Genetic, Physiological, and Industrial Aspects of the Fructophilic Non- Saccharomyces Yeast Species, Starmerella bacillaris

: Starmerella bacillaris (synonym Candida zemplinina ) is a non- Saccharomyces yeast species, frequently found in enological ecosystems. Peculiar aspects of the genetics and metabolism of this yeast species, as well as potential industrial applications of isolated indigenous S. bacillaris strains worldwide, have recently been explored. In this review, we summarize relevant observations from studies conducted on standard laboratory and indigenous isolated S. bacillaris strains.

S. bacillaris has been described as an acidogenic, fructophilic, psychrotolerant, and highly osmotolerant microorganism [9,23,24]. The great genetic biodiversity found in isolates from different environments and the vast physiological diversity encountered in metabolic characterizations have contributed to interest in using this species as an industrial co-starter in fermented beverage production [7,8,25,26]. S. bacillaris has been characterized as a safe microorganism, which may encourage to its use as a biocontrol agent of several food pathogens [27].
In this review article, we discuss recent results from research into the genetics and genomics, ecology, population genetics and geographical biodiversity, metabolism, and ethanol production, as well as the biocontrol potential of S. bacillaris.

Starmerella/Wickerhamiella Clade
S. bacillaris has been recognized as part of the W/S (Wickerhamiella and Starmerella) clade that branches close to Yarrowia lipolytica in the Saccharomycotina species tree [28]. A detailed phylogenetic tree of the W/S clade has recently been published [28]. Members of this clade are characterized by limited nutritional versatility, an unusually small cell size, and a strong association with the floral niche (i.e., flowers and insects that visit flowers) [28]. The Starmerella genera forms a well-supported subclade, with species usually presenting fermentative capacity, a trait normally absent in Wickerhamiella [28]. According to a comprehensive phylogenetic tree based on ITS1 sequences of 3942 fungal species, this Starmerella subclade does not include other enological non-Saccharomyces yeast species, suggesting a strong genetic divergence from the large wine yeast group [29]. The specific features present in Starmerella versus Wickerhamiella subclades probably result from the remodeling of important fluxes in central carbon metabolism as well as the reinstatement of other metabolic pathways [28]. Thus, the genera Starmerella harbors an unusually large number of genes of alien origin, which were shown to reconstruct the fermentative pathway in S. bacillaris [30], as well as other metabolic pathways, as shown in Starmerella bombicola [31,32]. For example, most of the genes in the thiamine salvage pathway in distinct subclades within the W/S lineage were originally acquired from bacteria by either horizontal gene transfer (HGT) or horizontal operon transfer (HOT) events [32]. In fact, S. bacillaris species lacks both THI5 and THI4 genes, required for de novo thiamine synthesis, but harbors the salvage pathway bacterial genes, THI6 and THI20, which allow the assimilation of thiamine derivatives from the environment [32]. Other potential HGT events in S. bacillaris populations, evolving in alternative non-conventional enological (i.e., non-V. vinifera; [11]), floral, and high sugar ecosystems, remain to be explored.

S. bacillaris Ploidy and Microsatellite Loci
S. bacillaris has shown no evidence of sporulation ability [9], and it is considered haploid [38]. Interestingly, allele characterization for microsatellite loci CZ15 and CZ59 in S. bacillaris strain 11-6 suggested apparent heterozygosis for both loci [38]. A recent detailed analysis of S. bacillaris microsatellite loci CZ15 and CZ59, however, highlighted the molecular basis for the observed apparent heterozygosity [39]. The study, which in- cluded the analysis of the 10 S. bacillaris reference polymorphic microsatellite loci (i.e., CZ1, CZ4, CZ11, CZ13, CZ15, CZ20, CZ33, CZ45, CZ54, and CZ59; [38]), revealed a higher degree of structural complexity than expected from previous descriptions of yeast microsatellite loci [40][41][42][43][44]. In fact, alleles of alternative S. bacillaris microsatellite loci contain, in addition to the expected, and/or sometimes absent, variable lengths at their internal tandem-repeated motifs (TRM), extensive variations consisting of additional SNPs and/or insertions/deletions (indels), largely contributing to allelic variations [39]. In the case of loci CZ11 and CZ59, these alternative non-TRM sequences may explain the observed apparent heterozygosity in certain strains [39]. Interestingly, extension of these studies to microsatellite loci of S. cerevisiae and other common enological non-Saccharomyces species (i.e., Brettanomyces bruxellensis, Hanseniaspora uvarum, Meyerozyma guilliermondii, Saccharomyces uvarum, and Torulaspora delbrueckii) showed the existence of similar sequence and structural variants, potentially contributing to allele diversity [39]. These studies indicated that allele sizing of TRM polymorphic yeast microsatellite loci using PCR, although valid for strain differentiation and population genetic studies, does not necessarily score the number of units at their TRM [39]. Moreover, sequence analysis of microsatellite loci alleles could be used in evolutionary and phylogeny studies of yeast species [39].

