Patricia Taillandier, Laboratoire de Génie Chimique, UMR5503 (CNRS/INPT/UPS), Department Bioprocedes & Systemes Microbiens, ENSIACET-INPT, 5, rue Paulin Talabot, BP 1301, 31106 Toulouse Cedex, France. E-mail: firstname.lastname@example.org
Aims: To study the interaction between Bacillus spp. and contaminating Aspergillus flavus isolated strains from Thai fermented soybean in order to limit aflatoxin production. To study the detoxification of aflatoxin B1 (AFB1) and ochratoxin A (OTA) by Bacillus spp. in order to find an efficient strain to remove these toxins.
Methods and Results: One A. flavus aflatoxin-producing strain and 23 isolates of Bacillus spp. were isolated from soybean and fresh Thua-nao collected from the north of Thailand. Inhibition studies of A. flavus and A. westerdijkiae NRRL 3174 (reference strain) growth by all isolates of Bacillus spp. were conducted by dual culture technique on agar plates. These isolates were also tested for AFB1 and OTA detoxification ability on both solid and liquid media. Most of the strains were able to detoxify aflatoxin but only some of them could detoxify OTA.
Conclusions: One Bacillus strain was able to inhibit growth of both Aspergillus strains and to remove both mycotoxins (decrease of 74% of AFB1 and 92·5% of OTA). It was identified by ITS sequencing as Bacillus licheniformis. The OTA decrease was due to degradation in OTα. Another Bacillus strain inhibiting both Aspergillus growth and detoxifying 85% of AFB1 was identified as B. subtilis. AFB1 decrease has not been correlated to appearance of a degradation product.
Significance and Impact of the Study: The possibility to reduce AFB1 level by a strain from the natural flora is of great interest for the control of the quality of fermented soybean. Moreover, the same strain could be a source of efficient enzyme for OTA degradation in other food or feeds.
Thua-nao is one of the oldest traditionally fermented soybean products produced for years by the people in many small villages up north of Thailand. Bacillus spp., Gram-positive, strict or facultative aerobe and endospore-forming bacteria, were found to be the dominant microflora of this product (Chantawannakul et al. 2002). However, during fermentation, the soybeans would be sometimes contaminated with toxigenic fungi for which the intrinsic and extrinsic factors may also induce the mycotoxin formation, leading to health risk for consumers (Garcia et al. 1997).
The mycotoxins of the greatest significance in food and feeds are aflatoxins which are of great concern because of their detrimental effects on the health of humans and animals, including carcinogenic, mutagenic, teratogenic, and immunosuppressive effects (Eaton and Gallagher 1994). They are produced mainly by Aspergillus flavus, A. parasiticus or A. nominus. The four main naturally produced aflatoxins are B1, B2, G1 and G2, with the former two being the aflatoxins found at the highest concentration in contaminated food and feed.
Another important mycotoxin which has received increasing interest from both scientific communities and food committees is ochratoxin A (OTA). OTA is produced by A. ochraceus and related species. Aspergillus section Nigri, for example, A. niger aggregate and A. carbonarius, were shown to be responsible for OTA production in grapes (Abarca et al. 2001; Cabaňes et al. 2002). Penicillium spp., particularly Penicillium verrucosum is OTA producer as well (Pitt 1987). OTA is a derivative of isocoumarin linked to l-phenylalanine and it has a potent nephrotoxic (Krogh et al. 1974; Mortensen et al. 1983), teratogenic (Arora and Fröelén 1981), genotoxic (Dirheimer 1998), immunosuppressive (Haubeck et al. 1981) and carcinogenic (Boorman 1989) properties. If the consumers take moderate to low concentrations of this mycotoxin in the long term, much more serious problem may arise.
The present study has been then performed with two objectives. The first one was to study the interaction between Bacillus spp., isolated strains from Thai fermented soybean product, and contaminating A. flavus in soybean in order to limit aflatoxin production. The second objective was to study the detoxification of AFB1 and OTA by Bacillus spp. in order to find an efficient strain to remove these toxins.
