Bile salt–dependent efflux of cellular phospholipids mediated by ATP binding cassette protein B4

Authors

  • Shin-ya Morita,

    1. Laboratory of Cellular Biochemistry, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Aya Kobayashi,

    1. Laboratory of Cellular Biochemistry, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Yasukazu Takanezawa,

    1. Department of Health Chemistry, Graduate School of Pharmaceutical Sciences, University of Tokyo, Tokyo, Japan
    Search for more papers by this author
  • Noriyuki Kioka,

    1. Laboratory of Cellular Biochemistry, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Tetsurou Handa,

    1. Department of Biosurface Chemistry, Graduate School of Pharmaceutical Sciences, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Hiroyuki Arai,

    1. Department of Health Chemistry, Graduate School of Pharmaceutical Sciences, University of Tokyo, Tokyo, Japan
    Search for more papers by this author
  • Michinori Matsuo,

    1. Laboratory of Cellular Biochemistry, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Kazumitsu Ueda

    Corresponding author
    1. Laboratory of Cellular Biochemistry, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto, Japan
    • Laboratory of Cellular Biochemistry, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    Search for more papers by this author
    • fax: (81) 75-753-6104


  • Potential conflict of interest: Nothing to report.

Abstract

Human ABCB4 (multidrug resistance [MDR]3 P-glycoprotein) is expressed in the canalicular membrane of the hepatocyte. ABCB4 has been shown to be required for phosphatidylcholine (PC) secretion into the bile and to translocate PC across the plasma membrane. To further investigate the function of ABCB4, we established a cell line stably expressing ABCB4 (human embryonic kidney [HEK]/ABCB4). The efflux of phospholipids from HEK/ABCB4 cells was remarkably increased by the addition of taurocholate. In addition, the cholesterol efflux from HEK/ABCB4 cells was also enhanced in the presence of taurocholate. Light scattering measurements suggested that the taurocholate monomer plays an important role in ABCB4-mediated lipid secretion. On the other hand, the efflux of phospholipids and cholesterol was not mediated by ABCB1 (MDR1) even in the presence of taurocholate. Taurocholate promoted the efflux of phospholipids and cholesterol from HEK/ABCB4 cells more efficiently than glycocholate and cholate. ABCB4-K435M and ABCB4-K1075M, Walker A lysine mutants, did not mediate the phospholipid and cholesterol efflux in the presence of taurocholate, suggesting that ATP hydrolysis is essential for the efflux. Verapamil completely inhibited the taurocholate-dependent efflux of phospholipids and cholesterol from HEK/ABCB4 cells. Mass spectrometry revealed that, in the presence of taurocholate, HEK/ABCB4 cells preferentially secreted PC compared to sphingomyelin. PC vesicles induced cholesterol diffusion from cell membrane, but did not accept cholesterol from ABCB4. Conclusion: ABCB4 mediates the efflux of phospholipids into the canalicular lumen in the presence of bile salts, and plays a crucial role in bile formation and lipid homeostasis. (HEPATOLOGY 2007.)

Bile formation is important for cholesterol excretion from body as well as cholesterol absorption from the intestine, and thus it is critical for overall cholesterol homeostasis.1 Several ABC proteins, expressed on the apical membranes of hepatocytes, are involved in the secretion of bile salts, phospholipids and cholesterol, and thus in canalicular bile formation. ABCB4, also called multidrug resistance 3 (MDR3) in humans or mdr2 in mice, is essential for the secretion of phosphatidylcholine (PC) into bile.2 ABCB11, also called bile salt export pump or sister of P-glycoprotein, secretes bile salts into bile.3, 4 ABCG5 and ABCG8 are considered to be the main transporters for secretion of biliary cholesterol.5–7

The function of ABCB4 is required for proper bile formation. Mice with homozygous disruption of the Abcb4 gene show almost complete absence of PC from their bile,2, 8–10 which causes segmental biliary strictures due to periductal fibrosis, fibro-obliteration of bile ducts, and spontaneous gallstone formation.10, 11 Human ABCB4 mutations result in a wide spectrum of phenotypes, ranging from progressive familial intrahepatic cholestasis type 3 to adult cholestatic liver disorders characterized by elevated γ-glutamyl transpeptidase levels.12 The primary function of biliary phospholipid excretion is suggested to be to protect the membranes of cells facing the biliary tree against bile salts, and PC in bile salt mixed micelles reduces the detergent activity of micelles.13

ABCB4 has an amino acid sequence with 76% identity and 86% similarity to that of ABCB1 (MDR1). Abcb4 knockout mice do not excrete any phospholipid into bile, despite the high expression of Abcb1 on the canalicular membranes of hepatocytes. ABCB1 has quite low, if any, ability to mediate phospholipid excretion, although it actively transports an enormously broad range of compounds to bile. ABCB4 is predicted to be a floppase for PC.13–17 ABCB4 flops PC from the inner to the outer leaflet of the canalicular membrane to make PC available for extraction into the canalicular lumen by bile salts,13 but the mechanism of PC efflux mediated by ABCB4 has not been characterized in detail.

