Antibacterial and antifungal activity of crude and freeze-dried bacteriocin-like inhibitory substance produced by Pediococcus pentosaceus

Pediococcus pentosaceus LBM 18 has shown potential as producer of an antibacterial and antifungal bacteriocin-like inhibitory substance (BLIS). BLIS inhibited the growth of spoilage bacteria belonging to Lactobacillus, Enterococcus and Listeria genera with higher activity than Nisaplin used as control. It gave rise to inhibition halos with diameters from 9.70 to 20.00 mm, with Lactobacillus sakei being the most sensitive strain (13.50–20.00 mm). It also effectively suppressed the growth of fungi isolated from corn grain silage for up to 25 days and impaired morphology of colonies by likely affecting fungal membranes. These results point out that P. pentosaceus BLIS may be used as a new promising alternative to conventional antibacterial and antifungal substances, with potential applications in agriculture and food industry as a natural bio-controlling agent. Moreover, cytotoxicity and cell death induction tests demonstrated cytotoxicity and toxicity of BLIS to human colon adenocarcinoma Caco-2cells but not to peripheral blood mononuclear cells, with suggests possible applications of BLIS also in medical-pharmaceutical applications.


cultivations for BLiS production
BLIS was produced in static P. pentosaceus cultivations carried out at 30 °C for 10 h in 500-mL Erlenmeyer flasks containing 300 mL of MRS medium placed in an incubator. BLIS-containing medium was separated from Pediococcus biomass by centrifugation (4,470 × g at 4 °C for 20 min), and the supernatant pH adjusted to 6.0-6.5 by addition of 1.0 N NaOH for use in the analyses. Crude BLIS (CB), i.e. BLIS without any purification, was tested for its activity against the above indicator strains after several dilutions with sterile deionized water, namely 1:2, 1:5, 1:10, 1:50 and 1:100 (v/v). Freeze-dried crude BLIS (FB) was tested at three different concentrations, namely 1.0, 2.5 and 5.0% (w/v), after its dilution with sterile deionized water.
Nisaplin (DuPont Danisco, Copenhagen, Denmark), having nisin in its formulation as an active compound at 2.5% (w/w), was also diluted with sterile deionized water up to the same concentrations as FB (1.0, 2.5 and 5.0%, w/v) and used for comparison with BLIS antimicrobial power.

Determination of BLiS antibacterial activity
The agar-well diffusion method was performed to evaluate the antimicrobial activity of CB, FB and Nisin against L. sakei. For this purpose, after overnight bacterial growth at 37 °C, 150 µL of each microbial suspension with 0.3 optical density at 600 nm (OD 600nm ), corresponding to 8 × 10 6 CFU/mL for L. sakei and 3 × 10 6 CFU/mL for E. faecium, L. innocua and L. seeligeri, were 1:100 (v/v) diluted with sterile deionized water, added to 15 mL of MRS soft agar-medium (0.75%, w/v) and poured into Petri dishes. After agar-medium solidification, 50 μL of BLIS or Nisin were added into each well, plates were incubated at 37 °C for 16-18 h, and the antagonistic activity was determined by measuring the diameter of the inhibition halos in millimeters 43 .
The antagonistic activity of CB without dilution or diluted several times [from 1:2 to 1:100 (v/v)] was also evaluated by counting colonies according to the pour plate methodology. All the assays were performed in triplicate.

Determination of BLiS antifungal activity
Fungi isolated from Austrian corn grain silage 41 were grown in PEG medium in an incubator at 30 °C for 3-5 days. After incubation, CB was applied directly on the different mycelia, and the seven fungi that proved to be sensitive to CB were isolated from the plates, named from F1 to F7 and grown separately. Sensitive fungi were evaluated by the pour plate technique, after addition of 1.0 mL of sample, consisting of 500 µL of fungus suspension and Scientific RepoRtS | (2020) 10:12291 | https://doi.org/10.1038/s41598-020-68922-2 www.nature.com/scientificreports/ 500 µL of CB (or 500 µL of sterile deionized water in the control), to 10 mL of Yeast Extract Glucose (YEG) agar medium (Merck Millipore, Darmstadt, Germany), and incubation at 30 °C for 24 h, 7 days, 15 days and 25 days. After these periods, mycelial growth in both treated and control Petri dishes were visually examined.

