Inactivation of Brettanomyces bruxellensis and Saccharomyces cerevisiae in dry and sweet wines by high hydrostatic pressure

The aim of the research was to investigate a potential application of high hydrostatic pressure (HHP) for reduction/elimination of Brettanomyces bruxellensis and Saccharomyces cerevisiae in wines. Dry red wine was inoculated with B. bruxellensis and sweet white wine was inoculated with S. cerevisiae yeast. Both wines were treated by HHP under 100 and 200 MPa for 1, 3, 5, 15 and 25 min. The culturability was determined immediately after the treatment and again after 30, 60 and 90 days of storage. The phenolic content and chromatic characteristics were evaluated spectrophotometrically immediately after the treatment and after 90 days of storage. The culturability of B. bruxellensis was not confirmed immediately after the most invasive treatment (200 MPa for 15 and 25 min). With the same parameters, only a decrease in the culturability of S. cerevisiae was observed. During storage, opposing results were observed for two yeasts treated with 200 MPa for 15 and 25 min: there was a complete reduction of S. cerevisiae in the wine treated, but the culturability of B. bruxellensis completely recovered in all wines, implying that B. bruxellensis yeast entered a viable but not culturable (VBNC) state after HHP exposure. Regarding the chemical analyses, applied process parameters induced a slight decrease of anthocyanins in red wine, while changes of total phenolics and total colour difference value were negligible. In conclusion, HHP could potentially be successful for microbial stabilisation of sweet wines and consequently assure a lower use of sulphur dioxide, while inactivation of B. bruxellensis could only be successful in the early stages of wine contamination.