Disruption of the Abcg5 and Abcg8 genes in mice greatly inhibits cholesterol secretion into bile, whereas the phospholipids and bile salts in bile are virtually normal.6 Thus, ABCG5 and ABCG8 are considered to be the main transporters secreting biliary cholesterol.5–7 However, the gallbladder bile of Abcb4 knockout mice contains only a trace amount of cholesterol,8–10, 18 and biliary cholesterol secretion of Abcb4 knockout mice is not increased by overexpression of human ABCG5 and human ABCG8.18 Therefore, the roles of ABCB4, ABCG5 and ABCG8 in biliary cholesterol excretion have not yet been clarified.

In the present study, we compared the function of ABCB1 and ABCB4 using the HEK293 cell line stably expressing ABCB1 and ABCB4 (HEK/ABCB1 and HEK/ABCB4). We showed phospholipid and cholesterol efflux from HEK/ABCB4 cells, but not from HEK/ABCB1 cells, in the presence of taurocholate. Mass spectrometry revealed that HEK/ABCB4 cells preferentially secreted PC compared to sphingomyelin (SM). The direct involvement of ABCB4 in the efflux of PC and cholesterol will be discussed.

Abbreviations

ABC, ATP binding cassette; MDR3, multidrug resistance 3; PC, phosphatidylcholine; HEK, human embryonic kidney; NaTC, sodium taurocholate; NaGC, sodium glycocholate; NaC, sodium cholate; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum; PBS, phosphate-buffered saline; BSA, bovine serum albumin; LDH, lactate dehydrogenase; cmc, critical micelle concentration; WT, wild-type; SM, sphingomyelin; NBD, 7-nitro-2,1,3-benzoxadiazol group.

Materials and Methods

Materials.

Monoclonal antibody C219 was purchased from Centocor (Philadelphia, PA). Sodium taurocholate (NaTC) was obtained from Wako Pure Chemicals (Osaka, Japan). Sodium glycocholate (NaGC), egg yolk PC, and verapamil hydrochloride were purchased from Sigma Chemical Co. (St. Louis, MO). Sodium cholate (NaC) was purchased from Calbiochem (San Diego, CA). All other chemicals used were of the highest reagent grade available.

Plasmids.

The human ABCB4 gene19 was obtained from the American Type Culture Collection (Manassas, VA). ABCB4 WalkerA lysine mutants (ABCB4-K435M and -K1075M) were prepared with the QuikChange II Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) as described by the manufacturer. The human ABCB4 gene and its mutant genes were inserted into the HindIII-XbaI site of pcDNA3.1/Hygro(+) (Invitrogen, Carlsbad, CA) to make an expression vector for pcDNA3.1/Hygro(+)/ABCB4, pcDNA3.1/Hygro(+)/ABCB4-K435M and pcDNA3.1/Hygro(+)/ABCB4-K1075M. The human ABCB1 gene was fused to His tag in the pCAGGSP vector (pCAGGSP/MDR1-His).

Cell Culture.

HEK293 cells were grown in a humidified incubator (5% CO2) at 37°C in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS).

Establishment of a Stable Transformant of ABCB4.

HEK293 cells were transfected with pcDNA3.1/ Hygro(+)/ABCB4, pcDNA3.1/Hygro(+)/ABCB4-K435M, or pcDNA3.1/Hygro(+)/ABCB4-K1075M using LipofectAMINE (Invitrogen) according to the manufacturer's instructions. Cells were selected with 350 μg/ml hygromycin, and then single colonies were isolated. HEK293 cells transfected with pCAGGSP/ABCB1-His were selected with 400 nM vinblastine and 100 nM adriamycin. The expression of ABCB1 or ABCB4 was examined by Western blotting with monoclonal antibody C219.

Glycosylation of ABCB4.

Cells were lysed with phosphate-buffered saline (PBS) containing 1% Triton X-100 and protease inhibitors [100 μg/ml 4-(amidino)-phenylmethanesulfonyl fluoride hydrochloride, 10 μg/ml leupeptin, and 2 μg/ml aprotinin]. Digestion with endoglycosidase H and peptide N-glycosidase F (New England Biolabs, Ipswich, MA) was done as described by the manufacturer. In brief, cell lysate (20 μg of protein) was treated with 1,000 units of endoglycosidase H or 1,000 units of peptide N-glycosidase F for 1 hour at 37°C. The deglycosylated proteins were electrophoresed on a 7% SDS-polyacrylamide gel and immunodetected using monoclonal antibody C219.

Immunocytochemistry.

Cells were cultured on glass cover slips and fixed in 70% ethanol for 15 minutes. To diminish the nonspecific binding of antibodies, the cells were incubated in 10% goat serum in PBS containing 0.9 mM CaCl2 and 0.5 mM MgCl2. Cells were incubated for 1 hour with monoclonal antibody C219 (10 μg/ml in PBS containing 10% goat serum, 0.9 mM CaCl2 and 0.5 mM MgCl2), and then incubated for 1 hour with fluorescent Alexa488-conjugated anti-mouse IgG (Molecular Probes, Eugene, OR). Cells were directly viewed with a 63× Plan-Neofluar oil immersion objective using a confocal microscope (LSM5 Pascal, Zeiss, Thornwood, NY).

Cellular Phospholipid and Cholesterol Efflux Assay.

