Integration of Transcriptome and Metabolome Reveals the Genes and Metabolites Involved in Bifidobacterium bifidum Biofilm Formation

Bifidobacterium bifidum strains, an important component of probiotic foods, can form biofilms on abiotic surfaces, leading to increased self-resistance. However, little is known about the molecular mechanism of B. bifidum biofilm formation. A time series transcriptome sequencing and untargeted metabolomics analysis of both B. bifidum biofilm and planktonic cells was performed to identify key genes and metabolites involved in biofilm formation. Two hundred thirty-five nonredundant differentially expressed genes (DEGs) (including vanY, pstS, degP, groS, infC, groL, yajC, tadB and sigA) and 219 nonredundant differentially expressed metabolites (including L-threonine, L-cystine, L-tyrosine, ascorbic acid, niacinamide, butyric acid and sphinganine) were identified. Thirteen pathways were identified during the integration of both transcriptomics and metabolomics data, including ABC transporters; quorum sensing; two-component system; oxidative phosphorylation; cysteine and methionine metabolism; glutathione metabolism; glycine, serine and threonine metabolism; and valine, leucine and isoleucine biosynthesis. The DEGs that relate to the integration pathways included asd, atpB, degP, folC, ilvE, metC, pheA, pstS, pyrE, serB, ulaE, yajC and zwf. The differentially accumulated metabolites included L-cystine, L-serine, L-threonine, L-tyrosine, methylmalonate, monodehydroascorbate, nicotinamide, orthophosphate, spermine and tocopherol. These results indicate that quorum sensing, two-component system and amino acid metabolism are essential during B. bifidum biofilm formation.


Introduction
Biofilm is the aggregation of microorganisms in which cells are frequently embedded in a self-produced matrix of extracellular polymeric substances (EPSs) that are adherent to each other and/or a surface [1]. These EPSs are mainly polysaccharides, proteins, lipids and nucleic acids that provide mechanical stability of biofilms and form a three-dimensional polymer network that interconnects and immobilizes biofilm cells [2]. A bacterial biofilm is an emergent form of life based on an EPS matrix representing a protected mode for bacterial cells, which exhibit a set of characteristic features such as desiccation tolerance, resource capture, social cooperation and antimicrobial resistance [3]. Biofilm formation is a complex process involving multiple steps, including initial attachment, accumulation, maturation and dispersion [4].
Bifidobacteria, particularly Bifidobacterium bifidum, are believed to be the first colonizers of the human gastrointestinal tract; they can form biofilms on mucosa and food residues in the gut lumen [5]. The ability of various strains of Bifidobacterium to form biofilms on abiotic surfaces, such as glass, stainless steel, polystyrene and complex food matrices, has been well documented [6][7][8]. Environmental stresses, including oxidative stress, bile, low pH value and nutrient starvation, can induce bacterial biofilm formation [9]. Oxygen exposure causes changes in fatty acids in the bifidobacteria cells and an extension of the lag phase of growth; the cells become elongated and develop a rough surface with many nodes due to abnormal or incomplete cell division, and the oxygen-tolerant bifidobacteria seem to defend against oxygen stress through the increase in the content of short fatty acids and cyclopropane fatty acids and the induction of an oxygen stress protein [10]. Transcriptomic analysis of Bifidobacterium longum subsp. longum BBMN68 in response to oxidative shock reveals that 3% oxygen treatment can cause oxidative stress, which contributes to biofilm formation [11]. High concentrations of bile (0.5% and above) may lyse bifidobacterial cells, leading to the release of intracellular signals such as oligopeptides or autoinducer-2 (AI-2) to trigger quorum sensing; extracellular DNA released from lysed cells may also coat the surface and result in additional electrostatic interactions, allowing the adherence of bifidobacteria [12].
Metabolomic analysis has been widely applied to study bacterial biofilm, providing information on metabolic state during biofilm formation [13,14]. The metabolomic results of Helicobacter pylori presented as two distinctly different groups: low-biofilm-formers produced more metabolites than high-biofilm-formers [13]. Liquid chromatography coupled with mass spectrometry (LC-MS) analysis reveals biological and metabolic processes essential for Candida albicans biofilm growth [15]. Untargeted metabolomics revealed that the metabolic state of B. bifidum in the biofilm and planktonic states were different, mainly in aminoacyl-tRNA biosynthesis; alanine, aspartate and glutamate metabolism; arginine and proline metabolism; citrate cycle and nitrogen metabolism; and exopolysaccharide biosynthesis (serves as a major component of biofilm matrix) [16]. The beneficial effects of probiotic biofilms, especially those of Bifidobacterium and Lactobacillus, have gained special interest from academia and the food industry [7,8,17].
The aim of this study was to identify the key genes and metabolites involved in B. bifidum biofilm formation. Here, we established a bifidobacterial biofilm fermentation system that contains wheat fiber (WF) as a carrier and combined a time series transcriptome sequencing and untargeted metabolomics analysis of both biofilm and planktonic cells. Field-emission scanning electron microscopy (FESEM) was used to observe the biofilm cell morphology and extracellular matrix. Our analysis provides insights into the mechanism of bifidobacterial biofilm formation.

