Utilization of Non-Saccharomyces to Address Contemporary Winemaking Challenges: Species Characteristics and Strain Diversity

Abstract

Winemaking is facing significant challenges caused by industrialization of the process, climate change, and increased consumer awareness regarding the use of chemical preservatives. Although several solutions have been proposed, the utilization of non-Saccharomyces species seems to be the most efficient one. Several non-Saccharomyces species have been employed for this purpose, with Hanseniaspora uvarum, H. vineae, Kluyveromyces marxianus, Lachancea thermotolerans, Metschnikowia pulcherrima, Pichia fermentans, P. kluyveri, Schizosaccharomyces pombe, Starmerella bacillaris, Torulaspora delbrueckii, and Wickerhamomyces anomalus being the most promising ones. However, only a restricted amount of metabolic activities can be reliably attributed to the species level, while most of them are characterized by strain variability and are also affected by the Saccharomyces cerevisiae strains used to carry out alcoholic fermentation, as well as the efficient supply of precursor molecules by the grape varieties and the conditions for their effective bioconversion. This variability necessitates the application of optimization strategies, taking into consideration all these parameters. This review article aims to assist in this direction by collecting the data referring to the winemaking practice of the most interesting non-Saccharomyces species, presenting clearly and comprehensively their most relevant features, and highlighting the effect of strain diversity.

1. Introduction

The utilization of industrialized starter cultures has transformed wine production. It removed the uncertainty and lack of reproducibility of spontaneous fermentation and introduced a controlled and predictable process. However, the loss of grape variety typicity and terroir was an unpredictable side effect; wines originating from different areas and made using different grape varieties and under different cultivation conditions share organoleptic features due to the utilization of the same starter cultures. Thus, the need to re-invent grape variety typicity and enhance differentiation between wines was evident. In addition, there is a growing consumer concern regarding the use of chemical preservatives and a demand for their reduction and the utilization of alternative methods, including bioprotection. Moreover, global warming resulted in the earlier ripening of grapes, reducing organic acid and increasing the carbohydrate content [1]. All issues can be resolved through the use of non-Saccharomyces yeasts. The occurrence of non-Saccharomyces species in the fermenting must was originally regarded as a source of spoilage, even though the production of beneficial metabolites by them was also early acknowledged [2]. Enhancement of the specific qualities of each grape variety can take place through enzymatic activities not commonly present in Saccharomyces cerevisiae strains. Among them, glycosidase activity may release compounds from their glycosidic complexes that improve wine aroma, particularly terpenes; pectinolytic activity may improve color intensity by releasing anthocyanins and phenols, and enhance optical clarity by reducing the size and concomitantly improve the solubility of the pectin polymers; acyltransferase and esterase activities may regulate the type and amount of esters; proteolytic activity may improve clarification and stabilization of wines and at the same time provide yeasts with assimilable nitrogen. In addition, non-Saccharomyces yeasts may also provide additional benefits such as enhanced glycerol production and mannoprotein release, both of which positively affect organoleptic properties [3]. Moreover, the capacity of certain non-Saccharomyces yeasts to increase acidity through the production of lactic acid and to produce metabolites with antimicrobial activity are desirable traits that ensure microbial stability of wine and balanced flavor [4,5]. The technological benefits that have been associated with specific non-Saccharomyces yeasts are shown in

Table 1. Metabolic activities of non-Saccharomyces yeast species and related technological benefits.

Yeast Species	Metabolic Activities	Technological Benefits
M. pulcherrima	Extracellular enzyme production (glycosidase, protease, esterase, pectinase, β-lyase, cellulase, xylanase) Antimicrobial compounds production (killer factor, pulcherrimic acid)	Improvement of wine stability Improvement of sensorial properties Enhancement of optical clarity Improvement of color intensity
H. uvarum	Extracellular enzyme production (glycosidase, protease, esterase, pectinase)	Improvement of wine stability Enhancement of optical clarity Improvement of sensorial properties
H. vineae	Extracellular enzyme production (glycosidase, protease) De novo production of terpenes
Kl. marxianus	Extracellular enzyme production (endo-poly galacturonases)	Enhancement of optical clarity Improvement of sensorial properties Improvement of color intensity
T. delbrueckii	Increased mannoprotein and polysaccharide release Extracellular enzyme production (glycosidase, β-lyase)
L. thermotolerans	Production of L-lactic acid	Improvement of wine stability Improvement of sensorial properties
St. bacillaris	Fructophilic Pyruvic acid production Extracellular enzyme production (glycosidase, esterase) Antifungal activity
W. anomalus	Antimicrobial compounds Extracellular enzyme production (glycosidase, esterase)
Sch. pombe	Catabolizes malic acid to ethanol Urease activity	Improvement of wine safety Improvement of sensorial properties
P. fermentans	Extracellular enzyme production (glycosidase, esterase)	Improvement of sensorial properties
P. kluyveri	Extracellular enzyme production (β-lyase)

H.: Hanseniaspora; Kl.: Kluyveromyces; L.: Lachancea; M.: Metschnikowia; P.: Pichia; Sch.: Schizosaccharomyces; St.: Starmerella; T.: Torulaspora; W.: Wickerhamomyces.

.

The capacity of several non-Saccharomyces yeast species to carry out the aforementioned activities, and many more, in laboratory media or during must fermentation, has been extensively studied. As a result, many of them are currently commercially available [6]. In nearly all cases, the ad hoc incorporation of the non-Saccharomyces yeast was accompanied by a S. cerevisiae strain able to carry out alcoholic fermentation. This intensive assessment led to the conclusion that the effectiveness of the non-Saccharomyces strains depends upon the interplay between five key factors, namely the capacity of the grape variety to provide the necessary precursors, the metabolic capacity of the non-Saccharomyces strain employed to perform specific actions, the interactions with the S. cerevisiae strain used to carry out alcoholic fermentation, the inoculation strategy and the incubation conditions. Grape variety, grape maturity stage, and agronomic practices play a decisive role in the chemical composition of the must and the amount of precursor molecules, such as carbohydrates, organic acids, and phenolic compounds, which affect wine quality [7]. The diversity among non-Saccharomyces and S. cerevisiae strains affects not only their individual metabolic activities but also their interactions, which are also governed by inoculation strategy and incubation conditions. Inoculation strategy refers to the time of inoculation of each strain and can be roughly divided into co-inoculation, when all strains are inoculated simultaneously, and sequential inoculation, in which the non-Saccharomyces yeast is inoculated first and incubated under optimized temperature and time, and then the S. cerevisiae strain is added.

The strain diversity, from a winemaking perspective, particularly of the non-Saccharomyces species, is very often underestimated as it has not been assessed within the broader context defined by the aforementioned key parameters. This creates confusion regarding the actual capacities of non-Saccharomyces yeasts at species and strain levels and complicates the interpretation of the results obtained. It is, therefore, of utmost importance to collect these data and accurately assign them to the appropriate taxonomic level. This manuscript aims to assist in understanding the true capacities of the non-Saccharomyces yeast species employed in winemaking by collecting the data referring to the winemaking practice of the most promising ones, namely Hanseniaspora uvarum, H. vineae, Kluyveromyces marxianus, Lachancea thermotolerans, Metschnikowia pulcherrima, Pichia fermentans, P. kluyveri, Schizosaccharomyces pombe, Starmerella bacillaris, Torulaspora delbrueckii, and Wickerhamomyces anomalus, present their features clearly and comprehensively, accurately attribute them to the respective taxonomic level, and highlight the effect of strain diversity on them.

2. Hanseniaspora uvarum

The interest in H. uvarum results from the range of extracellular enzymes with enological interest that it may produce. Although this capacity is strain-dependent, it seems to be more common compared to other non-Saccharomyces species. The enzymes that have dominated scientific interest are β-glycosidases, proteases, esterases, and pectinases, due to their technological significance and the ability to affect wine aroma and color. β-glycosidases hydrolyse glycosidic bonds between a sugar moiety and an aglycone, many of which possess distinguished aromas. Such compounds may be monoterpenes (e.g., linalool, citronellol), norisoprenoids (e.g., β-damascenone, vitispirane), phenylpropenes (e.g., eugenol, vanillin), as well as aliphatic compounds [8]. In addition, β-glycosidase activity may also result in loss of wine color through the hydrolysis of anthocyanins [9]. Proteolytic enzymes hydrolyze peptide bonds, resulting in the decomposition of haze-producing proteins, improving the clarity of wine [3]. Esterases are involved in ester formation and hydrolysis, regulating the balance between esters and their respective moieties. Thus, they have a very important role in the aroma profile of wines [3,10]. Finally, pectinases hydrolyze plant cell walls, facilitating the diffusion of phenolic and color compounds into the must and reducing the pectin haze [3]. The occurrence of these enzymatic activities in H. uvarum strains has been thoroughly assessed. For example, Wang et al. [11] reported that 13 out of a total of 36 strains that were examined exhibited some activity in at least three of these enzymes. One of these strains, namely GS38, exhibited significant activity in all four. Utilization of this strain in Cabernet Sauvignon must fermentation under sequential fermentation with S. cerevisiae (strain EXCELLENCE^® XR), which was inoculated 48 h after the inoculation of H. uvarum strain GS38, resulted in the production of wine with enhanced concentration of ethanol, titrable and volatile acidity, total tannins, acetate esters and terpenes, reduced concentration of higher alcohols, ethyl esters and carbonyl compounds, while no quantitative effect on total phenolic, flavonoid and anthocyanin content was recorded [11]. The strain-dependent character of this enzyme-producing capacity was also highlighted by the studies of Lai et al. [12] and Miranda et al. [13]. In the first study, H. uvarum strain Pi235 was able to produce a relatively high amount of ethyl acetate (217.48 mg/L) after seven days of fermentation of Kyoho must in monoculture but lacked β-glucosidase activity [12]. On the contrary, the Madeira wine made by H. uvarum was characterized by enhanced concentration of hydroxybenzoics, some hydroxycinnamates, namely caffeic acid, ferulic acid, sinapic acid, cis and trans-coutaric acid, trans-resveratrol, some flavan-3-ols, namely catechin and epigallocatechin, as well as flavonols, namely myricetin, quercetin, and rutin [13].

