalysis speed. Therefore, the C2 position of ThDP was used as the grid box center (X: 12.927, Y: 17.033, Z: 24.852). The small molecule compound (PPY) was connected to the binding site of Aro10p, as shown in (Fig. 3B and D). In this situation, the lowest Aro10p^HJ docking score was −14.33 kJ/mol, with Aro10p^HJ - PPY compounds revealing Glu334, Ile335, Thr444, Gly445, Gln448, Phe449, Leu469, Ile544, Ile548, Leu621, Met624, Val625, Ala628 residues as key residues. The lowest Aro10p^S288C docking score was −12.84 kJ/mol, with the key residues near the ligand in the docked Aro10p^S288C – PPY compounds being identical to those in the Aro10p^HJ - PPY compounds. These results could assist in understanding the enzyme's conformational specificity and illustrate the better spontaneous binding of the substrate (PPY) to Aro10p^HJ.

In summary, sequence analysis showed that the amino acid mutation in Aro9p^HJ and Aro10p^HJ might be one of the reasons for the better spontaneous binding of substrate (L-PHE or PPY) to enzyme. The results might be elucidated further through the study of enzymology properties in experimental analysis.

3.2. Gene expression and enzyme purification

ARO9 (1542 bp) and ARO10 (1928 bp) genes (Figs. S1A and D) were amplified and cloned into pET28a-ARO9 and pET28a-ARO10 plasmids, respectively, and transformed to E. coli BL21 (DE3). Plasmid maps are shown in Figs. S1B and E. The recombinant plasmids were verified by restriction endonucleases and sequenced by Shanghai Sangon.

The ARO9 and ARO10 gene encoding 514 and 635 amino acids respectively, are expressed in recombinant bacteria, resulting in proteins with molecular weights of 58.53 kDa and 70.79 kDa. In short, the prepared IPTGs were added to TB liquid mediums until the final concentrations were 0.5 mmol L⁻¹ and 0.4 mmol L⁻¹, induced at 25 °C for 14 h and 12 h, and then the bacteria were centrifuged to obtain the precipitated thallus. The results were shown in Figs. S2A and C, characteristic bands appeared at about 60 kDa and 72 kDa which were consistent with the expected molecular weight of the designed protein [36].

The crude enzyme solutions were crushed by centrifugation, and the supernatants were pretreated with 0.22 μm membrane. Affinity chromatography was used to separate and purify the target proteins, and the eluents at peak positions were collected (Figs. S2B and E). The protein distributions were analyzed using SDS-PAGE, showing molecular weights approximately 60 kDa and 72 kDa, which were close to the predicted molecular weights of the recombinant enzymes, indicating successful overexpression and purification of phenylalanine aminotransferase Aro9p and phenylpyruvate decarboxylase Aro10p in E. coli strains (Figs. S2C and F).

3.3. Biochemical characterization of Aro9p

The pH and temperature characteristics of the phenylalanine aminotransferase Aro9p were studied. The experimental results showed that the optimal reaction pH of phenylalanine aminotransferase Aro9p was 6.0. The results were shown in Fig. 4A. In the range of pH 3.0–6.0, the enzyme activity increased; when the pH was greater than 6.0, the enzyme activity gradually weakened. The most stable pH of Aro9p was 7.0. Fig. 4B showed that after 1h incubation of phenylalanine aminotransferase Aro9p in the range of pH 7.0–8.0, more than 80 % of the enzyme activity remained, and the enzyme activity of Aro9p^HJ was slightly higher than that of Aro9p^S288C. The residual enzyme activity of Aro9p^S288C was less than 40 % after 1 h incubation in a buffer of pH 3.0–5.0. It also indicated that phenylalanine aminotransferase Aro9p was relatively stable under neutral pH conditions.

Fig. 4.

Optimal pH and temperature for enzyme activity and stability of Aro9p and Aro10p

Relative enzyme activity under different pH reaction conditions of Aro9p(A); Stability of the enzyme Aro9p under different pH conditions (B); Relative enzyme activity under different pH reaction conditions of Aro10p(C); Stability of the enzyme Aro10p under different pH conditions (D); The relative Aro9p enzyme activity was measured at different temperature (E); The enzyme stability of Aro9p under different temperature (F); The relative Aro10p enzyme activity was measured at different temperature (G); The enzyme stability of Aro10p under different temperature (H).