S. bacillaris Species and Strain Identification
In addition to microsatellite loci, a large repertory of molecular strategies has been used to identify S. bacillaris isolates worldwide ( Figure 1). Among these, restriction fragment length polymorphism (RFLP) analyses of 5.8S-ITS regions (Table 1) enable the identification of S. bacillaris isolates to species level, as well as differentiation of this yeast species from the close species, Candida stellata (i.e., DraI and MboI enzymes) [33] and Starmerella bombicola [45] (Table 1). Other molecular strategies have been used to characterize S. bacillaris to strain level, including SAU-PCR, RAPD-PCR, micro/minisatellites, and Rep-PCR, as well as AFLP-fingerprinting, mt-DNA-RFLP, and TRtRNA-PCR analyses [8,16,17,23,37,38,46]. Fingerprinting analyses, using SAU-PCR and Rep-PCR, have enabled recognition of genetic similarity between isolates from different sources [17,23] (Figure 1). Similarly, RAPD-PCR and SAU-PCR analyses showed a relative genetic homogeneity within Italian strains, with no differences in terms of strain clustering or geographic distribution [8]. The combination of different molecular strategies (i.e., polymorphic minisatellite loci, RAPD-PCR fingerprinting, and microsatellite primer (GTG) 5 analyses) had a marked impact in population genetics analyses in S. bacillaris [16,46] (Figure 1). In addition, in situ fluorescence hybridization (FISH), targeting rRNA, has been optimized and validated as a culture-independent technique to monitor and identify S. bacillaris in biological samples [47].

S. bacillaris Ecology
A detailed intraspecific genotype analysis in a large number of indigenous isolates of S. bacillaris, using 10 informative microsatellite loci, revealed a high degree of genetic heterogeneity [38] (Figure 1). In this study, genotypic characterization of 157 strains from various enological regions (i.e., 28 vineyards/wineries of France, Greece, Hungary, Italy, Spain, Switzerland, and New Zealand), as well as 6 strains from nature, revealed that populations isolated from winemaking environments are quite diverse [38]. Interestingly, neither clonal-like behavior nor specific genetic signatures were associated with strains isolated from different vineyards and wineries, the genetic diversity of S. bacillaris strains being shaped by geographical localization [38]. A further study involving the same 163 S. bacillaris strains, plus 127 strains isolated from V. vinifera and V. labrusca ecosystems of.
Argentina (Colonia Caroya, Córdoba) [11,12] and Portugal (Azores Archipelago) [4,5], reinforced the impact of geographic localization on S. bacillaris genetic population structure. This study also showed that Argentinian S. bacillaris populations are more differentiated from European populations than S. bacillaris populations within Europe [37]. In addition, no evidence of genetic differentiation based on the Vitis species or vintages, nor an evolving S. bacillaris population during alcoholic fermentation was found [37]. Overall, no genetic signature of S. bacillaris strains was found associated with different vintages, Vitis species, vineyards, and/or wineries, indicating that winemaking-related factors (i.e., Vitis species, vintage, alcoholic fermentation, and/or wineries) do not impact S. bacillaris population structure [37,38]. Thus, S. bacillaris is not under selective pressure in winemaking environments, representing an interesting model of a non-domesticated ubiquitous wine yeast species [37].

