Sustainable Exploitation of Wine Lees as Yeast Extract Supplement for Application in Food Industry and Its Effect on the Growth and Fermentative Ability of Lactiplantibacillus plantarum and Saccharomyces cerevisiae

Abstract

Wine lees, the residue left behind after racking or bottling of wine, are predominantly composed of dead yeast cells, ethanol, phenolic compounds, and tartrates. Yeast extract (i.e., commercial yeast extract), a highly nutritious powder derived from commercially cultivated yeast biomass, is commonly used in nutrient media as a nitrogen source. In the context of by-product valorization, wine lees could potentially be used to produce a substitute for commercial yeast extract (CYE). In our study we investigated the growth and fermentative ability of two major winemaking microorganisms, Lactiplantibacillus plantarum and Saccharomyces cerevisiae, in culture media containing a wine lees yeast extract (WLYE) and a CYE. The effects of yeast extract type, concentration, and initial cell concentration (y₀) on key kinetic parameters—maximum specific growth rate (μ_max), lag phase duration (λ), and maximum cell concentration (y_max)—were evaluated. For L. plantarum, the results showed that using a WLYE led to similar kinetic parameters to those obtained with a CYE, with λ being unaffected by y₀ in samples containing a WLYE. For S. cerevisiae, simultaneous addition of both yeast extracts led to increased μ_max values (up to 0.136 h⁻¹) compared to individually added yeast extracts, although this negatively affected λ and y_max. Current research on wine lees is mainly focused on using them as a substrate to produce valuable metabolites through fermentation, overlooking the potential industrial applications of the nutrient-rich autolysate. The findings of this study appear promising for the holistic valorization of wine lees, contributing towards the concepts of sustainability and circular economy.

1. Introduction

Grapes (Vitis vinifera L.) are a widely recognized crop with a history of cultivation and diverse consumption spanning millennia. They are cultivated globally, with a production of 79.4 million tons in the year 2022 (47.4% dedicated to wine grapes), while the wine production for that year reached an estimated 25.8 billion liters [1]. Notably, the global wine sector is continually growing, and it is anticipated to attain a market value of 343.6 billion USD by the end of year 2027 [2]. In summary, the extensive cultivation of grapes and their subsequent conversion into wine play a significant role in the agro-industrial economic landscape.

Wine lees, characterized as the sedimentary residue found in vessels post-fermentation or during wine storage (EEC regulation No. 337/79), is a composite material comprised of both solid and liquid components. The solid fraction contains precipitates like microbial biomass (yeast and bacterial cells) and insoluble carbohydrates settled at the tank’s base [3]. On the other hand, the liquid fraction primarily consists of the fermented medium, which, in the case of wine lees, is wine itself, containing notable amounts of ethanol, phenolic compounds and organic acids [3,4]. Approximately 2–6% of the total wine produced and 14–25% of the winery by-products are estimated to be constituted of wine lees [5], with around 60 kg of wine lees generated per ton of crushed grapes [6]. Based on the above, it can be concluded that more than 2 million tons of wine lees were accumulated globally in 2022 [7,8]. The effective utilization and management of such materials has a direct impact on global environmental stewardship and sustainability. Interestingly, sustainability issues regarding the wine sector have been targeted by many researchers, having performed numerous environmental analyses, such as a carbon footprint analysis [9].

Unless deliberately employed, for instance during wine aging on lees to enhance sensorial attributes [10], the interaction between lees and wine can result in the migration of undesirable flavors. In an attempt to solve the waste disposal problem, wine lees have been used as a nutritional supplement in animal feed; however, given their lower nutrient content this application is limited. This limitation is likely attributed to the increased concentration of polyphenols linked with proteins, which greatly affects their availability, or the presence of toxic compounds (resulting from the treatment of residues), which subsequently accumulate in yeast lipids [11]. Typically, lees are treated together with other sediments in expensive treatment facilities [12]. Regarding their revalorization, wine lees have been traditionally utilized for extracting their key components (ethanol, organic acids, and phenolic compounds), while more contemporary applications include their use as a feed additive in broiler chicken diets, as a substrate for fermentation to harvest products of metabolic activity, as a supplement for microbial growth, and for the simultaneous production of energy and microalgae biomass for biofuels [13,14,15]. Additionally, the extraction of cell wall polysaccharides (mannoproteins and β-glucans) from wine lees has also been investigated, as these compounds can be utilized in numerous applications, e.g., as mouthfeel enhancers in wine, or as emulsifiers and fat replacers in food products [16]. Nonetheless, uncontrollable disposal of fermentation-derived yeast biomass poses a severe environmental threat (e.g., water pollution, soil contamination, waste management issues), especially considering the large volumes that are being industrially produced [17].

