Commercially Available Non- Saccharomyces Yeasts for Winemaking: Current Market, Advantages over Saccharomyces , Biocompatibility, and Safety

: About 42 commercial products based on non- Saccharomyces yeasts are estimated as available on the market, being mostly pure cultures (79%), with a predominance of Torulaspora delbrueckii, Lachancea thermotolerans , and Metschnikowia pulcherrima . The others are multi-starter consortia that include non- Saccharomyces / Saccharomyces mixtures or only non- Saccharomyces species. Several commercial yeasts have shown adequate biocompatibility with S. cerevisiae in mixed fermentations, allowing an increased contribution of metabolites of oenological interest, such as glycerol, esters, higher alcohols, acids, thiols, and terpenes, among others, in addition to a lower production of acetic acid, volatile phenols, biogenic amines, or urea. Multi-starter inoculations are also reviewed here, which show adequate biocompatibility and synergy between species. In certain cases, the aromatic proﬁle of wines based on grape varieties considered neutral is improved. In addition, several yeasts show the capacity as biocontrollers against contaminating microorganisms. The studies conducted to date demonstrate the potential of these yeasts to improve the properties of wine as an alternative and complement to the traditional S. cerevisiae .


Introduction
Traditionally, non-Saccharomyces yeasts have been considered as contaminants due to the production of undesirable metabolites by many of the species currently known. This aspect has been changing in recent years due to the growing interest in certain strains that contribute with metabolites that positively impact wine.
Most non-Saccharomyces yeasts are characterized by their low fermentative power and low ethanol tolerance, especially in the presence of SO 2 [1]; thus, to ensure a correct end of the fermentative process, its use necessarily requires the implementation of mixed fermentations (simultaneous or sequential) together with Saccharomyces cerevisiae [2].
In the last 20 years, this scenario has resulted in the search for and selection of new strains by the scientific community, as evidenced in the high number of publications related to non-Saccharomyces yeasts of oenological interest as well as in the development and market launch by commercial houses of products based on these selected non-Saccharomyces strains in various formats included in Resolution OIV-OENO 576B-2017 of the International Organization of Vine and Wine (OIV), which includes active dry yeast (ADY, dry matter > 92%), active frozen yeast (AFY, dry matter 40-85%), compressed yeast (COY, dry matter 30-35%), and cream yeast (CRY, dry matter 18-25%), in addition to encapsulated yeasts (pearls) or immobilized yeasts (ENY) with more than 86% dry matter. Additionally, several commercial products in the form of fresh liquid yeast have been identified from an online review.

Non-Saccharomyces Yeasts Available on the Market
Various reviews have addressed aspects such as the metabolic characteristics and the most important contributions of non-Saccharomyces to wine [3][4][5], improvement in wine properties such as acidity, and its influence on various oenological parameters [4], as well as statistical information regarding the providing companies, more commercialized species, quantity of commercial strains, regulations, and patents [5], among other aspects.
Based on an Internet search, as a part of this study, it is estimated that about 42 commercial products based on non-Saccharomyces yeasts are available for winemaking in different formats (Figure 1), of which 52% are represented by three species: Torulaspora delbrueckii, Lachancea thermotolerans, and Metschnikowia pulcherrima. In addition, 79% are marketed as pure cultures (monoculture) and the remaining available products are offered as multi-starters (blends of various yeast species). Four companies produce 52% of the supply.  According to the scientific literature, the use of non-Saccharomyces at the industrial level is still a pending issue since most applications have been conducted on the experimental scale. This indicates that the knowledge of these yeasts is still a field requiring development, considering the growing interest from the oenological sector in the production of wines with differentiated profiles, as reported by Roudil et al. [5], who described the evolution of the supply of these commercial yeasts in recent years. Several companies supply these yeasts for oenological applications (Table 1). Table 1. Commercial non-Saccharomyces yeasts available on the market. Information obtained from the website of the companies that commercialize them. More details of each yeast are provided in Supplementary Materials (Table S1). The general aspects of commercial non-Saccharomyces are described in Section 2. Other aspects such as the improvement in the fermentative and varietal aromatic profile with respect to S. cerevisiae, applications in sparkling wines, ternary fermentations, biocompatibility with S. cerevisiae, and aspects of wine safety are described from Section 3.

Torulaspora Delbrueckii
Torulaspora delbrueckii (previously known as Saccharomyces rosei or Saccharomyces delbrueckii) shows the capacity to ferment in monoculture. One case was reported in fermentations of Chenin Blanc and Chardonnay blend musts [7] by the two isolated strains from the Harmony and Melody multi-starters (CHR Hansen, Hørsholm, Denmark), reaching between 12% and 13% v/v of ethanol (S. cerevisiae reached 13% v/v), in addition to having low residual sugar content.
T. delbrueckii Biodiva TD291 (Lallemand, Montreal, QC, Canada) also shows good fermentative performance in monoculture in Chardonnay and Xarel.lo musts [8], as well as in mixed fermentations with S. cerevisiae in musts with a high sugar content (between 30 and 42 • Brix), used to produce the Italian wines Amarone and Santo [9,10], which makes it suitable for producing late-harvest wines. However, the commercial strain T. delbrueckii Zymaflore Alpha (Laffort, Bordeaux, France) does not show capacity to finish the fermentative process in Sauvignon Blanc musts, reaching only 6.2% v/v of ethanol and residual sugars > 100 g/L [11].
T. delbruekii shows the capacity to produce other metabolites of oenological interest, such as glycerol. Biodiva TD291 [12] and Oenoferm Wild & Pure (Erbslöh, Geisenheim, Germany) [13] show a high production of glycerol in sequential fermentations with Saccharomyces, indicating the potential of these yeasts to improve complexity and mouth-feel. Biodiva TD291 also shows less, or similar production, of acetic acid and volatile acidity compared to pure S. cerevisiae [10,12,14] as shown in Table 2.

