Diversity and Dissemination of Brettanomyces bruxellensis During Winemaking and Barrel Aging

Abstract

Brettanomyces bruxellensis is recognized as the main spoilage yeast in red wines, producing volatile phenols that negatively impact wine quality. However, few studies have investigated strain diversity within wineries. Understanding the diversity and distribution of B. bruxellensis strains in different wines can provide insights into the origin and timing of contamination. This study aimed to evaluate the presence and diversity of B. bruxellensis biotypes during the production of four red wines in the same winery and to identify critical contamination stages. The analysis covered the entire process, from grape to six months of aging. B. bruxellensis yeasts were isolated and identified, and representative strains were typified by RAPD analysis. The results suggest that B. bruxellensis contamination did not originate from a single source. The grapes harbored low levels of B. bruxellensis, yet all wines were positive before barrel filling. This study demonstrates that winery equipment can serve as a vector for Brettanomyces introduction. Two critical contamination stages were identified: the shared use of equipment during fermentation, facilitating strain dissemination across wines, and the reuse of barrels, introducing new strains during aging. Additionally, some winery practices further promote B. bruxellensis spread and proliferation.

1. Introduction

Brettanomyces spp. have been described as the main spoilage yeast in wine. Within this genus, the species B. bruxellensis is a well-known spoilage microorganism in winemaking. However, B. bruxellensis is not exclusive to wine; it also occurs in various other alcoholic beverages where it is not considered a spoilage yeast, such as certain specialty beers and ciders [1,2,3]. B. bruxellensis negatively alters the chemical composition of wine by producing metabolic volatile compounds that are detrimental to the organoleptic properties of the final product [4]. This defect primarily occurs when hydroxycinnamic acid precursors (p-coumaric, ferulic, and caffeic acids) are enzymatically decarboxylated by cinnamate decarboxylase, leading to the formation of vinyl derivatives, which are subsequently reduced by vinyl phenol reductase to produce ethyl derivatives such as 4-ethylphenol (4-EP), 4-ethylguaiacol (4-EG), and 4-ethylcatechol (4-EC) [5,6]. Additionally, other minor metabolic by-products of B. bruxellensis, such as isovaleric and isobutyric acids, contribute to the perception of “Brett character”. The phenolic off-odors associated with Brettanomyces contamination have been described with sensorial descriptors such as “medicinal”, “phenolic”, “rancid”, “sweaty”, “smoky”, “Band-Aid^®”, “barnyard”, or “horse sweat” [2].

Both names of this yeast, Brettanomyces and Dekkera, are often used as synonyms in the scientific literature. Moreover, according to the revised International Code of Nomenclature for algae, fungi, and plants (the Melbourne Code), fungal species should be assigned only a single valid name. However, the designation Brettanomyces is more commonly applied in the food and biotechnology industries, where the species B. bruxellensis is immediately associated with wine by all stakeholders in the sector [7,8,9]. Therefore, in this study, the term Brettanomyces is preferentially used.

Unlike Saccharomyces cerevisiae, B. bruxellensis exhibits slow growth and is not a strong competitor in high sugar substrates. B. bruxellensis displays low nutrient requirements and is well adapted to harsh environments such as wine. Indeed, it can grow in nutrient-poor media, under low pH conditions (3–4), in the presence of high ethanol concentrations, and at elevated sulphite levels [10]. A notable example of this adaptation is its role as a spoilage yeast in wine, where it is typically detected during the aging process, though it may also appear at lower frequencies during earlier winemaking stages, including grapes and must [11]. However, B. bruxellensis could develop at the end of alcoholic fermentation, when other microbial populations decline, and it has been mainly associated with sluggish fermentation and delayed malolactic fermentation. While this yeast is not dominant in must during fermentation, its presence can significantly impact the final product [12].

The origin of Brettanomyces contamination remains unclear. Some authors suggest that grapes could be a source of contamination in wines [2,13,14]. In fact, winemakers in commercial wineries frequently associate specific vineyards with the recurrent presence of Brettanomyces in wines produced from their grapes. However, due to the challenges in isolating B. bruxellenis from grapes, this hypothesis remains difficult to confirm [15]. On the other hand, several studies propose that the winery itself is the primary source of contamination, with the reuse of oak barrels being the most frequently identified contamination point [5,16]. Additionally, other B. bruxellensis contaminated surfaces in winery have been described as cellar walls, drains, pumps, crushing equipment, and any tanks or transfer lines that are difficult to be effectively sanitized [3,7].