The effect of the utilization of H. uvarum strains in winemaking has been thoroughly assessed and the effect of the inoculation protocol, namely simultaneous or sequential inoculation with S. cerevisiae, the metabolic capacity of the H. uvarum and S. cerevisiae strains employed, as well as the grape variety utilized, on the classical enological characteristics and the concentration of the volatile compounds in the resulting wine, have been adequately exhibited. In the case of simultaneous fermentation, a marginal, if any, effect on ethanol production was reported [14,15,16,17]. On the contrary, the capacity of the H. uvarum strains to persist was quite different. Hanseniaspora uvarum strain Yun268, which was employed in the studies by Hu et al. [14] and Xia et al. [17], was more tolerant to ethanol, compared to H. uvarum strain S8 that was employed in the study by Lee et al. [15], as it retained a population above 5 log CFU/mL, even at an ethanol volume of approximately 10% (v/v) [14]. On the contrary, the population of H. uvarum strain S8 diminished after seven days of fermentation, when ethanol production commenced. Unfortunately, no data regarding the persistence of H. uvarum HU4487 were provided by Li et al. [16]. In the studies by Hu et al. [14] and Xia et al. [17], in which H. uvarum Yun268 was employed, different S. cerevisiae strains and different grape varieties were used. More specifically, in the study by Hu et al. [14], fermentation was performed in Ecolly and Cabernet Sauvignon musts by S. cerevisiae strain Actiflore^®, and in the study by Xia et al. [17], fermentation was performed in Italian Riesling must by S. cerevisiae Excellence^® TXL. These differences were reflected in the concentration of the volatile compounds detected. More accurately, in both studies, the increase in the concentration of ethyl esters was reported. However, in the first study, this was mostly due to the increase in ethyl octanoate and ethyl decanoate, while in the second study, it was mostly due to the increase in ethyl hexylate and ethyl laurate. An increase in volatile fatty acids was reported by Xia et al. [17] and by Hu et al. [14]; however, in the latter case, only regarding Ecolly must fermentation. In all cases, this increase was mostly attributed to the increase in octanoic acid. An increase in acetate esters was only reported by Xia et al. [17], and only when the inoculum ratio was 10:1 (H. uvarum: S. cerevisiae), and was mostly due to the enhancement of ethyl acetate concentration. Similarly, an increase in higher alcohols concentration was only reported by Xia et al. [17], and only when the inoculum ratio was 1:1 (H. uvarum: S. cerevisiae), and was mostly due to the increase in phenethyl alcohol concentration. Finally, an increase in terpene concentration was reported by Hu et al. [14] in Ecolly must, mostly due to the increase in linalool oxide and a-terpineol. Finally, the increase in the concentration of C13-norisoprenoids to Cabernet Sauvignon must [14] and Riesling must, which was evident only when the inoculum ratio was 1:1 [17], could be principally attributed to the increase in the concentration of β-damascenone. On the contrary, the effect of the S. cerevisiae strain was not highlighted in the study by Li et al. [16]. In that study, S. cerevisiae strains EC1118^TM and ZYMAFLORE^TM VL3 were separately employed, simultaneously with H. uvarum HU4487, for the fermentation of Sauvignon Blanc must, in both cases, decrease in glycerol concentration along with the increase in acetate ester, organic acids, and terpenes was reported and could be more or less attributed to the increase in the same compounds, namely ethyl acetate, isoamyl acetate, hexyl acetate, 2-phenethyl acetate, octanoic acid, and linalool. On the contrary, differences were observed in the concentration of higher alcohols since co-inoculation of the H. uvarum strain with S. cerevisiae EC1118^TM increased the concentration of butanol, 3-methylthio propanol, octanol, benzyl alcohol, 2-phenyl ethanol, and decreased the concentration of 2-methyl-1-propanol, compared to the respective when S. cerevisiae EC1118^TM was used as a monoculture, while co-inoculation of the H. uvarum strain with S. cerevisiae ZYMAFLORE^TM VL3 increased the concentration of 2-methyl-1-propanol, isoamyl alcohol, pentanol, hexanol, 4-methyl-1-pentanol and octanol, compared to the respective when S. cerevisiae ZYMAFLORE^TM VL3 was used as a monoculture [16]. Finally, the decrease in the concentration of acetate esters and ethyl esters was reported by Lee et al. [15] after simultaneous fermentation of Campbell Early must by S. cerevisiae W-3 and H. uvarum S8 inoculated at a 1:9 ratio.

The effectiveness of H. uvarum Yun268 has also been studied in sequential fermentation with S. cerevisiae strains by Hu et al. [14] and Xia et al. [17]. In these studies, S. cerevisiae strains (Actiflore^® and Excellence^® TXL, respectively), were inoculated 48 h after the inoculation of the H. uvarum strain. In these studies, the final ethanol volume presented no statistically significant differences with the respective volumes obtained by the use of the respective S. cerevisiae strains as monocultures, which concurs with the results presented by Wang et al. [18] and Boban et al. [19], who employed different H. uvarum and S. cerevisiae strains as well as different grape varieties. On the contrary, in the study by Gao et al. [20], the ethanol production after sequential fermentation of Cabernet Sauvignon must by H. uvarum BF345 followed by S. cerevisiae Vintage Red^® after 24 h was reduced to 14.97% (v/v), compared to the 15.70% (v/v) that was achieved by the use of the S. cerevisiae strain as monoculture. In sequential fermentation, the tolerance of H. uvarum Yun268 seemed to be enhanced, since it remained at a population around 7 log CFU/mL throughout Ecolly must fermentation, and persisted around the same population for six days in Cabernet Sauvignon must fermentation [14]. Similarly, the population of H. uvarum Yun268 remained around 5 log CFU/mL for 11 and 18 d in Italian Riesling must fermentation, when inoculated at 1:1 and 1:10 ratios with S. cerevisiae Excellence^® TXL (S. cerevisiae/ H. uvarum), respectively [17]. In both cases, this persistence presented a significant improvement compared to the persistence of the same strain during simultaneous inoculation under the same conditions. On the contrary, the population of H. uvarum BF345 was negatively affected during Cabernet Sauvignon and Chardonnay musts fermentation, after S. cerevisiae Vintage Red^® and Aroma White^® inoculation, and was not detectable after seven and six days, respectively. Similarly, H. uvarum Z-7 persisted for 14 days, during Marastina must fermentation, until ethanol reached approximately 8% (v/v). In that case, S. cerevisiae EC1118^TM was inoculated after six days of fermentation, when ethanol content was 2–3% (v/v) [19]. Unfortunately, no such data were provided by Wang et al. [18]. As in the case of simultaneous fermentation, the concentration of the volatile compounds detected was affected by the H. uvarum and S. cerevisiae strains as well as the grape variety employed. In the majority of the cases, the concentration of higher alcohols was not affected, and the increase in acetate esters was observed. The latter could be principally attributed to the increase in different compounds in each case, namely, ethyl acetate [14,17], phenylethyl acetate [19], and isoamyl acetate [20]. As far as the rest of the classes of chemical compounds were concerned, no clear trend could be observed.

3. Hanseniaspora vineae

Hanseniaspora vineae is a very feasible option to increase the fruity and flowery notes of wines. This is achieved through the increased production of phenylpropanoids, benzenoids, and acetate esters, particularly 2-phenylethyl acetate [21,22,23,24,25,26,27,28]. In addition, the reduced production of short and medium chain fatty acids and ethyl esters compared to S. cerevisiae [21,22,23,24,25,26,27,28], the production of acetic acid and ethyl acetate within typical wine ranges [29], the de novo production of terpenes [30,31,32] as well as the extracellular protease and β-glucosidase enzyme activities [33,34], constitute also important features of this species, which are naturally subjected to variations according to the capacity of the strains [35]. The increased formation of 2-phenylethyl acetate and phenylpropanoids has been attributed to the gene duplications of aromatic amino acid aminotransferases (ARO8 and ARO9) and phenylpyruvate decarboxylases (ARO10), and their high level of expression at the beginning of the stationary growth phase. Similarly, the enhanced formation of acetate esters has been correlated to the occurrence of six novel proteins with alcohol acetyltransferase (AATase) domains. On the contrary, the reduced production of branched-chain higher alcohols, fatty acids, and ethyl esters has been attributed to the absence of the genes encoding for branched-chain amino acid transaminases (BAT2) and acyl coenzyme A (acyl-CoA)/ethanol O-acyltransferases (EEB1) [36].

