Exploitation of Yeasts with Probiotic Traits for Kefir Production: Effectiveness of the Microbial Consortium

Abstract

Kefir is a fermented milk made by beneficial lactic acid bacteria and yeasts inoculated as grains or free cultures. In this work, five yeast strains with probiotic aptitudes belonging to Candida zeylanoides, Yarrowia lipolytica, Kluyveromyces lactis, and Debaryomyces hansenii species were assessed in a defined consortium, in co-culture with a commercial strain of Lactobacillus casei, in order to evaluate the yeasts’ fermentation performance during kefir production, using different milks. The concentration of each yeast was modulated to obtain a stable consortium that was not negatively affected by the bacteria. Furthermore, all yeasts remained viable for five weeks at 4 °C, reaching about 8.00 Log CFU in 150 mL of kefir, a volume corresponding to a pot of a commercial product. The yeasts consortium showed a suitable fermentation performance in all milks, conferring peculiar and distinctive analytical and aromatic properties to the kefirs, confirmed by a pleasant taste. Overall, the panel test revealed that the cow’s and sheep’s kefir were more appreciated than the others; this evaluation was supported by a distinctive fermentation by-products’ content that positively influences the final aroma, conferring to the kefir exalted taste and complexity. These results allow us to propose the yeasts consortium as a versatile and promising multistarter candidate able to affect industrial kefir with both recognizable organoleptic properties and probiotic aptitudes.

1. Introduction

A healthy diet coincides with the actual consumer trend to prefer superfoods presented as enriched, fortified, or pre- and probiotic foods. Functional food provides health benefits in two ways: by providing essential nutrients (e.g., proteins, carbohydrates, fatty acids, vitamins, and minerals) and by containing live probiotic micro-organisms [1,2] or their metabolites (peptides, bacteriocins, organic acids, and so on), which may interact with multiple key targets in metabolic pathways, regulating the proliferation, differentiation, inflammation, apoptosis, angiogenesis, and metastasis of cells [3]. Recently, it was demonstrated that fermented dairy products consumed at recommended levels represent an excellent source of high-quality protein and calcium, improve well-being through enhanced immune system function, and provide antimicrobial effects and antioxidant properties protecting against free radical damage [4,5].

Kefir, a fermented milk-based dairy product, has recently been recognized as a beverage with functional characteristics [6,7,8]. Within the past few years, kefir has gained wide consumer acceptance in Europe and in Italy, aided by its image as a ‘healthy’ food. It differs from other fermented products in its production, which involves a complex mixture of lactic acid and acetic acid bacteria and lactose-fermenting and non-fermenting yeasts in a symbiotic ecosystem [9]. The total number of micro-organisms in a fermented milk should be at least 10⁷ colony forming units (CFUs)/mL, and the yeast number should not be less than 10⁴ CFU/mL [10]. They are responsible for the production of bioactive metabolites such as organic acids, antimicrobial peptides, polysaccharides, and bacteriocins, involved in the suppression of the growth of spoilage and pathogenic micro-organisms, immunostimulation, and antiviral effects [11], aspects that reinforce the beneficial value of the kefir.

Despite its importance as a food of broad significance, it has not yet been included in the Italian national database. Regarding its nutritional value and according to the Codex Alimentarius [12], a typical kefir should contain at least 2.7% protein, 0.6% lactic acid, and less than 10% fats. Kefir is mainly produced using cow’s milk, but recent reports indicate that goat, sheep, buffalo, camel, or even donkey milks can also be used to prepare kefir, with small nutritional differences [13,14].

The production of milk-based kefir could be done in two ways: traditional/homemade kefir and industrial kefir. Traditional manufacturing provides for the use of kefir grains, a polysaccharide matrix that contains inherent microflora producing the matrix itself. The origin of the kefir grains is unknown; they are traditionally passed from generation to generation among families. This is the cause of the unstable microbial composition [14], and thus is difficult to replicate at the industrial scale, although its beneficial properties are undoubted [15,16,17]. The various products available on the market are produced without granules, using commercial selected mixed cultures that represent a good compromise between tradition and innovation [11,18].

Although numerous health-promoting benefits of kefir have been described in the literature, many questions still need to be answered such as the microbial kefir composition [19,20], especially with regard to yeast species, and their list declared on the label, often enclosed in the expression “yeasts”. In this scenario, the present study proposes a selected microbial consortium of yeasts with probiotic aptitudes as starters to be used in co-culture with commercial lactic acid bacteria for the production of kefir by the industrial method. It was proposed to reinforce the probiotic value of kefir, to ensure the standardization of the final product with specific sensorial properties, and to be useful in different cattle milk fermentation.