Yeast extract and peptone are the most commonly used nitrogen sources for fermentations. The utility of yeast extract is particularly noteworthy, given its high purine and pyrimidine content, along with B-complex vitamins, peptides, amino acids, and minerals. The primary source for industrial-scale yeast extract production is considered to be high-protein yeast species within the genus Saccharomyces, specifically cultivated for this intention [4]. However, it is important to note that both peptone and yeast extract come with a significant cost, constituting nearly 30–40% of the overall expenses in fermentation procedures [4]. The reason for this is that during its commercial production different processes must be used, such as mechanical lysis of cell membranes, autolysis, plasmolysis at high NaCl concentrations, permeabilization, precipitation with chitosan, enzymatic or chemical treatment, and protein solubilization [18]; these processes are quite costly from an economic point of view. On an economic basis, the market price of yeast extract is estimated to be around 7 USD kg⁻¹ for industrial applications (e.g., food and beverage industry, cosmetics, animal feed, biotechnology) and between 70 and 120 USD kg⁻¹ for laboratory applications [19]. By contrast, the cost on untreated lees is merely 0.008 EUR kg⁻¹ [18]. Despite the high cost, there is a substantial and expanding global market for yeast extract, experiencing a remarkable 46% increase from 1.1 billion USD in 2015 to 1.6 billion USD in 2022. The market is also anticipated to continue growing at a compound annual growth rate (CAGR) of 5.5% between 2023 and 2032 [20].

Residual carbohydrates and nitrogen compounds in lees, coupled with metabolites and yeast-derived compounds, e.g., vitamins beneficial for other microorganisms like lactic acid bacteria, positions lees as a promising supplement comparable to yeast extract or corn steep liquor for the production of culture media [21]. However, research on utilizing wine lees as a nitrogen source for microorganisms often overlooks isolating the nutrient-rich autolysate, which has potential industrial applications. Instead, it primarily focuses on using the solid residue of wine lees as a substrate for producing useful metabolites through fermentation, such as poly(3-hydroxybutyrate) [22], lactic acid [4,12,23] and xylitol [18,21]. In a recent study, wine lees were used to promote lactic acid bacteria growth and malolactic fermentation in wine [24], but specific studies evaluating its effect on growth of wine bacteria and yeast are rare in the literature. The present study focuses on the use of wine lees yeast extract powder, derived from the winemaking of Greek, indigenous Assyrtiko grapes, as a novel economical substitute for commercial yeast extract in culture media. The effect of yeast extract concentration and initial cell concentration on the main growth parameters of two commonly used winemaking microorganisms, i.e., Saccharomyces cerevisiae (alcoholic fermentation) and Lactiplantibacillus plantarum (malolactic fermentation), was evaluated and compared for both yeast extracts. Moreover, the fermentative ability of these microorganisms was investigated using a model grape must, with each microorganism grown on culture media containing one of the yeast extracts (i.e., commercial yeast extract or wine lees yeast extract).

2. Materials and Methods

2.1. Materials

Wine lees originating from white winemaking of the Assyrtiko variety of grapes were collected during the 2020 vintage and were kindly provided by Ktima Gerovassiliou (Epanomi, Greece). Fermentation of the grape must was performed by the indigenous microflora, while no malolactic fermentation took place. Aliquots of wine lees were stored in airtight containers at −18 °C until further use.

All reagents, chemicals and solvents were purchased from Sigma–Aldrich Chemie GmbH (Taufkirchen, Germany), unless stated otherwise.

