The membrane protein ATPase class I type 8B member 1 signals through protein kinase C zeta to activate the farnesoid X receptor

Authors

  • Tamara Frankenberg,

    1. Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
    • These authors contributed equally to this study.

  • Tamir Miloh,

    1. Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
    • These authors contributed equally to this study.

  • Frank Y. Chen,

    1. Division of Gastroenterology, Hepatology and Nutrition, Children's Hospital of Pittsburgh, University of Pittsburgh Medical Center, Pittsburgh, PA
    2. Department of Pediatrics, University of Pittsburgh School of Medicine, Pittsburgh, PA
    Search for more papers by this author
  • Meena Ananthanarayanan,

    1. Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
  • An-Qiang Sun,

    1. Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
  • Natarajan Balasubramaniyan,

    1. Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
  • Irwin Arias,

    1. Cell Biology and Metabolism Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Kenneth D. R. Setchell,

    1. Department of Pathology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH
    2. Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH
    Search for more papers by this author
  • Frederick J. Suchy,

    1. Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
  • Benjamin L. Shneider

    Corresponding author
    1. Division of Gastroenterology, Hepatology and Nutrition, Children's Hospital of Pittsburgh, University of Pittsburgh Medical Center, Pittsburgh, PA
    2. Department of Pediatrics, University of Pittsburgh School of Medicine, Pittsburgh, PA
    • Division of Pediatric Gastroenterology, Hepatology, and Nutrition, Children's Hospital of Pittsburgh, University of Pittsburgh Medical Center, 3705 Fifth Avenue, Pittsburgh, PA 15213
    Search for more papers by this author
    • fax: 412 692 8906


  • Potential conflict of interest: Nothing to report.

Abstract

Prior loss-of-function analyses revealed that ATPase class I type 8B member 1 [familial intrahepatic cholestasis 1 (FIC1)] posttranslationally activated the farnesoid X receptor (FXR). Mechanisms underlying this regulation were examined by gain-of-function studies in UPS cells, which lack endogenous FIC1 expression. FXR function was assayed in response to wild-type and mutated FIC1 expression constructs with a human bile salt export pump (BSEP) promoter and a variety of cellular localization techniques. FIC1 overexpression led to enhanced phosphorylation and nuclear localization of FXR that was associated with FXR-dependent activation of the BSEP promoter. The FIC1 effect was lost after mutation of the FXR response element in the BSEP promoter. Despite similar levels of FIC1 protein expression, Byler disease FIC1 mutants did not activate BSEP, whereas benign recurrent intrahepatic cholestasis mutants partially activated BSEP. The FIC1 effect was dependent on the presence of the FXR ligand, chenodeoxycholic acid. The effect of FIC1 on FXR phosphorylation and nuclear localization and its effects on BSEP promoter activity could be blocked with protein kinase C zeta (PKC ζ) inhibitors (pseudosubstrate or small interfering RNA silencing). Recombinant PKC ζ directly phosphorylated immunoprecipitated FXR. The mutation of threonine 442 of FXR to alanine yielded a dominant negative protein, whereas the phosphomimetic conversion to glutamate resulted in FXR with enhanced activity and nuclear localization. Inhibition of PKC ζ in Caco-2 cells resulted in activation of the human apical sodium-dependent bile acid transporter promoter. Conclusion: These results demonstrate that FIC1 signals to FXR via PKC ζ. FIC1-related liver disease is likely related to downstream effects of FXR on bile acid homeostasis. Benign recurrent intrahepatic cholestasis emanates from a partially functional FIC1 protein. Phosphorylation of FXR is an important mechanism for regulating its activity. (HEPATOLOGY 2008;48:1896-1905.)

Mutations in ATPase class I type 8B member 1 [ATP8B1; familial intrahepatic cholestasis 1 (FIC1)] lead to a spectrum of liver diseases.1–4 The more mild end of the spectrum of FIC1 disease is termed benign recurrent intrahepatic cholestasis (BRIC),5 whereas the more severe disease is known as Byler disease or progressive FIC1.6 The range of liver disease is presumed in large part to be related to the severity of the functional defect associated with the specific mutation in ATP8B1, although this has not been formally assessed.4 The liver disease may be accompanied by extrahepatic manifestations. These problems do not improve after liver transplantation; the diarrhea may worsen considerably, and steatohepatitis may develop as a new problem after liver replacement.7 FIC1 is expressed broadly among tissues in the body, and this accounts in part for its varied extrahepatic manifestations.1, 8, 9

