DECONTAMINATION OF AFLATOXINS WITH A FOCUS ON AFLATOXIN B1 BY PROBIOTIC BACTERIA AND YEASTS: A REVIEW

Food and feed contamination by aflatoxins represents a great challenge for human and animal health. Aflatoxins detoxification using probiotic bacteria and yeasts has been introduced as an inexpensive and promising method. This article is organized with an overview of the potential application of probiotic bacteria and yeasts to eliminate, inactivate or reduce the bioavailability of aflatoxins, especially aflatoxin B1, in vitro and in vivo. Also, a fast glance to beneficial health effects and preservative properties of probiotics followed by the mechanism of binding of aflatoxins by probiotics, influence of different probiotic pretreatments, and the stability of aflatoxin-probiotic complexes are mentioned.


INTRODUCTION
Contamination of food and feed by mycotoxins is a severe problem in all countries; hence, decontamination of mycotoxins from food and feed is essential. The food and agriculture organization evaluates that approximately 25% of global food and feed are contaminated with mycotoxins (Zoghi et al., 2017). Mycotoxins are secondary metabolites of mycelia or filamentous fungi associated to the Penicillium, Aspergillus (A.), and Fusarium genera. Production of mycotoxins may happen during the process of production, harvesting, storage or processing, under suitable temperature (between 24 and 37 °C) and humidity (above 13%) conditions (Massoud et al., 2018;Perczak et al., 2018). Several pre-and postharvest methods in order to decrease mycotoxins level in raw materials have been suggested; but, while mycotoxins levels have attained to contamination limited level in a product, it is difficult to eliminate the total toxin amount. Directly or indirectly exposure to mycotoxins may cause teratogenic, mutagenic, estrogenic, haemorrhagic, carcinogenic, immunotoxic, nephrotoxic, hepatotoxic, neurotoxic and immunosuppressive impacts on the health of animals and humans (Haskard et al., 2000;Zoghi et al., 2019). Aflatoxins are a group of the most repeatedly found mycotoxins in a variety of foods and feeds commodities causing economic losses in industry, veterinary care costs enhancement, and livestock production decline. These toxins are secondary metabolite products of some Aspergillus species, especially A. flavus, A. parasiticus and A. nomius. Several factors affect the production of this toxin including water activity, temperature, available nutrients, competitive growth of other microorganisms, and pH-value (Ghofrani Tabari et al., 2018). Various agricultural products may be contaminated by aflatoxins such as cereal grains especially rice, corn, maize, wheat, soya, rye, oats, barley, sorghum, nuts (almonds, peanuts, Pistachio, chestnuts, pumpkin seeds, etc.) and oily seeds such as cottonseed (Fochesato et al., 2018). Aflatoxins can enter the human body directly or indirectly by consuming contaminated products or derived foods, such as dairy products and meats from contaminated livestock, respectively. Exposure to aflatoxins leads to severe effects on human and animal health including chronic intoxications and liver and kidney cancers (Karazhiyan et al., 2016). Once aflatoxins are ingested by animals, they get adsorbed rapidly in the gastro intestinal tract (GIT), because they have low molecular weight, and then appear in blood and milk quickly after 15 minutes and 12 hours of post-feeding, respectively (Martins et al., 2001). 18 types of aflatoxins are identified through toxicological studies, but the major aflatoxins are aflatoxins B1 (AFB1), B2 (AFB2), G1 (AFG1) and G2 (AFG2). These names are related to their fluorescence under UV light (blue (B) or green (G)) and comparative chromatographic migration patterns through thin layer chromatography (TLC) (Lizárraga-Paulín et al., 2011;Rahnama Vosough et al., 2013). A. flavus usually produces the B group of aflatoxins, while A. parasiticus produces both B and G groups of aflatoxins through several biochemical processes. Among four mentioned aflatoxins, AFB1 is considered as the most common and dangerous one and exposure to AFB1 leads to both acute and chronic hepatocellular injury (Jakhar and Sadana, 2004). Aflatoxin M1 (AFM1) and aflatoxin M2 (AFM2) are metabolic derivates of AFB1 and AFB2, respectively (Lizárraga-Paulín et al., 2011). When feed containing AFB1 is ingested by livestock, it can be bio-transformed into AFM1 (4-hydroxy-AFB1) in the liver and excreted in milk, tissues, and urine of animals (Iha et al., 2013;Karazhiyan et al., 2016). AFM1 is resistant to all stages of dairy processing including pasteurization or sterilization (Prandini et al., 2009;Assaf et al., 2018). Approximately 0.3 to 6.2% of ingested AFB1 by livestock appears as AFM1 