Antimicrobial activity of ozone. Effectiveness against the main wine spoilage microorganisms and evaluation of impact on simple phenols in wine

180 Assay of antimicrobial activity of ozone Australian Journal of Grape and Wine Research 19, 180 188, 2013 Antimicrobial activity of ozone. Effectiveness against the main wine spoilage microorganisms and evaluation of impact on simple phenols in wine R. GUZZON, T. NARDIN, O. MICHELETTI, G. NICOLINI and R. LARCHER Technology Transfer Centre, Edmund Mach Foundation, Via E. Mach 1, 38010 San Michele all Adige (TN), Italy Correspondence author: Dr Raffaele Guzzon, email raffaele.guzzon@fmach.it Abstract Background and Aims: Microbial contamination affects winemaking, especially after fermentation, e.g. wine ageing in barrels or in contact with oak pieces, when spoilage microbes find an environment favourable for their development. Ozone was evaluated as a sanitising agent in order to assess its potential to prevent microbial spoilage occurring during ageing of wine in barrels using a model system based on barrel wood. Methods and Results: Fifty microorganisms of oenological significance were evaluated for their spoilage potential in the barrel. Ethanol resistance, biofilm formation and production of volatile phenols were studied using physiological tests. The effectiveness of ozone in eliminating microorganisms was evaluated in aqueous solution at several cell and ozone concentrations. At a high cell concentration, the presence of organic matter reduced the effectiveness of ozone. At a cell concentration of under 10 3 CFU/mL, typical of wine cellars, ozone was able to eliminate microorganisms. Resistance to ozone was observed in diverse microorganisms, and this feature is linked to their ability to produce a biofilm. The reduction in simple phenols obtained from oak wood was tested by treating oak chips, routinely used in the wine industry, with an increasing dose of ozone. There was no statistical difference in the phenolic composition of wine treated with six commercial chips. Only a significant exposure of the chips to ozone caused a 33%reduction in the initial content of gentisic acid. Conclusions: Ozone was shown to be a highly effective sanitising agent without interfering with the profile of the phenolic substances extracted from oak. The application of ozone for barrel sanitising may be a feasible solution for the prevention of wine spoilage during ageing in oak barrels. Significance of the Study: A survey of the effect of ozone on a large number of microorganisms and phenolic compounds of oenological significance, considering some technological variables, is reported. Keywords: barrel, Brettanomyces/Dekkera, oak phenolic profile, ozone, wine spoilage Introduction In the last few years, the interest of consumers and food producers in low-impact and safe technologies has increased. The use of food additives with an antimicrobial effect is therefore not without complications, because of tightening of regulations and, no less importantly, to consumer opposition. In the wine industry, the use of some antimicrobials, such as sulfur dioxide and lysozyme (Usseglio-Tomasset 1992, Guzzo et al. 2011), has been restricted, with a significant impact on the microbiological stability of wines and the control of microbial contamination in the wine cellar. Ozone could be an alternative to traditional approaches to microbial control. This molecule, coming from a rearrangement of oxygen in the presence of intense electric discharge (Jin-Gab et al. 2003), has some attractive features. Because of recent advances in technology, i.e. ozone generators based on dielectric barrier discharge (Haverkamp et al. 2002), ozone may be a possible, cheap and in situ solution. Ozone has a non-specific action and has been shown to be active against all microbiological forms: fungi, bacteria, virus and spores (Khadre et al. 2001). Furthermore, its high reactivity ensures the complete disappearance of residues in the environment after just a few minutes of treatment (Khadre et al. 2001). Ozone acts as a dipole and has electrophilic and nucleophilic properties. It can react with numerous biological structures relevant to cell life. It oxidises unsaturated fatty acids, glycoproteins or glycolipids which make up the cell envelope (Komanapalli and Lau 1998, Erickson and Ortega 2006) and proteins of spore coats (Foegeding 1985, Khadre et al. 2001). The functioning of enzymatic systems and nucleic acids may also be compromised by the action of ozone (Komanapalli and Lau 1998). Several applications for ozone have been proposed at different stages in winemaking. Ozone is able to reduce spoilage microflora in grapes, barrels and tanks (Coggan 2003, Hester 2006, Guillen et al. 2010). Moreover, ozone has been recognised as an effective alternative to traditional approaches based on steam or chemical agents for sanitising bottling plants and bottles. Despite this practical experience, a detailed study of the effectiveness of ozone against wine microflora and the possible risks associated with its use is still lacking. In comparison with other fields in the food industry, the microflora associated with grapes, must and wine are much more complex, because of the absence of sterilisation treatments or pasteurisation of raw materials. The species of yeast, bacteria and mould vary during winemaking in response to the compositional changes of the fermenting substrate, as they have different doi: 10.1111/ajgw.12018