The contamination of wines by B. bruxellensis has increased in recent years due to evolving winemaking techniques that favor the production of wines with extended macerations that enhance precursor extraction, and wines that may be unsulphited, unfiltered, aged on the lees, or matured for longer periods in barrels [2]. All these factors are considered to promote B. bruxellensis growth [10]. Strategies to control growth of the spoilage yeast present a significant challenge for winemakers. Understanding the diversity of strains and their distribution in different wines can help reveal the origin and timing of contamination, allowing for the design of prevention and control strategies [4].

The aim of this study was to evaluate the occurrence and diversity of B. bruxellensis isolates during the production of four red wines, from grape to wine aging in a commercial winery, and to identify the critical points at which contamination occurs.

2. Materials and Methods

2.1. Vineyards and Winemaking Conditions

Four vineyards located in the Mendoza Province (Argentina) were selected for this study. The Merlot (Me) and Cabernet Sauvignon (Cs) vineyards were situated in Agrelo, Luján de Cuyo, within the viticultural region known as “Zona Alta del Rio Mendoza”, while the two Malbec (Mb2 and MbC) vineyards were in Villa Bastías, Tupungato, within the viticultural region known as “Valle de Uco”. Both viticultural regions belong to the climatic group IH + 2 IF + 1 IS + 1 [17], characterized by a hot climate with cold nights and moderate drought. The climate is dry, with annual precipitation around 200 mm [18]. The vineyards belonged to a commercial winery where the winemaking process took place. The winery was situated a considerable distance from the sampled vineyards, thus preventing any potential cross-anthropogenic contamination between the vineyards and the winery. The grape samples were taken from vineyards to produce middle to high-quality wines. The vines were trained in a low trellis system with proper disease control, resulting in grapes of high sanitary quality. The grapes were crushed, and fresh must was obtained. Initial sulfiting was carried out using potassium metabisulfite to achieve a final total SO₂ dose of 30 ppm in the fresh must. Fermentation occurred in 250 hL stainless steel tanks, inoculated with a commercial strain of S. cerevisiae (Lalvin EC1118, Lallemand Inc., Montreal, QC, Canada). Malolactic fermentation was carried out spontaneously. Alcoholic and malolactic fermentations were carried out at 26 ± 2 °C. The wine was then transferred to 200 L oak barrels for aging over a period of six months. During the aging period, the barrels were maintained in a room acclimated to 15 °C and 75–80% humidity. To maintain SO₂ levels during aging, the winery employs two correction practices: the addition of SO₂ to each barrel individually (SO₂ single addition), or a practice known as “rack and SO₂”, where the wine is removed from the barrels and placed in a holding tank, where SO₂ levels are corrected before being returned to the barrels. This practice combines wines from used barrels on the one hand and wines from new barrels on the other.

2.2. Grape and Wine Sampling

In each vineyard, ten vines were sampled along two diagonal transects, with three healthy bunches of grapes taken from each vine. Therefore, a total of 30 bunches from each vineyard were considered for sampling. To increase the likelihood of detecting Brettanomyces, additional samples were collected from damaged grapes (dehydrated, rotted, discolored, or spotted), as well as those affected by Planococcus ficus and bunches located near the vineyard edges, close to trees. These samples were analyzed separately. All samples were carefully placed in plastic bags and stored at 4 °C until laboratory analysis.

The musts from each of the four studied vineyards were fermented in a commercial winery and monitored for the presence of Brettanomyces from crushing until the end of barrel aging. Samples of 250 mL were provided by the winery in sterile containers for analysis throughout the winemaking process. During the aging period, five barrels were randomly selected, marked, and monitored over a period of six months. In addition, five samples from other Brettanomyces-positive wines aging in barrels from the same winery were included in the study. Samples provided by the winery were identified as follows: Beginning of Fermentation (BF), Mid-Fermentation (MF), End of Fermentation or 0° Baumé (0Be), Wine Before Barrel (BB), Barrel Aging-Month 1 (M1), Barrel Aging-Month 2 (M2), Barrel Aging-Month 3 (M3), Barrel Aging-Month 4 (M4), Barrel Aging-Month 5 (M5), Barrel Aging-Month 6 (M6) and Brettanomyces-positive wines aging in barrels: 16088, 16091, 7042, 7090, and 7129.

