Assessment of safety and in situ antibacterial activity of Weissella cibaria strains isolated from dairy farms in Minas Gerais State, Brazil, for their food application

Abstract

Weissella cibaria W21, W25, and W42 strains have previously been characterized for their antagonism against a range of foodborne pathogens. However, prior to their use as protective agents, further analyses such as their safety and in situ activity are needed. The safety of W. cibaria W21, W25, and W42 strains was predicted in silico and confirmed experimentally. Analyses of their genomes using appropriate software did not reveal any acquired antimicrobial resistance genes, nor mobile genetic elements (MGEs). The survival of each strain was determined in vitro under conditions mimicking the gastrointestinal tract (GIT). Thus, hemolysis analysis was performed using blood agar and the cytotoxicity assay was determined using a mixture of two cell lines (80% of Caco-2 and 20% of HT-29). We also performed the inflammation and anti-inflammation capabilities of these strains using the promonocytic human cell line U937. The Weissella strains were found to be haemolysis-negative and non-cytotoxic and did not induce any inflammation. Furthermore, these strains adhered tightly to intestinal Caco-2 cell-lines and exerted in situ anti-proliferative activity against methicillin-resistant Staphylococcus aureus (strain MRSA S1) and Escherichia coli 181, a colistin-resistant strain. However, the W. cibaria strains showed low survival rate under simulated GIT conditions in vitro. The unusual LAB-strains W. cibaria strains W21, W25, and W42 are safe and endowed with potent antibacterial activities. These strains are therefore good candidates for industrial applications. The results of this study provide a characterization and insights into Weissella strains, which are considered unusual LAB, but which prompt a growing interest in their bio-functional properties and their potential industrial applications.

Introduction

Several regions of Brazil produce their own artisanal cheeses, offering therefore access to diverse bacterial microbiota and beneficial microorganisms such as LAB that could be employed in the food and pharmaceutical sectors [1]. Handmade cheeses and raw milk are important sources of novel LAB strains [2], and each of these cheeses has its specific microbiota. Environmental microorganisms can also contaminate the product during the manufacturing process, further influencing the fermentation process and consequently impact the organoleptic and aromatic qualities of the final product [3]. Among these microorganisms, there is the genus Weissella. It consists of Gram-positive bacteria which are catalase-negative and asporogenous, with coccoid morphology (or are short bacilli) with heterofermentative metabolism and produce CO₂ from carbohydrate uptake, making them suitable for use in fermentation processes. Of note, the Weissella genus was created in 1993, after reconsidering taxonomically atypical Leuconostoc [4]. Its species are facultative anaerobic, obligately fermentative, and non-motile bacteria, except for W. beninensis, which is reported as being motile [5].

Weissella spp. inhabit a wide range of ecological niches including plants, vegetables, soil, water, and fermented foods, where they assume a probiotic role [4]. Weissella species can also be found in habitats associated with the human or animal body, e.g., the GIT [6], vaginal microbiota [7], or in human breast milk [8]. Furthermore, strains of this genus endowed with antagonistic activities against different foodborne pathogens have been isolated from different Brazilian regions producing artisanal cheeses [9–11]. Weissella is widely reported for diversity and high production of exopolysaccharides with potential of industrial applications and bio-functional properties [12, 13].

The ability of Weissella and other LAB to produce antimicrobial compounds is of great interest because of their potential applications in the food sector. Indeed, the last decade has seen a new challenge for the food industry, namely, to identify natural preservative compounds that will improve safety and security after their incorporation into foods [14]. Foods free of synthetic chemicals are new demands of consumers, though not exclusively of consumers of high living standards. Thus, alternatives based on the incorporation of beneficial microorganisms or their metabolites into food matrix might be a next solution.

However, the use of a novel strain as an ingredient in either human or animal feed requires strict adherence to safety criteria, ensuring the absence of adverse effects such as hemolysis, presence of genes encoding antibiotic resistance, or pro-inflammatory reaction [15, 16]. Thus, the aim of this study was to conduct a comprehensive characterization of W. cibaria strains (W21, W25, and W42), isolated in different Brazilian farms, in Campos das Vertentes (Minas Gerais State), and address their safety features and in situ antibacterial activity.

Materials and methods

Bacterial strains and their draft genomes

The strains W21 and W25 were previously isolated from pasture samples and strain W42 from a soil sample, all of them from dairy farms located in the Campos da Vertentes region, in the southeast of the Minas Gerais state, Brazil. These strains were named respectively as “isolate id 21,” “isolate id 25,” and “isolate id 42” by Teixeira et al. [9]. The strains were identified as Weissella cibaria after sequencing of the gene 16S rRNA.

