Food Additives & Contaminants
ISSN: 0265-203X (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/tfac19
Studies on a toxic metabolite from the mould Wallemia G. M. Wood , P. J. Mann , D. F. Lewis , W. J. Reid & M. O. Moss To cite this article: G. M. Wood , P. J. Mann , D. F. Lewis , W. J. Reid & M. O. Moss (1990) Studies on a toxic metabolite from the mould Wallemia , Food Additives & Contaminants, 7:1, 69-77, DOI: 10.1080/02652039009373822 To link to this article: http://dx.doi.org/10.1080/02652039009373822
Published online: 10 Jan 2009.
Submit your article to this journal
Article views: 14
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tfac20 Download by: [Universite Laval]
Date: 06 November 2015, At: 20:20
FOOD ADDITIVES AND CONTAMINANTS, 1990, VOL. 7, NO. 1, 6 9 - 7 7
Studies on a toxic metabolite from the mould Wallemia G. M. WOOD†, P. J. MANN†, D. F. LEWIS†, W. J. REID† and M. O. MOSS‡ † Leatherhead Food Research Association, Randalls Road, Leatherhead, Surrey KT22 7RY, UK ‡ Department of Microbiology, University of Surrey, Guildford, Surrey GU2 5XH, UK
Downloaded by [Universite Laval] at 20:20 06 November 2015
(Received 16 June 1988; revised 15 January 1989; accepted 29 July 1989) While monitoring the occurrence of toxigenic moulds in foods, using a bioassay screen, it was shown that an isolate of Wallemia sebi produced toxic effects in several of the bioassays. The toxic metabolite was isolated and purified using solvent extraction, TLC and HPLC coupled with the brine shrimp assay to monitor the toxic fractions. The purified toxin, which we propose to call walleminol A, has been partially characterized by mass spectroscopy, nuclear magnetic resonance, ultraviolet and infrared spectroscopy. It can be provisionally interpreted as a tricyclic dihydroxy compound, C 15 H 24 O 2 , with structural features characteristic of a sesquiterpene with an isolated double bond, but further work is required to characterize this compound unequivocally. The minimum inhibitory dose of walleminol A in the bioassays is approximately 50 µg/ml, which is comparable with a number of mycotoxins such as citrinin and penicillic acid. Keywords: mycotoxin; Wallemia; walleminol A; bioassay.
Introduction
The mould Wallemia sebi is a Deuteromycete capable of growth over an exceptionally wide range of aw from 0-69 to 0-997 (Pitt and Hocking 1977) and has been isolated from foods such as jam, cake, cereals and flour, as well as meat and dairy products. Although W. sebi is very widespread it is frequently overlooked because it forms small, dull brown colonies, which may be overgrown by other moulds on conventional isolation media. Pitt (1975) discussed the problems of studying the distribution and ecology of xerophilic fungi with the media then available, but the dichloran-glycerol medium subsequently described by Hocking and Pitt (1980) has proved to be suitable for the enumeration and isolation of xerophiles, including Wallemia, from low-moisture foods. Wallemia has been frequently found to cause brown discoloration of fish and to spoil condensed milk and rice. Pitt (1975) reported that it could almost always be isolated by moistening and incubating Australian cereals, or bread, and that partially rehydrated prunes, packed in plastic pouches at av up to 0-86, provided a favourable substrate for xerophilic moulds, including W. sebi. Saito et al. (1971,1974) found Wallemia to be a dominant contaminant of milled rice and flours, as well as of dried fish products. It has also been isolated from animal feeds (Takatori and Kondoh 1979), flood-damaged grains (Mills and Abramson 1981), several types of dried beans (Hitokoto et al. 1981), cakes and jam (Wood 1984), wheat stored in aerated bins (Hurlock et al. 1980), and brown rice during long-term storage under natural conditions (Tsuruta and Saito 1980). 0265-203X/90 $3.00 ©1990 Taylor & Francis Ltd.
70
G. M. Wood et al.
There are few data on the toxigenic potential of Wallemia, although Saito et al. (1971) showed that filtrates from cultures of Wallemia were toxic to HeLa cells and mice, and Wood (1984) reported on the toxicity of an isolate of Wallemia in several bioassays. Because of the widespread occurrence of Wallemia, this study was initiated to isolate and characterize any toxic metabolites produced by the mould.
