Insights into the genomic and phenotypic characteristics of Bacillus spp. strains isolated from biofilms in broiler farms

ABSTRACT

The characterization of surface microbiota living in biofilms within livestock buildings has been relatively unexplored, despite its potential impact on animal health. To enhance our understanding of these microbial communities, we characterized 11 spore-forming strains isolated from two commercial broiler chicken farms. Sequencing of the strains revealed them to belong to three species Bacillus velezensis, Bacillus subtilis, and Bacillus licheniformis. Genomic analysis revealed the presence of antimicrobial resistance genes and genes associated with antimicrobial secretion specific to each species. We conducted a comprehensive characterization of the biofilm formed by these strains under various conditions, and we revealed significant structural heterogeneity across the different strains. A macro-colony interaction model was employed to assess the compatibility of these strains to coexist in mixed biofilms. We identified highly competitive B. velezensis strains, which cannot coexist with other Bacillus spp. Using confocal laser scanning microscopy along with a specific dye for extracellular DNA, we uncovered the importance of extracellular DNA for the formation of B. licheniformis biofilms. Altogether, the results highlight the heterogeneity in both genome and biofilm structure among Bacillus spp. isolated from biofilms present within livestock buildings.

IMPORTANCE

Little is known about the microbial communities that develop on farms in direct contact with animals. Nonpathogenic strains of Bacillus velezensis, Bacillus subtilis, and Bacillus licheniformis were found in biofilm samples collected from surfaces in contact with animals. Significant genetic and phenotypic diversity was described among these Bacillus strains. The strains do not possess mobile antibiotic resistance genes in their genomes and have a strong capacity to form structured biofilms. Among these species, B. velezensis was noted for its high competitiveness compared with the other Bacillus spp. Additionally, the importance of extracellular DNA in the formation of B. licheniformis biofilms was observed. These findings provide insights for the management of these surface microbiota that can influence animal health, such as the use of competitive strains to minimize the establishment of undesirable bacteria or enzymes capable of specifically deconstructing biofilms.

INTRODUCTION

Livestock building surfaces are reservoirs of potential undesirable microorganisms and so are considered a vector for disease transmission in animals (1). Yet, scientific studies investigating the microbial composition, diversity, and structure of biofilms associated to animal production systems remain scarce. To minimize the presence of undesirable microorganisms on the surfaces of livestock buildings between successive batches of animals, cleaning and disinfection procedures using biocides are implemented. However, the efficacy of these procedures can be reduced by the formation of biofilm structures formed by the microbial community, which can adhere to surfaces (2).

The majority of microorganisms on Earth live in a biofilm lifestyle (3). Biofilms are spatially organized microbial communities embedded in a self-produced extracellular matrix (4). The extracellular matrix of biofilms consists mainly of water, exopolysaccharides, proteins, and extracellular DNA (eDNA) (5). Due to the matrix and resulting gradients, these communities possess emergent properties compared with the planktonic form, such as greater tolerance to environmental fluctuations like the application of biocides or antibiotics (6, 7).

Antibiotics are sometimes used on a massive scale in animal production systems as curative or preventive measures to control pathogens as well as growth promoter for animals on livestock farms, depending on the regulations in force to control their use. The use of antibiotics inevitably leads to the emergence of resistances. Antimicrobial resistance genes (ARGs) can be carried by pathogens or by other members of the microbial community, including the ones attached to surfaces (8, 9). Strains carrying mobilizable ARGs of clinical relevance are then considered undesirable because they can confer resistance to microbial pathogens and pose safety concerns for animals, humans, and the environment.

We recently described that Bacillus spp. can be very abundant on commercial poultry farm surfaces, even after cleaning and disinfection procedures and even prior to the entrance of animals (10). The Bacillus genus is ubiquitous in the environment and known for its ability to form biofilms (11, 12). Some Bacillus spp. may be undesirable and carry mobile ARGs or be pathogenic such as Bacillus cereus or Bacillus anthracis (13). Characterization of Bacillus spp. strains encountered on farms can therefore be important to (i) identify strains potentially posing safety concerns and find ways to limit their establishment on farm surfaces and (ii) select for safe strains with high biofilm-forming capacity, which could be used as competitive exclusion agents. Indeed, the biotechnological potential of beneficial strains of Bacillus spp. has already largely been investigated in various industrial applications and products have been developed, such as probiotics (14), biocontrol agents (15), or positive biofilms (1).