INTRODUCTION
Wine production is a complex microbiological and chemical process that involves the activity of different microorganisms -primarily yeasts. In order to produce high-quality wines, it is crucial to control the growth of microorganisms during the winemaking process. It is not only important to maximise the fermentative performance of Saccharomyces cerevisiae yeast and lactic acid bacteria during alcoholic and malolactic fermentation, but also to manage the growth of undesirable yeasts and bacteria (Fungelsang et al., 2007). Uncontrolled growth of specific wine microorganisms leads to significant wine spoilage, which could occur at several stages of winemaking: during alcoholic and malolactic fermentation, during the storage in barrels prior to packaging, and in the bottled wine (Agnolucci et al., 2017). One of these undesirable microorganisms is Brettanomyces bruxellensis, the main spoilage yeast in red wines. Contamination with this yeast negatively affects the sensory properties of wine through production of off-flavours that are often described as "barnyard", "horse sweat" and "medicinal". Moreover, this consequently results in wine not suitable for consumption and, subsequently, in great economic losses (Oelofse et al., 2009;Childs et al., 2015). The compounds most responsible for this off-flavour of red wine are volatile phenols (4-ethylphenol and 4-ethylguaiacol); these are formed by a B.
bruxellensis metabolism from hydroxycinnamic acids (p-coumaric and ferulic acid), which are their precursors naturally present in wine (Kheir et al., 2013). The main challenge in controlling B. bruxellensis growth is its strong ability to survive and grow in unfavourable conditions, such as low pH, high alcohol content, anaerobic conditions and very low amounts of fermentable sugars (Fabrizio et al., 2015). In stress conditions, this yeast has an ability to enter a viable but not culturable (VBNC) state, a physiological condition characterised by the inability of cells to divide on culture media, even if they are still alive and maintain metabolic activity (Agnolucci et al., 2010;Serpaggi et al., 2012;Agnolucci et al., 2017). This represents an additional challenge in controlling the B. bruxellensis population. Conversely, S. cerevisiae is considered the main yeast for alcoholic fermentation, but it could also negatively affect wine quality through refermentation of sweet wines and bottled wines with residual sugars (Loureiro and Malfeito-Ferreira, 2003). In addition to S. cerevisiae, refermentation could also be caused by other yeasts such as Zygosaccharomyces bailii, Schizosaccharomyces pombe and Saccharomycodes ludwigi, but it is Z. bailii and S. cerevisiae that most commonly cause refermentation in white sweet wines (Martorell et al., 2005;Enrique et al., 2007;Rai and Bridge, 2009;Alonso et al., 2015).
Microbiological stability is fundamental for preserving wine quality. Generally, this may be achieved by the use of chemical additives, physical removal through filtration, various biological procedures or other methods, along with the control of cellar/tank temperatures (Kheir et al., 2013;Lisanti et al., 2019). The proposed biological methods used to avoid yield losses caused by B. bruxellensis are related to antimicrobial peptides derived from natural proteins (Enrique et al., 2008), killer toxins produced by non-Saccharomyces yeasts (Santos et al., 2009;Mehlomakulu et al., 2014;Mehlomakulu et al., 2015), as well as tailored selected yeast and bacterial strains inoculated to promote alcoholic and malolactic fermentation (Berbegal et al., 2018). Nevertheless, the most popular, inexpensive and effective agent is the chemical preservative sulphur dioxide (SO 2 ). In terms of antimicrobial activity, molecular SO 2 is the most active form while bisulphite ions show low antimicrobial activity and bound SO 2 exerts only a low antibacterial action. In finished wine, the typical targets for preventing microbial spoilage are at least 0.6 and 0.8 mgL -1 of molecular SO 2 for dry and sweet wines, respectively (Waterhouse et al., 2016). However, the level of molecular SO 2 should be kept under its sensory threshold of 2 mgL -1 (Waterhouse et al., 2016). Previous investigations reported that a concentration of 1.4 mgL -1 of molecular SO 2 induced a VBNC state with up to 29 % of viable cells of B. bruxellensis, while viable cells were not detected at a concentration of 2.1 mgL -1 of molecular SO 2 (Agnolucci et al., 2013). Meaning that it can be expected that concentrations of total SO 2 in wine are likely to exceed legal limits and potentially also affect human health. According to Fabrizio et al. (2015), even if the health aspect is neglected the addition of SO 2 is not effective due to the presence of resistant strains or cross contamination in cellar operations.
Furthermore, the use of SO 2 is being reconsidered due to increasing allergenic Marina Tomašević et al. concerns, and researchers are investigating alternative methods to reduce the concentrations commonly added during winemaking (Santos et al., 2012;Guerrero and Cantos-Villar, 2015), wherein the use of high hydrostatic pressure (HHP) is highlighted as promising. Generally, HHP is a non-thermal batch process that utilises the application of pressures between 100 and 600 MPa to inactivate spoilage microorganisms in beverages and other foods (van Wyk et al., 2018). This is an interesting alternative to traditional food processing and preservation methods due to its limited effects on covalent bonds, and results in minimal modifications in the nutritional and sensory quality of treated products. Previous investigations regarding the effect of HHP on microbial stabilisation of wine are mainly related to controlling the growth of B. bruxellensis (González-Arenzana et al., 2016;van Wyk and Silva, 2017a;van Wyk and Silva, 2017b;van Wyk and Silva, 2019) and S. cerevisiae yeast (Tonello et al., 1998;Marx et al., 2011). Studies have demonstrated that the effectiveness of HHP on microbial inactivation is strongly dependent on applied pressure and treatment duration (González-Arenzana et al., 2016;van Wyk and Silva, 2017a,b), as well as on the type of wine treated (van Wyk and Silva, 2017b).
The aim of this research was to preliminary investigate the effectiveness of HHP process parameters (pressures and treatment duration) for inactivation of B. bruxellensis and S. cerevisiae yeasts. Its efficiency was examined in terms of the different types of wine, where dry red wine was used for inoculation by B. bruxellensis, and sweet white wine for inoculation by S. cerevisiae yeast. The effect of HHP on main quality parameters (total phenolics, total anthocyanins and colour change) was also examined.

Wines
The wines used in this study were vintage 2016 red wine Cabernet Sauvignon and sweet wine Graševina, obtained from winery Erdutski vinogradi d.o.o. (Erdut, Croatia). The wines were passed through a 0.4-µm pore filter. Next, the physicochemical parameters and microbial population were determined by spreading both wines on various solid mediums (WLN agar, Brettanomyces agar, Sabourand dextrose agar and Lysine agar). After confirming that there were no undesirable microorganisms, each wine was further inoculated with the tested yeast. The average physicochemical parameters of the wines were analysed by FTIR spectrophotometer Bacchus II (Microdom, Saint-Savinien, France) and the results are showed in Table 1.

