Potential production of 2-phenylethanol and 2-phenylethylacetate by non-Saccharomyces yeasts from Agave durangensis

Abstract

Introduction

The participation of non-Saccharomyces yeasts in fermentation processes is of great importance due to their participation in the formation of esters and superior alcohols, which confer characteristic aromas to beverages such as wine and mescal.

The aim

The aim of this study was identify and evaluate the potential aroma production of yeast native of Agave fermentation by the mescal production in Durango, Mexico. Isolated yeasts were molecularly identified by 5.8s ribosomal gene; the potential production of aromas was carried out in fermentations with the addition of l-phenylalanine and evaluated after 24 h of fermentation. Extraction and quantification of aromatic compounds by headspace solid-phase micro-extraction (HS-SPME) and gas chromatograph mass spectrometry (GC-MS).

Results

The isolated non-Saccharomyces yeasts could be classified into six different genera Saccharomyces cerevisiae, Clavispora lusitaniae, Torulaspora delbrueckii, Kluyveromyces dobzhanskii, Kluyveromyces marxianus, and Kluyveromyces sp. All probed strains presented a potential aroma production (ethyl acetate, isoamyl acetate, isoamyl alcohol, benzaldehyde, 2-phenylethyl butyrate, and phenylethyl propionate), particularly 2-phenylethanol and 2-phenylethylacetate; the levels found in the Kluyveromyces marxianus ITD0211 yeast have the highest 2-phenylethylacetate production at 203 mg/L and Kluyveromyces marxianus ITD0090 with a production of 2-phenylethanol at 1024 mg/L.

Conclusion

Non-Saccharomyces yeasts were isolated from the mescal fermentation in Durango; the Kluyveromyces genus is the most predominant. For the production of aromas, highlighting two strains of Kluyveromyces marxianus produces competitive quantities of compounds of great biotechnological interest such as 2-phenylethanol and 2-phenylethylacetate, without resorting to the genetic modification of yeasts or the optimization of the culture medium.

Introduction

The non-Saccharomyces yeasts are well recognized for their contribution to the aroma of fermentative beverages (Cordente et al. 2012; Ciani et al. 2016; Masneuf-Pomarede et al. 2016), especially wine. Their presence has also been reported in mescal and tequila. In Mexico, these alcoholic beverages are distinguished from each other, based on the agave species used in their production. For example, Agave tequilana Weber var. Azul (blue variety) is used for tequila, whereas Agave salmiana and Agave durangensis, (or Agave duranguensis) among others, are used for mescal production in various regions of Mexico (Lappe-Oliveras et al. 2008; Páez-Lerma et al. 2010; De los Rios-Deras et al. 2015; Kirchmayr et al. 2017). Mescal has elevated its economic importance in the last years (Kirchmayr et al. 2017). During the mescal production process, agave juice is naturally fermented by native yeasts, such as Saccharomyces, Pichia, Kluyveromyces, Candida, Debaryomyces, Hanseniaspora, Kloeckera, Schizosaccharomyces, Torulaspora, and Zygosaccharomyces (Lachance 1995; Díaz-Montaño et al. 2008; Escalante-Minakata et al. 2008). Previously published research on fermentations of agaves suggests that non-Saccharomyces yeasts have an important role in the initial fermentation process and influence the production of the volatile compounds (Lappe-Oliveras et al. 2008; Narváez-Zapata et al. 2010; Martell Nevárez et al. 2011). The potential use of these yeasts as inoculants has been described (Rodríguez-Sifuentes et al. 2014; Nuñez-Guerrero et al. 2016), as well as their participation in generating the volatile compounds in mescal, mainly esters (Martell Nevárez et al. 2011; Rutiaga-Quiñones et al. 2012; Hernández-Carbajal et al. 2013). Despite the increasing use of non-Saccharomyces yeasts in biotechnology, there are still many opportunities to improve native yeast exploration. These prospects have led to a great interest in further enhancing the number of non-Saccharomyces yeasts available, by selecting or developing strains with novel and attractive properties.

