Wine Spoilage Control: Impact of Saccharomycin on Brettanomyces bruxellensis and Its Conjugated Effect with Sulfur Dioxide

The yeast Brettanomyces bruxellensis is one of the most dangerous wine contaminants due to the production of phenolic off-flavors such as 4-ethylphenol. This microbial hazard is regularly tackled by addition of sulfur dioxide (SO2). Nevertheless, B. bruxellensis is frequently found at low levels (ca 103 cells/mL) in finished wines. Besides, consumers health concerns regarding the use of sulfur dioxide encouraged the search for alternative biocontrol measures. Recently, we found that Saccharomyces cerevisiae secretes a natural biocide (saccharomycin) that inhibits the growth of different B. bruxellensis strains during alcoholic fermentation. Here we investigated the ability of S. cerevisiae CCMI 885 to prevent B. bruxellensis ISA 2211 growth and 4-ethylphenol production in synthetic and true grape must fermentations. Results showed that B. bruxellensis growth and 4-ethylphenol production was significantly inhibited in both media, although the effect was more pronounced in synthetic grape must. The natural biocide was added to a simulated wine inoculated with 5 × 102 cells/mL of B. bruxellensis, which led to loss of culturability and viability (100% dead cells at day-12). The conjugated effect of saccharomycin with SO2 was evaluated in simulated wines at 10, 12, 13 and 14% (v/v) ethanol. Results showed that B. bruxellensis proliferation in wines at 13 and 14% (v/v) ethanol was completely prevented by addition of 1.0 mg/mL of saccharomycin with 25 mg/L of SO2, thus allowing to significantly reduce the SO2 levels commonly used in wines (150–200 mg/L).


Introduction
The indigenous microbiota of grape musts includes an immense variety of microorganisms that can grow and ferment sugars [1,2]. Nowadays, most wine fermentations are conducted by selected yeast starters, which are mainly composed of Saccharomyces cerevisiae strains, due to their fast fermentation rates and ability to survive in the harsh environmental conditions of wine [3][4][5]. Although S. cerevisiae strains usually dominate alcoholic fermentations, some microorganisms such as lactic and acetic acid bacteria and yeasts like Dekkera/Brettanomyces bruxellensis may remain in finished wines and proliferate under certain conditions (e.g., oxygen and/or sugars availability), spoiling the wine [6,7].
Amongst wine contaminants, B. bruxellensis (anamorphic form) and its ascosporeforming type D. bruxellensis (teleomorphic form) is considered the most dangerous spoilage microorganism and has been isolated from almost every wine-producing area of the world [8][9][10]. In red wines, but also in some white wines, B. bruxellensis produces volatile phenols such as 4-ethylphenol and 4-ethylguaiacol, which have characteristic aromas vintage white grapes (Vitis vinifera L. cv. Alvarinho, Viosinho and Encruzado) collected from an experimental vineyard of Instituto Superior de Agronomia (Lisbon, Portugal). The grapes were frozen at −70 • C and stored until the beginning of the assay (approximately 6 months). Grapes were manually crushed, and the obtained grape juice was centrifuged at 10,000× g for 15 min (twice) and filtered sequentially through the following pore-size membrane filters (Millipore): 8.0 µm; 1.2 µm; 0.45 µm (twice). Finally, the cleared juice was filter-sterilized twice again using 0.22 µm membrane and the pH adjusted to 3.5 with ortho-phosphoric acid.
2.3. Synthetic-Grape Must (SGM) and True-Grape Must (TGM) Fermentations Performed with B. bruxellensis in Single-and in Mixed-Culture with S. cerevisiae SGM and TGM fermentations were performed in 500 mL flasks containing 300 mL of each medium (supplemented with 10 mg/L of p-coumaric acid) that were inoculated with 5 × 10 4 cells/mL of B. bruxellensis (strain ISA 2211) in single-culture fermentations and with 5 × 10 4 cells/mL of each yeast species (i.e., of S. cerevisiae and B. bruxellensis) in the mixed-culture fermentations. All fermentations (i.e., single-and mixed-culture SGM and TGM fermentations) were carried-out in duplicates and incubated at 25 • C, under slow agitation (80 rpm). Daily samples were taken from each flask to determine yeasts culturability, B. bruxellensis viability, as well as sugars consumption and ethanol and 4-ethylphenol production.

