Penicillium Commune, P. Camembertii, the Origin of White Cheese Moulds, and the Production of Cyclopiazonic Acid J

Food Microbiology, 1986, 3, 363-371

Penicillium commune, P. camembertii, the origin of white cheese moulds, and the production of cyclopiazonic acid J. I. Pitt’, R. H. Cruickshankz and L. Leistners 1 CSIRO Division of Food Research, P.O. Box 52, North Ryde, N.S.W. 2113, Australia 2 Department of Agricultural Science, University of Tasmania, Hobart 7000, Australia, and 3 Federal Centre for Meat Research, 8650 Kulmbach, Federal Repub- lic of Germany Received 21 September 1986

The taxonomy of Penicillia producing the mycotoxin cyclopiazonic acid, including isolates classified as Penicillium aurantiogriseum and P. puberulum, is reviewed on the basis of morphology, physiology, mycotoxin production and isoenzyme profiles. It is concluded that P. puberulum, as neotypified by Pitt in his 1979 monograph, is a synonym of P. aurantiogriseum. The correct name for saprophytic Penicillia producing cyclopiazonic acid is P. commune with P. palitans as a synonym. The moulds used in the manufacture of white cheeses, which are all classified in P. camembertii, and which also produce cyclopiazonic acid, are domesticated fungi derived from P. commune.

Introduction cult to escape the conclusion that the The white moulds which are used in the strains with white conidia are mutants production of cheeses such as Camem- of the grey green: selected for, perpetu- bert, Brie, Weiljschimmelkase and Neuf- ated in, and apparently confined to, chatel have long been regarded as cheese manufacture’. domesticated fungi. Thorn (19061, in des- The origin of the moulds used in the cribing Penicillium camembertii, stated manufacture of white cheeses has ‘Persistent search has failed to find a aroused curiosity, for domesticated fungi single colony [of P. camembertiil in are rare. Pitt (1979) speculated that FRR America whose presence can be attri- 2160, an isolate from spoiled New buted to anything but Camembert Zealand cheddar cheese, was a ‘wild’ P. cheese imported from Europe’. These camembertii. This isolate differed in a moulds have been known by various number of morphological and micro- names in cheese industries around the scopic properties from the domesticated world: Thorn (1930) and Raper and Thorn species, but similarities suggested a com- (1949) accepted two species, P. caseicola, mon ancestor. which produces white conidia, and P. The origin of P. camembertii became of camembert& in which conidia are pale more than academic interest when it was grey. Samson et al. (1977) combined the discovered by Still et al. (1978) that two species under the older of the two white cheese moulds may produce the names, P. camembertii. Pitt (1979) mycotoxin cyclopiazonic acid. In a subse- agreed, considering that ‘the two species quent study of 69 P. camembertii isolates are so strikingly similar that it is diffi- of both white and grey spored types,