Cells were subcultured in poly-L-Lys-coated 6-well plates at a density of 1.0 × 106 cells in DMEM supplemented with 10% FBS. After incubation for 24 hours, the cells were washed with fresh medium and incubated with DMEM (1 ml) containing 0.02% bovine serum albumin (BSA) in the absence or presence of NaTC, NaC, NaGC, PC vesicles, or verapamil. The content of choline phospholipids and cholesterol in the medium was determined after 24 hours incubation as described.20 In brief, the medium was centrifuged to remove cellular debris and lipids in the medium were extracted with chloroform/methanol (2/1). The chloroform/methanol extract was back-extracted with water to remove contaminating proteins and phenol red, which would interfere with the subsequent assays. The amounts of choline phospholipids and free cholesterol were determined using enzymatic assay kits purchased from Wako. Cells were dissolved in 1% Triton X-100. The cell protein concentration measured using a BCA protein assay kit (Pierce, Rockford, IL) was determined to normalize lipid efflux.

Cell Viability Assay.

Cell viability was estimated by measuring the lactate dehydrogenase (LDH) activity in the media and total cells using a CytoTox 96 Non-Radioactive Cytotoxicity Assay Kit purchased from Promega (Madison, WI).

Mass Spectrometric Analysis.

Cells were subcultured in poly-L-Lys-coated 10-cm dishes in DMEM supplemented with 10% fetal bovine serum. The cells were washed with fresh medium and incubated with DMEM containing 0.02% bovine serum antigen in the presence of 0.5 mM NaTC for 24 hours. The medium was centrifuged twice for 15 minutes each at 7,000g to remove cell debris. The lipids were extracted from 12 ml medium by the method of Bligh and Dyer21 after the addition of 14:1-14:1 PC as an internal standard. The content of choline phospholipids in the medium was analyzed by mass spectrometry. Mass spectrometric analyses were performed with a triple quadrupole instrument model Quattro micro (Waters, Milford, MA) equipped with an electrospray source as described.22 The samples were provided by the UltiMate high-performance liquid chromatography system (Dionex Corp., Sunnyvale, CA) into the electrospray interface at a flow rate of 4 μl/minute in a solvent system of acetonitrile-methanol-water (2:3:1) containing 0.1% ammonium formate (pH 6.4). The mass spectrometer was operated in the positive and negative scan modes. The flow rate of the nitrogen drying gas was 12 l/minute at 80°C. The capillary and cone voltages were set at 3.7 kV and 30 V, respectively, argon at 3 × 104 to 4 × 104 Torr was used as the collision gas, and a collision energy of 30-40 V was used to obtain fragment ions for precursor ions.

Preparation of Vesicles.

Small unilamellar vesicles of PC were prepared from a thin film of pure PC obtained by evaporation of a chloroform solution of the lipids. The dried film was hydrated with PBS. The lipid dispersions were ultrasonicated (TOMY Ultrasonic Disruptor UD-200, Tomy Seiko Co., Tokyo, Japan) during 2 cycles of 5 minutes at 1-minute intervals under a nitrogen atmosphere in an ice bath. After the removal of contaminating multilamellar vesicles and metal particles from the ultrasonic tip by ultracentrifugation (198,000g, 2 hour) and filtration through a 0.22-μm filter, homogenous unilamellar vesicles were obtained. The mean particle diameter of vesicles was about 25 nm, as determined by dynamic light scattering measurements using FPAR-1000 (Otsuka Electronics Co., Osaka, Japan). The concentration of PC was determined using an enzymatic assay kit.

Light Scattering Measurements.

The right-angle light scattering of NaTC and PC vesicles in DMEM containing 0.02% bovine serum antigen without phenol red was measured in a spectrofluorometer (Hitachi F-4500, Hitachi High-Technologies Co., Tokyo, Japan) at 37°C using excitation and emission wavelengths of 400 nm.

Statistical Analysis.

The statistical significance of differences between mean values was analyzed using the non-paired t-test. Multiple comparisons were performed using the Dunnet test following ANOVA. Differences were considered significant at P < 0.05. Unless indicated otherwise, results are given as mean ± SE (n = 3).

Results

Expression of Human ABCB4 in HEK293 Cells.

To investigate the function of human ABCB4, cell lines stably expressing ABCB1 (HEK/ABCB1) and ABCB4 (HEK/ABCB4) were established. His-tag was fused to the C-terminus of ABCB1 to distinguish ABCB1 from ABCB4. His-tag fusion has no effect on the function of ABCB1.23, 24 ABCB1 and ABCB4 were expressed as proteins migrating at about 160 kDa and 140 kDa on SDS-PAGE (Fig. 1A). Neither ABCB1 nor ABCB4 was detected in the host HEK293 cells. Both ABCB1 and ABCB4 expressed in HEK293 cells migrated faster after treatment with N-glycosidase F but not after treatment with endoglycosidase H (Fig. 1B). These results suggest that both ABCB1 and ABCB4 were sorted through the trans Golgi and modified with complex oligosaccharides.

Figure 1.

Expression and glycosylation of ABCB4 in HEK293 cells. (A) Cell lysates (9.4 μg of proteins) from HEK293 cells (lane 1), HEK/ABCB1 cells (lane 2) and HEK/ABCB4 cells (lane 3) were separated by 7% polyacrylamide gel electrophoresis. The C-terminal of ABCB1 was fused to His tag. (B) Cell lysates (17.1 μg of proteins) from HEK/ABCB1 cells (lanes 1-3) and HEK/ABCB4 cells (lane 4-6) were treated without (lanes 1, 4) or with endoglycosidase H (H; lanes 2, 5) or peptide N-glycosidase F (F; lanes 3, 6). The samples were separated by 7% polyacrylamide gel electrophoresis. ABCB1 and ABCB4 were detected with monoclonal antibody C219.