The results showed that the optimum reaction temperature of phenylalanine aminotransferase Aro9p was 30 °C, and the enzyme activity of Aro9p^HJ (17.86 U g⁻¹) was similar to that of Aro9p^S288C (17.61 U g⁻¹). The results were shown in Fig. 4E. Within the range of 15 °C–30 °C, and the enzyme activity increased with the increase in temperature. When the temperature was higher than 30 °C, the enzyme activity decreased relatively quickly. The most stable temperature of Aro9p was 30 °C. Fig. 4F showed that after incubation at 15 °C–30 °C for 2 h, the enzyme activity of phenylalanine aminotransferase Aro9p was still more than 60 %. When the temperature was higher than 30 °C, the enzyme activity began to weaken, and when the temperature was higher than 60 °C, the remaining enzyme activity was less than 40 %.

According to some relevant studies, aromatic amino acid transaminase (AAATs) from Proteus mirabilis JN458, a key enzyme in the β-phenethyl alcohol synthesis pathway, exhibits high enzyme activity at temperatures ranging from 30 to 50 °C and pH levels between 6 and 9 [37]. The enzyme VpAT isolated from Variovorax paradoxus demonstrated high activity across a broad pH range (4–11.2) at 30 °C. At 55 °C, VpAT presents a maximum specific activity of 33 U mg⁻¹, approximately twice the specific activity at 30 °C [38]. For both PHE and LEU activities of aminotransferase from Lactococcus, the optimum temperature determined at pH8 ranged between 35 and 45 °C. Likewise, the activity of PDT, similar to that of S. cerevisiae, declined sharply in low pH conditions, suggesting that an acidic environment damages its activity [39].

3.4. Biochemical characterization of Aro10p

The experimental results showed that the optimal reaction pH of Aro10p was 7.0. The results were shown in Fig. 4C. The enzyme activity increased with the increase of pH within the range of 3.0–7.0 and gradually weakened when the pH was greater than 7.0. The most stable pH of Aro10p was 7.0. Fig. 4D showed that after incubation for 1h for phenylpyruvate decarboxylase Aro10p within the pH range of 7.0–8.0, more than 80 % of the enzyme activity remained, and the enzyme activity of Aro10p^HJ was relatively slightly higher than that of Aro10p^S288C. The residual enzyme activity of Aro10p^S288C was less than 40 % after 1h incubation in a buffer with pH 3.0–5.0.

The results showed that the optimum reaction temperature of Aro10p was 35 °C, and the specific enzyme activity of Aro10p^HJ (77.81 U g⁻¹) was 25.49 % higher than that of Aro10p^S288C (62.00 U g⁻¹). The results were shown in Fig. 4G. Within the range of 15 °C–40 °C, specific enzyme activity increased with the increase in temperature. When the temperature was higher than 40 °C, the relative enzyme activity decreased rapidly. The most stable temperature of Adh1p is 30 °C. Fig. 4H showed that after 1h incubation at 20–35 °C, more than 80 % of the enzyme activity remains. When the temperature was higher than 35 °C, the relative enzyme activity decreased rapidly, and when the temperature was higher than 60 °C, the remaining enzyme activity was less than 30 %.

The optimum pH of phenylpyruvate decarboxylase KDC4427 from Bacterial Enterobacter sp. CGMCC 5087 for substrate PPY was between 6 and 6.5, and the optimum temperature was 37 °C. KDC4427 was relatively stable between pH 6 and 8, but only 41 % of the residual activity was detected after incubation at 50 °C for 1 h [40].

3.5. Kinetic analysis

As depicted in Fig. 5, the reaction rate of Aro9p^HJ (17.61 μmol·(min∙g)⁻¹) was comparable to that of Aro9p^S288C (17.86 μmol·(min∙g)⁻¹) when the substrate concentration of L-PHE was 1 mM. However, when the substrate concentration of L-PHE reached 2 mM, the V_max reaction rate of Aro9p^HJ (23.89 ± 1.35 μmol·(min∙g)⁻¹) was 12.16 % higher than that of Aro9p^S288C (21.3 ± 1.12 μmol·(min∙g)⁻¹). The enzyme activity began to show some inhibition when the L-PHE concentration exceeded 2 mM.