Analysis of Growth 2.4.1. Culturability
Culturability of S. cerevisiae and B. bruxellensis was determined by the classical plating method. Briefly, 100 µL of samples were plated onto YEPD-agar plates, after appropriate dilution (decimal serial dilution method). Plates were incubated at 30 • C (Vertical Incubator, Infors HT, Anjou, QC, Canada) and the number of Colonies Forming Units (CFU) enumerated after 2-6 days. In the mixed-culture fermentations, CFU counts of B. bruxellensis were obtained on 0.01% (w/v) cycloheximide YEPD-agar plates and the CFU counts of S. cerevisiae as the difference between total CFU counts (corresponding to S. cerevisiae plus B. bruxellensis) on YEPD-agar plates and the CFU counts of B. bruxellensis. The detection limit of the CFU method was 1 CFU/mL for results given in Sections 3.1 and 3.2.1, since 1000 µL were directly inoculated onto YEPD-agar plates for samples where no colonies were detected in 100 µL.

Viability
Viability (live/dead) of B. bruxellensis cells in single-culture fermentations was determined by directly applying the Live/Dead staining (LDS) procedure, as described below. For mixed-culture samples, PI-stained cells were then subjected to the Fluorescence In Situ Hybridization (FISH) method, the so-called LDS-FISH method, using the protocol described in [32]. The species-specific FISH-probe used to hybridize B. bruxellensis cells (26S-D.brux. 5.1) was designed by [33] and comprises the following oligonucleotide sequence: 5 -CTTACTCAAATCCCTCCGGT-3 . This FISH-probe was synthesized and labelled with the fluorochrome Fluorescein Isothiocyanate (FITC) at the 5 -end by demand of external services (STAB VIDA, Lisbon, Portugal).
LDS procedure: Briefly, 1 mL of culture medium was collected daily from single-and mixed-culture fermentations and cells were concentrated by centrifugation at 10,000× g for 10 min. The pellet was then washed with Bovine Serum Albumin (BSA)-saline solution (0.25% BSA w/v, 0.1% NaCl w/v) by gently pipetting up and down several times. Afterwards, the cell suspension was centrifuged again at 10,000× g for 10 min and resuspended in 100-1000 µL of BSA-saline solution, depending on the cellular density. Then, 10 µL of Propidium Iodide (PI, supplied by Life Technologies, Carlsbad, CA, USA) working solution (5 µg/mL) was mixed with 100 µL of cellular suspension (ca 10 6 cells/mL) and incubated for 10-20 min at room temperature without light. LDS-FISH method: After applying the LDS procedure above-described, the PI-stained cellular suspension was centrifuged for 5 min at 5000× g, the pellet was washed once with 1× phosphate-buffered saline solution (PBS) and then incubated with 4% (v/v) of paraformaldehyde for 4 h at 4 • C under agitation. Afterwards, fixed cells (approx. 10 6 cells) were centrifuged for 2 min at 10,000× g and hybridized in 45 µL of hybridization buffer (0.9 M sodium chloride, 0.01 % w/v sodium dodecyl sulfate, 20 mM Tris-HCl and 5 % v/v formamide) together with 5 µL of FITC labelled probe (50 ng/µL). Incubation was performed at 46 • C for 3 h. Subsequently, the cell suspension was centrifuged again (5 min at 10,000× g) and cells resuspended in 100 µL of washing solution (25 mM Tris/HCl and 0.5 M NaCl). This mixture was incubated for 30 min at 48 • C. Before enumeration, the previous suspension was again centrifuged, and cells resuspended in 100 µL of 1× PBS.
Quantification of live/dead cells: after applying the LDS or the LDS-FISH treatment, approximately 5 µL of each cell suspension was mixed with 5 µL of Vecta Shield (Vector Laboratories, Burlingame, CA, USA), spotted onto a Neubauer chamber and cells enumerated using an epifluorescence microscope (Olympus BX-60, Tokyo, Japan). Total cells were visualized in the bright field of the microscope and fluorescent cells in the U-MWB filter. Figure 1 shows LDS-FISH treated cells from a mixed-culture sample, visualized in the bright field ( Figure 1a) and in the U-MWB filter (Figure 1b), where green cells correspond to live B. bruxellensis cells (FISH-hybridized cells/non-PI-stained), orange/yellow cells correspond to dead B. bruxellensis cells (FISH-hybridized/PI-stained) and red cells correspond to dead S. cerevisiae cells (not FISH-hybridized/PI-stained). cells) were centrifuged for 2 min at 10,000× g and hybridized in 45 µL buffer (0.9 M sodium chloride, 0.01 % w/v sodium dodecyl sulfate, 20 m % v/v formamide) together with 5 µL of FITC labelled probe (50 ng/µL performed at 46 °C for 3 h. Subsequently, the cell suspension was centrif at 10,000× g) and cells resuspended in 100 µL of washing solution (25 m 0.5 M NaCl). This mixture was incubated for 30 min at 48 °C. Before previous suspension was again centrifuged, and cells resuspended in 10 Quantification of live/dead cells: after applying the LDS or the LD approximately 5 µL of each cell suspension was mixed with 5 µL of Ve Laboratories, Burlingame, CA, USA), spotted onto a Neubauer chamber ated using an epifluorescence microscope (Olympus BX-60, Tokyo, J were visualized in the bright field of the microscope and fluorescent ce filter. Figure 1 shows LDS-FISH treated cells from a mixed-culture sam the bright field ( Figure 1a) and in the U-MWB filter (Figure 1b

