Proteomic response of Turicibacter bilis MMM721 to chicken bile and its bile acids

Bile and its individual components, mainly bile acids, are important for digestion and drive bacterial community dynamics in the upper gastrointestinal tract of chickens. However, specific responses to bile acids have been characterized in only a few commensal bacteria, and it is unclear how other members of the microbiota respond to biliary stress. Here, we used label-free LC–MS/MS to assess the proteomic response of a common inhabitant of the chicken small intestine, Turicibacter bilis MMM721, to 24 h of growth in anaerobic growth media supplemented with 0.1% whole chicken bile, 0.1% taurochenodeoxycholic acid (TCDCA), or 0.1% taurocholic acid (TCA). Seventy, 46, and 10 differentially expressed proteins were identified in Turicibacter bilis MMM721 cultured with supplements of chicken bile, TCDCA, and TCA, respectively, when compared to unsupplemented controls. Many differentially expressed proteins were predicted to be involved in ribosomal processes, post-translational modifications and chaperones, and modifications to the cell surface. Ultimately, the T. bilis MMM721 response to whole bile and bile acids is complex and may relate to adaptations for small intestine colonization, with numerous proteins from a variety of functional categories being impacted.

Turicibacter are anaerobic commensals of the upper GIT of numerous species [14][15][16][17][18]. Recently, Turicibacter bilis MMM721 (henceforth referred to as MMM721) was isolated from chicken eggshells during a study of bacterial succession and is capable of colonizing the small intestine where it likely interacts with bile and BAs [18].
Despite the importance of bacteria-BA interactions, few studies have measured the proteomic response of individual gut bacteria to bile and BAs. The ubiquitous nature of Turicibacter spp. in the upper intestinal tract of vertebrate species suggests the genus has mechanisms through which it can tolerate the stress induced by BAs. Here, we use survival assays and LC-MS/MS proteomics to interrogate the response of MMM721 to chicken bile and its BA components: TCDCA and TCA.
For growth assays, colony forming units (CFU)/mL were determined on BHIGL immediately after inoculation using previously described methods [20]. After 24 h, cultures were plated on BHIGL agar and incubated for 48 h to determine CFU/mL. Three biological replicates were conducted for each condition tested with two technical replicates each.
For proteomic analysis, cells from broth cultures described above, were harvested via centrifugation. Pellets were washed twice with 1•phosphate-buffered saline (PBS) and resuspended in 3 mL of cold 50 mM ammonium bicarbonate buffer (pH 8.0) prior to freezing at -80 °C. Three biological replicates of each strain were conducted for each condition tested. All centrifugation steps were 4000×g for 20 m at 4 °C.

LC-MS/MS sample preparation
Cell pellets were thawed and distributed into 2×1.5 ml screw-cap bead-beating tubes filled with 300 μl of 0.1 mm zirconia/silica beads (BioSpec Products, Inc.). Tubes were placed on a Vortex-Genie with tube-holder attachment (USA Scientific) and agitated at maximum speed for 3 min followed by 5 min on ice. This was repeated for 5 cycles, after which samples were spun for 10 m (16000 g) at 4 °C. Supernatant was removed and protein concentrations were determined using a Bradford protein assay (Bio-Rad). Samples with low protein concentrations were concentrated with Amicon Ultra-15 (10 kDa cut-off ) Centrifugal Filter Units (Millipore). Trypsin/Lys-C Mix, Mass Spec Grade (Promega) was used to digest 40 μg of protein from each sample. Samples were then desalted, dried, and resuspended using previously described methods prior to LC-MS/MS [23]. All preparation steps were conducted according to the manufacturer's instructions.

Data analysis
Raw data files were processed in MaxQuant v1.6.7.0 and Perseus v1.6.7.0 using previously described parameters [23][24][25]. The Prokka-annotated Turicibacter bilis MMM721 genome was used to construct a protein database for proteomic analysis. Protein expression levels were transformed to log2 quantities. Peptides were considered significantly differentially expressed peptides (DEPs) if they had a false discovery rate q < 0.05 and a log 2 difference of greater than 0.6 or less than −0.6. Statistical analyses and figure generation were conducted in R [26,27].

Preliminary growth curve
Growth was observed in 0.1% bile but not in 1.0% bile. Growth curves for MMM721 in BHIGL + Bile, BHIGL + TCDCA, BHIGL + TCA, and BHIGL suggested the strain reached stationary phase somewhere between 12-20 h after inoculation (Additional File 1). An incubation time of 24 h was chosen for the survival assay and proteomics to allow time for MMM721 to mount a response to supplemented bile and BAs.

Proteomic data confirm differential response of MMM721 to bile and TCDCA
Label-free LC-MS/MS was used to identify and quantitate DEPs in the supplemented media relative to BHIGL alone. Of the 70 DEPs identified in BHIGL + Bile compared to BHIGL, 35 were increased and 35 were reduced ( Fig. 2A, Table 1). Clusters of Orthologous Group (COG) category G (carbohydrate transport and metabolism) had the most increased DEPs ( Fig. 2A). COG category J (translation, ribosomal structure, and biogenesis) had the most decreased DEPs in MMM721 grown in BHIGL + Bile, though several COG J DEPs were also highly expressed (Table 1).
Ten DEPs (8 increased/2 reduced) were identified when comparing BHIGL + TCA to BHIGL. Both COG categories G and J had multiple proteins increased in BHIGL + TCA (Fig. 2C). Several of these proteins were undescribed cytoplasmic proteins (Table 1). A dehydrogenase (ONGNNMAF_01745) belonging to COG C (Energy production and conversion) and CplX (ONGN-NMAF_00641), an ATPase belonging to COG O, were both increased in BHIGL + TCA.
Two DEPs were increased in all treatments, peptidase dimerization protein (ONGNNMAF_02259) and d-alanine-d-alanine ligase (ONGNNMAF_01948) (Fig. 2D, Table 1). GMP synthase (ONGNNMAF_01593) and an intersubunit bridge protein (ONGNNMAF_01914) were decreased across all treatments ( Fig. 2E; Table 1). A full summarization of all the DEPs in the study can be found in Table 1.

