Lc. paracasei SP5 was previously isolated from kefir grains (Mantzourani et al., 2019), whereas Lp. pentosus L33 and Lp. plantarum L125 were isolated from traditional fermented meat products (Pavli et al., 2016) and provided by the Institute of Technology of Agricultural Products, Hellenic Agricultural Organization DIMITRA, Likovrisi, Attiki, Greece. All lactobacilli were cultivated O/N in de Man, Rogosa and Sharp (MRS) broth at 37°C, under anaerobic conditions for 20 h. Clinical isolates of Staphylococcus aureus, Salmonella enterica ssp. enterica serovar Enteritidis and Escherichia coli were isolated from clinical specimens of stools, at the Laboratory of Clinical Microbiology, University Hospital of Alexandroupolis. Bacterial identification was based on conventional methods, the automated system Vitek II (Biomerieux, France) and the VITEK^® MS automated mass spectrometry microbial identification system (Biomerieux, France). Pathogens were maintained in Tryptic Soy Broth (TSB, Condalab, Madrid, Spain) at 37°C under anaerobic conditions. Tryptone soya agar (TSA) (Condalab, Madrid, Spain) and McConkey agar (VWR, Radnor, PA, USA) were used to culture S. aureus or S. enterica and E. coli, respectively, for colony enumeration.

The human colorectal adenocarcinoma cell line HT-29 was purchased from ATCC. Cells were maintained in Roswell Park Memorial Institute GlutaMAX™ (RPMI)-1640 medium, containing 10% fetal bovine serum (FBS), 100 μg/mL streptomycin and 100 U/mL penicillin (all from Thermo Fisher Scientific, Waltham, MA, USA). For HT-29-bacteria co-incubations, a modified RPMI-1640 medium consisted of 10% FBS and 20mM 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES; all from Thermo Fisher Scientific) was used. Cells were incubated in a humidified atmosphere at 37°C, 5% CO₂ under sterile conditions.

The antimicrobial activity of the viable potentially probiotic cultures against the enteropathogens was investigated using the well diffusion assay (Balouiri et al., 2016). More specifically, TSA plates were inoculated with fresh O/N cultures of the pathogenic bacteria, using a sterile swab. Then, wells were punctured into the agar (0.7 cm) and filled with 100 μL of lactobacilli (approximately 10⁹ CFU/mL). Plates were left at RT for 1 h to facilitate diffusion and were, subsequently, incubated at 37°C for 20 h. Inhibition zones were measured the next day.

Fresh O/N cultures of lactobacilli and pathogens were utilized to investigate the auto-aggregation and co-aggregation capacity of strains, following a previously published protocol (Prabhurajeshwar and Chandrakanth, 2017), with minor modifications. To test for co-aggregation capacity, 10⁸ CFU/mL of lactobacilli were mixed with 10⁸ CFU/mL of pathogens. Lactobacilli (10⁸ CFU/mL) and pathogens (10⁸ CFU/mL) were separately tested for auto-aggregation. Samples were incubated for 4 h at RT under strictly static conditions. The upper layer (500 μL) was collected at 0 and 4 h of incubation to be subjected to spectrophotometry at 620 nm. Results are expressed as: Co-aggregation (%) = [(A₁ + A₂)/2 – A_mix(A₁ + A₂)/2] × 100, where A₁ and A₂ represent the absorbance at 620 nm of pathogen monocultures and lactobacilli (auto-aggregation), respectively, and A_mix the absorbance of the mixed culture (co-aggregation). Auto-aggregation was calculated using the formula: 1- (A_t=0h/A_t=4h) × 100, where A_t=0h and A_t=4h refer to absorbance at 620 nm of the monocultures at 0 and 4 h, respectively.

The ability of the potentially probiotic strains to inhibit pathogen adhesion on HT-29 cell monolayers was determined after 4 h co-incubations. Briefly, HT-29 cells were seeded in 24-well plates at a density of 4 ×10⁵ cells per well and were incubated until the formation of a monolayer (100% confluency), following a previously published protocol (Plessas et al., 2020). Then, cells were treated with lactobacilli and/or pathogens at a concentration of 10⁸ CFU/mL for 4 h. Control samples were incubated with pathogens alone for 4 h, under the same conditions. The monolayers were washed twice with Phosphate-Buffered Saline (PBS) (Thermo Fischer Scientific) to remove unattached bacteria and cells were detached using 1% v/v Trypsin (Thermo Fischer Scientific). The suspension was serially diluted in 1× Ringer’s solution (Lab M, Lancashire, United Kingdom) and plated onto agar plates: TSA plates for S. aureus and McConkey agar plates for S. enterica and E. coli enumeration. Plates were incubated at 37°C, under anaerobic conditions, until the formation of visible colonies. Attached bacteria on epithelial cells are expressed as Log CFU/mL.

