Báo cáo khoa học: Structural requirements for the apical sorting of human multidrug resistance protein 2 (ABCC2)

doi:10.1046/j.1432-1033.2002.02832.x

Eur. J. Biochem. 269, 1866–1876 (2002) (cid:211) FEBS 2002

Structural requirements for the apical sorting of human multidrug resistance protein 2 (ABCC2)

Anne T. Nies1, Jo¨ rg Ko¨ nig1, Yunhai Cui1, Manuela Brom1, Herbert Spring2 and Dietrich Keppler1 1Division of Tumor Biochemistry, Deutsches Krebsforschungszentrum, Heidelberg, Germany; 2Division of Cell Biology, Deutsches Krebsforschungszentrum, Heidelberg, Germany

showed that full-length GFP–MRP2 was localized to the apical membrane in 73% of transfected, polarized cells, whereas it remained on intracellular membranes in 27% of cells. Removal of the C-terminal TKF peptide and stepwise deletion of up to 11 amino acids did not change this pre- dominant apical distribution. However, apical localization was largely impaired when GFP–MRP2 was C-terminally truncated by 15 or more amino acids. Thus, neither the PDZ-interacting TKF motif nor the full seven-amino-acid extension were necessary for apical sorting of MRP2. Instead, our data indicate that a deletion of at least 15 C-terminal amino acids impairs the localization of MRP2 to the apical membrane of polarized cells.

Keywords: epithelial polarity; green fluorescent protein; multidrug resistance protein 2; protein tra(cid:129)cking.

The human multidrug resistance protein 2 (MRP2, symbol ABCC2) is a polytopic membrane glycoprotein of 1545 amino acids which exports anionic conjugates across the apical membrane of polarized cells. A chimeric protein composed of C-proximal MRP2 and N-proximal MRP1 localized to the apical membrane of polarized Madin–Darby canine kidney cells (MDCKII) indicating involvement of the carboxy-proximal part of human MRP2 in apical sorting. When compared to other MRP family members, MRP2 has a seven-amino-acid extension at its C-terminus with the last three amino acids (TKF) comprising a PDZ-interacting motif. In order to analyze whether this extension is required for apical sorting of MRP2, we generated MRP2 constructs mutated and stepwise truncated at their C-termini. These constructs were fused via their N-termini to green fluorescent protein (GFP) and were transiently transfected into polar- ized, liver-derived human HepG2 cells. Quantitative analysis

MRP2. The absence of MRP2 from the canalicular membrane of human hepatocytes is the molecular basis of the Dubin–Johnson syndrome [15,22–24], which is associ- ated with conjugated hyperbilirubinemia.

Epithelial cell polarity is a result of the domain-specific sorting of proteins. Neither apical nor basolateral trafficking seems to follow a (cid:212)default(cid:213) pathway, rather, specific signals or interactions are required for inclusion of proteins into apically or basolaterally destined transport vesicles within the trans Golgi network (TGN; reviewed in [25]). Basolat- eral sorting signals are most often tyrosine- or dileucine- based motifs in the cytoplasmic domains of proteins [26], however, other basolateral sorting signals have been also identified [27,28]. Several mechanisms have been described for apical sorting. These include apical localization signals in the extracellular, transmembrane, or cytoplasmic domains [29]. For several apical proteins, clustering into cholesterol- and sphingolipid-rich, detergent-insoluble microdomains has been demonstrated to be important for the formation of apical vesicles from the TGN [30].

Members of the multidrug resistance protein (MRP) family are integral membrane glycoproteins which mediate the ATP-dependent export of amphiphilic anions across the plasma membrane [1]. MRP1, the first cloned member of the MRP family [2], is present in the plasma membrane of several cell types [3–5]. After transfection of MRP1 cDNA in polarized cells, MRP1 is localized to the basolateral membrane [6]. Several MRP family members are known to be endogenously expressed in polarized cells. Whereas MRP3 [7,8] and MRP6 [9,10] are localized to the basolateral membrane of rat and human hepatocytes, MRP2 is the only isoform identified so far that is localized exclusively to the apical membrane of polarized cells, such as hepatocytes and renal proximal tubule cells [1,11,12]. MRP2 was initially cloned from rat liver [11,13,14], and subsequently from human liver [11,15,16] and human tumor cells [17]. Trans- port studies using inside-out oriented membrane vesicles from liver [18,19] or from cells stably transfected with human MRP2 cDNA [16,20,21] demonstrated the transport of conjugated and unconjugated lipophilic anions by

In addition to active sorting into specific transport vesicles within the TGN, selective stabilization of proteins in their respective membrane domains has been suggested [31]. One mechanism by which this may be achieved is the binding of membrane proteins via their C-termini to PDZ domain-containing proteins. The latter recognize a consen- sus sequence (T/S-X-V/I) at the C-termini of membrane proteins [32]. Interaction of these PDZ-interacting motifs with PDZ domain-containing proteins has been shown to be required for the membrane domain-specific sorting of some basolateral as well as of some apical membrane

Correspondence to A. Nies, Division of Tumor Biochemistry, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany. Fax: + 49 6221 422402, Tel.: + 49 6221 422403, E-mail: a.nies@dkfz.de Abbreviations: GFP, green fluorescent protein; MRP2, multidrug resistance protein 2 (human genome nomenclature symbol (cid:136) ABCC2); PDZ, PSD-95/DlgA/ZO-1-like. (Received 23 January 2002, accepted 6 February 2002)

Apical sorting of human MRP2 (Eur. J. Biochem. 269) 1867

(cid:211) FEBS 2002

proteins [33]. PDZ domain-containing proteins either bind directly or via adaptor proteins to the cytoskeleton [33].

Alexa Fluor488 were from Molecular Probes (Eugene, OR, USA). Donkey anti-(rat IgG) Ig coupled to TexasRed and Cy3-conjugated goat anti-(mouse IgG) Ig were from Jackson Immunoresearch (West Grove, PA, USA).

Generation of a cDNA encoding a MRP1/2 chimeric protein

Present knowledge on the mechanisms by which MRP isoforms are targeted to their respective membrane domain in polarized cells is limited. We recently showed that a six- nucleotide deletion within the human MRP2 gene causes Dubin–Johnson syndrome [24,34]. This mutation, leading to the loss of two amino acids from the second nucleotide- binding domain [24], results in defective MRP2 maturation and retention of MRP2 in the ER, so that sorting of MRP2 to the apical membrane is impaired [34]. The aim of the present study was to identify structural determinants required for apical sorting of human MRP2. Because MRP2 has a seven- amino-acid extension at its C-terminus, which is not found in the basolaterally localized isoforms MRP1, MRP3, and MRP6 [7], it was hypothesized that this C-terminal extension contains a signal for apical localization of MRP2. In addition, the C-terminal three amino acids of MRP2 were identified as a motif interacting with a PDZ domain- containing protein [35]. A recent study described that deletion of this PDZ-interacting motif leads to localization of MRP2 predominantly in the basolateral membrane of polarized Madin–Darby canine kidney (MDCK) cells [36]. This result may, however, be misleading because MRP2 was tagged at the C-terminus with GFP and interaction with PDZ domain-containing proteins may be disrupted by the addi- tion of amino acids to the C-terminal PDZ-interacting motif [37,38]. In addition, human proteins may localize differently in canine cells. In the present work, we therefore used human MRP2 tagged with GFP at the N-terminus, thus leaving the C-terminus free for possible binding of interacting proteins. With this experimental setup, we show that, in contrast to our expectations, the C-terminal 11 amino acids of MRP2, including the PDZ-interacting motif, were not necessary for apical sorting of MRP2 in polarized human HepG2 cells. However, truncation by more than 15 amino acids resulted in impaired delivery of MRP2 to the apical membrane.

