Influence of different yeast mannoprotein extracts on the evolution and stability of vitisins

Abstract

In this study, the effects of eight mannoprotein-rich extracts obtained from Saccharomyces and non-Saccharomyces yeasts on vitisin formation and color evolution were evaluated in a model wine system. Results showed that the accumulation of A-type and B-type vitisins depended on the yeast extract added. The mechanism underlying this effect does not seem to involve a direct contribution of mannoprotein-rich extracts in the oxidative step of vitisin formation, but rather their role in the physicochemical stabilization of these pigments. Color evolution also varied depending on the extract added. In this way, nucleic acids present in the autolysates of S. cerevisiae, M. pulcherrima and T. delbrueckii may have acted as copigments with anthocyanins, although stabilization of the complexes was only achieved by the autolysate of M. pulcherrima. The enhanced accumulation of B-type vitisins in the presence of the P. kluyveri autolysate resulted in a faster evolution of the hue of this wine.

Highlights

• •Eight MP extracts were obtained from S. cerevisiae and non-Saccharomyces yeasts.
• •MP extracts were added to a model wine and pigment and color evolution were studied.
• •The observed effects were highly dependent on the MP extract used.
• •The MP extracts obtained from P. kluyveri favoured vitisin accumulation.
• •The MP extract obtained from M. pulcherrima favoured anthocyanin stability.

1. Introduction

The color of red wine is mainly due to the presence of a class of phenolic compounds named anthocyanins, which are located on the skins of red grapes and extracted to the fermenting must during maceration. Moreover, in young red wines, a great percentage of the color is due to the phenomenon of copigmentation, in which anthocyanins form weak complexes with themselves or with other copigments, preventing anthocyanin hydration and the consequent formation of the colorless hemiacetal form (He et al., 2012).

Monomeric anthocyanins are not stable pigments and, during fermentation and, especially, during wine ageing, they tend to react with other wine constituents leading to the formation of anthocyanin-derived pigments, such as pyranoanthocyanins (PACs). Among the different classes of PACs, A and B-type vitisins (VitA and VitB) are formed, respectively, by the condensation of anthocyanins with pyruvic acid and acetaldehyde, secondary metabolites released by yeast during fermentation. The new pigment formed is characterized by possessing an additional pyranic ring (D ring). PAC formation in wine is of oenological relevance for two main reasons. Firstly, they are chemically more stable than anthocyanins, being more resistant to oxidation and discoloration due to pH changes and sulphur dioxide; and, secondly, they show red-orange hues, contributing to the color shift from red-purple hues of young red wines to red-orange tonalities of aged red wines (De Freitas & Mateus, 2011).

Recent studies suggest that the addition to wine of mannoproteins (MPs) and other yeast derivatives could influence the formation of anthocyanin-derived pigments (Dong et al., 2024; Oyón-Ardoiz et al., 2022; Rinaldi et al., 2019; Snyman, Mekoue Nguela, Sieczkowski, Marangon, & Divol, 2023; Zhai et al., 2025). MPs are glycoproteins that are located mainly in the external layer of the yeast cell wall and that are released to the wine during alcoholic fermentation and yeast autolysis (Guadalupe et al., 2014). These glycoproteins are heavily substituted with mannose residues, up to the point that the polysaccharide part can represent the 50–95% (w/w) of the molecule (Lipke & Ovalle, 1998). Numerous studies have focused on the characterization of MPs from Saccharomyces cerevisiae strains. However, other non-Saccharomyces yeast can also be present in wine, participating in the first stages of spontaneous fermentations. Moreover, nowadays, the use of these yeasts in mixed fermentations with S. cerevisiae is under research as a tool for improving some technological and organoleptic aspects of wines (Jolly et al., 2014), but few studies exist on the characterization of their MPs. It has not been until recently that studies addressing this topic have emerged, pointing out to significant structural and compositional differences in the MPs derived from different wine yeast species (Assunção Bicca et al., 2025; Domizio et al., 2014; Snyman, Mekoue Nguela, Sieczkowski, Divol, & Marangon, 2023).

Studies on the effect of MPs on wine properties, such as color stability, report contradictory findings (Alcalde-Eon, Ferreras-Charro, et al., 2019; Alcalde-Eon, Pérez-Mestre, et al., 2019; Guadalupe & Ayestarán, 2008; Oyón-Ardoiz et al., 2024a; Rodrigues et al., 2012; Snyman, Mekoue Nguela, Sieczkowski, Marangon, & Divol, 2023). Such discrepancies could result from multiple factors like the use of MP extracts of different degrees of purity, the great heterogeneity of this class of biopolymers and the differences in the wine matrix influencing MP effects, among others. Therefore, this study aims to provide more information about the characteristics of the MPs isolated from different wine yeasts. For this purpose, yeast extracts rich in MPs have been obtained from S. cerevisiae and from four non-Saccharomyces strains used in oenology. The impact of adding these extracts on pigment composition and color evolution has been assessed in a model wine containing pyruvic acid and acetaldehyde. Moreover, the observed effects on color and vitisin evolution have been related to the characteristics of the yeast extracts used.

2. Materials and methods

2.1. Yeast strains and materials

Five commercial oenological yeast strains were purchased for the extraction of MPs: Saccharomyces cerevisiae (EC1118 Organic, Lallemand Inc., Montreal Canada), Lachancea thermotolerans (LAKTIA, Lallemand Inc., Montreal, Canada), Metschnikowia pulcherrima (FLAVIA, Lallemand Inc., Montreal, Canada), Torulaspora delbrueckii (BIODIVA, Lallemand Inc., Montreal, Canada) and Pichia kluyveri (FROOTZEN, Chr Hansen, Hørsholm, Denmark). Select yeast extract, potassium persulfate, 1-phenyl-3-methyl-pyrazolone (PMP) (99%) and phosphate buffered saline 10× (pH 7,4) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Peptone from soybean was purchased from Merck-Millipore (St. Louis, MO, USA) and D-(+)-glucose from VWR Chemicals (Ohio, USA). HPLC-grade acetonitrile was purchased from Merck (Darmstadt, Germany), methanol from VWR Chemicals (Ohio, USA) and TFA from Carlo Erba Reagents (Val de Reuil, France). LiNO₃ (99 + %) was purchased from Thermo Fisher Scientific. HCl 37% and NaOH pellets were purchased from PanReac AppliChem ITW Reagents (Darmstadt, Germany).

2.2. Extraction of the MP-rich extracts

Ten MP-rich extracts were obtained from S. cerevisiae, L. thermotolerans, M. pulcherrima, T. delbrueckii and P. kluyveri by yeast induced autolysis followed by β-1,3-glucanase digestion according to Oyón-Ardoiz et al. (2024b). Briefly, yeasts were grown in Yeast Peptone Dextrose (YPD) until DO_600nm 12–16. Biomass was collected by centrifugation and subjected to induced autolysis in NaCl 3% (w/v) at 55 °C for 24 h. MP extracts obtained by autolysis of S. cerevisiae, L. thermotolerans, M. pulcherrima, T. delbrueckii and P. kluyveri were named, respectively, A-S.cer, A-L.the, A-M.pul, A-T.del and A-P.klu. Then, the autolyzed cells were recovered by centrifugation and subjected to enzymatic hydrolysis using Zymolyase 20 T (US Biological, Salem, MA, USA) during 90 min at 45 °C. MP-rich extracts obtained by Zymolyase digestion were named Z-S.cer, Z-L.the, Z-M.pul, Z-T.del and Z-P.klu.

