The bacterial culture of M. fulvus ANSM068 reduced AFB₁ by 80.7% after co-incubation at 30 °C for 72 h. The culture supernatant of M. fulvus ANSM068 was able to reduce AFB₁ by 76.6%. However, no significant reduction was observed in the treatments with the cells and cell extracts (Figure 1). Culture supernatant treated with proteinase K (Prok) displayed significantly decreased reduction ability, from 76.6 to 18.9%. No reduction of AFB₁ concentration was observed when the culture supernatant was treated with Prok plus SDS or with heating (boiling water bath for 10 min) (Figure 1).

Myxobacteria became well-known as prolific producers of secondary metabolites during the past few decades [37,38]. Thus far, no pathogenic myxobacteria have been observed and many bioactive components and drugs were isolated from myxobacteria [39–41]. Myxococcus fulvus was first documented and characterized in 1969 [42]. It is a species of myxobacteria, a group of Gram-negative eubacteria with rod-shaped cells. The colonial and cell morphology of M. fulvus ANSM068 are shown in Figure 2. It was previously found to be capable of reducing AFB₁ [22]. Up to now, no study has focused on mycotoxin control by using myxobacteria. This is the first report on mycotoxin biotransformation by a bacterial strain belonging to myxobacteria.

Reduction of AFB₁ by culture supernatant produced without pre-exposure to AFB₁ suggested that it was achieved during the normal growth of the bacterium, indicating that the reduction was a constitutive activity of M. fulvus ANSM068. Results from proteinase K, proteinase K plus SDS and heating treatments implied that the reduction of AFB₁ was due to biotransformation instead of cell binding. Protein or enzymes in culture supernatant might be involved in AFB₁ biotransformation by the bacterial strain.

Nitrogen sources have a dramatic effect on AFB₁ biotransformation by M. fulvus ANSM068. Yeast extract was found to be the best nitrogen source for AFB₁ biotransformation and resulted in 76.6% AFB₁ biotransformation (Figure 3). In comparison, AFB₁ biotransformation was significantly (P < 0.05) lower with ammonium nitrate (25.8%). Among the four levels of yeast extract tested (0.2, 0.5, 0.8 and 1.2%), 0.5% was optimal for AFB₁ biotransformation (76.6%). Nitrogen in substrates is one of the most important factors influencing myxobacteria growth and production of metabolites. Chun [43] studied the effect of different nitrogen sources on the production of antifungal metabolites by myxobacteria in liquid culture and found that yeast extract improved secretion of antifungal metabolites but fish meal, soybean powder and inorganic nitrogen treatments showed no function. A similar trend was also observed in this study. Yeast extract was most favored in AFB₁ biotransformation by M. fulvus ANSM068 and inorganic nitrogen was far less effective. The reasons could be that in natural environments these myxobacteria feed on other organisms such as eubacteria or yeasts by bacteriolysis or cellular lysis [44]. The nutrients ratio such as amino acid and vitamins in yeast extract might be more suitable to the bacterial strain for its growth and production of AFB₁ biotransformation components.

The initial pH value in culture medium showed a significant effect on AFB₁ biotransformation (Figure 4). The highest AFB₁ biotransformation rate was detected at pH values of 6.5 to 7.5 (P < 0.05). It has been well documented that the optimum pH range for myxobacteria is 6.8–7.8 in liquid culture [45]. Ahn et al. [46] studied the effect of pH value on antifungal metabolites produced by a Myxococcus isolate and found that a pH value between 6.5 and 7.6 was favored and the optimal pH value was 7.2.

The effect of temperature on AFB₁ biotransformation by M. fulvus ANSM068 is shown in Figure 5. The % biotransformation exhibited was significantly lower (P < 0.05) at 15 °C (11.0%) and 40 °C (18.9%) than at 30 °C (76.6%). Since the bacterial strain originated from deer feces, a temperature at 30 °C should be more suitable for the survival and growth of the bacterium, thus optimal for its enzyme system.

In the analysis by LCMS, AFB₁ standard was identified at m/z = 313 for the protonated cation [M + H]⁺, m/z = 335 for the sodium adduct of AFB₁ [M + Na]⁺, and at m/z = 351 for the potassium adduct [M + K]⁺ (Figure 6A); VY/2 medium supplemented with AFB₁ clearly showed the above three distinct ions (Figure 6B). However, the distinct AFB₁ ion at m/z =351 disappeared while the ions at m/z = 313 and m/z = 335 significantly decreased after AFB₁ was co-incubated with M. fulvus ANSM068 culture supernatant for 72 h (Figure 6C). These results suggest AFB₁ level was significantly reduced although it was still present.

Additionally, the HPLC chromatograph (Figure 7) confirmed that AFB₁ (peak eluting at 8.27 min) was biotransformed to a new product eluting at 3.17 min. The IR spectrum showed that there were three major modifications in absorption peaks between the standard AFB₁ and the product resulting from culture supernatant treatment (Figure 8). The absorption peak at 1728 cm⁻¹ in AFB₁ standard disappeared after the treatment, which indicates that the lactone attached to the benzene ring was modified. The two peaks in the area between 1658 and 1634 cm⁻¹ in AFB₁ standard changed into one peak after treatment. The change at 2930 cm⁻¹ indicates modification of methyl group.

