Taurolithocholate-induced MRP2 retrieval involves MARCKS phosphorylation by protein kinase Cϵ in HUH-NTCP Cells

Authors

  • Christopher M. Schonhoff,

    1. Departments of Biomedical Sciences, Tufts Cummings School of Veterinary Medicine, North Grafton, MA
    Search for more papers by this author
  • Cynthia R. L. Webster,

    1. Departments of Clinical Sciences, Tufts Cummings School of Veterinary Medicine, North Grafton, MA
    Search for more papers by this author
  • M. Sawkat Anwer

    Corresponding author
    1. Departments of Biomedical Sciences, Tufts Cummings School of Veterinary Medicine, North Grafton, MA
    • Department of Biomedical Sciences, Tufts Cummings School of Veterinary Medicine, 200 Westboro Road, North Grafton, MA 01536===

    Search for more papers by this author
    • fax: 508-839-8787


  • Potential conflict of interest: Nothing to report.

  • This study was supported in part by National Institutes of Health grants DK-33436 and DK-90010 (to M. Sawkat Anwer) and DK-65975 (to Cynthia R. L. Webster).

Abstract

Taurolithocholate (TLC) acutely inhibits the biliary excretion of multidrug-resistant associated protein 2 (Mrp2) substrates by inducing Mrp2 retrieval from the canalicular membrane, whereas cyclic adenosine monophosphate (cAMP) increases plasma membrane (PM)–MRP2. The effect of TLC may be mediated via protein kinase Cϵ (PKCϵ). Myristoylated alanine-rich C kinase substrate (MARCKS) is a membrane-bound F-actin crosslinking protein and is phosphorylated by PKCs. MARCKS phosphorylation has been implicated in endocytosis, and the underlying mechanism appears to be the detachment of phosphorylated myristoylated alanine-rich C kinase substrate (pMARCKS) from the membrane. The aim of the present study was to test the hypothesis that TLC-induced MRP2 retrieval involves PKCϵ-mediated MARCKS phosphorylation. Studies were conducted in HuH7 cells stably transfected with sodium taurocholate cotransporting polypeptide (HuH-NTCP cells) and in rat hepatocytes. TLC increased PM–PKCϵ and decreased PM-MRP2 in both HuH-NTCP cells and hepatocytes. cAMP did not affect PM-PKCϵ and increased PM-MRP2 in these cells. In HuH-NTCP cells, dominant-negative (DN) PKCϵ reversed TLC-induced decreases in PM-MRP2 without affecting cAMP-induced increases in PM-MRP2. TLC, but not cAMP, increased MARCKS phosphorylation in HuH-NTCP cells and hepatocytes. TLC and phorbol myristate acetate increased cytosolic pMARCKS and decreased PM-MARCKS in HuH-NTCP cells. TLC failed to increase MARCKS phosphorylation in HuH-NTCP cells transfected with DN-PKCϵ, and this suggested PKCϵ-mediated phosphorylation of MARCKS by TLC. In HuH-NTCP cells transfected with phosphorylation-deficient MARCKS, TLC failed to increase MARCKS phosphorylation or decrease PM-MRP2. Conclusion: Taken together, these results support the hypothesis that TLC-induced MRP2 retrieval involves TLC-mediated activation of PKCϵ followed by MARCKS phosphorylation and consequent detachment of MARCKS from the membrane. (HEPATOLOGY 2013;)

Multidrug-resistant associated protein 2 (MRP2; adenosine triphosphate–binding cassette C2), an adenosine triphosphate–binding transporter located at the canalicular membrane of hepatocytes, is involved in the biliary secretion of conjugated endogenous and exogenous organic anions.1, 2 MRP2 has been shown to undergo both transcriptional and posttranslational regulation in cholestasis. For example, the transcription of MRP2 is down-regulated in rodent models of cholestasis3 and during liver regeneration.4 Cholestatic agents such as taurolithocholate (TLC)5 and estradiol-17β-glucuronide (E217G)6 induce the retrieval of MRP2 from the canalicular membrane. More recent studies suggest that protein kinase Cs (PKCs) may be involved in the retrieval of MRP2 by TLC and E217G. On the basis of studies with chemical inhibitors, it has been proposed that the effect of E217G may be mediated via classic PKC-induced endocytosis7 and the phosphoinositide 3-kinase/Akt signaling pathway.8 Similarly, the TLC-induced retrieval of Mrp2 has been suggested to be mediated via a phosphoinositide 3-kinase- and PKCϵ-dependent mechanism.9, 10 However, the role of PKCϵ in TLC-induced MRP2 retrieval has not been directly evaluated. Moreover, signaling pathways by which PKCϵ may induce MRP2 retrieval have not been investigated.