Draft genomes of W. cibaria W21, W25, and W42 strains were previously sequenced and assembled [17, 18]. They were retrieved from the GenBank database under accession numbers JALRNM000000000, JAFNKE000000000, and JALRNN000000000, respectively. The annotation of these draft genomes was performed using Prokka software version 1.14.5 [19].

Safety of Weissella cibaria W21, W25, and W42 strains

Bioinformatics analyses of Weissella genomes

To locate gene coding for the well-known virulence factors in the genomes of W. cibaria W21, W25, and W42 strains, web tools recommended by the European Food Safety Authority were used. The search for MGE like plasmids was done by employing plasmidFinder [20], while the PathogenFinder web tool [21] identified their profile as potential human pathogens. Gene coding for the virulence factors were investigated by using the “Virulence Factors Database” (VFDB) (http://www.mgc.ac.cn/VFs/main.htm) [22], while the search for gene coding for antibiotics resistance was carried out using the Resistance Gene Identifier (RGI) tool and the “Comprehensive Antibiotic Resistance Database” (CARD) (https://card.mcmaster.ca/) [23]. To this end, the contigs were imported and the parameters “perfect and strict hits only” and “high-quality coverage” were selected. Finally, the ResFinder 4.1 [24] server (https://cge.cbs.dtu.dk/services/ResFinder/) was used to identify the acquired antimicrobial resistance genes with a selected % ID threshold of 90%, and the selected minimum length of 60%.

For the detection and annotation of prophage sequences within bacterial genomes, the PHASTER [25] (PHAge Search Tool Enhanced Release) (http://phaster.ca) webserver was used. The CRISPRCasFinder [26] tool (https://crisprcas.i2bc.paris-saclay.fr/CrisprCasFinder/Index) was also used to detect the CRISPR and Cas genes by importing the contig FASTA sequences and selecting “Perform cas detection,” while the remaining parameters were set as default.

Hemolysis and antibiotic susceptibility

In vitro tests to establish the hemolytic activity and antibiotics susceptibility of these strains were performed according to protocols reported by Colombo et al. [27]. To evaluate the hemolytic activity, these strains were cultured on Columbia Blood Agar Settle plate (Merck KGaA, Darmstadt, Germany). After 24 h at 37 °C, the plates were analyzed for microbial hemolytic activity, considering a total or β-hemolysis as clear halos around the colonies, partial or α-hemolysis as greenish halos around the colonies, and γ-hemolysis as the absence of hemolysis [27]. Staphylococcus aureus ATCC 29213 was used as the positive control.

Susceptibility to antibiotics was determined using the disk diffusion assay (Oxoid), according to the protocol reported by Colombo et al. [27]. Each strain was diluted in a solution of 0.85% NaCl (w/v) to obtain a turbidity equivalent to 0.5 McFarland standard and then was swabbed onto the surface of an MRS agar plate (where the antibiotic disks were placed). These antibiotic disks contained imipenem (carbapenem), vancomycin (glycopeptide), clindamycin (lincosamide), erythromycin (macrolide), ampicillin (penicillin), ciprofloxacin (fluoroquinolone), and amoxicillin (penicillin) + clavulanic acid (β-lactamase inhibitor). The plates were incubated at 37 °C for 18 h, after which the diameter of the inhibition halos was measured and the resistance profile was determined according to CLSI breakpoints [28].

Cytotoxicity and inflammatory activity of Weissella strains in human cell

Cytotoxicity assay

For the cytotoxicity test, a mix of human cell lines composed of 80% of human colon adenocarcinoma Caco2 cells and 20% of HT-29 was used. The preparation of the mix of human cell cultures was carried out according to Pouille et al. [29]. The mix of Caco2 cell and HT-29 line was cultured in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, Thermo Fisher Scientific), supplemented with 10% (v/v) of heat-inactivated fetal bovine serum (FBS) (Biowest, Nuaillé, France), 100 units per milliliter of penicillin, 100 µg per ml of streptomycin (Gibco, Life Technologies, Grand Island, NY, USA), and 10 mM nonessential amino acids (Gibco, Life Technologies, Paisley, UK). Cells were cultured under standard cell culture conditions (37 °C and 5% CO₂). The culture medium was changed regularly and when the cells reached sub-confluence (80–90%), they were sub-passaged. The cells were cultured at 1.5 × 10⁴ cells ml⁻¹ in 96-well plates for 24 h and then exposed to the three different Weissella strains.

