Introduction

The human intestinal tract houses an intricate microbial community that significantly impacts human health1,2,3. Gut bacteria are significantly associated with various diseases, such as cancer4, diabetes5, hypertension6,7, obesity8, hyperlipidemia9, and life expectancy10. Therefore, given its widespread impact on human health, it is essential to comprehend the factors influencing the microbiome, especially at the community and strain levels.

Bile acids (BAs) are essential endogenous molecules that originate from cholesterol in the liver, which are subsequently transported into the intestine via the gallbladder, and metabolized into secondary BAs by gut microbiome11,12. BAs are known to activate nuclear receptors and play a critical role in the regulation of metabolism and immune function in the body. Consequently, BAs have been implicated in the pathogenesis of various diseases, including diabetes, obesity, atherosclerosis, inflammatory bowel disease (IBD), colorectal cancer, liver cancer, and breast cancer2,13. BAs also play a crucial role in regulating the growth, metabolism, and composition of the gut microbiota14,15,16. BAs not only inhibit the growth of harmful bacteria such as Clostridium difficile17, cholera18, and Entamoeba histolytica19, but also modulate the antibiotic resistance of gut bacteria15. Additionally, BAs promote the maturation of newborn’s gut microbiota20 and promote intestinal cancer development through gut microbiota21. Therefore, the role of BAs in shaping gut microbiota has important implications for human health.

Our understanding of the effect of BAs on gut bacteria is limited, primarily due to the complex composition of both entities22,23. We employed a comprehensive approach encompassing bacterial growth rate measurements (OD600), 16S rRNA gene sequencing, culturing of complex communities, transcriptomics, and metabolomics. Our study has shed light on the intricate effects of BAs on intestinal bacteria, demonstrating that different types of BAs exhibit varying effects on gut microbiota both in vivo and in vitro. These effects are mediated in part by the effect of BAs on the ribosome, as well as the bioaccumulation of BAs by bacteria. These findings pave the way for future investigations into the role of BAs in host pathophysiology through their modulation of the intestinal microbiota.

Result

BAs perturbs the growth of human gut bacteria at physiological concentrations

To investigate the impact of different BAs on various gut bacteria, we constructed a panel of 65 strains from 6 phyla commonly found in the human gut microbiota (Supplementary Table 1). Each intestinal bacterial strain was exposed to 21 BAs at concentrations within the normal physiological range of the human intestinal tract, covering more than 80% of the BAs found in the intestinal tract24,25 (Supplementary Table 2). We screened the effect of BAs on strains in 384-well plates and monitored growth over time by measuring optical density (OD600) (Supplementary Fig. 1A).

Our study revealed a broad spectrum of activity among the 21 types of BAs tested, with 14 showing activity against at least three bacterial species (Fig. 1A, B). Interestingly, despite having similar structures, BAs exhibited significant differences in their effects on the growth of various strains (Supplementary Fig. 1B). Unconjugated BAs, particularly DCA and CDCA, exhibited stronger inhibitory effects on different gut microbiota strains compared to conjugated BAs. Interestingly, despite being isomers, DCA, CDCA, and UDCA displayed different inhibitory effects on the 65 strains of bacteria tested. For instance, 1 mM of DCA inhibited the growth of 55 strains, while 1 mM of UDCA inhibited the growth of only 22 strains, with a much weaker inhibitory effect than CDCA and DCA (Fig. 1B). Contrastingly, some BAs were found to promote the growth of some strains, such as the promoting effects of UDCA and TUDCA on the growth of Bifidobacterium pseudonumeratum and Bifidobacterium pseudopodium (Fig. 1C, D, Supplementary Fig. 1C).

Fig. 1: Effect of BAs on 65 strains of intestinal bacteria.
figure 1

A Heatmap showing the sensitivity or resistance of each bacterial strain to different concentrations of BAs. The redder the color, the stronger the inhibitory effect. B Number of bacterial strains exhibiting an inhibition rate >25% among 65 strains tested for each BAs. C The effect of UDCA, GUDCA, and TUDCA on the growth of Bifidobacterium pseudocatenulatum. Each group included six replicates (n = 6). D The effect of UDCA, GUDCA, and TUDCA on the growth of Bifidobacterium pseudolongum. Each group included six replicates (n = 6). E Heatmap showing the inhibitory effect of 21 BAs on phyla. C, D data with error bars represent mean ± SD. C, D Data were analyzed by two-tailed unpaired Student’s t test. *P < 0.05, **P < 0.01, ****P < 0.0001.

Full size image

The impact of BAs on various bacterial phyla was assessed, revealing distinct responses. Proteobacteria (8) and Verrucomicrobia (1) appeared to remain unaffected by BAs, whereas Firmicutes (34), Fusobacteria (1), Actinobacteria (7), and Bacteroidetes (12) exhibited significant inhibition in the presence of BAs (Fig. 1E). However, 21 BAs exhibited varying inhibitory strains within the same bacterial phylum, and the same BAs demonstrated different inhibitory effects across 6 bacterial phyla (Supplementary Fig. 1D). These results indicate that different BAs have distinct effects on various bacterial phyla.

Previous studies have reported the different inhibitory effects of isoalloLCA on intestinal bacteria in different culture media16. To further demonstrate whether DCA has similar effects in different media, we studied the effect of DCA on 2 strains in modified Gifu Anaerobic Medium (mGAM) and brain heart infusion medium (BHI). Our study confirmed that different concentrations of DCA in mGAM and BHI have varying effects on Streptococcus thermophilus and Lactobacillus rhamnosus (Supplementary Fig. 2A).

