Yeast from Distillery Plants: A New Approach

Nowadays, there is more and more interest in the microbiological resources from different ecosystems, not only because this would allow knowing more about the microbial biodiversity related with these substrata but also because it provides an opportunity to study their characteristics and technological properties which may be of potential interest. This knowledge may allow finding future biotechnological applications for these microorganisms on bio-conservation and reuse of agricultural by-products and may also lead to studies on the improvement of raw material processing. Some raw materials and processing plants in wine and related industries constitute a suitable place for yeast growth; for example, musts, wines in cellars, piquettes, bagasse, pomace, grape skins and yeast lees in the ethanol industry all provide an inexhaustible supply of yeasts. Few microbiological studies have been published so far about the biodiversity of the yeast population in distillery plants. For that reason, the aim of this research was to determine yeast biodiversity and their distribution in different distillery plants in the La Mancha region which are at least 100 years old.


Introduction
Agricultural residues from food industries are an important raw material involved in bioethanol production. Traditionally, residual juice, molasses and pomace from sugarcane, agave and sugar beet have been widely used in South America for obtaining distilled beverages such us cachaça, tequila and rum. The distillation process is used to isolate, select and concentrate pleasant volatile compounds from the previously fermented liquids and concentrate the alcohol content. Additionally, certain long esters from yeast cells are extracted by distillation and transferred to the final product [1].
Microbial communities from these raw materials and their fermented and distillate beverages not only are interesting due to their role in the aroma production, but their biodiversity and other biotechnological properties are also important. Yeast populations from these ecosystems have not been studied very much, and any studies on them have normally been focused on tequila [2][3][4][5], rum [6] and cachaça beverages [7][8][9][10].
Yeasts are able to spread from diverse niches to many environments, especially in the vegetable world [11]. Crops and processing plants provide a good niche for yeast growth. In fact, grape crops, musts and wines have been thoroughly studied [12][13][14], although distillate products and their industry have not been analyzed in Spain in spite of the fact that it is believed to be a new environment for yeast biodiversity study and its biotechnological applications.
In recent years, Spain has been established as the vineyard of the world, presenting the largest surface area (13%) dedicated to this crop [15]. The wine industry is an important sector in Spain which grew considerably throughout 2018. The number of cellars has increased by 6.8% with wine production also increasing (26%) and current production being 40.9 million hectolitres.
Castilla-La Mancha is the world's largest vine-growing region with an annual wine production of around 17 million hectolitres in the 2017-2018 vintage, which accounted for nearly 50% of the total Spanish production. Part of this large production is derived from the distillery industry; in the last year, nearly 250,000 hectolitres were transformed into alcoholic derivatives (16). There is a total of 33 authorized distilleries for wine by-product distillation, 13 of which are located in the La Mancha region. These industries process not only wine but also sweet grape pomaces and its fermented products, obtaining around 4-4.5 million hectolitres [16].
Wine production generates around 600,000 tons of grape derivatives annually such as fermented red skins, which still contain reducing sugars and ethanol, and sweet pomace (from white wine vinification). These by-products, as well as yeast lees and flocculated yeasts, are transported to distilleries where the ethanol is extracted. As Figure 1 shows, sweet pomaces are mixed and stored for 10-15 days, starting a spontaneous fermentation process. Then, pomace and grape skins are washed with water at 50°C in a heat diffusion system in order to extract the residual sugars and ethanol. After that, a liquid is obtained which is a mixture of alcohol (3-4%) (V/V), water and sugar and is called fermented or sweet piquette. On the other hand, a liquid is drained during the storage of solid organic waste which is mixed with the piquette and fermented for 2-3 days in a stainless-steel container, obtaining a higher alcohol  content (4-5% V/V). Finally, red fermented skins are washed at a lower temperature with the aim of extracting the residual ethanol [17,18].
The fermented piquettes and the drained liquid are distilled, producing a 93% (V/V) alcohol content product. Then, a dehydration process is carried out until the ethanol concentration of 99.9% is reached. This is mainly used in gasoline as an anti-detonating additive. Residues from distillation can be used as solid fuel (solid residue or "bagasse") or as fertilizer (liquid residue or "vinasse") [17].
Spontaneous fermentations during this last process are produced by non-Saccharomyces and Saccharomyces biota present in the environment whose biodiversity has not been widely studied [19].
The lack of information about yeast ecology in this habitat and, more specifically, in this territory has prompted the aim of this research.

