Production of PR Toxin and Roquefortine by Penicillium Roqueforti Isolates from Cabrales Blue Cheese

118

Journal of Food Protection, Vol. 48, No. 2, Pages 118-121 (February 1985) Copyright® International Association of Milk, Food, and Environmental Sanitarians

Production of PR Toxin and Roquefortine by Penicillium roqueforti Isolates from Cabrales Blue Cheese

MARGARITA MEDINA, PILAR GAYA and M. NUNEZ*

Departamento de Bioquimica y Microbiologia, Instituto Nacional de Investigaciones Agrarias Apartado 8111, Madrid 28040, Spain Downloaded from http://meridian.allenpress.com/jfp/article-pdf/48/2/118/1656557/0362-028x-48_2_118.pdf by guest on 25 September 2021 (Received for publication July 13, 1984)

ABSTRACT was the predominant species. The possibility of selecting non-toxic strains of P. roqueforti for use in large-scale PR toxin production in yeast extract-sucrose broth by 33 manufacture of standardized Cabrales cheese led us to in­ Penicillium roqueforti isolates from Cabrales blue cheese was vestigate mycotoxin synthesis by our isolates under dif­ quantified by a disc assay technique with Bacillus megaterium ferent cultural conditions. In the present paper we report NRRL B-1368 as the test organism. Isolates from the interior information on PR toxin and roquefortine production. of the cheese reached an average production of 1,89 mg PR toxin/100 ml, whereas the mean level of isolates from the sur­ MATERIALS AND METHODS face was 1.64 mg/100 ml. Roquefortine production in the same broth by these isolates was quantified by a similar technique, Organisms with Bacillus stearothermophilus DSM 22 as the test organism. Thirty-three isolates of P. roqueforti, obtained in previous Mean production of roquefortine was 0.18 mg/100 ml for P. work (9) from the interior (21 isolates) and the surface (12 iso­ roqueforti isolates from the interior and 0.09 mg/100 ml for lates) of Cabrales blue cheeses from two vats (A and B), each isolates from the surface of the cheese. If lactose or sodium from a different producer, were investigated. P. roqueforti lactate replaced sucrose in the growth medium, levels of both strains ATCC 6987, ATCC 6989 and NRRL 849, and P. toxins decreased considerably. The identity of PR toxin and roqueforti strains 2004, 2007, 2012, 2013 and 2014 from the roquefortine in crude extracts was confirmed by thin-layer Dairy Organisms Culture Collection of the Instituto Nacional chromatography. de Investigaciones Agrarias were also studied. Cultures were maintained at 4°C on YM agar (23). Bacillus megaterium NRRL B-1368 and Bacillus stearother­ Penicillium roqueforti is responsible for the greenish mophilus DSM 22, obtained from B. Kopp (Institut fur Mik- blue moldy aspect characteristic of cheeses such as robiologie, Miinster University, West Germany) were used as Roquefort, Bleu, Stilton, Gorgonzola, etc. Tolerance of test organisms for determination of PR toxin and roquefortine high salt and low oxygen concentrations enables the concentrations, respectively. They were maintained at 4°C on nutrient agar and on Kopp's sporulation agar (personal com­ selective development of this species in the interior of munication), respectively. blue-veined cheese varieties. Extracellular proteinases and lipases formed by P. roqueforti largely contribute to the Determination of PR toxin body and flavor changes which occur during ripening. Roux flasks (1 L) containing 150 ml of sterile YES broth Since the work by Kanota (5) on toxic metabolites of (2% yeast extract, 15% sucrose) were inoculated with 1 ml of P. roqueforti, interest and research on production of a 7-d-old culture of P. roqueforti in the same broth and incu­ mycotoxins by this species have increased considerably bated aerobically in the dark at 24°C for 14 d as stationary cul­ due to its importance for the dairy industry. Some of tures. Isolates showing the highest PR toxin production in YES these toxins, mainly roquefortine, isofumigaclavine A, broth were also grown in YEL (2% yeast extract, 15% lactose) mycophenolic acid and ferrichrome, have been detected and YESL (2% yeast extract, 5% sodium lactate) broths. After in blue cheese at low levels, whereas others, such as incubation, crude extracts of the cultures were obtained accord­ penicillic acid and PR toxin, are unstable in cheese, ac­ ing to Wei and Liu (21). cording to the conclusions drawn by Scott (75). B. megaterium NRRL B-1368 was grown in the sporulation Cabrales cheese is a blue-veined variety manufactured broth of Millet and Aubert (7) at 30°C for 40 h as a shake-cul­ in Northern Spain, from a mixture of raw cow's, ewe's ture. Seeded tryptone-glucose-yeast extract agar plates were pre­ and goat's milk, without inoculation of lactic starter or pared and 6-mm diameter paper disks containing known quan­ mold spores into milk or curd. Changes in its microflora tities of crude extract were placed on them as described by Cle­ throughout ripening have been studied by Nunez (8). ments (2). After incubation of the plates at 30°C for 24 h, di­ Mold isolates from the interior and the surface of the ameters of the inhibition zones were measured. Inhibition by cheese were identified (9), concluding that P. roqueforti disks containing known quantities of PR toxin standard (Makor