Lachancea Thermotolerans
Lachancea thermotolerans was previously known as Kluyveromyces thermotolerans. The main advantage of this yeast is its capacity to produce lactic acid, improving the acidity of wine, along with its intensity and aromatic complexity [3,15], especially in sequential fermentations with S. cerevisiae [15]. In Riesling [16] and Tempranillo wines [17], L. thermotolerans Concerto (CHR Hansen) has shown capacity to increase the production of lactic acid compared to the traditional process that integrates alcoholic fermentation and malolactic fermentation (AF + MLF) ( Table 2). The increase in lactic acid in Tempranillo contributed to improve the sensation of acidity and the sensory acceptability.
Another important contribution of L. thermotolerans Concerto is the highest production of pyruvic acid in sequential fermentations with S. cerevisiae and with Schizosaccharomyces pombe in Tempranillo wines [17] compared to the traditional AF + MLF process ( Table 2). The increase in pyruvic acid also contributed to improve the color of the wines due to the synthesis of vitisin A, which is characterized by its increased stability [18].
In sequential fermentation with S. cerevisiae, L. thermotolerans Concerto showed a higher production of glycerol than S. cerevisiae [2]. In ternary fermentations (simultaneous fermentations), the multi-starter Melody (CHR Hansen), which contains L. thermotolerans [2], showed the capacity to reduce the alcohol content in Syrah wines ( Table 2).
The improvement in acidity with L. thermotolerans can be a strategy that contributes to improving the composition and sensory profile in wines produced from grapes grown in warm regions where high temperatures may generate low acidity in the grape berries [19].
Both wines: highest scores in Verdejo typicity, fruity, intensity, and aromatic quality

Metschnikowia Pulcherrima
Metschnikowia pulcherrima is known for its capacity to reduce alcohol content [24,25], which has piqued commercial interest, especially for wines produced from grapes grown in warm regions that are characterized by a high sugar content [19,27].
Similar to most non-Saccharomyces yeasts, M. pulcherrima is characterized by a low ethanol tolerance, especially in the presence of SO 2 [1], limiting the ability to complete the transformation of grape sugars into ethanol [13]. This reduction in ethanol synthesis is the result of the redirection of the fermentation pathway toward the production of biomass [24] or toward the production of metabolites of oenological interest, mainly esters during the early stages of the fermentative process [21] and during mixed fermentations with Saccharomyces [14,23,26], and of glycerol, especially in sequential fermentations with Saccharomyces [2,13].
M. pulcherrima AWRI Obsession (AB Biotek, London, United Kingdom), in simultaneous fermentations with S. cerevisiae, reduced the alcohol content in Merlot wines by up to 1.0% v/v [24]. M. pulcherrima NS-EM-34 (reported as pre-commercial) in sequential fermentations with S. cerevisiae produced reductions of up to 0.63% v/v of ethanol in Verdejo wines [25] as well as increases in the glycerol content. A greater production of glycerol was obtained with M. pulcherrima Flavia MP346 (Lallemand, Montreal, QC, Canada) in sequential fermentation with S. cerevisiae in Syrah wines [2] as shown in Table 2.
Puškaš et al. [13], working with M. pulcherrima Flavia MP346, obtained reductions in the alcohol degree of Sila white wines of up to 0.62% and 0.91% v/v using sequential fermentations M. pulcherrima/S. bayanus/S. cerevisiae and M. pulcherrima/S. bayanus, respectively, in addition to a greater production of glycerol (Table 2). In addition, these fermentations improved the aromatic quality of the wine, in general.
Unlike most commercial non-Saccharomyces yeasts, which are commercialized as active dry yeast (ADY), Viniflora FrootZen is commercialized and stored in freezing conditions (−45 • C) and inoculated directly into the grape must without prior rehydration. The multi-starter Trillyeast is commercialized in the form of cream yeast (CRY) ( Table 1).
Regarding the contributions of this yeast, the existing evidence reflects the need for more studies to optimize the fermentative conditions to take advantage of the benefits mentioned in the technical datasheets of these commercial strains, in order to improve certain aspects such as the high production of metabolites that negatively affect wine.

Schizosaccharomyces Pombe
Schizosaccharomyces pombe is commercialized as an alternative to acid lactic bacteria for the biological deacidification of wine, converting malic acid into ethanol due to its capacity to achieve maloalcoholic fermentation (MAF) simultaneously with alcoholic fermentation (AF) [4]. According to Mylona et al. [29], S. cerevisiae may finish AF in 6 days, and then wine is submitted to the traditional malolactic fermentation (MLF) that may last up to 21 days, for a total of 27 days. S. pombe can finish the fermentation process in 10 days, performing AF and MAF simultaneously, with ethanol levels of up to 12% v/v and with a residual sugar content lower than 2 g/L.
The use of S. pombe for biological deacidification was approved by the OIV in 2013. However, few commercial strains of this species are available (Table 1). This low commercial availability is related to its high production of acetic acid, which can reach levels of higher than 1.0 g/L [29]. This aspect has been improved through the selection of strains with a low production of acetic acid (around 0.4 g/L [30]), or the implementation of suitable fermentative strategies. An interesting result was obtained by Benito et al. [17] in sequential fermentations between L. thermotolerans Concerto (CHR Hansen) and a non-commercial strain of S. pombe, reaching an acetic acid content of less than 0.36 g/L in Tempranillo wines.
This yeast shows a high capacity to release polysaccharides during ageing-on-lees (AOL), especially in red wines, due to its greater autolytic capacity compared to S. cerevisiae, contributing to improving the stability and quality of wine, whose benefits were addressed in a prior review [31]. It also has the capacity to produce higher levels of pyruvic acid compared to the traditional fermentation process AF + MLF [17], with the consequent potential for increased synthesis of vitisin A, contributing thus to improving the color and stability of red wines [18].