Wine preparation
B. bruxellensis CBS 2499 and S. cerevisiae DSM 70468 were used as test microorganisms. These strains were selected because the B. bruxellensis CBS 2499 strain has a high tolerance to stress and S. cerevisiae DSM 70468 strain has a high tolerance to ethanol, and because they were not previously tested for inactivation by HHP. Moreover, the CBS 2499 strain has a published genome and the DSM 70468 strain has a high-quality Illumina sequence read dataset, which should facilitate future studies at the transcriptome, proteome, and metabolome levels, hopefully leading to a complete understanding of HHP effects on these yeasts. Both cultures were stored at -80 °C in 20 % glycerol (v/v). Yeast peptone dextrose (YPD) medium (10 gL -1 yeast extract, 20 gL -1 peptone and 20 gL -1 glucose) was used for growing B. bruxellensis and S. cerevisiae inoculum. YPD medium was acidified to pH 3.5 with phosphoric acid.
The inocula were prepared as described by Delsart et al. (2016) and Križanović et al. (2019) in a four-step procedure, to ensure yeast adaptation to wine conditions. Both yeast cultures, B. bruxellensis and S. cerevisiae, were grown in separate YPD media and then in YPD media with 4, 8 and 12 % (v/v) ethanol, at 28 ºC without shaking. When the yeast population entered the stationary phase in the first medium, another medium was used and inoculated with 10 %, v/v of culture from the previous medium. The growth of both yeasts during the inoculum preparation in YPD media without or with ethanol was followed by measuring the absorbance at 600 nm and by plate counts (Table  2), but the cell population of both yeast in wines was only measured by plate counts.
Cells of B. bruxellensis and S. cerevisiae yeast were harvested by centrifugation at 4000 rpm for 10 min and inoculated into wine samples at approximately 6 and 5 log (CFUmL -1 ) for red and sweet wine, respectively. The initial cell populations of B. bruxellensis and S. cerevisiae yeast in wine samples were then confirmed by plate counts. The inoculation of wine samples was performed 24 h before HHP treatment and these were incubated at 20 ± 2 °C. The cell population of both yeasts in the wine samples immediately after inoculation and after 24 h, i.e. before HHP treatment, was identical.

High hydrostatic pressure processing
The experiment was performed in three biological replicates for each yeast. Each biological replicate was treated with high hydrostatic pressure system FPG7100 (Stansed Fluid Power, Harlow, UK) in triplicate for each set of processing parameters and then analysed.
After inoculation with tested yeast, 100 mL of wine was poured into a plastic bottle, packaged in an individual plastic bag, vacuum sealed, and placed in a working vessel of high hydrostatic pressure system. To assess the possible effects of HHP treatment, an experimental test was conducted with various pressures (100 and 200 MPa) and processing times (1, 3, 5, 15 and 25 min). A control sample represented the wine sample inoculated with B. bruxellensis or S. cerevisiae and not exposed to HHP treatment. The control sample and the samples obtained after HHP treatment were placed in sterile bottles and stored for 90 days at 20 ± 2 °C, to assess the long-term effects of applied treatment.
Inoculated wines were analysed before and immediately after HHP treatments as well as after storage for 30, 60 and 90 days. The samples were decimally diluted in physiological saline solution. The serial decimal dilutions and the undiluted sample of inoculated red wines were plated onto Brettanomyces agar in triplicate and incubated at 28 ºC for up to 7 days, and the samples of inoculated sweet wines were plated onto WLN agar (with added chloramphenicol) in triplicate, and incubated at 25 ºC for up to 2 days. The number of colony-forming units (CFU) was calculated according to the following formula (Eqn. 1):

