Mycotoxins are secondary metabolites of fungi, which are common pollutants of global food and feed chains. Change in global climate has increased the size of the areas suitable for fungal growth and cause considerable economic loss by infection and mycotoxins produced [1
]. The major mycotoxin-producing fungi genera are Aspergillus
, and Fusarium
, which produce most mycotoxins with different toxic effects such as hepatotoxicity, reproductive toxicity, nephrotoxicity, carcinogenicity, and immunotoxicity [3
]. Mycotoxins that mean considerable potential risk to human and animal health are aflatoxins, trichothecenes, zearalenone, ochratoxin A, fumonisins, and ergot alkaloids [3
and A. parasiticus
fungi species produce aflatoxin B1
(AFB1) and mainly infect peanuts, corn, wheat, rice, cottonseed, copra, nuts, and various foods. AFB1 has a carcinogenic effect and causes various diseases in animals and humans [3
]. Acute toxicological effects are liver damage, decreased egg- and milk production, feed refusal, decreased growth, anemia, jaundice, weight loss, anorexia, hemorrhage, embryotoxicity, and carcinoma in rainbow trout, pig, cattle, poultry, duckling, turkeys, and chickens [8
]. Chronic effects are hepatocellular carcinoma, lower reproductivity, decreased cellulose digestion, volatile fatty acid formation, and proteolysis in cattle, swine, and rainbow trout [9
]. The effects on humans include liver cancer, chronic hepatitis C infection, and Reye’s syndrome (encephalopathy and visceral degeneration in children). Incidence of these diseases have been recorded in India (1974) and Kenya (2004 and 2005) [3
and F. culmorum
fungi species produce zearalenone (ZEN) and mainly infect corn, hay, and pelleted commercial feed. ZEN has an estrogenic effect in humans and animals [3
]. In general, toxicological effects are swollen vulva, mammary gland enlargement, hyperestrogenism, feminization in young male animals (testicular atrophy, swollen prepuce), decreased libido, decreased spermatogenesis, infertility, and embryonic death in swine, cattle, poultry, and laboratory rodents [3
]. The effects on humans are premature puberty, premature thelarche, prepubertal breast enlargement in boys, pseudopuberty in girls, and involves cervical cancer and lasting effects on the endocrine system [3
]. These effects have been studied in Puerto Rico and Hungary [3
and F. sporotrichioides
fungi species produce T-2 and mainly infect corn, wheat, commercial feed, and mixed feed. Acute toxicological effects are dermatitis, feed refusal, vomiting, testis and ovary aberrations, hemorrhages and necrosis of stomach, depressed growth, early life stage toxicity in cats, dairy cattle, dogs, pigs, ducklings, zebrafish, and rainbow trout [3
]. Chronic exposure causes dizziness, excessive salivation, fatigue, secondary infections (pneumonia) and abdominal pain in chickens, mice, rats, and rhesus monkeys [25
]. The effects on humans could be connected to alimentary toxic aleukia (ATA; sepsis, agranulocytosis, atrophy of the bone marrow, mortality), however, it is not clear whether T-2 causes the disease alone or with other mycotoxins [3
]. Incidences of disease have been described in the USSR (Union of Soviet Socialist Republics; USSR) (1941–1947), China (1984–1985), and India (1987) [22
Mycotoxins are produced by different fungus species and some fungi are able to produce different toxins [3
]. The natural co-occurrence of mycotoxins increases the risk of exposure to several mycotoxins at the same time in humans and animals [32
]. The toxicity of mycotoxin mixtures is not always possible to predict based on their individual effects as interactions among mycotoxins could be additive, antagonistic, or synergistic [6
]. Earlier worldwide examination of mycotoxin levels in food and feed indicate that more than 70% [6
] of samples were contaminated with at least one mycotoxin [6
]; another global measurement showed that 48% [6
] of samples were contaminated with at least two mycotoxins. In a previous study, the mycotoxin content of AFB1, ZEN, and T-2 were investigated in compound animal feed and more than 57% of samples were contaminated with three types of mycotoxins [7