in milk. Diet type, amount of milk production, breed, health, and rate of digestion can affect the change rate of AFB1 to AFM1. A linear relationship between the AFM1 concentration in milk and AFB1 in contaminated feed is reported by Adibpour et al. (2016). The International Agency for Research on Cancer (IARC) categorized AFB1 and AFM1 as group 1 that leads to human cancer (IARC, 2016). However AFM1 is about ten times less toxigenic, mutagenic and genotoxic than AFB1, its carcinogenic effects have been demonstrated in several species (Elsanhoty et al., 2014). AFM1 is cytotoxic and can also cause DNA damage, gene mutation, chromosomal anomalies and cell transformation in mammalians cells. The Food and Drug Administration (2005) recommended that the maximum acceptable level of AFM1 in milk is 0.5 μg/kg, and the European Commission (2006) settled this limit to 0.05 μg/kg. Various strategies have been applied to remove aflatoxins from contaminated food and feed. Elimination of aflatoxins with chemical (addition of chlorinating, oxidizing or hydrolytic agents) and physical (UV light, heat, or ionizing radiation) approaches has some disadvantages, such as possible losses in nutritional value of treated commodities, insufficiency of toxin elimination, and requirement of expensive equipment (Zoghi et al., 2014). In addition, one of the most effective adsorbents for AFB1 is clay soil-based adsorbent. The layer structure of this type of adsorbents swells when it is placed in a liquid medium and it can adsorb AFB1 on its layers and prevent adsorption of AFB1 by cells in the GIT (Hadiani et al., 2018a). Nevertheless, this group of adsorbents is of low efficiency in adsorbing AFB1. Therefore, according to the researches, bioremoval method is an interesting alternative for inexpensive control or reducing of aflatoxins in foodstuffs without losses of nutritional quality or toxic compounds generation. Probiotics are the best candidate for aflatoxins detoxification due to their GRAS (Generally recognized as safe) status. Recently, several approaches to the removal of aflatoxins using probiotics are demonstrated. This article reviews the potential applications of probiotics in aflatoxin detoxification and the mechanism of aflatoxin binding by probiotics and the aflatoxin-probiotic complex stability are discussed.

PROBIOTICS AS AFLATOXIN BINDERS
Probiotics are described by FAO (2016) as 'viable microorganisms that, while ingested in sufficient amounts, exert health benefits on the host'. The main benefits for health include: lactose intolerance reduction, gut mucosal immunity support, a possible hypocholesterolemic effect, preventing the diarrheas or respiratory infections, colon cancer or inflammatory bowel disease inhibition, Helicobacter pylori or intestinal pathogens prevention, and antimutagenic and anticarcinogenic activities (Sanders et al., 2014;Yu, Chang and Lee, 2015). Recently, the use of microorganisms, especially probiotics, has been studied for their potential to aflatoxins elimination with an indirect health effect on the host (Bovo et al., 2012). Several probiotic strains have been investigated for their ability to bind aflatoxins (El-Nezami et al., 1998;Bueno et al., 2006;El Khoury et al., 2011;El-Nezami et al., 2002). Lactobacillus (L.) and Bifidobacterium (B.) species are the most known commonly probiotic bacteria, as well as the yeast Saccharomyces (S.) cerevisiae and Bacillus species and some strains of Escherichia (E.) coli. A functional classification of nontoxigenic, nonpathogenic, and fermentative probiotic bacteria are Lactic acid bacteria (LAB) witch are mainly related to the human gastrointestinal tract and widely used in food industry (Zoghi et al., 2017). LAB are Gram-positive, organotrophic, nonsporulating, fermentative rods or cocci, air and acid tolerant, which produce mostly lactic acid as the end-product of carbohydrate fermentation. All of them are anaerobic, but some of them can tolerate low levels of oxygen. Enterococcus, Lactococcus, Pediococcus, Oenococcus, Leuconostoc, Streptococcus, and Lactobacillus species are industrially important genera. The genus Bifidobacteria is also used as LAB, however they are phylogenetically unrelated and have unique sugar fermentation pathways. LAB are widely used in the world food production, vegetables, meat, and fermented dairy products. LAB play a significant role in improving the flavour, texture, and shelf-life of food products (Perczak et al., 2018). It is demonstrated that living and dead probiotics are able to decontaminate aflatoxins by attaching the toxin to their cell wall components. This phenomenon can be described as adsorption by components of the cell wall rather than by metabolism or covalent binding (Santos et al., 2006).