Hanseniaspora vineae may tolerate up to 10% (v/v) ethanol, a property that has been attributed to the increased number of alcohol dehydrogenase genes (ADH) that have been detected in the genome of sequenced strains [37]. Therefore, in the case of musts that may potentially reach higher alcohol levels, the presence of S. cerevisiae is necessary. Indeed, the production of wines with more than 10% (v/v) ethanol has been attributed to the occurrence of S. cerevisiae strains, either as part of the must native microbiota that has not been completely eliminated or as a contaminant from the winery environment [38]. Hanseniaspora vineae is also characterized by a reduced fermentative ability, compared to S. cerevisiae, which is reflected in the increased time that it needs to complete the fermentation. Indeed, Lleixa et al. [22] reported that more time was needed by the inoculated H. vineae to complete the fermentation of Macabeo and Merlot musts than the respective S. cerevisiae strain. However, this may not constitute a problem, since rapid fermentations are also characterized by increased cooling needs and, therefore, energy consumption, along with the loss of volatile compounds [21]. This reduced fermentative ability may be related to the reduced number of genes associated with hexose sensing, hexose transport, and pyruvate decarboxylation in the genome of H. vineae compared to the respective of S. cerevisiae [39].

In terms of competitiveness, H. vineae seems to be able to prevail over the native yeast microbiota of fermenting must, but only during the first days of fermentation, after which it is outcompeted by native yeasts and particularly native S. cerevisiae [22,40]. If persistence for more days is required, then the reduction in native yeast populations before inoculation with H. vineae is capable of providing this extra time [22]. A similar case is the case of co-inoculation with S. cerevisiae. The time during which the population of H. vineae exceeds that of S. cerevisiae is affected by initial aeration and the inoculum ratio. Yan et al. [41] reported that initial limited aeration of Colombard must sustain H. vineae population longer in the presence of S. cerevisiae. As far as the effect of the inoculum ratio was concerned, a ratio of 50:50 allowed H. vineae to dominate for only about a day [40]. An increase in the H. vineae proportion will add days of dominance [40,42]. Indeed, inoculation of Glera must with a ratio of 90:10 (H. vineae/ S. cerevisiae) and above resulted in the dominance of H. vineae even after six days of fermentation [42]. The increased percentage of H. vineae in the inoculum also resulted in the increased production of 2-phenylethyl acetate [42,43].

Based on the above, H. vineae may be used in both sequential and simultaneous fermentations with S. cerevisiae. In both cases, nutrient depletion may result in sluggish or stuck fermentations. In sequential fermentations, H. vineae may reduce the ability of S. cerevisiae to complete fermentation. On the other hand, in simultaneous fermentations, S. cerevisiae may reduce the capacity of H. vineae to grow and provide the desired additional benefits. Therefore, adequate nutrient levels should be ensured, especially nitrogen, amino acids, and vitamins [21,31,35]. Medina et al. [44] suggested that thiamine or pantothenic acid limitation is highly unlikely to result in sluggish or stuck fermentation, at least regarding the yeast strains employed. Increase in diammonium phosphate levels was consistently negatively correlated with the formation of benzyl alcohol, 2-phenylethyl alcohol, and their acetates, whereas inhibition of their production was observed at a YAN level of 250 mgN/L [31]. The effect of L-phenylalanine addition on growth and 2-phenylethyl acetate production by H. vineae has also been assessed to some extent. Zhang et al. [45] reported that the addition of L-phenylalanine had no clear effect on cell growth of H. vineae and S. cerevisiae during growth in pure cultures but enhanced the production of 2-phenylethyl acetate during sequential fermentation of chemically defined medium MS300. This was attributed to the increased accumulation of 2-phenylehtly alcohol by S. cerevisiae, which takes place through the Ehrlich pathway. However, this increased production gradually decreases with the addition of L-phenylalanine, suggesting that high levels of this amino acid may not promote 2-phenylethyl acetate production. The positive correlation between phenylalanine addition and an increase in benzenoids and phenylpropanoids during fermentation of Chardonnay musts was verified by Valera et al. [24].

4. Kluyveromyces marxianus

Kluyveromyces marxianus is a Crabtree-negative yeast species; thus, fermentation is affected only by the Pasteur effect. This could indicate poor fermentative performance; however, the occurrence of Kl. marxianus strains able to produce up to 12.52% (v/v) ethanol has been reported [46]. The property that makes this species interesting, from an enological perspective, is the production of endo-polygalacturonases that are active under the specific conditions of winemaking and significantly enhance phenolic and aroma compound extraction [47]. These properties are strain-dependent; however, endo-polygalacturonase production seems to be a common trait among the strains belonging to this species.

Nitrogen metabolism of Kl. marxianus, and particularly strain IWBT Y885 has been extensively studied, due to the effect on aroma compound production [48,49,50,51] and concomitantly the sensorial quality of the wine [52,53,54]. The nitrogen requirements and assimilation pattern of the Kl. marxianus strain was reported to be similar to the S. cerevisiae strain EC1118^TM, except for ammonium and arginine [48,50]. Regarding these cases, Kl. marxianus strain initiated growth consuming both nitrogen sources, and during growth, the contribution of arginine remained relatively stable, while the respective of ammonium decreased. Ultimately, Kl. marxianus strain depleted arginine before the maximum population was reached and assimilated only 50% of the available ammonium. On the other hand, S. cerevisiae strain first depleted ammonium and arginine, both before the maximum population was reached [50]. This could indicate a possible antagonism during must fermentation for the nitrogen sources that could lead to stuck fermentation. Indeed, sequential inoculation of these strains in synthetic medium, with S. cerevisiae strain inoculation following the respective strains of Kl. marxianus by 48 h, resulted in stuck fermentation with 48 g/L of residual sugar [49].

The effectiveness of polygalacturonase production by Kl. marxianus IWBT Y885 during cold maceration and alcoholic fermentation driven by S. cerevisiae EC1118^TM was assessed by Rollero et al. [55]. It was reported that the polygalacturonase produced by this strain was active during both. However, significant differences were reported between the compounds extracted during cold maceration and fermentation, due to the different conditions, resulting in significant qualitative and quantitative differences in the aroma compounds produced. Notably, the pectinase activity during cold maceration in the presence of the Kl. marxianus strain was higher than the respective of a commercial enzyme preparation used for the same purpose.

The effect of Kl. marxianus utilization as a starter culture on wine quality has only been marginally assessed. Barone et al. [56] provided a new insight and suggested that Kl. marxianus may have the capacity to initiate alcoholic fermentation, which can be completed by autochthonous S. cerevisiae strains. The sequential inoculation of Kl. marxianus strains followed by S. cerevisiae strains resulted in wines with enhanced isoamyl acetate and 2-phenethyl acetate and reduced isoamyl alcohol content, whereas, in the majority of the cases, hexanol and decanoic acid content remained without any statistically significant change compared to the wines fermented with the utilization of the respective S. cerevisiae strains as monocultures [51,55,56]. However, in order to reach safe conclusions, the assessment of more Kl. marxianus strain is necessary.

5. Lachancea thermotolerans

Lachancea thermotolerans (formerly Kluyveromyces thermotolerans) is a Crabtree-positive yeast species that is among the most fermentative members of the non-Saccharomyces group [57,58,59]. It is well known for its capacity to produce L-lactic acid and, thus, increase the acidity of musts and wines; a trait that is particularly important for the low-acidic musts of warm viticultural regions [60]. Taking into consideration the effect that climate change has on the cold climate viticultural regions, namely the earlier ripening of grapes that results in higher carbohydrate and lower organic acid content [1], it is reasonable to expect that this yeast will be equally important in these regions as well [61].

L-lactic acid and ethanol production are strain-dependent properties; their amount during must fermentation when L. thermotolerans has been used as a monoculture has been reported to range between 0.26 [62] and 12 g/L [63] and between 3.98 [57] and 10.9% (v/v) [64], respectively. Lactic acid production may affect the development of Oenococcus oeni and the subsequent malolactic fermentation, when applicable. Indeed, Snyder et al. [65] reported that the use of L. thermotolerans strains that produce low amounts of lactic acid allowed successful malolactic fermentation, whereas the use of L. thermotolerans strains that produce high amounts of lactic acid inhibited malolactic fermentation. The concentration of lactic acid required to inhibit malolactic fermentation seemed to be matrix-dependent and was determined at 1.5 g/L in Merlot and Sauvignon Blanc wines, while malolactic fermentation was completed in red chemically defined wine with lactic acid concentration exceeding 3 g/L.

The outcome of co-inoculation of S. cerevisiae and L. thermotolerans is strain-dependent. In most cases, co-fermentation has been reported to negatively affect the development of both yeasts, resulting in less ethanol and less lactic acid production, compared to the production by S. cerevisiae and L. thermotolerans monocultures, respectively [66,67,68,69]. On the other hand, in most cases, sequential fermentation negatively affected only the L. thermotolerans population, which presented a rapid population reduction upon S. cerevisiae inoculation, while S. cerevisiae population development was not affected. Sequential inoculation resulted in the majority of the cases in a reduction in the final ethanol volume and only in a few cases in ethanol volume without a statistically significant difference from that obtained by the same S. cerevisiae strain used as monoculture. The lactic acid produced has been reported to range from 0.2 to 8.1 g/L, which, in most cases, resulted in a reduction in the pH value. As far as the final glycerol concentration was concerned, in the majority of the cases, an increase has been reported, reaching 10.4 g/L. However, there are many cases in which the glycerol content has been reported to be equal to or lower than that obtained by the same S. cerevisiae strains used as monoculture [67,68,70,71,72,73,74,75,76,77,78,79]. Regarding the effect of S. cerevisiae inoculation time, i.e., 1, 2, 3, 4, or 6 days after L. thermotolerans inoculation, on ethanol, glycerol, and lactic acid production, the strain-dependent character seems to persist, as contradicting results have been reported [66,67].