Materials and methods
Isolation of Bacillus spp. from Thai fermented soybean product (Thua-nao)
Three samples of fresh Thua-nao were collected from the local market in ChiangMai and Maehongson province, which are located in the North of Thailand. One gram of each sample was suspended in 9 ml of phosphate buffer, mixed well by vortex and agitated for 30 min. Appropriate dilutions (up to 10−7) were made with phosphate buffer as diluent. One hundred microlitres of each dilution were spreaded on nutrient agar and incubated at 37°C for 24 h. Colonies appeared on plate were picked up and maintained on nutrient agar [3 g beef extract (Difco laboratories, USA), 5 g peptone (Fluka BioChemika, Buchs, Switzerland), 15 g agar (Difco laboratories) in 1 l of distilled water) at 4°C. Stock culture were cultivated on nutrient broth (NB) and stored at −20°C after addition of sterile glycerol at 20%.
Identification of interested isolates of Bacillus spp.
Twenty-three colonies were first characterized by Gram’s stain technique and then identified as Bacillus spp. by its sugar utilization pattern in the API 50CH system (API 50 CHB medium, BioMèrieux® SA 673 620 399 RCS Lyon, 69280 Marcy-l’Etoile, France). All isolates were maintained on nutrient agar.
Screening of the isolated Bacillus spp. for growth inhibition of Aspergillus flavus and Aspergillus westerdijkiae NRRL 3174
Aspergillus westerdijkiae NRRL 3174 was chosen as OTA producer reference strain. Aspergillus flavus strain used in this study was isolated from soybean in Thailand and its ability to produce aflatoxin was confirmed as described by Bragulat et al. (2001). Growth inhibition studies of isolated A. flavus and A. westerdijkiae by all isolates of Bacillus spp. were conducted by dual culture technique. The culture media used were nutrient agar and potato dextrose agar (PDA) for A. flavus and nutrient agar and Czapek yeast extract (CYA) agar for A. westerdijkiae NRRL 3174. They were poured into 100-mm Petri dishes. After agar solidification, the fungal strains were inoculated in the centre and then each isolate of Bacillus spp. was inoculated at 10 mm from the edge of the Petri dish. They were then incubated at 25°C and 37°C for 7 days. The experiment was carried out in triplicate. Radial growth reduction was calculated as follows:
where a = radial growth measurement of the pathogen in control and b is that of the pathogen in the presence of bacteria tested; these values were expressed as percentage inhibition of radial mycelial growth.
Degradation of AFB1 and OTA by 23 Bacillus spp. isolates
Standard AFB1 and OTA were used through this experiment. One hundred microlitres of Bacillus spp. from frozen stock suspension was inoculated in 5 ml of NB filled in 15-ml Falcon tube. Standard AFB1 or OTA at initial concentration of 5 mg l−1 was added. The cultures of Bacillus spp. with toxin were then incubated at 37°C in shaking incubator with speed of 150 rev min−1 for 1 week.
Kinetics of AFB1 and OTA degradation during growth of the two selected isolates (Bacillus CM 21 and Bacillus MHS 13)
One millilitre of each frozen Bacillus stock culture (Bacillus CM 21 and Bacillus MHS 13) was transferred to a 250-ml Erlenmeyer flask (Pyrex, USA) containing 100 ml of sterilized NB and shaking incubated at 37°C for 48 h. The optical density (OD) of the culture broth was measured by using a spectrophotometer at 660 nm every 6 h. After 48 h, the cultured broth was used to inoculate 100 ml of fresh sterilized NB. The culture was then diluted to an absorbance approximately 0·1 with fresh media. The initial concentration of 5 mg l−1 of each standard AFB1 or OTA were added into NB. All cultures were incubated at 37°C with agitation. Five millilitres of each culture were sampled twice a day for 5 days and then once a day for the following 10 days. OD at 660 nm was measured at each time of sampling.
