Exploitation of Three Non-Conventional Yeast Species in the Brewing Process

Consumers require high-quality beers with specific enhanced flavor profiles and non-conventional yeasts could represent a large source of bioflavoring diversity to obtain new beer styles. In this work, we investigated the use of three different non-conventional yeasts belonging to Lachancea thermotolerans, Wickerhamomyces anomalus, and Zygotorulaspora florentina species in pure and mixed fermentation with the Saccharomyces cerevisiae commercial starter US-05. All three non-conventional yeasts were competitive in co-cultures with the S. cerevisiae, and they dominated fermentations with 1:20 ratio (S. cerevisiae/non-conventional yeasts ratios). Pure non-conventional yeasts and co-cultures affected significantly the beer aroma. A general reduction in acetaldehyde content in all mixed fermentations was found. L. thermotolerans and Z. florentina in mixed and W. anomalus in pure cultures increased higher alcohols. L. thermotolerans led to a large reduction in pH value, producing, in pure culture, a large amount of lactic acid (1.83 g/L) while showing an enhancement of ethyl butyrate and ethyl acetate in all pure and mixed fermentations. W. anomalus decreased the main aroma compounds in comparison with the S. cerevisiae but showed a significant increase in ethyl butyrate and ethyl acetate. Beers produced with Z. florentina were characterized by an increase in the isoamyl acetate and α-terpineol content.


Introduction
In the brewing industry, Saccharomyces cerevisiae and Saccharomyces pastorianus are the most used yeast species in starter cultures ensuring certain advantages in the fermentation process and in the standard product quality. In the past few years, brewers have payed attention to yeast selection not only for their fermentation efficiency and technological advantages but also for the aromatic characters and the flavors they give to the final product. To this end, recent genetic investigations have focused on methods to enhance the fermentation efficiency and aromatic profile of selected S. cerevisiae strains [1][2][3]. Other studies proposed the isolation of new starter yeasts (S. cerevisiae) from natural matrices [4,5] and the selection of wine yeast strains [6]. In recent years, the rapid expansion in the number of microbreweries has led the brewers to differentiate their products through the use of alternative raw materials and the selection and use of non-conventional yeasts [7][8][9][10]. Indeed, bioflavor from metabolic pathways of yeasts may be a suitable strategy to obtain beer with a different aromatic taste [11,12]. In this context, the use of non-conventional yeasts could be an alternative way to enhance the aroma profile of beers. Within non-conventional yeasts, different genera and species have been proposed. Torulaspora delbrueckii is a yeast species widely studied in winemaking for its ability to produce fruitiness and positive aromatic flavor, and for this reason, it was also evaluated for beer production, both in pure and in mixed cultures with different S. cerevisiae starter strains [7,13]. Beers

Analytical Procedures
The Official European Union Methods [18] were used to determine the ethanol content, volatile acidity, and pH. Acetaldehyde, ethyl acetate, n-propanol, isobutanol, amyl, and isoamyl alcohols, and acetoin were quantified by direct injection into a gas-liquid chromatography system (GC-2014; Shimadzu, Kjoto, Japan). Each sample was prepared and analysed as reported by Canonico et al. [19].
The solid-phase microextraction (HS-SPME) method was used to determine the concentration of the volatile compounds. Five mL of beer was placed in a vial containing 1 g NaCl closed with a septum-type cap. HS-SPME was carried out with magnetic stirring for 10 min at 25 • C. After this period, the internal standard (3-octanol) at a concentration of 1.6 mg/L was added, and the sample was heated to 40 • C. Divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fibre (Sigma-Aldrich) was inserted into the vial headspace for 30 min. The compounds were desorbed by inserting the fibre into a Shimadzu gas chromatograph GC injector for 5 min. A glass capillary column was used: 0.25 µm Supelcowax 10 (length, 60 m; internal diameter, 0.32 mm). The fibre was inserted in the split-splitless mode as reported by Canonico et al. [19]. The compounds were identified and quantified by comparisons with external calibration curves for each compound.
Final gravity was determined by a densimeter (Polsinelli Enologia Srl, Isola Liri, Italy) at a temperature of 20 • C. Specific enzymatic kit (Megazyme, Wicklow, Ireland) was used to determine the concentration of lactic acid (kit K-DLATE) according to the manufacturer instructions.