Fig. 5.

Kinetic analysis of Aro9p and Aro10p

Non-linear regression of substrate and reaction rate of Aro9p(A); Non-linear regression of substrate and reaction rate of Aro10p(B); Relationship between 1/[S]-1/V of Aro10p^S288C(C); Relationship between 1/[S]-1/V of Aro10p^HJ(D).

Homology modeling results illustrated that CaAro9p from C. albicans was selected as the template for Aro9p. Unlike S. cerevisiae, CaAro9p demonstrated lower affinity for L-PHE (K_m: 4.76 ± 1.17 mM) than for L-TRP (K_m: 1.80 ± 0.18 mM). Moreover, the V_max reaction rate of CaAro9p (7.457 ± 457 μmol·(min∙g)⁻¹) to L-PHE was lower than that of Aro9p^HJ of S. cerevisiae HJ (23.89 μmol·(min∙g)⁻¹). These characteristics may explain why S. cerevisiae, especially HJ, is more adept at utilizing L-PHE to produce PPY [41,42].

In terms of structure, the N-terminal domain is preceded by a 50-residue loop that extends over the surface of the other subunit, covering part of its substrate-binding site. There are amino acid differences in the N-terminal domain of Aro9p^HJ, Aro9p^S288C and CaAro9p, located at the gating access position implicated in substrate entry into the active site. This might account for the varying affinity and reaction rates of the enzyme to the substrate.

In fact, there are also some studies on aromatic amino acid aminotransferase involved in β-phenylethanol biosynthesis in rose petal protoplasts. Under optimal pH and temperature conditions, recombinant RyAAAT3 revealed a V_max value for the conversion of L-PHE of 6.86 ± 0.04 nmol/mg protein/min [43]. Aro9p^HJ of S. cerevisiae HJ might be more promising than rose petal protoplasts for further synthesis of β-phenylethanol from L-PHE [44].

Despite studies suggesting Aro9p is inhibited by the substrate keto acid, there have been few studies on the substrate inhibition of aromatic amino acid transaminase in the β-phenylethanol pathway. A series of experiments showed that the substrate inhibition constant showed that the K_i value of Aro9p^HJ (0.28 μmol L⁻¹) was higher than the K_i value of Aro9p^S288C (0.26 μmol L⁻¹). This indicates that Aro9p^HJ has greater resistance to substrate inhibition during catalysis, and higher catalytic efficiency (Fig. 5A). On a similar note, Variovorax paradoxus isolated from soil exhibits an aromatic β-amino acid aminotransferase, substrate inhibition also exists [38]. The apparent K_m and k_cat values (per monomer) for (S)-β-phenylalanine in the presence of 10 mM α-ketoglutarate were 1.5 mM and 11.8 s⁻¹, respectively, with an apparent substrate inhibition constant (K_i) of 40.3 mM. When using 10 mM (S)-β-phenylalanine, the apparent K_m and k_cat for α-ketoglutarate were 0.3 mM and 10.6 s⁻¹, respectively, with an apparent K_i of 82.4 mM. The larger K_i of VpAT for (S)-β-phenylalanine confirms that VPAT has a superior catalytic capacity for β-amino acids than for α-amino acids. In short, the higher K_i value of Aro9p^HJ compared with Aro9p^S288C may indicate that Aro9p^HJ is more resistant to L-PHE in the substrate inhibition study. This could potentially be due to changes in amino acid positions in the gating access.

