Bacillaene Mediates the Inhibitory Effect of Bacillus subtilis on Campylobacter jejuni Biofilms

ABSTRACT

Biofilms are the predominant bacterial lifestyle and can protect microorganisms from environmental stresses. Multispecies biofilms can affect the survival of enteric pathogens that contaminate food products, and thus, investigating the underlying mechanisms of multispecies biofilms is essential for food safety and human health. In this study, we investigated the ability of the natural isolate Bacillus subtilis PS-216 to restrain Campylobacter jejuni biofilm formation and adhesion to abiotic surfaces as well as to disrupt preestablished C. jejuni biofilms. Using confocal laser scanning microscopy and colony counts, we demonstrate that the presence of B. subtilis PS-216 prevents C. jejuni biofilm formation, decreases growth of the pathogen by 4.2 log₁₀, and disperses 26-h-old preestablished C. jejuni biofilms. Furthermore, the coinoculation of B. subtilis and C. jejuni interferes with the adhesion of C. jejuni to abiotic surfaces, reducing it by 2.4 log₁₀. We also show that contact-independent mechanisms contribute to the inhibitory effect of B. subtilis PS-216 on C. jejuni biofilm. Using B. subtilis mutants in genes coding for nonribosomal peptides and polyketides revealed that bacillaene significantly contributes to the inhibitory effect of B. subtilis PS-216. In summary, we show a strong potential for the use of B. subtilis PS-216 against C. jejuni biofilm formation and adhesion to abiotic surfaces. Our research could bring forward novel applications of B. subtilis in animal production and thus contribute to food safety.

IMPORTANCECampylobacter jejuni is an intestinal commensal in animals (including broiler chickens) but also the most frequent cause of bacterial foodborne infection in humans. This pathogen forms biofilms which enhance survival of C. jejuni in food processing and thus threaten human health. Probiotic bacteria represent a potential alternative in the prevention and control of foodborne infections. The beneficial bacterium Bacillus subtilis has an excellent probiotic potential to reduce C. jejuni in the animal gastrointestinal tract. However, data on the effect of B. subtilis on C. jejuni biofilms are scarce. Our study shows that the B. subtilis natural isolate PS-216 prevents adhesion to the abiotic surfaces and the development of submerged C. jejuni biofilm during coculture and destroys the preestablished C. jejuni biofilm. These insights are important for development of novel applications of B. subtilis that will reduce the use of antibiotics in human and animal health and increase productivity in animal breeding.

INTRODUCTION

Campylobacter jejuni is the most common cause of bacterial foodborne infection in developed countries, resulting in high costs for the public health and economic sectors (1, 2). C. jejuni is typically found in animals (e.g., broiler chickens) (3) and contaminated meat products (4, 5). C. jejuni poses a serious threat to humans because it can adhere to, enter, and colonize the mucus layer of the human gastrointestinal tract (6). Such infection can cause various clinical manifestations, such as gastroenteritis, in the general population (7–9). Antimicrobial resistance is another emergent problem due to the increasing multidrug resistance of pathogenic Campylobacter species to antimicrobial drugs (1, 10–13). The development of novel strategies to reduce and eliminate C. jejuni in food is therefore essential (13).

Biofilms are multicellular structures and a dominant form of bacterial existence (14). Since multispecies biofilms can improve the survival of enteric pathogens, novel strategies are needed to control biofilm formation, especially in the poultry industry (15, 16). In particular, the ability of C. jejuni to adhere to surfaces and form mono- or multispecies biofilms promotes its persistence under stress conditions, which consequently enhances survival in the food chain (17–22). Recent studies indicate that interactions with other enteric pathogens may positively or negatively affect C. jejuni biofilm formation (23–26). There is evidence that Bacillus subtilis prevents biofilm formation of certain bacterial pathogens (27–34). However, data on the effect of B. subtilis on C. jejuni biofilms are scarce.

B. subtilis is a spore-forming probiotic that inhabits many natural habitats (35), including chickens (36, 37) and the human gastrointestinal tract (38, 39), and is known for its robust biofilm formation (40–42). B. subtilis synthesizes a rich arsenal of putative antimicrobial molecules (mainly antimicrobial peptides) against various pathogenic bacteria (43–47), including ribosomal synthesized peptides (bacteriocins) and nonribosomal peptides (lipopeptides, simple antibiotics, polyketides, or siderophores) (48). Previous in vitro and in vivo studies suggest that B. subtilis inhibits C. jejuni growth (46, 49–51) and provides additional benefits to broilers, such as increased populations of Lactobacillus spp. (52, 53) and improved growth performance (54) as well as improved mucosal immunity (55–57) and egg and meat quality (57, 58). It is therefore not surprising that at least 14 Bacillus species, including B. subtilis, with the status “Qualified Presumption of Safety” are used as additives (probiotics) in animal feed (59). However, little is known regarding the social interactions between probiotics and pathogens in biofilms and the mechanisms by which B. subtilis inhibits C. jejuni growth.

The aim of this study was to investigate the potential inhibitory effect of B. subtilis on C. jejuni growth and biofilm formation. We addressed this question by investigating the dynamics of multispecies submerged biofilm formation by using colony counts, visualizing the spatial distribution of B. subtilis and C. jejuni by confocal laser scanning microscopy (CLSM), testing coadhesion to abiotic surfaces, and testing the potential of B. subtilis to disrupt preestablished C. jejuni biofilms. The results showed that the B. subtilis natural isolate PS-216 prevents the development of submerged C. jejuni biofilm during coculture and destroys the preestablished biofilm. The transwell coculture experiments suggested that the inhibitory effect of B. subtilis does not entirely depend on direct cell contact with C. jejuni. Single-gene knockout mutants of B. subtilis PS-216 revealed that phosphopantetheinyl transferase (sfp) and genes involved in bacillaene production play important roles in preventing C. jejuni growth and biofilm formation. Overall, this study unravels the mechanism by which B. subtilis inhibits the growth of C. jejuni during biofilm formation.