Downloaded by [Universite Laval] at 20:20 06 November 2015
Experimental Organisms Wallemia sebi was isolated from a sample of cake (isolate A) and jam (isolate B) and maintained at 4°C on slopes of osmophilic agar (Scarr 1959). Isolate A has been deposited at CAB International Mycological Institute, Kew (IMI 333109). Brine shrimp eggs (Artemia salina) were obtained from Betta Aquaria (108 Shields Road, Newcastle) and stored at room temperature in a desiccator; baby hamster kidney cells (BHK) from Flow Laboratories; rat liver cells from MRC, Carshalton; and Tetrahymena pyriformis from the Freshwater Biological Association (Ferry House, Far Sawrey, Ambleside, Cumbria). BHK and rat liver cells were maintained in the Glasgow modification of Eagle's medium and the protozoan was maintained in 2% proteose peptone with 0.25% yeast extract. Bioassays The brine shrimp assay was based on that of Harwig and Scott (1971); the protozoal assay was that of Hayes and Wyatt (1970), and the animal cell culture assays were those described by Wood (1984). Animal cells were examined by both light microscopy and scanning electron microscopy using the methods of Gamliel et al. (1983). Production of toxic metabolites Toxin production was studied in media containing 2% yeast exact with either 4% or 20% sucrose. Flasks (4 X 250 ml) containing 100 ml of sterile medium were inoculated with a spore suspension of Wallemia and incubated at 25°C, without shaking, for 2-10 weeks. Mycelium was removed by filtration, washed with several portions of water and the culture filtrate was defatted by extraction with «-hexane. The toxic material was obtained by repeated extraction with chloroform (from the filtrate), combining the chloroform extracts and concentrating by evaporation under reduced pressure in a rotary evaporator. The procedure was scaled up to the inoculation of 2-litre flasks containing 500 ml of medium (2% yeast extract + 20% sucrose) for purification of toxic metabolites. Purification of toxic metabolites The concentrated chloroform extract was loaded as a streak on to TLC plates of silica gel (Merck 5553) and chromatographed with toluene:ethyl acetate:formic acid (5:4:1). All solvents used for extraction and chromatography were distilled prior to use to remove non-volatile residues. The chromatogram was divided into three fractions (Fn 1, Rf 0-4; Fn 2, 0-4 R{ 0-7; Fn 3, Rt 0-7) and each fraction was cut into small portions and shaken for 15 min with 75 ml methanol. The strips were removed, the extraction repeated and the methanol extracts pooled and evaporated to dryness on a rotary evaporator.
Mycotoxins from Wallemia
71
The three fractions were assayed for toxicity using brine shrimp larvae and were further examined by TLC after spraying with /7-anisaldehyde reagent (0-5% in methanol:acetic acid:conc. sulphuric acid (80:15:5)). Further purification was carried out by preparative HPLC using an amino column (Spherisorb 5 fim, 250 x 4-9 mm from Hichrom, Reading UK). The mobile phase was butan-1-ol:«hexane (10:90) at a flow rate of 2 ml/min and the detector was an LDC Spectromonitor D UV detector set at 230 nm. Samples were injected via a 100 /A loop (Rheodyne 7125).
Downloaded by [Universite Laval] at 20:20 06 November 2015
Structural characterization The purified toxin was analysed by GC-MS using the following conditions: Kratos MS 80 RFA GC conditions: 25 m OV—1701, 0-25 urn, 0-3 mm i.d., 40°C for 2 min, 10°C/min to 270°C; 2 ul injected. MS conditions: (i) Electron impact—electron energy—70 eV, beam current—100 /iA, source temperature—200°C, Scan rate—650-25 amu at 1 s/decade, resolution—1000. (ii) High resolution—above conditions except resolution—3000. (iii) Chemical ionization—reagent gas—NH3, beam current—500 pA, source temperature 200°C, electron energy—n 100 eV (tuned to maximum sensitivity), scan rate—650-100 amu at 1 s/decade, resolution—1000 or 3000. (iv) Trimethylsilyl (TMS) derivative—the toxin was converted into its TMS derivative by adding 250 /«I of reagent (N-methyl-A'-trimethylsilyl trifluoroacetamide) to 1 mg of toxin and warming for 15 min at 70°C. The sample was diluted to 5 ml in diethyl ether for GC-MS analysis. NMR spectra of the toxin (5 mg) were recorded in CDCI3 using trimethylsilane as the internal standard using a Bruker WH-90 90 MHz pulse Fourier transform NMR spectrometer; IR spectra were recorded in KBr discs using a Perkin Elmer model 1700 and UV spectra in ethanol using a Pye Unicam SP 800. Results and discussion
Growth and toxin production by isolate A were greater in 20% sucrose yeast extract medium than in the medium containing only 4% sucrose (table 1). Toxicity to brine shrimps reached a maximum after 6 weeks' incubation in both media and was significantly reduced after 10 weeks. Isolate B grew more slowly than isolate A; however, chloroform extracts of culture filtrates from isolate B were more toxic to brine shrimps and rat liver cells. Preliminary evidence suggests that the two isolates produce related, but distinct, toxins, which we propose to call walleminol A and B. Walleminol B showed a similar fragmentation pattern to walleminol A when analysed by GC-MS. It was shown, however, to have a higher molecular weight of 252. It also showed a colour difference when treated with /7-anisaldehyde and a small difference in R( when compared with the TLC of walleminol A. Purification of walleminol A Fractionation of metabolites in the chloroform extract of isolate A was followed using both brine shrimp and animal cell cultures to monitor toxicity, as well as TLC on silica gel plates developed with toluene:ethyl acetate:formic acid (5:4:1)
G. M. Wood et al.
72
Table 1. Effect of incubation time on the toxicity of Wallemia extracts. Growth medium
Downloaded by [Universite Laval] at 20:20 06 November 2015
Yeast extract + 4 % sucrose
Yeast extract +20% sucrose
Percentage kill of brine shrimps (dilution)
Incubation time (weeks)
Dry weight of mycelium (g)
1:50
1:200
2 3 4 6 10
0-22 0-35 0-69 1-58 1-27
92 100 100 100 100
0 0 4 100 32
Percentage kill of brine shrimps (dilution)
1:400
Dry weight of mycelium (g)
1:50
1:200
1:400
0 0 0 90 0
1-62 3-47 3-06 3-84 4-29
90 100 100 100 100
0 100 100 100 100
0 10 100 100 80
No effect was caused by extracts of media blanks.
