Characterization of bacteriocins produced by strains of Pediococcus pentosaceus isolated from Minas cheese

Abstract

Interest in obtaining bacteriocin-producing strains of lactic acid bacteria (LAB) from different sources has been increasing in recent years due to their multiple applications in health and food industries. This study focused on the isolation and characterization of metabolically active populations of bacteriocinogenic LAB and the evaluation of their antimicrobial substances as well as of some nutritional requirements of them. One hundred and fifty colonies of LAB from artisanal cheeses produced in Minas Gerais state (Brazil) were isolated and screened for their antimicrobial activity. According to their activity against Listeria monocytogenes, ten strains were selected and subsequently identified using biochemical and molecular techniques including 16s rRNA amplification and sequencing as Enterococcus faecalis, Lactobacillus spp., and Pediococcus pentosaceus. Antimicrobial substances produced by four of the selected strains, P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147, were biochemically characterized, and presented sensitivity to proteolytic enzymes (suggesting their proteinaceous nature) and to extreme pH. Antimicrobial activity showed stability after treatment with lipase, catalase, α-amylase, and chemicals. Growth kinetics of the P. pentosaceus selected showed maximal bacteriocin production at 37 °C during the end of the exponential growth phase (25,600 AU/mL) and stable production during 24 h of incubation. Dextrose, maltose, and a mixture of peptone, meat extract, and yeast extract increased bacteriocin production. This study demonstrated that dairy products provide a good alternative for obtaining LAB, with the ability to produce antimicrobial substances such as bacteriocins that have potential use as biopreservatives in food.

Introduction

Milk and dairy products represent important ecological niches that are sources of bacteriocinogenic strains of lactic acid bacteria (LAB) (Furtado et al. 2014). Minas cheese, produced in Brazil (Minas Gerais state), is an artisanal product which is a ripened cheese made mostly from raw cow’s milk. Producers require approximately 60 days to complete maturation of the product and, during this period, a reduction of the most common pathogen population (Martins et al. 2015; Perin et al. 2015) such as Listeria monocytogenes, Salmonella spp., Escherichia coli, and Staphylococcus aureus occurs (Freitas et al. 2013). After this time, cheeses reach quality standards according to Brazilian food production regulations of L. monocytogenes and Salmonella spp. absence and 10³ CFU/g as maximal count of coagulase-positive staphylococci (CPS) (Moraes et al. 2009). The LAB isolated from dairy products belong to genera Lactobacillus, Enterococcus, Pediococcus, and Lactococcus (Luiz et al. 2017) and have as important characteristics the production of organic acids, carbon dioxide, hydrogen peroxide, diacetyl, and bacteriocins (Ammor et al. 2006; Khan et al. 2010). Pediocins are class II bacteriocins produced by Pediococcus strains as a primary metabolite with antimicrobial activity against Listeria monocytogenes. Pediocins are generally small peptides (with 36–48 residues) and non-modified after translation, with some exceptions such as pediocin AcH/PA-1 (Papagianni and Anastasiadou 2009).

Different studies of Pediococcus strains have been focused on their antimicrobial activity. Numerous strains of Pediococcus spp. have been reported to be producers of various bacteriocins, including pediocin PA-1/AcH (P. acidilactici PAC 1.0, P. acidilactici H, E, F, and M), JD (P. acidilactici SJ-1), pediocin 5 (P. acidilactici UL5), pediocin A (P. pentosaceus FBB-61), pediocin N5p (P. pentosaceus), pediocin ST18 (P. pentosaceus), and pediocin PD-1 (P. damnosus) (Anastasiadou et al. 2008). A recent study reported a bacteriocinogenic strain Pediococcus pentosaceus ST65ACC from Minas cheese with activity against two strains of Listeria monocytogenes (Cavicchioli et al. 2017). Application of bacteriocins, such as pediocins from Pediococcus spp. strains, is an alternative means of controlling food-borne pathogenic bacteria and may lead to reduced use of chemical preservatives and the production of healthier food products (Udhayashree et al. 2012).

Knowledge of the optimal production conditions for bacteriocins is important for obtaining maximum activity. Information on inoculation conditions, environmental factors (pH and temperature), and nutritional requirements are key to obtain amounts of bacteriocins that of use in industrial applications (Malheiros et al. 2015). Studies investigating these requirements are necessary because some nutrients can stimulate or limit expression of bacteriocins (Todorov et al. 2012; Abbasiliasi et al. 2017).

In the present study, we report on the isolation, identification, and characterization of LAB with bacteriocinogenic potential from Minas cheese. Based on a preliminary screening, four strains of Pediococcus pentosaceus were selected and a biochemical and molecular characterization of their bacteriocins was conducted. Finally, the effect of modifications to the growth medium on bacteriocin production was studied.

Materials and methods

Isolation of bacteriocin-producing strains

Two different samples of Minas cheese were obtained from a dairy store selling artisanal products in Viçosa (Minas Gerais state, Brazil). Screening for LAB bacteriocin-producing strains was performed as previously described by Todorov et al. (2010). Eleven grams of Minas cheese were homogenized in 99 mL of physiological solution (0.85% NaCl, w/v). Serial dilutions of the homogenized cheese were prepared, plated onto man, rogosa, sharpe (MRS) agar (Difco, BD), and covered with a thin layer of bacteriological agar. Plates were incubated at 37 °C for 24–48 h and total microbial populations were counted. Plates with less than 50 separated colonies were covered with BHI medium containing 1.0% (w/v) agar (Oxoid) and inoculated with a culture of Listeria monocytogenes 104, L. monocytogenes 712, or L. monocytogenes ATCC 7644 (final concentration of 10⁶ CFU/mL). After incubation for an additional 24 h at 37 °C, 150 colonies that presented inhibition zones were selected and cultured in MRS broth (Difco, BD) for 24 h. Bioactivity of the selected strains against L. monocytogenes 104, L. monocytogenes 712, and L. monocytogenes ATCC 7644 was verified using the agar spot-test according to Murua et al. (2013). Briefly, cell-free supernatants (CFS) of isolated LAB were obtained by centrifugation (8000×g at 4 °C for 10 min). The pH of the supernatants was adjusted to 6.0 with sterile 1 N NaOH to eliminate the effect of lactic acid produced by the strains. Potential generation of proteolytic enzymes and H₂O₂ was prevented by heat treatment of CFS (10 min at 80 °C).

