Mechanism of Growth Inhibition by Free Bile Acids in Lactobacilli and Bifidobacteria

Abstract

The effects of the free bile acids (FBAs) cholic acid (CA), deoxycholic acid (DCA), and chenodeoxycholic acid on the bioenergetics and growth of lactobacilli and bifidobacteria were investigated. It was found that these FBAs reduced the internal pH levels of these bacteria with rapid and stepwise kinetics and, at certain concentrations, dissipated ΔpH. The bile acid concentrations that dissipated ΔpH corresponded with the MICs for the selected bacteria. Unlike acetate, propionate, and butyrate, FBAs dissipated the transmembrane electrical potential (ΔΨ). In Bifidobacterium breve JCM 1192, the synthetic proton conductor pentachlorophenol (PCP) dissipated ΔpH with a slow and continuous kinetics at a much lower concentration than FBAs did, suggesting the difference in mode of action between FBAs and true proton conductors. Membrane damage assessed by the fluorescence method and a viability decrease were also observed upon exposure to CA or DCA at the MIC but not to PCP or a short-chain fatty acid mixture. Loss of potassium ion was observed at CA concentrations more than 2 mM (0.4× MIC), while leakage of other cellular components increased at CA concentrations more than 4 mM (0.8× MIC). Additionally, in experiments with membrane phospholipid vesicles extracted from Lactobacillus salivarius subsp. salicinius JCM 1044, CA and DCA at the MIC collapsed the ΔpH with concomitant leakage of intravesicular fluorescent pH probe, while they did not show proton conductance at a lower concentration range (e.g., 0.2× MIC). Taking these observations together, we conclude that FBAs at the MIC disturb membrane integrity and that this effect can lead to leakage of proton (membrane ΔpH and ΔΨ dissipation), potassium ion, and other cellular components and eventually cell death.

Bile, which is one of the most important host-derived factors to affect the indigenous intestinal microbiota, contains conjugated bile acids, i.e., free bile acids (FBAs) that are esterified by taurine or glycine. In the intestine, conjugated bile acids facilitate lipid digestion and absorption by emulsifying lipids. However, several indigenous intestinal microorganisms can enzymatically liberate the FBAs from conjugated bile acids (1). These FBAs are known to inhibit the growth of various intestinal microbes (3). Although this phenomenon was reported quite a long time ago, the mechanism of growth inhibition by FBAs and the details of the effects of these compounds on bacterial physiology have not been clarified. Moreover, knowledge of the mechanism underlying growth inhibition by FBAs is crucial in light of the development of probiotics that are capable of growing in the intestinal tract, in which FBAs are present at high concentrations.

Recently, we carried out transport experiments to show that cholic acid (CA), which is one of the FBAs, is accumulated in energized Lactobacillus and Bifidobacterium strains, where the transmembrane proton gradient (ΔpH, alkaline interior) is the driving force (11, 12). Since this accumulation must occur during bacterial growth in the intestine, growth inhibition may be associated with FBA accumulation. It has been proposed that the accumulation of CA, which is a hydrophobic weak acid with a pK_a value of 6.4, is based on the diffusion of the protonated (neutral) CA molecules across the cell membrane, followed by their dissociation according to the ΔpH of the energized cells (11, 12). Since this process consumes ΔpH, the internal pH must be reduced. However, acidification of the internal pH by exogenously added FBAs has not been measured previously. Based on previous findings, growth inhibition by FBAs has been attributed to either acidification itself and/or ΔpH dissipation, which would lead to a deficiency in biological energy (proton motive force). Another possibility as a mechanism behind growth inhibition by FBAs is their potential membrane-damaging effect. Therefore, in this study, we examined the bioenergetic consequences and the effect on membrane integrity of the exposure of bacterial cells to FBAs, to elucidate the mechanism of growth inhibition by FBAs.

MATERIALS AND METHODS

Bacterial strains and chemicals.

The microorganisms used in this study were obtained from the Japan Collection of Microorganisms (JCM, Wako, Japan). The strains were grown in half-strength (1/2) MRS medium (Becton Dickinson and Company, Franklin Lakes, NJ) at 37°C (except for Lactobacillus sakei JCM 1157^T, which was cultured at 30°C) under anaerobic conditions (mixed gas, N₂:CO₂:H₂; 8:1:1). In the case of the bifidobacteria, the medium was supplemented with 0.025% l-cysteine hydrochloride. The sodium salts of the FBAs and pentachlorophenol (PCP) were purchased from Sigma-Aldrich Inc. (St. Louis, MO). Fluorescent dyes 5 (and 6-)-carboxyfluorescein diacetate succinimidyl ester (cFDASE) and 3,3′-dipropylthiadicarbocyanine iodide [DiSC₃(5)] were obtained from Molecular Probes, Inc. (Eugene, OR). Pyranine was a product of Acros Organics (Geel, Belgium). Other chemicals were of the highest quality commercially available.

Growth experiments.

Bacteria were cultured in 1/2 MRS broth that contained various concentrations of FBAs, or other chemicals, under the conditions described above. Bacterial growth was monitored periodically by measuring the absorbance of the culture broth at 660 nm. The MIC for growth of each compound was determined as the lowest concentration that inhibited by 100% the growth of the test microorganism.

Intracellular pH measurement.

