Isolation and selection of antifungal lactic acid bacteria to use as inoculants for improving quality of fermented plant beverages

Abstract

Yeast contamination invariably occurs with fermented plant beverages (FPBs). To understand what strains of yeast were, the most prevalent contaminants, 40 FPB samples collected from various sources were investigated. Fifty-nine contaminating yeast strains were isolated and identified into seven groups by conventional methods. A representative yeast strain from each group was then selected for identification using 26S rRNA gene sequences. The results of DNA sequence analysis indicated that the main group of contaminating yeasts was Rhodotorula mucilaginosa (14 isolates, 23.7%) followed by Pichia membranifaciens (11 isolates, 18.6%). Other minor groups were Pichia anomala (9 isolates, 15.3 %), Saccharomyces cerevisiae (6 isolates, 10.2 %), Issatchenkia occidentalis (5 isolates, 8.5 %) and Candida tropicalis (4 isolates, 6.8 %) and the remaining 10 isolates, (16.9%) could not be identified. All isolates except for S. cerevisiae were film yeasts. In order to control yeasts that might be contaminated, FPBs selected lactic acid bacteria (LAB), isolated from FPBs and fermented foods, were tested for their ability to inhibit growth of yeasts when used to assist with the FPB process. 500 isolates of LAB were investigated. 261 LAB isolates produced inhibitory effects against an isolate of Rhodotorula mucilaginosa DKA and previously isolated as an FPB contaminant, when production by the LAB of acid or hydrogen peroxide was not limited. However, only 6 of the LAB isolates showed good yeast inhibitory activity as the productions of both compounds were limited. Based on the spectrum of yeasts inhibited, isolate DW3 was selected for further studies due to its ability to inhibit 9 of the 14 contaminating yeast species tested, and it was identified as Lactobacillus plantarum by a commercial test kit API 50CH and confirmed with 16S rDNA sequence. A 10X concentrated neutralized culture filtrate (CNF or freeze dried supernatant: FDS) from isolate DW3 did not inhibit the molds and yeasts tested. However, the FDS gave good inhibition against Staphylococus aureus PSSCMI 0004, Vibrio parahaemolyticus PSSCMI 0064, Shigella sonneii PSSCMI 0032 and Escherichia coli PSSCMI 0001. The FDS was active against tested bacteria (S. aureus PSSCMI 0004 and E. coli PSSCMI 0001) in an order of acidity > neutralization > alkalinity and in the temperature range of less than 60ºC, for 30 min. The enzymes catalase, amylase and lipase had no effect on the FDS antimicrobial activity. However, trypsin, pepsin, α-chymotrypsin and pronase E, completely inhibited its antimicrobial activity. Based on these results - a class III bacteriocin - (large heatlabile protein), we suggest that the antimicrobial activity of L. plantarum DW3 was named bacteriocin DW3. The bacteriocin DW3 gave higher inhibitory effect against S. aureus PSSCMI 0004 (80 AU/ml) than that of E. coli PSSCMI 0001 (60 AU/ml). Besides, the bacteriocin DW3 was kept for 30 days in a form of FDS which had higher stability to control both organisms than that found in a concentrated solution form. In addition, the isolate DW3 also produced the antifungal substance that inhibited variety of microbes, particularly R. mucilaginosa DKA. The substance was isolated and identified by Gas chromatography-Mass spectrometer (GC-MS) and it was phenyllactic acid (PLA). The isolate DW3 produced 31 mg/L PLA in MRS medium and its MIC90 against the target yeast was 5 mg/ml. In order to identify whether the use of L. plantarum DW3 could control the amount of yeast in a fermented seaweed beverage (FSB), it was introduced as a starter culture at approximately 7.5 x 106 CFU/ml into partial sterilization (0.5% KMS: potassium metabisulfite) or no partial sterilization FSB substrates, together with appropriate controls (No Partial sterilization-No starter: N-N and Partial sterilizationNo starter: P-N). At the end of the fermentation (60 days) the amount of the yeast in the FSB with DW3 as starter cultures was less than 100 CFU/ml (N-S) and no yeast was found at day 45 in a set of P-S. The natural fermentation (N-N) did not pass a standard guideline for yeast count; however, partial sterilization without inoculation of DW3 (P-N) could pass the standard level. The amounts of LAB at day 60 in the NS and P-S sets were 4.7 and 4.0 log CFU/ml, whereas those in N-N and P-N sets were 2.4 and 2.3 log CFU/ml, respectively. At the end of the FSB fermentation with starter, there was more acidity as both sets of inoculated cultures obtained 0.95% acidity with 2.9 pH and remaining sugar at 4%. The FSB had similar changes of acidity, electrical conductivity (1.0 mS/cm) and elements (Na, K, Cu, Fe and Zn). In addition, all these values increased during the fermentation time and were below the recommended safety levels for drinking. Toxic elements, As and Pb, were not detected in any sets of the FSB. Moreover, the FSB was a good source of Fe with as much as 17 mg/L in the N-N and P-S sets. In the FSB, no methanol was detected throughout the fermentation and the organic acid levels in the partial sterilization treatment sets were significantly lower than those in the no partial sterilization sets (p = 0.05). Acetic acid levels at day 60 in the FSB were in the range of 1.8 to 2.5 g/L while lactic acid was between 3.2 to 7.0 g/L. The inhibitory effect of culture fluids from all inoculated sets on all target organisms (Staphylococcus aureus PSSCMI 0004, Escherichia coli PSSCMI 0001, Salmonela sp., Vibrio parahaemolyticus PSSCMI 0064 and Bacillus cereus ATCC 11778) was much higher than that of the uninoculated sets (P = 0.05) and V. parahaemolyticus PSSCMI 0064 was the most sensitive organism to the beverage. Results of the sensory tests for the FSB at days 30 and 60 were similar in both inoculated sets with higher scores from the factor tests (most acceptable) than those from the uninoculated sets although no significant difference was found (P>0.05). The maximum score in each sensory test was found in the N-S set. Moreover, sets partial sterilization using 0.5% KMS obtained lower scores than no partial sterilization treatment sets. A three month-storage time produced no effects on the sensory tests for the beverage because there is no change in the acceptable level. Each treatment set of the beverage obtained an average score of more than 3.00 which is moderately acceptable. The finished product of FSB at the end of fermentation and after storing for 3 months passed microbiological quality tests because there were no detectable coliforms and E. coli (fecal indicators). Moreover, none of sets found S. aureus, Clostridium perfringens and Salmonella spp. The storage time of the beverage provided sterile conditions. The inhibitory effects of the beverage against pathogenic bacteria were retained with only minor reductions. The inoculated treatment sets had a higher inhibiting activity than that in the uninoculated sets (P = 0.05).