Hanseniaspora Vineae
Hanseniaspora vineae was previously known as Hanseniaspora osmophila. In a recent study, a pre-commercial strain of H. vineae (currently under evaluation by Oenobrands, Montpellier, France) was used in monoculture fermentation to produce Albillo wines [26], reaching an alcohol degree of 11.9% v/v, and showing no major differences from the wine produced with S. cerevisiae, with the exception of the total acidity, which was slightly lower in the H. vineae wine, related to the precipitation of tartaric acid.
A recent study [33] found a higher production of higher alcohols by Zymaflore Alpha compared to Prelude and Biodiva TD291, all in monoculture in a commercial grape juice, in addition to a higher production of medium-chain fatty acids and total esters and an increased degradation of malic acid and sugar consumption.
In the same study, with the same grape juice enriched with N (based on inactivated yeasts), they obtained a higher production of total esters, especially 2-phenylethyl acetate, with Zymaflore Alpha, noting further that this strain was the only one to produce isoamyl acetate, amyl acetate, ethyl hexanoate, and ethyl octanoate. However, in the N-enriched medium, the three strains of T. delbrueckii produced higher amounts of H 2 S, especially Zymaflore Alpha, which may be related to the presence of sulfur amino acids in the enriched medium.
Regarding the improvement in the synthesis of esters, sequential and simultaneous fermentations in Sauvignon Blanc and sequential fermentations in Merlot, with T. delbrueckii Zymaflore Alpha and S. cerevisiae, produced an increase in isoamyl acetate (banana), isobutyl acetate (banana), 2-phenylethyl acetate (rose), ethyl isobutyrate (strawberry, red fruit), ethyl propanoate (strawberry), and ethyl dihydroxycinnamate (pineapple, almond) [11]. The authors considered that ethyl propanoate, ethyl isobutanoate, and ethyl dihydroxycinnamate may be considered aromatic markers for T. delbrueckii Zymaflore Alpha, since they are usually not synthesized at important levels by S. cerevisiae. All these esters contributed to improving the complexity and fruity note in Merlot wines, produced at a semi-industrial level ( Table 2).
The high production of isoamyl acetate and isobutyl acetate in Merlot wines indicates a positive interaction between both yeasts, also observed in sequential fermentations with T. delbrueckii Prelude and S. cerevisiae in Shiraz wines (Table 2), as well as 2-phenylethyl acetate and ethyl isobutyrate in sequential fermentations involving Zymaflore Alpha, Biodiva TD291, and Prelude, with a consequent improvement in the aromatic quality, especially the fruity notes [2].
However, other authors reported a decrease in the content of esters in sequential fermentations with Zymaflore Alpha and Biodiva TD291, especially isoamyl acetate in Chardonnay, Soave, Amarone, and Santo wines [9,10], although, in all cases, it was produced above its threshold (30 µg/L [32]), showing that the capacity to produce esters, in addition to the grape variety and the type of winemaking, depends on the intra-species variability of each yeast, based on its enzymatic esterase and acetyltransferase activities [11].
The production of aromatic compounds in the early stages of fermentation (alcohol degree about 2-3% v/v) with pure T. delbrueckii Biodiva TD291 was also evaluated [21]. The authors obtained an increase in phenethyl propanoate (rose aroma) ( Table 2), which is not commonly synthesized by S. cerevisiae. The authors indicated the need to evaluate whether this ester persists until the end of fermentative process, for example, in sequential fermentations with S. cerevisiae.
An increase in the content of lactones was obtained with Biodiva TD291 in Amarone wines, especially 4-carbethoxy-γ-butyrolactone (sweet coconut aroma), at concentrations higher than its threshold of 400 µg/L, in agreement with the sensory analysis in which wines were described with greater aromatic intensity and sweetness [9]. However, the authors emphasized that these results should be considered with caution since Amarone wines can be commercialized after two years of ageing.
Regarding the production of metabolites with negative connotations, sequential fermentations with the Zymaflore Alpha strain showed the capacity to reduce the content of 4-vinylphenol and 4-vinylguaiacol in Soave and Chardonnay wines [10], and in the case of 4-vinylguaiacol, to levels below its threshold (40 µg/L [34]). Biodiva TD291 showed the same effect in sequential fermentations in Santo sweet wine [10].
In contrast, sequential fermentations of Biodiva TD291 with S. cerevisiae in Sauvignon Blanc wines increased the production of 3-methylthio-1-propanol (sulfur, onion, raw potato), 3-(2-hydroxyethyl)thio-1-propanol (sulfur, onion), and ethyl ester of 3-methylthiopropanoic acid (metallic, pineapple, fruity, ripe pulpy tomato) [14]. The origin of these thiols may be the metabolism of amino acids such as methionine, or fermentations with grape musts poor in amino acids [35], such that the results indicate a grape must with low levels of amino acids, that T. delbrueckii catabolized methionine more easily, or that it generated an impoverishment in amino acids in the medium facilitating the subsequent synthesis of these thiols by S. cerevisiae [14].
However, the concentrations of these thiols were not high enough to be detected in sensory analysis [14], which indicates the need for further studies on the catabolism of amino acids by T. delbrueckii to elucidate the mechanism through which thiols are synthesized and to establish strategies to decrease their production.