2.3. Characterization of the MP-rich extracts

The apparent molecular weight (MW) distribution, protein content and monosaccharide composition of the MP-rich extracts were determined according to Oyón-Ardoiz et al. (2024b). Briefly, the apparent MW distribution was determined by HRSEC-RID following the method described by Manjón et al. (2020), using an Agilent 1260 Infinity (Agilent Technologies, Palo Alto, CA, USA) equipped with two serial Shodex OHPak SB-803 HQ and SB-804 HQ columns (8 mm × 300 mm) (Showa Denko Europe GmbH, Germany) thermosttated at 60 °C. A flow rate of 1 mL/min LiNO₃ 0.1 M was employed for elution. The protein content of the MP-rich extracts was assessed by Lowry method employing the DC protein assay kit (Bio-Rad Laboratories Inc.) (Lowry et al., 1951). Monosaccharide composition was determined by a modification of the method proposed by Ruiz-Garcia et al. (2014). After polysaccharide hydrolysis in TFA 2 M, derivatization of the monosaccharides was carried out with 1-phenyl-3-methyl-pyrazolone (PMP). The derivatized monosaccharides were then analyzed in an Agilent 1200 Series HPLC using an Agilent Poroshell 120 EC-18 column (2.7 μm, 4.6 mm × 150 mm) (Agilent Technologies, Waldbronn, Gernamy) coupled to an API 3200 Qtrap mass spectrometer (Applied Biosystems, Darmstdat, Germany). The conditions of the HPLC and the mass spectrometer for the analysis of the derivatized monosaccharides are those employed in our previous work (Oyón-Ardoiz et al., 2022). After MP characterization, A-L.the and Z-M.pul were discarded for the following assays due to their high content, respectively, in nucleic acids and β-glucan.

2.4. Evaluation of the antioxidant activity of the MP-rich extracts

In vitro antioxidant activity of the MP-rich extracts was evaluated by the FRAP (Ferric Reducing Antioxidant Power) assay following the method proposed by Benzie and Strain (1996) with minor modifications. Briefly, 10 mM 2,4,6-Tris(2-pyridyl)-s-triazine (Sigma-Aldrich, St. Louis, MO, USA) in 40 mM HCl and 20 mM FeCl₃·6H₂O (Sigma-Aldrich, St. Louis, MO, USA) were mixed with acetate buffer (300 mM, pH 3.6) in a proportion of 1:1:10. An aliquot of 250 μL of this solution was mixed with 10 μL of the MP extracts at two concentrations (0.4 and 1 mg/mL) in a 96-well microplate. The mixture was incubated at 37 °C for 6 min and absorbance at 593 nm (A_593nm) was measured in a FLUOstar Omega microplate reader (BMG Labtech, Ortenberg, Germany). A Trolox calibration curve was constructed.

The ABTS^·+ scavenging capacity of the MP-rich extracts was also evaluated following the ABTS/persulfate assay described by Re et al. (1999) with minor modifications. To produce the ABTS^·+ radical, 7 mM ABTS (Sigma-Aldrich, St. Louis, MO, USA) was mixed with 2.45 mM potassium persulfate and incubated in the dark for 12–16 h. Then, ABTS^·+ was diluted with phosphate-buffer saline (pH 7.4) until reaching an absorbance at 734 nm (A_734nm) of 0.70 ± 0.02. An aliquot of 200 μL of this solution was mixed with 7 μL of the MP extracts at two concentrations (0.4 and 1 mg/mL) in a 96-well microplate. After 4 min of incubation at room temperature, the absorbance was measured at 734 nm using the same equipment as above. The ABTS^·+ scavenging capacity (%) was calculated as follows:

$${\mathit{ABTS}}^{·+}\text{scavenging capacity}\left(\%\right)=\left[1-\left(\frac{A_{\mathit{sample}}}{A_{\mathit{blank}}}\right)\right]\ast 100$$

2.5. Isolation of anthocyanin 3-O-monoglucosides

Anthocyanin 3-O-monoglucosides were extracted from red grape skins (Vitis vinifera L. cv Tempranillo) with 95:5 (v/v) methanol/HCl 0.5 N by sonication in an ultrasound bath for 30 min followed by maceration at −20 °C for 24 h according to the method reported by García-Estévez, Alcalde-Eon, and Escribano-Bailón (2017). The extraction process was repeated three times and, then, the methanolic extracts were mixed and evaporated under vacuum. The final extracts were loaded onto a Sephadex LH-20 column (Sigma-Aldrich, St. Louis, MO) and the elution was carried out with acidified water (HCl 0,1 N; pH 1) as described by García-Estévez et al. (2013). Aliquots of 20 mL were collected from the column and analyzed by HPLC-DAD in an Agilent 1100 Series HPLC (Agilent Technologies, Waldbronn, Germany) using the conditions reported by Alcalde-Eon et al. (2006). A C18 reverse-phase column (5 μm, 150 mm × 4.6 mm) (Aqua®, Phenomenex, Torrance, CA) was employed and chromatograms were registered at a preferred wavelength of 520 nm. The fractions containing the anthocyanins of interest were then mixed and freeze-dried.

2.6. Preparation of the wine model solution

All assays were carried out in a solution resulting from the fermentative metabolism of glucose by S. cerevisiae to ensure the presence of acetaldehyde and pyruvic acid, secondary metabolites required for the formation of A and B-type vitisins (VitA and VitB). S. cerevisiae (EC1118 Organic, Lallemand Inc., Montreal, Canada) was pre-cultured in YPD until reaching an OD_600nm of 1 (exponential growth phase) to obtain the inoculum necessary to carry out the fermentation. Cells were then collected by centrifugation (1290 g, 10 min) and washed 3 times with distilled water. Yeast was inoculated at an OD_600nm of 0.2 in a solution containing 200 g/L of glucose, previously autoclaved, and added with 6.9 g/L of Yeast Nitrogen Base (Formedium Ltd., King's Lynn, England). Fermentations were carried out at room temperature until glucose exhaustion. Glucose consumption and ethanol formation were monitored by HRSEC-RID using the same equipment as described in Section 2.3. At the end of fermentation, the media were centrifuged three times (5170g, 10 min) and the yeast pellets were discarded. The pH of the resulting media, which contained 11.5% (v/v) of ethanol, was adjusted to 3.6 with NaOH 20% (w/v).

2.7. Influence of MP-rich extracts on pigment evolution in model wine

The model solution obtained in the previous section was added with 300 mg/L of the purified anthocyanin extract. This concentration of anthocyanins was chosen because it is a common concentration found in red wines (Flanzy, 2003). To evaluate the possible influence of the MP extracts on pigment evolution, 400 mg/L of each extract were added to the model wine (maximum dose recommended in wines by the EC Regulation No. 606/2009). The use of model systems allows the effect of MPs on vitisin formation to be studied in isolation, eliminating variables associated with the wine matrix, such as grape variety, while simulating conditions similar to those of fermentation. Samples were stored at room temperature (18–22 °C) in the dark and pigment composition was monitored periodically. A total of 16 sampling points were analyzed over the course of more than 9 months. The analysis of anthocyanins and vitisins was carried out by HPLC-DAD-MS using the same HPLC equipment as described in Section 2.5. Before injecting in the HPLC, 120 μL of the samples were diluted with 30 μL of HCl 0.25 N and filtered using a 2 μm pore size filter (Merk Millipore Ltd., Cork, Ireland). The HPLC was coupled to an API 3200 Qtrap mass spectrometer (Applied Biosystems). The conditions of the mass spectrometer were those described in the work of Alcalde-Eon, Pérez-Mestre, et al. (2019). Chromatograms were recorded at the wavelengths of 280, 360, 440, 505 and 520 nm, allowing the detection of pigments and the monitoring of the formation of degradation products in a single run. The quantification of the pigments was carried out at 520 nm using a malvidin 3-O-glucoside calibration curve. Quantification of the degradation products was carried out at 280 nm using calibration curves of gallic, protocatechuic and vanillic acids.

