Deciphering the synthesis in Saccharomyces cerevisiae of the bioactives melatonin, serotonin, indoleacetic acid, hydroxytyrosol and tyrosol from glucose by using 13C labelling precursors and UHPLC-MS

Deciphering the synthesis in Saccharomyces cerevisiae of the bioactives melatonin, serotonin, indoleacetic acid, hydroxytyrosol and tyrosol from glucose by using 13 C labelling precursors and UHPLC-MS Marina Gonzalez-Ramirez a , Domenico Masuero b , Ana B. Cerezo a , Ana M. Troncoso a , Urska Vrhovsek b , M. Carmen Garcia-Parrilla a,* a Departamento de Nutrici´ on y Bromatología, Toxicología y Medicina Legal. Facultad de Farmacia, Universidad de Sevilla, C/P Garcıa Gonz´ alez No. 2, Sevilla 41012, Spain b Metabolomics Unit, Research and Innovation Centre, Fondazione Edmund Mach (FEM), via E. Mach 1, San Michele all’Adige, Italy ARTICLE INFO Keywords: Glucose Hydroxytyrosol UHPLC-MS Tyrosol Yeast Fermentation Isotopic labelling ABSTRACT Saccharomyces cerevisiae produces bioactive compounds such as melatonin (MEL), serotonin (SER), indoleacetic acid (IAA), hydroxytyrosol (HT), and tyrosol during alcoholic fermentation. Previous research shows that tryptophan and tyrosine are precursors. This study aimed to investigate whether S. cerevisiae can synthetize these bioactives from glucose. Using [U 13 C]-glucose as the fermentation substrate and UHPLC-MS for analysis, the research traced unequivocally the formation of these compounds. Results confirmed that MEL, SER, IAA, tyrosol, and HT are partly derived from glucose. Different 13 C-labelled forms of HT, tyrosol, and tyrosine were identified which allows to propose a biosynthetic pathway leading to the formation of HT through reactions between intermediates from the pentose phosphate pathway, glycolysis, additionally to the Erlich pathway. The proposed pathway includes L-DOPA and catechol which were detected in their 13 C-labelled forms. Understanding HT synthesis opens opportunities to enhance the bioactive potential of fermented beverages. 1. Introduction During alcoholic fermentation (AF), yeast transforms glucose into ethanol and simultaneously, it synthetizes many secondary metabolites, such as fusel alcohols, aldehydes, organic acids, esters, organic sulphides, and carbonyl compounds with different functions. Generally, primary metabolism is considered essential for growth, cell division, and survival whilst secondary metabolites may serve as detoxifying agents to reduce the stress, or as quorum sensing molecules among other functions (Mas et al., 2014). In addition to the roles that secondary metabolites may exert for yeast, they determine the sensory characteristics of the product having strong impact on product quality (Hazelwood et al., 2008). More recently, yeasts have been studied for their ability to synthesize metabolites with biological properties such as melatonin (MEL), (Fernandez-Cruz et al., 2019) and hydroxytyrosol (HT) (´ AlvarezFern´ andez et al., 2019). MEL, N-acetyl-5-methoxytryptamine, is a wellknown neurohormone that participates in the regulation of human circadian rhythms which is also found in the plant kingdom where it is widely distributed and, consequently, present in several foods (Szafra´ nska et al., 2017). Several biological properties have been described for this compound such as antioxidant, neuroprotective, antiinflammatory and cardiovascular protection (Reiter et al., 2000). Due to its presence in wines, Rodriguez-Naranjo et al. (2012) explored the winemaking process and confirmed its production during the alcoholic fermentation by yeast. Likewise, HT is a bioactive compound produced also during the alcoholic fermentation process (´ Alvarez-Fern´ andez et al., 2018). Although it is present in wines, its major concentration is found in olives and olive oil (Mateos et al., 2001). As for MEL, HT has shown several bioactive properties such as anti-inflammatory, neuroprotective, and antiangiogenic effects (Gallardo-Fern´ andez, Cerezo, et al., 2022; Gallardo-Fern´ andez, Valls-Fonayet, et al., 2022). Additionally, HT is a well -known antioxidant (Achmon & Fishman, 2015). It can be considered that, as secondary metabolites, they can be synthetized de novo or from primary compounds available in the must such as amino acids (Baumes, 2009). The knowledge of metabolic pathways of synthesis and the required precursors for their production, * Corresponding author. E-mail address: [email protected] (M.C. Garcia-Parrilla). Contents lists available at ScienceDirect Food Research International journal homepage: www.elsevier.com/locate/foodres https://doi.org/10.1016/j.foodres.2025.116596 Received 27 January 2025; Received in revised form 24 April 2025; Accepted 4 May 2025 Food Research International 216 (2025) 116596 Available online 21 May 2025 0963-9969/© 2025 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). can help to develop better methods and design strategies for forecasting the quantities of different impact metabolites in final wines. Indeed, deciphering the mechanism involved in the production of MEL and HT by yeast and its biological role, as well as their precursors such as SER, and IAA for MEL, or tyrosol for HT, is of great interest. In fact, SER and MEL can be synthetized from tryptophan (Fern´ andez-Cruz et al., 2017). Therefore, Mu˜ niz-Calvo et al. (2019) proposed a putative pathway of MEL synthesis that includes the decarboxylation of tryptophan into tryptamine, which could be thereafter hydroxylated to SER. In other terms, for many years the Ehrlich pathway has been