The precise function of FIC1 and the pathophysiology of its variable disease manifestations are not well understood. Nucleotide homology analysis suggests that FIC1 could be a phospholipid flippase, potentially transferring aminophospholipids from the outer hemileaflet to the inner hemileaflet of the lipid bilayer.1, 10 A Chinese hamster ovary (CHO) cell line that lacks FIC1 has impaired lipid transport capacity.8, 11 Expression of FIC1 in this cell line enhances phosphatidylserine transport.8, 12 An analysis of a limited number of human ileal tissue samples suggested that FIC1 might signal through the farnesoid X receptor (FXR).13 Confirmation of these findings using human liver tissue has been controversial and problematic because of the limited number of samples analyzed and the potential effects of the intrinsic liver disease on gene expression.14, 15 In vitro studies revealed that nuclear localization of FXR was diminished when FIC1 was knocked down.13 Overexpression of FXR after FIC1 silencing did not rescue the effect, and this suggested that posttranscriptional regulation was operative. FXR plays a key role in a variety of biologically important processes.16–23 FXR-mediated transcriptional effects are of fundamental relevance in bile acid homeostasis, including regulation of ileal bile acid uptake by the apical sodium-dependent bile acid transporter (ASBT) and canalicular bile acid excretion via the bile salt export pump (BSEP).24–29 The following studies were performed with a gain-of-function model to further assess the potential role that FIC1 may play in modifying FXR function.

Abbreviations

ASBT, apical sodium-dependent bile acid transporter; ATP8B1, ATPase class I type 8B member 1; BRIC, benign recurrent intrahepatic cholestasis; BSEP, bile salt export pump; CDCA, chenodeoxycholic acid; cDNA, complementary DNA; CHO, Chinese hamster ovary; CTFCS, charcoal-treated fetal calf serum; FCS, fetal calf serum; FIC1, familial intrahepatic cholestasis 1; FXR, farnesoid X receptor; FXR-GFP, green fluorescent protein–fused farnesoid X receptor; GFP, green fluorescent protein; hASBT, human apical sodium-dependent bile acid transporter; hFXR, human farnesoid X receptor; hFXR-GFP, green fluorescent protein–fused human farnesoid X receptor; MBP, myelin basic protein; P-FXR, phosphorylated farnesoid X receptor; P-MBP, myelin basic protein; PI3K, phosphatidyl inositol 3-kinase; PKC, protein kinase C; PKC ζ PS, protein kinase C zeta pseudosubstrate; SHP, short heterodimer partner; siRNA, small interfering RNA; wt, wild type.

Materials and Methods

Cells and Cell Culture.

UPS cells (generously provided by Dr. Richard Pagano, Mayo Medical Center, Rochester, MN) were grown and maintained in Ham's F-12 medium supplemented with 10% fetal calf serum (FCS). CV-1 (monkey kidney),29 Caco-2, and HEK-293 cells (CRL-1573, American Type Culture Collection, Rockville, MD) were grown and maintained in Dulbecco's modified Eagle's medium containing 10% FCS. UPS cells were cultured at 33°C, whereas CV-1 and HEK-293 cells were cultured at 37°C, both in 5% CO2. The effect of the FXR ligand, chenodeoxycholic acid (CDCA), was investigated through the incubation of cells in 0.5% charcoal-treated fetal calf serum (CTFCS; Cocalico Biological, Inc., Reamstown, PA) with or without additional CDCA. The concentrations of serum total bile acid and the principal individual bile acids, CDCA, cholic acid, deoxycholic acid, lithocholic acid, and ursodeoxycholic acid, were measured in undiluted FCS and CTFCS by stable-isotope dilution selected ion monitoring gas chromatography–mass spectrometry with previously described and validated methods.30–32

Plasmid Constructs.