RESULTS

PS-216 inhibits C. jejuni growth and its adhesion to polystyrene surfaces.

C. jejuni represents a major health threat to the general human population. A key issue in food production is transmission through abiotic surfaces on which biofilms can support the growth and survival of C. jejuni. It is known that C. jejuni adheres to abiotic surfaces and forms biofilms (18, 19, 22, 26). Therefore, we have tested whether B. subtilis reduces the adhesion of C. jejuni to polystyrene surfaces. We quantified attached cells after B. subtilis and C. jejuni mixed at a ratio of 1:1 were grown for 24 h at 42°C. The number (CFU/ml) of adhered C. jejuni cells in multispecies biofilm was significantly lower (4.9 log₁₀ CFU/ml) than that in monoculture (7.4 log₁₀ CFU/ml) [P₍₂₄₎ = 0.002]. As such, the presence of B. subtilis reduced the colony counts of C. jejuni by 2.4 log₁₀ (Fig. 1A). In contrast, the colony counts of adhered B. subtilis PS-216 were similar in monoculture and coculture [P₍₂₄₎ = 0.54] (see Fig. S1A in the supplemental material).

FIG 1.

Three experimental designs to evaluate interactions between B. subtilis PS-216 and C. jejuni in broth suspension during 24-h microaerobic static cultivation. (A) Adhesion of C. jejuni to an abiotic surface during mono- and coculture with B. subtilis. Results are presented as colony counts, and a schematic representation of the experiment is shown on the left. (B) C. jejuni growth was measured as colony counts at different time points during static mono- and coculture. Samples containing biofilm and broth were vortexed prior to plating (disruptive sampling). (C) The spatial distribution of C. jejuni (green) and B. subtilis (red) in mono- and coculture in a broth volume of 100 μl with a height of 1,600 μm and a surface of 319 μm by 319 μm was visualized by confocal microscopy (left panel); quantification of C. jejuni as colony counts during static mono- and coculture in the pellicle and broth below the pellicle is shown in the right panel. We were not able to observe green dot areas representing C. jejuni GFP-labeled cells in the biofilm by CLSM due to their low numbers. (D) C. jejuni growth was determined in a transwell system, with two species separated by a membrane (with a pore size of 0.1 μm). After 24 h in static mono- and coculture, the colony counts of C. jejuni in the broth were determined after the inner chamber was removed. Three biological and up to three technical repeats were performed. The error bars represent ±standard deviation of the mean value. * represents statistically significant values (two-sample t test; see supplemental material for details).

Furthermore, we investigated the potential of B. subtilis PS-216 to antagonize C. jejuni growth in static coculture inoculated at a ratio of 1:10 (B. subtilis to C. jejuni). The colony counts of each species were determined at specific times by plating the samples on selective media. B. subtilis significantly inhibited C. jejuni after 14 h [P₍₁₄₎ = 0.0046], 16 h [P₍₁₆₎ = 0.009], and 24 h [P₍₂₄₎ = 4 × 10⁻⁴] of coincubation. After 24 h of incubation at 42°C, the colony counts of C. jejuni in coculture with B. subtilis were significantly lower (6 log₁₀ CFU/ml) than that of C. jejuni in monoculture (8.8 log₁₀ CFU/ml) [P₍₂₄₎ = 4 × 10⁻⁴]; the colony counts of C. jejuni cells in the presence of B. subtilis were 2.8 log₁₀ lower (Fig. 1B). At earlier time points of coincubation, 10 h [P₍₁₀₎ = 0.22] and 12 h [P₍₁₂₎ = 0.14], inhibition was noted but not significant. C. jejuni had no effect on the growth of B. subtilis PS-216 [P₍₁₀₎ = 0.33; P₍₁₂₎ = 0.64; P₍₁₄₎ = 0.55; P₍₁₆₎ = 0.87; P₍₂₄₎ = 0.50] (Fig. S1B).

C. jejuni and B. subtilis biofilms segregate in coculture.

B. subtilis is a ubiquitous bacterium that is commonly present in the gastrointestinal tract and environment of poultry (36, 37), in which it can potentially interact and interfere with C. jejuni. However, the interspecies interactions between C. jejuni and B. subtilis in biofilms are poorly understood. To further investigate multispecies biofilms, we incubated fluorescently tagged C. jejuni (wild type [WT]-GFP) in the presence or absence of fluorescently tagged B. subtilis PS-216 (WT-RFP₁) for 24 h at 42°C in 96-well microtiter plates (Table 1). CLSM revealed that C. jejuni predominantly occupied the bottom of the well, forming submerged cell aggregates with characteristic finger-like structures (C. jejuni submerged biofilm), when cocultured alone (Fig. 1C). In coculture with B. subtilis, however, even if C. jejuni was inoculated at 10-times-higher initial numbers than B. subtilis, no C. jejuni biofilm was observed after 24 h, which supports the conclusion of strong growth inhibition of the pathogen. In fact, we were not able to observe by CLSM green dots representing C. jejuni GFP-labeled cells, probably due to their low numbers. In contrast, B. subtilis PS-216 biofilms were not affected by the presence of C. jejuni. B. subtilis cells formed a biofilm at the bottom of the well (B. subtilis submerged biofilm) and at the air-liquid interface (B. subtilis pellicle) (Fig. 1C) regardless of the presence or absence of C. jejuni.