Antimicrobial activity was measured using the spot-on-the-law method. Twofold dilutions of the CFS were made in phosphate buffer (100 mM, pH 6.5). Aliquots (10 μL) of each dilution were spotted onto soft BHI agar (1% agar) inoculated with 10⁶ CFU/mL of L. monocytogenes 104. Tests were conducted in three independent repetitions. Antimicrobial activity was expressed as arbitrary units per milliliter (AU/mL) and defined as the reciprocal of the highest dilution showing a clear zone of growth inhibition and calculated according to the equation:

$$ \frac{\mathrm{AU}}{\mathrm{mL}}={a}^b\times 100 $$

where a = 2 (factor dilution) and b = value of the highest dilution showing at least 2-mm inhibition zone (Murua et al. 2013).

Morphology of the studied cultures was examined using Gram staining. Pure cultures were stored at − 20 °C in MRS broth supplemented with 20% (w/v) glycerol.

Differentiation and identification

Basing on preliminary screening for bacteriocin production, conducted with L. monocytogenes 104, L. monocytogenes 712, and L. monocytogenes ATCC 7644, 150 colonies with potential for bacteriocin production were isolated. However, 18 isolates were confirmed to be bacteriocin producers according to the applied agar spot-on-lawn test reported by Murua et al. (2013). Random amplification of polymorphic DNA (RAPD-PCR) analysis was performed in order to obtain differentiation of the selected 18 isolates with primers OPL-04, OPL-05, and OPL-20 (www.operon.com/products/downloads/OperonsRAPD10merSequences.xls). Total DNA from the 18 LAB isolates was extracted using ZR Fungal/Bacterial DNA Kit (Zymo Research, Irvine, CA, USA). The DNA concentration was estimated on a NanoDrop 2000 spectrophotometer (Thermo Scientific Inc., Waltham, MA, USA). Amplification reactions were performed according to Dos Santos et al. (2015). The 25 μL reaction volume contained the following: 2 μL total DNA, 5 μL of 10 mM primer, 2.5 μL of buffer (BioLab), 10 μL of 5 mM MgCl₂ (Fermentas), 1 μL Milli-Q water, 4 μL dNTP (Fermentas), and 0.5 μL Taq DNA polymerase (BioLab). Amplifications were performed on a DNA MasterCycler® (Eppendorf Scientific, Hamburg, Germany) as follows: 45 cycles of 1 min at 94 °C and 1 min at 28 °C, followed by an increase to 72 °C for 2 min. Extension of the amplified product was at 72 °C for 5 min. Amplified products were separated by electrophoresis on 1.2% (w/v) agarose gels in TAE buffer at 120 V for 1 h. Gels were stained with GelRed (Biotium Inc., Hayward, CA, USA). A 100-bp DNA ladder (Fermentas) was used as a molecular weight marker.

Bacteriocin-producing LAB strains were identified according to physiological and biochemical characteristics as previously described by Todorov et al. (2013). Carbohydrate fermentation profiles were recorded using APICHL50 (Biomérieux, Marcy-l′Etoile, France). In addition, molecular identification was confirmed by 16s rRNA sequencing. Total DNA was isolated and quantified as described previously. PCR was performed with primers 8F: 5′-AGTTTGATCCTGGCTCAG-3′ and 1512R: 5′-ACGGCTACCTTGTTACGACTT-3′, according to the method described by Felske et al. (1997). PCR amplification was performed using a DNA MasterCycler® with a 20-μL reaction volume containing 0.1 μL of each primer 10 mM, 2 μL buffer (BioLab), 8 μL of 5 mM MgCl2 (Fermentas), 1.95 μL dNTP (Fermentas), and 0.05 μL Taq DNA polymerase (BioLab). Amplification conditions were as follows: initial denaturation at 94 °C for 5 min, 35 cycles of 5 min at 94 °C, and 10 s at 61 °C, followed by an increase to 72 °C for 2 min. Final extension of the amplified product was at 72 °C for 75 min. The obtained amplicons were purified with a QIAquick PCR Purification Kit (Qiagen), following the manufacturer’s instructions, and submitted to sequencing at the Center for Human Genome Studies, Institute of Biomedical Sciences, University of Sao Paulo, Brazil. The sequences were compared to those deposited in GenBank, using the BLAST algorithm (http://www.ncbi.nlm.nih.gov/BLAST). After identification, four of the isolates were used for analysis.

Screening for the presence of bacteriocin genes

Total DNA was isolated as previously described and amplified by PCR using primers, targeting different bacteriocin genes (nisin, pediocin PA-1, enterocin A, enterocin B, enterocin L50B, enterocin P, plantaricin W, plantaricin S, and plantaricin NC8). PCRs were performed using a DNA MasterCycler® with conditions based on previous studies (Stephens et al. 1998; du Toit et al. 2000; Holo et al. 2001; Maldonado et al. 2003; Kruger et al. 2013; Todorov et al. 2016) and based on the specification of the primers, which are summarized in Table 1. The amplified products were visualized on agarose gel and stained with GelRed.