Flavor has a major impact on the quality perception of food and beverages, and fragrances are highly valued in the cosmetic and perfume industry. For natural aroma compounds that exist at low concentrations in their original sources, biotechnological processes represent an attractive alternative to the traditional preparation by extraction (Schrader et al. 2004). Due mainly to its sweet and rose-like taste and odor 2-phenylethanol (2-PE) and its more fruit-like form, acetate ester 2-phenylethylacetate (2-PEA), find use in various flavor compositions (Fabre et al. 1998). For food applications, the rising demand for natural products means natural flavor compounds are increasingly becoming a necessity (Etschmann and Schrader 2006, Morrissey et al, 2015).

Both 2-PE and 2-PEA can be produced by de novo synthesis or from l-phenylalanine (l-Phe) by non-Saccharomyces yeast whole-cell biocatalysis via the Ehrlich pathway (Etschmann et al. 2003), (Etschmann and Schrader 2006). 2-PE can also be metabolized to 2-PEA by a trans-esterification reaction, which involves the transfer of a group of acetyl-coenzyme A acetate to the hydroxyl group of 2-PE (Hazelwood et al. 2008; Pires et al. 2014). When l-Phe is the sole nitrogen source in the medium, large amounts of 2-PE are accumulated. Several biotechnological processes are known for producing 2-PE, based on this pathway, and considerable progress has been made on the development of this process. In this context, yeast biodiversity may be greatly impacted by the production of different aroma products derived from primary and secondary metabolism. The diversity of non-Saccharomyces yeasts responsible for many of the volatile compounds found in mescal, in the state of Durango, Mexico, has not yet been evaluated. This research aimed to identify the non-Saccharomyces microbiota present in fermentations in three different mescal-producing regions and assess the production potential of aromatic compounds the addition of l-Phe as an inductor.

Materials and methods

Yeast strains

Thirty-four native strains, identified as non-Saccharomyces from Agave durangensis fermentation and obtained from the Collection of the Instituto Tecnologico de Durango, were isolated from three mescal-producing regions of Durango State, Mexico: Mezquital (23° 28′ 22″ N, 104° 24′ 40″ W), Nombre de Dios (23° 51′ 00″ N, 104° 14′ 00″ W), and Durango (24° 01′ N, 104° 40′ W). All yeast strains were conserved, as culture stock at − 20 °C in 30% (v/v) glycerol.

Molecular identification

Growth conditions

Yeast cells preserved in glycerol were first activated on YDP solid medium (glucose 20 g/L, casein peptone 20 g/L, yeast extract 10 g/L, and agar 20 g/L). DNA was then extracted at 24-h growth, using the method detailed by Sambrook and Russell (2001).

Polymerase chain reaction and amplification

Polymerase chain reaction (PCR) was carried out in 50-μL volumes, using 2.0 μL of DNA with ITS1 (5′–TCC GTA GGT GAA CCT GCG G–3′) and ITS4 (5′–TCC TCC GCT TAT TGA TAT GC–3′) primers to amplify the rDNA repeat unit that includes the 5.8S rRNA gene and the two non-coding regions designated the internal transcribed spacers (ITS1 and ITS4) (White et al. 1990). Amplification began with an initialization step of one cycle at 95 °C for 5 min, then 35 cycles of 95 °C for 1 min, 52 °C for 2 min, and 72 °C for 2 min, followed by a final elongation at 72 °C for 10 min (White et al. 1990). The PCR product was electrophoresed on 1% agarose gel with TAE 0.5× buffer (Promega, Madison, WI, USA), at 95 V for 45 min, stained with ethidium bromide (Sigma–Aldrich. St. Louis, MO, USA) and visualized under UV light (Benchtop UV transilluminator, Upland, CA, USA); DNA fragment sizes were determined using a 100-bp DNA ladder (Promega, USA). The PCR product was purified using C₂H₇NO₂ and C₂H₆O (> 99%) (Sigma–Aldrich). The rDNA sequences were acquired using an ABI PRISM Model 3730XL sequencer (Applied Biosystems, Inc., Foster City, CA, USA) at the National Genomics for Biodiversity Laboratory (Langebio) of Cinvestav (Irapuato, Mexico).