Quantification of Sugars and Ethanol by High Performance Liquid Chrom
Sugars (glucose and fructose) and ethanol were quantified by High uid Chromatography (HPLC), using an HPLC system (Merck Hitachi many) equipped with a refractive index detector (L-7490, Merck Hitachi many). Fermentation samples were filtered through 0.22 µm Millipore f lipore, Algés, Portugal) and then injected (20 µL) in a Sugar-Pack column Milford, CT, USA). Samples were eluted using as mobile phase CaEDT

Quantification of Sugars and Ethanol by High Performance Liquid Chromatography
Sugars (glucose and fructose) and ethanol were quantified by High Performance Liquid Chromatography (HPLC), using an HPLC system (Merck Hitachi, Darmstadt, Germany) equipped with a refractive index detector (L-7490, Merck Hitachi, Darmstadt, Germany). Fermentation samples were filtered through 0.22 µm Millipore filters (Merck Millipore, Algés, Portugal) and then injected (20 µL) in a Sugar-Pack column (Waters Hitachi, Milford, CT, USA). Samples were eluted using as mobile phase CaEDTA (50 mg/L) at 90 • C, with a flow-rate of 0.5 mL/min. All samples were analysed in duplicate. Glucose, fructose, and ethanol standards at concentrations of 15, 7.5 and 3.75 g/L were used to construct calibration curves.

Quantification of 4-Ethylphenol by Gas-Chromatography
The concentration of 4-ethylphenol (4-EP) produced by B. bruxellensis during singleand mixed-culture SGM and TGM fermentations was quantified by gas-chromatography (GC) using filtered (0.22 µm Millipore filters) samples that were first frozen at −18 • C in 15 mL Falcon tubes (Orange Scientific, Braine-L'Alleud, Belgium) and kept frozen until use. 4-EP was quantified using the protocol described in [9,34]. The volatile phenol (4-EP) was extracted using ether-hexan from a 5 mL sample with pH adjusted to 8.0 with NaOH. The volatile 4-EP was separated by collecting the organic phase of the mixture. The quantification was achieved by gas chromatography using 3,4-dimethylphenol as internal standard. A GC-FID (Varian CP-3800 series, Walnut Creek, CA, USA) with a capillary column Factor-Four (internal diameter 0.25 mm, length 15 m, film thickness 0.25µm) was used. The injector was run in split less mode, at 230 • C and the volume of injection was 2 µL. The detector temperature was set to 250 • C. Hydrogen was used as gas carrier at a flow rate of 0.1 mL/min. The oven was initially set at 50 • C, then the temperature was raised to 215 • C at a 10 • C/min rate and finally increased up to 250 • C at a rate of 20 • C/min. Calibration curves were constructed using 4-EP standards with concentration values ranging from 0-10 mg/L.