The internal pH values of the lactobacilli and bifidobacteria were measured according to the method described previously (11), except that the experiments were performed with an external pH of 6.5, which is the pH of the 1/2 MRS broth. This method uses the internally conjugated fluorescent pH probe 5 (and 6-)-carboxyfluorescein succinimidyl ester (cFSE). The exponential-phase cells cultured by the method described under “Bacterial strains and chemicals” were washed twice with 50 mM potassium phosphate buffer (pH 6.5) containing 1 mM MgSO₄ and 0.1 U/ml horseradish peroxidase (buffer A). The cells were resuspended in 150 mM potassium phosphate buffer (pH 6.5) containing 1 mM MgSO₄ and 1.0 U/ml horseradish peroxidase (buffer B) at an A₆₆₀ of 0.5. The bacterial cells were preloaded with 4 μM of membrane-permeable precursor probe cFDASE and incubated at appropriate temperatures (the same as the culture temperatures) for 30 min. During the incubation the precursor probe was cleaved by intracellular esterases, and the resulting cFSE molecules were conjugated to bacterial proteins. After centrifugation, cells were resuspended in the same volume of buffer B, and the elimination of the unbound cFSE probe was conducted by the addition of glucose at a final concentration of 10 mM for 1 h at appropriate temperatures. The mixtures were centrifuged again, resuspended in buffer B at an A₆₆₀ of 0.5, and dispensed into a stirred and heated (at appropriate temperatures) cuvette holder of an LS50B fluorimeter (Perkin-Elmer Life and Analytical Sciences, Inc., Boston, MA). The internal pH of the bacteria was determined by measuring fluorescence intensities of the cell suspension with excitation and emission wavelengths of 490 and 520 nm, respectively (slit widths of 2.5 nm). During the measurement, energization with glucose and additions of various kinds of chemicals were conducted as described in the corresponding figures and their legends. During premeasurement preparations (e.g., probe loading and elimination of the unbound probe molecules) and in the cuvette throughout the internal pH measurements with the fluorimeter, an anaerobic atmosphere (mixed gas) was applied to prevent exposure of the cells to oxygen. Calibration of the fluorescent signal was carried out by measuring the fluorescence intensity of the deenergized cells at pH values between 4 and 10. The deenergization was conducted by the addition of 2 μM valinomycin plus 2 μM nigericin to the cell suspension.

Monitoring of the transmembrane electrical potential (ΔΨ).

Changes in the membrane potentials of the lactobacilli and bifidobacteria upon energization and the addition of various chemical compounds were monitored using the cationic potential-sensitive fluorescent dye DiSC₃(5), as reported previously (11), except that the experiments were carried out at an external pH of 6.5. The cells cultured in the same manner as in the case of intracellular pH measurement were washed twice with 50 mM potassium phosphate buffer (pH 6.5) containing 1 mM MgSO₄ and 65 U/ml catalase (buffer C) and resuspended in 150 mM potassium phosphate buffer (pH 6.5) containing 1 mM MgSO₄ and 65 U/ml catalase (buffer D) at an A₆₆₀ of 10. Then the cells were transferred to a stirred cuvette containing buffer D and 0.5 μM DiSC₃(5) at a final A₆₆₀ of 0.05. The fluorescence intensity was monitored with an LS50B fluorimeter (Perkin-Elmer) with excitation and emission wavelengths of 651 and 657 nm, respectively (slit widths of 4.0 nm). During measurements with the fluorimeter, mixed gas was introduced into the headspace of the cuvette, to ensure anaerobic conditions for the cells.

Membrane phospholipid extraction and membrane vesicle preparation.

Cellular lipids were extracted by the modified Bligh and Dyer method (4). Approximately 4 to 5 g cell biomass (wet weight) was suspended in 5 ml deionized water and incubated with 20 mg/ml lysozyme at 37°C for about 6 h. The following steps were performed under N₂ to minimize oxidation. A methanol:chloroform (2:1, vol/vol) mixture (90 ml) was added to the cell suspension and stirred overnight at 4°C in the dark. The mixture was centrifuged (2,000 × g, 5 min, 4°C), and the supernatant was collected and then stirred with 60 ml of a chloroform:water mixture (1:1) for 3 h. The mixture was allowed to stand until the phases separated. The chloroform phase was collected and centrifuged (2,000 × g, 20 min, 4°C) to separate residual water. The chloroform-rich phase was collected again with a Pasteur pipette, and the solvent was evaporated. The phospholipids from the total lipid extract were separated by a method that was based on the procedure described by Viitanen et al. (16). Dry lipids were dissolved in 1 ml chloroform; the mixture was dripped slowly into 80 ml ice-cold acetone and stirred overnight in the dark. The mixture was centrifuged (2,000 × g, 30 min, 4°C), the supernatant was removed, and the pellet was dried under an N₂ gas stream. The pellet was dissolved in 80 ml diethyl ether and stirred for 1 to 2 h at room temperature. The mixture was centrifuged, and the supernatant was collected and evaporated. The extracted phospholipids were suspended in 1 ml of 50 mM (for experiments in Fig. 6) or 150 mM (for experiments in Fig. 5) potassium phosphate buffer (pH 6.5) that contained 1 mM pyranine. Small unilamellar vesicles were prepared by sonication for 30 min at 50 W. The vesicle suspension was centrifuged (20,000 × g, 10 min, 4°C) to remove metal particles. Untrapped pyranine was removed by gel filtration on a PD-10 column (GE Healthcare Amersham Biosciences Corp., Piscataway, NJ) and washed with 50 mM (for experiments in Fig. 6) or 150 mM (for experiments in Fig. 5) potassium phosphate buffer (pH 6.5). The internal pH changes of the pyranine-entrapped membrane vesicles were measured by pyranine fluorescence (excitation wavelength of 455 nm and emission wavelength of 509 nm with both excitation and emission slit widths of 15 nm) using an LS50B fluorimeter (Perkin-Elmer). Calibration of the fluorescence signal was carried out as described previously (9).