The cDNA encoding the chimeric MRP1/2 protein (Fig. 1) was constructed by generating a XbaI restriction site in the cDNA sequence of human MRP1 in a PCR-based approach. In detail, a MRP1 cDNA fragment was amplified using the MRP1 cDNA, inserted into the vector pcDNA3.1(+), as template and the T7 vector primer as forward primer. The reverse primer ochimrp1.rev was used to generate the XbaI restriction site in the MRP1 cDNA. It has the sequence 5¢-AGAGGGGATCATCTAGAAG GTA-3¢ (position 2386–2365) and has three base-pair substitutions when compared with the MRP1 wild-type sequence: 2370G fi A, 2371A fi G, and 2373G fi T. These substitutions were necessary to generate the XbaI restriction site. A (cid:25) 2500 bp fragment was PCR amplified using the following cycles: 5 min 94 (cid:176)C, 5 cycles with 45 s at 94 (cid:176)C denaturation, 45 s 55 (cid:176)C annealing and 120 s 72 (cid:176)C elongation, 30 cycles with 45 s 94 (cid:176)C denaturation, 45 s at 65 (cid:176)C annealing, and 120 s at 72 (cid:176)C elongation, followed by 10 min at 72 (cid:176)C. The fragment was subcloned into the vector pCR2.TOPO (Invitrogen, Carlsbad, CA, USA) resulting in the plasmid pmrp1/XbaI.topo. Human MRP2 cDNA (GenBank/EMBL accession number X96395) was cloned into pcDNA3.1(+) as described previously ([16], pMRP2). For generating a full-length cDNA encoding the MRP1/2 chimera, pMRP2 was restricted with NotI/XbaI and the MRP1 cDNA fragment from the pmrp1/XbaI.topo plasmid obtained by NotI/XbaI restriction was inserted, thus generating the plasmid pmrp1/2chim.31. The correct sequence of the fragment and the cloning sites were verified by sequencing and restriction analysis.

M A T E R I A L S A N D M E T H O D S

Materials and antibodies

Generation of green fluorescent protein (GFP)–MRP2 constructs

Normal and C-terminally mutated GFP–MRP2 constructs were generated in the mammalian expression vector pcDNA3.1(+) (Invitrogen). After translation, GFP was attached to the N-terminus of the proteins, so that the GFP moiety was in the lumen of the ER or on the extracellular side (Fig. 2). Constructs were restriction-mapped and sequenced to verify correctness of the fragments.

Fetal bovine serum and agarose were from Sigma (St Louis, MO, USA). Pfu DNA polymerase, restriction enzymes, ligase, and modifying enzymes were from Stratagene (La Jolla, CA, USA) or Promega (Madison, WI, USA). Lysozyme and ampicillin were from Roche Molecular Biochemicals (Indianapolis, IN, USA). Rhodamine-conju- gated concanavalin A was from Vector Laboratories (Burlingame, CA, USA). All other chemicals were of analytical grade and obtained either from Merck (Darm- stadt, Germany) or Sigma.

GFP, optimized for maximal fluorescence [39] and mam- malian expression [40], was cloned into the BamHI and NotI restriction sites of the expression vector pcDNA3.1(+) (pGFP). GFP was PCR-amplified using the sense-primer 5¢-AGATCTGCCACCATGGTGAGC AAG-3¢, which introduced a BglII site (bold), and the antisense primer 5¢-CCGCGGCCGCTTGTATAGCTCGTCCATGCCG AG-3¢, which introduced a SacII (underlined) and a NotI site (bold), at the same time removing the stop codon and the BsrGI site at the 3¢ end of the GFP coding sequence. PCR- amplified GFP was cloned into the pDisplay vector (Invi- trogen) using the BglII and the SacII sites (plumGFP).

The polyclonal rabbit antibody directed against the C-terminus of human MRP2, EAG5, has been described previously [11,12]. The mouse mAb to dipeptidylpepti- dase IV (CD26; anti-DPPIV Ig; clone 202.36) was from Ancell (Bayport, MN, USA), and the mouse monoclonal antibody to protein disulfide isomerase (PDI; clone RL90) was purchased from Affinity Bioreagents (Golden, CO, USA). The mouse monoclonal anti-villin Ig was from Transduction Laboratories (Lexington, KY, USA). Rat anti-(ZO-1) Ig was from Chemicon (Temecula, CA, USA). Goat anti-(rabbit IgG) Ig coupled to Alexa Fluor546 or

pMRP2 was digested with NotI and BsrGI, and the fragment was replaced with a PCR-fragment that enabled

1868 A. T. Nies et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

erated by digesting pGFP-MRP2 with HindIII/BsrGI and by cloning this fragment into the respective HindIII/BsrGI- digested deletion construct.

Cell culture and transfection

Human hepatoma HepG2 and MDCK cells (strain II) were maintained in Dulbecco’s modified Eagle’s medium (Sig- ma), supplemented with 10% (v/v) fetal bovine serum, penicillin (100 UÆmL)1) and streptomycin (100 lgÆmL)1). For transient transfections, cells were seeded into 35-mm and 100-mm and dishes at a density of 5 · 105 and 5 · 106 cells per dish, respectively, 24 h prior to transfection. HepG2 cells were transfected with the FuGENE 6 transfection reagent (Roche Molecular Biochemicals) according to the manufacturer’s instructions using 5 and 25 lL transfection reagent and 1.5 and 7.5 lg DNA per 35- and 100-mm dish, respectively. MDCKII cells were transiently or stably [16] transfected using calcium phosphate precipitation or the FuGENE transfection reagent.