Phospholipid Secretion from HEK/ABCB4 Cells in the Presence of NaTC.

To examine whether ABCB4 is directly involved in cellular phospholipid efflux, we measured phospholipid contents in the medium of HEK293, HEK/ABCB1, and HEK/ABCB4 cells. Disappointingly, there was no significant difference in the phospholipid secretion from these three cell lines into the medium (containing 0.02% BSA) (Fig. 2A, open bars). Because bile salts are secreted across the canalicular membrane of the hepatocyte and exist in the canalicular space,4 NaTC was added to the medium to examine the effect on ABCB4-mediated lipid secretion. Phoshpolipid secretion from HEK293 host cells or HEK/ABCB1 cells was not affected by the addition of 0.5 mM NaTC to the medium. However, the phospholipid secretion from HEK/ABCB4 cells was markedly increased in the presence of 0.5 mM NaTC (Fig. 2A, closed bars).

Figure 2.

ABCB4-mediated secretion of cellular phospholipids (A) and cholesterol (B). HEK293 cells, HEK/ABCB1 cells and HEK/ABCB4 cells were incubated for 24 hours at 37°C with 0.02% BSA in the absence (control, open bars) or presence (filled bars) of 0.5 mM NaTC. Bars represent the mean ± S.E. of 3 measurements. *P < 0.05, significantly different from control. #P < 0.05, significantly different from HEK293 cells.

Cholesterol Secretion from HEK/ABCB4 Cells in the Presence of NaTC.

Cholesterol secretion from HEK293, HEK/ABCB1, and HEK/ABCB4 cells into the medium was also measured. The cholesterol secretion from these three cell lines was similar in the absence of NaTC (Fig. 2B, open bars). Cholesterol secretion from HEK293 host cells or HEK/ABCB1 cells was not influenced by the addition of 0.5 mM NaTC. However, the addition of NaTC led to a significant enhancement of the cholesterol secretion from HEK/ABCB4 cells (Fig. 2B, closed bars).

Effects of Bile Salt Structure on ABCB4-Mediated Lipid Secretion.

The majority of bile salts are conjugated in the liver to glycine or taurine. The glycine/taurine ratio is 2/1 in human gallbladder. The sum of NaGC and NaTC is roughly 50% of total bile salts in gallbladder bile.25 The hydrophilic taurine and glycine groups increase the polar surface of the bile salt. To examine the effect of the bile salt structure on the ABCB4-mediated lipid secretion, lipid secretion in the presence of NaC, NaGC and NaTC was examined. The phospholipid secretion from HEK/ABCB4 cells increased in the order of NaC < NaGC < NaTC when eash agent was included at 0.5 mM (Fig. 3A). The same tendency was observed for the cholesterol secretion from HEK/ABCB4 cells (Fig. 3B). It has been reported that NaTC induces greater secretion of phospholipids and cholesterol than NaGC in hamsters.26 The features of the PC and cholesterol secretion from HEK/ABCB4 cells were consistent with this observation.

Figure 3.

Effects of NaC, NaGC and NaTC on the secretion of phospholipids (A) and cholesterol (B) from HEK/ABCB4 cells. HEK/ABCB4 cells were incubated for 24 hours at 37°C with 0.02% BSA in the absence (control) or presence of 0.5 mM NaC, 0.5 mM NaGC or 0.5 mM NaTC. Bars represent the mean ± SE of 3 measurements. *P < 0.05, significantly different from control.

NaTC-Dependent Secretion of Phospholipids and Cholesterol from HEK/ABCB4 Cells.

The dependency of lipid secretion from HEK/ABCB4 cells on the NaTC concentration was examined. The secretion of phospholipids and cholesterol from HEK/ABCB4 cells were synchronously enhanced with increasing concentrations of NaTC, and showed concentration-dependence from 0.2 mM to 1 mM NaTC (Fig. 4A,B). The phospholipid and cholesterol secretion from HEK/ABCB4 cells were increased 3.4- and 3.5-fold by the addition of 0.5 mM NaTC, respectively. They increased up to 8.6- and 7.9-fold in the presence of 1 mM NaTC, respectively, while no increase in the lipid secretion was observed from HEK293 cells or HEK/ABCB1 cells. Incubation with 1 mM NaTC for 24 hours did not affect the viability of HEK293 or HEK/ABCB1 cells, as estimated by the LDH release (data not shown). The addition of 0.5 mM NaTC had no effect on the LDH release from HEK/ABCB4 cells. However, the LDH release from HEK/ABCB4 cells in the presence of 1 mM NaTC became 3.4-fold higher than that in the absence of NaTC (3.6%). The medium of HEK/ABCB4 cells in the presence of 1 mM NaTC might contain cellular debris that was not removed by centrifugation. The viability of HEK293 cells expressing nonfunctional ABCB4 mutants in the ATP binding domain, described later, was not affected by 1 mM NaTC (data not shown). The increased sensitivity to 1 mM NaTC was probably due to the phospholipid imbalance and the loss of membrane integrity27, 28 caused by PC and cholesterol secretion by ABCB4.

Figure 4.

