Alternaria, Tenuazonic Acid and Spoilage Yeasts Associated with Bunch Rots of the Southern Oasis of Mendoza (Argentina) Winegrowing Region

Abstract

A study was carried out to identify the filamentous fungi and yeasts present in rotten wine grapes from two subzones of the Southern oasis of Mendoza winegrowing region, to assess the occurrence of tenuazonic acid (TA), a mycotoxin produced by the Alternaria genus, and to evaluate the wine spoilage potential of the associated yeasts in vitro and during microvinifications. The main fungal genera present were Alternaria (69.3%), followed by Aspergillus (16.8%), Penicillium (9.3%), and Cladosporium (4.6%), while the dominant yeast species Metschnikowia pulcherrima (23.1%), Aureobasidium pullulans (20.2%) and Hanseniaspora uvarum (13.0%) were followed by H. vineae (11.6%), Zygosaccharomyces bailii (10.4%), and H. guilliermondii (9.2%). Additionally, 94.1% of the rotten samples were contaminated with TA, with the highest level found in the Cabernet Sauvignon variety. No geographic association was found in the incidence of the different fungal genera or yeast species, nor in the occurrence of TA. Almost all of the tested yeasts produced H₂S, the majority of the Hanseniaspora strains produced acetic acid, and only one M. pulcherrima strain produced off-flavours in in vitro tests. Wines co-fermented with H. uvarum L144 and S. cerevisiae showed higher volatile acidity and lower fruity aroma and taste intensity. Therefore, processing bunch rot could pose a toxicological and microbiological risk to winemaking due to the high incidence of Alternaria and TA, as well as the potential of the associated yeasts to spoil wine.

1. Introduction

Argentina ranks 5th among wine-producing countries, with 10.9 million hectolitres (mhL) of wine production in 2024 [1]. In particular, the winegrowing region of the Southern Oasis of Mendoza province in Argentina has a tradition of producing quality wines derived from its valuable terroir. However, adverse weather events, exacerbated by climate change, have affected vineyards worldwide and global wine production, estimated at 231 million mhL in 2024, is at one of its lowest levels in more than six decades. Particularly, the Mendoza wine region has experienced an increase in extreme rainfall in recent years [2,3].

Increased rainfall before or during the harvest favours the development of vine diseases, especially those affecting the fruit, collectively known as bunch rots [4]. Bunch rots are caused predominantly by fungi and, in many cases, form complex associations that vary depending on the winegrowing region considered. In addition to Botrytis cinerea, which is widely recognised as the cause of grey mould or botrytis bunch rot, a number of other filamentous fungi have also been associated with bunch rot, including Aspergillus, Penicillium and Alternaria, but its diversity has generally been underestimated [5,6]. Processing even low levels of rotten grapes during winemaking can pose technological and quality risks [4]. High-molecular-weight polysaccharides and small volatile molecules produced by fungi caused processing difficulties and earthy off-flavours, respectively. Additionally, the breakdown of anthocyanins and proanthocyanidins by laccases, as well as the oxidation of desirable aromatic compounds, negatively impact the quality attributes of wine, such as colour and aroma. More worryingly, some of the decaying or secondary intruder fungi have the potential to produce mycotoxins, which can compromise wine safety [4]. In previous studies, we have demonstrated a high incidence and prevalence of the Alternaria genus on healthy Malbec grapes across different years in the Southern oasis of Mendoza winegrowing region, as well as the ability of isolated Alternaria alternata strains to produce mycotoxins (alternariol -AOH, alternariol monomethyl ether -AME- and tenuazonic acid -TA) in specific media under temperature and humidity conditions commonly encountered during ripening [7,8]. We also reported the ability of A. alternata strains to produce TA on detached Malbec grapes and the natural occurrence of TA in Malbec, Cabernet Sauvignon and Syrah from Argentina [9,10]. Additionally, AOH, AME and TA have already been detected in grape juice and wine, certainly as a consequence of Alternaria presence in wine grapes [11,12,13,14,15]. As Alternaria toxins are increasingly regulated in food due to their numerous adverse effects on humans [16,17,18], and rotten wine grapes could be a significant source of contamination, monitoring the presence of Alternaria together with their toxins is crucial to defining the toxicological risk to wine and developing the most effective control strategies.

On the other hand, among the complex of organisms associated with bunch rot, yeasts could also imply a risk to winemaking since some of them are capable of surviving the fermentation process and can contribute to wine spoilage. According to Barata et al. [19], the availability of high concentrations of sugars on the grape surface, as a result of skin damage or rotting, favours the proliferation of highly fermentative ascomycetes yeasts, such as Pichia spp. and Zygoascus hellenicus, as well as some harmful wine spoilage yeasts (e.g., Zygosaccharomyces spp., Torulaspora spp.). The main yeast genera involved in wine spoilage can be grouped according to the specific defects they can cause. Some species of the genus Kloeckera/Hanseniaspora, called apiculate yeasts, are associated with excessive production of ethyl acetate and acetic acid; species of the genera Candida and Pichia, generally film-forming yeasts, are involved in the formation of pellicles on the wine’s surface and the production of off-flavour compounds (oxidised odours and volatile phenols) [20]. Finally, species of Brettanomyces/Dekkera bruxellensis, Saccharomyces cerevisiae, Saccharomycodes ludwigii, Schizosaccharomyces pombe, and Zygosaccharomyces bailii, have been catalogued as “sensu stricto” spoilage yeasts, since they can form visible sediment and cloudiness in dry wines and referment sweet wines, often with the production of off-flavours (e.g., acetic acid, acetaldehyde, and volatile phenols) [20]. However, the spoilage ability of yeasts appears to be species or even strain dependent [21]. Consequently, an assessment of the yeast population associated with rotten grapes and its spoilage potential could be used to predict the microbiological risks of the wine-growing region in focus and, therefore, to select the most appropriate prevention methods.

Recently, a digital, open-access platform has been developed showing the characterisation of Argentina’s Wine Regions based on the delimitation of homogeneous areas for zoning winegrowing regions in the different provinces [22]. This project was supported by the analysis and classification of edaphic and climatic data obtained through climate characterisation studies, as well as the geological, geomorphological and edaphic characterisation of the winegrowing regions. In particular, Mendoza’s winegrowing regions were grouped into four zones: the Northern Oasis, the Southern Oasis, the Lower Tunuyán Oasis and the Upper Tunuyán Oasis. The Southern Oasis zone covers 25,007 hectares of vine-cultivated land around the Atuel and Diamante rivers. In addition, this oasis has been divided into three climatic subzones, and each of these has been further divided into two or three edaphic subzones, making a total of eight subzones (1-1, 1-2, 1-3, 2-1, 2-2, 3-1, 3-2, 3-3) [22]. Since geographical associations have been previously reported in filamentous fungi and yeast populations as well as in mycotoxin occurrence in other winegrowing regions [23,24], whether this phenomenon can also be seen in rotten wine grapes from the Southern oasis of Mendoza should be addressed.

