Comparison of the Glycolytic and Alcoholic Fermentation Pathways of Hanseniaspora vineae with Saccharomyces cerevisiae Wine Yeasts

Hanseniaspora species can be isolated from grapes and grape musts, but after the initiation of spontaneous fermentation, they are displaced by Saccharomyces cerevisiae. Hanseniaspora vineae is particularly valuable since this species improves the flavour of wines and has an increased capacity to ferment relative to other apiculate yeasts. Genomic, transcriptomic, and metabolomic studies in H. vineae have enhanced our understanding of its potential utility within the wine industry. Here, we compared gene sequences of 12 glycolytic and fermentation pathway enzymes from five sequenced Hanseniaspora species and S. cerevisiae with the corresponding enzymes encoded within the two sequenced H. vineae genomes. Increased levels of protein similarity were observed for enzymes of H. vineae and S. cerevisiae, relative to the remaining Hanseniaspora species. Key differences between H. vineae and H. uvarum pyruvate kinase enzymes might explain observed differences in fermentative capacity. Further, the presence of eight putative alcohol dehydrogenases, invertase activity, and sulfite tolerance are distinctive characteristics of H. vineae, compared to other Hanseniaspora species. The definition of two clear technological groups within the Hanseniaspora genus is discussed within the slow and fast evolution concept framework previously discovered in these apiculate yeasts.


Introduction
One of the main characteristics of yeast affecting their oenological use is their capacity to ferment sugars. Non-Saccharomyces yeasts have traditionally been considered bad fermenters. For that reason, selected strains of S. cerevisiae have been used in the oenological industry to ensure that complete fermentation occurs [1,2]. Recently, however, wineries have been encouraged to apply new, non-Saccharomyces species in winemaking processes to provide distinguishable flavours within wines [3][4][5][6]. Non-Saccharomyces species have been used to produce different aromas and flavours, compared with Saccharomyces strains [7,8]. Therefore, many efforts have been made to identify non-conventional yeast strains for oenological purposes [4,8,9].
The selection of oenological yeasts is commonly accomplished by identifying species from raw material. Spontaneously fermented grape musts are the niche that is most commonly used to identify novel strains that are both capable of fermenting sugars and confer desirable flavours to wines [10,11]. Hanseniaspora is the most abundant genus on grapes and grape juices. Studies have shown that up to 75% of the yeast population during the early stages of fermentation is made up of Hanseniaspora species [12,13]. After the first 48-72 h of spontaneous fermentation, the percentage of

Fermentation in Natural Grape Must
Chardonnay grape must containing 300 mg N/L and 200 g/L of sugars at pH 3.5 was treated with 200 mg/L dimethyldicarbonate to prevent microorganism growth. Pre-cultures of H. vineae T02/19AF, H. vineae T02/05AF, H. uvarum AWRI1280, H. osmophila AWRI3579, and S. cerevisiae ALG804 were isolated from the Chardonnay grape must and incubated at 25 • C for 12 h in a rotary shaker at 150 rpm. Then, 125-mL Erlenmeyer flasks closed with cotton plugs used to simulate microaerobic conditions were inoculated with 75 mL of must containing 1 × 10 5 cells/mL. Static batch fermentations were conducted at 20 • C to simulate winemaking conditions.

Growth Kinetics in Different Types of Media
Six types of growth media were prepared using yeast nitrogen base (YNB) (Difco, Detroit, MI, USA) as a sole nitrogen source (6.7 g/L). Media were supplemented with the following carbon sources: Glucose, fructose, sucrose, xylose, glycerol, and maltose (2% w/w). YNB that lacked a carbon source was used as a negative control.
Chardonnay must used in the fermentation analysis was also used to measure the growth kinetics of yeast strains tested. Moreover, synthetic media that mimicked must fermentations at pH adjusted to 3.5 and ethanol concentrations of either 5% or 10% were used (20 g/L glucose, 4 g/L tartaric acid; 0.134 g/L sodium acetate; 5 g/L glycerol; and 1.7 g/L YNB) (v/v).
Pre-cultures of H. vineae T02/19AF, H. vineae T02/05AF, H. uvarum AWRI1280, H. osmophila AWRI3579, and S. cerevisiae ALG804 were prepared in yeast extract peptone dextrose (YPD) media (1% yeast extract and 2% peptone, 2% glucose) via incubation for 12 h in a rotary shaker at 150 rpm and 25 • C. These pre-cultures were used to inoculate fermentations carried out in microtitler plates at a final volume of 250 µL. Inoculates producing 1 × 10 5 cells/mL in media were used for all strains and treatments. All conditions were tested in triplicate. Absorbance at 620 nm was measured at 30-min intervals for 48 h at 25 • C using an automatic plate reader (Tecan, Männedorf, Switzerland) and data were acquired with the Magellan software for further statistical analyses.

