**1. Introduction**

Non-*Saccharomyces* yeasts have attracted increasing attention in recent years, with several studies providing evidence of their impact on the organoleptic characteristics and chemical-physical stability of wines. Despite their large intraspecific biodiversity, non-*Saccharomyces* yeasts often show species-specific metabolic features that contribute to the specific imprint of the resulting wines, when inoculated in mixed fermentation with *Saccharomyces cerevisiae* [1]. For instance, among the non-*Saccharomyces* yeasts, those belonging to the species *Torulaspora delbrueckii* result in the production of low volatile acidity, high terpenols, and 2-phenylethanol when utilized in mixed culture with *S*. *cerevisiae* [2–7]. In addition, the release of higher concentrations of thiols, with consequent increase of varietal

> 67

characters, was reported for *T. delbrueckii*/*S*. *cerevisiae* mixed starters [8,9]. *Starmerella bacillaris* (synonym *Candida zemplinina*) contributes to reduce the amount of acetic acid in mixed fermentation with *S*. *cerevisiae* [5,10–12]. Moreover, this yeas<sup>t</sup> is usually characterized by high glycerol production [11,13–17] and low ethanol yield [16,18–20] making it an interesting tool to increase the wine sweetness and modulate the ethanol content. *Lachancea thermotolerans* strains produce lactic acid during the alcoholic fermentation causing a decrease of wine pH while reducing its volatile acidity [4,21–24]. Moreover, an increase of 2-phenylethanol, glycerol, and polysaccharides in mixed fermentation *L. thermotolerans*/*S*. *cerevisiae* was reported [4,21,22,25]. Regarding *Metschnikowia pulcherrima*, some studies showed its high β-glucosidase activity [26–28] with consequent increase of volatile terpene content from glycosylated flavorless precursors present in grapes. Moreover, because of a high β-lyase activity, yeasts belonging to the species *M*. *pulcherrima* release high quantity of varietal thiols from grape precursors conjugated to cysteine or glutathione [29,30]. In the last few years there has been also a renewed interest in yeasts belonging to the genus *Schizosaccharomyces*. Indeed, besides reducing malic acid in grape juice and/or wine, these yeasts produce high quantities of pyruvic acid [31,32] and polysaccharides during the course of alcoholic fermentation [33–36], positively contributing to the chemical-physical stability of wines. Finally, among yeasts typically considered as potential spoilage, those belonging to the genus *Zygosaccharomyces* have also started to attract attention [37,38]. In particular, the species *Z. florentina* contributes to increase esters and glycerol concentration when used in co-culture with *S*. *cerevisiae*, thus producing wines with higher floral notes and lower perception of astringency [39].

A Scopus database search with the combination of terms "wine and non-*Saccharomyces*" as query statement to highlight the relevant literature in the last decade, indicates that an increasing number of peer-reviewed publications have considered the use of non-*Saccharomyces* yeasts as starters together with *S*. *cerevisiae*. In particular, of a total of 458 peer reviewed scientific articles published from 2010 to 2019, the average number of publications/year on this topic was 26 during the period in between 2010 and 2014, and reached 65 in the following years (2015–2019). It is worth pointing out that most of these publications refer to laboratory scale fermentations. In particular, considering publications starting from 2015, 76% of the fermentations were carried out at the laboratory scale (50% in up to 1 L, 15% in 1.2–5 L and 11% in 10–20 L). Instead, pilot plant and industrial scale fermentations, that regarded 24% of the trials, are still quite limited. Of these, 18% were carried out in grape must volumes ranging from 30 to 200 L (pilot plant fermentations) and 5% in 700 to 1000 L (industrial scale fermentations). One fermentation was performed in a 100,000 L vessel.

Moreover, while 12% of the publications starting from 2015 describe fermentations carried out in synthetic media, the majority of works report on the utilization of different grape varieties to evaluate the impact of the non-*Saccharomyces* yeasts on the chemical and physical characteristics of the relevant wine. Among these, Shiraz, Sauvignon Blanc, Barbera, Cabernet Sauvignon, Chardonnay and Merlot were the most frequently utilized. Few articles (4%) describe mixed fermentations in Sangiovese grape must despite the importance of this grape variety that in Italy represents 90% of total world Sangiovese vineyard area (http://www.oiv.int/public/medias/5888/en-distribution-of-theworlds-grapevine-varieties.pdf). In order to avoid any metabolic interference by other microorganisms, half of the studies evaluated the impact of pure and mixed starters on the final wine by using sterile grape juice.

Indeed, laboratory scale fermentation and synthetic media or sterile grape juice are important conditions to evaluate the specific metabolic traits of the non-*Saccharomyces* yeasts included in the mixed starter and also to establish their possible interactions with *S*. *cerevisiae*. However, the results obtained under these conditions are likely far away from those obtainable under technological conditions, also due to the unpredictability of the interactions that the inoculated starters may establish with wild grape must microflora.

Based on these observations, in the present work seven different non-*Saccharomyces*/*S*. *cerevisiae* mixed starters were inoculated in Sangiovese grape must at the pilot plant scale and their impact on the final product was evaluated through chemical and sensory analyses of the resulting wines after stabilization. Sangiovese is the most widely planted red grape variety in Italy, particularly in Tuscany where it represents the obligatory variety in the production of wines with a protected and guaranteed designation of origin (DOCG) such as Chianti Classico and Brunello di Montalcino.

#### **2. Material and Methods**

## *2.1. Yeast Strains*

Seven non-*Saccharomyces* strains from the yeas<sup>t</sup> culture collection of the Department of Agriculture, Food, Environment and Forestry (DAGRI, University of Florence, Italy) and of the Department of Life and Environmental Sciences (DiSVA, Polytechnic University of Marche, Ancona, Italy) were used (Table 1). The yeas<sup>t</sup> strains were sub-cultured on YPD (1% yeas<sup>t</sup> extract, 2% peptone, 2% glucose, 2% agar) at six months intervals, and maintained at 4 ◦C.

**Table 1.** Origin of the seven non-*Saccharomyces* strains and the commercial strain of *S. cerevisiae* used in this study.


a Department of Life and Environmental Sciences of the Polytechnic University of Marche (Italy), b Department of Agriculture, Food, Environment and Forestry, University of Florence (Italy). c Lallemand Inc. (Montreal, QC, Canada).

The strains reported in Table 1 were isolated from grapes and musts of different origins and characterized for their enological performances in mixed fermentations carried out in grape juice at laboratory scale [4,37,38]. A commercial *S*. *cerevisiae* starter, Lalvin EC1118 (Lallemand Inc., Montreal, QC, Canada), was used as reference strain and for comparison determination.

#### *2.2. Pilot Scale Fermentation*

The fermentation trials were carried out in 100 L steel tanks containing 70 L of Sangiovese grape must with the following characteristics: pH 3.66, 234 ±7 g/<sup>L</sup> sugars, 4.0 g/<sup>L</sup> total acidity (as tartaric acid), 1.2 g/<sup>L</sup> malic acid. Non-*Saccharomyces* yeasts were inoculated in 12 L filtered sterilized commercial red grape must, aliquoted within 2 L flasks, each containing 1.5 L and grown for 48-h at 25 ◦C under shaking conditions (150 rpm). Cell concentration was determined by microscope counting. Each tank was inoculated with 10<sup>7</sup> cell/mL of the non-*Saccharomyces* yeas<sup>t</sup> strain. After 3 days of fermentation, *S*. *cerevisiae* EC1118 was inoculated as active dry yeas<sup>t</sup> (ADY) at the final concentration of 10<sup>7</sup> cell/mL. Control trials were inoculated with 10<sup>7</sup> cell/mL of *S*. *cerevisiae* EC1118. Skin cap was punched down twice a day and fermenting must was sampled during the fermentation process immediately after inoculation (T0) and 3, 5, and 10 days after inoculation (T3, T5, T10, respectively) to evaluate the evolution of the yeas<sup>t</sup> populations as viable cell counts and to determine the residual sugars. Alcoholic fermentation was monitored by periodically measuring the density by a double scale hydrometer (density and Baumé). All trials were fermented at 25 ◦C, in duplicate. After completion of fermentation, the wines were naturally fined by three successive rackings over a month at 16–18 ◦C and added with SO2 up to 100 mg/<sup>L</sup> before bottling (0.75 L cork-capped glass bottles).
