**1. Introduction**

Aflatoxins are fungal secondary metabolites toxic to humans and animals, causing carcinogenic, mutagenic, teratogenic, and immunosuppressive effects [1]. Aflatoxins are produced by toxigenic *Aspergillus flavus*, *A. parasiticus*, and *A. nomius* isolates growing in a variety of food and feed commodities [2]. These metabolites are very stable to autoclaving, pasteurization, and other food processing procedures [3].

Aflatoxin M1 (AFM1) is a 4-hydroxy derivative of aflatoxin B1 (AFB1), which, although approximately ten-fold less toxigenic than aflatoxin B1, exerts cytotoxic, genotoxic, and carcinogenic effects in a variety of species [2], being classified as belonging to group 1 (i.e., carcinogenic to humans) by the International Agency for Cancer Research [4]. AFM1 is

**Citation:** Cruz, P.O.d.; Matos, C.J.D.; Nascimento, Y.M.; Tavares, J.F.; Souza, E.L.d.; Magalhães, H.I.F. Efficacy of Potentially Probiotic Fruit-Derived *Lactobacillus fermentum*, *L. paracasei* and *L. plantarum* to Remove Aflatoxin M1 In Vitro. *Toxins* **2020**, *13*, 4. https:// dx.doi.org/10.3390/toxins13010004

Received: 23 September 2020 Accepted: 18 November 2020 Published: 23 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/).

formed in the liver and excreted through the milk of lactating animals that have consumed feed contaminated with AFB1. Approximately 0.3–6.2% of AFB1 ingested by livestock is converted to AFM1 in milk [5]. In Brazil and the USA, the maximum allowable limit of AFM1 in raw milk is 0.5 μg/L [6,7]. The European Union has set a maximum limit of AFM1 of 0.05 μg/L for raw milk, heat-treated milk, and milk used in dairy products formulation [8].

Control of aflatoxin in food and feed can be primarily achieved by a prevention of mold contamination and growth with the adoption of improved agricultural practices and control of storage conditions, as well as by the detoxification of contaminated products through chemical (e.g., ammonia, hydrogen peroxide, alkalis, and acids) or physical methods (e.g., heat, radiations, ultraviolet, and microwave) [9]. Some methods used for aflatoxins decontamination, although they have been shown to be effective to a certain extent, may have some drawbacks, such as negative impacts on nutritional and sensory characteristics of foods, production of potentially toxic by-products, or non-suitability for use in solid foods [2,9].

Use of lactic acid bacteria (LAB) has been considered a safe and environmentally friendly biological method for the detoxification of aflatoxins in foods and feeds [10,11]. Studies have found a variable capability among probiotic *Lactobacillus* species or isolates to bind aflatoxins [12–14]. These studies have mostly used commercial *Lactobacillus* cultures or isolates from dairy origin. Although a number of *Lactobacillus* isolates recovered from fruit, vegetables, or their processing by-products have shown good performance in in vitro tests for the selection of probiotics [15–17], none of these isolates have been examined for their capacity to remove aflatoxins. The use of select probiotic *Lactobacillus* isolates has been considered a promising biological tool for removing aflatoxins from foods through adsorption when compared to chemical and physical treatments. Furthermore, although still the fastest method for retaining high detoxification efficacy [18,19], many chemical agents are nonedible materials and need to be eliminated after aflatoxin decontamination [20,21], while *Lactobacillus* species have been usually considered safe for use in foods [16,17].

Considering the available evidence, it was expected that fruit-derived *L. fermentum*, *L. paracasei*, and *L. plantarum* isolates with aptitudes to be used as probiotics would be able to remove AFM1 in a prospective view for application in food and feed detoxification. To test this hypothesis, this study evaluated the efficacy of these isolates as viable and non-viable (heat-killed) cells, in the removal of AFM1 in vitro, as well as the recovery of the AFM1 bound to bacterial cells.

#### **2. Results and Discussion**

Chromatograms for the quantification of AFM1 in positive control, negative control, as well as in samples with viable cells of *L. paracasei* 108, *L. plantarum* 49, and *L. fermentum* 111 are shown in Figure 1. Chromatograms for the quantification of AFM1 in assays evaluating the recovery of AFM1 from cells after 1 h of incubation are shown in Figure 2.