FIG. 6.

Effects of lower concentrations of CA (a) and DCA (b) on the internal pH of membrane vesicles of L. salivarius subsp. salicinius JCM 1044. Artificial ΔpH was established by the addition of 10 μl 2 N KOH to the membrane vesicle suspension in 2 ml 50 mM potassium phosphate buffer, which altered the external pH from 6.5 to about 6.8. The other experimental conditions were the same as those described in the legend to Fig. 5.

FIG. 5.

Effects of PCP (a), CA (b), and DCA (c) on the internal pH of membrane vesicles of L. salivarius subsp. salicinius JCM 1044. An artificial ΔpH was established by the addition of 28 μl of 2 N KOH to the membrane vesicle suspension in 2 ml of 150 mM potassium phosphate buffer, which altered the external pH from 6.5 to about 6.8. The chemical compounds were added at the indicated final concentrations. Nigericin (100 nM) was added to dissipate the preexisting ΔpH.

Determination of the viability of Bifidobacterium breve JCM 1192 upon exposure to FBAs, PCP, or SCFA mixture.

Cells were cultured until mid-exponential phase in 1/2 MRS medium containing 0.025% l-cysteine hydrochloride at 37°C under anaerobic conditions. After harvesting, the cells were washed twice with sterile buffer B and resuspended in the same buffer to an A₆₆₀ of 0.5. Portions of this cell suspension were incubated with 10 mM glucose and each of the following chemicals: various concentrations of CA or deoxycholic acid (DCA), 0.3 mM PCP, or 117 mM short-chain fatty acid (SCFA) mixture (sodium acetate, 66 mM; sodium propionate, 26 mM; sodium butyrate, 25 mM) for 1 and 3 h. Cell suspension dilutions were made with sterile 0.85% NaCl solution, and cells from appropriate dilutions were plated onto 1/2 MRS agar plates, which were incubated for 2 days in an anaerobic jar at 37°C. Colonies were counted, and the viabilities (%) were calculated based on the initial untreated cell suspension.

Determination of the membrane integrity of B. breve JCM 1192 upon exposure to FBAs, PCP, or SCFA mixture.

We assessed membrane integrity changes of JCM 1192 in the presence of FBAs, PCP, or an SCFA mixture by a fluorescent method based on membrane permeability of dead cells. This method can indicate the ratio of the cells with intact membranes in the population. A cell suspension (A₆₆₀ of 0.3) was prepared and treated under the same conditions as in the plating method. After the treatment cells were subsequently incubated with a fluorescent dye mixture (component A plus component B) of the LIVE/DEAD BacLight bacterial viability kit (Molecular Probes) according to the manufacturer's recommendation for 15 min at 37°C. This kit contains the green fluorescent DNA dye SYTO 9 for all kinds of cells and the red fluorescent DNA dye propidium iodide for cells with a compromised membrane. The cell suspensions were excited by 480-nm-wavelength light, and the emission spectra between 490 and 700 nm were measured using an LS50B fluorimeter (Perkin-Elmer) with both the excitation and emission slits set at 3.0 nm. Calibration used 100% live (no treatment) and 100% dead (treated with 100% isopropanol for 1 h) cells. The ratio of the integrated intensity of the portion of each spectrum between 500 and 530 nm (green) to that between 620 and 650 nm (red) was calculated. To obtain a calibration curve, the ratio of integrated green/red fluorescence was plotted versus the known percentage of live cells of the standard cell suspensions (10%, 50%, and 90% live cell suspensions, prepared using 100% live and 100% dead cells).

Measurement of the leakage of cellular material from B. breve JCM 1192 upon exposure to FBAs.

Cells were cultured until mid-exponential phase in 1/2 MRS medium containing 0.025% l-cysteine hydrochloride at 37°C under anaerobic conditions. After the harvest the cells were washed once with 50 mM sodium phosphate buffer (pH 6.5) containing 1 mM MgSO₄ and 1.0 U/ml horseradish peroxidase. Then the cells were resuspended in 150 mM sodium phosphate buffer (pH 6.5) containing 1 mM MgSO₄ and 1.0 U/ml horseradish peroxidase at an A₆₆₀ of 0.5. The cell suspension was incubated at 37°C in the presence of 10 mM glucose and CA or DCA at the following concentrations (mM): CA at 0.1, 0.5, 1, 2, 4, 6, 8, 15, and 20 or DCA at 0.01, 0.05, 0.1, 0.2, 0.4, 0.6, 0.8, 1, and 2. Samples at various time intervals (up to 3 h) were taken from the reaction mixtures and centrifuged twice at 20,000 × g for 5 min. Then the supernatants were used for further studies. The absorbances of the supernatants at 260 and 280 nm of the CA-treated samples were measured with a Beckman DU 640 spectrophotometer (Beckman Coulter, Inc., Fullerton, CA). For the measurement of potassium leakage portions of the supernatants were treated with 1 N HClO₄ (1:1) overnight at room temperature. Then the mixtures were diluted 2.5-fold with 0.36 N HCl and centrifuged at 1,350 × g for 10 min. The potassium amounts in the supernatants were then measured by flame atomic absorption spectroscopy (Z-5310 polarized Zeeman atomic absorption spectrophotometer; Hitachi High Technologies Corporation, Tokyo, Japan).

RESULTS

The MICs for FBAs correspond with their ΔpH-dissipating concentrations in many lactobacilli and bifidobacteria.

The MICs that were obtained for CA and DCA using selected lactobacilli and bifidobacteria are shown in Table 1. The MIC of CA ranged from 3.0 to 13.0 mM, which was about 10 times higher than the MIC of DCA (0.3 to 0.8 mM). In our previous report (11), we showed that an SCFA mixture reduced the ΔpH of B. breve JCM 1192. Therefore, we expected that FBAs, such as CA and DCA, which are also weak acids, might have similar effects on the internal pH of this microorganism. The finding that CA is accumulated in energized lactobacilli and bifidobacteria by a ΔpH-driven diffusion mechanism (11, 12) led us to expect that the FBAs would acidify the cytoplasm of these bacteria. Indeed, we observed a decrease in ΔpH following the addition of FBAs in JCM 1192, as shown in Fig. 1a and b. However, this reduction turned to total dissipation of ΔpH at certain bile acid concentrations, which was confirmed by the finding that the addition of nigericin (an ionophore known to dissipate ΔpH) had no further effect on the internal pH of JCM 1192. Surprisingly, the ΔpH-dissipating concentrations of CA and DCA closely paralleled the MICs of these FBAs in many lactobacilli and bifidobacteria (Table 1). Moreover, we observed an approximately 10-fold difference between the ΔpH-dissipating concentrations of CA and DCA, with a similar difference being observed between their MICs (Table 1). In addition, it was found that chenodeoxycholic acid (CDCA) dissipated the ΔpH of JCM 1192 at a concentration similar to that of DCA (0.4 mM; Fig. 1c), which was also similar to the MIC of CDCA (0.5 mM; data not shown in Table 1). In contrast, the SCFA mixture, which reduced the internal pH of JCM 1192 (11), had no growth-inhibitory effect on this bacterium, even when individual SCFAs alone were applied at concentrations up to 400 mM (data not shown).

TABLE 1.

CA and DCA concentrations that totally inhibit growth and dissipate the ΔpH and ΔΨ in lactobacilli and bifidobacteria^a

Strain	Origin	CA (mM)			DCA (mM)
		MIC	ΔpH diss.	ΔΨ diss.	MIC	ΔpH diss.	ΔΨ diss.
Lactobacillus acidophilus JCM 1034	Human intestine	8.5	5.5	ND	ND	ND	ND
L. salivarius subsp. salicinius JCM 1040	Human intestine	6.0	5.0	ND	0.6	0.8	ND
L. salivarius subsp. salicinius JCM 1044	Human intestine	6.0	7.0	3.0	0.7	0.8	0.3
L. salivarius subsp. salicinius JCM 1047	Swine intestine	4.5	8.0	ND	0.55	0.9	ND
L. gasseri JCM 1131T	Human intestine	7.0	6.5	3.5	ND	1.0	0.7
L. reuteri JCM 1112T	Intestine of adult	13.0	10.0	4.0	ND	0.7	0.4
L. salivarius subsp. salivarius JCM 1231T	Human saliva	5.5	6.5	ND	ND	ND	ND
L. buchneri JCM 1115T	Tomato pulp	6.5	6.0	ND	0.8	0.8	ND
L. sakei JCM 1157T	Starter of sake	4.0	9.0	3.0	0.7	0.9	0.2
Bifidobacterium breve JCM 1192T	Intestine of infant	5.0	6.0	2.0	0.5	0.6	0.3
B. breve JCM 7017	Human feces	3.0	6.0	ND	0.55	0.6	ND
B. pseudocatenulatum JCM 1200T	Feces of infant	3.0	6.0	3.0	0.45	0.6	0.3
B. gallicum JCM 8224T	Human feces	7.0	8.0	ND	0.45	0.8	ND
B. longum JCM 1217T	Intestine of adult	9.0	8.0	3.0	0.55	0.6	0.2
B. bifidum JCM 1255T	Feces of infant	5.0	4.0	ND	0.3	0.4	ND
B. adolescentis JCM 1275T	Intestine of adult	5.0	8.0	ND	0.4	0.6	ND

^aAll of the data are the means of two separate experiments. ND, not determined; diss., dissipating.

FIG. 1.

Effects of CA (a), DCA (b), and CDCA (c) on the internal pH of B. breve JCM 1192. Cells (A₆₆₀ of ∼0.5) were preloaded with cFSE and energized with 10 mM glucose in buffer B. The respective FBAs were then added to the indicated final concentrations. Nigericin (200 nM) was added to check the dissipation of the ΔpH. The data shown in this and the following figures are representative of at least three experiments that gave similar results.

FBAs but not SCFAs are able to dissipate ΔΨ.