0740-0020/86/040363 + 09 $02.00/O 0 1986 Academic Press Inc. (London) Limited 364 J. I. Pitt et al. drawn from a wide variety of sources, species accepted by Pitt (1979). Frisvad Leistner and Eckardt (1979) failed to (1986) placed the neotype of P. puberu- find a single isolate which was not lum in his subspecies ‘P. auruntio- capable of cyclopiazonic acid production. griseum Group II’. In his view, the Apparently all known starter cultures of cyclopiazonic acid producers were not P. cumembertii used in cheese manufac- related to P. puberulum. ture can produce cyclopiazonic acid Williams and Pitt (19861, working (Scott 1981). The discovery of a P. from traditional taxonomic bases, camembertii ancestor which was not enlarged the concept of P. auruntio- mycotoxigenic could be of great potential griseum so that it effectively included all value in producing new strains suitable P. puberulum isolates, but they did not for cheese manufacture. resolve the fundamental confusion The taxonomy of Penicillia producing between isolates which produced penicil- cyclopiazonic acid has proved to be diffi- lit acid, and those which produced cyclo- cult. Leistner and Pitt (1977) reported piazonic acid. Penicillium cyclopium Westling (= P. Two studies which together resolved auruntiogriseum Dierckx) to be the main this question, and established the origins producer of this toxin, but some isolates of P. camembertii, are reported in this identified as P. puberulum Bainier, P. paper. One study was carried out in the viridicatum Westling, P. crustosum Federal Republic of Germany by two of Thorn and P. patulum Bainier (= P. us, while at the same time the third griseofulvum Dierckx) produced it also. (R. H. C.) independently worked in Tas- Pitt (1979) regarded all of these species mania with cultures supplied from the as distinct, though with the name collection at the first author’s institute. changes indicated above in brackets. Of particular interest here is his retention Materials and Methods of P. puberulum as a separate species Studies in culture from P. aurantiogriseum on the basis of About 250 isolates producing penicillic acid relatively minor differences in growth and/or S-toxin (Leistner 19841, or cyclopi- rates, colony texture and conidial col- azonic acid was studied in pure culture using ours. Toxin production was not con- the morphological and gross physiological sidered. methods of Pitt (1979). Cultures were grown on the standard plating regime of Czapek In their study of secondary metabolite yeast extract agar (CYA) at 5, 25 and 37”C, production by species in subgenus Peni- and on malt extract agar (MEA) and 25% cillum, Frisvad and Filtenborg (1983) glycerol nitrate agar at 25°C. Cultures were ‘provisionally included’ isolates of P. examined macroscopically and microscopi- puberulum in P. aurantiogrisum, a cally from CYA and MEA at 25°C after 7 days incubation, and colony diameters were species they regarded as the major pro- measured from all five standard conditions. ducer of penicillic acid. Isolates produc- Colour names in capitals and other colour ing cyclopiazonic acid were all assigned nomenclature used below are from the to a single species, P. camembert& which ‘Methuen Handbook of Colour’ (Kornerup they divided into two ‘groups’. ‘P. and Wanscher 1978). Nearly all of the cultures examined were from the collection at camembertii Group I’ included all of the the Federal Centre for Meat Research, Kulm- cheese moulds placed in this species by bath. modern taxonomists, while ‘P. camembertii Group II’, centred on FRR 2160, Mycotoxin assays was introduced for creatine positive iso- The cultures indicated above had previously lates (Frisvad, 1981) drawn from several been examined for mycotoxin production at Cheese moulds and cyclopkonic acid 365

Kulmbach, by thin layer chromatography MEA at 25°C for 7 days were consistently and other methods outlined by Leistner and greyish blue to blue green (24-25D-E3- Pitt (1977) and Leistner and Eckardt (1979). 4). Under the high power microscope, Enzyme profiles stipes from colonies grown on both media were smooth to finely roughened. This Production of pectic enzymes, amylases and ribonucleases were studied by growing cul- feature was not recognized by Pitt (1979) tures in the presence of suitable substrates, but was emphasized by Williams and citrus pectin or wheat grains, followed by Pitt (1986). Conidial colours of isolates electrophoretic separation at low tempera- producing cyclopiazonic acid, however, ture and visualization by staining. Isoen- were more greenish in colour: on CYA, zymes were then photographed and com- pared. The methods are given in detail by colours ranged from grey blue (24D-E3- Cruickshank and Pitt (Mycologia, submit- 4) to green (27331, while on MEA, col- ted). Cultures examined included type, ours could almost always be considered authentic and other cultures classified in to be true greens (26-27D-33-4). Stipes of relevant speciesby the first author, including cyclopiazonic acid producers were usu- some regarded as synonyms in Pitt (1979). All cultures came from the FRR collection at ally distinctly, though not prominently, CSIRO Division of Food Research, North roughened when examined from colonies Ryde, N.S.W., but included some originally on both CYA and MEA. It became clear obtained from the collection at Kulmbach. that two species of very similar appear- ance were being observed, separated readily by clear cut differences in myco- Results toxin production and, with care, on mor- Studies at Kulmbach phological criteria as well (Table 1). At Kulmbach, detailed taxonomic stud- Examination of a culture ex neotype of ies on isolates producing either penicillic P. puberulum present in the collection at acid and/or S-toxin showed that the great Kulmbach, Sp 916 (= FRR 20401, indi- majority could confidently be placed in cated that this should be placed in P. Penicillium aurantiogriseum. Conidial aurantiogriseum on these revised mor- colours of colonies grown on CYA and phological grounds, as it had been on the

Table 1. Features distinguishing PenicUkm isolates producing cyclopiazonic acid from those producing penicillic acid.