Effects of NaTC on the secretion of cellular phospholipids (A) and cholesterol (B). HEK cells (open circles), HEK/ABCB1 cells (closed circles) and HEK/ABCB4 cells (closed triangles) were incubated for 24 hours at 37° C with 0.02% BSA in the presence of the indicated concentrations of NaTC. Values represent the mean ± SE of three measurements. *P < 0.05, significantly different from the efflux in the absence of NaTC. #P < 0.05, significantly different from HEK293 cells. (C) Light scattering intensity of NaTC in DMEM containing 0.02% BSA at 37°C. The section mark § indicates the cmc (2.5 mM). The light scattering intensity does not change below the cmc, and then increases above the cmc due to the micelle formation. The cmc was defined as the concentration at the intersection between the lines for the monomer and micelle. Values represent the mean ± SD of 3 measurements.

The critical micelle concentration (cmc) of NaTC is 1-5 mM, and is influenced by the salt concentration.29, 30 The light scattering method has been employed to determine the cmc of micellar solutions.29 The micelle formation of NaTC in DMEM containing 0.02% BSA at 37°C was assessed by measuring the light scattering intensity. The cmc for NaTC was estimated to be 2.5 mM (Fig. 4C), and below 2.5 mM, NaTC molecules were present as monomers. NaTC above the cmc could not be used in this study because of its cytotoxicity to HEK293 cells (data not shown). These results suggest that monomers of NaTC in the culture medium mainly support the phospholipid and cholesterol secretion from HEK/ABCB4 cells.

We also examined whether the volume of the medium containing NaTC affects the secretion of lipids from HEK/ABCB4 cells. The amount of phospholipid and cholesterol secreted from HEK/ABCB4 cells into 2 ml medium containing 0.5 mM NaTC was the same as that into 1 ml medium containing 0.5 mM NaTC (data not shown). These results indicate that the secretion of lipids mediated by ABCB4 is dependent on the concentration of NaTC rather than the total amount of NaTC under these conditions.

Effects of PC Vesicles on Secretion of Cellular Cholesterol.

Because abundant unilamellar lipid vesicles are present in the bile canalicular lumina,31 we next explored whether PC vesicles, in the absence or presence of NaTC, play a role in the cholesterol secretion mediated by ABCB4. As shown in Fig. 5A, the addition of PC vesicles resulted in a dose-dependent increase in the cholesterol secretion from HEK293 cells, HEK/ABCB1 cells and HEK/ABCB4 cells. However, in the presence of PC vesicles, the cholesterol secretion from HEK/ABCB4 cells was not significantly different from the secretion from HEK293 cells or HEK/ABCB1 cells. The cholesterol secretion in the presence of PC vesicles may have been due to the diffusion of cholesterol from the cell membrane to PC vesicles. These results suggest that PC vesicles cannot act as acceptors for cholesterol secreted by ABCB4.

Figure 5.

Effects of PC vesicles and NaTC on the secretion of cellular cholesterol. (A) HEK293 cells (open circles), HEK/ABCB1 cells (closed circles) and HEK/ABCB4 cells (closed triangles) were incubated for 24 hours at 37°C with 0.02% BSA in the presence of the indicated concentrations of PC vesicles. Values represent the mean ± SE of 3 measurements. There was no significant difference between HEK, HEK/ABCB1 and HEK/ABCB4 cells. (B) HEK293 cells (open circles), HEK/ABCB1 cells (closed circles) and HEK/ABCB4 cells (closed triangles) were incubated for 24 hours at 37°C with 0.02% BSA and 0.1 mM PC vesicles in the presence of the indicated concentrations of NaTC. Values represent the mean ± SE of 3 measurements. *P < 0.05, significantly different from the efflux in the absence of NaTC. #P < 0.05, significantly different from HEK293 cells. (C) Change of light scattering intensity of 0.1 mM PC vesicles with NaTC concentration in DMEM containing 0.02% BSA at 37°C. The vesicle-micelle transition results in light scattering changes. Values represent the mean ± SD of 3 measurements.

The effect of NaTC on the cholesterol secretion from HEK/ABCB4 cells was also examined in the presence of PC vesicles (Fig. 5B). In the presence of 0.1 mM PC vesicles, the cholesterol secretion from HEK/ABCB4 cells was further stimulated by the addition of NaTC in a dose-dependent manner. However, the addition of NaTC had no effect on the cholesterol secretion from HEK293 cells or HEK/ABCB1 cells in the presence of PC vesicles.

Vesicles are large particles that cause strong light scattering, whereas micelles are small ones, and thus weakly scatter light. Thus, the vesicle-micelle transition results in light scattering changes. Fig. 5C shows the change of the light scattering of PC vesicles in DMEM containing 0.02% BSA upon the addition of NaTC. The light scattering slightly increased at 1 mM NaTC due to the association of NaTC molecules with PC bilayers, indicating NaTC molecules were distributed in the PC bilayers and the aqueous medium. The light scattering strongly decreased from 1 to 4 mM NaTC due to the rupture of liposomes followed by formation of mixed NaTC/PC micelles. These results suggest that NaTC monomers mainly facilitated the secretion of cholesterol from HEK/ABCB4 cells in the presence of 0.1 mM PC and 1 mM NaTC.

ATP-Dependence of Lipid Efflux Mediated by ABCB4.