Effects of BAs on the composition of human gut microbiota in vitro

To scrutinize the influence of BAs on the gut microbiota composition, our study specifically targeted DCA, CDCA, and LCA. This selection is substantiated by two primary considerations: (1) These three BAs rank as the foremost constituents in the human intestine, collectively encompassing over 50% of the BAs profile within intestinal bacteria24,25. (2) Demonstrating the utmost potency, these three BAs exert profound effects on diverse strains of intestinal bacteria, underlining their pivotal role in shaping microbial communities. In vitro, we incubated DCA, LCA, and CDCA with fecal bacteria from healthy individuals in mGAM, a medium known to maintain the composition of gut microbiota26. Results revealed that DCA, LCA, and CDCA had significant effects on the gut microbiota, with 185, 230, and 290 different operational taxonomic units (OTUs), respectively, compared to the control group. Moreover, there were also substantial differences in OTUs among the different BAs treatment groups (Fig. 2A). DCA, CDCA, and LCA significantly increased the relative abundance of Proteobacteria compared to the control group, which was consistent with the effect of BA on single bacteria in vitro (Fig. 2B). No significant difference was observed in α diversity analysis between the DCA, LCA, CDCA group and control group (Supplementary Fig. 3A). In β diversity analysis, DCA and CDCA group was separated from the control group, and then there was no difference between LCA and the control group (Fig. 2D).

Fig. 2: BAs alters the fecal flora of healthy patients.
figure 2

A The Venn diagram illustrated the overlapped OTUs between the group of control, DCA, CDCA, and LCA. B Accumulation of bacterial phyla with their relative abundance in the gut microbiota between the group of control, DCA, CDCA, and LCA; C The abundance of Proteobacteria in the CDCA, DCA, LCA, and control groups. D Principal Coordinates Analysis (PCoA)showed β Diversity in CDCA, DCA, LCA, and control groups. E–G Linear discriminant analysis (LDA) coupled with effect size measurements identifies significant abundance. Green indicates the vehicle control group. Red indicates the BAs group (CDCA (E), LCA (F), DCA (G)). Only taxa with LDA score greater than 4 are shown. C Data with error bars represent min to max. C Data were analyzed by two-tailed unpaired Student’s t test. *P < 0.05, **P < 0.01.

Full size image

The analysis using the Metastat method and Student’s t test revealed that CDCA and DCA significantly increased the abundance of Proteobacteria at the phylum level compared to the control group (Fig. 2C, Supplementary Fig. 3D). Additionally, DCA significantly inhibited the abundance of Bacteroides (Supplementary Fig. 3B). The LEfSe analysis showed that at the genus level, Escherichia coli, which belongs to Proteobacteria, was the main enriching bacteria in the DCA and CDCA groups compared to the control group. Furthermore, CDCA also increased the abundance of Lachnoclostridium and Escherichia/Shigella. In contrast, Parabacteroides was the most abundant bacterial species in the control group, compared to the CDCA and DCA groups. Similarly, Paeniclostridium was the most abundant bacterial species in the control group, compared to the CDCA and LCA groups (Fig. 2E–G). These findings provide further evidence that DCA and CDCA inhibit the growth of other bacteria and promote the relative abundance of Proteobacteria.

We utilized 16S rRNA sequencing to predict metabolic pathways and observed that DCA and CDCA upregulated the expression of key genes involved in bacterial motility, ribosome biogenesis, phosphotransferase system, and other related pathways in gut bacteria. Moreover, DCA and CDCA down-regulated the ribosome and amino-acid energy metabolism pathways (Supplementary Fig. 3C). These findings suggest that BAs may modulate bacterial function through the regulation of these crucial pathways.

Effects of BAs on the composition of gut bacteria in vivo

Due to the complexity of the intestinal environment, the exact impact of BAs on gut bacteria in vivo remains elusive. To elucidate this relationship, rats were orally administered CDCA, DCA, LCA, and Solvent sodium carboxymethyl cellulose (CMC), and fecal samples were subsequently collected for 16S rRNA sequencing analysis. (Supplementary Fig. 4A). Result showed that compared with the CMC group, the α diversity of gut bacteria was significantly altered in the CDCA and DCA groups, particularly in the CDCA group (Fig. 3A). The β diversity between the CDCA and DCA was also significantly different compared to the control group, whereas there was no significant difference between the LCA group and the control group (Fig. 3B, C, Supplementary Fig. 4B). Notably, the β-diversity of DCA and CDCA were clearly distinguishable from that of the LCA group (Supplementary Fig. 4C). Furthermore, compared to the control group, the composition of the microbiota at the phylum level was altered by DCA, CDCA, and LCA, particularly with an increase in the abundance of Proteobacteria (Supplementary Fig. 4D), consistent with the trend observed in human fecal bacteria in vitro. The results of OTUs also showed significant differences between the BAs group and the control group, as well as between the BAs groups (Supplementary Fig. 4E). To assess changes in bile acid levels after administration, we analyzed the BAs profile in the cecum (Supplementary Table 3). Following administration of DCA, LCA, and CDCA to rats, we observed significant increases in DCA, LCA, and CDCA levels in the cecum by 3.9-fold, 1.7-fold, and 6.6-fold, respectively (Supplementary Fig. 4F, H). To determine whether the effects of CDCA on intestinal bacterial composition are mediated by its metabolites, we investigated the impact of 7 bile acids (each at 0.3 mM), which are significantly elevated following CDCA administration, on the growth of 42 intestinal bacterial strains. Our findings demonstrate that CDCA exerted the most substantial influence on these strains, while the other significantly altered bile acids exhibited partial inhibitory effects on specific strains (Supplementary Fig. 4I, J). After oral administration of CDCA, it showed the most significant changes in the cecum. Therefore, CDCA plays a major role in influencing the intestinal flora, while other bile acids partially contribute to altering the composition of gut bacteria.

