865 Journal of Food Protection, ol. 48, No. 10, Pages 865-869 (September 1985) Copyright 0 International Association of ilk, Food, and Environmental Sanitarians Factors Affecting the icrobiological Quality of Burgos and illalon Cheeses at the Retail Level F. J. CHAARRI, J. A. NUNEZ, L. BAUTISTA and. NUNEZ* epartamento de Bioquimica y icrobiologia, Instituto Nacional de Investigaciones Agrarias, Apartado 8111, adrid 28040, Spain (Received for publication November 30, 1984) ABSTRACT One hundred forty-four samples of Burgos and illalon cheeses collected during April and July 1983 from retail outlets in adrid, Spain, were analyzed for microbiological quality. Geometric mean counts were 6.8 10 5 staphylococci/g, 1.0 x 10 5 coliforms/g and 7.9 x 10 3 yeasts/g for Burgos cheese, whereas the respective mean counts for illalon cheese were 1.4 x 10 6 /g, 8.7 x 10 4 /gand 1.9 x l(f/g. Coagulase-positive staphylococci represented in April and July, 7.9 and 42.2%, respectively, of the isolates from egg-yolk tellurite glycine agar plates. Fecal coliforms accounted for 5.6 and 26.9% of the isolates from violet red bile agar plates in April and July, respectively. Presence of alkaline phosphatase activity in cheese was not significantly related to numbers of staphylococci or coliforms and should only be regarded as a presumptive test. The use of ph values or total counts as indicators of the microbiological quality of Burgos and illalon cheeses is also discussed. Burgos and illalon are the main white cheese varieties manufactured in Spain, with a combined production of over 10,000 tons/year. Both were traditionally made from raw ewe's milk. However, pasteurized milk is nowadays almost exclusively used and ewe's milk has been replaced for economic reasons by an undefined mix of cow's, ewe's and goat's milk. Calcium chloride (0.01 to 0.02%) and animal rennet, but no lactic cultures, are added to the milk. Coagulation takes place at 28 to 30 C and curds are cut 30 to 40 min later into 2-cm cubes. The temperature of the curd may be raised after cutting up to 35 C in illalon cheesemaking. The curd is scooped into cylindrical molds and allowed to drain. Burgos cheese is not pressed, but molds are turned 2 to 3 times at 1-h intervals, whereas illalon cheese is slightly pressed for 2 to 4 h. Salting in a 10% NaCl brine for 2 h is common in illalon cheesemaking, whereas Burgos cheese either is not salted or is brine-salted for 10 to 30 min. Final sizes are 14 to 16 cm in diameter and 6 to 8 cm high for Burgos cheese and 7 to 9 cm in diameter and 20 to 25 cm long for illalon cheese, with weights of 1 to 2 kg and moisture contents of 55 to 65% for Burgos cheese and 50 to 60% for illalon cheese. Cheeses are stored at 4 C, distributed and consumed unripened within the next 3 to 5 d. High moisture levels and low salt and lactic acid contents are some of the main characteristics of Burgos and illalon cheeses. These favorable environmental conditions enhance growth of pathogenic and undesirable bacteria. There have been some reports of microbial levels in various soft and white cheese varieties {2,3,5,6). Information on numbers of indicator or pathogenic microorganisms in Burgos and illalon cheeses has not been published. However, public health authorities have recognized a high degree of hazard due to these white cheese varieties, and have used the phosphatase test as an index of inadequate heat treatment of the milk and, subsequently, as a criterion for rejecting the product. In the present work, the results of a survey on the microbiological quality of Burgos and illalon cheeses at the retail level are reported. The influence of various parameters, such as temperature and ph of the cheese, day of the week and sampling period, on the microbiological quality of these varieties has also been studied. The convenience of using the phosphatase test as the basis for rejection of a lot is discussed. ATERIALS AN ETHOS Samples Field samples (200 g) of Burgos (94 samples) and illal6n (50 samples) cheeses were collected 3 d (onday, Tuesday and Wednesday) per wk during April and July 1983 at adrid groceries and supermarkets chosen at random. Temperature of the sample was determined at the time of collection. Samples were transported to the laboratory under refrigeration and held at 4 C until examined. icrobiological analyses and chemical determination Representative 10-g sample units were homogenized with 90 ml of a sterile 2% citrate solution at 45 C for 1 min by means of a Stomacher 400 (Seward Laboratory, London, England). ecimal dilutions were prepared in 0.1% sterile peptone water. Total viable counts were determined in plate count agar (12). Coliform counts were estimated from violet red bile agar (RBA) plates incubated at 30 C for 24 h (10), staphylococci