TABLE 1.

Strains used in this study

Strain name or plasmid	Strain abbreviation	Background	Genome description	Source or reference
Campylobacter jejuni subsp. jejuni strains
NCTC11168	WT	Domesticated strain	92 – 94
NCTC11168	WT-GFP	NCTC11168	pWM1007	95
Bacillus subtilis strains
PS-216	WT	Undomesticated strain	96
3610	WT	Ancestral strain	Bacillus Genetic Stock Center (BGSC); Kolter lab collection
168	WT	Domesticated strain trpC2	BGSC
BM1705	PS-216	ΔsrfAA::erm (Erm)	This work
BM1707	PS-216	ΔsrfAA	This work
BM1411	PS-216	ΔppsB::spec (Spec)	This work
BM1858	PS-216	Δsfp/1::erm (Erm)	This work
BM1859	PS-216	Δsfp/2::erm (Erm)	This work
BM1860	PS-216	ΔpksL::erm (Erm)	This work
BM1861	PS-216	ΔpksN::erm (Erm)	This work
BM1629	WT-RFP2	PS-216	sacA::P43-mkate2 (Kn)	90
BM1314	WT-RFP1	PS-216	amyE::Phyperspank- mKate2 (Cm)	91
DNA donors for transformation
BKE03480	168 trpC2	ΔsrfAA::ery (Erm)	88
PSK0156	3610	ppsB ΩTn10 (Spec)	89
BKE03569	168 trpC2	Δsfp/1::erm (Erm)	88
BKE03570	168 trpC2	Δsfp/2::erm (Erm)	88
BKE17190	168 trpC2	ΔpksL::erm (Erm)	88
BKE17210	168 trpC2	ΔpksN::erm (Erm)	88
Plasmids (from E. coli strains)
pWM1007	GFP expression plasmid (Kn)	95
pDR244	ECE274	pDR244 (ampicillin, Spec)	88

We also quantified both species in the pellicle and below the pellicle by colony counts. After 24 h of static growth in Mueller-Hinton broth (MHB), we sampled the pellicle and the submerged area (planktonic cells in the medium and the submerged biofilm) for colony counts. After 24 h of incubation in coculture with B. subtilis, the number of C. jejuni bacteria decreased by 4.2 log₁₀ CFU/ml. C. jejuni grew to 8.6 log₁₀ CFU/ml in monoculture but only to 4.4 log₁₀ CFU/ml in coculture [P₍₂₄₎ = 7 × 10⁻⁶] (Fig. 1C). The pellicle consisted mainly of B. subtilis (7.3 log₁₀ CFU/ml) (Fig. S1C) with only 0.3 log₁₀ CFU/ml of C. jejuni (Fig. 1C). B. subtilis growth remained stable, with no differences in colony counts between mono- and cocultures with C. jejuni [P₍₂₄₎ = 0.72] (Fig. S1C). There were also no differences in the pellicle of B. subtilis between mono- and cocultures with C. jejuni [P₍₂₄₎ = 0.51] (Fig. S1C).

Cell-cell contact is not necessary for the anti-Campylobacter activity of the PS-216 strain.

The inhibition of bacteria depends on either diffusible factors or cell-cell contact (60, 61). To test whether cell-cell contact between B. subtilis and C. jejuni is necessary for the inhibition of the latter, we physically separated the two species in a transwell experiment by inserting a 0.1-μm-pore membrane between two incubation chambers, of which each contained one species. An inlay of B. subtilis was submerged into the lower chamber that contained C. jejuni. Our results demonstrate that the inhibition of C. jejuni by B. subtilis does not depend on cell contact, as the colony counts of C. jejuni after 24 h of incubation in coculture with B. subtilis were significantly lower (5.7 log₁₀ CFU/ml) than that in monoculture (9.4 log₁₀ CFU/ml) [P₍₂₄₎ = 7.6 × 10⁻⁹], with a 3.7-log₁₀-CFU/ml reduction in C. jejuni cells in the presence of B. subtilis (Fig. 1D). Again, C. jejuni had no influence on the growth of B. subtilis, which reached similar numbers in mono- and coculture [P₍₂₄₎ = 0.59] (Fig. S1D).

B. subtilis PS-216 affects preestablished C. jejuni submerged biofilms.

C. jejuni is capable of forming biofilms on both biotic (6, 62) and abiotic (17–21) surfaces. Therefore, preventing and disrupting C. jejuni biofilms may represent an important strategy for pathogen control. Previous in vitro studies showed that different B. subtilis strains inhibit the growth of C. jejuni via cell contact (46, 49, 50) or the culture supernatant (45). However, these studies did not address the timing or mechanisms of inhibition (63).