Increased health risk due to co-contamination of mycotoxins confirms the elimination of mycotoxins from the food and feed chain [37
]. Several processes of removal and detoxification of mycotoxins have been investigated, however, most of these are ineffective, decrease nutritional values, or produce toxic derivates [31
]. Biological transformation may be an ideal approach to decrease mycotoxins. Several previous studies have described bacterial degradation and detoxification of AFB1 [44
], ZEN [46
], and T-2 [37
] with different strains, however, limited data are available about the degradation and detoxification of mycotoxin mixtures [49
]. In an earlier study, Ery4 laccase from Pleurotus eryngii
was used for the degradation of AFB1+ZEN and fumonisin B1
(FB1)+T-2 mixtures. AFB1+ZEN were degraded by 86% and 100%, while FB1+T-2 by 25% and 100% [50
]. Rumen fluid was able to degrade 90% of ZEN and 100% of T-2 toxin, but had no effect on AFB1 [51
]. Members of the genus Rhodococcus
were used for degradation of AFB1+ZEN+T-2 mixtures, the R. pyridinivorans
K408 strain was able to degrade AFB1+ZEN mixture by 99% and 96%, the R. rhodochrous
NI2 strain degraded AFB1+T-2 by 99% and 97%, and the R. erythropolis
NI1 strain degraded AFB1+ZEN+T-2 by 99% and 98% and 96% [52
]. Degradation does not mean detoxification in every case, because mycotoxins can transform into more toxic metabolites (such as AFB1—AFB1-8,9-epoxide; ZEN—α-zearalenol; T-2—3-hydroxy-T-2) and bacterial metabolites can also be toxic (such as bacterial metabolites of the Rhodococcus rhodochrous
NI2 strain) [49
]. European Food Safety Authority (EFSA) guidelines suggest that the toxicity of degradation metabolites needs to be examined with in vivo
toxicological methods [54
Previously, the Csenki-Garai three-step method ((1) determination of mycotoxin toxicity baseline, (2) examination of bacterial metabolites toxicity, and (3) identification of degradation products toxicity) was developed, which is a microinjection-based technique in a zebrafish model system, for qualification of degradation and detoxification efficiency of bacteria, and suitable for indirect testing of toxin metabolites [53
]. During the development of the method, it was demonstrated that the injection volumes alone do not cause mortality or other malformations in the treated embryos. In the case of a well optimized method, injection volume variations can be kept within ±20%, according to the OECD 236 (Organisation for Economic Cooperation and Development; OECD) test guideline’s recommendations and result reliability can be ensured [56
]. Csenki et al. described Cupriavidus basilensis
ŐR16 bacteria strain ochratoxin-A (OTA) degradation efficiency, which was used to developed a suitable microinjection test. The following results were used as the basis for the test: the ŐR16 strain could degrade almost 100% of OTA, the OTA major degradation metabolite was OTα, and neither of the samples (bacteria metabolites and degradation products) had any effect in the mice test, thus confirming that the zebrafish embryo—thanks to their sensitivity—proved to be a good model for this type of study. The results also showed that the effects observed in the treatments were derived only from the toxin and the normal and degradation metabolites of the microbe [56
]. With the help of the Csenki-Garai three-step method, seven different bacterial strains and T-2 toxin-degrading properties were examined for classifying the strains [53
]. The results confirmed that this microinjection technique may provide an opportunity for the selection of microbial strains that are able to degrade toxins and the identification of the most effective and environmentally safe microbes from the selected strains.
In this study, we investigated whether the Csenki-Garai three-step method is appropriate to evaluate the multimycotoxin-degrading efficiency of the Rhodococcus erythropolis NI1 strain. The objective of this experiment was to examine the toxic effects of AFB1, ZEN, and T-2 individually or in combination as well as that of their degradation products by the NI1 strain on zebrafish embryos. In addition, this study explored the interactions among these mycotoxins.