It has been reported that acetic acid is accumulated in Streptococcus bovis cells (15), which results in a reduction of the intracellular pH. In Clostridium thermoaceticum, acetate caused some decrease in ΔpH, but this decrease was partially compensated for by an increase in ΔΨ (2). However, the effect of SCFAs on ΔΨ, especially in the enteric species of lactobacilli and bifidobacteria, has not been demonstrated previously. The cationic fluorescent probe DiSC₃(5) was used to monitor ΔΨ, as indicated by the fluorescence quenching. The results show that the addition of any of three SCFAs (sodium acetate, sodium propionate, and sodium butyrate) to energized JCM 1192 cells decreased the fluorescent signal, which indicated an increase of ΔΨ (Fig. 2). The existence of ΔΨ after the addition of SCFAs was confirmed by the addition of valinomycin, which is an ionophore that dissipates ΔΨ, resulting in a sharp increase in probe fluorescence. All three FBAs had the effect opposite from that of the SCFAs, since they dissipated the ΔΨ of JCM 1192 at certain concentrations (Fig. 3). The dissipation of ΔΨ by bile acids was verified by the fact that subsequent addition of valinomycin did not cause any further increase in the fluorescence intensity. Similar to the ΔpH dissipation results, we observed that the concentration required to dissipate ΔΨ was approximately 10-fold lower for DCA and CDCA than for CA; this tendency also held true for other strains, at least in the case of DCA (Table 1). Although the ΔpH-dissipating concentrations of the FBAs were found to be somewhat higher than the ΔΨ-dissipating concentrations (Table 1), this may be attributed to the fact that the cell densities were higher during the ΔpH measurements (A₆₆₀ = 0.5) than during the ΔΨ measurements (A₆₆₀ = 0.05).

FIG. 2.

Effects of sodium acetate (a), sodium propionate (b), and sodium butyrate (c) on the membrane potential of B. breve JCM 1192. Washed cells (A₆₆₀ of ∼0.05) were added to the cuvette, which contained the DiSC₃(5) probe (0.5 μM) in buffer D. The cells were energized with glucose (10 mM), followed by the addition of the respective SCFAs at the indicated final concentrations. Valinomycin (5 nM) was added to check the dissipation of the ΔΨ. The fluorescence intensity is expressed in arbitrary units (a.u.).

FIG. 3.

Effects of CA (a), DCA (b), and CDCA (c) on the membrane potential of B. breve JCM 1192. The experimental procedures were the same as those described in the legend to Fig. 2, except that respective FBAs were added instead of SCFAs, at the indicated final concentrations.

The kinetics of ΔpH dissipation by a synthetic proton conductor, PCP, appear to be different from that by FBAs in Lactobacillus salivarius subsp. salicinius JCM 1044 and B. breve JCM 1192.

Compounds that can simultaneously dissipate ΔpH and ΔΨ are called proton conductors. We compared the effects of FBAs and a known proton conductor, PCP, on the bioenergetics of L. salivarius subsp. salicinius JCM 1044. As expected, PCP dissipated the ΔΨ and ΔpH of JCM 1044 but at lower concentrations than those of FBAs (Fig. 4). However, the kinetics of ΔpH dissipation was not stepwise as in the case of FBAs (Fig. 1) but rather followed a continuously decreasing curve (Fig. 4b). In addition, similar ΔΨ and ΔpH-dissipation kinetics were observed when PCP was added to energized JCM 1192 (data not shown). The observed relatively slower but gradual ΔpH dissipation kinetics is probably reflecting the PCP and PCP anion's recycling transmembrane movement, which ultimately leads to the dissipation of the transmembrane proton concentration gradient. Therefore, these data indicate that the apparent proton conductive mode of action of FBAs is possibly different from that of true proton conductors.

FIG. 4.

Effects of PCP on the membrane potential (a) and internal pH (b) of L. salivarius subsp. salicinius JCM 1044. (a) The experimental procedures were the same as those described in the legend to Fig. 2, except that PCP was added instead of SCFAs at the indicated final concentrations. (b) The experimental procedures were the same as those described in the legend to Fig. 1, except that PCP was added instead of FBAs at the indicated final concentrations.

Exposure to CA or DCA increases membrane damage and decreases viability in B. breve JCM 1192.

We investigated the membrane integrity and viability of JCM 1192 upon exposure to FBAs by a fluorescent method and the conventional plate-out method, respectively, with PCP and SCFAs as references. As shown in Table 2, CA and DCA showed decreasing membrane integrity and decreasing viability with increasing concentrations of CA or DCA. In contrast treatment with the proton conductor PCP or SCFAs caused no serious membrane damage and viability loss even over 3 h except that the viability dropped remarkably after 3 h in the case of PCP treatment (Table 2). This indicates the importance of proton motive force (cellular energy) for the viability of the bacterial cell. PCP, while possibly not affecting the membrane integrity, causes a loss of viability presumably by its proton conductance, leading to a serious cellular energy shortage (Fig. 4) after 3 h. It should be noted that in the case of CA and DCA at concentrations more than 4 mM and 0.4 mM, respectively, we detected much less viability by the plate-out method than we detected membrane integrity reduction by the fluorescent method (Table 2).

TABLE 2.