JOURNAL OF FOOD PROTECTION, VOL. 48, FEBRUARY 1985 TOXIN PRODUCTION BY PENICILLIUM ROQUEFORTI 119

Chemicals Ltd., Jerusalem) was also determined (3 replications) Statistical analyses and data obtained used to calculate the equation of the best Linear regression equations, correlation coefficients and sig­ fitting line by the least-squares criterion. Quantities of PR toxin nificance of differences between means according to the Stu­ in crude extracts from P. roqueforti cultures were estimated by dent's t test were calculated after Steel and Torrie (77). means of this equation. Thin-layer chromatography (TLC) of PR toxin in crude extracts was performed and the Rf value de­ RESULTS termined according to the procedure of Wei and Liu (21). Relationship between quantity of toxins and diameter of Determination of roquefortine inhibition zones P. roqueforti isolates were grown as previously described for Equations of the best fitting lines for the relationships determination of PR toxin, and crude extracts of roquefortine between the amounts of PR toxin or roquefortine and the were prepared as indicated by Wagener et al. (79). diameters of the inhibition zones of B. megaterium or B. B. stearothermophilus DSM 22 was grown in Kopp's spol­ stearothermophilus, respectively, and the correlation iation broth at 50°C for 24 h. Disks containing known quantities of crude extracts were placed on plates of sporulation agar coefficients are shown in Table 1. Both correlation coef­ Downloaded from http://meridian.allenpress.com/jfp/article-pdf/48/2/118/1656557/0362-028x-48_2_118.pdf by guest on 25 September 2021 seeded with 10% of the B. stearothermophilus culture. Plates ficients were highly significant (P<0.01). were incubated for 24 h at 45°C and diameters of the inhibition zones measured. An equation to calculate the amount of Production of PR toxin and roquefortine on sucrose roquefortine present in crude extracts was obtained as described Tables 2 and 3 report the amounts of PR toxin and for PR toxin, using purified roquefortine provided by B. Kopp roquefortine produced in YES broth by P. roqueforti iso­ as the standard. TLC of roquefortine in crude extracts was per­ lates from the interior and the surface of Cabrales cheese, formed using 3 solvents and Rf values determined following the respectively. Mean production of PR toxin was 1.89 mg/ procedures of Scott and Kennedy (75). 100 ml for 21 isolates from the interior of the cheese, 1.64 mg/100 ml for 12 isolates from the surface, 1.38 TABLE 1. Regression equations of the diameters of the inhibi­ mg/100 ml for 12 isolates from vat A and 2.03 mg/100 tion zones of Bacillus megaterium or B. stearothermophilus on ml for 21 isolates from vat B. Average levels of the quantities of PR toxin or roquefortine standards. roquefortine were 0.18, 0.09, 0.18 and 0.13 mg/100 ml D, = 27.68+ 9.65 log Q, r,=0.968 for isolates from the interior, the surface, vat A and vat D2 = 32.04 +10.73 log Q2 r2 = 0.938 B, respectively. No significant differences were detected between interior and surface or between vats for any of D,: diameter in mm of the inhibition zone of B. megaterium. the two toxins, according to the Student's t test. D2: diameter in mm of the inhibition zone of B. stearother­ mophilus. Correlation coefficients between amount of crude ex­ Qi: mg PR toxin/disk; Q2: mg roquefortine/disk. tract and amount of toxin were calculated for PR toxin r1( r2: correlation coefficients. (r = 0.140) and roquefortine (r = 0.345), the latter being