Immunofluorescence microscopy

(MRP2D20),

the in-frame insertion of GFP at the N-terminus of MRP2 (pMRP2.1). The sense primer for this PCR reaction was 5¢-GCGGCCGCTCATGCTGGAGAAGTTCTG-3¢ (NotI site in bold) and the antisense primer was 5¢-GTGCCACA GAGTATCGAG-3¢. plumGFP vector was digested with HindIII and NotI, and the resulting GFP-encoding frag- ment including the murine Ig j-chain leader sequence was cloned into HindIII/NotI-digested pMRP2.1 (pGFP- MRP2). For generation of C-terminal deletion constructs, a 2346-bp DNA fragment encoding the C-proximal part of MRP2 was generated by PCR with ApaI and SacII sites added at the 3¢ end during amplification. Primers used were 5¢-AGCGGATCAGCCTGG-3¢ (sense primer) and 5¢-GGGCCCGCGGCTAGAATTTTGTGCTGTTCAC-3¢ (antisense primer, ApaI site bold, SacII site underlined). This PCR fragment was ligated into ApaI-digested pMRP2 (pMRP2.2). C-Terminal deletion constructs were generated by cloning PCR-amplified fragments into the Bsu36I and the SacII sites of pMRP2.2. For these PCR reactions, the sense primer was 5¢-CCTGTTCTCTGGAAGCC-3¢ and the antisense primers were 5¢-CCGCGGCTAGCTGTTC ACATTCTCAATG-3¢ (MRP2D3), 5¢-CCGCGGCTACT (MRP2D7), 5¢-CCGCGG CAATGCCAGCTTCCTT-3¢ CTATTCCTTAGCCATAAAGTAAAA-3¢ (MRP2D11), 5¢-CCGCGGCTAAAAGTAAAAGGGTCCAGGG-3¢ (MRP2D15), 5¢-CCGCGGCTAAGGGATTTGTAGCA GTTCT-3¢ 5¢-CCGCGGCTATTCTTCA GGGCTGCCGC-3¢ (MRP2D25), 5¢-CCGCGGCTATTC CTTAGCCATTTCTTCAGGGCTGCCGC-3¢ (MRP2 D25MAKE), 5¢-CCGCGGCTACAGCCTGTGGGCGA TGG-3¢ (MRP2D50), 5¢-CCGCGGCTACAGCAGCTG CCTCTGGC-3¢ (MRP2D100), 5¢-CCGCGGCTAGAAT TTTGCGCTGTTCACATTC-3¢ (MRP2T1543 A), and 5¢-CCGCGGCTAGAATTTTGTAAAGTAAAAGGGT CCAGGG-3¢ (MRP2D15TKF). GFP constructs were gen-

HepG2 or MDCKII cells grown on glass cover slips were fixed with methanol at )20 (cid:176)C for 1 min and rehydrated in NaCl/Pi. Cells were incubated with the primary antibody for 60 min at room temperature, washed three times with NaCl/Pi, incubated with the secondary antibody for 60 min, and then washed again three times with NaCl/Pi. Cover slips were mounted in Moviol (Hoechst, Frankfurt, Germany) and observed on a confocal laser scanning microscope (LSM 510, Carl Zeiss, Jena, Germany) using the excitation wavelengths of the argon ion (488 nm) and the helium/neon laser (543 nm). Prints were taken of optical sections of 0.8-lm thickness. Antibodies were diluted in NaCl/Pi at

Fig. 1. Predicted topology models (A) and localization of MRP2 (B,C) and chimeric MRP1/2 (D,E) in polarized MDCKII cells. The chimeric MRP1/2 consists of the MRP1 sequence followed by the sequence of MRP2 starting at amino acid 791. For MRP2, only four transmembrane segments are predicted between both nucleotide-binding domains (NBD1 and NBD2 [43]), whereas six trans- membrane segments are predicted for MRP1 [44]. MDCKII cells stably synthesizing MRP2 or chimeric MRP1/2 were immunostained with the EAG5 antibody directed against MRP2 (green in B–E). Both proteins were localized to the apical membrane as observed in the x–y plane (B,D) and the x–z plane (C,E). Nuclei were stained with propidium iodide (red in B–E). Bar, 10 lm.

Apical sorting of human MRP2 (Eur. J. Biochem. 269) 1869

(cid:211) FEBS 2002

GFP fluorescence) and polarized (as observed by ring-like DPPIV fluorescence) cells were counted on a fluorescence microscope (Axioskop; Carl Zeiss, Jena, Germany). For each transfected and polarized cell, the localization of the respective GFP–MRP2 protein was analyzed and classified into one of three categories as follows: when GFP and DPPIV fluorescence merged in ring-like, microvilli-lined structures between adjacent cells, i.e. the apical membrane [42], the localization was defined as (cid:212)apical(cid:213), irrespective of additional intracellular GFP fluorescence. When GFP fluorescence was absent from these ring-like structures in polarized cells, but observed in vesicular structures, local- ization was defined as (cid:212)vesicular(cid:213). When DPPIV fluorescence was present in the ring-like structures and GFP fluorescence appeared exclusively reticular, localization was defined as endoplasmic reticulum (ER). Localization of the respective GFP–MRP2 in the ER was confirmed by colocalization with an antibody against an ER marker protein, protein disulfide isomerase (data not shown), as described previ- ously [34]. For each GFP–MRP2 construct, the percentage of each localization was calculated. At least four indepen- dent transfections were analyzed in this way. For analysis of the steady-state distribution of GFP–MRP2 proteins, cells were induced with 5 mM butyrate for 24 h [16] and observed 48 h after start of transfection. For analysis of the time- course of GFP–MRP2 protein localization, cells were observed after 1, 2, 3, and 4 days post-transfection without prior induction with butyrate.

For assessment of polarity, HepG2 cells were double- labeled with anti-DPPIV Ig (1 : 100) and EAG5 (1 : 100), or anti-villin Ig (1 : 100) and EAG5 (1 : 100), and the respective secondary antibodies as described above. Apical vacuoles staining positive for DPPIV and MRP2 or villin and MRP2 were counted on a fluorescence microscope (Axioskop).

R E S U L T S

Apical localization of a MRP1/2 chimeric protein in polarized MDCKII cells

The amino-acid identity of only 48% between the laterally localized isoform MRP1 and the apically localized isoform MRP2 [1] hampers the identification of apical sorting signals in the MRP2 sequence by direct comparison of both sequences. We therefore constructed a cDNA encoding a MRP1/2 chimeric protein and immunolocalized this chi- meric protein in MDCKII cells (Fig. 1). The chimeric MRP1/2 protein was localized in the apical membrane of polarized MDCKII cells as was full-length MRP2 (Fig. 1) suggesting that the C-proximal part of MRP2 contains information for apical sorting of MRP2.

the following dilutions: anti-(ZO-1) Ig (1 : 100), EAG5 (1 : 200), anti-PDI Ig (1 : 400), anti-DPPIV Ig (1 : 500), and the respective secondary antibodies at 1 : 300. For staining of lysosomes, LysoTracker Red (Molecular Probes) was used according to the manufacturer’s instructions. For staining of the apical membrane of MDCKII cells, rhod- amine-labeled concanavalin A was added to the apical chamber of a Transwell filter insert at 5 lgÆmL)1 according to a method described recently [41]. Live HepG2 cells expressing GFP were observed as described previously [42].