Therefore, the aims of this work were (i) to identify the filamentous fungi and yeasts associated with rotten wine grapes from two subzones of the Southern Oasis wine-growing region of Mendoza, (ii) to assess the natural occurrence of TA from the same subzones, and (iii) to evaluate the wine spoilage potential of the isolated yeasts under in vitro test and during microvinification trials.

2. Materials and Methods

2.1. Study Area and Sampling Procedure

The Southern Oasis winegrowing region of Mendoza (Argentina) covers an area between 34.5° and 35.2° S latitude, 67.4° and 68.6° W longitude, and altitudes between 400 and 900 metres. During the 2017 harvest, 17 Vitis vinifera L. vineyards exhibiting symptoms of rot were sampled from two subzones (3-1 and 3-3) of the Southern Oasis winegrowing region of Mendoza (Figure 1). These subzones have similar climatic conditions (semi-arid with an average annual precipitation of 365 mm, and continental thermal values), but their edaphic conditions differ [25]. The subzone 3-1, which is comprised between 450 and 600 m a.s.l., has deep soils (1.5 to 2 m) with a sandy texture and the highest clay and silt contents, and has a moderately alkaline pH and the highest salinity and total calcareous contents within the oasis (with no sodicity problems due to the gypsum present). Meanwhile, subzone 3-3, which is comprised between 600 and 800 m a.s.l., has shallow soils with sandy textures to the north and loamy textures to the south, with non-saline and non-sodic soils and a moderately alkaline pH. Additionally, samples collected in subzone 3-1 belong to different wine grape varieties (Bonarda, Malbec, Syrah, Criolla, Cereza, Cabernet Sauvignon and Torrontés), whereas samples from subzone 3-3 belong to the Bonarda variety only. Each independent sample consisted of different grape bunches with rotting symptoms (this included berries that were discoloured, shrivelled, smelled of vinegar, and/or had signs of fungal growth) distributed along each vineyard, reaching approximately 1 kg. A total of 17 samples were kept in individual plastic bags and placed in ice-cooled boxes for transportation to the laboratory, where they were subjected to further processing and analysis as described below.

2.2. Enumeration and Isolation of Filamentous Fungi and Yeasts from Rotten Grapes

The samples were submitted to plate count methodology for microbiological analysis [26,27]. In brief, 100 symptomatic berries were taken from each sample (approximately 100 g) and processed with 900 mL of 0.1% peptone water in a laboratory mixer. Decimal dilutions were then carried out (10⁻¹ to 10⁻⁶). The laboratory mixer was disinfected with hypochlorite solution (1%) and sodium thiosulfate solution (0.5%) at the beginning and between samples. The enumeration and isolation of filamentous fungi and yeasts was carried out by surface plating (0.1 mL) of the adequate dilutions onto plates (duplicates) of Dichloran-Rose-Bengal-Chloramphenicol Agar (DRBC) medium [28] followed by incubation at 25 °C for 7 days. All media plates containing countable filamentous fungi or yeast colonies (10 to 100) were examined, all different colony morphological types were registered, differentially counted and representative isolates from each morphology were selected for further identification. When colonies from the same type were identified as different genus or species, the respective proportion was changed accordingly. In addition, the relative abundance or relative density (RD) of each fungal genus or yeast species was calculated in each sample as the number of isolates of a genus or species/total number of isolated fungal genus or yeast species × 100 [7].

Filamentous fungi colonies were sub-cultured in Czapek-Yeast extract-Agar (CYA) medium [28] at 25 °C and single conidial colonies were obtained. The yeast isolates were streaked and purified on Yeast-Peptone-Dextrose (YPD) plates. All isolates were stored at 4 °C.

2.3. Identification of Filamentous Fungi, Yeast and Yeast-like Organisms

The filamentous fungi isolates were identified morphologically according to Pitt and Hocking [28]. Briefly, single spore isolates were grown on Czapek Yeast extract Agar (CYA), Malt Extract Agar (MEA) and 25% Glicerol Nitrate (G25N) media under temperature regimens of 5, 25 and 35 °C. After 7–10 days of incubation, identification to genus level was made based on the macroscopic characteristics of the colonies on agar medium (diameter, colour and texture) and microscopic characteristics (reproductive structures, conidiophores, types of conidia, fruiting bodies, among others).

The isolated yeasts and yeast-like organisms were identified by the molecular method of restriction fragment length polymorphisms (RFLPs). Firstly, DNA extraction was carried out following a procedure previously described by Querol et al. [29]. Then, DNA were used to amplify the region between the 18S rRNA and 28S rRNA genes with the specific internal transcribed spacers ITS1 and ITS4 primers [30]. This region contains the highly conserved region of ribosomal 5.8S, and a variable zone which is the region on the ITSs. The amplified genes were then treated with restriction enzymes CfoI, HindI, HaeIII and DdeI for identification of yeasts at species level [30,31]. The sizes of amplified and restriction fragments were analysed using YEAST-ID (www.yeast-id.org, accessed on 26 December 2024) to assign species.

2.4. Tenuazonic Acid Extraction, Detection and Quantification

For TA analysis, each sample was completely ground (whole grape bunches) in a laboratory mixer and three replicate aliquots of each (2.5 g) were collected in 50 mL PTFE plastic tubes and stored at −20 °C. Each replicate aliquot from a sample was submitted independently to the TA extraction procedure, following a high-throughput modified QuEChERS (quick, easy, cheap, effective, rugged and safe) method previously developed for TA extraction from wine grapes [9]. The final extract was resuspended in 0.5 mL mobile phase [(MeOH: 0.1 M NaH₂PO₄ (2:1 v/v), adjusted to pH 3.2] and 20 µL were injected in the HPLC MWD system (DionexSoftron GmbH, Thermo Fisher Scientific Inc., Germering, Germany). The working wavelength for the analyte was 279 nm. HPLC separations were carried out in a Kinetex XB-C18 column (4.6 mm × 150 mm, 5 µm) Phenomenex (Torrance, CA, USA) and TA mobile phase and running conditions were those described by Fontana et al. [9]. Samples were quantified by using a matrix-matched calibration. Limit of detection (LOD, signal-to-noise ratio 3) was 10 µg/kg and the quantification limit (LOQ) (lowest concentration of the analyte with recovery within the range 70–120% and relative standard deviation ≤ 20% by applying the complete analytical method) was 50 µg/kg.