Discussion
Bile and BAs are critical in shaping the GIT microbiota, especially in chickens where current production feed models result in long periods of bile flow [28][29][30][31][32]. Microbial responses to bile and BAs impact host immune regulation and weight gain, yet proteomic responses of individual microbes to bile and BAs are not well understood [1,2]. Turicibacter bilis, and other gut microbes, interact with bile and BAs, using them as an environmental cue for germination and biotransforming them in vitro [18,[33][34][35]. Here, we describe the growth and proteomic responses of Turicibacter bilis MMM721 to chicken bile and its main BA components: TCDCA and TCA. Bacterial CFU/mL after 24 h were determined on BHIGL agar. The dashed line denotes the mean CFU/mL at inoculation (hour 0). *differences between the initial CFU/mL immediately after inoculation vs. CFU/mL after 24 h as determined by the t.test function in R (p.adjusted < 0.05). Error bars represent standard error of the mean MMM721 grown in BHIGL + Bile or BHIGL + TCDCA showed a reduction in CFUs and similar proteomic responses after 24 h. These similarities were not unexpected, as the most abundant BA in chickens is TCDCA [4,[11][12][13]. Previous studies and work from our lab demonstrated the ability of Turicibacter, including MMM721, to deconjugate taurine from TCDCA and TCA, forming CDCA and CA, respectively [35]. Unconjugated BAs more easily cross into the bacterial cell, though CDCA does so more efficiently, disrupting membrane integrity to a greater extent than CA, partially explaining the differences in growth and proteomic response [9,28,36].
Many DEPs in COG category J, translation, ribosomal structure, and biogenesis, were significantly decreased in BHIGL + Bile and BHIGL + TCDCA. Decreased protein synthesis would manifest as an overall decrease in bacterial growth, matching the observations in this and other studies [37][38][39][40]. COG category O, post-translational modification, protein turnover, and chaperones, DEPs were enriched in BHIGL + Bile when compared to BHIGL. Both BHIGL + Bile and BHIGL + TCDCA had molecular chaperones increased, likely in response to the ability of BAs to unfold cytoplasmic proteins, ensuring the proper folding of proteins within the cell [8,41].
Only two DEPs were increased in all treatments. One was a peptidase dimerization domain protein that plays a protective role under BA stress [42,43]. A d-alanined-alanine ligase involved in peptidoglycan biosynthesis was also increased, which may result from the disruptive effects BAs have on the bacterial cell surface [44,45]. GMP synthase and an intersubunit bridge protein were decreased across all treatments. Subunit bridges are vital for the assembly and initiation of ribosomal translation, highlighting the decreased translational rate of MMM721 in response to BAs [46]. GMP synthase is responsible for the first step of de novo GMP synthesis. Decreased GMP synthase would reduce the GMP pool and potentially replication rate [47,48].
The reduction in CFU/mL observed in the BHIGL + Bile and BHIGL + TCDCA suggests that stress response proteins observed in the proteomic data could be related to the stress the CDCA causes MMM721 or due to MMM721 cell death. However, studies show the half-life for bacterial proteins is approximately 20 h, so many of the DEPs observed in this study likely reflect the MMM721 response to TCDCA and whole chicken bile [49,50]. Proteomic and transcriptomic analyses of earlier timepoints in the MMM721 growth cycle would      be beneficial to more fully characterize the MMM721 response to TCDCA and whole bile. The observed DEPs from the BHIGL + TCA treatment likely reflect the response of MMM721 to TCA. Of note, both a dehydrogenase (ONGNNMAF_01745) belonging to COG C (Energy production and conversion) and CplX (ONGNNMAF_00641), a subunit of an ATP-dependent Clp protease belonging to COG category O were significantly increased in BHIGL + TCA. CplX is an ATPase involved in several regulatory and proteolytic processes. Lactobacillus delbrueckii increases expression of CplX upon exposure to bile salts, suggesting this protein plays an important role in adaptation to BA stress [51,52]. Dehydrogenases represent a broad family of enzymes that catalyze reduction reactions, with some classes, like hydroxysteroid dehydrogenases, capable of modifying the steroid ring of BAs themselves [53]. More work should be done to characterize the activity of the various MMM721 dehydrogenases. Many of the DEPs identified in this study were unable to be identified or had unknown functions. Further work characterizing the genome of MMM721 would be beneficial in future transcriptomic and proteomic studies.

Conclusions
The present study constitutes the first characterization of the Turicibacter bilis response to chicken bile and individual bile acids using culturing and label-free LC-MS/MS strategies. Comparisons of the differentially expressed peptides were obtained from Turicibacter bilis MMM721 exposed to chicken bile, TCDCA, and TCA. Whole chicken bile and TCDCA both reduced MMM721 growth after 24 h of exposure, producing similar proteomic responses. Exposure of MMM721 to TCA did not substantially inhibit cell growth and up-regulated proteins involved in metabolism and post-translational modifications. Ultimately, the Turicibacter bilis MMM721 response to whole bile and bile acids is complex, involving proteins from several pathways. Understanding how members of the microbiota respond and modify BAs is critical to optimize animal nutrition and maintain efficient production.

Limitations
Many of the DEPs have not yet been identified or characterized, making it difficult to determine responsive proteins and pathways and highlighting the importance of further characterization work for Turicibacter bilis