physical characterization of BLiS
Fourier Transform Infrared Spectroscopy (FT-IR) (Frontier-PerkinElmer, Massachusetts, USA) was used for a partial physical characterization of BLIS, since it allows qualitatively identifying unknown functional chemical groups 43,44 in any different kind of material. For this purpose, 1.0 mg of lyophilized BLIS was milled with 100 mg of KBr at a pressure of 7,500 kg cm -2 per 30 s, and the resulting translucent disks were analyzed by FT-IR 45 in the 4,000-400 cm -1 wavenumber range with resolution of 4 cm -1 . CB dissolved in deuterated water was also analyzed by 1 H Nuclear Magnetic Resonance (NMR) with a Bruker AIII 500 spectrometer (Bruker, Rheinstetten, Germany) at 500 MHz, to find out different amino acids that form the molecular structure of BLIS 46 . The 1 H chemical shifts were expressed in ppm with respect to the standard solvent displacement.

Scanning electron microscopy
To examine P. pentosaceus cells, Scanning Electron Microscopy (SEM) was used as previously described 47 . For this purpose, samples were prepared as follows. After cultivation, the culture medium was removed by centrifugation at 5,000 rpm and 4 ºC for 20 min. In order to remove contaminants from the culture medium and free biomass from the viscous matrix formed during fermentation, the pellet obtained was redispersed in 100 mL of deionized water and subjected to several cycles of centrifugation under the same conditions described above and washing with an equal water volume, until a milky color liquid was obtained. The viscous matrix was then redispersed in 50 mL of deionized water and cryogenically frozen by immersion and cooling to a temperature of − 196 ºC using liquid nitrogen. The sample water was subsequently removed by lyophilization, and the resulting white solid with flocculated consistency was preserved in a desiccator at room temperature for further analysis. The samples were mounted on metal stubs with a conductive paste and coated with gold or gold-palladium in an argon-ion atmosphere, and examined in a Scanning Electron Microscope, model Neoscope JCM-5000 (JEOL, Peabody, MA, USA).

cytotoxicity of BLiS to human cells
To assess the cytotoxicity of BLIS to human cells, peripheral blood mononuclear cells (PBMC) and human colon adenocarcinoma cells (Caco-2) were used 48 . In 96-well plates, 2.5 × 10 5 cells per well were plated in 100 µL of culture medium [RPMI (Gibco, Grand Island, NY, USA) for PBMC cells or DMEN (Gibco) for Caco-2 cells] supplemented with 10% fetal bovine serum. Plates were incubated in a 5% CO 2 oven at 37 °C for 24 h to allow cell adhesion. After 24 h of incubation, crude BLIS without any dilution was added to the wells, and plates were incubated again for 48 h. The control was prepared in the same way but without adding the BLIS. Each test was performed in triplicate. After incubation, the content of each well was removed, and 100 µL of 20 µg/ml 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyl tetrazoline bromide (MTT) were added. Plates were incubated for 3 h under the same conditions mentioned above. The MTT solution was then removed from the wells, and 200 µL of dimethyl sulfoxide (DMSO) were added. After smooth homogenization of plate content to dissolve formazan crystals, the absorbance was read at a wavelength of 570 nm.

cell death induction assay
After cultivation, removal of the culture medium and treatment with BLIS as described in the previous section, cells were stained with 100 µL of 50 µg/mL acridine orange solution and 50 µg/mL propidium iodide prepared in phosphate saline buffer (PBS). After 5 min of incubation, acquisition of images was done using a fluorescence microscope (Axiovert 200, Zeiss, Oberkochen, Germany). According to this methodology, viable cells stained green, due to the presence of acridine orange, while dead cells stained red because of propidium iodide.

Statistical analysis
The results, expressed as mean values ± standard deviations, were submitted to one-way analysis of variance (ANOVA) by the Statistica Software 12 (TIBCO Software Inc., Palo Alto, CA, USA), compared using the Tukey's post-hoc test and considered significantly different when p < 0.05.