(TEF), and visualized by spraying with /7-anisaldehyde followed by heating for 5 min at 100°C. Toxicity was associated with fraction 2 from the preliminary fractionation procedure (see methods), and further fractionation and purification was achieved using the protocol shown in figure 1, the position of bands being visualized by spraying strips from each edge of the chromatograms with the /»-anisaldehyde reagent. The fraction F2 A2 (figure 1) was toxic to both brine shrimps and rat liver cells, although it was noticeable that purification seemed to result in a decrease in toxicity to brine shrimps. During the purification studies 'equivalent' amounts of the crude
F2C1 Non-toxic
Chloroform extract from aqueous media (6 weeks incubation in YE+ 20% sucrose)
Tolueneethyl acetate: formic acid 5-4:1
Chloroform-.methanol 97:3
Chloroform:methanol 99:1
Chloroform:mettianol 9:1
Figure 1. Separation of compounds from Wallemia by TLC purification.
Mycotoxins from Wallemia
73
1 100 908070-
I 6500 "
I " |
40-
cx. 3020-
Downloaded by [Universite Laval] at 20:20 06 November 2015
100 20:36
2203
23:31 24:58 Retention time (min:s) Figure 2. Analysis of extract F2 A2 by GC.
27:53
26:26
extracts were tested for toxicity to brine shrimps and rat liver cells. The decrease in toxicity could have been due to the removal of compounds such as lipids (which are known to affect brine shrimps) or to losses of toxin during purification. Further analysis of both crude extracts and purified toxin by HPLC was performed in the final stages of this work. This indicated that the observed decreased effect on toxicity was not solely due to toxin losses but the removal of other compounds, possibly acting synergistically with the toxin, could be contributing to the toxicity. Walleminol A was separated from interfering compounds by preparative HPLC; the major impurity compound (as seen in figure 2) was identifed as 3-indole ethanol (tryptophol) by GC-MS and by comparison with authentic standards. To our knowledge this is the first report of tryptophol from Wallemia. 69
100 90
70 g 60•- 5 0 -
79
41
I |
91 yl
105
98 40-
3053
20100
218
I llllkl.l,lllll,.I.I.L.l,,,I...J,, i * 50
100
150 m/z
200
Figure 3. Electron impact spectrum of F2 A2.
250
Downloaded by [Universite Laval] at 20:20 06 November 2015
74
G. M. Wood et al.
Characterization of walleminol A Analysis of F2 A2 by GC-MS showed only two peaks, the window containing these peaks is shown in figure 2, the major peak had a probable molecular ion at m\z 236 (figure 3). This was confirmed by the presence of ions at mjz 237, 219 and 201 in the spectrum using positive chemical ionization, corresponding to the sequential loss of two molecules of water from a molecule of mass 236 (figure 4) and by a major ion at mjz 235 in the negative chemical ionization spectrum. The trimethylsilylated derivative also showed two peaks in the same ratio as the underivatized GC trace. The major peak had a molecular ion at m\z 380 corresponding to a bis (trimethylsilyl) derivative of a molecule of mass 236. Further GC-MS analysis of the material purified by HPLC confirmed the observations obtained on F2 A2. High-resolution mass spectrometry in NICI mode at resolution 3000 gave M-l peak of 235-170 ± 0-002, suggesting that walleminol A has the molecular formula C15H24O2. The experimental error was determined by mass measurement of a standard ion of known composition. The NMR and IR spectra are shown in figures 5 and 6, and are consistent with the presence of an olefin bond (a single proton CDCI3 at 4-7 ppm, end absorption in the UV and a peak at 835 cm"1 in the IR) and four methyl groups, two of which may represent a gem-dimethyl system (1120 cm" 1 in the IR and four 3-proton signals: CDCI3 at 0-85, 1-06, 1-21 and 1-29 ppm), and two hydroxyl groups (3300 cm"1 in the IR and a broad 2-proton signal at CDCI3 4-9-5-3 ppm in the NMR). The four double bond equivalents implied by a molecular formula C15H24O2 could thus be accounted for as three rings and a trisubstituted double bond, although structures containing fully substituted unconjugated double bonds in place of one or more rings cannot be ruled out. All the evidence presented is consistent with walleminol A being a tricyclic dihydroxysesquiterpene.
219
237 .1... 300
Figure 4. Positive chemical ionization spectrum of F2 A2.
400
Downloaded by [Universite Laval] at 20:20 06 November 2015
Mycotoxins from Wallemia
75
-iM*^
6
5
4
3
2
6 ppm
Figure 5. NMR spectrum of pure toxin produced by Wallemia.
Toxicity of walleminol A Brine shrimp. Brine shrimps are normally extremely active swimmers, and it has been suggested that if shrimps can be seen to be 'moving but not making forward progress' they should be considered as 'dead' for the assay (Eppley 1974). Walleminol A at 80 ;tg/ml caused 41% kill of shrimps, with the remaining showing movement but not swimming. This would be equivalent to 100% 'dead'. Further dilutions of the toxin showed that a concentration of 40 /ig/ml caused 30% kill with a further 20% of weak shrimps. The LD50 (including weak shrimps) was therefore approximately 40 /*g/ml. Protozoa. Death of the protozoa (71 pyriformis) was caused by concentrations of 50 /tg/ml of the toxin and above. Further dilutions of the toxin caused restriction of • 40
"4000
3500
3000
2500
2000 1800 1600 Frequency/cm
1400
1200
1000
Figure 6. IR spectrum of pure toxin produced by Wallemia.
800
500
76
G. M. Wood et al.
growth and a lag in the growth phase. A concentration of 25 /tg/ml of the toxin allowed 50% growth compared with the controls and 10 /xg/ml had no effect on the protozoa.
