and P. crustosum isolated in Canada. Mark W. Sumarah1, J. David Miller1 & Barbara A. Blackwell2. 1Ottawa-Carleton Institute of Chemistry, Carleton University, ...
Springer 2005
Mycopathologia (2005) 159: 571–577 DOI: 10.1007/s11046-005-5257-7
Isolation and metabolite production by Penicillium roqueforti, P. paneum and P. crustosum isolated in Canada Mark W. Sumarah1, J. David Miller1 & Barbara A. Blackwell2 1
Ottawa-Carleton Institute of Chemistry, Carleton University, Steacie Building, K1S 5B6, Ottawa, Ontario, Canada; 2Eastern Cereal and Oilseed Research Centre, Agriculture and Agri-Food Canada, Central Experimental Farm, KW Neatby Building, K1A 0C6, Ottawa, Ontario, Canada Received 12 February 2005; accepted in revised form 8 April 2005
Abstract Penicillium roqueforti, P. crustosum and P. paneum grow on ensiled grain and recycled feed unless properly treated. The former two species occur also on cut lumber in Canada. These are known to produce a number of secondary metabolites including roquefortine. In cooler dairy production areas, including Scandinavia and North America, cattle toxicosis has been associated with silage contaminated by these fungi. We collected strains associated with cow or cattle toxicoses. The principal metabolites were determined making use of a new extraction method and analysis combining HPLC, LC/MS/MS, and LC/NMR. Penicillium roqueforti and P. crustosum required amino acid nitrogen for metabolite formation and their toxins were formed under conditions of low oxygen (20–30% saturation). Production of roquefortine C occurred on depletion of the available nitrogen and penitrem A on depletion of carbon source. Yield was reduced by excess carbon. Medium osmotic tension (aw) affected metabolite production by the two species differently. Penicillium paneum was associated with ill-thrift of dairy cows and P. roqueforti was associated with more serious symptoms. Our data suggest a physiological basis for the common occurrence of roquefortine C in silage without serious consequences and the alternative, the presence of roquefortine C and toxicoses. The strain isolated from lumber was the best producer of the toxins studied. This is the first report of the toxigenic potential of P. roqueforti and P. paneum from Canada. Key words: cattle toxicosis, lumber, Penicillium crustosum, Penicillium paneum, Penicillium roqueforti, roquefortine, silage
Introduction In areas where silage is an important component of cattle feed, a number of toxic phenomena have been occasionally observed associated with P. crustosum and the P. roqueforti group. Severe toxicoses in cows associated with the latter fungi growing on silage were first reported from Japan and the United States in the 1960s [1, 2]. This issue remained unresolved but largely ceased to be a practical problem. More recently, these taxa are associated with reports of two syndromes in cows: the serious toxicoses reported initially as
well as a general ill-thrift [3–6]. Recycled feed can be contaminated by these fungi and their toxins [7]. These Penicillium species produce a variety of secondary metabolites, including roquefortine, PR toxin, penitrem A and mycophenolic acid [8, 9]. None of these compounds appear sufficiently toxic on their own to explain the symptoms seen, although the toxicology is incomplete [10]. These species also occur in lumber and in indoor environments where very wet wood is present. The purpose of this study was to characterize the mycotoxin profiles of Penicillia isolated from
572 silage or recycled feed in eastern Canada associated with cow or cattle toxicoses as well as additional strains from feed and lumber. To achieve this, analytical methods were developed to permit HPLC-UV, LC/MS/MS, and LC/NMR analyses using the same chromatographic conditions. In addition, studies of the factors that affect the yield of roquefortine and penitrem A were conducted.