Results and discussion
Antibacterial activity. The ability of Pediococcus pentosaceus LBM 18 to produce bacteriocin-like inhibitory substance (BLIS) and its effectiveness as an antimicrobial agent were first evaluated by the agar well diffusion assay. As shown in Table 1, all the indicator strains were sensitive to BLIS, exhibiting large diameters of the inhibition halos (9.70 to 20.00 mm), with Lactobacillus sakei being the most sensitive one (13.50 to 20.00 mm), regardless of BLIS dilution. Even though BLIS was used in its crude form (CB), it showed higher antimicrobial activity (larger inhibition halos) against all indicator strains compared to 1% (w/v) nisin-based Nisaplin solution. BLIS was shown to be more effective when diluted in water, especially at 1:5 and 1:10 (v/v) dilutions (Table 1), likely because of better diffusion through the bacterial cell membrane. At the highest dilutions (1:50 and 1:100), BLIS still exerted a weak antibacterial effect against all strains, with significantly (p < 0.05) narrower inhibition halos (10.00 to 14.50 mm).
BLIS antibacterial effect was also checked by counting colonies according to the pour plate method, whose results are listed in Table 1. Both CB and BLIS at any dilutions were as effective as Nisaplin in suppressing L. sakei and E. faecium growth (logCFU/mL < 4), irrespective of dilution. Listeria strains proved to be less sensitive, in that L. innocua growth was completely suppressed (logCFU/mL < 4) only by CB and 1:2 (v/v) water-diluted BLIS and L. seeligeri by 1:2 and 1:5 (v/v) water-diluted BLIS. The strain-dependent efficiency of CB or water-diluted BLIS may be explained by interaction between BLIS and plasma membrane and consequent formation of pores, which is a common occurrence among most LAB bacteriocins. BLIS may have also interacted with receptors or docking molecules, in a specific way for a specific strain depending on its membrane lipid composition and membrane potential threshold 49 , abrogating essential cell pathways which account for BLIS potency 50 . At the other BLIS dilutions (1:10, 1:50 and 1:100, v/v), L. innocua showed growth (logCFU/mL = 8.22 ± 0.00, 8.11 ± 0.29, 8.11 ± 0.31, respectively) similar to that in the control (logCFU/mL = 8.25 ± 0.06). Also, L. seeligeri, when exposed   54 , which suggests the use of this microorganism to produce novel antimicrobials. The efficiency of BLIS as an antibacterial was also tested after freeze-drying (Table 2) to check whether its antimicrobial activity would be impaired by this operation. Regardless of BLIS concentration (1.0, 2.5 or 5.0%, w/v), its ability to reduce L. sakei counts (logCFU/mL) was in the range 0.5-0.6, which suggests the existence of a threshold amount for bacteriocin to exert a bacteriostatic or bactericidal effect against closely related species like L. sakei, while Nisaplin was much more effective, being able to completely suppress it (logCFU/mL < 4) regardless of its concentration (1.0, 2.5 and 5.0%, w/v) or dilution.
In summary, the results of this part of the study clearly show that P. pentosaceus LBM 18 BLIS exerted higher antimicrobial effect when directly used after its production rather than after freeze-drying. Moreover, BLIS proved its antimicrobial action even without any purification, showing potential as a novel biopreservative alternative to nisin-based preparations.