Downloaded by [Universite Laval] at 20:20 06 November 2015
Cell lines. High concentrations of the toxin (500 jig/ml) caused rapid death of both rat liver cells and BHK cells. Lower concentrations (50-125 /tg/ml) caused a restricted growth rate and affected the morphology of the cells. The toxin caused changes in the nuclei of both cell lines and, in the case of the BHK cells, the formation of giant cells was apparent. The rat liver cells were subject to cytoplasmic vacuolation. All changes were visible after examination by light microscopy and electron microscopy; the latter indicated that the nuclei of toxin-treated cells contained unusual membrane-like structures. The minimum inhibitory dose of the toxin was approximately 50 /tg/ml. Discussion This study outlines the isolation of a compound produced by the mould Wallemia sebi, to be named walleminol A. The compound causes toxic effects in a range of in vitro systems such as mammalian cell lines, protozoa and brine shrimp. The dose affecting these systems is comparable with such mycotoxins as penicillic acid and citrinin. The timescale of this study did not allow a complete characterization of the compound, but data collected have allowed the following information. The toxin has a molecular weight 236, composition C15H24O2. It contains two hydroxyl groups, four methyl groups, including a gem dimethyl, and there are two or three ring structures in the molecule. There is no literature available on this compound, suggesting that it has not been previously characterized. The work reported includes only a study of the mould growth in culture media, which does not assess the significance of this mould in the food environment. However, information suggests that this mould is a greater contaminant of foods than previously realized (Saito et al. 1971, 1974, Pitt and Hocking 1985), and we have found in our surveillance studies that Wallemia is a common contaminant of cake and jam samples. Further work on the analysis of foods for walleminol A will be reported at a later date. Acknowledgements This work was supported by funding from the Ministry of Agriculture, Fisheries and Food. Many thanks are due to Kathy Groves for electron microscopy studies, and Mr J. P. Bloxsidge and Dr G. J. Buist at the University of Surrey for obtaining the NMR and IR data. References EPPLEY, R. M., 1974, Mycotoxins: sensitivity of brine shrimp (Artemia salina) to trichothecenes. Journal of the Association of Official Analytical Chemists, 57, 618-620. GAMLIEL, H., GURFEL, D., LEIZEROWITZ, R. and POLLIACK, A., 1983, Air drying of human leucocytes for
scanning electron microscopy using the GTGO procedure. Journal of Microscopy, 131, 87-95. HARWIG, J. and SCOTT, P, M., 1971, Brine shrimp (Artemia salina) larvae as a screening system for fungal toxins. Applied Microbiology, 21, 1011-1016. HAYES, A. W. and WYATT, E. P., 1970, Survey of the sensitivity of micro-organisms to rubratoxin B. Applied Microbiology, 20, 164-165.
Mycotoxins from Wallemia
77
HITOKOTO, H., MOROZUMI, S., WAUKE, T., SAKAI, S. and KURATA, H., 1981, Fungal contamination and
mycotoxin producing potential of dried beans. Mycopathologia, 73, 33-38. HOCKING, A. D. and PITT, J. I., 1980, Dichloran-glycerol medium for enumeration of xerophilic fungi from low moisture foods. Applied and Environmental Microbiology, 39, 488-492. HURLOCK, E. T., ARMITAGE, D. M. and LLEWELLIN, B. E., 1980, Seasonal changes in mite ocari and
fungal populations in aerated and unaerated wheat stored for 3 years. Bulletin of Entomological Research, 70, 537-548. MILLS, J. T. and ABRAMSON, D. 1981, Microflora and condition of flood damaged grain in Manitoba, Canada, Mycopathologia, 73, 143-152. PITT, J. I., 1975, Xerophilic fungi and the spoilage of foods of plant origin. Water Relations of Foods, edited by R. B. Duckworth (London: Academic Press), pp. 273-307. PITT, J. 1. and HOCKING, A. D. 1977, Influence of solute and hydrogen ion concentration on the water relations of some xerophilic fungi. Journal of General Microbiology, 101, 35-40.
Downloaded by [Universite Laval] at 20:20 06 November 2015
SAITO, M., ISHIKO. T., ENOMOTO. M., OHTSUBO, K., UMEDA, M., KURATA, H., UDAGAWA, S.,
TANIGUCHI, S. and SEKITA, S. 1974, Screening test using HeLa cells and mice for detection of mycotoxin-producing fungi isolated from foodstuffs. An additional report on fungi collected in 1968 and 1969. Japanese Journal of Experimental Medicine, 44, 63-82. SAITO, M., OHTSUBO, K., UMEDA, J., ENOMOTO, M., KURATA, H., UDAGAWA, S., SAKABE, F. and
ICHINOE, M., 1971, Screening tests using HeLa cells and mice for the detection of mycotoxinproducing fungi isolated from foodstuffs. Japanese Journal of Experimental Medicine, 41, 1-20. SCARR, M. P., 1959, Selective media used in the microbiological examination of sugar products. Journal of the Science of Food and Agriculture, 10, 678-681. TAKATORI, K. and KONDOH, S., 1979, Fungal distribution of swine feeds and the mycotoxin producibility of Aspergillus flavus and Aspergillus versicolor. Japanese Journal of Zootechnology and Science, 50, 453-459. TSURUTA, O. and SAITO, M. 1980, Mycological damage of domestic brown rice during storage in warehouses under natural conditions. 3. Changes in mycofloras during the storage. Nikon Kin Gakkai Kaiho, 21, 121-125. WOOD, G. M., 1984, Assessment of toxigenic moulds in foods by means of a biological screening method. Mycotoxins in Animal and Human Health, edited by M. O. Moss and M. Frank (Guildford: University of Surrey), pp. 95-105.