To study the involvement of ATP hydrolysis in ABCB4-mediated secretion of phospholipids and cholesterol in the presence of NaTC, we established HEK293 cells stably expressing ABCB4-K435M (HEK/ABCB4-K435M) and ABCB4-K1075M (HEK/ABCB4-K1075M), in which the Walker A lysine in either nucleotide binding domain was substituted by methionine. NaTC-dependent efflux of phospholipids and cholesterol was not observed with HEK/ABCB4-K435M or HEK/ABCB4-K1075M cells (Fig. 6F,G), although the expression levels, glycosylation, and the surface expression of mutant ABCB4 were comparable to those of wild-type ABCB4 (Fig. 6A-E). These results suggest that the NaTC-dependent efflux of phospholipids and cholesterol is mediated by ABCB4 in an ATP-dependent manner.

Figure 6.

Effects of ABCB4 Walker A lysine mutations on the efflux of cellular phospholipids and cholesterol. (A) Cell lysates (32.0 μg of proteins) from HEK/ABCB4-WT cells (lane 1), HEK/ABCB4-K435M cells (lane 2) and HEK/ABCB4-K1075M cells (lane 3) were separated by 7% polyacrylamide gel electrophoresis. ABCB4-WT, -K435M and -K1075M were detected with mouse monoclonal antibody C219. (B-E) HEK293 cells (B), HEK/ABCB4-WT cells (C), HEK/ABCB4-K435M cells (D), and HEK/ABCB4-K1075M cells (E) were fixed in 70% ethanol and reacted with monoclonal antibody C219 and Alexa488-conjugated anti-mouse IgG. The bars represent 10 μm. (F-G) HEK/ABCB4 cells, HEK/ABCB4-K435M cells and HEK/ABCB4-K1075M cells were incubated for 24 hours at 37°C with 0.02% BSA in the absence (control, open bars) or presence (filled bars) of 0.5 mM NaTC. Bars represent the mean ± SE of 3 measurements. *P < 0.05, significantly different from control. #P < 0.05, significantly different from HEK/ABCB4-WT cells.

Effects of Verapamil on ABCB4-Mediated Lipid Efflux.

Verapamil is a transport substrate for ABCB1, and in addition is a potent inhibitor of ABCB1 and ABCB4.14, 17, 32–34 The effect of verapamil on the ABCB4-mediated lipid efflux was examined (Fig. 7). The addition of 50 μM verapamil, which inhibits the function of ABCB1, completely blocked the NaTC-dependent efflux of phospholipids and cholesterol mediated by ABCB4. In the absence of NaTC, verapamil showed no effect on lipid efflux from HEK/ABCB4 cells.

Figure 7.

Effects of verapamil on the efflux of phospholipids (A) and cholesterol (B) from HEK/ABCB4 cells. HEK/ABCB4 cells were incubated for 24 hours at 37°C with 0.02% BSA and 50 μM verapamil in the absence (control) or presence of 0.5 mM NaTC. Bars represent the mean ± SE of 3 measurements. *P < 0.05, significantly different from control. #P < 0.05, significantly different from the efflux in the absence of verapamil.

Molecular Species of Choline Phospholipids Secreted by ABCB4.

The predominant (>95%) biliary phospholipid is PC, while sphingomyelin (SM) is present only in trace amounts in bile.35 The profile of choline phospholipids secreted by ABCB4 in the presence of NaTC was analyzed by using electrospray ionization mass spectrometry. The medium of HEK/ABCB4 cells in the presence of NaTC contained higher amounts of various choline phospholipids (16:0-18:1 SM, 16:0-16:1 PC, 16:0-18:2 PC, 16:0-18:1 PC, and 18:0-18:2 PC) than that of the host HEK293 cells (Fig. 8). In the medium of HEK/ABCB4 cells, the peak height of 16:0-18:1 PC was the highest and followed by that of 16:0-16:1 PC. The peak height of 16:0-18:1 SM, the highest peak of SM, was lower than those of PC. Although the ion-peaks from a triple quadupole mass spectrometer do not allow for direct comparison between phospholipid species, we have successfully demonstrated by using mass spectrometry36 that ABCG1 and ABCA1 mediate the secretion of several species of SM and PC, and that ABCG1 preferentially mediates the excretion of SM, while ABCA1 preferentially mediates the excretion of PC. The rough total peak heights of PC in the medium of HEK/ABCB4 were calculated to be several fold higher than that of SM. These results suggest that ABCB4, like ABCA1, preferentially mediates the excretion of PC compared to SM.

Figure 8.

Positive-ion electrospray ionization MS spectra of phospholipid molecular species in lipid extracts of the culture medium of HEK/ABCB4 cells. HEK293 cells (A) and HEK/ABCB4 cells (B) were incubated for 24 hours at 37°C with 0.02% BSA in the presence of 0.5 mM NaTC. Aliquots of chloroform extracts were infused directly into the electrospray ionization ion source using an UltiMate HPLC system at a flow rate of 4 μl/min. Positive-ion electrospray ionization of lipid extracts of the cell medium was performed as described in Materials and Methods. Individual molecular species were identified using tandem mass spectrometry. The internal standard (I.S.) is 14:1-14:1 PC (m/z 647.5) and represented as 100%.

Direct Involvement of ABCB4 in Cholesterol Efflux.

