Screening of Lactic Acid Bacteria Isolated from Foods for Interference with Bacterial Quorum Sensing Systems ^†

Abstract

Quorum sensing (QS) is a cell-to-cell communication mechanism through which microorganisms can sense their population density and adjust their physiology by producing and detecting small signaling molecules called autoinducers (AIs). QS influences various aspects of microbial physiology, including virulence and pathogenesis by bacterial pathogens, biofilm formation, sporulation, antimicrobial resistance, etc. Lactic acid bacteria (LAB) have been used for centuries in food fermentation to improve sensory and nutritional profiles and preserve against spoilage and pathogenic microflora. This study investigated the potential of foodborne LAB of various genera, including Lactococcus, Lactobacillus, Leuconostoc, Streptococcus, and Enterococcus, to interfere with the QS system of bacterial pathogens. For this, cell-free supernatants (CFSs) of 89 LAB foodborne isolates were collected by centrifugation following a 20 h culture (at 30 °C) in quarter-strength Brain Heart Infusion (BHI) broth. The pH of all CFSs was adjusted to 6.5 and sterilized by filtration. The anti-QS activity of the sterilized and neutralized CFSs was initially screened using the biosensor strains Chromobacterium violaceum 026 and Agrobacterium tumefaciens NTL4 (pZLR4) through an agar well diffusion assay that can detect the inhibition of the QS system that is based on acylated homoserine lactones (AHLs), which are used as AIs by Gram-negative bacteria. Additionally, all the CFSs were also screened for interference with the autoinducer 2 (AI-2) QS system that is mostly used for interspecies communication by both Gram-positive and Gram-negative bacteria. This was assessed using a luminescence bioassay with the Vibrio harveyi BAA-1117 biosensor strain. The results indicate that none of the LAB CFSs could inhibit AHL-based QS. However, 61.8% (55/89) of the CFSs induced luminescence in V. harveyi BAA-1117, while the remaining 38.2% (34/89) of the samples were capable of inhibiting AI-2-based QS. In the next steps, the most representative of these latter AI-2 interfering LAB isolates will be investigated for possible inhibition of biofilm formation by some important foodborne bacterial pathogens.

1. Introduction

Many bacterial pathogens rely on quorum sensing (QS) to induce pathogenicity, including adhesion to host cells, invasion, and biofilm formation, by producing and detecting small signaling molecules called autoinducers (AI) [1]. QS systems vary among microorganisms, with AI-2 (produced via the luxS homologous genes) being a widespread signaling molecule in both Gram-negative and Gram-positive bacteria [2]. AI-2 mediates diverse phenotypes, serving as a universal language for interspecies bacterial communication [3,4,5]. In addition to the AI-2 QS system, Gram-negative bacteria typically use acylated homoserine lactones (AHLs) in LuxI/LuxR-type QS systems (AI-1) [4]. On the other hand, in Gram-positive bacteria, QS relies on modified oligopeptides (AIPs) and two-component histidine kinase systems [6]. In general, the various described QS mechanisms have been previously found to regulate biofilm formation and virulence in some important foodborne pathogens like Salmonella enterica, Campylobacter jejuni, Listeria monocytogenes, and Staphylococcus aureus [7,8].

Not surprisingly, quorum quenching (QQ) has emerged as a promising antibiofilm strategy, with both natural (e.g., halogenated furanones, lactonases) and synthetic QS inhibitors (QSIs) shown to disrupt QS systems without affecting planktonic cell growth, reducing this way the risk of resistance development [9,10]. Nowadays, the QQ approach is indeed gaining recognition as a promising alternative to conventional antimicrobial practices in industrial and clinical settings [11].

Lactic acid bacteria (LAB), such as Lactococcus, Lactobacillus, and Pediococcus, and their metabolites are not only important in food fermentations but have shown promise in inhibiting or even disrupting foodborne pathogenic bacterial biofilms [12]. Interestingly, it has been found that LAB metabolism may interfere with QS and the virulence of bacterial pathogens while posing less risk for resistance development by the latter. Thus, several LAB species have been shown to produce metabolites with QQ activity, inhibiting not only biofilm formation by bacterial pathogens but also suppressing some of their other virulence traits. For example, Lactobacillus sakei from kimchi reduced AI-2 production, motility, biofilm formation, and virulence gene expression in Escherichia coli O157:H7 without affecting bacterial viability [13]. Similarly, Weissella viridescens and Weissella confusa isolated from fermented grapes significantly reduced Salmonella biofilm formation and AI-2 activity [14]. These representative previous findings highlight LAB’s potential as biofilm inhibitors targeting QS in foodborne pathogens.