2.3. Detection and Isolation of Brettanomyces

The refrigerated bags containing the harvested grapes were opened in the laboratory under sterile conditions and aseptically crushed. A total of 45 mL of crushed grapes were placed into a 100 mL sterile flask, and 1 mL of a 50X Brettanomyces selective and differential enrichment solution (BSDES) was added. The 50X BSDES had the following composition: 100 g/L yeast extract (Oxoid Co., Ltd., Basingstoke, UK), 15 g/L KNO₃ (Sigma-Aldrich Co., Ltd., St. Louis, MO, USA), 5 g/L p-coumaric acid (Sigma-Aldrich Co., Ltd.), 25 g/L chloramphenicol (Sigma-Aldrich Co., Ltd.), and 10 g/L cycloheximide (Sigma-Aldrich Co., Ltd.). In addition 4 mL of ethanol (99%) was added to increase the selective condition of the medium. The prepared flasks were incubated at 28 °C with gentle agitation (150 rpm) for a period of 60 days. To monitor the evolution of the Brettanomyces population during enrichment, sampling was carried out every 10 days. After enrichment, viable and culturable Brettanomyces cells were determined by spreading onto Petri dishes containing the Brettanomyces selective differential medium (BSDM): 10 g/L glucose (Oxoid Co., Ltd.), 10 g/L yeast extract (Oxoid Co., Ltd.), 15 g/L agar, 0.4 g/L p-coumaric acid (Sigma-Aldrich Co., Ltd.), 0.5 g/L chloramphenicol (Sigma-Aldrich Co., Ltd.), and 0.05 g/L cycloheximide (Sigma-Aldrich Co., Ltd.). The pH was adjusted to 4.7 with HCl. After medium sterilization, 6% ethanol (99%) was added, the medium was homogenized, and the Petri plates were prepared. The selectivity of the enrichment solution and medium is provided by cycloheximide, ethanol, and chloramphenicol, while the differential characteristic of the medium is due to the addition of p-coumaric acid, a precursor in the production of ethyl phenols.

For the isolation of Brettanomyces during winemaking and aging, 250 mL samples were provided by the winery. Enrichment was carried out for samples from the beginning and middle of fermentation, by adding the BSDES described above. For the end of fermentation, wine before barrel and aging samples, 100 mL of wine was filtered through an acetate membrane (0.45 µm pore) and cultivated in BSDM. The plates were incubated at 28 °C for 20 days. Colonies were considered Brettanomyces-positive if they exhibited growth after five days, the production of a phenolic odor (ethyl phenol production), and typical cellular morphology under optical microscopy. In some cases, acid production in WLN medium (Oxoid Co., Ltd.) [3] was also determined. The five Brettanomyces spoilage wine samples from the winery were also analyzed by filtration. Counts were performed for all samples and representative colonies from each positive sample were isolated, purified, and preserved in 30% glycerol at −80 °C for subsequent molecular identification.

2.4. Brettanomyces Species Assignation by RFLP-PCR

DNA extraction was performed from an active pure culture of each isolated yeast. A loopful of the culture was inoculated into 5 mL of liquid YPD medium (40 g/L glucose, 5 g/L yeast extract, 5 g/L peptone) and incubated for 48 h at 28 °C with daily agitation. DNA extraction was conducted using a modified method from Hoffman and Winston [19], involving cell lysis through chemical and mechanical disruption, followed by purification with phenol/chloroform/isoamyl alcohol (25:24:1) and precipitation with absolute ethanol. DNA integrity was assessed by 0.7% agarose gel electrophoresis and DNA quantification was carried out by measurements in microvolumes (1 µL) of samples in NanoDrop spectrophotometer (Thermo Fisher Scientific Co., Ltd., Waltham, MA, USA)

Molecular identification at the species level was carried out by RFLP-PCR, as described by Cocolin et al. [20]. This method uses specific primers for the genus Dekkera (DB90F 5′-GAY ACT AGA GAG AGR RGG ARG GC-3′ and DB394R 5′-ACG AGG AAC GGG CCG CT-3′) (Macrogen Co., Ltd., Seoul, Republic of Korea) followed by digestion with the restriction enzyme DdeI (New England Biolabs Co., Ltd., Ipswich, MA, USA) to differentiate D. bruxellensis from D. anomala.

