Partial Characterization of Bacteriocins Produced by Lactobacillus reuteri 2-20B and Pediococcus acidilactici 0-11A Isolated from Fura, a Millet-Based Fermented Food in Ghana

Cell-free supernatants (CFS) produced by 369 LAB strains previously isolated from fura and identified based on sequencing of their 16S rRNA genes were screened for their antagonistic activities against pathogenic bacteria including Bacillus cereus PA24, Escherichia coli SKN 541, Enterococcus faecalis 103907 CIP, Staphylococcus aureus ATCC 19095 and Listeria monocytogenes Scott A, using the agar well diffusion method. Bacteriocins of Lactobacillus reuteri 2-20B and Pediococcus acidilactici 0-11A were further evaluated for their stability when subjected to a range of pH conditions, enzymatic and heat treatments. Growth and bacteriocin production rates as well as influence of media composition on bacteriocin production were also evaluated. Cell free supernatants of Lb. reuteri and Pd. acidilactici strains exhibited the widest inhibitory activities whereas CFS of Lb. fermentum exhibited the least inhibitory activity towards the tested pathogens. Bacteriocins of Lb. reuteri 2-20B and Pd. acidilactici 0-11A retained their antibacterial activities over a wide range of pH. Whereas the antimicrobial activity of Lb. reuteri 2-20B was lost after being subjected to temperature of 90°C for 1 h, CFS of Pd. acidilactici 0-11A remained stable after autoclaving at 121°C for 15 min. Again, the antimicrobial activities of both Lb. reuteri 2-20B and Pd. acidilactici 0-11A were lost when the CFSs were subjected to the action of proteolytic enzymes but remained active under the actions of catalase, lipase and α-amylase. Production of bacteriocins by both Lb. reuteri 2-20B and Pd. acidilactici 0-11A were growth associated and influeced by media composition.


Introduction
Millet (Pennisetum glaucum) is a cereal for the production of many common staple foods in West Africa.In Ghana, millet is used for the production of fermented foods such as koko and koko sour water, furo-furo, maasa, and fura.Fura a millet-based spontaneously fermented dumpling is produced in parts of West Africa including Ghana.The microorganisms associated with fura processing have been isolated and identified by both classical and molecular biology-based techniques to include lactic acid bacteria and yeasts (Owusu-Kwarteng et al., 2012).
Lactic acid bacteria (LAB) have a long history of application in fermented foods because of their beneficial influence on nutritional, organoleptic, and safety and preservative characteristics (Hammes & Vogel, 1995;Leroy & De Vuyst, 2004).During fermentation, LAB displays a range of antimicrobial activities due to the production of organic acids and bioactive molecules such as ethanol, formic acid, fatty acids, hydrogen peroxide (H 2 O 2 ), and diacetyl.The screening of LAB for antimicrobial activity has shown that, besides acidification, some LAB produce inhibitory substances other than organic acids.Thus, certain strains of LAB produce bacteriocins and bacteriocin-like molecules that display antimicrobial activities (De Vuyst & Vandamme, 1994) which inhibits the growth of pathogenic and food spoilage microorganisms thereby ensuring safety of the product.
Whereas the involvement of LAB in fura and similar other traditional fermented foods is based on spontaneous fermentations, a deliberate addition of the isolated microorganisms as starters to the food matrix would result in a high degree of control over the fermentation process to yield a more standardized end product.Owusu-Kwarteng et al. (2012) identified the LAB isolated during traditional fura processing in Ghana.However, no study has so far screened the isolated LAB isolated from fura for their ability to produce bacteriocins.The objective of this study was therefore to screen LAB isolated during traditional fura processing for their ability to produce bacteriocins against selected pathogens, and to assess their effectiveness when subjected to some food processing environments.This is geared towards the provision of rational basis for further research in selecting potential LAB strains for use as starter cultures, co-cultures, or bio-protective cultures, to improve the quality and safety of fura and other fermented food products.