Surely, the exploitation of LAB and/or their metabolites to combat detrimental biofilms within the food industry consists of an eco-friendly approach that may offer a cost-effective alternative to synthetic antimicrobial chemicals in enhancing food safety [15]. As a primary stage, the objective of this study was to investigate the potential of foodborne LAB isolates to interfere with the QS of bacterial pathogens. For this, sterilized and pH-neutralized cell-free supernatants (CFSs) of 89 LAB isolates of various genera were screened for their content in metabolites with the ability to inhibit either AI-1 or AI-2 QS systems. The existence of AI-2 interspecies signaling molecules in CFSs was also sought. All the analyses were conducted using classical biosensor-based assays.

2. Methods

2.1. Bacterial Strains and Culture Conditions

Eighty-nine foodborne LAB isolates were used in the present study. These were kept frozen (at −80 °C) in the microorganisms collection of the Laboratory of Food Microbiology and Hygiene (LFMH) of our department (Department of Food Science and Nutrition; DFSN). To resuscitate the LAB isolates, these were grown in de Man Rogosa Sharpe (MRS) broth (Condalab, Madrid, Spain) at 30 °C for 24 to 48 h (until visible turbidity appeared in the broth). Subsequently, they were cultivated twice in quarter-strength Βrain Heart Ιnfusion (BHI) broth (Lab M, Bury, UK) at 30 °C with agitation (160 rpm) for 24 and 20 h, respectively. For the AI-2 activity bioassays, the Vibrio harveyi BAA-1117 strain (sensor 1⁻, sensor 2⁺), which specifically detects AI-2, and the V. harveyi BAA-1119 strain which continuously produces AI-2, were utilized. Before their use, the two V. harveyi strains were grown twice in an autoinducer bioassay (AB) broth at 30 °C with agitation (160 rpm). The AB medium was prepared in accordance with the methodology outlined by Lu et al. (2004) [16]. The working cultures of the two other biosensor strains, Chromobacterium violaceum 026 and Agrobacterium tumefaciens NTL4 (pZLR4), were used to check for anti-AHL action of CFSs against either short or long-chain signaling molecules, respectively, and were prepared in accordance with the methodology previously outlined by Singh et al. (2016) [17].

2.2. Preparation of Neutralized and Sterilized LAB CFSs

The CFS of each LAB isolate was obtained by removing the cells from its 20 h working culture (in quarter-strength BHI broth) by centrifugation (at 4000× g for 10 min at 4 °C). Subsequently, the pH of each CFS was adjusted to 6.5 (with 5 N NaOH), using a digital pH meter equipped with a glass electrode (Consort C931, Turnhout, Belgium) and the CFS was then filter sterilized by passing it through 0.22-μm-pore-size filters (Labbox Labware S.L., Barcelona, Spain). Neutralized and sterilized CFSs were stored at −80 °C until screened for QSI and/or AI-2 producing activities. For each LAB isolate, two biological replicates of CFSs were prepared.

2.3. Screening of LAB CFSs for Inhibition of AHL-Based QS System

The anti-AHL QS activities of LAB CFSs were evaluated using the biosensor strains C. violaceum 026 and A. tumefaciens NTL4 (pZLR4) through agar well diffusion assays as previously described by Singh et al. (2016) [17].

2.4. Screening of LAB CFSs for Interference with AI-2-Based QS System

The LAB CFSs were screened for their AI-2 content according to the bioluminescence method of Surette and Bassler (1998) [18]. For this, an overnight (16 h) culture of V. harveyi BAA-1117 was diluted 1:5000 in fresh AB medium and 90 μL of this suspension were then mixed with 10 μL of each test sample (i.e., LAB CFC) in a 96-well polystyrene (PS) microtiter plate (Cell Culture Plate, white, 85.4 × 127.6 mm, flat bottom, SPL Life Sciences, Pocheon-si, Gyeonggi-do, Republic of Korea). Negative and media controls included sterile AB medium and quarter-strength BHI broth, respectively. The positive control was CFC from the V. harveyi BAA-1119 strain prepared in AB medium. Microtiter plates were incubated at 30 °C, and luminescence in each well was measured every 30 min using a Synergy HT microplate reader (Winooski, VT, USA) until the negative control showed a considerable increased luminescence. Relative AI-2-like activity (relative light units; RLU) was calculated as the ratio of luminescence of the test sample (LAB CFC) to that of the negative control sample.