S. bacillaris Physiology
S. bacillaris grows as ellipsoid to elongated (2.2-3.0 mm × 3.0-5.2 mm) cells, which divide by multilateral budding [9] (Figure 2). Indigenous strains of this yeast species have been isolated worldwide, from grapes and grape musts, using the general yeast growth media YPD-agar (Figure 3a), the differential media WL-nutrient-agar (Figure 3b), and the selective media YPD agar, supplemented with cycloheximide (Figure 3d-f) and/or lysine-agar [20,48]. In standard YPD-agar media, S. bacillaris form small, white, creamy, shiny colonies (Figure 3a), while similar colonies, but green with a white peripheral halo, develop in WL-nutrient agar media (Figure 3b,c), which enables it to be differentiated from other non-Saccharomyces species (Figure 3c). Enological species of the Hanseniaspora genera (i.e., H. opuntiae, H. osmophila, H. uvarum, and H. vineae) also form green colonies in WL-nutrient agar, although these are larger and flat (see Figure 3c). Argentina (Colonia Caroya, Córdoba) [11,12] and Portugal (Azores Archipelago) [4,5], reinforced the impact of geographic localization on S. bacillaris genetic population structure. This study also showed that Argentinian S. bacillaris populations are more differentiated from European populations than S. bacillaris populations within Europe [37]. In addition, no evidence of genetic differentiation based on the Vitis species or vintages, nor an evolving S. bacillaris population during alcoholic fermentation was found [37]. Overall, no genetic signature of S. bacillaris strains was found associated with different vintages, Vitis species, vineyards, and/or wineries, indicating that winemaking-related factors (i.e., Vitis species, vintage, alcoholic fermentation, and/or wineries) do not impact S. bacillaris population structure [37,38]. Thus, S. bacillaris is not under selective pressure in winemaking environments, representing an interesting model of a non-domesticated ubiquitous wine yeast species [37].