To test whether B. subtilis PS-216 can also disrupt preestablished C. jejuni biofilms, we introduced B. subtilis PS-216 at the time point t₀ to C. jejuni cultures that had already been maintained for 26 h. The preincubation time provided the opportunity for C. jejuni to develop an easily visible biofilm with a finger-like structure. The number of B. subtilis cells then added to the preestablished C. jejuni biofilm at t₀ was the same as that used for static cocultures in Fig. 1, which is ∼4.3 log CFU/ml. Therefore, at the time of B. subtilis inoculation the C. jejuni reached a 10,000-times-higher frequency than that of the added B. subtilis. Despite this dramatic disadvantage in cell numbers, B. subtilis was still able to disrupt the C. jejuni biofilm (Fig. 2). Using CLSM and bright-field microscopy, the characteristic C. jejuni submerged biofilm was no longer visible after 12 h of B. subtilis and C. jejuni coincubation (t₁₂) (Fig. 2). Similarly, B. subtilis PS-216 significantly reduced the colony counts of C. jejuni at t₁₂ [P₍₁₂₎ = 5 × 10⁻⁸] and t₂₂ [P₍₂₂₎ = 3.4 × 10⁻⁶] by 2.1 log₁₀ CFU/ml and 2.8 log₁₀ CFU/ml, respectively (Fig. 2). Conversely, no inhibition of B. subtilis by C. jejuni was observed at t₁₂ [P₍₁₂₎ = 0.45] and t₂₂ [P₍₂₂₎ = 0.85] (Fig. S2). Taken together, our results demonstrate that B. subtilis PS-216 can disintegrate the preestablished C. jejuni biofilm already after 12 h of coincubation under microaerobic conditions, leaving no or few biofilm aggregates and no finger-like structures characteristic of C. jejuni biofilms (Fig. S2B).

FIG 2.

B. subtilis PS-216 destroys preestablished submerged C. jejuni biofilms. (Top left corner) C. jejuni growth measured as colony counts during static mono- and coculture. B. subtilis was introduced (at t₀) to undisturbed C. jejuni biofilms precultivated for 26 h with a C. jejuni/B. subtilis ratio advantage of 10,000:1. Samples containing biofilm and broth were vortexed prior to plating at the indicated time points. Cocultures were sampled at 26 h (t₀; start of treatment with B. subtilis) and after 38 h (t₁₂) and 48 h of coincubation. (Top right corner) Schematic representation of C. jejuni submerged biofilm. (A to C) Brightfield and confocal microscopy before treatment (t₀; green cells) (A) and after 12 h of treatment (t₁₂) (B and C); (B) a C. jejuni monoculture biofilm (green cells); (C) a coculture with B. subtilis (red cells). This visualization revealed a strong effect of B. subtilis PS-216 on preestablished C. jejuni biofilms, as it excluded C. jejuni from its characteristic niche position. Three biological and up to three technical repeats were performed. The error bars represent ±standard deviation of the mean value. * represents statistically significant values (two-sample t test; see the supplemental material for details).

The B. subtilis strain PS-216 exhibits the strongest anti-Campylobacter activity.

To investigate the potential mechanisms by which B. subtilis affects the growth of C. jejuni, we assessed the inhibitory effect of three different B. subtilis strains. In addition to B. subtilis PS-216 (see above results), we also tested B. subtilis NCIB 3610, a strain with high genomic similarity to PS-216 (64), and the domesticated laboratory strain B. subtilis 168, which, due to drastic UV and X-ray exposure, lost important biofilm determinants, including a functional sfp gene (65, 66).

These three B. subtilis strains were coincubated with C. jejuni at a ratio of 1:10 in MHB, and the colony counts of both species were determined after 24 h. The strain PS-216 exhibited the greatest inhibition of C. jejuni growth, as the colony count of C. jejuni after 24 h of coculture with B. subtilis PS-216 was significantly lower (4.9 log₁₀ CFU/ml) than that in monoculture (9.2 log₁₀ CFU/ml) [P₍₂₄₎ = 3.3 × 10⁻⁵], with a 4.3-log₁₀-CFU/ml reduction. A weaker, yet still significant reduction (2.8 log₁₀) of C. jejuni cells [P₍₂₄₎ = 2.7 ×10⁻⁶] was observed in coculture with B. subtilis 3610. Interestingly, no inhibitory effect on C. jejuni cells was observed in coculture with B. subtilis 168 [P₍₂₄₎ = 0.92] (Fig. 3). Similar to the experiments described above, no significant differences in B. subtilis growth {PS-216 [P₍₂₄₎ = 0.81], 3610 [P₍₂₄₎ = 0.50], or 168 [P₍₂₄₎ = 0.60]} were observed between mono- and cocultures (Fig. S3).

FIG 3.

B. subtilis PS-216 exhibited the strongest anti-Campylobacter activity in static cocultures compared to the B. subtilis 3610 and 168 strains. C. jejuni growth was measured as colony counts after 24 h in static mono- and coculture. Samples containing biofilm and broth were vortexed prior to plating. Three biological and up to three technical repeats were performed. The error bars represent ±standard deviation of the mean value. * represents statistically significant values (two-sample t test; see supplemental material for details).

B. subtilis nonribosomal/polyketide synthesis is involved in C. jejuni inhibition.

As our results indicate that cell contact is not a prerequisite for the inhibition of C. jejuni growth and that B. subtilis 168 exerts no inhibitory effect on C. jejuni cells, we hypothesized that the antagonism of B. subtilis PS-216 is most likely due to the production of antimicrobial compounds. One of the genes involved in the production of these antimicrobial compounds in B. subtilis is the sfp gene (65–68). This gene is mutated in B. subtilis 168 (65) and might be responsible for the lack of any inhibitory effects on C. jejuni (Fig. 3). This functional phosphopantetheinyl transferase (sfp) transfers 4′-phosphopantetheine from coenzyme A to carrier proteins during the early nonribosomal synthesis of antimicrobial compounds, such as surfactin, plipastatin, and bacillaene (67, 69–73). First, we tested whether the B. subtilis sfp gene mutation increases the survival of C. jejuni in direct coculture assays. We generated mutations by inserting a resistance cassette into genes coding for the C-terminal (Δsfp/1) and N-terminal (Δsfp/2) part of the sfp gene in PS-216. Second, we constructed PS-216 mutants with disrupted genes coding for surfactin (srfAA), plipastatin (ΔppsB), and bacillaene (ΔpksL and ΔpksN) (Table 1).

