Resistance of Yeast Species to Benzoic and Sorbic Acids and to Sulfur Dioxide

5 Journal of Food Protection, Vol. 8, No. 7, Pages 5-5 (July 85) Copyright International Association of Milk, Food, and Environmental Sanitarians Resistance of Yeast to Benzoic and Sorbic Acids and to Sulfur Dioxide ALAN D. WARTH CSIRO Division of Food Research, P.O. Box 5, North Ryde, NSW, Australia (Received for publication November 5, 8) ABSTRACT Effects of sorbic and benzoic s and S on the growth and survival of strains of yeasts, differing widely in their preservative resistance, were studied. Exponential phase cultures not adapted to preservative were tested under anaerobic conditions at ph.5. In general, species tolerant of one preservative were also tolerant of the others, but significant differences in the relative effectiveness of the preservatives were found in some species. Maximum tolerated levels of benzoic ranged from.5 mm for Hansenula anomala to mm for Zygosaccharomyces bailii. The range in tolerance to S was.5 mm free S for Klockera apiculata to.8 mm for Z. bailii. The principal effect of sorbic and benzoic s was to reduce cell yield. At higher concentrations, growth rates and lag times were affected. Benzoic generally inhibited growth less than sorbic, but had a greater effect on lag time and so had a similar overall degree of effectiveness. Cultures treated with S characteristically showed long lag times of up to h. Reductions in growth rate and final yield were often not apparent, mainly because S became bound by metabolic products. Spoilage of foods and beverages containing preservatives is caused by relatively few species of yeast (,5,). It is apparent that resistance to one of the common preservatives is often associated with resistance to one or more of the others. In particular, species that are resistant to benzoic are often resistant to sorbic and appear to tolerate acetic also. There is very little published information on the sensitivity to S of benzoic and sorbic -resistant yeasts, but as spoilage occurs even when both types of preservative are used together, a common resistance must occur. Much research has been done on the mechanism of action of these three preservatives individually (5,,), but few studies have directly considered their common effects (e.g., ). Since these preservatives are all weak s with a lipophilic form, they have predictable common effects on cells in media. Cells generally are very permeable to the undissociated form of preservatives. If the cell is impermeable to the preservative anion, then at equilibrium, the ratio between the internal and external concentrations of anion will be in inverse proportion to the hydrogen ion ratio, e.g., a concentration of mm preservative anion in a medium of ph.5 would equilibrate with M anion in cytoplasm at ph.5. Alternatively, if anions leave the cell either by leakage or transport, there will be a net flow of protons into the cell which will either equilibrate the cytoplasmic ph with that of the medium, or impose a heavy energy load on the cell in expelling protons. Preservatives have important effects on cell yield (,,), ATP levels (8), transport (7,8) and cytoplasmic ph (), which are all explainable in terms of the stresses resulting from preservative uptake and loss from the cell. In addition to these non-specific effects, preservatives may have specific actions on the cell which enhance their effectiveness. Bisulfite reacts with various cell constituents including, for example, acetaldehyde, thus blocking the regeneration of NAD required for glycolysis, and the organic preservatives can react with CoA (). This study examines the effects of the three common preservatives, sorbic, benzoic and S, on the growth parameters of a variety of yeasts. Differences in effects between preservatives indicate either a specific action of that preservative or a specific resistance mechanism to that preservative. On the other hand, similarities in the resistance to all preservatives suggest that common mechanisms of action and of cell resistance are important. A number of yeast strains having different resistances to preservatives have been included in this study to provide data for further studies on the mechanism of action of preservatives, and on the mechanism of resistance in the tolerant species. The species chosen include common wild yeasts, baker's and wine strains of Saccharomyces cerevisiae and several preservative resistant species isolated from spoiled foods and beverages containing preservatives. MATERIALS A METHODS Yeasts Zygosaccharomyces bailii (Saccharomyces bailii) FRR 7, isolated from a winery, was obtained from R. R. Davenport JOURNAL OF FOOD PROTECTION, VOL. 8, JULY 85