866 CHAARRI ET AL. from egg-yolk tellurite glycine agar (EYTGA) plates incubated at 37 C for 48 h (4), and yeasts and molds from potato dextrose agar plates (72). uplicate plates were used for all determinations. Ten isolates per sample picked at random from RBA plates were tested for oxidase activity (77) and inoculated into brilliant green lactose bile broth tubes (70) that were incubated for 48 h at 45 C. Ten colonies per sample picked at random from EYTGA plates were tested for coagulase production by the tube method (14). Alkaline phosphatase activity in cheese was determined with the Lactognost kit (Heyl, Berlin, West Germany) on sodium citrate cheese homogenates. The ph of cheeses was determined by immersing the electrode into the sample (12). Statistical treatment of data Effects of cheese variety (), presence of active phosphatase (P), month () and day () of sampling on counts of the various microbial groups were studied, after a log transformation, by a factorial analysis of variance (75) using program BP2 (epartment of Biomathematics, UCLA, Los Angeles, CA). Interactions of three and four effects were not included in the model, excepting x x interaction, due to the lack of data. Influence of ph and temperature of the samples on microbial counts was determined by means of a linear regression analysis (75) using program BP1R (epartment of Biomathematics, UCLA). RESULTS icrobial levels recorded for Burgos and illalon cheeses were high for cheeses made from pasteurized milk, with mean log ]0 counts for staphylococci, coliforms and yeasts of 5.94, 4.99 and 4.03, respectively (Tables 1 and 2). Coagulase-positive staphylococci represented 24.8% of the isolates from EYTGA plates, whereas fecal coliforms accounted for 16.4% of the isolates from RBA plates. A low proportion (4.5%) of the isolates from RBA plates exhibited oxidase activity. A significant (P<0.05) effect of cheese variety on yeast counts was observed (Table 3). Alkaline phosphatase activity was detected in 27 samples (Table 1). onday samples (9 samples) more frequently had a strong positive reaction than Tuesday (2 samples) or Wednesday (2 samples) samples. Phosphatase activity of the samples showed a significant (P<0.01) effect on yeast counts, with higher levels in cheeses with a strong positive reaction (Tables 1 and 3). Sampling period did not influence significantly the microbial contents of Burgos and illalon cheeses, excepting coliforms (Table 3), though July mean levels of total counts and staphylococci were twice those obtained in April (Table 2). Coagulase-positive staphylococci represented 7.9% of the isolates from EYTGA plates in TABLE 1. ean log ]0 microbial counts in 144 Burgos and illalon cheese samples, according to cheese variety and alkaline phosphatase activity. Cheese variety Phosphatase activity" icrobial group Total counts Staphylococci Coliforms Yeasts olds Burgos (n c = 94) 7.34 5.83 5.01 3.90 2.31 illalon (n = 50) 7.43 6.15 4.94 4.28 2.33 "ata from Burgos and illalon cheeses have been pooled. b ±, weak positive. c n, number of samples. + (n=13) logio/g 8.11 6.45 5.52 4.96 2.52 ± b (n=14) 7.34 5.72 5.31 3.84 2.29 - (n=117) 7.29 5.91 4.89 3.95 2.30 TABLE 2. ean log I0 microbial counts in 144 Burgos and illalon cheese samples", according to the sampling period and the day of the week samples were collected. Sampling period ay of the week icrobial group Total counts Staphylococci Coliforms Yeasts olds April (n = 72) 7.23 5.80 4.78 3.97 2.43 July (n = 72) 7.50 6.08 5.19 4.09 2.21 a ata from Burgos and illal6n cheese have been pooled. onday login/g 7.68 6.18 5.07 4.24 2.28 Tuesday 7.25 5.70 4.77 3.86 2.27 Wednesday 7.16 5.94 5.12 3.99 2.41