The present study aimed at characterizing 11 spore-forming strains isolated from poultry farm surfaces (10) using genomic analysis and biofilm phenotyping using different laboratory models. In order to increase our knowledge of the microbial ecology of these biotopes, we studied the compatibility between these strains with a macro-colony biofilm model used to observe the potential of biofilms to merge and cover the surface (16). Confocal laser scanning microscopy (CLSM) coupled with image analysis was used to visualize and quantify biofilm formation with a focus on eDNA quantification and its involvement in the formation of biofilms.

RESULTS

Genomic analysis of Bacillus spp.

Genome characteristics

Genome analysis was carried out after the sequencing of the 11 strains (Table 1). The assignment of species through genome analysis was conducted, and the strains were sorted according to their phylogenetic distance. Three strains were identified as Bacillus velezensis, two as Bacillus subtilis, and six as Bacillus licheniformis.

TABLE 1.

Genomic characterization of the strains used in this study^a

Name	Completeness	Genome size (bases)	N50	Genome size (bp)	%GC	CDS	rRNA	rRNA	tmRNA	Building number	BioSample accession
Bacillus velezensis B18	99.5	3915595	315607	3915595	46	3771	3	82	1	1	SAMN30015524
Bacillus velezensis B8	99.67	3938456	390404	3938456	46	3738	3	79	1	1	SAMN30015523
Bacillus velezensis B1	99.67	4158515	381251	4158515	46	4143	4	84	1	1	SAMN30015521
Bacillus subtilis C15	99.62	4047208	1E+06	4047208	44	4032	3	73	1	2	SAMN30015532
Bacillus subtilis B2	99.6	4188957	654724	4188957	44	4241	3	77	1	1	SAMN30015522
Bacillus licheniformis C3	99.57	4167229	349485	4167229	46	4247	3	79	1	2	SAMN30015528
Bacillus licheniformis C2	99.46	4309318	313710	4309318	44	4445	3	80	1	2	SAMN30015527
Bacillus licheniformis C1	98.5	4204947	44804	4204947	45	4306	4	52	1	2	SAMN30015526
Bacillus licheniformis C10	99.45	4319637	347658	4319637	46	4486	4	80	1	2	SAMN30015530
Bacillus licheniformis C11	99.17	4176285	1E+06	4176285	46	4234	5	79	1	2	SAMN30015531
Bacillus licheniformis C5	99.17	4256195	399216	4256195	45	4349	4	79	1	2	SAMN30015529

^a All strains were isolated from two separate broiler buildings (10).

ARG

The genome-based identification of ARG reveals patterns linked to the species and is represented in a presence/absence heat map (Fig. 1; File S1).

Fig 1.

Identification of Bacillus spp. classified according to phylogenetic distance with identification of ARG and biosynthetic gene clusters related with antimicrobials in genomes. The Bootstrap value is shown in the tree. The presence or absence of ARG in the genome identified by the tree tools ARG-ANNOT, CARD, and ResFinder is represented in the heat map on the left. The identification of genes involved in antimicrobial secretion in Bacillus spp. genomes is represented on the right. Highly conserved hits (>90%) of antimicrobial secreted molecules were identified through genome analysis using the antiSMASH tool. NCBI blast tool was used to identify genes related with biofilm formation. The presence of genes is represented by a yellow box in the heat map.