2. Materials and Methods

2.1. Probiotic Bacteria, Yeast Consortium in Different Milks

A Lactobacillus casei Shirota commercial strain (LbS) (isolated from Yakult beverage, Yakult Honsha Co., Ltd., Tokyo, Japan) was used as a unique probiotic bacterium, while a mixture of five wild yeast strains belonging to the Department of Life and Environmental Sciences (DiSVA) of the Polytechnic University of Marche (Ancona, Italy) was used as probiotic yeasts. In particular, two strains of Debaryomyces hansenii (strains 36 and 78) and one strain each of Candida zeylanoides (strain 13), Yarrowia lipolytica (strain 92), and Kluyveromyces lactis (strain 80) were evaluated for use during kefìr production. The yeast strains were previously isolated from artisan dairy environments and cheeses, and they were reported because of their preliminary probiotic aptitudes [21]. LbS was cultivated at 37 °C under anaerobic conditions using MRS agar (LIOFILCHEM^® S.r.l., Roseto degli Abruzzi, Teramo, Italy). The yeasts were cultivated at 25 °C using YPD agar (10 g/L yeast extract, 20 g/L dextrose, 20 g/L peptone, and 18 g/L agar). The bacteria and yeasts were stored for short periods at 4 °C on MRS agar or YPD agar, respectively, while they were stored for long periods at −80 °C using MRS broth and YPD broth containing glycerol as a cryo-preservant. Different types of whole pasteurized milk were used: cow’s milk (C), highly digestible cow’s milk (HD), sheep’s milk (S), and goat’s milk (G). Their main nutritional characteristics are reported in

Table 1. Main nutritional characteristics of whole cow’s milk (C), lactose-free cow’s milk (HD), goat’s milk (G), and sheep’s milk (S) used as different matrices for kefír fermentation.

Milks	Nutritional Values for 1 L of Product
	Energy (Kcal)	Total Fat (g)	Saturated Fat (g)	Total Carbohydrates (g)	Sugars (g)	Fibers (g)	Protein (g)	Salt (g)	Calcium (g)
Whole cow’s milk	640	36	26	49	49	0	31	1	1.2
Lactose-free Cow’s milk	340	35	24	51	51	0	32	1	1.2
Goat’s milk	450	16	12	44	44	0	32	2	1.2
Sheep’s milk	1030	69	48	51	51	0	53	3	1.8

.

2.2. Preliminary Kefir Fermentation for Yeast Consortium Setup and Evaluation of Yeasts’ Coexistence with LbS

Whole pasteurized cow’s (C) milk was used to carry out the preliminary fermentation. The bacteria and yeasts used were precultured as follows: bacteria were grown in MRS broth at 37 °C for 24 h, while the yeasts were grown in YPD broth at 25 °C for 24 h. Afterwards, LbS was centrifuged at 14,000 rpm for 5 min, and the yeasts were centrifuged at 4000 rpm for 5 min to collect the microbial biomass. The bacteria and yeasts were washed twice with 9 g/L NaCl saline solution before use. Whole pasteurized cow’s milk was aliquoted (150 mL) into sterile glass jars, inoculated with the micro-organisms, and sealed with a sterile screw cap. All jars were inoculated with LbS at 1 × 10⁸ CFU/mL followed by the inoculation of each yeast separately and with a yeast consortium (D. hansenii 36 and 78, C. zeylanoides 13, Y. lipolytica 92, and K. lactis 80) at a final concentration of 1 × 10³ CFU/mL. The 5 log orders of magnitude inoculum ratio of LbS compared with the yeasts was selected on the basis of previous works on kefir production [22]. A sample inoculated only with LbS was used as a control to evaluate the contribution of yeasts. Kefir fermentation was carried out at 25 °C for 24 h, and then the jars were stored at 4 °C for 5 weeks to simulate the shelf life of the product, following the procedure reported by Grønnevik et al. [22]. The trial was carried out in triplicate. The presence of yeasts and their growth kinetics, the coexistence of all the yeasts species in the consortium, and the coexistence of yeasts and LbS were monitored through viable cell counts.

2.3. Yeasts and LbS Detection and Growth Kinetics

The detection and growth kinetics of yeasts (in pure culture and in the consortium) and LbS were monitored through viable cell counts using different media: Wallerstein Lab (WL) nutrient agar (LIOFILCHEM^® S.r.l., Roseto degli Abruzzi, Teramo, Italy) supplemented with 0.05 g/L chloramphenicol (Sigma-Aldrich, Saint Louis, MI, USA) to suppress bacterial growth was used as differential medium for yeast detection and recognition, while MRS agar medium (LIOFILCHEM^® S.r.l., Roseto degli Abruzzi, Teramo, Italy) supplemented with 0.02 g/L cycloheximide (Sigma-Aldrich, Saint Louis, MI, USA) to suppress yeasts growth was used for LbS detection. Each sample was subjected to appropriate decimal serial dilutions and was spreading on both media described above. Plates containing WL agar were incubated at 25 °C for 3–5 days, while plates containing MRS agar were incubated at 37 °C for 5 days under anaerobic conditions (using specific jars) before micro-organism enumeration. Monitoring of yeasts and LbS was done at the end of the fermentation (24 h at 25 °C) and every week for five consecutive weeks (storage period), with the aim of monitoring any changes in the microbial population composition of the kefir obtained.