Fig. 3: The effect of BAs on the composition of gut bacteria in vivo.
figure 3

A Alpha diversity box plot (Ace) between the group of control, DCA, CDCA, and LCA. B Principal coordinate analysis (PCoA) of beta diversity between the group of CDCA and control. C PCoA of beta diversity between the group of DCA and control. D–F Feces microbiome communities are significantly different between the group of control, DCA, CDCA and LCA. LDA score computed features differentially abundant between the control and CDCA (D), LCA (F), or DCA (E) gavaged rat. The criteria for feature selection was LDA Score >4. A Data with error bars represent min to max. A Data were analyzed by two-tailed unpaired Student’s t test. *P < 0.05, #P < 0.0001.

Full size image

After administration of CDCA, LEfSe analysis indicated a significant increase in the abundance of Bacteroidales, Alloprevotella, Prevotella_9, and Muribaculaceae (Fig. 3D). Similarly, administering DCA led to a significant increase in the abundance of Enterrhabdus, Alistipes, Dubosiella, Candidatus Saccharimonas, Muribaculaceae, and Akkermansia (Fig. 3E). Administration of intragastric LCA resulted in a significant increase in the abundance of Muribaculaceae (Fig. 3F). DCA, LCA and CDCA increased the abundance of Muribaculaceae in vivo. DCA had the greatest impact on intestinal bacteria, followed by CDCA and LCA, which is consistent with the trend of impact observed in vitro.

Intestinal bacteria inhibit the growth of bacteria through the bioaccumulation of BAs

Compounds can be accumulated by gut bacteria, thereby affecting the growth and composition of gut bacteria27. To investigate whether intestinal bacteria accumulate BAs, we incubated DCA, CDCA, CA, and UDCA with 22 strains of intestinal bacteria and measured the accumulation of the four BAs. To quantify the bacteria, we measured their OD values, as the concentration of bacteria varied between strains. The results revealed that each strain had a distinct capacity for BAs accumulation (Supplementary Fig. 5A–C, Fig. 4A). Notably, spearman correlation analysis confirmed that the accumulation of DCA and UDCA by the 22 strains of gut bacteria was significantly positively correlated (r = 0.61, p = 0.0013), while the accumulation of CDCA and CA was also significantly positively correlated (r = 0.60, p = 0.0016). In contrast, the correlation between DCA and CDCA and CA, and the correlation between UDCA and CDCA and CA, was weak (Supplementary Fig. 5G). This finding were attributed to the fact that some bacteria can metabolize CDCA and CA, while they cannot metabolize DCA and UDCA (Supplementary Table4). Moreover, we incubated Enterococcus faecalis, Providencia stuartii, and Escherichia coli K-12 with different concentrations of DCA and found that the accumulation of DCA by these bacteria increased with increasing DCA concentrations (Supplementary Fig. 5D–F).

Fig. 4: Intestinal bacteria are inhibited by accumulation of BAs.
figure 4

A Quantification of DCA concentrations in 22 strains by UHPLC-MS (n = 5–6). B Linear fitting analysis between DCA accumulation of 22 strains and inhibition rate of DCA (0.5 mM) to the 22 strains. C Correlation between DCA inhibition rate and accumulation by Spearman analysis. D–F Effects of different concentrations of PAβN and DCA on the growth curves of EDTE (D), E.coli-ED1a (E), and E.coli-k12 (F) in mGAM medium. PAβN(H):0.15 mg/ml; PAβN(M):0.1 mg/ml; PAβN(L):0.05 mg/ml (n = 6). G-I Effects of different concentrations of PAβN and DCA on the area under the curve of EDTE (G), E.coli-ED1a (H), and E.coli-k12 (I). A, D–I Data with error bars represent mean ± SD.

Full size image

The inhibitory effect of DCA on the 22 strains bacteria increased with increasing concentration in mGAM medium (Supplementary Fig. 6A). Linear fit analysis showed a significant positive correlation between DCA accumulation and inhibition rate, with a correlation coefficient of 0.60 (Fig. 4B). Moreover, spearman correlation analysis confirmed the significant correlation between DCA accumulation and inhibition rate, with a correlation coefficient of 0.63 (p = 0.0031) (Fig. 4C). These results indicate that the accumulation of DCA within bacteria is positively correlated with their sensitivity to DCA inhibition.

Both LCA and DCA are secondary BAs known for their high toxicity, with DCA exhibiting stronger antibacterial effects compared to LCA28. To investigate whether the differential sensitivity of bacterial strains to DCA and LCA is related to their accumulation by these strains, we compared 22 bacterial strains incubated with equivalent concentrations of DCA and LCA. Interestingly, certain strains, such as Bacteroides, showed greater sensitivity to DCA despite accumulating higher levels of LCA than DCA (Supplementary Fig. 5H). This suggests that strain sensitivity to DCA and LCA is not solely determined by their accumulation. To assess the proportion of BAs accumulated by different strains relative to the total BAs, we calculated the accumulation percentage of DCA and LCA across 22 strains (Supplementary Fig. 5I). Additionally, correlation analysis of BAs accumulation by different bacterial strains revealed a significant positive correlation between the accumulation of DCA and LCA (Supplementary Fig. 5J).