A relationship between the presence of ARGs and the species was observed. The three B. velezensis has two hits (cfrB, clbA) and in addition satA for the strain B1 and tetL and tet45 for the strain B8. The two B. subtilis strains each have 10 hits, with rphD only found in C15. B. licheniformis strains all showed hits against rphD and ermD for strains C3, C2, C5, and C11. No Mobile Genetic Elements associated with antibiotic resistance genes were found in the genomes of any of the 11 isolated Bacillus strains analyzed using the online tool CGE-MobileElementFinder (File S2). When blasting the ARGs in the NCBI, the coverage and identity of these putative ARGs match with typical genes commonly encountered into the genomes of the respective species and characterized as chromosomally located intrinsic resistance genes (File S3). For B. velezensis, these include genes encoding a streptothricin N-acetyltransferase related to streptothricin resistance, a Methyltransferase resulting in resistance via methylation of the 50S ribosomal complex and genes related to tetracycline as well as a gene conferring broader resistance to lincosamides, streptogramin, oxazolidinones, phenicols, and pleuromutilins. These resistance genes are common in B. velezensis, Bacillus amyloliquefaciens, and others Bacillus spp. and are reported to be located on the chromosome with no indication of mobile elements. For B. subtilis, these include notably genes encoding resistance to lincomycin (lmrB), streptomycin (aadK/aadE), tetracycline (tet(45)/tetL), macrolide antibiotics (mph(K)), and a broad specificity multidrug (ykkC/ykkD) resistance genes. These resistance genes are common in B. subtilis and are found located on the chromosome with no indication of mobile elements. Finally, for B. licheniformis, these include rphD encoding resistance to rifamycin and ermD encoding broad-spectrum resistance to macrolides (e.g., erythromycin), streptogramin b (e.g., virginiamycin, quinupristin, and pristinamycin), and lincosamides (e.g., clindamycin and lincomycin). These resistance genes are common in B. licheniformis and Bacillus paralicheniformis and are also reported to be located on the chromosome with no indication of mobile elements.

Biosynthetic gene clusters related with antimicrobials

Antimicrobial secretion genes were searched in the genomes, and species patterns were revealed (Fig. 1; File S4).

Bacilysin, bacillaene, and bacillibactin were found in B. subtilis and B. velezensis, while lichenicidin and lichenysin were detected in B. licheniformis. Subtilosin A was identified only in the B. subtilis species, while bacillomycin D and surfactin were specific to the B. velezensis species. However, some strain-specific hits were also identified: mersacidin for B. velezensis B8, plantazolicin for B. velezensis B1, the sporulation killing factor for B. subtilis C15, and difficidin and fengycin for B. subtilis B2.

Genes related with biofilm formation

Genes identified as being involved in biofilm formation (17) were searched in the genomes of the study strains with a threshold of 80% nucleotide identity and 70% length coverage with the reference genes. The presence of genes is grouped by species; all the genes were found in B. subtilis, whereas only spo0A and sinR were found in B. licheniformis, with the addition of pgsB for B. velezensis (Fig. 1; File S5).

Description of Bacillus spp. biofilms isolated from livestock building surfaces

Biofilm formation capability

All the Bacillus spp. isolates were able to form macroscopic and structured biofilms in the different laboratory models used (Fig. 2).

Fig 2.

Representative images of the 11 Bacillus spp. grown on macro-colony, pellicles, and swarming biofilm models. Macro-colony and swarm biofilm models enable bacterial growth at the agar-air interface. Pellicles form at the interface between liquid and air. Representative images of the three biological replicates performed are shown.

Structural heterogeneity was observed among the different isolates and within the same species. Macro-colony staining differed between strains, demonstrating the diversity of extracellular matrix composition. The red colony contains amyloid fibers because of the staining with Congo red. A lighter coloration, such as the one observed at the center of the B. licheniformis C11 colony, was associated with concentrated protein because of Coomassie Brilliant Blue labeling. Some colonies, such as the ones from B. subtilis strains and B. velezensis B1, formed tall, wrinkled structures, while others, such as B. velezensis B18, were flat but showed the ability to colonize the entire surface of the well. Likewise, remarkable phenotypes were observed with the pellicles model. B. licheniformis strains C1, C2, and C10 formed a fine, highly structured pellicles, while all other strains showed flat pellicles. B. velezensis B8 was the strongest colonizer in the swarming model. Phenotypes differed between strains, with the remarkable presence of pronounced halo-like growth for B. licheniformis strains C1, C2, and C11. As shown in other biofilm models, a significant degree of heterogeneity in 3D structures was likewise observed in submerged biofilms. SYTO9 staining visualized the entire population, while TOTO3 only enabled detection of eDNA or bacteria with damaged cell walls (Fig. 4A). The biofilms of B. velezensis and B. subtilis strains were homogeneous, whereas B. licheniformis formed biofilms with a large number of dense structures, making them highly heterogeneous. Certain structures exhibited a notable abundance of eDNA. The amount of eDNA was higher in B. licheniformis biofilms than in other Bacillus spp. except for B. velezensis B8.