Table 1 Primers

Effect of enzymes, temperature, pH, and surfactants on bacteriocin activity

Strains were grown in MRS broth for 18 h at 37 °C. Cells were separated by centrifugation (8000 g, 10 min, 4 °C), and the CFS was adjusted to pH 6.0 with 1 M NaOH. One milliliter of CFS was incubated for 1 h at 37 °C in the presence of 1 mg/mL (1%) proteinase K, papain, pepsin, and lipase and 0.1 mg/mL α-amylase and catalase (all from Sigma-Aldrich). In a separate experiment, 1% (w/v) sodium dodecyl sulfate (SDS), Tween 80, Triton X-100, and NaCl (all from Sigma-Aldrich) were added to the CFS; these were also were incubated for 1 h at 37 °C. Untreated CFS and chemicals at their respective concentrations in water were used as controls. Effect of different pH on the activity of bacteriocins was tested by correcting pH of the CFS, prepared as described before, to pH 2.0, 4.0, 6.0, 8.0, and 10 adjusted with sterile 1 M NaOH or 1 M HCl. Samples were incubated for 1 h at 25 °C, and after incubation, they were re-adjusted to pH 6.5 with sterile 1 M NaOH or 1 M HCl. Effect of temperature on the bacteriocin activity was tested by incubating CFS at 4, 25, 30, 37, 45, 60, 80, and 100 °C for 1 h and at 121 °C for 20 min. After each treatment, antimicrobial activity was tested by using the agar spot test method, as previously described and L. monocytogenes 104 was used as the target strain. Results were expressed as percentages of reduction of activity by comparing the diameters of the inhibition zones of treated CFS with untreated CFS (control). Tests were conducted in three independent repetitions.

Adsorption of bacteriocin on producer cells

Determination of the adsorbed bacteriocin onto the surface of the producer cells was performed as previously proposed by Yang et al. (1992). Briefly, after incubation in MRS broth for 18 h at 37 °C, the cultures were adjusted to pH 6.0 with 1 M NaOH and the cells then harvested (10,000 g, 15 min, 4 °C) and washed with 10 mL of sterile phosphate buffer (0.1 M, pH 6.5). The CFS samples were stored for use as controls. The cells were re-suspended in 10 mL 100 mM NaCl (pH 2.0), stirred for 1 h at 4 °C, and then harvested (12,000×g, 15 min, 4 °C). The CFS supernatant obtained was neutralized to pH 7.0 with sterile 1 N NaOH and tested for activity using the agar spot-test (Murua et al. 2013). Tests were conducted in three independent repetitions.

Growth dynamics and bacteriocin production

Growth dynamics and bacteriocin production were evaluated using the turbidity and spot-on-the-law methods, respectively. MRS broth (100 mL) was inoculated with 2% overnight culture and incubated at 37 °C for 24 h. Changes in optical density at 600 nm (OD600) and pH were monitored hourly for 24 h. Antimicrobial activity was measured every 3 h using the spot-on-the-law method (Murua et al. 2013). Twofold dilutions of the CFS were made in phosphate buffer (100 mM, pH 6.5). Aliquots (10 μL) of each dilution were spotted onto soft BHI agar (1% agar) inoculated with 10⁶ UFC/mL of L. monocytogenes 104.

Growth of Listeria monocytogenes 104 in the presence of CFS

One hundred milliliters of BHI was inoculated with 2 mL overnight culture of L. monocytogenes 104 and incubated at 37 °C. After 3 h of incubation, 20 mL aliquots of CFS (pH 6.5) of P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, or P. pentosaceus 147 were filter-sterilized (0.20 mm, Millipore) and added. The incubation was continued. Control without addition of CFS served as a comparison of L. monocytogenes 104 growth. Optical density measurements (600 nm) were recorded at 1-h intervals during the subsequent 12 h according to Todorov et al. (2010). Tests were conducted in three independent repetitions.

Adsorption onto target cell

The adsorption of the bacteriocins produced by P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147 onto L. monocytogenes104, Lb. sakei ATCC 15521, and Enterococcus faecalis ATCC 19443 was measured according to Biscola et al. (2013). The target microorganisms were grown overnight in 10 mL of BHI (for L. monocytogenes 104) and MRS broth (for Lb. sakei and E. faecalis) at 37 °C. Biomass was recovered by centrifugation (8000×g, 15 min, 4 °C). Cells were washed twice with sterile 5 mM phosphate buffer (pH 6.5) and re-suspended in the same buffer to reach an equal to 1.0 of OD600. One milliliter of each cell suspension was mixed with 1 mL of CFS prepared as described before and incubated at 37 °C for 1 h. The antimicrobial activity, using the spot-on-the-law method against L. monocytogenes 104 as previously described, of unbound bacteriocin in the CFS was measured after removal of cells (8000×g, 15 min, 4 °C). Reduction of bacteriocin activity results in adsorption of bacteriocin onto cell surface of target cells and being unavailable for detection in cell-free supernatant. In addition, the effect of pH (4.0, 6.0, 8.0, and 10.0), temperature (4, 25, 30, and 37 °C), and the presence of 1% (w/v) of NaCl, Tween 80, glycerol, and SDS on the adsorption of the bacteriocin was determined (Biscola et al. 2013). The adsorbed bacteriocins were determined as follows:

$$ \%\mathrm{Adsorption}=\left(100-\frac{\mathrm{AU}/\mathrm{m}{\mathrm{L}}_1}{\mathrm{AU}/\mathrm{m}{\mathrm{L}}_0}\right)\times 100 $$

where AU/mL₀ is the bacteriocin activity before treatment, and AU/mL₁ is the bacteriocin activity after treatment. Tests were conducted in three independent repetitions.

Effect of medium composition on the production of bacteriocins

To investigate the effect of nitrogen and carbon sources and also the requirements of micronutrients on the growth and antimicrobial activity of the studied strains, different modified MRS broths were developed. Strains were grown in 10 mL MRS broth at 37 °C for 18 h. Aliquots (100 μL) of the cultures were used to inoculate 10 mL of the following media: (a) MRS broth without organic nutrients, supplemented with peptone (25 g/L), meat extract (25 g/L), and yeast extract (25 g/L) or supplemented with combinations of peptone (12.5 g/L) plus meat extract (12.5 g/L), peptone (15 g/L) plus yeast extract (7.5 g/L), meat extract (15 g/L) plus yeast extract (7.5 g/L), or peptone (10 g/L), meat extract (10 g/L), and yeast extract (5 g/L); (b) MRS broth, replacing the carbon source with fructose, sucrose, lactose, mannose, raffinose, mannitol, or maltose (20 g/L); (c) MRS broth modified to contain 0, 2, 5, or 10 g/L K₂HPO₄; (d) MRS broth modified to contain 0, 0.1, or 0.5 g/L of MgSO₄ and 0, 0.05, or 0.20 g/L of MnSO₄; (e) MRS broth supplemented with 0, 0.5, 1, 2, 5, or 10 g/L glycerol; (f) MRS broth modified to contain 0, 2, or 5 g/L of ammonium citrate; (g) MRS broth modified to contain 0, 1, 2, or 5 g/L of Tween 80; and (h) MRS broth with pH adjusted to 2, 4, 6, 8, 10, or 12. Incubation in all tests was at 37 °C for 24 h. Activity levels of bacteriocins were determined as described before in the “Isolation of bacteriocin-producing strains” section. Tests were conducted in three independent repetitions.