Copper salt of TA (Sigma-Aldrich, Steinheim, Germany) was converted into its free form as described in the literature [32]. Stock solutions of TA were prepared in methanol (MeOH). Further dilutions were prepared monthly in MeOH and stored in brown bottles at −20 °C to ensure stability.

2.5. In Vitro Evaluation of Spoilage Features in Isolated Yeasts

Fifty-nine representative yeasts and yeast-like isolates from different species were evaluated together with the reference fermentative strain S. cerevisiae IOC 18-2007^® (Institut Œnologique de Champagne, Epernay, France) and a native strain of D. bruxellensis V3-4-11, a typical “Brett” character producer previously isolated in the Southern Oasis of Mendoza winegrowing region [33].

2.5.1. Acidity Production

Young cultures (24 h) of each yeast isolate were spot-inoculated in Petri dishes containing calcium carbonate medium (g/L: glucose 5, calcium carbonate 0.5, yeast extract 0.5, agar–agar 2) [34]. Dishes were kept at 25 °C for 2 weeks. The test was positive when a clear halo was developed around the colony and the width of clarification halos were calculated as the difference between the total diameter of the clarification halo and the colony diameter. The experiments were performed in duplicate.

2.5.2. Hydrogen Sulphide Production

The ability of the isolates to produce different levels of hydrogen sulphide (H₂S) was evaluated on BIGGY agar [35]. The medium was spot-inoculated and plates were incubated at 25 °C for 48 h. Finally, the resulting colour of the colony was related to H₂S production, from white (no production) to dark brown (high production) [36]. An additional criterion was established by considering the reference strain S. cerevisiae IOC 18-2007^® as a negative H₂S producer, because of its low production on BIGGY agar medium and its reported oenological properties. Therefore, darker colony colour than that obtained by S. cerevisiae IOC 18-2007^® was considered as a positive H₂S producer. The experiments were performed in duplicate.

2.5.3. Unpleasant Aroma Production

Each indigenous isolate was inoculated at 1 × 10⁶ cells/mL in into 15 mL screw-capped tubes containing 13 mL of pasteurised Sauvignon Blanc must supplemented with p-coumaric acid 100 mg/L following Sangorrin et al. [34] and Portugal et al. [37] with modifications. Unpleasant visual and aromatic characteristics were recorded after 15 days at 25 °C. The experiments were performed in duplicate. A sensory analysis was conducted by a tasting panel of three expert assessors (oenologists from the FCAI-UNCUYO and members of the National Institute of Vitiviniculture in San Rafael, Mendoza) to evaluate the undesirable growth features and the generation of altered aromas [37]. The following criteria were used: in the visual phase, description of unpleasant characteristics (film, flocculation and cloudiness) and in the aromatic phase, classification of the aroma, as follows: absence of off-odour; generation of a typical Brettanomyces off-odour (“Brett” character); or presence of another off-odour, different to the typical “Brett” character, in which case the aroma was specified. The resulting fermented musts (13 mL samples) were presented at 13–15 °C in ISO wine glasses (ISO, 1977) [38], labelled with three-digit random codes and covered with plastic lids to trap volatiles. All samples were evaluated in ten sessions at controlled temperature (20 °C) in the wine tasting room at the FCAI-UNCUYO, which has individual booths, each equipped with a sink and both white and red lighting.

2.6. Microvinification Trials

Malbec red grapes (Vitis vinifera L.) were obtained from a commercial vineyard located in the Rama Caída district (34.66° South latitude and 68.38° West longitude) in the Southern Oasis of Mendoza winemaking region. The grapes were crushed and destemmed immediately at the FCAI-UNCUYO winery. The resulting must (reducing sugar 230.3 g/L, titratable acidity 3.71 g/L of tartaric acid, and pH 4.16) was fractionated by aliquoting equal volumes into individual containers maintaining a consistent solid-to-liquid ratio across all replicates, and each replicate was subsequently pasteurised separately [39]. The pasteurised must in the presence of skins was treated with 1 g/Kg tartaric acid (correcting the pH to 3.93) without the addition of sulphite.

Microvinifications were carried out in 1 L Erlenmeyer flasks containing 800 mL of Malbec must per flask in duplicate at 25 °C. Each potential spoilage yeast (pre-adapted in 12° Bx sterile Malbec must at 25 °C for 3 h) and the commercial S. cerevisiae strain (IOC 18-2007, Institut Œnologique de Champagne, Epernay, France) were inoculated simultaneously at 10⁶ cells/mL and 10⁴ cells/mL, respectively. Fermentation control consisted of inoculating S. cerevisiae alone [39].

The progress of the alcoholic fermentation was monitored by daily measurements of weight loss of the flasks until constant weight for two consecutive days. During the skin contact period, punch-downs were carried out once a day for two minutes. After fermentation, the wines were devatted and clarified after 14 days. They were then transferred to 250 mL glass bottles and stored in the dark at 12 °C until analysis.

2.6.1. Physicochemical Analyses of Wines

Main oenological parameters (ethanol, %v/v; residual sugars, g/L; total acidity, g/L tartaric acid; volatile acidity, g/L acetic acid; and pH) were measured in wines at the end of fermentation following the official methods of the International Organisation of Vine and Wine [40].

2.6.2. Sensory Analysis of Wines

A descriptive sensory analysis of wines was carried out three months after bottling by a tasting panel consisting of eight expert assessors (four males and four females) consisting of oenologists from the FCAI-UNCUYO and members of the National Institute of Vitiviniculture (San Rafael, Mendoza), with ages ranging between 26 and 60 years old [41]. All assessors provided voluntary verbal consent prior to participation. The experimental protocol involving sensory evaluation was in accordance with the relevant operation specification in Argentina. Approximately 30 mL samples of wine were presented at 16–18 °C in ISO wine glasses (ISO, 1977), labelled with three-digit random codes and covered with plastic lids to trap volatiles, following a fully randomised serving order. Two consecutive sessions were performed on different days. During the first session, the assessors reached a consensus on the sensory descriptors to be used, selecting the following properties: colour intensity, violet hue, aroma intensity, aroma quality, fruity aroma, floral aroma, acidity, bitterness, astringency, body, taste intensity, taste quality, and harmony-balance [39]. During the second session, panellists rated the intensity of each descriptor on a scale from 0 (not perceivable) to 5 (very strong). To minimise the effects of sensory carry-over, panellists were asked to rinse their mouths with mineral water and eat unsalted crackers between samples, following a ‘sip and spit’ procedure. All samples were evaluated at a controlled temperature (20 °C) in individual booths equipped with a sink and both white and red lighting. The evaluation followed standard sensory analysis procedures adapted from Stone et al. [41].