With 0 mmol L⁻¹-5 mmol L⁻¹ PPY as substrate, the initial reaction speed of phenylpyruvate decarboxylase Aro10p was measured, and the curve fitting was carried out with substrate concentration [S] as the X axis and speed V as the Y axis. As shown in Fig. 5B, with the increase of substrate concentration, phenylpyruvate decarboxylase Aro10p was saturated. Furthermore, the enzymatic kinetic results of Aro10p^HJ and Aro10p^S288C were analyzed, and the results were shown in Fig. 5. Finally, the Michaelis equation was obtained according to the data, and K_m, V_max, and k_cat/K_m values were obtained. As shown in Table 2, when PPY was used as the substrate, the K_m value of the characteristic constant of Aro10p^S288C (102.24 μmol L⁻¹) was larger than that of Aro10p^HJ (86.36 μmol L⁻¹), indicating that Aro10p^HJ had greater affinity with PPY than Aro10p^S288C. The V_max value of Aro10p^HJ (2106 μmol·(min∙g)⁻¹) was 6.8 % lower than that of Aro10p^S288C (2260 μmol·(min∙g)⁻¹). The value of k_cat/K_m (35.70 L (μmol∙min)⁻¹) of Aro10p^HJ was equivalent to Aro10p^S288C (35.80 L (μmol∙min)⁻¹), which indicated that the catalytic efficiency of Aro10p^HJ was equivalent to Aro10p^S288C under the same substrate PPY. However, Aro10p^HJ had a stronger affinity for PPY than Aro10p^S288C. Research indicates that during the grape juice fermentation process, for the substrate PPY, the V_max of ScAro10p was 3-fold higher than that of SkAro10p [21.7 ± 0.5 vs. 6.95 ± 0.15 nmol min⁻¹ (mg protein) ⁻¹] during wine fermentation. This mirrors our findings with Aro10p^HJ, where ScAro10p displayed robust catalytic capacity for PPY, a precursor of β-phenylethanol. The discrepancy could arise from the complexity of the synthetic must and the potential influence of other compounds on the formation of higher alcohols. Hence, Aro10p^HJ from S. cerevisiae HJ seems to showcase higher potential for the synthesis of aromatic alcohols such as β-phenylethanol from PPY [45].

Table 2.

Analysis of Aro10p enzymatic parameters.

	Aro10pS288C	Aro10pHJ
Vmax (μmol·(min∙g)−1)	2260.49 ± 7.54	2106.46 ± 5.14
Km(μmol·L−1)	102.24 ± 2.15	86.36 ± 6.18
kcat/Km (L·(μmol∙min)−1)	35.8 ± 1.23	35.7 ± 2.01

3.6. The effects of ethanol and β-phenylethanol on enzyme activity and stability

The experimental results were shown in Fig. 6A. By exogenously adding 0 mg L⁻¹ - 200 mg L⁻¹ β-phenylethanol, the enzyme catalytic activity of phenylalanine aminotransferase Aro9p decreased with the increase of β-phenylethanol concentration, but the residual relative enzyme activity was more than 60 % and the expression of Aro9p^HJ with the increase of β-phenylethanol concentration. The tolerance to β-phenylethanol was stronger than that of Aro9p^S288C. When the concentration of β-phenylethanol was 200 mg L⁻¹, the specific enzyme activity of phenylalanine aminotransferase Aro9p^HJ (18.69 U g⁻¹) was 30.52 % higher than that of Aro9p^S288C. However, the catalytic activity of phenylalanine aminotransferase Aro9p on the substrate L-PHE decreased rapidly with the increase of alcohol content in the exogenous addition of 0%vol - 40%vol ethanol (Fig. 6B). The relative enzyme activity of Aro9p^HJ remained 90.84 % in the addition of 40%vol ethanol. The relative enzyme activity of Aro10p^S288C remained 78.02 %. In conclusion, β-phenylethanol and ethanol could inhibit the catalytic activity of phenylalanine aminotransferase Aro9p, but Aro9p was more stable under ethanol and β-phenylethanol.

Fig. 6.

Effect of exogenous end products on enzyme activity of Aro9p and Aro10p

β-phenylethanol was added to the assay system of enzyme Aro9p with L-PHE as substrate(A); Ethanol was added to the assay system of enzyme Aro9p with L-PHE as substrate(B); β-phenylethanol was added to the assay system of enzyme Aro10p with PPY as substrate(C); Ethanol was added to the assay system of enzyme Aro10p with PPY as substrate(D); *P < 0.05**P < 0.01*** It means P < 0.001**** It means P < 0.0001.