TABLE 2. Production of PR toxin and roquefortine by 21 P. roqueforti isolates from the interior of Cabrales cheese Crude extract* Crude extract" Isolate of PR toxin PR toxin" of roquefortine Roquefortine 2026 6.67 0.59 16.47 0.07 2031 7.67 1.46 22.07 0.16 2032 7.00 0.57 19.07 0.45 2033 6.93 2.52 16.73 0.10 2034 6.47 0.09 18.27 0.17 2035 5.60 0.07 13.07 0.33 2036 9.47 0.48 10.33 0.21 2037 9.07 0.93 20.27 0.18 2038 8.07 0.80 10.47 0.12 2039 17.13 0.01 10.00 0.20 2040 8.20 0.00 16.93 0.48 2043 13.70 1.99 12.47 0.11 2046 21.73 0.05 32.80 0.26 2047 18.73 7.05 19.27 0.40 2048 20.73 2.55 7.20 0.04 2049 17.13 0.11 13.87 0.00 2050 13.00 0.01 9.53 0.09 2051 7.33 0.01 10.13 0.15 2057 12.73 20.30 6.20 0.06 2058 7.07 0.01 4.27 0.09 2059 3.87 0.00 7.27 0.06 "nig crude extract/100 ml YES broth. bmg toxin/100 ml YES broth.

JOURNAL OF FOOD PROTECTION, VOL. 48, FEBRUARY 1985 120 MEDINA, GAYA AND NUNEZ

TABLE 3. Production of PR toxin and roquefortine by 12 P. roqueforti isolates from the surface of Cabrales cheese. Crude extract*1 Crude extract" Isolate of PR toxin PR toxinb of roquefortine Roquefortineb 2028 3.53 2.27 21.73 0.00 2029 9.93 0.30 14.33 0.32 2030 6.73 6.47 4.13 0.02 2041 9.80 0.37 20.60 0.17 2042 10.73 0.01 14.93 0.07 2044 3.87 0.35 17.47 0.06 2045 10.73 0.00 16.53 0.16 2052 10.27 0.00 7.87 0.00 2053 2.87 0.18 10.53 0.12 2054 5.73 4.08 10.40 0.12 2055 5.60 1.40 10.67 0.03 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/48/2/118/1656557/0362-028x-48_2_118.pdf by guest on 25 September 2021 205(5 6^00 4^20 21.53 0.00 "trig crude extract/100 ml YES broth. bmg toxin/100 ml YES broth. statistically significant (P<0.05). The correlation coeffi­ TABLE 4. Production of PR toxin in YES, YEL and YESL cient between quantities of PR toxin and roquefortine of broths by P. roqueforti isolates from Cabrales cheese. the different isolates (r= -0.128) was not significant. Isolate mg PR toxin/100 ml broth Presence of PR toxin was detected by TLC in most YES" YELb YESLC crude extracts (79%) from cultures of isolates producing over 2 mg/100 ml, whereas it was not detected in any 2004 24.62 0.00 0.10 of the extracts from cultures of isolates producing under 2007 18.40 1.13 0.03 2012 49.17 1.44 0.00 1 mg/100 ml. All the isolates were positive for roquefor­ 2013 13.51 0.00 0.05 tine production after examination of crude extracts by 2014 34.49 1.03 0.08 TLC procedures. 2057 20.30 0.23 0.00 aYES broth: 2% yeast extract, 15% sucrose. Production of PR toxin and roquefortine on lactose and bYEL broth: 2% yeast extract, 15% lactose. lactate CYESL broth: 2% yeast extract, 5% sodium lactate. Quantities of PR toxin and roquefortine produced in TABLE 5. Production of roquefortine in YES, YEL and YESL YEL and YESL broths by P. roqueforti isolates which broths by P. roqueforti isolates from Cabrales cheese. exhibited the highest toxin production on YES broth are given in Tables 4 and 5, respectively. If lactose replaced Isolate my roquefortine/100 ml broth a a sucrose in the culture medium, PR toxin synthesis was YES YEL YESL considerably reduced, with an average production for the 2035 0.33 0.07 0.03 six isolates 42 times lower than with sucrose. PR toxin 2038 0.12 0.15 0.02 production with sodium lactate as carbon source de­ 2039 0.20 0.04 0.04 creased further, with a mean value 617 times lower than 2040 0.48 0.03 0.03 with sucrose. A similar effect was recorded when P. 2051 0.15 0.00 0.04 roqueforti strains ATCC 6987, ATCC 6989 and NRRL 2053 0.12 0.11 0.02 849 were tested for PR toxin, with average production aYES, YEL and YESL broths: see Table 4 for composition. for the three strains of 41.73 mg/100 ml in YES broth, 0.48 mg/100 ml in YEL broth and undetectable amounts from Cabrales cheese were usually lower than values re­ in YESL broth. ported for isolates from blue cheese by Scott et al. (7(5), Roquefortine production was influenced to a lesser de­ in the range 11.5-17.5 mg/100 ml, or for ATCC strains, gree than PR toxin synthesis by the carbon source. Mean which produced 21-30 mg/100 ml, according to Wei and levels of roquefortine on lactose and sodium lactate were Liu (21). Nevertheless, some of the isolates studied in 3.5 and 7.8 times lower, respectively, than on sucrose. the present work, Nos. 2030, 2047, 2054, 2056 and 2057, were able to excrete considerable amounts of PR DISCUSSION toxin under favorable conditions, and five isolates previ­ ously obtained from Cabrales cheese (see Table 4) also Patterns of standard curves similar to the equations pre­ exhibited a high PR toxin production. No relationship be­ sented in Table 1 have been proposed for determination tween the origin of the isolate and the ability to synthe- of concentrations of patulin with B. megaterium (18) and tize PR toxin was found, as differences between isolates of roquefortine with B. stearothermophilus (6), as test or­ from the interior and the surface of the cheese or between ganisms. isolates from cheeses of various producers were not Levels of PR toxin excreted by P. roqueforti isolates statistically significant. The observed influence of carbon