Guzzon et al. Assay of antimicrobial activity of ozone 181 environmental and nutritional requirements (Ribéreau-Gayon et al. 2006). It is therefore reasonable that these species will show a variable resistance to the action of ozone, as found for other antimicrobial agents (Bartowsky 2009, Guzzon et al. 2011). Furthermore, the use of ozone in the wine cellar has met with reluctance on the part of winemakers, because of the potential risk of uncontrolled oxidation of some compounds that are important to the sensory characteristics of wine. In this context, the study aimed to define the application and the possible risks associated with the use of ozone as a sanitising agent in wine cellars. Considering that some of the most serious microbial contamination occurs during the ageing of wine in barrels, the study focused on the effectiveness of ozone in combating wine spoilage microorganisms that have adapted well to the barrel environment and on the impact of ozone on simple phenols, which represent a significant fraction of the extractable components transferred to wine by barrels or oak pieces. Materials and methods Microorganisms and microbiological tests The microorganisms studied (Table 1) were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Leibniz-Institut, Braunschweig, Germany), the Industrial Yeasts Collection (DBVPG, University of Perugia, Perugia, Italy) and the ARS Culture Collection (NRRL, National Centre for Agricultural Utilization Research, Peoria, IL, USA). Some species were isolated at wineries by researchers from the Edmund Mach Foundation (San Michele all Adige, Italy) and identified by sequencing the D1/D2 domain of the large subunit 26S ribosomal ribonucleic acid, following the protocol proposed by Kurtzman and Robnett (1998). The sequences (600 base pairs) were compared with those available in the GenBank DNA database (http://www.ncbi.nlm.nih.gov/). According to the needs of the different species, microorganisms were cultured in Wallerstein Laboratory Agar Medium (WL, Oxoid Limited, Basingstoke, England) or in De Man Rogosa Sharpe Agar Medium (MRS, Oxoid Limited). Cell density was determined using the plate count method, as specified by the Organisation Internationale de la Vigne et du Vin (Organisation Internationale de la Vigne et du Vin 2011). The optical density (OD) of microbial cultures was measured at 460 nm with a Scientific Evolution 300 ultraviolet-visible spectrophotometer (Thermo Fischer Scientific, Waltham, MA, USA). Biofilm production and adhesion activity were evaluated as proposed by Joseph et al. (2007). The results were expressed as the OD reading at the end of the tests, which is directly correlated to biofilm formation or the adhesion of each species. The ability of microorganisms to produce volatile phenols was tested by growth in a synthetic medium containing cinnamic acids, as proposed by Chatonnet et al. (1995). The volatile phenols were quantified by high-performance liquid chromatography (HPLC)-electron capture detection (Larcher et al. 2007). The results are expressed as the substrate conversion yield present in the culture media. The ability of microorganisms to grow in the presence of ethanol was evaluated by measuring OD, read after 10 days of growth at 25 C in the appropriate medium with the addition of the 15% v/v of ethanol (Sigma-Aldrich Chemie GmbH, Steinheim am Albuch, Germany). Table 1. List of microorganisms selected for the preliminary test, with the objective of identifying species causing the most spoilage in wine barrels. Adhesion activity Biofilm production Volatile phenol yield (%) Ethanol resistance Brettanomyces bruxellensis 0.093 0.562 86 1.984 Acetobacter aceti 0.121 0.684 1.222 Candida ishiwadae 0.327 0.564 100 2.236 Candida vini 0.215 0.312 6 2.459 Cryptococcus rajasthanensis 0.390 0.708 100 1.468 Gluconobacter oxidans 0.085 0.353 0.795 Hanseniaspora uvarum 0.311 0.375 100 0.587 Issatchenkia terricola 0.097 0.564 9 1.122 Lactobacillus planatarum 0.366 0.483 0.633 Metschnikowia pulcherrima 0.149 0.424 9 0.431 Oenococcus oeni 0.114 0.341 1.289 Pediococcus acidilattici 0.166 0.220 1.683 Pediococcus pentosaceus 0.198 0.241 1.589 Pichia fermentans 0.195 0.197 100 1.922 Pichia stiptis 0.219 0.377 11 1.986 Rhodotorula mucilaginosa 0.426 0.525 100 0.433 Saccharomyces cerevisiae 0.093 0.174 7 2.683 Schizosaccharomyces pombe 0.193 0.248 17 1.822 Torulaspora delbrueckii 0.136 0.100 8 2.456 Zygosaccharomyces bailii 0.139 0.107 12 1.982 Expressed as optical density measured at the end of the respective tests. Expressed as the ratio between the sum of ethyl and vinyl phenols produced and the initial amount of cinnamic acid present in the media., Test not performed for this microorganism.

182 Assay of antimicrobial activity of ozone Australian Journal of Grape and Wine Research 19, 180 188, 2013 Ozone test Ozone was produced by an Ossigen A3 Cold Plasma Generator Air (Aslan srl, Padova, Italy) with a maximum production rate of 12 g O 3/h coupled with an ATI Q45H Ozone probe (Analytical Technology Inc., Collegeville, PA, USA) for the detection of ozone concentration in water or air. For tests aimed at evaluating the effect of temperature (range 7 27 C), ozone was dissolved in sterile distilled water (volume 15 L), the ozone concentration being measured at intervals of 5 min using the ATI Q45H probe. In order to verify the effect of organic matter on ozone activity, an increasing dose of heat-inactivated dry yeast was dissolved in 15 L of water at 17 C. The organic matter was expressed as chemical oxygen demand (COD), measured by Spectroquant Test Kits (Merck Millipore, Darmstadt, Germany). After dissolution of the ozone (initial concentration of O 3 12.8 mg/l) its residual concentration in the medium was measured at intervals of 5 min using the ATI Q45H probe. A pure culture of each microorganism (initial cell concentration 8 10 8 CFU/g), after the appropriate decimal dilution, was tested in 1 L of water at 17 C at four ozone concentrations (range 1 7 mg/l) and at five initial cell densities (range 10 2 10 6 CFU/mL). Microorganisms were contacted with aqueous O 3 for 30 min, after which the residual cell concentration was evaluated using the plate count method. The results were expressed as the reduction in cell concentration because of the action of ozone, calculated as the ratio between the initial microbial concentration and