the most studied and best-known pathway to produce fusel alcohols, which basically consists of a three-step pathway. Firstly, amino acids are deaminated; then, they are decarboxylated and finally reduced to their respective alcohol derivatives, as described for the synthesis of tyrosol from tyrosine (Hazelwood et al., 2008; Mas et al., 2014). Additionally, ´ Alvarez-Fern´ andez et al. (2019), highlighted the production of HT by yeast, possibly through the hydroxylation of tyrosol highlighting the interest in this pathway and the possibility to raise the concentration of the bioactive HT by modulating the Ehrlich pathway. In this account, Rebollo-Romero et al. (2020), studied how the concentration of the corresponding aromatic amino acid precursor in the initial substrate could influence the final HT concentration. For this purpose, different initial concentrations of tyrosine concentrations were added to a synthetic must (10 mg L −1 and 60 mg L −1 ). After the alcoholic fermentation, they noticed that a higher tyrosine concentration did not imply a higher HT content. In this regard, Gallardo-Fern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, Valls-Fonayet, et al. (2022), studied the bioactive compounds synthetized by yeast from aromatic amino acids by using labelled L-Tryptophan15 N 2 and L-Tyrosine-(phenyl413 C) as the only sources of tyrosine and tryptophan in the must. Nevertheless, unlabelled metabolites were found, identifying HT, MEL and SER, labelled and unlabelled, revealing that yeast could form these bioactive compounds (tyrosol, HT, MEL or SER) from other sources different from the initial amino acids’ precursors present in the must (Gallardo-Fern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, VallsFonayet, et al. (2022)). Their results showed that there are other pathways involved in the synthesis of these bioactive compounds. One step further is the study of their synthesis from glucose that it is the purpose of the present work. Nisbet et al. (2014) studied the production of volatile compounds from glucose by calculating the ratio between 13 C/ 12 C and analysing it by GC–MS. They evaluated the significance of the anabolic pathway in the synthesis of fusel alcohols, their acetate esters and fatty acids. They concluded that fusel alcohols were 75 % hexoses derived. Notably, the anabolic pathway could be a strong contributor to fusel alcohol production, throughout the novo synthesis from hexoses (Ugliano & Henschke, 2009). Furthermore, the role of the initial concentration of sugar (glucose and fructose) as a factor that influences HT concentration was previously explored by our group. Alcoholic fermentation was performed using two different must glucose concentrations (100 g L −1 and 240 g L −1 ). The study concluded that the higher the glucose concentration in must, the higher HT production by the yeasts (Gonzalez-Ramirez et al., 2024), revealing the relevance of glucose in the media for the final concentration of the bioactive HT. A significant number of fusel alcohols are believed to be formed during the alcoholic fermentation by one of two pathways connected with amino acid metabolism: (i) the Ehrlich pathway (Hazelwood et al., 2008); or (ii) the anabolic pathway in which the production of α -keto acids during amino acid biosynthesis came from hexoses, which may then be degraded to fusel alcohols as described above (Ugliano & Henschke, 2009; Hazelwood et al., 2008, Garde-Cerd´ an & Ancín-Azpilicueta, 2008). Tracers’ studies with stable isotopes can be used to analyse the relation and evaluation of precursor-product. Hence, tracer studies have been widely used in food chemistry (Schieberle et al., 2005) and have occasionally been extended to studies of the fate and origin of wine components, such as, for instance, deuterated isobutyric acid and ethyl isobutyrate spikes, used to study their formation in wine (Díaz-Maroto et al., 2005). All in all, Su et al. (2020), used 15 N-labelled NH 4 Cl, arginine, and glutamine, and 13 C-labelled leucine and valine to study the nitrogen metabolism pattern of two non-Saccharomyces species. Labelled glucose has been used to track the production of fusel alcohols in grain mash fermentations (Reazin et al., 1973). Therefore, the aim of this paper is to decipher whether several bioactive compounds (MEL, SER, IAA, HT, tyrosol and tyrosine, among others) could be synthetized by yeast from glucose by using [U – 13 C]-glucose which is pioneering for studying the synthesis of these compounds by S. cerevisiae. For this purpose, we performed alcoholic fermentation experiments with synthetic must that includes labelled [U13 C]-glucose as potential precursor of those bioactive compounds and analysed them by UHPLC-MS for an unequivocal identification. 2. Materials and method 2.1. Reagents and materials HT standard was purchased from Extrasynthese (Genay, France), tyrosol, tyrosine, MEL, SER, IAA from Sigma-Aldrich (St Louis, United States), and HPLC-MS grade methanol and formic acid were acquired from Merck (Darmstadt, Germany). Uniformly labelled [U13 C]-glucose was acquired from Cambridge Isotope Laboratories (Andover, MA, USA). 