Fermentation Ability in Different Carbon Sources
The carbohydrate fermentation capacity was tested using Durham tubes immersed in media to detect gas production. Each type of medium tested was inoculated to produce a final concentration of 1 × 10 6 cells/mL in a final volume of 8 mL performed in triplicate. Results were visually assessed after a 48 and 96 h static incubation period at 28 • C.

Genomic Analysis
Genomic DNA was obtained from H. vineae cultures grown in a YPD medium at 30 • C using the Wizard Genomic DNA Purification Kit (Promega, NY, USA), according to the manufacturer's instructions.
The Illumina Genome Analyzer Iix platform in paired end mode was used to perform genomic sequencing as described previously [22]. Gene prediction was carried out using Augustus [30] trained with S. cerevisiae gene models. Peptide predictions were then annotated using BLASTp (cutoff for e-value 1 −10 ) against S. cerevisiae proteins, obtained from the Saccharomyces Genome Database [31].
A dendrogram was constructed using the sequences of nine genes encoding components of pathways related to glycolysis and fermentation from the Hanseniaspora species and S. cerevisiae. The genes assessed were CDC19, FBA1, PGI1, PFK1, PFK2, HXK2, ENO1, PGK1, and PDC1. Schyzosaccharomyces pombe was used as an external group. Neighbour joining and Kimura 2-parameter methods were carried out using the MEGA version 4 software [32,33].

Transcriptomic Analysis
Fermentations were performed in triplicate using chemically defined grape (CDG) must with a composition similar to that of natural grape juice, but devoid of grape precursors. Components of CDG must were defined as described in Carrau et al. [34], with modifications. Briefly, glucose and fructose were added in equimolar concentrations until a total sugar concentration of 200 g/L was reached. Vitamins and salts were added as previously described [35]. Yeast available nitrogen (YAN) content was adjusted to 100 mg N/L. Of this total, 50 mg N/L corresponded to amino acids and 50 mg N/L corresponded to diammonium phosphate (DAP) supplementation, as described previously [35]. The pH of the media was adjusted to 3.5 using HCl and a final concentration of 10 mg/L ergosterol was the only lipid provided.
Pre-cultures of H. vineae T02/19AF were prepared in a CDG medium and incubated 12 h in a rotary shaker at 150 rpm and 25 • C. The pre-cultures were subsequently used to inoculate fermentation reactions carried out in 250 mL Erlenmeyer flasks that were closed with cotton plugs to simulate microaerobic conditions. For all strains, fermentations were performed using 125 mL CDG and an inoculum to produce 1 × 10 5 cells/mL in the final medium. Static batch fermentations were conducted at 20 • C to simulate winemaking conditions.
Wine samples for transcriptomic analyses were taken during the fermentation process at day 1 (exponential growth), day 4 (end of exponential phase), and day 10 (stationary phase of fermentation). For transcriptomic studies, total RNA obtained from H. vineae T02/19AF isolated from three replicates sampled from three different fermentation stages (days 1, 4, and 10) were analysed independently. The nine samples were paired-end sequenced using Illumina MySeq. Trinity software was used to assemble raw reads from transcriptomic analyses and further statistical analyses were performed as specified by Giorello et al. [22].

Statistical Analysis
All the treatments were performed in triplicate and the statistical error was calculated as the standard deviation of all data analysed. To compare growth and fermentation kinetics, variance comparison was performed by the ANOVA test carried out with STATISTICA 7.0 software. Differences in the mean absorbance or weight loss were evaluated using the Tukey test.