2.4. Improved Yeast Consortium-LbS in Milks Fermentation

Based on the analysis of the data regarding the preliminary kefir fermentation, an improved yeast consortium regarding the modulation of each yeast species’ concentration was tested with the purpose of testing a yeast consortium in which each species coexists without the absolute supremacy of any of them. In particular, the inoculum of D. hansenii 36 and 78 was increased to 5 × 10³ CFU/mL, while the inoculums of C. zeylanoides 13 and Y. lipolytica 92 were reduced to 5 × 10² CFU/mL. K. lactis 80 was maintained at 1 × 10³ CFU/mL and the LbS inoculum was maintained at 1 × 10⁸ CFU/mL. Each yeast was also tested in pure culture with LbS, maintaining the concentration reported above. A sample inoculated only with LbS was used as a control to evaluate the contribution of yeasts. All fermentation trials were carried out in milks as previously described (C, HD, S, and G milk), and inoculated jars were placed at 25 °C for 24 h (fermentation step) and then transferred at 4 °C for 5 weeks (storage period/shelf-life period). All trials were carried out in triplicate and analyzed for yeasts and LbS growth kinetics and their coexistence through viable cell counts, as reposted for the preliminary kefir fermentation. Moreover, at the end of the storage period, all samples were analyzed for the main analytical characters, by-products of fermentation, and sensorial properties with the aim to evaluate the yeasts’ contribution to the kefir, when compared with the control trial (only LbS), and their fermentation performance using different milks.

2.5. Main Analytical Characters and By-Products of Fermentation of Kefir

At the end of the storage period, the kefirs fermented by yeasts (whose concentration of each species has been remodeled) and LbS were subjected to analytical chemistry measurements, including pH, lactose, acetic acid, lactic acid, and ethanol. Specific enzyme kits (Megazyme International Ireland, Wicklow, Ireland) were used for lactose, acetic acid, lactic acid, and ethanol determinations. The pH of the samples was determined using a pH metre (Meter S400, Mettler, Toledo).

Regarding the main by-products of fermentation, acetaldehyde, ethyl acetate, n-propanol, isobutanol, amyl alcohol, isoamyl alcohol, and acetoin content were analyzed. Each compound was quantified by the solid-phase microextraction (HS-SPME) method following the procedure described by Dertli and Çon [23] with some modifications: 5 mL of homogenized kefir, 1 g of NaCl, and 162 mg/L of 1-pentanol (used as a standard) were placed into a 20 mL glass vial and sealed. The sample was mixed at 120 rpm for 10 min at room temperature. Then, the vial was placed at 55 °C with the HS-SPME fibre exposed to the headspace for 50 min. The compounds were desorbed by inserting the fibre Divinylbenzene/Carboxen/Polydimethylsiloxane (Sigma-Aldrich, St. Louis, MI, USA) into a Shimadzu gas chromatograph GC injector (GC-2014; Shimadzu, Kyoto, Japan). The column was a Zebron ZB-WAX Plus polyethylene glycol column (30 m × 0.32 n mm ID × 0.25 µm film-thickness; Phenomenex, Torrance, CA, USA). Each compound was quantified by comparison with external calibration curves.

2.6. Sensorial Analyses of Kefìr

The kefirs were subjected to sensorial analysis by serving tasters disposable plastic cups each containing 10 mL of refrigerated (10 ± 1 °C) samples. Ten trained tasters expressed their score using a 10-point hedonic scale (10 = like extremely; 0 = dislike extremely) regarding overall acceptability texture, smell, viscosity, acidity, bitterness, and sweetness for each sample. The sensory analysis was conducted similarly to that reported by Yıldız-Akgül et al. [24]. Specifically, the tasting was conducted blindly using a panel of 12 tasters, 6 males and 6 females aged between 24 and 60, habitual consumers of fermented milk.

2.7. Statistical Analyses

Experimental data regarding the main analytical characters and the by-products of fermentation were subjected to analysis of variance (ANOVA). Significant differences were determined using Duncan tests with associated p-values < 0.05. Furthermore, mean values of the same data, normalized to eliminate the influence of hidden factors, were analyzed by principal component analysis (PCA). Both statistical analyses were carried out using JMP^® 11 statistical software (Statistical discovery from SAS, New York, NY, USA).

3. Results

3.1. Preliminary Kefir Fermentation to Test Yeast Consortium–LbS Coexistence

The growth kinetics of the inoculated lactic acid bacteria and yeasts in the kefir trials prepared with cow’s milk are given in Figure 1. As expected, the LbS inoculated at 10⁸ CFU/mL increased more than one log order of magnitude within 24 h of fermentation at 25 °C and then remained constant throughout the storage period at 4 °C. The presence of yeasts inoculated both separately and together in a consortium did not modify the growth kinetics of LbS, which remained comparable to that of pure culture (Figure 1a, top).

Regarding all inoculated yeasts (separately and in the consortium without distinction between the species), they showed a similar trend (Figure 1a, bottom), characterized by a slight increase in growth during the first 24 h at 25 °C, followed by an equally slow but constant increase of approximately 2 Log orders during the storage period, with the exception of the two D. hansenii strains in pure culture with LbS: D. hansenii 78 decreased by 2 Log orders after the fermentation step, then remained almost constant for the next 2 weeks of storage at 4 °C. Finally, it gradually decreased until the third week, when it was not detectably present. In contrast, the other D. hansenii 36 showed a completely different trend, keeping the initial concentration almost constant until the fourth week of storage and then decreasing by one log order.