BURGOS AN ILLALON CHEESES 867 April, whereas they accounted for 42.2% of the isolates in July. A similar increase was recorded for fecal coliforms, which reached 5.6% of the isolates from RBA plates in April and 26.9% in July. The day of the week samples were collected did not influence significantly any of the microbial groups studied (Table 3), although their mean logio counts were higher on ondays than on Tuesdays or Wednesdays (Table 2). ean ph value for Burgos and illalon cheeses (pooled data) was ph 6.2. Linear regression analysis showed statistically significant relationships between ph and levels of most microbial groups (Table 4), but coefficients of determination remained low in all cases. Slightly higher coefficients of determination were obtained by quadratic regression. At least 97% of samples with ph <6.0 exhibited heavy contamination, with logio total counts >7.0 CFU/g, logi 0 staphylococci counts >6.0 CFU/g, logjo coliform counts >5.0 CFU/g or log 10 yeast counts >4.0 CFU/g. On the contrary, cheeses with ph >6.0 showed a large variation of their microbial content, with only 41% of the samples having log 10 total counts <7.0 CFU/g, 24% of samples having logio staphylococci counts <5.0 CFU/g, 27% of samples having logio coliform counts <4.0 CFU/g and 12% of samples having logio Y east counts <3.0 CFU/g. The average temperature of Burgos and illalon cheeses (pooled data) was 11.4 C. Linear regression analysis between temperature and microbial levels (Table 4) detected some significant relationships, but coefficients of determination were lower than those obtained for ph. Significant relationships between total counts and staphylococci, coliform or yeast counts were also detected by linear regression analysis (Table 5), with higher coefficients of determination than those obtained for ph or temperature. Cheeses with log 10 total counts >6.0 TABLE 3. Levels of significance (tail probability) of main effects and interaction effects of cheese variety (), phosphatase activity (P), month () and day () of sampling on log, 0 counts of various microbial groups in Burgos and illalon cheeses" obtained by factorial analysis of variance. Effect Total counts Staphylococci Coliforms Yeasts olds p P P p 0.930 0.111 0.157 0.441 0.933 0.804 0.340 0.131 0.911 0.117 0.493 0.404 0.210 0.194 0.633 0.859 0.368 0.257 0.383 0.914 0.085 0.750 a ata from Burgos and illal6n cheeses have been pooled. b P<0.05. C P<0.01. 0.459 0.055 0.036 b 0.855 0.789 0.968 0.641 0.581 0.894 0.321 0.463 0.020 b 0.008 c 0.242 0.617 0.412 0.249 0.529 0.377 0.272 0.512 0.851 TABLE 4. Linear regression analysis of total counts, staphylococci, coliforms, yeasts and molds on ph and temperature of Burgos and illalon cheeses". Regression equation ^ ph logio total counts/g log,o staphylococci/g log 10 yeasts/g logio molds/g 16.422-1.448 12.766-1.092 11.864-1.100 8.885-0.777 2.290 +0.005 61.742 f 24.882 f 28.306 f 17.762 f 0.001 NS g 0.767 0.420 0.224 0.243 0.691 0.445 0.365 0.846 0.100 0.058 0.267 0.303 0.149 0.166 0.111 0.000 Temperature logi 0 total counts/g logio staphylococci/g logi 0 yeasts/g log, 0 molds/g a ata from Burgos and illal6n cheeses have been pooled. b F, regression mean square / residual mean square. c r 2, coefficient of determination (multiple r square). d P<0.05. e P<0.01. f P<0.001. 8 NS, not significant. y = y = 6.474 +0.079 5.062+ 0.077 3.996 + 0.087 3.794 + 0.021 2.238+0.007 6.860 6 5.664 d 8.036 e 0.582 NS g 0.116 NS g 0.046 0.038 0.054 0.004 0.001
868 CHAARRI ET AL. TABLE 5. Linear regression analysis of staphylococci, coliforms sts and molds on total counts of Burgos and illalon cheeses". * y Regression equation F b r 20 Logio total counts/g logio staphylococci/g logio yeasts/g logio molds/g y = 0.754 +0.704 y = 0.359 + 0.628 1.194 + 0.385 y = 1.876 + 0.061 106.494 d 85.033 d 33.099 d 1.016 NS e 0.429 0.375 0.189 0.007 "ata from Burgos and illalon cheeses have been pooled. b F, regression mean square / residual mean square. c r 2, coefficient of determination (multiple r square). d P<0.001. e NS, not significant. CFU/g (126 samples) generally had log 10 staphylococci counts >4.0 CFU/g (125 samples), log 10 coliform counts >3.0 CFU/g (124 samples) and log 10 yeast counts >2.0 CFU/g (125 samples). ISCUSSION Levels of microorganisms found in most samples of Burgos and illalon cheeses exceeded standards suggested in Spain (75) for cheese made from pasteurized milk, which specify limits of 100 Staphylococcus aureus/ g, 100 Enterobacteriaceaelg and 10 Escherichia colilg. A large proportion also surpassed the proposed Canadian standards for cheese made from pasteurized milk (5), based on a 3-class acceptance plan with n = 5 and