Fermentative Capacity of H. vineae in Different Media
Hanseniaspora species used a limited number of carbon sources, which may have been related to the reduced competitiveness of the species throughout fermentations [27]. Regarding growth in different carbon sources ( Figure 1A), growth of all the Hanseniaspora strains tested on both glucose and fructose had kinetics similar to that of S. cerevisiae ALG804. The media supplemented with sucrose was fermented by S. cerevisiae in a similar manner as that of media containing simple hexose. H. uvarum AWRI1280, however, did not grow on media containing sucrose. H. vineae T02/05AF, H. vineae T02/19AF, and H. osmophila AWRI3579 were able to grow on and ferment sucrose to an extent. Invertase gene (SUC2) is present in the genome of H. vineae. SUC2 is highly expressed on day 4 of fermentation reactions, but not day 1 or 10. However, other invertase homologs were not observed in the genomes of any other Hanseniaspora species except H. osmophila [25]. Recently, Steenwyk et al. [36] grouped H. vineae and H. osmophila within the slower-evolving linage of Hanseniaspora. In this branch, the SUC2 gene is present. This is different in species of the fast evolving linage including H. uvarum, H. opuntiae, H. valbyensis, and H. guilliermondii, which might have lost the gene as a result of rapid mutation rates [36]. The same fact was detected with another key gene that show Saccharomyces wine yeast adaptations. Increased sulfite tolerance conferred by SSU1 (Table 2) is present in H. vineae and H. osmophila and it is absent in the other Hanseniaspora species. The presence of SUC2 and SSU1 genes are indicators of adaptations to alcoholic fermentation in yeast [37].
Glycerol was not used as a unique carbon source for H. vineae in accordance with data reported by Albertin et al. [38]. However, HvGUT1 and HvGUT2 genes were present in the genomes of both H. vineae strains analysed. In addition, xylose was not used by the H. vineae strains as expected. A finding that was likely due to the lack of enzymes needed to carry out the xylose conversion. The group of genes were also determined to be absent in H. guilliermondii, H. uvarum, and H. opuntiae [27]. However, H. vineae T02/05AF and T02/19AF have the ability to grow weakly when maltose is provided as a sole carbon source, despite the fact that they were not able to ferment the sugar. The same behaviour was also observed for H. jakobsenii [36]. different carbon sources ( Figure 1A), growth of all the Hanseniaspora strains tested on both glucose and fructose had kinetics similar to that of S. cerevisiae ALG804. The media supplemented with sucrose was fermented by S. cerevisiae in a similar manner as that of media containing simple hexose. H. uvarum AWRI1280, however, did not grow on media containing sucrose. H. vineae T02/05AF, H. vineae T02/19AF, and H. osmophila AWRI3579 were able to grow on and ferment sucrose to an extent. Invertase gene (SUC2) is present in the genome of H. vineae. SUC2 is highly expressed on day 4 of fermentation reactions, but not day 1 or 10. However, other invertase homologs were not observed in the genomes of any other Hanseniaspora species except H. osmophila [25]. Recently, Steenwyk et al. [36] grouped H. vineae and H. osmophila within the slower-evolving linage of Hanseniaspora. In this branch, the SUC2 gene is present. This is different in species of the fast evolving linage including H. uvarum, H. opuntiae, H. valbyensis, and H. guilliermondii, which might have lost the gene as a result of rapid mutation rates [36]. The same fact was detected with another key gene that show Saccharomyces wine yeast adaptations. Increased sulfite tolerance conferred by SSU1 (Table 2) is present in H. vineae and H. osmophila and it is absent in the other Hanseniaspora species. The presence of SUC2 and SSU1 genes are indicators of adaptations to alcoholic fermentation in yeast [37].
Glycerol was not used as a unique carbon source for H. vineae in accordance with data reported by Albertin et al. [38]. However, HvGUT1 and HvGUT2 genes were present in the genomes of both H. vineae strains analysed. In addition, xylose was not used by the H. vineae strains as expected. A finding that was likely due to the lack of enzymes needed to carry out the xylose conversion. The group of genes were also determined to be absent in H. guilliermondii, H. uvarum, and H. opuntiae [27]. However, H. vineae T02/05AF and T02/19AF have the ability to grow weakly when maltose is provided as a sole carbon source, despite the fact that they were not able to ferment the sugar. The same behaviour was also observed for H. jakobsenii [36]. and the capacity of the species ferment (F) when six different carbon sources (2% w/w) were provided. Yeast nitrogen base (YNB) that lacked a carbohydrate was used as a negative control. Black filled circles indicate that full growth and and the capacity of the species ferment (F) when six different carbon sources (2% w/w) were provided. Yeast nitrogen base (YNB) that lacked a carbohydrate was used as a negative control. Black filled circles indicate that full growth and fermentation were observed, grey circles indicate the moderate capacity of species to grow and ferment, and white circles indicate that the species was not able to grow or ferment. (B) Growth kinetics of Hanseniaspora spp. and S. cerevisiae on the Chardonnay grape juice measure as increased absorbance over a period of 48 h. (C) Fermentation kinetics of the three strains in the Chardonnay grape juice after 12 days are shown. Growth and fermentation experiments were performed using independent triplicate samples and error bars express standard deviation. As expected, Saccharomyces was able to grow and ferment sugars faster than Hanseniaspora species and significant differences between the species occurred after 16 h. The growth kinetics of the three Hanseniaspora strains tested were similar on grape must ( Figure 1B), however fermentation kinetics of H. vineae and H. osmophila revealed that these species consume sugars significantly faster than H. uvarum ( Figure 1C).