2.7. Statistical Analysis

Data on the relative abundance of each filamentous fungi genus and each yeast species as well as the natural occurrence of TA in rotten samples, were subjected to analysis of variance (ANOVA), followed by the LSD Fisher test to determine significant differences (p < 0.05) between them. Pearson correlation coefficients were used to evaluate the relationship between Alternaria relative abundance (%) and TA (µg/kg) occurrence in rotten samples. The width (mm) of the clarification halo in YPD-CaCO₃ medium of acetic acid-producing yeast and yeast-like strains was analysed by ANOVA after transformation was applied, followed by the LSD Fisher test. Different letters showed significant differences (p < 0.05). The physicochemical parameters and sensory analysis of the Malbec wines were subjected to either ANOVA and LSD Fisher test (p < 0.10) or Kruskal–Wallis analysis (p < 0.05) to determine significant differences among the wines. Statistical analyses were performed using the Software Infostat^® (InfoStat version 2013, FCA, Universidad Nacional de Cordoba, Cordoba, Argentina).

3. Results and Discussion

3.1. Filamentous Fungi and Yeasts Present in Wine Grape Bunch Rots

The mycological study of the grape bunch rot samples collected in the Southern Oasis winegrowing region of Mendoza (Argentina) during the 2017 harvest showed that filamentous fungi were recovered from almost all samples (76.9% in subzone 3-1 and 75% in subzone 3-3). Alternaria spp. showed predominance in terms of relative abundance (62.3% in subzone 3-1 and 88.0% in subzone 3-3), followed alternatively by Aspergillus (23.0% in subzone 3-1 and 0.4% in subzone 3-3), Penicillium (11.2% in subzone 3-1 and 4.3% in subzone 3-3) and Cladosporium (3.5% in subzone 3-1 and 7.4% in subzone 3-3), including almost all the samples with filamentous fungi (except two samples in subzone 3-1, one of Syrah and one of Criolla, whose population could not be recovered) (Figure 2A). In addition, a small proportion (less than 2.5%) of Phoma and other unidentified genera was found through the two analysed subzones. No significant differences were found in the relative abundance of each fungal genus between the different grape varieties present in subzone 3-1 or between the two subzones studied (3-1 and 3-3) (p > 0.05, ANOVA, LSD-Fisher test).

Although Alternaria is considered to be an opportunistic rather than a true pathogen, several cases of Alternaria rot in grape berries have been reported [42,43,44,45,46]. In addition, a first study of rotten wine grapes from south-west Western Australia during 1976–77 has already reported the presence of Alternaria [47]. Also, in a survey of rotten wine grape berries carried out in vineyards from Pennsylvania, Maryland and Virginia from the US during 2014–2020, Alternaria was reported as the 4th most common genera isolated after Botrytis, Colletotrichum and Aspergillus [48]. Alternaria spp. was also the predominant fungus isolated from wine grape bunch rot during post-harvest withering [49,50]. Interestingly, the results from the present work are similar to those previously obtained on healthy Malbec grapes from the same winegrowing region with regard to the major genus Alternaria (58% in 2011 vintage, 81% in 2012 and 2013 vintages, except that this was followed by Cladosporium (22% 2011, 19% 2012 and 7% 2013) and with a lower incidence of Aspergillus (11% 2011 and 3% 2013) and Penicillium (3% 2011 and 4% 2013) [7]. On the other hand, no influence of wine grape variety or subzoning (with edaphic differences) was observed on the relative abundance of fungal genera present in rotten grapes in the Southern Oasis of Mendoza winegrowing region. Although no other works have explicitly linked the micobiota in relation to the different edaphic conditions of a winegrowing region, a geographic association has been previously reported for some fungal genera or species in certain winegrowing regions. In partial agreement with our work, the geographical origin of grapes in Portugal did not significantly influence the incidence of Alternaria, Cladosporium and Epicoccum, but had a significant effect in A. niger, B. cinerea and Penicillium species [23]. In addition, A. niger as well as A. carbonarius seems to prevailed in vineyards with Mediterranean climates (zone CI-IIb), while Botrytis, as well as T. roseum in vineyards with more temperate climates (zone CIa) [23,51]. In the northern vineyards of France, Penicillium spp. predominated, whereas Aspergillus spp., including A. niger and A. carbonarius, were found in the southern vineyards with Mediterranean or tropical climates [52]. In different wine-growing regions in Argentina, a positive correlation was found between the isolation percentage of A. carbonarius in grapes and temperature [53]. However, no statistically significant differences were observed in the occurrence of Aspergillus section Nigri species in seven vineyards located mainly along the Mediterranean coast of Spain [54]. In addition, considering different wine grape varieties, Alternaria and Cladosporium were more frequently isolated from red than from white grape varieties in Portugal [26], but the differences in the isolation frequency of ochratoxinogenic fungi were uncorrelated to berry colour in Spain [55]. Meanwhile, the highest percentage of A. carbonarius (50% of the isolates) was found on Cabernet Sauvignon grown in Argentina [53]. A. carbonarius and A. niger were also more abundant on red than on white grape varieties grown in Chile [56]. By contrast, Melki Ben Fredj et al. [57] reported that the development of Aspergillus spp. and Penicillium spp. on Tunisian wine grapes would depend more on the climate conditions than on the grapevine genotype.

On the other hand, yeasts and yeast-like organisms were recovered from almost all of the rotting samples collected from the Southern Oasis winegrowing region of Mendoza (Argentina) during the 2017 vintage (88.2% in subzone 3-1 and 100% in subzone 3-3). In addition, the molecular identification profile of the representative isolates, based on PCR-RFLP analysis of the ITS1-5.8S-ITS2 region of the rRNA gene, revealed the presence of six relevant species (

Table 1. Different patterns obtained from the amplification product and the restriction length size of the ITS1-5.8S-ITS2 region of the rRNA gene of the representative yeast isolates.