The preparation of the Weissella cultures to contact the mix of human cell culture was realized as described below. One percent of frozen Weissella strains was propagated in MRS broth (Oxoid) for 24 h, centrifuged at 9000 g for 5 min, and the supernatant was removed. The obtained pellet was washed with 5 ml of phosphate-buffered saline (PBS, pH 7.4), and resuspended in 2 ml of DMEM without antibiotics and fetal bovine serum and diluted to a final concentration of 10⁷ CFU per ml. The suspension of cells was incubated at 37 °C until contact with human cells. For the contact, the mix of human cell culture fixed in the bottom of the 96-well plate was washed with DMEM and the suspensions of Weissella strains were added in a proportion of 1:1 and 1:10 (MOI; multiplicity of infection), after which the plate was incubated for 24 h at 37 °C.

Cell viability of Caco2 and HT-29 after 24 h of incubation with Weissella strains was measured using a cell counting kit-8 assay (CCK-8, Dojindo Molecular Technologies, Inc. Japan). After incubation, the 96-well plate was washed with DMEM with antibiotics (100 U ml⁻¹ penicillin and 100 µg ml⁻¹ streptomycin). After the washing procedure and removing all Weissella strains cells, 150 µl of CCK-8 solution (5%) was added to each well of the 96-well plate, followed by incubation for 1.5 h at 37 °C. Then, 10 ml of stop solution provided in the CCK-8 kit was added to each well. The absorbance of the mixture was measured at 450 _nm using a microplate reader (Synergy H1, BioTek, USA). Cell viability was calculated based on the relative absorbance compared with the control group.

Inflammatory effects of Weissella strains

To investigate the inflammatory and anti-inflammatory effects of Weissella strains, the promonocytic human cell line U937 was used according to the protocol described by Pouille et al. [29]. The first step is to verify that the Weissella strains are not cytotoxic for these cell lines. For that we performed the cytotoxicity assay as described before in the previous section. The cell line U937 was cultured in Roswell Park Memorial Institute 1640 Medium (RPMI, Gibco, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS), 100 U ml⁻¹ penicillin, 100 µg ml⁻¹ streptomycin, and 2 mmol glutamine in a humidified 5% CO₂ atmosphere at 37 °C. The culture medium was replaced regularly and when the cells reached sub-confluence (80–90%), they were sub-passaged. For the differentiation of the cell line in macrophages, U937 cells were seeded at approximately 1 × 10⁵ cells/well in 96-well plates with 60 ng ml⁻¹ of phorbol-12-myristate-13-acetate (PMA, Thermo Fisher Scientific) for 48 h. The adherent cells were washed with PBS and incubated for 2 h with LPS (10 µg ml⁻¹; LPS from E. coli O26:B6, Millipore Corporation) and Weissella strains. Cell supernatants were collected on ice and centrifuged (250 g, 5 min) to eliminate cell debris. The supernatants were aliquoted and stored at − 80 °C until further analysis. Weissella strains samples were prepared as described above, resuspended in unsupplemented RPMI, and added to each well in 1:10 MOI. The inflammatory control was U937 differentiated with LPS (50 µg ml⁻¹) and the anti-inflammatory control was U937 differentiated with LPS (50 µg ml⁻¹) and dexamethasone (20 µM).

Multiple cytokines in culture supernatants of the U937 cell line were detected with a customized Milliplex Map kit (Human High Sensitivity T Cell Magnetic Bead Panel, Millipore, Billerica, MA), following the manufacturer’s instructions. Briefly, antibody-immobilized beads for detection of interleukin-1β (IL-1β), IL-2, IL-4, IL-5, IL-6, IL-7, IL-10, IL-12 (p70), IL-13, tumor necrosis factor-alpha (TNF-α), granulocyte–macrophage colony-stimulating factor (GM-CSF), and interferon-γ (IFNγ) were incubated with 5-times-diluted culture supernatants of U937. The beads were subsequently treated with detection antibodies as well as PE-conjugated streptavidin and the quantification was carried out using the Luminex® 100/200 (Luminex Corporation, Austin, TX, USA) system and the Luminex xPONENT® for LX100/200 software. The cytokine IL-8 was quantified using a human IL-8/CXCL8 Quantikine® ELISA kit (R&D Systems, Inc., Minneapolis, MN, USA). Medium samples were diluted 100-fold according to the kit recommendations.