Production and Purification of Saccharomycin
The natural biocide (saccharomycin) secreted by S. cerevisiae (strain CCMI 885) was obtained from a SGM-fermentation performed at 25 • C without agitation for 7 days. The 7 day-old fermented broth was filtered through 0.22 µm Millipore membranes (Merck Millipore, Algés, Portugal) and the supernatant was first ultrafiltrated using 10 kDa centrifugal units (Vivaspin 15R, Sartorius, Germany) and then the permeate (<10 kDa) was concentrated (40-fold) in similar centrifugal units equipped with 2 kDa membranes. Finally, 100 µL of this concentrated peptidic fraction (2-10 kDa) was fractionated by size-exclusion chromatography using a Superdex-Peptide column (10/300 GL, GE Healthcare, Buckinghamshire, UK). The HPLC system was equipped with an UV-detector (Merck Hitachi, Darmstadt, Germany) and samples were eluted with ammonium acetate 0.1 M at a flow rate of 0.7 mL/min. The chromatographic pick with retention-time 26-27 min, previously found to contain saccharomycin [28,30] was collected, lyophilized, and stored frozen at −20 • C until required.

Effectiveness of the Natural Biocide to Prevent B. bruxellensis Growth in Wine
300 mL of TGM were fermented by S. cerevisiae at 25 • C without agitation for 20 days. Then, the fermented broth was filtered through 0.22 µm Millipore filters (Merck Millipore, Algés, Portugal) and the 2-10 kDa peptidic fraction of this cell-free supernatant was ultrafiltrated and concentrated (40-fold) as described in the previous sub-section. The 20-day-old fermented supernatant (pH 3.5), containing 118 g/L ethanol and no sugars, was supplemented with 8 g/L of fructose to simulate a wine with residual sugars that allow microbial growth, i.e., the "simulated wine". 2 mL of the above-mentioned peptidic fraction was added to this "simulated wine" that was then inoculated with 5 × 10 2 CFU/mL of B. bruxellensis (strain ISA 2211). A control-assay was performed in the same "simulated wine" but without addition of the 2-10 kDa peptidic fraction, which was used as Control. Culture-assays were incubated at 25 • C without agitation. Culturability of B. bruxellensis was followed by plate counts (CFU/mL), as described in Section 2.4.1, and viability by the LDS procedure described in Section 2.4.2.

Conjugated Effect of Saccharomycin with Sulfur Dioxide (SO 2 ) on B. bruxellensis Culturability
Simulated wines were prepared using the SGM medium (pH 3.5) mentioned in Section 2.2, modified in its sugars solution to contain just 4.5 g/L of fructose. Ethanol was added to this modified-SGM to obtain simulated wines with 10%, 12%, 13% and 14% (v/v), respectively, with final pH values of 3.5. Each simulated wine was artificially contam-inated with 5 × 10 3 cells/mL of B. bruxellensis in a final volume of 300 µL. First, the inhibitory effects of ethanol and SO 2 were analyzed in separate, i.e., simulated wines without SO 2 but with 10%, 12%, 13% and 14% (v/v) ethanol, respectively, were used to evaluate the ethanol effect on B. bruxellensis growth; simulated wines without ethanol but with 25, 50, 100 and 150 mg/L of potassium metabisulfite (PMB) (Sigma-Aldrich, Missouri, EUA) (concentrations equivalent to 0.16, 0.33, 0.66 and 1 mg/L of molecular SO 2 , at pH 3.5) were used to assess the SO 2 effect on B. bruxellensis growth. Then, the synergistic effect of SO 2 with ethanol was tested using simulated wines at all ethanol levels (i.e., at 10%, 12%, 13% and 14% (v/v) ethanol), each of them supplemented with 25, 50, 100 and 150 mg/L of PMB (Sigma-Aldrich, Missouri, EUA). Finally, the conjugated effect of saccharomycin (obtained as described in Section 2.7) with SO 2 was evaluated on B. bruxellensis growth using the simulated wines at all ethanol levels (i.e., at 10%, 12%, 13% and 14% (v/v) ethanol), each of them supplemented with 0.25, 0.5 and 1 mg/mL of saccharomycin together with PMB at 25 and 50 mg/L, respectively. All growth-assays were performed in triplicates in 96 well-microplates and incubated in a Multiskan-GO spectrophotometer (Thermo-Fisher Scientific Inc., Waltham, MA, USA) at 30 • C, under strong agitation. Cell growth was followed by optical density measurements (at 590 nm) in a Microplate Reader (Dinex Technologies Inc., Chantilly, VA, USA) and by CFU counts. For CFU counts, 10 µL of samples were taken and after appropriate dilution (decimal serial dilution method) 100 µL were plated onto YEPD-agar plates, as described in the Section 2.4.1. Whenever colonies could not be detected in agar-plates inoculated with diluted samples, 100 µL of sample were directly plated onto YEPD-agar plates. Thus, the detection limit of the CFU method for results presented in Section 3.2.2 was 10 CFU/mL.