Partial bacteriocin purification and determination of approximate molecular mass by SDS-PAGE

Partial bacteriocin purification was performed according to Martinez et al. (2013), with some modifications. Strains were cultured in 1 L of MRS for 18 h at 37 °C and CFS then obtained by centrifugation for 15 min at 12000×g at 4 °C. Proteins from the CFS were precipitated by 80% saturation with ammonium sulfate at 4 °C (overnight), and the precipitate was then centrifuged for 60 min at 12,000 g at 4 °C. The pellets were resuspendend in 10 mL of 25 mM phosphate buffer (pH 6.5), and antimicrobial activity against L. monocytogenes 104 was determined as described before. In the next step, the resulting material was loaded on an activated SepPakC18 column (Waters, Millipore, MA, USA) and different fractions were eluted using 20, 40, 60, and 80% isopropanol in 25 mM phosphate buffer (pH 6.5). Antimicrobial activity of the obtained fractions was determined as described previously, using L. monocytogenes 104.

SDS-PAGE electrophoresis was performed according to Laemmli (1970), and sample preparation was performed according to Schagger (2006). All examined fractions were loaded in duplicate, and SDS-PAGE electrophoresis was performed at 200 V and 60 mA for first 10 min and then at 200 V and 35 mA. One part of the gel was stained with Coomassie Blue, as described by Schagger (2006), and the second part was used for an overlay assay, according to Cytryńska et al. (2001). Overlay gel was irradiated with UV for 30 min to prevent potential antimicrobial contamination and covered with a soft BHI agar (1%) inoculated with Listeria monocytogenes 104 (approx. 10⁵ CFU/mL) in order to localize the protein bands with antibacterial activity.

Results and discussion

Isolation, differentiation, and identification

One hundred and fifty LAB grown on MRS agar, and that had formed clear inhibition zones against Listeria spp. incorporated in the third agar layer, were isolated from two samples of artisanal cheese produced in Viçosa municipality (Minas Gerais, Brazil). According to the results of additional antimicrobial tests (Murua et al. 2013) using the CFS (pH 6.5) of the 150 isolated colonies, on spot agar test 18 of them (isolates 54, 56, 59, 63, 64, 65, 66, 67, 68, 70, 87, 91, 127, 145, 146, 147, 148, and 149) produced more than 10 mm of inhibition zones using the Listeria spp. strains (the other strains did not produce inhibition after pH correction) and were selected for further analysis. Based on RAPD-PCR performed with 18 selected isolates, 10 presented unique profile and were selected for further studies. From them, seven presented cocci and three rods morphology. Analysis of the 16s rRNA amplified fragments showed that isolates 54, 87, and 91 presented homology with Enterococcus faecalis (Enterococcus faecalis 54, Enterococcus faecalis 87, and Enterococcus faecalis 91), isolates 56 and 127 with Lactobacillus plantarum (Lb. plantarum 56 and Lb. plantarum 127), and isolate 70 with Lactobacillus rhamnosus (Lb. rhamnosus 70). Isolates 63, 145, 146, and 147 presented homology with Pediococcus pentosaceus (P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147). Biochemical characterization of the 10 selected strains was performed using carbohydrate fermentation reactions and was recorded according to the API50CHL® test.

Artisanal cheeses produced in Minas Gerais state (Brazil) are considered to be a cultural heritage and are traditional products made with raw milk and serum collected from cheeses prepared the previous day (Lima et al. 2009). Lima et al. (2009) reported on Lactococcus lactis, Enterococcus spp., Enterococcus faecalis, and Streptococcus agalactiae, isolated from Minas cheese. Another type of Brazilian cheese made with raw milk is the coalho cheese, a traditional product of the North-West region of Brazil. Different species of Lactobacillus spp. such as Lb. acidophilus, Lb. casei, Lb. fermentum, and Lb. rhamnosus and Lactococcus spp. such as Lc. lactis and Lc. raffinolactis were reported to be isolated from this type of cheese (Neto et al. 2005). LAB belonging to the genera Pediococcus have rarely been isolated from dairy products, generally being isolated from meat products. Nevertheless, there are some reports on the occurrence of Pediococcus spp. strains in Minas cheese in Brazil (Cavicchioli et al. 2017; Luiz et al. 2017). Strains of P. acidilactici and P. pentosaceus, isolated from South African farm-style cheese (pasteurized Gouda, young and matured; un-pasteurized aged Bouquet, aged and matured Gouda), were also reported (Gurira and Buys 2005). In another study, strains of P. acidilactici were isolated from traditional Colombian double-cream cheese (non-matured acid cheese), prepared from a mixture of fresh and acidified cow milk. The process of milk acidification of the Colombian cheese as well as maturation of Minas cheese occurs naturally as a result of native microbiota containing LAB, which promotes the organoleptic, physico-chemical, and microbiological characteristics of the finished product (Londoño-Zapata et al. 2017).