Survival of Weissella cibaria strains under conditions mimicking a human gastric digestion

The survival of Weissella strains in simulated human gastric and intestinal juices was assessed as described by Fonseca et al. [30] with modifications. The composition of the simulated gastric juice was pepsin (2 g l⁻¹) (Sigma-Aldrich, St Louis, MO, USA) and pH 3.5 PBS, adjusted with HCl 0.5 mol. Each isolate was grown at 37 °C overnight in sterile MRS broth (Oxoid) and inoculated in the solution until the final concentration of ~ 1 × 10¹⁰ CFU ml⁻¹ and followed by incubation at 37 °C for 90 min with constant shaking at 160 rpm. After incubation, the samples were taken and serially diluted, then plated on MRS agar plates and incubated at 37 °C for 24 h. Then, the simulated duodenal juice was prepared by adding pancreatin (Dinâmica, Brazil) and bile salt solutions to obtain final concentrations of 0.2% and 0.5% (w/v), respectively, and the pH was then adjusted to 7.0 by adding 1 mol sodium hydroxide (NaOH). After mixing, samples were incubated at 37 °C for 180 min with agitation (160 rpm), and then viable cell counts were determined. All sample counts were determined by plating on MRS (Oxoid) agar. The experiments were repeated at least three times. Results were expressed as mean log colony-forming units per ml (CFU ml⁻¹). The rate of survival was calculated as the percentage of Weissella colonies grown on MRS agar relative to the initial bacterial concentration:

$$Survivalrate\left(\%\right)=\frac{\text{log}CFUN1}{\text{log}CFUN0}\times 100$$

where N1 is the viable count of isolates after incubation and N0 is the initial viable count.

Adhesion of Weissella cibaria strains to the human Caco-2 cell line

The adhesion capacity test for the three selected strains to the human colon adenocarcinoma cell lines (Caco-2) was performed according to Fonseca et al. [30] with slight modifications. The Caco-2 cells were sub-cultured (2 × 10⁵ cells ml⁻¹) in 24-well tissue culture plates (Sarstedt, Germany) and grown at 37 °C in a humidified atmosphere of 5% CO₂ for 21 days to promote differentiation in cell media. The culture medium was changed on alternate days. For the adhesion assay, bacteria were cultured in MRS broth overnight at 32 °C, and after washing twice with PBS, the cultures were resuspended in the DMEM at a concentration of approximately 10⁷ CFU ml⁻¹. After the DMEM of the 24-well tissue culture plates was removed, the culture cells were washed with the same medium. Then 400 µl of DMEM and 100 µl of Weissella strains were added to each well. The plate was incubated at 37 °C overnight in a 5% CO₂ atmosphere. Subsequently, the cells were washed three times with 1 ml of PBS to remove non-adherent bacteria cells and then lysed with 300 µl of trypsin. After 15 min of incubation at 37 °C, the solution with released bacteria cells was transferred to 1.5-ml Eppendorf and the bacteria cells were centrifuged (9000 g, 5 min), the supernatant was removed, and the pellet was resuspended in 500 µl of PBS. After recovery, bacterial cells were serially diluted and plated on MRS plates. The plates were incubated at 32 °C for 24 h. Then, enumerations of W. cibaria strain colonies were performed. Experiments were performed in duplicate and repeated three times.

In situ inhibition of pathogenic bacteria

For this analysis, Caco-2 cell line cultures were cultivated and Weissella strains were propagated as described above. Before the experiment, the DMEM of the 24-well tissue culture plates was removed and the culture cells were washed with the same medium. For the pathogen inhibition test, the exclusion assay was performed according to the protocol described by Fonseca et al. [30], with modifications. The pathogens used in this experiment, Escherichia coli 184 and Staphylococcus aureus MRSA SA1, were obtained from the bacteria culture collection of Institut Charles Viollette (ICV; Lille, France). They were propagated separately in BHI medium (Oxoid) at 37 °C for 24 h. Caco-2 cells were first preincubated with 100 µl of Weissella strains (10⁷ CFU ml⁻¹) for 2 h at 37 °C in a 5% CO₂ atmosphere and then 100 µl of E. coli (10⁷ CFU ml⁻¹) or S. aureus (10⁷ CFU ml⁻¹) was added to each well. The plates were incubated at 37 °C overnight in a 5% CO₂ atmosphere. After the incubation period, the cells were washed three times with 1 ml of PBS to remove non-adherent bacteria cells and then lysed with 300 µl of trypsin. After 15 min of incubation at 37 °C, the solution with released bacteria cells was transferred to 1.5-ml Eppendorf tubes and the bacterial cells were centrifuged (9000 g, 5 min), the supernatant was removed, and the pellet was resuspended in 500 µl of PBS. After recovering, the bacterial cells were serially diluted and plated. E. coli was plated on Eosin Methylene Blue agar (EMB) and S. aureus on BHI supplemented with the antibiotic erythromycin (100 µl ml⁻¹) that prevented the growth of Weissella strains. The plates were incubated at 37 °C for 24 h when enumerations of pathogen colonies were performed. Experiments were performed in duplicate and repeated three times.