Figure 1b showed in detail the trend of each inoculated yeast within the consortium. The low viability of the two strains of D. hansenii (which were not distinguished from each other) was confirmed: their concentration decreased by about one Log order at the end of the fermentation, and they were not detectably present starting from the second week of the storage at 4 °C. K. lactis 80 maintained a 10³ CFU/mL concentration after the fermentation step and within 2 weeks of storage, but it was not present after the third week at 4 °C. In contrast, the C. zeylanoides and Y. lipolytica strains showed a constant growing trend, with an increase in their initial concentration of over 4 log orders, reaching a final concentration of about 10⁷ CFU/mL within 5 weeks of storage.

3.2. Assessment of the Improved Yeast Consortium in Different Milks

Figure 2 shows the growth kinetics of the remodulated yeast strain concentrations to compensate for the too low or high vitality of each yeast strain within the consortium during kefir fermentation in different milks. Although the concentration of D. hansenii was increased compared with the preliminary fermentation, this species showed the lowest growth kinetics compared with the other yeasts tested, independent of the type of milk in which they were inoculated. However, comparing the two D. hansenii yeasts, strain 78 showed lower kinetics than strain 36, especially in goat milk, which showed the greatest decline (Figure 2c).

Conversely, Y. lipolytica and C. zeylanoides, despite having been inoculated with 1 order Log less, dominated the other yeasts in the cow’s milk matrix (with and without lactose), exhibiting the same trend (Figure 2a,b). In goat’s milk, they dominated the other yeasts only in the first 24 h of fermentation at 25 °C (Figure 2c), while in sheep’s milk, the most evident effect was a reduction in their initial concentration (Figure 2d).

The remodulated concentrations of D. hansenii (increased), C. zeylanoides, and Y. lipolytica (decreased) reflect an increase in the total concentration of yeasts in the consortium of approximately 2 log orders during the 24 h of fermentation; this concentration then remained constant during storage. The only exception was the sheep’s milk matrix, in which the total concentration of yeasts in the consortium remained constant with respect to the inoculum until the third week and then slowly increased until the end.

3.3. Analytical Characters and Fermentation By-Products of Different Kefir Fermented by the Improved Yeast Consortium–LbS

The results of the chemical determinations of the kefirs fermented by the improved yeast consortium with LbS are shown in

Table 2. Main analytical characters of kefirs from cow’s milk (K_C), lactose-free cow’s milk (K_HD), goat’s (K_G), and sheep’s (K_S) milk obtained by LbS and yeast fermentation, at the end of the storage period (5 weeks). Data means ± standard deviations and values showing different superscript letters (^{a,b,c,d,e,f,g}) within each line are significantly different according to Duncan tests (p < 0.05).

		LbS (Control)	Dh 36	Dh 78	Cz 13	Kl 80	Yl 92	Y. Consortium
K_C	pH	3.95 ± 0.02 ab	3.94 ± 0.02 a	3.92 ± 0.04 a	3.96 ± 0.01 a	3.92 ± 0.01 a	3.96 ± 0.03 a	3.92 ± 0.03 ab
	Lactose % w/w	3.02 ± 0.01 c	3.95 ± 0.14 a	3.74 ± 0.07 a	1.01 ± 0.05 e	1.98 ± 0.04 d	3.88 ± 0.08 a	3.27 ± 0.01 b
	Lactic acid g/L	9.32 ± 0.01 b	8.59 ± 0.03 d	9.52 ± 0.02 a	9.13 ± 0.07 c	8.62 ± 0.03 d	9.39 ± 0.01 b	8.36 ± 0.02 e
	Acetic acid g/L	0.71 ± 0.01 a	0.54 ± 0.01 d	0.62 ± 0.00 c	0.53 ± 0.01 d	0.43 ± 0.01 f	0.49 ± 0.01 e	0.66 ± 0.00 b
	EtOH % v/v	nd	nd	nd	nd	nd	nd	nd
K_HD	pH	3.97 ± 0.03 a	4.02 ± 0.01 a	4.03 ± 0.00 a	4.05 ± 0.01 a	3.97 ± 0.01 a	3.99 ± 0.02 a	4.03 ± 0.01 a
	Lactose % w/w	nd	nd	nd	nd	nd	nd	nd
	Lactic acid g/L	7.95 ± 0.01 cd	8.33 ±0.07 a	8.36 ± 0.04 a	7.95 ± 0.05 d	8.07 ± 0.04 bc	8.07 ± 0.11 b	8.43 ± 0.06 a
	Acetic acid g/L	0.47 ± 0.01 a	0.28 ± 0.01 d	0.40 ± 0.01 b	0.40 ± 0.01 b	0.38 ± 0.01 b	0.32 ± 0.01 c	0.32 ± 0.00 c
	EtOH % v/v	nd	nd	nd	nd	nd	nd	nd
K_G	pH	3.97 ± 0.00 bc	3.99 ± 0.01 bc	3.91 ± 0.00 c	3.96 ± 0.01 bc	3.95 ± 0.01 bc	4.21 ± 0.01 a	4.08 ± 0.02 b
	Lactose % w/w	4.02 ± 0.01 c	4.49 ± 0.04 a	3.45 ± 0.00 e	4.09 ± 0.01 b	4.09 ± 0.03 b	2.69 ± 0.01 f	3.95 ± 0.04 d
	Lactic acid g/L	7.24 ± 0.01 c	8.04 ± 0.07 a	8.04 ± 0.06 a	7.56 ± 0.01 b	7.11 ± 0.01 d	6.22 ± 0.01 e	7.08 ± 0.02 d
	Acetic acid g/L	0.94 ± 0.00 a	0.70 ± 0.02 c	0.66 ± 0.01 d	0.78 ± 0.00 b	0.62 ± 0.00 e	0.56 ± 0.00 f	0.56 ± 0.01 f
	EtOH % v/v	nd	nd	nd	nd	nd	nd	nd
K_S	pH	3.82 ± 0.02 ab	3.87 ± 0.02 a	3.90 ± 0.01 a	3.78 ± 0.03 ab	3.91 ± 0.02 a	3.78 ± 0.00 ab	3.77 ± 0.01 b
	Lactose % w/w	2.09 ± 0.01 e	3.00 ± 0.01 a	2.01 ± 0.01 f	2.20 ± 0.02 d	2.36 ± 0.02 c	2.70 ± 0.01 b	1.86 ± 0.01 g
	Lactic acid g/L	11.76 ± 0.03 c	13.14 ± 0.03 b	10.06 ± 0.00 f	10.45 ± 0.01 e	10.54 ± 0.02 d	14.26 ± 0.00 a	11.73 ± 0.00 c
	Acetic acid g/L	0.53 ± 0.01 c	0.41 ± 0.00 f	0.70 ± 0.01 a	0.49 ± 0.01 d	0.55 ± 0.01 b	0.47 ± 0.00 e	0.40 ± 0.01 f
	EtOH % v/v	nd	nd	nd	nd	nd	nd	nd