To further confirm that the inhibition of gut bacteria by BAs is due to their accumulation within the cells, we incubated three Escherichia coli strains (E.coli-ED1a, E.coli-ETEC, E.coli-K12), Bacteroides vulgatus, Bacteroides ovatus and Clostridium difficile with different concentrations of DCA and Phe-Arg-β-naphthylamide dihydrochloride (PAβN) (H- PAβN; M- PaβN; L- PAβN), an inhibitor of efflux pumps that can increase the intracellular accumulation of substrates29,30,31. Our results showed that DCA alone did not significantly inhibit the growth of the three Escherichia coli strains, but when used in combination with PAβN, we found that as the concentration of DCA increased, the inhibitory effect of PAβN on the three strains of E. coli demonstrated a corresponding escalation (Fig. 4D–I). Similarly, DCA had a inhibitory effect on the growth of Bacteroides vulgatus, Bacteroides ovatus and Clostridium difficile. However, when PAβN and DCA are used in combination, they completely inhibited the growth of three strains (Supplementary Fig. 5K, L). These results provide further evidence that BAs inhibit bacterial growth by entering the cells and accumulating within them. To further demonstrate that compound PAβN enhances the antibacterial effect of DCA by increasing DCA accumulation, we tested the accumulated DCA content in E.coli strains (E.coli-ED1a, ETEC, E.coli-K12) and Clostridium difficile in the presence or absence of compound PAβN. Our results show that compound PAβN significantly increased the DCA accumulation in these four strains. (Supplementary Fig. 5M, N).

DCA impacts conserved pathways necessary for bacteria growth

To investigate the impact of BAs on bacterial metabolism, we conducted RNA sequencing (RNA-seq) analysis to identify differentially expressed transcripts in the presence or absence of DCA. We selected two strains with different sensitivities to DCA, Bacteroides ovatus and Escherichia coli K-12 (Supplementary Fig. 7A, B). Our results showed that DCA significantly altered the transcriptional profile of both strains compared to the control group by principal components analysis (Fig. 5A, B). Notably, the sensitive strain Bacteroides ovatus exhibited a more pronounced transcriptional response, with 83% of genes affected by DCA, compared with 76% of genes affected in Escherichia coli K-12 (Fig. 5C, D). Specifically, DCA exposure induced transcriptional changes in 1020 genes in Bacteroides ovatus, while only 229 genes were affected in Escherichia coli K-12, highlighting the differential impact of DCA on the two bacterial strains (Fig. 5C, D).

Fig. 5: Effect of DCA on transcriptome of Bacteroides ovatus and Escherichia coli K-12.
figure 5

A, B PCA of RNA-seq data from Bacteroides ovatus (A) and Escherichia coli K-12 (B) after 6 hours of incubation with solvent and DCA. C, D Volcano plots showed significance change of genes in response to DCA in Bacteroides ovatus (C) and Escherichia coli K-12 (D) after 6 hours of exposure. E, F Heatmaps of Gene Ontology (GO) pathway enrichments of differentially expressed genes in bacterial isolates treated with 0.4 mM of DCA compared to the vehicle control. Colors represent false discovery rate (FDR), and size represents the number of genes in Bacteroides ovatus (E) and Escherichia coli K-12 (F).

Full size image

We then performed gene ontology (GO) and KEGG pathway analysis to explore the functional pathways affected by DCA. GO pathway analysis revealed that most of the enriched pathways in Bacteroides ovatus were related to ribosomes, including the structural construction of ribosomes, rRNA binding, and ribosomes (Fig. 5E). In Escherichia coli K-12, the enrichment of the GO pathway was mainly concentrated in L-arginine metabolism (Fig. 5F). The first 20 KEGG pathways with FDR values revealed that ribosome and oxidative photosynthesis were the most affected pathways in Bacteroides ovatus, which is consistent with the enrichment results of GO. Moreover, changes were observed in Taurine and Tryptophan metabolism in Bacteroides ovatus. In addition to the amino acid metabolism that was enriched in Escherichia coli K-12, which is consistent with the enrichment results of GO, changes in the ribosome pathway were also observed. Among these amino acid metabolic pathways, the main ones affected were Arginine proline Alanine, aspartate and glutamine metabolism, and Arginine, Valine, leucine, and isoleucine biosynthesis (Supplementary Fig. 7C, D). In short, DCA may impact bacterial growth by affecting ribosome function and amino acid metabolism after entering the bacteria.

Effects of BAs on the ribosome and amino-acid metabolism

To investigate the impact of BAs on ribosome subunits, we evaluated their effect on mRNA expression levels in Escherichia coli K-12 and Bacteroides ovatus. The result showed that DCA and CDCA inhibited the mRNA expression of ribosomal L2, L3, L4, and S3 subunits in Escherichia coli K-12, while L22 expression remained unaffected (Fig. 6A). Interestingly, CDCA had a pronounced inhibitory effect on the transcription of ribosomal subunits L2, L3, L4, S3, and L22 in Bacteroides ovatus, whereas DCA up-regulated the transcription of ribosomal L2, L3, L4, S3, and L22 in Bacteroides ovatus (Fig. 6B). As the ribosomal gene is a crucial component of bacterial housekeeping genes, many antibiotics, such as aminoglycosides, tetracyclines, chloramphenicol, macrolides, lincomycin, etc. can affect ribosomal function32,33.

Fig. 6: Effect of BAs on ribosome and amino-acid metabolism of intestinal bacteria.
figure 6

A Differential gene expression in Bacteroides ovatus treated with DCA, CDCA, or solvent. B Differential Gene Expression in E. coli K12 Treated with DCA, CDCA, or solvent. C AAs Concentration in Bacteroides ovatus Treated with DCA, CDCA, or solvent. D AAs concentration in Ecoli-K12 Treated with DCA, CDCA, or solvent. E AAs in rat plasma treated with DCA or solvent. A–E Data with error bars represent mean ± SD. A–E data were analyzed by two-tailed unpaired Student’s t test. C–E data were analyzed by Multiple t tests-one per row. *P < 0.05; **P < 0.01, ***P < 0.001, #P < 0.0001.