The antimicrobial potential of viable lactobacilli was determined using a previously published protocol with minor modifications (Ayala et al., 2019). Briefly, pathogens (10⁸ CFU/mL) and lactobacilli (10⁸ CFU/mL) were co-incubated for 24 h, at 37°C under anaerobic conditions in TSB (Condalab). The next day, the bacterial suspension was serially diluted in 1× Ringer’s solution (Lab M) and spread in agar plates, for colony enumeration. The investigation of the antimicrobial activity of secreted metabolites was performed as described previously, with minor modifications (Forestier et al., 2001). Lactobacilli were grown for 22 h in MRS and CFCS (10 mL) was collected by centrifugation at 4,000 × g for 10 min and filtered through an 0.2 μm filter. Native CFCS (pH 4.2) or CFCS adjusted to pH 6.2 with NAOH 2M, were used to determine the contribution of pH to the antimicrobial activity of the tested strains. Similarly, native (RT) or heat-treated CFCS (incubation at 100°C for 30 min), were used to determine the stability of the bioactive antimicrobial compounds at denaturing temperature. To this aim, 10⁸ CFU/mL of pathogens were added to the CFCS and were left to incubate for 24 h in sterile tubes. MRS adjusted at the appropriate pH or heat-treated MRS at 100°C were used as a negative control. After the end of the incubation period, samples were vortexed and transferred to a 96-well plate for absorbance reading at 620 nm, using a microplate reader (EnSpire Multimode Plate Reader, PerkinElmer, Waltham, MA, United States). Results are expressed as the percentage (%) = [(Sample OD₆₂₀-Media blank OD₆₂₀)/(Mean control OD₆₂₀-Media blank OD₆₂₀)] × 100.

The ability of CFCS to limit biofilm formation was investigated using a microbiological assay, as described previously, with minor modifications (Sabbatini et al., 2020). Briefly, 10⁸ CFU/mL of fresh O/N cultures of pathogens (100 μL) were co-incubated with 100 μL adjusted CFCS (pH 6.2) in a 96-well plate for 24 h. For the estimation of viable cells, biofilms were washed twice post-treatment with PBS, and were mechanically disrupted. The suspension was serially diluted in 1× Ringer’s solution (Lab M) and spread on agar plates for colony enumeration. Plates were incubated at 37°C under anaerobic conditions until the formation of visible colonies. Viable bacteria are presented as Log CFU/mL.

Confocal microscopy was used for the visualization of the inhibitory effect of CFCS on biofilm formation. To this end, pathogens (10⁸ CFU/mL) were seeded on No. 1.5 coverslips and were treated with either MRS (control) or CFCS (50% v/v) adjusted to pH 6.2 for 24 h. Then, biofilms were washed twice with PBS and incubated with 10 μM carboxyfluorescein succinimidyl ester (CFSE) stain (BD Biosciences, Franklin Lakes, NJ, USA) and 1 μg/mL Hoechst 33342 dye (Biotium, San Francisco, CA, USA) for 1 h. Coverslips were washed three times with PBS and fixed in 4% paraformaldehyde (PFA) (AppliChem, Darmstadt, Germany) in PHEM solution [25 mM HEPES, 10 mM EGTA (Merck Millipore, Burlington, MA, USA), 60 mM PIPES, 2 mM MgCl2 (Applichem) pH 6.9], for 12 min at RT, followed by three washes with PBS. Finally, coverslips were mounted in mowiol 4-88 (AppliChem) medium.

Image acquisition was performed on a customized Andor Revolution Spinning Disk Confocal system (Yokogawa CSUX1; Yokogawa, Tokyo, Japan), built around an Olympus IX81 (Olympus Shinjuku, Tokyo, Japan), with 60x 1.42NA oil lens (UPlanXApo; Olympus Shinjuku, Tokyo, Japan) and a digital camera (Andor Zyla 4.2 sCMOS; Andor Technology Ltd., Belfast, Northern Ireland). The system was controlled by Andor IQ3.6 software (Andor Technology). Images were acquired as z-stacks with a z-step of 0.5 μm, through the entire volume of the biofilm. For each image, the maximum projection of z-stacks was generated, and the background was subtracted using a custom script in ImageJ (National Institute of Health, United States).