Lachancea Thermotolerans
L. thermotolerans, in addition to acidity, can improve the aromatic profile of wine. The use of the Concerto strain (CHR Hansen, Hørsholm, Denmark) in monoculture was reported to result in a high production of 2-phenylethanol, phenethyl propionate (rose), and other esters in the initial stages of fermentation of Sauvignon Blanc and Syrah musts and wines (alcohol degree between 2.0% and 3.0% v/v) [21]. Phenethyl propionate is commonly not synthesized by S. cerevisiae; thus, its synthesis constitutes a strategy to improve the aromatic profile in wine through mixed fermentations with S. cerevisiae.
The Concerto strain, in sequential fermentations with S. cerevisiae, also showed a higher production of 2-phenylethanol in high-alcohol Syrah wines [2] and ethyl esters in Riesling wines [16]. In Sauvignon Blanc wines, it produced high amounts of isoamyl acetate and citronellol acetate [14] and isobutyl acetate in Syrah wines [2] (Table 2).
In Syrah wines, a higher content of isoamyl acetate and isobutyl acetate was obtained with the Melody multi-starter (CHR Hansen, Hørsholm, Denmark), which contains a strain of L. thermotolerans as well as a higher production of 2-phenylethanol in Syrah wine produced from over-ripe grapes (29 • Brix) [2].
Another metabolite of interest is acetaldehyde, whose production was decreased in sequential fermentations of L. thermotolerans Concerto with S. cerevisiae in Riesling wines [16], contributing to higher preference and Riesling typicity, higher fruit perception (peach and apricot) and aromatic quality, and lower oxidation, acetaldehyde, and ethyl acetate perception.
However, in Beckner Whitener et al. [21], L. thermotolerans Concerto produced higher amounts than the control wine of 3-methylthio-1-propanol (Table 2), which negatively impacts wine (sulfur aroma, onion). The persistence of this compound can be evaluated in sequential fermentations with S. cerevisiae, as its presence was detected in the early stages of fermentation (alcohol degree between 2.0% and 3.0% v/v).

Metschnikowia Pulcherrima
In the early stages of fermentation (between 2.0% and 3.0% v/v ethanol) in Syrah musts [21], M. pulcherrima Flavia MP346 (monoculture) showed the capacity to synthesize isoeugenil phenylacetate and phenethyl propionate (rose aroma), which are not usually produced by S. cerevisiae (Table 2). However, it is necessary to assess whether these esters persist until the end of fermentative process, for example, in sequential fermentations.
In sequential fermentations with S. cerevisiae [14], Flavia MP346 also showed the capacity to produce methyl-butyl, methyl-propyl, and phenylethyl esters in Sauvignon Blanc wines (Table 2). On the contrary, in Riesling wines, a decrease in acetate esters was obtained [16]. In ternary fermentations L. thermotolerans/M. pulcherrima Flavia MP346 + S. cerevisiae, Vaquero et al. [27] obtained a higher production of esters in Airén wines.
M. pulcherrima AWRI Obsession (AB Biotek, London, United Kingdom) showed increased production of total esters in simultaneous fermentations with S. cerevisiae in Merlot wines [24], obtaining high scores in aroma and fruity flavor, and a sensory profile similar to the wine produced with S. cerevisiae.
Regarding higher alcohol content, in sequential fermentations of M. pulcherrima Flavia MP346 with S. cerevisiae, variable results have been obtained, with an increase in 2-phenylethanol in high-alcoholic Syrah wines [2] or a decrease in the content of 2phenylethanol and higher alcohols in general in Riesling wines [16]. M. pulcherrima NS-EM-34 (reported as precommercial) in sequential fermentation with S. cerevisiae demonstrated a lower production of higher alcohols in Verdejo wines [25].
M. pulcherrima AWRI Obsession showed a higher production of higher alcohol in simultaneous fermentations with S. cerevisiae in Merlot wines [24]. In ternary fermentations of L. thermotolerans/M. pulcherrima Flavia MP346 + S. cerevisiae, Vaquero et al. [27] reported a higher production of higher alcohol in Airén wines.
The increased synthesis of compounds with negative, such as 2-methoxy-4-vinylphenol in Sauvignon Blanc and Syrah musts and wines [21], was also reported with M. pulcherrima Flavia MP346 in monoculture (Table 2), which indicates the presence of hydroxycinnamate decarboxylase activity, in addition to 3-methylthio-1-propanol in Syrah wines (sulfur, onion aroma). It will be necessary to evaluate their evolution in mixed fermentations with S. cerevisiae, as well as to study the evolution of these compounds on a larger scale to optimize fermentation conditions that help reduce their production and impact on the wine.
Regarding other sulfur compounds, in simultaneous fermentations of M. pulcherrima AWRI Obsession and S. cerevisiae in Merlot wines, an increase in H 2 S, dimethyl sulfide, ethanethiol, and methanethiol was obtained ( Table 2) with respect to the control wine [24], being the first study to report the production of these compounds by M. pulcherrima. However, in the sensory analysis, the presence of these compounds was not detected, highlighting, on the contrary, fruity aromas. The perception of fruity aromas may be related to the lower alcohol content and higher levels of esters and higher alcohol. Previously, it was reported that lower levels of ethanol can contribute to a better expression of fruit aromas [36].