Statistical Analyses
The minimum significant difference between results presented in Table 1 and in figures was calculated to allow comparison of mean values, as described by Fry et al. [35]. To check the assumption of equal variances the Levene's test was used and then, one way ANOVA (if the variances were equal) or Welch tests (if the variances were unequal) were applied to determine the significance of the difference between means. The statistical analysis was performed in Microsoft Excel. Table 1. Independent effect of ethanol and sulfur dioxide on the culturability (CFU/mL) of B. bruxellensis (strain ISA 2211) inoculated in simulated wines (modified-SGM) with 10, 12, 13 and 14% (v/v) of ethanol, pH 3.5, and in the same modified-SGM without ethanol but with 25, 50, 100 and 150 mg/L of potassium metabisulfite (PMB) that correspond to concentrations of molecular SO 2 of 0.16, 0.33, 0.66 and 1 mg/L, respectively. Values presented correspond to means ( ± SD) of duplicate measurements of three independent biological experiments. Different letters located before the CFU/mL indicate significantly different values (p < 0.05).

Synthetic-Grape Must (SGM) and True-Grape Must (TGM) Fermentations Performed with B. bruxellensis in Single-and in Mixed-Cultures with S. cerevisiae
Metabolic and yeasts growth profiles during SGM fermentations performed with B. bruxellensis in single-culture and in mixed-culture with S. cerevisiae are represented in Figure 2. During mixed-culture fermentations (Figure 2a,b) S. cerevisiae increased its cell density from an initial cell density of 5 × 10 4 CFU/mL up to 4 × 10 7 CFU/mL after 3 days, remaining at about 10 7 CFU/mL until the end of fermentation (day-10), while B. bruxellensis grew during the first 3 days (from 5 × 10 4 CFU/mL up to 4 × 10 6 CFU/mL) but then began to die-off, decreasing its culturability in the next 5 days (to 4 CFU/mL at day-8) (Figure 2a). The loss of culturability of B. bruxellensis during the mixed-culture fermentation was accompanied by an increase of the number of dead cells (PI-stained cells) (Figure 2a) that represented 99% of the population at day-8. Since the number of culturable cells is extremely low at day-8 (4 CFU/mL) and 99% of the total cell population was dead, the percentage of viable but non-culturable (VBNC) cells should be less than 1%. Conversely, during the single-culture fermentation (Figure 2c,d) B. bruxellensis culturability increased from 5 × 10 4 CFU/mL at day-0 up to 4 × 10 8 CFU/mL at day-7, remaining at about 10 8 CFU/mL until the end of fermentation (10 days) (Figure 2c). During the singleculture fermentation (Figure 2c,d) B. bruxellensis cell viability (live/dead cells) correlated with its culturability, since the number of viable cells (non-PI-stained cells) remained high throughout fermentation (ranging from 92-98% during the first 8 days) and decreased to only 65% at the end of fermentation (day-10) (Figure 2c), when sugars were already completely consumed (Figure 2d). Metabolic profiles (i.e., sugars consumption, and ethanol and 4-ethylphenol production) during the mixed-culture fermentation (Figure 2b) show that sugars (glucose and fructose) were almost completely consumed within the first 5 days (4.7 g/L of residual fructose), when ethanol attained its highest level (92 g/L), and 4-ethylphenol was produced in very low amounts, attaining a maximal concentration of 0.25 mg/L at day-3. The negligible levels of 4-ethylphenol produced during the mixedculture fermentation (Figure 2b) correlate with the loss of B. bruxellensis viability (Figure 2a). On the contrary, during B. bruxellensis single-culture fermentation (Figure 2c,d) sugars were consumed at a much slower rate (the same amount of sugars was consumed only after 10 days) and ethanol attained its highest concentration (93 g/L) after 10 days (Figure 2d), showing that B. bruxellensis metabolism is much slower than that of S. cerevisiae. Regarding 4-ethylphenol, results show that this phenolic compound was produced at significantly higher levels in the single-culture fermentation (Figure 2d) attaining 6.44 mg/L at day-7, what can be explained by the high culturability of B. bruxellensis during this fermentation (Figure 2c). Comparing the culturability/viability profiles of B. bruxellensis in single-culture fermentation (Figure 2c) with that in mixed-culture fermentation (Figure 2a), it becomes clear that S. cerevisiae exerted a strong antagonistic effect against B. bruxellensis growth and 4-ethylphenol production.
To check if the antagonistic effect exerted by S. cerevisiae against B. bruxellensis would also be effective in TGM, mixed-and single-culture fermentations were performed at the same growth conditions. Yeasts growth and metabolic profiles during the mixed-and singleculture TGM-fermentations are shown in Figure 3. Results show that S. cerevisiae exerted an antagonistic effect against B. bruxellensis also in the TGM-fermentation (Figure 3a,b), although the effect was less pronounced than in the SGM-fermentation. In fact, while B. bruxellensis completely lost its culturability and viability within 8 days (<10 CFU/mL and >99% dead-cells) in the mixed-culture SGM-fermentation (Figure 2a), in the TGMfermentation B. bruxellensis was able to grow in the first 2 days (up to 4.7 × 10 5 CFU/mL) but then its culturability decreased to 1.7 × 10 4 CFU/mL at day-13, as well as its viability (from 92% at day-0 to 77% at day-13) (Figure 3a). In the single-culture TGM-fermentation (Figure 3c,d), B. bruxellensis was able to grow in the first 6 days, increasing its cell density from 5 × 10 4 CFU/mL at day-0 to 3 × 10 8 CFU/mL at day-6 and keeping this value (10 8 CFU/mL) for 17 days, while dead cells remained at low numbers (ranging 6-15% of PI-stained cells) (Figure 3c). Once again, one can conclude that S. cerevisiae inhibited B. bruxellensis metabolism since a much lower level of 4-ethylphenol (1.3 mg/L) was produced during the mixed-culture TGM-fermentation (Figure 3b) by comparison with 2.82 mg/L of 4-ethylphenol produced during the single-culture fermentation (Figure 3d). To check if the antagonistic effect exerted by S. cerevisiae against B. bruxellensis would also be effective in TGM, mixed-and single-culture fermentations were performed at the same growth conditions. Yeasts growth and metabolic profiles during the mixed-and single-culture TGM-fermentations are shown in Figure 3. Results show that S. cerevisiae exerted an antagonistic effect against B. bruxellensis also in the TGM-fermentation ( Figure  3a,b), although the effect was less pronounced than in the SGM-fermentation. In fact, while B. bruxellensis completely lost its culturability and viability within 8 days (<10 CFU/mL and >99% dead-cells) in the mixed-culture SGM-fermentation (Figure 2a), in the TGM-fermentation B. bruxellensis was able to grow in the first 2 days (up to 4.7 × 10 5 CFU/mL) but then its culturability decreased to 1.7 × 10 4 CFU/mL at day-13, as well as its viability (from 92% at day-0 to 77% at day-13) (Figure 3a). In the single-culture TGM-fermentation (Figure 3c,d), B. bruxellensis was able to grow in the first 6 days, increasing its cell density from 5 × 10 4 CFU/mL at day-0 to 3 × 10 8 CFU/mL at day-6 and keeping this value (10 8 CFU/mL) for 17 days, while dead cells remained at low numbers (ranging 6-15% of PI-stained cells) (Figure 3c