Antifungal activity of BLiS
LAB were shown to produce, during growth, antifungals, inhibitors of mycotoxin synthesis and mycotoxin detoxifying agents, which are in great demand by the agricultural and food sectors 9 . It has been proven their ability to produce fatty acids 55 , cyclic dipeptides 56 , proteinaceous compounds 57 , organic acids 58-60 , bacteriocins and BLIS 61 with potent antifungal effect 62 , whose action seems to occur, in some cases, through damage of fungal membrane 63 , but it is still unknown in other cases.
In general, BLIS production by P. pentosaceus begins during the exponential growth phase 64 , reaches a maximum in the early stationary phase, and then quickly decreases; in contrast, Dalie et al. 65 observed that the www.nature.com/scientificreports/ production of antifungal metabolites by the strain P. pentosaceus L006 was not growth dependent and achieved a maximum at the end of the stationary growth phase. Aspergillus spp. are spoilage microorganisms that are able to release mycotoxins in cereal grains 66 . To check the potential antifungal activity of P. pentosaceus LBM 18 BLIS, CB was applied directly to one of the fungal isolates (F1) from Austrian corn grain silage, which was apparently an Aspergillus (Fig. 1). Fungus damage was evident after 2 days of contact with CB from the prevention of mycelium expansion from the place of its application. This observation agrees with the antifungal effect observed against Aspergillus niger and Aspergillus flavus after exposure to cell-free culture of P. pentosaceus isolated from barley sourdough 67 .
Fungal spoilage and mycotoxin contamination are the greatest risks during storage of feed like silage. Silage is a traditional way to preserve farm animal feed for extended periods of time when feed is limited or unavailable 68,69 . It consists in green forage preservation by spontaneous lactic fermentation under anaerobic conditions 70,71 , whose primary purpose is to maximize nutritive value with minimum loss 71 . However, forage used for making silage is naturally in contact with yeasts and filamentous fungi present in the harvested biomass, or contamination may occur during harvesting, transport and storage. Growth of yeasts occurs under aerobic conditions, leading to loss of nutrients and dry matter, favoring the formation of butyric acid, reducing palatability and, consequently, shattering silage consumption 72 .
Most of studies published on this issue demonstrated a certain incidence of fungi in silage, which results in alteration or even deterioration of its nutritional value, being fungal distribution quite similar among different countries. However, the main fungi isolated from silage belong to Aspergillus, Penicillium and Fusarium genera, with high incidence of potentially toxicogenic species such as A. flavus, Aspergillus parasiticus, Aspergillus fumigatus, Penicillium roqueforti, Fusarium verticilloides and Fusarium graminearum 71 .
In the present study, seven CB-sensitive fungi were isolated from Austrian corn grain silage 41 . Since CB was more effective as antibacterial agent after 1:5 and 1:10 (v/v) dilution with sterile deionized water (Table 1), both dilutions were tested against fungi. The 1:10 (v/v) water-diluted BLIS was very effective to inhibit the growth of two fungal isolates, named as F3 and F4, after 3-months of treatment (Fig. 1), while the 1:5 (v/v) one was effective against four other isolates, named as F2, F5 (Fig. 2), F6 and F7 (Fig. 3) until 7 and 15 days. After 25 days  www.nature.com/scientificreports/ of 1:5 (v/v) water-diluted BLIS treatment, the F2, F5 (Fig. 2), F6 and F7 (Fig. 3) isolates were able to grow, but their growth was visibly impaired compared to the control (water). In addition, fungi that grew after 25 days of treatment had completely different morphology compared to the control, indicating a severe damage caused by the contact with BLIS. Such a damage may be ascribed to some effect of BLIS on fungal membrane and the observed variations to a sensitivity dependence on the species and/or strain, which may be related to different capacities to change the cell metabolism in response to stress conditions 63,67,72 . These results agree with literature reports, in which Pediococcus genus has proven to be promising in the production of antifungal molecules 57,73 . For instance, P. pentosaceus L006 isolated from maize leaves was able to control the growth of mycotoxigenic molds such as a fumonisin-producing fungus 65 , while Pediococcus acidilactici LAB 5 isolated from meat was able to suppress the growth of food and feedborne molds and plant-pathogenic fungi 73 .

Sensitivity of BLiS to hydrolytic enzymes
To assess the antibacterial activity of CB either before or after treatment with proteolytic enzymes, L. sakei was used as an indicator strain after 1:100 (v/v) dilution of a suspension with 0.3 OD 600nm in sterile deionized water. Non-treated CB exhibited an inhibition halo as large as 19.00 mm, while the treated one completely lost (100.0%) his activity (results not shown) after treatment with all proteolytic enzymes. Whereas the treatment with trypsin, pepsin and papain confirms the proteinaceous nature of BLIS, that with α-amylase suggests a complex structure 74 with a saccharide moiety playing an important role in biological activity 75 . In addition, since there are only few examples of circular bacteriocins or BLIS resistant to proteolysis [76][77][78] , these results taken together indicate that BLIS produced by Pediococcus pentosaceus LBM 18 may be a linear proteinaceous glycoactive compound.

