INFLUENCE OF MEDIUM CHAIN FATTY ACIDS ON SOME BOTRYTISED WINE-RELATED YEAST SPECIES AND ON SPONTANEOUS REFERMENTATION OF TOKAJ ESSENCE

Medium chain fatty acids are candidates of partial sulphur dioxide replacement in wine, as a solution to the growing consumer concerns about chemical additives. In botrytised sweet wine specialties, large amount of sulphur dioxide addition is one of the eﬀ ective practices to stop alcoholic fermentation. Increasing medium chain fatty acid levels up to 80 mg l –1 was tested as a sole inhibitor on solid agar surface. S. bacillaris seemed to be the most sensitive, S. cerevsisiae and S. bayanus were more tolerant, while Z. bailii showed the highest tolerance. Then, increasing medium chain fatty acid levels up to 40 mg l –1 combined with 100 mg l –1 sulphur dioxide was introduced into a Tokaj Essence under refermentation. After 56 days, the highest dosage had pronounced eﬀ ect on the yeast population, but the refermentation was not inhibited completely. Medium chain fatty acids have varying inhibitory eﬀ ect on botrytised wine-related yeasts, moreover, it could be used eﬀ ectively in media with high ethanol content, unlike Tokaj Essence.

Nowadays, consumers are particularly concerned about health aspects, connected with sulphite toxicity in wine, therefore, the current general tendency is to reduce the use of sulphite in winemaking. Sulphur dioxide is the most frequently used chemical additive in winemaking, employed for multiple benefi ts as antiseptic, antioxidant, colour-, fragrance-, and taste protector (S et al., 2012). Based on current knowledge, none of the studied alternatives can totally replace SO 2 , which remains a useful, sometimes indispensable agent.
In Tokaj botrytised wine specialties, it is a widespread practice to stop alcoholic fermentation and save the residual sugars with a considerable SO 2 addition in combination with cooling, racking, and microfi ltration (M , 2011). In some cases, it is a challenge to meet the upper limit of total SO 2 concentration set by EU Commission legislation 607/2009/ EC (EC, 2009). Consequently, any eff ective SO 2 replacement could facilitate botrytised winemaking.
Recent candidates for partial substitution of SO 2 in wine are medium chain fatty acids (MCFA). MCFA and their esters are common yeast secondary metabolites usually produced in small quantities (B et al., 2018). Earlier studies revealed that artifi cially added MCFA could be used to stop an alcoholic fermentation carried out by S. cerevisiae (B et al., 2012), inhibit refermentation (B , 2014), and consequently decrease the necessary SO 2 addition. MCFA application in wines is studied only by a few (e.g. B et al., 2012; B et al., 2017) in normal winemaking environments. However, the inhibitory eff ect might be signifi cantly diff erent for various yeast species and in special winemaking environments, like botrytised winemaking.
In case of botrytised wine fermentation, the original yeast biota of the grape berry is altered considerably (reviewed by e.g. S , 2019), the harsh fermentation conditions are tolerable only for the well-adapted species. Beside others, Saccharomyces cerevisiae has great importance, Saccharomyces bayanus is also well presented (M -P et al., 2010). Starmerella bacillaris (syn. Candida zemplinina) was originally described in Tokaj wine region and connected to botrytised, sweet wine fermentation (S , 2003). Due to the signifi cant amount of remaining sugars, refermentation of these wine specialties by the tolerant spoilage yeast Zygosaccharomyces bailii is a major threat (A et al., 2015). In this study, fi rst we focused on the characterisation of the general tolerance of S. cerevisiae, S. bayanus, S. bacillaris, and Z. bailii against MCFA as a sole additive in the medium. Furthermore, various MCFA concentrations were tested in combination with SO 2 to inhibit spontaneous refermentation in a Tokaj Essence.