Sugar Transport
The transport of sugars into the cytosol of cells is a key step of the glycolytic pathway. S. cerevisiae is able to detect extracellular nutrients and make metabolic adjustments that rapidly facilitate the use of extracellular compounds [39].
Of the multiple sensors described in S. cerevisiae, H. vineae possessed the following genes HvSNF3, HvGPA2, HvGPR1, and HvASC1, which were determined to be associated with the hexose sensing capacity of both T02/19AF and T02/05AF strains (Table 2). ScSNF3 encodes a low glucose sensor present in the plasma membrane that is involved in the regulation of glucose transport and also has the capacity to sense fructose and mannose in S. cerevisiae. Expression of the gene in H. vineae increases throughout fermentation ( Figure 1). ScGPA2, ScGPR1, and ScASC1 are hexose sensors that have been reported to be necessary for fermentation and are part of the "fermentome" in S. cerevisiae. Deletion of the genes was previously reported to induce protracted fermentation [28]. HvGPA2 and HvGPR1 have similar expression patterns throughout the fermentation process. The genes are maximally expressed on day 4 and their expression levels decrease at day 10. On the other hand, HvASC1 is most highly expressed on the first day of fermentation and levels were drastically reduced both on day 4 and 10 relative to day 1 (Figure 2).
S. cerevisiae possesses 20 sequences putatively associated with the hexose transport [40]. H. guilliermondii UTAD222 possess 22 sugar transporters, and based on their DNA sequences, ten were predicted to be associated with the hexose transport, all of them were most similar to HXT2 [27]. A comparison of sugar transporters of both sequenced strains of H. vineae with S. cerevisiae revealed that T02/05AF had two copies of the HvHXT6 gene and one copy of HvHXT1. Sequences homologous to ScHXT2 were not found in the species. However, strain T02/19AF was determined to have a single copy of HvHXT6. Expression levels of the gene increased after day 10 of fermentation. No sequences homologous to HXT1 were identified. ScHXT1 is a low affinity hexose and pentose transmembrane transporter and is paralogous to ScHXT6 [41]. The ScHXT6 gene encodes a high affinity hexose transmembrane transporter that transports glucose, fructose, and mannose [42]. Tondoni et al. [43] revealed that in S. cerevisiae and Torulaspora delbrueckii, HXT6 is most highly expressed throughout late stages of fermentation ( Figure 2). In addition, both T02/05AF and T02/19AF strains have one sequence that is homologous to the S. cerevisiae ScHXT7 gene. ScHXT7 is a high-affinity glucose transporter that is very similar to ScHXT6 [42,44]. This gene is maximally expressed at day 4 at the end of the exponential phase of fermentation in H. vineae. Both H. vineae strains sequenced lacked polyol transporters (such as ScHXT13, ScHXT17, or ScHXT16 of S. cerevisiae) needed for the uptake of sorbitol and mannitol. S. cerevisiae possesses 20 sequences putatively associated with the hexose transport [40]. H. guilliermondii UTAD222 possess 22 sugar transporters, and based on their DNA sequences, ten were predicted to be associated with the hexose transport, all of them were most similar to HXT2 [27]. A comparison of sugar transporters of both sequenced strains of H. vineae with S. cerevisiae revealed that T02/05AF had two copies of the HvHXT6 gene and one copy of HvHXT1. Sequences homologous to ScHXT2 were not found in the species. However, strain T02/19AF was determined to have a single copy of HvHXT6. Expression levels of the gene increased after day 10 of fermentation. No sequences homologous to HXT1 were identified. ScHXT1 is a low affinity hexose and pentose transmembrane transporter and is paralogous to ScHXT6 [41]. The ScHXT6 gene encodes a high affinity hexose transmembrane transporter that transports glucose, fructose, and mannose [42]. Tondoni et al. [43] revealed that in S. cerevisiae and Torulaspora delbrueckii, HXT6 is most highly expressed throughout late stages of fermentation (Figure 2). In addition, both T02/05AF and T02/19AF strains have one sequence that is homologous to the S. cerevisiae ScHXT7 gene. ScHXT7 is a high-affinity glucose transporter that is very similar to ScHXT6 [42,44]. This gene is maximally expressed at day 4 at the end of the exponential phase of fermentation in H. vineae. Both H. vineae strains sequenced lacked polyol transporters (such as ScHXT13, ScHXT17, or ScHXT16 of S. cerevisiae) needed for the uptake of sorbitol and mannitol.