Two hundred thirty-one base pairs of the BSEP promoter (−145 to +86), linked to luciferase expression vector pSV0ALΔ5′ (p-145/Luc),29 were used as a readout of the FXR activity. For analyses in Caco-2 cells, the wild-type human apical sodium-dependent bile acid transporter (hASBT) promoter and the retinoic acid receptor cis-element mutant (hASBTμ) were used as previously described.33 pEF-FIC1, encoding a full-length wild-type human FIC1 coding region, was used to express FIC1 in UPS cells.8 pEF-FIC1/G308V and pEF-FIC1/δ795-7 contained mutant FIC1 genes that had been characterized in the Byler and BRIC diseases, respectively.34 Four additional human FIC1 mutant constructs, pEF-FIC1/D554N (1660 G>A ccaggcagcctctcccaatgaaggtgccctgg; mutated nucleotides in bold italics) and pEF-FIC1/G1040R (3118 G>A gcttgttgcatagggtcc; both described in Byler disease) and pEF-FIC1/R600W (1798 C>T ggacttcaacagtgactcgaagcg) and pEF-FIC1/R600Q (1799 G>A ggacttcaacagtgaccagaagcg; both described in BRIC),4 were prepared with the Quick-Change site-directed mutagenesis kit (Stratagene, La Jolla, CA).28 These specific mutations were chosen because their clinical phenotypes are relatively well described. The point mutations were confirmed by DNA sequencing. FIC1 silencing was accomplished with a previously described construct, siFIC1, whereas a scrambled antisense construct, siScr, was used as a control.13

A protein kinase C zeta (siPKC ζ) small interfering RNA (siRNA) expression construct was synthesized (GenScript Corp., Piscataway, NJ) with a PKC ζ complementary DNA (cDNA) sequence derived from UPS cells. The coding sequence for hamster PKC ζ cDNA was determined by reverse-transcription polymerase chain reaction of the hamster gene with the following oligonucleotide primers: PKC ζ 5′-ACCTCGTCCCGCTGACCTGCAGG-3′ and 5′-ACTCTGCCTCTGCATGTGGAAC-3′ (nucleotides 504-525 and 1000-1028, respectively, in the human PKC ζ gene, GenBank accession number Z15108, and the hamster partial PKC ζ gene, GenBank accession number DQ913741). The entire region was used for northern analyses of the expression of PKC ζ in UPS cells. The selected sequence for the PKC ζ antisense was 5′-AUGAUCAGAUCUAUGCCAUGA-3′. The siRNA expression construct was cotransfected into UPS cells along with a BSEP-promoter luciferase construct and FIC1 expression plasmid as a measure of FIC1 activity. A scrambled RNA antisense construct was used as a negative control. FXR was silenced with a previously described construct.35

A fragment of human farnesoid X receptor (hFXR; accession number U68233), inserted into pCMX (resulting in pCMX-hFXR),24 was used in transfections to express FXR in CV-1 cells. Mutant pCMX-hFXR constructs were generated with the QuickChange mutagenesis kit (Stratagene). The T442A mutant was synthesized by a point mutation, ACA → GCA, with the following forward primer: 5′-cgcctgactgaattacgggcattcaatcatcacc-3′ (mutated nucleotides in bold italics). The T442E mutant was synthesized by a point mutation from the T442 A mutant, ACA → GAA, with the following forward primer: 5′-cgcctgactgaattacgggaattcaatcatcacc-3′.

Transient Transfection Assays.

The procedures for transient transfection and luciferase analysis of cells were performed as described previously.36 Confluent cells (5×106) were transfected with 3 μg of human BSEP-luciferase reporter and 0.3 μg of a control plasmid (pRL-TK; Promega). Luciferase activities were determined by the dual luciferase reporter assay system (Promega) with a Turner 20/20 luminometer (Turner BioSystems). All transfections were performed in triplicate and repeated in two separate sets of experiments.

FIC1 Anti-Fusion Protein Antibodies.

A fusion protein of human FIC1 (amino acids 748-851) coupled to maltose binding protein was generated by amplification of the corresponding cDNA from human intestinal cDNA and ligation into the vector pmalE C2 (New England Biolabs, Beverly, MA). Anti-FIC1 antibodies were generated in rabbits (Covance Research Products, Denver, PA). Antibodies were assessed by western blotting of protein extracts from cells shown to either express or not express FIC1.

Protein Kinase Inhibitor Treatments.

Staurosporine (100 nM),37, 38 bisindolylmaleimide I (10 μM),39, 40 wortmannin (25 nM),41 protein kinase C zeta pseudosubstrate (PKC ζ PS) inhibitor (myristoylated; 100 μM),39, 40 and Gö6976 (10 μM)39 were all purchased from EMD Biosciences, Inc. (San Diego, CA). These inhibitors were added at the stated concentrations to the cells 38 hours after transfection, and the cells were harvested 2 hours after the treatment was added to the medium.

Generation of Green Fluorescent Protein–Fused Human Farnesoid X Receptor (hFXR-GFP).