ISSN: 0265-203X (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/tfac19
Studies on a toxic metabolite from the mould Wallemia G. M. Wood , P. J. Mann , D. F. Lewis , W. J. Reid & M. O. Moss To cite this article: G. M. Wood , P. J. Mann , D. F. Lewis , W. J. Reid & M. O. Moss (1990) Studies on a toxic metabolite from the mould Wallemia , Food Additives & Contaminants, 7:1, 69-77, DOI: 10.1080/02652039009373822 To link to this article: http://dx.doi.org/10.1080/02652039009373822
Published online: 10 Jan 2009.
Submit your article to this journal
Article views: 14
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tfac20 Download by: [Universite Laval]
Date: 06 November 2015, At: 20:20
FOOD ADDITIVES AND CONTAMINANTS, 1990, VOL. 7, NO. 1, 6 9 - 7 7
Studies on a toxic metabolite from the mould Wallemia G. M. WOOD†, P. J. MANN†, D. F. LEWIS†, W. J. REID† and M. O. MOSS‡ † Leatherhead Food Research Association, Randalls Road, Leatherhead, Surrey KT22 7RY, UK ‡ Department of Microbiology, University of Surrey, Guildford, Surrey GU2 5XH, UK
Downloaded by [Universite Laval] at 20:20 06 November 2015
(Received 16 June 1988; revised 15 January 1989; accepted 29 July 1989) While monitoring the occurrence of toxigenic moulds in foods, using a bioassay screen, it was shown that an isolate of Wallemia sebi produced toxic effects in several of the bioassays. The toxic metabolite was isolated and purified using solvent extraction, TLC and HPLC coupled with the brine shrimp assay to monitor the toxic fractions. The purified toxin, which we propose to call walleminol A, has been partially characterized by mass spectroscopy, nuclear magnetic resonance, ultraviolet and infrared spectroscopy. It can be provisionally interpreted as a tricyclic dihydroxy compound, C 15 H 24 O 2 , with structural features characteristic of a sesquiterpene with an isolated double bond, but further work is required to characterize this compound unequivocally. The minimum inhibitory dose of walleminol A in the bioassays is approximately 50 µg/ml, which is comparable with a number of mycotoxins such as citrinin and penicillic acid. Keywords: mycotoxin; Wallemia; walleminol A; bioassay.
Introduction
The mould Wallemia sebi is a Deuteromycete capable of growth over an exceptionally wide range of aw from 0-69 to 0-997 (Pitt and Hocking 1977) and has been isolated from foods such as jam, cake, cereals and flour, as well as meat and dairy products. Although W. sebi is very widespread it is frequently overlooked because it forms small, dull brown colonies, which may be overgrown by other moulds on conventional isolation media. Pitt (1975) discussed the problems of studying the distribution and ecology of xerophilic fungi with the media then available, but the dichloran-glycerol medium subsequently described by Hocking and Pitt (1980) has proved to be suitable for the enumeration and isolation of xerophiles, including Wallemia, from low-moisture foods. Wallemia has been frequently found to cause brown discoloration of fish and to spoil condensed milk and rice. Pitt (1975) reported that it could almost always be isolated by moistening and incubating Australian cereals, or bread, and that partially rehydrated prunes, packed in plastic pouches at av up to 0-86, provided a favourable substrate for xerophilic moulds, including W. sebi. Saito et al. (1971,1974) found Wallemia to be a dominant contaminant of milled rice and flours, as well as of dried fish products. It has also been isolated from animal feeds (Takatori and Kondoh 1979), flood-damaged grains (Mills and Abramson 1981), several types of dried beans (Hitokoto et al. 1981), cakes and jam (Wood 1984), wheat stored in aerated bins (Hurlock et al. 1980), and brown rice during long-term storage under natural conditions (Tsuruta and Saito 1980). 0265-203X/90 $3.00 ©1990 Taylor & Francis Ltd.
70
G. M. Wood et al.
There are few data on the toxigenic potential of Wallemia, although Saito et al. (1971) showed that filtrates from cultures of Wallemia were toxic to HeLa cells and mice, and Wood (1984) reported on the toxicity of an isolate of Wallemia in several bioassays. Because of the widespread occurrence of Wallemia, this study was initiated to isolate and characterize any toxic metabolites produced by the mould.
Downloaded by [Universite Laval] at 20:20 06 November 2015
Experimental Organisms Wallemia sebi was isolated from a sample of cake (isolate A) and jam (isolate B) and maintained at 4°C on slopes of osmophilic agar (Scarr 1959). Isolate A has been deposited at CAB International Mycological Institute, Kew (IMI 333109). Brine shrimp eggs (Artemia salina) were obtained from Betta Aquaria (108 Shields Road, Newcastle) and stored at room temperature in a desiccator; baby hamster kidney cells (BHK) from Flow Laboratories; rat liver cells from MRC, Carshalton; and Tetrahymena pyriformis from the Freshwater Biological Association (Ferry House, Far Sawrey, Ambleside, Cumbria). BHK and rat liver cells were maintained in the Glasgow modification of Eagle's medium and the protozoan was maintained in 2% proteose peptone with 0.25% yeast extract. Bioassays The brine shrimp assay was based on that of Harwig and Scott (1971); the protozoal assay was that of Hayes and Wyatt (1970), and the animal cell culture assays were those described by Wood (1984). Animal cells were examined by both light microscopy and scanning electron microscopy using the methods of Gamliel et al. (1983). Production of toxic metabolites Toxin production was studied in media containing 2% yeast exact with either 4% or 20% sucrose. Flasks (4 X 250 ml) containing 100 ml of sterile medium were inoculated with a spore suspension of Wallemia and incubated at 25°C, without shaking, for 2-10 weeks. Mycelium was removed by filtration, washed with several portions of water and the culture filtrate was defatted by extraction with «-hexane. The toxic material was obtained by repeated extraction with chloroform (from the filtrate), combining the chloroform extracts and concentrating by evaporation under reduced pressure in a rotary evaporator. The procedure was scaled up to the inoculation of 2-litre flasks containing 500 ml of medium (2% yeast extract + 20% sucrose) for purification of toxic metabolites. Purification of toxic metabolites The concentrated chloroform extract was loaded as a streak on to TLC plates of silica gel (Merck 5553) and chromatographed with toluene:ethyl acetate:formic acid (5:4:1). All solvents used for extraction and chromatography were distilled prior to use to remove non-volatile residues. The chromatogram was divided into three fractions (Fn 1, Rf 0-4; Fn 2, 0-4 R{ 0-7; Fn 3, Rt 0-7) and each fraction was cut into small portions and shaken for 15 min with 75 ml methanol. The strips were removed, the extraction repeated and the methanol extracts pooled and evaporated to dryness on a rotary evaporator.