Gene expression analysis at the transcriptome level was employed to examine changes in the production of biofilm-formation-related genes after exposure to CFCS. To this aim, 10⁸ CFU/mL of fresh O/N cultures of the pathogenic bacteria were seeded in a 100 mm well. Adjusted CFCS (50% v/v, pH 6.2) was added simultaneously, and the cells were left to incubate for 24 h. The next day, the bacterial suspension was discarded, and the adhered bacteria were washed with PBS. Then, biofilms were scraped off, and the resulting suspension was centrifuged at 8,000 × g for 5 min. S. aureus cells were disrupted using a lysis buffer containing 1 M Tris-Cl (pH 8), 0.5 M EDTA (pH 8), 10% v/v Triton-X100 and 100 mg/mL lysozyme. S. enterica and E. coli were resuspended at the same lysis buffer, without the addition of lysozyme, and were incubated at 37°C for 30 min. Subsequently, cells were lysed by sonication. Finally, 1 mL of ice-cold Trizol (Sigma-Aldrich, Saint Louis, MO, USA) was added to each sample. RNA extraction was performed according to the manufacturer’s instructions. RNA integrity was investigated with agarose gel electrophoresis. 100 ng of whole RNA was used as a substrate for cDNA synthesis (Invitrogen, Waltham, MA, USA). Gene expression analysis was performed using Real-time PCR. Briefly, reactions were performed on a StepOne PCR System in MicroAmp^® Fast Optical 48-Well Reaction Plates (both from Thermo Fisher Scientific), using the KAPA SYBR^® FAST qPCR Kit (Kapa Biosystems, Wilmington, MA, USA), following manufacturer’s instructions. Each reaction mixture (20 μL) consisted of 10 μL SYBR Premix, 0.4 μL of forward and reverse primer, 2 μL cDNA and 7.2 μL ddH₂O. The PCR conditions were: 95°C for 3 min followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min. All reactions were performed in duplicates and each experiment included two non-template, negative controls for each primer used. For the relative quantification of gene expression, the formula RQ=2^-ΔΔCt was used. Primer sequences are shown in Table 1 .

To investigate the genetic basis of the antimicrobial potential of the strains of interest, a comprehensive bioinformatic analysis with established predictive algorithms was performed. Putative bacteriocin clusters were predicted using BAGEL4 (van Heel et al., 2018), BLASTp and local BLASTp+ (Camacho et al., 2009). More specifically, these algorithms were used to identify homologous genes to putative bacteriocin peptides, immunity proteins, transporters, and pheromone peptides. Additionally, the database Bactibase was utilized to mine sequences of functionally characterized bacteriocins (Hammami et al., 2010). Amino acid sequences of closely related species were aligned using CLUSTALW (Thompson et al., 1994) and T-COFFEE (Notredame et al., 2000) and phylogenetic trees were built using MEGA-X (version 11) (Tamura et al., 2021). More specifically, the aligned sequences were used as input to construct a tree with the Neighbor-Joining method and bootstrap replication values of 1000. The publicly available iTol server (Letunic and Bork, 2016) was used for the visualization of the resulting phylogenetic tree. The prediction of the topology and physicochemical properties of proteins was performed using DeepTHMM (Hallgren et al., 2022) and ProtParam (Gasteiger et al., 2005), respectively. InterPro (Blum et al., 2021) and Signal 6.0 (Teufel et al., 2022) were employed to investigate the family of the identified proteins, related motifs, and N-terminal signals. Structure predictions were performed using CollabFold, the publicly available version of AlphaFold2 (Mirdita et al., 2022) and Chimera version 1.16 was employed for structure superimpositions (Pettersen et al., 2004). Finally, assignment into Kyoto Encyclopedia of Genes and Genomes (KEGG) orthologous (KO) groups (Kanehisa et al., 2016) was performed to annotate pathways involved in the antimicrobial phenotype.

For the statistical analysis of the experimental data, Student’s T-test was performed using GraphPad PRISM 9 (GraphPad Software Inc., CA, USA). All experiments were performed in triplicate unless otherwise stated. Results are represented as mean ± standard deviation. A p-value < 0.05 was considered statistically significant.