2.8. Colorimetric measurements

Before color measurement, samples were filtered using a 2 μm pore size filter (Merk Millipore Ltd., Cork, Ireland) to avoid the presence of particles in suspension. The absorption spectra (190–770 nm) were registered in a Hewlett-Packard 8453 UV–Vis spectrophotometer (Agilent Technologies, Waldbronn, Germany) using 2 mm path length quartz cuvettes. The fermented media without anthocyanins and MPs was used as blank. The CIELAB parameters (L^⁎, a^⁎ and b^⁎) of the samples were calculated from the visible spectra (380–770 nm) with the software CromaLab (Heredia et al., 2004). The values of chroma (C^⁎_ab), hue (h_ab) and the color differences (ΔE_ab^⁎) between the control sample and the samples supplemented with MPs were also calculated. ΔE_ab^⁎ was calculated with the following formula:

$$∆E_{\mathit{ab}}^{\ast }=\sqrt{{\left(L_2^{\ast }-L_1^{\ast }\right)}^2+{\left(a_2^{\ast }-a_1^{\ast }\right)}^2+{\left(b_2^{\ast }-b_1^{\ast }\right)}^2}$$

The total color of the samples was calculated as the color difference (ΔE_ab^⁎) between each sample and distilled water (L^⁎ = 100; a^⁎ = 0; b^⁎ = 0). A possible copigmentation effect of the MP extracts was assessed by calculating the percentage of the anthocyanin color solution that is due to copigmentation (CCI) (Gordillo et al., 2012):

$$\mathit{CCI}=\left(\frac{E_{\mathit{MP}}-E_{\mathit{control}}}{E_{\mathit{control}}}\right)\times 100$$

being E_MP the total color of the samples added with the MP extract and E_control the total color of the control wine.

2.9. Statistical analyses

All statistical analyses were carried out using the software IBM-SPSS statistics v. 28. The statistical significance of the differences in the composition of MP extracts and in MP extract antioxidant activity were determined by a one-way analysis of variance (ANOVA) followed by a post-hoc Tukey-b test. The same test was performed to determine the statistical differences among the different wine model solutions in the pigment and degradation product concentrations, as well as in color parameters. Differences were considered significant when p < 0.05.

3. Results and discussion

3.1. Characterization of the MP-rich extracts

3.1.1. Monosaccharide composition

Monosaccharide composition analysis of the MP-rich extracts (Table 1) showed that mannose was the main monosaccharide present (>70%) indicating an effective release of MPs from the yeast cell wall by the two treatments applied. It should be noted that yeast extracts obtained by enzymatic digestion with Zymolyase contained a higher proportion of glucose than those obtained by induced autolysis. In this sense, Z-S.cer, Z-L.the and Z-P.klu showed around 28% of glucose. This arises from the liberation of cell wall β-glucan by the β-1,3-glucanase activity of Zymolyase. Therefore, MP extracts obtained by Zymolyase digestion contained fragments of β-glucan and/or molecules consisting of a MP bound to a β-glucan chain. In contrast, MP extracts isolated by induced autolysis of yeast cells contained ribose in their composition, especially A-S.cer and A-T.del (∼22% ribose). Leakage of cellular components like nucleic acids can be expected to happen during the induced autolysis of yeast cells (Takalloo et al., 2020). However, nucleic acid release was dependent on the yeast strain, since A-M.pul and A-P.klu showed little presence of ribose (<5%). In fact, these two MP extracts showed the highest content of mannose of all the extracts obtained (>93%) suggesting that induced autolysis of M. pulcherrima and P. kluyveri could be an effective method to obtain extracts of high purity in MPs from these two yeast species.

Table 1.

Monosaccharide composition (%) of the MP extracts.

	A-S.cer	A-M.pul	A-P.klu	A-T.del	Z-S.cer	Z-L.the	Z-P.klu	Z-T.del
Mannose	73 ± 2 e	93 ± 1 b	95.7 ± 0.3 a	75.4 ± 0.5 d	71.6 ± 0.1 e	72.7 ± 0.5 e	72 ± 1 e	87.2 ± 0.1 c
Glucose	4.5 ± 0.1 d	1.62 ± 0.02 f	3.1 ± 0.1 e	2.2 ± 0.2 ef	28.4 ± 0.1 a	27.3 ± 0.5 b	28 ± 1 ab	12.6 ± 0.2 c
Ribose	22 ± 2 a	5 ± 1 b	1.2 ± 0.3 c	22.4 ± 0.3 a	–	–	–	0.17 ± 0.08 c

Different letters within each row indicate significant differences (p < 0.05).

3.1.2. Protein content

Protein determination by Lowry method showed significant differences in the protein content of the MP-rich extracts (Table 2). Both MP extracts isolated from P. kluyveri, A-P.klu and Z-P.klu, contained the highest percentage of protein, 11.3% and 8.7%, respectively, indicating that the MPs of this yeast could have a larger protein fraction than the MPs extracted from the other yeasts. On the contrary, MPs isolated from T. delbrueckii, A-T.del and Z-T.del, showed the lowest protein proportion, 3.4% and 3.5%, respectively. In view of these results, the yeast species could be a factor with greater influence in the protein content of the MP extracts than the method of MP isolation. However, overall, the protein content of all the MP-rich extracts obtained in this study was low (<11%). During the extraction process, some degree of protein degradation could have occurred, since, in the one hand, during autolysis yeast proteases are activated and, in the other hand, Zymolyase shows side protease activity. Despite this, our results are consistent with the literature, given that a great variability in the protein content of the MPs can be found in the different studies (Domizio et al., 2014; Giovani et al., 2012; Manjón, Brás, García-Estévez and Escribano-Bailón, 2020, Manjón, Recio-Torrado, Ramos-Pineda, García-Estévez and Escribano-Bailón, 2021; Snyman, Mekoue Nguela, Sieczkowski, Divol, & Marangon, 2023).

Table 2.

Molecular weight (Mw) distribution, average Mw and protein percentage (w/w) of the MP extracts.

MP	Mw (kDa)	%	Average Mw	% Protein
A-S.cer	558–189	19.3	82 ± 2 d	7.8 ± 0.6 c
	189–103	6.6
	103–35	9.5
	35–18	3.2
	18–4	39.5
	< 4	21.9
A-M.pul	1614–794	2.3	220 ± 10 a	6.0 ± 0.2 e
	794–90	60.4
	90–40	6.6
	40–9	23.7
	9–4	4.9
	< 4	2.1
A-P.klu	1087–650	1.2	132 ± 3 b	11.3 ± 0.4 a
	650–388	64.7
	388–81	6.8
	81–48	4.2
	48–29	3.6
	29–19	11.8
	19–5	3.5
	< 5	4.2
A-T.del	620–184	23.0	98 ± 1 c	3.4 ± 0.3 f
	184–42	10.4
	42–17	14.0
	17–4	40.0
	< 4	12.6
Z-S.cer	330–130	12.6	56.5 ± 0.3 e	6.7 ± 0.1 d
	130–9	81.9
	9–4	5.5
Z-L.the	474–108	41.0	99 ± 2 c	5.70 ± 0.02 e
	108–50	29.7
	50–7	29.3
Z-P.klu	364–24	73.9	52.0 ± 0.3 e	8.7 ± 0.2 b
	24–10	15.8
	10–5	7.6
	< 5	2.8
Z-T.del	418–132	24.3	93 ± 1 c	3.5 ± 0.2 f
	132–24	65.9
	24–5	9.8

Different letters within the same column indicate significant differences (p < 0.05).