Mycotoxins from Wallemia
71
The three fractions were assayed for toxicity using brine shrimp larvae and were further examined by TLC after spraying with /7-anisaldehyde reagent (0-5% in methanol:acetic acid:conc. sulphuric acid (80:15:5)). Further purification was carried out by preparative HPLC using an amino column (Spherisorb 5 fim, 250 x 4-9 mm from Hichrom, Reading UK). The mobile phase was butan-1-ol:«hexane (10:90) at a flow rate of 2 ml/min and the detector was an LDC Spectromonitor D UV detector set at 230 nm. Samples were injected via a 100 /A loop (Rheodyne 7125).
Downloaded by [Universite Laval] at 20:20 06 November 2015
Structural characterization The purified toxin was analysed by GC-MS using the following conditions: Kratos MS 80 RFA GC conditions: 25 m OV—1701, 0-25 urn, 0-3 mm i.d., 40°C for 2 min, 10°C/min to 270°C; 2 ul injected. MS conditions: (i) Electron impact—electron energy—70 eV, beam current—100 /iA, source temperature—200°C, Scan rate—650-25 amu at 1 s/decade, resolution—1000. (ii) High resolution—above conditions except resolution—3000. (iii) Chemical ionization—reagent gas—NH3, beam current—500 pA, source temperature 200°C, electron energy—n 100 eV (tuned to maximum sensitivity), scan rate—650-100 amu at 1 s/decade, resolution—1000 or 3000. (iv) Trimethylsilyl (TMS) derivative—the toxin was converted into its TMS derivative by adding 250 /«I of reagent (N-methyl-A'-trimethylsilyl trifluoroacetamide) to 1 mg of toxin and warming for 15 min at 70°C. The sample was diluted to 5 ml in diethyl ether for GC-MS analysis. NMR spectra of the toxin (5 mg) were recorded in CDCI3 using trimethylsilane as the internal standard using a Bruker WH-90 90 MHz pulse Fourier transform NMR spectrometer; IR spectra were recorded in KBr discs using a Perkin Elmer model 1700 and UV spectra in ethanol using a Pye Unicam SP 800. Results and discussion
Growth and toxin production by isolate A were greater in 20% sucrose yeast extract medium than in the medium containing only 4% sucrose (table 1). Toxicity to brine shrimps reached a maximum after 6 weeks' incubation in both media and was significantly reduced after 10 weeks. Isolate B grew more slowly than isolate A; however, chloroform extracts of culture filtrates from isolate B were more toxic to brine shrimps and rat liver cells. Preliminary evidence suggests that the two isolates produce related, but distinct, toxins, which we propose to call walleminol A and B. Walleminol B showed a similar fragmentation pattern to walleminol A when analysed by GC-MS. It was shown, however, to have a higher molecular weight of 252. It also showed a colour difference when treated with /7-anisaldehyde and a small difference in R( when compared with the TLC of walleminol A. Purification of walleminol A Fractionation of metabolites in the chloroform extract of isolate A was followed using both brine shrimp and animal cell cultures to monitor toxicity, as well as TLC on silica gel plates developed with toluene:ethyl acetate:formic acid (5:4:1)
G. M. Wood et al.
72
Table 1. Effect of incubation time on the toxicity of Wallemia extracts. Growth medium
Downloaded by [Universite Laval] at 20:20 06 November 2015
Yeast extract + 4 % sucrose
Yeast extract +20% sucrose
Percentage kill of brine shrimps (dilution)
Incubation time (weeks)
Dry weight of mycelium (g)
1:50
1:200
2 3 4 6 10
0-22 0-35 0-69 1-58 1-27
92 100 100 100 100
0 0 4 100 32
Percentage kill of brine shrimps (dilution)
1:400
Dry weight of mycelium (g)
1:50
1:200
1:400
0 0 0 90 0
1-62 3-47 3-06 3-84 4-29
90 100 100 100 100
0 100 100 100 100
0 10 100 100 80
No effect was caused by extracts of media blanks.