S. bacillaris Physiology
S. bacillaris grows as ellipsoid to elongated (2.2-3.0 mm × 3.0-5.2 mm) cells, which divide by multilateral budding [9] (Figure 2). Indigenous strains of this yeast species have been isolated worldwide, from grapes and grape musts, using the general yeast growth media YPD-agar (Figure 3a), the differential media WL-nutrient-agar (Figure 3b), and the selective media YPD agar, supplemented with cycloheximide (Figure 3d-f) and/or lysineagar [20,48]. In standard YPD-agar media, S. bacillaris form small, white, creamy, shiny colonies (Figure 3a), while similar colonies, but green with a white peripheral halo, develop in WL-nutrient agar media (Figure 3b,c), which enables it to be differentiated from other non-Saccharomyces species (Figure 3c). Enological species of the Hanseniaspora genera (i.e., H. opuntiae, H. osmophila, H. uvarum, and H. vineae) also form green colonies in WLnutrient agar, although these are larger and flat (see Figure 3c).  S. bacillaris can ferment glucose, sucrose, fructose, and raffinose, but not galactose, maltose, and lactose [9]. It shows a marked preference for fructose over glucose when both sugars are present simultaneously [7,10,25,50,51]. This fructophilic character is also associated with other yeast species found in high sugar environments (e.g., Candida apicola, Candida magnoliae, Candida versatilis, S. bombicola, Zygosaccharomyces bailii, and Zygosaccharomyces rouxii) [10,30,52]. Genetic evidence indicates that the fructophilic character of members of the W/S clade is dependent on a specific, low affinity, high capacity fructose transporter named "Ffz1" [52]. In silico analyses of S. bacillaris draft genome sequences showed the presence of two distinct FFZ1 genes at a distance of approximately 4 kb from each other [52]. These transporters (i.e., Ffz1a and Ffz1b) proved to enable growth on fructose and mannose when expressed as sole hexose transporters in a S. cerevisiae hxt-null mutant strain [52]. Kinetic parameters of these two transporters revealed that they are not functionally identical: Ffz1a more closely resembles the Ffz1 transporter from Z. rouxii, Fermentation 2021, 7, 87 6 of 15 which seems to be indispensable for fructophily [53], while Ffz1b supports weaker growth on fructose and supports growth on mannose [52]. S. bacillaris can ferment glucose, sucrose, fructose, and raffinose, but not galactose, maltose, and lactose [9]. It shows a marked preference for fructose over glucose when both sugars are present simultaneously [7,10,25,50,51]. This fructophilic character is also associated with other yeast species found in high sugar environments (e.g., Candida apicola, Candida magnoliae, Candida versatilis, S. bombicola, Zygosaccharomyces bailii, and Zygosaccharomyces rouxii) [10,30,52]. Genetic evidence indicates that the fructophilic character of members of the W/S clade is dependent on a specific, low affinity, high capacity fructose transporter named "Ffz1" [52]. In silico analyses of S. bacillaris draft genome sequences showed the presence of two distinct FFZ1 genes at a distance of approximately 4 kb from each other [52]. These transporters (i.e., Ffz1a and Ffz1b) proved to enable growth on fructose and mannose when expressed as sole hexose transporters in a S. cerevisiae hxt-null mutant strain [52]. Kinetic parameters of these two transporters revealed that they are not functionally identical: Ffz1a more closely resembles the Ffz1 transporter from Z. rouxii, which seems to be indispensable for fructophily [53], while Ffz1b supports weaker growth on fructose and supports growth on mannose [52].
An interesting finding concerning fermentation in members of the W/S clade is the apparent absence of a typical pyruvate decarboxylase (PDC) enzyme [30]. In fact, decarboxylation of pyruvate to acetaldehyde is a key step in the alcoholic fermentative pathway, catalyzed by Pdc1 in S. cerevisiae [54][55][56]. Although orthologs of this gene appear to be absent in W/S-clade genomes, a modification of specificity in the enzyme Aro10, allowing this enzyme to accept pyruvate in addition to phenylpyruvate as a substrate, appears to be involved in the remodeling of alcoholic fermentation in W/S clade yeasts [30]. Furthermore, phylogenetic and kinetic analyses of putative alcohol dehydrogenase proteins, . "Multi-species" reconstitution experiment of common non-Saccharomyces yeasts (Hanseniaspora uvarum, Pichia membranifaciens, Metschnikowia pulcherrima, and Torulaspora delbrueckii), S. cerevisiae (strain EC1118), and C. zemplinina (strain CBS 9494) (c). The photography shows the particular morphology and color aspects, on WL-nutrient agar, of the various plated yeast species [49]. Bar in (c) represents 1 mm. Identical magnification for all colonies is shown. S. bacillaris (C. zemplinina CBS 9494; Cz) is resistant to cycloheximide (d-f), a phenotype that allows selective recognition of indigenous non-Saccharomyces compared to cycloheximide-sensitive S. cerevisiae strains (Sc) [20].
An interesting finding concerning fermentation in members of the W/S clade is the apparent absence of a typical pyruvate decarboxylase (PDC) enzyme [30]. In fact, decarboxylation of pyruvate to acetaldehyde is a key step in the alcoholic fermentative pathway, catalyzed by Pdc1 in S. cerevisiae [54][55][56]. Although orthologs of this gene appear to be absent in W/S-clade genomes, a modification of specificity in the enzyme Aro10, allowing this enzyme to accept pyruvate in addition to phenylpyruvate as a substrate, appears to be involved in the remodeling of alcoholic fermentation in W/S clade yeasts [30]. Furthermore, phylogenetic and kinetic analyses of putative alcohol dehydrogenase proteins, Adh1 and Adh6, present in the genomes of W/S-clade species, revealed that the corresponding ADH1 and ADH6 genes seem to have been horizontally transferred from bacteria [30]. Ethanol production in members of the W/S clade, conducted by alcohol dehydrogenases of bacterial origin, allow W/S clade species to maintain redox homeostasis (i.e., NAD+ regeneration) when growing under anaerobic conditions [30]. S. bacillaris, in particular, harbors one copy of an ADH1 xenolog; the ADH6 xenolog was apparently duplicated several times, as this yeast species harbors four ADH6 paralogs [30].
S. bacillaris uses higher sugar quantities (i.e., up to 40 g/L) than S. cerevisiae to produce 1% (v/v) ethanol [7,23,50,57]. This low ethanol yield, in addition to its low acetic acid production, of S. bacillaris compared to S. cerevisiae, reveals low activity of the acetaldehyde pathway in S. bacillaris, leading to a redistribution in the fluxes of the central carbon metabolism network [58]. Differently from other members of the W/S clade, which convert fructose directly into mannitol [31], S. bacillaris overproduces glycerol to maintain the NADH/NAD+ redox balance in the cells [31,58]. Thus, high glycerol levels are frequently reported for fermentations involving this yeast species [17,[58][59][60][61]. Interestingly, different levels of glycerol production may be associated with alternative alleles of the GPP1 gene, encoding glycerol-3-phosphate phosphatase [29].
A reduced lag phase has been observed in S. bacillaris strains growing under low nitrogen conditions, suggesting a limited nitrogen requirement of this yeast species [51]. A preferential uptake of ammonium, tryptophan, and arginine, versus other poorly assimilated amino acids, has been observed in fermenting S. bacillaris cells [62]. The consumption of nitrogen sources by S. bacillaris revealed the strong inability of this species to take up most amino acids in the presence of ammonium [62]. However, nitrogen provided as ammonium versus a mixture of amino acids showed that organic nitrogen compounds supported more efficiently the growth of S. bacillaris [62].
Extracellular enzymes produced by S. bacillaris include β-glucosidase [50], proteases [26,50], and chitinases [26]. No pectinase, xylanase, lipase, or cellulase activities have yet been reported. The production of these enzymes may be finally dependent on the analyzed strain, the composition of the growing media, and/or the growth conditions [50]. For specific fermentation processes, it may be important to perform a detailed characterization of the extracellular enzymes secreted by S. bacillaris, to determine the final chemical profile of wines and/or to use this species as a biocontrol agent [26,27].
Subtle but significant differences have been observed for the various metabolic fermentative traits of S. bacillaris strains [16,17,50]. Fermentation vigor, tolerance to ethanol and acetic acid, and H 2 S production have been reported as more diverse than ethanol production [16,17,50]. In addition, these differences are affected by abiotic (e.g., nutrient availability, pH, oxygen levels, and temperature) and biotic (e.g., initial cell density and presence of other yeast species) factors [16,63]. Thus, as a warning, genetic similarities found among strains, following genotypic characterization, do not necessarily imply physiological similarities in S. bacillaris, and this should be taken into consideration when analyzing genotypic and phenotypic profile correlations [16,17].