Table 2 shows the bioactivity of the 10 identified strains against Listeria monocytogenes 104 using the agar spot-test. The most active were the Pediococcus strains; Lactobacillus presented the lowest activity, while the Enterococcus presented intermediate activity. Similar results have been reported for these genera. Two strains of Pediococcus acidilactici HA-6111-2 and HA-5692-3 were isolated from alheira and showed 1600 AU/mL of antimicrobial activity against Listeria innocua N27 (Albano et al. 2007). Another study reported higher activity (6400 AU/mL) for the previously mentioned P. acidilactici HA-6111-2 under high pressure (Castro et al. 2015). Cavicchioli et al. (2017) isolated Enterococcus hirae ST57ACC and P. pentosaceus ST65ACC from Minas cheese; these two bacteriocinogenic strains showed antimicrobial activity against 101 different strains of Listeria spp., 8 Enterococcus spp., 9 Lactobacillus spp., 1 Leuconostoc spp., 2 Pediococcus spp., and 2 Streptococcus spp. In another study, P. pentosaceus FBBGl (ATCC 43200) presented antimicrobial activity of 3200 AU/mL (Piva and Headon 1994). Pediococcus strains isolated in this study presented activity of 51,200 AU/mL, recorded against L. monocytogenes 104. E. faecalis 54, E. faecalis 87, and E. faecalis 91 showed activity of 3200 AU/mL, recorded against L. monocytogenes 104. Activity of E. faecium SD1, SD2, SD3, and SD4 strains, isolated from goat’s milk, was reported as 51,200 AU/mL for strains SD1 and SD2, 3200 AU/mL for SD3, and 800 AU/mL for SD4 (Schirru et al. 2012). Casaburi et al. (2016) described activity of 6400 AU/mL for Lactobacillus curvatus 54 M16, isolated from traditional fermented sausages of Campania region (Italy). In the present study, activities of 200, 800, and 3200 AU/mL were reported for Lb. plantarum 56, Lb. rhamnosus 70, and Lb. plantarum 127, respectively. Similar results of 800 AU/mL were reported for Lb. rhamnosus EM253 (dos Santos et al. 2015) and less than 800 AU/mL for Lb. plantarum HKN01 isolated from dairy products (Sharafi et al. 2013). However, these levels of activity may be unreliable, since bacteriocin activity depends on the specificity of the expressed antibacterial protein and on the specific characteristics of the microorganisms investigated. The optimal scenario would be if the same test microorganisms were used in all studies, which would facilitate comparison of the investigated bacteriocins.

Table 2 Antimicrobial activity (AU/mL) of isolates recorded against L. monocytogenes 104

Screening for the presence of bacteriocin genes in total DNA

When total DNA was screened for presence of genes related to bacteriocin production, positive results were only generated for the presence of pediocin PA-1 gene in DNA obtained from P. pentosaceus 63, 145, 146, and 147 strains. There was no evidence of the presence of genes related to nisin, enterocin A, enterocin B, enterocin L50B, enterocin P, plantaricin NC8, plantaricin S, or plantaricin W. Figure 1 shows the bands obtained with the Pediococcus strains using the primer to amplify the gene of PA-1 (1044 bp).

Fig. 1

Amplification of total DNA from Pediococcus strains using a primer of PA-1 gen. P63 P. pentosaceus 63, P145 P. pentosaceus 145, P146 P. pentosaceus 146, P147 P. pentosaceus 147

P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147 harbor a 1044 bp fragment corresponding to that reported for pediocin PA-1 (Fig. 1). The size of the obtained amplicon was consistent with that reported for pediocin PA-1 by Marugg et al. (1992). Pediocin PA-1 biosynthesis involves a DNA fragment of approximately 3.5 kb with the presence of four genes pedA, pedB, pedC, and pedD (Marugg et al. 1992). However, amplicon sequencing can confirm the fact that P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147 studied are producers of pediocin PA-1.

Effect of enzymes, temperature, pH, and surfactants on bacteriocin activity

All tests were performed with CFS from each strain in MRS broth incubated at 37 °C for 24 h and pH was corrected (pH 6.5) each time. Table 3 shows percentage reduction of activity for each isolate. CFS from P. pentosaceus 63 lost at least 50% of its activity by treatment with proteinase K, and pepsin. Papain produced a reduction almost of the 100% and lipase 74%. An antimicrobial activity reduction of 95% of CFS from P. pentosaceus 145 was produced by proteinase K and papain, less reduction was obtained with pepsin (63%). Lipase, catalase and α-amylase produced a reduction almost of the 50%. P. pentosaceus 146 and P. pentosaceus 147 lost almost 100% of their activity with proteinase K and pepsin and the rest of the enzymes caused 50% or least reduction. The effect of α-amylase was very low for all isolates. The CFS of each strain presented a small partial loss of activity at 25, 30, and 37 °C, remaining active after 1 h at 60, 80, and 100 °C, also with the treatment at 121 °C for 20 min. This heat tolerance, characteristic of the bacteriocins, obeys to their small size and makes them a good option as biopreservatives in foods ((Karumathil et al. 2016; Parada et al. 2007). Similar results have been reported by different authors (Todorov and Dicks 2005a; Albano et al. 2007; Murua et al. 2013; Seo et al. 2014). Low pH, such as 2.0 and 4.0, had little effect on antimicrobial activity, as did pH 8.0. Ghanbari et al. (2013) reported this tolerance to low pH with bacteriocins produced by Lb. casei AP8 and Lb. plantarum H5, isolated from the intestinal bacterial flora of beluga (Huso huso) and Persian sturgeon (Acipenser persicus); inactivation at pH 10.0 was reported to be due to proteolytic degradation, protein aggregation, and instability of proteins.

Table 3 Percentages of reduction of activity after different treatments

Treatment with Triton X-100, Tween 80, SDS, NaCl, or skimmed milk had no significant effect on the antimicrobial activity. Sharafi et al. (2013) reported the lack of effect of these treatments in bacteriocins from Lb. plantarum HKN01 isolated from Iranian traditional dairy products. Todorov and Dicks (2005a, b) reported that pediocin ST18 produced by Pediococcus pentosaceus ST18, isolated from boza (a cereal-fermented non-alcoholic beverage from Bulgaria), was not sensitive to SDS, Tween 20, Tween80, urea, N-lauroylsarcosine, or Triton X-100. The effect of different chemicals, pH, and temperature is dependent on the specific structure and amino acid sequence of the bacteriocins studied. Moreover, these results may have a practical application in subsequent experiments, including in their planning, and investigations of bacteriocin use in food biopreservation.