3.1.3. Apparent molecular weight distribution

HRSEC-RID analysis showed great differences in the apparent MW distribution of the MPs (Table 2). The chromatographic profiles of the MP extracts obtained by autolysis showed a higher complexity than those of the Zymolyase extraction. In this sense, the former ones presented a wider MW distribution and contained several fractions of different MW while the latter showed a predominant fraction with a narrower MW range (Fig. S1 in the Supplementary Material). In consequence, MP-rich extracts obtained by autolysis exhibited a greater structural complexity which could be due to the wider range of yeast endogenous enzymes acting upon the cell wall in this process (Takalloo et al., 2020). Moreover, the average MW of the MPs extracted by autolysis was greater than that of the MPs obtained by Zymolyase digestion (Table 2). The smaller size of the MPs extracted by Zymolyase, together with the simpler chromatographic profile, could indicate some degree of degradation of the MPs structure by the Zymolyase, which, besides its main activity as β-1,3-glucanase, also shows side mannanase and protease activities.

A-M.pul showed the highest average MW of the extracted MPs (220 kDa), followed by A-P.klu (132 kDa). A-M.pul showed a main population of MPs, that represented the 60.4% of the total, with a MW comprised between 90 and 794 kDa. A-P.klu also presented a main peak representing 64.7% of the total and comprising MPs of MW between 388 and 650 kDa. In addition, both MP extracts showed a minoritarian high MW peak (among 794 and 1614 kDa in A-M.pul and among 650 and 1087 kDa in A-P.klu) which represented less than 2.3% of the total. The profile of A-T.del presented similarities to that of A-S.cer in the sense that both MPs showed a majoritarian peak (∼40%) of low MW MPs (4–18 kDa in A-S.cer and 4–17 kDa in A-T.del). Furthermore, both MPs showed a peak (representing around 20%) comprising MPs of MW between ∼180 and ∼ 600 kDa. In the case of the MPs isolated by Zymolyase digestion, Z-L.the and Z-T.del showed a similar average MW (99 and 93 kDa, respectively) with no significant differences among them. The sizes of these MPs were larger than that of Z-S.cer and Z-P.klu, whose average MW were 57 and 52 kDa, respectively, with no significant differences among them. Altogether, these results indicate a high polydispersity of the obtained MPs, which agree with the wide range of MW reported in the literature for MPs extracted from different yeast species by the application of different treatments (Gao et al., 2025).

3.2. Antioxidant activity of the MP-rich extracts

Given that the transformation of anthocyanins into VitA and VitB involves an oxidation step (De Freitas & Mateus, 2011), the ability of the MP-rich extracts to participate in redox reactions was evaluated. This could help clarify the possible mechanisms by which MPs could influence the formation and evolution of vitisins. All the MP-rich extracts tested negative for the FRAP assay at both concentrations (i.e. the A_593nm of the samples was similar to that of the blank), suggesting that these extracts had no components that could participate in electron transfer reactions. The ABTS^·+ scavenging capacity of the extracts was low, less than 6% for the low concentration assayed (0.4 mg/mL) and less than 11% for the highest (1 mg/mL) (see Fig. S2 in the Supplementary Material). This agrees with other studies that report a weak radical scavenging capacity of MPs (Wu et al., 2015). A-M.pul, A-P.klu and Z-Pklu were the MP extracts with the highest ABTS^·+ scavenging percentage, yet their effects on vitisin accumulation differ, as will be discussed later. Therefore, MP extract influence on pigment evolution is unlikely to be related to their participation in oxidative processes.

3.3. Influence of MP-rich extracts on pigment evolution in model wine

3.3.1. Evolution of the content of anthocyanins and vitisins

The concentrations of VitA, VitB and native anthocyanins were monitored periodically by HPLC-DAD-MS (Fig. 1). The qualitative evolution of the pigments was the same in all samples and, at the end of the experiment, a total of 6 different vitisins (2 VitA and 4 VitB) were identified (Table S1 in the Supplementary Material). Anthocyanin disappearance followed a similar trend in all samples at the first stages of the experiment, except for the sample added with A-T.del (Fig. 1A), in which a rapid loss of anthocyanins was observed. The addition of A-T.del appears to induce the partial precipitation of anthocyanins, resulting in a significantly lower concentration of these pigments the day after MP extract was added. This is in agreement with our previous work, in which the addition to red wine of the MP-rich extract obtained by the autolysis of T. delbrueckii caused an enhanced precipitation of grape native anthocyanins (Oyón-Ardoiz et al., 2024a). The monosaccharide composition of this extract, which contains an important proportion of ribose, does not differ from that of other extracts used in this study, such as A-S.cer. Therefore, characteristics of the MP extract beyond the presence of ribose and the monosaccharide composition must influence its interaction with anthocyanins. It should be noted that, after the initial anthocyanin precipitation that caused a rapid loss of these pigments, anthocyanin concentration was stabilized in the wine added with A-T.del (Fig. 1A). In this sense, the rate of anthocyanin disappearance was much slower in this wine than in the other samples studied and, at the end of the experiment (day 282), its content in anthocyanins was significantly higher than in the other samples.

Fig. 1.

Evolution of the mean concentration (n = 3) of anthocyanin 3-O-monoglucosides (A), total vitisins (B), A-type vitisins (C) and B-type vitisins (D) in the model wine solutions.

Regarding the model wine added with A-M.pul, a significantly higher concentration of anthocyanins could be observed from day 46 onwards (Fig. 1A), indicating that the addition of this MP extract could favour the stability of anthocyanins overtime. The addition of A-S.cer also slightly favoured the stability of anthocyanins from day 204 onwards, resulting in a significantly higher concentration of these pigments at the end of the experiment, although the effect of this MP-rich extract on anthocyanin content was lower than that of A-T.del and A-M.pul.

The formation of vitisins in the model wines was low, reaching a maximum concentration of around 1 mg/L (Fig. 1B). Vitisin concentration was in the range of that reported in the work of Morata et al. (2003), who studied pyruvic acid and acetaldehyde production by different S. cerevisiae strains and its relationship to VitA and VitB formation in red wine. However, other authors have described higher vitisin levels in both red and model wines (Alcalde-Eon, Ferreras-Charro, et al., 2019; García-Estévez, Alcalde-Eon, Puente, & Escribano-Bailón, 2017), although vitisin content can vary greatly among different red wines (Blanco-Vega et al., 2014).

The addition of A-P.klu seemed to favour the accumulation of VitB (Fig. 1D), resulting in a significantly higher concentration of these pigments compared to the control wine from day 4 until day 138, in which the concentration started to decrease. The higher content of VitB translated into an increased total vitisin content, being this difference of concentration with the control wine statistically significant from day 4 until day 75. In the model wine supplemented with Z-P.klu, VitA levels were significantly higher than in the control from day 113 to day 204, suggesting that this MP extract may slightly enhance the accumulation of these pigments. However, the concentration of VitA in this model wine did not differ significantly from that found in model wines added with the other MP extracts obtained by Zymolyase digestion.