Abbreviations

cAMP, cyclic adenosine monophosphate; CON, control; CPT, 8-(4-chlorophenylthio); DN, dominant-negative; E217G, estradiol-17β-glucuronide; E-Cad, E-cadherin; EV, empty vector; GFP, green fluorescent protein; HA, hemagglutinin; HuH-NTCP cell, HuH7 cell stably transfected with sodium taurocholate cotransporting polypeptide; MARCKS, myristoylated alanine-rich C kinase substrate; MRP2, multidrug-resistant associated protein 2; PD, phosphorylation-deficient; PKC, protein kinase C; PM, plasma membrane; PMA, phorbol myristate acetate; pMARCKS, phosphorylated myristoylated alanine-rich C kinase substrate; TLC, taurolithocholate; WT, wild type.

PKCs mediate effects by phosphorylating their substrates. Myristoylated alanine-rich C kinase substrate (MARCKS) is one such substrate and plays a key role in cytoskeletal dynamics.11, 12 MARCKS is an F-actin crosslinking protein and is phosphorylated by cPKCα, PKCδ, and PKCϵ in vitro.13, 14 Phosphorylation of MARCKS by PKCδ and PKCϵ has been shown to be involved in exocytosis and endocytosis in nonhepatic cells. Thus, MARCKS phosphorylation by PKCδ is involved in airway mucin secretion15, 16 and gut peptide secretion.17 MARCKS phosphorylation by PKCϵ has been shown to stimulate vesicle translocation in chromaffin cells18 and basolateral fluid-phase endocytosis in T84 cells.19 Phosphorylation of MARCKS by PKCs results in the retrieval of MARCKS from the plasma membrane (PM) to the cytosol and in F-actin disassembly.18 It may be noted that actin plays an important role in hepatobiliary transporter translocation20-22 and that TLC induces F-actin accumulation around bile canaliculi.23 Phosphorylation of MARCKS by PKCs requires the translocation of PKCs to MARCKS located in the PM, and as a result, MARCKS phosphorylation and the consequent effect are dependent on subcellular targeting of PKC.24, 25 These studies raise the possibility that TLC-induced endocytic retrieval of Mrp2 may result from PKCϵ-dependent MARCKS phosphorylation.

In the present study, we determined whether TLC-induced MRP2 retrieval is mediated via PKCϵ and whether the effect of PKCϵ is mediated via MARCKS phosphorylation. The results of our studies with dominant-negative (DN)–PKCϵ and phosphorylation-deficient (PD)–MARCKS in HuH7 cells stably transfected with sodium taurocholate cotransporting polypeptide (HuH-NTCP cells) are consistent with the following signaling pathway: TLC → PKCϵ → MARCKS phosphorylation → MRP2 retrieval.

Materials and Methods

Materials.

8-(4-Chlorophenylthio)–cyclic adenosine monophosphate (CPT-cAMP), wortmannin, and the antibody for human MRP2 were purchased from Sigma-Aldrich (St. Louis, MO). The commercial sources of other antibodies were Cell Signaling [phosphorylated myristoylated alanine-rich C kinase substrate (pMARCKS) and hemagglutinin (HA)], Calbiochem (actin), Clontech [green fluorescent protein (GFP)], Upstate (PKCϵ), and BD Transduction Laboratories (E-cadherin). Sulfosuccinimidyl-6-(biotin-amido)hexanoate was purchased from Pierce (Rockford, IL). Streptavidin beads were purchased from Novagen (Madison, WI). Lipofectamine 2000 was obtained from Invitrogen (Carlsbad, CA). Plasmid constructs for wild-type (WT)–MARCKS and PD-MARCKS (with the effector domain phosphorylation sites at S152, S156, and S163 replaced by alanine) were kind gifts from Dr. Saito.26 Kinase-dead DN-PKCϵ plasmids were purchased from Addgene (Cambridge, MA). HuH-NTCP cells were generously provided by Dr. Gores.27

Rat Hepatocyte Preparation.