The experimental results were shown in Fig. 6C. By exogenously adding 0 mg L⁻¹ - 200 mg L⁻¹ β-phenylethanol, the enzyme catalytic activity of phenylpyruvate decarboxylase Aro10p decreased rapidly with the increase of β-phenylethanol concentration. When the concentration of β-phenylethanol was 100 mg L⁻¹, the enzyme activity of phenylpyruvate decarboxylase Aro10p decreased to less than 20 %. In contrast, the tolerance ability of Aro10p^HJ of S. cerevisiae HJ to β-phenylethanol was slightly higher than that of Aro10p^S288C. As shown in Fig. 6D with the increase of alcohol content, the catalytic activity of phenylpyruvate decarboxylase Aro10p for the decarboxylation reaction of the substrate phenylpyruvate also decreased rapidly. Under the condition of the addition of 5%vol ethanol, the relative enzyme activity of Aro10p^HJ remained at 21.61 %. The relative enzyme activity of Aro10p^S288C was only 12.45 %. In conclusion, β-phenylethanol and ethanol could inhibit the catalytic activity of phenylpyruvate decarboxylase Aro10p, and the ability of Aro10p^HJ to tolerate β-phenylethanol and ethanol was slightly higher than that of Aro10p^S288C. The template ScPDC(2W93) of Aro10p from the previous homology modeling results has shown that there are two identified pathways for signal transfer during allosteric enzyme activation. The direct pathway begins at C221 and leads to the substrate binding site at ThDP via H92, E91, and W412. Structured loops 104–113 and 288–304 protect the active site from solvents, aided by the C-terminal helix. Intriguingly, the difference in amino acids at location 377 (A-V) is situated near their structural loop ScPDC(2W93) 288–304 and the enzyme regulatory site at C221 (Fig. 3F). It is plausible to hypothesize that this amino acid change restricts the effect of solvents (ethanol and β-phenethyl alcohol) on the enzyme Aro10p [5,[25], [26], [27],45].

4. Discussion and conclusion

One of the reasons for the difference in aromatic alcohol content is the use of different yeast strains. Previous studies have shown that when the aromatic amino acid L-PHE is used as the sole nitrogen source, S. cerevisiae HJ produces more β-phenylethanol than S288C, indicating that there are differences in the enzyme activity of the pathways involved in the production of aromatic compounds [46].

The Ehrlich pathway is the main pathway for the production of β-phenylethanol during Huangjiu fermentation [14]. The phenylalanine transaminases, encoded by ARO8 and ARO9, are crucial reversible enzymes that convert α-keto acids into aromatic amino acids [26]. However, the mechanism of how the catalytic direction changes unexpectedly, which plays an important role in the synthesis of aromatic alcohols, remains unclear. Previous studies have found that the promoter's strength for amino acid transaminase in S. cerevisiae HJ01 is weaker than that in the model strain S288C [6]. Aminotransferases are PLP-dependent enzymes that catalyze reversible transamination reactions between amino acids and α-ketoglutarate. Previous reports have shown that the enzyme activities of aminotransferases were affected by the product or substrate concentrations [38]. By analyzing the sequence information, we found four distinct amino acids (12(T-A), 79(N-D), 454(L-V), and 460(V-L)) in Aro9p^S288C and Aro9p^HJ. These differences might be located at the gated channel of the N-terminal domain, potentially influencing substrate entrance into the active site. This finding may partially explain why Aro9p^HJ, has a higher tolerance for substrate L-PHE and terminal metabolites (ethanol and β-phenylethanol) than Aro9p^S288C. Kinetic studies revealed that when the concentration of substrate L-PHE reaches 2 mM, the reaction rate of Aro9p^HJ (23.89 μmol·(min∙g)⁻¹) is 12.16 % higher than that of Aro9p^S288C (21.3 μmol (min∙g)⁻¹). Also, the substrate inhibition constants show that the K_i value of Aro9p^HJ (0.28 mol L⁻¹) is higher than that of Aro9p^S288C (0.26 mol L⁻¹), indicating Aro9p^HJ more tolerant to substrate inhibition during the catalytic process and having a higher catalytic efficiency (Fig. 6A). This is also the first discovery of the potential of Aro9p^HJ, an aromatic amino acid transferase of industrial S. cerevisiae HJ from Huangjiu fermentation, for step-by-step conversion of L-PHE to produce aromatic alcohols such as β-phenylethanol [35]. In summary, Aro9p^HJ has a certain degree of tolerance to the substrate L-PHE and the product (ethanol and β-phenylethanol). Substrate inhibition may be an important mechanism for regulating ARO9 enzyme activity and thus regulating the synthesis of aromatic alcohol metabolism in the Ehrlich pathway.