Species	AP a (bp)	Restriction Lengths (bp)
		Cfo I b	Dde I c	Hae III d	Hinf I e
Aureobasidium pullulans	600	100 + 180 + 190	ND	150 + 450	130 + 180 + 290
Hanseniaspora guilliermondii	775	105 + 320 + 340	90 + 120 + 160 + 360	775	160 + 200 + 360
Hanseniaspora uvarum	775	100 + 320 + 340	80 + 100 + 170 + 300	775	160 + 200 + 370
Hanseniaspora vineae	775	90 + 150 + 180 + 270	80 + 230 + 460	110 + 660	370 + 390
Metschnikowia pulcherrima	400	95 + 100 + 205	ND	100 + 280	190 + 200
Zygosaccharomyces bailii	790	95 + 95 + 270 + 320	ND	90 + 690	55 + 160 + 225 + 340

^a 5.8S-ITS-amplified product size. ^b,c,d,e restriction enzymes used. ND, not determined.

). Among them, Metschnikowia pulcherrima (19.4% in subzone 3-1 and 33.3% in subzone 3-3), the yeast-like Aureobasidium pullulans (25.6% in subzone 3-1 and 5.4% in subzone 3-3) and Hanseniaspora uvarum (9.4% in subzone 3-1 and 23.0% in subzone 3-3) were predominant in terms of relative abundance (Figure 2B). They were followed by Hanseniaspora vineae (15.8%), Zygosaccharomyces bailii (14.2%) and Hanseniaspora guilliermondii (12.5%) in subzone 3-1. In addition, three minor morphotypes (each accounting for less than 6.4%) were found through the two analysed subzones, but did not survive to be identified. Furthermore, no significant differences were found in the relative abundance of any of the yeast species between the different grape varieties present in subzone 3-1 or between the two subzones studied (3-1 and 3-3) (p > 0.05, ANOVA, LSD-Fisher test).

Similarly to our work, Metschnikowia spp., Hanseniaspora spp. and A. pullulans have previously been described as predominant inhabitants of rotten grapes from New York (USA) vineyards during the 1999 to 2002 vintages [58]. In particular H. uvarum has been reported as the most common species in grey and noble rot from Athens (Greece) vineyards during the 2005 vintage, as well as from Hunter Valley, New South Wales, Australia [59,60]. Furthermore, despite the variation in proportions, M. pulcherrima, A. pullulans and H. uvarum were the three most abundant species reported on healthy Malbec grapes from the same wine-growing region that had been studied previously (A. pullulans 44.4% in 2011, 100% in 2012 and 84.8% in 2013; H. uvarum 22.2% in 2011 and 4.5% in 2013; and M. pulcherrima 33.3% in 2011) [61]. However, H. vineae, Z. bailii and H. guilliermondii were not found in healthy Malbec wine grapes during our previous study in 2011, 2012 and 2013 vintages. According to Barata et al. [19], the proportion of each microorganism constituting the microbial consortium of wine grapes depends on the grape ripening stage and nutrient availability, and they can be grouped according to similar physiological characteristics. Among them, the third group, commonly described in the high nutrient availability environment of rotting grapes, is composed of lower proportions of basidiomycetes yeasts, the yeast-like fungi A. pullulans, and the weakly fermentative yeasts Hanseniaspora spp., Metschnikowia spp. and Starmerella bacillaris (named as Candida zemplinina), and higher proportions of strongly fermentative yeasts such as Pichia spp., Zygosaccharomyces spp., Zygoascus spp., Torulaspora spp.). In line with these statements, Z. bailli was reported in the present work and in other studies, as an inhabitant of rotten grapes, consistently isolated from grapes affected by noble rot, sour rot and honeydew from different viticultural regions worldwide [60,62,63,64]. Interestingly, a diversity of species of the genus Hanseniaspora was found in rotten wine grapes from the Southern Oasis of Mendoza winegrowing region, including H. uvarum, H. guilliermondii and H. vineae. H. guilliermondii has previously been reported as an inhabitant of healthy and Botrytis-infected grapes of two red Vitis vinifera varieties (Mavroliatis and Sefka) from a vineyard in Attica, Greece [60] and in healthy wine grapes of Bangalore Blue, Zinfandel, Cabernet, Chenin Blanc, Sauvignon Blanc and Syrah varieties from vineyards in India where H. vineae was also found (Cabernet and Shiraz) [65]. However, this is the first report of H. vineae in rotten wine grapes. Finally, no influence of wine grape variety or site effect (sub-zones) was observed on the relative abundance of yeast species present in rotten grapes in the Southern Oasis of Mendoza winegrowing region. Previous work indicates that there are no obvious influences of grape variety on yeast species and populations in healthy wine grapes, but that some vineyard or geographical associations were observed, although these were not further explored to be related to edaphic differences. Raspor et al. [66] previously reported that the number of yeasts and yeast species isolated varied according to different grape varieties of V. vinifera (Žametovka, Modra frankinja and Kraljevina) and according to different sampling location of the Dolenjska vine-growing region, Slovenia. However, as in our work, the yeast communities remain similar in Cabernet Sauvignon and Merlot grapes, indicating that the analysed parameters (grape variety/vintage) are not interfering with the yeast populations found in the region’s highlands of Rio Grande do Sul, Brazil [67]. Also, a study carried out through 4 Denomination of Origin from Galicia (Spain) showed that wine grape yeast community differences were more related to location than grapevine variety and/or farming system, supporting the existence of yeast geographic patters (microbial terroir) on the region [24].

3.2. TA Natural Occurrence

Almost all the rotten wine grape samples from the Southern Oasis of Mendoza winegrowing region collected during the 2017 vintage showed TA contamination by the QuEChERS-HPLC-UV method applied, with an incidence of 100% (13/13) in subzone 3-1 and 75.0% (3/4) in subzone 3-3 (

Table 2. Occurrence of TA in rotten wine grapes from the Southern Oasis of Mendoza (Argentina) during the 2017 harvest.