Rat hepatocytes were isolated from male Wistar rats (200-250 g) and cultured as previously described,28 and they were used to determine the effect of TLC on PKCϵ, Mrp2, and the phosphorylation of MARCKS. Male Wistar rats were obtained from Charles River Laboratories and the protocol for harvesting livers was approved by the Institutional Animal Care and Use Committee.

HuH-NTCP Cell Culture and Transfections.

HuH-NTCP cells were cultured in Eagle's minimum essential medium supplemented with 10% fetal bovine serum, 1.2 g/L G418, 100,000 U/L penicillin, 100 mg/L streptomycin, and 25 μg/mL amphotericin B at 37°C in a 5% CO2 and 95% O2 air incubator. For transfection experiments involving DN-PKCϵ, WT-MARCKS, and PD-MARCKS, the cells were cultured in six-well plates for 24 hours and then transiently transfected with 1 to 3 μg of the desired plasmid with Lipofectamine as previously described.29 After 24 hours of incubation in the transfection medium, the cells were cultured for an additional 24 hours in a culture medium. The expression of these plasmids was confirmed by immunoblot analysis with anti-HA (for PKCϵ) and anti-GFP antibodies (for WT-MARCKS and PD-MARCKS). Cells were transfected at 70% to 80% confluence, and nontransfected cells were at 80% to 90% confluence before treatments. For all experiments, cells were then incubated in serum-free media for 3 hours at 37°C before treatments (as described in the figure legends).

PM-MRP2 and PM-PKCϵ.

As previously described by us,20, 30-32 a cell surface protein biotinylation method was used to assess MRP2 and PKCϵ translocation to PMs. Briefly, after various treatments, cell surface proteins were biotinylated by the exposure of hepatocytes to sulfosuccinimidyl-6-(biotin-amido)hexanoate and then the preparation of a whole cell lysate. Biotinylated proteins were isolated with streptavidin-agarose beads and then subjected to immunoblot analysis to determine PM-MRP2, PKCϵ, and E-cadherin. The amounts of MRP2 and PKCϵ present in the PM were expressed as relative values versus E-cadherin (a PM protein), which was used as a loading control.

Other Methods.

Phosphorylation of MARCKS was determined with a pMARCKS (Ser152/156) antibody. The Lowry method33 was used to determine cell proteins. The blots were scanned with Adobe Photoshop (Adobe Systems, Inc., San Jose, CA), and the relative band densities were quantitated with Sigma Gel (Jandel Scientific Software, San Rafael, CA). All values were expressed as means and standard errors. An analysis of variance followed by Fisher's least significant difference test was used to statistically analyze the data, with P < 0.05 considered significant.

Results

TLC-Induced MRP2 Internalization Is Mediated Via PKCϵ.

TLC has been shown to activate PKCϵ9, 10 and internalize Mrp2 in rat hepatocytes.5 This was further confirmed by our studies showing that TLC, but not cAMP, increased PM-PKCϵ in rat hepatocytes (Supporting Fig. 1). Furthermore, TLC decreased and cAMP increased PM-MRP2 in rat hepatocytes (Supporting Fig. 2). In the present study, we tested the hypothesis that this effect of TLC is mediated via PKCϵ with DN-PKCϵ in HuH-NTCP cells, which constitutively express MRP2.32 To ascertain that HuH-NTCP cells constitute a valid model, we first determined whether TLC activates PKCϵ and internalizes MRP2 in this cell line. To determine the effects of PKCϵ and MRP2, cells were treated with TLC for 15 or 25 minutes, respectively. These time points are based on previous studies reporting the effects of TLC on PKCϵ activation in HuH-NTCP cells34 and biliary excretion of the Mrp2 substrate in perfused rat livers.5 TLC increased PM translocation of PKCϵ and decreased PM-MRP2 in HuH-NTCP cells (Fig. 1). Phorbol myristate acetate (PMA), used as a positive control, also increased PM-PKCϵ. cAMP, used as a negative control, did not affect PM-PKCϵ; cAMP does not activate PKCϵ in rat hepatocytes.31 cAMP also increased PM-MRP2 in HuH-NTCP cells (Fig. 1). Thus, HuH-NTCP cells were considered a valid model for studying the role of PKCϵ in TLC-induced MRP2 internalization.