Evaluation of strain interactions using a macro-colony model

Interaction experiment was performed between Bacillus spp. strains (Fig. 3).

Fig 3.

Compatibility assessment of Bacillus spp. according to the phylogenetic distance of strains in a macro-colony biofilm model. A representative image of the contact and gap phenotypes is shown as an example. Scale bar = 1 cm. One square represents three biological replicates, each showing an interaction.

A relationship between the phylogenetic distance of the strains and their ability to form mixed biofilms was observed. B. velezensis B8, B18, and B1 primarily demonstrate a tendency toward negative recognition, as evidenced by the spatial separation between their colonies and other neighboring colonies. This suggests that these strains likely secrete a diffusible molecule into the agar medium. The colonies of all the other strains were able to form contacts with each other, except for strains B. subtilis B2 and B. subtilis C15. Interestingly, some interactions between the same strain from different cultures showed antagonism with a distance between the two colonies. This was the case for B. velezensis B1, B. subtilis B2, or B. subtilis C15.

eDNA differentially influences bacillus biofilm organization

DNAse I had an effect on the structure of all B. licheniformis strains and B. subtilis C15 (Fig. 4A and B). The enzymatic treatment significantly reduced (P < 0.01) biofilm formation for all strains of B. licheniformis save strain C11 (Fig. 4C), and this was associated with a significant reduction of eDNA (Fig. 4D).

Fig 4.

Effect of DNAse I on the formation of Bacillus spp. biofilms isolated from the surface of poultry farm buildings. (A) Representative images captured using CLSM of 24-h Bacillus spp. biofilms within the culture medium following adhesion. Chemical labeling using SYTO9 to highlight the entire population and TOTO3 to mark the DNA were used. (B) Same experiment with DNAse I incorporated in the medium after the adhesion step. Scale bar = 40 µm. (C) Quantification of the biovolume of the SYTO9 signal. (D) Quantification of the biovolume of the TOTO3 signal. Error bars correspond to standard deviation.

DISCUSSION

Eleven sporulated strains belonging to QPS (qualified presumption of safety) status bacterial species were isolated from two poultry farms. We investigated some of their genomic and biofilm-forming features. All the strains were able to form biofilms in different laboratory models with a high diversity of morphology between strains. However, it is important to note that a larger strain collection is needed to draw general conclusions at the species level, especially for B. subtilis and B. velezensis, where only three and two strains, respectively, were isolated and studied.

A detailed in silico analysis was conducted to identify putative ARG genes in the genomes of the Bacillus strains. The hits identified are concluded to be related to genes inherent and intrinsic to the respective Bacillus species of the 11 Bacillus strains studied. Indeed, these genes were all found to be located in chromosomic regions and detected in published genomes of strains from the same bacterial species including the respective type strains. The absence of any mobile genetic elements associated with these genes in the genomes of the 11 strains confirms that they do not pose a safety problem. However, phenotypic testing for resistance to clinically relevant antibiotics would still be required to complete the safety assessment of these strains in terms of ARG.

The search for genes involved in biofilm formation was conducted on the 11 strains of Bacillus spp. to demonstrate a higher abundance of these genes in B. subtilis compared with the other species. A strict detection threshold was used to determine the presence or absence of these genes in the genome based on reference sequences of B. subtilis. In some strains, genes were not identified as present according to the criteria, even though they were very close to the thresholds, indicating that the absence of the gene function cannot be conclusively demonstrated. Moreover, the studied strains, which form different biofilms in the models used here, probably possess other genes involved in biofilm formation that are specific to species (i.e., matrix composition, motility, and quorum sensing) (6). In addition, a gene may be present but differentially expressed depending on the biofilm culture conditions (18). It would be interesting to study the implication of these genes on biofilm phenotypes.