Aflatoxin B1 analysis by HPLC
AFB1 extraction was made according to Sánchez et al. (2005) modified method. Ten millilitres of acetone and 15 ml of dichloromethane were added to NB. Following the addition of each solvent, the cultures were shake-agitated at 150 rev min−1 for 30 min. The filtrate was partitioned in a preparatory funnel into an aqueous phase and a dichloromethane phase that contained the most of the aflatoxin. The aqueous phase was partitioned again with 15 ml of fresh dichloromethane. The two dichloromethane fractions were combined. Residual H2O was removed from the final dichloromethane solution with 1 g Na2SO4 and the solution was filtered with Whatman paper No.1. The filtrate was evaporated at 40°C (90 rev min−1) until dry. The residue was dissolved with 1 ml acetonitrile, filtered (Millex® HV PVDF 0·45 μm; Millipore, Billerica, MA) and analysed by HPLC.
The HPLC method for AFB1 analysis was modified from Chan et al. (2004). The HPLC apparatus consisted of a solvent delivery system with both fluorescence (λex = 364 nm; λem = 440 nm) and ultraviolet (UV) detectors (λ = 225 nm and 362 nm). The spectra range is from 200 to 500 nm. The analytical column used was a 150 × 4·6 mm Uptisphere 5 μm C18 ODB fitted with a guard column of 10 × 4 mm. The column temperature was 25°C. An aliquot of sample (80 μl) was injected using an autoinjector (BIO-TEK, Milan, Italy). The mobile phase was 0·1% phosphoric acid (A) and methanol/acetonitrile (50 : 50) (B) delivered at flow rate of 1 ml min−1 for 30 min. The sample was eluted with a linear gradient for 30 min. AFB1 and its derivatives were detected by comparing the elution time and maximum absorption of UV with the standards. The relative standard deviation of this analysis method is of 6%.
Ochratoxin A analysis by HPLC
Thirty-four microlitres of 12 N HCl (NormaPur; Prolabo, R.P.) was added to the culture followed by 3·5 ml of chloroform (Prolabo, R.P. NormaPur). The solution was partitioned in a preparatory funnel into an aqueous phase and a chloroform phase that contained most of the OTA. The chloroform phase was recuperated and then evaporated under nitrogen gas at 65°C (90 rev min−1) until dry. The samples were then dissolved in 1 ml methanol (Fisher Scientific), filtered (Millex®HV PVDF 0·45 μm) and analysed by HPLC as described by Awad et al. (2005).
The HPLC apparatus consisted of a solvent delivery system with fluorescence detector (λex = 332 nm; λem = 466 nm). The analytical column used was a 150 × 4·6 mm Uptisphere 5 μm C18 ODB fitted with a guard column of 10 × 4 mm. The column temperature was 30°C. An aliquot of sample (80 μl) was injected using an autoinjector (Bio-Tek). The mobile phase was acetic acid in water 0·2% (A) and acetonitrile (B) delivered at flow rate of 1 ml min−1 for 45 min. The sample was eluted with a linear gradient from 10% to 50% of B over the first 30 min followed by a linear gradient to 90% of B from 30 to 35 min and then a steady flow of 90% of B through 8 min and then reduced to 10% of B through 2 min. OTA and its derivative products were detected by comparing the elution time with the standards. The relative standard deviation of this analysis method is of 5·4%.
All experiments were carried out in triplicate or duplicate. spss software for Windows was used for the statistical analysis of the data. For Bacillus strains comparison, the differences were considered significant if the associated P-values were <5%. The significant differences between strains have been pointed out by anova analysis and Duncan’s multiple range test.
Bacillus isolation and identification
Three samples of fresh Thua-nao, respectively two and one of them, were collected from the local market in ChiangMai and Maehongson province, which are located in the north of Thailand, and they were labelled with the code FTN 01, FTN 02 and FTN 03.