Tolerance test
Yeast strains: Yeast strains are shown in Table 1. Natural isolates were previously identifi ed by their rDNA and ITS regions (based on the methods of Z and co-workers, 2010) except Z. bailii strains, which were formerly identifi ed by classical methods, upon characteristic sporulation and physiological traits. Inoculum was prepared in YEPD broth (20 g l -1 glucose, 10 g l -1 peptone, and 10 g l -1 yeast extract), incubated (25 °C/48 h) without agitation.
Fermentation conditions: Fermentation (preculturing) was carried out under semianaerobic condition at 20 °C, in test tubes containing 5 ml aliquots of model media, without shaking. Tubes were inoculated to a level of 1×10 6 cell ml -1 with 48-hour-old yeast cultures grown in YEPD broth. Cell concentration was measured by Bürker chamber cell counting after 72 h of fermentation.
Drop test: After 72-hour fermentation without preservative, the inhibitory eff ect of MCFA was tested on solid agar surface. The test was carried out with 5 μl of serial dilutions from the cultures (10 -1 , 10 -3 , 10 -5 ), in triplicate, according to P -T and co-workers' (2016) modifi ed method. Into the agar 0, 10, 20, 40, and 80 mg l -1 MCFA mixture was introduced. Upon the results of B (2014), the MCFA mixture contained C 8 :C 10 :C 12 in 2:7:1 ratio, solved in 70 v/v% ethanol. After 7 days of incubation at 20 °C, drop test images were recorded in a fi x vision system with a Sony Exmor RSIMX315 camera (Sony Corp., Minato, Japan). Growth area analysis with ImageJ software (S et al., 2012) was used to assess the capability of the strains to grow under various MCFA conditions. Growth values given as percentage are raw colony area-means of triplicate drop-tests, normalised with the control growth of each strain. Data were evaluated with ANOVA after checking the assumptions, using IBM SPSS 23.0. Armonk, NY, USA.

Inhibition of refermentation in Tokaj Essence by MCFA
Culture media: A Tokaj Essence from Vintage 2005 was bottle aged by the producer at 12 °C in cellar conditions, then re-bottled in 2018 into Tokaj-shape 0.33 l bottles. After two weeks of bottling the Essence was sent to our laboratory in a spontaneous refermentation state with 1.86×10 5 CFU ml -1 initial cell concentration. Total yeast count was determined by culturing on DRBC agar (Sigma Aldrich), and the population was found heterogeneous upon colony and microscopic morphology (no further identifi cation was performed). Basic parameters of the Tokaj Essence were determined according to the offi cial OIV methods: 0/38 mg l -1 free/ total SO 2 (OIV-MA-AS323-04B), 54.19 °Brix total soluble solids (OIV-MA-AS2-02), 2.32 v/v% ethanol (OIV-MA-AS312-01A).
Refermentation conditions: To stop the spoilage, 0, 10, 20, and 40 mg l -1 of MCFA mixture was applied to 150 ml of Tokaj Essence in 200 ml fl asks, in duplicates, incubated at 15 °C. After 24 h of the MCFA dosage 100 mg l -1 SO 2 was added to each of these treatments, (except for an absolute control, where no SO 2 and no MCFA were used) according to the recommendation of an earlier study (B , 2014). Composition of the MCFA mixture was identical with the above described (see drop test section). Population dynamic changes of the MCFA-treated Tokaj Essence was followed with traditional plating of serial dilutions on DRBC agar surface, sampling at day 0, 1, 2, 7, 14, 21, 28, and 56.
All chemicals were purchased from Sigma-Aldrich Chemie Gmbh (Munich, Germany).

Tolerance test
Four botrytised wine-related yeast species were evaluated in terms of their tolerance in growth towards increasing amounts of MCFA. Considering the suggested practical application of MCFA, that is to stop fermentation (B , 2017), the conditions of the test included a small amount (5%) of ethanol present in the medium. All species under study showed some intraspecifi c variation between 5.8% and 25.6%, without correlation among better MCFA tolerance, better fermentation ability (e.g. M & T , 2011), and geographical origin (Table 1).
Due to the increasing MCFA levels, considerable diff erences were detected among the investigated species in their tolerance ( Fig. 1). At 10 mg l -1 MCFA, all S. bacillaris strains were slightly inhibited, while the other three species were not infl uenced signifi cantly (Fig.  2). In earlier studies, this MCFA level seemed to be eff ective in combination with SO 2 (100 mg l -1 total) and a higher concentration of ethanol (12 v/v%) in fermenting wine (B et al., 2017). From our results it could be seen that the MCFA mixture without SO 2 and with the presence of only 5 v/v% ethanol cannot inhibit the growth of the investigated strains at this low concentration. At 20 and 40 mg l -1 MCFA, S. bacillaris strains were further inhibited. S. cerevisiae strains were slightly reduced in growth at 20 mg l -1 MCFA, while the eff ect of the 40 mg l -1 MCFA was more pronounced. S. uvarum growth was not infl uenced considerably by 20 mg l -1 MCFA, while 40 mg l -1 MCFA resulted in noticeable intraspecifi c variance. E105, SB42, and TKH1 did not seem to be infl uenced, while S103 decreased moderately and CBS395 showed the highest sensitivity, comparable with that of S. bacillaris (Fig. 2). Z. bailii strains were able to tolerate these concentrations without signifi cant reduction (Fig. 2)  At 80 mg l -1 MCFA, all S. uvarum and S. bacillaris strains were inhibited completely, while some very limited growth was detected in the case of S. cerevisiae strains. Z. bailii strains showed still signifi cant growth and excellent tolerance. This could be a limitation of industrial MCFA application against refermentation, since this wine yeast is often responsible for spoilage of sweet wines (A et al., 2015). In this investigation, MCFA mixture as a sole yeast-inhibitor seemed to be eff ective only at considerably higher levels than in combinations used in earlier works (B et al., 2012; B et al., 2017).