The full-length wild-type hFXR cDNA was inserted downstream in frame into the Bgl II and Xba I sites of an ecdysone-inducible expression vector, pIND (Invitrogen, Carlsbad, CA), containing the expression sequence of a green fluorescent protein (GFP) to produce the GFP-fused hFXR plasmid construct.

hFXR-GFP Transient Transfection and Fluorescence Microscopy.

HEK-293 cells were transiently transfected with hFXR-GFP cDNA with the Lipofectamine reagent (Gibco/BRL) as describe previously.29 Briefly, cells were cotransfected for 3 hours with 0.5 μg/well of both hFXR-GFP and pVgRXR (Invitrogen) with Lipofectin at a DNA/lipid ratio of 1:3. After 48 hours of growth in serum-containing media, cells were treated for 24 hours with an inducer, 5 μM ponasterone A (Invitrogen; dissolved in 100% EtOH). Cells were then untreated or treated with 100 μM PKC ζ PS inhibitor for 4 hours before they were fixed for fluorescent microscopy of confluent cells.

Protein Preparation, Immunoprecipitation, and Western Blotting.

Cytoplasmic and nuclear extracts were prepared from Caco-2 and/or UPS cells cotransfected with FIC1 and/or siPKC ζ expression constructs with NE-PER (Pierce, Rockford, IL). Control untransfected cells were left in minimum essential medium. The immunoprecipitation method used was described by Anderson and Blobel.42 Total nuclear, cytoplasmic, or FXR antibody immunoprecipitated proteins were analyzed by western blot analysis.36 FIC1 and FXR proteins were detected with the antibodies described previously, whereas the immunoprecipitated (IP) FXR was blotted with an anti-phosphothreonine antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The sample loading was examined with a mouse monoclonal anti-actin or anti-histone H1 antibody (Santa Cruz Biotechnology).

PKC ζ Assay.

Whole cell extracts were prepared from 5 × 107 UPS cells that were transfected with a wild-type FXR construct and siPKC ζ. FXR was immunoprecipitated and then treated directly with either purified recombinant PKC ζ (Upstate, Temecula, CA) or bovine serum albumin as a negative control. Dephosphorylated myelin basic protein was used as a positive control for PKC ζ–mediated phosphorylation. Phosphorylation of either FXR or myelin basic protein was assessed by western blot analysis with a phosphothreonine antibody.

Statistical Analysis.

Data were expressed as the means ± standard deviation from at least six experimental measurements. Differences between experimental groups were evaluated for statistical significance with the Student t test; P < 0.05 was considered to be statistically significant (GraphPad Software, Inc., San Diego, CA).

Results

Gain-of-Function Studies in UPS Cells.

UPS cells were selected because of their defective nonendocytic uptake of fluorescent phosphatidylserine analogs.11 UPS cells lack endogenous expression of FIC1, unlike the parent CHO cell line.8, 11 Western blotting using the FIC1 anti-fusion antibody (Fig. 1 M) detected a 140-kDa protein in wild-type Caco-2 cells, wild-type CHO cells, and UPS cells transfected with the wild-type, G308V Byler, and δ795-7 BRIC FIC1 constructs. The FIC1 antibody did not detect antigen in FIC1 antisense–treated Caco-2 cells or in untransfected UPS cells. Preincubation of the anti-fusion FIC1 antibody with the glutathione S-transferase–FIC1 fusion peptide abolished the antibody reactivity to FIC1 protein (data not shown). A comparison of the anti-fusion FIC1 antibody with the antibody previously described by Ujhazy et al.8 revealed that proteins of identical mobility were detected by both antibodies (data not shown).

Figure 1.

Western blot of FIC1 protein expression in whole cellular extracts. FIC1 protein expression was detected by anti-fusion FIC1 antibody generated from the immunization of rabbits as described in the Materials and Methods section. FIC1 protein was endogenously expressed in Caco-2 cells (lane 1), in CHO cells (wild-type CHO in lane 3), and in UPS cells transfected with wild-type FIC1 (lane 5), with G308V, a Byler's construct (lane 6), and with δ795-7, a benign recurrent intrahepatic cholestasis construct (lane 7). FIC1 protein expression was not found in Caco-2 cells treated with FIC1 antisense (lane 2) or in untransfected UPS cells (lane 4). CHO, Chinese hamster ovary; FIC1, familial intrahepatic cholestasis 1.