2.2. Synthetic must (SM) The SM was prepared for the experiment as instructed on Riou et al. (1997) with slight differences. The content was the same in the following ingredients: acids (malic acid 5 g L −1 , citric acid 0.5 g L −1 and tartaric acid 3 g L −1 ), minerals (KH 2 PO 4 0.75 g L −1 , K 2 SO 4 0.5 g L −1 , MgSO 4 7H 2 O 0.25 g L −1 , NaCl 0.2 g L −1 and CaCl 2 0.155 g L −1 ), 1 mL of trace elements (CoCl 2 ⋅ 6H 2 O 0.4 g L −1 , CuSO 4 ⋅ 5H 2 O 1 g L −1 , H 3 Bo 3 1 g L -1 , KI 1 g L −1 , MnSO 4 ⋅ H 2 O 4 g L -1 , (NH 4 ) 6 Mo 7 O 24 1 g L −1 and ZnSO 4 ⋅ H 2 O 4 g L −1 and 10 mL of a vitamins solution (myoinositol 2 g L −1 , calcium pantothenate 0.15 g L −1 , thiamine hydrochloride 0.025 g L −1 , nicotinic acid 0.2 g L −1 , pyridoxine 0.025 g L −1 and biotin 3 mL). The pH was adjusted to 3.31 with NaOH. The YAN (Yeast Assimilable Nitrogen) was obtained by the mix of ammonium from the amino acids and the ammonium from NH 4 Cl. The final YAN was 140 mg L −1 . From the total YAN, 84 mg L −1 corresponds to the nitrogen from the amino acids and 56 mg L −1 to the ammonium (NH 4 Cl). An amino acids stock solution was prepared following the instructions presented by Riou et al. (1997): tryptophan 13.4 g L −1 , isoleucine 2.5 g L −1 , aspartic acid 3.4 g L −1 , glutamic acid 9.2 g L −1 , arginine 28.3 g L −1 , leucine 3.7 g L −1 , threonine 5.8 g L −1 , glycine 1.4 g L −1 , glutamine 38.4 g L −1 , alanine 11.2 g L −1 , valine 3.4 g L −1 , methionine 2.4 g L −1 , phenylalanine 2.9 g L −1 , serine 6 g L −1 , histidine 2.6 g L −1 , lysine 1.3 g L −1 , cysteine 1.6 g L −1 and proline 46.1 g L −1 . The tyrosine was added independently, 10 mg L −1 , following GallardoFern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, Valls-Fonayet, et al. (2022). The sugar content was 200 g L −1 (glucose-fructose, 1:1). Two simultaneous batches of fermentation were prepared, one treatment served as a control with 100 g L −1 of glucose (no [U13 C]- glucose added), and a second batch in which from the final concentration of glucose (100 g L −1 ), the 50 % was non [U13 C]-glucose and the 50 % uniformly labelled [U13 C]-glucose (Cambridge Isotope Laboratories, Andover, MA, USA) due to the high cost of labelled compounds. Each fermentation condition was tested in flasks of 100 mL with 80 mL of SM. In order to ensure the utmost accuracy of outcomes and account for the inherent biological variability, six biological replicates were taken into account for each fermentation. The SM was sterilized with M. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 2 bottle-top vacuum filters of 0.2 μ m (Nalgene PES membrane, Thermo Fisher Scientific, Waltham, Massachusetts, USA). 2.3. Yeast strains The alcoholic fermentation was performed with S. cerevisiae Red Fruit strain (Enartis; RF) (Navarrete, La Rioja). They were selected based on our previous experiments in which Red Fruit produces the highest quantity of HT (Rebollo-Romero et al., 2020). 2.4. Inoculation of the alcoholic fermentation S.cerevisiae Red Fruit was provided as active dried yeasts and were rehydrated for 30 min at 37 ◦C and plated on yeast extract peptone dextrose (YPD) agar (2 % peptone, 2 % glucose, 1 % yeast extract and 2 % agar). Then, they were incubated at 28 ◦C in an oven for 48 h. Secondly, two flasks of 250 mL were used with 100 mL of YPD (1 % yeast extract, 2 % glucose and 2 % peptone) and shaken at 150 rpm (New Brunswick Innova® 40/40R, Hamburg, Germany), 28 ◦C, overnight, to let yeast grow before inoculation. Once the strains were activated, each fermentation flask of 100 mL with 80 mL of SM was inoculated with 10 6 cells/mL and capped with taps equipped with a capillary to release carbon dioxide. The AF were elapsed at 25 ◦C and 150 rpm (New Brunswick Innova® 40/40R, Hamburg, Germany). The two alcoholic fermentations, the one with labelled glucose (U 13 C-G-fermentation) and the one with unlabelled glucose (control), were performed in sextuplicate. The fermentation was daily monitored; the flasks were weighed before and after sampling and the density was controlled. The density was measured thanks to a density meter (Anton Paar, Graz, Austria). The course of the fermentation, expressed as rate of CO 2 versus fermentation time, and the density graphic is given in Fig. 1. The cells were counted with a Neubauer Camera and cytometer (Beckman Coulters, Pasadena, California, USA). The end of alcoholic fermentation was reached when the differences between daily weights were not significant (See Fig. 1). 2.5. Sampling The alcoholic fermentation lasted for 10 days, and the samples were collected according to the exponential and stationary phase of the growth curve of the yeast. Samples collected correspond to days 1,2,3,4,7 and 10. From day 1 to 4, 4 mL samples were collected daily, matching with the different stages of the exponential phase, on day 7 samples were collected as a representation of the start of the stationary phase and at the final day, day 10. Samples were centrifuged at 4500 rpm (Allegra X-12-R, Beckman Coulters, California, USA) at a temperature of 4 ◦C for 3 min. The supernatants were preserved at −80 ◦C until the analysis. A total of 72 samples were collected. Different samples were subjected for intracellular and extracellular methods and a total of 120 samples were analysed, 72 extracellular and 48 intracellular samples. 