The WGS of Lc. paracasei SP5, Lp. pentosus L33, Lp. plantarum L125 is available at the DDBJ/ENA/GenBank under the accession numbers: JAHKRU000000000.1, JAIGOE000000000.1 and JAKJPP000000000.1, respectively.

This study was approved (24-10/12/2022) by the Institution Review Board of the General University Hospital of Alexandroupolis, Greece.

A preliminary examination of the antimicrobial potential of Lp. pentosus L33, Lp. plantarum L125 and Lc. paracasei SP5, against clinical isolates of S. aureus, S. enterica ser. Enteritidis and E. coli, was performed, using the agar well diffusion assay. It was shown that the three strains inhibited the growth of the pathogens with variable efficiency ( Figures 1A, B ). The most profound effect was induced by Lc. paracasei SP5 against E. coli in the well diffusion assay, while in suspension Lc. paracasei SP5 was most effective against S. aureus ( Figure 1C ). In this context, all putative probiotic strains exerted antimicrobial effects after 24 h co-incubation with S. aureus or E. coli (p < 0.05), in suspension. The highest inhibitory activity against S. aureus was induced by co-incubations with Lp. pentosus L33 or Lp. plantarum L125 (~1.5 log reduction), while co-incubation of E. coli with Lp. pentosus L33, limited pathogen viability by ~1.7 log. Of note, S. enterica was resistant to Lp. pentosus L33 and Lc. paracasei SP5 co-incubation, whereas a significant reduction was observed with Lp. plantarum L125.

Next, we determined lactobacilli-pathogen competition for adherence on HT-29 cell monolayers. It was found that Lc. paracasei SP5 significantly limited adherence of all pathogens on the eukaryotic monolayers (p < 0.05, Figures 2A, C, E ). On the contrary, Lp. pentosus L33 promoted the adhesion of S. aureus on HT-29 cells ( Figure 2A ). Next, we sought to determine the co-aggregation capacity of lactobacilli with the three pathogens. As shown in Figures 2B, D, F , Lc. paracasei SP5 demonstrated similar co-aggregation efficiency with all pathogenic bacteria, while Lp. pentosus L33 and Lp. plantarum L125 showed higher preference for the gram-positive S. aureus ( Figure 2B ). Lp. plantarum L125 exhibited the highest auto-aggregation capacity among the lactobacilli. The auto-aggregation capacity of the three lactobacilli and pathogens is presented in Supplementary Figure 1 .

The antimicrobial capacity of metabolites produced by the lactobacilli was tested against planktonic cells of S. aureus, S. enterica and E. coli. More specifically, pathogens were treated with CFCS at native pH (4.2) or adjusted pH (6.2). As shown in Figures 3A–C , native CFCS from all tested lactobacilli, reduced pathogen viability (p < 0.01), compared to control (MRS adjusted at pH 4.2). Similarly, adjusted CFCS derived from Lc. paracasei SP5 and Lp. plantarum L125, retained its antimicrobial activity against all pathogens, while treatments with CFCS derived from Lp. pentosus L33, inhibited the proliferation of S. aureus and E. coli, showing no significant effect against S. enterica.

Heat-treated CFCS (pH 4.2) was utilized to determine the stability of the bioactive compounds responsible for the antimicrobial effects at denaturing temperature. It was found that the antimicrobial activity of Lp. pentosus L33 heat-treated CFCS was not significantly influenced ( Figures 3D–F ). On the other hand, heat treated CFCS derived from Lp. plantarum L125 limited pathogen viability significantly less than non-treated CFCS. Interestingly, heat-treated Lc. paracasei SP5 CFCS retained its antimicrobial activity against S. enterica and E. coli, exhibiting decreased activity against S. aureus (p < 0.05).

Next, we determined the ability of CFCS-derived metabolites to inhibit biofilm formation of pathogens after co-incubation for 24 h at static conditions, utilizing a CFU determination assay ( Figure 4 ), confocal microscopy ( Figures 5 - 7 ), and the crystal violet assay ( Supplementary Figure 2 ). S. aureus biofilm viability was significantly impaired by Lc. paracasei SP5 CFCS treatment (p < 0.05, Figure 4A ), while it was only marginally decreased by Lp. pentosus L33 and Lp. plantarum L125. On the other hand, biofilm viability of S. enterica and E. coli was significantly reduced by treatments with CFCS derived from all three lactobacilli ( Figures 4B, C ). The most prominent effect was induced by Lc. paracasei SP5 against S. enterica, reaching a 3-log reduction (p < 0.05) ( Figure 4B ). Confocal microscopy provided visual evidence of the inhibitory activity of the tested CFCS on biofilm formation, as demonstrated in Figures 5 – 7 . Biofilm biomass was also significantly reduced after treatments with CFCS, derived from the lactobacilli ( Supplementary Figure 2 ).