Phylogenetic analysis

The obtained sequences were aligned using the MUSCLE program (https://www.ebi.ac.uk/Tools/msa/muscle), and regions of local similarity between sequences were identified from the National Center for Biotechnology Information (NCBI) database of GenBank using the BLAST program (https://blast.ncbi.nlm.nih.gov/Blast). Phylogenetic analyses were conducted in MEGA7 Program. The sequences were deposited in GenBank.

Production of volatile organic compounds

Chemicals and reagents

l-Phe (< 98%), 2-PE (> 99%), and 2-PEA (> 99%) were purchased from Sigma–Aldrich. Na₂HPO₄.2H₂O, MgSO₄.7H₂O (Caisson Laboratory In., Smithfield, UT, USA), and citric acid were obtained from Fermont (Mexico City, Mexico). Glucose, yeast extract and casein peptone came from BD Bioxon (Mexico City, Mexico).

Bioconversion

The strains were pre-grown in 125-mL baffled Erlenmeyer flasks (Corning, Inc., USA) with vented top, containing a 50-mL operative volume of standard yeast medium YPD broth (20 g/L glucose, 20 g/L casein peptone, and 10 g/L yeast extract), at 30 °C for 12 h and 120 rpm. For fermentation, the strains were inoculated at a concentration of 10⁷ cells/mL and incubated at 30 °C for 24 h and 120 rpm. Duplicate experiments were done for induction with l-Phe (9 g/L), in which the culture medium contained 30 g/L glucose, 35 g/L Na₂HPO₄.2H₂O, 10.5 g/L citric acid, 0.5 g/L MgSO₄.7H₂O, and 0.17 g/L yeast extract, in a 50-mL medium, in a 125-mL Erlenmeyer flask (Etschmann et al. 2004). The yeast Kluyveromyces marxianus CBS 600 (KY103809.1) was included as a reference.

Gas chromatography–mass spectrometry analysis

The volatile organic compounds were extracted by headspace solid-phase micro-extraction (HS-SPME) with a divinylbenzene/carboxen/polydimethylsiloxane fiber (Supelco, Bellefonte, PA, USA). One milliliter of the sample was taken from each fermentation at 24 h, placed inside a 4-mL vial, sealed tightly with a screw-top septum-containing cap, and allowed to stand at 35 °C for 1 h. The SPME needle was then inserted through the septum, the holder was secured, and the fiber was exposed to the headspace. After 1 h of sampling at 35 °C, the fiber was retracted and immediately inserted into the inlet of a HP 5890 Series II GC instrument directly coupled to an HP 5972 mass-selective detector (Hewlett–Packard, Palo Alto, CA, USA) and equipped with an HP-FFAP capillary column (25 m × 0.320 mm i.d., film thickness 0.50 m; Hewlett–Packard), for thermal desorption. The injection was accomplished by desorption of the fiber at 230 °C for 6 min with the injector operated in the splitless mode for 1 min. An additional 5-min exposure in the injection port allowed the fiber to be cleaned of any compound that may not have been desorbed during the initial minute (Calvo-Gómez et al. 2004). Helium was used as the carrier gas, at a linear flow of 2 mL/min. The injector and detector temperatures were 230 and 260 °C, respectively. The oven temperature was increased from 40 to 240 °C, using the following program: the initial temperature was maintained for 3 min, ramped at 4 °C/min to 100 °C, held for 1 min, and then ramped at 4 °C/min to 240 °C and held for 10 min. The ionization voltage was 70 eV. All the assays were performed twice. The analyzed compounds were identified by comparing their mass spectra with those in the NIST database (Calvo-Gómez et al. 2004). In addition, the volatile compounds of interest (2-PE and 2-PEA) were quantified by standard curves.

Statistical analysis

Data of the volatile compounds, 2-PE and 2-PEA, were evaluated by the HSD–Tukey–Kramer comparison test, at α = 0.01 All statistical analyses were done using JMP software version 13.2 (SAS Institute, Inc., NC, USA).