Sample collection
Six of the largest distilleries in Europe, which are at least 100 years old and are found in the towns of Argamasilla de Alba (A), Campo de Criptana (B), Madridejos (C), Villarrobledo (D), Daimiel (E) and Tomelloso (F) in the La Mancha region (Figure 2), were selected to carry out the study. La Mancha is the principal area for the production of bioethanol and distillates in Spain.
A total of 47 samples were randomly collected from sweet piquettes [20], fermented piquettes [19], flocculated lees [7] and plant oil [1] throughout the pomacebased ethanol production process, and they were transported to the laboratory under aseptic and refrigerated conditions.
Samples and/or their dilutions were spread on YPD agar plates (10 g/L yeast extract, 20 g/L peptone, 20 g/L glucose and 20 g/L agar); chloramphenicol and sodium propionate were added to inhibit bacteria and mold growth, respectively. Plates were incubated at 28°C/72 hours. Then, samples displaying fewer than 30 colonies were centrifuged to concentrate the cells, and the pellet was directly spread Advances in Grape and Wine Biotechnology 4 on YPD agar. Plates with sufficiently separated colonies were replicated onto lysine agar medium (Oxoid, Basingstoke, UK) to distinguish between Saccharomyces sp. and non-Saccharomyces sp.
The isolates were obtained from 19 samples. A sample was not taken from distillery F, which is possibly due to the hot washing of the skins which would drastically decrease the number of cells.
A total of 210 purified isolates were obtained, 144 Saccharomyces and 66 non-Saccharomyces, and were stored in 15% glycerol at −80°C until they were studied.

Yeast classification by genetic identification
Saccharomyces spp. yeasts were the predominant profile in all distillery plants. However, the number of non-Saccharomyces species varied between distilleries.

Non-Saccharomyces yeasts
Genetic species identification was done using the polymerase chain reaction/ restriction fragment length polymorphism (PCR-RFLP) technique, by amplifying the 5.8S rRNA gene using ITS1 and ITS4 [20]. Amplified products were digested (37°C for 7 h) with the three restriction endonucleases Hinf I, Hae III and Cfo I.
Both PCR products and their restriction fragments were separated on agarose gel with GelGreen™ (Biotium), and the results were visualized using a GeneFlash documentation system. For those isolates that could not be identified by PCR-RFLP analysis, the region D1/D2 from the domain 26D rRNA gene was sequenced using NL1 and NL4 primers. If any variation existed due to the action of the NL4 primer, LR6, NL3A and NL2A primers were used as alternatives. Finally, for those samples in which the percentage of identity at species level was less than 99%, the ITS region was sequenced using ITS1 and ITS4 primers [17]. In Table 1 all isolates are shown, classified at the species level with 99% similarity and the NCBI accession number obtained. A percentage of similarity lower than 99% was obtained with isolates 23, 33, 48 and 62 using the primers NL1/NL4. Sequencing of the 5.8S rRNA + ITS region confirmed this with a similarity of 99%.
As can be observed in Figure 4, non-Saccharomyces yeasts were more present in plant C (14%) and in plant D (47%) due to the difference in the age of the distilleries and the specific elaboration process followed.
There were four predominant species that were identified as Pichia galeiformis, Torulaspora delbrueckii, Hanseniaspora osmophila and Candida lactis-condensi. All these results showed that a considerable diversity exists in this environment, unlike in grape must fermentations [21].
With regard to the substrata of isolation (Figure 3), T. delbrueckii, H. osmophila, P. kudriavzevii, C. lactis-condensi and P. anomala were isolated from sweet piquettes, while P. galeiformis and C. ethanolica were found in fermented piquettes, from which other species, such us S'codes ludwigii, P. bimundalis, Zygosaccharomyces bailii and C. sake, were also isolated but at a very low percentage. Only two species, L. thermotolerans and T. delbrueckii, were found equally frequently in both sweet and fermented piquettes.
Having analysed all these results, a large biodiversity of yeasts was found in the studied substrata, as was documented for grape marc by Bovo et al. [22,23].
On the other hand, the distribution of genera (Figure 4a) and species (Figure 4b) in the studied distilleries was also analysed. Candida and Pichia genera were found in almost all of them, and Torulaspora and Hanseniaspora were found in three of the five plants in which yeasts were isolated. P. galeiformis, P. kudriavzevii, T. delbrueckii and H. osmophila were the species identified in most of the ethanol plants, with plant A being the only one where no major species were found, which contrasts with the results for the other plants (Figure 4b).  The presence of Candida species (C. sake, C. sorbosa, C. stellata, C. guilliermondii, C. karawaiewii and C. citrea), P. membranaefaciens, P. guilliermondii, K. marxianus and large Saccharomyces spp. populations has been previously documented in Brazilian distilleries [7,8]. These results confirmed that the yeast profiles in the distilleries of the two regions are very different and it is evident that the Spanish industry is an interesting yeast niche. Additionally, some of these genera and species were also found by Amaya-Delgado et al. [5] and Lappe-Oliveras et al. [4] in tequila and agave beverages.

Saccharomyces yeasts
For Saccharomyces isolate characterization, a PCR-RFLP analysis was done, and the results showed that 95% of the isolates belonged to Saccharomyces cerevisiae, while only 3% and 2%, respectively, were identified as S. paradoxus and S. bayanus.
However, to discriminate isolate samples within the Saccharomyces sensu stricto group, a mitochondrial DNA restriction analysis [13] was carried out by digestion with the restriction endonuclease enzyme Hinf I. Restriction fragments were separated by electrophoresis on agarose gel with GelGreen™ (Biotium), and the results were visualized using a GeneFlash documentation system.
The Saccharomyces isolates were clustered in 105 different mtDNA patterns ( Table 2), reflecting a variability of nearly 73% which is very high if it is compared to their variety in cellars [24][25][26].
Genetic patterns which involved at least 20% of the isolates were named as the "majority profile". At plants A and C, two majority profiles were characterized; at B and D, there was only one; and none was found at plant E. In addition, sweet and fermented piquettes were the substrata from which the most profiles were identified. Although patterns tended to be typical of each plant, the majority profiles accounted for 57% of the isolates at plant B and 33% and 30% at plants C and A, respectively.
Fermented piquettes presented the greatest degree of Saccharomyces variability, although several strains coexisted in both lees and sweet piquettes.

Biotechnological properties of non-Saccharomyces: fermentation and assimilation of carbon compounds
Fermentation of carbon compounds is particularly useful for identifying isolates with new fermentation profiles for potential applications in various fields. The carbon compounds assayed were D-glucose, D-galactose, L-arabinose, L-rhamnose, melibiose, lactose, raffinose, xylose, maltose, mannose, saccharose and cellobiose.
The tests were carried out on a 96-well microtiter plate. Each well was filled with sugar solution, bromocresol green and cell suspensions (exhausting the endogenous carbon compound reserves). Finally, the wells were sealed with sterile vaseline, and the plates were incubated at 28°C/5 days. Depending on the time of the change and the intensity of colouration (from blue to yellow or yellow green), a classification system was established [27,28].
The majority of the isolates (Torulaspora, Lachancea and Saccharomycodes species and C. lactis-condensi) fermented D-glucose either in the first 12 h or on the 5th day. D-mannose and saccharose were fermented to a lesser extent.
None of the isolates fermented xylose, lactose, arabinose, melibiose and rhamnose, and some only weakly fermented galactose, maltose and raffinose.
C. lactis-condensi fermented the majority of the sugars at a major or minor intensity. On the other hand, for galactose, raffinose and saccharose fermentation, variability was observed in species such as T. delbrueckii, C. lactis-condensi, P. galeiformis and C. ethanolica.
Only one H. uvarum isolate and one H. vinae isolate weakly fermented cellobiose, which is a sugar of great biotechnological interest in the production of bioethanol from agricultural and forest by-products.
The tests were carried out in agar plates containing the carbon source and YNB without amino acids (Difco™). The assimilation profile was noticed as (++) abundant growth, (+) normal growth and (−) absence of growth.
Assimilation of carbon compounds, glucose and maltose were the most commonly used and, to a lesser extent, xylose and methanol. Three species of Candida, C. viswanathii, C. ethanolica and C. sake, and one P. galeiformis isolate assimilated carboxymethyl cellulose, while three Pichia isolates used starch. The majority of Torulaspora isolates and a few isolates of P. kudriavzevii, P. galeiformis and H. osmophila assimilated xylose. All of the H. osmophila, H. uvarum and S'codes ludwigii isolates effectively assimilated cellobiose. Ethanol was assimilated by a few P. galeiformis and P. anomala isolates. Finally, only some L. thermotolerans, P. kudriavzevii, C. sake and C. viswanathii isolates assimilated methanol. Thus, differences between isolates of the same species were observed, as can be seen in the fermentation tests.

Biotechnological properties of Saccharomyces: cell vitality and growth rate at different temperatures
Cell vitality and growth rates at different temperatures were carried out with the 105 strains. These properties were selected because they are considered a relevant characteristic in a fermentation process.
Cell vitality was evaluated as a measure of fermentative activity by an indirect electrical method [29].
Detection time (DT), expressed in hours, was obtained by impedance measured. It was considered that strains with lower DT presented high vitality.
DT results were clustered in five groups, as can be observed in Figure 5. In the interval 0.61-0.95 h, 10% of the strains studied were included, suggesting the highest vitality. Other yeasts (27%) were involved in the range between 0.96 and 1.29 h, indicating a fast cell vitality. Nevertheless, most stains (40%) were comprised between 1.30 and 1.64, and only 3% showed a low cell vitality (1.93-2.33 hours). These results indicate that yeasts from distillery plants have adequate vitality and probably they can displace the slower strains. In studies carried out by Ortíz et al. [13] and Barrajón et al. [29], it was noticed that DT of Saccharomyces wine strains oscillated from 0.67 to 1.80 h, although most of the strains showed a DT higher than 1.5 h.
The kinetic parameters (the maximum growth rate, generation time and maximum optical density) were studied at different temperatures (18,24,28,38,40 and 42°C) using a hurdle selection criteria. All strains were evaluated at 28°C and, depending on their specific growth rate (h −1 ), were distributed into three groups: higher rate values correspond to the first group and the lowest to the third group.
Strains in the top range were assayed at 38 and 24°C. Likewise, strains with the best rate at 38°C were then tested at 40°C, and those which showed the best rate were again tested at 42°C. Similarly, the best strains at 24°C were also tested at 18°C.
At 28°C, 41 of the 105 evaluated strains were in the first group with the bestperforming growth rates (0.25-0.32 h −1 ), and 46 and 21 strains, respectively, were categorized in the second and third groups.
At 38°C, the groups presented 14, 13 and 14 strains, respectively, with homogenous results. Afterwards, assays were performed at 40 and 42°C, based on the same criteria. It is remarkable that at 42°C the duration of lag phase was higher than 45.5 hours. Nevertheless, at 40°C, 13 strains from the 25 studied gave the worst growth rate, which constitutes an expected result since this temperature is suboptimal.
In Figure 6 the percentage of strains in every group was showed. It can be observed that at 38°C, strains were dispersed among three groups. Nevertheless at 18 and 40°C, most strains were included in the worst group.
In the majority of the cases, growth rates at low temperatures (≤0.2 h −1 ) were worse than those gotten at 40°C (≥0.25 h −1 ); this fact confirms that the microbial growth (outside the optimal temperature interval) is better at higher temperatures ( Table 3). This is a logical outcome, because in distillery plants the substrate is  washed with warm water, and the yeasts isolated from there will grow better at higher temperatures.
The thermal washing process for the extraction of alcohol contributes to the presence of Saccharomyces strains with technologically interesting properties, especially in terms of vitality and resistance to high temperatures.

Biocontrol activity of yeast against epiphytic molds
The molds were provided from the culture collection of the University of Castilla-La Mancha (UCLM) and IVICAM (Grapevine and Wine Institute of Castilla-La Mancha). They were Phaeomoniella (Pa.) chlamydospora, Neofusicoccum parvum, Diplodia seriata, Phaeoacremonium (Pm.) aleophilum and Aspergillus niger.
Fungi were grown in YM agar, and pieces of agar with fungal mycelium were inserted in wells excavated in the YM agar which had been previously inoculated with yeast strains.
The results showed that there were both inter-and intraspecific variabilities. H. meyeri, H. uvarum, H. vineae and H. valbyensis scarcely controlled fungal growth, and mycelium grew as in the control except for six H. osmophila which showed a good action against them.
However, P. anomalous, P. galeiformis and P. kudriavzevii effectively controlled all fungal strains including A. niger. Also, all S. cerevisiae strains except one presented good    fungal growth control behaviour towards all the molds, and A. niger was inhibited effectively by only one of these strains. Additionally, the different C. ethanolica and C. sake have an effective action on the fungal growth, except in the case of C. lactis-condensi. Finally, T. delbrueckii and S'codes ludwigii strains proved to have a large biocontrol effect not only because of their action against the growth but also because they affected every mold.
Most of the yeasts grew rapidly, forming a very dense lawn after 2 days of growth, suggesting that the mechanism of control might be based on a competition for space and nutrients. To qualitatively analyse the degree of competition between yeast and mold, the 0-day test was carried out afterwards. The assay was carried out with the yeast species which presented the best result in the previous experiment (Figure 7), allowing the detection of a high degree of competition between the two microorganisms. DOI: http://dx.doi.org/10.5772/intechopen.86291 The Pichia species and the only S'codes ludwigii assayed offered a high degree of control. One of the conclusions given by these trials is that the competition between yeast and mold for nutrients and space appeared from the first moment of contact, probably due to the very different growth rates, i.e. the yeasts have a high rate and rapidly colonize the medium preventing the development of molds. However, the inhibition mechanism may be associated with other antagonistic or enzymatic activities occurring via the production of some active compounds.
With the aim of verifying if the inhibition mechanism was produced by cell metabolites or cell wall components, the biocontrol assays were carried out with viable yeast cells, cell extract and filtered supernatant. To carry out the experiment, four wells were excavated at different points on growth fungal plates and were filled with each faction and a negative control (lysis buffer). All of them were incubated at 30°C for a maximum of 5 days in a wet chamber [30].
In most of the tests, an inhibition halo was observed with cell extracts, but when compared to the control (lysis buffer), it was difficult to identify a clear discrimination. Nevertheless, with some cell extracts, an inhibition halo slightly larger than that of the control was observed but only related to Pm. aleophilum. No supernatant showed antifungal activity except H. uvarum against A. niger (Figure 8). Finally, whole cells inhibited the molds in most cases, which is consistent with previous results except for A. niger which was tested with H. osmophila.
On the other hand, enzymatic activity such as in pectinolytic enzymes and chitinase was studied. The tests were carried out to know if the yeasts were able to degrade polygalacturonic acid and chitin. For both activities, the presence of a hydrolysis halo around the colony was considered a positive result; nevertheless, chitinolytic and pectinolytic activities were not observed in the yeasts assayed in the conditions tested.

Bioaccumulation of heavy metals
For bioremediation proposals, a selective elimination of metals using yeasts combined with other processes could be a feasible strategy.
Metallic ion determination was performed by means of an inductively coupled plasma optical emission spectrometer (ICP-OES: Varian Vista-Pro, Mulgrave, VIC, Australia). Tests were semiquantitative.
Very different results were obtained depending on the yeast species as well as the metal tested for the bioaccumulation experiment ( Table 4). The greatest metal elimination took place for Pb (II) with H. meyeri, Z. bailii, P. membranaefaciens, P. kudriavzevii and S'codes ludwigii, which presented an elimination range of around 20%, reaching 30% in some cases.
This percentage diminished by nearly half for Cd (II), with P. kudriavzevii having produced the highest elimination, followed by Z. fermentati.
Cr (VI) was eliminated in a much lower proportion, highlighting only P. membranaefaciens with 10% elimination, followed by the majority of the yeasts in which adsorption was not detected or was very low.
In general, the metal removal was instantaneous, and during the first 10 min of contact, no additional adsorption was observed. However, in some cases, S'codes ludwigii for Pb (II) and H. uvarum for Cd (II), the adsorption was progressive, possibly due to the different compositions of polysaccharides and proteins in the cell wall [31]. Unfortunately, S. cerevisiae, a by-product of the wine industry and suitable for this type of process, offered a low percentage of elimination for Pb (II) and a medium percentage for the other two metals compared with the rest of the yeasts of the same group. Appreciable desorption processes were not observed, although P. kudriavzevii released Cr (VI) into the media after 6 h of contact.

Conclusions
This initial study of yeast populations isolated from very old distilleries reflects the great existing biodiversity of this valuable yeast niche. This contrasts with what occurs in wine cellars, where the intra and interspecific variability of yeasts have been reduced drastically due to the starter use. Saccharomyces, Pichia and Candida are the genera found in large proportions. Some species were only isolated for certain substrates, like T. delbrueckii in sweet piquettes and P. galeiformis in fermented piquettes.
The yeast biota of these environments is varied, so these ecological niches are microbial reserves of undoubted biotechnological interest.
In fact, a great number of thermophilic Saccharomyces strains with a great cell vitality were found to have potential use as starters in distillery plants.
On the other hand, yeasts coming from very old distilleries might be used as biocontrol and bioremediation agents. Pichia sp. inhibited all molds effectively and might be produced in an aerated fermentation process and used as an antifungal postharvest treatment of fruits. In the case of S'codes ludwigii, P. membranaefaciens and P. kudriavzevii, the elimination of Pb (II) was achieved, with the adsorption being almost instantaneous.
P. kudriavzevii is a good candidate for both biocontrol and bioremediation because it efficiently inhibited molds and had the highest accumulation average of the tested metals.

Conflict of interest
None of the authors have any conflict of interest with respect to the material contained in this manuscript.