The LAB CFSs were also screened for their content in metabolites with the ability to inhibit the AI-2 QS system. For this, equal volumes (5 μL) of each LAB CFS and the CFS from the known AI-2 producer V. harveyi BAA-1119 strain were combined, and the resulting mixture was assayed for AI-2 activity to determine whether the tested LAB CFS could inhibit luminescence induction in the V. harveyi BAA-1117 reporter strain. As a positive control in this assay, equal volumes (5 μL) of AB medium and the CFS from the V. harveyi BAA-1119 were combined and used in the assay. The inhibition of the AI-2-like activity was expressed as a percentage of luminescence relative to the corresponding positive control: 100 − [RLU_sample/RLU_{positive control} × 100] [16].

3. Results and Discussion

3.1. Ability of LAB CFSs to Inhibit AHL-Based QS

None of the 89 LAB CFSs could inhibit violacein production or the blue-green pigment production (in the C. violaceum 026 and A. tumefaciens NTL4 (pZLR4) bioassays, respectively). To improve the diffusion of the CFSs into the solid medium, lower agar concentrations (0.5%, 0.8%, and 1.0% w/v) were also tested, but no inhibition zones were observed. The CFSs of the LAB cultures were also re-prepared by replacing ¼ BHI broth with MRS broth as the growth medium. Once more, no inhibition zones were observed.

The lack of AHL-QS inhibition in this study contrasts with prior findings where certain LAB strains have been shown to possess such activity [19,20]. This discrepancy could be attributed to differences in strain specificity, experimental procedures (e.g., method used to extract LAB metabolites), and the different molecular structures of the specific metabolites produced [21]. Additionally, the absence of such activity could probably indicate that the LAB-derived metabolic compounds may preferentially target AI-2-mediated interspecies communication rather than intraspecies AHL signaling.

3.2. Ability of LAB CFSs to Interfere with AI-2 QS

Table 1. Relative AI-2-like activity (RLU) of the CFSs of the 89 LAB isolates. Values represent mean values ± standard deviations. The 55 isolates whose CFS had the ability to induce AI-2 QS of the reporter V. harveyi BAA-1117 strain are shown in bold (RLU ≥ 10). The strain code LFMH_B86 refers to AI-2 producer V. harveyi BAA-1119 strain (used as positive control).