Protein quantification
To obtain additional information on BLIS structure, CB protein content without any enzymatic treatment was determined after increasing dilutions (up to 1: 500 v/v) in sterile deionized water. As shown in Table 3, the 1:50 (v/v) and 1:100 (v/v) dilution allowed reducing CB protein content from 3.00 ± 0.74 g/mL to values (0.06 ± 0.01 g/ mL and 0.04 ± 0.01 g/mL, respectively) close to that of commercial pediocins (0.05 g/mL), which are low molecular weight (2.7-4.0 kDa) peptides. Therefore, it can be inferred that the BLIS produced in this study may be a larger protein than commercial pediocins. Indeed, there are reports in the literature of bacteriocins with very large molecular size (> 10 kDa), which should then be classified as proteins rather than peptides [79][80][81][82][83] .  FT-IR is a useful technique that allows identifying functional groups and then elucidating the chemical structure of unknown molecules. The FT-IR spectrum of BLIS produced by P. pentosaceus LBM 18 after 10 h of cultivation (Fig. 4) points out the presence of bands located at 1691 and 1745 cm -1 corresponding to carbonyls of carboxylic acids and amides, respectively, while those in the region of 1,700-1,550 cm -1 are characteristic of  www.nature.com/scientificreports/ polysaccharides and are usually attributed to C=O stretching 84 . The absorption band at 3,730 cm -1 can be assigned to C-NH 2 stretching 85 . Possibly, the stretching vibration at 1745 cm -1 may be attributed to the carbonyl bonded to the C-NH 2 group, which would correspond to the presence of an amide in the biomolecule under study. Chemical shifts of carbon and hydrogen atoms in NMR spectra give significant information about structures 85,86 . Some peaks (signals) in the 1 H NMR spectrum (0.9, 1.3-1.5 and 2.0 ppm) correspond to the signals of protons of the aliphatic chains R-CH 3 , R-CH 2 -R and R-CH-R, respectively, while peaks between 3.3-4.0 ppm to CH 2 -O and H-C-OH groups (Fig. 5). Peaks between 7.0-7.5 ppm and 5.0-9.0 ppm correspond to the aromatic ring and amides, respectively 87 .
While the amidic groups identified by FT-IR may be easily ascribed mainly to BLIS, the nature of polysaccharides required additional information, for which we resorted to Scanning Electronic Microscopy. One can see in Fig. 6 that P. pentosaceus LBM cells tended to form aggregates, which suggests that at least a significant portion of them may be constituted by exopolysaccharides typically involved in the formation of biofilms by species of the same genus 88 . Figure 7 shows that BLIS was not cytotoxic to peripheral blood mononuclear cells (PBMCs), the absorbance of viable cells treated with BLIS (0.9-1.0) being very close to that of untreated ones (control) (1.0-1.2). In contrast, it demonstrated cytotoxicity to human colon adenocarcinoma Caco-2cells, since the treatment led to a drop in the absorbance of viable cells from 2.7-3.0 in the control to only 2.1-2.3. Moreover, unlike PBMCs, the treatment with BLIS impaired the morphology of Caco-2 cells in culture.

cell death induction assay
It is evident in Fig. 8a that BLIS was not toxic to PBMCs, as the treated cells remained alive after 48 h of treatment with no relevant difference with the control. In contrast, as suggested by a significantly higher number of dead cells than in the control, BLIS was found to be toxic to Caco-2 cells after the same time (Fig. 8b). This result, which is quite encouraging since the targets are tumor cells, suggests possible applications of BLIS not only in food preservation, but also in medical-pharmaceutical applications.
conclusions Novel antimicrobial molecules able to act as biopreservatives, instead of chemical preservatives, are of great interest for the feed and food industry mainly from the microbiological safety and food security viewpoints. Pediococcus pentosaceus LBM 18 BLIS inhibited the growth of spoilage bacteria belonging to Lactobacillus, Enterococcus and Listeria genera and effectively suppressed the growth of fungi isolated from corn grain silage for up to 25 days, impairing the morphology of their colonies likely affecting fungal membranes. The results of the present study showed the effectiveness of BLIS produced by the strain P. pentosaceus LBM 18 as an antibacterial