RESISTANCE OF YEASTS TO PRESERVATIVES 55 (strain BS), Long Ashton Research Station, England, and all of the others were obtained from A. D. Hocking, CSIRO Division of Food Research. S. cerevisiae FRR 7 is a baker's yeast; FRR 8 was from a winery. Z. bailii FRR and FRR 7 and Schizosaccharomyces pombe were isolated from spoiled beverages containing benzoic. Saccharomycodes ludwigii was from apple cider. FRR denotes the culture collection of the CSIRO Division of Food Research, North Ryde. Culture conditions Yeasts were grown in 5 ml of fructose yeast extract medium (FYE) at 5 C in X -mm screw cap tubes fitted with butyl rubber septa, and shaken at. Hz. Air was displaced by bubbling N at ml/min for min. Fructose was used in preference to glucose as the fermentable substrate because it binds S less strongly, and for Z. bailii it is the preferred substrate (). FYE medium contained, per L: 5 g fructose; g yeast extract (Oxoid); g (NH ) S ; g KH P ; 5 mg CaCl ; 5 mg MgCl ; ml lipid ( mg/ml ergosterol and mg/ml oleic in ethanol); and ml.5 M potassium citrate buffer. The buffer was adjusted to ph.5 at mm and 5 C using an Activon research grade ph meter. The ph of sterile medium was.5 ±.5 ph units. Preservative resistance Tubes were inoculated with early exponential phase cells growing anaerobically in FYE medium without added preservative to give a calculated initial absorbance of. ( cells). After min at 5 C, preservative was added. Absorbance at nm was read directly using a Perkin Elmer model 55 spectrophotometer, and the result corrected for non-linearity with a cubic equation. Cultures that had not shown visible growth by h, were tested for viability on malt agar and found to have less than colony-forming unit per ml. Sorbic and benzoic s were assayed at the conclusion of the experiment. In all cases, recovery was 8 to %. Sulfur dioxide concentrations were measured after d and then periodically until growth was evident. Benzoic and sorbic levels are reported as the amount added and S as the analytical value after d. Ethanol production was measured in chemostats at a specific growth rate of.5 h "'. Chemical assays Cultures containing sorbic or benzoic s were diluted - fold in.5 M sodium acetate, ph 5.5, heated at C for 5 min, and centrifuged. Preservative concentration was determined on - xl samples by HPLC using an RP8 (Merck) column, eluted with methanol.5 M sodium acetate, ph., (:, vol/vol), with detection at 55 or 5 nm. S concentration was determined by gas chromatography on ml of the headspace gas drawn directly from the culture tubes (). This analysis gives a true measure of the molecular S concentration in the medium. Free S (HS ~ S ) concentrations were calculated using the ph and a pk a of.8. Ethanol was determined by gas chromatography, using Graphpac GC8/ at 5 C, and % isopropanol as the internal standard. Carbohydrate remaining after completion of growth was determined on culture supernatant fluids by the anthrone method () using fructose as the standard. Carbon dioxide evolution rates were measured using Warburg manometers. RESULTS Growth parameters under anaerobic conditions were determined for each yeast at several concentrations of preservative. Exponentially growing cells from preservative-free media were used as inocula. The common yeast species chosen as controls were all sensitive to preservatives. Klockera apiculata, Hansenula anomala, Kluveromyces fragilis and the baker's strain of S. cerevisiae tolerated.5 to.5 mm sorbic or benzoic s, and showed no growth at mm (Tables and ), whereas the yeasts isolated from preservatized food tolerated levels of to mm sorbic and benzoic s. The wine strain of S. cerevisiae was more tolerant than the baker's strain, especially to S, but was not as resistant as the spoilage yeasts. For each species, there were no major differences between the levels of sorbic and benzoic s tolerated. isolated from products containing benzoic were also resistant to sorbic. Resistance to SO was closely associated with resistance to the organic preservatives. However, Candida krusei was sensitive to SO but relatively tolerant to benzoic and sorbic s, and Z. bailii FRR 7 was as tolerant of S as the other Z. bailii strains, but did not share their very high tolerance of organic preservatives. TABLE. Minimum concentrations of preservative preventing anaerobic growth at ph.5 and 5 C within h. ] FRR No. Klockera apiculata Hansenula anomala Kluveromyces fragilis Saccharomyces cerevisiae S. cerevisiae Candida krusei Saccharomycodes ludwigii Schizosaccharomyces pombe Zygosaccharomyces bailii Z. bailii Z. bailii a, value not determined 8 8 7 8 555 55 7 7 Sorbic..5.5 Benzoic a Free S..8.......5. TABLE. Maximum concentrations of preservative tested permitting anaerobic growth at ph.5. S. cerevisiae 7 5. cerevisiae 8 Z. bailii 7 Z. bailii Z. bailii 7 Sorbic Benzoic.5.5.7 Free S.5.8. <..5.8...8..8 JOURNAL OF FOOD PROTECTION, VOL. 8, JULY 85

5 WARTH Inhibition characteristics of sorbic and benzoic s Representative data are shown in Tables and and Figure. Relatively low amounts of preservative generally had a small (<5%) effect on the growth rate. Fermentation rates (ethanol production), however, were stimulated considerably (Table ) at these preservative levels, but were inhibited at higher concentrations (data not shown). Lag times were often long, especially at preservative concentrations near the maximum tolerated, which is consistent with cell death or very extended adaptation periods. A notable effect of these preservatives was to reduce the cell yield. With the exceptions of K. apiculata and Z. bailii FRR 7, the carbohydrate substrate was fully utilized at maximum growth. Evidently, TABLE. Inhibitory responses to sorbic. Concentration Growth iate a S. cerevisiae 7 S. cerevisiae 8 S. cerevisiae 8 Z. bailii 7 Z. bailii Z. bailii Z. bailii 7 Z. bailii 7.5 7 7 57 8 5 8 8 Lag time (h) 58 5 7 5 7 5 "Percentage of value in the absence of preservative. Yield" 5 5 5 7 8 nd 8 5 Time (h). Figure. Growth curves for Z. bailii FRR 7. Anaerobic growth at ph.5 and 5 C in media containing 5% fructose. The inoculation level at zero time was equivalent to. absorbance. O, no preservative; ^, mm sorbic ;. mm sorbic ; A, mm benzoic ;, mm benzoic. in these two species, growth is limited by other factors, therefore, maximum growth does not quantitatively indicate growth yield. However, growth yield was clearly affected by sorbic and benzoic s because these compounds reduced both the maximum growth level and the amount of residual carbohydrate. TABLE. Inhibitory responses to benzoic. Concentration S. cerevisiae 7 S. cerevisiae 8 S. cerevisiae 8 bailii 7 bailii bailii bailii 7 bailii 7.5.7 Growth rate" 8 8 5 5 7 87 7 Lag time (h) "Percentage of value in the absence of preservative. b Rate of ethanol production per mg cells as percentage of value in the absence of benzoic. C, value not determined. 8 7 5 8 Yield" 7 5 7 7 5 55 57 55 85 5 5 Ethanol production b 7 5 57 5 C 87 55 7 7 JOURNAL OF FOOD PROTECTION. VOL. 8, JULY 85

RESISTANCE OF YEASTS TO PRESERVATIVES 57 Reduction in yield was correlated with inhibition of growth rate (correlation coefficient =.78, N = ), but was unrelated to lag time. Differences between sorbic and benzoic s Although the general level of effectiveness was similar in all cases, some species showed significant differences in their responses to sorbic and benzoic s (Table 5, Fig. ). In all strains and at each concentration tested, sorbic inhibited growth as much as or more than benzoic did. This was most evident in. On the other hand, benzoic most often had a greater effect on lag times. The apparent exceptions were all in species for which sorbic was appreciably more effective in inhibiting the growth rate. Effect of sulfur dioxide on growth parameters In contrast to sorbic and benzoic s, S did not greatly inhibit the measured growth rates, and had almost no effect on cell yields, but produced greatly extended lag times (Table, Fig. ). was exceptional in showing short lag times. Free S levels typically remained constant until growth was visually apparent, when they rapidly diminished. Where S was determined at the time of visible cell growth, levels were maintained up to an absorbance of. ( x 5 cells/ml), were half depleted near. absorbance, and were very low thereafter. An example is shown in Figure. In most cases, growth curves in the S cultures were linear, not showing an increase in growth rate corresponding to the exhaustion of free S. DISCUSSION The results show that yeast species that are resistant to one weak type of preservative tend also to be resistant to the others. There were clear differences in the responses to individual preservatives, but these differ- TABLE. Inhibition characteristics for S Concentration Growth Lag rate" time (VM) (h) S. cerevisiae 8 Z. bailii 7 Z. bailii Z. bailii Z. bailii 7 Z. bailii 7 c.8 7 5. 