The human BSEP promoter is activated by FXR, and its activity is directly relevant to bile flow.29, 43, 44 As such, a human BSEP promoter fragment linked to a luciferase reporter gene (p-145/Luc) was used as a readout of FXR activity. This BSEP promoter is known to be positively regulated by FXR and contains one FXR cis-element (IR-1 at bp −50/−63).29 BSEP promoter activity was diminished by 72% in UPS cells in comparison with the parent CHO cells (UPS, 100% ± 8%; CHO, 384% ± 22%; P < 0.0001). FXR and short heterodimer partner (SHP) protein expression were reduced in UPS cells compared to CHO cells (Fig. 2). FIC1 transfection of UPS cells led to a 6-fold activation of the BSEP promoter, which was abrogated when the FXR-response element was mutated (Supplementary Table 1). None of the Byler disease constructs activated BSEP, whereas the BRIC disease constructs yielded a 2-fold activation (Supplementary Table 1 and Fig. 3A). Equivalent amounts of the appropriately sized FIC1 protein were generated by each of the mutant constructs (Fig. 3B). FXR and SHP protein expression was increased in cells transfected with wild-type FIC1, whereas minimal induction of these proteins was seen after transfection with the Byler disease constructs (Fig. 3B). Intermediate FXR and SHP protein induction was observed after transfection with the BRIC disease constructs. The effects of FIC1 were dependent on the presence of either 10% FCS or the addition of 50 μM CDCA to CTFCS (Fig. 3C). Total bile acid concentration in FCS was 6.7 μM, whereas it was 0.7 μM in CTFCS (Supplementary Table 2).

Figure 2.

Western blot of FXR signaling in whole cellular extracts. Three sets of UPS or CHO cells were analyzed by western blotting of total cellular proteins. FIC1 protein was repeatedly not detectable in UPS cells. FXR and SHP protein expression was reduced in UPS cells with respect to CHO cells. The actin signal shows equivalent loading. C, Chinese hamster ovary (CHO); FIC1, familial intrahepatic cholestasis 1; FXR, farnesoid X receptor; SHP, short heterodimer partner; U, UPS.

Figure 3.

(A) BSEP promoter activity in UPS-transfected cells. UPS cells were transfected with wt FIC1, one of three Byler constructs (G308V, D554N, and G1040R), or one of three BRIC constructs (R600W, ∂795-7, and R600Q). Six-fold activation of the BSEP promoter was observed with wt FIC1 in comparison with untransfected UPS cells. None of the Byler disease constructs activated the BSEP promoter, whereas all the BRIC disease constructs yielded a 2-fold activation (mean of 236%). The Renilla activity was the same in all transfected cells (not shown). (B) Western analysis of protein expression after FIC1 transfection. The equivalent expression and size of the translated FIC1 protein were observed. FXR and SHP expression was minimal in the untransfected cells and those cells transfected with Byler disease constructs. Transfection with the BRIC disease constructs yielded a level of expression of FXR and SHP that was intermediate between the Byler disease and wt constructs. (C) Analysis of the effect of bile acids on the activation of BSEP promoter activity by FIC1 in UPS cells. BSEP promoter activity was minimal in the cells cultured in 0.5% CTFCS. Promoter activity for the various FIC1 constructs was markedly increased in the presence of either 10% FCS or supplementation with 50 μM CDCA. BRIC, benign recurrent intrahepatic cholestasis; BSEP, bile salt export pump; CDCA, chenodeoxycholic acid; CTFCS, charcoal-treated fetal calf serum; FCS, fetal calf serum; FIC1, familial intrahepatic cholestasis 1; FXR, farnesoid X receptor; SHP, short heterodimer partner; wt, wild type.

Effect of Protein Kinase Inhibitors on an FIC1-Mediated Increase of BSEP Promoter Activity in UPS Cells.

Protein kinase and PKC inhibitors were tested on UPS cells cotransfected with FIC1 plasmid and the BSEP promoter as a readout of FXR activity. All results are expressed with respect to BSEP promoter activity in untreated UPS cells (Table 1). FIC1 activation of BSEP was inhibited by wortmannin, staurosporine, and bisindolylmaleimide I but not Gö 6976. PKC ζ PS, a more specific inhibitor of PKC ζ, also inhibited FIC1 stimulation of BSEP activity (Table 1).

siPKC ζ Effects on FIC1-Mediated Induction of BSEP Promoter Activity in UPS Cells.