nd = not detected.

. After the period of storage, in each milk matrix used as a kefir fermentation substrate, single inoculated yeasts showed significant differences in lactose residue and lactic acid and acetic acid production compared with the control (only LbS). However, the production of each compound also varied among the different milks, while the pH value did not differ as a function of the yeasts or in the different types of milk. Ethanol was not detected in all trials.

In cow’s milk, both D. hansenii strains (Dh 36 and Dh 78) and Y. lipolytica (Yl 92) showed higher residual values of lactose than the control, while C. zeylanoides (Cz 13) and K. lactis (Kl 80) showed the opposite behaviour when fermenting milk in pure culture with LbS. In contrast, in goat’s and sheep’s milks, all yeast strains in pure culture with LbS exhibited a higher residual value of this sugar than the control (LbS), with the only exception being D. hansenii 78 and Y. lipolytica, which showed an inconstant utilization of lactose and consequent lactic acid production. Without distinctions among milks, D. hansenii showed a general increase in lactic acid production, while contrary behaviour was shown by C. zeylanoides and K. lactis. The acetic acid concentration showed a general decrease (each yeast and each matrix), with the only exception being D. hansenii 78 in sheep’s milk.

Within each milk matrix, the comparison between the yeast consortium and single pure cultures of yeast exhibited a constant trend: lactose reduction and lactic and acetic acid production showed significant differences in kefir fermented with the yeast consortium (Table 2).

Although, in this preliminary study, it was difficult to evaluate the relative contribution of each yeast to the analytical characteristics of the kefir obtained after consortium fermentation, these results highlighted the greater complexity deriving from multiple yeast–yeast interactions during the fermentation processes compared with the fermentation carried out by yeast in pure culture. In this regard, principal component analysis (PCA) (Figure 3) showed a clear diversification of the kefir obtained by the yeast consortium with respect to those carried out with pure culture strains, regardless of the fermentation matrix. In particular, the yeast consortium showed a unique and intermediate production of the main analytical compounds and by-products of fermentation compared with single yeasts; it is especially diversified from the other samples for the production of aromatic compounds such as amyl and isoamyl alcohol. These compounds are involved in conferring greater aromatic complexity to the final fermented product such as pleasant vegetable aroma.

The concentrations of the fermentation by-products detected at the end of storage are reported in

Table 3. Main fermentation by-products detected at the end of the storage at 4 °C (5 weeks) in the kefir obtained by whole cow’s milk (K_C), lactose-free cow’s milk (K_HD), goat’s milk (K_G), and sheep’s milk (K_S). Data means ± standard deviations and values showing different superscript letters (^{a,b,c,d,e,f,g}) within each line are significantly different according to Duncan tests (p < 0.05).