The world mycotoxin survey showed that 68% of tested feed materials contained more than one mycotoxin and natural co-occurrence of mycotoxins increased health risk [107
]. Co-contamination of mycotoxins confirms the elimination of mycotoxins from the food and feed chain [37
]. Removal of mycotoxins with biological transformation may be an ideal approach, and bacterial degradation and detoxification of individual toxins have been described in several previous studies [44
]. Limited data are available on the biodegradation of multiple mycotoxins, however, testing of these bacteria is important due to their increasing co-occurrence [49
AFB1 has two main detoxification pathways: modification of difuran ring or coumarin structure. At first, the AFB1-8,9-epoxide formed, then hydrolysis resulted in dihydrodiol-derivatives. Second, the lactone ring can be changed in the coumarin moiety [49
]. ZEN has two main detoxification mechanisms: both cleave a ring structure. First, hydrolysis of the ester bond in the lactone ring, followed by a spontaneous decarboxylation. Second, cleavage at the C6-ketone group resulted in lactone intermediate and subsequent activity by unspecified a/b-hydrolase, without decarboxylation [49
]. The T-2 toxin detoxification pathway is de-acylation into HT-2, then T-2 triol, which is followed by de-epoxidation or de-acylation into T-2 tetraol, then de-epoxidation into de-epoxy T-2 tetraol [49
]. These three mycotoxins have a few similarities: the lactone ring (aromatic ester) is the main cause of the toxicity of the AFB1 and ZEN, and carboxyl ester groups in the T-2 toxin also play an important role in the toxicity [49
]. Hypothetically, this chemical structure (ester) analogue could be the common point in the biodegradation pathways at the same strain. Multiple strains of Rhodococcus
cells can degrade both the AFB1 and T-2 toxin, and can utilize ZEN, and none can convert OTA and FB1. Therefore, these bacteria could possess active multitarget enzyme(s) detoxifying mycotoxins [108
The detoxification efficiency of microbes is usually tested by biotests that measure only one specific effect, therefore, not suitable for testing various mycotoxin effects. Effects of AFB1 degradation products are mostly measured with SOS-Chromotest, which is a genotoxicity test. According to the earlier study by our institute, the AFB1 degradation products by NI1 are not genotoxic, and enzymes (or enzyme groups) that are responsible for the biodegradation of AFB1 are constitutive intracellular [109
]. Specific enzymes of NI1 bacteria are not available as the genus of Rhodococcus
has one hundred different aromatic ring proteases, which may be responsible for AFB1 degradation as aromatic ring degrading enzymes [109
]. The effects of ZEN degradation products are commonly investigated with the BLYES test, which is an estrogenicity test. An earlier study described that ZEN degradation products by NI1 had no estrogenic effects, however, during the deeper enzymatic investigation, constitutive and indicated intracellular enzymes were not able to degrade ZEN (6-h experiment) [110
]. Degrading enzymes of T-2 mycotoxins are not available. A biotest was not available to examine the effects of T-2 degradation products until the Csenki–Garai three-step method, which is an in vivo test on zebrafish embryos [53
]. The advantages of this method are that mycotoxins and both types of metabolites (degradation products and bacterial metabolites) can be tested in a complex, synergistic and antagonistic, can be also detected, and the more efficient and safe strains can be selected.
Toxicological effects of AFB1, ZEN, T-2 in individual and combination and their degradation products were examined on zebrafish embryos in this study. Results showed that individual exposure of AFB1, ZEN, and T-2 mycotoxins significantly increased the mortality and caused different phenotypic deformities in zebrafish embryos. Based on the results of earlier published and recent studies, mycotoxins and mixtures caused the same symptoms following classical exposure and microinjection. Therefore, the microinjection-based Csenki–Garai three-step method ((1) determination of mycotoxin toxicity baseline, (2) examination of bacterial metabolites toxicity, and (3) identification of degradation products toxicity) can be used to study the effect of toxins and mixtures as the results were highly comparable with the results of classical methods. The outcomes showed that different combined exposures of AFB1, ZEN, and T-2 increased the mortality rate and caused different malformations in zebrafish embryos. The combined exposure of mycotoxins was synergistically toxic, except for ZEN+T-2 and AFB1+ZEN+T-2, which had a very strong antagonistic effect. Zhou et al. described the effects of AFB1+ZEN mixtures [62
], in addition to these, we found that the seriousness of lens and head distortion and pericardial edema increased with injected volume. Toxicological effects of AFB1+T-2 on zebrafish embryos, ZEN+T-2 in vivo
, AFB1+ZEN+T-2 mixtures in vivo
and in vitro had not been reported in previous studies.
Results showed that the Rhodococcus erythropolis NI1 strain was able to degrade mycotoxins and their mixtures to different ratios (85–100%), and mycotoxins in combination were reduced to a higher degree than single ones. The NI strain reduced the toxic effects of mycotoxins and mixtures on mortality, except for the AFB1+T-2 mixtures. Degradation products of the eAFB1+T-2 mixture by the NI1 strain were more toxic than AFB1+T-2, while the analytical results showed very high degradation, which means that the NI1 strain degraded this mixture to toxic degradation products.