Each B. subtilis mutant strain was tested for its ability to inhibit the growth of C. jejuni. After 24 h of coincubation and starting from a ratio of 1:10 (B. subtilis to C. jejuni), the colony counts of both species were determined. In comparison to wild-type (WT) B. subtilis PS-216, both Δsfp mutants showed significantly lower inhibition of C. jejuni [Δsfp/1, P₍₂₄₎ = 9.3 × 10⁻⁵; Δsfp/2, P₍₂₄₎ = 1 × 10⁻⁴]; WT B. subtilis inhibited the growth of C. jejuni by 4.2 log₁₀ CFU/ml, whereas Δsfp/1 and Δsfp/2 mutants inhibited C. jejuni by only 1.3 log₁₀ CFU/ml (Fig. 4). Similarly, the ΔpksL and ΔpksN mutants also showed significantly lower inhibition of C. jejuni than did WT PS-216, with an inhibition of 2.3 log₁₀ CFU/ml [P₍₂₄₎ = 0.001] and 1.3 log₁₀ CFU/ml [P₍₂₄₎ = 1 × 10⁻⁴], respectively (Fig. 4). However, mutations in the srfAA and ppsB genes did not exert significant C. jejuni growth inhibition compared to that of B. subtilis PS-216 WT [P₍₂₄₎ = 0.81; P₍₂₄₎ = 0.70], with an inhibition of 4.15 log₁₀ CFU/ml and 4.0 log₁₀ CFU/ml, respectively (Fig. 4).

FIG 4.

Inhibition of C. jejuni growth by B. subtilis PS-216 mutants in loci involved in nonribosomal/polyketide synthesis. Colony counts were measured in static mono- and coculture after 24 h; the log₁₀ CFU/ml inhibition was calculated as (C. jejuni monoculture log₁₀ CFU/ml) − (C. jejuni coculture log₁₀ CFU/ml). Samples containing biofilm and broth were vortexed prior to plating. Three biological and up to three technical repeats were used. The error bars represent ± standard deviation of the mean value. * represents statistically significant values (two-sample t test; see supplemental material for details).

DISCUSSION

In this work, we have revealed that the B. subtilis natural isolate PS-216 is a strong antagonist against C. jejuni in multispecies biofilms. B. subtilis PS-216 prevented C. jejuni adhesion to abiotic surfaces and biofilm formation and disrupted preestablished C. jejuni biofilms. Furthermore, this work highlights the important role of sfp and pks genes in the antimicrobial effect of B. subtilis on C. jejuni by disrupting growth and biofilm formation. Overall, these results suggest that B. subtilis PS-216 could be potentially used as a probiotic strain against C. jejuni.

C. jejuni is the most frequently reported bacterial foodborne pathogen in developed countries (1, 2) and is known to adhere to surfaces and form multispecies biofilms, improving its survival outside its natural environment (16–22). Persistent C. jejuni populations in poultry environments facilitate their circulation in the broiler gastrointestinal tract (GIT) (3), resulting in contaminated food products (4, 5). The use of antibiotics to reduce Campylobacter spp. during animal breeding is banned by both European Union (13) and U.S. (12) policy. Efficient strategies that reduce C. jejuni transmission are needed to prevent or reduce the presence of C. jejuni in poultry breeding and food processing (13). Probiotics are natural antibiofilm agents that act through secretion of biosurfactants and antimicrobial peptides, which are widely used in the treatment of bacterial biofilms (13, 48, 74). Microbial interactions represent an increasingly popular natural strategy of pathogen control, which have also been demonstrated to be beneficial to chicken hosts (52, 59). Probiotic supplements at the primary food production level could potentially minimize the risk of Campylobacter contamination (13, 75). Antimicrobial peptides can selectively destroy existing biofilms by reducing the bacterial biomass, affect different phases of the biofilm formation (27–34, 63, 74), and cause pore formation which usually results in cell lysis (45, 48, 63, 74). For example, one study predicted a 30-fold reduction in the incidence of campylobacteriosis in humans if Campylobacter species in contaminated meat is reduced by 2 log₁₀ (76), and another study predicted a 58-fold risk minimization if Campylobacter species is reduced by 3 log₁₀ in the cecum of birds (13).

This study demonstrated a dramatic antimicrobial effect of B. subtilis PS-216 against C. jejuni (with a reduction of 4.3 log₁₀ CFU/ml) during static coculture and under microaerobic conditions. This is especially interesting because microaerobic conditions are similar to those found in animal and human GITs (77) where B. subtilis probiotics need to act. This is in contrast to the findings of reference 78, which demonstrated that oxygen supply is a key factor for higher antimicrobial substance expression in cocultures of B. subtilis MA139 and Escherichia coli K88, as aerobic cocultures under shaking conditions exhibited higher inhibition than those under static conditions (78). Therefore, our result suggests the potential of B. subtilis PS-216, which also functions well under microaerobic conditions, as a probiotic strain in animal food production.