Other homologs of sugar transporters have been found in H. vineae. Three tandem copies of HvSTL1 were identified in both T02/19AF and T02/05AF strains that shared homology with a glycerol proton symporter of the plasma membrane, which has been shown to be inactivated in response to glucose in S. cerevisiae [45]. Other Hanseniaspora species sequenced, such as H. osmophila, H. opuntiae,  H. guilliermondii, H. uvarum, and H. valbyensis also possessed between two and four copies of the gene. Other homologs of sugar transporters have been found in H. vineae. Three tandem copies of HvSTL1 were identified in both T02/19AF and T02/05AF strains that shared homology with a glycerol proton symporter of the plasma membrane, which has been shown to be inactivated in response to glucose in S. cerevisiae [45]. Other Hanseniaspora species sequenced, such as H. osmophila, H. opuntiae, H. guilliermondii, H. uvarum, and H. valbyensis also possessed between two and four copies of the gene. According to the transcriptomic analyses, just one of the copies identified was differentially expressed throughout fermentation in H. vineae (Figure 2). HvFPS1, a putative plasma membrane channel involved in glycerol and xylitol movement, is present in the genome of both T02/19AF and T02/05AF. Expression of the gene is elevated near the beginning of fermentation reactions (days 1 and 4) and decreased at day 10 ( Figure 1). One copy of HvGUP1 was present in each strain analyzed as well as HvJEN1. Moreover, it was suggested that ScGUP1 participates in glycerol transport and ScJEN1 mediates the high-affinity uptake of lactate, pyruvate, and acetate so that they can be used as carbon sources in S. cerevisiae [46,47].

Glycolytic Pathway in H. vineae Strains
The first enzyme of the glycolysis pathway is a hexokinase (Figure 3). ScHXK2 phosphorylates glucose in the cytosol. In S. cerevisiae, this isoform is principally responsible for glucose activation, which is needed to initiate glycolysis when glucose is provided as a carbon source and inhibits ScHXK1 [41]. However, in H. vineae, HvHXK2 was the only enzyme identified with putative hexokinase activity, amino acid homology was higher compared to other species of this genus.  Predicted amino acid sequences of phosphoglucose isomerase (PGI1) from H. vineae and H. osmophila were 86% similar to that of S. cerevisiae. Predicted PGI1 amino acid sequences from H. uvarum, H. valbyensis, H. guilliermondii, and H. opuntiae were approximately 71% similar to S. cerevisiae. This tetrameric enzyme is involved in the interconversion of glucose-6-phosphate and fructose-6phosphate. Phosphoglucose isomerase activity has also been associated with the regulation of the cell cycle and gluconeogenic events of sporulation in S. cerevisiae [49,50]. Phosphofructokinase activity was determined to be the second-most important glycolytic enzyme. The enzyme determines fermentation capacity and is indispensable for anaerobic growth. In S. cerevisiae, the enzyme is composed of two alpha and beta subunits that are encoded by ScPFK1 and ScPFK2, respectively. Hanseniaspora strains possess sequences homologous to both ScPFK1 and ScPFK2 subunits and similar to S. cerevisiae, the subunits form a hetero-octameric complex [29]. Protein Fermentation 2020, 6, 78 9 of 17 sequences of both HvPFK1 and HvPFK2 were most similar to S. cerevisiae (76.78% and 79.24%) and H. osmophila (76.46% and 77.50%) relative to the other Hanseniaspora species assessed ( Figure 4A). Phosphofructokinase only works in the forward direction and is not involved in gluconeogenesis. In fact, three activities are required for gluconeogenesis: Pyruvate carboxylase, phosphoenolpyruvate carboxykinase, and fructose-1,6-bisphosphatase. No genes encoding the key gluconeogenic enzymes have been identified in H. vineae, H. guilliermondii, H. uvarum, H. osmophila, or H. valbyensis [27]. This explains why Hanseniaspora species are not able to grow when non-carbohydrate precursors such as pyruvate, amino acids, or glycerol are provided as energy sources. This is different than S. cerevisiae, which is able to grow on a variety of carbon sources including ethanol and lactate [48]. Predicted amino acid sequences of phosphoglucose isomerase (PGI1) from H. vineae and H. osmophila were 86% similar to that of S. cerevisiae. Predicted PGI1 amino acid sequences from H. uvarum, H. valbyensis, H. guilliermondii, and H. opuntiae were approximately 71% similar to S. cerevisiae. This tetrameric enzyme is involved in the interconversion of glucose-6-phosphate and fructose-6-phosphate. Phosphoglucose isomerase activity has also been associated with the regulation of the cell cycle and gluconeogenic events of sporulation in S. cerevisiae [49,50].