Lactic Acid Bacteria
Lactic acid bacteria used in this study were previously isolated during traditional fura processing in Ghana.They were identified by a combination of phenotypic and genotypic methods including conventional morphological characteristics and carbohydrate fermentation profiling, (GTG) 5 -based rep-PCR fingerprinting and 16S rRNA gene sequencing as described elsewhere (Owusu-Kwarteng et al., 2012).

Indicator Strains (Pathogens)
Indicator organisms used in this study and their sources are shown in Table 1.

Preparation of Cell-free Supernatant (CFS)
Each LAB isolate was grown in MRS broth at 30°C for 48 h in a shaking incubator.The culture extracts were obtained by centrifugation at 5,000 g for 10 min.the supernatants were decanted and the pHs adjusted to 6.5 with NaOH (1 mol L -1 ), to eliminate the effect of acidity (Mathieu et al., 1993).Inhibitory activity of hydrogen peroxide was eliminated by the addition of catalase (5 mg ml -1 , Sigma, C-100 bovine liver) (Daba et al., 1993) and filtered through 0.20 µm pore size syringe filters (Sartorius, Minisart®, Göttingen, Germany).

Screening of LAB for Antagonistic Activity against Pathogenic Bacteria
The agar-well diffusion method was employed in the screening LAB for bacteriocin activities.Indicator lawns were prepared by mixing (inoculating) 20 ml of BHI molten agar media with 100 µl (approximately 10 7 cfu/ml) of an overnight culture of each indicator organism and allowing to solidify in a Petri dish.Wells were cut into the agar with a sterile 6 mm diameter cork borer and sealed with two drops of sterile agar.Fifty microlitres (50 µl) of the filtered cell-free supernatant of test strains were separately placed into the wells.The plates, prepared in duplicate, were kept at 4°C for 24 h (Bonade et al., 2001) to allow pre-diffusion of the CFS into the agar and then incubated at 37°C for 24 h.They were then observed for possible clearing of zones (inhibition zones).The antimicrobial activity was determined by measuring the diameter of the inhibition zones around the well.Results were recorded as no inhibition (-), weak inhibition (+), moderate inhibition (++) and strong inhibition (+++) when the diameter is <1 mm, 1-5 mm, 6-10 mm and >10 mm respectively.

Sensitivity of CFS to pH Changes, Enzyme and Heat Treatment
Sensitivity studies were conducted with CFSs obtained from Lb. reuteri 2-20B and Pd.acidilactici 0-11A.The effect of pH on the bacteriocins was determined by adjusting the cell-free supernatant to pH 2.0 to 9.0 (at increments of one pH unit) with sterile 1 N HCl or 1 N NaOH.After 2 h of incubation at room temperature, the samples were readjusted to pH 6.5 with sterile 1 N HCl or 1 N NaOH and the activity determined by the agar well diffusion method.For enzyme treatment, 2 ml of CFS were incubated for 1 h in the presence of 1.0 mg ml -1 (final concentration) trypsin, pronase E, proteinase K, lipase and α-amylase and then tested for antimicrobial activity.For heat treatment, the CFS were heated in water bath at 30°C to 100°C (at increments of 10°C) for 1 h, or autoclaved (121°C) for 15 min and tested for their bacteriocin activities using the agar well diffusion method described in section 2.4.

Growth Characteristics and Bacteriocin Production
Growth experiments were carried out in 500 ml ErlenMeyer flask containing 250 ml MRS broth.An overnight pre-culture of Lb. reuteri 2-20B and Pd.acidilactici 0-11A were used for the inoculations of MRS broths at initial densities of ca 10 3 cfu/ml.At 2 h intervals, samples were removed from the culture and used for viable plate count (cfu/ml), OD (600nm), pH measurements, and antimicrobial activity.The antimicrobial concentration of each sample was estimated using the critical method of dilution.Arbitrary Unit ml -1 (AU/ml) was calculated as the inverse of the highest two-fold dilution which induced definite inhibition.

Statistical Analysis
Data obtained were subjected to one-way analysis of variance (ANOVA) and means were separated by Tukey's family error rate multiple comparison test (p<0.05)using the MINITAB statistical software package (MINITAB Inc.Release 14 for windows, 2004).