Similar to its inhibitory effect in liquid media, B. subtilis PS-216 inhibited C. jejuni adhesion to polystyrene surfaces and reduced C. jejuni colony counts by 2.4 log₁₀. Moreover, PS-216 also disintegrated (reduced biomass of) the preestablished submerged biofilm of C. jejuni. To date, very few studies have addressed direct surface inhibition of one model organism by another model organism. However, it has been shown that extracts of overnight B. subtilis cultures affect the adhesion of Staphylococcus aureus (29, 30), Escherichia coli (30), Listeria monocytogenes (31), and uropathogenic bacteria (27). The antiadhesive properties of B. subtilis extracts were identified as lipopeptides, namely, as biosurfactants (e.g., surfactins, iturins, and fengycins). However, there is a lack of literature specifically dealing with the interaction of B. subtilis and C. jejuni during biofilm formation on abiotic surfaces.

The mechanism by which B. subtilis inhibits C. jejuni biofilm formation and growth has been unknown. Our results indicate that the inhibition of C. jejuni growth and promotion of biofilm disintegration by B. subtilis PS-216 are most likely due to the production of diffusible antimicrobial molecules, as B. subtilis reduced C. jejuni colony counts, even when the species were separated by a 0.1-μm-pore membrane. B. subtilis is well known for its arsenal of antimicrobial molecules against various pathogenic bacteria (33–45). For example, surfactin, besides acting as an antiadhesive, also disperses biofilms of Staphylococcus aureus, E. coli, Salmonella enterica, and Proteus mirabilis strains (28–30, 34); the spoilage bacterium Shewanella putrefaciens (33); the water contaminant Legionella pneumophila (32); and different uropathogenic bacteria (27). This antibiofilm property of surfactin might have been due to its killing effect on the pathogen. Similarly, purified antimicrobial compounds, structurally related to amicoumacin A and produced by a B. subtilis 3 strain, which is a component of the commercial probiotic Biosporin (Biofarm, Dniepropetrovsk, Ukraine), inhibit the growth of Helicobacter pylori and C. jejuni LMBA 2568 in vitro (79). Other extracellular molecules produced by B. subtilis, such as α-amylase and bacteriocins, inhibit methicillin-resistant Staphylococcus aureus (MRSA), Vibrio cholerae, Pseudomonas aeruginosa, Staphylococcus epidermidis, and E. coli biofilm formation (80, 81).

B. subtilis hosts a variety of secondary metabolite gene clusters that encode polyketide synthases and nonribosomal peptide synthetases (44). These enzymes produce active secondary metabolites against various pathogenic bacteria (33, 45). In our search for metabolites affecting C. jejuni biofilm formation, we compared three different B. subtilis strains: the two highly related natural strains B. subtilis PS-216 and NCBI 3610, which show high genomic similarity (64), and the domesticated laboratory strain B. subtilis 168, which has lost important genetic biofilm determinants due to drastic UV and X-ray exposure (65). These include the sfp gene, which is involved in polyketide production and the nonribosomal peptide biosynthesis of surfactin, plipastatin, and bacillaene (67, 68, 71, 82, 83). Since B. subtilis 168 did not inhibit C. jejuni growth or C. jejuni biofilm formation, we took advantage of the genetic differences between B. subtilis 168 and PS-216/NCIB3610 to identify potential genetic determinants of PS-216 antagonism. Since B. subtilis PS-216 carries intact surfactin, plipastatin, and bacillaene genes (45, 84), we hypothesized that these genes potentially contribute to the antimicrobial activities during coculture with C. jejuni. By assessing C. jejuni inhibition by the PS-216 mutants lacking the above-mentioned genes, we confirmed that the genes that are part of the pksA-pksS cluster (78 kb) and encode the machinery that produces bacillaene (85) reduce the inhibitory effect of B. subtilis on C. jejuni. Bacillaene therefore plays an important role in C. jejuni inhibition. Conversely, mutations in genes coding for the nonribosomal peptides surfactin and plipastatin (srfAA and ppsB) did not significantly alter C. jejuni inhibition.

Campylobacter infections are a major burden on health care industries and national economies, and because C. jejuni is a common avian commensal, efforts are being made to challenge the pathogen at its primary source, the chicken reservoir. Probiotics can have beneficial effects on poultry, such as growth promotion, immunomodulation, and inhibition of pathogens (46, 49–51, 54–58). Different B. subtilis strains have been shown to improve feed conversion and body weight in chickens and lower the count of pathogens (52–54, 86). B. subtilis PS-216 also exhibits probiotic potential for Campylobacter and other enteropathogenic and foodborne pathogenic bacteria, including E. coli, Salmonella, and Listeria, and these observations are described in our patent application (87). The use of the B. subtilis PS-216 strain could therefore become widespread in the fight against foodborne bacteria. We anticipate that this microbial interaction study on PS-216 and C. jejuni will facilitate and provide ample scope for further studies in this direction. In conclusion, we have demonstrated that B. subtilis PS-216 interferes with the adhesion of C. jejuni biofilms to abiotic surfaces and prevents and disrupts the development of C. jejuni biofilms. The mechanism by which B. subtilis inhibits C. jejuni biofilm development depends on diffusible factors such as nonribosomal/polyketide bacillaene.

MATERIALS AND METHODS

Bacterial strains and strain construction.