Figure 1.

TLC induces PM translocation of PKCϵ and retrieval of MRP2. HuH-NTCP cells were incubated with 10 μM TLC, 1 μM PMA, or 100 μM CPT-cAMP for 15 minutes, and this was followed by the biotinylation of cell surface proteins and an immunoblot analysis of biotinylated PKCϵ (PM-PKCϵ, 88 kDa) and E-cadherin (E-Cad) as a loading control (CON; 135 kDa). For the MRP2 assay, cells were treated with 10 μM TLC (25 minutes) or 100 μM CPT-cAMP (15 minutes), and this was followed by an immunoblot analysis of biotinylated MRP2 (PM-MRP2; 195 kDa) and E-Cad. Typical PM-MRP2, PM-PKCϵ, and E-Cad immunoblots are shown in the upper panel, and the results of a densitometric analysis, presented as means and standard errors of the mean (n = 5), are shown in the bar graph. *Significantly different (P < 0.05) from the respective CON values.

The transfection of HuH-NTCP cells with HA-tagged DN-PKCϵ resulted in the overexpression of total PKCϵ by 2- to 3-fold (Fig. 2). DN-PKCϵ did not affect the basal expression of MRP2 in the PM versus an empty vector. TLC decreased PM expression of MRP2 in cells transfected with an empty vector. However, this effect was reversed in cells transfected with DN-PKCϵ. cAMP, which has been shown to increase PM expression of MRP2 by activating PKCδ,31, 32 was used as a negative control. The ability of cAMP to increase PM-MRP2 was not affected by DN-PKCϵ. These results support the hypothesis that TLC-induced internalization of MRP2 is mediated via PKCϵ and that cAMP-mediated translocation of MRP2 to PM does not involve PKCϵ.

Figure 2.

DN-PKCϵ inhibits TLC-induced MRP2 retrieval. HuH-NTCP cells transfected with HA-tagged DN-PKCϵ were treated with TLC for 25 minutes or CPT-cAMP for 15 minutes, and this was followed by the biotinylation of cell surface proteins and an immunoblot analysis of HA, PKCϵ, biotinylated MRP2 (PM-MRP2), and biotinylated E-cadherin. Typical immunoblots are shown in the upper panel, and the results of a densitometric analysis, presented as means and standard errors of the mean (n = 4), are shown in the bar graph. *Significantly different from the respective control (CON) values and #significantly different from the respective empty vector (EV) values.

TLC Phosphorylates and Translocates MARCKS Into Cytosol.

Because MARCKS is a substrate for PKC and has been implicated in endocytosis,19 it is possible that TLC-induced MRP2 internalization involves TLC/PKCϵ-mediated phosphorylation of MARCKS. To test this hypothesis, we first determined whether TLC can phosphorylate MARCKS. In these studies, actin instead of MARCKS was used as the loading control because the MARCKS antibody gave inconsistent results on stripped blots. A time-dependent study showed that TLC increased MARCKS phosphorylation as early as 5 minutes, with significant phosphorylation observed until 25 minutes (Fig. 3). On the other hand, cAMP, which stimulates MRP2 translocation to the PM, did not phosphorylate MARCKS during the same time period. Similar results were obtained in rat hepatocytes (Fig. 3B), and this indicates that this is not an effect specific to transformed cells. Thus, MARCKS phosphorylation may be involved in MRP2 retrieval and not MRP2 translocation to the membrane.

Figure 3.

TLC phosphorylates MARCKS. (A) HuH-NTCP cells and (B) cultured rat hepatocytes were treated with 10 μM TLC or 100 μM CPT-cAMP for the indicated times, and this was followed by the determination of pMARCKS (80 kDa) and actin as a loading control (CON). Typical immunoblots are shown in the upper panel, and the results of a densitometric analysis, presented as means and standard errors of the mean (n = 5), are shown in the bar graph. *Significantly different (P < 0.05) from the CON values.