The importance of the L. thermotolerans and S. cerevisiae strains employed in winemaking seemed to be also the case regarding the production of volatile compounds. When strains of both species were simultaneously inoculated in musts, the production of several volatile compounds, such as isobutanol, 2-phenylethanol, 1-propanol, 1-hexanol, 3-methyl-1-butanol, 1-butanol, ethyl 2-methylbutyrate, ethyl 2-butenoate, ethyl butyrate, ethyl hexanoate, 2-phenethyl acetate, hexyl acetate and diethyl succinate seemed to be largely strain-dependent, as their production, compared to the respective when the S. cerevisiae strains were used as monocultures, did not follow any particular trend. On the contrary, the increase in isobutyric acid and ethyl lactate concentration, the decrease in 1-octanol, hexanoic acid, octanoic acid, decanoic acid, ethyl octanoate, and ethyl hexanoate concentration, as well as the unchanged concentration of linalool, acetaldehyde, ethyl decanoate, isoamyl acetate, and ethyl acetate has been reported in the majority of the studies [57,67,68,69,70,71]. Similarly, when strains of L. thermotolerans and S. cerevisiae were sequentially inoculated in musts, in the majority of the studies the concentration of α-terpineol, nerolidol, β-ionone, β-damascenone, acetaldehyde, 2-nonanol, 1-propanol, 1pentanol, trans and cis-3-hexen-1-ol, 2-ethyl-1-hexanol, ethyl 3-methylbutyrate, 2-methyl-1-propanol, 1-hexanol, 3-methyl-1-butanol, 1-butanol, butyric acid, 2-methylbutyric acid, nonanoic acid, 2-methylpropionic acid, ethyl 2-methylbutyrate, ethyl nonanoate, isoamyl acetate, γ-butyrolactone, ethyl 2-furoate, ethyl 2-phenylacetate, acetophenone, dodecanal, 4-methylbenzaldehyde, methyl benzoate and benzyl alcohol remained without any statistically significant change; the concentration of geraniol, menthol, acetoin, isobutyric acid, ethyl lactate, and isoamyl lactate was increased and the concentration of isoamyl alcohol, 1-octanol, hexanoic acid, octanoic acid, ethyl 3-hydroxybutyrate, ethyl hexanoate, ethyl dodecanoate, ethyl decanoate, ethyl octanoate, isoamyl octanoate, methyl octanoate, δ-decalactone, γ-nonalactone, γ-octalactone and vanillin was decreased, compared to the respective obtained by the S. cerevisiae strains when employed as monocultures. On the contrary, no particular trend was evident for volatile compounds such as citronellol, linalool, benzaldehyde, decanal, 3-methyl butanal, 2-heptanol, 1-heptanol, isobutanol, 3-methylpentanol, 2-phenylethanol, 4-methylpentanol, 1-decanol, 2-methyl-1-butanol, isovaleric acid, isobutyric acid, propanoic acid, heptanoic acid, decanoic acid, ethyl butyrate, ethyl hexanoate, isobutyl acetate, ethyl acetate, 2-phenethyl acetate, hexyl acetate, diethyhl succinate, diethyhl glutarate, γ-decalactone, γ-hexalactone, γ-undecalactone, styrene, 4-vinylphenol, 4-vinylguaiacol, and guaiacol [19,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86].

The detrimental effect of lactic acid on malolactic fermentation, along with the possible deviations that may occur during the latter, which are triggered by musts with high pH value and may have a detrimental effect on wine quality and safety, due to the production of volatile acidity and biogenic amines, resulted in a quest for alternatives. The combination of L. thermotolerans with Sch. pombe has attracted scientific attention as one such alternative. The capacity of L. thermotolerans to produce lactic acid and direct alcoholic fermentation, along with the capacity of Sch. pombe to convert L-malic acid to ethanol, could lead to completion of must fermentation without the addition of S. cerevisiae. Indeed, Benito et al. [87] reported that the sequential fermentation of Tempranillo must by L. thermotolerans Concerto™ followed by Sch. pombe V2 led to a wine with 14.03% (v/v), which was significantly less than the 14.56% (v/v) obtained by S. cerevisiae 87, which also contained 2.96 g/L lactic acid and 0.01 g/L malic acid. As far as the concentration of acetic acid, histamine, tyramine, putrescine, and cadaverine was concerned, no statistically significant differences were observed with the wine made by S. cerevisiae 87. Similar results were obtained with the use of other strains of L. thermotolerans, Sch. pombe, and S. cerevisiae [88,89]. In addition, the sensorial evaluation revealed that the sequential use of L. thermotolerans and Sch. pombe resulted in the most preferred final product, most likely due to a more acidic and fruity character as well as the balanced mouthfeel [88,90,91]. The latter was attributed to the release of high mannose-containing polysaccharides by Sch. pombe [91]. Color differences were also observed and assigned to the differences in the anthocyanin profiles of the final products [90].

The co-inoculation of L. thermotolerans with H. vineae and H. opuntiae, aiming to acidify and improve the sensory profile of the resulting wine by introducing floral notes, also seemed a promising concept. Indeed, Vaquero et al. [92] fermented Airen must with different combinations of L. thermotolerans strains and H. opuntiae A56 or H. vineae Hv, followed by the addition of S. cerevisiae 7VA on day seven of fermentation, and verified the production of unique organoleptic profiles.

6. Metschnikowia pulcherrima

Metschnikowia pulcherrima is characterized by low to moderate fermentative power and tolerance to ethanol, which usually ranges between 3.5 and 4.8% (v/v) ethanol, glycerol production that may reach 13.2 g/L, and total and volatile acidity development that may reach 7.5 g tartaric acid/L and 0.73 g acetic acid/L, respectively [93,94].

The interesting enological traits that this species possesses are the production of a wide range of extracellular enzymes as well as metabolites with antimicrobial activity. Indeed, the production of β-glucosidases, proteases, esterases, β-lyases, cellulases, pectinases, and xylanases has been reported and exploited either by including strains of this species as adjunct cultures in winemaking or by using their growth supernatants [57,95,96,97,98,99,100,101,102,103]. Regarding the production of antimicrobial compounds, the strong antimicrobial activity against wine spoilage yeasts and fungi, such as Brettanomyces/Dekkera, Hanseniaspora, Pichia, Schizosaccharomyces, Penicillium, Aspergillus, and Fusarium, as well as against lactic acid bacteria, such as Levilactobacillus brevis, Lactiplantibacillus plantarum, and Pediococcus acidilactici, and pathogenic bacteria such as Staphylococcus aureus, Salmonella, and Escherichia coli, has been reported [94,96,104,105,106,107] and primarily attributed to the production of killer toxins and pulcherriminic acid [108]. The capacity of several M. pulcherrima strains to inhibit the development of spoilage biota, combined with the capacity to protect against browning, has been considered as a method for the reduction of sulfur dioxide use with quite promising results [101,109]. Naturally, all these properties are subjected to strain-dependent variability [93,110,111,112].

In order to exploit these properties, strains of M. pulcherrima have been utilized in winemaking, either in co-inoculation or sequential inoculation format with S. cerevisiae. Due to the low ethanol tolerance and the inability of M. pulcherrima strains to compete with S. cerevisiae strains, co-inoculation has not been extensively studied. Indeed, the M. pulcherrima population declines rapidly upon co-inoculation and diminishes after 3–5 days of fermentation. Studies employing co-inoculation concur that in the majority of the cases the concentration of acetaldehyde, acetoin, 1-hexanol, ethyl lactate, 2-phenethyl acetate, hexyl acetate and diethyl succinate does not present any statistically significant differences, the concentration of hexanoic acid, octanoic acid and decanoic acid increases and the concentration of ethyl hexanoate and ethyl decanoate decreases, compared to wines made solely by the respective S. cerevisiae strains. The rest of the volatile compounds, as well as the basic enological characteristics, i.e., ethanol and glycerol content, pH value, total and volatile acidity, presented no particular trend and seemed to be largely affected by the M. pulcherrima and S. cerevisiae strains employed [57,96,113,114]. The sequential inoculation scheme has been more thoroughly assessed. In this case, two strategies have been employed. The first one included the initial inoculation with M. pulcherrima strains and, after at least 48 h, inoculation with the S. cerevisiae strain that is able to carry out the alcoholic fermentation. In that case, the final ethanol and acetic acid content as well as total acidity of the wine presented no statistically significant differences or decrease, compared to the wines made solely by the respective S. cerevisiae strains. On the contrary, the pH value presented no statistically significant change, and glycerol content increased in the majority of the cases. Regarding volatile compounds, the concentration of linalool, 1-butanol, hexanoic acid, octanoic acid, decanoic acid, ethyl 3-methylbutyrate, ethyl lactate, ethyl butyrate, ethyl hexanoate, ethyl decanoate, ethyl octanoate, isobutyl acetate, benzyl alcohol and benzaldehyde presented no statistically significant changes, the concentration of 2-phenylethanol, 2-methyl-1-propanol, ethyl propanoate and ethyl acetate increased and the concentration of 1-hexanol decreased, in the majority of the cases [72,94,98,115,116,117,118,119,120,121,122,123,124,125,126]. Interestingly, the use of M. pulcherrima led to reduced SO₂ and medium-chain fatty acid levels at the end of alcoholic fermentation, creating conditions more favorable for malolactic fermentation compared with using S. cerevisiae alone. This effect was highly dependent on the grape variety and the bacterial strain used for malolactic fermentation [127,128].