Effect of Saccharomycin on B. bruxellensis Culturability and Viability
To evaluate the effectiveness of the natural biocide (saccharomycin) to prevent B. bruxellensis proliferation in wine, a simulated wine (118 g/L of ethanol and 8 g/L of residual fructose, pH 3.5) supplemented with 1 mg/mL of the peptidic fraction containing the natural biocide was artificially contaminated with 5 × 10 2 cells/mL of B. bruxellensis. Culturability (CFU/mL) and viability (PI-staining) profiles of B. bruxellensis in the biocideassay and in the control-assay (without biocide) are shown in Figure 4. Results show that while in the control-assay B. bruxellensis was able to grow after the second day of inoculation, attaining 3.3 × 10 7 CFU/mL at day-7, in the biocide-assay B. bruxellensis culturability continuously decreased upon inoculation attaining a cell density of 10 CFU/mL at day-9. The loss of B. bruxellensis culturability in the biocide-assay was accompanied by an increase of the percentage of dead cells that reached 85% at day-9 and 100% at day-12, while in the control-assay, viability of B. bruxellensis remained high even after 12 days (15% of cells dead) (Figure 4).      . Effect of saccharomycin on the culturability (CFU/mL) and viability (PI-stained cells) of B. bruxellensis during the biocide assay (simulated wine with 1 mg/mL of saccharomycin), and in the respective control-assay (simulated wine without saccharomycin). The detection limit of the CFU method was 1 CFU/mL. Values presented correspond to means (± SD) of duplicate measurements of two independent biological experiments.