S. bacillaris as a Co-Starter in Grape Must Fermentations
In mixed alcoholic fermentations, S. bacillaris preferentially consume fructose, providing the evolving S. cerevisiae cell population with the use of glucose at both middle and later fermentation stages [60,73]. Thus, co-inoculation of S. bacillaris and S. cerevisiae strains can result in complete fermentation of the major sugars present in musts [59,63,73]. Moreover, mixed fermentations with S. bacillaris can also alleviate osmotic stress for the prevailing S. cerevisiae cells, improving fermentation kinetics and reducing acetic acid production [8,65,73,74]. S. bacillaris strains can also reduce the final contents of malic acid in wines [6,72,75,76]. This phenomenon appears to be dependent on the use by S. bacillaris of malic acid and/or on the stimulation of the malolactic activity of O. oeni, thus playing an indirect role in driving malolactic fermentation [75]. It has also been reported, however, that the inoculation of a S. bacillaris strain inhibited malolactic fermentation, possibly by the presence of inhibitory compounds that negatively affected the yeast-bacteria interaction [77].
S. bacillaris can normally maintain relatively high cell population levels up to the middle [25,60,78] or even to the final [73,79] stages of fermentation. This may have negative and/or antagonistic consequences of S. bacillaris in S. cerevisiae growth. In fact, mixed inoculations of S. cerevisiae with S. bacillaris can lead to a reduction in maximum S. cerevisiae  [59,60,63,80]. This could be related to a decrease in nutrient concentrations in the must [25,60,78]. S. bacillaris death in mixed inoculated fermentations has also been investigated. Englezos et al. [79] demonstrated that high ethanol concentrations (~11.4% v/v) did not influence viability loss of S. bacillaris. What is more, it has been shown that S. bacillaris strains were able to grow at ethanol concentrations as high as 14% v/v, which could contribute to the successful implantation and good performance of this species during fermentations [10,17,50]. Other S. bacillaris isolates were reported to have low tolerance to alcohol levels (up to 5% v/v ethanol) [11,12]. Because the production of ethanol and other toxic metabolites by S. cerevisiae (such as killer toxins, SO 2 , and short-to medium-chain fatty acids) have not resulted in a negative co-existence of S. bacillaris' populations [11,78], cell-to-cell contact mechanisms may be associated with S. bacillaris cell death [78]. Finally, it should be stressed that all reported S. bacillaris and S. cerevisiae interactions could be strain-specific more than species-specific [25,59].