Apical localization of GFP–MRP2 in polarized HepG2 and MDCKII cells

Quantitative analysis of the subcellular localization of C-terminally mutated and truncated GFP-MRP2 proteins in polarized HepG2 cells

A sequence alignment of the C-terminal ends of human MRP1, MRP2, MRP3, and MRP6 (Fig. 2) shows that the apical MRP2 has a seven amino-acid extension in compar- ison to the basolateral family members MRP1, MRP3, and MRP6. Recombinant MRP1 was localized to the basolat- eral membrane in polarized porcine cells [6]. MRP3 and MRP6 are endogenously synthesized in polarized cells such

HepG2 cells were transiently transfected and immuno- stained with the anti-DPPIV Ig as described above. For each transfection, at least 100 transfected (as observed by

Fig. 2. Alignment of the C-termini of members of the human MRP family (A) and predicted topology models of MRP2, GFP–MRP2, and lumGFP (B). According to the prediction of the TMHMM program [45], and experimentally confirmed [16], the N-terminus of MRP2 has an extracellular location. Therefore, a cDNA was constructed which encoded a fusion protein of GFP and MRP2 with the GFP moiety targeted to the lumen of the ER, followed by the complete sequence of human MRP2 (GFP–MRP2). Expression of GFP from the pDisplay vector (lumGFP for (cid:212)lumenal GFP(cid:213)) resulted in a GFP which was targeted to the lumen of the ER because of a murine Ig j-chain leader sequence ([47]; black box) at the N-terminus of GFP and which was anchored in the plasma membrane due to the platelet-derived growth factor receptor transmembrane domain at the C-terminus of GFP ([48]; cross-hatched box).

1870 A. T. Nies et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

indicating that

as hepatocytes and localized in the basolateral membrane [7–10]. Because the extension of MRP2 might represent a signal for apical localization of MRP2, we generated MRP2, which was mutated or stepwise truncated at its C-terminus, and analyzed quantitatively the localization of these MRP2-derived proteins in polarized HepG2 cells. In order to distinguish between endogenous MRP2 in HepG2 cells [42,46] and C-terminally mutated MRP2 in these cells, we constructed cDNAs coding for fusion proteins of MRP2 and GFP. Because a (cid:212)free(cid:213) C-terminus may be necessary for proper apical sorting of MRP2, e.g. by binding of interact- ing proteins, GFP was fused to the N-terminus of MRP2. The N-terminus of MRP2 is located on the extracellular side [16], therefore a cDNA was constructed which led to translation of a GFP inserted into the lumen of the ER by the murine Ig j-chain leader sequence, a sequence described to target proteins to the secretory pathway [47], followed by the sequence of MRP2 (Fig. 2). This GFP–MRP2 fusion protein was localized to the apical membrane of polarized HepG2 cells (Fig. 3). When lumenal GFP (lumGFP) was lumGFP was not expressed from the pDisplay vector, secreted into the medium but anchored to the plasma membrane due to the platelet-derived growth factor recep- tor (PDGFR) transmembrane domain at the C-terminus of GFP (Fig. 2, [48]). This PDGFR domain is not present in the GFP–MRP2 constructs (Fig. 2). LumGFP was equally distributed in the apical and the basolateral membrane of polarized HepG2 cells, and, in addition, in intracellular vesicular structures (Fig. 3) indicating that neither the murine Ig j-chain leader sequence nor the PDGFR transmembrane domain contained a specific signal for apical localization. To exclude an effect of GFP on MRP2 targeting, the distribution of GFP in polarized HepG2 cells was analyzed (Fig. 3). The soluble GFP was present within the cells without any localization in the plasma membrane. As a control, GFP–MRP2 was also observed in MDCKII cells where it localized to the apical membrane (Fig. 4). The polarity of the MDCKII cells was confirmed by immunostaining with an antibody detecting the tight- junctional protein ZO-1 (Fig. 4), the MDCKII cells were polarized under our experimental conditions. MDCKII cells synthesizing GFP–MRP2 were also immunostained with the EAG5 antibody resulting in identical fluorescence as the GFP fluorescence (Fig. 4). Because the EAG5 antibody was raised against the 15 C-terminal amino acids of human MRP2 [11,12], this result demonstrates that the observed GFP fluorescence reflects localization of a complete GFP–MRP2 protein.

The C-terminal PDZ-interacting motif is not required for apical sorting of MRP2

Because apical vacuoles form between adjacent HepG2 cells as vesicle-like structures lined with microvilli [49], they can be stained with antibodies either to cytoskeletal proteins such as villin [49,50] or with antibodies to canalicular membrane proteins such as DPPIV and MRP2 [42]. To assess the validity of DPPIV as a marker for polarity, HepG2 cells were double-stained for DPPIV and MRP2. The majority (98.9%) of DPPIV-positive, microvilli-lined ring-like structures were also positive for MRP2 (540 apical vacuoles counted). Similarly, 99.6% of villin-positive, microvilli-lined ring-like structures were also positive for MRP2 (535 apical vacuoles counted). This result indicates that staining for all three proteins, villin, DPPIV, and MRP2, can be used as marker for cell polarity in HepG2 cells.

The C-terminal three amino acids of the human MRP2 sequence (TKF, Fig. 2) have been reported to interact with a PDZ domain-containing protein [35] and may thus be necessary for apical sorting of MRP2. We therefore deleted the C-terminal three amino acids or substituted threonine with alanine at position 1543. The respective, mutated GFP–MRP2 was observed in polarized HepG2 cells. For quantitative analysis, localization of GFP–MRP2 proteins were classified into one of three categories as shown in the representative images of Fig. 5 and described in Materials and methods.

Fig. 3. Localization of GFP–MRP2, lumGFP, and GFP in polarized HepG2 cells. HepG2 cells were transiently transfected with GFP– MRP2 (A,B) or lumGFP (C,D), fixed 48 h after transfection, and immunostained with an antibody against dipeptidylpeptidase IV (DPPIV) in order to visualize apical vacuoles (B,D). GFP-transfected cells (E,F) were visualized by fluorescence microscopy (E) or by phase- contrast (F). In GFP–MRP2-transfected cells (A), fluorescence was observed in ring-like structures, i.e. the apical (vacuolar) membrane, and, in addition, in intracellular vesicular structures of varying size. In contrast, lumGFP (C) was observed in the basolateral and in the apical membrane in equal amounts, and, additionally, in intracellular vesic- ular structures, most likely vesicles of the secretory pathway. GFP (E) was distributed throughout the cells without localization to the plasma membrane. Asterisks mark the lumen of apical vacuoles. Bars, 10 lm.

Apical sorting of human MRP2 (Eur. J. Biochem. 269) 1871

(cid:211) FEBS 2002

In 73% of transfected and polarized HepG2 cells GFP– MRP2 reached the apical membrane (Table 1). In the remaining 27% of transfected and polarized cells, GFP– MRP2 did not reach the apical membrane, but was present in intracellular compartments, such as vesicular structures and the ER. Deletion of the C-terminal three amino acids TKF or substitution of threonine with alanine led to proteins that were as efficiently sorted to the apical membrane of polarized HepG2 cells as was full-length MRP2 (Table 1). Furthermore, GFP–MRP2D15 which was predominantly localized in the ER was not (cid:212)rescued(cid:213) from this localization by addition of the TKF motif (Table 1).

Asacontrol,localizationofGFP–MRP2,GFP–MRP2D3, and GFP–MRP2-T1543A was also analyzed in MDCKII cells grown polarized on Transwell filter membranes (Fig. 6). The apical membrane was visualized by rhod- amine-conjugated concanavalin A added to the upper chamber of the Transwell insert. GFP–MRP2, GFP– MRP2D3, and GFP–MRP2-T1543A were almost exclu- sively present in the apical membrane with some GFP fluorescence also present in intracellular compartments. None of the three analyzed proteins were observed in the basolateral membrane.