The second strategy, which has only been marginally employed but provided very interesting results, included the inoculation of M. pulcherrima strains during pre-fermentative cold maceration, aiming at the production of the aforementioned enzymes that will collectively improve the quality of the final product as well as the production of antimicrobial compounds that will ensure bioprotection. Based on the results obtained so far, no statistically significant differences in ethanol yield, total acidity and pH value, compared to wines made without cold maceration but with the incorporation of the same M. pulcherrima and S. cerevisiae strains in sequential format and compared to wines made by cold maceration under the same conditions but without the addition of M. pulcherrima, seem to be a common trend. Regarding the production of volatile compounds, more studies are necessary in order to detect common trends, as the effect of M. pulcherrima and S. cerevisiae strains, as well as the duration of cold maceration, seems to be the source of variability [99,100,129,130,131].

7. Pichia fermentans

Pichia fermentans is among the yeasts that produce moderate amounts of ethanol, approximately 5% (v/v), low amounts of acetic acid, and large amounts of polysaccharides (up to 278 mg/mL) [132]. More importantly, P. fermentans is known for its high glucosidase and esterase activities, which release varietal odorants from their precursors and enhance wine aroma. All these properties are strain-dependent and affected by winemaking conditions [133,134].

The most studied P. fermentans strain is Z9Y-3, which was isolated from a Sichuan liquor pit and was distinguished by its high glucosidase activity. This strain was co-inoculated with S. cerevisiae (Actiflore^® F5) at different ratios in Ecolly must. In addition, the extracellular enzyme extract resulting from the development of this strain in chemically defined medium was also combined with the S. cerevisiae strain, for reasons of comparison [135]. Co-fermentation resulted in a significant increase in ethyl esters, acetates, fatty acids, and phenyl ethyls, which correlated positively with the increased inoculation ratio of the P. fermentans strain. On the other hand, the use of the extracellular enzyme extract resulted in a significant increase in varietal aroma compounds, such as terpenols, C13-norisoprenoids, and C6 compounds. The authors concluded that moderate wine aroma enhancement could be achieved using P. fermentans to S. cerevisiae ratios between 1:4 and 4:1, as increasing the P. fermentans ratio further would result in the development of negative earth odors [135]. The capacity of the extracellular extract of this strain to enhance the content of varietal and fermentative volatiles during Pinot Noir winemaking in a monsoon climate was also verified by a follow-up study by Kong et al. [136]. Co-inoculation with a high antagonistic S. cerevisiae strain resulted in higher glucosidase activities, compared to the respective obtained by co-inoculation with a low antagonistic S. cerevisiae strain and sequential inoculation with any of them. This resulted in an increase in the concentration of terpenes and C13-norisoprenoids, while C6 compounds presented no statistically significant change [137]. Especially regarding the latter, during storage of the wine made by co-fermentation with the high antagonistic S. cerevisiae strain, the hydrolysis of fruity esters was retarded by the increased amount of polysaccharides that were released by the spent yeast [138]. The improvement of the floral and fruity characteristics of the wines, as well as their persistence, was verified by sensory analysis [137,138]. In order to achieve new insights and further unravel the capacities of this strain, its genome was sequenced, assembled, and annotated [139]. Regarding the properties that could be associated with winemaking, two beta-glucosidase genes were detected, namely bglH and bglC, as well as two secondary metabolic gene clusters, a terpene one consisting of 14 genes and a non-ribosomal peptide synthetase one consisting of 23 genes. Finally, the composition and the structure of its cell wall have been studied, revealing a high concentration of hydrophilic amino acids that could contribute to the reduction in ester volatility and persistence of wine aroma over time, which has been observed [140,141].

Apart from strain Z9Y-3, the capacity of other strains has also been examined. In general, there is a limited number of studies addressing the effect of P. fermentans strains on wine aroma, in either simultaneous or sequential inoculation with S. cerevisiae strains. However, some common traits have been identified. When P. fermentans was co-inoculated with S. cerevisiae, the reduction in the final ethanol content by less than 1% (v/v) was reported in most of the cases [17,142,143,144], whereas glycerol content has been reported to increase [17], decrease [143], or remain without any statistically significant change. Regarding the volatile compounds, in most of the studies the increase in linalool, β-damascenone, octanoic acid, and decanoic acid concentration has been reported, while no statistically significant change has been observed in acetaldehyde, 1-propanol, isoamyl alcohol, ethyl acetate, and 2-phenethyl acetate concentration, compared to the respective achieved by using the S. cerevisiae strains as monocultures [17,142,143,144,145,146].

Currently, there are only two studies employing P. fermentans strains sequentially with S. cerevisiae strains. They concur only with the decrease in the ethanol content and the increase in 2,3-butanediol concentration, compared to the respective values achieved when using S. cerevisiae strains as monocultures [17,147]. Regarding the effect that utilization of P. fermentans may have on malolactic fermentation, co-inoculation with S. cerevisiae, P. fermentans H5Y-28, and Lv. brevis 26 [144] or O. oeni SD-2a [145] resulted in the induction of simultaneous alcoholic and malolactic fermentation, significantly reducing the total fermentation time.

Despite the lack of studies, the great potential of P. fermentans strains to enhance wine aroma has been highlighted. However, attention should be given to the concurrent enhancement of urea concentration, which may lead to ethyl carbamate development. Indeed, Leca et al. [148], reported that the P. fermentans strain used yielded the second-highest urea concentration among the strains studied, which reached 4.2 mg/L. However, Zhao et al. [149] reported that P. fermentans J13 possessed the capacity to hydrolyze ethyl carbamate.

8. Pichia kluyveri

Pichia kluyveri is an oxidative yeast with a rather limited fermentative capacity. Therefore, it preferentially grows on the surface of the fermenting must, resulting in the formation of a film that is characteristic of its successful development. It has not been extensively studied, compared to other non-Saccharomyces species; however, the results that are currently available indicate that this yeast may be a promising option, at least regarding certain improvements that are discussed below.

Pichia kluyveri is mostly recognized for its capacity to increase varietal thiols. Indeed, Anfang et al. [150] reported that co-fermentation of S. cerevisiae ZYMAFLORE^TM VL3 with P. kluyveri JT3.71, at an inoculum ratio of 1:9, resulted in the production of more than 1500 ng/L of 3-mercaptohexyl acetate in Sauvignon Blanc, which is a significant improvement compared to the production of approximately 500 ng/L when S. ceverisiae ZYMAFLORE^TM VL3 was used as a single culture. This increase was not simply attributed to the increase in the population of potent thiol producers, since S. cerevisiae ZYMAFLORE^TM VL3 was also a potent thiol producer, but to an interaction of yet unknown nature. Similarly, Boban et al. [19] reported that P. kluyveri Z-3 was the most potent thiol producer in Marastina wines among ten indigenous non-Saccharomyces isolates and three commercially available yeasts, namely Hyphopichia pseudoburtonii N-11, M. chrysoperlae, M. sinensis/shanxiensis, M. pulcherrima, L. thermotolerans P-25, H. uvarum Z-7, H. guillermondii N-29, H. pseudoguilliermondii V-13, Starmerella apicola VP-8, S. cerevisiae EC1118^TM, L. thermotolerans Viniflora^TM Octave, and M. pulcherrima Flavia^TM. More specifically, in monocultures of P. kluyveri Z-3 in Marastina grape must, the furfurylthiol concentration reached 1.7 μg/L, which was approximately two times higher than the concentration reached by the other yeasts included in that study.

When P. kluyveri was used in sequential or simultaneous fermentation with S. cerevisiae, apart from the increased production of thiols, the enhancement of ester production was also reported, at least in the majority of the cases [18,19,72,82,149,151]. This was usually attributed to the enhancement of the concentration of acetates of higher alcohols, such as ethyl acetate, 2-phenylethlyl acetate, isoamyl acetate, and hexyl acetate. On the other hand, the concentration of ethyl esters of fatty acids was not always enhanced [70,72,152,153]. As far as the effect of P. kluyveri on the concentration of the rest classes of chemical compounds was concerned, namely, terpenic compounds, volatile acids, C13-norisoprenoids, higher alcohols, aldehydes, ketones, volatile phenols, and fatty acids, the results that are currently available in the literature were rather contradictory. Indeed, the increase in linalool concentration was reported by Gao et al. [151], Scansani et al. [153], and Ge et al. [152], while no effect on the concentration of terpenic compounds was observed by Boban et al. [19], Hu et al. [154], Dutraive et al. [72], and Benito et al. [70]. The increase in the concentration of volatile acids was reported by Hu et al. [154] and Blanco et al. [82], while Wang et al. [18] reported no statistically significant change. Similarly, no statistically significant change in C13-norisoprenoids and aldehydes concentration was reported by Ge et al. [152]; on the contrary, Boban et al. [19] reported their increase. The case of higher alcohols was more complicated, as the increase in their concentration was observed by Boban et al. [19] and Dutraive et al. [72], the decrease in their concentration was reported by Gao et al. [151], Scansani et al. [153], Benito et al. [70], and Blanco et al. [82], while Hu et al. [154], Ge et al. [152], and Wang et al. [18] reported that their concentration was not affected. Similar was the case of fatty acids, with Blanco et al. [82] reporting the increase in their concentration, Dutraive et al. [72], Gao et al. [151], and Scansani et al. [153] reporting the decrease in their concentration, whereas Benito et al. [70], Ge et al. [152], and Wang et al. [18] reported that their concentration presented no statistically significant change. Regarding volatile phenols, the decrease in their concentration was reported by Boban et al. [19], while Ge et al. [152] observed that their concentration was not affected. Finally, Boban et al. [19] and Ge et al. [152] reported the reduction in the concentration of ketones, whereas Wang et al. [18] reported that their concentration was not affected.