To determine the specific role of PKC ζ, an siRNA hamster PKC ζ expression construct was synthesized with the hamster PKC ζ as a template. PKC ζ messenger RNA was knocked down by 87% in UPS cells after siRNA treatment (northern analysis of PKC ζ messenger RNA levels: control, 9400 ± 300; siRNA, 1000 ± 100; and scrambled siRNA, 9400 ± 100; P < 0.0001 for siRNA versus either control or scrambled). siPKC ζ treatment yielded a reduction in basal BSEP promoter activity, and FIC1 induction of BSEP activity was abrogated (Table 1).

Table 1. Effect of Protein Kinase Inhibition on FIC1-Mediated Activation of the BSEP Promoter
+FIC1InhibitorSubstrateBSEP Promoter Activity (%)
MeanStandard Deviation
  1. UPS cells were transfected with human BSEP promoter as a reporter for farnesoid X receptor activity. The activity is reported with respect to BSEP promoter activity in UPS cells without cotransfection with the FIC1 expression construct (row 1). Various protein kinase inhibitors and the substrates of the inhibitors are displayed. The Renilla activity was the same in all transfected cells (not shown).

  2. Abbreviations: BSEP, bile salt export pump; FIC1, familial intrahepatic cholestasis 1; PI3K, phosphatidyl inositol 3-kinase; PKC, protein kinase C; siRNA, small interfering RNA.

NoNoneNone10011
YesNoneNone56074
YesWortmanninPI3K10312
YesStaurosporineGeneral protein kinase939
YesBisindolylmaleimide IPKC9710
YesGö 6976Classical PKC57231
YesMyristoylated PKC ζ pseudosubstratePKC ζ9144
YesPKC ζ siRNAPKC ζ1125
YesScrambled siRNANone60539

Effect of PKC ζ Inhibition on Human ASBT Promoter Activity.

siRNA-mediated silencing and pseudosubstrate inhibition led to a marked increase in human ASBT promoter activity in Caco-2 cells (Supplementary Table 3). Promoter activity was also enhanced after siFIC1 treatment. Basal activity of the retinoic acid receptor mutant promoter was reduced, and this promoter construct was not activated after inhibition of PKC ζ or silencing of FIC1.

Effects of FIC1 and PKC ζ Knockdown of FXR Phosphorylation and Localization.

Under basal conditions, which included the bile acids found in FCS, total FXR and phosphorylated FXR were distributed between the cytoplasm and nucleus (Fig. 4A). FIC1 overexpression led to an increase in FXR phosphorylation and nuclear translocation. In contrast, FIC1 overexpression in the context of PKC ζ knockdown showed a marked decrease in FXR protein phosphorylation and nuclear localization (Fig. 4A).

Figure 4.

(A) Effect of siPKC ζ on nuclear localization and phosphorylation of FXR in UPS cells. UPS cells were transfected with a human FIC1 expression construct and an siPKC ζ antisense construct. Cytosolic and nuclear proteins were analyzed by western blotting. Actin and histone H1 antibodies served as loading controls for cytosolic and nuclear proteins, respectively. In the basal state (FIC1 −, siPKC ζ −), FXR was found in both the cytosol and nucleus. FIC1 transfection (FIC1 +, siPKC ζ −) led to nuclear localization of FXR, whereas siPKC ζ treatment of the FIC1-transfected cells (FIC1 +, siPKC ζ +) blocked translocation to the nucleus. FXR antibodies were used to immunoprecipitate FXR, which was then blotted for phosphothreonine (P-FXR). FXR phosphorylation was increased after FIC1 transfection, which was blocked by the siPKC ζ cotreatment. (B,C) Effects of PKC ζ pseudosubstrate inhibitor treatment on the cellular distribution of FXR-GFP in transfected HEK-293 cells. HEK-293 cells were grown on glass coverslips and transiently transfected with hFXR-GFP complementary DNA. The transfected HEK-293 cells were (B) untreated or (C) treated with 100 μM PKC ζ pseudosubstrate inhibitor for 4 hours before they were fixed for fluorescent microscopy. The fluorescent pictures were merged from blue (4′,6-diamidino-2-phenylindole–stained nucleus) and green (FXR-GFP) images. FIC1, familial intrahepatic cholestasis 1; FXR, farnesoid X receptor; FXR-GFP, green fluorescent protein–fused farnesoid X receptor; GFP, green fluorescent protein; hFXR-GFP, green fluorescent protein–fused human farnesoid X receptor; P-FXR, phosphorylated farnesoid X receptor; PKC, protein kinase C.