Subzone	Wine Grape Variety	TA in Wine Grapes
		Number of Positive Samples 1	Range (µg/kg) 2
3-1	Bonarda b	5/5	<LOQ 3; 249.29 ± 0.00
	Cabernet Sauvignon a	1/1	1395.68 ± 32.95
	Cereza b	2/2	<LOQ 3; 75.61 ± 17.68
	Criolla b	2/2	<LOQ 3; 103.47 ± 25.00
	Malbec b	1/1	<LOQ 3
	Syrah b	1/1	76.90 ± 0.00
	Torrontés b	1/1	140.14 ± 14.10
3-3	Bonarda	3/4	<LOQ; 67.27 ± 19.36

¹ Number of samples with TA values over the LOD versus total samples, LOD (Limit of detection): 10 μg/kg. ² Range of TA concentration in positive samples, each sample value includes the average of TA concentrations (µg/kg) with their standard deviations, n = 3 replicates. ³ LOQ (Limit of quantification): 50 μg/kg. Different letters indicate statistically significant differences (p < 0.05, LSD Fisher, ANOVA) between the analysed wine grape varieties in subzone 3-1. No significant differences were found between the analysed subzones concerning Bonarda variety according to Fisher’s LSD test (p > 0.05, ANOVA).

). The highest TA value in subzone 3-1 was 1395.68 ± 32.95 µg/kg, while the highest TA value in subzone 3-3 was 67.27 ± 19.36 µg/kg. The Cabernet Sauvignon variety showed higher TA values than the other varieties present in subzone 3-1 (p < 0.05, ANOVA, LSD-Fisher test), while no significant differences were found between subzones 3-1 and 3-3 for either the Bonarda variety or all varieties studied (p > 0.05). In addition, no correlation was found between Alternaria relative abundance (%) and TA concentration of the samples (Pearson, p > 0.05).

In our first survey of Malbec wine grapes from the Southern Oasis of Mendoza winegrowing region during the 2015 vintage, the only visibly infected sample (M14) showed a TA value of 595 ± 50 µg/kg [9]. In another study conducted during the 2016 vintage in the same region, the incidence of TA in rotting wine grapes of Malbec, Cabernet Sauvignon, Syrah, Torrontés and Chenin varieties was 42% (5/12), with a maximum value of 778 ± 15 μg/kg in a Cabernet Sauvignon sample [10]. In the present study of the 2017 vintage, the Cabernet Sauvignon variety showed the maximum TA value, as in the 2016 vintage, but the value (1395.68 ± 32.95 µg/kg) was the highest ever recorded in the Southern Oasis of Mendoza winegrowing region, and the TA incidence (94.1%) considering all samples was also the highest. Importantly, this study strengthens the hypothesis that rotten wine grapes provide a more favourable environment for mycotoxin production than healthy grapes, given that the incidence and levels of TA are significantly higher [9,10]. In addition, a varietal effect was observed in the TA values of rotten grapes, with Cabernet Sauvignon showing higher TA values, but no influence of the different sub-zones (with edaphic differences) from the Southern Oasis of Mendoza winegrowing region was found. Previous works have shown that meteorological conditions and geographic area, with climatic differences, can contribute not only to the variation in the incidence of some OTA-producing fungi, but also to OTA contamination in grapes [23,52,68,69]. In particular, Chiotta et al. [53] have founded that OTA levels in grapes and rain at harvest time correlated positively. However, the influence of winegrowing regions with different edaphic conditions on the natural occurrence of OTA or other mycotoxins in vineyards was not previously studied. Furthermore, certain wine grape varieties such as Cabernet Sauvignon, Trebbiano and Verdeca from two vineyards in Puglia (southern Italy) were found to be more susceptible to A. carbonarius infection and OTA production during in-uva experiments [68]. Meanwhile, the mean OTA production in the Vinhão variety at ripening was higher than in the other grape varieties (Alvarinho, Loureiro, Tinta Barroca, Touriga Franca and Cabernet Sauvignon) from different Portuguese vineyards [70].

On the other hand, no correlation between Alternaria relative abundance (%) and TA production was found in the present study. This could raise to the first question of whether the Alternaria present in rotten wine grapes belong to mycotoxin-producing species. Although the Alternaria spp. isolates were not further identified, they could be included in the Alternaria section due to their similarity in microscopic characteristics to the small-spored Alternaria species [71]. Even today, the assignment of species within this section remains difficult, due to the absence of coherent morphological features and the limited variability of the molecular markers [72]. In a previous study carried out in the same winegrowing region, all Alternaria spp. strains isolated from healthy wine grapes were included into A. alternata species-group according to morphological and molecular markers (Alt a1), a species-group defined by Pavon et al. [73] equivalent to the latter defined Alternaria section [71]. Moreover, almost all (97%) of the analysed strains showed toxicogenic ability (alternariol, alternariol monomethyl ether, TA), with TA as the toxin produced in the highest frequency (97%) and at the highest levels [7]. Similarly, all Alternaria strains from chickpeas submitted to phylogenetic analysis (tef1, gpd, and Alt a1 genes) clustered into the section Alternaria, and no matter the particular species (A. alternata or A. arborescens), most of the strains were able to produce at least one of the mycotoxins tested (alternariol, alternariol monomethyl ether, TA) [72]. Therefore, beyond species-level identification, isolates belonging to the Alternaria section seem to display a high toxicogenic potential. However, it could not be expected that this would guarantee the occurrence of mycotoxins under field conditions. Similarly to our work, a lack of correlation between mycotoxin producing fungi presence (Aspergillus section Nigri) and mycotoxin incidence (OTA) in wine grape samples was founded, meanwhile a positive correlation with rainfall levels and OTA incidence was reported [53]. In addition, no correlation between the OTA amount in musts and the contamination by Aspergillus species in Tunisian vineyards was found [57]. Our previous studies have shown that favourable environmental conditions (temperature and water activity or relative humidity) for Alternaria growth or grape infection do not necessarily coincide with those for mycotoxin production (alternariol, alternariol monomethyl ether and TA) and vice versa [8,10]. Particularly, high amounts of TA were obtained at 0.96 a_W and 15 or 30 °C, showing that certain stressful conditions for growth also promote mycotoxin production. The decaying environment found in rotten grapes might provide a stressful condition for Alternaria, with an increased number of other organisms competing for nutrient resources, which in turn could respond by increasing TA production, as mycotoxins can act as antibiotics to suppress the growth of commensal microorganisms [74].

3.3. Evaluation of the Wine Spoilage Potential of the Isolated Yeasts

Of the total number of yeasts and yeast-like organisms analysed (59), 37.3% produced acetic acid in YPD-CaCO₃ medium and all producing species belonged to the genus Hanseniaspora (

Table 3. Detrimental oenological features of isolated yeasts and yeast-like strains from rotten wine grapes in the Southern Oasis winegrowing region of Mendoza during the 2017 harvest.