Inhibition of aflatoxin biosynthesis by LAB
A few authors also reported the antifungal properties of LAB. The main LAB recognized for this ability belong to Lactococcus and Lactobacillus (L.) genera. In contrast, it is reported that some LAB strains such as L. lactis can motivate aflatoxin accumulation. Coallier-Ascah and Idziak (1985) demonstrated a significant inhibition of aflatoxin accumulation by LAB and reported that this inhibition was not related to a pH decrease or a hydrogen peroxide production but rather to producing a heat stable and low molecular weight metabolite by LAB at the beginning of its growth phase. Gourama and Bullerman (1997) also reported that prevention of aflatoxin synthesis by Lactobacillus strain was due to specific bacterial metabolites. Several effective parameters related to antifungal properties of LAB have been investigated including growth medium, temperature, incubation time, pH, and nutritional factors. It was revealed that temperature and period of incubation were significantly affecting the amounts of antifungal metabolite production (Dalié et al., 2010). Gonzalez Pereyra et al. (2018) found that six Bacillus sp. strains were capable of decrease aflatoxigenic A. parasiticus growth rate significantly and could also decrease AFB1 concentration.

FACTORS AFFECTING AFLATOXIN BIOREMOVAL BY PROBIOTICS
Several criteria affect the aflatoxins removal using probiotics such as probiotic strain concentration and specificity, toxin concentration, pH, and incubation time.

Effect of probiotic strain specificity and concentration
In addition to bacterial strain specificity, the bacterial concentration can also affect the aflatoxin removal. Detoxification of aflatoxins by viable or nonviable probiotic cells is strain dependent (Topcu et al., 2010). In some studies, LAB were considered to be inappropriate binders of AFB1. This may be due to the specific LAB strains used in those studies (Shetty and Jespersen, 2006). Similarly, Peltonen et al. (2001) assayed 20 LAB strains and reported that the differences in AFB1 binding were because of different bacterial strain specificity. So, differences between aflatoxin ability of strains of LAB indicate that binding ability is highly strain dependent. El-Nezami et al. (1998) showed that L. rhamnosus strains GG and LC 705 can significantly remove AFB1 in comparison to other strains of LAB and the removal process was bacterial concentration dependent.

Effect of pH-value
Some investigation showed that binding process is not pH dependent exclusively. According to Zinedine et al. (2005) all the assayed Lactobacillus spp. removed AFB1 from 5% to 40% when pH increased from 3 to 5.5. Also, Pranoto et al. (2007) demonstrated that amount of bound AFB1 by LAB was higher at low pH (< 5) in compare with pH 6 and 7. In another study, Rayes (2013) stated that at pH 8.5 the highest decrease percentage of AFB1 by a pool of LAB occurred, while at pH 4.5 the lowest removal observed. On the other hand, the highest and lowest AFB1 removal was at pH 4.5 and 8.5, respectively, when the pool was included a S. cerevisiae strain. Hernandez-Mendoza et al.
(2009) investigated the binding of L. reuteri and L. casei with AFB1 at different pH (6, 7.2, and 8) and incubation time (4 and 12 h). They showed that the highest AFB1-binding capacity was at pH 7.2. Furthermore, Topcu et al. (2010) found that the binding of AFB1 by Enterococcus faecium was a pH and incubation time dependent process. In contrast, Bovo et al. (2014) showed no significant differences in the AFB1 reduction between L. rhamnosus strains conditions (spray, in solution or freeze-dried) at pH 3 and 6. So, it can be concluded that the pH dependence of AFB1 binding vary between bacterial strains. In addition, binding of AFB1 in a study, was not affected by pH, but binding of AFB2 considerably influenced by pH. It indicates that different metabolites of the same mycotoxin may show significant differences depend on binding mechanisms. reported that Enterococcus faecium M74 and EF031 strains at 1 h removed almost 65% of the total AFB1 removed during the whole incubation period (48 h). Bovo et al. (2012) stated that some probiotic strains bound AFM1 from skimmed milk in 15 min within a range from 13.51 to 37.75%. In another study, it was reported that the percentage of AFB1 removal was not significantly different between the 0 h and 72 h incubation period . In addition, El-Nezami et al. (1998) showed that the AFB1 removal was fast and no significant different was observed between different incubation times. Motawee and El-Ghany (2011) noted that the percentage of AFM1 and AFB1 reduction after 5 h by eight dairy strains of LAB in yoghurt was not considerably less than the whole of storage time. These results suggest that the binding of AFB1 by probiotics is a rapid process and the removal does not increase with the incubation time, considerably.

BINDING OF AFLATOXINS BY LAB
Specific strains of LAB are generally the most known probiotics for reducing aflatoxins. It has been reported that different strains of LAB have different effect on AFB1 removal in vitro. This removal is due to binding of bacterial cell wall to the aflatoxin, not bacterial metabolism. It was described that in vitro binding of AFB1 by LAB is a fast (less than 1 min), strain specific, and reversible process (Bueno et (2013) reported that L. plantarum was more efficient than L. lactis in removing AFB1 from liquid culture (46% and 27%, respectively), but when the two strains were combined, AFB1 removal reached 81%. Corassin et al. (2013) also revealed that a combination of LAB (L. rhamnosus, L. delbrueckii, and B. lactis) and S. cerevisiae could reduce AFM1 from UHT skim milk, completely. In contrast, El-Khoury et al. (2011) stated that L. bulgaricus, Streptococcus thermophilus and a mixture of these two bacterium reduced AFM1 content of milk to 58.5, 37.7 and 46.7%, respectively. It can be concluded that combination of specific probiotic strains may lead to a more aflatoxin removal efficiency than a single one, but may reduce their toxin removal capacity. Some authors have reported a mathematical model to illustrate the in vitro AFB1 binding to the LAB cell wall. A theoretical model has been suggested by Bueno et al. (2006). This model takes two possible processes into investigation: adsorption (binding) and desorption (release) of AFB1 to and from the binding site on the LAB surface. This model shows that AFB1 binds to a number of sites in LAB and allows us to evaluate the number of AFB1 binding sites and the efficacy of cells to reduce AFB1 from a liquid medium. So, this model demonstrates that the different abilities of probiotic strains to bind AFB1 are directly link to the number of binding sites of each probiotic.

BINDING OF AFLATOXINS BY YEASTS
Data found in the literature indicate that in addition to LAB, other organisms such as S. cerevisiae have the potential to bind aflatoxins. Yeast cells can bind to different molecules such as toxins as complexes on their cell wall surfaces (Baptista et al., 2004). Corassin et al. (2013) evaluated the AFM1 binding ability of L. rhamnosus, L. delbrueckii spp. bulgaricus, and B. lactis in combination with heat-treated S. cerevisiae. This mixture could bind with 100% of AFM1. In a study, S. cerevisiae and Candida krusei were tested for AFB1 binding and they could bind more than 60% (w/w) of the added mycotoxins in PBS. They emphasized the AFB1 binding was highly strain specific (Shetty and Jespersen, 2006). In another research, when dried yeast and yeast cell wall (include mannan-oligosaccharides) with AFB1 were added to rat-ration feed, a significant decrease in the toxicity was observed (Baptista et al., 2004).