Monitoring of membrane integrity and viability of B. breve JCM 1192 upon exposure to bile acids, PCP, and SCFA mixture^a

Treatment and concn (mM)	Membrane integrity by fluorescent method (%)		Viability by plate-out method (%)
	1 h	3 h	1 h	3 h
CA
0.1	87.90 ± 5.29	93.12 ± 13.51	112.5 ± 12.02	106.4 ± 16.33
1	92.68 ± 5.73	90.71 ± 1.22	49.85 ± 8.84	14.50 ± 1.41
2	88.53 ± 12.41	89.04 ± 1.36	14.67 ± 2.83	3.96 ± 2.06
4	47.56 ± 6.07	62.00 ± 13.66	0.86 ± 0.06	0.14 ± 0.04
6	14.19 ± 3.54	18.73 ± 0.19	0.16 ± 0.11	0.03 ± 0.01
8	13.63 ± 0.52	18.31 ± 3.42	0.03 ± 0.02	0.006 ± 0.003
DCA
0.01	75.10 ± 5.87	68.51 ± 3.80	111.0 ± 17.0	101.6 ± 5.23
0.1	67.15 ± 18.35	58.97 ± 2.30	82.39 ± 14.83	67.65 ± 3.61
0.2	67.60 ± 18.34	60.24 ± 8.89	42.42 ± 4.41	27.46 ± 14.35
0.4	27.08 ± 4.06	32.09 ± 5.52	4.06 ± 2.50	0.19 ± 0.09
0.6	6.11 ± 2.67	6.44 ± 1.81	0.10 ± 0.04	0.05 ± 0.02
0.8	4.47 ± 0.19	7.77 ± 0.17	0.007 ± 0.003	0.002 ± 0.001
PCP
0.3	72.10 ± 17.35	92.20 ± 2.19	88.86 ± 27.29	0.83 ± 0.75
SCFA
117	91.73 ± 2.34	84.17 ± 3.50	33.18 ± 16.33	39.39 ± 3.89

^aResults are shown as means ± standard deviations where n ≥ 2.

The different effects of PCP and FBAs in isolated membrane lipid vesicles indicate different modes of action.

Proton conductors share the characteristic that they can cross membrane bilayers in both the protonated (neutral) and anionic forms. However, previous studies indicated that bile acid anions traverse phosphatidylcholine membranes very slowly (5). Therefore, we extracted the membrane phospholipids from JCM 1044, to test the proton conductive properties of FBAs on these membrane vesicles. In these experiments, we added KOH to the vesicle suspension, creating a ΔpH of approximately 0.3 pH unit across the vesicle membrane. The addition of KOH instantly increased the external pH of the vesicle suspension from about 6.5 to 6.8. The external addition of KOH resulted in a fast, sharp increase in intravesicular pH, followed by a slower pH increase, which indicated slow leakage of H⁺/OH⁻ across the membrane bilayer (Fig. 5). When PCP was added after the KOH, it caused a small decease in intravesicular pH, followed by a decrease in the preexisting ΔpH to near zero within a few minutes (Fig. 5a). The internal acidification immediately after the addition of PCP may be attributed to transbilayer movement of the neutral (protonated) form, followed by dissociation inside the vesicles, which results in an increase in the intravesicular proton concentration. Similar events were observed by adding CA at 3.5 mM (Fig. 5b) or DCA at 0.4 mM (Fig. 5c) to the vesicles, which also led to a substantial decrease of ΔpH (increasing internal pH). These results suggest that FBAs can increase the H⁺/OH⁻ leakage across the membrane bilayer, although at a lower rate than PCP. However, when ΔpH-dissipating concentrations (for JCM 1044) of CA (7 mM) or DCA (0.8 mM) were added (Fig. 5b and c), preexisting ΔpH instantly disappeared, similar to the effect of nigericin (Fig. 5a). This latter sudden dissipation of ΔpH was not observed in the case of PCP, suggesting a mechanism of action of CA and DCA different from that of proton conductors. In contrast, when lower concentrations of CA (1 to 2 mM) or DCA (0.1 to 0.2 mM) were added to membrane vesicles the H⁺/OH⁻ leakage did not visibly increase, as shown in Fig. 6. Furthermore, our experiments with membrane vesicles from JCM 1192 produced similar results (data not shown). Additionally, when membrane vesicles of JCM 1192 were ultracentrifuged (135,000 × g, 30 min) after 7.0 mM CA or 0.8 mM DCA treatment, almost all of the total pyranine fluorescence was found in the supernatant, indicating severe membrane damage, while about 50% of the total pyranine fluorescence was detected in the supernatant of untreated vesicles. These data imply that after the addition of 7.0 mM CA or 0.8 mM DCA parallel with H⁺/OH⁻ leakage pyranine leakage possibly also occurred (Fig. 5b and c). Therefore, our results suggest that FBAs at certain concentrations ([CA] > 2 mM, [DCA] > 0.2 mM) disturb membrane integrity in Lactobacillus and Bifidobacterium cells, which leads to the collapse of the transmembrane proton gradient.

Exposure to FBAs leads to leakage of cytoplasmic potassium in a concentration-dependent manner and can result in leakage of cytoplasmic proteins when CA concentration increases further.