Results

Weissella cibaria W21, W25, and W42 strains are safe

Based on the ResFinder data, no acquired antimicrobial resistance genes were found in the genomes of the three tested Weissella strains. Moreover, the research for plasmid evaluated by plasmidFinder [20] revealed the absence of this MGE. The CRISPRCasFinder analysis showed that these three Weissella strains harbor in their genomes the CAS-TypeIIA array, and CAS-TypeIIU for the W21 strain (Table 1).

Table 1.

CRISPR regions and cas genes found in genomes of the three Weissella cibaria strains isolated from dairy farms located in Campos das Vertentes (Minas Gerais State, Brazil) using the CRISPRCasFinder tool

CRISPR region							CAS protein
Strain	Region	Start	End	DR consensus	DR length	Spacers	Gene name	Start	End	Orientation
W25	C4	12,728	13,621	GTTTTAGTGTCATGTTGAATAGAATGCTTCTCAAAC	36	13	Type: CAS-TypeIIA Start: 6032; end: 12,578
							cas9_TypeII	6032	10,498	+
							cas1_TypeII	10,710	11,582	+
							cas2_TypeI-II-III	11,579	11,914	+
							csn2_TypeIIA	11,907	12,578	+
W42	C22	95,084	95,977	GTTTGAGAAGCATTCTATTCAACATGACACTAAAAC	36	13	Type: CAS-TypeIIA Start: 96,127; end: 102,673
							csn2_TypeIIA	96,127	96,798	−
							cas2_TypeI-II-III	96,791	97,126	−
							cas1_TypeII	97,123	97,995	−
							cas9_TypeII	98,207	102,673	−
W21	C25	12,686	13,579	GTTTTAGTGTCATGTTGAATAGAATGCTTCTCAAAC	36	13	Type: CAS-TypeIIA Start: 11,865; End: 12,536
							csn2_TypeIIA	11,865	12,536	+
							Type: CAS-TypeIIU Start: 5990; End: 12,536
							cas9_TypeII	5990	10,456	+
							cas1_TypeII	10,668	11,540	+
							csn2_TypeIIA	11,865	12,536	+

The analysis conducted using the CARD database only identified genes involved in the resistance to vancomycin (vanY gene in the vanB cluster, and vanT gene in the vanG cluster) for all three tested strains (Table 2), and this resistance was confirmed experimentally in vitro (Table 3), demonstrating the convergence of these two methods.

Table 2.

Antibiotic resistance genes identified in Weissella cibaria strains using CARD (Comprehensive Antibiotic Resistance Database) webserver

Strain	RGI criteria	ARO term	Detection criteria	AMR gene family	Drug class	Resistance mechanism	% identity of matching region	% length of reference sequence
W21 W25 W42	Strict	vanY gene in vanB cluster	Protein homolog model	vanY, glycopeptide resistance gene cluster	Glycopeptide antibiotic	Antibiotic Target alteration	34.95	97.39
	Strict	vanT gene in the vanG cluster	protein homolog model	Glycopeptide resistance gene cluster, VanT	Glycopeptide antibiotic	Antibiotic target alteration	33.33	69.38

Table 3.

Results from in vitro antibiotic resistance test for Weissella cibaria strains (W21, W25, W42)

Strain	Antibiotics spectrum (mm)
	Ampicillin	Amoxicillin + clavulanic acid	Imipenem	Ciprofloxacin	Vancomycin	Erythromycin	Clindamycin
W21	29	29	35	24	0	24	24
W25	31	31	33	20	0	27	26
W42	28	32	32	22	0	27	26

Results are the diameter of inhibition zones, expressed in mm, obtained against different antibiotics

Moreover, only two prophages (1 entire and 1 incomplete) could be located using the PHASTER tool in the genomes of the three strains as shown on Table 4. Additionally, our study revealed the absence of hemolysis-associated genes in the genomes of the W. cibaria strains W21, W25, and W42. This was further confirmed through phenotypic analysis using blood-agar medium (Fig. 1). Therefore, slight haemolysis that could be graded as an α-hemolysis was detected.