Full size image

To further demonstrate the effect of BAs on amino acid metabolism, we investigated the effects of BAs on the 39 types of amino acids and their metabolites (AAs) of Escherichia coli K-12, Bacteroides ovatus, Bacteroides vulgatus, Ecoli-ED1a, and Providencia stuartii in mGAM medium by UHPLC-MS (Supplementary Table5). The results indicated that DCA (0.1 mM) affected the metabolism of six AAs in Bacteroides ovatus, while CDCA (0.2 mM) had an obvious impact on four AAs in Bacteroides ovatus (Fig. 6C). DCA (0.5 mM) and CDCA (0.5 mM) affected the metabolism of two and seven AAs, respectively, in Escherichia coli K-12 (Fig. 6D). In Bacteroides vulgatus, DCA (0.1 mM) and CDCA (0.2 mM) affected the metabolism of fifteen and five AAs, respectively (Supplementary Fig. 8A). In Ecoli-ED1a, DCA (0.5 mM) and CDCA (0.5 mM) impacted the metabolism of two and one AAs, respectively (Supplementary Fig. 8B). Lastly, in Providencia stuartii, DCA (0.5 mM) and CDCA (0.5 mM) affected the metabolism of six and eight AAs, respectively (Supplementary Fig. 8C).

To investigate the effect of DCA on AAs in vivo, we administered DCA to rats and measured the plasma amino acid levels and cecal contents composition. Our results showed that DCA significantly altered the composition of 17 amino acids in rats AAs (Fig. 6E). Additionally, we found that DCA significantly changed the content of 21 amino acids in the cecum compared to the control group (Supplementary Fig. 8D). Given that bile acid significantly impacts both amino acid metabolism in various strains and the composition of intestinal bacteria, the observed changes in amino acid composition within the body may result from a combination of these effects.

Discussion

In the intestinal milieu, a plethora of BAs types and quantities abound, wielding a pivotal influence on human pathophysiology through their intricate interactions with gut bacteria14,15,16. However, the effect of BAs on gut bacteria has not been systematically studied. Our results indicated that BAs not only had different effects on 65 intestinal strains but also had important effects on the overall composition of intestinal bacteria in vitro and in vivo. Additionally, the more intestinal strains accumulated BAs, the stronger the inhibitory effect of BAs on intestinal strains. Once inside the gut bacteria, BAs affect ribosomal function and amino acid metabolism, further contributing to the inhibition of bacteria. The findings contribute to the current understanding of the structure and hierarchical organization of the gut microbiome, as well as the development of BA-based antibiotics. Additionally, this work has the potential to deepen our understanding of the effects of BAs on the body’s pathophysiology by affecting gut bacteria. However, Further research is needed to fully elucidate the molecular mechanisms underlying these effects and to explore potential clinical applications.

Our results demonstrated that BAs had broad and significant effects on the human gut microbiota, which extends beyond what was previously appreciated18. Although the composition of gut bacteria differs between rats and humans, common trends in the microbiota’s response to BAs were observed from the phylum level in both in vivo and in vitro experiments34,35. Specifically, BAs can increase the abundance of Proteobacteria in vivo and in vitro, which is concerning as higher levels of this phylum have been linked to metabolic diseases, cancer, inflammation, and other adverse health outcomes36,37,38. The rise of DCA and CDCA, which have been associated with enteritis, colorectal cancer, and liver cancer, may contribute to these negative effects by increasing the abundance of Proteus39,40. Meanwhile DCA, LCA, and CDCA have been observed to reduce the abundance of Firmicutes, which is the predominant phylum in the gut. This reduction of Firmicutes by BAs may disrupt the composition of the gut microbiota. However, regarding Bacteroidetes, DCA and CDCA demonstrate substantial reductions in vitro, yet interestingly, they paradoxically increase the relative abundance of Bacteroidetes in vivo. LCA, conversely, shows slight impact on Bacteroidetes levels in both in vivo and in vitro. Moreover, BAs also have a notable impact on probiotics. For example, UDCA and TUDCA can promote the growth of Bifidobacterium pseudonumeratum and Bifidobacterium pseudolongum, which have been shown to improve gut microbiota and immunity41,42. Additionally, recent studies have reported that 3-succinylated cholic acid, a BAs produced by intestinal bacteria, can promote the abundance of Akkermansia muciniphila and improve metabolic dysfunction-associated fatty liver disease43. Interestingly, we also observed that CDCA and DCA can promote the growth of Akkermansia muciniphila in vivo, which has been linked to improved glucose and lipid metabolism44,45,46, which shows that BAs may have the same effect on increasing the abundance of Akkermansia muciniphila bacteria. Overall, the host’s metabolism may be affected by different BAs through their distinct impacts on the gut microbiota.

BAs have similar structures but can exhibit different effects on intestinal bacteria. DCA, CDCA, and UDCA are isomers in unconjugated BAs, with the only difference between CDCA and UDCA being the isomerization of hydroxyl groups at position 7. However, DCA was found to be significantly stronger than CDCA, and CDCA was significantly stronger than UDCA in inhibiting the growth of 65 strains of bacteria. Similar trends were observed in the effects of conjugated BAs. Clinically, CDCA and UDCA have been employed as therapeutic agents for biliary tract and liver diseases. Notably, a prevalent side effect associated with CDCA use is diarrhea, occurring at an incidence rate ranging from 30% to 50%47,48. In contrast, UDCA also presents diarrhea as a common side effect, but with a markedly lower incidence rate of only 2%48. This manifestation of diarrhea is likely attributed to the distinct effects exerted by the two BAs on the composition and dynamics of the intestinal microbiota. The strikingly different effects of structurally similar BAs on gut bacteria warrant further investigation.