B. bifidum FHB150 Form Biofilm on WF
In this study, we used WF as a carrier in the fermentation system and evaluated the biofilm formation by the biofilm formation rate, carrier particle size and FESEM. The pH values in the control and WF cultures of B. bifidum were decreased during fermentation and lower than 4.0 at 32 h (Figure 1a). The pH value in WF culture was significantly lower than that in control culture at 10 h (p < 0.05). There were 1.95 × 10 8 and 1.03 × 10 9 CFU/mL in the control culture and the WF culture at 10 h (Figure 1b), respectively (p < 0.05). These results suggest that cells attached to carriers in the early fermentation stage and proliferated rapidly, producing more acidic substances than control culture. Figure 1c shows WF with an average particle size of around 50 µm. However, the average particle size was over 150 µm (Figure 1d) with the biofilm rate higher than 85% (Figure 1b) at 22 h. The FESEM results showed that B. bifidum cells adhered to the surface of WF ( Figure 1e) and secreted extracellular substances to form the biofilm (Figure 1f). The biofilm rate, carrier particle size and FESEM results indicate that the formation of B. bifidum biofilm includes the adsorption of cells to the carrier at the initial stage (0-10 h), the growth and development of the biofilm (22-32 h) and the dispersion of the biofilm.   Biofilm formation is a complex dynamic process, generally established through surface attachment, biofilm maturation and biofilm dispersion [18]. To identify the key differentially expressed genes (DEGs) during B. bifidum biofilm formation, transcriptomic analysis was conducted on both biofilm and planktonic samples collected at 22

The Interaction of DEGs during B. bifidum Biofilm Formation
To further investigate the interaction of these nonredundant DEGs during B. bifidum biofilm formation, we used the STRING database to identify potential interactions between them. A protein-protein interaction (PPI) network between DEGs was constructed, and five hub genes (degP, groL, groS, hrcA, sigA) were identified by MCODE ( Figure 5). Despite there being 235 nonredundant genes that were differentially expressed, the final PPI network only had 18 genes, which means that the interaction of most DEGs during biofilm formation is still unknown. degP was related to two-component system, upregulated during biofilm formation ( Figure 6b shows that the enriched pathways of these differentially expressed metabolites (arachidic acid, butyric acid, eicosapentaenoic acid, L-cystine, Llysine, L-threonine, melatonin) included valine, leucine and isoleucine biosynthesis; biosynthesis of unsaturated fatty acids; biotin metabolism; butanoate metabolism; aminoacyl-tRNA biosynthesis; lysine degradation; galactose metabolism; glycine, serine and threonine metabolism; cysteine and methionine metabolism; and tryptophan metabolism.