Species	Acetic Acid Production		N° of Positive H2S-Producing Strains 2/Total	Defect in Grape Must
	N° of Positive Strains/Total	Width 1 (mm) of the Clarification Halo
M. pulcherrima	0/17	nd	17/17	M. pulcherrima L1: solvent and acetate odour; cloudy
A. pullulans	0/9	nd	5/9	nd
H. guilliermondii	7/8	(0.986 ± 0.305) b	8/8	nd
H. uvarum	9/10	(1.694 ± 0.919) ab	10/10	nd
H. vineae	6/10	(1.104 ± 0.279) ab	10/10	nd
Z. bailii	0/5	nd	5/5	nd
D. bruxellensis *	1/1	(5.375) a	0/1	“Brett” defect: smell of horse sweat, rubber; rotten; cloudy
S. cerevisiae *	0/1	nd	0/1	nd

¹ The width (mm) of the clarification halo in YPD-CaCO₃ medium (average ± SD) of acetic acid producer strains. Different letters show significant differences (p < 0.05) according to the ANOVA and LSD-Fisher test. nd, not detected. ² The production of H₂S was considered positive when the evaluated strains generate a darker colour than S. cerevisiae IOC18-2007^® in the BIGGY agar medium. * Reference strains: Dekkera bruxellensis V3-4-11, previously isolated as a “Brett” character strain in grape must; S. cerevisiae IOC 18-2007^®, a commercial oenological strain.

). The 90.9% of the yeasts of the H. uvarum species were positive, while 87.5% and 66.7% of the H. guilliermondii and H. vineae species, respectively, were positive. Furthermore, the control yeast S. cerevisiae IOC 18-2007^® was not an acetic acid producer, whereas D. bruxellensis strain V3-4-11, showed the highest acetic acid production. The clarification halo produced by H. guilliermondii strains was smaller than that produced by D. bruxellensis V3-4-11 (p < 0.05, ANOVA, LSD-Fisher test), but no differences were found with those produced by H. uvarum and H. vineae strains (p > 0.05, ANOVA, LSD-Fisher test). On the other hand, the majority (93.2%) of the yeast and yeast-like strains analysed showed higher levels of H₂S production in BIGGY agar than those of the S. cerevisiae IOC 2007^® reference strain and were therefore considered positive (Table 3). All species were able to produce hydrogen sulphide and almost all the species showed 100% H₂S-producing isolates (excepting A. pullulans with a 55.6% of incidence). Finally, only one strain belonging to the M. pulcherrima species (L1) showed an off-odour in grape must, characterised by solvent and acetate odours as well as cloudiness according to the sensory analysis panel (Table 3). Meanwhile, D. bruxellensis V3-4-11 showed the characteristic “Brett” defect, including horse sweat, rubber and a rotten smell as well as cloudiness.

Although the in vitro tests are preliminary, they resulted useful for screening spoilage features in a considerable number of isolates and for describing their general characteristics. Interestingly, the majority (78.6%) of the Hanseniaspora strains isolated (H. uvarum, H. guilliermondii and H. vineae) showed acetic acid production in YPD-CaCO₃. This result was in line with previous association of the apiculate yeasts (Hanseniaspora spp.) to the excessive production of acetic acid and ethyl acetate, which are considered detrimental to wine flavour [20]. On the other hand, in agreement to our work, Mestre et al. [35] have reported that non-Saccharomyces yeast species are capable of producing high levels of H₂S, while Comitini et al. [36] detected a large variability in H²S levels produced by non-Saccharomyces species. Interestingly, the present work showed no correlation between H₂S producers in BIGGY agar or acetic acid producers in YPD-CaCO₃ agar with flavour defects in pasteurised must. Only one strain (1/17) of M. pulcherrima was found to cause some defects in pasteurised must supplemented with p-coumaric acid. Finally, these in vitro tests enabled us to select two yeast strains with the potential to spoil wine: M. pulcherrima L1 (generation of odour defects) and H. uvarum L144 (maximum production of acetic acid), which underwent microvinification trials, since modulation of spoilage yeast metabolism in co-culture with S. cerevisiae has already been reported [75].

The physicochemical properties of the Malbec wines resulting from microvinification trials with two potential spoilage yeasts (M. pulcherrima L1 or H. uvarum L144) in co-culture with S. cerevisiae, and S. cerevisiae alone (control) are shown in

Table 4. Physicochemical properties of Malbec wines obtained in co-culture of potential spoilage yeasts and the commercial S. cerevisiae strain (Mp + Sc and Hu + Sc wines) and in monoculture of the S. cerevisiae strain (Sc wine, control wine).

Physicochemical Parameters	Wines
	Mp + Sc	Hu + Sc	Sc (Control)
Ethanol (%, v/v)	14.40 ± 0.14	14.10 ± 0.71	14.10 ± 0.62
Residual sugars (g/L)	1.05 ± 0.21	0.80 ± 0.57	1.23 ± 0.38
Total acidity (g/L tartaric acid)	4.55 ± 0.21	4.65 ± 0.07	4.90 ± 0.20
Volatile acidity (g/L acetic acid)	0.60 ± 0.00 a	0.69 ± 0.02 b	0.66 ± 0.01 ab
pH	3.51 ± 0.02	3.49 ± 0.02	3.48 ± 0.03

Data represent mean values of two experiments ± standard deviation. ^a,b Different subscript letters in the same row represent statistically significant differences, according to ANOVA analysis and LSD Fisher or Kruskal–Wallis tests (p < 0.05). Mp, M. pulcherrima L1; Hu, H. uvarum L144; Sc, S. cerevisiae IOC 18-2007.

. The residual sugars in all the wines were completely consumed, resulting in a dryness level (below 2 g/L) and standard ethanol concentrations (around 14% v/v), with no statistically significant differences among the wines. No significant differences were observed in the total acidity or pH of wines produced through co-culture or monoculture microvinifications. However, the volatile acidity of the Hu + Sc wine was significantly higher than that of the Mp + Sc wine, although it did not differ significantly from that of the control wine (Sc).