Pichia Kluyveri
In sequential fermentations with S. cerevisiae, P. kluyveri Viniflora FrootZen (CHR Hansen, Hørsholm, Denmark) showed the capacity to increase the 2-phenylethyl acetate content and release high amounts of amino acids in Riesling wines [16] while reducing the contents of acetaldehyde and isoamyl acetate. However, these properties do not seem to affect the acceptability of wines, showing increased preference by the sensory panel, a lower perception of oxidation, acetaldehyde, and ethyl acetate, and greater value for the peach and apricot attribute.
In a subsequent study with Viniflora FrootZen, in sequential fermentations with S. cerevisiae in Sauvignon Blanc musts [14], high production of 3-methyl-butanoic acid (isovaleric acid) was obtained (sour, sweaty, cheese-like aroma) ( Table 2), which derives from the catabolism of L-leucine. However, the concentration of this acid was not high enough to be detected in the sensory analysis. It was suggested that the synthesis of isovaleric acid can be considered a criterion for the selection of P. kluyveri [28] and that its esterification can produce the ethyl ester of 3-methyl-butanoic acid, which has a pleasant fruity aroma.
Based on those results, together with the absence of positive sensory attributes, Beckner Whitener et al. [14] suggested that Viniflora FrootZen strain would not be a good candidate to produce Sauvignon Blanc wines, in addition to the high levels of phenylethylamine detected in these wines.
Since few commercial products are based on P. kluyveri, the available field of study to select new strains with commercial potential is wide, as well as to develop fermentative strategies that, by using the strains currently available, take advantage of the benefits reported for this yeast in its technical datasheet (Table S1).
The capacity of this yeast to form films on the surface of wine may also be used [28], for example, through its industrial application as a "flower-film yeast", as an alternative to the traditional Saccharomyces used for the production of Sherry wines.

Hanseniaspora Vineae
The strain H. vineae T02/5AF (of Uruguayan origin) was used in monoculture to produce Macabeo wine with increased contents of 2-phenylethyl acetate, isobutyl acetate, and ethyl lactate with respect to S. cerevisiae [37], in addition to a lower acetoin content ( Table 2). In this study, a lower synthesis of higher alcohols was obtained, in addition to the synthesis of N-acetiltiramine and 1H-indole-3-ethanol acetate (not synthesized by S. cerevisiae). In addition, the wine produced with H. vineae T02/5AF received a higher preference score and a higher score for the fruity and floral attribute in the sensory analysis, which indicated the positive contribution of the esters.
More recently, an increase in the production of 2-phenylethyl acetate was obtained in Albillo wines with a pure pre-commercial strain of H. vineae [26] as shown in Table 2.

Commercial Non-Saccharomyces Yeasts in Sparkling Wines
The current literature reports few studies with commercial non-Saccharomyces yeasts in sparkling wines, both at the base wine level (first fermentation) and during the stage of second fermentation and bottle ageing.
One of the pioneering studies was conducted by González-Royo et al. [12] with T. delbrueckii Biodiva TD291 (Lallemand, Montreal, QC, Canada), in sequential fermentation with S. cerevisiae to produce base wine from Macabeo grapes, resulting in an increase in glycerol, a decrease in volatile acidity, and better foam properties ( Table 2). The sensory acceptability of this wine was also higher. However, a higher production of volatile phenols with T. delbrueckii Biodiva TD291 and M. pulcherrima Flavia MP346 was obtained, although, in all cases, within the desired sensory limits. In addition, 2,6-dimethoxyphenol was produced by Flavia MP346 at levels higher than its threshold (OAV > 1.0), producing a marked smoky aroma in the sensory analysis.
Despite the improvements over the base wines, authors [12] highlighted the need to assess the long-term impact on the corresponding sparkling wine to determine whether the properties are maintained, or whether they are modified by the action of second fermentation and bottle ageing.
More recently, the same yeasts (Biodiva TD291 and Flavia MP346), in addition to three strains of S. cerevisiae, were used in monoculture to produce base wine from Chardonnay and Xarel.lo grape musts [8]. All wines showed residual sugar levels below 0.4 g/L. An increased amount of proteins was also obtained, especially in wines fermented with Flavia MP346, conferring better foam properties (Table 2) as reported by González-Royo et al. [12]. All the wines were subsequently fermented and bottle aged for 18 months with a commercial strain of Saccharomyces bayanus. T. delbrueckii Biodiva TD291 wines showed a higher content of esters, especially ethyl isovalerate (aroma of pineapple, apple, pear, anise, and flowers) in both Cava wines (Chardonnay and Xarel.lo), in addition to isoamyl acetate and hexyl acetate. This indicates the contribution of Biodiva TD291 to the fruity character, in agreement with the sensory analysis (Table 2). Furthermore, unlike the study of González-Royo et al. [12], undesirable compounds such as volatile phenols were not detected.
Based on the results of T. delbrueckii Biodiva TD291 and M. pulcherrima Flavia MP346 in both studies, its biocompatibility in ternary fermentations could be evaluated, in addition to designing and implementing fermentative strategies that take advantage of the positive effects of both yeasts on the quality of foam and sensory profile in sparkling wines, as both studies were conducted at a semi-industrial level.