2.6. Intracellular extraction procedure The pellets were washed twice with distilled water and collected for the intracellular metabolite extraction. The volume collected was calculated to have 10 9 cells. The cells were subjected to a quenching and extraction procedure following the protocol by ´ Alvarez-Fern´ andez et al. (2018). Firstly, they were resuspended in 1 mL of distilled water and transferred into a centrifuge tube containing 4 mL of cold-glycerol saline solution [3:2 (vol/vol) glycerol: saline solution] to stop enzymatic activity. The treated samples were centrifuged at 36,086 g for 20 min at −20 ◦C (Sorvall LYNK 6000, Thermo Fisher Scientific, Waltham, MA USA). The supernatant was removed, and the pellet was washed again with 2.5 mL of cold washing solution [1:1 (vol/vol) glycerol/saline solution]. A new centrifuge cycle was applied. The supernatant was discarded, and the pellet was stored at −80 ◦C before extraction procedure. For the extraction procedure, 2.5 mL of cold methanol-water solution (50 % v/v) was added to the pellet. After that, the samples were subjected to two freeze-thaw cycles (frozen at −80 ◦C for 30 min and then thawed in an ice bath for 5 min). One finished, they were subjected to a sonication for 1 min in an ice bath using an ultrasonic liquid processor with 20 kHz homogeneous sound (Sonicator Sonoplus HD 2070, Bandelin Electronic GmbH & Co. KG, Berlin, Germany). A centrifuge cycle was applied at 36,086 g for 20 min at −20 ◦C. This cycle was repeated one more time with another 2.5 mL of cold methanol-water solution, subsequently stored at −80 ◦C until clean up. The aim of this extraction procedure is to break the cell walls in order to make them more Fig. 1. Monitorization of alcoholic fermentation with U13C-G labelled and non-labelled glucose. Left: Daily weight loss; Medium: Daily density. Right: Daily cell growth. Fermentations were carried out in sextuplicate (n =6). M. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 3 permeable to the organic solvent and consequently, intracellular metabolites can be released, while preserving their chemical integrity (Smart et al., 2010). 2.7. Clean-up and concentration procedure All extracellular extracts were cleaned up as reported by ´ AlvarezFern´ andez et al. (2019) some modifications. C18 SPE cartridges (1 g, Variant, Agilent, Santa Clara, California, USA) were conditioned with 2 mL of methanol and 2 mL of milli Q water. For cleaning extracellular samples, an aliquot of 2 mL of extracellular sample was loaded, followed by a washing step with 2 mL of water. The analytes were eluted with 1 mL of 70:30 methanol: water. Solvents were evaporated until dryness by using a vacuum concentrator (HyperVAC-LITE, GYOZEN, Korea) at 30 ◦C and 2000 rpm for 8 h. Each sample was reconstituted with 100 μ L of methanol. For the intracellular samples, they were cleaned up using 1 mL zirconia coated Phree™ cartridge (Phenomenex, Torrance, California, USA) to avoid the presence of phospholipids and proteins due to cell wall fragmentation (Carmical & Brown, 2016; van der Rest et al., 1995). For the procedure to load the samples, the protocol from the manufacturer was followed. Finally, the samples were dried as extracellular ones and reconstituted in 100 μ L methanol. 2.8. UHPLC-MSMS instrumental analysis The analysis was performed in an UHPLC ExionLC system coupled with AB6500+Qtrap mass spectrometer and an electrospray ionization system (ESI) (AB Sciex LLC,Framingham, MA, U.S.A.). Waters Acquity column (Milford, Massachusetts, USA), HSS T3 (2.1 ×150 mm, 1.8 μ m), Fig. 2. Labelled Hydroxytyrosol fragments. M. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 4 was used. The column was kept at 40 ◦C. The flow rate was set at 0.28 mL/min. The UHPLC method was adapted from ´ Alvarez-Fern´ andez et al. (2019). Mobile phase A was water with 0.1 % formic acid; mobile phase B was methanol with 0.1 % formic acid. The gradient was programmed as follows: 5 % B (0 min); 20 % B (2 min); 20 % B (3 min); 28 % B (3.3 min); 42 % B (8 min); 100 % B (8–12 min); 5 % B (14.5 min). The injection volume was 2 μ L. The MS parameters were optimized individually for each standard by the direct infusion of their solutions (10 μ g mL −1 ). The most abundant fragments were identified for each compound to be used as the quantifier, the other fragments as qualifiers. Fragmentations for labelled compounds were calculated theoretically based on the monoisotopic mass of the labelled compound and the possible fragmentations achievable based on the position of the labelled element. The curtain gas was set at 35 psi. The Ion Spray Voltage was set at −4500 V in the negative mode and 5000 V in the positive mode. The source temperature was set at 500 ◦C, the nebulizer gas (Gas1) and heater gas (Gas2) at 50 and 60 psi, respectively. Declustering potential (DP) and entrance potential (EP) were optimized for each precursor ion and collision energy (CE) and Collision Cell Exit Potential (CXP) for each product ion. The retention time of the labelled compounds was the same as the unlabelled standard compound (Watkins et al., 2019). For quantification purposes the reference standard of the unlabelled compound was used (Gallardo-Fern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, VallsFonayet, et al. (2022)). The quantification of each compound was performed using external calibration curves prepared in methanol. 2.9. Statistical analysis Data were subjected to analysis of variance (ANOVA). Results are expressed as mean ±standard deviation (SD). Differences at p <0.05 were considered statistically significant. Statistica version 14.00 was used for data analysis. 