Localization of C-terminally truncated GFP–MRP2 proteins

Because the PDZ-interacting motif was not necessary for apical sorting of MRP2, the C-terminus of GFP–MRP2 was further truncated. Truncation of the C-terminus by seven or 11 amino acids led to proteins that reached the apical membrane of polarized HepG2 cells as full-length

Fig. 4. Localization of GFP–MRP2 in polarized MDCKII cells. MDCKII cells transiently transfected with GFP–MRP2 were fixed 48 h after transfection and immunostained with an antibody against the tight-junctional protein ZO-1 (C,D), or with the EAG5 antibody (G,H) which is directed against the 15 C-terminal amino acids of human MRP2 [11,12]. The GFP fluorescence (A,B,E,F) shows that GFP–MRP2 is localized to the apical membrane, as observed in the x–y plane (A,E) and the x–z plane (B,F). ZO-1 staining lines the cells in the x–y view (C), however, ZO-1 is restricted to the tight-junctions appearing as dots in the vertical section (D). EAG5 fluorescence (G,H) was identical to the GFP fluorescence (E,F) showing synthesis of a complete GFP–MRP2 protein. Bars, 10 lm.

Fig. 5. Representative fluorescence images of subcellular localization of GFP–MRP2 con- structs in polarized HepG2 cells as quantified in Tables 1–3. When GFP fluorescence (A) and DPPIV fluorescence (B) merged to yellow in the apical membrane (C) the localization of the GFP–MRP2 construct was designated as (cid:212)apical(cid:213). When the GFP–MRP2 construct was present in intracellular vesicles (D) without reaching the apical membrane (E), no yellow color was observed (F). Some GFP–MRP2 constructs remained in reticular structures, i.e. the ER (G), and no GFP fluorescence of the apical vacuolar membrane (H) was observed (I). Bars in A–I, 10 lm. Asterisks mark apical vacuoles.

1872 A. T. Nies et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

Table 1. Quantitative analysis of the subcellular localization of C-terminally mutated GFP–MRP2 constructs in polarized HepG2 cells. Data are percentages of cells in which the respective localization of recombinant protein was observed as described in Materials and methods. Cells were observed 2 days after transfection. Data are means (cid:139) SD of six transient transfections using butyrate-induced cells as described under Materials and methods.

GSPEELLQIPGPFYFMAKEAGIENVNSTKF GSPEELLQIPGPFYFMAKEAGIENVNS

GSPEELLQIPGPFYFMAKEAGIENVNSAKF GSPEELLQIPGPFYF

Construct % Apical % Vesicles % ER C-Terminal sequence (1516–1545)

GSPEELLQIPGPFYFTKF

GFP–MRP2 (Table 2). However, delivery to the apical membrane was largely impaired when GFP–MRP2 was C-terminally truncated by 15, 20, 25, 50 or 100 amino acids. The percentage of polarized and transfected cells in which the respective protein reached the apical membrane decreased to 16% (GFP–MRP2D15), 15% (GFP– MRP2D20), 8% (GFP–MRP2D25), and 1% (GFP– MRP2D50, GFP–MRP2D100) with a concomitant accumulation of the proteins in intracellular compartments, such as the ER and intracellular vesicles (Table 2). Because

deletion of the tetrapeptide MAKE, i.e. amino acids 1531– 1534, resulted in a shift in the percentage of cells with an apical (GFP–MRP2D11) to an intracellular localization (GFP–MRP2D15), this sequence might be involved in the apical delivery of MRP2. However, addition of this tetrapeptide onto GFP–MRP2D25, which had an intracel- lular localization in most of the cells, did not increase the number of cells in which GFP–MRP2D25MAKE reached the apical membrane. This result indicates that it is not the co-linear sequence of the tetrapeptide that is required for apical delivery of MRP2. The intracellular vesicles contain- ing the respective GFP–MRP2 construct were not lyso- somes as shown by the lack of colocalization with the lysosomal marker LysoTracker Red (Fig. 7). Similarly, GFP–MRP2 constructs were not present in intracellular vesicles that contained DPPIV (Fig. 7). These results suggest that GFP–MRP2 was present in endosomes of yet unidentified nature.

Because intracellular accumulation of GFP–MRP2 trunc- ated by 15–25 amino acids may be due to a delay in intracellular transport to the apical membrane we analyzed localization of GFP–MRP2, GFP–MRP2D15, and GFP– MRP2D25 from 1 to 4 days after the start of transfection (Table 3). There was no difference in the intracellular distribution of the respective GFP–MRP2 protein over time.

GFP–MRP2 GFP–MRP2D3 GFP–MRP2-T1543A GFP–MRP2D15 GFP–MRP2D15TKF 73 (cid:139) 9 64 (cid:139) 9 67 (cid:139) 6 16 (cid:139) 7 21 (cid:139) 11 18 (cid:139) 9 13 (cid:139) 5 16 (cid:139) 2 17 (cid:139) 7 21 (cid:139) 7 9 (cid:139) 5 23 (cid:139) 9 17 (cid:139) 6 67 (cid:139) 14 58 (cid:139) 11

D I S C U S S I O N

conductance

transmembrane

MRP2 is the only MRP isoform known so far which localizes to the apical membrane of polarized cells [1,10]. Recently, the C-terminal three amino acids (TRL) of the cystic fibrosis regulator (CFTR), which comprise a PDZ-interacting motif, were identified as a signal for apical localization [51]. Because CFTR is a member of the MRP (ABCC) family with 27% amino-acid identity to MRP2 [1], we investigated whether the C-terminal tail of MRP2 is also involved in apical sorting. Interaction of a PDZ domain-containing protein with the C-terminal three amino acids of MRP2 (TKF, Fig. 2) has been described previously [35].

shows

The epithelial MDCKII cell line is often used to study the polarized sorting of proteins to different plasma membrane domains, however, some proteins are sorted differently in the canine MDCKII cells as compared to polarized kidney cells from other species [52], therefore sorting of human proteins might be different in a canine cell line. We therefore used human hepatoma HepG2 cells that polarize after several days in culture and form apical vacuoles reminiscent of bile canaliculi [49]. Because HepG2 cells endogenously

Fig. 6. Localization of GFP–MRP2 (green in A,B), GFP–MRP2D3 (green in C,D), and GFP–MRP2-T1543A (green in E,F) in polarized MDCKII cells. MDCKII cells grown on Transwell filter membranes were transiently transfected with the respective construct and fixed 24 h after transfection. The apical membrane was visualized by staining with rhodamine-conjugated concanavalin A (red fluores- cence). In the x–y planes (A,C,E), the GFP signals of all three con- structs give a pattern typical for apical localization. The intense yellow color in the x–z planes, due to merging of the green GFP and the red concanavalin A fluorescence, that GFP–MRP2, GFP– MRP2D3, and GFP–MRP2-T1543A are almost exclusively localized in the apical membrane. Bars, 10 lm.