Similar was the case of the classical enological parameters. In the majority of the cases, ethanol content was reported as unaffected by P. kluyveri; however, its reduction has also been reported [18,72,152]. Regarding glycerol, Dutraive et al. [72], Scansani et al. [153], and Benito et al. [70] reported the increase in its concentration, while Gao et al. [151], Blanco et al. [82], and Serafino et al. [155] reported that there was no statistically significant change. Similar was the case of the pH value; Dutraive et al. [72] and Scansani et al. [153] reported its increase, Ge et al. [152] its decrease, while Benito et al. [70], and Serafino et al. [155], reported no statistically significant change. Based on these results, it seems that further assessment is necessary in order to elucidate the effect of P. kluyveri and S. cerevisiae strains and their interactions, the grape variety, and technological practices on the production of metabolites that affect sensorial perception of wine.

The effect of simultaneous and sequential fermentation on the growth and persistence of the two yeasts has been studied to some extent. The response of P. kluyveri strains in simultaneous fermentation seemed to be uniform. Indeed, the co-inoculation of S. cerevisiae ZYMAFLORE^TM VL3 and P. kluyveri JT3.71 (1:9 initial cell ratio) [150] as well as the co-inoculation of S. cerevisiae F33 and P. kluyveri HSP14 (at approximately 6 log CFU/mL each), were followed by growth of both yeasts at comparable populations for 2–3 days, and after that, the decline of the P. kluyveri population. On the contrary, in sequential fermentation, three types of response by the P. kluyveri strains employed have been recorded. Benito et al. [70] and Serafino et al. [155] reported the immediate decline of their population (P. kluyveri strains Viniflora^® FrootZen^TM and X3-5, respectively), upon inoculation of S. cerevisiae EC1118^TM, which took place 48 h after the inoculation of P. kluyveri strains. Dutraive et al. [72] and Gao et al. [151] reported the persistence of the P. kluyveri strains Viniflora^® FrootZen^TM and DG2, for two and four days, after inoculation of the S. cerevisiae strains Level2^® and MT, which took place after 96 and 24 h, respectively. Finally, Boban et al. [19] inoculated S. cerevisiae EC1118^TM six days after the inoculation of P. kluyveri Z-7 and reported the dominance of the latter until the 14th day of fermentation, in ethanol concentration exceeding 10% (v/v), followed by a small decline of its population. In an attempt to gain insights into the interactions between these yeasts Hu et al. [154] reported that cell-to-cell contact between P. kluyveri, Viniflora^®, FrootZen™, and S. cerevisiae Viniflora^® Jazz™, during simultaneous fermentation of Pinot Noir, resulted in the significant decrease in their cell viability, which was accompanied by the decrease in the production of higher alcohols, and the increase in the production of acetate and ethyl esters along with specific amino acids, such as histidine, glycine and proline that are associated with cell growth.

The significance of mannoproteins as protective colloids in wine has been adequately highlighted [156,157,158,159,160]. Valuable insights regarding the structure of P. kluyveri call wall mannoproteins and their interactions with wine fruity esters were provided by Kong et al. [140,141]. These studies highlighted the differences in the chemical composition and structure of the mannoproteins of Pichia spp., including P. kluyveri, which directly affect the volatility of fruity esters and revealed that the complexes formed between P. kluyveri and P. fermentans mannoproteins and wine fruity esters were more stable than the respective of S. cerevisiae, which could greatly affect the organoleptic quality of wine. However, further study is still necessary in order to elucidate the nature of these interactions and their potential to diversify and improve wine aroma profiles.

9. Schizosaccharomyces pombe

Schizosaccharomyces pombe is a yeast with fermentative catabolism comparable to the respective of S. cerevisiae in terms of final ethanol volume and the level of residual carbohydrates. Additional features that make this yeast an interesting alternative to S. cerevisiae are the capacity to catabolize malic acid to ethanol and the decomposition of urea through urease activity. The first may render malolactic fermentation unnecessary, which may relieve wine from the additional biogenic amines produced, while the second minimizes ethyl carbamate formation. Both actions are highly desired, from a food safety perspective. However, the fermentation rate of Sch. pombe is usually lower, which may increase the risk of contamination, and the fermentation is usually accompanied by high volatile acidity and acetaldehyde production. Although all these metabolic features are subjected to strain variability [161,162,163], the utilization of Sch. pombe in winemaking has drawn scientific and technological attention, especially for regions of short grape growth cycles, with concomitant accumulation of high levels of malic acid, such as Northern Spain.

The aforementioned features, along with the strain variability that characterizes them, were particularly highlighted by the studies of Benito et al. [87,88,164,165], Mylona et al. [166], Del Fresno et al. [167] and Vicente et al. [89], by comparing the concentration of these metabolites after must fermentation by the Sch. pombe strains employed to the respective of the S. cerevisiae strains under the same conditions. Indeed, in these studies, the ethanol volume achieved by the Sch. pombe strains ranged between 11.72 and 14.3% (v/v) while the respective by the S. cerevisiae strains ranged between 11.63 and 14.7% (v/v). In the majority of the studies, the ethanol volume produced by the Sch. pombe strains was slightly lower than the one achieved by the S. cerevisiae strains; the greatest difference was observed in the study by del Fresno et al. [167] between Sch. pombe V1 and S. cerevisiae 3VA, the latter produced 1% (v/v) more ethanol after fermentation of Tempranillo must. Regarding the residual carbohydrates, in the wines made by Sch. pombe strains 0.54–2.13 g/L were determined, while in the respective wines made by S. cerevisiae strains 0–2.08 g/L [87,88,89,164,165,166,167]. Only in the study by del Fresno et al. [167] were statistically significant differences observed; both S. cerevisiae strains employed managed to consume nearly all carbohydrates, while the Sch. pombe strains did not. As far as the concentration of lactic, citric and malic acids were concerned, the results presented agree to the following: no statistically significant differences in the lactic and citric acids concentration were observed [87,88,89,164,165,166,167]; on the contrary, utilization of Sch. pombe strains resulted in the consumption of malic acid to a final concentration ranging between 0 and 0.51 g/L, while in the wines made by the S. cerevisiae strains the concentration of malic acid was significantly higher and ranged between 0.92 and 4.46 g/L [87,88,89,164,165,166,167]. As a result, the pH value of the wines made by Sch. pombe strains ranged between 3.46 and 4.32, which was in all cases higher than the respective values when the S. cerevisiae strains were employed, namely 3.11–3.94. Another property that was common among the Sch. pombe strains used was urea decomposition. Indeed, fermentation by Sch. pombe strains resulted in wines with urea concentration of less than 0.4 mg/L, while the respective by S. cerevisiae strains exceeded 1.22 mg/L [87,88,164,165]. Concerning the concentration of acetic acid, glycerol, and acetaldehyde, the results presented highlighted the effectiveness of proper strain selection. Indeed, acetic acid concentration ranged between 0.31 and 1.12 g/L and 0.08–0.51 g/L, glycerol concentration between 5.99 and 10.35 g/L and 5.30–9.25 g/L, and acetaldehyde concentration between 7.28 and 76.00 mg/L and 9.21–85.00 mg/L, in wines made by the Sch. pombe and the S. cerevisiae strains, respectively. Similar was also the case of pyruvic acid production [87,88,89,164,165,166,167].

As far as the production of volatile compounds was concerned, fermentation by Sch. pombe strains resulted, in the majority of the cases, in the production of reduced concentrations of acetate esters, ethyl esters, and alcohols, compared to the respective by S. cerevisiae strains [88,89,164,165,166,167]. Notable exceptions constitute the production of higher concentration of ethyl esters and alcohols by Sch. pombe 938 compared to the respective produced by S. cerevisiae Crue Blanc, during fermentation of Airen must, which can be attributed to the production of elevated concentration of ethyl lactate and 2,3-butanediol, respectively [164], and the production of higher concentration of alcohols by Sch. pombe Atecrem^TM 12H compared to the respective produced by S. cerevisiae AG006, during fermentation of Tempranillo must, which can be attributed to the enhanced production of 2-nonanol [89].