2.5. Molecular Typing of B. bruxellensis Isolates

B. bruxellensis isolates were analyzed at the strain level using RAPD with three different primers: M13 (5′-TTA TGA AAC GAC GGC CAG T-3′) [14], COC (5′-AGCAGCGTGG-3′) [21], and OPK03 (5′-CCAGCTTAGG-3′) (Macrogen Co., Ltd., Seoul, Republic of Korea) [14]. The amplification protocols were carried out according to the conditions described by Miot-Sertier and Lonvaud-Funel [21] and Crauwels et al. [22]. To evaluate the discriminatory power of the selected molecular markers, seven reference strains previously identified as B. bruxellensis were initially typed. These included six locally isolated and identified strains deposited in the INTA Mendoza Microorganism Collection (CoMIM) (F1, E3, B12, PM14, PM15, CH30) and one B. bruxellensis reference strain from the Spanish Type Culture Collection (CECT11045). These molecular markers were then used to genotype 85 representative isolates that had previously been identified as B. bruxellensis.

2.6. Cluster Analysis

The banding patterns obtained after gel electrophoresis were normalized using PyElph software version 1.4 [23] to construct a presence/absence matrix. The matrix combining the three markers was constructed and the Dice coefficient was calculated to estimate the similarity between isolates. A dendrogram was constructed using the Unweighted Pair Group Method with Arithmetic Mean (UPGMA) with PAST 3.21 and FIG TREE version 1.4.4 [24]. The cophenetic correlation coefficient was calculated as a measure of clustering reliability using the same software package.

3. Results

3.1. Occurrence of Brettanomyces in Grapes

The difficulty in isolating Brettanomyces from grapes is widely recognized. For this reason, an extensive grape sampling was conducted. Representative samples for vineyards were taken and different types of grapes were collected. The grape samples were classified as healthy grapes, damaged grapes, dehydrated grapes, discolored or spotted grapes, grapes collected from shaded areas near trees, and grapes damaged by Planococcus ficus. The results showed that only one grape sample exhibited the presence of Brettanomyces, despite conducting 24 enrichment cultures of the grapes over a 60-day period. Damaged grapes from the Malbec C vineyard (MbC) showed a count of 1.5 × 10⁴ CFU/mL after enrichment.

3.2. Occurrence of Brettanomyces During Winemaking

Musts from each of the four vineyards studied were fermented in a commercial winery using individual stainless-steel tanks. The presence of Brettanomyces was monitored from crushing through to the end of aging. To monitor the populations of Brettanomyces in each of the wines produced and to understand the impact of the winemaking practices employed, the following results are presented separately for each wine.

3.2.1. Merlot (Me)

For the fermentation conducted with grapes from the cv. Merlot (Me) vineyard, no positive samples for Brettanomyces were found during the early stages of alcoholic fermentation. However, the sample taken at the end of alcoholic fermentation showed positive results for Brettanomyces, with populations of 3.3 × 10² CFU/100 mL. These populations continued to increase, showing a population of 6.2 × 10² CFU/100 mL in the wine sample collected a few days prior to barrel filling (Figure 1A).

Then, the wine was placed in barrels without the addition of SO₂, as spontaneous malolactic fermentation had not yet concluded. Figure 1B shows the Brettanomyces counts for the five selected barrels that were monitored over six months of aging. It can be observed that the barrels exhibited high counts of Brettanomyces during the first three months of aging. In the first month, Brettanomyces populations were similar to those observed in the wine prior to barrel filling. However, by the second month of aging, the wine in the barrels showed an increase in Brettanomyces populations by approximately two logarithmic orders, reaching populations of around 10⁴ CFU/100 mL. After this elevated count, the winery applied Chitosan (No Brett Inside^®, Lallemand Inc.) at a dose of 4 g/hL, reducing Brettanomyces populations by 1.66 logarithmic units in the wine samples collected during the third month of aging. By the fourth month, populations were either undetectable or very low (Figure 1B). Notably, before sample collection at four months of aging, the winery added SO₂ (33 ppm). This may have enhanced the effect of chitosan, keeping culturable Brettanomyces populations at undetectable levels (<1 CFU/100 mL) until the end of the aging period.

3.2.2. Cabernet Sauvignon (Cs)

No positive counts for Brettanomyces were observed during alcoholic and malolactic fermentation in the Cabernet Sauvignon produced in the winery. Shortly after completing malolactic fermentation, but before the addition of SO₂ (30 ppm) prior to barrel filling, a positive result for Brettanomyces was observed, although the count was low (8 CFU/100 mL) (Figure 2A). During barrel aging, the wines showed negative or very low Brettanomyces counts (2 CFU/100 mL) until the fifth month of aging. At the end of the second month of aging, the winery adjusted the SO₂ levels (36 ppm) by directly adding the compound to each barrel. In the fifth month of aging, two barrels showed the presence of Brettanomyces, with populations increasing the following month, accompanied by the appearance of new contaminated barrels, surpassing values of 10² CFU/100 mL in two of them (Figure 2B). It is important to highlight that between the fifth and sixth months of aging, the winery performed the practice known as “rack and SO₂”, as was previously described in material and methods section. The results suggest that this practice may promote the spread of contamination with Brettanomyces.