Mycotoxin Penicillic acid Cyclopiazonic acid Conidial colour, CYA Bluish green Bluish green to green 24-25D-33-48 24-27D-33-4 Conidial colour, MEA Bluish green Green 24-25D-E3-4 26-27D-33-4 Colony diameters, mmb absolute range, CYA 20-44 24-40 absolute range, MEA 16-41 18-35 80% of isolates, CYA 3637 30-37 80% of isolates, MEA 24-37 2330 Stipe roughening, CYA Smooth to finely rough Finely rough to rough (rarely rough) (rarely smooth) Conidia, length Not exceeding 4 pm Often up to 4.5 pm

* Colour codes from Kornerup and Wanscher (1978). b Data from 167 isolates producing penicillic acid, and 94 producing cyclopiazonic acid. Cultures examined included many old and deteriorating isolates, with colony diameters outside the ranges to be expected from fresh isolates. 366 J. I. Pitt et al. basis of secondary metabolism by Fris- in detail elsewhere (Cruickshank and vad and Filtenborg (1983). Examination Pitt, Mycologia, submitted). With regard of the FRR strain of the P. puberulum to the species of interest here, these neotype (FRR 2040) confirmed this. It studies independently produced the was now clear that the name P. puberu- same conclusions as those outlined Zum, neotypified by Pitt (1979) using an above: that isolates producing cyclopi- isolate (NRRL 1889) regarded as rep- azonic acid gave zymograms distinct resentative of the species by Raper and from those isolates ofP. aurantiogriseum Thorn (1949), is a synonym of P. aurun- which produced penicillic acid; that the tiogriseum. An equally important corol- neotype of P. puberulum produced zymo- lary is that the name P. puberulum is grams characteristic of P. aurantio- unavailable for the producers of cyclo- griseum; and that the type of P. palitans piazonic acid. produced zymograms characteristic of The collection at Kulmbach also con- the isolates forming cyclopiazonic acid tained a culture ex type of p. palitans (Fig. 1). Furth ermore, the cultures of P. Westling 1911, Sp 915 (= CBS 107.11, camembertii examined all produced IMI 40215, FRR 20331, which has been zymograms identical with, or very established to be a cyclopiazonic acid similar to, those of the other cyclopi- producer. Based principally on the exam- azonic acid producers. Further study of ination of several isolates regarded by type isolates of some species placed by Raper and Thorn (1949) as authentic for Pitt (1979) in synonymy with P. aurun- this species, Pitt (1979) had placed P. tiogriseum and p. puberulum showed palitans in synonymy with P. uiridi- that the type of P. commune Thorn 1910 catum; however, in the opinion of R. A. produced zymograms characteristic of Samson (pers. comm.) the type of P. the cyclopiazonic acid producers also palituns could not be classified in the (Fig. 1). This isolate (FRR 890) was latter species. Morphological examina- examined subsequently in the North tion of Sp 915 showed Samson’s opinion Ryde laboratory: its morphology was also to be correct: conidia on both CYA and characteristic of the cyclopiazonic acid MEA were dull green (2734 and 26fE4 producers. Moreover, it produced a low respectively), with definitely rough level of cyclopiazonic acid, despite hav- stipes. The early original publication ing been maintained in culture for most date for P. palitans, and its morphology, of the past 80 years. It can be confidently together with the production of cyclo- concluded that the earliest identifiable piazonic acid, indicated that it was an valid names for Penicillium isolates pro- appropriate name for the cyclopiazonic ducing cyclopiazonic acid are P. camem- acid producers. The possibility that an bertii and P. commune. earlier valid and typified name existed could not be dismissed, but deterioration of old types in culture made accurate Discussion assessment of earlier species names diffi- It has been shown above that the earliest cult. recognizable species producing cyclo- At the same time as the studies out- piazonic acid are P. camembertii and P. lined above were in progress, studies on commune. It is considered both logical the enzyme patterns of these and other and expedient to reserve the name P. species in subgenus Penicillium were camembertii for the domesticated moulds being undertaken at the University of with which cheeses such as Camembert Tasmania. These studies will be reported and Brie are produced, and to revive the Cheese moulds and cyclopiazonic acid 367