3. Results and discussion 3.1. Alcoholic fermentation monitoring It was necessary to check that yeast could metabolize [U13 C]- Glucose and that the fermentations were similar to those performed with unlabelled glucose. To this end, both alcoholic fermentations (control and [U13 C}-Glucose) were performed in sextuplicate, and the monitoring was based on the daily weight of the flasks, density of the must and cell growth. The flasks were weighted daily before and after sampling, showing a decrease in weight in terms of CO 2 expelled due to the metabolic reaction, as displayed in Fig. 1. During the alcoholic fermentation, yeast converts glucose into ethanol and carbon dioxide according to the following chemical equation: C₆H₁₂O₆ → 2C₂H₅OH +2 CO₂. This reaction demonstrates that one molecule of glucose yields two molecules of carbon dioxide. Considering the molar masses of glucose (180 g mol −1 ) and carbon dioxide (44 g mol −1 ), it can be calculated that the fermentation of one gram of glucose results in the production of approximately 0.489 g of CO₂. In our case we have initially 200 g L −1 , and observed an avarage total weight loss of approximately 42 g L −1 , which could correspond to the loss of CO₂. The density of the medium was also measured every day of sampling as an indicator of the glucose consumed (Fig. 1). Furthermore, cell growth was determined by flow cytometry, increasing at day 1 and remaining constant during alcoholic fermentation (Fig. 1). As the figure shows, both fermentations took place in a similar way, lasting 10 days and according to ANOVA not statistical differences were found (p <0.05) for the same day along the alcoholic fermentation between the control and U13 C-G fermentations for weight lost, density and cell growth. 3.2. MRM and MS optimization Secondly, it was necessary to optimize the analytical method for the determination of the bioactive compounds. Samples taken at different days of alcoholic fermentations were analysed for 6 bioactive compounds (MEL, SER and 3-IAA, and HT, tyrosine, and tyrosol). These compounds were determined by UHPLC/MS-MS and undoubtably identified thanks to the comparison of retention times, exact mass and the spectra with the respective commercial standards. For the analysis of the mass, the MRM method was selected. The different parameters for MS conditions were optimized for each compound in positive and negative mode, which are summarized in Table S.1 and S.2. Declustering potential (DP) and entrance potential (EP) were optimized for each precursor ion and collision energy (CE) and Collision Cell Exit Potential (CXP) for each product ion. The optimization was performed through the infusion of the standards of each compound. For the labelled compounds, a previous analysis of the molecule fragmentation was necessary. The retention time and the different voltages remained the same. The MRM method for the labelled compounds was structured based on the fragmentation pattern of the unlabelled compounds. As the medium contains both unlabelled and [U13 C]-Glucose, a deeper study of the different possible fragments was required since multiple number and positions of labelled carbons can be expected. All the labelled compound possibilities range from 1 labelled carbon ( 13 C 1 ) to the total number of carbons of the compound. For example, 13 C 1 to 13 C 8 were explored for HT. At the same time labelled carbons may be present in different positions in the molecule. In the case of HT, they can be distributed between the ring and the lateral chain (Fig. 2). HT molecular mass is 154.16 Da, therefore, 153.16 Da corresponds to its molecular ion in negative mode. According to the fragmentation, the most abundant transition is m/z at 153 to 123, which supposed a loss of 30 amu, which corresponds with the fragment CH 2 O − , part of the lateral chain. Table S.1 shows all the transitions found in labelled and unlabelled HT. Fig. 2 shows the different HT transitions depending on the number and positions of the labelled carbons. If [ 13 C 4 ]-HT is considered, the transitions can be as follows: m/z 157–127, (if the 13 C 4 labelled carbons are in the ring) or m/z 157–126 (if one of the labelled carbons is in the chain) (Fig. 2). Tyrosol, the most abundant fragment corresponded to m/z 137–92, which explains the loss of the lateral chain (45 amu). The number and positions of the possible labelled carbons are shown in Fig. S.1. The most abundant fragment for tyrosine was m/z 182–165; the loss occurring in the amino group (NH 2 ) (Table S.1). The transition for the labelled tyrosine corresponds to m/z 188–171 since all the carbons remain in the fragments (Table S.1 and Fig. S.2). A similar pattern is found for SER (Table S.2 and Fig. S.3). MEL and IAA most abundant fragments were m/ z 233–174 and m/z 176–130, respectively. Fig. S.4 and S.5 show the different labelled fragments for MEL and IAA (Table S.2). 3.3. Analysis of labelled compounds in fermented samples and possible pathways As can be seen in Table S.3, the six compounds under study (HT, tyrosol, tyrosine, MEL, SER and 3-IAA) were detected in their 13 Clabelled forms in U 13 C-G fermented samples. Indeed, it can be undoubtedly concluded that all these compounds are in part glucosederived, indicating the importance of anabolic pathways for their formation as the novel contribution of this work. Table S.3 presents the different concentrations of each compound (labelled and unlabelled) at each point of the fermentation process. Results are expressed as means of the six biological replicates. There were no significant differences between the concentrations of HT, MEL, SER, 3-IAA, tyrosine and tyrosol in control samples and samples from 13 C-G fermentations (sum of unlabelled and labelled forms). Moreover, the total amount of HT produced agrees with the results previously reported for the same strain and fermentation conditions (GallardoM. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 5 Fern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, Valls-Fonayet, et al. (2022)). Indeed, it can be concluded that 13 C labelled glucose has not affected HT production by the yeast; consequently, labelling does not disturb yeast metabolism. Although compounds under study ranged from 13 C 8 to 13 C 13, as MEL has 13 carbons in total, only compounds with 13 C 1 to 13 C 6 labelled carbons were detected. It is worth highlighting that certain labelled compounds with 13 C 1 carbon were also detected in the control fermentations, since native isotopes such as 13 C are naturally present (Spitzke & Fauhl-Hassek, 2010). In fact, there were not significant differences in the concentration of 13 C 1 labelled carbon compounds (HT, tyrosol, tyrosine, MEL, IAA and SER) between samples from the 13 C-G and the control fermentations. Therefore, we discarded these data for further discussion and focused on the differences between the two fermentations. Unlabelled MEL, SER and IAA were detected and quantified in some of the samples, it can be concluded that these bioactive compounds can be synthesised from glucose in addition to other biosynthesis pathways. Su´ astegui and Shao (2016) concluded that during glycolysis, glucose is Fig. 3. Different pathway from glucose to bioactive compounds. PEP, phosphoenolpyruvate; E4P, erythrose-4-phosphate; DAHP, 3-deoxy-D-arabinoheptulosonate-7phosphate; DHQ, dehydroquinate; DHS, dehydroshikimate; SA, shikimic acid; S3P, shikimitate-3-phosphate; EPSP, e-enolpyruvyl-3-shikimate phosphate; CA, chorismic acid; AA, anthranilic acid. Fig. 4. Abundance of the different labelled tyrosine possibilities excluding 113 C. M. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 6 transformed into phosphoenolpyruvate (PEP) and also glucose through the pentose phosphate pathway can yield erythrose-4-phosphate (E4P) (Fig. 3). These two compounds bond to finally form anthranilic acid, which will later add a phosphorylated ribose to produce tryptophan (Su´ astegui & Shao, 2016). Tryptophan can yield SER and IAA, SER can finally produce MEL (´ Alvarez-Fern´ andez et al., 2019). Unlabelled tyrosine was detected and quantified in the intracellular and extracellular media. In fact, it is present as such in the starting synthetic must. However, labelled tyrosine was detected and quantified in the intracellular media but not in the extracellular one. Indeed, yeast can synthesize de novo aromatic amino acids, such as tyrosine (Su´ astegui & Shao, 2016). Our data show that this newly synthetised tyrosine from glucose is not excreted but, most likely, used as precursor of other compounds. It is well known (Su´ astegui & Shao, 2016) that during glycolysis, glucose is transformed into PEP, and it is bonded to E4P (Fig. 3). These two compounds bond to finally form prephenate, which will later produce tyrosine (Su´ astegui & Shao, 2016). Furthermore, tyrosine is metabolized to tyrosol by the Ehrlich pathway which might be subsequently hydroxylated to HT. Hence, de novo synthetised tyrosine was found with different labelled carbons as shown Fig. 4 being 13 C 3 tyrosine the most abundant followed by 13 C 4 tyrosine as shown in Fig. 4. If 13 C E4P takes part, it will result in tyrosine labelled in four carbons. However, if 13 C PEP is involved, then tyrosine with three labelled carbons will be formed. In the case tyrosine is produced from two 13 C PEP which afterwards loses 1 13 C carbon, it will result in tyrosine with five 13 C (Fig. 5). Labelled tyrosol is detected from 13.11 to 221.55 ng ml −1 as shown in Table S.3, however most of tyrosol determined (from 1 to 16.47 mg L −1 ) is unlabelled consistent with a majority production from the initial tyrosine from the medium. Indeed, Gallardo-Fern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, Valls-Fonayet, et al. (2022) in an experiment with labelled tyrosine observed that the major proportion of tyrosol (75 %) came from labelled tyrosine. Our results reinforce the fact that tyrosol is mostly formed from the initial tyrosine concentration in the synthetic must. Moreover, in agreement with tyrosine results, the most abundant tyrosol form was the one with three 13 C followed by four 13 C. HT was detected and quantified in control fermentations in both intracellular and extracellular media. Labelled HT was detected and quantified in all its different labelled possibilities both in extracellular and intracellular media. Gallardo-Fern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, Valls-Fonayet, et al. (2022) demonstrated that HT can be formed from tyrosine, but it was in a low extension. Consequently, a S. cerevisiae strain was genetically modified to enhance the enzyme which hydroxylates tyrosol to HT, obtaining a moderate increase of HT concentration (Gonzalez-Ramirez et al., 2024). In the Fig. 5. Different possibilities for the formation of tyrosine from glucose. Fig. 6. Abundance of the different labelled HT possibilities excluding 113 C. M. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 7 present work, we compared the ratio of tyrosol to HT obtained in control fermentation with the ratio for the labelled compounds in the 13 C-G fermentation resulting that this ratio is 20 times higher for the labelled compounds. This fact reinforces that an additional pathway to produce HT must exist, apart from the putative hydroxylation of tyrosol. Taking into account that the most abundant HT forms are the ones with two and four 13 C (32 and 30 % respectively) (Fig. 6), we hypothetised a possible pathway as shown in Fig. 3 by putting together pieces of information in literature and the matching with our data. It is known that PEP and E4P are formed through glycolysis and the pentose phosphate pathways, respectively. Thus, from E4P and PEP, the shikimic pathway gives rise to dehydroshikimate (DHS), which by the muconic acid pathway yields catechol by losing 1 carbon atom (Su´ astegui & Shao, 2016). Additionally, Min et al., (2014) described that catechol, together with ammonium, that is present in the medium, and pyruvic acid, formed by glycolysis, produces L-DOPA ((L)-3,4-dihydroxyphenylalanine). Furthermore, it was proven that S. cerevisiae can metabolize L-DOPA to finally produce HT (Behringer et al., 2024) (Fig. 3). These authors showed that nonproteinogenic aromatic amino acids such as L-DOPA, can be substrates for degradation via the Ehrlich pathway. 3-(3,4-dihydroxyphenyl)-pyruvic acid is transaminated to give L-DOPA and then converted to the final HT by losing another additional carbon atom. The carbon atoms lost come from PEP and pyruvic acid. To confirm this possibility, an analysis proved that 13 C catechol and 13 C LDOPA are present. A Parallel Reaction Monitoring (PRM) analysis was performed and 13 C 5 L-DOPA and 13 C 2 catechol are detected, supporting the proposed pathway which is depicted in Fig. 3. It is worth to remember that both labelled and not labelled glucose were present in the medium, therefore labelled and not labelled precursors (E4P, PEP and pyruvic acid) are produced, thus, explaining the different possibilities of combination which result in the different labelled HT forms found (Figs. 2 and 7). For instance, the combination of 13 C E4P and 13 C PEP results in HT with six 13 C in the ring. If 13 C E4P, non labelled PEP and 13 C pyruvic acid are combined, HT with four 13 C in the ring and two 13 C in the lateral chain can be formed. Finally, if just 13 C PEP or 13 C pyruvic acid are present, it can result in HT with two 13 C in the ring or lateral chain respectively. This pathway explains the two, four and six 13 C of carbons of HT that are in the higher proportion as Fig. 6 shows. Furthermore, the presence of HT with three 13 C can be explained from the three carbons 13 C tyrosine synthetized by yeast from glucose. Our data show that the sum of HT labelled in two, four or six carbons explains most of the HT produced revealing the proposed pathway should be the main favored with regard to the pathway from tyrosine. The pentose phosphate pathway is upregulated under stressful conditions (Bertels et al., 2021), and yeast may prioritize the production of compounds needed to counteract stress (Postaru et al., 2023). This could be the case for HT synthesis, a well-known antioxidant, produced from E4P through this pathway. Fig. 7. Different possibilities for the formation of HT from glucose. M. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 8 Regarding the results presented by Gallardo-Fern´ andez, Cerezo, et al. (2022); Gallardo-Fern´ andez, Valls-Fonayet, et al. (2022), HT, MEL, and SER were found in higher concentrations in their unlabelled forms; indeed, it was concluded that there was a pathway that had more impact than the one from amino acids. However, the labelled tyrosol from the labelled tyrosine was found in a greater concentration. These results agree with our findings, as we quantify tyrosol at low concentrations and it is mostly found in unlabelled form coming from initial tyrosine. 4. Conclusions All in all, this study demonstrates that S. cerevisiae can synthesize bioactive compounds such as HT, MEL, SER, IAA and tyrosol from glucose during alcoholic fermentation. Although tryptophan and tyrosine are known precursors, the results show that glucose is a significant source for the synthesis of these compounds, especially HT. Using a UHPLC-MS method and [U 13 C]-glucose, the synthesis pathways were traced and confirmed. The study proposes a mechanism where HT is formed from glycolysis (PEP) and pentose phosphate (E4P) pathway intermediates, through catechol and L-DOPA intermediates. Additionally, tyrosol is mainly formed from tyrosine through the shikimic and Ehlirch pathways. This finding opens new opportunities to enhance the bioactive potential of fermented beverages by exploiting these synthetic pathways. CRediT authorship contribution statement Marina Gonzalez-Ramirez: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Domenico Masuero: Writing – review & editing, Writing – original draft, Visualization, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ana B. Cerezo: Writing – review & editing, Writing – original draft, Visualization, Supervision, Investigation, Formal analysis, Data curation, Conceptualization. Ana M. Troncoso: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Investigation, Formal analysis, Conceptualization. Urska Vrhovsek: Writing – review & editing, Writing – original draft, Visualization, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. M. Carmen Garcia-Parrilla: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Investigation, Formal analysis, Conceptualization. Funding This research was funded by grant PID2022–137807OB-C22 funded by MICIU/AEI/ 10.13039/501100011033 and by grant PID2019108722RB-C32 funded by MICIU/AEI/ 10.13039/501100011033, as appropriate, by “ERDF A way of making Europe” by the “European Union” and Junta de Andalucía Grant P18-RT-3098. 