Two copies of the S. cerevisiae ScENO1 gene that encodes an enolase were identified in H. osmophila, while only one copy was identified in other sequenced Hanseniaspora species. This enzyme catalyzes the conversion of 2-phosphoglycerate to phosphoenolpyruvate during glycolysis. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is a tetramer that catalyzes the conversion of glyceraldehyde-3-phosphate to 1,3 bis-phosphoglycerate. Three unlinked genes, ScTDH1, ScTDH2, and ScTDH3, encode related, but not identical, polypeptides that form catalytically active homotetramers with different specific glyceraldehyde 3-phosphate dehydrogenase activities in S. cerevisiae [51,52]. In H. vineae, only TDH2 and TDH3 homologues have been identified, and both were differentially expressed throughout fermentation ( Figure 4C).
H. vineae strains T02/19AF and T02/05AF possess HvGUT1 and HvGUT2 genes. Both ScGUT1 and ScGUT2 are associated with glycerol kinase activities in the cytoplasm and mitochondria, respectively. Glycerol degradation is a two-step process that is mediated by GUT1 and/or GUT2. Under aerobic conditions, S. cerevisiae is able to utilize glycerol as a sole carbon and energy source [53]. Both of the enzymes have homologs that have been identified in H. vineae and H. osmophila, other Hanseniaspora species such as H. uvarum, H. guilliermondii, and H. opuntiae lack homologous of these genes [27].
Several specific activities associated with glycolytic enzymes of S. cerevisiae and H. uvarum have high degrees of similarity, which highlights the general conservation of glycolytic pathways and the downstream reactions involved in ethanol production [29]. Pyruvate kinase is a key enzyme that catalyses an irreversible step of the glycolytic pathway. The position of the enzyme at the branchpoint between fermentation and respiration makes it a key determinant energy metabolism [54]. Recent work revealed that the pyruvate kinase activity enhanced the capacity of S. cerevisiae to ferment sugars versus H. uvarum [29]. The predicted proteins, Cdc19p, of H. vineae and H. osmophila are more homolog to the corresponding Cdc19p of S. cerevisiae than those of H. uvarum and other Hanseniaspora species ( Figure 4A). When residues of the catalytic domain of ScCdc19p [55] are compared with those of H. uvarum and H. vineae, only one amino acid difference was identified; Asp 265 was substituted with Gly 269 in H. uvarum and H. vineae ( Figure 4B). However, in the binding site of the allosteric activator, fructose 1,6-bisphosphate, two amino acid differences between H. uvarum and S. cerevisiae and one between H. vineae and S. cerevisiae were identified. The two differences identified between H. uvarum and S. cerevisiae are at the same positions ( Figure 4B) as those identified in the PYK2 gene of S. cerevisiae, a paralog of CDC19 that is characterized by its low pyruvate kinase activity compared with the pyruvate kinase protein encoded by CDC19 (formerly PYK1) [54].
Expression levels of 13 H. vineae genes involved in the glycolytic pathway mainly decreased from day 1 to day 4 of fermentation and were maintained throughout the stationary phase ( Figure 4C). This finding is in agreement to previous observations in S. cerevisiae [43]. However, levels of HvTDH2 expression remained high at both days 1 and 4 and decreased expression levels were observed at day 10. Additionally, expression of HvGUT2 peaked at days 4 and 10, and increased expression levels of the gene were not detected at day 1. Finally, HvGPM2 was not expressed under the conditions assessed.