I 2 3 4 5 6 7 8 9 lo 11 12 13 14 15 16 17 Fig. 1. Pectic (polygalacturonase) zymograms of Penicillium isolates. Lanes l-10, cyclopiazonic acid producers; lanes 11-17, penicillic acid producers. 1, P. pulituns, FRR 2033, type isolate; 2, P. commune, NRRL 890a, type isolate; 3-7, various isolates grouped in P. commune by zymogram; 8, P. camembertii, FRR 2160, considered to be a ‘wild type’ by Pitt (1979), now placed in P. commune; 9, P. ccznembertii, recently isolated from Camembert cheese; 10, P. camembertii, FRR 877, type isolate; 11, P. auruntiogriseum, FRR 971, neotype isolate; 12, P. puberulum, FRR 2040, neotype isolate; 13-17, isolates grouped in P. aurantiogriseum by zymogram. name P. commune for isolates which are been confused. Thorn (1910,193O) regar- not specifically cheese moulds but which ded P. commune as a ubiquitous species. occur as ubiquitous saprophytes. The Raper and Thorn (19491, however, classi- common origin of these two species is fied it in their subsection Lanatu, to evident, both from their production of a which floccose (and therefore deteriorat- single mycotoxin, rarely produced by ing) species were consigned. It is prob- other Penicillium species and, perhaps able that, in their classification, produc- more convincingly, by unique and virtu- ers of cyclopiazonic acid were assigned to ally identical patterns of zymograms P. cyclopium, P. puberulum, P. palitans (Fig. 1). If it is accepted that P. camem- (perhaps only the type), and also other bertii is a domesticated species, then P. floccose species as well as P. commune. commune must logically be its ancestral Samson et al. (1977) maintained P. com- wild type. It is interesting and relevant mune as a rare species characterized by that the type isolate of P. commune was its floccose habit and ellipsoidal conidia. isolated from cheese, and that FRR 2160, In their taxonomy, most freshly isolated regarded by Pitt (1979) as a wild P. cyclopiazonic acid producers would be camembertii, was also isolated as a identified, with a variety of other myco- cheese spoilage fungus. Taking into toxigenic isolates, as P. verrucosum var. account the extreme age and floccose cyclopium. Pitt (1979) assigned the cyclo- habit of FRR 890, the type of P. com- piazonic acid producers to either P. mune, the morphological and physio- aurantiogriseum or P. puberulum, or in logical resemblance of this isolate to one case, FRR 2160, to P. camembertii, FRR 2160 is quite striking. while Frisvad and Filtenborg (1983) logi- Previously, the taxonomy of the spe- cally created ‘P. camembertii Group II cies referred to here as P. commune has for them. Williams and Pitt (1986) as- 368 J. I. Pitt et al.