c = 2, which state m= 100 and = 1000 for S. aureus, m = 500 and =1500 for total coliforms and m=100 and = 500 for fecal coliforms. High counts of coliforms in cheese at the retail level have been reported previously. Collins-Thompson et al. (5) found 23.4% samples of Cheddar cheese, 13.6% of semisoft cheeses and 18.1% of soft cheeses, all made from pasteurized milk, exceeding 1600 coliforms/g. Frank and arth (6) detected coliform levels over 10 4 /g in 17.0% samples of semisoft and soft cheeses, whereas Asperger and Brandl (3) determined that 22.8% samples of semisoft cheese and 16.9% samples of soft cheese had more than 10 5 coliforms/g. ore recently, Arispe and Westhoff (2) reported in soft enezuelan queso bianco cheese, levels of coliforms, S. aureus and yeasts and molds very similar to those found in the present work for Burgos and illalon cheeses. anufacturing procedures of queso bianco cheese, where no starter culture is used, resemble those of Burgos and illalon cheeses. icrobial growth during storage may be enhanced by favorable environmental conditions in these white cheese varieties. No significant differences between microbial contents of Burgos and illalon cheeses, with the exception of yeast counts, were detected. Higher yeast levels present in illalon cheese may originate from contamination during brine-salting, as Burgos cheese either is not salted or is salted for a shorter period. Phosphatase activity found in Burgos and illalon cheeses does not seem to be related to the use of raw milk in cheesemaking, according to the low levels of significance of the phosphatase effect on most microbial groups. In case the presence of active phosphatase was due to unpasteurized milk total counts, staphylococci and coliforms should reach significantly higher numbers in samples with a positive phosphatase reaction than in samples with a negative reaction. Presence of active phosphatase may be explained either by reactivation of milk enzyme after heat treatment (76) or by production of phosphatase by cheese microflora (9). Reactivation of phosphatase in pasteurized milk products with a high fat content can occur when held at or above 10 C (12), which are temperatures frequently recorded for Burgos and illalon cheeses at retail outlets. It is unlikely that microbial phosphatase was produced because there was no significant effect of the presence of active phosphatase on total counts, staphylococci or coliforms. However, yeasts may be involved in phosphatase production, as a significant effect of the presence of active enzyme on counts of these microorganisms was observed. According to these results, the phosphatase test should only be interpreted as a presumptive test for evaluating the microbiological quality of white cheese. The 2.6-fold increase in coliform counts from April to July was statistically significant. Furthermore, the high percentages of coagulase-positive staphylococci and fecal coliforms observed in July imply a great degree of concern for public health. Storage temperature was shown to be the most important factor in controlling microbial growth in queso bianco cheese, as there is nothing in the manufacturing procedure able to ensure its microbiological safety (7). Temperatures recorded for Burgos and illalon samples reflect inadequate storage conditions at retail outlets, favorable for the development of contaminating microflora, with serious risks of cheeseborne intoxications and product spoilage. Length of storage and temperature values throughout the storage period probably have a greater effect on levels of most microbial groups than the actual temperature of the cheese at the time of sample collection. Cheese ph was the factor showing the closest relationship with microbial population in the samples. Because no starter culture is used in Burgos or illalon cheesemaking, the decrease of ph must be due to the metabolic activity of contaminating microflora. Arispe JOURNAL OF FOO PROTEC, OL. 48, OCTOBER 1985