Conjugated Effect of Saccharomycin with Sulfur Dioxide (SO 2 )
The single effect of ethanol and potassium metabisulfite (PMB) on B. bruxellensis growth was evaluated in simulated wines (pH 3.5), artificially contaminated with 5 × 10 3 cells/mL of B. bruxellensis. Results (Table 1) showed that B. bruxellensis was able to grow in the presence of 10%, 12%, 13% and 14% (v/v) of ethanol, reaching 3 × 10 8 CFU/mL after 72 h. Likewise, SO 2 at 25, 50, 100 and 150 mg/L of PMB was not able to inhibit growth of B. bruxellensis in simulated wines without ethanol, with cultures reaching similar cell density levels (i.e., ca 10 8 CFU/mL) after 72 h ( Table 1). The combined effect of ethanol (10%, 12%, 13% and 14% (v/v)) with PMB (25, 50, 100 and 150 mg/L of PMB) was also assessed. Results ( Figure 5) revealed that in simulated wines at 10% and 12% (v/v) ethanol, B. bruxellensis growth was completely inhibited by 100 and 150 mg/L of PMB (i.e., 0.66 and 1.0 mg/mL of molecular SO 2 ), respectively (Figure 5a,b), whereas in simulated wines at 13% and 14% (v/v) ethanol, B. bruxellensis was only able to proliferate in the presence of 25 mg/L of PMB (0.16 mg/mL of molecular SO 2 ) (Figure 5c,d). Our results are in accordance with the probabilistic model developed by Sturm et al. [36] for B. bruxellensis growth as a function of pH, ethanol and free SO 2 , which predicts that B. bruxellensis is not able to grow in a simulated wine with 50 mg/L of free SO 2 (ca 150 mg/mL of PMB) when conjugated with ethanol levels between 10% and 15% (v/v) and pH values between 3.3 to 4.1. The inhibitory effect of saccharomycin was tested at concentrations of 0.25, 0.5 and 1.0 mg/mL in simulated wines at 10% and 12% (v/v) of ethanol together with 25 mg/L of PMB (Figure 6a,c) and 50 mg/L of PMB (Figure 6b,d). Results showed that in both wines inhibition of B. bruxellensis growth was only achieved with addition of 1.0 mg/mL of saccharomycin together with SO2 at both 25 and 50 mg/mL PMB (Figure 6a-c). However, even addition of 1.0 mg/mL saccharomycin was not sufficient to induce total loss of B. bruxellensis culturability with cultures remaining at ca 10 3 -10 4 CFU/mL after 72 h (Figure The inhibitory effect of saccharomycin was tested at concentrations of 0.25, 0.5 and 1.0 mg/mL in simulated wines at 10% and 12% (v/v) of ethanol together with 25 mg/L of PMB (Figure 6a,c) and 50 mg/L of PMB (Figure 6b,d). Results showed that in both wines inhibition of B. bruxellensis growth was only achieved with addition of 1.0 mg/mL of saccharomycin together with SO 2 at both 25 and 50 mg/mL PMB (Figure 6a-c). However, even addition of 1.0 mg/mL saccharomycin was not sufficient to induce total loss of B. bruxellensis culturability with cultures remaining at ca 10 3 -10 4 CFU/mL after 72 h (Figure 6a-c).In simulated wines at 13% and 14% (v/v) ethanol, the inhibitory effect of saccharomycin (at 0.25, 0.5 and 1.0 mg/mL) together with 25 mg/mL of PMB (Figure 7a,b) revealed that 0.5 mg/mL of saccharomycin prevented B. bruxellensis growth above 5 × 10 3 CFU/mL in the first 24 h in the simulated wine at 14% (v/v) ethanol, while addition of 1.0 mg/mL saccharomycin induced loss of B. bruxellensis culturability (to less than 10 CFU/mL) in both simulated wines (i.e., wines at 13% and 14% (v/v) ethanol) (Figure 7a,b). This demonstrate that 1.0 mg/mL of saccharomycin together with 25 mg/L of PMB (0.16 mg/mL of molecular SO 2 ) is sufficient to reduce B. bruxellensis culturability below 10 CFU/mL within 48 h in wines at 13% and 14% (v/v) ethanol ( Figure 7).