the residual cell density at the end of test, both expressed as log CFU/mL. Degradation of simple phenols by ozone Six types of oak chips (French and American oak chips with low, medium and high toasting; medium size 3 5 5 mm) were purchased (Tonelerìa Nacional, Santiago, Chile). Wood chips were used in order to maximise the ozone/wood and wood/ wine contact surface, and to minimise the heterogeneity of the inner surface of the traditional, fire-toasted barrels. Homogenous, 500 g samples of chips were treated inside a stainless steel chamber (a hermetically closed cylinder with an internal volume of 5 L, rotating at 60 rpm) with defined doses of gaseous O 3. The first treatment protocol exposed the chips to fluxing ozone for 30 s until an initial concentration of 10 mg O 3/L was reached. Then, once the flux was stopped, the chip samples remained in an ozone atmosphere for 9.5 min, mixing continuously, before removing them from the chamber. This procedure was repeated three times for each chip sample, preparing 18 single-treated samples (T1 group). A subaliquot of each of these samples was treated identically, obtaining a further 18 double-treated samples (T2 group). The chip samples were subjected to a second more rigid protocol (four times) in the rotating chamber with a continuous ozone flow of 12 g O 3/h for 40 min. Considering the production rate of the apparatus and the reactor volume, we estimate that ozone reached a maximum concentration of 20 mg/l in 5 min and remained at that level until the end of the experiment. After the preparation phase, all the untreated and treated chip samples were stored in closed, clean plastic boxes in air at 25 C for one month before the analytical tests. The possible oxidative action of ozone on the components extracted from oak was investigated by immersing the control and O 3-treated chips in wine. A homogeneous, 5 g sample of chips was added to 50 ml of a white wine (alcohol content 12% v/v, ph 3.5) in a 100 ml polyethylene bottle (Kartell, Noviglio, Italy). The headspace inside the container was saturated with nitrogen gas before capping. The mixture was shaken daily for 7 days before removal of the chips. Fifty-four wines were prepared for the first test (n = 18 for control T0, 18 T1 and 18 T2) and 48 for the second test (24 control samples and 24 treated samples). The wines were 0.45 mm filtered and transferred under nitrogen into clear 2 ml vials for analysis. Twenty-nine wine phenols and cyclotene (vanillin was supplied by Carlo Erba Reagenti (Arese, Italy); hydroxytyrosol and tyrosol by Chemos GmbH (Regenstauf, Germany); 4- ethylphenol, 4-hydroxybenzoic acid, 4-methylcatecol, ellagic acid, gentisic acid, isoeugenol, protocatechuic acid, scopoletin, syringic acid, tryptophol and vanillic acid by Fluka (Buchs, Switzerland); 4-vinylphenol and syringaldehyde by Safc (Modena, Italy); 4-allylsyringol, 4-methylsyringol, acetosyringone, coniferyl aldehydes, cyclotene, esculetin, guaiacol, homovanillic acid, homovanillic alcohol, isopropiovanillone, protocatechuic aldehydes, synapaldehyde and syringol by Sigma-Aldrich Chemie GmbH; isopropiosyringone by TransMIT GmbH (Gießen, Germany)) were quantified with an HPLC 2695 Alliance system (Waters Corporation, Milford, MA, USA), equipped with an eight-electrode coulometric array electrochemical detector 5600A (ESA, Bedford, MA, USA) adapting the method of Larcher et al. (2007). The compounds were separated on a LiChroCART RP-18e Purospher column (125 mm 3 mm, 5 mm particle size (Merck). The quantification limit and precision for all the tests (as for Relative Standard Deviation) were equal to or lower than 0.01 mg/l and 3%, respectively. Statistical analysis Statistical analysis of the data was carried out using Statistica 7.1 software (StatSoft, Inc., Tulsa, OK, USA). Results and discussion Validation of the effectiveness of ozone under wine cellar conditions The effectiveness of ozone and consequently its antimicrobial efficacy are influenced by several environmental parameters: ph, temperature and the presence of ozone-consuming compounds (Khadre et al. 2001, Guillen et al. 2010). The present work takes into account two parameters: temperature and the amount of organic matter present in the water used as the ozone carrier. The first test measured the rate of ozone degradation in sterile water over a temperature range normally found in wine cellars, i.e. between 7 and 27 C (Figure 1a). In Figure 1b, the residual ozone concentration was reported at three times. Following the ideal gas law, the water temperature regulates the initial concentration of ozone which, under the experimental conditions, reached a maximum of 14 mg/l in the test at 7 C. Over and beyond the absolute concentration value, however, a linear correlation between the rate of ozone degradation and the water temperature was found over the tested interval. Considering that barrels are usually stored in cold rooms inside wineries, at a temperature range between 15 and 20 C, the results are encouraging because in the test performed at 17 C, the ozone maintained a concentration with relevant biological activity (Guzzon et al. 2011) after 30 min. The effect of the presence of organic matter in water when ozone was dissolved is the second parameter which was considered. The reduction in the efficiency of sanitising agents based on oxidative mechanisms exerted by organic matter, and consequently the protective action on microbial forms, is generally known (Kitada 2010, Stanga 2010). Ozone does not escape this general trend (Figure 2). In our tests, the ozone concentration