Results of the capability of viable and heat-killed (non-viable) cells of *L. paracasei* 108, *L. plantarum* 49, and *L. fermentum* 111 for removing AFM1 in PBS are presented in Table 1. Viable and heat-killed cells of all examined *Lactobacillus* isolates were able to remove AFM1 in PBS, with removal percentage values in the range of 73.0 ± 1.2–80.0 ± 1.7% and 72.9 ± 1.1–78.7 ± 1.2%, respectively. Viable and heat-killed cells of the three examined isolates had similar values (*p* > 0.05) of AFM1 removal. Only *L. paracasei* 108 had higher values (*p* ≤ 0.05) of AFM1 removal after 24 h for both viable and heat-killed cells compared to 1 h. Higher values of AFM1 removal (*p* ≤ 0.05) after 1 h were found for *L. plantarum* 49 and *L. fermentum* 111, but the three examined isolates had similar values of AFM1 removal (*p* > 0.05) after 24 h.

**Figure 1.** Chromatograms of aflatoxin M1 (AFM1) quantification in positive and negative control. (**I**) Positive control: phosphate buffer solution (PBS) with AFM1. Rt = Retention time of AFM1 in phosphate buffer solution; chromatographic peak area corresponding to AFM1; (**II**) Negative control after 1 h of incubation: PBS + *L. paracasei* 108; (**III**) Negative control after 1 h of incubation: PBS + *L. plantarum* 49; (**IV**) Negative control after 1 h of incubation: PBS + *L. fermentum* 111.

**Figure 2.** Chromatograms of aflatoxin M1 (AFM1) quantification in PBS. (**I**) Chromatogram of assays after 1 h of incubation: PBS + AFM1 + *L. paracasei* 108; (**II**) Chromatogram of assays after 1 h of incubation: PBS + AFM1 + *L. plantarum* 49; (**III**) Chromatogram of assays after 1 h of incubation: PBS + AFM1 + *L. fermentum* 111; (**IV**) AFM1 recovery chromatogram of *L. paracasei* 108 and AFM1 complex after 1 h of incubation; (**V**) AFM1 recovery chromatogram of *L. plantarum* 49 and AFM1 complex after 1 h of incubation; (**VI**) AFM1 recovery chromatogram of *L. fermentum* 111 and AFM1 complex after 1 h of incubation. (**A**) Retention time (min) of aflatoxin M1 in phosphate buffer solution; (**B**) chromatographic peak area corresponding to aflatoxin M1.


**Table 1.** Percentage (average values ± standard deviation) of aflatoxin M1 (AFM1) removal in phosphate buffer solution by *L. paracasei* 108, *L. plantarum* 49, and *L. fermentum* 111.

Different small letters in the same row (a,b) denote a significant difference (*p* ≤ 0.05) among values, based on Tukey's test; different capital letters in the same column (A,B) denote a significant difference among values (*p* ≤ 0.05), based on Tukey's test.

Previous studies have also verified that the capacity of LAB—either as viable or non-viable cells, of binding aflatoxins (e.g., aflatoxin B1, ochratoxin, trichothecene, and AFM1) in PBS, laboratory media, or dairy matrices (e.g., milk and yoghurt)—varies in an isolate-dependent manner [2,11,22,23]. Aflatoxins bind to the surface components of LAB

cells and variations in aflatoxin's binding capacities among LAB species or isolates could be associated with differences in the bacterial cell wall and cell envelope structures [7]. Early investigations have found lower capacity of AFM1 removal by viable and/or heatkilled cells of different LAB (e.g., *L. plantarum*, *L. acidophilus*, *L. reuteri*, *L. johnsonii*, *L. rhamnosus*, *L. bulgaricus*, and *Streptococcus thermophilus*) [2,22,23], including probiotic *L. casei* [10], compared to *L. paracasei* 108, *L. plantarum* 49, and *L. fermentum* 111. The efficacy of AFM1 removal from PBS as high (>60%) as those found for *Lactobacillus* isolates examined in this study was reported to *L. plantarum* MON03 and *L. rhamnosus* GAF01 after 6 or 24 h of incubation [24].

Results of the AFM1 retention capacity of the viable and heat-killed cells of *L. paracasei* 108, *L. plantarum* 49, and *L. fermentum* 111 after washing with PBS are presented in Table 2. Percentage values of recovered AFM1 from viable and heat-killed cells were in the range of 13.4 ± 1.5–60.6 ± 1.6% and 10.9 ± 1.2%–47.9 ± 1.5%, respectively. The highest values of recovered AFM1 after 1 and 24 h were found for *L. fermentum* 111 and *L. paracasei* 108, respectively, for both viable and heat-killed cells. Only for *L. fermentum* 111 did the values of recovered AFM1 decrease after 24 h for viable and heat-killed cells; for *L. paracasei* 108 and *L. plantarum* 49, these values varied with the viability/non-viability of cells and incubation time period. Overall, *L. plantarum* 49 had the higher AFM1 retention capacity after washing. Variations in aflatoxin release have been linked to the differences in binding sites in different LAB isolates, or even in these binding sites being very similar. They could have minimal differences depending on each isolate [13,25,26].

**Table 2.** Percentage (average values ± standard deviation) of recovered aflatoxin M1 (AFM1) in solution after washing with phosphate buffer solution.


Different small letters in the same row (a–c) denote a significant difference (*p* ≤ 0.05) among values, based on Tukey's test; different capital letters in the same column (A,B) denote a significant difference among values (*p* ≤ 0.05), based on Tukey's test.

For all examined isolates, the values of recovered AFM1 decreased after 24 h of incubation, indicating that AFM1 retention capacity increased when the length of the contact time increased. There was no clear association between the capability of removing AFM1, initially, and of retaining AFM1 after washing among examined isolates. Interestingly, a study with different *Lactobacillus* species found lower AFM1 removal values than those found in this study, although the recovery of AFM1 from bacterial cells was lower in the former [11].

Heat treatment positively affected the capability of retaining AFM1 in *L. paracasei* 108 after 1 h of incubation, as well as of *L. plantarum* 49 and *L. fermentum* 111 after 24 h of incubation. Heating could increase the interaction capacity of bacterial cells/aflatoxin complexes by causing an increased exposure of the cell wall components, primarily polysaccharides and peptidoglycans, which act as binding sites to aflatoxin [14]. However, the destruction of specific components of the bacterial cell wall by heating, causing the denaturation of proteins and increased cell surface hydrophobicity, has been cited to result in a decreased capability of LAB cells of binding AFM1 [7]. An increased capability of removing aflatoxin B1 was also found in *L. rhamnosus* after heating [27].

The recovery of the AFM1 bound to the cells of examined *Lactobacillus* isolates after washing indicates that the binding was not strong and could not involve a non-covalent weak bond, but probably a physical association of AFM1 with hydrophobic sites in the

bacterial cell wall [13,20,25]. The lower AFM1 recovery values found for the examined isolates could be linked to the interaction of AFM1 molecules retained in the bacterial cell wall with other AFM1 molecules retained in adjacent cells, forming a type of cross-linked matrix that avoids aflatoxin release during washing [10]. Probably, the efficacy of this type of cross-linked matrix decreased over time for *L. paracasei* 108 and *L. plantarum* 49. Although some authors have reported that a part of non-recovered AFM1 might be degraded or biotransformed by a *Lactobacillus* metabolism [2,7], most of the available literature has indicated that aflatoxins are not removed by the metabolism of LAB, but because of a physical bound to the molecular components of bacterial cells, primarily peptidoglycans from the cell wall [19,21,25].

In agreemen<sup>t</sup> with available literature, the results of this study showed that the cell viability of the examined isolates is not a prerequisite for the removal and retaining of AFM1 [13,28]. Cell concentration as high as 108–109 CFU/mL of viable or non-viable LAB is typically needed to reach a level of aflatoxins removal of ≥ 50% [22,28].