(TEF), and visualized by spraying with /7-anisaldehyde followed by heating for 5 min at 100°C. Toxicity was associated with fraction 2 from the preliminary fractionation procedure (see methods), and further fractionation and purification was achieved using the protocol shown in figure 1, the position of bands being visualized by spraying strips from each edge of the chromatograms with the /»-anisaldehyde reagent. The fraction F2 A2 (figure 1) was toxic to both brine shrimps and rat liver cells, although it was noticeable that purification seemed to result in a decrease in toxicity to brine shrimps. During the purification studies 'equivalent' amounts of the crude
F2C1 Non-toxic
Chloroform extract from aqueous media (6 weeks incubation in YE+ 20% sucrose)
Tolueneethyl acetate: formic acid 5-4:1
Chloroform-.methanol 97:3
Chloroform:methanol 99:1
Chloroform:mettianol 9:1
Figure 1. Separation of compounds from Wallemia by TLC purification.
Mycotoxins from Wallemia
73
1 100 908070-
I 6500 "
I " |
40-
cx. 3020-
Downloaded by [Universite Laval] at 20:20 06 November 2015
100 20:36
2203
23:31 24:58 Retention time (min:s) Figure 2. Analysis of extract F2 A2 by GC.
27:53
26:26
extracts were tested for toxicity to brine shrimps and rat liver cells. The decrease in toxicity could have been due to the removal of compounds such as lipids (which are known to affect brine shrimps) or to losses of toxin during purification. Further analysis of both crude extracts and purified toxin by HPLC was performed in the final stages of this work. This indicated that the observed decreased effect on toxicity was not solely due to toxin losses but the removal of other compounds, possibly acting synergistically with the toxin, could be contributing to the toxicity. Walleminol A was separated from interfering compounds by preparative HPLC; the major impurity compound (as seen in figure 2) was identifed as 3-indole ethanol (tryptophol) by GC-MS and by comparison with authentic standards. To our knowledge this is the first report of tryptophol from Wallemia. 69
100 90
70 g 60•- 5 0 -
79
41
I |
91 yl
105
98 40-
3053
20100
218
I llllkl.l,lllll,.I.I.L.l,,,I...J,, i * 50
100
150 m/z
200
Figure 3. Electron impact spectrum of F2 A2.
250
Downloaded by [Universite Laval] at 20:20 06 November 2015
74
G. M. Wood et al.
Characterization of walleminol A Analysis of F2 A2 by GC-MS showed only two peaks, the window containing these peaks is shown in figure 2, the major peak had a probable molecular ion at m\z 236 (figure 3). This was confirmed by the presence of ions at mjz 237, 219 and 201 in the spectrum using positive chemical ionization, corresponding to the sequential loss of two molecules of water from a molecule of mass 236 (figure 4) and by a major ion at mjz 235 in the negative chemical ionization spectrum. The trimethylsilylated derivative also showed two peaks in the same ratio as the underivatized GC trace. The major peak had a molecular ion at m\z 380 corresponding to a bis (trimethylsilyl) derivative of a molecule of mass 236. Further GC-MS analysis of the material purified by HPLC confirmed the observations obtained on F2 A2. High-resolution mass spectrometry in NICI mode at resolution 3000 gave M-l peak of 235-170 ± 0-002, suggesting that walleminol A has the molecular formula C15H24O2. The experimental error was determined by mass measurement of a standard ion of known composition. The NMR and IR spectra are shown in figures 5 and 6, and are consistent with the presence of an olefin bond (a single proton CDCI3 at 4-7 ppm, end absorption in the UV and a peak at 835 cm"1 in the IR) and four methyl groups, two of which may represent a gem-dimethyl system (1120 cm" 1 in the IR and four 3-proton signals: CDCI3 at 0-85, 1-06, 1-21 and 1-29 ppm), and two hydroxyl groups (3300 cm"1 in the IR and a broad 2-proton signal at CDCI3 4-9-5-3 ppm in the NMR). The four double bond equivalents implied by a molecular formula C15H24O2 could thus be accounted for as three rings and a trisubstituted double bond, although structures containing fully substituted unconjugated double bonds in place of one or more rings cannot be ruled out. All the evidence presented is consistent with walleminol A being a tricyclic dihydroxysesquiterpene.
219
237 .1... 300
Figure 4. Positive chemical ionization spectrum of F2 A2.
400
Downloaded by [Universite Laval] at 20:20 06 November 2015
Mycotoxins from Wallemia
75
-iM*^
6
5
4
3
2
6 ppm
Figure 5. NMR spectrum of pure toxin produced by Wallemia.
Toxicity of walleminol A Brine shrimp. Brine shrimps are normally extremely active swimmers, and it has been suggested that if shrimps can be seen to be 'moving but not making forward progress' they should be considered as 'dead' for the assay (Eppley 1974). Walleminol A at 80 ;tg/ml caused 41% kill of shrimps, with the remaining showing movement but not swimming. This would be equivalent to 100% 'dead'. Further dilutions of the toxin showed that a concentration of 40 /ig/ml caused 30% kill with a further 20% of weak shrimps. The LD50 (including weak shrimps) was therefore approximately 40 /*g/ml. Protozoa. Death of the protozoa (71 pyriformis) was caused by concentrations of 50 /tg/ml of the toxin and above. Further dilutions of the toxin caused restriction of • 40
"4000
3500
3000
2500
2000 1800 1600 Frequency/cm
1400
1200
1000
Figure 6. IR spectrum of pure toxin produced by Wallemia.
800
500
76
G. M. Wood et al.
growth and a lag in the growth phase. A concentration of 25 /tg/ml of the toxin allowed 50% growth compared with the controls and 10 /xg/ml had no effect on the protozoa.