3.2.3. Malbec 2 (Mb2)

During the alcoholic fermentation of the Malbec 2 (Mb2) must, the presence of Brettanomyces was not detected. However, after completing fermentation, the wine was kept in the tank until spontaneous malolactic fermentation (MLF) finished over a 19-day period. A wine sample taken at the end of MLF showed elevated counts of Brettanomyces, with populations of 3.3 × 10³ CFU/100 mL. Subsequently, the wine was racked and pressed, and SO₂ (30 ppm) was added. These actions reduced the Brettanomyces populations by 1.5 log units but still maintained viable populations of 1 × 10² CFU/100 mL. As a corrective measure, the winery added SO₂ again (25 ppm) before barrel filling (Figure 3A). No positive samples for Brettanomyces were observed during the first five months of barrel aging. However, from the sixth month, most barrels showed a positive count for this yeast, with populations exceeding 10² CFU/100 mL (Figure 3B). It should be noted that between the fifth and sixth months of aging, the winery performed in this wine the practice of “rack and SO₂”, previously mentioned. These results further highlight the impact of this practice on the spread of Brettanomyces contamination. The only barrel (B5) that did not show positive counts during the sixth month was a wine to which chitosan had been added before refilling the barrel (Figure 3B).

3.2.4. Malbec C (MbC)

The samples from the fermentation of Malbec C musts showed no presence of Brettanomyces. This was contrary to expectations, as it should be noted that the grapes from this vineyard were the only ones to yield a positive result for Brettanomyces isolation. On the other hand, the sample taken at the end of spontaneous malolactic fermentation (16 days after the completion of alcoholic fermentation) exhibited populations of around 5 × 10³ CFU/100 mL in the wine, which were significantly reduced after the sulfiting conducted during racking and pressing of wine (Figure 4A). Monitoring Brettanomyces populations during the aging of this wine showed negative counts for this yeast during the first five months of aging. Similar to the previously results obtained for Malbec 2 (Mb2), the wines showed positive Brettanomyces counts in the sixth month of aging, following the “rack and SO₂” practice performed by the winery (Figure 4B).

A general analysis of the four wines examined revealed that they all presented Brettanomyces populations prior to barrel filling. Wines showed positive counts mainly during the spontaneous malolactic fermentation. Before barrel aging, the winery applies sulfiting, which helps maintain low culturable Brettanomyces populations. However, the subsequent “rack and SO₂” practice promotes an increase and dissemination of Brettanomyces populations within the barrels.

3.3. Molecular Identification and Typing of Brettanomyces Isolates

To understand the diversity of Brettanomyces populations in the positive samples collected and consequently identify the origin and point of contamination during winemaking and aging, the isolates were first identified at the species level and then molecularly typed.

The restriction analysis of the amplicons obtained by the PCR assay with the specific primer pair DB90F/DB394R, using DdeI as the restriction enzyme, showed two distinct restriction fragments of approximately 154 bp and 129 bp [20], thus confirming that all the isolates belonged to B. bruxellensis ().

The molecular typing of B. bruxellensis isolates was performed using the M13, COC, and OPK03 RAPD markers. Firstly, to evaluate the discriminatory power of the selected markers, seven reference strains were amplified. The markers used produced molecular profiles ranging from 8 to 15 bands of different molecular weights (Figure 5A–C). The UPGMA dendrograms showed different degrees of relatedness among the strains, depending on the molecular marker used (Figure 5D–F). The molecular marker M13 and COC RAPD produced a similar clustering for most of the strains, with the Spanish reference strain being more distantly related. The OPK03 RAPD produced a completely different clustering of the strains, placing the Spanish reference strain within a cluster alongside local reference strains. Despite this discrepancy, OPK03 RAPD was included for the typing of the isolates in this study, as it demonstrated good discriminatory power and had been recommended for such analyses by Oro et al. [14].