Guzzon et al. Assay of antimicrobial activity of ozone 183 (a) 8,0 7,0 Ozone concentration (mg/l) 6,0 5,0 4,0 3,0 2,0 (b) Ozone concentration (mg/l) 1,0 0,0 5 4 3 2 1 0 5 10 15 20 Time (min) 25 30 35 40 Figure 1. Ozone degradation in sterile water. (a) Effect of temperature on the rate of ozone degradation at ( ) 7 C, (D) 12 C, (X) 17 C, ( ) 22 C and ( ) 27 C. (b) Linear correlation between ozone concentration and the temperature of aqueous solution after ( ) 10 min, ( ) 20 min and ( ) 30 min. 0 0 5 10 15 20 Temperature ( C) 25 30 was measured in water with an increasing dose (expressed as COD) of organic matter (non-active yeast). A COD equal to or greater than 9 mg/l led to the immediate degradation of ozone to a value without biological significance after just 5 min. In contrast, at a COD value between 1 and 0.01 mg/l, the ozone concentration remained around 2 mg/l after 20 min, guaranteeing potential antimicrobial activity. The use of cellular biomass with a nominal concentration of 10 11 CFU/g as a source of organic matter allowed consideration of the effectiveness of ozone as a sanitising treatment applied to the winery environment. In general, the concentration of microorganisms found on the surface of winery equipment (and therefore suspended in the water involved in the washing procedures) does not exceed 10 3 CFU/mL (Renouf et al. 2006, Guzzon et al. 2011). In the current tests, this cell density corresponds to an OD value of 1 mg/l, conditions under which ozone maintains a concentration with relevant antimicrobial impact for at least 20 min. The results of preliminary tests agreed with previous experience (Kim 1998, Khadre et al. 2001) confirming the dependence of ozone activity on environmental conditions, but at the same time demonstrating that under wine cellar conditions, ozone continues to be an effective biocide. Characterisation of potential spoilage by microorganisms in barrels During the winemaking process, a succession of species of yeast and bacteria is found in must and wine, the succession evolving according to the modification of the wine environment and particularly in response to the appearance of limiting factors, such as ethanol, sulfur dioxide and a lack of nutrients (Deliu et al. 2010). The microorganisms tested were chosen carefully, focusing on the risk of deterioration of wines during ageing in barrels or in contact with oak pieces because of microbial spoilage. Therefore, characteristics of the microorganisms investigated included adhesion and biofilm activity, production of volatile phenols and ethanol resistance (Renouf et al. 2006, 2007). Microbial resistance to ethanol is important given the significant content in wine, on average in the range between 12 and 16% v/v. Considering the porous nature of oak wood, the ability of microorganisms to adhere to the substrate, producing an exocellular matrix, a biofilm, represents a characteristic that favours microbial spoilage because of the resistance of the biofilm to cleaning treatments (Tristezza et al. 2010). The process leading to biofilm formation involves several steps. Of these, the adhesion of cells to the substrate through weak

184 Assay of antimicrobial activity of ozone Australian Journal of Grape and Wine Research 19, 180 188, 2013 14,0 12,0 Ozone concentration (mg/l) 10,0 8,0 6,0 4,0 2,0 Figure 2. Effect of chemical oxygen demand at ( ) 860, (D) 86, ( ) 9,( ) 1,( ) 0.1 and ( ) 0.01 mg/l on the rate of ozone degradation in sterile water. COD, chemical oxygen demand. 0,0 0 5 10 15 20 Time (min) 25 30 chemical bonds is the main limiting step (Joseph et al. 2007). Measurement of both parameters provides information about the kinetics of substrate colonisation by different microorganisms. Enzymatic conversion of cinnamic acids into vinyl or ethyl phenols is one of the most insidious phenomena involved in microbial wine spoilage, because it is not easily controllable with standard oenological tools (Chatonnet et al. 1995, Renouf et al. 2007). Fifty species of yeast, bacteria and mould were selected for the physiological tests (data not shown). Table 1 lists the data for the 20 microorganisms showing above-average results in at least two of the tests. Saccharomyces cerevisiae and Oenococcus oeni were considered as reference species because of their winemaking relevance. As expected, Brettanomyces bruxellensis showed significant resistance to ethanol and high decarboxylase activity, allowing the accumulation of a significant concentration of ethylphenols in the medium. The ability to produce a biofilm was also relevant, although comparison with adhesion activity confirmed the slow growth rate of this species. Candida genera showed both significant biofilm formation and adhesion activity, as reported by many authors, which have associated this yeast with microbial contamination in food-producing plants (Timke et al. 2008, Brugnoni et al. 2011); Candida ishiwadae confirmed its significant ability to produce vinylphenols (Guzzon et al. 2011). Some other species of yeasts, such as Hanseniaspora uvarum, Metschnikowia pulcherrima and Rhodutorula mucilaginosa, were able to rapidly adhere to the substrate, producing a biofilm. Their poor ethanol resistance, however, guaranteed a low tendency for spoilage during the ageing of wine. Cryptococcus rajasthanensis gave results of considerable concern in terms of sanitisation, with biofilm activity, good resistance to the ethanol concentration and a significant yield of vinylphenols. It has recently been isolated from wine (Brezna et al. 2010, Ocon et al. 2010), therefore careful control of the proliferation of this genera is required. Torulaspora delbrueckii is one of the few yeasts shown to be of oenological interest, because of the high ethanol resistance observed during the fermentation of highly concentrated musts (Bely et al. 2008). Acetobacter and Pediococcus showed marked ethanol resistance, combined, in the case of acetic bacteria, with biofilm activity. Sensitivity of selected microorganisms to ozone As described in the previous section, 20 species of microorganisms were subjected to the ozone test. For each of them, the residual cell concentration after an exposure of 30 min in an Table 2. Summary of results of ozone treatments. The ozone concentration causing complete cell death for each microorganism after 30 min exposure to ozone in aqueous solution. Initial cell concentration (CFU/mL) 10 6 10 5 10 4 10 3 10 2 Brettanomyces bruxellensis 5 1 1 1 1 Acetobacter aceti VC 5 5 2.5 1 Candida ishiwadae 7 5 2.5 1 1 Candida vini 5 2.5 1 1 1 Cryptococcus rajasthanensis 7 5 5 1 1 Gluconobacter oxidans 7 2.5 2.5 1 1 Hanseniaspora uvarum 7 2.5 1 1 1 Issatchenkia terricola VC 1 1 1 1 Lactobacillus planatarum 2.5 1 1 1 1 Metschnikowia pulcherrima VC VC VC 1 1 Oenococcus oeni 2.5 1 1 1 1 Pediococcus acidilattici 2.5 1 1 1 1 Pediococcus pentosaceus 5 2.5 1 1 1 Pichia fermentans 2.5 2.5 2.5 1 1 Pichia stiptis 2.5 2.5 1 1 1 Rhodotorula mucilaginosa VC 2.5 1 1 1 Saccharomyces cerevisiae 7 2.5 1 1 1 Schizosaccharomyces pombe 2.5 1 1 1 1 Torulaspora delbrueckii 7 5 1 1 1 Zygosaccharomyces bailii 7 7 2.5 2.5 1 Ozone concentration in mg/l causing complete cell death. VC, viable cells detected after O 3 treatment. aqueous O 3 solution was evaluated, varying the initial cell and/or ozone concentration. Table 2 summarises the results, indicating the ozone concentration necessary to obtain complete cell inactivation after 30 min, in relation to the initial cell concentration in the solution. At a higher cell concentration (10 6 CFU/mL, Figure 3), non-specific protective action was carried out by the abundant organic matter present in the solution. Indeed, in the majority of samples, 7 mg/l of O 3

Guzzon et al. Assay of antimicrobial activity of ozone 185 Reduction in cell concentration (log CFU/mL) 10,0 0,8 0,6 0,4 0,2 Candida ishiwadea Dekkera bruxellensis Metschnikowia pullcherima Rhodutorula mugilaginosa Candida vini Hanseniaspora uvarum Pichia fermentans Saccharomyces cerevisiae Cryptococcus rajastanensis Issatchenkia terricola Pichia stiptis Schizosaccharomyces pombe 0,0 0 Torulaspora delbrueckii Zygosaccharomyces bailii 2 4 6 8 Figure 3. Reduction in cell concentration (RC) of microbial cultures with an initial concentration of 10 6 CFU/mL in the presence of an increasing ozone concentration in the medium. Data were determined after 30 min contact between cells and ozone. RC was calculated as the ratio between the initial cell density and the residual cell density, both expressed as logarithmic units. Initial ozone concentration (mg/l) Reduction in cell concentration (log CFU/mL) 10,0 0,8 0,6 0,4 0,2 Candida ishiwadea Dekkera bruxellensis Metschnikowia pullcherima Rhodutorula mugilaginosa Candida vini Hanseniaspora uvarum Pichia fermentans Saccharomyces cerevisiae Cryptococcus rajastanensis Issatchenkia terricola Pichia stiptis Schizosaccharomyces pombe Torulaspora delbrueckii Zygosaccharomyces bailii 0,0 0 2 4 6 8 Figure 4. Reduction in cell concentration (RC) of microbial cultures with an initial concentration of 10 5 CFU/mL in the presence of an increasing ozone concentration in the medium. Data were determined after 30 min contact between cells and ozone. RC was calculated as the ratio between the initial cell density and the residual cell density, both expressed as logarithmic units. Initial ozone concentration (mg/l) was required to eliminate microorganisms and, despite the intensity of treatment, in some cases, residual viable cells were found. Interestingly, a correlation between ozone resistance and biofilm activity was found. Comparing the result of the ozone tests with those obtained in the biofilm tests, it is possible to observe that the microorganisms with above-average biofilm activity (Table 1) also required the highest O 3 concentration in order to be eliminated or, in four cases, maintained viable cells after treatment (Table 2). Two microorganisms, B. bruxellensis and Lactobacillus plantarum, did not follow this general trend. This apparent contradiction may be explained by considering the affinity of these microbes to anaerobic conditions (Yahara et al. 2007). Reducing the initial cell concentration, and consequently the organic matter content in the medium, the action of ozone was more uniform and 2.5 mg/l of ozone was generally sufficient to kill the microbiota present in the samples (Figure 4). These results agree with tests performed to validate ozone activity under wine cellar conditions, and have significant technological implications, because the concentration of spoilage microbial forms in the wine cellar is normally in this order of magnitude (Renouf et al. 2006, Guzzon et al. 2011). Over and beyond these general considerations, a brief description of the behaviour of species with the most technological relevance would appear necessary. Saccharomyces cerevisiae, widely employed in industrial fermentation, showed average resistance to the ozone treatment for yeasts, being inactivated by 1 mg/l of O 3 at a cell density below 10 5 CFU/mL. In contrast, O. oeni had poor O 3 resistance, being eliminated by 2.5 mg/l of O 3 at the highest cell concentration. This behaviour is probably because of anaerobic characteristics not guaranteeing tolerance of the oxidative stress caused by direct contact with more active forms of oxygen. The similar results obtained by bacteria belonging to the Lactobacillus and Pediococcus genera support this theory, and this hypothesis was also indirectly confirmed by the results obtained for the aerobic bacteria, Acetobacter aceti and Gluconobacter oxidans, whose ozone resistance also agrees with the intense biofilm formation observed in the specific test (Table 1). Of the spoilage yeasts, B. bruxellensis showed low ozone tolerance, requiring only 5 mg/l of O 3 for complete inactivation of a culture with a concentration of 10 6 CFU/mL, and was shown to be completely eliminated by 1 mg/l of O 3 in the other tests. This evidence confirms previous work by the same authors, which has suggested greater sensitivity to ozone by some spoilage yeasts typical of the wine environment, on the basis of a survey of the effect of ozone on barrel microflora (Guzzon et al. 2011).

186 Assay of antimicrobial activity of ozone Australian Journal of Grape and Wine Research 19, 180 188, 2013 Table 3. Phenolic composition of wines after contact with oak chips subjected to treatments with gaseous ozone at three levels of ozone application. Phenolic compounds (mg/l) 10th Control (n = 24) Treated (n = 24) Mann Whitney U-test result Median 90th 10th Median 90th 4-allylsyringol 0.01 0.22 0.83 0.01 0.16 0.92 n.s. 4-ethylphenol 0.01 0.01 0.49 0.01 0.01 0.37 n.s. 4-hydroxybenzoic acid 0.01 0.06 0.30 0.01 0.06 0.35 n.s. 4-methylcatecol 0.01 0.01 2.31 0.01 0.01 2.05 n.s. 4-methylsyringol 0.01 0.35 1.21 0.01 0.31 0.98 n.s. 4-vinylphenol 0.01 0.01 0.94 0.01 0.01 0.93 n.s. Acetosyringone 0.15 2.37 7.20 0.15 3.27 6.79 n.s. Coniferyl aldehyde 0.95 5.21 7.85 0.98 4.76 8.20 n.s. Cyclotene 0.01 0.28 1.18 0.01 0.10 0.79 n.s. Ellagic acid 3.37 6.30 10.20 3.55 7.03 13.02 n.s. Esculetin 0.01 0.01 0.45 0.01 0.01 0.38 n.s. Gentisic acid 0.11 0.37 0.42 0.06 0.25 0.39 P < 0.05 Guaiacol 0.08 0.20 0.30 0.05 0.10 0.31 n.s. Homovanillic acid 0.53 1.75 2.61 0.30 1.79 2.81 n.s. Homovanillic alcohol 0.01 2.10 3.73 0.01 1.95 3.68 n.s. Hydroxytyrosol 0.01 0.11 0.27 0.01 0.19 0.44 n.s. Isoeugenol 0.01 0.01 0.18 0.01 0.01 0.16 n.s. Isopropiosyringone 0.01 3.64 14.20 0.01 2.23 15.70 n.s. Iisopropiovanillone 1.84 5.64 10.99 1.88 6.15 10.10 n.s. Protocatechuic acid 0.56 0.87 1.65 0.55 0.92 1.34 n.s. Protocatechuic aldehyde 0.01 0.19 0.35 0.01 0.14 0.34 n.s. Scopoletin 0.01 0.13 0.59 0.01 0.11 0.51 n.s. Synapaldehyde 25.80 51.40 69.78 36.20 49.90 78.00 n.s. Syringaldehyde 3.84 9.63 18.45 2.44 8.72 13.95 n.s. Syringic acid 4.33 9.44 21.14 2.95 10.03 20.32 n.s. Syringol 0.01 0.38 1.94 0.01 0.32 2.32 n.s. Tryptophol 1.15 1.64 2.50 1.06 1.73 3.38 n.s. Tyrosol 4.45 7.55 21.50 3.70 6.90 8.60 n.s. Vanillic acid 4.64 7.30 11.65 3.87 8.04 11.31 n.s. Vanillin 2.41 5.25 10.90 2.50 5.16 10.38 n.s. n.s., not statistically significant. Impact of ozone sanitisation treatment of oak chips on the phenolic composition of wines Barrels are a controversial tool in wine production, they participate in two phenomena: microoxygenation of wines because of the porous nature of oak wood and, no less importantly, complex patterns of reactions occurring between the phenolic compounds of the wood and the wine during ageing (Singleton 1995). For these reasons, we carefully investigated, according to two experimental protocols, the effect of ozone sanitisation of oak used in winemaking on the final content of 29 simple phenols and cyclotene in wine. The first approach exposed oak chips to O 3 for a fixed time (T1) and to repeated O 3 treatments (T2). No statistical difference was found, using either nonparametric (analysis of variance (ANOVA) by ranks, Kruskal Wallis test) or parametric approaches (ANOVA-Tukey test), for the concentration of the 29 simple phenols and cyclotene in wine made with the addition of untreated (T0) and treated chips (T1 and T2) (Table 3). The second approach submitted the oak chips to more intensive treatment, exposing them to a continuous flow of O 3 for an extended length of time. Table 4 shows a comparison between wines made with the addition of untreated (control) and treated chips. A statistically significant difference was observed only for gentisic acid using both non-parametric (Mann Whitney U-test) and parametric approaches (ANOVA, Tukey test). Gentisic acid is well known for its pharmaceutical uses because of its antioxidant and free radical scavenging activity (Joshi et al. 2012), and we found that it was oxidisable by our coulometric detector under 100 mv, more easily than all the other analytes studied, with the exception of gallic acid (100 mv). While these last severe treatment conditions were able to destroy roughly one third of the initial median content of gentisic acid, it should

Guzzon et al. Assay of antimicrobial activity of ozone 187 Table 4. Phenolic composition of wines after contact with oak chips that had been subjected to 40 min continuous gaseous ozone treatment (ozone concentration 20 mg/l). Phenolic compounds (mg/l) 10th T0 (n = 18) Median 90th 10th T1 (n = 18) Median 90th 10th T2 (n = 18) Median 90th Kruskal Wallis ANOVA by ranks sign. 4-allylsyringol 0.01 0.20 0.90 0.01 0.15 0.92 0.01 0.19 0.99 n.s. 4-ethylphenol 0.01 0.01 0.27 0.01 0.01 0.25 0.01 0.01 0.26 n.s. 4-hydroxybenzoic 0.01 0.14 1.33 0.01 0.21 0.96 0.01 0.17 1.55 n.s. acid 4-methylcatecol 0.01 0.01 1.27 0.01 0.01 1.82 0.01 0.01 2.06 n.s. 4-methylsyringol 0.01 0.43 1.46 0.01 0.43 1.37 0.01 0.46 1.51 n.s. 4-vinylphenol 0.01 0.01 0.95 0.01 0.01 0.81 0.01 0.01 0.83 n.s. Acetosyringone 0.10 2.68 5.05 0.10 3.11 6.80 0.10 3.52 7.20 n.s. Coniferyl aldehyde 2.02 5.16 8.38 1.62 4.83 8.84 1.78 5.80 7.10 n.s. Cyclotene 0.01 0.18 0.50 0.01 0.11 0.54 0.01 0.15 0.71 n.s. Ellagic acid 3.27 5.65 9.98 3.64 5.68 10.90 3.55 5.69 10.61 n.s. Esculetin 0.01 0.01 0.44 0.01 0.01 0.39 0.01 0.01 0.42 n.s. Gentisic acid 0.15 0.29 0.42 0.11 0.27 0.43 0.11 0.28 0.38 n.s. Guaiacol 0.05 0.11 0.26 0.06 0.15 0.29 0.07 0.11 0.33 n.s. Homovanillic 0.51 1.56 2.29 1.19 1.60 2.16 0.52 1.72 2.45 n.s. acid Homovanillic 0.01 1.82 3.31 0.01 1.65 3.32 0.01 1.71 3.83 n.s. alcohol Hydroxytyrosol 0.01 0.31 0.58 0.01 0.30 0.60 0.01 0.27 0.59 n.s. Isoeugenol 0.01 0.01 0.13 0.01 0.01 0.13 0.01 0.01 0.14 n.s. Isopropiosyringone 0.01 3.63 14.10 0.01 4.50 13.60 0.01 3.33 15.72 n.s. Isopropiovanillone 1.70 4.89 9.70 2.03 5.85 12.54 1.88 4.93 16.90 n.s. Protocatechuic 0.49 0.75 2.40 0.47 0.96 2.48 0.48 0.84 1.28 n.s. acid Protocatechuic 0.01 0.21 0.41 0.01 0.23 0.45 0.01 0.21 0.45 n.s. aldehyde Scopoletin 0.01 0.05 0.44 0.01 0.15 0.64 0.01 0.06 0.51 n.s. Synapaldehyde 38.90 49.70 75.16 40.00 48.62 70.38 29.30 48.59 76.04 n.s. Syringaldehyde 2.76 9.08 15.60 3.33 10.67 17.90 2.39 9.90 19.40 n.s. Syringic acid 1.37 8.35 20.04 1.91 8.73 17.63 1.85 9.39 22.01 n.s. Syringol 0.01 0.31 2.35 0.01 0.35 2.41 0.01 0.29 2.25 n.s. Tryptophol 1.10 1.53 2.05 1.10 1.57 1.86 1.21 1.61 2.60 n.s. Tyrosol 3.40 3.90 12.13 3.30 4.00 12.25 3.25 4.04 11.14 n.s. Vanillic acid 0.74 6.15 9.83 4.15 6.95 8.69 1.07 7.19 9.61 n.s. Vanillin 2.23 5.55 9.00 2.28 5.44 10.05 2.52 5.98 11.17 n.s. ANOVA, analysis of variance; n.s., not statistically significant. be underlined that it is generally present in wine only in trace amounts. (Ribéreau-Gayon et al. 2006). Conclusion The future of the wine industry is now clearly directed towards the production of wines with a low ecological footprint. In this context, the implementation of technological solutions able to prevent microbial spoilage without side-effects on the environment or on the healthfulness of wine would appear to be essential. At the same time, the increasing competence of customers has forced winemakers to ensure a high standard of wine production. The use of ozone for the prevention of microbial spoilage would appear to meet both requirements. The physiological tests performed in the first part of this work demonstrate the potential spoilage activity of a large number of microorganisms of oenological significance. Of the large number of yeast and bacteria tested, 20 were shown to be of particular concern for wine quality, although the mechanism of ozone action varies between species. Considering the specific intention of this work, the control of microbial spoilage during wine ageing in barrels or in contact with oak pieces, the formation of biofilm and the production of volatile phenols were the most serious aspects

188 Assay of antimicrobial activity of ozone Australian Journal of Grape and Wine Research 19, 180 188, 2013 characterising a large number of other yeasts belonging to the Candida and Cryptococcus genera, as well as to Brettanomyces/ Dekkera spp. Aqueous ozone solution was shown to be effective in eliminating viable cells present in solution, although its activity would appear to be strongly related to environmental variables, such as temperature, COD and cell concentration. Nevertheless, analysing the results of the tests carried out, it can be stated that the environmental conditions typical of a winery do not reduce significantly the ozone effectiveness. The use of ozone for the sanitisation of oenological wood materials under reasonable treatment conditions showed no significant alteration in terms of simple phenols in wine and only long exposure caused a reduction of 30% in gentisic acid, one of the most effective antioxidants naturally present in wine, although at low concentration. References Bartowsky, E.J. (2009) Bacterial spoilage of wine and approaches to minimize it. Letters in Applied Microbiology 48, 149 156. Bely, M., Stoeckle, P., Masneuf-Pomarede, I. and Dubourdieu, D. (2008) Impact of mixed Torulaspora delbrueckii-saccharomyces cerevisiae culture on high-sugar fermentation. International Journal of Food Microbiology 122, 312 320. Brezna, B., Zenisova, K., Chovanova, K., Chebenova, V., Krakova, L., Kuchta, T. and Pangallo, D. (2010) Evaluation of fungal and yeast diversity in Slovakian wine-related microbial communities. Antonie van Leeuwenhoek International Journal of General and Molecular Microbiology 98, 519 529. Brugnoni, L.I., Cubitto, M.A. and Lozano, J.E. (2011) Biofilm formation under laminar flow conditions of yeast isolated from an apple juice processing plant. Journal of Food Process Engineering 34, 49 66. Chatonnet, P., Dubourdieu, D. and Boidron, J.N. (1995) The influence of Brettanomyces/Dekkera sp. yeasts and lactic acid bacteria on the ethylphenol content of red wines. American Journal of Enology and Viticulture 46, 463 468. Coggan, M. (2003) Ozone in wineries part 1 getting beyond myths and mistakes. Vineyard & Winery Management 29, 1 13. Deliu, I., Brinduse, E., Giosanu, D. and Aurelia, T. (2010) Researches concerning the evolution of wine microbiota during the spontaneous fermentation of red grapes juices. Annals Food Science and Technology 11, 103 108. Erickson, M.C. and Ortega, Y.R. (2006) Inactivation of protozoan parasites in food, water, and environmental systems. Journal of Food Protection 