The strains and genotypes of C. jejuni and B. subtilis used in this study and the construction of their mutant derivatives are described and listed in Table 1, including the strains used for the construction of the B. subtilis mutants described previously (88–91). In the multispecies biofilm experiments, C. jejuni NCTC11168 (WT) (92–94) and its derivative tagged with a gfp gene expressed on a plasmid (95) (WT-GFP) were used together with a WT B. subtilis PS-216 strain isolated from soil (96) and with an already-sequenced genome (64). Its derivatives were tagged with an mKate2 fluorescent protein (red fluorescent protein [RFP]) linked to a constitutive promoter integrated into two different loci (PS-216 amyE::P_hyperspank-mKate2 Cm and PS-216 sacA::P₄₃-mkate2 Kn) (WT-RFP₁ and WT-RFP₂) (Table 1). C. jejuni 11168 with a constitutively expressed GFP on the plasmid pWM1007, described by reference 95, was obtained from the Food Safety and Health Research Unit, Agricultural Research Service, U.S. Department of Agriculture (Albany, CA, USA). The recombinant B. subtilis strains were constructed by transforming the DNA of B. subtilis donor strains (Table 1) into B. subtilis PS-216 recipient strains using standard transformation techniques. Briefly, the PS-216 recipient strain was grown in competence medium at 37°C, and chromosomal DNA or PCR products were added to competent cells after 6 h of incubation. The transformants were selected on LB agar plates containing the appropriate antibiotics with the following concentrations: 5 μg/ml chloramphenicol (Cm), 100 μg/ml spectinomycin (Spec), and 20 μg/ml erythromycin (Erm). The PS-216 Δsfp/1, Δsfp/2, ΔpksL, ΔpksN, and ΔsrfAA mutants were constructed by transforming PCR products derived from B. subtilis 168 mutants from the (BKE) single-gene inactivation library (88) as DNA donors. To remove the antibiotic resistance cassette, the PS-216 ΔsrfAA::erm was transformed with plasmid DNA pDR244 as described by reference 88. The strain that grew on LB agar plates, but not on the LB plates containing antibiotics, had lost the pDR244 plasmid and antibiotic resistance cassette. The PS-216 ΔppsB mutant was constructed by transformation using chromosomal DNA isolated from the B. subtilis 3610 (PSK0156) (89) mutant. For PCR transformation, chromosomal DNA from the B. subtilis 168 mutant strain from the single-gene inactivation library (88) was isolated and amplified by specific primers (5pL/3pR [88]), and the obtained PCR products were transformed into the PS-216 WT strain.

Bacterial growth conditions.

The WT C. jejuni strain was subcultured from the stock culture (−80°C) on Karmali agar (Oxoid) plates with the selective supplement SR1607E (Oxoid), while WT-GFP C. jejuni was cultivated on Müller-Hinton agar (MHA) supplemented with kanamycin (Kn) (50 μg/ml) for 24 h. After revitalization from −80°C and before the experiments, the WT and WT-GFP C. jejuni strains were subcultured on MHA with or without kanamycin for an additional 24 h. All C. jejuni cultures were maintained at 42°C under microaerobic (85% N₂, 5% O₂, 10% CO₂) conditions. All B. subtilis strains were subcultured from the stock (−80°C) on MHA with or without the appropriate antibiotics (100 μg/ml Spec, 20 μg/ml Erm, 5 μg/ml Cm, 50 μg/ml Kn) (WT-RFP_1,2 strain derivates) for 24 h. After 24 h, the colonies were subcultured on MHA with or without the appropriate antibiotics at 28°C for 24 h under aerobic conditions.

C. jejuni colony counts (CFU/ml) were determined on Karmali agar plates incubated at 42°C for 24 h under microaerobic conditions. B. subtilis colony counts were determined on MHA plates incubated at 28°C for 15 to 18 h under aerobic conditions selective against C. jejuni growth. All B. subtilis-C. jejuni coculture experiments were routinely performed in a controlled atmosphere under static microaerobic conditions at 42°C in Müller-Hinton broth (MHB).

Multispecies biofilms.

To test the antagonistic activity of the B. subtilis strain PS-216 against C. jejuni in multispecies biofilm cultures, we first diluted the overnight cultures of each species to an optical density at 600 nm (OD₆₀₀) of 0.1 and used 0.1% inoculum (each) in 5 ml MHB. This approach ensured a ratio of 1:10 between B. subtilis and C. jejuni with ∼4.3 log₁₀ CFU/ml and ∼5.3 log₁₀ CFU/ml, respectively. Before each sampling, cultures were vortexed, sampled, and then statically incubated until the next disruptive sampling. Monocultures of B. subtilis and C. jejuni were used as controls (0.1% inoculum). In addition to WT B. subtilis PS-216, we also tested B. subtilis 3610, B. subtilis 168, and six PS-216 single-gene knockout mutants (Table 1) for antagonistic activity against C. jejuni. The two species were mixed at a ratio of 1:10 in MHB as described above and incubated under microaerobic conditions at 42°C under static conditions that support biofilm development of both species when in monocultures; the colony counts were determined at 0 h and after 24 h. The biofilms were disrupted by vortexing and strong pipetting before plating.

Transwell cocultures.

To test whether B. subtilis produces diffusible substances that inhibit the growth of C. jejuni, the two species were cocultured in a system that allows the exchange of molecules but not direct contact between cells. The 12-well plates (Cellstar; Greiner, Bio-One) contained wells with two chambers separated by a membrane with 0.1-μm pores. First, C. jejuni was inoculated into the well with 3 ml MHB, and then an inlay (2 ml) containing B. subtilis was submerged into the well with C. jejuni. An inoculum of B. subtilis and C. jejuni was prepared as described above. Monocultures of B. subtilis and C. jejuni were used as controls (0.1% inoculum). C. jejuni cultures were inoculated into wells, after which an inlay containing only MHB was submerged, whereas B. subtilis was inoculated into inlays submerged into wells containing only MHB. The colony counts of B. subtilis and C. jejuni mono- and cocultures were determined at 0 h and 24 h.