Phenylpyruvate decarboxylase exhibits independent α-keto acid decarboxylase activity, and in the reaction process, PPY is preferentially catalyzed by the enzyme encoded by ARO10. The promoter intensity of the Ehrlich pathway-related ARO10 gene in S. cerevisiae HJ01 is significantly higher than that in the model strain of S288C [6]. Meanwhile, purified enzymatic characterization showed that the k_cat/K_m value of Aro10p^HJ (35.7 L (μmol∙min)⁻¹) was comparable to that of Aro10p^S288C (35.8 L (μmol∙min)⁻¹) under the same substrate PPY. However, Aro10p^HJ has a stronger affinity for PPY than Aro10p^S288C. Analysis of the effects of exogenous addition of terminal metabolites on enzyme activity showed that although β-phenylethanol and ethanol could seriously inhibit the catalytic activity of phenylpyruvate decarboxylase Aro10p, the ability of Aro10p^HJ to tolerate β-phenylethanol and ethanol was higher than that of S. cerevisiae Aro10p^S288C. The reason for this, perhaps the difference in amino acids 377 (A-V) is located near their structural loop ScPDC(2W93) 288–304 and enzyme regulatory site C221(Fig. 3F) [8,14,15,[24], [25], [26],45].

Since aromatic metabolism of yeast is strictly regulated, the tolerance mechanism of key enzymes can provide an effective method for the accumulation of related aromatic compounds. In addition, the synergistic effect of ARO9 binding to ARO10 is crucial and needs further study, such as the co-regulation of ARO9 and ARO10 on the effect of phenols reported by ZHU. In addition, after Sebastián M. Tapia proposed that changes in the ARO80 gene may be meaningful for characterizing the regulatory domain of Aro80p, he intended to develop new strains that showed higher expression in ARO9 and ARO10 genes, thus increasing the yield of PE and PEA in wine. It can be seen that the key enzymes encoded by ARO9 and ARO10 genes are very important for regulating the metabolic synthesis of aromatic alcohols such as β-phenylethanol. To our knowledge, this report is the first to investigate the structural and enzymatic properties of the key enzymes Aro9p and Aro10p in the β-phenylethanol synthesis metabolic pathway of Huangjiu yeast. We also explored their tolerance mechanisms and found that these two key enzymes are less sensitive to inhibition by substrates or products, which may be one of the reasons why Huangjiu yeast produces a larger amount of β-phenylethanol. It lays a theoretical foundation for exploring the regulation of β-phenylethanol production by Huangjiu yeast. Furthermore, this study can also be used for research on the regulatory mechanism of other aromatic compounds.

CRediT authorship contribution statement

Qilin Yang: Investigation, Methodology, Data curation, Formal analysis, Writing – original draft. Shuangping Liu: Data curation, Formal analysis, Funding acquisition, Writing – review & editing. Yuzong Zhao: Resources, Project administration. Xiao Han: Methodology, Project administration. Rui Chang: Methodology, Resources. Jian Mao: Formal analysis, Funding acquisition.

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.

Abbreviations

L-PHE

l-Phenylalanine;

PPY

Phenylpyruvate

PLP

Pyridoxal 5′-phosphate

ThDp

Thiamine pyrophosphate

Aro9p^HJ

Phenylalanine aminotransferase Aro9p from Saccharomyces cerevisiae HJ

Aro9p^S288C

Phenylalanine aminotransferase Aro9p from Saccharomyces cerevisiae S288C

Aro10p^HJ

Phenylpyruvate decarboxylase Aro10p from Saccharomyces cerevisiae HJ

Aro10p^S288C

Phenylpyruvate decarboxylase Aro10p from Saccharomyces cerevisiae S288C

Appendix A. Supplementary data

The following is the Supplementary data to this article: