Comparison of Methods for Enumeration of Yeasts and Molds in Shredded Low-Moisture, Part-Skim Mozzarella Cheese

Transcription

1 529 Journal of Food Protection, Vol. 63, No. 4, 2000, Pages Copyright, International Association for Food Protection Comparison of Methods for Enumeration of Yeasts and Molds in Shredded Low-Moisture, Part-Skim Mozzarella Cheese DANA S. SPANGENBERG 1 A STEVEN C. INGHAM 2 * 1 Northland Laboratories, 1110 North Main Street, Fort Atkinson, Wisconsin 53538; and 2 University of Wisconsin Madison, Department of Food Science, 1605 Linden Drive, Madison, Wisconsin , USA MS : Received 13 July 1999/Accepted 13 November 1999 ABSTRACT Two studies were conducted to compare established and new methods for enumerating yeasts and molds in shredded lowmoisture, part-skim mozzarella cheese stored under refrigeration and temperature-abuse conditions. Yeast and mold counts covered a range of 6 log 10 units. In study 1, the potato dextrose agar plus chlortetracycline () pour plate, dichloran rose bengal chloramphenicol () spread plate,, and hydrophobic grid membrane filtration methods were used to analyze samples after 1 day of unopened storage at 7 C and after opening, resealing, and 2 days of storage at 25 C. The results of all methods were highly correlated (r 2 ). In study 2, the,, and methods were compared with the 2-day method in an analysis of 42 samples stored for various times at 8, 11, 15, and/or 22 C. The results of all methods except the method were again highly correlated (r 2 ), although yeasts and molds were not always detected by all methods. Compared with the,, and methods, the method most often (10 of 42 samples, 23.8%) failed to detect yeasts and molds when at least one other method did, and the results were less highly correlated with those of other methods (r to 0.90). Our results suggest that the,,, and methods are equivalent for enumerating yeasts and molds in shredded low-moisture, part-skim mozzarella cheese samples. Enumeration of yeasts and molds is routinely performed to assess the sanitary condition of shredded lowmoisture, part-skim mozzarella cheese. A variety of methods are available to the analyst, although no one method is universally recommended by food regulators and mycologists. Acidified potato dextrose agar () pour plating has traditionally been used, and this method was accepted as a class O standard method in the 16th edition of Standard Methods for the Examination of Dairy Products (5). However, acidification of can suppress fungal growth, and in the Standard Methods for the Examination of Dairy Products (5), pour plating with containing added chloramphenicol and chlortetracycline (to prevent bacterial growth) was preferentially recommended. In the 1992 edition of the United States Food and Drug Administration Bacteriological Analytical Manual (7), pour plating with containing 40 ppm of added chlortetracycline was cited as a preferred method. Criticisms of the various versions of pour plating include the following: (i) The thermal stress associated with exposure to molten agar based medium may kill or injure cells; (ii) pour plating inhibits growth of strictly aerobic fungi except when these organisms are on the medium surface; (iii) does not contain enough nutrients to support the growth of many common fungi; and (iv) the incubation period of 5 days at 25 C is too long. In the 1998 edition of the Bacteriological Analytical * Author for correspondence. Tel: ; Fax: ; scingham@facstaff.wisc.edu. Manual (8), spread plating on dichloran rose bengal agar containing added chloramphenicol () was recommended as a preferred method for enumerating yeasts and molds in foods. This method is also listed in the Standard Methods for the Examination of Dairy Products (5) as a class B method. The medium contains chloramphenicol and has a ph of to inhibit bacterial growth. It also contains rose bengal to control the size and height of mold colonies and to enhance visual recognition of colonies and provides more nutrients than (6). Because this medium is inoculated by spread plating, all obligate aerobes are exposed to oxygen. Inoculated plates are incubated at 25 C and examined for colonies after 5 days. Other alternatives to the pour plating methods include the,, and methods. According to the manufacturer (3M Microbiology Products, St. Paul, Minn.), the yeast and mold count film is coated with a dehydrated medium that contains antibiotics to inhibit bacterial growth, a compound that gels in cold water, and a dye to enhance viewing of colonies. The primary advantage of the method is that preparation of a plating medium is not necessary, thus reducing labor costs. The method requires incubation at 20 to 25 C with examination for colonies at both 3 and 5 days. The method (QA Life Sciences, Inc., San Diego, Calif.) involves preparation of a sample dilution, prefiltration if necessary, filtration through a hydrophobic grid membrane filter (HGMF), placement of the HGMF on YM-11 agar, and incubation at 25 C for 48 to 52 h. The YM-11 agar contains chloramphenicol and chlortetracy-

2 530 SPANGENBERG A INGHAM J. Food Prot., Vol. 63, No. 4 cline to inhibit bacterial growth and trypan blue to aid in visualizing colonies. The hydrophobic grid lines on the m (pore size) filter act as barriers against spreading fungal colonies and thus divide the filter into 1,600 individual compartments. A most probable number (MPN) is calculated from the number of compartments containing fungal growth. The method is AOAC International first-action official method 991 for enumerating yeasts and molds in foods (1). Major advantages of the method are its short incubation time and the ability to detect a wide range of cell numbers using a single HGMF. The method (IDEXX Laboratories, Westbrook, Maine) recommends a 2-day incubation period. An aliquot of diluted sample is mixed with a liquid medium and is then transferred to a multiwell container underlaid with an absorbent liner. The container is then covered and incubated for 48 h at 25 C. Proprietary medium constituents inhibit bacterial growth, and the activity of multiple fungal enzymes results in fluorescence under UV illumination. After incubation, an MPN is calculated from the number of fluorescent wells. The major advantages of the method are a relatively short incubation time and the ability to detect a wide range of cell numbers using a single testing container. The objective of this project was to compare new and established methods for enumerating yeasts and molds in shredded low-moisture, part-skim mozzarella cheese. Various methods have been compared for enumeration of yeasts and molds in cheeses. Entis and Lerner (4) found no statistically significant difference between the method and pour plating on with added chlortetracycline for enumerating fungi in grated parmesan and cream cheese. Beuchat et al. (2) reported correlation coefficients between and pour plating or spread plating on acidified and plate count agar with added chloramphenicol in an analysis of 30 lots of hard and soft cheeses. Vlaemynck (9) reported a similarly high correlation coefficient between and a glucose yeast extract plus oxytetracycline pour plate method used to analyze fresh cheeses and yogurt. In the present work, two studies were conducted. In study 1, triplicate plus chlortetracycline pour plating (7), spread plating,, and Iso- Grid methods were compared. In study 2, duplicate plus chlortetracycline pour plating, spread plating,, and methods were compared. MATERIALS A METHODS Samples. Packages of shredded low-moisture, part-skim mozzarella cheese were purchased from local supermarkets. To evaluate a wide range of fungal populations, cheese samples were analyzed after a variety of storage treatments. In study 1 (conducted from 1994 to 1995), samples were purchased at local markets, transported within 20 min to the Department of Food Science, University of Wisconsin Madison, and stored at 7 C for 1 day. After storage, duplicate 25-g subsamples were removed for analysis. Then, the sample bag was taped shut and stored at 25 C for 2 days before duplicate 25-g subsamples were taken for analysis. In study 2 (conducted in 1999), samples were either stored at 8 C or at various combinations of 22, 8, 15, and 11 C to represent temperature-abuse conditions. To simulate package seal TABLE 1. Shredded low-moisture, part-skim mozzarella cheese samples and sample storage conditions after arrival in the laboratory Study Sample and brand Storage a 1 30A,31A,48E, 7 C, 1 day 49 E, 50 B, 51 B, 52 F, 53 F, 54 G, 55 G, 56 H, 57 H, 59 C, 72 G, 73 G, 74 H, 75 H, 76 C, 77 C 38 A, 39 A, 43 B, 25 C, 2 days, open 44 C, 45 C, 46 D, 62 B, 63 B, 64 F, 65 F, 71 C, 81 G, 85 C, 86 D 2 1I,4A 8 C, 26 days 2I,3A 15 C, 26 days 5C,7B 8 C, 19 days 6C,8B 15 C, 19 days 9C,11J,41C,43A 8 C, 12 days 10 C, 12 J 15 C, 12 days 13 C, 15 A 8 C, 6 days 14 C, 16 A 15 C, 6 days 17 A, 19 I 8 C, 33 days 18 A, 20 I 15 C, 20 days; 8 C, 13 days 21 A, 23 C 8 C, 25 days 22 A, 24 C 15 C, 12 days; 8 C, 13 days 25 B, 27 I 22 C, 38 h; 8 C, 19 days 26 B, 28 I 22 C, 38 h; 15 C, 7 days; 8 C, 12 days 31 B, 32 B 22 C, 28 h, open; 8 C, 11 days 37 B, 39 A 22 C, 24 h; 8 C, 18 days 38 B, 40 A 22 C, 24 h; 15 C, 12 days; 11 C, 8 days 42 C, 44 A 15 C 6 days; 11 C, 6 days 45 I, 47 C 8 C, 5 days 46 I, 48 C 11 C, 5 days a Sequential temperatures and times are separated by a semicolon. failure, some sample bags were opened for part of the storage time. A single 25-g subsample was analyzed at the end of each storage treatment. Sample descriptions are listed in Table 1. Subsample preparation. In study 1, each 25-g subsample was aseptically removed, diluted 1:10 in 0.1% (wt/vol) peptone (Difco Laboratories, Detroit, Mich.) water, and homogenized using a stomacher for 90 s. In study 2, samples were transported in refrigerated, insulated containers within 1 h to Northland Laboratories, Fort Atkinson, Wis. On receipt, samples were stored at 2 C until analysis the next day. Each 25-g subsample was diluted 1:10 in 0.1% peptone water and homogenized using a stomacher for 30 s. In both studies, subsequent serial dilutions were made, as needed, in 0.1% peptone water. Analysis was performed within 5 min of homogenization. Methods of analysis. supplemented with 40 ppm of chlortetracycline (Sigma Chemical Co., St. Louis, Mo.) was used to prepare triplicate (study 1) and duplicate (study 2) pour plates of appropriately diluted subsamples. Plates were incubated at 25 C for 5 days, after which yeast and mold colonies were counted.

3 J. Food Prot., Vol. 63, No. 4 YEASTS A MOLDS IN MOZZARELLA CHEESE 531 Subsamples were also analyzed by duplicate spread plating onto agar (Oxoid, Inc., Ogdensburg, N.Y.) supplemented with 100 ppm of chloramphenicol (Oxoid). To attain the same detection limit as the pour plating method, 0.5 ml of the initial (10 1 ) subsample dilution was placed onto each of two plates of the medium. Plates were incubated at 25 C for 5 days, after which yeast and mold colonies were counted. The method calls for a -ml aliquot of the appropriate subsample dilution and 15 ml of sterile 0.1% peptone water to be vacuum filtered together through a HGMF. Once all liquid had passed through the filter, it was aseptically transferred, grid side up, to the surface of a YM-11 agar plate (QA Life Sciences). The filter and plate were then incubated at 25 C for 48 to 52 h. Grid squares containing growth were counted, and the MPN of fungal cells in the sample was calculated from a table provided by the manufacturer. Each of two films was inoculated with ml of the appropriate subsample dilution according to the manufacturer s instructions. Colonies were counted after 5 days of incubation at 25 C. The analysis was started by adding a -ml aliquot of the appropriate subsample dilution to a vial of SYM test medium (IDEXX Laboratories) previously reconstituted with 9.0 ml of sterile water. The inoculated medium was then dispensed into a and incubated at 25 C for 72 h. wells fluorescing under a 6-watt, 365-nm UV lamp were counted, and the MPN of yeast and mold cells in the original sample was determined using a table provided by the manufacturer. Statistical analysis. In study 1, data from a given subsample were included in the analysis when all four methods detected fungi. The same criteria for inclusion were initially used with study 2 data. Next, all study 2 sample data were analyzed twice, first with a value of 0 and then with a value of 10 MPN or CFU per g assigned when a given method did not detect viable fungi. The latter value is the detection limit for the,, and Iso- Grid methods, whereas the detection limit is 2 MPN per g. When all compartments or wells contained fungal cells, the maximum MPN given by the table for that method was used in the statistical analysis. All data values were converted to log 10 CFU or MPN per g. For each pair of methods, a simple regression line was plotted using Cricket Graph III software (Computer Associates International, Inc., Islandia, N.Y.). From each regression line, the slope, y intercept, and variance of method Y attributable to linear regression on method X (r 2 -value) were calculated using the same software. RESULTS A DISCUSSION In study 1, differences between methods in the number of yeasts and molds enumerated in shredded low-moisture, part-skim mozzarella cheese were generally quite small (Table 2). The difference between methods exceeded 0.5 log 10 unit in only 4 of 34 (1%) samples. In study 2, 4 of 42 samples (9.5%) had between-method differences log 10 unit, and 14 of 42 (3%) samples had between-method differences of 0.50 log 10 unit (Table 3). For 10 samples, the method failed to detect yeasts and molds, whereas at least one other method did detect viable fungi. For most of these samples, 2.0 log 10 CFU or MPN per g of yeasts and molds were detected by another method, but there were three samples for which at least one method detected 2.0 log 10 CFU or MPN per g. In contrast, there were only two samples for which the TABLE 2. Enumeration of yeasts and molds on shredded lowmoisture, part-skim mozzarella cheese in study 1 a Sample and brand 30 A 31 A 38 A 39 A 43 B 44 C 45 C 46 D 48 E 49 E 50 B 51 B 52 F 53 F 54 G 55 G 56 H 57 H 59 C 62 B 63 B 64 F 65 F 71 C 72 G 73 G 74 H 75 H 76 C 77 C 81 G 85 C 86 D Enumeration method (see text) (MPN/g) a Storage conditions are listed in Table 1. All values are log 10 values. method detected fungi and at least one other method did not. Both of these samples contained 2.0 log 10 MPN per g as measured by the method. The,, and methods did not detect fungi in four, three, and five samples, respectively, for which fungi were detected by at least one other method. All of these samples contained 2.0 log 10 MPN or CFU per g. Throughout study 2, the vast majority of colonies counted on and plates were yeasts. In study 1, results of the,,, and methods were all highly correlated (r 2 ) and practically equivalent (regression equation slope 0.91 to 6; y intercept to 0.51 log 10 unit; Table 4). Regardless of the criteria for data inclusion and value assignment used in study 2, results of the,, and methods were highly correlated (r 2 to ; Table 5) and had regression equation slope and y intercept values similar to those of study 1. Results of the

4 532 SPANGENBERG A INGHAM J. Food Prot., Vol. 63, No. 4 TABLE 3. Enumeration of yeasts and molds on shredded lowmoisture, part-skim mozzarella cheese in study 2 a Sample and brand 1I 2I 3A 4A 5C 6C 7B 8B 9C 10 C 11 J 12 J 13 C 14 C 15 A 16 A 17 A 18 A 19 I 20 I 21 A 22 A 23 C 24 C 25 B 26 B 27 I 28 I 31 B 32 B 37 B 38 B 39 A 40 A 41 C 42 C 43 A 44 A 45 I 46 I 47 C 48 C b Enumeration method a (MPN/g) a Storage conditions are listed in Table 1. All values are log 10 values. b, none detected. Detection limits were as follows: and, 0.7 log 10 CFU/g;, log 10 MPN/g;, 0.3 log 10 MPN/g. method were not as well correlated with those of the other methods (r to 0.90, depending on data inclusion and value assignment criteria; Table 5). The relatively poorer correlation between results obtained with the method and those obtained with other methods was strongly influenced by the results of five samples (1 I, 8 B, 9 C, 10 C, and 23 C), for which the method detected 0.8 TABLE 4. Correlation between results of the,,, and methods for enumerating yeasts and molds in shredded low-moisture, part-skim mozzarella cheese in study 1 (n 33) Method X Method Y Regression line Slope y-intercept a r a In units of log 10 CFU or MPN per g to log 10 MPN per g fewer cells than the other methods. All of these samples were stored for at least 12 days at 8 or 15 C, conditions that may have selected in favor of certain organisms not detected by the method. The close similarity of results obtained with the,, and methods reported here is consistent with the findings of other comparisons of methods of enumeration of fungi in cheese (2, 4, 9). When all methods detected yeasts and molds in sam- TABLE 5. Correlation between results of the,, Iso- Grid, and methods for enumerating yeasts and molds in shredded low-moisture, part-skim mozzarella cheese in study 2 Method X Method Y Regression line r 2 for given data treatment All tests DL a 0 b 10 c a In this test, only data for samples having fungi present at levels greater than or equal to the detection limit (DL) of 10 MPN or CFU per g for,, and or 2 MPN per gram for were analyzed. b A value of 0 MPN or CFU per g was assigned when no fungi were detected () by a given method. c A value of 10 MPN or CFU per g was assigned when no fungi were detected () by a given method.

5 J. Food Prot., Vol. 63, No. 4 YEASTS A MOLDS IN MOZZARELLA CHEESE 533 ples stored unopened for 1 day at 7 C in study 1, the counts were typically between and log 10 log CFU or MPN per g. These yeast and mold levels were consistent with those reported for shredded mozzarella cheese at the beginning of 10 C storage under air with 90% relative humidity (3). Only five samples, all stored under air at 25 C for 2 days, had yeast and mold counts exceeding log 10 CFU or MPN per g (Table 2). These elevated yeast and mold counts were similar to counts in shredded mozzarella cheese stored under air for 14 days at 10 C (3). In study 2, counts in samples stored unopened at 8 C for up to 26 days were usually lower than the detection limits of the methods used. Samples 1 I and 21 I, stored unopened at 8 C for 26 and 25 days, respectively, had yeast and mold counts (mean of the four methods) of 4.0 log 10 CFU or MPN per g (Table 3). The food microbiology testing industry is swiftly adopting rapid and simple methods for a variety of analyses. Enumeration of yeasts and molds has traditionally been one of the slowest laboratory analyses. Ideally, more rapid yeast and mold enumeration methods should not have less sensitivity or colony differentiation (i.e., between yeasts and molds). We found that each yeast and mold enumeration method tested had notable advantages and disadvantages, listed below, that should be considered when choosing a method for use in a quality-assurance program for shredded low-moisture, part-skim mozzarella cheese. The traditional pour plate method allows easy differentiation of the yeast and mold groups, although individual genera within either group are not readily differentiated. Another drawback of this method is that spreading mold colonies can sometimes hinder colony counting. The method is relatively easy but requires preparation of medium before the analysis begins. The agar medium inhibits spreading mold colonies and allows easy differentiation of the yeast and mold groups, but not genera within a group. Medium preparation time for the method is comparable to that of the method. The recommended incubation time for the and methods is 5 days, but for a majority of samples in our studies, colonies could be accurately enumerated after 3 days of incubation. As with each of the other methods tested, the and methods estimate only the total fungal populations present in a sample. This estimation allows the analyst to make a subjective evaluation, often based on experience, of cheese plant hygiene. The results of enumerating yeasts and molds on freshly made mozzarella cheese do not necessarily tell the analyst which fungi will subsequently spoil the cheese if it is improperly stored or handled. But, results from subsequent analyses of stored cheese may show whether yeasts or molds have grown during storage. The method requires a shorter incubation time than the and methods (48 to 52 h compared with 3 to 5 days), but differentiation of the yeast and mold groups is not possible, and counting large numbers of colonies on the HGMF can be time consuming. The method requires less medium preparation than the and methods but has one-time costs for filtration units, a clamp, tubing, a vacuum source, and an optional manifold for performing multiple analyses. Labor costs are incurred for washing and sterilizing the filtration units between uses. Compared with pour plating and spread plating, the filtration and membrane transfer steps in the method take little time. However, these steps can take more time than plating when multiple dilutions must be filtered. The method requires no preparation of growth medium, and its incubation time is comparable to that of the and methods. Some differentiation of the yeast and mold groups is possible with this method. The method was convenient to perform because the test medium and holder are supplied by the manufacturer. Furthermore, this method had the lowest detection limit of the methods tested. Fluorescence in the wells after 48 h of incubation at 30 C was weak and difficult to detect, so incubation was continued for another 24 h. Even with this additional incubation, results were sometimes equivocal. Differentiation of the yeast and mold groups was not possible. In summary, the choice among the,, Iso- Grid, and methods for enumerating yeasts and molds in shredded low-moisture, part-skim mozzarella cheese can be based on such factors as cost, convenience, and expected number of samples. Our results indicate that these four methods will produce the same results. Additional testing of the method will be necessary to establish its equivalence to the other methods. ACKNOWLEDGMENT The authors gratefully acknowledge the laboratory assistance of Jee- Hoon Ryu in study 1. REFERENCES 1. AOAC International Official methods of analysis of AOAC International, 16th ed., 3rd revision. AOAC International, Gaithersburg, Md. 2. Beuchat, L. R., B. V. Nail, R. E. Brackett, and T. L. Fox Evaluation of a culture film ( YM) method for enumerating yeasts and molds in selected dairy and high-acid foods. J. Food Prot. 53: Eliot, S. C., J.-C. Vuillemard, and J.-P. Emond Stability of shredded mozzarella cheese under modified atmospheres. J. Food Sci. 63: Entis, P., and I. Lerner Two-day yeast and mold enumeration using the ISO-GRID membrane filtration system in conjunction with YM-11 agar. J. Food Prot. 59: Marshall, R. T. (ed.) Standard methods for the examination of dairy products, 16th ed. American Public Health Association, Washington, D.C. 6. Oxoid The Oxoid manual, 6th ed. Unipath Ltd., Basingstoke, UK. 7. U.S. Food and Drug Administration Yeasts, molds, and mycotoxins, p In United States Food and Drug Administration bacteriological analytical manual, 7th ed. AOAC International, Gaithersburg, Md. 8. U.S. Food and Drug Administration Yeasts, molds, and mycotoxins, p In United States Food and Drug Administration bacteriological analytical manual, 8th ed., revision A. AOAC International, Gaithersburg, Md. 9. Vlaemynck, G.M Comparison of and plate count methods for enumerating molds and yeasts in cheese and yogurt. J. Food Prot. 57: