Antifungal Activity of Biocontrol Agents In Vitro and Potential Application to Reduce Mycotoxins (Aflatoxin B1 and Ochratoxin A)

Food bio-preservatives are requested as substituents of chemical pesticides in food. The aim of this study was to carry out a screening of twenty biocontrol agents (BCAs) for their potential fungicidal activity in vitro. Twenty BCAs were tested against ten pathogenic fungi. Some of the cell-free supernatants (CFS) tested showed in vitro antifungal activity versus pathogenic fungi. The highest fungicidal activity was observed in the fermented CFS of Paenibacillus chibensis CECT 375, Bacillus amyloliquefaciens CECT 493, and Pantoea agglomerans CECT 850, which showed a minimum inhibitory concentration (MIC) and minimum fungicidal concentration (MFC) values of 125 and 250 g/L, respectively. The compounds responsible for the antifungal activity, such as organic and phenolic acids, were determined. Lactic acid, acetic acid, benzoic acid, and phenyllactic acid among others can be related to antifungal activity. HPLC-MS/MS analysis showed a reduction of ochratoxin A (OTA) and aflatoxin B1 (AFB1) up to 26% (Paenibacillus alvei CECT 2) and 55% (Paenibacillus polymyxa CECT 155), respectively. The present study prompts that metabolism products of BCAs are propitious for the bioconservation of food, due to their ability to reduce the proliferation of mycotoxigenic fungi and mycotoxins production.


Introduction
Fungi contamination remains a significant issue for the food industry due to their ability to degrade food and produce mycotoxins [1]. Fungal contamination of crops occurs at various stages during pre-harvest and postharvest, producing economic losses for farming and health problems in animals and humans [2][3][4]. It is estimated that about 10% of world basic staples are lost due to fungal contamination [5].
From the toxicological perspective, the most significant mycotoxins are aflatoxins. Aflatoxins are toxic metabolites mainly produced by Aspergillus flavus and Aspergillus parasiticus. Among them, aflatoxin B 1 (AFB 1 ) is the most worrisome. It has been classified as a group 1 compound, due to its carcinogenic effect, by the International Agency for Research on Cancer (IARC) [6]. In long-term exposure, this mycotoxin is associated to liver carcinoma in humans [7,8]. Another concerning mycotoxin is ochratoxin A (OTA), which is a nephrotoxic substance. It also possesses teratogenic, immunotoxic, and possibly neurotoxic properties. IARC classifies OTA as a possibly carcinogenic compound in humans

In Vitro Antifungal Activity of Biocontrol Strains
The overlay assay showed a high antifungal activity of twenty isolated BCAs versus Penicillium, Aspergillus, Fusarium, and Alternaria species (ten in total) expressed as percentage of inhibition after seven days with respect to the control. The strains Paenibacillus polymyxa CECT 155 and Metschnikowia pulcherrima CECT 10408 showed percentage of inhibition (up to 20%) for all pathogens tested, even reaching 50% of inhibition in most cases with the exception of Aspergillus flavus (ITEM 8111) ( Table 1). P. polymyxa CECT 155, Pantoea agglomerans CECT 850, M. pulcherrima CECT 1691, and CECT 10408 evidenced the highest antifungal activities against almost all Fusarium pathogenic fungi tested. P. polymyxa CECT 155 showed percentage inhibition of 50.5%, 48.2%, 40.3%, and 46% for Fusarium graminerarum, Fusarium poae, Fusarium verticilloides, and Fusarium langsethiae, respectively. P. agglomerans inhibited 47.4% of Fusarium sporotrichoides, whereas strains of M. pulcherrima (CECT 1691 and CECT 10408) showed inhibition of 50.4% and 50.1% for Fusarium proliferatum, respectively (Table 1). On the other hand, some of the CFS (cell-free supernatants) of biocontrol microorganisms tested proved antifungal activity against pathogenic fungi in the solid medium diffusion agar test ( Table 2). The highest antifungal activity was detected in the CFS of Paenibacillius chibensis and P. agglomerans, which inhibited the growth of almost all the pathogens tested, and Bacillus amyloliquefaciens, which inhibited the growth of F. Poae and F. graminearum by the diffusion agar method ( Table 2). The antifungal capacity of P. chibensis against Fusarium verticilloides, F. graminearum, and F. langsethiae was proved, as can be clearly seen in Figure 1. In accordance with these results, other authors evidenced the antifungal activity of bacteria against different fungi species by using the diffusion agar method [1,16,19] and overlay method [17]. However, none of these studies used such selection of BCAs including non-lactic acid bacteria, yeast, and fungi. With the aim to quantify the potential antifungal activity of the three bioactive CFS (P. chibensis, P. agglomerans, and B. amyloliquefaciens), the minimum inhibitory concentration (MIC) and minimum fungicidal concentration (MFC) values were determined. The results of MIC-MFC tests are described in Table 3. In line with qualitative results, the three CFS tested showed inhibitory activity reaching MIC-MFC values from 125 to 250 g/L against almost all the fungi tested (Penicillium spp., Aspergillus spp., Fusarium spp., and Alternaria spp.). The three biocontrol strains tested (P. chibensis, B. amyloliquefaciens, and P. agglomerans) showed MIC and CFS values of 125 and 250 g/L, respectively, against F. langsethiae. Meanwhile, the CFS of strain PA850 showed MIC and CFS values (125 and 250 g/L, respectively) against Penicillium and Alternaria strains. P. chibensis showed such values against F. graminearum and A. flavus. These results were in accordance with Luz et al. (2020) [16], who evidenced that the Fusarium genus was the most sensitive specie to the bioactive compounds present in the fermented CFS of Lactobacillus plantarum. In the same work, the authors showed MFC values of 250 g/L from LABs against Aspergillus [16]. In the same line, Izzo et al. (2020) investigated the inhibitory effect of sweet whey fermented by Lactobacillus plantarum strains against fungal growth (Fusarium and Aspergillus), and the results of the MIC-MFC varied between 1.95 and 250 g/L, whereas the value of MFC was in the range of 4 to 250 g/L [12]. Hence, MIC and MFC values obtained in this work are similar comparing with values previously reported in which LABs were tested to increase the shelf life of food [12,16,19]. These results suggest the possible use of BCAs as a source of new preservatives of natural origin to incorporate in food matrices for the purpose of improving the shelf life. However, due to the Good Hygienic and Manufacturing Practices required for food production and their low microbial load, the use of concentrations significantly below the MIC and MFC values could also result in a significant increase in shelf life.   Table 2. Antifungal activity of twenty cell-free supernatants (CFS) against Penicillium, Aspergillus, Fusarium, and Alternaria species by the diffusion agar method. The antifungal activity was expressed as follows: (+) means 8 mm of inhibition zone between the well and fungal growth, (++) means 8-10 mm of inhibition zone between the well and fungal growth, (+++) means > 10 mm of inhibition zone between the well and fungal growth. The well radius was 5 mm.

Identification of Antifungal Compounds in CFS
In this study, two organic acids and seven phenolic acids were identified in CFS fermented by twenty and nine biocontrol strains, respectively (Tables 4 and 5). These compounds were determined by HPLC-qTOF-MS. The organic acids in lyophilized CFS fermented by BCAs are indicated in Table 4. Two organic acids were determined: lactic acid and acetic acid. All isolated BCAs produced lactic acid, reaching concentrations from 129 to 3491 mg/L. The BCAs with the highest lactic acid production were Bacillus thuringiensis (3491 mg/L), P. polymyxa CECT 155 (3372 mg/L), P. polymyxa CECT 153 (2622 mg/L), Bacillus subtilis (2553 mg/L), and Bacillus megaterium (2350 mg/L). The production of acetic acid (203-1873 mg/L) was determined only in five CFS, the highest concentration produced being P. polymyxa CECT 153 and P. polymyxa CECT 155 (1873 mg/L and 1271 mg/L, respectively). Table 5 lists the nine fermented CFS in which phenolic acids were detected. These phenolic acids detected and quantified in the CFS were DL-3-phenyl lactic acid (PLA), benzoic acid, 1-2-dydroxybenzene, hydroxycinnamic acid, p-coumaric acid, 3-4-dihydroxycinnamic acid, and protocatechuic acid. Some of these compounds have been previously reported as antifungal agents produced by LABs [18]. Benzoic acid was quantified in each CFS tested, and its concentration range was 0.004-0.067 g/L. The highest concentrations of benzoic acid were detected in P. alvei and P. polymyxa CECT 153 (0.067 and 0.060 g/L, respectively).
PLA was detected only in five CFS, and concentrations ranged from 0.058 to 0.844 g/L, among which the B. thuringiensis strain was the highest producer of PLA. In addition, 1-2-dyhdroxybenzene was quantified only in two CFS (B. amyloliquefaciens and P. chibensis), with values of 0.03-0.04 g/L, respectively. These results are in line with Rodriguez et al. (2008), who noticed the presence of 1-2-dyhdroxybenzene generated by enzymes of Lactobacillus plantarum [22]. The fact that P. polymyxa CECT 155 and P. chibensis were the only CFS that produced dydroxyhydrocinnamic (0.022 ± 0.06 g/L) and protocatechuic acid (0.008 ± 0.01 g/L) respectively, should be highlighted.  Our results pointed out that strains of P. polymyxa CECT 153, P. polymyxa CECT 155, P. chibensis, B. amyloliquefaciens, and P. alvei showed the highest capacity to produce organic and phenolic compounds. Like organic acid production, the phenolic acids data were associated with the observed antifungal activity of CFS. In fact, previous authors evidenced synergism between lactic acid and acetic acid [23], and between these and PLA regarding the potential antifungal activity [24].
Moreover, some of these phenolic compounds with antifungal activity, above described, have been linked to the ability to prevent mycotoxin production [1,19,25,26].

Antimycotoxigenic Activity of CFS
Along with evaluating the qualitative and quantitative antifungal activity of CFS, the ability to reduce mycotoxins was investigated. The antimycotoxigenic activity was evaluated by using 20 biocontrol strains against AFB 1 (1 mg/L) and OTA (1 mg/L) after 48 h of exposure (Table 6). To our knowledge, this is the first work in which the antimycotoxigenic activity of 20 different BCAs against ten pathogens has been investigated. As evidenced in Table 6, 11 out of 20 and 14 out of 20 BCAs tested showed AFB 1 and OTA reduction above 10%. The reduction of AFB 1 was in the range of 1.2 to 55%, and the reduction percentage of OTA ranging from 4.7 to 26.5%. P. polymyxa CECT 155 showed the highest reduction (55%) of AFB 1 after 48 h of exposure. These values are slightly higher than those found in other studies testing biocontrol strains. Topcu et al. (2010) tested some non-lactic acid bacteria, specifically probiotic Enterococcus faecium M74 and EF031 strains, and observed a reduction of the AFB 1 content of aqueous solution by 19-38% [27]. However, other authors noticed that a Bacillus subtilis strain decreased the AFB 1 level of contaminated feed and food by 60-95% [28,29]. These differences may be related to the type of biocontrol strain, days of exposure, and the presence or absence of a food matrix.
Moreover, in our study, P. alvei showed the highest reduction of OTA (26.5%) after 48 h of exposure. Similar results were obtained by Shukla et al. (2018) who isolated Bacillus subtilis KU-153 from Kimchi (a traditional Korean fermented food) and investigated its ability to lessen OTA content in culture medium. They observed a reduction of OTA content in viable cells by 22%. However, B. subtilis heat-treated cells proved a higher OTA reduction (45%) [30].
As previously mentioned, these biocontrol strains were among the great producers of phenolic compounds; hence, the production of metabolites with antifungal activity may explain their anti-mycotoxigenic activity.
The present results suggested that metabolism products of BCAs, due to their potential to minimize the growth of mycotoxigenic fungi and the production of the mycotoxins, could be promising for the bioconservation of food.