Screening of LAB for Antimicrobial Activities
A total of 369 previously identified strains of LAB (Owusu-Kwarteng et al., 2012) were first screened for their antimicrobial activities.These comprised 176 strains of Lb. fermentum, 82 strains of W. confusa, 60 strains of Lb. reuteri and 51 strains of Pd. acidilactici (Table 2).Antagonistic activities of CFSs of the LAB is shown in Table 2.The result demonstrated a diversity of the strains, within and between species, in their ability and the extent to which they inhibit the growth of different pathogenic bacteria.Strains of Lb. reuteri and Pd.acidilactici displayed the widest inhibitory activities against the indicator pathogens whereas Lb. fermentum species showed the least inhibitory activity spectrum against the pathogens.Staphylococcus aureus enterotoxin A producer was the most susceptible pathogen with about 57% and 52% strains of Lb. reuteri and Pd.acidilactici respectively showing inhibition towards S. aureus ATCC 19095.Generally, the Gram-positive pathogenic bacteria were more susceptible to the inhibitory actions of the isolated LAB than the Gram-negative indicator organisms.
Several bacteriocin producing LAB have been isolated from various traditional spontaneous fermented foods such as bosa (Todorov & Dicks, 2006), kenkey (Olsen et al., 1995), and ogi and fufu (Sanni et al., 1999;Olukoya et al., 1993).Similarly, strains of Lb. reuteri and Pd.acidilactici in this study produced bacteriocins which showed inhibition against both the Gram-positive and Gram-negative bacteria (E. coli SKN541) although the Gram-positive bacteria were generally more susceptible.Bacteriocins of LAB have been defined as bioactive peptides or proteins that are active against Gram-positive bacteria and usually against species closely related to the producer strain (De Vuyst & Vandamme, 1994).There are however, reports on bacteriocins of LAB with activity against a broad spectrum of Gram-positive and Gram-negative bacteria, including Klebsiella pneumoniae, E. coli, and Pseudomonas spp.(Todorov & Dicks 2005) which are not necessarily closely related to the producer strains.

Effect of Heat, Enzymes and pH Treatments on Activity of Bacteriocins
The effect of different treatments on inhibitory activity of bacterioncins of Lb. reuteri 2-20B and Pd.acidilactici 0-11A against S. aureus ATCC 19095 is shown in Table 3. Inhibitory activity of these bacteriocins reduced significantly (p<0.05) with increasing temperature and was completely inactivated at temperatures above 90°C.