Adsorption on the cell surface of producer cells

Secretion of the bacteriocins normally is performed via ABC transporter system or sec-dependent (Cintas et al. 2000; Kumar et al. 2011). Yang et al. (1992) showed that some bacteriocins can be secreted and then be adsorbed onto the cell surface of the producer cells. This adsorption could be a result of some affinity or because of charge-specific interaction. High levels of adsorbed bacteriocins on the cell surface of producer cells could be considered an opportunity to facilitate the purification process of produced bacteriocins, and this was applied by Yang et al. (1992). However, in the case of the bacteriocins studied here, only low levels were found to be adsorbed on the surface of P. pentosaceus 63, P. pentosaceus 147, P. pentosaceus 146, and P. pentosaceus 147 (Fig. 2). This was found to be the case for most of the investigated bacteriocins. For instance, similar results were reported for bacteriocins produced by Lactococcus lactis subsp. lactis B14 isolated from boza (Ivanova et al. 2000) and for bacteriocin bacST8KF produced by L. plantarum ST8KF isolated from kefir (Powell et al. 2007). Two bacteriocins from Lb. curvatus and Lb. sakei, isolated from salpicao, a traditional fermented pork sausage produced in Portugal, also presented low levels of bacteriocin adsorption onto the cell surface of producer cells (Todorov et al. 2013).

Fig. 2

Adsorption of the bacteriocins produced to the own surfaces of the studied strains of Pediococcus. Antimicrobial activity (AU/mL) of the isolates. Light gray column: recovered from cell surface (after desadsorption) and black column: in cell-free supernatant. Titles presented reduction along the development of the study. All data represent an average of three repeats. The values recorded in each experiment did not vary by more than 5%, and single data points are presented in the figures without standard deviation bars

Growth dynamics and bacteriocin production

Figure 3 shows the relationship between bacterial growth of the selected strains and produced bacteriocin with activity against L. monocytogenes 104 during a 24-h period of culture in MRS broth at 37 °C. P. pentosaceus 63 reached its stationary phase at 15 h with OD600 of 4.82. Antimicrobial activity was reported early in the exponential growth phase (3 h), with bacteriocin levels of 1600 AU/mL and at 6 h increased to 3200 AU/mL. Maximum activity was recorded after 9 h of incubation (12,800 AU/mL) and remained stable until 24 h, with an OD600 of 4.692 (Fig. 3a). Antimicrobial activity of bacteriocin produced by P. pentosaceus 145 started during the exponential phase, with 1600 AU/mL and OD600 of 0.298 at 3 h; maximum activity (25,600 AU/mL) was after the beginning of the stationary phase after 12 h of incubation, with OD600 of 3.368; at 24 h, OD600 was 3.49 (Fig. 3b). Similar dynamics were observed with P. pentosaceus 146, presenting 25,600 AU/mL at 24 h and OD600 of 3.588 (Fig. 3c). P. pentosaceus 147 reached maximum activity in the middle of the exponential phase with OD600 of 1.13 and continued until 24 h with OD600 of 3.58 (Fig. 3d). The results are according to other studies that report optimal production on stationary phase, for example, of bacteriocins EM485 and EM925 (produced by E. faecium EM485 and E. faecium EM925 isolated from Brazilian cheese) (dos Santos et al. 2014) and bacteriocins produced by E. faecium ET05, ET12, and ET88 isolated from smoked salmon that were produced during stationary growth (Tomé et al. 2009). Maximal production occurs during the stationary phase, which suggests that bacteriocins are secondary metabolites, according to other studies (Albano et al. 2007). Another study reported pediocin PD-1 production by P. damnosus NCFB 1832 during logarithmic phase (1600 AU/mL) and an increment during the stationary phase (Nel et al. 2001). P. acidilactici P9, isolated from pickles, started production at 8 h and, during the stationary phase (after 16 h of incubation), reached maximum production, and remained constant until 24 h of incubation (Wang et al. 2014). For strains of Lactobacillus spp., production was reported during the logarithmic phase of growth, as in the case of Lb. plantarum ST71KS, with maximum production (6400 AU/mL) during the stationary phase (Martinez et al. 2013), similar results were reported for Lb. curvatus 54 M16 (Casaburi et al. 2016).

Fig. 3

Growth dynamics. a P. pentosaceus 63. b P. pentosaceus 145. c P. pentosaceus 146. d P. pentosaceus 147. Optical density (OD600) (filled diamond); pH (filled square). Bars represent antimicrobial activity (kAU/mL) (1 kAU/mL = 1000 AU/mL)

Growth of Listeria monocytogenes 104 in the presence of CFS

Visualization of the effect of bacteriocin containing CFS on actively growing L. monocytogenes 104 is presented in Fig. 4. After 3 h of incubation of L. monocytogenes 104, values of OD600 reached an average of 0.64. The addition of bacteriocin containing CFS of P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147 to the L. monocytogenes 104 actively growing in culture resulted in growth inhibition after 1 h (hour 4, Fig. 4) with P. pentosaceus 63 seeing almost no change of OD600 from 0.511 (hour 3) to 0.585 (hour 4), the same happened with P. pentosaceus 146 (OD600 from 0.574 to 0.594) and P. pentosaceus 147 (OD600 from 0.749 to 0.620). P. pentosaceus 145 allowed initial growth of L. monocytogenes 104, and the values of OD600 were very close to the control (without CFS). However, 2 h after the addition of CFS, growth was limited and similar OD600 values were observed in all cases and remained at this level. The control reached a maximum OD600 after 8 h of incubation (6.128); this decreased to 4.296 at the end of the test. The results suggest that CFS of isolates can inhibit growing cultures of L. monocytogenes 104. Similar results have been reported for P. pentosaceus ST65ACC against L. monocytogenes 211 and L. monocytogenes 422 (Cavicchioli et al. 2017) and for L. casei AP8 isolated from sturgeon fish against L. monocytogenes ATCC 19115 (Ghanbari et al. 2013).