These findings were confirmed by the transfection of a GFP-tagged FXR protein expression construct into the HEK-293 cell and the treatment of these cells with or without the PKC ζ PS inhibitor. As shown in Fig. 4B, in the cells untreated with the PKC ζ inhibitor, the vast majority of the GFP-tagged FXR was found in the nucleus. In contrast, the addition of PKC ζ PS inhibitor redistributed the FXR to the cytoplasm (Fig. 4C). PKC ζ PS inhibitor treatment did not lead to an alteration in histone H1 distribution (data not shown). Treatment of immunopurified FXR with recombinant PKC ζ resulted in threonine-based phosphorylation (Fig. 5).

Figure 5.

PKC ζ treatment of immunopurified FXR. FXR was immunopurified from UPS cells that were transfected with both wild-type FXR (to enhance expression) and siPKC ζ (to minimize endogenous phosphorylation). FXR or MBP was treated with PKC ζ (+) or bovine serum albumin (−) and then was analyzed by western blot with a phosphothreonine antibody. PKC ζ but not albumin treatment yielded P-FXR or P-MBP. FXR, farnesoid X receptor; MBP, myelin basic protein; P-FXR, phosphorylated farnesoid X receptor; P-MBP, myelin basic protein; PKC, protein kinase C.

Effects of T442 Mutations on FXR Function and Localization.

Site-directed mutagenesis was performed to assess the importance of FXR phosphorylation. Studies were performed in CV-1 cells, which lack endogenous FXR, and the readout of FXR activity was assessed with the BSEP promoter. T442, one of the predicted threonine phosphorylation sites, was chosen for the initial analysis based on the Scansite 2.0 algorithm (http://scansite.mit.edu). Alanine substitution was used to prevent phosphorylation, whereas glutamate was used to simulate phosphorylation. T442A was associated with diminished baseline activity, whereas T442E was associated with enhanced activation of the BSEP promoter (Table 2). Activation of BSEP by T442E was dependent on the presence of CDCA as a ligand. The wild-type and two mutant FXR proteins were equally expressed in the CV-1 cells (Supplementary Fig. 1). Cellular localization of the mutated FXR constructs was performed by western blotting of cytoplasmic or nuclear proteins. The T442A mutant construct was localized to either the cytosol or nucleus in proportions similar to wild-type FXR (Supplementary Fig. 2). In contrast, the T442E mutant was localized exclusively to the nucleus, as might be predicted if phosphorylation of this residue controlled nuclear localization (Fig. 6).

Figure 6.

Cellular localization of T442E FXR. CV-1 cells were untransfected (n), transfected with a wild-type FXR (wt), or transfected with a T442E FXR mutant (m+). FXR was found in both the cytoplasmic and nuclear extracts in wild-type transfected cells, whereas FXR was found only in the nucleus in the T442E-transfected cells. FXR, farnesoid X receptor.

Discussion

Prior loss-of-function studies using siRNA-mediated knockdown of FIC1 in Caco-2 cells demonstrated diminished FXR function that was associated with a reduction of its nuclear localization.13 Overexpression of FXR did not rescue the effect, and this suggested that the effect was not simply based on FXR abundance but instead involved posttranslational modification. The current studies extend these observations to a gain-of-function model, in which exogenous overexpression of FIC1 results in enhanced FXR activity that is associated with its increased phosphorylation and nuclear localization. FXR-mediated downstream activation of the BSEP promoter is absent in cells lacking FIC1 (UPS), cells transfected with FIC1 constructs that harbor Byler-type mutations, and UPS cells cotransfected with wild-type FIC1 and an FXR-response element mutated BSEP promoter. FXR activity is present in cells that endogenously express FIC1 and is reconstituted after FIC1 transfection. Interestingly, the BRIC-type mutant partially activated BSEP, and this indicated diminished but not absent functional activity of the FIC1 gene in BRIC. This novel functional assay of FIC1 activity corresponds with previously reported FIC1 disease phenotypes.1, 4

The effects of FIC1 on FXR function lead to a plausible hypothesis of the pathophysiology of FIC1-related liver disease. Severe cholestasis may result from a combination of diminished canalicular excretion and enhanced intestinal reabsorption of bile acids.13 FXR transcriptionally activates BSEP, so FIC1 deficiency would be expected to lead to reduced canalicular bile acid excretion. ASBT (SLC10A2) expression was increased in human ileal samples from patients with FIC1 disease.13 Negative feedback regulation of human ASBT by bile acids is mediated via the FXR-SHP cascade on the retinoic acid receptor response element.33 Thus, in FIC1 deficiency, there is reduced inhibition of ASBT resulting in induction of its expression and enhanced intestinal reclamation of bile acids. The more “benign” disease seen in BRIC appears to be related to incomplete inactivation of FIC1. The reason for the periodic nature of BRIC remains unknown, although this novel model system may be useful in examining the effects of exogenous factors on FIC1 activity.