Fermentation By-Products (mg/L)		LbS (Control)	Dh 36	Dh 78	Cz 13	Kl 80	Yl 92	Y. Consortium
K_C	Esters
	Ethyl acetate	nd	nd	nd	nd	nd	nd	nd
	Alcohols
	n-propanol	12.00 ± 0.04 a	9.91 ± 0.03 f	10.26 ± 0.03 e	11.70 ± 0.02 b	10.69 ± 0.03 d	10.26 ± 0.02 e	11.40 ± 0.02 c
	Isobutanol	nd	nd	nd	nd	nd	nd	nd
	Amyl alcohol	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.22 ± 0.01 a	0.00 ± 0.00 b
	Isoamyl alcohol	2.83 ± 0.02 a	2.68 ± 0.04 c	0.00 ± 0.00 d	2.68 ± 0.01 c	0.00 ± 0.00 d	2.75 ± 0.01 b	2.70 ± 0.02 c
	Carbonyl Compounds
	Acetaldehyde	0.49 ± 0.01 c	0.00 ± 0.00 e	0.00 ± 0.00 e	0.06 ± 0.01 d	0.00 ± 0.00 e	13.44 ± 0.03 a	2.96 ± 0.03 b
	Acetoin	0.39 ± 0.01 a	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b
K_HD	Esters
	Ethyl acetate	0.11 ± 0.01 f	2.52 ± 0.04 c	3.45 ± 0.03 b	0.00 ± 0.00 g	1.05 ± 0.03 e	1.55 ± 0.02 d	15.97 ± 0.01 a
	Alcohols
	n-propanol	14.88 ± 0.07 b	12.11 ± 0.01 f	12.68 ± 0.04 d	12.42 ± 0.04 e	12.86 ± 0.03 c	18.11 ± 0.01 a	10.44 ± 0.07 g
	Isobutanol	nd	nd	nd	nd	nd	nd	nd
	Amyl alcohol	0.04 ± 0.00 e	2.10 ± 0.02 f	0.00 ± 0.00 f	7.59 ± 0.02 a	2.43 ± 0.02 c	0.00 ± 0.00 f	4.91 ± 0.01 b
	Isoamyl alcohol	2.70 ± 0.02 ab	2.73 ± 0.04 ab	0.00 ± 0.00 c	2.69 ± 0.25 b	0.00 ± 0.00 c	2.82 ± 0.02 a	2.99 ± 0.08 a
	Carbonyl Compounds
	Acetaldehyde	3.99 ± 0.02 d	0.36 ± 0.02 f	1.09 ± 0.02 e	0.00 ± 0.00 g	5.43 ± 0.02 b	4.32 ± 0.03 c	11.11 ± 0.02 a
	Acetoin	nd	nd	nd	nd	nd	nd	nd
K_G	Esters
	Ethyl acetate	13.98 ± 0.07 a	6.66 ± 0.21 d	0.00 ± 0.00 g	0.24 ± 0.01 f	4.04 ± 0.03 e	4.04 ± 0.03 e	8.87 ± 0.15 c
	Alcohols
	n-propanol	15.95 ± 0.14 a	11.92 ± 0.35 c	9.83 ± 0.28 f	10.98 ± 0.14 d	10.18 ± 0.01 e	10.18 ± 0.01 e	12.06 ± 0.12 b
	Isobutanol	nd	nd	nd	nd	nd	nd	nd
	Amyl alcohol	0.01 ± 0.00 c	3.62 ± 0.28 a	0.00 ± 0.00 c	0.00 ± 0.00 c	1.62 ± 0.03 b	1.62 ± 0.03 b	0.00 ± 0.00 c
	Isoamyl alcohol	0.00 ± 0.00 c	2.73 ± 0.14 b	0.00 ± 0.00 c	2.72 ± 0.03 b	2.79 ± 0.02 b	2.79 ± 0.02 b	0.00 ± 0.00 c
	Carbonyl Compounds
	Acetaldehyde	0.00 ± 0.00 c	0.00 ± 0.00 c	0.00 ± 0.00 c	0.00 ± 0.00 c	0.17 ± 0.01 c	0.17 ± 0.01 c	5.97 ± 0.42 a
	Acetoin	28.39 ± 0.07 a	16.79 ± 0.21 b	0.00 ± 0.00 d	0.00 ± 0.00 d	0.00 ± 0.00 d	0.00 ± 0.00 d	1.54 ± 0.01 c
K_S	Esters
	Ethyl acetate	0.00 ± 0.00 b	12.31 ± 0.01 a	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b
	Alcohols
	n-propanol	10.12 ± 0.03 f	13.73 ± 0.02 b	11.19 ± 0.02 e	12.84 ± 0.36 c	15.99 ± 0.16 a	9.29 ± 0.03 g	11.95 ± 0.26 d
	Isobutanol	2.66 ± 0.04 c	0.00 ± 0.00 d	2.69 ± 0.03 bc	2.76 ± 0.03 b	2.95 ± 0.03 a	2.64 ± 0.04 bc	2.65 ± 0.03 c
	Amyl alcohol	0.15 ± 0.01 e	0.54 ± 0.03 b	0.04 ± 0.00 f	0.30 ± 0.03 d	0.39 ± 0.01 c	0.78 ± 0.00 a	0.06 ± 0.00 f
	Isoamyl alcohol	2.73 ± 0.03 c	2.70 ± 0.02 c	3.04 ± 0.03 b	2.73 ± 0.04 c	3.04 ± 0.06 b	3.23 ± 0.01 a	2.74 ± 0.02 c
	Carbonyl Compounds
	Acetaldehyde	0.00 ± 0.00 b	0.74 ± 0.02 a	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b
	Acetoin	0.00 ± 0.00 b	41.67 ± 0.22 a	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b	0.00 ± 0.00 b

nd = not detected. Bold format: distinguish from compounds.