Apical sorting of human MRP2 (Eur. J. Biochem. 269) 1873

(cid:211) FEBS 2002

Table 2. Quantitative analysis of the subcellular localization of C-terminal deletion constructs in polarized HepG2 cells. Data are percentages of cells in which the respective localization of recombinant protein was observed as described in Materials and methods. Cells were observed 2 days after start of transfection. Data are means (cid:139) SD of n (cid:136) 6 (GFP–MRP2D25MAKE, GFP–MRP2D50, GFP–MRP2D100, n (cid:136) 4) transient trans- fections using butyrate-induced cells as described in Materials and methods.

GKIIECGSPEELLQIPGPFYFMAKEAGIENVNSTKF GKIIECGSPEELLQIPGPFYFMAKEAGIE

GKIIECGSPEELLQIPGPFYFMAKE GKIIECGSPEELLQIPGPFYF

GKIIECGSPEELLQIP GKIIECGSPEE

GKIIECGSPEEMAKE

Construct % Apical % Vesicles % ER C-Terminal sequence (1510–1545)

at the N-terminus in order to leave the C-terminus free for possible binding of interacting proteins. Interaction of the C-terminal PDZ-interacting motif with PDZ domain-con- taining proteins seems to require a free C-terminus [37,38]. A comparable approach of N-terminal GFP-tagging was taken for the identification of apical localization signals in the C-termini of CFTR [51] and of the type IIb Na+/Pi co-transporter [53].

GFP–MRP2 GFP–MRP2D7 GFP–MRP2D11 GFP–MRP2D15 GFP–MRP2D20 GFP–MRP2D25 GFP–MRP2D25 MAKE GFP–MRP2D50 GFP–MRP2D100 73 (cid:139) 9 69 (cid:139) 7 65 (cid:139) 7 16 (cid:139) 7 15 (cid:139) 7 8 (cid:139) 3 9 (cid:139) 4 1 (cid:139) 1 1 (cid:139) 1 18 (cid:139) 9 18 (cid:139) 6 20 (cid:139) 4 17 (cid:139) 7 64 (cid:139) 9 59 (cid:139) 14 59 (cid:139) 5 64 (cid:139) 8 35 (cid:139) 7 9 (cid:139) 5 13 (cid:139) 3 15 (cid:139) 4 67 (cid:139) 14 20 (cid:139) 4 33 (cid:139) 15 32 (cid:139) 5 35 (cid:139) 6 65 (cid:139) 6

In contrast to CFTR [51] and the type IIb Na+/Pi co-transporter [53], the N-terminus of MRP2 is located extracellularly [16]. Therefore, a GFP–MRP2 was con- structed in which the GFP moiety was extracellular due to the murine Ig j-chain leader sequence preceding the GFP sequence [47]. This sequence does not function as a signal for apical localization because GFP, when expressed from the pDisplay vector, was targeted to the apical and to the basolateral membrane in equal amounts (Fig. 3). Synthesis of extracellular GFP was also reported for other signal sequences known to direct proteins to the lumen of the ER [54,55]. As expected, GFP–MRP2 was localized to the apical membrane of polarized HepG2 cells whereas GFP was not (Fig. 3).

With this experimental setup, the effect of C-terminal mutations and truncations on apical sorting of MRP2 was investigated. In contrast to our expectations, neither the

synthesize MRP2 [42,46], we used GFP-tagged MRP2 to distinguish between endogenous and recombinant MRP2. Although MRP2 tagged with GFP at its C-terminus localized correctly to the apical membrane in polarized HepG2 cells [1,34] we constructed MRP2 tagged with GFP

Fig. 7. Localization of GFP–MRP2 constructs in vesicular structures in polarized HepG2 cells. HepG2 cells transiently synthesizing GFP– MRP2 (green in A,B) were incubated with LysoTracker Red to stain lysosomes (red in A), or immunostained with an antibody against DPPIV to stain DPPIV-containing vesicles (red in B). Absence of colocalization indicates that GFP–MRP2 is neither present in lyso- somes nor in DPPIV-containing vesicles. Bars, 2.5 lm.

Table 3. Quantitative analysis of the subcellular distribution of GFP–MRP2, GFP–MRP2D15, and GFP–MRP2D25 at different times after trans- fection in polarized HepG2 cells. Data are percentages of cells in which the respective localization of recombinant protein was observed as described in Materials and methods. Data are means (cid:139) SD of four transient transfections. Experiments were performed without butyrate induction.

Time (days) % Apical % Vesicles Construct % ER

GFP–MRP2

GFP–MRP2D15

GFP–MRP2D25

1 2 3 4 1 2 3 4 1 2 3 4 70 (cid:139) 6 81 (cid:139) 4 77 (cid:139) 3 71 (cid:139) 3 9 (cid:139) 2 8 (cid:139) 1 7 (cid:139) 3 5 (cid:139) 3 2 (cid:139) 1 2 (cid:139) 1 2 (cid:139) 1 3 (cid:139) 1 21 (cid:139) 8 8 (cid:139) 4 11 (cid:139) 6 7 (cid:139) 1 55 (cid:139) 10 47 (cid:139) 7 54 (cid:139) 11 47 (cid:139) 9 78 (cid:139) 10 69 (cid:139) 6 69 (cid:139) 8 63 (cid:139) 3 9 (cid:139) 4 11 (cid:139) 3 12 (cid:139) 4 22 (cid:139) 3 36 (cid:139) 10 45 (cid:139) 6 38 (cid:139) 9 47 (cid:139) 8 20 (cid:139) 9 28 (cid:139) 7 29 (cid:139) 7 35 (cid:139) 2

1874 A. T. Nies et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

A C K N O W L E D G E M E N T S

for excellent

We thank Dr Tobias Cantz for contributions to this work and helpful discussion, Dr Blanche Schwappach for helpful discussions on GFP tagging, Dr Wolfgang Hagmann for MRP1 cDNA, and Marion Pfannschmidt technical assistance. This work was supported in part by grants from the Deutsche Forschungsgemein- schaft through SFB 352/B3.

R E F E R E N C E S

PDZ-interacting motif TKF nor the seven-amino-acid extension of MRP2, which is not present in basolaterally localized MRP family members (Fig. 2), was required for apical sorting of GFP–MRP2 in polarized HepG2 cells (Tables 1 and 2). A similar result was obtained with the type IIb Na+/Pi co-transporter, whose C-terminal three amino acids (TVF) strongly resemble a PDZ-interacting motif. However, deletion of these amino acids did not affect the apical localization of the type IIb Na+/Pi co-transporter [53]. Similarly, mutants of the basolateral GABA trans- porter lacking the PDZ-interacting motif were still targeted to the basolateral membrane [56]. Although the C-terminal PDZ-interacting motif of MRP2 is not required for apical sorting, it may be necessary for linking additional regulatory proteins to MRP2 or for clustering of MRP2 in the apical membrane in order to modulate function, as recently discussed for CFTR [57]. In addition, interaction of PDZ domain-containing proteins with internal PDZ-interacting motifs within the MRP2 protein may occur [58,59].