Finally, we examined whether the cholesterol secretion from HEK/ABCB4 cells was directly mediated by ABCB4 or was due to diffusion from the cell membrane to mixed particles of NaTC and PC excreted by ABCB4. The medium of HEK/ABCB4 cells contained 2.5 μM PC after a 24-hour incubation in the presence of 0.5 mM NaTC. However, no significant increase in the cholesterol secretion from HEK293 host cells was observed in the presence of 0.5 mM NaTC and 2.5 μM PC compared to the secretion in the absence of NaTC and PC (Fig. 9). This experiment suggests that ABCB4 directly mediates the efflux of cholesterol together with PC from cells. However, because the local concentration of PC exported by ABCB4 would be higher in the vicinity of the plasma membrane, the diffusion of cholesterol from the cell membrane to the NaTC/PC vesicles might also be involved.

Figure 9.

Effects of 0.5 mM NaTC and 2.5 μM PC on the cholesterol secretion from HEK293 cells. HEK293 cells were incubated for 24 hours at 37°C with 0.02% BSA in the absence (control) or presence of 0.5 mM NaTC and 2.5 μM PC. The culture medium of HEK/ABCB4 cells contained 2.5 μM PC after a 24-hour incubation in the presence of 0.5 mM NaTC. Bars represent the mean ± SE of 3 measurements. There was no significant difference between the cholesterol secretion from HEK293 cells in the absence and presence of 0.5 mM NaTC and 2.5 μM PC.

Discussion

In this study, we established HEK293 cells stably expressing human ABCB4, and showed that ABCB4 mediates the excretion of phospholipids (preferentially PC) and cholesterol in the presence of bile salts. Because the excretion of phospholipids and cholesterol was not observed with ABCB4-K435M or ABCB4-K1075M, in which the Walker A lysine in either nucleotide binding domain was substituted by methionine, ABCB4 mediates lipid efflux in an ATP-dependent manner, like other ABC transporters.

It has been suggested that ABCB4 promotes the transfer of PC from the inner to the outer leaflet of the plasma membrane, mainly based on experiments using a fluorescent PC analog containing a 7-nitro-2,1,3-benzoxadiazol group (NBD).14, 16 However, ABCB1, which is not physiologically involved in PC secretion into bile, also translocates a short-chain PC analog (C6-NBD-PC).16 Smith et al. have demonstrated that, in fibroblasts from transgenic mice expressing the human ABCB4 gene, ABCB4 translocates long-chain radioactive PC that is exchanged by PC-transfer protein shuttling between the outer leaflet of plasma membrane and acceptor liposomes in the medium.15 This experiment has suggested that liposomes themselves were unable to accept long-chain PC translocated by ABCB4 and some acceptors such as PC-transfer proteins might be required for excretion of PC to the medium. Thus, the mechanism of PC secretion mediated by ABCB4 has not yet been clarified.

Recently, we reported that ABCG1 mediates the secretion of phospholipids (preferentially SM) and cholesterol in the presence of BSA.36 In this case, BSA most likely works as an acceptor for lipids secreted from the cells expressing ABCG1. However, no lipid efflux by ABCB4 was observed in the presence of BSA (Fig. 2). The addition of bile salts was required for the ABCB4-mediated excretion of phospholipids and cholesterol. While 1 mM NaTC showed cytotoxicity to HEK/ABCB4 cells, 0.5 mM NaTC did not affect their viability. The efflux of phospholipids and cholesterol in the presence of 0.5 mM NaTC is considered to be mediated by ABCB4.

Several lines of evidence in this study suggest that bile salt monomers, but not PC vesicles, support the lipid efflux mediated by ABCB4. (1) The presence of NaTC below the cmc promoted the lipid efflux mediated by ABCB4 (Fig. 4). (2) The addition of PC vesicles increased the cholesterol efflux from HEK293 cells in a dose-dependent manner, but the differences between HEK293, HEK/ABCB1, and HEK/ABCB4 cells were not significant (Fig. 5A). (3) A synergistic effect of PC vesicles and NaTC was not observed (Fig. 5B). (4) Within 4 hours of incubation with NaTC below the cmc (0.5 mM), a significant amount of [3H]cholesterol was secreted from HEK/ABCB4 cells, but not HEK293 cells or HEK/ABCB1 cells (data not shown). Therefore, the monomer form of bile salts most likely function, at least initially, in supporting the ABCB4-mediated lipid efflux. However, because bile salts may form mixed micelles, even at concentrations below the cmc, with phospholipids secreted by ABCB4, the mixed micelle may also support the lipid efflux mediated by ABCB4.

The medium of HEK/ABCB4 cells contained 2.5 μM PC after a 24-hour incubation in the presence of 0.5 mM NaTC. However, no significant increase in the cholesterol secretion from HEK293 host cells was observed in the presence of 0.5 mM NaTC and 2.5 μM PC compared to the secretion in the absence of NaTC and PC as shown in Fig. 9. Therefore, ABCB4 might directly mediate the efflux of cholesterol together with PC from cells. However, because the local concentration of PC exported by ABCB4 would be higher in the vicinity of the plasma membrane, the diffusion of cholesterol from the cell membrane to the NaTC/PC vesicles may also occur. Further studies are necessary to show the direct involvement of ABCB4 in cholesterol efflux out of cells. Even in the case of ABCA1, which mediates apoA-I-dependent PC and cholesterol efflux, the direct involvement of ABCA1 in cholesterol movement is still controversial.37, 38