Table 4.

Prophage regions identified in Weissella cibaria strains genome using the PHASTER webserver

Strain	Region	Length (kb)	Completeness*	Score	GC %	Most common phage
W21	1	17.9	Incomplete	20	41.52	PHAGE_Lactoc_PLgT_1_NC_031016(3)
	2	32	Intact	110	42.64	PHAGE_Entero_phiFL1A_NC_013646(14)
W42	1	17.9	Incomplete	20	41.51	PHAGE_Lactoc_PLgT_1_NC_031016(3)
	2	35.8	Intact	110	42.30	PHAGE_Entero_phiFL1A_NC_013646(14)
W25	1	17.9	Incomplete	20	41.54	PHAGE_Lactoc_PLgT_1_NC_031016(3)
	2	35.8	Intact	110	42.30	PHAGE_Entero_phiFL1A_NC_013646(14)

^*Intact (score > 90); questionable (score 70–90); incomplete (score < 70)

Fig. 1.

Hemolytic activity of Weissella cibaria strains W21, W25, and W42

Weissella cibaria W21, W25, and W42 strains are not cytotoxic towards cells from Eucarya domain and do not induce inflammatory effects

For the cytotoxicity test, a mix of human cell cultures with 80% of Caco2 and 20% of HT-29 (Fig. 2A) and with the human U937 cells (Fig. 2B) was used. As shown, no cytotoxicity effect was registered for W. cibaria strains W21, W25, and W42 in comparison to the untreated control on both the tested eukaryotic cell lines.

Fig. 2.

Effect of Weissella cibaria strains on cytotoxicity of cell-culture composed of 80% of Caco2 + 20% of HT-29 cells (A) and U937 cells (B). In panel A, (1:10) and (1:1) indicate the proportion added of strains in the cell culture well plates. Error bars represent standard error of the mean

Survival of W. cibaria strains in the simulated gastro-intestinal tract

During the assessment of the survivability of Weissella strains in the GIT, a small loss of the viability of the cells was observed after 1.5 h in conditions mimicking the stomach compartment, but no survival was observed upon 3 h of incubation in conditions simulating those of the duodenal compartment (Table 5). The strain W21 presented the upmost rate of survivability after following its submission to an acidic pH and pepsin solution, which indicates higher tolerance to stomach conditions. On the other hand, strain W42 presented the smallest rate with a difference of 2.5% from strain W21 under similar conditions. However, when the duodenal conditions were simulated and then used, by increasing the pH to 7 and adding pancreatin and bile salts solutions, none of these three strains was able to survive.

Table 5.

Survival of Weissella cibaria strains expressed as log₁₀ of colony-forming units per milliliter (log₁₀ CFU ml⁻¹) under conditions simulating gastric digestion

Strains	Time						Survival rate (%) of time 1.5 h
	0 h		1.5 h		4.5 h
	pH	Survival	pH	Survival	pH	Survival
W25	3.5	10.04	3.5	9.68	7	ND	96.4
W42	3.5	9.80	3.5	9.37	7	ND	95.6
W21	3.5	9.80	3.5	9.60	7	ND	98.0

ND not detected—below the limit of detection

Adhesion of W. cibaria strains to the human Caco-2 cell line and exclusion assay

The adhesion of W. cibaria strains to the Caco-2 cell lines was almost 100% effective for W21 and W25 strains (Table 6). The initial inoculum was 7–8 log CFU ml⁻¹ and all the W. cibaria strains W21, W25, and W42 tested strains were able to adhere in the Caco-2 cell line, with a lower adhesion for W42.

Table 6.

Results expressed as log₁₀ of colony-forming units per milliliter (log₁₀ CFU ml⁻¹) obtained for adhesion of Weissella cibaria strains to Caco-2 cells and their competitive exclusion against pathogenic bacteria

	Adhesion	Exclusion
Strains		S. aureus MRSA S1	E. coli 184
Control		6.53	6.47
W25	7.34	6.65	6.50
W42	6.33	6.85	6.42
W21	7.33	ND	6.44

ND not detected—below the limit of detection

The adhesion of W. cibaria W21 and W25 was 7.33 log CFU per ml⁻¹ considered almost 100% effective once their values are between the initial inoculum. The strain W42 presented a less adherent score than the other and the counting after the adhesion analysis was 6.33 which is lower than the initial inoculum. Regarding the exclusion assay, it was revealed that only W. cibaria W21 was able to impede completely the adhesion of S. aureus MRSA S1 to the Caco-2 cells.