Subsequently, these three molecular markers were used for the typing of 72 B. bruxellensis isolates from the four vineyards analyzed, including isolates from the grapes to six months of wine aging, as well as 13 isolates from the five contaminated wines obtained in the same winery. A combined binary matrix was constructed by incorporating the bands generated by all three markers, resulting in a presence/absence band matrix with 178 bands (54 for M13, 59 for COC, and 65 for OPK03). The level of discrimination among the isolates was high, with the first dendrogram branching occurring at a similarity coefficient lower than 0.05 (Figure 6). The overall typing using the combination of the molecular markers revealed 20 genetic clusters at a similarity coefficient of 0.30 (Figure 6).

In the analysis of the UPGMA dendrogram (Figure 6), there can initially be observed the presence of several clusters that group samples of the same grape varietal at the same stage of aging. The clearest example can be seen in four clusters that group isolates obtained during the sixth month of aging. Cluster 1 grouped five isolates from Cabernet Sauvignon wine (CsM6), cluster 6 grouped 11 isolates from Malbec C and Malbec 2 wines (MbCM6 and Mb2M6), and clusters 7 and 8 grouped six and two isolates from Malbec 2 wine, respectively (Mb2M6). A deeper analysis of these samples also revealed that the isolates within each cluster originated from the same wine aged in different barrels. Additionally, wines during barrels aging revealed the presence of multiple B. bruxellensis biotypes. For instance, Merlot isolates were included in six different clusters (cluster 3, cluster 4, cluster 10, cluster 15, cluster 16, and cluster 19); Cabernet Sauvignon isolates were found in five different clusters (cluster 1, cluster 3, cluster 5, cluster 14, and cluster 15); Malbec C isolates were grouped into three different clusters (cluster 6, cluster 11, and cluster 20); and Malbec 2 isolates were found in four different clusters (cluster 6, cluster 7, cluster 8, and cluster 17) (Figure 6). This finding confirms that B. bruxellensis populations in wines during aging are composed of multiple strains.

On the other hand, another cluster grouped isolates from wines during fermentation and the first month of aging. This grouping of isolates shows that the strains present during fermentation may persist in the wine once it is transferred to the barrel. For example, cluster 16 grouped two isolates from the end of fermentation (Me0Be) and the first month of aging (MeM1) for Merlot wine. Additionally, cluster 3 grouped eight isolates from different varietal wines during fermentation, including four isolates from Merlot at the end of fermentation (Me0Be), before barrel filling (MeBB), and during the first month of aging (MeM1), as well as two isolates from Malbec C and Malbec 2 at the end of fermentation (MbC0Be and Mb20Be). The presence of isolates from fermentation samples of different wines (Merlot, Malbec C, and Malbec 2) suggests some level of strain dissemination between the fermentation tanks, potentially mediated by shared equipment in the winery.

A deeper analysis of the Merlot wine isolates, which showed contamination with B. bruxellensis from fermentation until the third month of aging, reveals that, as previously mentioned, isolates from fermentation reappeared during the first month of aging (Figure 6, clusters 3 and 16). However, isolates from this wine during the second and third months of aging were grouped into different clusters. Cluster 4 grouped six isolates from Merlot in the third month of aging (MeM3), and cluster 10 grouped six isolates from the same wine from the second and third months of aging (MeM2 and MeM3). This finding suggests that the different B. bruxellensis strains detected during aging may have originated from reused barrels or from pre-existing populations that remained undetected due to their low initial numbers.

This same behavior can also be observed when analyzing the Cabernet Sauvignon isolates. Although this wine showed a low level of contamination, showing only a few barrels contaminated during aging, isolates from the wine before barrel filling (CsBB) and during aging (CsM1 and CsM5) clustered together in cluster 5. However, after the “rack and SO₂” practice performed by the winery before the sixth month of aging, B. bruxellensis isolates clustered into different groups, such as cluster 1 (5 isolates), cluster 3 (2 isolates), and cluster 15 (3 isolates) (Figure 6). This finding once again confirms that the strains present in contaminated wines could remain in the barrels during aging, and that the “rack and SO₂” practice could disperse the contamination and introduces new strains.

The 13 B. bruxellensis isolates obtained from the five contaminated wines during aging provided by the winery (16088; 16091; 7042; 7090; and 7129), grouped into two related clusters (cluster 11 and cluster 12) and were clustered separately from the isolates of the four studied wines (Me, Cs, MbC, and Mb2). Interestingly, the only isolate from damaged Malbec grapes (MbCD5) was included in cluster 12 (Figure 6).