S. bacillaris and the Reduction in Ethanol Levels in Wines
Due to its low ethanol yield, S. bacillaris is a promising yeast species to reduce ethanol contents in wines [25,50,57]. In recent years, for social, industrial, marketing, and healthassociated reasons [81], there has been an increasing interest in reducing the final ethanol concentration of wines. With this aim, different technological and microbiological approaches, including the use of non-Saccharomyces starters, have been proposed [17,26,61,72]. Mixed culture fermentations of S. bacillaris and S. cerevisiae are normally differentiated from S. cerevisiae pure culture fermentations because of poor ethanol yields and high glycerol contents [25,50,51,59]. Interestingly, glycerol levels higher than 15 g/L have a positive effect on wine quality and sensory perception by contributing to wine structure and body perception [58].

Chemical Complexity of S. bacillaris and S. cerevisiae Fermented Beverages
Different complexities of fermented products can be obtained when performing single versus combined S. cerevisiae and/or S. bacillaris fermentations. Several authors have shown that, in laboratory scale fermentations, the concentrations of some aromatic compounds decline when using S. bacillaris and S. cerevisiae co-inoculums versus S. cerevisiae monocultures (Table 2) [27,58,60,63,65,82]. Other authors, however, have found an increase in volatile compounds when using mixed S. bacillaris and S. cerevisiae versus single yeast species fermentations [59]. S. bacillaris has been reported to overproduce compounds such as H 2 S, acetoin, ethyl acetate, and terpenes, which may have a negative impact on the wine organoleptic profile [23,49,61,63]. The apparent contradictory results regarding the sensory characteristics and chemical complexity of these studies could be dependent on: (i) the use of different S. bacillaris and S. cerevisiae strains, (ii) the inoculation procedures (i.e., simultaneous or sequential inoculations), (iii) the fermentation conditions (i.e., inoculum density, temperature, SO 2 , nitrogen and ethanol levels), and/or (iv) the grape must varieties analyzed [63,80,83] (Table 2). In some cases, these outcomes could be the consequence of either negative metabolic or synergistic interactions between S. bacillaris and S. cerevisiae strains [60,63,83]. Thus, a well-characterized set of co-starter strains and a proper design of the co-fermentations are essential factors to enhance or reduce the presence of particular metabolites [61,64,70]. Under these ideal co-fermentation conditions, the final wines would mimic the organoleptic profile of beverages obtained by spontaneous fermentations, where the local and/or regional sensorial identity of wines is enhanced [61,65,68].
Concerning the production of acetate and ethyl fatty esters, contradictory results have been obtained when using S. bacillaris in mixed fermentations (Table 2). In fact, mixed inoculations may result in increased [60,65,82,83] or reduced [58,63] overall levels of the various esters analyzed. Similar results were observed when analyzing higher alcohols. In these studies, S. bacillaris has been associated with either increased levels of total [82,83] and specific [58,59] or reduced [27,65] overall levels of these compounds (Table 2). Genome comparisons between two S. bacillaris strains (i.e., PAS13 and FRI751) and S. cerevisiae strain EC1118 revealed that the S. bacillaris branched-chain amino acid aminotransferase (BCAT) enzyme was strongly divergent from that of S. cerevisiae [29]. These differences in BCAT enzymes could influence valine, leucine and isoleucine degradation, and potentially the corresponding higher alcohol productions [29]. In the case of other aromatic compounds, like terpenes and C-13 norisoprenoids, their presence and relative concentration levels are related to the fermentation matrix (i.e., fruits and/or fruit varietals) and strain tested [27,58,63,83]. Again, either an increase or no change in their presence and relative levels was observed in S. cerevisiae and S. bacillaris mixed fermentations using alternative grape varietals and fruits and/or yeast strains ( Table 2).
Production of H 2 S by S. bacillaris seems to also be variable and strain-specific. Different authors have reported high [17], medium [12,17,23], and low [23,84] H 2 S production from S. bacillaris strains. In some studies, the temperature of the fermentations has been considered, because production of H 2 S seems to increase at higher temperatures [50]. Mixed inoculations using S. cerevisiae and S. bacillaris result in wines with higher levels of sulfur compounds [74]. Other undesirable compounds, like volatile fatty acids, showed a reduction in mixed fermentations with S. cerevisiae and S. bacillaris [58,82,83]. Low production of acetic acid by S. bacillaris strains, either in pure or in mixed fermentations, has been reported [6,17,23,25,51,60,61,63,74]. Other S. bacillaris strains, however, have been shown to produce relatively high levels of acetic acid [50].