Discussion

This study evaluated various criteria which showed that the W. cibaria W21, W25, and W42 strains are promising probiotic candidates. First, their safety was evaluated through genotypic and phenotypic analyses. Then, the strains were confronted with the harsh conditions of the GIT. According to the results obtained using the CRISPRCasFinder tool, each of the three Weissella strains’ genomes harbors a single CRISPR array, with evidence level 4 [26]. The CRISPR-associated protein Csn2_TypeIIA is a mandatory protein of the type IIA system and the gene is present in the three sequenced genomes. According to Roberts and Barrangou [31], the type II systems are rare in nature compared to types I and III, which are also more well-characterized. The CRISPR-Cas system is a defense mechanism used by bacteria to protect themselves against infection by foreign genetic elements. Thus, strains endowed with the CRISPR-Cas system could acquire fewer sequences derived from other bacteria preventing therefore horizontal gene transfer and leading to a low level of drug resistance [32].

The number of spacers in the CRISPR-Cas system represents the ability to eliminate heterologous genes, such as prophage. Thus, the more spacer sequences that are present, the more the strain has a limited capacity to acquire foreign sequences like prophages [32]. Furthermore, the PHASTER tool located only two prophages in the genome of these strains. Dong et al. [32] demonstrated that the correlation between spacers and prophages is negative. Prophages are MGE which is considered the primary factor in genetic diversity enabling niche adaptation in bacteria [33]. As abovementioned, two prophages were located in the genomes of the three strains W. cibaria W21, W25, and W42; one of them was intact, whereas the second was incomplete in each of the tested strains. The length of the intact prophage varies from 32 to 35.8 Kb with the most common phage: PHAGE_Entero_phiFL1A_NC_013646. The presence of intact regions could be a selective advantage and help to protect bacteria against other prophage infections [34, 35].

The analysis conducted using the CARD database revealed the presence of vancomycin resistance genes in the genomes of the studied Weissella strains. Muñoz-atienza et al. [36] reported that 60% of Weissella species are resistant to several antibiotics. Additionally, Tenea and Hurtado found putative resistance genes in the genome of W. cibaria UTNGt21O [37]. Therefore, the results obtained from both the detection of genes using the CARD database and the in vitro antibiotic tests conducted in this study corroborate the low level of resistance, once the three tested strains presented resistance to vancomycin, which is intrinsic for this genus [13]. Indeed, Weissella is known to be intrinsically resistant to vancomycin and has high minimum inhibitory concentration (MIC) of ≥ 256 µg per milliliter [13, 38]. However, antimicrobial susceptibilities of other species like W. confusa are not fully understood. Currently, there are no standard methods and interpretation criteria of antimicrobial susceptibilities established for Weissella spp. by the Clinical and Laboratory Standards Institute (CLSI). Besides, intrinsic resistance to antibiotics is not redhibitory or a break in the procedure of selection of these strains for probiotic applications [15].

Several genes were predicted as putative virulence factors by VFDB (data not shown). However, none of the well-known virulence genes gelE (gelatinase), hyl (hyaluronidase), asa1 (aggregation substance), esp (enterococcal surface protein), cylA (cytolysin), efaA (endocarditis antigen), and ace (adhesion of collagen) were detected in the genomes of W. cibaria W21, W25, and W42 strains. Besides, most of the genes detected by the VFDB webserver are related to other cellular functions such as genes for putative enolase (eno) phosphopyruvate hydratase. Enolase is a key glycolytic enzyme in the cytoplasm of prokaryotes, essential for the degradation of carbohydrates via glycolysis [39]. Some researchers reported the presence of virulence-encoded genes such as collagen adhesin proteins and genes associated with toxin production systems (including botulinum neurotoxin homolog) in some Weissella species like W. cebi, W. confusa [40], and W. oryzae [41], respectively. None of these genes were found in our W. cibaria strains tested. Also, Tenea and Hurtado [37] reported the presence of two hemolysis-associated genes (hlyD, hlyIII) in some strains of W. cibaria. Our investigation revealed the absence of these genes in the genomes of the W. cibaria strains. Despite several virulence factors detected by the VFDB webserver, the results indicated that none of the known virulence factors were present among them, and the analysis of PathogenFinder [21] indicates that the three input organisms were predicted to be non-human pathogens. Therefore, W. cibaria W21, W25, and W42 strains can be considered safe for use as probiotics. Of note, to evaluate the safety of probiotics, there are guidelines that consider several factors in advance, like excessive immune stimulation in sensitive individuals, systemic infection, gene transfer, or deleterious metabolic effects [42]. To the best of our knowledge, W. cibaria is not currently used as a probiotic ingredient, despite its frequent presence in fermented foods and human feces, and is well-known for its beneficial effects [43–45].