On the contrary, the addition of A-S.cer and A-M.pul resulted in a significant reduction in the concentration of total vitisins in the model wine (Fig. 1B). This lower concentration of total vitisins can be related to the effect of these MP-rich extracts on VitA (Fig. 1C), decreasing the accumulation of these pigments, while having no significant impact on VitB concentration (Fig. 1D).

The accumulation of vitisins followed a different kinetic in the model wine added with A-T.del. At the beginning of the experiment, this model wine showed a decreased concentration of vitisins, which can be explained by the lower anthocyanin concentration observed in this sample, likely resulting in a lower formation of these pigments. Throughout the experiment, the rates of accumulation and disappearance of VitA and VitB in this sample differed. VitB reached the maximum concentration at day 75 and, afterwards, their concentration started to gradually decrease (Fig. 1D). However, the rate of disappearance of VitB was lower than in the other model wines and, at the end of the experiment, the residual concentration of VitB remained significantly higher in this sample compared to the other model wines. In the case of VitA, the addition of this yeast extract seemed to slow down the formation of these pigments, as the concentration of VitA was still increasing on day 282, unlike the decreasing trend observed in the rest of the samples at this sampling point (Fig. 1C). The slower accumulation of VitA, together with the slower disappearance of VitB, resulted in this wine having the highest concentration of total vitisins at the end of the experiment (Fig. 1B).

Regarding the MP extracts obtained by Zymolyase digestion of the cell wall, the addition of Z-S.cer, Z-L.the and Z-T.del had no substantial impact on either the rate of disappearance of anthocyanin monoglucosides or in the accumulation of vitisins, as the evolution of these model wines did not differ from that of the control samples.

In view of these results, it is notable that vitisin accumulation in the model wines is highly dependent on the MP-rich extract used. The enhanced accumulation of vitisins observed in the presence of A-P.klu and Z-P.klu is a phenomenon already described by some authors when other MP extracts have been assayed (Dong et al., 2024; Oyón-Ardoiz et al., 2022; Rinaldi et al., 2019; Snyman, Mekoue Nguela, Sieczkowski, Marangon, & Divol, 2023; Zhai et al., 2025). However, in those works, it remains unclear if the observed effect could be attributed to MPs increasing the stability in solution of these pigments or to MPs accelerating their formation. For example, Zhai et al. (2025) observed the accumulation of PACs in red wine after the addition of a commercial MP and suggested that it was due to the polysaccharides acting as steric stabilizers thus reducing the possibility of PAC aggregation and precipitation. In our previous study, we observed that some MP-rich extracts could increase the colloidal stability of some wine pigments, being this effect especially notable in the case of VitA and VitB. We also observed that, after a short period of storage, the accumulation of some anthocyanin-derived pigments in red wine was favoured in the presence of some MP-rich extracts (Oyón-Ardoiz et al., 2022). Other authors have suggested that MPs could favour the formation of polymeric pigments probably by favouring the reaction between anthocyanins and flavanols (Rinaldi et al., 2019).

3.3.2. Evolution of the content of anthocyanin degradation products

To try to understand the mechanisms behind the observed effects, besides quantifying the pigments, we also monitored the formation in the model wines of anthocyanin degradation products. A total of 9 compounds were identified as degradation products (see Table S2 in the Supplementary Material). The degradation pathway of anthocyanins involves the deglycosylation and opening of the pyrylium ring to form the chalcone. Differences have been reported in the order of occurrence of these two processes (Sun et al., 2011). In the present study, several chalcone glycosides were identified (see Table S2 in the Supplementary Material), suggesting that opening of the pyrylium ring occurs first followed by deglycosylation, as reported by Alcalde-Eon et al. (2022) in wine-like conditions. The chalcone aglycon is then cleaved into 2,4,6-trihydroxybenzaldehyde (THB) (A ring) and the corresponding hydroxybenzoic acid (B ring). THB can undergo further oxidation producing 2,4,6-trihydroxybenzoic acid (THBA) (Alcalde-Eon et al., 2022; Piffaut et al., 1994; Sun et al., 2011). In our model wines, THB was identified but no THBA was found. In addition, protocatechuic, vanillic and 3-O-mehtyl gallic acids were detected (as the degradation products of cyanidin, peonidin and petunidin 3-O-glucosides, respectively) as well as the ethylated ester of gallic acid (which is the benzoic acid produced by the degradation of delphinidin 3-O-glucoside). The cis and trans chalcone aglycones of malvidin were also identified in the model wines. Two additional compounds were detected and considered as degradation products although their structures could not be elucidated from the mass spectra and UV spectrum data.

The total concentration of these degradation products (Fig. 2A) increased throughout the experiment. The levels of hydroxybenzoic acids followed a trend similar to that of the total degradation products, showing increasing concentrations until the end of the experiment (Fig. 2C). However, the THB concentration increased rapidly in the early stages of the experiment, followed by the stabilization and a subsequent decrease of its levels (Fig. 2B). The content of malvidin chalcone remained stable over time which could be related to its role as an intermediate product in the anthocyanin degradation pathway (Fig. 2D).

Fig. 2.

Evolution of the mean concentration (n = 3) of the main anthocyanin degradation products in the model wine systems. A: total concentration of degradation products; B: concentration of 2,4,6-trihydroxybenzaldehyde (THB); C: total concentration of hydroxybenzoic acids; D: concentration of malvidin chalcone.

Vitisin accumulation was slightly favoured in the presence of A-P.klu and Z-P.klu. The formation of PACs in wine involves a last oxidation step necessary to yield the additional D ring (De Freitas & Mateus, 2011). The accumulation of anthocyanin degradation products in the model wines added with A-P.klu and Z-P.klu followed a similar trend to that of the control (Fig. 2), with few differences among them. This, altogether with the negative results obtained in the antioxidant activity assays, could suggest that these MP-rich extracts do not probably participate as oxidant agents. Consequently, the enhanced accumulation of vitisins could be more related to the ability of these MP-rich extracts to act as colloidal stabilizers, reducing the possibility of vitisin aggregation and precipitation. Liu et al. (2023) showed the ability of a commercial MP to interact with the vinylphenol pyranoanthocyanin of cyanidin 3-O-glucoside, reducing pigment aggregation and promoting its stability in solution. In line with these findings, other studies have also reported a certain promotion of the colloidal stability of VitA and VitB by the addition of MPs (Alcalde-Eon, Pérez-Mestre, et al., 2019; Oyón-Ardoiz et al., 2022). On the other hand, the interaction of A-P.klu and Z-P.klu with anthocyanins influencing their reactivity and, consequently, vitisin formation cannot be discarded. There are already studies that demonstrate the ability of MPs to interact with anthocyanins (Gonçalves et al., 2018; Wan et al., 2023), although the impact of these interactions on anthocyanin reactivity is unknown.