Spectrophotometric and chromatic analyses
The spectrophotometric analyses were carried out using a Specord 50 Plus spectrophotometer (AnalytikJena, Jena, Germany). Total phenolics (TP) content was determined by the Folin-Ciocalteu method as described in detail by Singleton and Rossi (1965). The results were expressed as mgL -1 of gallic acid equivalents (mg GAEL -1 ). The total anthocyanins (TA) content was determined using the SO 2 bleaching method as previously described by Ribéreau-Gayon and Stonestreet (1965). All spectrophotometric analyses were conducted in triplicate.
The chromatic characteristics of the wine samples were measured with a Specord 50 Plus AnalytikJena spectrophotometer (Jena, Germany) using the CIELab space (OIV, 2006). Chromatic characteristics were expressed as CIE coordinates of L (lightness), a (redness/ greenness) and b (yellowness/blueness), with illuminant D65 and observer 10° standardisation, using the software WinASPECT PLUS (Jena, Germany). All measurements were performed in triplicate and the results presented as average value of total colour difference. The total colour difference value (∆E) between the control and treated wine samples was calculated with the following formula (Eqn. 2):

Data analysis
Significant differences between samples for chemical and microbiological analyses were determined by one-way analysis of variance (ANOVA) using the Statistica V.10 software (StatSoft Inc., Tulsa, USA). Tukey's honestly .05) was used for comparison when samples differed significantly after ANOVA. Figure 1 shows the reduction of B. bruxellensis and S. cerevisiae yeasts in wines subjected to pressures of 100 and 200 MPa for 1, 3, 5, 15 and 25 min. The initial cell population of B. bruxellensis yeast in wine before HHP treatment was 5.5 log (CFUmL -1 ) and S. cerevisiae yeast was 4.4 log (CFUmL -1 ). With 100 MPa HHP treatment for all treatment durations, no reduction in cell culturability was observed for the yeasts. Conversely, with 200 MPa HHP treatment a reduction in cell culturability was observed for both yeasts even for the shortest treatment duration of 1 min.

Short-term effect of HHP
However, at 200 MPa the treatment duration had a significantly different impact between the two yeasts. After treatment for 1 min, a similar reduction in culturability was observed for the yeasts. At longer treatment durations (3 and 5 min), the observed culturability reduction was much higher for B. bruxellensis and the difference increased with treatment duration. In S. cerevisiae the largest culturability reduction was 1.5 log (CFUmL -1 ), observed after the 15and 25-min treatments (the reduction was 1.0 log (CFUmL -1 ) after the 1-min treatment). In B. bruxellensis the cell culturability decreased rapidly, achieving complete reduction after the 15-min treatment, with the observed culturability reduction of 5.5 log (CFUmL -1 ).

Long-term effect of HHP
The long-term effect of HHP was investigated during 90 days of storage at three different sampling points (after 30, 60 and 90 days), and While monitoring sweet white wine treated with 200 MPa during storage, the cell culturability of S. cerevisiae was observed only in samples treated for 1, 3 and 5 min, but not in samples treated for 15 and 25 min. A gradual increase in cell culturability during storage was observed in samples treated for 1, 3 and 5 min. However, after 30 and 60 days, cell culturability in the control sample was higher than in samples treated for up to 5 min, while after 90 days the culturability for those samples was roughly equal to the control sample.

Changes in wine phenolics and chromatic characteristics
To evaluate the potential effect of HHP on wine quality, total phenolics and total anthocyanins were determined in all samples, and the results are shown in Table 3. In the red wine, immediately after treatment the concentrations of both total phenolics and total anthocyanins were significantly different in treated and untreated Marina Tomašević et al. wines, but those differences were not significant. The greatest differences were observed for total anthocyanins in wines treated with 200 MPa for 15 and 25 min, but the decrease in these compounds was less than 10 % in comparison to control wine. During the storage period, the difference in the concentration of total phenolics were also statistically significant, but rather small between the control and treated wines. Conversely, the differences in the concentration of total anthocyanins were statistically significant between the samples treated with different pressures, but not statistically different between the samples treated with the same pressure concentrations. Herein, the lower concentrations were observed in the treated samples, with a larger decline determined in the samples treated with 200 MPa. Moreover, longer treatment durations resulted in a larger decrease of these compounds. For example, the most invasive treatment -200 MPa for 25 minresulted in a more than 26 % lower concentration of total anthocyanins after 90 days of storage. In terms of total phenolics sweet wine demonstrated a similar trend to that observed in red wine, where the differences in concentration of total phenolics in the control and treated wines were statistically significant, but also quite small at less than 5 %. After 90 days of storage a decrease was observed for all analysed parameters, but there were no significant differences in the concentrations of total phenolics between treated and control samples.
In addition, the total colour difference value (∆E) between the control and treated wine samples was calculated and the results are also shown presented in