Each strain showed different biofilm characteristics. Submerged biofilms were subjected to more comprehensive characterization. CLSM enables quantification of structural parameters, such as the biovolume. There seems to be no link between the presence of eDNA in the biofilm and the biovolume. It is plausible that eDNA may influence the early stages of biofilm formation, possibly preventing strains from aggregating during the initial phases. This could hinder the establishment of cell-to-cell density-dependent communication systems such as Quorum Sensing (19). Considering DNase treatment for surface decontamination, its potential to reduce biofilm formation in certain strains suggests that such treatments should be applied preventively before introducing animals. This approach could inhibit the initial stages of biofilm development and prevent its establishment.

A correlation between the ability of strains to merge their macro-colonies and the presence of genes associated with antimicrobial secretion would be interesting to study. For instance, B. subtilis B2 and C15 exhibit a compatibility between each other and with the strains of B. licheniformis. Interestingly, the B. velezensis strains studied have the potential to encode distinct antimicrobials compared with the other strains. Essential compounds such as bacillomycin D, bacilysin, bacillaene, surfactin, and bacillibactin are common to all three strains of B. velezensis. In addition, strain B18 has more compatibility traits than strains B1 and B8. The specificity of B1 lies in its production of plantazolicin, while strain B8 synthesizes mersacidin. On the other hand, all B. licheniformis strains are compatible with each other and secrete mainly lichenicidin, with the additional presence of lichenysin in strains C3, C2, C11, and C5. It would be interesting to investigate further the ability of strains to form mixed biofilms. Indeed, based on their compatibilities, it has been shown that colonies able to merge can form stable mixed biofilms and colonize environments such as plant roots (20). Interestingly, compatibility assessment showed the “gap” phenotype for B1, C15, and B2 strains when interacting between themselves. Indeed, some secreted compounds can become toxic above certain concentrations. Enzymes such as exoproteases or siderophores may be secreted by Bacillus spp., leading to nutritional competition that prevents the two colonies from meeting due to local nutrient deficiencies along the migration fronts. Additionally, signaling molecules might diffuse through the agar, halting colony expansion and instead promoting biofilm matrix production (21).

The selection of autochthonous competitive strains of Bacillus spp. based on their biofilm properties and resistance to DNase could be a promising approach. Indeed, developing products combining such strains together with DNase enzymatic treatment could favor the establishment of positive biofilm while limiting the development of biofilm by certain members of the microbial community already in place.

In conclusion, this research has illuminated various aspects of bacterial strains isolated from the broiler breeding environment, revealing their genetic diversity and biofilm structures. B. velezensis appears to be highly competitive in these environments compared with other Bacillus species. Characterization of submerged biofilms has uncovered the role of eDNA in Bacillus species biofilm structures. In particular, we have demonstrated that DNAse treatment prevents B. licheniformis biofilm formation but not those of B. velezensis. Finally, the lack of mobilizable ARGs enables the consideration of some of these strains for biotechnological applications.

MATERIALS AND METHODS

Isolation of strains from farm surface biofilms

Eleven strains were isolated using a coupon sampling method (22) from the surfaces of two commercial poultry farms described in a previous study [i.e., building 1 and 2 in reference (10)]. After sampling, the polyvinyl chloride coupons (2.5 cm × 6 cm ×3 mm) were placed in individual tubes containing 30 mL of a saline solution (NaCl 9 g/L). The microbiota attached to the surface was mechanically detached using a pipette cone in successive round trips. In order to count spores in the environmental biofilms, 1 mL of detached biofilm suspension was placed in a glass tube that was immersed in a water bath for 10 min at 80°C, before enumeration on TSA [trypticase soy broth (TSB) Agar, BioMérieux, France]. Representative colonies were picked and grown overnight (∼16 h) in TSB (TSB; BioMérieux, France) before being stored in glycerol tubes at −80°C.