2.2. Production of Wine Lees Yeast Extract Powder

Wine lees yeast extract was produced from the autolysate of wine lees. More specifically, wine lees were initially centrifuged (2716× g, for 10 min) (Universal 16A, Hettich, Tettlingen, Germany) to obtain the solid fraction and subsequently an aqueous suspension was prepared by mixing the solid fraction with distilled water. Adjustment of pH was performed by the addition of sodium hydroxide (1 M) and the suspension was placed in an orbital shaker (Lab-Line Orbit Environ-Shaker, LAB-LINE INSTRUMENTS INC., Melrose Park, IL, USA) to perform autolysis. During this procedure the following conditions were used: temperature, 45.1 °C; pH, 5.01; solid/liquid ratio, 120.94 g L⁻¹; shaking, 200 rpm. The supernatant was harvested after centrifugation (2716× g, 10 min) (Universal 16A, Hettich, Tettlingen, Germany) after 24 h and was further concentrated using a rotary vacuum evaporator (Rotovapor R114, Waterbath B480, Büchi, Flawil, Switzerland) thermostatted at 30 °C, until the volume in the flask was significantly reduced (1/3 of initial value). Finally, concentrated wine lees yeast extract was dried using a freeze dryer (Alpha 1–2 LDplus, Martin Christ Gefriertrocknungsanlagen GmbH, Osterode am Harz, Germany) paired with a rotary vane pump (RZ 2.5, Vacuubrand GmbH + Co. KG, Wertheim, Germany) for 48 h to obtain a fine powder of wine lees yeast extract (WLYE) for further use in microbiological experiments.

2.3. Inoculum Preparation

A bacterial and a yeast strain were evaluated in the present study. Lactiplantibacillus plantarum 2035 (from the collection of the Laboratory of Food Microbiology and Hygiene, Department of Food Science and Technology, Aristotle University of Thessaloniki, Greece) and Saccharomyces cerevisiae NCYC187 (National Collection of Yeast Culture, Norwich, UK) were used in experiments. Their activation and growth were performed from glycerol stock cultures, which were stored at −80 °C. The strains were activated by two consecutive, 24-h subcultures at 30 °C in De Man, Rogosa and Sharpe (MRS) and Yeast Malt (YM) broths, for L. plantarum 2035 and S. cerevisiae NCYC187, respectively. The inoculum cell concentration was determined by plating on nutrient agar and was further standardized by serial dilution to 10⁷ CFU mL⁻¹.

2.4. Test Culture Media, Growth Monitoring and Cell Enumeration

For the growth of L. plantarum 2035, an MRS broth was formulated as follows: 20 g L⁻¹ glucose, 10 g L⁻¹ peptone, 10 g L⁻¹ meat extract, 3 g L⁻¹ sodium acetate, 2 g L⁻¹ dipotassium phosphate, 2 g L⁻¹ triammonium citrate, 1 g L⁻¹ polysorbate 80, 0.2 g L⁻¹ magnesium sulfate, 0.05 g L⁻¹ manganese sulfate and consequently 3 distinct culture media were formulated adding 5 g L⁻¹ of either a commercial yeast extract (CYE) (Oxoid Ltd., Hampshire, UK), a WLYE (See Section 2.2), or an equal mixture of both (

Table 1. Different nutrient broths examined for the growth of Lactiplantibacillus plantarum 2035 and Saccharomyces cerevisiae NCYC187 and their coding.

Culture Media Base	Type of Yeast Extract	Concentration (g L−1)	Coding
Lactiplantibacillus plantarum 2035
MRS broth	CYE	5	C
	WLYE	5	WL
	CYE + WLYE	2.5 + 2.5	C + WL
Saccharomyces cerevisiae NCYC187
Yeast Carbon Base	CYE	0.5	C0.5
	CYE	1	C1
	CYE	2	C2
	CYE	5	C5
	CYE	10	C10
	WLYE	1	WL1
	WLYE	2	WL2
	WLYE	5	WL5
	WLYE	10	WL10
	CYE + WLYE	0.5 + 5	C0.5 + WL5

CYE: commercial yeast extract; WLYE: wine lees yeast extract.

). For the growth of S. cerevisiae NCYC187, a Yeast Carbon Base (YCB) was used (DIFCO LABORATORIES, Detroit, MI, USA), which contained all essential nutrients and vitamins necessary for the cultivation of yeasts except a source of nitrogen. To formulate the nutrient broths, an aqueous suspension of YCB was prepared at 11.7 g L⁻¹ and a CYE and/or a WLYE was added at different concentrations (0.5, 1, 2, 5 and 10 g L⁻¹) (Table 1).