Bacteria can resist BAs by using efflux pumps to remove the BAs from the cells, which reduces the toxicity of BAs to the bacteria49,50,51. Different bacterial strains have different types and quantities of BAs transporters, leading to varying levels of BAs accumulation in each strain. Ability of different bacterial strains to accumulate BAs is a key factor in their sensitivity to BAs. By inhibiting the efflux pump of intestinal bacteria with PAβN, even strains that were not originally inhibited by BAs become significantly inhibited by BAs. This provides a new approach for developing BAs as antibiotics, particularly for inhibiting harmful bacteria. For instance, DCA can inhibit Clostridium difficile, while the combination of DCA and PAβN can significantly reduce the inhibitory concentration of DCA on Clostridium difficile52. This provides new therapeutic options for treating Clostridium difficile infection.

BAs have been shown to inhibit Intestinal bacteria growth through mechanisms such as DNA damage, ribosome function inhibition, and cell membrane damage53,54. RNA-seq results showed that DCA had the greatest impact on ribosome and amino acid metabolism, which was consistent with data from untargeted metabolomics of BAs against intestinal bacteria in vitro55. Studies have revealed that in the transcriptome of Bifidobacterium longum, the combined impact of DCA and CA on bacterial ribosomes is less pronounced compared to that observed in Bacteroides ovatus56. Nonetheless, these BAs still induce significant changes in the 50S ribosomal protein. Furthermore, in Bifidobacterium longum, DCA and CA increase the transcription of choloylglycine hydrolase, whereas DCA reduces its transcription in Bacteroides ovatus. These findings underscore the strain-specific and variable effects of BAs on bacterial transcriptomes. While BAs do not inhibit the growth of some gut bacteria, they can still affect the metabolism of intestinal bacteria, highlighting the importance of considering the broader impacts of BAs on gut microbial physiology and metabolic activity, even in the absence of marked changes in community structure or colonization level. BAs have been shown to impact the metabolism of ribosomes and amino acids in bacteria. However, the mechanism by which they act on ribosomes remains unclear and warrants further investigation. Moreover, BAs may affect other metabolic pathways, such as antibiotic resistance genes, bacterial energy metabolism, and protein and nucleic acid metabolism. Therefore, studying whether BAs alter the composition of intestinal microbiota by modulating these metabolic pathways deserves further attention.

Our study highlights the significance of considering the unintended impacts of BAs on our associated microbial communities and the need for a broader pharmacological view. Our findings demonstrate the usefulness of integrated in vitro and in vivo studies to elucidate the causality and mechanisms of complex interactions between BAs and microbiome. Investigating the effects and mechanisms of BAs on intestinal microorganisms could lead to the discovery of the important role of BAs in pathophysiology.

Methods

Bacterial culture conditions

The bacteria were cultured under anaerobic conditions at a constant temperature of 37 °C in a chamber (Coy Laboratory Products) (2% H2, 12% CO2, rest N2). To initiate the experiment, each bacterial strain was inoculated onto mGAM or BHI agar plates and subsequently cultured in the corresponding mGAM or BHI medium. When the bacteria reached the plateau stage, they were subcultured (Supplementary Table1).

Effects of BAs on gut bacterial strains

Accurately weighing BAs of varying masses into 1.5 mL EP tubes, DMSO was added for dissolution and then diluted with mGAM or BHI medium. 50 μL of the resulting solution was transferred to pre-reduced 384-well plates that were placed in the anaerobic chamber overnight.

Frozen bacterial solution was then applied to the plate to grow colonies, and monoclonal strains were selected and cultured in liquid medium. After the growth of the strain reaches the plateau phase, the strain is passaged. After one day, the bacteria were subcultured and introduced into the pre-reduced medium at an inoculation proportion of 6%. The subcultured bacterial solution was added to the pre-reduced 384-well plate with a final drug concentration diluted by 200 times, and each well contained 0.5% DMSO. After inoculation, plates were immersed in PBS to prevent edge effects and then incubated at 37 °C. OD600 measurements using a Microplate Reader were employed to derive growth curves, with quantification of the area under the curve (AUC) until unperturbed growth controls transitioned to the stationary phase. Both manual and dynamic detection methods were employed, with measurements taken every 1–3 hours for a period of 10–30 hours57. Normalization of the growth curves involved identifying the endpoint of the control group, which was reached 1–2 time points after the plateau phase. This endpoint was used for calculating the AUC and comparing the growth curves between the control and BA-treated groups. BAs that reduced the AUC of strains by more than 25% compared to controls were classified as hits with anti-commensal activity. If any abnormal values were detected during the measurement process (such as unusually high initial OD values, sudden OD increases, or OD values outside the normal range), the growth curve was discarded58. Each strain was screened at least three biological replicates.

Incubation of BAs with mixed gut bacteria

Human stool-derived material was approved by the medical ethics committee of Xiangya Hospital, Central South University (Approval Number: 2019040116), ensuring adherence to ethical standards. Informed consent was obtained from all donors who were 7 healthy male volunteers meeting specific inclusion criteria. These criteria included being aged 18 and above, having a body mass index between 18 and 25 kg/m2, and not experiencing malnutrition, metabolic disorders, or gastrointestinal diseases. Additionally, the volunteers had not taken antibiotics or probiotics in the three months leading up to or during the sample collection period. To preserve the integrity of the samples, fresh fecal samples were promptly transferred to a temperature of −80 °C for the preparation of the gut microbiota reserve solution. These procedures were carried out in compliance with the rigorous guidelines set by the medical ethics committee of Xiangya Hospital, Central South University, ensuring scientific integrity and ethical considerations in line with the standards required for publication in scientific journals.