Result and discussion

Molecular identification and phylogenetic analyses

Table 1 indicates the molecular identification of the studied strains, for each geographic region. These strains corresponded to six different genera: Clavispora lusitaniae, Kluyveromyces sp., Kluyveromyces dobzhanskii, Kluyveromyces marxianus, Saccharomyces cerevisiae, and Torulaspora delbrueckii. Previous investigations of the yeasts associated with mescal production in Mexico, described the presence of non-Saccharomyces strains, such as K. marxianus, C. lusitaniae, and Pichia fermentans from Agave salmiana fermentation, in San Luis Potosí State (Escalante-Minakata et al. 2008). In another Vinata, from the same region, the non-Saccharomyces yeasts were: K. marxianus, Pichia kluyveri, Zygosaccharomyces bailii, C. lusitaniae, T. delbrueckii, and Candida ethanolica (Verdugo-Valdez et al. 2011). In mescal produced using the species Agave durangensis in Durango, the predominant non-Saccharomyces yeasts belonged to Candida genus, including Candida lusitaniae, Candida kefir, Candida glabrata, Candida laurentii, and Candida tropicalis (Páez-Lerma et al. 2010). Equally, in Durango State, Páez-Lerma et al. (2013) observed diverse microorganisms at the beginning of fermentation: S. cerevisiae, T. delbrueckii, K. marxianus, Candida diversa, P. fermentans, and Hanseniaspora uvarum, but only T. delbrueckii and S. cerevisiae were found at the end of the fermentations. Recently, Kirchmayr et al. (2017) mentioned K. marxianus, Zygosaccharomyces rouxii, Z. bisporus, T. delbrueckii, and Pichia membranifaciens as the main microbiota present, after S. cerevisiae, during mescal production in Oaxaca State. In our study, Mezquital region presented the greatest number and diversity of isolated non-Saccharomyces, which included K. marxianus, T. delbrueckii, and C. lusitaniae. Both K. marxianus and C. lusitaniae were also detected in mescal from Nombre de Dios. In fermentation of agave in Durango, the species identified were K. marxianus and K. dobzhanskii. This article is the first report where the strain K. dobzhanskii has been found in natural fermentation processes. This genus has been cataloged as the closest Kluyveromyces lactis relative of wild or native strains, so it has been used for modeling population genetics (Belloch et al. 1997, 2002; Sukhotina et al. 2006; Lane and Morrissey 2010).

Table 1 Strains used in this study

In phylogenetic studies of Kluyveromyces strains (Fig. 1), three groups were recognized. The first two groups comprised strains directly related to the genus K. marxianus and K. dobzhanskii, respectively (Fig. 1a, b). The third group had direct relationship to the genera of the Kluyveromyces family (Fig. 1c). These strains were present in all the regions, accounting for 35% (Mezquital), 50% (Nombre de Dios), and 30% (Durango) of the total of the isolated Kluyveromyces strains and can represent a particular genetic diversity for K. marxianus strains isolated from the fermentation process during the production of mescal. In a recent study of the genetic diversity of the genus K. marxianus, all the isolates from a lactic environment were either diploid or triploid, whereas non-lactic isolates were haploid (Ortiz-Merino et al. 2018). Additionally, the authors distinguished three clades, of which the strain UFS-Y2791, isolated from American agave juice and representing the third clade, proved to be more diverse than the others (Ortiz-Merino et al. 2018). So far, only the presence of K. marxianus strains from different mescal production has been reported in the literature, indicating that the current work is the first to show that K. marxianus strains isolated from agave fermentation (mescal or tequila) have distinct genetic differences between them. Páez-Lerma et al. (2013) noted these differences with S. cerevisiae strains in wine.