4. Discussion

The present study aims to elucidate the origin of Brettanomyces contamination in a commercial winery with a history of phenolic off-odors in its wines. To achieve this, four vineyards from two important wine-producing regions in the province of Mendoza were selected, and their grapes were processed in this commercial winery. First, a thorough sampling of grapes from each vineyard was conducted, focusing on areas or grape characteristics that have been linked to a higher likelihood of Brettanomyces detection [15,25,26]. Periodic sampling was subsequently carried out throughout the winemaking process, encompassing the fermentation and aging stages. To assess the distribution of Brettanomyces in the winery, samples from other contaminated wines from the same winery were also included. In all samples, the isolation, identification, and molecular typing of the isolates were performed.

Detection of Brettanomyces in the present study was conducted using culture media; therefore, the results reflect culturable populations, which are consequently metabolically active. It is widely known that a Viable But Non-Culturable (VBNC) state in Brettanomyces can be induced by SO₂ addition [27,28]. However, previous studies have not clearly established whether the production of volatile phenols is linked to the physiological state, the growth phase of B. bruxellensis, or both, and findings have often been contradictory [29,30,31]. Some studies suggest that enzymatic activity in VBNC Brettanomyces cells could lead to sensory defects in wine. However, these studies have primarily been conducted on highly active populations before the induction of the VBNC state [26,29]. Our previous study clearly demonstrated the relationship between the capacity of Brettanomyces to grow in wine and reach critical populations, which leads to the appearance of sensory defects in the wines [32]. Additionally, Godoy et al. [33] demonstrated that both enzymes responsible for phenol production exhibit some instability in the presence of ethanol, with p-coumarate decarboxylase (CD) activity declining sharply. It is important to remark, in our study the Brettanomyces culturable counts in the wines studied did not exceed the critical population threshold (1000 cells/mL), which is associated with the appearance of perceptible sensory defects within a short period [34]. Consequently, even if residual enzymatic activity had been present in the VBNC cells in our study, the low Brettanomyces population detected before SO₂ addition was insufficient to produce volatile phenols above the perception threshold. As a result, the wines in this study did not develop organoleptic defects.

Our findings confirmed that B. bruxellensis was the only species belonging to the Brettanomyces genus associated with the wine and winery; this fact has been previously demonstrated [16,29]. Several studies have explored genetic diversity of B. bruxellensis using different fingerprinting techniques such as RAPD, Amplified Fragment Length Polymorphism (AFLP), microsatellite markers, mtDNA restriction analysis and others [16,21,35,36]. These studies highlight an important intraspecific diversity of B. bruxellensis, which makes the prediction of its occurrence and behavior in industrial fermentations difficult [11,37,38]. In our study, the construction of a binary matrix combining the three markers used for molecular typing resulted in a dataset with a high number of bands. The high molecular distance observed among the isolates in this study is consistent with previous research that has shown that the RAPD technique can generate variability between molecular profiles of strains, even when DNA concentrations and reagents are standardized [14,39,40,41]. For the interpretation of the results, the isolates grouped within the same cluster were considered genetically closely related biotypes, suggesting that these isolates may be variants of the same strain or a closely related group of strains [14,42].

Despite efforts to isolate Brettanomyces from grapes, only one sample was positive after a 60-day enrichment period. A single isolate was obtained from damaged grapes with no visible signs of rot. This finding limited our ability to link vineyard strains to the isolates detected during vinification and aging. The extremely low numbers of Brettanomyces in grapes, the random distribution of strains in vineyards, and the challenges of isolating them from sugar-rich substrates have been previously reported [13,14,26]. Similarly, Garijo et al. [34] carried out a study conducted over three consecutive harvests evaluating grapes and the initial steps of fermentation demonstrated that the presence of B. bruxellensis on grapes is generally low, supporting the conclusion that grapes themselves are not a critical point for yeast infection. On the other hand, Albertin et al. [43] compared B. bruxellensis strains detected from grape and wine by microsatellite analysis, showing that strains isolated from grape clustered together with strains coming from wine, suggesting a strong connection between grapes and the cellar.