Dong et al. (2024) have observed that the formation of the VitB of malvidin in a model wine spiked with acetaldehyde was enhanced in different proportion by the addition of commercial MPs. According to these authors, this was related to a higher mannan/glucan ratio, a higher proportion of hydrophobic and hydrophilic aminoacids and a low/medium MW of the MP. A-P.klu showed a significantly higher MW and mannan/glucan ratio than Z-P.klu and both MPs favoured vitisin accumulation. Indeed, in the present study, no simple correlations could be observed between MP extract effects on pigment stability and mannose/glucose ratio and apparent MW. A-P.klu and Z-P.klu showed the highest protein content, which could indicate that a higher protein proportion may favour the interaction with vitisins. However, A-S.cer had a slightly lower protein content than Z-P.klu and their effect towards VitA was the opposite to that of Z-P.klu. Therefore, if MP-pigment interactions are influencing vitisin and anthocyanin stability, the monosaccharide composition, apparent Mw and protein content may not be the only factors affecting these interactions. For example, Assunção Bicca et al. (2023) reported that the interaction of MPs with malvidin 3-O-glucoside is probably influenced by the negative net charge of the molecule and the degree of branching and compactness of the structure conditioning the accessibility of the pigments to the interaction sites of the MP. Other authors reported similar findings when studying anthocyanin interactions with pectins, with higher proportions of side branches limiting anthocyanin interactions likely due to steric hindrance (Fernandes et al., 2021). Additionally, Assunção Bicca et al. (2025) found important differences in the polysaccharide structure and net charge of MPs belonging to different wine yeast species (including S. cerevisiae, Hanseniaspora vinae, Schizosaccharomyces japonicus and T. delbrueckii). Therefore, differences in the structure of the polysaccharide moiety, molecular compactness and net charge of the MPs assayed in the present study may partially explain the differences observed in the effects of MPs towards pigment stability. However, to our knowledge, there are no studies to date describing the polysaccharide structure and net charge of MPs derived from L. thermotolerans, M. pulcherrima and P. kluyveri, making it difficult to establish a direct relationship between MPs structure and activity. Other aspects like amino acid composition may also impact MPs functionality, so a more profound characterization of the MPs contained in the extracts would facilitate the establishment of clear structure-activity relationships.

The model wines supplemented with A-M.pul and A-T.del showed a higher concentration of anthocyanins while tending to accumulate degradation products. This is especially notable for the sample added with A-T.del, in which a greater accumulation of all types of degradation products could be observed from the middle part of the experiment until the end. In this point, in which this wine started to accumulate more degradation products, coincided approximately with the moment in which the concentration of anthocyanins stabilized, becoming greater than in the other model systems (Fig. 1A). Moreover, at the end of the experiment, a higher content of vitisins was found in the sample added with A-T.del. The fact that, at the end of the experiment, this wine accumulated a greater concentration of degradation products but also of vitisins could suggest that the enhanced concentration of anthocyanins may be related to this MP-rich extract acting as a colloidal stabilizer of these pigments. However, as mentioned before, immediately after the addition of A-T.del, a precipitation of anthocyanins was observed, which is likely related to their aggregation with the components of the extract, thereby reducing the concentration of soluble MPs in the medium. Consequently, these results suggest two possible hypotheses: i) this yeast extract may contain different components exerting opposite effects on anthocyanin stability; ii) the effect of MPs is highly dependent on their concentration. In this context, Snyman, Mekoue Nguela, Sieczkowski, Marangon, and Divol (2023) reported a dose-dependent effect of a MP extracted from T. delbrueckii. According to these authors, decreasing MP concentrations led to a reduction in L* and an increase in color intensity of a Cabernet-Sauvignon wine, which could be associated with the stabilization of the pigment composition. Therefore, further studies could be useful to determine the influence of the dosage of this MP extract in the stability of anthocyanins.

Regarding the wine added with A-M.pul, a slightly higher tendency to accumulate THB and malvidin chalcone (Fig. 2B and D) was observed, although there were no significant differences in the total concentration of degradation products when compared to control samples. As pointed before, the accumulation of VitA in this model wine was lower (Fig. 1C), which in combination with the greater content of some degradation products, could suggest that, in the presence of A-M.pul, the degradation pathway of anthocyanins may be slightly favoured over anthocyanin transformation to vitisins.

The addition to the model wine of A-S.cer resulted in a lower formation of hydroxybenzoic acids and of malvidin chalcone (Fig. 2C and D), although the total concentration of degradation products did not differ significantly to that of the control. Moreover, this wine showed a greater content of anthocyanins at the end of the experiment and a decreased accumulation of VitA. Consequently, the slightly lower formation of degradation products and the promoted accumulation of anthocyanins may be related to the capacity of this yeast extract to slightly reduce the oxidative degradation of these pigments and/or their oxidative transformation to vitisins. In this sense, the lower concentration of malvidin chalcone could indicate that, in presence of A-S.cer, opening of the C-ring to form the chalcone glucosides and/or chalcone deglycosylation might be slightly reduced. Nevertheless, calculation of the thermodynamic constants governing these processes would be required to confirm this hypothesis.

3.4. Colorimetric measurements

Tables 3 shows the evolution of L^⁎, h_ab and C^⁎_ab values of the model wines over the course of the experiment (see also Fig. S3, S4 and S5 in the Supplementary Material). In all the model systems, L^⁎ and h_ab increased and C^⁎_ab decreased over time due to the loss of anthocyanins and, consequently, to a greater contribution to color of vitisins. Despite this tendency, the addition of the yeast extracts exerted different effects on CIELAB parameters. Regarding lightness (Table 3), the sample added with A-S.cer and, specially, that supplemented with A-M.pul showed lower L^⁎ values than the control wine, indicating the darkest color of these samples. Regarding the model wines added with the extracts obtained by Zymolyase digestion, L^⁎ values followed a similar trend to that of the control wine, with slightly lower lightness values on days 46 and 282.

Table 3.

Evolution of the values of lightness (L^⁎), hue (h_ab) and chroma (C^⁎_ab) in the model wines.