To prepare the samples, the surface of the feces was carefully scraped, and 1 gram of fecal matter from the central portion was weighed and suspended in a 50 mL sterile centrifuge tube containing 15 mL of pre-reduced phosphate-buffered saline (PBS) solution with 0.1% cysteine. The suspension was thoroughly mixed by vortexing to ensure proper homogenization. Next, the suspension was promptly transferred to an anaerobic box with a controlled environment of 37 °C carbon dioxide. The tube cap was unscrewed, allowing the suspension to stand for 5 minutes to facilitate the precipitation of insoluble particles. The supernatant was then mixed with an appropriate quantity of 40% glycerol solution, previously combined with PBS. The suspension was dispensed into sterile cryovials in 1 mL aliquots and underwent a gradient-freezing process at −80 °C, with a maximum storage duration of 7 months.

The frozen bacterial solution was thawed in a 37 °C water bath and transferred to a 10 mL mGAM medium, which had been subjected to high-temperature sterilization and pre-reduction. The solution was then evenly blown using a sterile straw and cultured under anaerobic conditions for 24 hours. Bacteria are introduced into a new mGAM medium to obtain a bacterial solution59.

Incubation of BAs with fecal bacteria

To incubate DCA, LCA, and CDCA with gut bacterial solution, we first prepared the three BAs to a concentration of 100 mM using DMSO. Using a pipette gun, we transferred 10 μL of DMSO, LCA, CDCA, and DCA, respectively, into a 2 mL EP tube, followed by the addition of 990 μL of mGAM medium. After vortexing and thorough mixing, take 100 µL of the prepared BAs solution and add it to 900 µL of cultured fecal bacteria solution for incubation. After incubation for 24 hours, the samples were tested for 16S rRNA.

Effects of BAs on gut bacteria in rats

Male Sprague-Dawley rats aged 7 weeks and weighing ~200 g (Changsha, China.) were purchased and housed in a specific pathogen-free environment at the Experimental Animal Department of Central South University, with free access to water. The ambient temperature was maintained at 22 ± 2 °C and the relative humidity was 60 ± 5%. The rats were subjected to a 12-hour light/dark cycle. Experimental Animal Department of Central South University approved under protocol CSU-2022-0087.

After a one-week acclimation period, the rats were randomly divided into control groups according to their weight, using a digital scale. The control groups were given an intragastric solvent (sodium CMC), while the experimental groups were given CDCA (150 mg/kg), DCA (150 mg/kg), and LCA (150 mg/kg) via gavage, with 6–7 rats in each group for a duration of 7 days. Fecal samples were collected before and after gavage.

16S-seq of gut microbiota

Genomic DNA was extracted from samples using the CTAB/SDS method and its concentration and purity were assessed on 1% agarose gels. DNA was diluted to a concentration of 1 ng/μL using sterile water. Specific primers with barcodes were used to amplify distinct regions of the 16S rRNA gene. PCR reactions were performed with Phusion® High-Fidelity PCR Master Mix (New England Biolabs). The PCR products were analyzed by electrophoresis on a 2% agarose gel with 1× loading buffer (containing SYB green) and samples with a bright main strip between 400–450 bp (16S) were selected for further experiments. The purified PCR products were subjected to sequencing libraries using the TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA) following the manufacturer’s instructions and index codes were added. The quality of the library was evaluated using the Qubit® 2.0 Fluorometer (Thermo Scientific) and Agilent Bioanalyzer 2100 system. Finally, the library was sequenced on an Illumina NovaSeq 6000 platform to generate 250 bp paired-end reads.

BAs accumulation detection

Incubation of BAs with bacteria

Add four types of bile acids (CA, CDCA, DCA, and UDCA) to the bacterial solution during the logarithmic growth phase. The final concentration of each bile acid is 10 μg/mL, with a total volume of 1.2 mL. For the co-accumulation of LCA and DCA, prepare a stock solution with a concentration of 2.4 mg/mL. Then, use this stock solution to prepare a bacterial solution with a final volume of 1.2 mL and a concentration of 20 μg/mL. The mixture was incubated for 24 h, and then 200 μL of the solution was used to measure the absorbance at OD600. The remaining bacterial solution was centrifuged at 14,000 rpm for 15 min at 4 °C, and the supernatant was discarded. The bacterial pellet was resuspended in PBS and centrifuged again under the same conditions. The supernatant was removed, and the pellet was washed again with PBS before being centrifuged to obtain the final pellet.

Sample pretreatment

To extract the BAs, 100 μL of water was added to the pellet, followed by the addition of 300 μL of methanol precipitation agent containing an internal standard. The mixture was sonicated for 15 min and vortexed for 5 min before being centrifuged at 14,000 rpm for 15 min at 4 °C. Then, 100 μL of the supernatant was transferred to a sample injection vial for UPLC-MS analysis. BAs analysis was conducted using an AB SCIEX UHPLC system coupled with a Triple Quad 6500+ mass spectrometer, both from USA. The flow rate of the mobile phase was 0.3 mL/min. Separation of BAs was performed using a Waters T3 C18 column (1.7 μm, 100 mm × 2.1 mm). The mobile phase A was water (containing 0.1% formic acid and 5 mM ammonium acetate) and phase B was acetonitrile. The gradient program consisted of intervals: 30% to 50% B (first minute), 50% to 70% B (3 minutes), 70% to 95% B (0.3 minutes), holding at 95% B (1 minute), re-equilibration at 30% B (0.1 minutes), and maintaining at 30% B (1.6 minutes). Column temperature was kept at 45 °C. Mass spectrometer capillary voltage was −4.5 kV. Acquisition data were assessed with Analyst Software for peak integration, calibration equations, and individual BA quantification.