The contradictory results presented regarding the organoleptic perception of wines deacidified by Sch. pombe [168,169,170], and the capacity of S. cerevisiae, even with partial contribution, to improve sensorial quality of wines, led Benito et al. [164,165] to study the effect of simultaneous and sequential fermentation of Airen and Garnacha musts by Sch. pombe and S. cerevisiae, on the sensorial quality of the resulting wines. In the case of Garnacha musts, another case was also studied; fermentation by S. cerevisiae 796 was followed by malolactic fermentation by O. oeni Alpha^TM. In both studies, Sch. pombe 938 was employed. In both studies, when fermentation was performed by S. cerevisiae strains Cru Blanc or 796, and by Sch. pombe 938, either as monocultures or as co-cultures in simultaneous or sequential fermentation, the concentration of lactic, acetic, and citric acids presented no statistically significant differences. Co-inoculation with S. cerevisiae strains did not affect malic acid catabolism by Sch. pombe 938 and total acidity and only had a marginal negative effect on urea decomposition by Sch. pombe 938, and only when the two yeasts were inoculated simultaneously. On the contrary, pyruvic acid production was significantly affected by S. cerevisiae strains inclusion; less pyruvic acid was produced, and this production was affected by the S. cerevisiae strain and inoculation protocol. Regarding the production of volatile compounds, it was significantly affected by the S. cerevisiae strain employed, and the results presented did not reveal any particular trend. When malolactic fermentation followed the fermentation of Garnacha must by S. cerevisiae 796, the removal of malic and citric acids was more effective, the production of lactic and acetic acids, as well as acetate esters, ethyl esters, alcohols, acetaldehyde, and diacetyl, was enhanced, while no statistically significant effect on ethanol and glycerol content was recorded. Sensorial evaluation of both wines revealed that the wines made by Sch. pombe 938 presented no perceptible organoleptic problems. The Airen wines produced by mixed fermentation and the Garnacha wine made by Sch. pombe 938 as a monoculture were the most highly appreciated ones [164,165].

The deacidification of wines due to malic acid catabolism to ethanol by Sch. pombe reduces the total acidity of wines and concomitantly increases the pH value, compromising, thus, the microbiological stability of the product. Attempts to address this issue included the co-fermentation with L. thermotolerans strains that are capable of producing L-lactic acid, through sequential [87,88,89,91] or simultaneous [167] inoculation. The sequential inoculation strategy was employed by inoculating L. thermotolerans strain first and Sch. pombe after some time that ranged from 3 [91] to 5 [89] days. In all cases, a significant amount of L-lactic acid was produced, ranging from 1.52 to 3.41 g/L, resulting in the reduction in the pH value to 3.53–3.69 from the pH value of 3.91–4.06 that was achieved when Sch. pombe strains were used as monocultures. In the majority of the cases, the utilization of L. thermotolerans strains had no or a marginal effect on the final malic, acetic, and citric acids, as well as urea concentration. Notable exceptions were the further decrease in malic acid concentration and the increase in urea concentration that were reported by Vicente et al. [89] and Benito et al. [88], respectively. In the majority of the studies, the decrease in pyruvic acid concentration was also remarked [87,88,91]. The effect on ethanol, glycerol, acetate esters, ethyl esters, alcohols, as well as color intensity, was variable and most likely depended on the Sch. pombe and L. thermotolerans strains used. Notably, no statistically significant difference in biogenic amine concentration was reported [87]. From a sensorial perspective, the wine made by sequential fermentation of L. thermotolerans Concerto^TM followed by Sch. pombe V2 was highly appreciated for its overall impression. The effect of simultaneous inoculation with Sch. pombe 938 and L. thermotolerans on Tempranillo must fermentation and the quality of the resulting wine was assessed by del Fresno et al. [167]. The L. thermotolerans strain was inoculated at a rate of 1:1 and 1:3 (Sch. pombe/ L. thermotolerans). In both cases, the production of L-lactic acid was very weak and did not exceed 0.71 g/L, which resulted in only a slight reduction in the pH value to 4.18. Moreover, this inoculation strategy resulted in the reduction in ethanol, glycerol, total anthocyanin, and pyroanthocyanin concentrations, as well as the increase in ethyl ester, acetate ester, and alcohol concentrations, compared to the respective concentrations obtained when Sch. pombe was used as a monoculture. No effect was observed regarding malic and pyruvic acid concentrations.

Additional features that make the utilization of Sch. pombe strains in winemaking very interesting are the high autolytic release of polysaccharides as well as the enhanced formation of pyroanthocyanins through the significant hydroxycinnamate decarboxylase activity. Romani et al. [171] studied the release of polysaccharides by 3 S. cerevisiae and 89 non-Saccharomyces strains, including Sch. pombe strain Sp1, and reported that the Sch. pombe strain managed to release 712 mg/L polysaccharides, which was the largest amount among the strains examined. This capacity may not only affect the organoleptic quality of the wine positively but also assist in color stabilization, especially when over-lees aging is exercised [172]. Anthocyanins are associated with color quality, improved mouthfeel, and a better aging potential [166]; therefore, their occurrence in the final product is desired. The enhanced formation of pyroanthocyanins was verified by many studies [166,167,173]. However, it should be noted that this is also a strain-dependent property, and there are S. cerevisiae strains that may produce higher concentrations [166], which is reduced by malolactic fermentation. Indeed, in the study by Mylona et al. [166], S cerevisiae 7VA produced a total of 36.40 mg/L anthocyanins, which was significantly higher than the respective amounts produced by the Sch. pombe strains 2139, 938, V1, and 4.2 that were included in that study (22.37–29.77 mg/L), but diminished to 12.38 mg/L after malolactic fermentation. Finally, in the study by del Fresno et al. [167], Sch. pombe strains 938 and V1 managed to produce higher concentrations of total anthocyanins and pyroanthocyanins than S. cerevisiae strains 7VA and 3VA, which were significantly reduced when the Sch. pombe 938 was co-inoculated with L. thermotolerans.

10. Starmerella bacillaris

Starmerella bacillaris (synonym Candida stellata reclassified as Candida zemplinina) [174] is a very interesting species for winemaking due to its fructophilic character, which means that it preferentially ferments fructose in the presence of glucose, its acidification capacity, resulting from the excessive production of pyruvic acid, the production of proteases that seems common among the strains assigned to this species, with a few of them being able to produce also esterases and β-glucosidases, as well as the capacity to tolerate ethanol up to 14% (v/v), which may result in its persistence until the end of fermentation [57,62,93,111,113,155,175,176,177,178,179,180,181,182,183,184]. In addition, St. bacillaris strains may also exhibit antifungal activity [184]. This capacity, and particularly the antifungal activity against Botrytis cinerea, has been utilized by Nadai et al. [185] by spraying the yeast on Pinot Grigio bunches one week before harvest. The yeast retained its viability during this time, and its presence resulted in the increased glycerol content during the subsequent fermentation [185].

Wines produced with St. bacillaris as an adjunct culture are usually characterized by enhanced color, improved thermal stability, and reduced haziness. Color improvement has been attributed to the increased levels of compounds, such as A-type vitisins [186,187]. In addition, the polyphenol levels were further increased when biofilm-detached cells of St. bacillaris were employed [188]. On the other hand, the higher concentration of polysaccharides and mannoproteins improved thermal stability and reduced haziness [189,190,191]. Moreover, these wines usually contain elevated glycerol concentration, most likely due to the need to regenerate NAD+ to support glycolysis. The strain-dependent character of all these properties has been adequately highlighted, except for two characteristics that can be attributed to the species, namely the fructophilic character and the enhanced production of glycerol, which has been reported to reach 16.2 g/L [192]. Regarding the effect that St. bacillaris strains may have on S. cerevisiae strains, as well as the lactic acid bacteria strains carrying out malolactic fermentation, it is also strain-dependent. Early population development of St. bacillaris has been reported to negatively affect the development of the S. cerevisiae population, in the case of sequential inoculation [113,193,194]. On the other hand, physical contact has been reported to play a role in the early disappearance of St. bacillaris during co-cultivation with S. cerevisiae, an action that is dependent on the strains of both species [195]. Finally, malolactic fermentation has been reported to be induced, inhibited, or unaffected in mixed fermentations [196,197,198,199].

The effect of St. bacillaris strains as a starter culture, in combination with S. cerevisiae strains, on the physicochemical properties, chemical composition, and organoleptic quality of the wine has been assessed in both simultaneous and sequential formats. When strains of both species were inoculated simultaneously, the reduction in acetaldehyde concentration was reported in all cases, in the majority of the cases the increase in glycerol and trans-3-hexen-1-ol concentration was reported, while total acidity as well as the concentration of octanoic acid and ethyl decanoate in the final product was reported without statistically significant changes, in comparison to the wines made solely by the respective S. cerevisiae strains. In addition, the concentration of ethyl hexanoate was reported in the majority of the cases to be constant or presented a statistically significant increase. Finally, the pH value, volatile acidity, as well as the concentration of isobutanol, isoamyl alcohol, hexanoic acid, ethyl 9-decenoate, isoamyl acetate, ethyl acetate, and 2-phenethyl acetate were reported to decrease or to remain constant, compared to the respective values obtained when the S. cerevisiae strains were employed as the sole starter culture. No particular trend was identified for the remaining volatile compounds [57,113,177,179,200,201,202,203,204,205].