Days	0	4	17	46	94	138	282
L⁎
Control	87.1 ± 0.6 a	86.1 ± 0.2 a	86.1 ± 0.3 b	86.2 ± 0.2 b	86.2 ± 0.1 bc	87.2 ± 0.1 b	89.9 ± 0.2 a
A-S.cer	83.6 ± 0.4 a	82.6 ± 0.4 b	83.7 ± 0.9 c	84.5 ± 0.2 d	85.6 ± 0.2 c	86.8 ± 0.2 b	88.1 ± 0.6 c
A-M.pul	80.0 ± 0.6 b	78. 9 ± 0.4 c	79.8 ± 0.3 d	80.5 ± 0.4 e	80.5 ± 0.5 d	82.6 ± 0.9 c	86.0 ± 0.3 d
A-P.klu	86.4 ± 0.3 a	85.7 ± 0.4 a	85.7 ± 0.5 b	85.7 ± 0.3 bc	86.7 ± 0.3 b	87.7 ± 0.3 b	89.8 ± 0.5 a
A-T.del	75 ± 4 c	86.0 ± 0.3 a	90.8 ± 0.1 a	90.46 ± 0.07 a	91.3 ± 0.3 a	89.9 ± 0.2 a	86.3 ± 0.3 d
Z-S.cer	86.3 ± 0.3 a	86.1 ± 0.2 a	86.3 ± 0.3 b	85.38 ± 0.04 c	85.8 ± 0.3 bc	87.0 ± 0.3 b	89.5 ± 0.6 ab
Z-L.the	85.9 ± 0.5 a	85.9 ± 0.3 a	85.8 ± 0.6 b	85.5 ± 0.2 c	85.8 ± 0.4 c	86.8 ± 0.2 b	89.6 ± 0.2 ab
Z-P.klu	86.1 ± 0.2 a	86.0 ± 0.3 a	85.9 ± 0.4 b	85.4 ± 0.2 c	85.6 ± 0.1 c	86.5 ± 0.6 b	88.5 ± 0.5 bc
Z-T.del	85.5 ± 0.3 a	85.6 ± 0.2 a	85.8 ± 0.4 b	85.6 ± 0.4 c	86.4 ± 0.4 bc	86.9 ± 0.3 b	88.5 ± 0.5 bc
hab
Control	−7.7 ± 0.2 ab	−6.63 ± 0.09 a	−4.78 ± 0.08 a	−2.9 ± 0.4 d	0.9 ± 0.1 b	5.40 ± 0.07 b	17.7 ± 0.1 b
A-S.cer	−13.7 ± 0.2 c	−12.9 ± 0.2 b	−10.3 ± 0.3 d	−7.8 ± 0.2 e	−2.97 ± 0.02 d	2.9 ± 0.1 c	16.5 ± 0.3 c
A-M.pul	−16.7 ± 0.7 d	−16.4 ± 0.4 c	−14.8 ± 0.3 e	−14.4 ± 0.2 f	−12.7 ± 0.5 e	−8 ± 1 e	5.62 ± 0.03 d
A-P.klu	−8.4 ± 0.1 b	−6.6 ± 0.1 a	−4.51 ± 0.06 a	−0.27 ± 0.05 a	1.96 ± 0.07 a	6.23 ± 0.07 a	18.5 ± 0.7 a
A-T.del	−21.7 ± 0.6 e	−16 ± 1 c	−10.2 ± 0.3 d	−2.4 ± 0.2 c	−1.07 ± 0.08 c	−0.27 ± 0.04 d	3.7 ± 0.4 e
Z-S.cer	−7.4 ± 0.3 ab	−7.4 ± 0.1 a	−6.4 ± 0.2 c	−2.02 ± 0.04 bc	1.11 ± 0.04 b	5.04 ± 0.07 b	18.6 ± 0.2 a
Z-L.the	−7 ± 1 a	−6.6 ± 0.2 a	−5.5 ± 0.1 b	−2.24 ± 0.05 bc	1.13 ± 0.06 b	5.15 ± 0.07 b	18.94 ± 0.08 a
Z-P.klu	−6.7 ± 0.2 a	−6.6 ± 0.6 a	−5.3 ± 0.1 b	−1.8 ± 0.2 b	1.17 ± 0.03 b	5.23 ± 0.03 b	19.22 ± 0.09 a
Z-T.del	−8.0 ± 0.4 ab	−8.0 ± 0.6 a	−6.1 ± 0.1 c	−2.9 ± 0.1 d	0.745 ± 0.003 b	4.60 ± 0.05 b	18.49 ± 0.09 a
C⁎ab
Control	23 ± 1 c	23.7 ± 0.4 c	22.4 ± 0.5 b	20.5 ± 0.1 bc	17.91 ± 0.02 b	14.9 ± 0.1 bc	9.8 ± 0.2 d
A-S.cer	25.4 ± 0.7 bc	26.0 ± 0.4 d	23.2 ± 0.9 b	20.8 ± 0.1 bc	17.5 ± 0.2 b	14.7 ± 0.2 bc	10.9 ± 0.5 c
A-M.pul	27.6 ± 0.7 b	28.1 ± 0.3 a	26.3 ± 0.2 a	24.6 ± 0.2 a	22.4 ± 0.5 a	18.6 ± 0.9 a	12.2 ± 0.1 b
A-P.klu	23.1 ± 0.6 c	23.6 ± 0.5 c	21.8 ± 0.9 b	20.1 ± 0.5 c	17.5 ± 0.5 b	14.3 ± 0.3 cd	9.9 ± 0.5 d
A-T.del	32 ± 4 a	20.3 ± 0.5 d	15.1 ± 0.2 c	13.6 ± 0.2 d	12.4 ± 0.5 c	13.6 ± 0.2 d	16.4 ± 0.2 a
Z-S.cer	23.6 ± 0.5 c	24.3 ± 0.5 c	22.5 ± 0.7 b	21.35 ± 0.06 b	18.5 ± 0.5 b	15.03 ± 0.31 bc	10.1 ± 0.5 cd
Z-L.the	23.8 ± 0.4 c	24.1 ± 0.7 c	22.6 ± 0.6 b	21.1 ± 0.5 b	18.6 ± 0.5 b	15.2 ± 0.2 bc	10.1 ± 0.1 cd
Z-P.klu	23.4 ± 0.6 c	24.2 ± 0.8 c	22.6 ± 0.4 b	21.3 ± 0.4 b	18.6 ± 0.2 b	15.5 ± 0.6 b	11.0 ± 0.4 c
Z-T.del	24.1 ± 0.7 c	24.2 ± 0.7 c	22.6 ± 0.5 b	20.8 ± 0.6 bc	17.7 ± 0.6 b	15.1 ± 0.2 bc	10.9 ± 0.4 c

Different letters within each column indicate significant differences (p < 0.5).

Hue evolution also differed among the samples supplemented with the MP-rich extracts (Table 3). The model wines added with A-S.cer, A-M.pul and A-T.del showed lower h_ab than the control at all the sampling points analyzed, which could be related to a copigmentation effect exerted by these extracts, as will be explained below. On the contrary, the addition of A-Pklu led to higher h_ab values from day 46 onwards, which were probably related to the promoted accumulation of VitB observed in this sample. However, the faster evolution of the hue of this wine was not sufficient to produce a color that could be visually distinguishable from the control (ΔE^⁎_ab < 3). Generally, the samples added with Z-S.cer, Z-L.the, Z-P.klu and Z-T.del exhibited similar hue values than the control wine.

Regarding the chroma of the model wines, it should be noted that the addition of A-M.pul resulted in higher C^⁎_ab values during the whole duration of the experiment, indicating that this sample had a higher color purity (i.e. a more vivid color) than the other wines (Table 3). This effect could also be observed in the sample added with A-S.cer, although it was significantly less important in the latter. The model wine added with A-T.del showed an initial increase of C^⁎_ab followed by its reduction after anthocyanin precipitation. In these three cases, the higher values of chroma could also be attributed to the copigmentation effect. The addition of the rest of the yeast extracts had little impact on the chroma of the model wines being the evolution of C^⁎_ab very similar to that of the control.