The growth of cultures was monitored by measuring optical density (OD) using a 96-well microtitration plate (BRANDplates, transparent, F-bottom, BRAND GmbH + Co. KG, Wertheim, Germany) and an absorbance microplate reader (BioTek 800TS, BioTek, Winooski, VT, USA). Initially, microtitration plate wells were prefilled with 180 μL of the culture media. The first column was inoculated with 20 μL of 10⁷ CFU mL⁻¹, while decimal dilutions of the inoculum were obtained across the microtitration plate by transferring 20 μL from one well to the other. For the fifth dilution, 20 μL were discarded to maintain an equal culture volume (180 μL). Therefore, different initial concentrations were obtained across the microtitration plate (10²–10⁶ CFU mL⁻¹). OD measurements were performed at standard time intervals (10 min) at 610 nm and 30 °C, for a total period of 48 h, during which the stationary phase for all five decimally diluted cultures was observed. The microplates were agitated for 10 s prior to each OD measurement. Positive control samples (inoculum and culture media without yeast extract) and negative control samples (uninoculated culture media) were also tested for means of comparison and validation. All experiments were performed in triplicate. A calibration curve was also performed to transform the obtained OD values to log(CFU) mL⁻¹.

2.5. Data Analysis

Maximum specific growth rate (μ_max, h⁻¹), lag phase duration (λ, h), initial cell concentration (y₀, log(CFU) mL⁻¹), and maximum cell concentration (y_max, log(CFU) mL⁻¹) were estimated by fitting the primary model of Baranyi and Roberts [25] to the growth data using the DMFit add-in (v. 3.5, Institute of Food Research, Norwich, UK, https://combase.errc.ars.usda.gov/, Accessed on 21 May 2024) in Microsoft Excel (v. 2408, Microsoft, Redmond, WA, USA), which provided an excellent fit for the vast majority of our experimental results (R² from 0.919 to 0.996, Supplementary Material, Tables S1 and S2). In addition, the standard error (SE) and coefficient of determination (R²) of the fit were calculated.

2.6. Evaluation of Fermentative Ability

For the evaluation of fermentative ability, both microorganisms were grown on culture media with either a commercial or wine lees yeast extract, as explained in detail below. For the growth of S. cerevisiae, the strain was activated by three consecutive, 24-h subcultures at 30 °C in a synthetic medium containing: 40 g L⁻¹ glucose, 4 g L⁻¹ yeast extract (commercial or wine lees), 1 g L⁻¹ (NH₄)₂SO₄, 1 g L⁻¹ KH₂PO₄, and 5 g L⁻¹ MgSO₄ [26]. For the growth of L. plantarum, the strain was activated by two consecutive, 24-h subcultures at 30 °C in an MRS broth (with either commercial or wine lees yeast extract). The activated strains were consequently inoculated in a synthetic grape juice (pH = 3.5), as proposed by Wang et al. [27] with slight modifications (60 g L⁻¹ glucose, 60 g L⁻¹ fructose, 1 g L⁻¹ yeast extract, 2 g L⁻¹ (NH₄)₂SO₄, 0.3 g L⁻¹ citric acid, 5 g L⁻¹ L-malic acid, 5 g L⁻¹ L-tartaric acid, 0.4 g L⁻¹ MgSO₄, 5 g L⁻¹ KH₂PO₄, 0.2 g L⁻¹ NaCl, 0.05 g L⁻¹ MnSO₄). S. cerevisiae was inoculated at 10 g L⁻¹ (wet weight) and two consecutive 7-day fermentations were performed; for L. plantarum, 1 mL of inoculum (~2 × 10⁹ CFU mL⁻¹) was added to 10 mL of synthetic grape juice and fermentation was carried out for a total of 7 days.

Evaluation of fermentative ability was based on the kinetic study of reducing sugars while, in the case of S. cerevisiae fermentations, final ethanol content was also determined.

2.7. Determination of pH, Reducing Sugars and Ethanol Content in Fermentation Media

The pH values were determined by portable, electronic pH-meter (SensoDirect pH 110, AQUALYTIC, Dortmund, Germany). Reducing sugars (RS) were determined using the DNS (3,5-dinitrosalicylic acid) spectrophotometric method and expressed as g glucose equivalents L⁻¹, as proposed by Miller [28]. Ethanol content was calculated using a Dujardin-Salleron ebulliometer (Laboratories Dujardin-Salleron, Arcueil Cedex, France), which is based on the difference in boiling points of water and wine.

2.8. Statistical Analysis

The effects of independent variables on growth parameters were analyzed using the statistical software Minitab (v. 15.1.20) (Minitab Inc., State College, PA, USA). An analysis of variance (ANOVA) and Tukey’s post hoc test were used to identify the significant effects and interactions between variables. Significance level was set at p < 0.05.