Discussion
In previous work we found that S. cerevisiae secretes a natural biocide (saccharomycin) during alcoholic fermentation that mediates the early death of Hanseniaspora guilliermondii in mixed-culture alcoholic fermentations [28] and inhibits the growth of wine-related non-Saccharomyces yeasts, including B. bruxellensis [28,30]. The effect of saccharomycin was evaluated against the growth of six B. bruxellensis strains (i.e., ISA 1649, ISA 1700, ISA 1791, ISA 2104, ISA 2116 and ISA 2211) in YEPD medium (at pH 3.5) demonstrating to it inhibits all those strains, although the minimal inhibitory concentration varied amongst strains, from 1-2 mg/mL [28]. Besides, S. cerevisiae CCMI 885 also demonstrated to exert an antagonistic effect against all the six B. bruxellensis strains during synthetic grape must (SGM) mixed-culture fermentations [29]. Those results [28][29][30]37] strongly suggested that saccharomycin is, at least in part, responsible for the antagonism exerted by S. cerevisiae against B. bruxellensis during mixed-culture alcoholic fermentations. In fact, results obtained in the present work (Figures 2 and 3) support that assumption, since B. bruxellensis rapidly lost its culturability (i.e., from 4.1 × 10 6 CFU/mL at day-3 to 4 CFU/mL at day-8 in SGM and from 4.7 × 10 5 CFU/mL at day-3 to 1.7 × 10 4 CFU/mL at day-13 in TGM) during the mixed-culture fermentations (Figures 2a and 3a) but kept its culturability at high levels (ca 10 8 CFU/mL) during the single-culture fermentations (Figures 2c and  3c), namely after total sugars exhaustion ( Figure 3c). Thus, neither nutrients depletion nor

Discussion
In previous work we found that S. cerevisiae secretes a natural biocide (saccharomycin) during alcoholic fermentation that mediates the early death of Hanseniaspora guilliermondii in mixed-culture alcoholic fermentations [28] and inhibits the growth of wine-related non-Saccharomyces yeasts, including B. bruxellensis [28,30]. The effect of saccharomycin was evaluated against the growth of six B. bruxellensis strains (i.e., ISA 1649, ISA 1700, ISA 1791, ISA 2104, ISA 2116 and ISA 2211) in YEPD medium (at pH 3.5) demonstrating to it inhibits all those strains, although the minimal inhibitory concentration varied amongst strains, from 1-2 mg/mL [28]. Besides, S. cerevisiae CCMI 885 also demonstrated to exert an antagonistic effect against all the six B. bruxellensis strains during synthetic grape must (SGM) mixed-culture fermentations [29]. Those results [28][29][30]37] strongly suggested that saccharomycin is, at least in part, responsible for the antagonism exerted by S. cerevisiae against B. bruxellensis during mixed-culture alcoholic fermentations. In fact, results obtained in the present work (Figures 2 and 3) support that assumption, since B. bruxellensis rapidly lost its culturability (i.e., from 4.1 × 10 6 CFU/mL at day-3 to 4 CFU/mL at day-8 in SGM and from 4.7 × 10 5 CFU/mL at day-3 to 1.7 × 10 4 CFU/mL at day-13 in TGM) during the mixed-culture fermentations (Figures 2a and 3a) but kept its culturability at high levels (ca 10 8 CFU/mL) during the single-culture fermentations (Figures 2c and 3c), namely after total sugars exhaustion (Figure 3c). Thus, neither nutrients depletion nor oxygen requirements can explain the early death of B. bruxellensis during mixed-culture fermentations.
to reduce the SO 2 levels to 25 mg/L PMB (i.e., ca 0.16 mg/L molecular SO 2 ), induing the loss of B. bruxellensis culturability to less than 10 CFU/mL (Figure 7).
Thus, our work shows that saccharomycin is a promising wine biopreservative that allows reducing the levels of SO 2 usually used in winemaking. However, the present results should be considered as preliminary results since they were obtained at micro-scale growth conditions and not under true wine production conditions. Besides, the impact of other parameters, such as the initial level of B bruxellensis contamination, wine pH and cells adaptation to ethanol, on the inhibitory efficiency of saccharomycin should also be further assessed.  Data Availability Statement: All data reported in this study is included in the manuscript.