69, 2786 2808. Foegeding, P.M. (1985) Ozone inactivation of bacillus and clostridium spore populations and the importance of the spore coat to resistance. Food Microbiology 2, 123 134. Guillen, A.C., Kechinski, C.P. and Manfroi, V. (2010) The use of ozone in a CIP system in the wine industry. Ozone: Science & Engineering 32, 355 360. Guzzo, F., Cappello, M.S., Azzolini, M., Tosi, E. and Zapparoli, G. (2011) The inhibitory effects of wine phenolics on lysozyme activity against lactic acid bacteria. International Journal of Food Microbiology 148, 184 190. Guzzon, R., Widmann, G., Malacarne, M., Nardin, T., Nicolini, G. and Larcher, R. (2011) Survey of the yeast population inside wine barrels and the effects of certain techniques in preventing microbiological spoilage. European Food Research and Technology 233, 285 291. Haverkamp, R.G., Miller, B.B. and Free, K.W. (2002) Ozone production in a high frequency dielectric barrier discharge generator. Ozone Science & Engineering 24, 321 328. Hester, T. (2006) New Brett solution arrives in Australia. Australian and New Zealand Grape Grower and Winemaker 512, 77 79. Jin-Gab, K., Ahmed, Y.E. and Mohammed, K.A. (2003) Ozone and its current and future application in the food industry. Advances in Food and Nutrition Research 45, 167 218. Joseph, C.M.L., Kumar, G., Su, E. and Bisson, L.F. (2007) Adhesion and biofilm production by wine isolates of Brettanomyces bruxellensis. American Journal of Enology and Viticulture 58, 373 378. Joshi, R., Gangabhagirathi, R., Venu, S., Adhikari, S. and Mukherjee, T. (2012) Antioxidant activity and free radical scavenging reactions of gentisic acid: in-vitro and pulse radiolysis studies. Free Radical Research 46, 11 20. Khadre, M.A., Yousef, A.E. and Kim, J.G. (2001) Microbiological aspects of ozone applications in food: a review. Journal of Food Science 66, 1242 1252. Kim, J.G. (1998) Ozone as an antimicrobial agent in minimally processed foods. DPhil Thesis, Ohio State University, Columbus, OH, USA:, p 50 199. Kitada, K. (2010) Introduction to CIP (clean-in-place) washing technology. Journal of Antibacterial and Antifungal Agents 38, 843 848. Komanapalli, I.R. and Lau, B.H.S. (1998) Ozone-induced damage of Escherichia coli K-12. Applied and Environmental Biotechnology 46, 610 614. Kurtzman, C.P. and Robnett, C.J. (1998) Identification and phylogeny of ascomycetous yeasts from analysis of nuclear large subunit (26S) ribosomal DNA partial sequences. Antonie Van Leeuwenhoek International Journal of General and Molecular Microbiology 73, 331 371. Larcher, R., Nicolini, G., Bertoldi, D. and Nardin, T. (2007) Determination of 4-ethylcatechol in wine by high-performance liquid chromatographycoulometric electrochemical array detection. Analytica Chimica Acta 609, 235 240. Ocon, E., Gutierrez, A.R., Garijo, P., Lopez, R. and Santamaria, P. (2010) Presence of non-saccharomyces yeasts in cellar equipment and grape juice during harvest time. Food Microbiology 27, 1023 1027. Organisation Internationale de la Vigne et du Vin (2011) Recueil des méthodes internationales d analyse des vins et des moûts (Organisation International de la Vigne e du Vin: Paris, France). Renouf, V., Claisse, O., Miot-Sertier, A., Perello, M.C., De Revel, G. and Lonvaud-Funel, A. (2006) Study of the microbial ecosystem present on the barrels surface used during the winemaking. Sciences des Aliments 26, 427 445. Renouf, V., Lonvaud-Funel, A. and Coulon, J. (2007) The origin of Brettanomyces bruxellensis in wines: a review. Journal International des Sciences de la Vigne et du Vin 41, 161 173. Ribéreau-Gayon, P., Glories, Y., Maujean, A. and Dubourdieu, D. (2006) Handbook of enology. Volume 2: the chemistry of wine and stabilization and treatments, 2nd edn (John Wiley & Sons: New York). Singleton, V.L. (1995) Maturation of wines and spirits comparisons, facts, and hypotheses. American Journal of Enology and Viticulture 46, 98 115. Stanga, M. (2010) Sanitation cleaning and disinfection in the food industry (Wiley VCH: Weinheim). Timke, M., Wang-Lieu, N.Q., Altendorf, K. and Lipski, A. (2008) Identity, beer spoiling and biofilm forming potential of yeasts from beer bottling plant associated biofilms. Antonie van Leeuwenhoek International Journal of General and Molecular Microbiology 93, 151 161. Tristezza, M., Lourenco, A., Barata, A., Brito, L., Malfeito-Ferreira, M. and Loureiro, V. (2010) Susceptibility of wine spoilage yeasts and bacteria in the planktonic state and in biofilms to disinfectants. Annals of Microbiology 60, 549 556. Usseglio-Tomasset, L. (1992) Properties and use of sulphur-dioxide. Food Additives and Contaminants 9, 399 404. Yahara, G.A., Javier, M.A., Tulio, M.J.M., Javier, G.J. and Guadalupe, A.U.M. (2007) Modeling of yeast Brettanomyces bruxellensis growth at different acetic acid concentrations under aerobic and anaerobic conditions. Bioprocess and Biosystems Engineering 30, 389 395. Manuscript received: 16 July 2012 Revised manuscript received: 10 September 2012 Accepted: 24 November 2012