Twenty-three colonies were isolated from these three samples on PDA-nutrient agar. Seven and eight isolates of Bacillus spp. were isolated from the two samples collected from ChiangMai province and named respectively CM 11 to CM 17 and CM 21 to CM 28. The eight isolates collected on the sample from Maehongson province were named as MHS 11 to MHS 18. Morphological and physiological characteristics of all isolates were identified as Bacillus spp. and confirmed by the API 50CH system.
Screening of the isolated Bacillus spp. for growth inhibition of Aspergillus flavus and Aspergillus westerdijkiae NRRL 3174
Twelve isolates out of twenty-three isolates of Bacillus spp. could inhibit the growth of A. flavus (data not shown) at a level significantly different from the others isolates. The average percentage of inhibition for these twelve isolates was 52%. According to these results, the growth of A. flavus was inhibited by Bacillus MHS 13 significantly more than by other isolates from the same origin. Regarding the results for A. westerdijkiae NRRL 3174, eight isolates can inhibited its growth with an average percentage of inhibition of 34%. Moreover, Bacillus CM 21 is responsible for the significantly highest growth inhibition percentage.
Screening of the isolated Bacillus spp. for removal of AFB1 and OTA
The 23 isolates were tested for AFB1 and OTA removal. Results are shown in Figures 1 and 2, respectively. Most of the strains were able to detoxify aflatoxin but only some of them could detoxify OTA. After 1 week, at least 60% AFB1 was degraded by 11 Bacillus strains. Among them, CM 17, CM 23 and MHS 13 were the most efficient. The highest percentage of OTA removal (81%) was obtained by Bacillus CM 21. As this last strain was able to degrade more than 60% AFB1, it has been retained like strain MHS 13 for the continuation of work.
These two interesting Bacillus isolates, Bacillus CM 21 and Bacillus MHS 13, were identified by ITS sequencing. The results of sequence analysis with GenBank database showed that Bacillus CM 21 is B. licheniformis (99% identity) and Bacillus MHS 13 is B. subtilis (100% identity).
Kinetics AFB1 and OTA removal during growth of the Bacillus licheniformis CM 21 and Bacillus subtilis MHS 13
Detoxification kinetics of AFB1 and OTA in Bacillus CM 21 and Bacillus MHS 13 liquid culture were investigated. We could observe that AFB1 decreased during the first 5–6 days (Figs 3 and 4). The percentage of AFB1 removal by Bacillus CM 21 and Bacillus MHS 13 was 74% and 85%, respectively.
Regarding OTA degradation (Fig. 5), Bacillus CM 21 can degrade OTA within 48 h of incubation, whereas ochratoxin α was increasing during the same time (data not shown). The percentage of OTA degradation by this isolate was 92·5%. On the opposite, there was no OTA degradation during the growth of Bacillus MHS 13 as observed in previous experiment on agar plates.
Table 1 shows some characteristics of Bacillus growth for both strains: duration of growth phase and produced ammonium.
Table 1. Some characteristics for Bacillus growth in liquid medium
Duration of growth phase (h)
Time for maximal production of ammonium (h)
Maximal concentration of ammonium (mg l−1)
B. licheniformis CM 21
B. subtilis MHS 13
In the present study, Bacillus spp. was investigated for Aspergillus growth inhibition for mycotoxins producing strains and for AFB1 and OTA reduction ability. Bacillus spp. were found to be the dominant microflora of Thai fermented soybean (Thua-nao) (Chantawannakul et al. 2002). These fermentations are characterized by extensive hydrolysis of protein to amino acids, peptides and ammonia and a rise of the pH. The pH increase observed during fermentation presumably resulted from proteolysis and the release of ammonia because of utilization of amino acids for growth (Terlabie et al. 2006).