Alcoholic Fermentation in H. vineae Strains
The pyruvate decarboxylase activity plays a key role in the alcoholic fermentation pathway. Three different pyruvate decarboxylase isozymes have been identified in the genome of S. cerevisiae: ScPDC1, ScPDC5, and ScPDC6. The function of pyruvate decarboxylase is the degradation of pyruvate into acetaldehyde and carbon dioxide. The enzyme is responsible for transferring the final product of glycolysis (pyruvate) to ethanol production [56]. In H. vineae, no sequences homologous to ScPDC5 and ScPDC6 were found and HvPDC1 was the only pyruvate decarboxylase isozyme identified in the species. In S. cerevisiae, ScPDC1 was strongly expressed in fermenting cells. The enzyme is conserved among yeast, bacteria, and plants. It is regulated by glucose and ethanol concentrations and also by itself [57]. The active enzyme has a homotetrameric structure and the enzyme has two known cofactors: Thiamin diphosphate (ThDP) and Mg 2+ [58][59][60]. In H. vineae, genes involved in thiamine biosynthesis have not been identified and a similar finding was also reported in H. guilliermondii [27] and most other Hanseniaspora species [36]. It has been suggested that this may contribute to the low alcoholic fermentative capacity of Hanseniaspora species, the phenotype has been shown to be related to the weak pyruvate kinase activity of H. uvarum [29]. S. cerevisiae genes associated with thiamine production are upregulated in the stationary phase of growth. Oenological strains with improved expression levels of the genes have corresponding elevated rates of fermentation [61]. This phenomenon may result from vitamin depletion that occurs after the exponential phase.
Alcohol dehydrogenases, which catalyse the conversion of acetaldehyde to ethanol are key fermentative enzymes. Many alcohol dehydrogenases have been identified in S. cerevisiae including ScADH1, ScADH2, ScADH3, ScADH4, ScADH6, and ScADH7. Many homologues of S. cerevisiae alcohol dehydrogenases have been found in the H. vineae genome. H. vineae has the same number of copies of the genes as S. cerevisiae. Eight alcohol dehydrogenase genes are present in H. vineae species, compared to six in H. osmophila, and four in other sequenced species of Hanseniaspora such us H. uvarum, H. guilliermondii, H. valbyensis, and H. opuntiae. This may explain the improved adaptation of H. vineae to alcohol fermentation relative to other Hanseniaspora. It is noteworthy that of the eight HvADH sequences found in the genome of H. vineae, at least three HvADH6 genes are encoded in tandem. Increased copies of the gene may be associated with increased fermentation capacity, indicating that the alcohol dehydrogenase activity might be a key feature of alcoholic fermentation adaptations [62]. H. vineae has an enhanced tolerance to ethanol ( Figure 5B) versus H. uvarum and H. osmophila, which are unable to grow in media containing 10% ethanol.
H. vineae and H. osmophila genes encoding putative alcohol dehydrogenases were grouped in two main clusters that contained either ADH1, ADH2 and ADH3 or ADH6 and ADH7 ( Figure 5A), this is in agreement with the two multigenic families reported by Giorello et al. [22]. The clusters were formed according to the clustal alignment of predicted protein sequences, however, regarding adscription by a single homology with S. cerevisiae ADHs in the databases [22] produced some discrepancies. Therefore, HvADH6 homologs from H. vineae and H. osmophila were removed from the HvADH6 and HvADH7 cluster. Moreover, the HvADH1 homologous sequence of H. vineae is grouped in the cluster of ScADH6 and ScADH7.
HvADH genes display different expression patterns ( Figure 5C). Two of four paralogous copies of HvADH6 were not differentially expressed at the time points analysed. Expression of one copy of ADH6 significantly declined between days 1 and 4 of fermentation. In addition, the expression of one copy of ADH3 was elevated on day 4 relative to day 1 ( Figure 5C). These behaviours are similar to those of aryl alcohol dehydrogenases that facilitate the production of increased levels of alcohol by S. cerevisiae [63]. Therefore, HvADHs may be important for reducing levels of fusel aldehydes by producing increased levels of alcohol in H. vineae [22]. H. vineae and H. osmophila genes encoding putative alcohol dehydrogenases were grouped in two main clusters that contained either ADH1, ADH2 and ADH3 or ADH6 and ADH7 ( Figure 5A), this is in agreement with the two multigenic families reported by Giorello et al. [22]. The clusters were formed according to the clustal alignment of predicted protein sequences, however, regarding adscription by a single homology with S. cerevisiae ADHs in the databases [22] produced some discrepancies. Therefore, HvADH6 homologs from H. vineae and H. osmophila were removed from the A B C

Hanseniaspora Genus as an Evolution Model for Alcoholic Fermentation Adaptations
The glycolytic potential of two strains of H. vineae were analysed using genetic, transcriptomic, and phenotypic data. Results explained the good performance of the species with respect to fermenting wine [7,21]. Findings also showed that the H. vineae behaviour was similar to traditional S. cerevisiae strains used in winemaking. Due to the outstanding capacity of H. vineae to produce aromatic metabolites, it was necessary to compare the capacities of the H. vineae strains to produce ethanol with S. cerevisiae. The high degree of similarity between glycolytic and alcoholic fermentation enzymes of H. vineae and H. osmophila with S. cerevisiae showed that the two species should be classified as fermenters, while the remaining Hanseniaspora species assessed were adapted to the fruit niche and were correspondingly included in the fruit group. In our experience, H. vineae strains cannot be isolated from the fresh grape fruits [19]. A dendrogram of concatenated DNA sequences from seven glycolytic and fermentation genes ( Figure 6) indicated the presence of two clades of Hanseniaspora species, similar to findings of Steenwyk et al. [36] determined using genes from the DNA repair processes present within the genus. Interestingly, the fruit and fermentation clades shown in Figure 6 were correlated with the slow and fast evolution lineages defined by these authors. Branches were in agreement with phylogenetic classifications that were based on ribosomal genes [19]. It might be interesting to use the group as an evolution model to determine the mechanism by which the fermentation group diverged separately from the fruit group [36], giving less species diversity probably due to slow evolution mechanisms. Further work will be needed to understand whether the process might be an example of domestication, as has been proposed for S. cerevisiae wine and beer strains [64].