Adhesion to abiotic polystyrene surfaces.

The adhesive capability of B. subtilis PS-216 and WT C. jejuni was determined by the number of cells (CFU/ml) that adhered to the abiotic polystyrene surface. Inocula of B. subtilis and C. jejuni were prepared from overnight cultures by diluting the culture to an OD₆₀₀ of 0.1. Cells of each species were mixed at a ratio of 1:1 (in 100 μl), with a final concentration of both strains of ∼6 log₁₀ CFU/ml in each well of the 96-well microtiter plate (Nunc, Roskilde, Denmark). Monocultures of B. subtilis and C. jejuni with the same initial cell numbers were used as controls. After 24 h of incubation under microaerobic conditions at 42°C, planktonic/unattached cells were removed by repeated (3 times) rinsing of the polystyrene surface with 100 μl of sterile phosphate-buffered saline (PBS). Attached cells were removed by sonication (room temperature, 10 min; 28-kHz frequency; 300-W power; Iskra Pio, Šentjernej, Slovenia) and resuspended in 100 μl of PBS.

Disruption of preestablished biofilms.

Next, the ability of B. subtilis to disrupt the biofilm of C. jejuni was tested. First, C. jejuni was allowed to form a biofilm. WT C. jejuni was inoculated at ∼6 log₁₀ CFU/ml (OD₆₀₀ = 0.1) in fresh MHB and incubated for 26 h at 42°C under microaerobic conditions, and then after C. jejuni concentration in biofilms already reached 8.5 log₁₀ CFU/ml, B. subtilis PS-216 (∼4.2 log₁₀ CFU/ml) was added. This experiment was performed in 5-ml 6-well polystyrene microtiter plates (TPP) for colony counts, which were determined for C. jejuni at the beginning of incubation and for both species at t₀ (the time point when B. subtilis was added), t₁₂ (12 h after coincubation), and t₂₂ (22 h after coincubation). To visualize the effect of B. subtilis on preestablished C. jejuni biofilms, we cocultured fluorescently labeled strains, C. jejuni WT-GFP (green) and B. subtilis WT-RFP₂ (red), in 96-well microtiter plates with a glass bottom (TPP) at 42°C. The biofilms were visualized by CLSM at the indicated times (t_0,12,22).

Spatial distribution of B. subtilis and C. jejuni cells in coculture.

Mono- and multispecies biofilms of B. subtilis PS-216 WT-RFP₁ and C. jejuni WT-GFP were grown in MHB in 96-well microtiter plates (TPP) under microaerobic conditions and at 42°C. Strains in coculture were mixed at a ratio of 1:1 (in 100 μl) with a final concentration of both strains of ∼6 log₁₀ CFU/ml; monocultures contained the same number of cells (∼ 6 log₁₀ CFU/ml). After 24 h of incubation, we performed biofilm imaging by CLSM. For colony counts of both species, we cultured biofilms in 5 ml of MHB (in 6-well microtiter plates) and sampled the pellicle (the biofilm formed at the air-liquid interface) and the broth below the pellicle at 0 h and 24 h. Before plating, pellicles were rinsed by repeated (3 times) immersion in the sterile PBS, after which pellicles were resuspended in 100 μl PBS and repetitively vortexed before and after sonication (room temperature, two times for 10 min; frequency, 28 kHz; power, 300 W; Iskra Pio, Šentjernej, Slovenia). B. subtilis and C. jejuni monocultures were used as controls.

Bright-field microscopy and CLSM.

The spatial distribution and structural properties of B. subtilis and C. jejuni biofilms in mono- and coculture were investigated using CLSM (with the inverted microscope AxioVision Z1, LSM800; Zeiss, Germany). Biofilms were grown as described above, and imaging was performed directly in microtiter wells containing GFP-tagged C. jejuni strains (green) and RFP_1,2-tagged B. subtilis (red). Excitation of GFP was performed at 488 nm with an argon laser, and the emitted fluorescence was recorded at 400 to 580 nm. Excitation of the red fluorescence protein RFP (mKate) was performed at 561 nm, and the emitted fluorescence was recorded at 580 to 700 nm. The laser intensities and GaAsP detector gain were 4% and 800 V and 4.5% and 650 V for mKate (RFP) and GFP, respectively. The pinhole size was 55 μm. To generate images of the biofilms, 10-μm z-stacks (height) were generated for each biological sample. The size of the acquired images was typically 1,024 by 1,024 pixels with 16-bit color depth, and microtiter wells were scanned using the 20×/0.4-numerical-aperture (NA) objective. The bright-field images were acquired using the AxioCam MRm rev. 3 (Zeiss) camera and HAL 100 light source (Zeiss). Zen 2.3 software (Carl Zeiss) was used for image acquisition and visualization. The noise on the acquired CLSM images was reduced by applying a single pixel filter (threshold = 1.5).

Statistical analysis.

To evaluate the influence of biofilm cocultivation on the growth of B. subtilis and C. jejuni, statistical significance was assessed by the two-sample t test using raw data. Probability values smaller than 0.05 (P < 0.05) were considered statistically significant (using the Welch correction in which equal variance is not assumed). Three biological and up to three technical repeats were performed for all experiments. The data are displayed as mean ± standard deviation. All analyses were performed using OriginPro 2020 (OriginLab Corporation, Northampton, MA, USA).