1. Ko¨ nig, J., Nies, A.T., Cui, Y., Leier, I. & Keppler, D. (1999) Conjugate export pumps of the multidrug resistance protein (MRP) family: localization, substrate specificity, and MRP2- mediated drug resistance. Biochim. Biophys. Acta 1461, 377–394. 2. Cole, S.P.C., Bhardwaj, G., Gerlach, J.H., Mackie, J.E., Grant, C.E., Almquist, K.C., Stewart, A.J., Kurz, E.U., Duncan, A.M.V. & Deeley, R.G. (1992) Overexpression of a transporter gene in a multidrug-resistant human lung cancer cell line. Science 258, 1650– 1654.

3. Hipfner, D.R., Gauldie, S.D., Deeley, R.G. & Cole, S.P.C. (1994) Detection of the Mr 190,000 multidrug protein, MRP, with monoclonal antibodies. Cancer Res. 54, 5788–5792.

4. Flens, M.J., Izquierdo, M.A., Scheffer, G.L., Fritz, J.M., Meijer, C.J., Scheper, R.J. & Zaman, G.J. (1994) Immunochemical detection of the multidrug resistance-associated protein MRP in human multidrug-resistant tumor cells by monoclonal antibodies. Cancer Res. 54, 4557–4563.

5. Flens, M.J., Zaman, G.J.R., van der Valk, P., Izquierdo, M.A., Schroeijers, A.B., Scheffer, G.L., van der Groep, P., de Haas, M., Meijer, C.J.L.M. & Scheper, R.J. (1996) Tissue distribution of the multidrug resistance protein. Am. J. Pathol. 148, 1237–1247. 6. Evers, R., Zaman, G.J.R., van Deemter, L., Jansen, H., Calafat, J., Oomen, L.C.J.M., Oude Elferink, R.P.J., Borst, P. & Schinkel, A.H. (1996) Basolateral localization and export activity of the human multidrug resistance-associated protein in polarized pig kidney cells. J. Clin. Invest. 97, 1211–1218.

Whereas the C-terminal 11 amino acids were not required for apical sorting of MRP2, a C-terminal deletion of 15 or more amino acids markedly reduced the percentage of cells in which MRP2 reached the apical membrane (Table 2). Because MRP2 is still observed in the apical membrane in a very low percentage of cells, MRP2 is at least in part delivered into apically-destined vesicles. A truncation of the C-terminus of MRP2 by at least 15 amino acids may cause the loss of a motif required either for efficient fusion of MRP2-containing vesicles with the apical membrane or for stabilization of MRP2 within the apical membrane. More- over, a MRP2 protein truncated by at least 15 amino acids may alter the conformation of the transport protein to such an extent that the misfolded protein is retained in the ER. A single leucine residue was recently shown to be part of a, yet unidentified, motif required for delivery of the type IIb Na+/Pi co-transporter to the apical membrane [53]. Stabi- lization of the GABA transporter in the basolateral membrane has been demonstrated to be mediated by a PDZ-interacting motif [56]. Whereas GABA transporters lacking the PDZ-interacting motif were still targeted to the basolateral membrane they were not retained, but internal- ized into an endosomal recycling compartment.

7. Ko¨ nig, J., Rost, D., Cui, Y. & Keppler, D. (1999) Characterization of the human multidrug resistance protein isoform MRP3 loca- lized to the basolateral hepatocyte membrane. Hepatology 29, 1156–1163.

8. Kool, M., van der Linden, M., de Haas, M., Scheffer, G.L., de Vree, J.M., Smith, A.J., Jansen, G., Peters, G.J., Ponne, N., Scheper, R.J., Elferink, R.P., Baas, F. & Borst, P. (1999) MRP3, an organic anion transporter able to transport anti-cancer drugs. Proc. Natl Acad. Sci. USA 96, 6914–6919.

9. Madon, J., Hagenbuch, B., Landmann, L., Meier, P.J. & Stieger, B. (2000) Transport function and hepatocellular localization of mrp6 in rat liver. Mol. Pharmacol. 57, 634–641.

When the present work was in progress, a study was published describing the PDZ-interacting motif as a signal for apical localization of MRP2 [36]; deletion of the C-terminal three amino acids resulted in localization of MRP2 predominantly to the basolateral membrane of MDCK cells. These observations are in disagreement with our results. However, the differences may be attributable to the expression in the canine MDCK cells of unspecified origin and to tagging of MRP2 at the C-terminus [36] rather than expression of N-terminally tagged MRP2 in human HepG2 cells (Tables 1 and 2) or polarized MDCKII cells (Fig. 6) as described in the present study.

10. Keppler, D., Ko¨ nig, J. & Nies, A.T. (2001) Conjugate export pumps of the multidrug resistance protein (MRP) family in liver. In The Liver: Biology and Pathobiology (Arias, I.M., Boyer, J.L., Chisari, F.V., Fausto, N., Schachter, D. & Shafritz, D.A., eds), pp. 373–382. Lippincott, Williams & Wilkins, New York.

11. Bu¨ chler, M., Ko¨ nig, J., Brom, M., Kartenbeck, J., Spring, H., Horie, T. & Keppler, D. (1996) cDNA cloning of the hepatocyte canalicular isoform of the multidrug resistance protein, cMRP, in hyper- reveals a novel conjugate export pump deficient bilirubinemic mutant rats. J. Biol. Chem. 271, 15091–15098. 12. Schaub, T.P., Kartenbeck, J., Ko¨ nig, J., Spring, H., Do¨ rsam, H., Sta¨ hler, G., Sto¨ rkel, S., Thon, W.F. & Keppler, D. (1999) Expression of the MRP2 gene-encoded conjugate export pump in human kidney proximal tubules and in renal-cell carcinoma. J. Am. Soc. Nephrol. 10, 1159–1169.

In conclusion, the C-terminal 11 amino acids of human MRP2, including the PDZ-interacting motif, are not required for apical sorting in polarized HepG2 cells. However, a C-terminal deletion of at least 15 amino acids prevents efficient delivery of the conjugate export pump MRP2 to the apical membrane either because part of a motif required for apical sorting is lost or because of a conformational change in the transport protein impairing MRP2 maturation.

13. Paulusma, C.C., Bosma, P.J., Zaman, G.J.R., Bakker, C.T.M., Otter, M., Scheffer, G.L., Scheper, R.J., Borst, P. & Oude Elfer- ink, R.P.J. (1996) Congenital jaundice in rats with a mutation in a

Apical sorting of human MRP2 (Eur. J. Biochem. 269) 1875

(cid:211) FEBS 2002

multidrug resistance associated-protein gene. Science 271, 1126– 1127. basolateral determinant unrelated to clathrin-coated pit localiza- tion signals. J. Biol. Chem. 273, 186–193.