To test the increase of the permeability of ions other than proton caused by membrane damage, we measured the amount of potassium released into NaPO₄ buffer upon CA exposure from JCM 1192. The results showed that 15 min (Fig. 7a) of exposure to 4 to 6 mM CA caused the leakage of almost all cytoplasmic potassium. Similar results were also obtained at 5 min and 3 h of exposure (data not shown). A similar pattern of potassium leakage was obtained with DCA, where 0.4 to 0.6 mM DCA resulted in the maximum potassium permeability of the cell membrane (data not shown). These results imply that all ions are possibly instantly permeable to the membrane in the presence of 4 to 6 mM CA (or 0.4 to 0.6 mM DCA), and they are in good agreement with MIC and ΔpH-dissipating concentrations of CA (or DCA) in JCM 1192. On the other hand, this CA concentration range did not lead to near-maximum UV absorbance values (at 260 and 280 nm) by macromolecules, which were observed at exposures to 15 to 20 mM CA (Fig. 7b). We measured near-zero UV absorbance values up to 2 mM CA, and the absorbances started to increase substantially at higher concentrations. In contrast to potassium ion leakage, UV absorbance values at every CA concentration were increasing as a function of time (data not shown), resulting in the highest absorbance values at the longest exposure time, 3 h. Additionally, we observed detectable protein bands on the sodium dodecyl sulfate-polyacrylamide gel of cell-free supernatant at CA concentrations of 15 and 20 mM after 3 h of exposure (data not shown).

FIG. 7.

Measurement of the leakage of K⁺ (a) and other cellular materials (b) from B. breve JCM 1192 upon exposure to CA for 15 min (a) and 3 h (b). Cells (A₆₆₀ of ∼0.6) were incubated with various concentrations of CA in 150 mM sodium phosphate buffer (pH 6.5) containing 1 mM MgSO₄, 1.0 U/ml horseradish peroxidase, and 10 mM glucose. The K⁺ concentration (a) and the UV absorbance (b) at 260 nm (•) and 280 nm (○) of the cell-free supernatant were determined spectroscopically.

DISCUSSION

In recent years, lactobacilli and bifidobacteria have been studied intensively and used frequently as probiotics. When used as probiotics and as members of the human indigenous intestinal microbiota, these lactic acid bacteria have to cope with the challenge of FBAs. In our previous reports, we noted that energized bifidobacteria accumulated CA, which is one of the most abundant FBAs in humans, when the extracellular concentration of CA was within the physiological range of up to 2 mM (11). The mechanism of accumulation of the weak acid CA was found to be similar to the accumulation of other weak acids, e.g., acetate, in some bacteria (15). The accumulation of weak acids in bacteria generally involves transbilayer diffusion of the neutral, protonated weak acid, followed by its intracellular dissociation, thereby lowering the internal pH. However, internal pH drop as a consequence of weak acid accumulation, as shown in previous studies (2, 6), has not been demonstrated previously with CA in lactobacilli or bifidobacteria. The results presented here (Fig. 1) demonstrate a decrease in the internal pH in response to the addition of FBAs; they also show total dissipation of the ΔpH. Moreover, the concentrations of FBAs required for ΔpH dissipation correspond with the minimum concentrations of FBAs that abrogate the growth in vitro of lactobacilli and bifidobacteria (MIC, Table 1). In addition, MIC and ΔpH-dissipating concentrations determined in this study are (with one exception) lower than the critical micellar concentrations of either CA (11 mM), DCA (3 mM), or CDCA (4 mM) (14). Our results also showed that, unlike acetate, propionate, and butyrate (Fig. 2), FBAs can also dissipate ΔΨ (Fig. 3) of intact cells. To our knowledge, this is the first time that growth inhibition by FBAs has been linked to a bioenergetic factor, suggesting that dissipation of proton motive force is involved in this mechanism.

Our data also showed a significant (about 10-fold) difference between the MICs and ΔpH-dissipating concentrations of CA and DCA, which cannot be attributed simply to the subtle difference in their pK_a values (6.4 and 6.58 for CA and DCA, respectively). It has been reported that transbilayer movement (flip-flop) in small unilamellar vesicles is at least 10 times faster for the more hydrophobic (8) DCA and CDCA compounds (both have two OH functional groups) than for CA (three OH groups) (10). This implies that 10 times more DCA and CDCA molecules than CA molecules can cross the membrane in a certain time. Thus, if CA is present at a concentration that is 10 times higher than that of DCA or CDCA, the same number of molecules of all three species will permeate the membrane within the same time period. This may explain the almost 10-fold differences between CA and DCA in terms of MIC and ΔpH-dissipating concentrations.

The mechanism of the proton conductance of FBAs (Fig. 1 and 3; Table 1) was suggested to be different from that of PCP, a synthetic proton conductor, by the observed different kinetics of ΔpH dissipation. The kinetics of ΔpH dissipation in energized cells of B. breve JCM 1192 by FBAs was stepwise (Fig. 1), which is in contrast to PCP, with kinetics of continuous, steady decrease (Fig. 4) probably resulting from its recycling mode of action. In experiments using extracted membrane phospholipid vesicles (Fig. 5), we found that at certain concentrations (3.5 mM CA or 0.4 mM DCA) FBAs behave similarly to the PCP, but at higher concentrations (7.0 mM CA or 0.8 mM DCA) FBAs instantly dissipated ΔpH (Fig. 5b and c), while ΔpH dissipation by PCP was continuous and not immediate (Fig. 5a). These observations did not contradict the experiments using small unilamellar phosphatidylcholine vesicles, in which FBAs did not show proton conductor-like behaviors (10). In those experiments, where the eukaryotic membrane constituent phosphatidylcholine was used, bile acids were applied only up to concentrations of 0.2 mM, at which we observed no proton conductor-like behavior by CA or DCA (Fig. 6). Nuclear magnetic resonance data have also indicated that the transmembrane movement of bile acid anions is very slow in unilamellar vesicles (5).