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. Acknowledgements The author thanks the Spanish Government for its financial assistance (PID2022-137807OB-C22) Junta de Analucía (P18-RT-3098); and the University of Seville for the M.G.R. predoctoral contract PIF and for the fellowship to stay abroad. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.foodres.2025.116596. Data availability Data will be made available on request. References Achmon, Y., & Fishman, A. (2015). The antioxidant hydroxytyrosol: Biotechnological production challenges and opportunities. Applied Microbiology and Biotechnology, 99, 1119–1130. https://doi.org/10.1007/s00253-014-6310-6 ´ Alvarez-Fern´ andez, M. A., Fern´ andez-Cruz, E., Cantos-Villar, E., Troncoso, A. M., & García-Parrilla, M. C. (2018). Determination of hydroxytyrosol produced by winemaking yeasts during alcoholic fermentation using a validated UHPLC–HRMS method. Food Chemistry, 242, 345–351. https://doi.org/10.1016/j. foodchem.2017.09.072 ´ Alvarez-Fern´ andez, M. A., Fern´ andez-Cruz, E., Valero, E., Troncoso, A. M., & GarcíaParrilla, M. C. (2019). Efficiency of three intracellular extraction methods in the determination of metabolites related to tryptophan and tyrosine in winemaking yeast’s metabolism by LC-HRMS. Food Chemistry, 297, Article 124924. https://doi. org/10.1016/j.foodchem.2019.05.198 Baumes, R. (2009). Wine aroma precursors. In Wine chemistry and biochemistry (pp. 251–274). New York, NY: Springer New York. Behringer, K. I., Kapeluch, J., Fischer, A., & Hellwig, M. (2024). Metabolization of free oxidized aromatic amino acids by Saccharomyces cerevisiae. Journal of Agricultural and Food Chemistry, 72(11), 5766–5776. https://doi.org/10.1021/acs.jafc.3c09007 Bertels, L. K., Fern´ andez Murillo, L., & Heinisch, J. J. (2021). The pentose phosphate pathway in yeasts-more than a poor cousin of glycolysis. Biomolecules, 11(5), 725. https://doi.org/10.3390/biom11050725 Carmical, J., & Brown, S. (2016). The impact of phospholipids and phospholipid removal on bioanalytical method performance. Biomedical Chromatography, 30(5), 710–720. https://doi.org/10.1002/bmc.3686 Díaz-Maroto, M. C., Schneider, R., & Baumes, R. (2005). Formation pathways of ethyl esters of branched short-chain fatty acids during wine aging. Journal of Agricultural and Food Chemistry, 53(9), 3503–3509. Fern´ andez-Cruz, E., ´ Alvarez-Fern´ andez, M. A., Valero, E., Troncoso, A. M., & GarcíaParrilla, M. C. (2017). Melatonin and derived L-tryptophan metabolites produced during alcoholic fermentation by different wine yeast strains. Food Chemistry, 217, 431–437. https://doi.org/10.1016/j.foodchem.2016.08.020 Fernandez-Cruz, E., Gonz´ alez, B., Mu˜ niz-Calvo, S., Morcillo-Parra, M.´ A., Bisquert, R., Troncoso, A. M., … Guillam´ on, J. M. (2019). Intracellular biosynthesis of melatonin and other indolic compounds in Saccharomyces and non-Saccharomyces wine yeasts. European Food Research and Technology, 245, 1553–1560. https://doi.org/ 10.1007/s00217-019-03257-5 Gallardo-Fern´ andez, M., Cerezo, A. B., Hornedo-Ortega, R., Troncoso, A. M., & GarciaParrilla, M. C. (2022). Anti-VEGF effect of bioactive indolic compounds and hydroxytyrosol metabolites. Foods, 11(4), 526. https://doi.org/10.3390/ foods11040526 Gallardo-Fern´ andez, M., Valls-Fonayet, J., Valero, E., Hornedo-Ortega, R., Richard, T., Troncoso, A. M., & Garcia-Parrilla, M. D. C. (2022). Isotopic labelling-based analysis elucidates biosynthesis pathways in Saccharomyces cerevisiae for melatonin,serotonin and hydroxytyrosol formation. Food Chemistry, 374, Article 131742. https://doi.org/ 10.1016/j.foodchem.2021.131742 Garde-Cerd´ an, T., & Ancín-Azpilicueta, C. (2008). Effect of the addition of different quantities of amino acids to nitrogen-deficient must on the formation of esters, alcohols, and acids during wine alcoholic fermentation. LWTFood Science and Technology, 41(3), 501–510. https://doi.org/10.1016/J.LWT.2007.03.018 Gonzalez-Ramirez, M., Gallardo-Fernandez, M., Cerezo, A. B., Bisquert, R., Valero, E., Troncoso, A. M., & Garcia-Parrilla, M. C. (2024). The production of bioactive Hydroxytyrosol in fermented beverages: The role of must composition and a genetically modified yeast strain. Fermentation, 10(4), 198. https://doi.org/10.3390/ fermentation10040198 Hazelwood, L. A., Daran, J. M., van Maris, A. J., Pronk, J. T., & Dickinson, J. R. (2008). The Ehrlich pathway for fusel alcohol production: A century of research on Saccharomyces cerevisiae metabolism. Applied and Environmental Microbiology, 74(8), 2259–2266. https://doi.org/10.1128/AEM.02625-07 Mas, A., Guillamon, J. M., Torija, M. J., Beltran, G., Cerezo, A. B., Troncoso, A. M., & Garcia-Parrilla, M. C. (2014). Bioactive compounds derived from the yeast metabolism of aromatic amino acids during alcoholic fermentation. BioMed Research International, 2014, Article 898045. https://doi.org/10.1155/2014/898045 Mateos, R., Espartero, J. L., Trujillo, M., Ríos, J. J., Le´ on-Camacho, M., Alcudia, F., & Cert, A. (2001). Determination of phenols, flavones, and lignans in virgin olive oils by solid-phase extraction and high-performance liquid chromatography with diode array ultraviolet detection. Journal of Agricultural and Food Chemistry, 49(5), 2185–2192. https://doi.org/10.1021/jf0013205 Mu˜ niz-Calvo, S., Bisquert, R., Fern´ andez-Cruz, E., García-Parrilla, M. C., & Guillam´ on, J. M. (2019). Deciphering the melatonin metabolism in Saccharomyces cerevisiae by the bioconversion of related metabolites. Journal of Pineal Research, 66 (3), Article e12554. https://doi.org/10.1111/jpi.12554 M. Gonzalez-Ramirez et al. Food Research International 216 (2025) 116596 9