Fig. 1

Neighbor-joining trees were constructed from the evolutionary distance data for ITSI-5.85 rDNA-ITS2. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates). aKluyveromyces marxianus tree of group one. bKluyveromyces dobhzankii tree. cKluyveromyces sp of group two tree. The accession numbers of reference sequences used in this tree are as follows: K. nonfermentans (AB011512.1), K. lactis (AB011515.1), K. wickerhamii (AB011521.1), K. aestuarii (AB011513.1), K. marxianus (AB011518.1), K. marxianus (MH045720.1), K. marxianus (MH045721.1), K. marxianus (MH045719.1), K. marxianus (MG966429.1), K. marxianus (JX174415.1), K. dobzhanskii (ABO11514.1), and D. hansenii (JQ912667.1). Evolutionary analyses were conducted in MEGA7

Additionally, phylogenetic analysis among the Clavispora lusitaniae yeasts from this study (Fig. 2a) evidenced the genetic variability between strains of Clavispora lusitaniae. These yeasts were found predominantly in fermentations from Nombre de Dios, with 45% of the strains identified as C. lusitaniae. Pérez-Brito et al. (2015) characterized the great genetic diversity of C. lusitaniae strains isolated from the fermentation of Agave fourcroydes Lem. There are numerous accounts of this species during the different stages of processing and fermentation of Agave to obtain traditional Mexican beverages, such as “pulque,” mescal, and tequila (Rodrigues de Miranda 1979; Lachance 1995; Lappe et al. 2004; De León Rodríguez et al. 2008; Lappe-Oliveras et al. 2008; Páez-Lerma et al. 2010; Kurtzman et al. 2011; Verdugo-Valdez et al. 2011), where its presence has been associated with the sensory qualities of these beverages (Escalante-Minakata et al. 2008).

Fig. 2

Neighbor-joining trees were constructed from the evolutionary distance data for ITSI-5.85 rDNA-ITS2. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates). aClavispora lusitaniae tree. bS. cerevisiae–T. delbrueckii tree. The accession numbers of reference sequences used in these trees are as follows: C. lusitaniae (AY321465.1), C. lusitaniae (KP131848.1), C. lusitaniae (KP131835.1), C. lusitaniae (HQ693786.1), C. lusitaniae (KU729100.1), C. lusitaniae (KP131836.1), C. lusitaniae (AY174102.1), C. lusitaniae (KY102564.1), C. lusitaniae MG518174.1), C. lusitaniae (MG183704.1), S. cerevisiae (JX141369.1), S. cerevisiae (JX141368.1), T. delbrueckii (FJ838774.1), and D. hansenii (JQ912667.1). Evolutionary analyses were conducted in MEGA7

The species present in relatively low quantity was T. delbrueckii, found only in the region of the Mezquital. Figure 2 b shows the phylogenetic tree for strains ITD0110 and ITD0129. These strains have been linked to a high production of volatile compounds that impart unique characteristics to beverages, such as mescal, and also other flavor compounds, including terpenoids, esters, higher alcohols, glycerol acetaldehyde, acetic acid, and succinic acid (Moreira et al. 2005; Jolly et al. 2014). Rutiaga-Quiñones et al. (2012) profiled the volatile compounds in Agave duranguensis juice supplemented with NH₄Cl and fermented with the yeast T. delbrueckii ITD0110. However, the genetic diversity present in this genus was not established. Nuñez-Guerrero et al. (2016) isolated S. cerevisiae, T. delbrueckii, and K. marxianus from A. duranguensis fermentation and proposed the use of a mixture of 75% S. cerevisiae and 25% T. delbrueckii as an inoculant to make mescal.

Production of volatile organic compounds

Table 2 presents the volatile compounds produced by the non-Saccharomyces yeasts studied in this work. In general, all strains were producers of esters, fatty acids esters, and higher alcohols. Esters are key flavor compounds in fermented beverages, like mescal. Among the acetate esters, the synthesis of ethyl acetate, which is responsible for the bouquet and desirable fruity flavors, depends on the ethanol concentration while the synthesis of isoamyl acetate, isobutyl acetate and 2-PEA, relies on the concentration of their corresponding higher alcohol, by the action of an alcohol acetyltransferase (Gethins et al. 2015; Loser et al. 2014).