s/n	LAB’s Code	Relative AI-2-like Activity (RLU)
1	DFSN_B42	25.86 ± 7.38
2	DFSN_B43	8.81 ± 7.63
3	DFSN_B44	1.60 ± 0.10
4	DFSN_B50	32.45 ± 14.48
5	DFSN_B51	48.74 ± 9.40
6	DFSN_B54	9.11 ± 2.42
7	DFSN_B55	24.01 ± 10.94
8	DFSN_B58	25.67 ± 8.57
9	DFSN_B63	22.64 ± 11.75
10	DFSN_B64	43.08 ± 11.33
11	LFMH_B1	29.89 ± 5.26
12	LFMH_B2	2.85 ± 2.18
13	LFMH_B3	33.39 ± 5.03
14	LFMH_B6	27.70 ± 9.71
15	LFMH_B7	18.36 ± 8.69
16	LFMH_B8	112.21 ± 57.48
17	LFMH_B9	83.76 ± 37.71
18	LFMH_B10	2.89 ± 1.27
19	LFMH_B11	54.5 ± 20.87
20	LFMH_B14	67.22 ± 37.13
21	LFMH_B16	35.25 ± 11.80
22	LFMH_B17	1.50 ± 0.79
23	LFMH_B18	35.47 ± 25.01
24	LFMH_B19	1.98 ± 1.41
25	LFMH_B20	25.24 ± 23.48
26	LFMH_B21	13.90 ± 3.86
27	LFMH_B23	60.48 ± 42.62
28	LFMH_B24	71.34 ± 54.09
29	LFMH_B25	6.56 ± 2.31
30	LFMH_B26	8.98 ± 5.91
31	LFMH_B29	1.82 ± 0.79
32	LFMH_B30	229.93 ± 99.35
33	LFMH_B31	120.98 ± 82.07
34	LFMH_B32	109.17 ± 55.58
35	LFMH_B33	71.95 ± 23.98
36	LFMH_B34	2.22 ± 1.93
37	LFMH_B35	2.04 ± 1.25
38	LFMH_B36	2.82 ± 1.26
39	LFMH_B37	107.67 ± 53.55
40	LFMH_B38	15.36 ± 7.85
41	LFMH_B39	198.56 ± 71.93
42	LFMH_B40	132.50 ± 82.22
43	LFMH_B41	38.58 ± 23.73
44	LFMH_B42	2.32 ± 0.81
45	LFMH_B43	2.09 ± 1.50
46	LFMH_B44	1.19 ± 0.33
47	LFMH_B45	1.41 ± 0.40
48	LFMH_B46	1.14 ± 0.59
49	LFMH_B47	1.40 ± 0.41
50	LFMH_B48	43.35 ± 28.19
51	LFMH_B49	55.34 ± 37.16
52	LFMH_B50	94.86 ± 16.88
53	LFMH_B51	155.10 ± 69.73
54	LFMH_B52a	1.33 ± 0.10
55	LFMH_B52b	136.51 ± 73.09
56	LFMH_B53	47.05 ± 31.94
57	LFMH_B54	1.47 ± 0.32
58	LFMH_B55	16.13 ± 1.13
59	LFMH_B56	42.24 ± 23.22
60	LFMH_B57a	1.88 ± 0.69
61	LFMH_B57b	62.79 ± 32.76
62	LFMH_B58	2.58 ± 2.82
63	LFMH_B59	1.37 ± 0.26
64	LFMH_B60	74.83 ± 14.64
65	LFMH_B61	41.44 ± 16.61
66	LFMH_B62	39.44 ± 11.26
67	LFMH_B63a	1.16 ± 0.47
68	LFMH_B63b	43.41 ± 16.42
69	LFMH_B64	55.74 ± 19.62
70	LFMH_B65a	2.22 ± 1.59
71	LFMH_B65b	1.46 ± 0.38
72	LFMH_B66	1.51 ± 0.55
73	LFMH_B67	114.18 ± 81.22
74	LFMH_B68	353.23 ± 66.57
75	LFMH_B69	1.89 ± 0.68
76	LFMH_B70	101.15 ± 46.30
77	LFMH_B71	42.07 ± 15.14
78	LFMH_B72	1.70 ± 0.41
79	LFMH_B73	34.89 ± 20.80
80	LFMH_B74	105.12 ± 28.38
81	LFMH_B75	1.62 ± 0.13
82	LFMH_B76	1.59 ± 0.33
83	LFMH_B77	82.71 ± 38.69
84	LFMH_B78	2.57 ± 1.17
85	LFMH_B79a	1.86 ± 0.06
86	LFMH_B79b	71.41 ± 13.73
87	LFMH_B81	63.20 ± 22.02
88	LFMH_B82	45.83 ± 25.60
89	LFMH_B83	49.13 ± 12.42
	LFMH_B86	197.77 ± 118.57

presents the relative AI-2-like activity for each LAB CFS. Sixty-two percent (55/89) of the CFSs could induce luminescence by the V. harveyi BAA-1117 (sensor 1⁻, sensor 2⁺) with RLU ≥ 10, while the rest of the CFSs (34/89) did show only a low signal induction (RLU < 10).

Table 2. Relative AI-2 activity (RLU) of the CFSs of the selected 34 LAB isolates and the percentage (%) inhibition of AI-2-like activity. Values represent mean values ± standard deviations. The 25 isolates whose CFS had the ability to inhibit more than 90% the biolumiscence of the reporter V. harveyi BAA-1119 (positive control) strain are shown in bold.