Detection of bile-acid profile

Sample pretreatment

Precisely weigh ~100 mg of cecal contents and add 1 mL of solution (methanol: water 1:1), followed by thorough grinding. After centrifugation, extract 50 μL of the supernatant and combine it with 150 μL of methanol solution containing the internal standard. Vortex the mixture thoroughly and centrifuge it at 14,000 rpm, 4 °C, for 10 minutes. Finally, collect 80 μL of the supernatant and inject it into the sample vial for analysis.

Detection conditions

The elution program employed a gradient with a flow rate of 0.35 mL/min as follows: starting with 30% B from 0 to 9 minutes, transitioning to 35% B from 9 to 11 minutes, increasing to 44% B from 11 to 16 minutes, ramping up to 100% B from 16 to 21 minutes, maintaining 100% B from 21 to 23 minutes, decreasing to 30% B from 23 to 23.1 minutes, and finally holding at 30% B from 23.1 to 25 minutes. All other conditions remain consistent with the prior detection of bile acids.

Detection of amino acids

Sample pretreatment

To obtain a bacterial culture solution extract, a ten-fold dilution was performed on the bacterial solution. Next, 20 μL of the diluted solution was combined with 60 μL of a methanol solution containing an internal standard. The mixture was vortexed for 5 minutes, followed by centrifugation at 12,000 rpm for 10 minutes at 4 °C. 10 μL of the supernatant solution was transferred into a new EP tube, and sodium borate solution and 6-aminoquinolinyl-N-hydroxysuccinyl carbamate were added to each centrifuge tube. The solutions were mixed with a pipette gun 5–10 times and then incubated in a 55 °Cwater bath for 10 minutes. The supernatant of each sample was collected, and diluted twice with pure water, and then vortexed for 5 minutes. The diluted samples were centrifuged at 12,000 rpm for 5 minutes at 4 °C. Finally, 80 μL of the supernatant was transferred for injection analysis and detection59. An ultra-high-performance liquid chromatography (AB SCIEX UHPLC, USA) coupled with a triple-quadrupole mass spectrometer (Triple Quad 5600, USA) was utilized for amino acid analysis. Amino acids were separated using a Waters T3 C18 column (1.7 μm, 100 mm × 2.1 mm). Phase A is water and Phase B is acetonitrile, both contain 0.1% formic acid. The gradient program was structured as follows: 2% to 22% B (initial 4 min), 22% to 95% B (0.1 min), 95% to 100% B (0.3 min), 100% to 2% B (0.2 min), and maintained at 2% B (1.4 min). Column temperature was held at 60 °C. Mass spectrometer capillary voltage was set at +5.5 kV. Analyst Software was employed for acquisition data analysis, encompassing peak integration, calibration equations, and individual amino acid quantification.

RNA extraction and RT‑qPCR

We investigated the transcriptomic responses of bacteria exposed to DCA and CDCA using the reverse-transcriptase quantitative PCR (RT-qPCR) technique. The treated group consisted of cell suspensions treated with 0.2 mM of DCA and CDCA, while the control group was treated with DMSO of the same volume. After 4 hours of treatment, total RNA was extracted following the Direct-zol RNA Miniprep Plus Kit (Zymo Research) protocol. One microgram of RNA from each sample was used for cDNA cloning by the PrimeScriptTM RT reagent Kit (Takara). To verify the effect of BAs on ribosomal transcription, we tested the expression of ribosomal L2, L3, L4, L22, and S3 genes of Bacteroides ovatus and Escherichia coli K-12. The list of primers used for qRT-PCR is presented in the Supplementary Table6. Quantification of transcripts was performed by real-time PCR (QuantStudio 3) using the 2x SYBR Green qPCR Master Mix (biomake). The relative fold changes of gene expression were calculated using the cycle threshold (Ct) method.

RNA sequencing

Bacteroides ovatus and Escherichia coli K-12 underwent 6-hour treatment with 0.5 mM DCA or DMSO. Total RNA extraction utilized Trizol Reagent (Invitrogen Life Technologies), followed by NanoDrop spectrophotometer (Thermo Scientific) and Bioanalyzer 2100 system (Agilent) assessment for quality and integrity. Zymo-Seq RiboFree Total RNA Library Kit was employed to deplete ribosomal RNA from total RNA. cDNA synthesis involved random oligonucleotides and SuperScript III, followed by second-strand cDNA synthesis using DNA Polymerase I and RNase H. The resulting blunt-ended DNA fragments were hybridized with Illumina PE adapter oligonucleotides. Library fragments were size-selected (400–500 bp) using the AMPure XP system (Beckman Coulter, Beverly, CA, USA). DNA fragments with adaptors on both ends were enriched using Illumina PCR Primer Cocktail in a 15-cycle PCR reaction. Amplified products were purified with the AMPure XP system and quantified using the Agilent high-sensitivity DNA assay on a Bioanalyzer 2100 system (Agilent). Subsequently, the sequencing library was sequenced on the NovaSeq 6000 platform (Illumina).

Statistical analysis

Quantitative data were expressed as mean ± standard error of the mean. The difference between the two groups was compared using a t test, and among multiple groups, an analysis of variance was performed. Statistical significance was determined at a two-tailed P value of less than 0.05. All statistical analyses and graphing were conducted using GraphPad Prism 8.0 and Origin 9.1.