Materials and methods Fungi Samples of corn and grass silage were collected from farms where cattle were experiencing unexplained toxicoses in Quebec, Canada. The samples were plated on 2% malt extract agar (MEA) plates (20 g/l Difco malt extract and 18 g/l agar) and incubated in darkness at 25 C for 1 week. The plates were densely colonized; representative isolates were transferred and preliminary identifications made. Representative single spore isolates of each strain were transferred to MEA slants for storage and deposited in the culture collection at Agriculture and Agri-Food Canada (DAOM). All of the isolates were identified as either P. roqueforti or P. paneum [Nielson et al. submitted]. Six isolates of P. paneum (DAOM 232120, 232121, 232122, 232123, 232124, 232125), and three isolates of P. roqueforti (DAOM 232127, 232128, 232126) were obtained from silage associated with ill-thrift and reduced milk production (farm 1). Four isolates of P. roqueforti (DAOM 234135, 234136, 234137, 234138) were recovered from silage associated with cows with complete reproductive failure (farm 2). One isolate of P. roqueforti (DAOM 234139) was obtained from silage associated with cow mortalities and reduced milk production (farm 3). One isolate of P. roqueforti (DAOM 229761) was isolated from recycled feed (bread) associated with severe cattle toxicosis in New Brunswick, Canada (farm 4). The other 3 strains tested were obtained from the DAOM collection: P. roqueforti DAOM 194930 [deposited as P. crustosum] was isolated from barley in Manitoba, Canada associated with undefined livestock toxicosis (farm 5); P. roqueforti DAOM 225571 was isolated from corn silage in Quebec associated with
undefined toxicosis in dairy cows (farm 6); and P. crustosum DAOM 215343 was isolated from lumber in Ontario. Fermentations Previous studies have demonstrated Czapek-Dox broth supplemented with yeast extract (CDYE) for the production of both penitrem and roquefortine [11, 12]. This medium was used as the reference medium in a series of shake and surface culture fermentations testing the relative productivity of media intended to stimulate the expression of acetate- and amino-acid derived metabolites. The accumulation of roquefortine by P. crustosum DAOM 215343 and for some trials, P. roqueforti DAOM 229761 was studied to determine the impact of C/N/P ratio. These experiments supported CDYE as a suitable medium for roquefortine production. Several ingredient modifications were tested, including (a) eliminating yeast extract or inorganic nitrogen, respectively, (b) reducing the phosphate content to 50% and 10%, respectively, (c) increasing by 2 and 5 times the concentration of sucrose ([11] used the higher value) and (d) changing the osmotic tension using the nonmetabolizable sugar alcohol glycerol at a rate of 10, 20, 30, 40, and 50 mg/l. In each case, experiments were done at least in triplicate compared to the reference culture. On this basis, the following protocol was developed for strain screening, fermentation time-course studies and biosynthetic experiments: Slants were macerated in 50 ml sterile distilled water and inoculated at 5% (v/v) into 250 ml flasks containing 50 ml Czapek-Dox broth supplemented with yeast extract (30 g/l sucrose, K2HPO4 1 g/l, yeast extract 5 g/l, MgSO4 0.5 g/l, NaNO3 3 g/l, FeSO4 0.01 g/l, and KCl 0.5 g/l). After 48 h incubation in darkness at 25 C on a rotary shaker (220 rpm, 3.8 cm throw), the cultures were combined and macerated, inoculated at 5% (v/v) into 200 ml of CDYE in one l Roux bottles or 500 ml Erlenmeyer flasks, and incubated in darkness at 25 C for 12 days (see Table 1). Analysis All manipulations were conducted under low light conditions and materials were stored in glassware
573 Table 1. Species DAOM P. P. P. P. P. P. P. P. P. P. P. P. P.
paneum 232120* paneum 232124* paneum 232125* roqueforti 232126* roqueforti 232127* roqueforti 234135 roqueforti 234136 roqueforti 234137 roqueforti 234138 roqueforti 234139 roqueforti 229761 roqueforti 194930 roqueforti 225571
Farm number
Substrate
Roquefortine C
1 1 1 1 1 2 2 2 2 3 4 5 6
Grass silage Grass silage Grass silage Corn silage Grass silage Grass silage Grass silage Grass silage Grass silage Grass silage Recycled feed (bread) Barley feed Corn silage
++++ ++++ ++++ ++++ ++++ +++ ++++ +++ +++ ++++ ++++ +++ +++
Mycophenolic acid
Marcfortine A & B ++ ++ ++
+ ++ + + + ++ + ++ + +
++++ = >10 mg/g, +++ = 1–10 mg/g detected, ++ = lg/g, + = trace, *grown in Erlenmeyer flasks.