. Overall, kefir fermented with yeasts (both in single culture and in consortium) showed significant differences compared with those inoculated with only LbS. In particular, Y. lipolytica produced the highest amount of acetaldehyde (13.44 mg/L), followed by yeast consortium (2.96 mg/L), in comparison with the other trials in which acetaldehyde almost was not detected in cow’s kefir (K_C). Acetaldehyde positively characterizes the dairy fermented products, conferring a yogurt-like aroma. In lactose-free cow’s kefir (K_HD), yeast consortium produced ethyl acetate in the order of 15.97 mg/L, contrary to all other theses, where the ethyl acetate content ranged from 0.00 to 3.50 mg/L. Meanwhile, in sheep’s kefir (S_K), the unique producer of ethyl acetate was D. hansenii 36 in pure culture with LbS (12.31 mg/L). Acetoin was abundantly produced in goat’s kefir (K_G) by LbS in pure culture (28.39 mg/L), followed by D. hansenii 36 in pure culture with LbS (16.79 mg/L), and slightly produced by yeast consortium (1.54 mg/L); it was completely absent in the other thesis.

Moreover, to evaluate the influence of different milks on the yeast consortium fermentation performance, data regarding analytical characters and fermentation by-products were elaborated in the PCA graphic (Figure 4). The results confirmed that sheep milk has differentiated from the others in isobutanol (together with lactic acid) production, kefir from lactose-free cow’s milk (K_HD) could be distinguished by its amyl alcohol content, while kefir from goat’s milk is differentiated from the others in acetoin production.

3.4. Sensorial Analysis

The preliminary organoleptic evaluation of all kefirs is reported in Figure 5. In general, the kefirs prepared with cow’s and sheep’s milk were positively evaluated by testers. Indeed, the data comparing the overall acceptance properties highlighted kefirs from cow’s and sheep’s milk as preferred, obtaining higher acceptability values than the others, especially for those obtained from LbS and both strains of D. hansenii, which, in Figure 3, were closely located in down quadrants, far from the others. The yeast consortium, regardless of the milk matrix, reached an acceptability showing almost intermediate values.

Overall, texture was the descriptor that obtained high and constant scores (between 6 and 8) for each type of milk, with the exception for yeast consortium, K. lactis, and Y. lipolytica 92 in sheep’s matrix.

Sweetness, viscosity, odor, and bitterness are discordant both in the comparison with the single yeasts and the consortium and between the various types of milk.

The descriptor acidity was closely related to the type of yeast inoculated in each milk matrix. A high score was reached by Y. lipolytica (c.a. 8) in cow’s and sheep’s milk. Similar acidity scores were obtained by D. hansenii 36 in HD milk and LbS and D. hansenii 78 in goat’s milk. On the contrary, C. zeylanoides reached the lowest acidity score in cow’s and HD milks, while the yeast consortium reached the lowest score in goat’s milk.

Regarding the yeast consortium, all evaluated descriptors obtained a similar score in cow’s milk, with and without lactose. In goat’s and sheep’s milk, it showed lower values for bitterness, texture, and viscosity then other trials. An opposite trend was shown for acidity in goat’s and sheep’s milk.

4. Discussion

Yeasts are commonly used in the worldwide food industry, with Saccharomyces and Kluyveromyces being the most representative genera [25]. Although lactic acid bacteria and bifidobacteria are the micro-organisms most widely studied for their probiotic and functional properties, the use of yeast as a bioactive food supplement is gaining relevance [26]. Many advances in the understanding of the beneficial properties of yeast have been presented [27] and the search for new strains and their characterization for food industry applications is an expanding area of investigation. For this reason, in a previous work [21], an isolation campaign in artisan dairies to search for new yeast strains with functional and/or probiotic aptitudes to be applied in industrial fermentation was carried out. In the present work, a yeast consortium was used in coculture with a commercial strain of Lb. casei Shirota to simulate the natural composition of kefir using different milk types.

Among fermented milk-based products, kefir has attracted interest recently owing to reports of its beneficial properties [28]. Traditionally, kefir is produced by fermenting milk with kefir grains, but large-scale industrial production forces the discovery of an alternative solution to the use of grains. With this study, for the first time, the fermentation efficiency and the stability of various yeast strains during the production of kefir were tested at the laboratory scale. Most of the previous studies about the involvement of yeasts in dairy products are focused on the maturation of cheeses [29]. Other studies have focused on the characterization of yeasts isolated from kefir and other fermented milk-based beverages without identifying a consortium of yeasts to be applied for the production of kefir on a large scale. For example, Koutinas et al. [30] reported studies on kefir yeast-based technology for the production of kefir yeast biomass for use as baker yeast, protein livestock feed for animals and food emulsifiers [30], and an economic analysis of the industrial scale-up of whey fermentation to produce drinks that are similar to kefir [31].

More recently, Guzel-Seidim et al. [32] evaluated the biomass increments of kefir biomass (lactic acid bacteria and yeasts) after different treatments with whey protein or inulin, supporting industrial kefir production using natural kefir grains. In this regard, although the definition of kefir cannot ignore the presence of grains, handling them at an industrial level is operatively impossible. The industrial management of kefir grains is difficult both because of the high probability of spoilage culture in the grains and because of the need for management of the exopolysaccharides in the grains themselves [33]. Moreover, some studies have shown that kefir grains cannot be used as starter cultures in industry because the microbiota of kefir grain is unstable and the bacteria and yeast species vary significantly during each refresh [34,35].