Downloaded by [Universite Laval] at 20:20 06 November 2015
Cell lines. High concentrations of the toxin (500 jig/ml) caused rapid death of both rat liver cells and BHK cells. Lower concentrations (50-125 /tg/ml) caused a restricted growth rate and affected the morphology of the cells. The toxin caused changes in the nuclei of both cell lines and, in the case of the BHK cells, the formation of giant cells was apparent. The rat liver cells were subject to cytoplasmic vacuolation. All changes were visible after examination by light microscopy and electron microscopy; the latter indicated that the nuclei of toxin-treated cells contained unusual membrane-like structures. The minimum inhibitory dose of the toxin was approximately 50 /tg/ml. Discussion This study outlines the isolation of a compound produced by the mould Wallemia sebi, to be named walleminol A. The compound causes toxic effects in a range of in vitro systems such as mammalian cell lines, protozoa and brine shrimp. The dose affecting these systems is comparable with such mycotoxins as penicillic acid and citrinin. The timescale of this study did not allow a complete characterization of the compound, but data collected have allowed the following information. The toxin has a molecular weight 236, composition C15H24O2. It contains two hydroxyl groups, four methyl groups, including a gem dimethyl, and there are two or three ring structures in the molecule. There is no literature available on this compound, suggesting that it has not been previously characterized. The work reported includes only a study of the mould growth in culture media, which does not assess the significance of this mould in the food environment. However, information suggests that this mould is a greater contaminant of foods than previously realized (Saito et al. 1971, 1974, Pitt and Hocking 1985), and we have found in our surveillance studies that Wallemia is a common contaminant of cake and jam samples. Further work on the analysis of foods for walleminol A will be reported at a later date. Acknowledgements This work was supported by funding from the Ministry of Agriculture, Fisheries and Food. Many thanks are due to Kathy Groves for electron microscopy studies, and Mr J. P. Bloxsidge and Dr G. J. Buist at the University of Surrey for obtaining the NMR and IR data. References EPPLEY, R. M., 1974, Mycotoxins: sensitivity of brine shrimp (Artemia salina) to trichothecenes. Journal of the Association of Official Analytical Chemists, 57, 618-620. GAMLIEL, H., GURFEL, D., LEIZEROWITZ, R. and POLLIACK, A., 1983, Air drying of human leucocytes for
scanning electron microscopy using the GTGO procedure. Journal of Microscopy, 131, 87-95. HARWIG, J. and SCOTT, P, M., 1971, Brine shrimp (Artemia salina) larvae as a screening system for fungal toxins. Applied Microbiology, 21, 1011-1016. HAYES, A. W. and WYATT, E. P., 1970, Survey of the sensitivity of micro-organisms to rubratoxin B. Applied Microbiology, 20, 164-165.
Mycotoxins from Wallemia
77
HITOKOTO, H., MOROZUMI, S., WAUKE, T., SAKAI, S. and KURATA, H., 1981, Fungal contamination and
mycotoxin producing potential of dried beans. Mycopathologia, 73, 33-38. HOCKING, A. D. and PITT, J. I., 1980, Dichloran-glycerol medium for enumeration of xerophilic fungi from low moisture foods. Applied and Environmental Microbiology, 39, 488-492. HURLOCK, E. T., ARMITAGE, D. M. and LLEWELLIN, B. E., 1980, Seasonal changes in mite ocari and
fungal populations in aerated and unaerated wheat stored for 3 years. Bulletin of Entomological Research, 70, 537-548. MILLS, J. T. and ABRAMSON, D. 1981, Microflora and condition of flood damaged grain in Manitoba, Canada, Mycopathologia, 73, 143-152. PITT, J. I., 1975, Xerophilic fungi and the spoilage of foods of plant origin. Water Relations of Foods, edited by R. B. Duckworth (London: Academic Press), pp. 273-307. PITT, J. 1. and HOCKING, A. D. 1977, Influence of solute and hydrogen ion concentration on the water relations of some xerophilic fungi. Journal of General Microbiology, 101, 35-40.
Downloaded by [Universite Laval] at 20:20 06 November 2015
SAITO, M., ISHIKO. T., ENOMOTO. M., OHTSUBO, K., UMEDA, M., KURATA, H., UDAGAWA, S.,
TANIGUCHI, S. and SEKITA, S. 1974, Screening test using HeLa cells and mice for detection of mycotoxin-producing fungi isolated from foodstuffs. An additional report on fungi collected in 1968 and 1969. Japanese Journal of Experimental Medicine, 44, 63-82. SAITO, M., OHTSUBO, K., UMEDA, J., ENOMOTO, M., KURATA, H., UDAGAWA, S., SAKABE, F. and
ICHINOE, M., 1971, Screening tests using HeLa cells and mice for the detection of mycotoxinproducing fungi isolated from foodstuffs. Japanese Journal of Experimental Medicine, 41, 1-20. SCARR, M. P., 1959, Selective media used in the microbiological examination of sugar products. Journal of the Science of Food and Agriculture, 10, 678-681. TAKATORI, K. and KONDOH, S., 1979, Fungal distribution of swine feeds and the mycotoxin producibility of Aspergillus flavus and Aspergillus versicolor. Japanese Journal of Zootechnology and Science, 50, 453-459. TSURUTA, O. and SAITO, M. 1980, Mycological damage of domestic brown rice during storage in warehouses under natural conditions. 3. Changes in mycofloras during the storage. Nikon Kin Gakkai Kaiho, 21, 121-125. WOOD, G. M., 1984, Assessment of toxigenic moulds in foods by means of a biological screening method. Mycotoxins in Animal and Human Health, edited by M. O. Moss and M. Frank (Guildford: University of Surrey), pp. 95-105.