MECHANISM OF AFLATOXIN BINDING BY PROBIOTICS
Several researchers studied the mechanism of binding of aflatoxins to probiotics. A review by Shetty and Jespersen (2006) stated that aflatoxin removal by probiotics is due to adhesion to cell wall components, because nonviable and viable probiotics are able to remove aflatoxins in vitro with similar efficiency. Possible binding sites include carbohydrates, proteins or a combination of both. It has been shown that two main components responsible for the binding of AFB1 by L. rhamnosus GG are cell wall polysaccharide and peptidoglycan. In addition, since LAB strain treatment with lipases did not lead to a significant increase in AFB1 binding, it was supposed that no fatty acids were involved in this adsorption (Lahtinen et al., 2004). Similarly, other authors have suggested that the peptidoglycan of LAB is the most likely site of aflatoxins binding (Haskard et al., 2000; Niderkorn et al., 2009). Yiannikouris et al. (2006) found that a cell wall component of many microorganisms named beta-d-glucans, played a key in the binding of aflatoxins. Recently, it was reported that the binding characteristics of a probiotic strain are possibly depend on the exopolysaccharides produced by the probiotics (Taheur et al., 2017). Also, Haskard et al. (2001) indicated superior involvement of hydrophobic interactions and main role of teichoic acids in aflatoxin binding mechanism. Similarly, Hernandez-Mendoza et al. (2009) showed that teichoic acids as well as peptidoglycans were important parts of the cell wall which could bind aflatoxin. Another report indicated the main role of teichoic acids in aflatoxin binding by probiotics. Teichoic acids may contribute mainly to hydrophobicity of wall contributed by anionic carbohydrates (Gratz et al., 2004). It can be concluded that binding of probiotics to aflatoxins is a function of fibril network of teichoic acids, peptidoglycans, and polysaccharides. Another mechanistic study conducted by Fochesato et al. (2018), which demonstrated that polysaccharides of L. rhamnosus attached aflatoxins. These polysaccharides are in three principal forms: peptidoglycan, cell wall polysaccharide, and teichoic or lipoteichoic acids. The environmental conditions such as pH-value or enzymes would be affecting the three-dimensional structure of the cell wall and the binding sites for aflatoxins. Therefore, it can be concluded that aflatoxin removal is due to the physical binding rather than metabolism, because peptidogylcan is one of the three principal carbohydrate forms of bacterial cell wall. When acid or heat treatments were used for LAB, it has been demonstrated that LAB ability to remove AFB1 increased. Also, inserting some basic compounds such as NaOH, Na2CO3, and isopropanol had negative influence on this binding (El-Nezami et al., 1998). Haskard et al. (2000) investigated the mechanism of binding of L. rhamnosus to aflatoxins. They used pronase E and periodate treatments (using periodate causes oxidation of cis OH groups to aldehydes and carbon acid groups) on viable, heat and acid-inactivated probiotic strains and suggested that binding was due to carbohydrate and protein components in cell wall, because a considerable decrease in AFB1 binding was observed. Heat and acid treatments cause protein denaturation and lead to the exposure of more hydrophobic surfaces. They also reported that AFB1 binding reduction by urea-treated LAB indicated the key role of hydrophobic interactions in binding. On the other hand, treatments with metal ions such as Na + and Ca 2+ showed that electrostatic interactions and hydrogen bonding played only minor role in AFB1 binding by LAB, because this process was not affected by mono and divalent ions or by changes in pH (2.5-8.5). Heterogeneous bacteria of LAB, posses a typical gram positive cell wall containing the peptidoglycan matrices, organic acids (teichoic and lipoteichoic acid), proteinacious surface (S) layer and neutral polysaccharides. These components play various functions including adhesion to macromolecules such as toxins (Perczak et al., 2018). Cell wall polysaccharides are produced by LAB with large variation between different strains (Zoghi et al., 2014). The peptidoglycan consists of polymerized disaccharide N-acetyl-glucosamine-beta (1-4)-N-acetyl muramic acid chains cross-linked by pentapeptide bridges. Disaccharide units of peptidoglycan have three different amendments, including acetyl groups of both N-acetyl-glucosamine and N-acetyl-muramic acid. Some LAB strains such as Enterococcus faecium, Pediococcus pentosaceus, L. plantarum, and L. casei have a diverging amino acid sequence of pentapeptide bridge where c-terminal d-alanine is replaced by d-lactate (Grohs et al., 2004). Teichoic acids are anionic polymers which bind to the peptidoglycan layer via a linkage unit and contribute more than 50% (w/w) of total weight of cell wall. The structure of the linkage unit is glycerol-phospho-N-acetyl mannosaminyl-beta (1-4)-glucosamine. Two types of teichoic acids which are detected from LAB, including poly glycerol phosphate and poly ribitol phosphate teichoic acids. Lipoteichoic acids are structurally similar to teichoic acids but they attach to the plasma membrane instead of peptidoglycan by a glycolipid anchor. The most frequently identified lipoteichoic acid in LAB is the poly glycerol phosphate lipoteichoic acid, which is almost similar to the structure of poly glycerol phosphate teichoic acid (Ambrosini et al., 1996). Some LAB strains such as Lactobacillus, Enterococcus, Sterptococcus, Bifidobacterium and Propionibacterium produce exopolysaccharides containing glucose, rhamnose, galactose, mannose, Nacetylgalactosamine, and N-acetylglucosamine (Grohs et al., 2004). Many LAB from the genus Lactobacillus are able to produce S-layer proteins. The size of these proteins is 25-50 kDa with calculated pI's ranging from 9.35 to 10.88, and they are highly basic. LAB which cannot produce S-layer proteins have a negative surface charge at neutral pH. Also, it has been reported that the surface charge of S-layer producing Lactobacillus are negative. This phenomenon may be due to the involvement of positively charged areas of S-layer proteins in their adhesion to peptidoglycan (Zoghi et al., 2014).

S. cereviciae cell wall
Except LAB, S. cereviciae is reported to the most used yeast as a probiotic strain in order to aflatoxins removal. S. cerevisiae cell wall represents about 30% (w/w) of total weight of the cell and made up of a network of back bone of β-1,3 glucan with β-1,6 glucan side chains, which is covalently linked to glycosylated mannoproteins. The cell wall mannoprotein includes a very heterogeneous class of glycoproteins. Carbohydrate fraction represents as much as 90% (w/w) of mannoproteins and oligosaccharide of mannan constitutes approximately 50% w/w of the total carbohydrates (Hadiani et al., 2018b). The core contains mannoproteins and branched mannose side chains as well as short and rigid rods like clusters of oligomannosyl chains extend out. Phosphodiester bridges in mannosyl side chains contribute negative charges on the cell surface. In addition, the cell wall of S. cereviciae is a highly dynamic structure which quickly replies to changes in the environmental stresses (Zoghi et al., 2014). Based on chemical combination and physical nature of cell wall of S. cerevisiae, it can be concluded that cell surface presents limitless sites on it in order to physical adsorption of aflatoxins. According to certain research, it is confirmed that mannan components of the cell wall play a main role in aflatoxin binding by S. cerevisiae (Shetty and Jespersen, 2006). The proteins and glucans provide accessible adsorption sites with ability to adsorb aflatoxins through various mechanisms such as hydrogen bonds and ionic or hydrophobic reactions. Heat treatment of S. cerevisiae increases permeability of the outer layer of cell wall, due to dissolution of cell-surface mannan and development of adsorption regions (Shetty et al., 2007). Haskard et al. (2001) revealed that using heat treatment for L. rhamnosus GG and LC 705 strains led to significant increase in AFB1 removal from contaminated defined medium and the stability of LAB-AFB1 complex. Similarly, Elsanhoty et al. (2014) found that heat treatment of L. rhamnosus can significantly enhance its binding to AFM1 in yoghurt. Reported literature indicates that heat treatment of LAB exhibit higher removal capacity, because of changes on the cell surface (Perczak et al., 2018). Other researchers showed that heat treated yeast reduce aflatoxins more than viable cells. Heating is responsible for protein denaturation or the formation of Maillard reaction products in the cell wall (Shetty et al., 2007; Rahaie et al., 2010). Several researchers showed that the acid treatment of yeast or LAB caused the highest adsorption of aflatoxin compared with viable and heattreated probiotic (Haskard et al., 2001; Rahaie et al., 2010; Hegazy et al., 2011). El-Nezami et al. (1998) reported that the binding ability of LAB increased by acid pretreatment. They also stated that acid treatment might break amine linkage in peptides and proteins, producing peptides and amino acids. Moreover, accessible aflatoxin binding sites increase and allow the aflatoxins to bind to the cell wall or its associated components (El-Nezami et al., 2002). According to Haskard et al. (2001) acid treatment may affect cell wall components such as peptidoglycan and polysaccharide by releasing monomers and further fragmentation into aldehydes after the glycosidic linkages break down. The acidic conditions could make AFB1 to be easily and repidly bound by constituents of cytoplasmic membrane (Bejaoui et al., 2004). Furthermore, Haskard et al. (2000) noted that hydrophobic interactions were expected in acid-treated LAB; because the protein denaturation may exhibit more hydrophobic binding areas to aflatoxins. In another study, significant increase in the ability of L. rhamnosus GG to bind AFB1 was observed after treatment with sodium dodecyl sulphate, whereas, treatment with urea showed no effect. One of the probable reasons could be the denaturation of protein by sodium dodecyl sulphate and cell wall isolation consist of peptidoglycan. The exposure of L. rhamnosus GG to divalent cations such as Ca 2+ and Mg 2+ or chelators such as EDTA and ethylene glycol tetra-acetic acid, as well as sonication and enzymatic treatments include different specific proteases, did not affect the binding of AFB1 may be due to the release of molecules bound to the surface of the bacteria (Lahtinen et al.,  2004). In fact, probiotic pretreatments which lead to protein denaturation, release of some components, and increase of pore size, probably act on the charge distribution change and hydrophobic nature of the bacterial surface and therefore enhance the efficiency of probiotics as adsorbent of aflatoxin (Karazhiyan et al., 2016; Ahlberg et al., 2015).

PROBIOTIC-AFLATOXIN COMPLEX STABILITY
Several researchers have reported the partial reversibility of the process of probiotics binding by probiotics (Peltonen et al., 2001; Hernandez-Mendoza et al., 2009); Haskard et al. (2001) studied the stability of 12 LAB-AFB1 complexes in both viable and nonviable forms (heat and acid treated LAB) after five washing steps with water. They exhibited that up to 71% of the total AFB1 remained bound and binding of aflatoxins to cell surface is significantly strong. In their investigation, viable cells of L. rhamnosus strains LGG and LC105 retained 38 and 50% (w/w) of the bound AFB1, respectively. Whereas, non-viable (acid and heat treated) cells retained the highest amount of AFB1 (66-71% (w/w)). Also, they revealed that autoclaving and sonication treated probiotic bacteria did not release any detectable AFB1. The authors concluded that the binding was reversible, but the stability of the complexes depended on strain, treatment and environmental conditions. Hernandez-Mendoza et al. (2009) reported that about 60-70% of AFB1 remained bound to the probiotic cells after washing by PBS; so, AFB1 attached to the bacteria by almost weak and partially reversible bound. Pizzolitto et al. (2012) stated that after five washings with PBS, different LAB cells retained AFB1 bound close to 50%, and the washing time (1-60 min) did not affect the release percentages. Among a panel of native LAB isolated from Iranian sourdough and dairy products, L. casei was reported to have the strongest binding of aflatoxin compared to other L. plantarum and L. fermentum strains (Fazeli et al., 2009). According to the findings obtained from the washing of AFB1-Enterococcus faecium complex, the binding of AFB1 to bacterial cell surface was a reversible process and the stability of the complexes was strain specific (Topcu et al., 2010). Similarly, it was noted that after washing the AFB1-Lactobacillus complexes, variable amounts of AFB1 were released back into the solution (Peltonen et al., 2001). Also, the stability of AFB1-Enterococcus faecium strains (MF4 and GJ40) complexes was found to be high after three washes with PBS (Fernandez Juri et al., 2014). In addition, a stable AFM1-LAB (L. rhamnosus and L. plantarum) complex was showed by Elsanhoty et al. (2014). Moreover, Bevilacqua et al. (2014) described the proportionality of the amount of aflatoxin released into the medium by the number of treatments performed. According to the above discussion, it is clear that any in vitro results must be supported by in vivo experiments, because aflatoxins may be released by the continual washing of the bacterial surface in the GIT and negative health implications may be observed. Thus, several studies have attempted to evaluate the stability of the aflatoxin-probiotic complexes in the GIT conditions. It is revealed that defined LAB that show significant adhesion to intestinal cells lose this property when they bind to aflatoxins. Therefore, in the gastrointestinal tract, the bacteriaaflatoxin complex is rapidly excreted (Gratz et al., 2004).

IN VIVO STUDIES
Many recent studies revealed that AFB1 intake can change the morphological and immune function of the intestinal mucosa due to decreasing the percentage of T-cell subsets and the expression level of cytokine mRNA in the small intestine. The mechanism of intestinal tissue poisoning of the host by AFB1 includes the prevention of oxygen production and inhibition of the free radicals of oxygen (Jiang et al., 2015). Intestinal cells can absorb aflatoxins at high rates (>80%), regardless of the species (Grenier and Applegate, 2013;Wan et al., 2016). Some experimental evidences reported that probiotics could bind aflatoxins within the lumen, so, reducing the negative impacts of aflatoxins and improving gut and liver health (Niderkorn et  showed that L. reuteri and L. casei were able to considerably reduce malondialdehyde concentration in the kidney and liver. As aflatoxin toxicity is mainly related to the liver, using probiotics could improve the histological picture and architecture of the liver and serum biochemical parameters. Besides in vitro studies, the AFB1 binding ability of probiotics was evaluated ex vivo in the intestinal lumen of chicken using the chicken duodenum loop technique (El-Nezami et al., 2000). The authors stated that L. rhamnosus GG, L. rhamnosus LC705, and Propionibacterium freudenreichii removed 54, 44, and 36% of the AFB1, respectively from the soluble fraction of the luminal fluid within 1 min. It can be concluded from these findings that AFB1 binding by LAB appears in physiological conditions in animals, which may represents a way to reduce AFB1 bioavailability in the organism. El Nezami et al. (2006) continued their research in Egypt and investigated the effect of a combination of L. rhamnosus LC705 and Propionibacterium freudenreichii on AFB1 levels in human feces samples from 20 healthy volunteers. The mentioned probiotic strains were administered two times per day (at a dosage of 2-5×10 10 CFU/day) for five weeks by volunteers and the control group received a placebo. The marker for biologically effective dose of AFB1 was the adduct AFB1-N7-guanine. High level of this adduct in the urinary excretion is associated to a high risk of liver cancer (Vinderola and Ritieni, 2015). The fecal samples were positive for AFB1 with a range from 1.8 to 6 μg AFB1/kg feces for 11 volunteers. A significant reduction in urinary excretion of AFB1-N7-guanine and fecal aflatoxin levels was observed for volunteers after receiving the probiotic mixture compared to volunteers receiving a placebo. Kankaanpaa et al. (2000) showed that aflatoxin binding by L. rhamnosus LGG and LC105 considerably reduced adhesion properties of the probiotic strains and facilitates excretion of immobilized AFB1. Similarly, Gratz et al. (2004) reported that pre-exposure of L. rhamnosus GG to AFB1 decreases its binding with intestinal mucus and leads to faster removal. Also, it was shown that addition of S. cerevisiae to the animal diet reduced aflatoxin toxicities; thus, possible stability of the yeast-afllatoxin complex was indicated through the GIT (Shetty and Jespersen, 2006;Armando et al., 2012). Similar results reported by Gratz et al. (2006) who found that L. rhamnosus GG was able to modulate AFB1 uptake in rats, increased fecal AFB1 excretion in rats and reduced liver injury. As demonstrated, L. casei Shirota can decrease AFB1 absorption in the GIT even after a long period of toxin exposure (Hernandez-Mendoza et al., 2010). Nikbakht Nasrabadi et al. (2013) also found that L. casei Shirota could reduce the blood serum level of AFB1 in rats and improved the adverse effect of AFB1 on rats' body weight and plasma biochemical parameters. This result is consistent with Hernandez-Mendoza et al. (2009) who stated that L. reuteri was able to bind to AFB1 in all intestinal sections under normal conditions of the GIT. On the other hand, another study revealed that the probiotic mixture could only retard the AFB1 absorption in duodenal loops and considerably decrease the AFB1 adsorption in the intestinal mucus (Gratz et al., 2005). Fochesato et al. (2018) reported that dynamics of AFB1 adsorption and desorption by L. rhamnosus RC007 were strongly affected by the salivary environment. The knowledge of the adsorption dynamics of AFB1 with a probiotic strain will allow predicting its behavior at each stage of the GIT.

CONCLUSION
Aflatoxins frequently contaminate the food and feed at various levels. So, for the food industry, it has always been an uphill task to control the aflatoxins level in the products. It is suggested that probiotic strains with high aflatoxin binding abilities can be used in food industries as additives in small quantities without compromising the characteristics of the final product and thus can avoid accumulation of this toxic compound and decrease its toxic effects. Many studies have demonstrated varying efficiency of some selected probiotics in removing aflatoxins. Tables 1 and 2 demonstrate several kinds of probiotics applied for decontamination of food and feed from AFB1 and AFM1, respectively.