To delve deeper into the ability of CFCS to inhibit biofilm formation, we determined the possibility of the lactobacilli-derived metabolites to modulate the expression levels of biofilm-related genes at the transcriptome level, after 24 h of co-incubation. The selected genes modulate adhesion of pathogens on hydrophobic surfaces and auto-aggregation, as well as the production and transport of polysaccharides that comprise the protective extracellular capsule of biofilms ( Table 1 ). Concerning S. aureus, CFCS derived from all three strains significantly lowered the expression of enolase, an adhesion-related moonlighting protein that mediates attachment on inorganic and organic surfaces, inducing no significant effect on the expression levels of the biofilm formation regulator complex icaA/icaD or the FnbpA adhesin ( Figure 8A ). In the case of S. enterica, significant downregulation on the expression of csgA, a gene involved in the production of the major curlin subunit, was recorded, after treatments with CFCS derived from Lp. pentosus L33 and Lp. plantarum L125 ( Figure 8B ). Concomitantly, the same treatments induced a significant upregulation of csgD and cpxR. Both genes code for transcriptional regulators responsible for adhesion on hydrophobic surfaces and biofilm formation ( Table 1 ). Incubation of E. coli with CFCS derived from Lp. pentosus L33 upregulated the production of luxS, involved in quorum sensing, and of the adhesin-coding gene cpxA ( Figure 8C ) and downregulated the negative biofilm formation regulator, csrA. Lp. plantarum L125 CFCS upregulated pgaA, a gene that stimulates the production of the exopolysaccharide biofilm matrix.

Comparative genomics and annotation algorithms were utilized to expand the bioinformatic analysis on the antimicrobial capacity of the putative probiotic strains. The KEGG database was utilized to investigate the presence of non-proteinaceous compounds that could contribute to pathogen inhibition. More specifically, genes coding for enzymes regulating the production of antimicrobial metabolites, including lactic acid (L-/D- lactate dehydrogenase), ethanol (decarboxylase, alpha-acetolactate decarboxylase, diphosphomevalonate decarboxylase), hydrogen peroxide (NADH oxidase, multicopper oxidase), were identified. Interestingly, competence gene clusters, responsible for nucleic acid uptake after cell disruption were also found conserved in the genome of Lp. pentosus L33, Lp. plantarum L125 and Lc. paracasei SP5 ( Figure 9A ).

Furthermore, we sought to detect genes coding for antimicrobial proteins and/or peptides in the genome of the strains. To this aim, analysis with BAGEL4 revealed clusters containing putative bacteriocins ( Supplementary Figure 3 ). To further determine the functionality of peptides and their accessory proteins, comparative analysis with characterized peptides, ensued. More specifically, Lp. pentosus L33 carries two loci, whose structure resembles that of bacteriocin clusters ( Supplementary Figure 3A ). One of these clusters includes genes for Plantaricin NC8 chains alpha and beta (core peptides) and Plantaricin A, the pheromone regulating the expression of the cluster. Downstream to these loci, a bacteriocin immunity protein (orf00024) and genes coding for proteins responsible for GG-leader motif cleavage and transport (orf00047, orf00048, orf00053), were identified. Sequence analysis of the core peptides revealed that chain alpha carries a SxxxG motif and chain beta a GxxxG motif possibly involved in the dimerization of the two partners. To predict the functionality of the peptides, Plnc8a and Plnc8b were aligned against the functionally characterized sequences of strain Lp. plantarum NC8 (Maldonado et al., 2003). The mature peptides (cleaved at the GG motif) of Plnc8a share 100% sequence and structural conservation. Similarly, Plnc8b mature peptides share a 79.31% sequence identity and structural conservation ( Figures 9B, C ). Of note, Plnc8b encoded by Lp. pentosus L33 carries only one dimerization motif, while its homologous sequence contains two. Thus, dimer formation for the active form of the peptide could be impaired. Concerning the physicochemical properties of the peptides, Plnc8a weighs 3.5 kDa and Plnc8b 3.4 kDa with a theoretical pI of 9.87 and 9.7, respectively ( Figure 9I ).