One of the consequences of MARCKS phosphorylation is the retrieval of MARCKS from the PM to the cytosol, which results in F-actin disassembly.18 Thus, we determined whether TLC increases cytosolic pMARCKS. TLC increased cytosolic pMARCKS 2.5-fold in comparison with controls (Fig. 4). PMA, used as a positive control, increased cytosolic pMARCKS more than 7-fold. The more pronounced effect of PMA was likely to be due to the activation of other PKCs. The observed increases in cytosolic pMARCKS were associated with decreases in PM-MARCKS (Fig. 4), which indicated the translocation of MARCKS from the membrane to the cytosol after phosphorylation. This result suggests that TLC-induced phosphorylation and the subsequent removal of MARCKS from the PM may be related to MRP2 retrieval by TLC.

Figure 4.

TLC increases cytosolic pMARCKS. HuH-NTCP cells were treated with 10 μM TLC or 1 μM PMA for 15 minutes, and this was followed by the biotinylation of cell surface proteins and an immunoblot analysis of biotinylated PM-MARCKS and E-cadherin or the isolation of cytosol (100,000g supernatant) and determination of pMARCKS. Typical immunoblots are shown in the upper panel, and the results of a densitometric analysis, presented as means and standard errors of the mean (n = 3), are shown in the bar graph. *Significantly different (P < 0.05) from the control values.

TLC-Induced MARCKS Phosphorylation Is Mediated Via PKCϵ.

MARCKS is a substrate for PKCs, and TLC may activate PKCs other than PKCϵ. To determine whether TLC-induced MARCKS phosphorylation is mediated via PKCϵ, we studied the effect of DN-PKCϵ on MARCKS phosphorylation (Fig. 5). DN-PKCϵ did not affect the basal level of MARCKS phosphorylation. TLC increased MARCKS phosphorylation in cells transfected with an empty vector but failed to do so in cells transfected with DN-PKCϵ. The effect of PMA, which was used as a positive control, on MARCKS phosphorylation was also significantly decreased by DN-PKCϵ. The residual MARCKS phosphorylation by PMA was likely due to the activation of other PKCs. As expected, the effect of cAMP, which was used as a negative control, was not affected by DN-PKCϵ. These results are consistent with the hypothesis that TLC-induced MARCKS phosphorylation is mediated via PKCϵ.

Figure 5.

TLC-induced MARCKS phosphorylation is mediated via PKCϵ. HuH-NTCP cells transfected with HA-tagged DN-PKCϵ were treated with 10 μM TLC, 1 μM PMA, or 100 μM CPT-cAMP for 15 minutes, and this was followed by an immunoblot analysis of HA, PKCϵ, pMARCKS, and actin. Typical immunoblots are shown in the upper panel, and the results of a densitometric analysis, presented as means and standard errors of the mean (n = 5), are shown in the bar graph. *Significantly different from the empty vector (EV) control (CON) values, &significantly different from the DN-PKCϵ CON values, and #significantly different from the respective EV values.

PD-MARCKS Inhibits TLC-Induced MRP2 Retrieval.

MARCKS phosphorylation has been implicated in fluid-phase endocytosis in T84 cells.19 Thus, it is possible that MARCKS phosphorylation may be involved in TLC-induced MRP2 retrieval. We tested this hypothesis by determining the effect of TLC on PM-MRP2 in cells transfected with GFP-tagged WT-MARCKS and PD-MARCKS. First, we determined the effect of WT-MARCKS and PD-MARCKS on TLC-induced MARCKS phosphorylation (Fig. 6). Because transfected MARCKS was tagged with GFP (26.9 kDa), we could distinguish between transfected (GFP-MARCKS) and endogenous MARCKS (endo-MARCKS) at the same time when we probed with the MARCKS antibody; GFP-MARCKS (107 kDa) appeared above endo-MARCKS (80 kDa). Transfection with GFP-MARCKS did not affect the level of endogenous MARCKS (Fig. 6), and GFP-MARCKS represented 50% to 80% of total MARCKS (GFP plus endogenous MARCKS). Phosphorylation of GFP-MARCKS was detected in cells transfected with WT-MARCKS. In contrast, no phosphorylation of GFP-MARCKS was detected in cells transfected with PD-MARCKS, and this confirmed the inability of PD-MARCKS to be phosphorylated. TLC increased phosphorylation of endogenous and transfected MARCKS in cells transfected with an empty vector or WT-MARCKS. However, TLC failed to increase phosphorylation of endogenous MARCKS in cells transfected with PD-MARCKS. The ability of PMA to increase MARCKS phosphorylation decreased significantly in cells transfected with PD-MARCKS.