Statistical Analysis
Analysis of variance (ANOVA) was applied to the experimental data for the main characteristics of the beers. The means were analysed using STATISTICA 7 software. Significant differences were determined by the means of Duncan tests, and the results were considered significant if the associated p values were <0.05. Principal component analysis (PCA) was applied to discriminate between the means of the contents of esters, higher alcohols, and carbonyl compounds in the beers from the pure and mixed fermentations PCA, which was carried out using the statistical software package JMP 11 ® . The mean data were normalised to neutralize any influence from hidden factors. The PCA provides a graphical representation of the overall differences due to the non-Saccharomyces in terms of fermenting products of the final beers.

Preliminary Screening
Initial screening of the strains belonging to the different species was carried out to determine their ability to ferment maltose, the most abundant fermentable sugar in the brewing wort, to select for their potential use in beer production ( Table 1). The results of the fermentation capacity showed that all of the strains tested exhibited the ability to ferment maltose, although even at different grades. From the strains tested, we chose L. thermotolerans DISVA 322, Z. florentina, DiSVA 263, and W. anomalus DiSVA 2 since they showed the best fermentative performance with the highest final CO 2 evolved. Regarding the species W. anomalus, the data reported in Table 1 highlighted that different strains belonging to this species exhibited a good fermentation performance. The choice to select W. anomalus DiSVA 2 has been evaluated taking into account the screening and the results reported by Oro et.al [20] regarding the metabolic activities and aromatic compounds production. For these reasons, these three strains were selected for the subsequent micro-fermentation trials in pure and in co-culture at different inoculation ratios S. cerevisiae/non-conventional yeasts (1:1; 1:10; 1:20) using the S. cerevisiae starter strain US-05. Data are the means ± standard deviations. CO 2 g evolved after 20 days of fermentation (in 100 mL of 10% malt extract). Fermentation rate: CO 2 g/day (over the first 6 days of fermentation).

Evaluation of Population Dynamics
The growth kinetics of L. thermotolerans in pure and in mixed fermentation are reported in Figure 1. The S. cerevisiae US-05 pure cultures achieved 10 8 CFU/mL after four days of fermentation (maximum viable cells) and maintained 10 7 CFU/mL until the end of the fermentation. Differently, in the mixed cultures with the inoculation ratio of 1:1 (Figure 1a), S. cerevisiae US-05 reached a lower biomass, achieving only 10 7 CFU/mL on the fourth day of fermentation. The same trend was also exhibited for S. cerevisiae US-05 with the inoculation ratios of 1:10 and 1:20 (Figure 1b,c). S. cerevisiae was not inhibited at all in the inoculation ratios and completed fermentation as showed by residual sugar and ethanol content in the final beers ( Table 2).
L. thermotolerans in mixed fermentation at the 1:10 and 1:20 inoculation ratios showed a biomass evolution comparable with that of the pure culture, while at the 1:1 inoculation ratio, L. thermotolerans showed a very limited increase in biomass compared with that exhibited by S. cerevisiae under the same conditions (Figure 1a). Under these conditions, both strains suffered a mutual inhibition but showed no differences compared with the control trial (US 05 strain) ( Table 2).
The biomass evolution of the S. cerevisiae and Z. florentina mixed fermentations at the 1:1 inoculation ratio ( Figure 2a) exhibited a similar growth kinetic (with a slight prevalence of S. cerevisiae), and they reached a maximum cell concentration of 10 7 mL/L after four days, showing a reciprocal inhibition compared with pure cultures.   The biomass evolution of the S. cerevisiae and Z. florentina mixed fermentations at the 1:1 inoculation ratio (Figure 2a) exhibited a similar growth kinetic (with a slight prevalence of S. cerevisiae), and they reached a maximum cell concentration of 10 7 mL/L after four days, showing a reciprocal inhibition compared with pure cultures. At the 1:10 and 1:20 inoculation ratios (Figure 2a, b), the two yeast species did not show the antagonistic effect on each other, however, Z. florentina prevailed over S. cerevisiae during the fermentation processes.
The cell evolution of W. anomalus in pure and mixed fermentations with S. cerevisiae showed the same trend exhibited by Z. florentina (Figure 3), indicating a good competition with S. cerevisiae in wort fermentation, especially when the inoculation level was 10-or 20-fold higher (1:10 and 1:20 trials). Despite the competition shown by the three non-conventional yeast species toward S. cerevisiae, all mixed fermentation trials (at different inoculation ratios) did not show relevant differences in the main analytical characteristics in comparison with the S. cerevisiae control trial At the 1:10 and 1:20 inoculation ratios (Figure 2a,b), the two yeast species did not show the antagonistic effect on each other, however, Z. florentina prevailed over S. cerevisiae during the fermentation processes. The cell evolution of W. anomalus in pure and mixed fermentations with S. cerevisiae showed the same trend exhibited by Z. florentina (Figure 3), indicating a good competition with S. cerevisiae in wort fermentation, especially when the inoculation level was 10-or 20-fold higher (1:10 and 1:20 trials). Despite the competition shown by the three non-conventional yeast species toward S. cerevisiae, all mixed fermentation trials (at different inoculation ratios) did not show relevant differences in the main analytical characteristics in comparison with the S. cerevisiae control trial (Table 2). Therefore, all the mixed fermentations have completed the fermentation as attested by the values of final gravity that are comparable with S. cerevisiae pure culture (Table 2). At the 1:10 and 1:20 inoculation ratios (Figure 2a, b), the two yeast species did not show the antagonistic effect on each other, however, Z. florentina prevailed over S. cerevisiae during the fermentation processes.
The cell evolution of W. anomalus in pure and mixed fermentations with S. cerevisiae showed the same trend exhibited by Z. florentina (Figure 3), indicating a good competition with S. cerevisiae in wort fermentation, especially when the inoculation level was 10-or 20-fold higher (1:10 and 1:20 trials). Despite the competition shown by the three non-conventional yeast species toward S. cerevisiae, all mixed fermentation trials (at different inoculation ratios) did not show relevant differences in the main analytical characteristics in comparison with the S. cerevisiae control trial (Table 2). Therefore, all the mixed fermentations have completed the fermentation as attested by the values of final gravity that are comparable with S. cerevisiae pure culture (Table 2).