Function of Differentially Expressed Metabolites
To further study the key metabolites during biofilm formation, we used MetaboAna-

Thirteen Pathways Were Identified during the Integration of Both Transcriptomics and Metabolomics Data
To provide a more comprehensive understanding of biofilm formation, the transcriptmetabolite interaction network was generated for DEGs and the differentially expressed metabolites. The gene-metabolite interaction consists of 36 nodes connected via 41 edges (Figure 8). Thirteen pathways were identified during the integration of both transcriptomics and metabolomics data, including ABC transporters; ascorbate and aldarate metabolism; biosynthesis of cofactors; cysteine and methionine metabolism; glutathione metabolism; glycine, serine and threonine metabolism; oxidative phosphorylation; pantothenate and CoA biosynthesis; phenylalanine, tyrosine and tryptophan biosynthesis; quorum sensing; two-component system; valine, leucine and isoleucine biosynthesis; and valine, leucine and isoleucine degradation. The DEGs that relate to the integration pathways included asd, atpB, degP, folC, ilvE, metC, pheA, pstS, pyrE, serB, ulaE, yajC and zwf. The differentially accumulated metabolites included L-cystine, L-serine, L-threonine, L-tyrosine, methylmalonate, monodehydroascorbate, nicotinamide, orthophosphate, spermine and tocopherol. Overall, the integration analysis successfully identified pathways that may be associated with biofilm formation.

Discussion
In this study, we firstly collected samples from B. bifidum under biofilm (cells on the WF surface) and planktonic (cells in the control group) conditions during fermentation and then conducted a combined transcriptomic and metabolomic analysis to determine the key genes and metabolites involved in B. bifidum biofilm formation. Two hundred thirty-five nonredundant DEGs (including vanY, pstS, degP, groS, infC, groL, yajC, tadB and sigA) and 219 nonredundant differentially expressed metabolites (including L-threonine, L-cystine, Ltyrosine, ascorbic acid, niacinamide, butyric acid and sphinganine) were identified during the biofilm formation. Thirteen pathways were identified during the integration of both transcriptomics and metabolomics data, including ABC transporters; quorum sensing; two-component system; oxidative phosphorylation; ascorbate and aldarate metabolism; cysteine and methionine metabolism; glutathione metabolism; glycine, serine and threonine metabolism; and valine, leucine and isoleucine biosynthesis. These results indicate that quorum sensing, two-component system and amino acid metabolism are essential during B. bifidum biofilm formation.
Various methods that rely on molecular and microbiological methods or on chemical or physical properties of the biofilm can be used to assess biofilm biomass and viability [19]. Microscopy methods are important tools for assessing biofilm biomass properties in a more direct way, allowing the description of biofilm heterogeneities, spatial organization and links with the community functions [20,21]. In addition, the diameter change of the biofilm in a dynamic fermentation system can reflect the growth rate of the biofilm [22]. Here, we used WF (with an average particle size of around 50 µm) as a carrier in the fermentation system and evaluated the biofilm formation by the biofilm formation rate, carrier particle size and FESEM. The average biofilm particle size was over 150 µm with the biofilm rate higher than 85% at 22 h (Figure 1). The number of viable bacteria on the WF indicates that the formation of B. bifidum biofilm includes the adsorption of cells to the carrier at the initial stage, the growth and development of the biofilm (22-32 h) and the dispersion of the biofilm. Carriers with porous structures, such as grape seeds, may contribute to biofilm formation on their surfaces by adsorbing cells through their porous structures, and the biofilm formation can be judged by FESEM and the change of particle size [23]. Moreover, choosing a material with a density less than that of culture media as a carrier (such as hollow glass microspheres) in the fermentation system can intuitively reflect the biofilm formation process [24].
Biofilm formation is a complex dynamic process, and the attachment of cells to a carrier is one of the most important processes, representing a turning point from planktonic to the biofilm life mode [18]. Cell appendages such as flagella, fimbriae and pili are involved in this stage. The second messenger cAMP, generated by adenylate cyclase CyaA, has been shown to regulate tad IV pili [25]. Tad IV pili are important surface appendages that are central to the surface-sensing mechanism in the early stages of biofilm formation [26]. High type IV pili/cAMP levels might result in a greater attachment tendency [26]. The expression levels of cyaA and Bbi37|peg.1398 (cAMP receptor) were high during biofilm formation, and Bbi37|peg.1398 was upregulated during the biofilm growth stage. Bifidobacteria have been shown to encode type IV pili, which are associated with biofilm formation and colonization [27][28][29]. Moreover, sortase-dependent pili identified in B. bifidum PRL2010 were shown to promote self-aggregation and aggregation with other gut bacteria [30,31]. Type IV tad pili and quorum sensing are important for the early stages of biofilm formation [26,32].
Interestingly, the functions of DEGs were mainly related to abiotic stimulation, stress response, protein and polysaccharide biosynthesis, two-component system, quorum sensing and global regulatory factors. The genes upregulated were a global phenomenon during B. bifidum biofilm formation : in the early stage of the biofilm, genes related to stress response were upregulated; after the adhesion was completed, cells begin to secrete EPSs leading to biofilm growth, and the related genes were upregulated. YidC, which functions as an integral membrane chaperone/insertase associated with the SecYEG translocon, was identified as a target that can inhibit biofilm formation [33]. Elimination of yidC paralogs in Streptococcus disrupts EPS composition and biofilm development [34]. The tad IV pili gene (tadB) and YidC protein gene (yidC) were both upregulated during biofilm formation (Figure 2d). The two-component system, consisting of a histidine kinase and a cognate response regulator [35], plays an important role in monitoring internal or environmental signals and then translating these stimuli into appropriate cellular responses, and it is also associated with bacterial biofilm formation [36]. Notably, four DEGs (Bbi37|peg.1341, degP, pstS, vanY) associated with the two-component system were found during B. bifidum biofilm formation (Figure 3). The ∆degP probiotic Escherichia coli strain exhibited 20-fold lower biofilm formation than the wild-type strain and lost the ability to inhibit pathogenic biofilm formation via a DegP-mediated interaction [37].
The function of differentially expressed metabolites indicates that amino acid, shortchain fatty acids and vitamins were the main metabolites and their expression levels were different during biofilm formation (Figure 7). Amino acids serve as precursors for energy generation with gluconeogenesis, and less abundance of amino acids in the biofilm state of B. bifidum reflects less energy production, which indicates the biofilm had already attained maturity [16]. Biofilm formation requires cell adhesion, surface conditioning and EPS production, which are energetically expensive processes; however, it is evolutionary justified given the great benefits for bacteria to live embedded in the EPS matrix under oligotrophic conditions; therefore, nascent biofilms in eutrophic systems have reduced lag phases and higher growth rates than biofilms from oligotrophic systems [3,38,39]. In bacterial biofilms, organic acids, including short-chain fatty acids, are mainly produced by the fermentation of sugars and can reduce the pH value [40]. The rate of biofilm formation by Actinomyces naeslundii cells was upregulated by 6.25 m M butyric acid compared to the rate in the control (no short-chain fatty acids) in 96-well microtiter plates, and this upregulation was mediated by GroEL (the heat shock protein) [41]. Another report indicated that the number of biofilms consisting of A. naeslundii cells generated from initial attachment cells in a flow cell system was increased by treatment with 60 mM butyric acid [42]. These reports suggested that the effects of short-chain fatty acids, including butyric acid, could induce the potentiation of the cell status required for initial cell attachment, colonization and biofilm formation [42,43]. Ascorbic acid (vitamin C) has previously been reported to enhance oxidative stress tolerance in Pichia caribbica [44].