s/n	LAB’s Code	Relative AI-2-like Activity (RLU)	% Inhibition of AI-2-like Activity
1	DFSN_B43	12.82 ± 4.12	87.48 ± 4.81
2	DFSN_B44	13.58 ± 2.90	84.09 ± 3.40
3	DFSN_B54	31.26 ± 14.96	63.37 ± 14.53
4	LFMH_B2	7.99 ± 1.93	92.09 ± 2.04
5	LFMH_B10	8.35 ± 6.37	90.94 ± 8.10
6	LFMH_B17	6.89 ± 0.60	91.92 ± 0.70
7	LFMH_B19	8.39 ± 0.26	90.17 ± 0.31
8	LFMH_B25	5.13 ± 0.92	96.22 ± 0.68
9	LFMH_B26	21.50 ± 0.39	84.14 ± 0.29
10	LFMH_B29	7.52 ± 1.94	94.45 ± 1.43
11	LFMH_B34	6.29 ± 3.05	89.56 ± 5.06
12	LFMH_B35	3.80 ± 0.81	97.20 ± 0.60
13	LFMH_B36	5.46 ± 0.39	95.97 ± 0.29
14	LFMH_B42	4.91 ± 1.39	96.38 ± 1.02
15	LFMH_B43	7.57 ± 5.94	94.41 ± 4.39
16	LFMH_B44	2.35 ± 0.03	98.27 ± 0.02
17	LFMH_B45	1.84 ± 0.32	98.71 ± 0.34
18	LFMH_B46	2.15 ± 1.41	98.46 ± 1.11
19	LFMH_B47	9.33 ± 9.79	93.11 ± 7.23
20	LFMH_B52a	5.56 ± 1.98	94.59 ± 1.93
21	LFMH_B54	4.89 ± 0.16	95.25 ± 0.16
22	LFMH_B57a	5.18 ± 0.02	94.97 ± 0.02
23	LFMH_B58	7.59 ± 2.10	91.11 ± 2.46
24	LFMH_B59	5.66 ± 3.16	93.37 ± 3.71
25	LFMH_B63a	7.76 ± 5.25	90.91 ± 6.15
26	LFMH_B65a	8.05 ± 1.79	90.57 ± 2.09
27	LFMH_B65b	10.81 ± 4.77	88.97 ± 4.87
28	LFMH_B66	7.01 ± 1.14	91.78 ± 1.33
29	LFMH_B69	6.22 ± 2.22	92.09 ± 3.96
30	LFMH_B72	4.39 ± 3.66	94.89 ± 4.32
31	LFMH_B75	6.76 ± 1.23	92.08 ± 1.44
32	LFMH_B76	14.78 ± 2.34	84.93 ± 2.39
33	LFMH_B78	13.05 ± 4.99	84.71 ± 5.84
34	LFMH_B79a	6.85 ± 1.49	89.09 ± 6.41

shows the percentage of AI-2-like activity inhibition of the 34 LAB isolates that could not produce strong AI-2 signals (RLU < 10) when tested in the V. harveyi BAA-1117 AI-2 induction assay. These results highlight the robust AI-2 inhibitory capacity of most of these tested LAB CFSs. Thus, among them, they included the CFSs of 25 isolates (73.5%) that exhibited strong AI-2 inhibitory activity, achieving at least 90% inhibition of the luminescence of the positive control (V. harveyi BAA-1119 strain).

The robust AΙ-2 inhibitory activity of LAB isolates is consistent with findings in the literature. For instance, Park et al. (2014) reported that L. sakei from kimchi significantly reduced AI-2 production and inhibited biofilm formation and motility in E. coli O157:H7 [13]. Similarly, Pelyuntha et al. (2019) demonstrated that Weissella spp. from fermented grapes effectively reduced Salmonella biofilm formation and AI-2 signaling [14]. These results suggest that LAB metabolites, particularly those targeting AI-2 pathways, have significant potential as QQ agents. LAB’s ability to disrupt QS systems is increasingly recognized as an eco-friendly and sustainable approach to combat microbial virulence [21]. Unlike traditional antimicrobials, QSIs interfere with microbial communication without affecting cellular viability [22]. This latter seems to reduce the likelihood of resistance development and may offer a novel strategy for managing biofilms in the food industry. Metabolites derived from LAB have indeed shown promise in disrupting biofilm formation in foodborne pathogens such as L. monocytogenes, Salmonella spp., and S. aureus by interfering with QS mechanisms [23]. The LAB isolates whose CFS exhibited strong AI-2 inhibitory activity in the present study warrant further investigation to identify the specific metabolites responsible for this effect. Advanced analytical techniques such as mass spectrometry could probably be employed to identify these compounds. Furthermore, the most representative of these LAB isolates could be investigated in the next steps for their potential to inhibit biofilm formation by some important foodborne bacterial pathogens.

4. Conclusions

Foodborne LAB isolates showed significant potential to interfere with AI-2-mediated QS, with 28.1% (25/89) of them producing extracellular metabolites that showed strong inhibitory activity. Another 61.8% (55/89) of the LAB CFSs induced luminescence in V. harveyi BAA-1117 indicating that these should contain AI-2 signals. Although no inhibition of AHL-mediated QS was observed, the results underline the strain-specific nature of LAB extracellular metabolites and their preferential targeting of AI-2-based QS systems. These findings pave the way for the development and application of environmentally friendly antimicrobial strategies that may minimize the risk of resistance development.