Main Analytical Profiles
The data regarding the analytical compositions of the beers produced are reported in Table 2.
The pure cultures with non-conventional yeasts showed significant reductions in ethanol content compared with the S. cerevisiae starter strain (4.03% v/v). Indeed, beers produced by pure cultures of L. thermotolerans, Z. florentina, and W. anomalus exhibited an ethanol content of 3.12% v/v, 3.47% v/v, and 1.53% v/v, respectively. Consequently, pure cultures of the non-conventional yeasts

Main Analytical Profiles
The data regarding the analytical compositions of the beers produced are reported in Table 2.
The pure cultures with non-conventional yeasts showed significant reductions in ethanol content compared with the S. cerevisiae starter strain (4.03% v/v). Indeed, beers produced by pure

Main Analytical Profiles
The data regarding the analytical compositions of the beers produced are reported in Table 2.
The pure cultures with non-conventional yeasts showed significant reductions in ethanol content compared with the S. cerevisiae starter strain (4.03% v/v). Indeed, beers produced by pure cultures of L. thermotolerans, Z. florentina, and W. anomalus exhibited an ethanol content of 3.12% v/v,

Main Analytical Profiles
The data regarding the analytical compositions of the beers produced are reported in Table 2.
The pure cultures with non-conventional yeasts showed significant reductions in ethanol content compared with the S. cerevisiae starter strain (4.03% v/v). Indeed, beers produced by pure cultures of L. thermotolerans, Z. florentina, and W. anomalus exhibited an ethanol content of 3.12% v/v,

Main Analytical Profiles
The data regarding the analytical compositions of the beers produced are reported in Table 2.
The pure cultures with non-conventional yeasts showed significant reductions in ethanol content compared with the S. cerevisiae starter strain (4.03% v/v). Indeed, beers produced by pure cultures of L. thermotolerans, Z. florentina, and W. anomalus exhibited an ethanol content of 3.12% v/v,

