Search

The bile salt export pump (BSEP, ABCB11) plays an essential role in the formation of bile. In hepatocytes, BSEP is localized within the apical (canalicular) membrane and a deficiency of canalicular BSEP function is associated with severe forms of cholestasis. Regulation of correct trafficking to the canalicular membrane and of activity is essential to ensure BSEP functionality and thus normal bile flow. However, little is known about the identity of interaction partners regulating function and localization of BSEP. In our study, interaction partners of BSEP were identified in a complementary approach: Firstly, BSEP interaction partners were co-immunoprecipitated from human liver samples and identified by mass spectrometry (MS). Secondly, a membrane yeast two-hybrid (MYTH) assay was used to determine protein interaction partners using a human liver cDNA library. A selection of interaction partners identified both by MYTH and MS were verified by in vitro interaction studies using purified proteins. By these complementary approaches, a set of ten novel BSEP interaction partners was identified. With the exception of radixin, all other interaction partners were integral or membrane-associated proteins including proteins of the early secretory pathway and the bile acyl-CoA synthetase, the second to last, ER-associated enzyme of bile salt synthesis.

Bile plays an essential role in the nutrient uptake of the human body. It consists mostly of amphipathic bile salts, phospholipids and cholesterol in mixed micelles. The bile solubilizes lipophilic nutrients, such as lipids and fat-soluble vitamins, and facilitates their uptake. The formation of bile depends on the activity of various ATP-binding cassette (ABC) transporters localized in the apical membrane of hepatocytes. The main components of bile, phosphatidylcholine, cholesterol and bile salts, are transported by multidrug resistance protein 3 (MDR3, ABCB4), ABCG5/G8 and the bile salt export pump (BSEP, ABCB11), respectively. BSEP, as the transporter of the main solute in bile, is considered to be the primary driving force of bile formation. BSEP belongs to the family of ABC transporters, which constitute one of the largest protein families. They are primary active transporters with a common functional unit: two nucleotide–binding domains and two transmembrane domains. The nucleotide–binding domains fuel the transport by hydrolyzing ATP, while the transmembrane domains provide the translocation pathway over the membrane and are thought to be responsible for substrate recognition. Eukaryotic ABC transporters are usually encoded by one gene as in the case of BSEP (full-size transporters) or by two genes coding for one nucleotide-binding domain and one transmembrane domain each. In the latter case the transporter is called a half-size transporter and the two polypeptides homo- or heterodimerize to form a functional unit. An example for such a half-size transporter is ABCG5/G8. Mutations in ABC transporters can lead to severe diseases. For example an impairment of targeting or activity in BSEP leads to the accumulation of bile salts in the hepatocytes. Bile salts, in addition to their detergent function, can furthermore act as signaling molecules and thereby cause different forms of cholestasis, such as progressive familial intrahepatic cholestasis type 2 (PFIC2) or benign recurrent intrahepatic cholestasis type 2 (BRIC2). An additional influence that can affect the development of diseases is the post-translational regulation of ABC transporter functionality. The regulation can occur directly in the form of chemical modifications or can be mediated for example by protein-protein interactions. For several ABC transporters of the apical hepatocyte membrane an influence of protein-protein interactions on targeting, turnover or activity could be shown. MRP2 for example interacts with NHERF-1, a scaffold protein that is responsible for linking its interacting partners to the cytoskeleton. It has furthermore been reported that the interaction of MRP2 with NHERF-1 influences the apical membrane expression of MRP2. For example, NHERF-1 knock–out mice show reduced abundance of MRP2 in the membrane [1]. For BSEP two interaction partners are known to date. One of them is the AP2 adaptor related protein complex, which is involved in internalization of BSEP [2]. To study ABC transporters in their isolated form a heterologous overexpression system is required to obtain sufficient amounts of protein. Initially, toxicity of the BSEP cDNA for Escherichia coli was an obstacle. Therefore a Saccharomyces cerevisiae based cloning technique has been developed, which is independent of E. coli[3]. Subsequently, with the optimized cloning procedure an overexpression of BSEP could be established in the methylotrophic yeast Pichia pastoris. With the help of a GFP-fusion protein BSEP could be traced to the P. pastoris plasma membrane. Additionally, sucrose density centrifugation experiments confirmed the correct localization of BSEP. A subsequent detergent screen led to the identification of detergents suitable for BSEP solubilization. The transporter appears to be in a detergent-resistant environment as only the lipid–like, zwitter–ionic detergents of the Fos–choline and Cyclofos series, comparatively “harsh” detergents, could efficiently extract BSEP. Finally, a purification protocol was established to obtain BSEP in a detergent-solubilized form. With a dual–affinity tag tandem affinity purification could be employed to obtain the ABC transporter with approximately 75% purity [4]. In summary, the optimized cloning strategy for BSEP cDNA in S. cerevisiae as well as the efficient expression and purification protocols enable us to investigate the interaction of BSEP with other proteins in vitro.