During alcoholic fermentation, Brettanomyces populations were not detected in any of the wines studied, despite the samples being enriched for 60 days. After the completion of alcoholic fermentation, the winery kept the wines for 19 to 25 days awaiting the spontaneous malolactic fermentation (MLF). During this time, the wines were maintained in the tanks at room temperature and without the addition of SO₂, generating the conditions for Brettanomyces populations to increase. This was evidenced by the positive Brettanomyces counts obtained from all the wines analyzed in this stage. These cultivable populations decreased to undetectable levels after sulfitation and before barrel filling. It is interesting to highlight this waiting period for the MLF as a critical period for the winery, as it allows for the growth of Brettanomyces populations in a favorable medium without competition from S. cerevisiae [34]. Additionally, when considering the biotype clustering results, it was observed that cluster 3 included isolates from the end of fermentation and pre-barrel stages in three of the studied wines (Merlot, Malbec C, and Malbec 2). This grouping may suggest, as previously mentioned, that contamination could originate from shared winery equipment used during stages prior to aging, which disseminates the contamination. However, other Brettanomyces isolates from samples obtained at pre-aging stages were also included in clusters grouping isolates from a single varietal wine. Therefore, at this stage of fermentation, a mixed origin of the present strains can be considered, with some potentially originating from the grape must and others being introduced via contaminated winery equipment. The need to consider earlier stages of cellar operations as potential sources of contamination have been suggested in previous studies [14,15,34].

In the present study, after the completion of MLF, the wines were racked and pressed, with most undergoing sulphite addition before being transferred to barrels, except for the Merlot wine, previously described. This process resulted in either negative or very low counts of culturable cells in the wines during the initial months of aging. As previously mentioned, the addition of SO₂ controls the population of culturable cells and could be induced in the VBNC state. Most of the positive samples for Brettanomyces during aging corresponded to the sixth month. This is consistent with the winery’s practice of SO₂ correction (known as “Rack and SO₂”), which involves pooling wines from the same batch stored in different barrels into a holding tank, adjusting SO₂ levels, and then refilling the barrels. Additionally, following this practice, isolates from the same wine aging in different barrels grouped within the same clusters, suggesting the potential spread of contamination.

On the other hand, the isolates from wines testing positive during aging were grouped into different clusters, suggesting the presence of more than one biotype in the wines. This finding aligns with the existing literature that highlights the diversity of Brettanomyces populations during the barrel-aging process [3,7,37]. Barrel porosity, residual wine, and inconsistent cleaning protocols may create micro-niches that enable certain B. bruxellensis strains to persist and proliferate [44,45]. Additionally, some B. bruxellensis strains have been shown to form biofilms in wine, further contributing to their persistence [46]. A study addressing the geographic dispersal and persistence of Brettanomyces isolates in wineries revealed that specific genotypes could be repeatedly isolated within the same winery over decades, demonstrating an unexpected capacity for persistence [16]. The detection of multiple clusters in this study indicates that more than one strain has successfully established itself within the winery environment, further emphasizing the importance of monitoring and controlling Brettanomyces at every stage of the production process.

Additionally, to evaluate whether a generalized contamination existed within the winery, five wines testing positive for Brettanomyces during aging were included in the study. The isolates obtained from these wines were grouped into two related clusters, which were distinct from the isolates from the four wines originally studied. This finding suggests that the winery does not harbor a widely disseminated strain or group of strains that recurrently contaminates all wines. Instead, distinct groups of strains appear to be associated with specific wines aging in particular groups of barrels. Several studies have highlighted the role of winery equipment, barrels, and environmental reservoirs in the dissemination of Brettanomyces [7,14,34]. However, the lack of a single dominant strain across all wines suggests that contamination events may be localized or linked to specific processes, such as wine transfer, barrel usage, or differences in vinification practices.

5. Conclusions

Although the primary origin of Brettanomyces is the grape, having initially entered the winery through this route, wines contaminated with Brettanomyces may not necessarily have their source of contamination in the vineyard. Other sources of contamination are more closely related to winery practices and equipment. This study demonstrates that winery equipment can act as a vector for introducing Brettanomyces strains. Two critical stages were identified: shared use of equipment during fermentation stages, which facilitates the dissemination of Brettanomyces strains across different wines, and the reuse of barrels, which introduces new strains to the wines during aging.

Moreover, this study highlights the significant impact of winery practices on the growth and dissemination of Brettanomyces. Practices that encourage population growth, such as conducting spontaneous malolactic fermentation (MLF), create prolonged periods during which wines are kept at room temperature without SO₂ addition, providing favorable conditions for Brettanomyces. Additionally, practices that spread contamination, such as the rack and SO₂ procedure, where wines from various barrels are combined in a holding tank, facilitate the dissemination of strains from contaminated barrels to previously uncontaminated ones. Furthermore, wine movement during such practices introduces oxygen, which can stimulate the growth of Brettanomyces. While it is impossible to eliminate Brettanomyces entirely from the winery environment, its growth can be prevented, and practices that facilitate its development and dissemination should be minimized.