Fig. 2. Penicillium aurantiogriseum: (A) colonies on Czapek yeast extract agar and malt extract agar at 25”C, 7 days; (C) penicillus, x750; CD) conidia, x 1875. Penicillium commune: 03) colonies on Czapek yeast extract agar and malt extract agar at 25”C, 7 days; iE) penicillus, Cheese moulds and cyctopiazonic acid 369 signed them to an enlarged P. aumntio- (Westling) Samson et al., Stud. Mycol., griseum. Baarn 11: 37,1976. More than 160 isolates known to pro- Penicillium commune Thom duce cyclopiazonic acid, but not identifi- Bull. Bur. Anim. Ind. US Dep. Agric. able as P. camembertii, have been exam- 118: 56, 1910. ined in this study. A high proportion of Penicillium palitans Westling, Ark. Bot. them, in excess of 80%, can‘be identified 11: 83, 1911. morphologically as a single species, now Penicillium lanosum Westling, op. cit. 11: recognized as P. commune. This species 97, 1911. is very similar morphologically to P. Penicillium lanosogriseum Thorn, Penicil- aumntiogriseum, usually differing from lia: 327, 1930. the latter by the features detailed in Colonies on CYA 30-37 mm diam, Table 1, although some degree of mor- radially sulcate, moderately deep, tex- phological overlap occurs. P. commune ture velutinous to fasciculate; mycelium also resembles P. uiridicatum: P. viridi- white, usually inconspicuous; coni- cutum produces brighter green conidial diogenesis moderate to heavy, Greyish colours and more delicate penicilli. A Turquoise to Dull Green (24-25D-33-4); small number of isolates have been exudate usually conspicuous, clear or examined which morphologically are pale brown; soluble pigment produced by indistinguishable from P. uiridicutum, some isolates, brown to reddish brown; but which produce cyclopiaronic acid. In reverse pale, light to brilliant orange, or our present state of knowledge it is reddish to violet brown. Colonies on considered preferable to maintain the MEA 24-37mm diam, plane or rarely morphological basis to this species, and radially sulcate, low and relatively to accept that a minority of P. uiridi- sparse, surface texture velutinous to cutum isolates produce cyclopiasonic fasciculate; mycelium usually subsur- acid rather than to attempt to ‘force’such face, occasionally conspicuous and then isolates into P. commune. bright yellow; conidiogenesis usually Descriptions of P. uuruntiogriseum, as moderate to heavy, Greyish Turquoise to now emended, and P. commune, as now revived, follow. Both species are illus- Dull Green (24-25D-34-5); soluble pigment sometimes produced, yellow brown trated in Fig. 2. to reddish brown; reverse pale, orange, or reddish brown. Colonies on G25N Penicillium aurantiogriseum Dierckx 16-24mm diam, usually radially sul- Annls Sot. Sci. Brux. 25: 88, 1901. cate, moderately deep, dense, velutinous Penicillium puberulum Bainier, Bull. to fasciculate; reverse pale, yellow or trimest. Sot. mycol. Fr. 23: 16, 1907. brown. At 5”, colonies 2-5 mm diam, of white mycelium. No growth at 37”. Penicillium cyclopium Westling, op cit. Conidiophores borne singly or in fas- 11: 90, 1911. cicles, mostly from subsurface hyphae, Penicillium aurantiovirens Biourge, Cell- stipes 200-400 pm long, or of indetermi- ule 33: 119, 1923. nate length in fascicles, with walls Penicillium marten&i Biourge, op. cit. 33: smooth to finely roughened, only rarely 152,1923. rough, bearing terminal terverticillate or less commonly biverticillate penicilli; Penieillium lanoso-coeruleum Thorn, Peni- rami 15-25(-30) pm long; metulae cillia: 322, 1930. lO-15(-l@ pm long; phialides slender, Penicillium verrucosum var. cyclopium ampulliform, mostly 7-10 pm long; coni- 370 J. I. Pitt et al.

dia spherical to subspheroidal, less com- Conidiophores borne singly or in fas- monly ellipsoidal, usually 3.W.0 ym cicles, mostly from subsurface hyphae, long, with smooth walls, mostly borne in stipes 200-400 urn long or of indetermi- long, well defined columns. nate length in fascicles, with walls finely Colonies on CYA 30-37 mm diam, to conspicuously roughened, typically radially sulcate, usually fasciculate, less bearing terminal terverticillate peni- commonly velutinous; mycelium white, cilli; rami X-20(-30) pm long; metulae usually inconspicuous; conidiogenesis lO-15(-18) urn long; phialides ampulli- moderate, of variable colour, Greyish form, 9-11 pm long; conidia spherical, Turquoise to Dull Green (24-27D-F3-5); uncommonly subspheroidal, 35-4.0(-5) exudate usually present, clear to pale urn diam, smooth walled, borne in dis- brown; soluble pigment not produced; ordered chains. reverse usually pale, occasionally yellow, brown or purple. Colonies on MEA 23-30 ‘mm diam, plane or lightly sulcate, low and dense, surface velutinous or Acknowledgements lightly fasciculate; mycelium inconspi- Support from the Adalbert Raps Founda- cuous, white; conidiogenesis moderate, tion, Kulmbach, Federal Republic of Dull Green (26-27D-33-4); exudate Germany, is gratefully acknowledged by and soluble pigment absent; reverse J. I. P. during studies at the Federal usually uncoloured. Colonies on G25N Centre for Meat Research, Kulmbach. 18-22 mm diam, plane, sulcate or wrink- Support for R. H. C. came from the Rural led, low to moderately deep, dense, usu- Credits Development Fund of the ally fasciculate; mycelium white to yel- Reserve Bank of Australia, and a lowish; reverse pale to orange brown. At research grant from the University of 5”, at least microcolony formation; typic- Tasmania. Ailsa D. Hocking supplied ally colonies of 24 mm diam formed. No cultures from the FRR culture collection growth at 37”. for R. H. C.