Stop of refermentation in Tokaj Essence by MCFA
At the start, the Tokaj Essence had a considerable yeast concentration of 1.86×10 5 CFU ml -1 , which heterogeneous yeast population presented an excellent overall tolerance towards low cellar temperature and extreme amount of sugars (54.19 °Brix). In general, MCFA addition had a prompt eff ect on the population, since after 1 day, the cell concentration decreased by one order of magnitude in the case of 10 and 20 mg l -1 MCFA, and by two orders of magnitude in the case of 40 mg l -1 MCFA. The SO 2 addition in combination with the MCFA did not have additional short-term eff ect on the yeast population (Fig. 3). Comparing the absolute and sulphited control, it could be seen that this SO 2 concentration itself did not inhibit refermentation (Fig. 3), which is in accordance with the high sulphite binding capacity of botrytised wines and general yeast characterisation (R S , 1993).  : 20 mg l -1 MCFA+ 100 mg l -1 SO2; : 40 mg l -1 MCFA+ 100 mg l -1 SO2 The 10 mg l -1 MCFA+100 mg l -1 SO 2 had negligible inhibitory eff ect on the yeast population in the fi rst 28 days, while in the case of 20 mg l -1 MCFA+100 mg l -1 SO 2 cell concentrations were lower than both controls, but still rather limited inhibition was noticed. The 40 mg l -1 MCFA+100 mg l -1 SO 2 had more pronounced eff ect on the refermentation, gradual decline was observed (Fig. 3). After 28 days, the cell concentration decreased to 10 1 CFU ml -1 level, but the Essence still cannot be regarded as stable, free from possible refermentation.
After 56 days, the 10 and 20 mg l -1 MCFA+100 mg l -1 SO 2 reduced the living yeast cell concentration with only two orders of magnitude (Fig. 3), which is a considerable decrease regarding the initial cell number, although the 100 mg l -1 SO 2 alone had the same inhibitory eff ect. In the case of the 40 mg l -1 MCFA+100 mg l -1 SO 2 , the cell concentration remained in the 10 1 CFU ml -1 range, however, from an oenological point of view, the lowest remaining cell amount is still not acceptable in bottled wine. These results are diffi cult to compare with earlier works, since the parameters of the botrytised wine specialties, particularly Tokaj Essence, are considerably diff erent from those in a normal wine. The limited inhibitory eff ect of the MCFA dosage must be infl uenced by the reduced ethanol content of the Tokaj Essence, which is in accordance with an earlier fi nding about S. cerevisiae (V et al., 1989).

Conclusions
Due to the increasing MCFA concentrations, considerable diff erences were detected among the investigated yeast species in growth. S. bacillaris seemed to be the most sensitive, S. cerevsisiae and S. bayanus were more tolerant, while Z. bailii showed the highest tolerance. It could be concluded that at low ethanol content (5%) and without SO 2 , the MCFA mixture as a sole additive needs to be implied in considerably higher amounts than suggested for normal wines. The inhibitory eff ect of MCFA should be thoroughly tested in the future with a wider strain set of the currently investigated and other species. The MCFA-SO 2 combinations had rather limited inhibitory eff ect on the Tokaj Essence refermentation, possibly due to the low alcohol content of the botrytised wine specialty, but the excellent general stress tolerance of the spoilage yeasts should not be excluded. Consequently, the future MCFA application should be reduced only to wine-media, where signifi cant amount of ethanol is present to reach acceptable inhibition.