DNA extraction and sequencing

The protocol described by Guéneau et al. (23) was used to extract and sequence bacterial genomes. Concisely, from a −80°C glycerol stock, the strains were cultured in TSB under agitation at 37°C overnight. Genomic DNA was extracted and purified using the Monarch genomic DNA purification kit (New England Biolabs, Ipswich, MA, USA). DNA sequencing (DNA-seq) was performed at the GeT-PlaGe core facility (INRAE, Toulouse, France; www.genotoul.fr/en/). DNA-seq libraries were prepared according to Illumina’s protocols using the TruSeq Nano DNA high-throughout (HT) library kit (Illumina, San Diego, CA, USA). DNA-seq experiments were performed on an Illumina NovaSeq 6000 system, using a paired-end read length of 2 × 150 bp with the Illumina NovaSeq 6000 reagent kits (Illumina).

Phylogenetic analysis of Bacillus spp. and species assignments

De novo assembly was performed using Unicycler (Galaxy version 0.4.8.0) (24), and genome annotation was performed using PROKA (rapid prokaryotic genome annotation, Galaxy version 1.14.6) (25). The phylogenetic tree analysis of the 11 Bacillus spp. strains was performed using the bacterial Phylogenetic Tree Service of BV-BRC (26). The whole genome sequence FASTA file of each of the 18 Bacillus spp. strains was uploaded to BV-BRC, set to the “contigs” object type. The phylogenetic tree was then built with the default Codon Tree method. For species assignments, the genome sequences were compared with the genome sequences of the closest species within Bacillus spp. genus in the BV-BRC database, with the requested use of 1,000 genes (up from the default 100), with five allowed deletions.

Identification of ARGs

ARGs were searched in genomes using three different bioinformatic tools. ARG-ANNOT (Antibiotic Resistance Gene-ANNOTation) (27) was used with the default threshold percentages for identity and minimum sequence length. A blast of the WGS of each Bacillus spp. strains in FASTA format was applied against the latest database ARG-ANNOT-NT-V6-July2019. CARD (Comprehensive Antibiotic Resistance Database) (28) analysis was performed with the same genomes. For this, each genome sequence was screened with CARD’s Resistance Gene Identifier version 5.2.0 and CARD database version 3.1.4. “Perfect,” “Strict,” and “Loose” hits were evaluated for the provided genome sequence and for complete and partial gene predictions with the default filter parameters.

ResFinder (29) was used to screen for ARGsin each of the Bacillus spp. genome sequences. The genome sequence in FASTA format was analyzed online at https://cge.cbs.dtu.dk/services/ResFinder/ in ResFinder 4.1 software version and database version 2021-09-23. The analysis was run against all available antibiotics with the following settings: 80% threshold for ID and 60% for minimum sequence length (coverage). The results presented are a compilation of the hits found by the three tools.

The intrinsic or acquired nature of the ARGs found in each of the Bacillus spp. was evaluated by checking the presence of such genes in strains of the same species (e.g., blasting each ARGs against the NCBI genomes database of the same species) and the search of any genetic mobile element associated to such ARGs to evaluate their non-transferability (e.g., using the online tool MobileElementFinder available at https://cge.cbs.dtu.dk/services/MobileElementFinder).

Identification of biosynthetic gene clusters related with antimicrobial

The whole genome sequence FASTA file of each of the 11 Bacillus spp. strains was uploaded to the antiSMASH 5.0 tool (https://antismash.secondarymetabolites.org/#!/start) (30). The detection strictness > 90% was chosen to detect well-defined clusters containing all required parts.

Identification of genes related with biofilm formation

The whole genome sequence FASTA file of each of the 11 Bacillus spp. strains were uploaded to the NCBI Blast tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi). A nucleotide query was used with genes identified as being involved in biofilm formation (17). For the epsA-O operon (15 genes), tapA, tasA, sipW, spo0A, srlR, and sinR, the reference genes were taken from Bacillus subtilis subsp. subtilis str. 168, complete genome (NCBI Reference Sequence: NC_000964.3). For the pgsB and bslA genes that are not present in B. subtilis str. 168 (31), the Bacillus subtilis strain NCIB 3610 chromosome, complete genome (NCBI Reference Sequence: NZ_CP020102.1) was used. The presence of genes was determined on the criterion of the coverage and identity values mentioned in the EFSA guideline (32). A gene will be present if its percentage of nucleotide identity is at least 80% and if its length coverage is at least 70%.