Planktonic and Biofilm Culture
B. bifidum FHB150 was obtained from the Jiangnan University (Wuxi, China) and grown anaerobically at 37 • C in MRS broth supplemented with 0.5 g/L L-cysteine hydrochloride monohydrate. Overnight B. bifidum FHB150 culture was inoculated (4%, v/v), and the culture was grown at 37 • C and 120 rpm for 32 h. WF with an average diameter of 50 µm was added to culture media with 4% (w/v) for biofilm culture [45,46]. No WFsupplemented groups were used as planktonic culture. The CFU and pH values in the WF and control culture group were determined at 10, 22 and 32 h during fermentation; for total cell counting in the WF culture group, samples were vortexed (30 s), sonicated (10 s) and vortexed again (30 s) to disperse biofilm cells into the suspension (Vibra Cell Model VCX150PB, Sonics & Materials Inc., Danbury, CT, USA) [24,47].

FESEM
Particles in the WF culture were filtered by a cell sieve (with a diameter of 40 µm), dehydrated with ethanol for 10 min, coated with gold and examined using FESEM (SU8220, Hitachi High-Technologies, Tokyo, Japan) [48]. The low-magnification images acquired by scanning electron microscopy can show the 3D architecture of the biofilm, while the highmagnification images can show the single-cell morphology and EPS organization [16,48] Transcriptome data of B. bifidum FHB150 during biofilm formation have been deposited in the National Center for Biotechnology Information (NCBI) database under BioProject accession code PRJNA733339. The paired-end reads were preprocessed using fastp [49], and reads were aligned to B. bifidum FHB150 genomes using HISAT2 v2.20 [50]. Raw read counts were created using featureCounts [51].

Functional and Pathway Analysis of DEGs
Differential expression analysis of B. bifidum biofilm and planktonic cells was performed using the DESeq2 [52]. Specified pairwise transcriptome comparisons ( [24,53]. Venn diagram was used to show the comparison and overlap between DEGs in different biofilm formation stages [54]. KEGG pathway and GO analyses were conducted using clusterprofiler (R package) [55] for DEGs obtained from different stages.

PPI Network Analysis
The search tool for retrieval of interacting genes (STRING) (https://string-db.org) (accessed on 20 June 2021) database, which integrates both known and predicted PPI networks, can be applied to predict functional interactions of proteins [56]. To seek potential interactions between DEGs in biofilm formation stages, the STRING tool was employed [57]. Active interaction sources, including experiments, text mining, databases and co-expression; species limited to "Bifidobacterium bifidum"; and an interaction score >0.4 were applied to construct the PPI network. Cytoscape software version 3.8.0 was used to visualize the PPI network [58].

LC-MS-Based Metabolome Assay
The column temperature was 40 • C with a flow rate of 0.5 mL/min; the injection volume was 5 µL; mobile phases were water (A) and acetonitrile (B) containing each 0.1% (v/v) of formic acid in positive mode; mobile phases were water with 1 mM ammonium fluoride and 0.1% formic acid (A) and acetonitrile (B) in negative mode; for both modes, the elution gradient (A:B, v/v) was as follows: 80:20 from 0 to 1 min, 0:100 in 7 min and kept 4 min, and then 80:20 at 11.5 min and kept 2 min [13,59]. Raw data handling was done using Compound Discoverer software. MetaboAnalyst 5.0 (https://www.metaboanalyst.ca) (accessed on 20 June 2021) was used for metabolic pathway and function analysis.

Conclusions
In conclusion, a time series global transcription and metabolite profiling of B. bifidum FHB150 showed transcriptional and metabolic changes when this strain was cultivated under biofilm and planktonic conditions. Using the combined transcriptomic and metabolomic analysis, the present study suggests that B. bifidum biofilm formation is associated with quorum sensing; two-component system and amino acid metabolism. However, further studies are needed to confirm this and to elucidate the exact underlying mechanism of B. bifidum biofilm formation.  Data Availability Statement: All raw data for RNA-seq were deposited into NCBI (BioProject accession code PRJNA734093).

Conflicts of Interest:
The authors declare no conflict of interest.