Conclusions
In this study, twenty different BCAs were tested against ten mycotoxigenic fungi.
In vitro experiments proved that CFS fermented by isolated BCAs has significant antifungal activity versus P. chibensis, B. amyloliquefaciens, and P. agglomerans. The organic (lactic and acetic acid) and phenolic acids (benzoic acid and PLA, among others) produced by BCAs showed antifungal activity against Fusarium spp., A. flavus, Penicillium verrucosum, and Alternaria alternata.
Furthermore, the use of BCAs partially reduced the presence of AFB 1 and OTA, two mycotoxins of great concern to animal and human health that have been declared carcino-genic to humans. Specifically, the biocontrol strain of P. polymyxa CECT 155 decreased the in vitro incidence of AFB 1 and OTA up to 55% and 20%, respectively. The reduction of mycotoxins production is mainly due to the inhibition of fungal proliferation by the bioactive compounds produced during the fermentation process by BCAs. The present work suggests that both the studied BCA strains and their metabolism products, due to their potential to bring down the growth of the toxigenic fungi and the production of the mycotoxins, could be promising for the bioconservation of food. These strains could be used in combination with starters in fermented foods, and the fermented freeze-dried solution could be used in both formulation and coating of foods. Therefore, these results provide new knowledge for the biotechnological use of BCAs, since they could increase the postharvest shelf-life of food, contributing to meet the demand of consumers, reducing the agricultural use of chemicals and increasing natural alternatives.
Further investigations are needed to clear up whether the efficacy of BCAs might be enhanced by adding more ingredients and other bio-products that can have synergistic effects.
Finally, the application of BCAs for the biocontrol of mycotoxins should be implemented combined with suitable farming practices and appropriate postharvest management.

Microbiological Culture
The strains of Penicillium verrucosum CECT 2913 and Botytis cinerea CECT 20973 were obtained from the Spanish Type Culture Collection (Science Park of the University of Valencia, Paterna, València, Spain), which is the public reference center for microbial resources in Spain. The strains of Aspergillus flavus ITEM 8111, Alternaria alternata ITEM 8121, and six Fusarium strains (Fusarium graminearum ITEM 126, Fusarium poae ITEM 9151, Fusarium langsethiae ITEM 11031, Fusarium verticillioides ITEM 12043, Fusarium proliferatum ITEM 12072, and Fusarium sporotrichioides ITEM 12168) were obtained from the Agro-Food Microbial Culture Collection of the Institute of Sciences and Food Production (ISPA, Bari, Italy). These fungi were cryopreserved in sterile liquid medium PDB with 30% glycerol at −80 • C. Prior to antifungal studies on solid and liquid medium, the strains were defrosted and grown in solid medium PDA at 25 • C, with periodic transfers to new PDA plates. These pathogens were selected to cover a representative sample of fungi of the genera Aspergillus, Penicillium, Fusarium, and Botrytis.
A total of twenty biocontrol strains:  (Table 7) and kept in frozen glycerol (30%) stocks at −80 • C until analysis. The antifungal activity of these 20 biocontrol strains was studied against the 10 toxigenic fungi belonging to the genera Aspergillus, Penicillium, Fusarium, and Alternaria described above. These biocontrol strains were selected mainly because of the scarce data in the literature, which has focused almost exclusively on the use of LABs. Prior to the fermentation process, the strains were cultivated in different mediums at different temperatures according to the best conditions for their growth. P. polymyxa (CECT 153 and CECT 155), B. thuringiensi, P. chibensis, P. alvei, B. megaterium, B. subtilis, and P. agglomerans were inoculated with NB (13 g in 1000 mL of distilled water) at 30 • C, while B. licheniformis and B. amyloliquefaciens were inoculated in the same medium at 37 • C. S. griseus and S. calvus were inoculated with yeast extract-malt extract (4 g, malt extract 10 g, glucose 4 g) at 30 • C. P. syringae (CECT 312, CECT 4390, and CECT 4393) were inoculated with TSB (15 g of triptone soy broth in 1000 mL of distilled water) at 26 • C and C. sake (CECT 1044 and CECT 10034), C. oleophila and M. pulcherrima (CECT 1691 and CECT 10408) were inoculated with glucose peptone-yeast extract (40 g of glucose, 5 g of Triptone, 5 g of yeast extract in 1000 mL of distilled water) at 24 • C.

Process of Obtaining Cell-Free Supernatant (CFS)
After their defrost and recovery, the biocontrol strains were cultured for 12 h in the media described in 4.2 until exponential phase growth. Then, the strains were placed in 50 mL of NB liquid medium (final concentration 10 7 CFU/mL) and maintained at 30 • C for 72 h. After this time, the liquid medium was centrifuged for 10 min at 7000 rpm to obtain CFS. These CFS were frozen for 24 h at −80 • C prior to lyophilization (FreeZone 2.5 L Benchtop Freeze Dry System, Labconco, Kansas City, MO, USA) and then stored at −19 • C [16,19].