It is noteworthy that no cytotoxicity was found for these Weissella strains against human cell lines tested during our study, matching with other studies that underpinned the safety of W. cibaria strains on the RAW 264.7 macrophage cell line [46] and human mouth epithelial cell line KCLB 10017 [47]. The inflammatory and anti-inflammatory capacity of Weissella strains against the human U937 cells was also assessed; these strains, notably, showed neither inflammatory nor anti-inflammatory responses. Taken all together, these data confirm the safety of these strains and promote their potential in the intestinal health of humans and animals. Similarly, Do et al. [48] showed significant therapeutic efficacy of W. cibaria CMU and W. cibaria CMS1 in reducing lung inflammation provoked by DEPM administration in a murine model.

The lack of survival of the Weissella strains during passage through the simulated GIT conditions could be a problem. Indeed, the ability of bacteria to survive under gastric digestion conditions is considered a key criterion for probiotic selection [49]. A recent study showed that the HigBA TA system of W. cibaria 018 responded to bile salt stress, but not to acid stress, which might offer novel perspectives to understand the tolerance mechanism of probiotics to the GIT environment [50]. Therefore, to overcome this potential drawback, delivery strategies such as encapsulation, could help the passage through the GIT by protecting bacteria from acidic stress and delivering directly in the intestine where they can act as probiotics and provide health benefits to the host [51, 52]. Various delivery systems employing different trigger mechanisms have been designed to effectively introduce a multitude of highly viable probiotics into the intestines [53]. Moreover, assessing GIT survival by agar enumeration does not detect the viable but nonculturable cells (VBNC) [54]. Previous studies used the cytometry method, which includes both viable bacteria and VBNC and as expected, the survival percentages obtained by this method were significantly higher in comparison to the data obtained after their growth and enumeration by the agar plating method [49].

The adhesion to the intestinal cell leads to a hurdle against enteropathogenic bacteria, avoiding their adhesion to the intestinal cells and competing with them for nutrients [55, 56]. Remarkably, the data from the exclusion assays revealed that only W. cibaria W21 was able to impede completely the adhesion of S. aureus MRSA S1 to the Caco-2 cells (Table 6), suggesting intraspecific differences between these strains. In the present work, only two pathogenic bacteria were tested in situ, i.e., S. aureus MRSA S1 and E. coli 184, but the W. cibaria strains have a broad spectrum as they are active against L. monocytogenes and Salmonella enterica [9]. Impeding pathogens to adhere to human cells may result from different mechanisms, including production of antimicrobial substances by LAB, and competition for binding sites [57]. Some pathogens have different levels of adhesion and invasion on the eukaryotic cells, and notably some pathogens are more or less stable in mucus-producing intestinal cell models (HT-29-MTX) compared to non-mucus-producing cells (Caco-2) [58]. Such ability of resilience is directly related to the expression of specific proteins, pili, fimbriae, and flagella [59].

Conclusions

Bioinformatic analyses, combined with phenotypic evaluations of functional properties, provided valuable information about Weissella strains. Indeed, W. cibaria W21, W25, and W42 strains were identified as safe according to the bioinformatic analyses, as these strains do not contain acquired antimicrobial resistance genes nor plasmids in their genomes. Only one intact prophage was detected, which can also help the bacteria against other prophage infections. These strains harbor a CRISPR-Cas system which constitutes a defense system against infections caused by heterologous genes. The strains do not contain any antibiotic resistance gene, except the vancomycin resistance gene, a phenotype confirmed experimentally. Such resistance is considered intrinsic for the Weissella genus and other LAB with GRAS status. Finally, none of the known virulence factors were detected in these genomes and no in vitro hemolytic activity was found. Further, no cytotoxicity and no inflammatory effects were detected for these strains. Regarding probiotic traits, the W. cibaria W21 strain showed higher potential, as it demonstrates a high exclusion of S. aureus MRSA S1 in the intestinal cells. However, none of the three strains survived through the simulated GIT. To overcome this, further study regarding protective strategies to help the passage of these strains to the gastric system is needed and more studies with other enteropathogenic strains should be explored.

Declarations

Conflict of interest

LA Nero is one of the editors-in-chief for Brazilian Journal of Microbiology and this article was independently handled by an editorial board member.