Biofilm experiments

Macro-colony

All experiments began with 5-mL cultures in trypticase soy broth (BioMérieux, France), which were made from glycerol stock at −80°C and placed at 30°C overnight without shaking. After 5 seconds of rapid vortexing to homogenize the cultures, 3 µL of two cultures was collected and deposited in one well of a six-well plate containing 4 mL of TSA 1.5% agar supplemented with Congo red 40 µg/mL (Sigma-Aldrich, France) to observe amyloid fiber production and 20 µg/mL Coomassie Brilliant Blue to contrast protein production (33, 34). For strain compatibility experiments, drops were deposited on a well containing TSA 1.5% agar so that their edges were about 2 mm in distance from each other. Samples were dried for 10 min under a hood in case of agar plates and then incubated 6 days at 30°C.

Swarming

In swarming experiments, 20 mL of TSB supplemented with agar to obtain 0.8% final (TSB Agar, BioMérieux, France) was prepared in a Petri dish and dried in a hood for half an hour. Ten microliters of culture was deposited in the center of the plate and left to dry for 10 min before 1 day of incubation at 30°C. For each model, a representative picture from three biological replicates was selected.

Pellicles

Experiments began with 5-mL cultures in trypticase soy broth (Biomérieux, France) made from glycerol stock at −80°C and placed at 30°C overnight without shaking. After 5 seconds of rapid vortexing to homogenize the cultures, 3 µL of two cultures was collected and deposited in one well of a six-well plate containing 4 mL of TS. Samples were incubated at 30°C for 2 days.

Submerged biofilm

Overnight cultures of strains were diluted 100 times in TSB. Two hundred microliters of the suspension was poured for 1.5 h at 30°C for an adhesion step into the wells of polystyrene 96-well microtiter plates with a μClear base (Greiner Bio-One, France) compatible with microscopy. Supernatants were removed, 200 µL of fresh media were added with or without DNAse I (Thermo Scientific, VF304452) at 98 U/mL (35), and the plate was incubated 24 h at 30°C. A solution of 3 µL/mL of SYTO 9, a cell-permeant green dye that labels nucleic acid (Invitrogen, Carlsbad, CA, USA), mixed with 3 µL of TOTO3, a red dye that labels extracellular nucleic acid (Invitrogen, Carlsbad, CA, USA), was prepared, and 50 µL of this solution was added to each well before CLSM acquisitions.

Confocal laser scanning microscopy

Acquisitions were performed using a Leica SP8 AOBS inverted high-content screening confocal laser scanning microscope (HCS-CLSM, LEICA Microsystems, Germany). Image acquisitions were performed on the MIMA2 platform (https://www6.jouy.inra.fr/mima2_eng/). Images every micrometer to capture the full height of the biofilm of 512 × 512 pixels (147.62 × 147.62 µm, pixel size 0.361 µm) were taken at a 600-Hz frequency using the 63× water objective lens (numerical aperture = 1.2). The SYTO9 was excited with an argon laser set at 488 nm, and the emitted fluorescence was collected with a hybrid detector (HyD LEICA Microsystems, Germany) in the range of 500–550 nm. The TOTO3 was excited with the helium-neon laser at 633 nm, and the emitted fluorescence was collected with a hybrid detector in the range of 650–800 nm. For each experiment, three biological replicates were carried out, representing at least 12 technical values per condition.

CLSM image analysis

2D projections of biofilms and movies were obtained using IMARIS 9.3.1 software (Bitplane, Zurich, Switzerland). BiofilmQ software was used to extract quantitative values from image stacks (36). Channels were analyzed separately using the OTSU thresholding method, and “global biofilm properties” were selected to extract the biovolume of the binarized images.

Statistical analysis

Two-way analysis of variance (ANOVA) using uncorrected Fisher’s least was used with PRISM software (GraphPad, USA, California). Data were considered significant when a P value was smaller than 0.05.

DATA AVAILABILITY

The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding author. The whole genome shotgun project has been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA863099.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.00663-24.

Supplemental material. aem.00663-24-s0001.docx.

Supplemental files S1 to S5.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.