The chemical approach to the detoxification of aflatoxins that have received considerable attention is ammoniation (Smith and Harran 1993; Bhatnagar et al. 2002). Ammoniation under appropriate conditions results in a significant reduction in the level of aflatoxins in contaminated peanut and cottonseed meals. The mechanism for this action appears to involve hydrolysis of the lactone ring and chemical conversion of the parent compound AFB1 to numerous products that exhibit greatly decreased toxicity. Two major products identified as compounds with molecular weight 286 (AFD1) and molecular weight 206, have been isolated and tested in various biological systems (Park et al. 1988). The first step in the reaction is reversible if the ammoniation process is carried out under mild conditions. This is a disadvantage of the method. However, when the reaction is allowed to proceed past the first step, the products formed do not revert back to AFB1. Reaction products of ammoniation are dependent on temperature, pressure and the source of ammonia. This method is quite energy consuming and presents some risks from security point of view.
In order to find biological alternative methods, micro-organisms (including yeasts, mold and bacteria) have been screened for their ability to modify or inactivate aflatoxins. Flavobacterium aurantiacum (NRRL B-184) was shown to remove aflatoxin from a liquid medium significantly without the production of toxic by-products (Ciegler et al. 1966). The same investigators also found that certain acid-producing molds could catalyse the hydration of AFB1 to B2a (a less toxic product). In the case of some Aspergillus strains that can be at the same time aflatoxin producer and able to degrade it, peroxydases activities where shown to catalyse aflatoxin degradation (Smith and Harran 1993). For Bacillus strains, this has never been reported. In our case, no degradation product of AFB1 could be detected by HPLC analysis. Growth and ammonium production occurred during the first hours of culture, whereas AFB1 decrease took several days (Table 1). So, the mechanism remains unknown.
Regarding OTA, several reports describe OTA degrading activities of the microbial flora of the mammalian gastrointestinal tract including rumen microbes of cow and sheep (Galtier and Alvinerie 1976; Hult et al. 1976; Park et al. 1988) and microbes living mainly in the caecum and large intestine of rats (Madhyastha et al. 1992). The human intestinal microflora can also partially degrade OTA (Akiyama et al. 1997). The species responsible for OTA detoxification have not yet been identified although mainly protozoa were suggested to take part in the biotransformation process in ruminants (Kiessling et al. 1984). Degradation of OTA was observed in milk as a result of the action of Lactobacillus, Streptococcus and Bifidobacterium spp. (Skrinjar et al. 1996), whereas two other bacteria, Acinetobacter calcoaceticus (Hwang and Draughon 1994) and Phenylobacterium immobile (Wegst and Lingens 1983), could also convert OTA to the much less toxic ochratoxin α in liquid cultures. Furthermore, recent reports describe the OTA degrading activities of some Aspergillus and Pleurotus isolates and/or their enzyme (Varga et al. 2000). In our case, we observed that OTA is also efficiently detoxified by some Bacillus isolates and specifically B. licheniformis CM21 and this has not been reported before for this species. Similarities between OTA degradation kinetics by A. niger and Bacillus isolates and the detection of the degradation product, ochratoxin α, in the ferment broth of B. licheniformis suggest that a carboxypeptidase A activity may be responsible for OTA decomposition by these isolates.
The possibility to reduce AFB1 level by a strain of B. licheniformis from the natural flora is of great interest for the control of the quality of fermented soybean. Moreover, the same strain could be a source of efficient enzyme for OTA biodegradation in other food or feeds. Accordingly, further studies are in progress to determine which enzymes take part in the detoxification process in Bacillus isolates, and to establish the absence of by-products and with any residual toxicity, and probably also genetic modification, are required to harness the potential of these strains.
AFB1 and OTA are a toxic compound that needs to be manipulated with care and with appropriate safety precautions. Decontamination procedures for laboratory wastes have been reported by the International Agency for Research on Cancer (IARC).
The present work was supported by the grants from The Franco-Thai Cooperation in Higher Education and Research in Tropical Agronomy and Agro Industry (Joint MSc and PhD Programs with Thai Universities in Agricultural Sciences) and The French Embassy (Bangkok, Thailand).
The authors also thank for project coordinator assistance, Dr Sunthorn Kanchanatawee.