In addition, a descriptive sensory analysis was performed to evaluate the impact of potentially spoilage-associated non-Saccharomyces yeasts on the organoleptic properties of Malbec wines. Wines co-fermented with M. pulcherrima L1 or H. uvarum L144 in combination with S. cerevisiae (Mp + Sc and Hu + Sc, respectively) were compared with a control wine that was mono-fermented with S. cerevisiae (Sc). The sensory evaluation focused on key descriptors related to colour, aroma, taste, mouthfeel, and overall harmony/balance (Figure 3). In the visual phase, there was no statistically significant difference in the colour intensity and tonality of wines. From an aromatic standpoint, wines inoculated with either M. pulcherrima or H. uvarum exhibited lower aroma intensity, quality intensity, fruity and floral aromas than the control, although only the fruity aroma was statistically different. The Hu + Sc wine scored significantly lower for fruity aroma than the control wine (Sc) (ANOVA and LSD Fisher, p = 0.075). There were no statistical differences in terms of taste (acidity and bitterness) and mouthfeel attributes (body and astringency) between the wines. And although the co-fermented wines scored lower for taste intensity, taste quality and harmony-balance, than de mono-fermented wine, only the Hu + Sc wine displayed significantly lower taste intensity (Kruskal–Wallis, p < 0.05).

Similarly to our findings, some previous studies have reported that the presence of the same non-Saccharomyces species (H. uvarum and M. pulcherrima) in mixed cultures with S. cerevisiae does not affect ethanol production during wine fermentation compared to S. cerevisiae control wines [76,77]. However, Wang et al. [78] found that co-fermenting H. uvarum with S. cerevisiae significantly increased the ethanol content of the wine, except for one strain for which the ethanol level remained unchanged. These discrepancies may be due to the specific strains used. Importantly, the volatile acidity in Hu + Sc wine was significantly higher than in the other co-fermented wine, approaching the sensory threshold of 0.70 g/L acetic acid [77]. Beyond this threshold, the wine becomes unpleasant and the quality is negatively affected. These findings indicate that H. uvarum L144 can influence the volatile acidity of the final wine when used in mixed fermentation with S. cerevisiae, in consistency with the YPD CaCO₃ in vitro testing obtained with this strain. The ability of H. uvarum to produce high levels of acetic acid has led to it being considered a potential spoilage yeast in wine, as reported by Malfeito-Ferreira [20]. Some studies have shown that H. uvarum can specifically increase acetic acid levels in wines. Wang et al. [78] reported that H. uvarum strains significantly increased the volatile acidity of co-fermented Cabernet Sauvignon wines (0.63–0.73 g/L acetic acid) compared to 0,50 g/L in the control wine. Conversely, other studies have reported no significant differences in volatile acidity when co-fermenting with H. uvarum and S. cerevisiae [77,79]. Therefore, the production of volatile acidity appears to be strain-dependent, reflecting the variable production of acetic acid by H. uvarum strains when co-cultured with S. cerevisiae [80,81,82].

Finally, the sensory analysis results are consistent with the physicochemical analysis of wines displaying higher levels of volatile acidity in wine co-fermented with H. uvarum L144. Excessive levels of this parameter (in amounts greater than 0.70–0.80 g/L acetic acid) impart a vinegary smell and a tart taste to the wine, rendering it spoiled [83]. Such levels can also negatively impact the perception of fruity aromas and overall flavour intensity by masking desirable, fruit-derived flavours that contribute to the aroma and flavour profile of wine. Previous studies have reported the capacity of H. uvarum species to produce elevated levels of compounds such as acetic acid, acetaldehyde, and ethyl acetate, which are associated with wine spoilage [84]. Nevertheless, certain H. uvarum strains have also been successfully used in mixed-culture fermentations alongside Saccharomyces starter strains, producing beneficial effects on the sensory and chemical profile of the final wine [77,85]. On the other side, M. pulcherrima L1, which was selected as a spoilage yeast in in vitro assays due to its production of solvent and acetate odours in white grape fermenting must, did not modify statistically the colour, aroma, taste or mouthfeel attributes when co-fermented with S. cerevisiae Malbec grapes. Numerous studies have shown that the metabolic activity of yeast during must fermentation can be mutually influenced in the presence of other yeast species [75,86]. It has been observed that compounds typically produced in high concentrations by non-Saccharomyces yeasts in pure cultures, compounds often associated with negative impacts on wine quality, tend to remain below sensory perception thresholds when these yeasts are involved in mixed fermentations [75,77].

4. Conclusions

Knowledge of the microbial inhabitants of grape bunch rot, as well as the implications of their presence, is crucial for winemakers to identify potential risk. The present work showed a comprehensive analysis of rotten wine grapes from the Southern Oasis of Mendoza winegrowing region, including filamentous fungi and their related micotoxins, as well as the yeast population and its potential to spoil wine, which were evaluated during in vitro testing and microvinification trials.

In the present work, no geographic association was observed on the relative abundance of fungal genera or yeast species, or in TA contamination in rotten grapes from two subzones of the Southern Oasis of Mendoza winegrowing region that differ in edaphic conditions (3-1 and 3-3). Considering that the Southern Oasis of Mendoza winegrowing region has recently undergone subzoning based on climatic and edaphic parameters, this study is the first to evaluate such subzones. The lack of differences could suggest that the micro-conditions present in the subzones are not sufficiently different, or that soil conditions are not a significant factor influencing the studied variables. However, further research involving more subzones from the Southern Oasis of Mendoza winegrowing region and/or other wine-growing regions would be necessary to confirm these hypotheses.

On the other hand, the finding that Alternaria prevails as a major component of the mycobiota in rotten wine grapes from the Southern Oasis of Mendoza winegrowing region, together with the fact that the same environment favours TA production, highlights the importance of this study in the context of the known toxicological hazards of Alternaria toxins, and their imminent regulation.

Moreover, certain yeast strains associated with bunch rot showed potential to spoil wine. The strain H. uvarum L144, which showed the highest production of acetic acid in YPDCaCO₃, was found to influence the volatile acidity of the final wine when used in mixed fermentation with S. cerevisiae, and to significantly reduce fruity aroma and taste intensity. Conversely, M. pulcherrima L1, which produced solvent and acetate odours in fermenting white grape must, did not affect any sensory attributes when co-fermented with S. cerevisiae on Malbec grapes. Therefore, it is extremely important to evaluate the potential for spoiling wine at the strain level and, fundamentally, to validate the in vitro findings under conditions that more closely resemble the winemaking environment, such as microvinification trials.

Finally, the findings of the present study emphasised the importance of using grapes in good condition to minimise the risk of contamination by mycotoxigenic fungi and subsequent mycotoxin occurrence in wine, as well as reducing the risk of spoiling yeasts. However, since limitations in the GMPs implementation in high-volume production and/or weather incidents could also be present, efforts should be made to further study such microbiological and toxicological hazards and find environmentally friendly alternatives to control them.