The increase in C^⁎_ab and the decrease in L^⁎ and h_ab observed in the model wines added with A-S.cer, A-M.pul and A-T.del (only in the beginning of the experiment for this last MP-rich extract) could be indicative of a copigmentation interaction between the yeast extracts and anthocyanins. In fact, the color measurements at day 0 show a bathochromic shift of the maximum absorption wavelength of these wine samples (Table S3 in the Supplementary Material). Moreover, the calculation of CCI yielded values of 16.7, 31.2 and 67.2% for the wines supplemented with A-S.cer, A-M.pul and A-T.del, respectively, suggesting that an important part of the color of these samples may be due to copigmentation interactions (Table S4 in the Supplementary Material). This shift in color could be visually perceptible (ΔE^⁎_ab > 3), as the model wines appeared bluer and darker (Fig. S6 in the Supplementary Material). The main contribution to the calculated color differences with the control was quantitative (i.e. the main changes were produced in L^⁎ and C^⁎_ab), while the change in hue represented between 17 and 24% of ΔE^⁎_ab (data not shown). However, this effect was only maintained until the end of the experiment in the sample supplemented with A-M.pul, being the bluer hue and darker color of this wine visually detectable (ΔE^⁎_ab > 3) during the whole duration of the experiment (Fig. S6 in the Supplementary Material). Many studies have shown that MPs do not seem to be good copigments of anthocyanins (Alcalde-Eon et al., 2024; Assunção Bicca et al., 2023). Therefore, it is likely that other components of the extracts are the ones exerting this effect. If we look at their composition (Table 1), we could see the presence of ribose in different proportions (5% in A-M.pul; and around 22% in A-S.cer and A-T.del), indicating the presence of nucleic acids. Therefore, the formation of copigmentation complexes might be occurring between anthocyanins and the nucleic acids contained in these extracts, since copigmentation between anthocyanins and DNA, RNA and other purine and pyrimidine derivatives is a documented phenomenon (Brouillard et al., 1991; Mistry et al., 1991; Sarma & Sharma, 1999). The fact that only the wine treated with A-M.pul maintained this effect during more than 9 months may indicate that the MPs contained in this extract were able to stabilize over time the copigmentation complexes formed. Indeed, a similar phenomenon was observed by Alcalde-Eon et al. (2024), in which a commercial MPs improved the stability of anthocyanin-flavonol copigmentation complexes over time. Furthermore, Guo et al. (2024) observed that MPs contributed to retain higher concentrations of malvidin 3-O-glucoside and quercetin under accelerated ageing conditions, probably trough the formation of binary/ternary complexes. Thus, the mixture of MPs and nucleic acids from the yeast in A-M.pul extract might be key for the stabilization of color. In addition, the lower content of ribose (5%) and, consequently, of nucleic acids, together with the higher proportion of mannose (93%) found in A-M.pul compared to A-S.cer and A-T.del could have contributed to this effect. Moreover, as previously discussed, other features of the MPs besides monosaccharide composition (i.e. polysaccharide structure, branching degree, molecular compactness, net charge, aminoacid composition, among others) may be influencing the stabilizing effect observed for A-M.pul. Nevertheless, further studies are needed in order to confirm a copigmentation interaction between anthocyanins and nucleic acids from yeasts in wine-like conditions as well as the possible role of MPs in the stabilization of these complexes.

Regarding the addition of A-T.del, after the initial increase in C^⁎_ab and decrease in L^⁎ and h_ab, the precipitation of anthocyanins caused by this extract translated into a clear decrease of C^⁎_ab and increase of L^⁎ leading to a loss of color that was detectable by the human eye (ΔE^⁎_ab > 3) (Fig. S6 in the Supplementary Material). However, lower values of hue were maintained in this wine over the course of the experiment resulting in this solution displaying more bluish tones, which, as explained before, could be related to the copigmentation effect exerted by the components of this extract. It is important to note that, despite this model wine showed a greater content of anthocyanins from the middle part of the experiment onwards, its lightness was significantly higher than that of the other samples making this wine to appear lighter (Fig. S6 in the Supplementary Material). Alcalde-Eon et al. (2024) suggested that several commercial MPs exhibited a slight anti-copigmentation effect, which was hypothesized to be associated with the ability of MPs to disrupt anthocyanin self-association. Therefore, the lighter color observed in this sample could be explained by the disruption of the self-association complexes of anthocyanins in the presence of A-T.del. On the other hand, a possible stabilization of the colorless hemiketal forms of anthocyanins by the addition of this yeast extract could also explain this fading of the color. In addition, this possible stabilization of hemiketal forms may also contribute to explain the higher concentrations of anthocyanin degradation products found in this sample (see Section 3.3.2.), since displacement of the equilibrium towards the formation of the hemiketal forms may facilitate their transformation into chalcones, which, as explained before, are intermediates in the anthocyanin degradation pathway. Further studies on the effect of MPs on anthocyanin self-association processes, as well as on the thermodynamic equilibrium among the different anthocyanin forms, would be useful to elucidate the mechanisms underlying these color changes.

Some authors have hypothesized that the copigmentation may slow down the oxidation and polymerization rates of anthocyanins, as these reactions are mainly attributed to the free forms of the anthocyanins (Boulton, 2001). Moreover, some authors have shown that interaction of anthocyanins with several polysaccharides (i.e. MPs, pectins and gum arabic) seemed to protect the anthocyanins from oxidation and condensation with nucleophiles such as ascorbic acid (Chung, Rojanasasithara, Mutilangi and McClements, 2015, Chung, Rojanasasithara, Mutilangi and McClements, 2016; Wan et al., 2023). In addition, in the work of Guo et al. (2024), the addition of MPs to a model wine seemed to protect malvidin 3-O-glucoside under accelerated ageing conditions. Therefore, it is possible that the slower rate of accumulation of vitisins and the higher contents of anthocyanins found at the end of the experiment in the wines treated with A-S.cer, A-M.pul and A-T.del could be associated to the interaction of the MPs contained in these extracts with anthocyanins and/or to the inclusion of anthocyanins into copigmentation complexes with nucleic acids, decreasing their susceptibility to condensate with pyruvic acid. It is necessary to stress out that, besides MPs, nucleic acids contained in A-S.cer, A-M.pul and A-T.del could have a role in pigment interactions and, consequently, in the pigment evolution of these model wines.

4. Conclusion

In conclusion, the rate of accumulation of VitA and VitB seemed to be affected by the addition of the MP-rich extracts, being this effect highly dependent on the yeast extract used. The mechanisms behind this effect remain to be clarified, but they do not seem to be related to the MP extracts participating directly as antioxidant or oxidant agents. No simple relations could be established between the MP extract characteristics (in terms of composition and apparent MW) and their effects on pigment evolution, since several structural and compositional features of MP extract could play an important role on that sense. Addition of the MP-rich extracts obtained from P. kluyveri, either A-P.klu or Z-P.klu, promoted vitisin accumulation, which could be related to some specific characteristic of the yeast extracts obtained from this yeast species. The nucleic acids present as other co-extracted components in A-S.cer, A-M.pul and A-T.del might be acting as copigments of anthocyanins, although the copigmentation effect was only maintained (for over more than 9 months) in the sample supplemented with A-M.pul, which could suggest that the MP characteristics of this extract allowed a greater stabilization of copigmentation complexes. Therefore, the characterization of the non-MP components of yeast extracts could help understand their effects and contribute to explain the different results found in the literature. The enhanced accumulation of anthocyanins and promoted color stability produced by A-M.pul suggest that its addition to wine could contribute to maintain a “younger” color. On the contrary, the addition of A-P.klu and Z-P.klu, in combination with other ageing techniques, may be useful to promote the chemical stability of the coloring matter and to produce red wines with a more aged color profile.

CRediT authorship contribution statement

María Oyón-Ardoiz: Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Elvira Manjón: Writing – review & editing, Supervision, Methodology, Investigation, Conceptualization. María Teresa Escribano-Bailón: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization. Ignacio García-Estévez: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization.

Funding sources

This work was supported by MCIN/AEI/10.13039/501100011033 and by “ERDF A way of making Europe” (Grant PID2021-127126OB-C21); and by Junta de Castilla y León-FEDER Programme (Project Ref. SA106P24 and ‘Escalera de Excelencia’ CLU-2025-2-04). M. Oyón-Ardoiz thanks Junta de Castilla y León, cofunded by Consejería de Educación and Fondo Social Europeo Plus (FSE+), her pre-doctoral contract.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.