Previous studies have compared the fermentation capacity of two species belonging to the fruit group: H. guillermondi and H.uvarum [27,29], and the work presented here is the first assessment of a member of the fermentation group of Hanseniaspora.
The glycolytic potential of two strains of H. vineae were analysed using genetic, transcriptomic, and phenotypic data. Results explained the good performance of the species with respect to fermenting wine [7,21]. Findings also showed that the H. vineae behaviour was similar to traditional S. cerevisiae strains used in winemaking. Due to the outstanding capacity of H. vineae to produce aromatic metabolites, it was necessary to compare the capacities of the H. vineae strains to produce ethanol with S. cerevisiae. The high degree of similarity between glycolytic and alcoholic fermentation enzymes of H. vineae and H. osmophila with S. cerevisiae showed that the two species should be classified as fermenters, while the remaining Hanseniaspora species assessed were adapted to the fruit niche and were correspondingly included in the fruit group. In our experience, H. vineae strains cannot be isolated from the fresh grape fruits [19]. A dendrogram of concatenated DNA sequences from seven glycolytic and fermentation genes ( Figure 6) indicated the presence of two clades of Hanseniaspora species, similar to findings of Steenwyk et al. [36] determined using genes from the DNA repair processes present within the genus. Interestingly, the fruit and fermentation clades shown in Figure 6 were correlated with the slow and fast evolution lineages defined by these authors. Branches were in agreement with phylogenetic classifications that were based on ribosomal genes [19]. It might be interesting to use the group as an evolution model to determine the mechanism by which the fermentation group diverged separately from the fruit group [36], giving less species diversity probably due to slow evolution mechanisms. Further work will be needed to understand whether the process might be an example of domestication, as has been proposed for S. cerevisiae wine and beer strains [64].
Previous studies have compared the fermentation capacity of two species belonging to the fruit group: H. guillermondi and H.uvarum [27,29], and the work presented here is the first assessment of a member of the fermentation group of Hanseniaspora.

Conclusions
The results suggest that H. vineae is clearly better adapted to the fermentation niche compared to what we named as the Hanseniaspora fruit clade. These results are in agreement with a separately evolution divergence between the two clades of the genus Hanseniaspora as was proposed previously. Phenotypic behavior of H. vineae growth, ethanol tolerance, and fermentation kinetics are in agreement with the genetic and transcriptomic data provided. The results obtained demonstrate that H. vineae and a genetically closely related species, H. osmophila, behave similarly. Homologies of glycolytic and alcoholic fermentation enzyme sequences of both species were compared to S. cerevisiae, and the similarities observed allowed the differentiation of H. uvarum from H. osmophila and H. vineae. High sequence homology in these latter two species was observed for key genes involved in glycolysis such as HXK2, which encodes hexokinase, PFK1/PFK2 subunits of phosphofructokinase, and CDC19 that encodes pyruvate kinase. This homology could explain the improved fermentative performance observed for H. vineae compared with other Hanseniaspora species. The elevated number of copies of ADH genes in H. vineae might be associated with increased ethanol tolerance in the species. The presence of active genes typically related to wine fermentation capacities in H. vineae and H. osmophila such as sulfite tolerance (SSU1) and sucrose hydrolyzing invertase (SUC2) differentiate both species from the other sequenced species of the genus. Taken together, findings reported here support the characterization of the Hanseniaspora genus into two different groups that are adapted to two different niches, fruit and juice fermentation. These results have contributed to the improved characterization of the genus and furthermore might support the importance of it as a model for further studies related to the genetic and evolutionary phenomena of yeast domestication processes.