Figure 6.

PD-MARCKS inhibits TLC-induced phosphorylation of endogenous MARCKS. HuH-NTCP cells transfected with empty vector (EV) GFP-tagged WT-MARCKS or PD-MARCKS (107 kDa) were treated with 10 μM TLC or 1 μM PMA for 15 minutes, and this was followed by an immunoblot analysis of MARCKS, pMARCKS, and actin. Typical immunoblots are shown in the upper panel, and the results of a densitometric analysis, presented as means and standard errors of the mean (n = 5), are shown in the bar graph. Control (CON) values for EV, WT-MARCKS, and PD-MARCKS were set at 1. *Significantly different from the respective CON values and #significantly different from the respective EV values.

TLC also decreased PM-MRP2 in cells transfected with an empty vector or WT-MARCKS (Fig. 7). The basal level of PM-MRP2 was not affected in cells transfected with WT-MARCKS. However, PM-MRP2 decreased by 30% in cells transfected with PD-MARCKS, and this raises the possibility that MRP2 may be stabilized in the membrane by unphosphorylated MARCKS (see the Discussion section). In addition, TLC failed to further decrease PM-MRP2 in cells transfected with PD-MARCKS. These results suggest that phosphorylation of MARCKS is necessary for the TLC-induced retrieval of MRP2.

Figure 7.

PD-MARCKS decreases PM-MRP2 and inhibits TLC-induced MRP2 retrieval. HuH-NTCP cells transfected with GFP-tagged WT-MARCKS or PD-MARCKS were treated with 10 μM TLC for 25 minutes, and this was followed by the biotinylation of PM proteins and an immunoblot analysis of PM-MRP2 and E-cadherin. Typical immunoblots are shown in the upper panel, and the results of a densitometric analysis, presented as means and standard errors of the mean (n = 5), are shown in the bar graph. *Significantly different from the respective empty vector (EV) values and #significantly different from the EV control (CON) values.

Discussion

The aim of the present study was to further define the mechanism by which TLC induces the retrieval of MRP2. The present study showed that TLC increased PM localization of PKCϵ, and a kinase-dead DN-PKCϵ inhibited TLC-induced MRP2 retrieval. In addition, DN-PKCϵ inhibited TLC-induced increases in the phosphorylation of MARCKS, and PD-MARCKS inhibited TLC-induced MRP2 retrieval. These results suggest that TLC-induced MRP2 retrieval involves the activation of PKCϵ followed by the phosphorylation of MARCKS, as discussed later.

PKCϵ has been suggested to be involved in TLC-induced cholestasis.9 However, this conclusion is based on indirect evidence. The strongest evidence in favor of this hypothesis is the reversal of TLC-induced membrane translocation of PKCϵ and cholestasis by tauroursodeoxycholate.9 In the present study, we tested this hypothesis more directly by using DN-PKCϵ. As previously reported in rat hepatocytes,5, 10 TLC induced the translocation of PKCϵ to the PM and the retrieval of MRP2 from the PM in HuH-NTCP cells as well as rat hepatocytes. TLC failed to induce MRP2 retrieval when cells were transfected with kinase-dead DN-PKCϵ, and this indicates that the PKCϵ kinase activity is needed for TLC-induced MRP2 retrieval. This is the first direct demonstration of a role for PKCϵ in MRP2 retrieval by TLC.

Our studies also provide evidence for PKCϵ-mediated phosphorylation of MARCKS by TLC. MARCKS is a PKC substrate and binds noncovalently to PM.12 MARCKS phosphorylation leads to its translocation to the cytosol in chromaffin cells.18 A previous study35 reported that PMA translocated MARCKS from the PM to the cytosol in HepG2 cells, and this effect, based on inhibition by chemical inhibitors of PKCs, appeared to be mediated via Ca2+-dependent and Ca2+-independent PKCs. However, whether PMA phosphorylated MARCKS was not determined. In the present study, we observed that TLC induced phosphorylation of MARCKS, increased the cytosolic levels of pMARCKS, and decreased PM-MARCKS. Thus, TLC-mediated phosphorylation of MARCKS results in the dissociation of MARCKS from the membrane. In addition, TLC-induced MARCKS phosphorylation was inhibited in cells transfected with DN-PKCϵ. These results suggest that TLC, acting via PKCϵ, phosphorylates MARCKS and results in the dissociation of MARCKS from the PM.

