Sorbic Acid as a Selective Agent in Cucumber Fermentations

Transcription

1 Sorbic Acid as a Selective Agent in Cucumber Fermentations II. Effect of Sorbic Acid on the Yeast and Lactic Acid Fermentations in Brined Cucumbers' 2 R. N. COSTILOW, F. M. COUGHLIN, E. K. ROBBINS, AND WEN-TAH Hsu Department of Microbiology and Public Health, Michigan State University, East Lansing, Michigan Previous studies (Costilow et al., 1955) demonstrated that the yeasts found to be most prevalent in cucumber fermentations were completely inhibited by.1 per cent sorbic acid in an 8 per cent salt (NaCl) medium at ph 4.6. As the ph was increased and/or the salt concentration decreased, more sorbic acid was required for complete inhibition of some of the yeasts tested. Cultures of Pediococcus cerevisiae, Lactobacillus plantarum and Lactobacillus brevis isolated from cucumber fermentations (Costilow et al., 1956) were not affected appreciably by sorbic acid in concentrations up to.1 per cent. Phillips and Mundt (195) and Jones and Harper (1952) reported that.1 per cent sorbic acid would prevent the growth of surface yeasts on cucumber fermentations without interfering with acid fermentation. The former authors found that concentrations less than.1 per cent were not completely effective. Borg et al. (1955) observed that.1 per cent sorbic acid would inhibit the growth of fermentative yeasts in cucumber fermentations; but, also, markedly inhibited the acid fermentation. Furthermore, the treatment was found to impair the curing and the color of the fermented cucumbers. As a result of our laboratory experiments, it was decided to test a number of sorbic acid concentrations in experimental cucumber fermentations to determine their effect on the yeast and acid fermentations. Received for publication May 13, 1957 Three varieties of cucumbers were brined in 1954; namely, SR-6, MR-17 and National. The cucumbers used in 1955 and 1956 were SR-6, with the exception of one station each year where a mixture of SR-6 and MR-17 was used. All lots were of no. 3 size (1Y8 to 2 in. diameter) fruit, except for one replication in 1955 and two in 1956 which were mixed no. 2 and no. 3 sizes (1'8 to 2 in. diameter), and two replications in 1956 with oversize cucumbers (2 to 21 in. diameter). Sorbic acid powder was used in all fermentations studied in 1954 and in one replication of the 1955 series. In the remaining fermentations, the sorbic was added in the form of 25 per cent sodium sorbate solution. Concentrations of sorbic acid in the barrel lots were all calculated on the basis of per cent of the weight of all the water in the fermentation; that is, the weight of the water added in the brine plus 9 per cent of the cucumber weight. The small scale operations permitted very accurate determinations of these weights. It was not possible to obtain comparable accuracy in most of the experimental lots in large tanks. Therefore, a ratio of sorbic acid treatment per 1 bushels of cucumbers approximately equivalent to a given percentage was developed from available data. This ratio was found to be as follows: 1 lb of sorbic acid per 1 bu of cucumbers equals approximately.1 per cent based on total water weight. It is obvious that this figure will be in error to some extent in most fermentations. If the cucumbers are packed tightly into the tank, the concentration of sorbic acid would be higher than estimated and vice versa. However, it is believed that the normal variations occurring were not of great significance in METHODS AND MATERIALS Table 1 shows the number of fermentations of each sorbic acid treatment studied and their distribution. The 1954 fermentations were all in 45-gallon barrels except for one series of treatments in 3-bushel tanks. this respect. The concentrations were checked in one In 1955, three series of treatments were conducted in of the replications where accurate measurements of barrel lots, one series in 3-bushel casks and the remainder in commercial scale lots (25- to 8-bushel quite close to the estimated levels. cucumbers and brine were possible and found to be tanks). All fermentations in 1956 were conducted in The experimental fermentations were brined at commercial scale lots. various salt levels each year but all treatments per 1 Journal Article No replication were brined at the same level. Table 2 2 This study was supported financially by a grant from The gives a breakdown of the number of replications brined Carbide and Carbon Chemicals Company, 3 East 42nd at different salt concentrations each year. Street, New York, N. Y. The average brine temperatures during the months of 373

2 374 August and September respectively were as follows: 1954, 2.4 and 19.3 C; 1955, 25.1 and 2.9 C; and 1956, 21.8 and 17.6 C. All fermentations were sampled at frequent intervals during the early phase of fermentation. The techniques for sampling, determination of acid and salt concentrations, and the counting of yeasts, coliforms and acidproducing bacteria were as outlined by Costilow and Fabian (1953). Brine temperatures and salometer readings were recorded at the time of sampling. RESULTS Effect of sorbic acid on subsurface yeast activity in cucumber fermentations. Sorbic acid proved to be a very effective inhibitor of yeasts in cucumber fermen- TABLE 1 Distribution and number of fermentations of each sorbic acid treatment studied COSTILOW, COUGHLIN, ROBBINS, AND HSU Brining Season Totals No. of locations where studies * were made No. of fermentations studied per treatment: Controls Sorbic acid (%) treatments :t Total no. of fermentations * A total of 11 different locations was used during the 3-yr period. t All sorbic acid concentrations in 1954 were calculated on the basis of the weight of the water in the brine plus 9 per cent of the cucumber weight. In the 1955 and 1956 studies with large tanks, the concentrations were estimated on the following basis: 1 lb of sorbic/1 bu cucumbers equals approximately.1 per cent. TABLE 2 Initial brine strengths of fermentations studied Brine Strength (Salometer) No. of Replications Per Salt Treatment* t 4t * All brine strengths raised 1 to 2 salometer per week unless otherwise noted. t One series held 2 to 3 weeks and then raised. t Two series held 3 to 4 weeks and then raised. Held 2 to 3 weeks and then increased 1 saloineter per week. tations. Figure 1 shows the log average yeast populations in the fermentations studied for each treatment. All populations were averaged at 5-day intervals for the first 3 days of fermentation and at 25-day intervals from 3 to 8 days of fermentation. The 1954 data for yeast counts are based on 17 replications per treatment on which a total of 618 counts were run, the 1955 data from 1 replications on which 391 counts were run, and that of 1956 from 1 replications on which 34 counts were run. It was evident at the conclusion of the 1954 studies that.5 per cent sorbic acid was approximately as effective as.1 per cent in the inhibition of yeasts in these fermentations. Therefore, in an effort to find the minimum effective levels for use,.1,.25, and.5 per cent concentrations were used in It is obvious from these results that all three levels of sorbic acid reduced yeast activity greatly. However, the yeast populations did increase to a fairly high level after an extended period when.1 per cent sorbic was used. 4- / Controls\ o-a.5% Sorbic Acid -a.1. % Sorbic Acid %ob 5 c Acid "I 6--* Controls o % S bc Aid.~ ~ ~ ~ ~ - %. E 24-e- 2lgs_ $ - :+ op *o [ VOL. 5 5 w d' Figure 1. Control of subsurface yeasts by sorbic acid treatment.

3 SORBIC ACID IN CUCUMBER FERMENTATIONS 375 N signiificaint yeast activ-ity developed when.25 and.5 per cenit sorbic acid was used. The 1956 data indicate that yeast cointrol was not (uite as good with.2 per cent sorbic acid as with.3 per ceint, although yeast growth was relatively low in each case. In two of the ten replications of the.2 per cent treatment, yeast populations attainied relatively high levels while in the other eight yeast growth was well controlled. This indicates that this level of treatment is somewhat marginal, and that slightly higher levels (.25 to.3 per cent) should be used to obtain more uniform control. Effect of sorbic acid on lactic fermentation in cucumber brines. The lactic fermentation occurring in the brines was followed by titration of acid and by determining populations of acid-forming bacteria. Average values for the titratable acidities (calculated as per cent lactic) of the various treatments are presented in figure 2. As with the yeast populations, the values plotted are averages at 5-day intervals through 3 days of fermentatioii and at 25-day intervals thereafter. It is obvious from these results that the average lactic acid development was not significantly affected by aniy concentration of sorbic acid other than.1 per cenit. As one would expect, the effect of.1 per ceint sorbic acid was more pronounced in some fermentations thani in others. This was particularly true whein the brine strength was varied. This is illustrated in figure 3 which shows the average lactic acid development in duplicate lots of treatments which were brined at 2 and 4 salometer and of four lots per treatment brined at 3 salometer. It should be noted, also, that a reduction in acid development was evident with the.5 per cent sorbic acid treatment in the 4 salometer brines. In three of the ten replications studied in 1955, a trend was noted which indicated that acid fermentation was inhibited slightly by increasing concentrations of sorbic acid. However, these differences were not great enough to be reflected in the averages. Therefore, they are of questionable significance. Data with respect to total populations of acid-forming bacteria correlated very well with that of acid development in the fermentations. In some instances, greater differences between treatments were evident in titratable acidity curves than in populations of acid-forming bacteria. Figure 4 illustrates the average acid-forming bacteria populations in the same fermentations for 4.. -J CX 4-j._ a ' Figuire 2. Effect of sorbic acid on lactic acid development in cucumber fermentations. Figure S. Effect of sorbic acid on lactic acid development in cucumber fermentations brined at 2, 3, and 4 salometer.

4 376 COSTILOW, COUGHLIN, ROBBINS, AND HSU [VOL. 5 wvhich acid production rates are presented in figure 3. On comparing these two figures, it is obvious that the measuremenit of acid reflects greater differences due to brine strength and sorbic acid treatment than bacterial population studies. Control of sutrface-yeasts by sorbic acid. Most of the experimental fermentations were not protected from the sunlight and, therefore, no surface yeasts developed. However, two replications in 1954 and 1956 and one series of treatments in 1955 were set up under cover to check on the control of such yeasts. Complete inhibition of film development was obtained with sorbic acid treatments of.25 per cent and higher (figure 5). As with the subsurface yeasts, the growth of film-forming yeasts was delayed by.1 per cent sorbic acid but it was not completely controlled. The.2 per cent treatment inhibited these yeasts completely in 32 salometer brines but a thin film was formed in low salometer brines (2). Measurement of sorbic acid in cucumber brines. Samples of the fermenting brines were taken during the 1954 and 1955 brining seasons and analyzed for sorbic L-I ) -. E z c 8 *4Q Brining Schedule 4 O Figure 4. Effect of sorbic acid on the growth of acid-forming bacteria in cucumber fermentations. acid content.3 The results of this study are presented in table 3. The concentration of sorbic acid present in the fermenting brines was from 35 to 65 per cent less than the amount added initially. The two sorbic acid concentrations are on a somewhat different basis, that is, the treatment was based on the per cent of water weight and the measurement on per cent of total brine weight. One might expect up to 2 per cent difference in the two, but this cannot explain the larger differences noted here. A Beckman Model DU spectrophotometer4 was obtained prior to the 1956 studies. Therefore, the measurement of the sorbic acid concentration in the brines was initiated in our laboratory. The method proposed by the Chemical and Physical Methods Laboratory, Carbide and Carbon Chemicals Company (1955) was selected for use. This is essentially the same as the method proposed by Melnick and Luckmann (1954) and by Luckmann and Melnick (1955) except that the sample is not distilled. The brine is filtered, diluted, acidified, and the absorbancy determined at 263 m,u wave length. Preliminary studies showed that there was some unknown material(s) in the brines to which no sorbic had been added which absorbed light at the same wave length as sorbic acid. Furthermore, the amount of this irrelevant absorption varied from one fermentation to another. In fact, in one replication of the 1956 treatments, the absorbancy of the control brine was approximately equal to that of the brine treated with.2 per cent sorbic acid. The average miscellaneous absorption at 263 m,u for 21 determinations run on the brine from one control tank was equivalent to.6 per cent sorbic acid, while the average of 21 measurements of another control at the same location was equivalent to.11 per cent sorbic acid. The average miscellaneous absorption in the brines from 1 control tanks in 1956 was equivalent to.8 per cent sorbic acid and the standard deviation was.3. Because there is no way to be sure of the amount of irrelevant absorption in the treated brines, a very large error in determination is possible. Efforts were made to separate the sorbic acid from the unknown material(s) in the brine which absorb at the same wave length. Distillation by the technique described by Melnick and Luckmann (1954) or extraction with organic solvents failed to effect any separation. DIscussioN The results of this study demonstrate that sorbic acid in concentrations of much less than.1 per cent 3Samples were pasteurized in flowing steam and then shipped to Carbide and Carbon Chemicals Company, South Charleston, W. Va., for analysis. 4Beckman Instruments, Inc., Fullerton, Calif.

5 1957] SORBIC ACID IN CUCUMBER FERMENTATIONS *5_Ct -_ E s a Figure 5. will effectively inhibit both subsurface and film-forming types of yeasts in cucumber fermentations. These results, in general, had been predicted from studies conducted in the laboratory with pure cultures of the yeasts which are most frequently found in these fermentations (Costilow et al., 1955). While the sorbic acid treatment required for complete control of yeasts in actual fermentations was greater than the level predicted from laboratory studies (.1 to.2 per cent), the actual sorbic acid concentrations in the fermenting brines were in many instances lower than this. Phillips and Mundt (195) and Jones and Harper (1952) found.1 per cent sorbic acid necessary for complete control of film yeasts in cucumber fermentations. However, a later report by Jones et al. (1956) indicated complete control of yeasts with a concentration of.1 per cent sorbic based on the weight of the brine in the fermentation. This would amount to a final concentration after equalization of.3 to.4 per cent and would be in line with the results of this study. No other reasons are apparent for the differences in observations reported here and the foregoing with respect to the efficiency of this compound as a yeast inhibitor. Control of film-yeasts by sorbic acid treatment TABLE 3 Measured sorbic acid concentrations in the fernienting brine Sorbic Fermenta- No. of Sorbic Acid Year Acid tion Age pisa Rnemvrae % days % % * Calculated on the basis of water weight. t Per cent by weight. The effect of sorbic acid on the lactic acid fermentation of cucumbers is very dependant on brine strength, that is, the higher the salt (NaCl) concentration, the greater the inhibition of lactic fermentation. Therefore, higher concentrations of sorbic acid may be used in conjunction with low salt levels than with high. However, in cucumbers salted between 2 and 35 salometer, a sorbic acid treatment of 2.5 lb per 1 bu of

6 378 cucumbers (about.25 per cent) should not impair lactic development and yet inhibit yeast growth. Phillips and Mundt (195) reported that lactic acid development in fermentations treated with.1 per cent sorbic acid was delayed slightly, but that eventually more acid was developed in the treated lot than was ever attained in the control. However, their data show that the highest per cent acid in the control was attained after 16 to 19 days, fermentation while this level was not attained in the treated lot until from 5 to 6 days. The maximum acid level attained in the control was only about.3 per cent which is quite low even for high salometer brine fermentations. Their results were complicated by the fact that the lactic acid concentration in the control should have continued to increase but, instead, decreased due to rapid utilization by film yeasts. This is always encountered when small containers are used for fermentations and nothing is used to prevent the growth of the surface yeasts. Borg et al. (1955) observed an even greater effect of.1 per cent sorbic acid on the lactic fermentations in brines ranging from 35 to 37 salometer than was observed in this work. Here again, there was a difference in the calculation of percentages. These workers based the percentage of sorbic acid treatment on the total weight of cucumbers and brine, while in this work the percentage was based on weight of the total water in the fermentation. Therefore, the concentration used in the work reported here would only amount to about 8 per cent of that used by Borg and his associates. Luckmann and Melnick (1955) have shown that approximately 1 per cent recoveries of sorbic acid are possible from a number of food products by spectrophotometric measurement. In conducting this work, they supplemented the foods with sorbic acid and retained an unsupplemented fraction as a control. This results in a true control in which the light absorption due to substances other than sorbic acid will be the same as in the sample containing sorbic acid. It was shown that this irrelevant absorption could amount to as much as 46.8 per cent of the absorption due to sorbic acid even when a.1 per cent sorbic acid supplement was used. This was found to vary greatly from food to food. However, Luckmann and Melnick did not present data for the variation in irrelevant absorption between samples of the same product. In the present study, the between-samples variation in irrelevant absorption was found to be very high in cucumber brines and one might expect a great amount of variation in other biological products. Certainly, this is deserving of further investigation. ACKNOWLEDGMENT The cooperation of the following companies in making their facilities and personnel available for the conduction of experiments is appreciated sincerely: H. W. COSTILOW, COUGHLIN, ROBBINS, AND HSU Madison Company, Cleveland, Ohio; J. J. Gielow and Sons, Detroit, Mich.; H. J. Heinz Company, Pittsburgh, Pa.; Squire Dingee Company, Chicago, Ill.; Budlong Pickle Company, Chicago, Ill.; Michigan Pickle Company, Lakeview, Mich.; Canadian Canners Limited, Hamilton, Ontario; and the Calhoun County Pickle Company, Marshall, Mich. We wish to express our particular gratitude to Mr. V. A. Weaver, H. W. Madison Company, Mason, Mich., for his great interest and cooperative effort during the course of these studies. Finally, our sincere thanks goes to Carbide and Carbon Chemicals Company for their support of this work, supplying the sorbic acid and sodium sorbate, and running sorbic acid determinations. SUMMARY The results of a 3-year study of the effect of sorbic acid on microbiological activity in cucumber fermentations are presented. This study involved 121 experimental cucumber fermentations distributed among 11 different locations in the northern cucumber production area. Sorbic acid proved to be a very effective inhibitor of yeast activity in cucumber fermentations. All levels tested from.1 to.1 per cent were very inhibitory. However, some yeast activity, both subsurface and film-forming, developed in the fermentations treated at the.1 per cent level and occasionally with the.2 per cent treatment. Small populations of yeasts were noted in some samples from fermentations treated with sorbic acid levels of.25,.3,.5, and.1 per cent, but they were too low to constitute a significant fermentation. Lactic acid fermentation was inhibited by.1 per cent sorbic acid treatments when the initial brine strengths ranged from 2 to 4 salometer; also, it was inhibited by.5 per cent sorbic acid with initial brining at 4 salometer. When.3 per cent levels or less were used, no significant effect on the acid development was noted. The sorbic acid levels as measured spectrophotometrically were always lower after equalization than calculated in the treatment. While it was shown that fairly large errors may occur in measurement, this does not account for the consistently low measured concentrations. No explanation for this is available at this time. REFERENCES [VOL. 5 BORG, A. F., ETCHELLS, J. L., AND BELL, T. A The influence of sorbic acid on microbial activity in commercial cucumber fermentations. Bacteriol. Proc., Abstract 19. Chemical and Physical Methods Laboratory 1955 Works Laboratory Manual. Carbide and Carbon Chemicals Company, South Charleston, W. Va. COSTILOW, R. N. AND FABIAN, F. W Microbiological

7 19571 LACTOSE NEGATIVE ENTEROBACTERIACEAE IN FOODS 379 studies of cucumber fermentations. Appl. Microbiol., 1, COSTILOW, R. N., FERGUSON, W. E., AND RAY, S Sorbic acid as a selective agent in cucumber fermentations. I. Effect of sorbic acid on microorganisms associated with cucumber fermentations. Appl. Microbiol., 3, COSTILOW, R. N., COUGHLIN, F. M., ROBACH, D L., AND RAGHEB, H. S A study of the acid-forming bacteria from cucumber fermentations in Michigan. Food Research, 21, JONES, A. H. AND HARPER, G. S A preliminary study of factors affecting the quality of pickles on the Canadian market. Food Technol., 6, JONES, A. H., FERGUSON, W. E., AND LYSTER, N. J Studies on the microbiology of cucumber brine stock fermentation. Can. Food Inds., 27, (6, 7). LUCKMANN, F. H. AND MELNICK, D Sorbic acid as a fungistatic agent for foods. X. Spectrophotometric determination of sorbic acid in foods in general. Food Research, 2, MELNICK, D. AND LUCKMANN, F. H Sorbic acid as a fungistatic agent for foods. III. Spectrophotometric determination of sorbic acid in cheese and in cheese wrappers. Food Research, 19, PHILLIPS, G. F. AND MUNDT, J Sorbic acid as an inhibitor of scum yeast in cucumber fermentations. Food Technol., 4, The Presumptive Enumeration of Lactose Negative as well as Lactose Positive Enterobacteriaceae in Foods D. A. A. MOSSEL Central Institute for Nutrition Research T.N.O. Utrecht, The Netherlands Seeliger (1952) and Henriksen (1955) have rightly pointed out that in the current methods of detection or enumeration of coliform bacteria in foods, that is, with the help of solid or liquid lactose-base media, an important class of Enterobacteriaceae of public health significance, namely, the Paracolobacteriaceae (Stuart et al., 1943; Hobbs et al., 1949; Murphy and Morris, 195; Edwards and Ewing, 1952; Heyl, 1954; is missed. Henriksen (1955) has suggested that this situation be corrected by substituting in current formulae mannitol for lactose because mannitol is a polyol that is fermented with gas formation by many Enterobacteriaceae including most Paracolobacteriaceae. So far, Henriksen's excellent suggestion has neither been applied in food microbiology, nor even received proper attention. We have investigated therefore, the possibilities for applying mannitol base media in the hygienic evaluation of foods. MATERIALS AND METHODS Preliminary Experiments As a first step, 65 strains of Enterobacteriaceae present in the culture collection of the Central Institute for Nutrition Research were seeded in a medium containing brilliant green bile, 1 per cent mannitol, peptone and water. This was done to check their expected ability to ferment mannitol in the presence of these current inhibitors. Among the cultures investigated were 16 strains of the genus Escherichia, 13 strains of Klebsiella, 6 strains of Paracolobacter, 2 cultures of Proteus, 27 strains of the genus Salmonella and one species of Serratia. All strains. except those Received for publication May 23, 1957 belonging to Proteus, Serratia, Group Providencia, and two of the five anaerogenic Salmonella species tested (Mossel et al., 1957) formed copious amounts of gas in this medium, and mostly so within 24 hr at 37C. Use of a Solid Medium It is preferred generally in diagnostic bacteriology to use solid instead of liquid media because, in the former, competition phenomena play virtually no role, while the result that isolation of primary cultures is a far more reliable procedure. This consideration led to using the well-established crystal violet neutral red bile agar (violet red bile agar of Bartram and Black, (1936)) for the present purpose. Though in this formula lactose could have been replaced by mannitol, the complete medium supplemented with mannitol is preferred to allow laboratories, who so far have used the dehydrated culture medium for the enumeration of coliform bacteria, to continue to do so. The new medium is prepared by adding 1 per cent of d-mannitol to the dehydrated medium and dissolving it together with the other ingredients. The formula of the medium, for which we propose the name CNBLM agar, is therefore: Yeast extract, dehydrated, 3 g; peptone, 7 g; sodium chloride, 5 g; lactose (optional), 1 g; d-mannitol, 1 g; bile salts, dehydrated, 1.5 g; crystal violet, 2 mg; neutral red, 3 mg; agar, 15 g; water, 1 L (ph = 7.4 i.1). The medium is not sterilized but, rather, as is usual with the classical violet red bile agar, pasteurized by heating the ingredients suspended in water to about 1 C immediately before use. This heat treatment is