JOURNAL OF FOOD PROTECTION, VOL. 48, FEBRUARY 1985 TOXIN PRODUCTION BY PENICILLIUM ROQUEFORTI 121 source on PR toxin synthesis is in accordance with previ­ REFERENCES ously published data (4,10,16), sucrose being by far the 1. Arnold, D. L., P. M. Scott, P. F. McGuire, J. Harwig, and E. best substrate. A. Nera. 1978. Acute toxicity studies on roquefortine and PR Surveys of commercial blue cheese samples (3,11) or toxin, metabolites of Penicillium roqueforti, in the mouse. Food studies on cheeses manufactured with P. roqueforti Cosmet. Toxicol. 16:369-371. strains able to produce the toxin when grown on YES 2. Clements, N. L. 1968. Note on a microbiological assay for aflato- broth (3) report the absence of PR toxin from cheese. xin B,: a rapid confirmatory test by effects on growth of Bacillus megaterium. J. Assoc. Off. Anal. Chem. 51:611-612. The instability of PR toxin in some acidic solvent solu­ 3. Engel, G., and D. Prokopek. 1979. Kein Nachweis von Penicil­ tions or in blue cheese itself, where it may react with lium roqueforti-Toxin in Kase. Milchwissenschaft 34:272-274. ammonia or ammonium salts forming PR imine (14,22), 4. Guiraud, I., M.-T. Piva, J. Crouzet, and P. Galzy. 1977. Action accounts for these negative results. PR imine and PR de quelques facteurs du milieu sur l'excretion de la P.R. toxine. toxin analogues obtained by mixing PR toxin with L- Ann. Nutr. Alim. 31:551-555. leucine or L-a-alanine showed a lower toxicity in mice 5. Kanota, K. 1970. Studies on toxic metabolites of Penicillium roqueforti. pp. 129-132. In M. Herzberg (ed.), Proc. first U.S.- following intraperitoneal administration than PR toxin lapan conf. toxic microorganisms. U.S. Dept. Interior, Wash­ Downloaded from http://meridian.allenpress.com/jfp/article-pdf/48/2/118/1656557/0362-028x-48_2_118.pdf by guest on 25 September 2021 V). ington, DC. Roquefortine was produced at lower levels than PR 6. Kopp, B., and H.-J. Rehm. 1979. Ein biologischer Test zur quan- titativen Bestimmung von Roquefortin. Z. Lebens. Unters. Forsch. toxin by P. roqueforti isolates from Cabrales cheese. 169:90-91. Nevertheless, synthesis of roquefortine by 29 of the 33 7. Millet, J., and J.-P. Aubert. 1960. In J. R. Norris and D. W. isolates investigated was detected by the disc assay tech­ Ribbons (eds.). Methods in Microbiology Vol. 6A p. 334. nique, and all isolates were roquefortine producers after Academic Press, New York. examination by TLC procedures. Moreover, carbon 8. Nunez, M. 1978. Microflora of Cabrales cheese: changes during source requirements for roquefortine synthesis were less maturation. J. Dairy Res. 45:501-508. 9. Nunez, M., M. Medina, P. Gaya, and C. Dias-Amado. 1981. Les restrictive than for PR toxin, with a mean production of levures et les moisissures dans le fromage bleu de Cabrales. Le roquefortine in YEL broth only 3.5 times lower than in Lait 61:62-79. YES broth. Levels of roquefortine formed by our isolates 10. Piva, M.-T., J. Guiraud, J. Crouzet, and P. Galzy. 1976. Influence were generally lower than values reported by Scott et al. des conditions de culture sur l'excretion d'une mycotoxine par quel­ (16), in the range 0.5-12 mg/100 ml broth, or than the ques souches de Penicillium roqueforti. Le Lait 56:397-406. 