3. Results and Discussion

3.1. Growth Kinetics of Lactiplantibacillus plantarum 2035

According to the growth data obtained, the highest maximum specific growth rate (μ_max) was observed for the sample with a commercial (C) yeast extract (0.105 ± 0.008 h⁻¹). By comparison, the highest μ_max obtained in experiments where wine lees (WL) or a mixture of both a commercial and a wine lees yeast extract were used (C + WL) were 0.074 ± 0.001 h⁻¹ and 0.066 ± 0.016 h⁻¹, respectively (Figure 1). Statistical analysis indicated that the effect of the interaction between yeast extract source and initial cell concentration (y₀) was significant (p < 0.05), therefore the change in μ_max for a level of yeast extract source depends on the level of y₀. Because of that, the influence of each factor on the specific growth rate cannot be distinguished, as they are involved in significant higher-order interactions. However, regarding the yeast extract source, it appears that mean μ_max values of C and WL samples did not significantly differ, suggesting that the simultaneous addition of these two different yeast extract sources had a negative impact on the growth rate of L. plantarum 2035. This can possibly be attributed to the different amino acid profiles of these extracts which, in combination, may hinder the adaptation of L. plantarum 2035. Interestingly, combining yeast extracts has been proven to lead to both an increased and decreased growth, depending on the microorganism [29].

Studies on the growth of L. plantarum report different μ_max values, as there is no consistency between studies regarding the experimental factors, e.g., strains, y₀, enumeration method, culture medium. For example, Horn et al. [30] studied the substitution of yeast extract with cod (Gadus morhua L.) viscera hydrolysates in MRS, for the growth of L. plantarum NC8 using the Bioscreen C automatic turbidimetric system. The reported μ_max values were in the range 0.35–0.41 h⁻¹; however, to produce the hydrolysates, the substrate was pretreated with enzymes. In a different study, Silva et al. [31] focused on the growth of L. plantarum ATCC 8014 in MRS and reported μ_max values of 0.01–0.65 h⁻¹ for temperatures ranging between 4 and 30 °C. These values appear higher compared to the μ_max values found in the present study, although it should be noted that Silva et al. [31] carried out growth experiments with y₀ = 9 log(CFU) mL⁻¹ and used the plating method to enumerate cells. Nonetheless, studies reporting lower μ_max values compared to the ones found in the present study also exist, such as Di Cagno et al. [32], who reported a μ_max value of 0.04 h⁻¹, for the growth of L. plantarum DPC2741 in reconstituted skim milk, with y₀ = 6 log(CFU) mL⁻¹.

Regarding the effects on the lag phase duration (λ), it should be noted that all samples inoculated with 6 log(CFU) mL⁻¹ presented immediate growth, therefore λ could not be calculated (Figure 2a). However, at different y₀ levels it appears that increasing y₀ led to significantly lower λ (p < 0.05), which has already been proven by other studies [33]. Similarly with the effect of yeast extract source on μ_max, it should be noted that means of λ values of C and WL samples for all y₀ levels did not significantly differ, however the correlation between y₀ and λ is more noticeable in the cases of C and C + WL samples, in which cases a considerable decrease of λ is observed at higher y₀. This finding can be explained by the better adaptation of L. plantarum at higher cell concentrations in media containing a CYE (either alone or in combination with a WLYE). Regarding the addition of wine lees yeast extract in an MRS broth (C + WL), it can be concluded that it promoted the growth of L. plantarum 2035, as this sample presented the lowest λ (0.59 ± 0.12 h at y₀ = 5 log(CFU) mL⁻¹). To the best of our knowledge, there is no literature on the growth of L. plantarum in MRS supplemented with wine lees yeast extract; however, data agree with the existing literature on conventional culture media, with Silva et al. [31] reporting λ values for the growth of L. plantarum ATCC 8014 in MRS in the range 0.3–127.9 h for temperatures ranging between 4 and 30 °C, and Di Cagno et al. [32] reporting a λ value of 6.2 h for the growth of L. plantarum DPC2741 with y₀ = 6 log(CFU) mL⁻¹ in reconstituted skim milk.