Table 2 Volatile metabolites produced by non-Saccharomyces strains

The ethyl esters of short-chain fatty acids present are 2-phenylethyl butyrate and Phenylethyl propionate, synthesized from 2-PE and short-chain fatty acids. Phenylethyl propionate is an ester desirable in wine, due to its floral aroma (Beckner Whitener et al. 2015; Padilla et al. 2016). The formation of benzaldehyde from Phe has been studied in several microorganisms, such as Pseudomonas putida and the white rot fungi, Tremetes suaveolens, Polyporus tuberaste, Bierkandera adusteand, and Phanerochaete chrysosporium (Rojas et al. 2001; Liu et al. 2004). Hence, these yeasts seem to have benzaldehyde production potential. The principal volatile compounds found were 2-PE, which is considered to be one of the most important aromatic alcohols, and 2-PEA. The higher alcohols are predominantly formed by yeast during fermentation by the Ehrlich pathway, involving amino acid degradation, particularly Phe (Hazelwood et al. 2008; Styger et al. 2011) but can also be synthesized from glucose via pyruvate (Cordente et al. 2012). The Ehrlich pathway involves three steps: phenylpyruvate is decarboxylated to phenylacetaldehyde, then reduced to 2-PE (Etschmann and Schrader 2006), and finally esterified to 2-PEA.

The production of volatile organic compounds (2-PE and 2-PEA) presented significant statistical differences between by non-Saccharomyces yeasts is shown in Table 3.The most productive yeasts, in terms of 2-PE, were as follows: Kluyveromyces marxianus (ITD0090, ITD0091), as well as the yeasts Kluyveromyces sp. (ITD0046, ITD0089 and ITD0040), Kluyveromyces dobzhanskii ITD0157, and Clavispora lusitaniae ITD0107. K. marxianus yeast ITD0090 can be classified as the largest producer of 2-PE, even when compared with the most studied yeast K. marxianus CBS600 (Table 3). The concentrations obtained were similar to that reported by Etschmann et al. (2003) (890 mg/L in a molasses-based medium). Eshkol et al. (2009) evaluated the potential of stress-tolerant Saccharomyces strains to produce 2-PE under inductive conditions (Phe addition) and detected the concentrations between 340 and 460 mg/L at 48-h incubation, but these concentrations increased to 540 and 850 mg/L with selected yeast, when conditions were optimized with 10 g/L Phe addition, which were very similar quantities to those reported here, without the optimization process. In a related study, with the K. marxianus strain CBS6556, the optimization of the grape must culture medium with 3 g/L of l-Phe improved the 2-PE titer of 0.39 g/L after 84 h of culture to 0.47 g/L (Garavaglia et al. 2007). Mei et al. (2009) also used a yeast Saccharomyces cerevisiae BD and reported in situ product adsorption techniques, to obtain a better performance regarding the biotransformation of l-phenylalanine to 2-phenylethanol, reaching a concentration of 4.65 g/L of 2-PE with a content of 10 g/L of l-Phe in the medium. Chreptowicz et al. (2016) with yeast not genetically modified strain Saccharomyces cerevisiae JM2014 was isolated from a fermented milk drink (Turkey), producing a total concentration of 3.60 g/L of 2-PE after 72-h incubation at 30 °C batch culture with a medium containing 5 g/L of l-Phe in a 4-L bioreactor at laboratory scale. Recently, De Lima et al. (2018) evaluated the potential of yeast strain K. marxianus CCT 7735 in the 2-PE production and reported a production of 2.47 g/L of 2-PE, with the optimization in the medium through the optimal conditions achieving thus a production of 3.44 g/L of 2-PE. Lu et al. (2016) showed the 2-PE titer in a batch fermentation with the stress-tolerant yeast Candida glycerinogenes WL2002-5, reaching 5 g/L from l-Phe, under optimized culture conditions. Genetic modification strategies have also been considered, to further increase 2-PE production, such as ARO8 and ARO10 overexpression in S. cerevisiae SPO810 yeast the 2-PE reached 2.61 g/L after 60 h of cultivation (Yin et al. 2015).