Growth and Bacteriocin Production by Lb. reuteri 2-20B and Pd. acidilactici 0-11A
Growth kinetics and bacteriocin production by Lb. reuteri 2-20B and Pd.acidilactici 0-11A were studied in MRS broth at 30°C with initial pH 6.5.Bacteriocin activities were first detected at the exponential growth phase for both Pd.acidilactici 0-11A and Lb.reuteri 2-20B.Bacteriocin production increased with increasing cell concentration, reaching a maximum of 12800 AU/ml after 12 h and 14 h of growth for Pd.acidilactici 0-11A and Lb.reuteri 2-20B, respectively (Figure 1).The highest activities were recorded at the stationary growth phases.Bacteriocins of both Pd.acidilactici 0-11A and Lb.reuteri 2-20B remained active during the stationery growth phase but a reduction of acitivity was observed after 20 h and 22 h for Pd.acidilactici 0-11A and Lb.reuteri 2-20B respectively.Other studies have similarly found production of plantaricin ST31 in detectable amounts during the exponential growth phase, while maximum activities (3200 AU/ml) were observed in the stationary phase (Todorov et al., 1999).Again, BLIS C57 activity was detected after 6 h of incubation with a drastic increase at the stationary growth phase.A loss of activity was, however observed after 72 h of growth (Corsetti et al., 1996).A loss of bacteriocins activity after extended period of incubation has been attributed to factors such as proteolytic degradation, protein aggregation, adsorption to cell surfaces, and feedback regulation (Parente & Ricciardi, 1994;Aasen et al., 2000).Growth in basal medium containing tryptone as sole nitrogen source yielded bacteriocin activities of 6400 AU/ml and 12800 AU/ml for Pd.acidilactici 0-11A and Lb.reuteri 2-20B respectfully.For Pd. acidilactici, the use of yeast extract, meat extract, or a combination of both reduced bacteriocin acitivies levels to 800, 3200 and 3200 AU/ml respectively.Similarly, for Lb.reuteri 2-20B, the use of yeast extract, meat extract, or a combination of both resuted in reduced bacteriocin acitivies levels to 1600, 6400 and 3200 AU/ml respectively (Table 4).Tryptone therefore seemed to be the major driver for optimal bacteriocin production by both Pd.acidilactici 0-11A and Lb.reuteri 2-20B as tryptone alone or in combination with other nitrogen sources yielded the highest bacteriocin acitivity (12800 AU/ml).Different bacteriocin producing LAB may require specific nitrogen cources for optimal activities.Like reported in this paper, tryptone has been a key nitrogen source required for optimal production of bacteriocins ST151BR (Todorov & Dicks, 2004), ST712BZ (Todorov &Dicks, 2007), andplantaracin 432 (Verellen et al., 1998).Different bacteriocin activity levels were observed for both Pd.acidilactici 0-11A and Lb.reuteri 2-20B when different carbohydrates were used as the sole carbon sources (Table 4).Glucose as a carbon source stimulated the highest bacteriocin production (12800 AU/ml) for both bacteria.Whereas mannose stimulated the production of bacteriocin by Pd. acidilactici 0-11A, bacteriocin production by Lb. reuteri 2-20B was reduced to 800AU/ml.Owing to its molecular size, rapid uptake, utilization and cellular energy conversion, glucose has been considered as the usual carbon source for bacterial growth media.However, some bacteria possess abilities owing to their enzymatic activities, to utilize other sugars as carbon sources for growth and for the production of bacteriocins, although bacteriocin activities may vary depending on the carbon source (Todorov & Dicks, 2006).There was a general increase in bacteriocin activities by Pd. acidilactici 0-11A and Lb.reuteri 2-20B with increasing addition of tween 80 (Table 4).In a tween 80 free medium, bacteriocin activity levels of 400 and 800 AU/ml were observed for Pd.acidilactici 0-11A and Lb.reuteri 2-20B respectively.However, the addition of 0.1 and 0.2 ml of tween 80 per liter of medium achieved the highest bacteriocin activity (12800 AU/ml) for Pd.acidilactici 0-11A and Lb.reuteri 2-20B respectively.The influence of tween 80 on bacteriocins activity has been attributed to changes in charges on the cell wall surface of producer strains, thereby preventing adsorption of bacteriocins onto the cells (Mørtvedt-Abilgaard et al., 1995).A reduction in bacteriocin activities were observed with the addition of glycerol (Table 4), an observation which may be due to osmotic stress or reduced water activity.Maximum bacteriocin activities were recorded in the absence of glycerol for both Pd.acidilactici 0-11A and Lb.reuteri 2-20B.

Conclusions
Some strains of LAB isolated from traditional millet-based fermented food (fura) have demonstrated the ability to inhibit the growth of some food-borne pathogens through the production of bacteriocins and therefore could be exploited for their applications in controlling food-borne pathogenic bacteria.Bacteriocins produced by Lb. reuteri 2-20B and Pd.acidilactici 0-11A were partially characterized based on their responses to some physicochemical treatments and were influenced by media composition.Since the effectiveness of bacteriocins are dependent on the chemical and physical properties of foods, they must be fully characterized and tested in appropriate food systems before commercial applications.

Table 1 .
Indicator organisms and their sources/origin

Table 2 .
Antagonistic activities of predominant lactic acid bacteria isolated from fura

Table 3 .
Effect of different treatments on inhibitory activity of cell-free supernatant of Lb. reuteri 2-10B and Pd.acidilactici 0-11A against Staphylococcus aureus enterotoxin A producer