The gradual decrease of membrane integrity as demonstrated by fluorescent dye staining with increasing concentrations of FBAs but not with PCP or an SCFA mixture (Table 2) may suggest that the proton conductor-like action of FBAs is associated with membrane damage. This notion was strengthened by the results of vesicle experiments in which instant dissipation of ΔpH was observed by challenge with 7.0 mM CA or 0.8 mM DCA (Fig. 5b and c). Under these conditions FBAs damage the membrane to the extent that it becomes freely permeable to H⁺/OH⁻ and possibly even to pyranine. Detection of nearly maximum leakage of potassium ion from intact cells of JCM 1192 upon exposure to 4 to 6 mM CA (Fig. 7a) or 0.4 to 0.6 mM DCA also indicated that the increased ion permeability is not specific to protons but belongs to ions in general. A recent finding that CA at a concentration near our MICs increased the uptake of the aminoglycoside gentamicin by 18-fold in Lactobacillus plantarum WCFS-1 (7) supports the notion that, beside ions, at least low-molecular-weight metabolites (those having molecular weights similar to those of gentamicin and pyranine, ca. 400 to 500) might also cross the membrane when CA is present in MIC range. The increase of UV absorbance (Fig. 7b) was more pronounced at CA concentrations of 15 to 20 mM, where membrane damage seems severe enough to let the cell proteins leak out. Although the precise mechanism that results in membrane permeability is not clear, we can assume that free bile acids integrate into the phospholipid bilayer to disorder membrane functions leading to the permeability increase. In the equilibration, DCA can be integrated into the membrane much more effectively than CA since its partition coefficient (weight percent of DCA bound to phosphatidylcholine vesicles/weight percent of DCA in water) was found to be 3.5 times higher than that of CA (10).

An important finding with the membrane integrity and viability check was that viabilities obtained with the plate-out method gave much lower values than membrane damage assessed by the fluorescent method at CA and DCA concentrations more than 4 mM and 0.4 mM, respectively (Table 2). This discrepancy may suggest that increased ion permeability due to the decreased membrane integrity below threshold value may determine the viability of the cell. Under such conditions (Table 2, CA at 4 mM), while membrane integrity is not very low (47 to 62%), the viability is lost dramatically (0.14 to 0.86%) due to the dissipation of proton motive force (Fig. 1a and 3a) and loss of homeostasis of potassium ion distribution across the cell membrane (Fig. 7a). Taking these observations together, the mechanism of FBAs to inhibit bacterial growth is concluded to be their membrane-damaging effect leading to the dissipation of concentration gradient of proton, potassium, and probably other ions across the membrane and also the leakage of low-molecular-weight metabolites from the cell. As the consequence, the bacterial cell may lose not only proton motive force (energy) but also important ions and solutes from the cytosol, thereby leading to the growth inhibition and viability loss as the incubation time increases. Based on this hypothesis, the observed increase of UV absorbance (Fig. 7b) probably due to the leakage of cellular proteins is not directly involved in the mechanism of growth inhibition. In sharp contrast, our data clearly demonstrated that PCP and SCFAs do not affect membrane integrity (Fig. 2; Table 2) and hence viabilities. This would also explain the observation of the absence of growth inhibition by SCFAs.

Our previous findings (11, 12) that energized Lactobacillus and Bifidobacterium strains accumulate CA intracellularly are not in conflict with the membrane integrity disturbance mechanism outlined above. The extracellular concentration of CA in our earlier transport experiments was 0.1 mM (up to 2 mM), which is substantially lower than that for complete dissipation of ΔpH (Fig. 1a) with high membrane integrity (Table 2). Under these concentrations, the CA molecule generally behaves as a weak acid, being distributed across the membrane according to the membrane ΔpH. Based on the above evidence, we can propose a concentration-dependent mode of action of CA for cells of lactobacilli and bifidobacteria: CA is accumulated in the energized cells at external concentrations at around 0.1 mM (up to 2 mM). CA provokes membrane damage at a 2 to 4 mM range, leading to increased permeability of ions and metabolites. Complete dissipation of proton motive force and the collapse of other ion concentration gradient(s) as well as probably the leakage of cellular metabolites take place at concentrations around the MIC, where the viability of the cell is dramatically decreased. When a much higher concentration of CA is applied (up to 20 mM), UV-absorbing cellular materials as represented by proteins are leaking out.

The results presented in this study strongly suggest that in the human intestine the populations of lactobacilli and bifidobacteria are controlled in part by the concentrations of FBAs. Probiotic lactic acid bacteria accumulate CA (or probably DCA as well) when its concentration is not so high with mild viability decrease. On the other hand, apparent proton conductance and severe membrane permeability disturbance can occur when CA or DCA concentrations in the human gut become higher than threshold values. For example, concentrations of up to 0.3 mM DCA have been detected in the fecal water of healthy subjects (13) and estimation from viability and membrane damage data in Table 2 suggests a possibility that this compound causes considerable inhibition of bacterial growth in the human intestine.