When strains of both species were inoculated sequentially, i.e., S. cerevisiae strains were inoculated usually 48 h after St. bacillaris inoculation. In all cases, the increase in glycerol, trans-3-hexen-1-ol, and ethyl lactate concentration and the decrease in octanoic acid, ethyl dodecanoate, and methyl decanoate concentration, compared to the respective concentrations obtained by S. cerevisiae strains, was reported. In addition, the increase in total acidity and the concentration of citronellol, geraniol, linalool, isobutanol, 1-hexanol, 2-methyl-1-propanol, isovaleric acid and benzyl alcohol and the decrease in the pH value and the concentration of 3-methyl-1-butanol, decanoic acid, ethyl 9-decenoate, ethyl heptanoate, ethyl decanoate, ethyl octanoate, ethyl acetate, and hexyl acetate was reported in the majority of the cases. On the other hand, the concentration of acetic acid, 4-terpineol, and 2,3 butanediol remained without any statistically significant changes, compared to the respective concentrations obtained when the S. cerevisiae strains were employed as the sole starter. Finally, the concentration of ethanol, β-damascenone, acetaldehyde, and 2-phenylethanol either remained stable or decreased, while the concentration of γ-butyrolactone and benzaldehyde either remained stable or increased. No particular trend was identified for the remaining volatile compounds [80,179,190,191,192,202,204,205,206,207].

11. Torulaspora delbrueckii

Torulaspora delbrueckii has been characterized as the most suitable non-Saccharomyces yeast species for winemaking [208]. This is justified by a comparatively good fermentation performance, in terms of alcohol production that may reach 13.6% (v/v), which is usually accompanied by low production of acetic acid (0.1–0.4 g/L), enhanced production of glycerol that may reach 8.6 g/L, as well as high levels of mannoprotein and polysaccharide release [209,210,211,212,213]. Notably, the higher mannoprotein concentration has been correlated with increased color intensity [214]. In addition, its strong glycosidase and β-lyase activities result in the release of aroma compounds with a positive impact on the aromatic profile of the final product [213,215,216]. Moreover, the capacity of T. delbrueckii to enhance malolactic fermentation performance by O. oeni, has been repeatedly reported [217,218,219]. However, this characterization is compromised by the slower growth rate and reduced fermentation rigor of T. delbrueckii under winemaking conditions, compared to S. cerevisiae, which may lead to the substitution of its population by indigenous S. cerevisiae strains [220,221,222]. Although differences at the strain level do occur, the aforementioned characteristics have been observed in the majority of the cases [93,223,224].

Apart from the metabolic interactions between T. delbrueckii and S. cerevisiae, physical interactions have also been reported to affect the fermentation. Indeed, the physical interactions between T. delbrueckii CVE-TD12 and S. cerevisiae Lalvin^® D254™ have been reported to enhance the production of most aroma compounds, especially acetate esters and volatile fatty acids [225]. As far as the metabolic interactions were concerned, specific metabolites produced by T. delbrueckii enhanced the production of specific metabolites by S. cerevisiae. More specifically, 3-amino-4-methylpentanoic acid produced by T. delbrueckii enhanced the production of ethyl octanoate and ethyl decanoate by S. cerevisiae, and 4-methylaminobutyric acid and citraconic acid enhanced the production of ethyl hexanoate and isoamyl acetate [226].

The effect of T. delbrueckii strains as a starter culture, in combination with S. cerevisiae strains, on the physicochemical properties and organoleptic quality of the wine has been assessed in both simultaneous and sequential formats. When T. delbrueckii and S. cerevisiae strains were simultaneously inoculated in musts, the population of the former was countable up to the seventh day of fermentation [15,227,228]. The wines that were produced were, in most cases, characterized by unaffected ethanol content, pH, total and volatile acidity, compared to the ones made solely by the respective S. cerevisiae strains. In addition, the concentration of citronellol, geraniol, D-limonene, α-terpineol, linalool, geranyl acetate, neryl alcohol, farnesyl alcohol, β- and α- ionone, octanal, nonanal, decanal, acetoin, 1-heptanol, 2-phenylethanol, 1-decanol, 2-ethyl-1-hexanol, cis-3-hexen-1-ol, isopentanol, isovaleric acid, isobutyric acid, hexanoic acid, octanoic acid, decanoic acid, trans-2-hexenoic acid, valeric acid, ethyl nonanoate, ethyl isovalerate, ethyl hydroxybutyrate, isobutyl acetate, diethyl succinate, 2-phenylacetaldehyde, benzyl alcohol, phenol, and 4-ethylguaiacol was not affected in the majority of the cases by T. delbrueckii strains addition. On the contrary, the changes in the concentration of glycerol, acetic acid, isobutanol, 3-methyl-1-pentanol, 1-propanol, 1-hexanol, 4-methyl pentanol, isoamyl alcohol, 1-octanol, 1-pentanol, trans-3-hexen-1-ol, 1-butanol, ethyl lactate, ethyl butyrate, ethyl hexanoate, ethyl decanoate, ethyl octanoate, isoamyl acetate, 2-phenethyl acetate, hexyl acetate, methyl octanoate, ethyl methylbutyrate, ethyl butyrate, ethyl laurate, ethyl palmitate and methyl salicylate exhibited no particular trend as statistically significant increases and decreases as well as statistically insignificant differences have been reported between the wines made by co-inoculation of T. delbrueckii and S. cerevisiae strains and the wines made solely by the respective S. cerevisiae strains [15,16,227,228,229]. When T. delbrueckii strains were sequentially inoculated with S. cerevisiae, the reduction in the population of the T. delbrueckii strains, if any, only took place at later stages of fermentation [83,227,230]. The wines that were produced were, in most cases, characterized by reduced ethanol content and unaffected volatile acidity, compared to the ones made solely by the respective S. cerevisiae strains. In addition, the concentration of citronellol, linalool, and 1-propanol was not affected; the concentration of β-damascenone increased, and the concentration of acetaldehyde and benzaldehyde decreased. On the contrary, the changes in the concentration of glycerol, acetic acid, geraniol, nerol, 2-phenylethanol, isobutanol, isoamyl alcohol, 1-hexanol, hexanoic acid, ethyl butyrate, ethyl hexanoate, ethyl octanoate, isoamyl acetate, 2-phenethyl acetate and ethyl acetate exhibited no particular trend as statistically significant increases and decreases as well as statistically insignificant differences have been reported between the wines made by sequential inoculation of T. delbrueckii and S. cerevisiae strains and the wines made solely by the respective S. cerevisiae strains [83,227,230,231].

12. Wickerhamomyces anomalus

Wickerhamomyces anomalus (synonym Pichia anomala, Hansenula anomala) is frequently isolated from a wide variety of microecosystems ranging from soil and plant surfaces to fermented food. Its capacity to withstand harsh environmental conditions and utilize an extended variety of carbon and nitrogen sources is indicative of its highly competitive nature [232,233]. Regarding winemaking, persistence of W. anomalus until the end of fermentation and tolerance above 15.5% (v/v) ethanol have been reported [234,235,236,237,238,239]. In addition, many strains have been reported to produce antimicrobial metabolites, including killer toxins and volatile organic compounds, with ethyl acetate and 2-phenylethanol considered as the most prominent ones [232,240,241,242,243]. The applicability of these antimicrobial compounds for the biocontrol of wine spoilage yeasts has also been exhibited [243,244]. Apart from these volatile compounds, another property of enological interest that characterizes this species is the production of glycosidases and esterases [245,246,247,248]. Despite the strain-dependent character of this property, the characterization of W. anomalus strains as good sources of adequate glycosidase activity is quite common [249,250]. Moreover, several β-glycosidases from W. anomalus strains have been reported to retain significant activity at elevated glucose concentration (up to 20%), ethanol concentration up to 20% (v/v), and pH value of 3.5 [251,252,253].

Only a few studies are currently available on the effect of W. anomalus on wine quality, in either a simultaneous or sequential inoculation strategy with S. cerevisiae. In both cases, the enhancement of aromatic complexity has been reported [18,246,254,255,256]. As far as the chemical characteristics of the final products were concerned, the specific characteristics of the W. anomalus strains employed seem to be very important. For example, the ethanol content has been reported to decrease, remain without statistically significant change, or even increase during co-inoculation of W. anomalus strains with S. cerevisiae [18,246,254,255,256,257,258]. Similar is the case of pH value, titratable and volatile acidity, as well as color intensity and hue, and the total phenolic, flavonoid, anthocyanin, ester, etc., content of the wines. This is also affected by the rather restricted number of studies; a larger number of studies would have facilitated the unveiling of the most common strain-dependent properties of this species. However, it seems that the higher the glucosidase and esterase activities, the more complex the aromatic profile of the final product.

13. Conclusions

Non-Saccharomyces yeasts can be very helpful in addressing the challenges caused by global warming, in enhancing the grape variety typicity of wine, and in assisting in its bioprotection. However, these activities are characterized by strain variability. In addition, the effectiveness of the non-Saccharomyces strains employed is also affected by the S. cerevisiae strains used to carry out alcoholic fermentation. This variability necessitates further research towards two directions, namely, the discovery of new strains, both Saccharomyces and non-Saccharomyces, capable of carrying out the aforementioned metabolic activities, and their optimization, taking also into consideration the vine grape variety and technological parameters, such as time of inoculation and incubation temperature. By elucidating the molecular and cellular mechanisms that govern the interactions between the yeast strains employed and between them and their environment, in terms of the occurrence of precursor molecules and the conditions for their efficient bioconversion, the way towards precision fermentation in winemaking will be paved.