The Csenki–Garai three-step method is an appropriate tool to evaluate the multimycotoxin-degrading efficiency of the Rhodococcus erythropolis NI1 strain and can be used for other microbial strains with similar characteristics. Limited data are available regarding the degradation and detoxification of mycotoxin mixtures, and the resulting degradation products and degradation enzymes, therefore, these should be identified in the future. The microinjection method can also be helpful in these studies by indirectly examining the effects of metabolites present in microbe degradation products.
4. Materials and Methods
4.1. Animal Protection
The Animal Protocol (2013) was approved under the Hungarian Government Regulation on animal experiments (42/2013. (II.4.)) and all studies were completed before the treated individuals reached free-feeding stage.
4.2. Mycotoxin and Mixture Degradation Experiments
The Rhodococcus erythropolis NI1 strain (stored at −80 °C) was maintained on Luria-Bertani (LB) agar plates (10 g tryptone, 5 g yeast extract, 9 g sodium-chloride and 18 g bacteriological agar (Biolab Ltd., Budapest, Hungary) in 1 L (pH 7.0) ion-exchanged water) and incubated at 28 °C for 72 h. Then, a single colony of the strain was inoculated into 50 mL 100% LB medium (10 g tryptone, 5 g yeast extract, and 9 g sodium-chloride in 1 L (pH 7.0) ion-exchanged water) in 250 mL flasks and cultures were grown for 120 h at 28 °C, 170 rpm in a shaking incubator (Sartorius Certomat BS-1, Germany). Liquid cultures were centrifuged at 3220× g, 4 °C for 20 min (Eppendorf 5810R, Germany), the pellets were resuspended in 50 mL 20% sterile LB medium (100% LB medium diluted with ion-exchanged water), and then were centrifuged again in the same conditions (repeated twice). After resuspension, the optical density of the cultures was measured at 600 nm (OD600) (GENESIS 10S UV-VIS, Thermo Fischer Scientific) and adjusted to 0.6 ± 0.05 to prepare bacterial inoculum. Five mL of the bacterial suspensions were inoculated into 45 mL of sterile 20% LB medium to test the effects of bacterial metabolites. Similar inocula were prepared in parallel, which contained AFB1, ZEN, T-2 (1 mg/L final concentration (Fermentek Ltd., Israel)), and mycotoxin mixtures (1 mg/L final concentration per toxin). The microbe-free control was uninoculated 20% LB medium contaminated with AFB1, ZEN, T-2 (1 mg/L final concentration), and mycotoxin mixtures (1 mg/L final concentration per toxin). Experiments were incubated on a laboratory shaker at 28 °C, 170 rpm for 168 h in triplicate. Cultures were centrifuged at 3220× g, 4 °C, for 20 min. Supernatants for microinjection (1 mL) were filtered with 0.2-µm syringe filters (VWR International Ltd., Hungary) to gain bacteriologically sterile samples and stored at −20 °C. Pellets and supernatants samples were stored separately at −20 °C until analytical measurements.
4.3. Measurement of Mycotoxin Concentrations
UHPLC-MS/MS (ultra-high-performance liquid chromatography with a tandem mass spectrometer) was applied for the measurement of AFB1, T-2, and ZEN concentrations. First, pellets were extracted with an acetonitrile/water/formic acid (79/20/1, v/v%) mixture, then an aliquot of 500 µL extracts was taken into 1.5-mL dark vials. Supernatants in LB medium were taken directly and an aliquot of 500 µL was put into 1.5 mL dark vials. Afterward, both sample types (LB broth and pellet) were evaporated until dryness under a gentle N2 stream. The residues were reconstituted in 50:50 v/v% A:B mobile phases (A: water, 5 mM ammonium formate, 0.1% formic acid; B: methanol, 5 mM ammonium formate, 0.1% formic acid) and were filtered through a 0.22 µm PTFE (polytetrafluoroethylene) filter. An Agilent 1290 Infinity II UHPLC system (Agilent Technologies, USA) equipped with an Agilent Zorbax Eclipse Plus chromatographic column (2.1 × 50 mm, 1.8 μm) was used. Five μL prepared samples were injected into the mobile phase, which initially contained 95% A and 5% B eluents. Four hundred μL/min flow rate and 40 °C column temperature was set. A triple-quadruple mass spectrometer (Ultivo, Agilent Technologies, Santa Clara, CA, USA) with an ESI (electrospray) ion source was used for the determination of mycotoxin concentrations of the samples. The mass spectrometer was operated in MRM (multiple reaction monitoring) scan mode and monitored two transitions (1 qualifier, 1 quantifier) of mycotoxin precursor ions in positive ion mode. The applied analytical method was validated for LB medium. The correlation coefficient (R2) of the matrix-matched calibration was >0.9936, the recovery from LB medium spiked with the T-2 standard was 78 ± 13%, AFB1 was 114 ± 19.2% and ZEN 79 ± 4.3%, LOD (limit of detection) for T-2 was 3 μg/L, AFB1 was 0.5 μg/L, and ZEN was 0.2 μg/L. The LOQ (limit of quantification) value for T-2 was11 μg/L, AFB1 was 2 μg/L, and ZEN was 1 μg/L.