DISCUSSION
Brettanomyces bruxellensis is often considered the main spoilage microorganism in red wines. It causes deterioration of wine quality through the production of volatile phenols, which are the compounds responsible for off-flavours in red wine (Oelofse et al., 2009;Childs et al., 2015). In the wine industry, contamination by B. bruxellensis is very hard to control due to its ability to survive and grow in conditions that are inhibitory for most other microorganisms (low pH, anaerobic conditions, high alcohol level, and very low amounts of fermentable sugars), and its ability to enter a VBNC state in stress conditions (Agnolucci et al., 2010(Agnolucci et al., , 2017Serpaggi et al., 2012;Fabrizio et al., 2015). Several other yeast species are also associated with wine quality degradation. Primarily, these yeasts are related to the refermentation of sweet wines and bottled wines with residual sugars. Yeasts such as Saccharomyces cerevisiae, Zygosaccharomyces bailii, Schizosaccharomyces pombe and Saccharomycodes ludwigi are reported to cause refermentation in various types of wine, while S. cerevisiae and Z. bailii are most commonly reported to cause refermentation in sweet white wines (Loureiro et al., 2003;Martorell et al., 2005;Enrique et al., 2007;Rai and Bridge, 2009;Alonso et al., 2015).
Among the different methods employed for microbial stabilisation, SO 2 is most common, but as it is not always effective due to its potential negative effect on human health and the presence of resistant yeast strains or the occurrence of cross contaminations in cellar operations (Fabrizio et al., 2015), new technologies for preventing and controlling undesirable yeasts are being developed. One of these technologies is high hydrostatic pressure (HHP), wherein pressures between 100 and 600 MPa are employed to inactivate spoilage microorganisms in beverages and other foods (van Wyk et al., 2018). Until now, this was successfully applied for growth control of several strains of B. bruxellensis (AWRI 1499, AWRI 1608, AWRI 1613 and two strains isolated from Rioja red wines) and several strains of S. cerevisiae (CBS 1171, KCCM 1224 and different commercial strains) in must and in white, red and rosé wines (Tonello et al., 1998;Perrier-Cornet et al., 2005;Mok et al., 2006;Marx et al., 2011;González-Arenzana et al., 2016;Van Wyk and Silva, 2017a,b;Van Wyk et al., 2019).