Alberts et al. [19] explored the AFB₁ biotransformation product with Rhodococcus erythropolis extracellular fractions by electro spray mass spectrometry and liquid chromatography mass spectrometry analysis. By comparing the three distinct AFB₁ peaks among treatments, AFB₁ was still present but at a lower concentration. No breakdown products were revealed, which suggested that AFB₁ was most likely metabolized to biotransformation products with chemical properties different from that of AFB_1. Liu et al. [12] compared the infrared spectrum between AFB₁ standard and fungal enzyme treated AFB₁. Opening of the difuran ring was proposed since the group of five bands in the area between 1000 cm⁻¹ and 1200 cm⁻¹ did not appear in the spectrum of the enzyme treated AFB₁. Marisa and Das [4] analyzed the biotransformed AFB₁ structure by fluorescence spectra measurements and TLC. Treatment of AFB₁ with purified enzyme resulted in a decrease in fluorescence intensity, suggesting the enzymatic cleavage of the lactone ring. The lactone ring was confirmed to be responsible for toxicity and mutagenicity in AFB₁ [47,48]. In the current study, results from LCMS, HPLC and IR analysis confirmed AFB₁ biotransformation by M. fulvus ANSM068 instead of cell binding. Modification of the lactone ring on the AFB₁ molecule may result in detoxification although further investigation needs to be done. This result implies the potential applications of M. fulvus ANSM068 or its metabolites in detoxifying AFB₁ in contaminated food or feed.

Since the culture supernatant of M. fulvus ANSM068 exhibited high AFB₁ biotransformation activity, it was selected for active components extraction. Precipitates from 80% saturated ammonium sulfate showed higher AFB₁ biotransformation activity, with 54.1% of AFB₁ biotransformed. Chromatography was used to purify enzyme from the precipitates. Two peaks were obtained by ion exchange chromatography on DEAE A-50 column and UV detection at 280 nm (Figure 9A). The second peak, which showed AFB₁ biotransformation activity (35.5%), was applied to molecular sieve chromatography on a Sephadex G-100 column (Figure 9B). SDS-PAGE showed that the active components consist of at least three proteins with a relatively small molecular weight between 20 and 66 kDa (Figure 9C).

To date, there are only a few reports on extraction or purification of mycotoxin biotransformation active components from microbial metabolites. Liu et al. [27] extracted and purified an intracellular enzyme from fungi Armillariella tabescens by ammonium sulfate precipitation, ion exchange chromatography and chromatofocusing chromatography. The AFB₁ detoxification enzyme was then characterized to have a molecular mass of 51.8 kDa by SDS-PAGE and it exhibited a specific activity of 7.09 nmol min/mg at pH 6.0 and 28 °C. Motomura et al. [6] purified an aflatoxin biotransformation enzyme from the edible mushroom Pleurotus ostreatus by two chromatographies on DEAE-Sepharose and Phenyl-Sepharose. The apparent molecular mass of the purified enzyme was estimated to be 90 kDa by SDS-PAGE. This study has confirmed that the AFB₁ biotransformation components exist in the culture supernatant and are very likely enzyme(s). Further purification and identification of the enzyme(s) are undergoing.

Myxococcus fulvus ANSM068 was isolated and purified from deer feces collected from the Beijing Zoo in July, 2006. It was identified using physiological and biochemical tests and 16S rRNA gene sequence analysis by the China General Microbiological Culture Collection Center (CGMCC) and has been deposited there as CGMCC #3194. It was preserved at −80 °C before use (China Agriculture University, Beijing, China).

AFB₁ was obtained from Sigma Chemical Co. (Bellefonte, USA). The medium used for liquid culture, VY/2 medium, contains 5 g yeast extract, 1 g CaCl₂, 0.5 g MgSO₄, 0.5 mg VB₁₂ in 1 L distilled water at room temperature.

For AFB₁ analysis, the HPLC procedure by AOAC [49] was used with slight modifications. The reaction mixtures were extracted three times with chloroform. The chloroform extracts were evaporated under nitrogen at room temperature, the residue was dissolved in 50% methanol in water (1:1, v/v) and analyzed by HPLC. HPLC analysis was performed using a LiChroCART RP-C18 (250-4 Hypersil ODS (5 μm), Merck) column with a guard column (LiChroCART 4-4 RP-C18 (5 μm), Merck). The mobile phase was methanol: water (1:1, v/v) isocratic at a flow rate of 1 mL/min. AFB₁ was derivatized by a photochemical reactor (AURA, USA) and measured by a fluorescence detector. The excitation and detection wavelengths were set at 360 and 440 nm, respectively. The percentage of AFB₁ biotransformation was calculated using the following formula:

Biotransformation of AFB₁ by M. fulvus ANSM068 was carried out in liquid culture. The bacterial isolate was cultured in VY/2 medium for 24 h and used to inoculate VY/2 (50 mL) in a 300 mL flask, which was incubated at 30 °C with agitation at 160 rpm for 50 h in a Gyrotary shaker incubator without AFB₁ (Haerbin Donglian Electronic Equipment Inc., China). AFB₁ standard solution (Sigma Chemical Co., Bellefonte, USA) was diluted with methanol (Beijing Chemical Inc., Beijing, China) to a stock solution of 500 ppb, of which 0.2 mL was added to bacterial cultures of 0.8 mL for a final concentration of 100 ppb. The biotransformation tests were conducted in the dark at 30 °C without shaking for 72 h. After incubation, bacterial cells were removed by centrifugation at 9,300 g for 10 min (TGL-16C centrifuge, Beijing Medical Centrifuge Inc., China). Sterile VY/2 medium was used to substitute the bacterial culture in the control.

Biotransformation of AFB₁ by cells, culture supernatant and intracellular cell extracts was conducted according to the procedure of Shu et al. [22]. Cells were pelleted using a centrifuge at 9,300 g, 4 °C for 10 min. The pellets were washed twice with a phosphate buffer (50 mM; pH 7.5) before being resuspended in the phosphate buffer (5 mL). Culture supernatant was obtained by centrifuging 50 h liquid culture at 9,300 g, 4 °C for 10 min. Intracellular cell extracts were harvested by the following procedures. The cell suspension was disintegrated twice (work every other 5 s for 33 min) by using ultrasonic cell disintegrator on ice. The disintegrated cell suspension was centrifuged at 9,300 g for 20 min at 4 °C. The cell extracts were collected by filtering the supernatant aseptically using 0.2 μm pore size sterile cellulose pyrogen free filters. The phosphate buffer solution (50 mM; pH 7.5) was used to substitute intracellular cell extracts and cells in the controls. The effect of protease treatment was determined by exposing the culture supernatant to 1 mg/mL proteinase K (Roche Diagnostics, Basel, Switzerland; specific activity ≥ 30 U/mg) for 1 h at 37 °C; 1 mg/mL proteinase K plus 1% SDS for 6 h at 37 °C. The effect of heat treatment was determined by dipping the culture supernatant into a boiling water bath for 10 min. 0.2 mL of AFB₁ stock solution was added to 0.8 mL of each reaction liquid for a final concentration of 100 ppb. The biotransformation tests were conducted in the dark at 30 °C without shaking for 72 h. After incubation, bacterial cells were removed by centrifugation at 9,300 g for 10 min (TGL-16C centrifuge, Beijing Medical Centrifuge Inc., China). The untreated culture supernatant was used as control.

Unless specifically indicated, all biotransformation experiments were conducted by using culture supernatant at 30 °C for 72 h with aeration. Culture supernatant was obtained by centrifuging 50 h culture with 9,300 g at 4 °C for 20 min.

To detect the appropriate nitrogen source for the bacterial growth with optimum AFB₁ biotransformation, 0.5% of tryptone, peptone, beef extract, yeast extract, fish peptone and ammonium nitrate (NH₄NO₃) were chosen as the nitrogen source in the culture medium.

Initial pH value in VY/2 medium was adjusted to 5.0, 5.5, 6.0 by using citrate acid buffer, and to 6.5, 7.0, 7.5, 8.0, 9.0 by sodium phosphate buffer.

Cultivation was performed at 15, 20, 25, 30, 35 and 40 °C respectively for 50 h before the culture supernatants were harvested.

The biotransformation products of AFB₁ after 72 h incubation with culture supernatants were determined by liquid chromatography mass spectrometry (LCMS) and infrared analysis (IR). For LCMS, the following samples were analyzed: (a) AFB₁ standard, (b) VY/2 medium supplemented with AFB₁, (c) AFB₁ treated with culture supernatants for 72 h. The samples were extracted with chloroform, dried under nitrogen, suspended in methanol: water (1:1, v/v) and analyzed by LCMS using a Phenomenex 2.0 × 150 mm C₁₈ column and methanol: acetonitrile: water (1:1:2, v/v/v) as solvent, at a flow rate of 100 μL/min. The HPLC system was equipped with a Finnigan Spectra System UV6000LP ultraviolet (UV) detector. Atmospheric pressure chemical ionization (APCI) positive ion mode was used for MS detection (ThermoFinnigan, San Jose, CA, USA).

For IR analysis, 2 mL chloroform was added to the incubation mixture and vortexed for 30 seconds. The phases were separated by low-speed centrifugation and the organic phase was evaporated to dryness at 20 °C under N₂. The samples were purified by HPLC before IR analysis. IR analysis was carried out with a KBr pellet using an IR spectrophotometer (NEXUS-470F TIR, Nicolet, Japan). Sterile VY/2 medium was used to substitute culture supernatant in the control.

The results were average of triplicates and were expressed as mean ± standard error. Data were analyzed as a completely randomized single factor design by ANOVA using the general linear models procedure in SAS. Significant F tests at the 0.05 levels of probability are reported. When a significant F-value was detected, Duncan’s Multiple Range Test was used to determine significant differences among means.