Main Analytical Profiles
The data regarding the analytical compositions of the beers produced are reported in Table 2. The pure cultures with non-conventional yeasts showed significant reductions in ethanol content compared with the S. cerevisiae starter strain (4.03% v/v). Indeed, beers produced by pure cultures of L. thermotolerans, Z. florentina, and W. anomalus exhibited an ethanol content of 3.12% v/v, 3.47% v/v, and 1.53% v/v, respectively. Consequently, pure cultures of the non-conventional yeasts showed higher final gravities. On the other hand, all mixed cultures produced beers with an ethanol content not significantly different to that of the S. cerevisiae US-05 control.
For the volatile acidity, the results did not show significant differences for the trials carried out with L. thermotolerans in comparison with S. cerevisiae pure culture, while Z. florentina at inoculation ratio 1:20 exhibited a significant increase in acetic acid content in comparison with S. cerevisiae and the other trials. Regarding W. anomalus significant increase was exhibited in pure and at inoculation ratio 1:20.
All of the beers exhibited pH values comparable to that of the S. cerevisiae starter strain, with the only exception being L. thermotolerans in pure and mixed fermentations. In particular, L. thermotolerans in pure culture exhibited a pH reduction of ca. 0.6 points. Regarding the lactic acid content, L. thermotolerans in pure and at different inoculum ratio exhibited a significantly higher concentration in comparison with the S. cerevisiae and the other two non-conventional yeasts. On the other hand, W. anomalus and Z. florentina in pure cultures exhibited lactic acid amounts significantly higher than those shown by the S. cerevisiae. Moreover, Z. florentina also at inoculation ratio 1:1 and 1:20 (S. cerevisiae/Z. florentina) exhibited a significant increase in comparison with S. cerevisiae, while at 1:10 ratio showed a comparable amount of lactic acid of S. cerevisiae. All mixed fermentation carried out with W. anomalus significantly increased the lactic acid content in comparison with S. cerevisiae.

Main Volatile Compounds
The main volatile compounds produced by L. thermotolerans in pure and mixed fermentations are reported in Table 3. The results highlighted a significant increase in ethyl acetate (fruit notes) content in pure and mixed fermentations compared with S. cerevisiae. L. thermotolerans in pure and mixed fermentations showed a significant increase in ethyl butyrate content, which is the ester responsible for the fruity or solvent aroma of beer [21], compared with S. cerevisiae US-05. A significant increase in isoamyl acetate content (banana aroma) was exhibited by L. thermotolerans at the 1:20 S. cerevisiae: L. thermotolerans inoculation ratio compared with S. cerevisiae in pure culture, while the concentration of ethyl hexanoate (i.e., apple, fruit flavour) decreased with increasing the inoculation ratios. The same trend was also exhibited for the linalool content. Furthermore, the citronellol content decreased in mixed fermentation at the 1:10 and 1:20 inoculation ratios, while at the 1:1 its concentration was comparable with that exhibited by the S. cerevisiae starter strain. Moreover, the mixed cultures showed significant increases in higher alcohols levels with the exception of β-phenyl ethanol. In regard to the acetaldehyde content, the mixed fermentations at the 1:1 inoculation ratio exhibited a reduction compared with the other beers.
In mixed fermentation trials, Z. florentina showed a significant decrease in acetaldehyde content compared with the beer obtained with the S. cerevisiae starter strain (Table 4).
Regarding the ester compounds, Table 4 showed a significant increase in ethyl butyrate content at 1:20 inoculation ratio in comparison with the other trials, while ethyl acetate increases in all mixed fermentation, phenyl ethyl acetate, ethyl hexanoate, and isoamyl acetate decreases when raising the inoculation ratio. Moreover, the mixed fermentations showed a significant increase in some higher alcohols content, with the only exception of β-phenyl ethanol that decreased in mixed fermentation in comparison with S. cerevisiae pure culture. Regarding terpens compounds, no significant differences were exhibited for α-terpineol, while citronellol content significant decrease in mixed fermentation in comparison with S. cerevisiae pure culture.