11. Polonelli, L., G. Morace, F. Delle Monache, and R. A. Samson. production obtained by Schoch et al. (12), in the range 1978. Studies on the PR toxin of Penicillium roqueforti. 2.4-80 mg/kg. No relationship between origin of the iso­ Mycopathologia 66:99-104. late and level of roquefortine was observed. 12. Schoch, U., J. Liithy, and C. Schlatter. 1984. Mykotoxine von Presence of roquefortine in cheese has been detected P. roqueforti und P. camemberti in Kase. I. Vorkommen chemisch by various workers. A maximum of 6.8 mg/kg in a blue identifizierter Mykotoxine. Milchwissenschaft 39:76-80. 13. Scott, P. M. 1981. Toxins of Penicillium species used in cheese cheese sample was recorded by Scott and Kennedy (75), manufacture. J. Food Prot. 44:702-710. with roquefortine present in all 16 blue cheese samples 14. Scott, P. M., and S. R. Kanhere. 1979. Instability of PR toxin examined, whereas Ware et al. (20) found 0.42 mg/kg in blue cheese. J. Assoc. Off. Anal. Chem. 62:141-147. as an average level for 12 samples and Schoch et al. (12) 15. Scott, P. M., and B. P. C. Kennedy. 1976. Analysis of blue a mean content of 0.93 mg/kg in 13 samples of blue vei­ cheese for roquefortine and other alkaloids from Penicillium ned cheeses. roqueforti. J. Agric. Food Chem. 24:865-868. 16. Scott, P. M., B. P. C. Kennedy, J. Harwig, and B. J. Blanchfield. Environmental conditions (temperature, aeration, car­ 1977. Study of conditions for production of roquefortine and other bon source, etc.) in cheese are far from optimal for PR metabolites of Penicillium roqueforti. Appl. Environ. Microbiol. toxin and roquefortine production. If the instability of PR 33:249-253. toxin in blue cheese (14,22), the concentrations of 17. Steel, R. G. D., and J. H. Torrie. 1980. Principles and procedures of statistics, 2nd ed. McGraw-Hill Book Co., Inc., New York. roquefortine detected in cheese (12,15,20) and the LD50 18. Stott, W. T., and L. B. Bullerman. 1975. Microbiological assay values of roquefortine in mice (1) are taken into account, of patulin using Bacillus megaterium. J. Assoc. Off. Anal. Chem. health hazards due to P. roqueforti metabolites seem con­ 58:497-499. siderably reduced. Moreover, use in large-scale Cabrales 19. Wagener, R. E., N. D. Davis, and U. L. Diener. 1980. Penitrem cheese manufacture of P. roqueforti strains unable to pro­ A and roquefortine production by Penicillium commune. Appl. En­ duce mycotoxins under favorable culture conditions elimi­ viron. Microbiol. 39:882-887. 20. Ware, G. M., C. W. Thorpe, and A. E. Pohland. 1978. Determi­ nates any potential risk of mycotoxicosis caused by nation of roquefortine in blue cheese and blue cheese dressing by cheese consumption. HPLC and UV and electrochemical detectors. Abstr. 92nd Ann. Meeting Assoc. Off. Anal. Chemists, Washington, DC, p. 50. ACKNOWLEDGMENTS 21. Wei, R.-D., and G.-X. Liu. 1978. PR toxin production in different Penicillium roqueforti strains. Appl. Environ. Microbiol. 35:797- The authors are indebted to M. Muzquiz and I. Rodenas for their 799. helpful assistance with TLC techniques. 22. Wei, R.-D., H. K. Schnoes, P. A. Hart, and F. M. Strong. 1975. The structure of PR toxin, a mycotoxin from Penicillium roqueforti. Tetrahedron 31:109-114. 23. Wickerham, L. J. 1951. Taxonomy of yeasts. Tech. Bull. U.S. Dept. Agriculture No. 1029, Washington, DC.

JOURNAL OF FOOD PROTECTION, VOL. 48, FEBRUARY 1985