A cluster coding for Plantaricin EF was identified in the genome of Lp. plantarum L125 ( Supplementary Figure 3B ). This locus also contains a gene to produce the pheromone Plantaricin K. Furthermore, it contains an immunity protein (plnL; orf00025) and other transport/immunity proteins (orf00002, HlyD, orf00016). Sequence analysis of the core peptides revealed the presence of the dimerization motif GxxxG in both mature peptides. To predict their functionality, sequences derived from Bactibase of Lp. plantarum C11 (Diep et al., 1996), were used. It was found, that the mature PlnE and PlnF peptides coded by Lp. plantarum L125, present 100% sequence and structural identity with the respective peptides derived from Lp. plantarum C11 ( Figures 9D, E ). PlnE is a 3.5 kDa protein, with a theoretical pI of 11.57, and PlnF weighs 3.7 kDa, with a theoretical pI of 10.27 ( Figure 9I ).

Lc. paracasei SP5 harbors multiple core peptides with putative antimicrobial activity ( Supplementary Figure 3C ). More specifically, the putative bacteriocin LSEI 2386 was annotated using BAGEL4 alongside putative transport and immunity proteins LanT, and Enterocin A immunity proteins. LSEI 2386 (2.5 kDa, pI 8.34) presents 100% sequence and structural conservation to the respective peptide annotated in the genome of Lc. casei ATCC 334 (Kuo et al., 2013), now belonging to the Lc. paracasei species (Zheng et al., 2015). To determine the presence of proteins specific for the modification and export of the LSEI 2386 peptide, sequences derived from the same strain were queried against the WGS of Lc. paracasei SP5. Genes coding for an ABC transporter (LSEI 2384) and an accessory secretion protein (LSEI 2381) were annotated upstream of the core peptide with 99% sequence identity ( Figures 9F–H ). Furthermore, the putative bacteriocin peptides LSEI 2392 (4.28 kDa, pI 4.64) and LSEI 2393 (4.6 kDa, pI 9.99), were also identified in the genome of the strain, presenting 98% and 100% sequence identity with the characterized peptides coded by Lc. casei ATCC334 (Kuo et al., 2013). The accessory secretion protein LSEI 2389 was also annotated upstream of the genes coding for the core peptides, with 100% identity with the sequence derived from Lc. casei ATCC334. Lastly, two genes coding for N-acetylmuramoyl-L-alanine-amidase, an enzyme responsible for peptidoglycan degradation were pinpointed in the genome of Lp. paracasei SP5. Although BAGEL4 annotated these loci as Enterolysin A, further analysis revealed greater similarity with the enzyme N-acetylmuramoyl-L-alanine-amidase coded by closely related lactobacilli (> 90%). Furthermore, domain analysis with InterPro showed that both annotated sequences carry an amidase domain, including a catalytic site, a substrate binding site and a Zinc binding site, characteristic of this enzyme family. Both proteins weigh around 31 kDa, possessing a theoretical pI of 5.41 ( Figure 9I ). The putative bacteriocin peptides annotated in the genome of the novel strains, and homologous proteins from other lactobacilli were aligned and utilized for the construction of an unrooted tree, using the neighbor joining method and 1000 bootstrap replications ( Figure 10A ). Furthermore, the structure of the putative bacteriocin pro-peptides was predicted using AlphaFold2, and their tertiary structure was superimposed to that of homologous proteins, whose function was experimentally validated ( Figure 10B ).

Finally, the genome sequences of the three LAB strains were scoured for the presence of genes involved in quorum quenching. The Uniprot database was searched for annotated sequences coding for N-acyl homoserine (AHL) lactonase, AHL acylases and AHL oxidoreductases. Among these enzymes, only AHL lactonases were characterized in lactobacilli. Thus, two sequences (A0A7Z2PEZ0 and A0A0M3QBV3) identified in Lp. plantarum were queried against the genome of the strains. It was found that Lp. plantarum L125 possesses one locus presenting sequence identity of 100% and 98% with the queries, while Lp. pentosus L33 also codes for a homologous gene, that however, presents lower percentage identity (77% and 78%) ( Supplementary Table 1 ).