BURGOS AN ILLALON CHEESES 869 and Westhoff (7) attributed the significant increases in queso bianco cheese acidity during storage to growth of native lactic microflora. Burgos and illalon samples with low ph value contained the highest numbers of total viable microorganisms, staphylococci, coliforms and yeasts, reflecting a parallel development of all microbial groups. Inoculation of a starter culture into the milk used in Burgos cheesemaking would give rise to high total counts but low levels of staphylococci and coliforms, because of the inhibition of these two groups by lactic acid bacteria (7). The determination of ph value in Burgos and illalon cheese samples may be of help for the analyst, as ph values <6.0 suggest severe microbial contamination of the samples. However, reliance on ph as the sole indicator of the microbiological quality of these cheeses does not seem convenient, due to the low coefficients of determination indicated in Table 4. As previously mentioned, total counts were significantly related to numbers of staphylococci, coliforms and yeasts in Burgos and illalon cheeses and, provided that lactic cultures are not added to the milk, they are a better indicator of the microbiological quality of these varieties than the result of the phosphatase test. Presently, limits for total counts could be included in a standard for Burgos and illalon cheeses, as low levels indicate good manufacturing practices and adequate storage conditions at the retail outlet. Garcia et al. (7) have recently reported a significant (P<0.001) effect of the inoculation of lactic starter, even at rates as low as 0.01%, on the development of S. aureus and Enterobacter cloacae during manufacture and storage of Burgos cheese at various temperatures. Present work at our laboratory deals with the possibility of improving the microbiological quality of Burgos and illalon cheeses, without altering their organoleptic characteristics, by inoculating selected strains of lactic acid bacteria, as previously achieved for anchego cheese (8). Should the use of lactic cultures in Burgos and illalon cheesemaking become established, total viable counts could no longer be accepted as an indicator of microbiological quality for these cheese varieties. ACKNOWLEGENTS We thank. P. Bermejo and R. Calvo for their helpful assistance. REFERENCES 1. Arispe, I., and. Westhoff. 1984. anufacture and quality of enezuelan white cheese. J. Food Sci. 49:1005-1010. 2. Arispe, I., and. Westhoff. 1984. enezuelan white cheese: composition and quality. J. Food Prot. 47:27-35. 3. Asperger, H., and E. Brandl. 1982. The significance of coliforms as indicator organisms in various types of cheese. Antonie van Leeuwenhoek J. icrobiol. 48:635-639. 4. Baird-Parker, A. C. 1962. An improved diagnostic and selective medium for isolating coagulase positive staphylococci. J. Appl. Bacteriol. 25:12-19. 5. Collins-Thompson,. L., I. E. Erdman,. E. illing,.. Burgener, U. T. Purvis, A. Loit, and R.. Coulter. 1977. icrobiological standards for cheese: survey and viewpoint of the Canadian Health Protection Branch. J. Food Prot. 40:411-414. 6. Frank, I. F., and E. H. arth. 1978. Survey of soft and semisoft cheese for presence of fecal coliforms and serotypes of enteropathogenic Escherichia coli. J. Food Prot. 41:198-200. 7. Garcia, B. E., L. E. Gaytan, F. J. Chavarri, and. Nunez. 1984. Comportamiento de Staphylococcus aureus y Enterobacter cloacae en queso de Burgos. I R. C. icrobiologia de los Alimentos, Pamplona, p. 105. 8. Gaya, P.,. edina, and. Nunez. 1983. Accelerated decrease of Enterobacteriaceae counts during ripening of raw milk anchego cheese by lactic culture inoculation. J. Food Prot. 46:305-308. 9. Hammer, B. W., and H. C. Olson. 1941. Phosphatase production in dairy products by microorganisms. J. ilk Technol. 4:83-85. 10. International airy Federation. 1974. ilk and milk products-count of coliform bacteria. International Standard FIL-IF 73: 1974, International airy Federation, Brussels, p. 9. 11. Kovacs, N. 1956. Identification of Pseudomonas pyocyanea by the oxidase reaction. Nature 178:703. 12. arth, E. H. (ed.). 1978. Standard methods for the examination of dairy products, 14th ed. American Public Health Association, Washington, C. 13. Pascual,. R. 1982. Tecnicas para el analisis microbiologico de alimentos y bebidas. inisterio de Sanidad y Consumo, adrid. p. 187. 14. Sperber, W. H., and S. R. Tatini. 1975. Interpretation of the tube coagulase test for identification of Staphylococcus aureus. Appl. icrobiol. 29:502-505. 15. Steel, R. G.., and J. H. Torrie. 1980. Principles and procedures of statistics, 2nd ed. cgraw-hill Book Co., New York. 16. Wright, R. C, and J. Tramer. 1953. Reactivation of milk phosphatase following heat treatment. I. J. airy Res. 20:177-188.