Reducing fumonisin exposure in African fermented maize-based foods with lactic acid bacteria

Abstract

Maize, rice, wheat and barley are the world’s four major cereal crops. In Africa, maize is the major staple cereal crop for the majority of the population and most of the traditional fermented products are maize-based (e.g. ogi, kenkey, mawe and mahewu). Similar to other cereals, maize is also at risk of infestation by mycotoxigenic fungi. International studies have indicated that fumonisins (64%) take the lead in mycotoxins contamination worldwide followed by deoxynivalenol (59%), zearalenone (45%), aflatoxin (33%) and ochratoxin A (28%). Various African studies have indicated fumonisins occur in maize at levels that can exceed 10 μg/g. Fumonisins are most widely known to cause leucoencephalomalacia in horses and high levels of fumonisin B1 (FB1) has been linked to increased incidence of oesophageal cancer in Eastern Cape formerly known as Transkei. Further research has shown the consumption of FB1 by pregnant women can cause birth defects. Due to their structural similarity to sphingonine and sphingosine, fumonisins interfere with sphingolipid metabolism and this result in liver disease and tumours in the liver and kidneys. Many methods have been investigated for the detoxification of mycotoxins. A new method of detoxification has recently been developed wherein some mycotoxins have a natural affinity to bind to microorganisms. Recent studies have indicated that probiotics, specifically lactic acid bacteria (LAB), have a natural ability to bind to fumonisins and introducing a detoxifying effect on the fermented product. In this project, we investigated the ability of the following dominant probiotic LAB strains isolated from traditional fermented maize-based products (ogi and mahewu) namely, Lactobacillus plantarum FS2, L. delbrueckii subsp. delbrueckii CIP 57.8T and Pediococcus pentosaceus D39, against a positive control strain, L. plantarum R1096, to reduce the toxicity of fumonisins (B1 and B2). In completion of the main objective, each LAB strain was evaluated for its binding ability. Visualization of the binding interaction was achieved by derivatization of the fumonisins and interaction with the non-fluorescent LAB cells (viable and non-viable) under confocal laser scanning microscopy (CLSM) with the appropriate controls in place. Results provided physical evidence that the fluorescent fumonisins (B1 and B2) bound to the LAB cells externally, as both the viable and non-viable cells fluoresced green under CLSM. To the best of our knowledge, this is the first study to visualize the interaction between LAB and fumonisins. Each LAB strain was evaluated to quantify the binding of fumonisin molecules to a certain number of bacterial cells, proof of adsorption as well as reproducibility and repeatability. The high binding percentages obtained indicated successful binding strength being achieved across all LAB strains for both FB1 and FB2. For viable cells, FB1 bound the most to L. plantarum FS2 but the least to P. pentosaceus D39, whereas, the highest binding for FB2 was with L. plantarum R1096 but the least with L. delbrueckii CIP 57.8 T. Similarly, non-viable cells of L. plantarum FS2 indicated the highest binding to FB1 and FB2, while P. pentosaceus D39 and L. delbrueckii CIP 57.8 T bound the least to FB1 and FB2, respectively. Overall, non-viable cells bound more to FB1 and FB2 in comparison to viable cells due to the heat shock causing the cell wall to disrupt exposing more favourable binding sites to FB1 and FB2. Binding occurred with viable and non-viable cells which confirmed the results seen with the visualization of binding. The binding strength was repeated in two independent trials to establish repeatability and reproducibility where the statistical analyses indicated no significant differences. To test the stability of the LAB – fumonisin complex, the main parameters were based on the fermentation process of ogi: incubation temperature of 30 oC and decreasing pH from 6 to 4 that occurs in a natural ogi fermentation process; same decreasing pH values and optimal growth temperature of the LAB strains which is 37 oC; simulated gastric condition of a lower pH of 2 and temperature of 37 oC and storage period of 7 days at pH 4 at 30 oC were selected. Results indicated that whether ogi fermentation process occurs at 30oC or 37 oC, the binding strength for both fumonisins was still very high, in addition, as the pH decreases it can be seen that the binding for both the fumonisins increased. Even at the lower pH of 2 and incubation temperature of 37 oC, binding was quite high and the complex was still stable. For the storage conditions (7 days at pH 4 and 30 oC), contrary to previous binding strength results in this project, L. plantarum R1096 was the only strain that increased in binding strength whereas the other strains decreased in binding. Generally, as pH decreased the binding increased, but across all strains, FB2 was bound more than FB1. Due to the liver being the main site of toxicity, the bound complex was evaluated on HepG2 cell line with controls and the LAB strains in order to give a comparative analysis and have a clear understanding of its toxicity in vitro for the objective of reducing the toxicity of fumonisins by use of LAB strains. HepG2 cells were observed in the presence of fumonisin only, LAB cells only and then with a combination of fumonisin and LAB cells to assess toxicity using the sulforhodamine B (SRB) assay. The results indicated that the LAB cells did not attach, aggregate or disrupt the HepG2 cells as the absorbance readings remained constant throughout the 24-48 h incubation period. Observation of the HepG2 cells exposed to the fumonisins alone indicated the IC50 value of 308.6 μg/ml. The comparative observation of the HepG2 cells exposed to a combination of the fumonisins and LAB cells indicated that the LAB cells “protected” the HepG2 cells by binding to the fumonisins and increasing the IC50 values. L. plantarum R1096 “protected’ the HepG2 cells the most by increasing the IC50 value to 903.1 μg/ml followed by L. delbrueckii CIP57.8T at 857.3 μg/ml, similarly with L. plantarum FS2 at 856.5 μg/ml and the least by P. pentosaceus D39 at 701.4 μg/ml. All LAB strains had more than doubled the IC50 value by being present and binding to the fumonisins. In conclusion, L. plantarum FS2, L. delbrueckii CIP 57.8T and P. pentosaceus D39, test strains isolated from African traditional fermented foods, have been successfully tested against the positive control strain L. plantarum R1096 and can be used as detoxifying agents in reducing the toxicity of fumonisin B1 and B2 in fermented cereal based products.