Discussion
In winemaking, the controlled use of non-conventional yeast species has been investigated to enhance complexity and give distinctive flavour profiles of wines [22][23][24].
Indeed, fermented alcoholic beverages with peculiar and distinctive flavours can be obtained through the yeast inoculation during the fermentation process [11,21,25]. In this regard, the potential use of non-conventional yeasts in brewing has been poorly explored. Only recently, some studies focused on flavouring by non-conventional yeasts in brewing [7,9,14,16,26].
In this work, we investigated three non-conventional yeast species widely found in the wine environment [23,27,28]. The three strains chosen within the three species were competitive versus the S. cerevisiae starter strain at the 1:1 inoculation ratio exhibiting similar numbers of viable cells. The competition of the non-conventional species in wort under semi-anaerobic conditions was confirmed for the 1:10 and 1:20 ratios (S. cerevisiae brewers' yeast: Non-conventional species). In these last conditions, all three non-conventional yeast species coexisted with brewers' yeast dominating the fermentation process. Similar results were found by Van Rijswijck et al. [17] where Candida fabianii at a 1:100 ratios (S. cerevisiae: C. fabianii) dominated the fermentation. However, compared to the results of the present work, a large amount of residual maltose was found.
This behaviour is of interest from a practical application point of view. Indeed, the dominance allows displaying the metabolism of the non-conventional yeast in co-culture with the brewing starter, obtaining, at the same time, a distinctive aromatic impact on final product and a correct and complete evolution of the fermentation. In this regard, differences in the analytical composition using the three non-conventional yeasts in pure and in co-culture were found (see Figure 4). All three non-conventional species determined a general decrease of acetaldehyde in all mixed fermentations (1:1, 1:10, 1:20 S. cerevisiae/non-conventional yeasts ratios).
L. thermotolerans, a yeast species widely investigated in wine fermentation, [28,29] showed a large reduction in pH compared with the S. cerevisiae starter strain. Domizio et al. [15] and Osburn et al. [8], investigating L. thermotolerans in pure fermentation with the goal of producing sour beer, and found a compared trend in pH reduction. In the wort used in this work, L. thermotolerans in pure culture produced a larger amount of lactic acid (1.83 g/L) compared with the quantities shown in the study of Domizio et al. [15] (0.24 g/L). Regarding the compounds influencing the bioflavor, L. thermotolerans produced beer with a higher ethyl butyrate content and an increase in ethyl acetate (i.e., floral, honey, sweet) in pure and in all of the mixed fermentations compared with S. cerevisiae. Moreover, a significant increase in isoamyl acetate content (banana aroma) was exhibited in the trial S. cerevisiae/L. thermotolerans at an inoculum ratio of 1:20. These results highlighted that this strain was interesting not only for sour beer production due to its lactic acid production and consequent pH decrease during primary fermentation but also to tailor the aroma profile of the beer. The low ethanol content exhibited by L. thermotolerans pure cultures was also relevant for the production of low alcohol beer. W. anomalus generally studied in apple wine and hard cider production [30,31], it was poorly investigated in brewing because it is associated with the concept of "spoilage yeast" [32]. Only Osburn at al. [8] showed that a strain of W. anomalus exhibited a good fermentation performance and led beers with fruit notes such as apple, pear, and apricot. Results of this study showed an increase in ethyl butyrate and ethyl acetate and a reduction in acetaldehyde in all mixed fermentations that positively affect the aromatic profile of the beer. As reported above, there are no studies on the use of this species in beer, thus, for this reason, it is difficult to understand the general behaviour of this non-conventional yeast in the brewing process and further investigations are needed. Furthermore, there is only one study that evaluated the use of the non-conventional yeast Z. florentina in brewing. In contrast to the report of Holt et al. [16] that showed increases in higher alcohols, we observed a different trend for these compounds. On the other hand, Z. florentina showed increased isoamyl acetate and α-terpineol content (floreal aroma) that positively influenced the final flavour of the beer. In conclusion, the use of non-conventional yeasts in mixed fermentations with S. cerevisiae is a suitable strategy to tailor flavour production during beer fermentation, thus making it possible to obtain products with aroma compounds that are different from those of beers brewed using pure S. cerevisiae starter strains. These data confirm that the brewing yeast used can modulate the production of the aroma compounds in the final beers. The modalities (pure or co-cultures with S. cerevisiae) and the inoculation ratio in mixed fermentations should be further investigated, even if the non-conventional yeasts coexist, and in some cases, dominate the process, the fermentations are almost completed. In this regard, further work is needed to understand the behaviours of these non-conventional yeasts in the brewing process and to know their possible uses in different beer styles. The next necessary step will be evaluating the selected combination of S. cerevisiae/non-conventional yeasts on a large scale to evaluate through sensory analysis the beer style more suitable for these yeasts.
Author Contributions: L.C., E.G., E.C., M.C. and F.C. contributed equally to this manuscript. All authors participated in the design and discussion of the research. L.C., E.G. and E.C. carried out the experimental part of the work. L.C., E.G., M.C. and F.C. carried out the analysis of the data and wrote the manuscript. All authors have read and approved the final manuscript.