Figure 2b indicates the effect of different yeast extract sources and y₀ on the maximum cell concentration (y_max) of L. plantarum 2035. It can be observed that C samples presented the highest overall y_max values (maximum 6.68 ± 0.07 log(CFU) mL⁻¹ at y₀ = 6 log(CFU) mL⁻¹), although not significantly higher (p > 0.05) than C + WL samples. Statistically, the effect of the interaction between yeast extract source and y₀ was significant (p < 0.05), indicating that the change in y_max for a level of yeast extract source depends on the level of y₀. Therefore, individual effects should not be taken into consideration, as they play a part in significant higher-order interactions. By comparison, Di Cagno et al. [32] reported a y_max of 6.73 log(CFU) mL⁻¹ for the growth of L. plantarum DPC2741 with y₀ = 6 log(CFU) mL⁻¹ in reconstituted skim milk.

The general picture emerging from the aforementioned results indicates that MRS with a commercial yeast extract provided the best overall results for L. plantarum growth, with the highest μ_max and y_max. In addition, C + WL samples presented the lowest λ, suggesting that L. plantarum cells present more active growth in culture media with individual yeast extract sources. However, no experiments with defined media have been conducted on the growth of the specific L. plantarum strain used in this study, thus the amino acid requirements for growth are not known. It is also interesting that C and WL samples did not produce significantly different overall results regarding μ_max and λ, while C + WL presented the lowest μ_max and y_max. This finding may be attributed to the distinct amino acid profiles of these nitrogen sources, which may not act synergistically when in combination, and may delay microbial adjustment. Thus, it can be concluded that the use of wine lees yeast extract resulted in the satisfactory growth of L. plantarum and can possibly be proposed as an alternative to a commercial yeast extract, considering its abundance, its low relative cost, and the benefits from by-products valorization.

3.2. Growth Kinetics of Saccharomyces cerevisiae NCYC187

According to the growth data (Figure 3), the highest specific growth rate (μ_max = 0.232 ± 0.022 h⁻¹) was observed for the sample with a commercial yeast extract at 10 g L⁻¹ (C10). Focusing on the effect of yeast extract addition (0.5–10 g L⁻¹) within C samples, a positive correlation emerges, while statistical analysis performed indicated that this effect is significant (p < 0.05). Meanwhile, it appears that increase of y₀ from 2 to 4 log(CFU) mL⁻¹ did not significantly increase the μ_max in most cases. Regarding the WL samples, the same pattern emerges with increased μ_max values for higher yeast extract concentrations. Regarding the mixed sample (C0.5 + WL5), it can be concluded that utilizing both yeast extract sources in the same culture media significantly increased μ_max compared to the individually used yeast extracts, from 0.094 h⁻¹ (C0.5) and 0.129 h⁻¹ (WL5) to 0.136 h⁻¹ (mean values). The current literature on the growth kinetics of S. cerevisiae strains reports different μ_max values, which are highly dependent on the culture medium used, yeast strain, temperature, nitrogen concentration, and other factors. By comparison, García-Ríos et al. [34] reported a μ_max value of 0.089–0.193 h⁻¹ for four different commercial wine yeasts belonging to the S. cerevisiae species, during secondary modelling of data on optimum growth temperature.

Linear regression analyses were performed to investigate and model the linear relationship between μ_max (response), yeast extract concentration, and y₀ (independent variables) for WL samples. The regression equation is estimated to be:

$$\begin{gathered} {\mu }_{max}\left(h^{-1}\right)=0.0159+0.0170\times yeastextractconcentration\left(g{L}^{-1}\right) \\ +0.00164\times y_0\left(\mathit{log}\left(CFU\right){mL}^{-1}\right) \end{gathered}$$

An analysis indicated that only the association between the response and the yeast extract concentration was statistically significant (p < 0.05), in contrast with initial cell concentration (p > 0.05), while the coefficient of determination (R²) was calculated at 91.3% (modified R² = 90.9%). Interestingly, according to the Histogram of Residuals (Figure 4a) no evidence of skewness or outliers can be observed, as there are no long tails or bars distinctly far away from others. In addition, the Normal Probability Plot of Residuals (Figure 4b) suggests a normal distribution of residuals, as they appear to follow a straight line, while no evidence of nonnormality, skewness, or unidentified variables exists.