protected from light. Fermentations were filtered through Whatman #1 paper and the mycelium was freeze-dried and weighed. The cells were extracted directly with 2 · 200 ml of chloroform (CHCl3). The culture filtrate was extracted with two equal volumes of chloroform and the filtrate pH adjusted to 12 with 1 N NaOH. More than 95% of the extractable roquefortine was obtained in the first two chloroform extractions from the filtrate. Preliminary experiments adjusting the culture filtrate to pH 4 with 1 N HCl as well as pH 8 and 12 demonstrated maximum recovery of roquefortine at pH 12. Extracts were dried using a rotary evaporator and taken up in 5 ml of acetonitrile. Toxin analysis in the crude extracts was performed using an Agilent 1100 series quartenary pump with a diode array detector connected to a Phenomenex Synergi Max 4l reverse phase C12 column (250 mm · 4.6 mm). The mobile phase involved an acetonitrile/water gradient with 0.1% ammonium acetate (adapted from [13]). These conditions produced clearly resolved peaks and run times of 30 min. Typically, 10 ll was injected on the column. Authentic standards of roquefortine and mycophenolic acid were obtained from Sigma (St. Louis, USA). Penitrems A and B were isolated from P. crustosum DAOM 215343 and were characterized by MS and NMR compared to the literature [14]. In the case of marcfortine A and B, determinations were made based on high resolution MS and NMR data. MS spectra were obtained using a Hewlett Packard 1050 HPLC attached to a triple quadropole Micromass electrospray ionization (EI) mass spectrometer (MS)
set for analysis in positive ion mode. Companion proton NMR spectra were obtained using a Varian 9010 HPLC attached to a Varian 600 MHz INOVA NMR spectrometer equipped with a 5 mm broad-band cryo flow probe with 60 ll active volume. For the time course experiment, 12 Roux bottles were inoculated with P. crustosum DAOM 215343 and incubated as above. This strain was chosen because it produced higher yields of both roquefortine and penitrem A. At days 3, 6, 9 and 12, three Roux bottles were removed from the incubator and placed flat in a laminar flow hood without disturbing the surface growth. Small aliquots of broth (5 ml) were withdrawn into 25 ml vials with a 10 ml pipette taking care to minimize disturbance to the liquid. Dissolved oxygen was immediately determined using a YSI model 57 dissolved oxygen meter (Yellow Springs, OH). The pH of the cultures was also recorded. The cultures were filtered and the filtrate and mycelium were retained. The culture filtrate from each bottle was analyzed for residual sugars, total nitrogen, mycelium dry weight, roquefortine and penitrem A. Sugars were determined with a Varian Vista 5500 HPLC with a Waters differential refractometer R401 detector connected to a Waters WAT0443550 high performance carbohydrate column [15]. Samples were analyzed in duplicate using an isocratic mobile phase of 75:25 acetonitrile:water with a flow rate of 1.4 ml/min. The total remaining nitrogen was determined after persulfate digestion by a method based on the second derivative
574 absorption spectrum of NO3) (peak at 224 nm) measured with a Varian DMS 200 spectrophotometer [16, 17]. The mycelium and filtrate were analyzed for roquefortine and penitrem A as above.
Results and discussion Over the period 1998–2002, we collected samples of ensiled materials and a single sample of recycled feed reliably associated with cow or cattle toxicoses. In each case, the dominant fungi isolated were in P. roqueforti section roqueforti (Table 1). The strains of P. paneum produced roquefortine C, marcfortine A and marcfortine B. In contrast, the strains of P roqueforti produced roquefortine C and mycophenolic acid. Further taxonomic information, data on strains not tested in these experiments and additional data on minor metabolite production are given elsewhere [Nielson et al. submitted]. The majority of isolates from grass silage were P. paneum, a species only recently circumscribed [18]. Most isolates of P. roqueforti occurred on corn silage. Boysen et al. [3] studied a large collection obtained from incidents of silage toxicoses in Sweden from 1988–1998. They also reported that P. paneum was associated with ill-thrift whereas P. roqueforti was associated with more serious disease. Ill-thrift was most common in the reports associated with the feed samples we examined from dairy herds in Quebec the majority of which yielded P. paneum. The toxicosis associated with P. roqueforti DAOM 229761 from recycled feed from NB led to the rejection by Canadian federal meat inspectors of the affected herd for slaughter. The growth of P. roqueforti in spent grain used as feed has been implicated previously with severe cow toxicoses in Ontario (Miller and Hussar, unpublished data). Based on the presence of PR toxin [18], P. roqueforti was inferred to be common in silage from some dairy farms in Wisconsin, Vermont and Florida. Seglar and colleagues studied 21 dairy herds with ill-thrift (veterinarian-diagnosed; based on milk production, reproduction and mastitis) and 18 healthy herds. Silage samples were analyzed for deoxynivalenol (DON), fumonisin, zearalenone, PR toxin and a number of other mycotoxins by commercial immune assays [4, 19].
The samples contained little or no zearalenone nor fumonisin. Eight sick herds were excluded based on a diagnosis of Johne’s disease, salmonellosis or internal parasites. When the remaining feed samples were stratified, there was no difference in the DON concentrations which were all greater than 1 lg/g. The concentration of PR toxin tended to be higher in herds manifesting ill-thrift (ca. 100 ng/ g). However, the concentrations of PR toxin observed were lower overall than have been reported in European studies [6, 10]. The metabolites of the Canadian strains studied from silage and other feeds (Table 1) are consistent with published chemotypes [3, 18, 20]. As noted, mycelial extracts contained the majority of roquefortine (70%) along with lesser amounts of the other compounds listed. When present, mycophenolic acid tended to be most concentrated in culture filtrate, accompanied by roquefortine. The identities of these metabolites were confirmed by stopped flow LC/NMR and molecular weights and fragmentation patterns from LC/MS/MS. While examining extraction conditions for the study, cultures exposed to light had significantly lower mycotoxin recovery (