Although some commercial lyophilized kefir starter cultures are available for industrial production, they usually have very few yeast species [36]. Therefore, a significant decline in taste, aroma, and textural properties is usually observed because commercial kefir starter cultures do not contain an authentic population of micro-organisms. Indeed, fermented products, such as dairy products, develop their nutritional and organoleptic qualities as a result of the complex metabolic activity of consortia of different micro-organisms, where more than one type of interaction may occur simultaneously [37].

The microbial diversity of kefir described in the literature varies greatly. Recently, Rosa et al. [10] reviewed a complete list of the bacteria and yeasts that have been identified in kefir, estimated to be more than 300. The microbial composition of kefir varies according to several intrinsic or extrinsic factors, such as the origin of the kefir grains, the different techniques employed during processing, the different fermentation temperatures, the type and composition of the milk used, and the storage conditions of the kefir and kefir grains.

Beshkova et al. [38] explained the symbiotic equilibrium among different yeast species in natural kefir grains, and Chen et al. [39] developed a stable multi-inoculum of different starter yeasts using a system of microspheres. In contrast to this goal, our study, through the choice of suitable yeasts and careful investigation of their optimal concentrations, allowed us to obtain a final product comparable with industrial analogues through the exploitation of an initial multiple-species inoculum without the aid of any technology. In this regard, Smid and Lacroix [40] stated that complex microbial consortia perform more complex activities (versatility) and tolerate greater variations in the environment (robustness) than pure cultures.

Generally, in fermented food matrices, simple knowledge about the microbial composition of a community is not enough to understand the way each species or strain contributes to the formation of functional properties. Microbial communities are usually known as stable systems with mutual interactions between the different metabolic networks [41]. To achieve similar features in synthetic microbial consortia for industrial applications, it is necessary to gain a deeper understanding of the allocated functions within the consortium and of the relationships between the microbiota and the process performance.

The yeast consortium proposed here showed a stable viability of each GRAS (generally recognized as safe) yeast and no negative interactions among them. Moreover, the selected probiotic yeasts did not appear to interact negatively with the commercial starter lactic acid bacteria, and they seemed to adapt easily to any substrate [42].

The yeast’s viability was guaranteed until the end of the fifth week of storage, a period that can simulate the real shelf life of a commercial kefir and the aromatic characteristics of the overall kefir. However, an exception was detected for D. hansenii. Although this yeast naturally colonizes dairy environments, in this work, specific strain differences have been highlighted, as already demonstrated by Petersen et al. [29]. Indeed, strain 78, selected by virtue of its proven probiotic aptitudes [15], suffered from the presence of other yeasts in the consortium, showing a slight vitality reduction during the last storage week. Finally, considering the final concentration of all yeasts in the consortium, the value reached at the end of experiment was approximately between 10⁵ and 10⁶ living cells per mL, and thus, in a hypothetical 100 mL serving consumed, the total concentration of 10⁷ is reached. Furthermore, if LbS concentration is also considered, the final quantity of probiotics, understood as the sum of bacteria and yeasts, greatly exceeds the value of 10⁹ per portion, generally considered an adequate probiotic claim.

Regarding the overall acceptability by the panel test, the kefir from cow’s and sheep’s milk was rated the best. The high score was corroborated by quantitative analytical and aromatic results. Indeed, both kefirs showed different but balanced notes of sweetness and acidity, probably owing to the presence of an equal amount of organic acids and aromatic compounds. This is a crucial parameter for determining the sensory character, as also previously reported by Duitschaever et al. [43].

Furthermore, the results obtained in this work show that all of the analyzed parameters (main fermentative characteristics, aromatic profile, and sensory evaluation) are expressed in a balanced way when fermentation is carried out by the yeast consortium. In particular, the results showed that the yeast consortium had a significant effect on higher alcohols in all milks tested, in accordance with Sulmiyati et al. [44]. In particular, acetaldehyde levels were always increased after consortium fermentation, in line with those reported by Grønnevik et al. [22]. Moreover, ethyl acetate and amyl alcohol also increased in HD milk fermented by yeast consortium.

Although at this stage, it is difficult to establish the role of each yeast within the consortium as well as the synergistic relationships that are established among them, the mixed inoculum improves the complexity of the finished product. Indeed, microbial interactions are generally mediated by molecular and physiological mechanisms.

Among the yeast species considered here for kefir production, D. hansenii and K. lactis are considered safe micro-organisms as they are included in the QPS list [45]; Y. lipolytica was recently described by Groenewald and coworkers [46] as a “safe-to-use” organism based on the evidence that they are naturally present on cheese and/or other human foods with positive technological features [47,48]. Furthermore, Yalçin and Ozbas [49] proposed C. zeylanoydes strains originated from traditional cheeses, as starter culture candidates for food production.

In conclusion, this study demonstrated the synergistic effect of a consortium of appropriated selected yeasts that actually contributed to the co-fermentation process with LbS in the production of a kefir. Together, C. zeylanoides, Y. lipolytica, K. lactis, and D. hansenii could be considered as an effective functional multistarter able to ferment different milks to produce, in a standardized and controlled way, a kefir similar to the natural one, overcoming the technological difficulties of the use of grains. Although the results obtained are encouraging, further research needs to be carried out to determine if this yeast consortium can be applied under pilot and/or industrial scales, in order to place on the market a standardized and safe product, similar to artisanal kefir.