The human multidrug resistance protein 3 (MDR3/ABCB4) belongs to the ATP binding cassette (ABC) transporter family found in all kingdoms of life. MDR3 is located in the canalicular membrane of hepatocytes and translocates phosphatidylcholine (PC) from the inner to the outer leaflet of the canalicular membrane energized by ATP hydrolysis. PC is one major component of bile and is essential to protect the biliary duct from bile salts, which are translocated by the bile salt export pump (BSEP/ABCB11). Phosphatidylcholine, bile salts and cholesterol translocated by ABCG5/G8 form mixed micelles. These mixed micelles have a lower capacity to extract lipids from the membrane and prevent the crystallization of cholesterol in the biliary duct. Different types of liver diseases e.g. progressive familial intrahepatic cholestasis type 3 (PFIC 3) are caused by dysfunction of MDR3. On the one hand mutations within the MDR3 gene can abolish the ATPase activity and PC lipid transport, respectively. On the other hand defects of the regulation pathway with respect to the targeting and retention of MDR3 to the canliculare membrane can occur. Currently, it is not known, how MDR3 is regulated and targeted to the canalicular membrane. Further, MDR3 shares over 75 % identity with the multidrug resistance protein 1 (MDR1/ABCB1). MDR1 is well-characterized and plays an enormous challenge in chemotherapy. If we identify the regulation mechanism of MDR3, we will as well contribute to the understanding of MDR1 regulation. Several adaptor proteins are identified, which are important for the incorporation, retention and degradation of ABC transporters into the canalicular membrane of hepatocytes. For example, the adaptor protein NHERF-1 is crucial for the internalization of MRP2 and MRP4, Radixin (RDX) for MRP2 and HAX-1 for BSEP. NHERF-1 contains two PDZ domains, which binds amongst others the highly conserved motive (S/T)X(V/I/L/A/F/M) (X = any amino acid). The primary sequence of the second MDR3 nucleotide binding domain (NBD) contains several times this binding motive. This is a first hint that NHERF-1 binds also MDR3. RDX is a member of the Ezrin/Radixin/Moesin (ERM) protein family and is mainly localized at the canalicular membrane of hepatocytes. The N-terminale FERM (4.1-protein, Erzin, Radixin, Moesin) domain of RDX binds the C-terminus of MRP2. The amino acid sequence alignment of MRP2 and MDR3 showed no conserved amino acids at the C-terminus. Ikebuchi Y. et al. used a yeast-two hybrid screen with a cytoplasmatic linker region of MDR3 and a human liver cDNA library to identify interaction proteins of MDR3. They determined that the localization of MDR3 is regulated by the receptor for activated C-kinase 1 (RACK1) [1]. We used RACK1 as an internal control for our experiments. To understand how MDR3 is targeted to the apical membrane, we aimed to identify further interaction partners of this transporter. For this purpose we want to use in vitro pulldown assays, size exclusion chromatography and blue native page with purified MDR3 and the potentially interaction proteins. First, we cloned the human cDNA of MDR3 in Saccharomyces cerevisiae and expressed MDR3 in the methylotrophic yeast Pichia pastoris in amounts suitable for a detailed structure-function analysis. We tested over 100 different detergents via dotblot analysis and found that only lipid-like detergents as Fos-cholines (FCs) were able to extract MDR3 in adequate yields. MDR3 was purified with FC-16 via tandem-affinity purification (TAP) consisting of an immobilized metal ion affinity chromatography and a calmodulin binding peptide purification step. About 6 mg protein per 100 g wet cells were obtained with high purity and homogeneity. The functionality was determined by measuring the ATPase activity of purified MDR3 in detergent solution. Purified MDR3 exhibited significant ATP hydrolysis activity that could be stimulated by DPPC and DOPC. To eliminate any possibility of unspecific ATPase activity due to contaminants, we cloned and purified an ATPase inactive mutant, where the glutamic acid is changed to a glutamine in the highly conserved Walker B motive of both NBDs. This mutant (E558Q/E1027Q) showed basal ATPase activity comparable to the wild-type protein. But more important, no ATPase activity stimulation by adding PC lipids to the reaction was obtained. So far our work demonstrates that heterologously expressed MDR3 can be purified in a functional state with respect to its capability to bind and hydrolyze ATP [2]. Second, we cloned the cDNA of the adaptor proteins NHERF-1 and RDX combined with a StrepII affinity tag in the Escherichia coli expression system. RACK1 cDNA was fused to a hexa-histidine affinity tag and expressed also in E. coli. Recently, we purified the interaction proteins NHERF-1 and RDX with Strep-tactin sepharose to high purity and homogenity. RACK1 was purified via immobilized metal ion affinity purification and size exclusion chromatography. We obtained approximately 20 mg out of 2 L cell culture. To determine the interaction of MDR3 with the putative interaction partners NHERF-1, RDX and RACK1 we want to use the in vitro pull-down assay, size exclusion chromatography (SEC) and blue native page. For the pull-down assay, the purified interaction partner was mixed with the purified and functional MDR3. Then one of the proteins was immobilized on the corresponding matrix by the affinity tag. After several washing steps, the complex was eluted and identified via immunoblotting. Currently, we have done initial pull-down assays with MDR3 and the positive control RACK1. MDR3 was immobilized by the calmodulin binding peptide (CBP) to calmodulin resin and eluted after adding EGTA. RACK1 has a size of 35 kDA and is obtained at the 35 kDA marker band in the SDS-PAGE gel. It binds unspecifically to the CBP resin. In the immunoblot against RACK1 we determined in the reaction of MDR3 and RACK1 an additional band of about 130 kDA for RACK1, while RACK1 alone ran at 35 kDa. One explanation is that RACK1 forms a highly stable complex with MDR3, which is resistent against SDS. To confirm this result we tried blue native page, but unfortunately we did not succeed so far. With this first results we can verify that MDR3 interacts with RACK1. Further, we want to analyze NHERF-1 and RDX concerning the potential interaction with MDR3.