The present study suggests that MARCKS phosphorylation by PKCϵ is involved in MRP2 retrieval by TLC. This is supported by the fact that TLC failed to induce MRP2 retrieval in cells transfected with PD-MARCKS (Fig. 7). Although the role of MARCKS phosphorylation has been investigated in other cell types, little is known about its effect in hepatocytes. PMA has been shown to phosphorylate and translocate MARCKS to lysosome in rat hepatocytes.36 Studies in most other cell types suggest a role of MARCKS in the exocytosis and exocytotic insertion of membrane proteins. Thus, the phosphorylation of MARCKS has been implicated in neurotransmitter release,37 glucose-induced secretion in isolated rat pancreatic islets,38 the release of adrenocorticotropin in ovine anterior pituitary cells,39 thrombin-induced serotonin release from platelets,40 insulin-induced Glut4 translocation to the PM in rat skeletal muscle cells,41 and mucin secretion in bronchial epithelial cells.15 However, MARCKS phosphorylation, most likely by PKCϵ, has also been suggested to be involved in basolateral fluid-phase endocytosis in T84 cells.19 MARCKS phosphorylation has also been suggested to be involved in an abnormal endocytic pathway in Alzheimer disease.42 On the basis of these studies and the results of the present study, we suggest that MARCKS phosphorylation leads to endocytic retrieval of MRP2 in hepatocytes. To our knowledge, this is the first study implicating MARCKS phosphorylation in membrane transporter retrieval in hepatocytes.

The precise intracellular mechanisms by which MARCKS regulates endocytosis and exocytosis have not been fully elucidated.12, 43 The finding that MARCKS can bind directly to actin and crosslinks it to PM44 has led to the suggestion that actin is essential to the overall functioning of MARCKS. The binding of MARCKS to the membrane requires the electrostatic interaction of basic (serine) residues of MARCKS in its effector domain with acidic lipids of the membrane and the hydrophobic insertion of myristate into the core of the membrane. Both of these interactions are necessary for significant membrane binding.12, 43, 44 When the serine residues in the effector domain of MARCKS are phosphorylated by PKC or replaced by alanine as in PD-MARCKS, the electrostatic interaction between MARCKS and the acidic lipids is abolished, and this results in the dissociation of MARCKS from the membrane. Because of the proximity of MARCKS phosphorylation sites to the actin binding site,45 MARCKS phosphorylation also results in the release of actin and a local softening (disruption) of the actin cytoskeleton with increased plasticity and endocytosis.11, 12 Thus, it can be speculated that by binding and tethering actin, unphosphorylated MARCKS stabilizes MRP2 in the membrane, as has been suggested for other actin crosslinking proteins, radixin46, 47 and Na+/H+ exchanger regulatory factor 1.48 Consistent with this hypothesis is a recent study in rats showing that taurochenodeoxycholate-induced retrieval of MRP2 is associated with changes in the actin cytoskeleton.49 Because three serine residues are replaced by alanine in PD-MARCKS, such a mechanism can also explain decreased PM-MRP2 in cells transfected with PD-MARCKS (Fig. 7), presumably because of the inability of PD-MARCKS to bind the membrane and thereby crosslink actin and result in actin cytoskeletal changes. It should, however, be noted that the endocytic retrieval of a transporter is a complex process requiring the participation of a number of regulatory proteins,42, 50 and MARCKS phosphorylation may also affect these regulators. Thus, further studies are needed to define the mechanism by which MARCKS phosphorylation leads to MRP2 retrieval.

In summary, the results of the present study support the hypothesis that TLC-induced retrieval of MRP2 from PM involves the activation of PKCϵ followed by PKCϵ-mediated phosphorylation of MARCKS. Unlike in most other cell types, MARCKS may be involved in endocytosis in hepatic cells.

Acknowledgements

The authors thank Holly Jameson and Ariel Hobson for their excellent technical assistance.

Ancillary