4.4. Zebrafish Maintenance and Egg Collection
Wild type laboratory-bred AB strain zebrafish were held in breeding groups of 30 females and 30 males at the Department of Aquaculture, Szent István University, Hungary, in a Tecniplast ZebTEC recirculation system (Tecniplast S.p.a., Buguggiate, Italy) at 25.5 °C ± 0.5 °C, pH 7.0 ± 0.2, conductivity 550 ± 50 µS (system water), and a light:dark period of 14 h:10 h. Fish were fed twice a day with dry granulate food (Zebrafeed 400–600 µm, Sparos Lda., Olhão, Portugal) supplemented with freshly hatched live Artemia salina once a day. Fish were placed in breeding tanks (Tecniplast S.p.a.) late in the afternoon the day before the experiment and allowed to spawn by removing the dividing walls the next morning. Spawning of individual pairs was delayed through time to allow a continuous supply of one-cell embryos.
Microinjection of zebrafish embryos (microinjector, capillary puller, and parameters of capillary) was conducted as described by Csenki et al. [56
]. Briefly, one-cell embryos were injected with different volumes: sphere diameter of 75 µm corresponded to an injection volume of 0.22 nL, 100 µm to 0.52 nL, 150 µm to 1.77 nL, and 200 µm to 4.17 nL. These injected volumes were selected so that the mortality values of toxins and mixtures were interpretable above and below the baseline in every dose [54
]. These doses were used for each test solution (1 mg/L AFB1, ZEN, T-2, mycotoxins mixtures, bacterial metabolites, and degradation products of toxins and mixtures). After 2 h, coagulated and/or non-fertilized eggs were removed and well-divided eggs were transferred in groups of twenty into 6-cm diameter Petri dishes. Each treatment group contained 20 eggs in three replicates. Embryos were then incubated (Sanyo MIR-154) in system water at 26 °C ± 1 °C and a 14 h light and 10 h dark period and checked for lethal and sublethal effects under a microscope. System water was replaced every 24 h until 120 hpf. Digital images of larvae (120 hpf) in lateral orientation were taken under a stereomicroscope at 30× magnification (Leica M205 FA, Leica DFC 7000T camera, Leica Application Suite X, Leica Microsystems GmbH, Germany).
4.6. Toxicological Endpoints
Mortality values of injected embryos were determined at 120 hpf on the basis of egg coagulation, the lack of somite formation, and the lack of heart function. Sublethal effects were examined at 120 hpf, the endpoints were hook-like tail, tail deformed, pericardial- and yolk edema, lens- and head distortion, and lack of swim bladder. The frequency of deformities was compared to the number of live embryos at 120 h.
Results were analyzed and graphs were plotted by GraphPad Prism 6.01 for Mac (GraphPad Software, San Diego, CA, USA). Data were checked for normality with the Shapiro–Wilk normality test. Significant differences were verified by Kruskal–Wallis analysis with Dunn’s multiple comparisons test and the Mann-Whitney test. Lethal and sublethal results were compared to the non-injected control (non-inj-c); mortality values of initial toxins were compared to the NI1 degradation products. CompuSyn software (Paramus, NJ, USA) was applied for the determination of interactions between mycotoxins and lethal dose (LD50
) values [111