References Cruickshank, R. H. and Pitt, J. I. (1986) Identification of terverticillate Penicillium speciesby enzyme electrophoresis. Mycologia (submitted). Frisvad, J. C. (1981) Physiological criteria and mycotoxin production as aids in identification of common asymmetric Penicillia. Appl. Environ. Microbial. 41, 568-579. Frisvad, J. C. (1986) Protiles of primary and secondarymetabolites of value in classification of Penicillium viridicatum and related species.In Advances in Penicillium and Aspergillus systematic% (Eds Samson, R. A. and Pitt, J. I.) pp. 311-325. New York, Plenum Press. Frisvad, J. C. and Filtenborg, 0. (1983) Classification of terverticillate Penicillia based on profiles of mycotoxins and other secondary metabolites. Appl. Environ. Microbial. 46, 1301-1310. Kornerup, A. and Wanscher, J. H. (1978) Methuen handbook ofcolour. 3rd ed. London, Eyre Methuen. Leistner, L. (1984) Toxigenic Penicillia occurring in foodsand feeds: a review. Food Z’echnol. Aust. 36,404-406, 413. Leistner, L. and Eckardt, C. (1979) Vorkommen toxigener Penicillien bei Fleischerzeugnis- sen. Fleischwirtschaft 59, 1892-1896. Leistner, L. and Pitt, J. I. (1977) Miscellaneous Penicillium toxins. In Mycotmins in human and animal health. (Eds Rodricks, J. V., Hesseltine, C. W. and Mehlman, M. A.) pp. 639-653. Park Forest South, Illinois, Pathotox Publishers. Pitt, J. I. (1979) The genus Penicillium and its teleomorphic states Eupenicillium and Talaromyces. London, Academic Press. Cheese moulds and cyclopiazonic acid 371

Raper, K. B. and Thorn, C. (1949) A manual ofthePeniciZZia. Baltimore, Williams and Wilkins CO. Samson, R. A., Eckardt, C. and Orth, R. (1977) The taxonomy of Penicillium speciesfrom fermented cheeses.Antonie van Leeuwenhoek 43,341-350. Scott, P. M. (1981) Toxins of Penicillium speciesused in cheesemanufacture. J. Food Protect. 44,702-710. Still, P., Eckardt, C. and Leistner, L. (1978) Bildung von Cyclopiazonsaure durch Penicillium camembertii-Isolate von K&e. Fleischwirtschaft 58, 87G-877. Thorn, C. (1906) Fungi in cheeseripening: Camembert and Roquefort. Bull. Bur. Anim. Znd. U.S. Dep. Agric. 82, l-39. Thorn, C. (1910) Cultural studies of speciesof Penicillium. Bull. Bur. Anim. Znd. U.S. Dep. Agric. 118, l-109. Thorn, C. (1930) The Penicillia. Baltimore, Williams and Wilkins Co. Williams, A. P. and Pitt, J. I. (1986) A revised key to Penicillium subgenusPenicillium. In Advances in Penicillium and Aspergillus systematics. (Eds Samson,R. A. and Pitt, J. I.) pp. 129-134. New York, Plenum Press.