Commercial Non-Saccharomyces to Improve Varietal Aromatic Profile
The design and implementation of appropriate fermentation strategies may allow exploiting the capacity of non-Saccharomyces yeasts to improve the aroma of wines produced from grape varieties with low aromatic expression considered neutral, for example, white varieties such as Airén [15], Macabeo [38], Chenin Blanc, and Colombar [39], among others, as well as red varieties such as Tannat [40], Parraleta [38], Viura, etc. The expression of varieties with aromatic typicity can also be improved.

Improvement in Terpenes Content
Torulaspora delbrueckii contributed terpenes in different types of wines (Table 2), demonstrating a good β-glucosidase activity, in addition to its capacity to initiate interconversion reactions between terpenes [41].
Cuš and Jenko [20] used T. delbrueckii Biodiva TD291 (Lallemand, Montreal, QC, Canada) in sequential fermentation with S. cerevisiae to produce Gewürztraminer wine, obtaining variable results for each terpene. The contents of citronellol (tropical fruit) and geraniol (floral: rose) were lower than when pure S. cerevisiae was used, although, in all cases, it was higher than the threshold (OAV > 1.0). α-terpineol (lily), despite increasing, did not exceed the threshold (OAV < 1.0). Linalool (floral: rose) was the only one that, in addition to increasing in the presence of Biodiva TD291, approached its threshold (OAV = 0.97). These results led to a higher overall sensory score, indicating a balance between terpenes.
Linalool was also reported at the beginning of the fermentation of Sauvignon Blanc and Syrah musts in the presence of T. delbrueckii Biodiva TD291 [21]. α-terpineol, Hodiendiol I, and endiol increased in concentration in Amarone wines in the presence of this yeast strain in mixed fermentations with S. cerevisiae [9], thereby improving the aromatic intensity and diminishing the vegetal notes.
The use of T. delbrueckii Biodiva TD291 produced improvements in the total terpenes content in Syrah wines in sequential fermentations with S. cerevisiae, as well as other commercial strains such as Zymaflore Alpha (Laffort, Bordeaux, France) and Prelude (CHR Hansen, Hørsholm, Denmark), even in musts with a high sugar content (29 • Brix, Table 2) [2].
A pure culture of L. thermotolerans Concerto (CHR Hansen, Hørsholm, Denmark) produced an increase in nerol and terpinen-4-ol contents in the initial stages of fermentation of Sauvignon Blanc and Syrah musts [21]. In sequential fermentations with S. cerevisae, it showed a high production of farnesol, geraniol, α-ionene, and cosmene in Sauvignon Blanc wines [14], as well as terpenes in general in Riesling wines, contributing to a higher valuation of Riesling typicity and aromatic quality during the sensory analysis (Table 2) [16] and demonstrating β-glucosidase activity in L. thermotolerans [42].
Metschnikowia pulcherrima also showed β-glucosidase activity. A pure culture of M. pulcherrima Flavia MP346 produced linalool at the beginning of the fermentative process in Syrah musts and wine (Table 2) [21].
Another species capable of releasing terpenes is Hanseniaspora vineae. H. vineae T02/5AF, in monoculture, produced a high amount of α-terpineol in Macabeo wines [37], in accordance with the highest preference and highest floral score obtained in sensory analysis.
One aspect to consider with respect to efficiency in the release of terpenes is that their aromatic precursors must enter the yeast cell. As such, the permeability of the cell membrane plays an important role. Permeability can be induced, for example, by applying treatments that modify its structure, increasing porosity [43], thereby improving the diffusion of precursors into the yeast, where enzymes can release the odorant terpenes.
More studies can be conducted to identify more non-Saccharomyces strains with a high capacity to release terpenes, in which treatments are also applied to improve their cellular permeability and thus their capacity to hydrolyze terpenic precursors. Among the possible treatments, ultrasound and pulsed electric fields, among others, may be evaluated [31,44,45], although it will also be necessary to optimize the intensity of the treatment to not affect the viability of the treated yeasts.
The β-lyase enzyme of T. delbrueckii cleaves the precursor Glut-3SH [23], although it has not shown the capacity to cleave Cys-3SH. Conversely, S. cerevisiae showed the capacity to synthesize Cys-3SH [51]. T. delbrueckii may also synthesize Cys-3SH from Glut-3SH [23]. In addition, S. cerevisiae can metabolize both Glut-3SH and Cys-3SH. Thus, in sequential fermentations, T. delbrueckii may convert the precursor Glut-3SH into 3SH and into Cys-3SH, which, in addition to the contribution of Cys-3SH from the must, increases the availability of Cys-3SH to be metabolized later by S. cerevisiae, contributing to an increase in 3SH [23,48], as well as 3SHA, because S. cerevisiae also has increased acetylation capacity (conversion of 3SH into 3SHA by acetyltransferase alcohol) [50,52].
These results demonstrate the adequate biocompatibility between T. delbrueckii Zymaflore Alpha and S. cerevisiae, which can improve the overall supply of aromatic thiols in the wine. Additional works with other commercial strains will help to confirm the synergy between these two species in high-thiol varieties, such as Sauvignon Blanc, and to improve the aromatic profile in varieties considered neutral, such as Verdejo.
An interesting result in Verdejo wines was obtained by Ruiz et al. [25] through sequential fermentation of Metschnikowia pulcherrima NS-EM-34 (reported as pre-commercial by the authors) with both S. cerevisiae Viniferm Diana and Viniferm Revelacion (both from Agrovin, Alcázar de San Juan, Spain). A higher content of 4MSP was obtained through both sequential fermentations as well as the highest scores in Verdejo typicity in the sensory analysis (Table 2), demonstrating β-lyase activity in M. pulcherrima [49].
Other species with the capacity to produce significant quantities of thiols, especially 3SH and 3SHA in Sauvignon Blanc wines, are L. thermotolerans and Pichia kluyveri, respectively [53,54]. However, this capacity is not evident in commercial strains.
Finally, similar to the metabolism of terpenes [43], the conversion of thiol precursors is dependent on the absorption capacity of the yeast through its cell membrane and subsequent cleavage to the corresponding odorant thiols [50]. As such, treatments to improve this absorption capacity will contribute to improving the varietal aromatic profile of wines, both at the terpenes and thiols levels as mentioned in Section 4.1, for example, in wines produced from Verdejo grapes in which 3SH, 3SHA, and 4MSP contribute to the typicity of this grape variety [25].

Biocompatibility between Commercial Non-Saccharomyces and Saccharomyces
Antagonism between certain strains of M. pulcherrima and other yeasts, including S. cerevisiae, was also reported due to the killer activity of M. pulcherrima [55] and the production of the pulcherrimin pigment [56]. However, the commercial strains of M. pulcherrima used in the studies reported in Table 2 showed a good compatibility with S. cerevisiae, including synergistic effects on the reduction in the alcohol degree (M. pulcherrima AWRI Obsession [24]; M. pulcherrima NS-EM-34 [25]; M. pulcherrima Flavia MP346 [13,27]), improving the production of metabolites of oenological interest, such as aromatic thiols (M. pulcherrima NS-EM-34 [25]), glycerol (M. pulcherrima Flavia MP346 [13]), and esters (M. pulcherrima AWRI Obsession [24]), and improving the foam quality of sparkling wines (M. pulcherrima Flavia MP346 [8]), among others.

Multi-Starter Inoculations with Commercial Non-Saccharomyces
The multi-yeast consortia represent 21% of the current market of non-Saccharomyces yeasts ( Figure 1c and Table 1). One of these products is commercialized under the brand Melody (CHR Hansen, Hørsholm, Denmark). This multi-starter has the capacity to reduce the alcohol degree by up to 0.6% v/v in Syrah wines (Table 2), in addition to improving the content of 2-phenylethanol and esters [2].
Since these results were obtained at the laboratory level, their validation under real winemaking conditions is necessary, with a more complex yeast community, associating the results with sensory evaluations, and thus verifying the impact of this strategy on the varietal aromatic profile of the wine.
The effect of ternary fermentations was also investigated to improve the acidity of Airén wines [27]. Co-inoculation of L. thermotolerans/M. pulcherrima Flavia MP346 and S. cerevisiae increased the concentration of lactic acid, reduced the alcohol degree, increased the content of higher alcohols, and led to a higher acidity assessment and better overall sensory valuation (Table 2). This provides a strategy to improve the quality of wines produced from the Airén grape variety, which is produced in the central-south zone of Spain, where high temperatures can cause accelerated accumulation of sugars and low acidity during ripening of the grape berries.

Non-Saccharomyces as Biocontrollers: Compatibility with Saccharomyces
Non-Saccharomyces yeasts can act as biocontrol agents against contaminating microorganisms. In the market, we identified eight products with this characteristic for dosage in the pre-fermentative phase, either directly on the grape, during crushing, or in the must before fermentation, or in the early stages of fermentation. These products include pure cultures or blends of two species (Table 1 and S1).
Simonin et al. [57] conducted an industrial-level study in two different wineries, adding the yeast Torulaspora delbrueckii BBMV 3FA5 (AEB Group, Brescia, Italy) to musts from Aligoté white grape during crushing. After 24 h, they inoculated the yeast S. cerevisiae. In parallel, they used the traditional method of addition of SO 2 as the control. Three hours after adding T. delbrueckii, population decreases in Hanseniaspora, Metschnikowia, Debaryomyces, Candida, Wickerhamomyces, Malassezia and Cryptococcus were observed, possibly due to the killer activity previously reported in T. delbrueckii [58]. Regarding biocompatibility, the addition of T. delbrueckii BBMV 3FA5 had no effect on S. cerevisiae since the fermentation developed normally.
Although T. delbrueckii showed an antimicrobial effect, the results indicate that it can be used only to partially replace SO 2 , since this additive, in addition to its antimicrobial effect, is the most effective antioxidant agent used to protect wine from oxidation. An alternative may be a combined use of both treatments.
Another species with biocontrol potential is Pichia kluyveri, although its killer activity has not been evidenced in the commercial strains currently available ( Table 1). The first killer toxin described was P. kluyveri 1002, with high activity at pH values between 3.8 and 4.0, which is within the usual vinification range. Its mechanism of action is based on the permeabilization of the target yeast membrane [59]. The other killer toxin, Pkkp, has an optimal pH of 4.0-4.5, with the cell wall of the target yeast being its main receptor, although its mechanism of action is not fully known [60]. As such, the selection of P. kluyveri strains with commercialization potential should consider its killer activity.
It is therefore necessary to identify a larger number of biocompatible S. cerevisiae and non-Saccharomyces strains. Through mixed fermentations, the advantages of both genera can be exploited to improve the aromatic profile of wine, in addition to the killer activity against contaminating microorganisms, partially reducing the doses of SO 2 , consistent with the current trend in the consumption of foods with a lower content of chemical additives [61].

Safety of Commercial Non-Saccharomyces Yeasts: Production of Toxic Metabolites
Certain commercial non-Saccharomyces strains release a high amount of amino acids [16]. These amino acids can provide yeast assimilable nitrogen (YAN) to S. cerevisiae during mixed fermentations [31]. However, the high production of amino acids, such as histidine, phenylalanine, or tyrosine, is related to the synthesis of biogenic amines, which pose a risk to the health of the consumer.
One strategy to reduce these amines is the identification of yeast strains that, in addition to providing metabolites of oenological interest, have a low capacity to produce amines. L. thermotolerans in mixed fermentations with S. cerevisiae have produced amounts of biogenic amines similar to or lower than those produced by S. cerevisiae (<0.85 and <0.50 mg/L for putrescine and histamine, respectively) [15], which, in both cases, is below the limits considered as critical (2.0 mg/L for histamine) [62].
However, most biogenic amines are produced during malolactic fermentation (MLF). The use of yeasts such as Schizosaccharomyces pombe is an appropriate alternative to decrease the malic acid content [4,29], reducing the risk of the synthesis of biogenic amines by malolactic bacteria. As such, sequential fermentations were implemented with L. thermotolerans Concerto (CHR Hansen, Hørsholm, Denmark) and a non-commercial strain of S. pombe to balance the contents of lactic and malic acids to desirable levels, as well as ensure a low production of biogenic amines, mainly histamine, up to four times lower than in the same wine produced by the traditional process of AF + MLF [63].
P. kluyveri FrootZen (CHR Hansen) in sequential fermentation with S. cerevisiae produces high amounts of phenylethylamine in Sauvignon Blanc wines [14]. This amine is derived from phenylalanine, such that more studies are needed to identify non-Saccharomyces strains producing these amines as well as to establish protocols to reduce the production of amines and their precursors.
Regarding other toxic compounds, S. pombe showed capacity to degrade urea (urease activity), for example, in Tempranillo wines produced by sequential fermentation together with L. thermotolerans Concerto (CHR Hansen, Hørsholm, Denmark) [17], with residual levels of urea 97% lower than the traditional process of AF + MLF (Table 2), thus reducing the risk of further synthesis of the carcinogenic agent ethyl carbamate, derived from urea [64].

Critical Appreciation and Conclusions
The reviewed studies demonstrated the biocompatibility between the commercial non-Saccharomyces strains and S. cerevisiae.
Since most of the results were obtained at the laboratory level, they must be validated on a larger scale in which the knowledge of performance and interactions between species allows the establishment of strategies that take advantage of the potential of these commercial yeasts for the production of more complex wines with higher aromatic quality.
One aspect to highlight is the potential of non-Saccharomyces yeasts to ferment in monoculture. Some commercial strains of T. delbrueckii can reach up to 13% v/v ethanol [7]. That is, wines can be produced only with this yeast, which can also produce high amounts of 2-phenylethanol, isoamyl acetate [7], and phenethyl propanoate [21]. Another yeast that has high fermentation capacity in monoculture is H. vineae, reaching 11.9% v/v of ethanol [26], high levels of 2-phenylethyl acetate and other esters, and lower contents of acetoin [37].
Mixed fermentation of S. cerevisiae with commercial strains of M. pulcherrima show potential to reduce the ethanol content by up to 1.0% v/v, in addition to improving the glycerol content and aroma [13,24,25]. L. thermotolerans mainly produces improvements in the acidity [16,17,27], in addition to improving the contents of phenethyl propionate and 2-phenylethanol [2,21].
Ternary fermentations [27] also improve wine aroma and acidity, and reduce the alcohol degree, especially in wines produced from grape varieties cultivated in warm regions, in addition to exploiting the potential of multi-starter commercial products (Table 2), which stand out mainly in their capacity to reduce alcohol content and increase the aromatic compounds [2].
Another aspect of interest is the potential of non-Saccharomyces yeasts to maintain or improve varietal aromas in aromatic grapes, or increase the release of odorants from neutral varieties. Several authors have reported this capacity in commercial strains of T. delbrueckii, L. thermotolerans, M. pulcherrima, and H. vineae [20,23,25,37,42,43,48,49].
Another area of interest is the production of sparkling wines. The few studies that applied these commercial yeasts reported improvements in the contents of glycerin and esters, a decrease in volatile acidity, and an increase in the quality of foam, highlighting T. delbrueckii and M. pulcherrima [8,12].
Therefore, the potential of these yeasts to produce differentiated wines is high. Biocompatibility between species is a fundamental aspect to obtain all the described improvements. In addition, the use of non-Saccharomyces killer species biocompatible with Saccharomyces facilitates the implantation of species of interest against undesirable microorganisms, including at the industrial level [57]. The market offers non-Saccharomyces strains with this characteristic, especially for use in the pre-fermentation stage ( Table 1). The bioprotective effect of these strains also has the potential to partially replace the use of SO 2 [45,57].
Another aspect that should be addressed in future studies is the effect of non-Saccharomyces yeasts on tannins, which, at the moment, has only been addressed using commercial strains of T. delbrueckii, L. thermotolerans, and M. pulcherrima [2], producing a decrease in tannins content in Syrah wines (Table 2).

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/fermentation7030171/s1, Table S1: Summary of commercial non-Saccharomyces yeasts datasheets available on the market. Information obtained from the website of the companies that commercialize them ("commercial website") for winemaking.
Author Contributions: Conceptualization, writing-review and editing, R.V.; writing-original draft preparation, A.G.-C. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by Universidad Privada del Norte (Peru), project UPN-20201003 "Non-Saccharomyces yeasts: potential applications at the industrial level".