Cholesterol excretion into the bile is almost completely impaired in the Abcb4 knockout mice.8–10, 18 In the absence of either ABCG5 or ABCG8 (or both), cholesterol excretion in mice is strongly reduced.6 However, residual cholesterol secretion is still observed in Abcg5/Abcg8 knockout mice.6 Furthermore, expression of the human ABCG5 and ABCG8 transgenes does not increase biliary cholesterol in Abcb4 knockout mice.18 Based on these results, it has been proposed that ABCG5 and ABCG8 require ABCB4 for the secretion of cholesterol into the bile,18 and that Abcg5/Abcg8-independent routes can significantly contribute to total hepatobiliary cholesterol output (≈20% of maximal output).39 However, there were no data to suggest that ABCB4 is involved in the ABCG5/ABCG8-independent routes. It has been also demonstrated that biliary cholesterol excretion in Abcb4 knockout mice can be completely restored by infusion of a sufficiently hydrophobic bile salt, taurodeoxycholate, to allow solubilization of cholesterol in the absence of phospholipids,8 and that cholesterol excretion in heterozygous Abcb4+/− mice is equal to that in wild-type mice although phospholipid excretion in these mice is 50% of normal levels.8,40 Hence, the reason for absence of cholesterol excretion in the Abcb4 knockout mice is considered to be the absence of phospholipids in bile without which cholesterol cannot be properly solubilized. In this study, we demonstrated the bile salt-dependent excretion of cholesterol from HEK/ABCB4 cells in addition to phospholipids. However, PC vesicles evoke the diffusion of cellular cholesterol as shown in Fig. 5A. At present, it is not clear whether ABCB4 is directly involved in translocation or efflux of cholesterol.

ABCB4 has a 76% overall amino acid sequence identity with ABCB1.19 Therefore, it is conceivable that ABCB4 has conserved domains for substrate recognition in common with ABCB1. Indeed, it has been reported that both ABCB1 and ABCB4 confer resistance to aureobasidin A, an antifungal cyclic depsipeptide antibiotic, when expressed in yeast.32 The resistance of yeast cells to aureobasidin A conferred by ABCB4 can be overcome by vinblastine, verapamil, and cyclosporin A. Transepithelial transport of C6-NBD-PC and digoxin by ABCB4 through LLC-PK1 cells are inhibited by vinblastine, verapamil, and cyclosporin A, indicating an interaction between these compounds and ABCB4.33 The translocation of NBD-PC in yeast secretory vesicles is mediated by both ABCB1 and ABCB4 and abrogated by verapamil.14 Our results showed that verapamil almost completely abolished the NaTC-dependent efflux of phospholipids and cholesterol mediated by ABCB4 (Fig. 7). These results indicate that ABCB1 and ABCB4 have quite similar substrate binding domains. However, ABCB1 cannot transport PC even in the presence of NaTC, wheras ABCB4 can.

Recently, the crystal structure of sav1866, a bacterial homolog of ABCB1, was reported.41 The predicted outward-facing conformation contains a central cavity that is shielded from the inner leaflet of the lipid bilayer and from the cytoplasm, but exposed to the outer leaflet and the extracellular space. PC could be presented by ABCB4 to the outside of the cell but still remain on the protein surface or in the cavity. Only after binding of bile salt molecules, phospholipids could be released into the extracellular space, and then a mixed bile salt/phospholipid micelle is formed. The association of bile salt monomers with phospholipid reduces the activation energy required to move substrates from the protein surface of ABCB4 to the aqueous environment. The increase in the intramicellar concentration of PC leads to the conversion of mixed micelles into vesicles. Vesicles or micelles may be unable to access to the cavity of ABCB4, probably due to their larger sizes compared with bile salt monomers.

The orders of bile salt-induced efflux of phospholipids and cholesterol, from the highest to the lowest, were as follows: NaTC > NaGC > NaC (Fig. 3), which was not consistent with the order of maximum solubility of cholesterol in bile salt micelles, NaGC > NaTC > NaC.29 The hydrophilic-hydrophobic balance of bile salt has been quantified as the bile salt monomeric hydrophobicity index.42, 43 The hydrophobicity indices, from least to most hydrophobic, are NaTC (0) < NaGC (+0.07) < NaC (+0.13).42 Therefore, the bile salt-dependent efflux of phospholipids and cholesterol by ABCB4 was inversely correlated with the bile salt hydrophobicity index. It is possible that bile salt monomers may associate not only with the phospholipid secreted by ABCB4 but also with the amino acids lining on the surface of the cavity of ABCB4, and that this helps the release of phospholipid.

Mass spectrometry revealed that the medium of HEK/ABCB4 contained more PC than SM in the presence of NaTC. Species of choline phospholipids secreted by ABCB4 were quite similar to those secreted by ABCA1 in the presence of apoA-I, but different from those secreted by ABCG1. These results may suggest that phospholipid specificity as substrates are similar between ABCB4 and ABCA1 but different from ABCG1. Alternatively, ABCB4 and ABCA1 preferentially function in SM-poor non-raft domains, while ABCG1 preferentially functions in SM-rich raft domains in the plasma membrane.

In conclusion, ABCB4 mediates the secretion of phospholipids, preferentially PC, from hepatocytes into the canalicular lumen in the presence of bile salts in an ATP-dependent manner. We propose that bile salt monomers function as acceptors for PC. Furthermore, the phospholipid secretion mediated by ABCB4 results in the formation of lipid vesicles, which would induce cholesterol diffusion from the canalicular membrane. Therefore, ABCB4 plays an important role in bile formation and lipid homeostasis.

Ancillary