Table 3 Production 2-phenylethanol and 2-phenylethylacetate obtained by different yeast strains non-Saccharomyces by HS-SPME

Chreptowicz et al. (2018) reported yeast strains capable of producing over 2 g/L 2-PE through the l-Phe biotransformation in standard medium for 72-h batch cultures. Clavispora lusitaniae WUT17 strain reached the levels of 2.04 g/L of 2-PE in a standard medium and 0.95 g/L of 2-PE in an organic waste-based medium, which is superior to the one reported by Etschmann et al. (2003) of 0.33 g/L. It is well known that 2-PE synthesis is carried out by the Ehrlich pathway in yeast, such as K. marxianus and Yarrowia lipolytica (Fabre et al. 1998; Celińska et al. 2013). In a recent study, González et al. (2018) screened the 2-PE production potential of some non-Saccharomyces yeasts and discovered a 2-PE productive yeast (T. delbrueckii). However, in all cases, non-Saccharomyces species produce lower quantities than S. cerevisiae, indicating that the Ehrlich pathway may not be as active in non-Saccharomyces species as in Saccharomyces, at least under nitrogen-limiting conditions. Rutiaga-Quiñones et al. (2012) revealed the non-Saccharomyces yeasts potential for volatile compounds, particularly in A. duranguensis juice for mescal production; in this study, the strains T. delbrueckii ITD0110 and K. marxianus ITD0211 showed to be more productive of 2-PE under nitrogen-limiting conditions that the strain S. cerevisiae ITD0109. A possible theory for our observations, when the fermentations of different strains induced with l-Phe as the only source of nitrogen were evaluated, is that the Ehrlich route is working on these strains, but the metabolic plasticity differs for each of the strains studied. These results allow to raise genetic and biochemical differences between the strains of wine production and mescal, but additional studies are required to elucidate and describe them.

The Table 3 illustrates that 2-PEA production presents significant difference for each strain where that highlighting the K. marxianus strains (ITD0040, ITD0090, ITD0102, and ITD0211). A previous research by Rojas et al. (2001) described a very productive H. guilliermondii yeast, with a 2-PEA production of 83.83–163.8 mg/L, when using 2-PE as induction conditions and in the presence of extraction solvent. The present results describe a difference in the production potential from l-Phe induction, among all the strains studied, highlighting two strains, K. marxianus ITD0090 and K. marxianus ITD0211, due to the potential to overproduce 2-PE and 2-PEA, respectively. Etschmann et al. (2005) described that the yeast K. marxianus CBS600 produced 1.3 g/L of 2-PEA, and a maximum of 4 g/L, using an organophilic pervaporation technique for continuous in situ product removal (ISPR) in a previously optimized medium. The production of 2-PEA by starter of Hanseniaspora vinae–Saccharomyces cerevisiae has been reported, where the concentration was 0.81 to 1.70 mg/L in wine; 2-phenylethyl acetate levels in wine vary from traces to 0.96 mg/L whereas its aroma threshold is around 0.25 mg/L (Viana et al. 2011). Also, a patent has been granted for the production of 2-PEA in anaerobic conditions with the Kluyveromyces marxianus KY3 strain with a production of 435 mg/L (Chang et al. 2014). Recently, it has been reported solid-state fermentation processes (SSF) for 2-PEA and 2-PE production using agroindustrial residue sugarcane bagasse as sole carbon source for the biotransformation of l-phenylalanine using Kluyveromyces marxianus strain as inoculum, showing effective results as in other systems submerged fermentation (Martínez et al. 2018).

Guo et al. (2017) designed and expressed a 2-PEA biosynthetic pathway in E. coli and in shake flask cultures with l-Phe (1 g/L) and recorded the generation of 268 mg/L of 2-PEA. This amount is very similar to the one reported in yeast ITD 00211, highlighting that the production by the strains of this study has not gone through the process of optimization.

Conclusions

This study presents preliminary evidence of differences between non-Saccharomyces yeasts found during fermentation of A. durangensis for the production of mescal. Particularly, Kluyveromyces yeasts have a high variability among them with respect to the production of volatile organic compounds, where it was evidenced that these have the extraordinary potential to produce aromas, particularly, 2 PE and 2 PEA.