Figure 5 depicts the effect of different yeast extract sources and y₀ on the lag phase duration (λ) of S. cerevisiae NCYC187. The results suggest that y₀ strongly influence λ, with higher y₀ leading to lower λ. Regarding the lowest λ observed, WL1 and WL2 samples with y₀ = 6 log(CFU) mL⁻¹ presented immediate growth, thus λ values were so low that they could not be calculated. However, for the remaining sample comparisons (C5-WL5, C10-WL10), there were no significant differences between means, indicating that wine lees yeast extract consists a compatible source of nitrogen for the growth of S. cerevisiae. Focusing on the effect of yeast extract concentration, it appears that within the C samples, means of λ values for 2 (11.75 h), 5 (11.62 h) and 10 g L⁻¹ (11.80 h) did not significantly differ, indicating that an increase of yeast extract addition did not favor the growth acceleration of S. cerevisiae. The same pattern can also be observed for the WL samples, as 1 g L⁻¹ (9.05 h) and 2 g L⁻¹ (9.74 h) samples presented the overall lowest mean λ values, even among all the samples, thus demonstrating that substitution of a commercial yeast extract with a wine lees yeast extract seems promising. Regarding the mixed sample (C0.5 + WL5), mean λ values 13.51 h) were significantly higher in comparison with both C0.5 (9.97 h) and WL5 (11.66 h) samples, proposing that the addition of two different nitrogen sources may not act in favor of S. cerevisiae growth. In a similar manner with μ_max, the literature on the lag phase duration of S. cerevisiae reports different values, as λ significantly depends on strain, initial cell concentration and culture medium. For example, Arroyo-López et al. [35] reported λ values of 0.89–8.39 h for the growth of S. cerevisiae T73 with different temperature, pH and sugar concentrations, and y₀ = 4.5 log(CFU) mL⁻¹.

Linear regression analyses were performed to investigate and model the linear relationship between λ (response), yeast extract concentration, and y₀ (independent variables) for both yeast extract sources. Regarding the C samples, the regression equation is estimated to be:

$$\begin{gathered} \lambda \left(h\right)=-0.055+0.0743\times yeastextractconcentration\left(g{L}^{-1}\right) \\ +3.75\times y_0\left(\mathit{log}\left(CFU\right){mL}^{-1}\right) \end{gathered}$$

An analysis indicated that the association between the response and both independent variables was statistically significant (p < 0.05), while R² was calculated at 96.8% (modified R² = 96.7%), indicating that the model fits the data well, and a significant proportion of variation in the response data can be explained by the independent variables. This finding is also supported by the Residuals versus Fits Plot (Supplementary Material, Figure S1), in which residuals appear to be randomly scattered around zero and no fanning or uneven spreading of residuals across fitted values occurs, implying that there is constant variance.

Regarding the WL samples, the regression equation is estimated to be:

$$\lambda \left(h\right)=-4.34+0.318\times yeastextractconcentration\left(g{L}^{-1}\right)+4.50\times y_0\left(\mathit{log}\left(CFU\right){mL}^{-1}\right)$$

Analysis also indicated that the association between the response and both independent variables was statistically significant (p < 0.05), while R² was calculated at 95.4% (modified R² = 95.1%), supporting the fact that the model adequately predicts the value of the response variable for any combination of values of the independent variables. However, in this case, the Normal Probability Plot of Residuals (Supplementary Material, Figure S2) indicates that residuals are not normally distributed, as a curve in the tails is observed, which evinces skewness. Moreover, deviation from a normal distribution can also be detected by the lack of a straight line.

Figure 6 presents the effect of yeast extract source and y₀ on y_max of S. cerevisiae. On the one hand, it is obvious that the C samples presented higher y_max values at lower y₀ (maximum 7.92 ± 0.04 log(CFU) mL⁻¹ at C10, y₀ = 2 log(CFU) mL⁻¹). However, on the other hand, the effect of y₀ was not significant (p < 0.05) and did not follow a specific pattern in the WL samples, with the highest mean y_max values being observed in WL5 (7.54 log(CFU) mL⁻¹) and WL10 (7.56 log(CFU) mL⁻¹) samples, while being lower than those of the C samples. The addition of both yeast extract sources in a culture medium (C0.5 + WL5) resulted in significantly lower mean y_max values (7.24 log(CFU) mL⁻¹) than both individual samples (C0.5: 7.57 log(CFU) mL⁻¹, WL5: 7.54 log(CFU) mL⁻¹), a fact that, similarly with the effect on λ, suggests that the addition of two different nitrogen sources may not act in favor of S. cerevisiae growth.