Qualitative Antifungal Activity of Biocontrol Strains
The effect of the BCAs on the growth inhibition of different fungi (listed in Table 2) was studied by the agar diffusion method and overlay assay. In the agar diffusion assay, PDA medium plates were inoculated with the 10 fungi described in Section 4.2 and after making 1 cm wells, these were loaded with 50 µL of lyophilized and resuspended CFS at a concentration of 100 g/L. After incubating the plates, the appearance of inhibition halos was observed [14,16].
In the overlay assay, the center of the PCA plates was contaminated with a suspension of the different BCA, and after solidification, they were covered with PDA medium previously inoculated with spores of the different toxigenic fungi. After incubating the plates, the appearance of inhibition halos was observed and measured [17,20].

Quantitative Antifungal Activity of Biocontrol Strains
Sterile 96-well microplates were used for this assay as described by Luz et al. (2020) [16]. The objective was to determine the MIC and minimum fungicide concentration (MFC). In the wells, 100 µL of CFS suspended at doses between 0.1 and 100 g/L were deposited together with 100 µL of PDB liquid medium contaminated with 5 × 10 4 spores/mL of the fungi described in Section 4.2. Microplates were incubated in the dark at 25 • C for 72 h, and MIC was considered as the lowest concentration of CFS at which no fungal growth was visually detected. To determine the MFC, 10 µL from each well with concentrations above the MIC were seeded onto PDA medium plates and incubated (25 • C, 72 h.), and the presence or absence of fungal growth was observed.

Determination of Organic and Phenolic Acids in CFS
A high-performance liquid chromatography (HPLC) system (Agilent 1100 Series HPLC System, Agilent Technologies, Palo Alto, CA, USA) was used to analyze the organic acids. The samples were diluted in distilled water and injected using a Spherisorb S5 ODS2 reversed-phase column (4.6 mm × 250 mm, 5 µm) (Waters Corp., Milford, MA, USA). The mobile phase was acidified water at pH 2.1 (0.6 mL/min). The diode array detector (DAD) was set to a wavelength of 210 nm.

Antimycotoxigenic Activity of CFS
The antimycotoxigenic activity of the CFS by biocontrol microorganisms was evaluated. This assay was carried out by using 20 biocontrol strains (previously listed). For each BCAs, sterile 15 mL Falcon tubes were provided with 5 mL of NB contaminated with 1 mg/L of AFB 1 and 1 mg/L OTA. Then, 50 µL of a NB suspension with recently grown bacteria (48 h ago) were added. To make this NB with the mycotoxins, 1 L of sterile NB was prepared to which the appropriate amounts of mycotoxin standard (Sigma-Aldrich (St. Louis, MO, USA) were added. Tubes were kept in an oven at 30 • C for 48 h while control tubes (0 h) were analyzed immediately after preparation. Finally, tubes were centrifuged at 4000 rpm for 15 min at 4 • C, and supernatants were filtered (0.22 µm), and vialized for mycotoxin determination by LC-MS/MS-QTRAP, as explained below.

Determination of AFB 1 and OTA by LC-MS/MS Spectrometry
For mycotoxin analysis, the samples were injected into an HPLC coupled to a 3200QTRAP mass spectrometer (Applied Biosystems, Foster City, CA, USA). The column used to separate mycotoxins was a Gemini NX C18 column (150 × 2.0 mm I.D, 3.0 mm, Phenomenex, Palo Alto, CA, USA). The mobile phases consisted of water (A) and ACN (B), both with 0.1% formic acid and 5 mM ammonium formate at 0.25 mL/min with a linear gradient. The ions transitions used for the AFB 1 and OTA identification and quantification were m/z 313.1/241.3 and 284.9 (AFB 1 ) and m/z 404.3/102.1 and 358.1 (OTA) [33].

Statistical Analysis
Statistical analysis was performed using the InfoStat software version 2008. The assays were realized in triplicates, and the differences between control and treated groups were analyzed by Student's t-test, while the differences between the groups were analyzed by one-way ANOVA test. The significance levels were set at p ≤ 0.05.