Short-term effect of HHP
The trends observed (a higher reduction with higher pressures and longer treatment durations) are in line with previous investigations, which showed that microbial inactivation during HHP processing strongly depends on applied pressures and treatment duration (González-Arenzana et al., 2016;van Wyk and Silva, 2017a,b). Additionally, a recent investigation suggested a minimum threshold pressure below which no inactivation occurs (van Wyk and Silva, 2019), with a pressure of 100 MPa considered to be the threshold as it did not result in significant inactivation of B. bruxellensis yeast (González-Arenzana et al., 2016;Van Wyk and Silva, 2017a). Our study additionally confirms that hypothesis, for reduction of both yeasts, as there were no significant differences observed after application of 100 MPa pressure. Although application of the same pressure resulted in statistically significant reduction of B. bruxellensis cell culturability (Figure 1a), these changes are negligible as they are under the same decimal dilution. Hence, the results suggest that a more effective reduction could be achieved by applying higher pressures, such as 200 MPa.
Regarding the yeast species treated, a higher reduction was achieved for B. bruxellensis (Figure 1a) than for S. cerevisiae (Figure 1b). Treatments at 200 MPa for 15 and 25 min resulted in the highest reduction for both yeasts, but these treatments were more effective for B. bruxellensis as its culturability was not confirmed after the treatment (Figure 1a). As previously stated, complete reduction of S. cerevisiae was not achieved by applied pressures, and the most intensive treatment (200 MPa during 25 min) resulted in only 1.5 log reduction. According to data in the literature (Tonello et al., 1998), higher applied pressures (300 and 500 MPa) resulted in a higher reduction in S. cerevisiae. Marx et al. (2011) showed that application of 600 MPa for 7 min resulted in complete inactivation of this yeast. The discrepancy between the B. bruxellensis and S. cerevisiae reduction could be due to several factors. First, different yeast species and strains could behave differently under HHP treatment (van Wyk and Silva, 2017a). Tonello et al. (1996) showed that HHP inactivation depends on type and size of microorganisms targeted, wherein different efficacy was obtained for S. cerevisiae and S. ludwigii yeasts. Taken together, we can expect that optimal set of processing parameters for preventing wine spoilage will depend not only on yeast species, but also on the strain. Second, the sugar present in wine could also act protectively, as determined for B. bruxellensis yeast treated by thermosonication (Križanović et al., 2020). Third, alcohol concentration also proved an important factor for microbial inactivation, with higher reduction achieved in wines with alcohol levels above 13 % (Tonello et al., 1996(Tonello et al., , 1998van Wyk and Silva, 2017b;Križanović et al., 2020), which were significantly higher than those in treated sweet wine (11 %) (Table 1). Generally, previous investigations also established that suspending medium could significantly affect pressure resistance of microorganisms. Moreover, the resistance could be greatly increased when treated in nutritionally rich media containing substances that can act protectively against damage or contain nutrients essential for microorganism repair (van Opstal et al., 2003;Molina-Höppner et al., 2004;Chauvin et al., 2006;Goh et al., 2007;Sokołowska et al., 2013). In other words, the results obtained in this study indicate the baroprotective effect of sugar in sweet wine on S. cerevisiae cells during pressurisation.

Long-term effect of HHP
Our results (Figure 2) suggest that HHP treatment resulted in B. bruxellensis entering an VBNC state; the culturability was not confirmed immediately after 15-min and 25-min treatments (Figure 1a), but after the stress was removed the cells were able to grow on culture media again. Unlike the given results, van Wyk et al. (2018) showed that HHP processing at 400 MPa for 5 s resulted in complete inactivation of B. bruxellensis yeast. The authors showed that culturability was not recovered during 12 months of storage, meaning that the VBNC state after HHP exposure was not observed, as is the case in our research. The obtained differences could be primarily due to the variation of applied pressures. Hence, the lower pressures used in this research potentially cause B. bruxellensis to enter a VBNC state, but higher pressures (400 MPa) used in van Wyk et al. (2018) cause complete inactivation. Additionally, a potential explanation could be that the strain (AWRI 1499) used in that study is more susceptible to HHP treatment in comparison to the CBS 2499 strain we used. Generally, there is a large diversity of genotypes and phenotypes of B. bruxellensis strains isolated from wine, and previous studies showed that metabolic activity, sulphur dioxide tolerance and behaviour in the VBNC state are strain dependent (Capozzi et al., 2016;Crauwels et al., 2016;Avramova et al., 2019;Dimopoulou et al., 2019). In terms of non-thermal technologies, it has already been demonstrated that HHP inactivation efficiency strongly depends on the yeast strain used, wherein AWRI 1499 was more resistant than AWRI 1608 and AWRI 1613 (van Wyk and Silva, 2017a). Although analyses regarding high tolerance to stress showed that genomic characteristics are distributed in the genome of both CBS 2499 and AWRI 1499 strains (Godoy et al., 2017), it is possible that CBS 2499 could be more resistant to HHP treatments. However, before making any general conclusions it is necessary to conduct further, more extensive studies. Nevertheless, van Wyk et al. (2018) showed that, among higher sulphur dioxide, pulsed electric field (PEF) and HHP, the latter was the best and the most reliable treatment for B. bruxellensis inactivation, meaning that HHP has a great potential for the wine industry.
Besides provoked stress conditions, the main theory regarding the microbial inactivation by HHP is related to changes in the cell membrane associated with macromolecules such as proteins, as well as to compression of the membrane's bilayer due to changes in permeability (Ross et al., 2003). These changes may not necessarily lead to cell death, although surviving cells showed enhanced sensitivity to different inhibitors (Mackey, 2000). In fact, several investigations showed the occurrence of sublethal injury after HHP exposure using different plating techniques (Alpas et al., 2000;Chilton et al., 2001;Jordan et al., 2001;Somolinos et al., 2008). In other words, and contrary to the VBNC state, sub-lethally injured cells still possess the ability to grow on growth media (Li et al., 2014). This effect is the most probable explanation for trend estimated in the case of S. cerevisiae yeast: immediately after the longest HHP treatments (15 and 25 min) sublethal injury occurred and, consequently, cells were able to grow on selective medium. However, after a certain period of storage, enhanced sensitivity resulted in cell death, Marina Tomašević et al. potentially due to higher levels of sugar (osmotic pressure), presence of ethanol, sulphur dioxide, or their combination. Nevertheless, these results imply that HHP treatment could be successfully applied in sweet wine production, in terms of avoiding the risk of a secondary fermentation and subsequent degradation of wine.