Fig. 4

Growth kinetics of L. monocytogenes 104 on BHI with added CFS of the studied P. pentosaceus strains. Optical density (at 600 nm) measurements of the medium with the following: circle: P. pentosaceus 63, diamond: P. pentosaceus 145, triangle: P. pentosaceus 146, square: P. pentosaceus 147, and asterisk: control without CFS. All data represent an average of three repeats. The values recorded in each experiment did not vary by more than 5%, and single data points are presented in the figures without standard deviation bars

Adsorption onto target cell

The aim of this test is to find how much bacteriocin was able to bind to the target cell surface comparing antimicrobial activity before and after contact of CFS with target cells. Bacteriocin adsorption is considered as first step for bacteriocin mode of action. This information about potential efficacy of the bacteriocin and its ability to bind on the surface is important for the technological applications of bacteriocins exploration. CFS from Pediococcus strains were incubated with Lb. sakei and E. faecalis during 1 h on 5 mM phosphate buffer, a short time and poor nutritional conditions that did not allow the target strains to produce bacteriocins to interfere with the test. Figure 5 shows the effect of different conditions on the adsorption of bacteriocins onto L. monocytogenes 104, E. faecalis ATCC 19443, and Lb. sakei ATCC 15521. Under natural conditions (pH 6.5 and 25 °C), the highest adsorption for P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147 was with L. monocytogenes 104 (98.4, 96.9, 96.9, and 98.4%, respectively). Adsorption onto E. faecalis ATCC 19443 surface was 93.8% for all isolates, except P. pentosaceus 145 which presented a lower value (87.5%). P. pentosaceus 63 showed 96.9% of adsorption to Lb. sakei ATCC 15521, and P. pentosaceus 145 and P. pentosaceus 146 presented 93.8%. The lowest adsorption value was for P. pentosaceus 147 with 70%.

Fig. 5

Percentage of adsorption of bacteriocins onto target cells under different treatments of the CFS. a L. monocytogenes 104. b E. faecalis ATCC 19443. c Lb. sakei ATCC 15521. □: P. pentosaceus 63, : P. pentosaceus 145; : P. pentosaceus 146, ≡: P. pentosaceus 147

Very low influence of temperature over adsorption of bacteriocins was observed in tests with L. monocytogenes 104. An increase in adsorption of P. pentosaceus 63 and P. pentosaceus 146 at 37 °C onto E. faecalis ATCC 19443 and a reduction at 4 °C were observed. For Lb. sakei ATCC 15521, P. pentosaceus 63, P. pentosaceus 145, and P. pentosaceus 146, the lowest adsorption was at 37 °C and 4 °C, and for P. pentosaceus 147, the lowest adsorption was at 25 °C. Low pH affected the adsorption of all isolates onto L. monocytogenes 104. The same effect occurred with E. faecalis ATCC 19443 and, at pH 10.0, adsorption also decreased. Adsorption onto Lb. sakei ATCC 15521 increased at low pH with P. pentosaceus 63 and P. pentosaceus 146 and decreased with P. pentosaceus 147. P. pentosaceus 146 had decreased adsorption onto Lb. sakei ATCC 15521 at pH 8.0. Percentage of adsorption of all isolates decreased in the presence of chemicals. SDS was the chemical that most affected adsorption onto target cells, especially onto L. monocytogenes 104. Glycerol only affected adsorption onto L. monocytogenes 104. Adsorption onto E. faecium ATCC 19443 was affected by all chemicals, except glycerol with P. pentosaceus 63 and P. pentosaceus 145.

It is important to note that different conditions common in the food industry can affect the ability of bacteriocins to bind to the microorganism surface; nevertheless, the results reported demonstrate the high affinity of bacteriocins for target cells and indicate that bacteriocins have a potential use in industry for controlling growth of microorganisms because they showed that bacteriocins continue active. Other studies have also investigated the effect of pH, temperature, and chemical agents and concluded that effect of temperature is minimal, similar to pH with values close to neutrality, and that chemicals may affect adsorption the most (Biscola et al. 2013; Furtado et al. 2014).

Effect of medium composition on the production of bacteriocins

Table 4 shows the results of bacteriocin production, expressed in arbitrary units per milliliter, with each modified MRS broth. P. pentosaceus 63 produced the maximum antimicrobial activity using dextrose and maltose as carbon sources (12,800 and 25,600 AU/mL, respectively); with raffinose and mannitol, production was minimal. Antimicrobial activity of 6400 AU/mL was obtained using peptone, meat extract, and yeast extract; in combination, these generated 128,000 AU/mL. Without K₂HPO₄ or with K₂HPO₄ at more than 2 g/L, production decreased to 6400 AU/mL. The absence of MnSO₄ or MnSO₄ at more than 0.05 g/L also caused decreased bacteriocin activity by P. pentosaceus 63. The same was observed with different concentrations of sodium acetate. The absence of MgSO₄ and glycerol had no effect, and bacteriocin production was 12,800 AU/mL. High amounts of Tween 80 had no effect on production, but its absence caused a decrease. The absence of ammonium citrate had no effect on bacteriocin production, but a high amount increased production. Extreme pH decreased the production of antimicrobial compound by P. pentosaceus 63.