7.5 7 8 8 8 8 7 7 7 7 8 8 5 "Percentage of value in absence of preservative. b, value not determined. c After depletion of free S. 7 8 5 57 5 7-58 Yield 8 75 b TABLE 5. Differences in the effects of sorbic and benzoic s on growth parameters. S. cerevisiae 7 S. cerevisiae 8 Z. bailii 7 Z. bailii Z. bailii 7 Rate inhibition a Lag time Yield reduction ",,, degree to which sorbic had a greater effect than benzoic ;, little difference between sorbic and benzoic s, -,,, degree to which sorbic hada lesser effect than benzoic. The differences were apparent at all concentrations tested. Time (h) 8 Figure. Growth curves for Z. bailii FRR 7. Anaerobic growth at ph.5 and 5 C. Free S (S HS ): O, none;.. mm; A,.8 mm; Q.5 mm. ences were relatively small. This suggests that the weak preservatives have a common mechanism of action, and that the factors which enable some species to tolerate high levels of preservative apply to all three preservatives. Earlier studies (,,7) support the view that across the full range of yeast species, resistance to the major preservatives is highly correlated, but that in groups of yeasts already selected for preservative resistance, significant differences in the effectiveness of individual preservatives can be demonstrated. JOURNAL OF FOOD PROTECTION, VOL. 8, JULY 85

58 WARTH Time (h) Figure. Molecular sulfur dioxide levels during growth of Z. bailii FRR. Anaerobic growth at ph.5 and 5 C in media containing 5% fructose., molecular S ; O, absorbance. Sorbic was generally more effective than benzoic in inhibiting growth rate and cell yield, and therefore may be a more effective preservative in products subject to recontamination and where the inoculum is adapted to growth in preservative. Benzoic and S tended to cause longer apparent lag times, implying that they are more toxic initially, either by causing cell death or by inhibiting the adaptation to growth in the presence of preservatives. Lag times determined from growth kinetics after growth was visually apparent, do not distinguish between cell death and a true lag phase. Viable counts on several cultures treated with benzoic or sorbic s showed low initial losses of <% (Warth, unpublished). Furthermore, several of the lag times were greater than would result from the survival of one organism. Restaino et al. () showed that sorbic killed yeasts very slowly. Benzoic could be more effective than sorbic in situations where the stability of the product depends upon the eventual death rather than inhibition of the yeast. Growth in the presence of several weak s induced the formation of a transport system for preservative anions in Z. bailii () and in most of the other yeast species (Warth, unpublished). Operation of this system, which allows the cell to avoid accumulation of large amounts of preservative, probably is very important in the cell's tolerance to a preservative. Adapted cells could therefore have shorter lag times and show growth at higher preservative levels than cultures not induced for the anion transport system. Bills et al. () found that S. rouxii could adapt to tolerate higher levels of sorbic. In this study, preservative was not added to the inocula to adapt the cells, but production of volatile fatty s, which is common in yeast cultures, could partly induce the system. The present results were obtained under anaerobic conditions, using a 5% carbohydrate substrate. It is likely that slightly greater tolerances would be found for aerobic conditions and for species-optimized carbohydrate levels. Pitt () found increased resistance in several species at % glucose compared with.5% glucose. However, further experiments (Warth, unpublished) have shown that the upper limits of benzoic for growth of adapted cultures and for semi-aerobic cultures were not significantly above the values in Table. A notable characteristic of the action of the organic preservatives was the marked reduction in cell yield. Although a reduction in final cell numbers has been noted previously (,,), its importance and generality has not been recognized. Reduction in cell growth yield is consistent with the operation of the preservative anion transport system and with simple leakage of preservative anions. In both cases, the cell must expend energy to export protons in order to maintain its internal ph (). This phenomenon is of some significance industrially, where a small number of yeast cells may produce large amounts of gas. The apparent lack of effect on S on growth rate and yield is largely because of production of S -binding compounds by the yeasts. The major part of growth occurred after depletion of free S and the lack of a major effect on cell yield and on the late growth rate may therefore be expected. Many growth curves did not show an increase corresponding to depletion of free S. This result suggests that once adapted to growth in S, these cells were not greatly inhibited by its presence. Some growth curves for Z. bailii did show evidence of an increase in growth rate corresponding to the depletion of S (Fig. ). The examples where growth rate remained inhibited after the depletion of free S indicate that some product of S can be toxic to yeasts, or some nutrient depleted by S. Headspace gas chromatography enables monitoring of free S levels while the cultures are still growing. Values obtained are thermodynamic values for molecular S, and should be preferred for mechanistic studies over those obtained by the traditional methods involving ification and distillation, which may be less specific. In contrast to the organic preservatives, S is reactive chemically, and there are several ways in which bisulfite specifically could inhibit metabolism (,). Nevertheless, S also behaves as a weak lipophilic, and in an environment the cell must accumulate bisulfite, or lower its internal ph, or expend energy as with the other preservatives. The relative importance of the various reactions of S in the killing and inhibition of yeast cells remains to be evaluated. It is very likely that the stresses of anion accumulation and ph maintainance are important, and that physiological adaptations that allow cells to cope with these stresses will enhance their tolerance both to the organic preservatives and to S. REFERENCES. Baird-Parker, A. C., and W. J. Kooiman. 8. Soft dinks, fruit juices, concentrates, and fruit preserves, pp. -8. In Microbial ecology of foods, vol.. Academic Press, New York. JOURNAL OF FOOD PROTECTION, VOL. 8, JULY 85

RESISTANCE OF YEASTS TO PRESERVATIVES 5. Balatsouras, G. D., and N. G. Polymenacos.. Chemical preservatives as inhibitors of yeast growth. J. Food Sci. 8:7-75.. Bell, T. A., J. L. Etchells, and A. F. Borg. 5. Influence of sorbic on the growth of certain species of bacteria, yeasts and filamentous fungi. J. Bacteriol. 77:57-58.. Bills, S., L. Restaino, and L. M. Lenovich. 8. Growth response of an osmotolerant sorbate-resistant yeast, Saccharomyces rouxii, at different sucrose and sorbate levels. J. Food Prot. 5:-. 5. Chipley, J. R. 8. Sodium benzoate and benzoic. pp. - 5. In A. L. Branen and P. M. Davidson (eds.) Antimicrobials in foods. Marcel Dekker, New York.. Davis, E. G., D. Barnett, and P. M. Moy. 8. Determination of molecular and free sulphur dioxide in foods by headspace gas chromatography. J. Food Technol. 8:-. 7. Deak, T., and E. K. Novak. 7. Effect of sorbic on the carbohydrate fermentation and oligosaccharide splitting enzymes of yeast. Acta Aliment. :7-8. 8. Freese, E., C. W. Sheu, and E. Galliers. 7. Function of lipophilic s as antimicrobial food additives. Nature :- 5.. Hammond, S. M., and J. G. Carr. 7. The antimicrobial activity of SO - with particular reference to fermented and non-fermented fruit juices, pp. 8-. In F. A. Skinner and W. B. Hugo (eds.) Inhibition and inactivation of vegetative microbes. Academic Press, London.. Harada, K., R. Higuchi, and I. Utsumi. 8. Studies on sorbic, part. Inhibition of the respiration in yeast. Agr. Biol. Chem. :-.. Herbut, D., P. J. Phipps, and R. E. Strange. 7. Chemical analysis of microbial cells, pp. -. In J. R. Norris and D. W. Ribbons (eds.) Methods in microbiology, vol. 5. Academic Press, London.. Krebs, H. A., D. Wiggins, M. Stubbs, A. Sols, and F. Bedoya. 8. Studies on the mechanism of the antifungal action of benzoate. Biochem. J. :57-.. Oka, S.. Mechanism of antimicrobial effect of various food additives, pp. -. In N. Molin (ed.) Microbial inhibitors in food. Almquist and Wiksell, Stockholm.. Pitt, J. I. 7. Resistance of some food spoilage yeasts to preservatives. Food Technol. Aust. :8-. 5. Pitt, J. I., and K. C. Richardson. 7. Spoilage by preservativeresistant yeasts. CSIRO Food Res. Q. :8-85.. Restaino, L., L. M. Lenovich, and S. Bills. 8. Effect of s and sorbate concentrations on the growth of four osmophilic yeasts. J. Food Prot. 5:8-. 7. Sand, F. E. M. J.. Into virgin territory of soft drink microbiology. Proc. Soc. Soft Drink Technol. :-7. 8. Schmiz, K.-L. 8. The effect of sulphite on the yeast Saccharomyces cerevisiae. Arch. Microbiol. 5:8-5.. Smittle, R. B., and R. S. Flowers. 8. Acid tolerant microorganisms involved in the spoilage of salad dressings. J. Food Prot. 5:77-8.. Sofos, J. N., and F. F. Busta. 8. Antimicrobial activity of sorbate. J. Food Prot. :-.. Sols, A., C. Gancedo, and G. Delafuente. 7. Energy yielding metabolism in yeasts, pp. 7-7. In A. H. Rose and J. S. Harrison (eds.) The yeasts, vol.. Academic Press, London.. Splittstoesser, D. F., D. T. Queale, and L. R. Mattick. 78. Growth of Saccharomyces bisporus var. bisporus, a yeast resistant to sorbic. Am. J. Enol. Viticult. :7-7.. Walker, H. W., and J. C. Ayres. 7. Yeasts as spoilage organisms. pp. -57. In A. H. Rose and J. S. Harrison (eds.) The yeasts, vol.. Academic Press, London.. Warth, A. D. 77. Mechanism of resistance of Saccharomyces bailii to benzoic, sorbic and other weak s used as food preservatives. J. Appl. Bacteriol. :5-. JOURNAL OF FOOD PROTECTION, VOL. 8, JULY 85