Changes in wine phenolics and chromatic characteristics
In terms of the phenolic content, the results obtained in this research (Table 3) are in accordance with previous investigations where it was demonstrated that HHP treatment did not result in immediate changes of total phenolics in red and white wines, while changes in concentrations were estimated after 9 and 12 months of storage (Santos et al., 2013a,b). As with sweet wine, these differences were observed in both, control and treated wines, and most probably were due to different chemical and biochemical reaction that occur during wine ageing.
Furthermore, the colour difference value (∆E) is important because it shows whether the changes in the chromatic characteristics were visually relevant. According to the literature, the difference in colour is clearly perceptible when the value of ∆E between two samples is in the range 2 -10, while for values higher than 10 the colours are more opposite than similar (Brainard, 2003). Moreover, a recent study showed that all the colour differences higher than 6 should be taken as considerable, clearly visible (Ramirez-Navas and Rodriguez de Stouvenel, 2012), while values higher than 3 were visible to the human eye (Puértolas et al., 2010).
These results confirm the potential of the HHP technique for microbial stabilisation of sweet wines, wherein the yeasts are inactivated without detrimental effects on wine quality parameters, such as visual colour perception and phenolic content. A previous investigation also reported that HHP treatment resulted in low values of ∆E, meaning that although changes in chromatic characteristics occurred, they could not be perceived by the human eye (Tao et al., 2012). Santos et al. (2013a) showed that immediately after the HHP treatment and during first six months of storage the ∆E values of red wine were below 3, but increased after 9 and especially after 12 months of storage. A similar trend was observed in white wines, but the values above 3 were observed much earlier, after 3 months of storage (Santos et al., 2013b). Even though obtained results suggest application of HHP on wine without effect on main quality parameters, it is necessary to conduct further research on longer storage periods and overall wine quality.

CONCLUSIONS
The results here showed that effectiveness of HHP for microbial stabilisation of wine depends on process parameters, such as applied pressure and treatment duration. Immediately after the treatment, HHP was more effective for reducing B. bruxellensis in red wine, wherein its culturability was not confirmed after pressurising by 200 MPa for 15 and 25 min. Conversely, the same process parameters only resulted in a decrease not complete reduction of culturability of S. cerevisiae yeast in sweet wine. The opposite result was established during 90 days of storage. Primarily, the culturability of B. bruxellensis was confirmed in all analysed samples, indicating that cells entered a VBNC state after HHP exposure, but after a certain period they were able to grow on culture media and were completely recovered. However, culturability of S. cerevisiae in samples treated by 200 MPa for 15 and 25 min was not confirmed after 30, 60 and 90 days of storage. The results suggest that HHP is a promising technique for microbial stabilisation of sweet wine, in terms of avoiding refermentation, and consequently it can assure lower concentrations of sulphur dioxide in treated wines. In the case of B. bruxellensis, HHP could have a positive effect only in the first stages of wine contamination, when the number of yeast cells is considerably lower. In terms of the quality parameters, applied treatments slightly affected the concentration of total anthocyanins in red wine, while changes in total phenolics and total colour difference value were negligible and unrelated to treatment.