Biocontrol Potential of S. bacillaris
S. bacillaris strains have been studied as a safe and eco-friendly method to control several diseases affecting fruit crops and their associated products [26,27]. Lemos Junior et al. (2020) reported the absence of pathogenicity factors for human health of S. bacillaris strains, including growth at 37 • C, pseudohyphae formation, invasive growth, and proteolytic activity, which guarantees that these strains do not represent a risk for human health [27]. Even when S. bacillaris has been associated with table grape sour rot [85], the use of S. bacillaris may control several fungal diseases and may also present a potential positive impact on subsequent fermentations [26,27].
The biocontrol activity of selected S. bacillaris strains has been studied against the gray mold disease agent, Botrytis cinerea, in apples and grapes [26,27]. These studies showed that the possible antifungal mode of action of this species is volatile organic compound (VOC) production, which, in turn, present inhibitory effects both in vivo and in vitro [26,27]. VOCs are suggested as the main compounds responsible for the reduction in fungal radial mycelial growth and B. cinerea gray mold decay [26,27], possibly due to the antimicrobial action of benzyl alcohol. In addition, Alternaria alternata grape infections and toxin production [86] have been successfully controlled by the use of S. bacillaris strains. The biocontrol of A. alternata could be the result of S. bacillaris' ability to colonize wound sites, which implies competitive mechanisms [86].

Conclusions
Starmerella bacillaris (syn., C. zemplinina) is a fructophilic non-Saccharomyces yeast species ubiquitously present in grapes, grape musts, and flowers. Surprising recent findings concerning the genetic diversity and metabolism of S. bacillaris have positioned this yeast species as an important model microorganism for evolutionary and metabolic studies, as well as the potential industrial and biocontrol uses. Detailed population genetic analyses of the S. bacillaris species, and comparative genomic studies in the genera Starmerella, have revealed the rich diversity of S. bacillaris worldwide, as well as the existence of complex HGT events that have exquisitely redesigned some metabolic pathways. These observations open a path for further studies on ecological and evolutionary aspects of the metabolism in S. bacillaris. Finally, the selection of unique, enologically advantageous S. bacillaris co-starter strains showing desired fermentation profiles would contribute to satisfy winemakers' and the consumer's expectations.