FIC1 has been proposed and experimentally shown to function as an aminophospholipid flippase/transporter.1, 8, 12 Because phosphatidylserine may activate protein kinases, initial experiments focused on this signaling pathway.45, 46 PKC ζ is of particular interest because the atypical protein kinases can be activated directly by phosphatidylserine. PKC ζ is expressed in the liver and has been implicated in mediating other bile acid–related processes.39, 47, 48 PKC ζ is downstream of phosphatidyl inositol 3-kinase (PI3K), and its activation depends on PI3K products49 and phosphorylation by phosphoinositide-dependent protein kinase 1 followed by autophosphorylation in some cells.50 Wortmannin (a general PI3K inhibitor), staurosporine (a general protein kinase inhibitor), and bisindolylmaleimide I (a PKC inhibitor) all abrogated the activation of BSEP by FIC1, implicating the aforementioned signaling pathway. These inhibitors also reduced basal BSEP promoter activity, and this suggested an element of basal induction of BSEP through this pathway. Gö 6976, a classical PKC inhibitor, had no effect on either basal activity or FIC1-mediated induction of BSEP, and this suggested involvement of an atypical PKC. The specificity of the response via PKC ζ was determined with a PKC ζ PS and siRNA-mediated silencing, both of which reduced basal BSEP promoter activity and eliminated activation by FIC1. In light of known nonspecific effects of PKC inhibitors, the siRNA studies provide the strongest specific support for the role of PKC ζ in mediating the FIC1 effect. Pseudosubstrate inhibition of PKC ζ markedly reduced nuclear targeting of a GFP-tagged FXR molecule.

Human ASBT is under negative feedback regulation by bile acids primarily mediated by the FXR-SHP cascade acting on the retinoic acid receptor.33 Inhibition of PKC ζ in Caco-2 cells leads to activation of the human ASBT promoter in a manner akin to that observed with FIC1 silencing.13 This activation is dependent on transcriptional effects mediated by the retinoic acid receptor cis-element.

Recombinant PKC ζ directly phosphorylated FXR, and this supported the potential functional importance of FXR phosphorylation in mediating the FIC1 effect. FIC1 activity, manipulated either by knockdown or exogenous overexpression, correlates positively with FXR phosphorylation and nuclear localization. There are multiple potential phosphorylation sites in the FXR protein. Initial studies focused on threonine sites as the immunoprecipitation studies were positive with a phosphothreonine antibody. The importance of FXR phosphorylation was modeled by site-directed mutagenesis of FXR at one of the predicted threonine phosphorylation sites, T442. Alanine substitution, which prevents phosphorylation, yielded a dominant negative protein. This protein could still translocate to the nucleus, and this was presumably related to phosphorylation of other sites in FXR (for example, T219; T.F. et al., unpublished data, 2008). Glutamate substitution yielding a phosphomimetic amino acid change generated an activated FXR molecule with accentuated nuclear localization. Activation was dependent on the presence of the FXR ligand, although the exact mechanism of activation deserves further study. T442 is conserved among the rat, mouse, and human FXR genes and has an intriguing pattern of conservation and substitution among other nuclear receptors.51 Its position in helix 11 underscores the potential critical relevance of this residue in conferring agonist or antagonist properties upon FXR.51–54

In summary, these studies suggest that FIC1 function leads to activation of PKC ζ with subsequent phosphorylation, nuclear localization, and activation of FXR. Future studies are needed to determine if there is a direct link between the aminophospholipid flippase activity of FIC1 and the observed changes in the PKC ζ activity and the phosphorylation status of FXR. Effects on downstream targets of FXR, including the canalicular bile salt transporter and ASBT, may underlie the cholestasis observed in individuals with FIC1 deficiency. Milder FIC1-related disease is associated with partial functionality of this protein. Posttranslational modifications may be an important mechanism for regulation of FXR function.

Ancillary