Table 4 Antimicrobial activity (AU/mL) on different modified MRS

P. pentosaceus 145 also produced the maximum of antimicrobial activity (12,800 AU/mL) using dextrose and maltose as carbon source. Maximum production (12,800 AU/mL) was obtained with a mixture of yeast (15 g/L) and meat extract (7.5 g/L) or a mixture of peptone (10 g/L), meat extract (10 g/L), and yeast extract (5 g/L) as nitrogen sources. The amount of K₂HPO₄ and MnSO₄ had no effect on production and changes in sodium acetate, ammonium citrate, and Tween 80 decreased production. Only pH 6.0 of the range of pH values tested registered antimicrobial activity (6400 AU/mL). The absence of glycerol increased production. P. pentosaceus 146, in addition to a preference for dextrose and maltose, exhibited antimicrobial activity of 12,800 AU/mL with fructose as the carbon source. The use of peptone alone or meat and yeast mixture caused decrease in production until (6400 AU/mL). Increasing amounts of K₂HPO₄ or absence of K₂HPO₄ caused decrease in antimicrobial activity; similar effects occurred with 0.05 g/mL of MgSO₄ or its absence. MnSO₄ in 0.05 g/L favored bacteriocin production and the opposite occurred with sodium acetate. The absence of Tween 80 decreased antimicrobial activity and the same occurred with extreme pH (2.0 or 12). P. pentosaceus 147 had no activity when sucrose, mannitol, or raffinose was used as the carbon source, and a mixture of peptone (10 g/L), meat extract (10 g/L), and yeast extract (5 g/L) generated maximum activity. The absence of K₂HPO₄ or ammonium citrate or more than 2 g/L of each of these chemicals caused decreased antimicrobial activity; the same occurred without MgSO₄ or with MgSO₄ at more than 1 g/L and without MnSO₄ or with MnSO₄ at more than 0.05 g/L. Different amounts of sodium acetate generated the same activity (6400 AU/mL). Less or more than 1 g/mL of Tween 80 added to the broth caused a decrease in activity, as well as extreme pH values.

In all cases, extremely, pH limited bacterial growth generating very low or no antimicrobial activity. Similar results have been reported in a study of optimization of bacteriocin ST22Ch production by Lb. sakei isolated from salpicao in which glucose, as the carbon source, was found to promote production of the antimicrobial substance. The same study reported that a combination of different sources of nitrogen (meat and yeast extract or tryptone and meat extract) stimulated production. The same happened with high concentrations of MgSO₄ and Tween 80. The absence of MgSO₄ decreased production, and the presence of glycerol had no effect (Todorov et al. 2012). Another study reported that optimal production of P. acidilactici LAB5 isolated from a fermented meat product was obtained with a mixture of tryptone, yeast extract as a nitrogen source, glucose as a carbon source, and a buffer composed of sodium citrate, sodium acetate, and K₂HPO₄ (0.2 g/L of each) (Mandal et al. 2008). Suganthi and Mohanasrinivasan (2015) used a process of optimization to obtain maximal production (25,600 AU/mL) of the bacteriocin from P. pentosaceus KC692718, isolated from mixed vegetable pickles (India), using sucrose (24 g/L) as a carbon source and soyatone (10.3 g/L) as a nitrogen source. Kaur et al. (2013) enhanced pediocin BA28 production by P. acidilactici using peptone (10 g/L), beef extract (10 g/L), meat extract (10 g/L), tryptone (10 g/L), KH2PO4 (2 g/L), potassium sodium tartrate (2 g/L), dextrose (50 g/L), and Tween 80 0.1 g/L.

Partial bacteriocin purification and determination of approximate molecular mass by SDS-PAGE

Precipitation with 80% ammonium sulfate saturation was successful in obtaining all antimicrobial proteins produced by the investigated strains. However, when proteins were separated using SepPack chromatography, almost all fractions presented activity against L. monocytogenes 104. Nevertheless, the most active of the isopropanol-eluted fractions was with 60% isopropanol presenting activities of 25,600, 12,800, 5600, and 25,600 AU/mL, respectively, for P. pentosaceus 63, P. pentosaceus 145, P. pentosaceus 146, and P. pentosaceus 147. Miteva et al. (1998) reported a difference with this study with activity of 50% fractions obtained with a strain of Lactobacillus spp. 1043 against Gram-positive and Gram-negative indicator strains.

The results presented in Fig. 6, representing the Tricine-SDS-PAGE gel, indicate that the approximate molecular weight of the bacteriocins studied was between 3.5 and 6.5 kDa. The antimicrobial activity was confirmed by inhibition zones against both L. monocytogenes 104 in the same place as the proteins bands. Similar weights of peptides were reported for bacteriocin PA-1 produced by P. pentosaceus NCDC 273 (Vijay Simha et al. 2012); for pediocin ST71KS produced by Lb. plantarum ST71KS, isolated from homemade goat feta cheese (Martinez et al. 2013); for pediocin ST44AM produced by P. pentosaceus ST44AM (Todorov and Dicks 2009); and for bacteriocins BacHA-6111-2 and bacHA-5692-3 produced by strains of P. acidilactici (Albano et al. 2007).

Fig. 6

Separation of the proteins obtained after precipitation by ammonium sulfate and separation by SepPack and subjected to SDS-PAGE electrophoresis. Stained electrophoresis gel (left) and inhibition zone observed using L. monocytogenes 104 as indicator strain with the non-stained electrophoresis gel (right)

Conclusions

LAB isolated from dairy products are a good alternative for obtaining antimicrobial substances such as bacteriocins. LAB that occur naturally in dairy products generally belong to species with well-proven GRAS status. However, additional research is required to confirm safety aspects of isolated LAB in order to recommend their application or their expressed bacteriocins as non-hazardous agents in food production. Although bacteriocins are recognized to be non-toxic proteinaceous molecules, their safety needs to be carefully examined prior to their use as food additives or therapeutic agents. Biochemical characteristics of bacteriocins allow better design for their possible application in the food industry. Pediocins have been reported to be a good option for food biopreservation, instead of conventional treatments used to preserve food products (Papagianni and Anastasiadou 2009). In our study, strains isolated from Minas cheese presented remarkable antimicrobial activity against three L. monocytogenes strains from different serological groups. Based on the specific characteristics of the bacteriocins studied, produced by four P. pentosaceus strains, it is necessary to be conducted a future research in order to explore the possibilities of the application of the strains as protector cultures or the expressed bacteriocins in the control of food spoilage in fermented food products.