To summarize the above findings, the first conclusion that can be drawn is that the addition of both yeast extract sources negatively affected λ and y_max of S. cerevisiae, in comparison with individual samples (C0.5 and WL5). Contrary to L. plantarum, the superior performance of a commercial yeast extract for S. cerevisiae growth is more apparent; however, the use of wine lees yeast extract cannot be considered discouraging, as growth is evident and results for μ_max and λ are comparable, especially for WL10 samples. As a matter of fact, the potential advantages of added wine lees yeast extract have recently been proven in grape must fermentation. Onetto et al. [36] applied mechanical, enzymatic, and accelerated autolysis on wine lees, reaching a 14-fold increase in yeast assimilable nitrogen. Subsequent fermentation of grape must by Saccharomyces cerevisiae AWRI 796 and supplementation with up to 10% v/v wine lees autolysate decreased fermentation time and modified the aromatic profile of produced wines.

It should be noted that the literature does not contain any studies on the growth of the specific S. cerevisiae strain used in this study, thus the amino acid requirements for growth are not known. Moreover, the enhanced nitrogen content of a commercial yeast extract should also be taken into consideration, as this material is appropriately engineered for such purposes through a variety of nitrogen-enhancing processes, such as enzyme-assisted hydrolysis. Given the competitive advantage of lower cost, wine lees yeast extract cannot be excluded as a potential yeast extract source for the growth of S. cerevisiae, as it can be used even in higher concentrations, without affecting the overall cost.

3.3. Fermentative Ability

Regarding the fermentation ability of L. plantarum grown on culture media containing either a commercial or wine lees yeast extract, Figure 7 shows the consumption of sugars during the fermentation of synthetic grape juice. It can be concluded that minor differences can be observed in sugar consumption between the samples. A slight increase in pH values was detected and can be explained by the biotransformation of malic acid into lactic acid, with the latter being less acidic [37]. It should also be highlighted that the pH increase was similar in both samples, which, in combination with the similarities in the sugar consumption rate, may also indicate that the ability of L. plantarum to metabolize malic acid has been unaffected by the yeast extract type used in the growth media. This conclusion can be supported by the fermentative behavior of L. plantarum, which is able to consume sugars and malic acid simultaneously [38].

According to the experimental design, S. cerevisiae stock culture was activated on culture media containing either a commercial or wine lees yeast extract and was further used in fermentations of synthetic grape juice (Figure 8). The results indicated that biomass grown on a commercial yeast extract consumed sugars at a slightly higher rate, especially during the first days of fermentation. However, this difference was equilibrated in the second fermentation, during which no significant differences in sugar consumption rate of the samples was observed for most fermentation intervals. Nonetheless, even during the first fermentation, any differences observed can be considered negligible. Interestingly, the final ethanol content of the samples did not appear to be affected by the yeast extract type, with biomass grown on a commercial yeast extract producing 7.0 ± 0.1 and 7.2 ± 0.1% v/v of ethanol, while biomass grown on wine lees yeast extract produced 6.8 ± 0.2 and 7.0 ± 0.1% v/v of ethanol, for the first and second fermentation, respectively. These findings suggest that S. cerevisiae can successfully grow in media containing wine lees yeast extract, without affecting its ability to ferment sugars present in grape juice.

4. Conclusions

The present study focused on the utilization of wine lees to produce a novel, low-cost yeast extract for the growth of the two major winemaking microorganisms, i.e., Saccharomyces cerevisiae (alcoholic fermentation) and Lactiplantibacillus plantarum (malolactic fermentation). Regarding the growth of L. plantarum, the results indicated that the use of wine lees yeast extract did not significantly affect any growth parameter in most cases (Figure 1 and Figure 2); whereas, on the growth of S. cerevisiae, only lag phase duration was not negatively affected (Figure 5). In addition, modelling of certain parameters regarding the growth of S. cerevisiae, using different yeast extract concentrations and initial cell concentrations, serves as a first step towards understanding the potential interactions between microorganism and yeast extracts and applying this knowledge industrially. Our findings support the possible use of wine lees yeast extract as an alternative to commercial yeast extract, however higher concentrations should and can be considered, given the low cost of wine lees. Although, to the best of our knowledge, there is no literature on the effect of wine lees yeast extract on the growth parameters of these microorganisms, the findings from this study on wine lees yeast extract contribute to this broader narrative of transforming waste into valuable resources and highlight the importance of sustainable practices in modern industries. Future studies should include other strains and microorganisms, as well as winemaking trials.