29. Dunbar, L.A. & Caplan, M.J. (2001) Ion pumps in polarized cells: sorting and regulation of the Na,K-and H,K-ATPases. J. Biol. Chem. 276, 29617–29620. 30. Simons, K. & Ikonen, E. (1997) Functional rafts in cell mem- 14. Ito, K., Suzuki, H., Hirohashi, T., Kume, K., Shimizu, T. & Sugiyama, Y. (1997) Molecular cloning of canalicular multi- specific organic anion transporter defective in EHBR. Am. J. Physiol. 272, G16–G22. branes. Nature 387, 569–572. 31. Matter, K. (2000) Epithelial polarity: sorting out the sorters. Curr. Biol. 10, R39–R42.

15. Paulusma, C.C., Kool, M., Bosma, P.J., Scheffer, G.L., ter Borg, F., Scheper, R.J., Tytgat, G.N.J., Borst, P., Baas, F. & Oude Elferink, R.P.J. (1997) A mutation in the human canalicular multispecific organic anion transporter gene causes the Dubin– Johnson syndrome. Hepatology 25, 1539–1542.

32. Songyang, Z., Fanning, A.S., Fu, C., Xu, J., Marfatia, S.M., Chishti, A.H., Crompton, A., Chan, A.C., Anderson, J.M. & Cantley, L.C. (1997) Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 275, 73–77.

33. Fanning, A.S. & Anderson, J.M. (1999) PDZ domains: funda- mental building blocks in the organization of protein complexes at the plasma membrane. J. Clin. Invest. 103, 767–772. 16. Cui, Y., Ko¨ nig, J., Buchholz, U., Spring, H., Leier, I. & Keppler, D. (1999) Drug resistance and ATP-dependent conjugate trans- port mediated by the apical multidrug resistance protein, MRP2, permanently expressed in human and canine cells. Mol. Pharma- col. 55, 929–937.

34. Keitel, V., Kartenbeck, J., Nies, A.T., Spring, H., Brom, M. & Keppler, D. (2000) Impaired protein maturation of the conjugate export pump MRP2 as a consequence of a deletion mutation in Dubin–Johnson syndrome. Hepatology 32, 1317–1328.

17. Taniguchi, K., Wada, M., Kohno, K., Nakamura, T., Kawabe, T., Kawakami, M., Kagotani, K., Okumura, K., Akiyama, S. & Kuwano, M. (1996) A human canalicular multispecific organic anion transporter (cMOAT) overexpressed in cisplatin-resistant human cancer cell lines with decreased drug accumulation. Cancer Res. 56, 4124–4129.

35. Kocher, O., Comella, N., Gilchrist, A., Pal, R., Tognazzi, K., Brown, L.F. & Knoll, J.H. (1999) PDZK1, a novel PDZ domain-containing protein up-regulated in carcinomas and mapped to chromosome 1q21, interacts with cMOAT (MRP2), the multidrug resistance-associated protein. Lab. Invest. 79, 1161– 1170. 18. Ishikawa, T., Mu¨ ller, M., Klu¨ nemann, C., Schaub, T. & Keppler, D. (1990) ATP-dependent primary active transport of cysteinyl leukotrienes across liver canalicular membrane: Role of the ATP- dependent transport system for glutathione S-conjugates. J. Biol. Chem. 265, 19279–19286.

36. Harris, M.J., Kuwano, M., Webb, M. & Board, P.G. (2001) Identification of the apical membrane-targeting signal of the multidrug resistance-associated protein 2 (MRP2/MOAT). J. Biol. Chem. 276, 20876–20881. 19. Nies, A.T., Cantz, T., Brom, M., Leier, I. & Keppler, D. (1998) Expression of the apical conjugate export pump, Mrp2, in the polarized hepatoma cell line, WIF-B. Hepatology 28, 1332– 1340.

37. Reczek, D. & Bretscher, A. (2001) Identification of EPI64, a TBC/ rabGAP domain-containing microvillar protein that binds to the first PDZ domain of EBP50 and E3KARP. J. Cell Biol. 153, 191–206.

20. Evers, R., Kool, M., van Deemter, L., Janssen, H., Calafat, J., Oomen, L.C.J.M., Paulusma, C.C., Oude Elferink, R.P.J., Baas, F., Schinkel, A.H. & Borst, P. (1998) Drug export activity of the human canalicular multispecific organic anion transporter in polarized kidney MDCK cells expressing cMOAT (MRP2) cDNA. J. Clin. Invest. 101, 1310–1319. 38. Muth, T.R., Ahn, J. & Caplan, M.J. (1998) Identification of sorting determinants in the C-terminal cytoplasmic tails of the gamma-aminobutyric acid transporters GAT-2 and GAT-3. J. Biol. Chem. 273, 25616–25627.

39. Cormack, B.P., Valdivia, R.H. & Falkow, S. (1996) FACS-opti- mized mutants of the green fluorescent protein (GFP). Gene 173, 33–38. 21. Ito, K., Suzuki, H., Hirohashi, T., Kume, K., Shimizu, T. & Sugiyama, Y. (1998) Functional analysis of a canalicular multi- specific organic anion transporter cloned from rat liver. J. Biol. Chem. 273, 1684–1688.

22. Kartenbeck, J., Leuschner, U., Mayer, R. & Keppler, D. (1996) Absence of the canalicular isoform of the MRP gene-encoded conjugate export pump from the hepatocytes in Dubin–Johnson syndrome. Hepatology 23, 1061–1066.

40. Yang, T., Cheng, L. & Kain, S.R. (1996) Optimized codon usage and chromophore mutations provide enhanced sensitivity with the green fluorescent protein. Nucleic Acids Res. 24, 4592–4593. 41. Rowe, J., Calegari, F., Taverna, E., Longhi, R. & Rosa, P. (2001) Syntaxin 1A is delivered to the apical and basolateral domains of epithelial cells: the role of munc-18 proteins. J. Cell Sci. 114, 3323– 3332. 23. Keppler, D. & Kartenbeck, J. (1996) The canalicular conjugate export pump encoded by the cmrp/cmoat gene. In Progress in Liver Diseases (Boyer, J.L. & Ockner, R.K., eds), pp. 55–67. Saunders, Philadelphia, PA, USA.

42. Cantz, T., Nies, A.T., Brom, M., Hofmann, A.F. & Keppler, D. (2000) MRP2, a human conjugate export pump, is present and transports Fluo-3 into apical vacuoles of HepG2 cells. Am. J. Physiol. 278, G522–G531. 24. Tsujii, H., Ko¨ nig, J., Rost, D., Sto¨ ckel, B., Leuschner, U. & Keppler, D. (1999) Exon-intron organization of the human multidrug-resistance protein 2 (MRP2) gene mutated in Dubin– Johnson syndrome. Gastroenterology 117, 653–660.

43. Keppler, D. & Ko¨ nig, J. (1997) Expression and localization of the conjugate export pump encoded by the MRP2 (cMRP/cMOAT) gene in liver. FASEB J. 11, 509–516. 25. Ikonen, E. & Simons, K. (1998) Protein and lipid sorting from the trans-Golgi network to the plasma membrane in polarized cells. Semin. Cell Dev. Biol. 9, 503–509.

26. Matter, K., Hunziker, W. & Mellman, I. (1992) Basolateral sort- ing of LDL receptor in MDCK cells: the cytoplasmic domain contains two tyrosine-dependent targeting determinants. Cell 71, 741–753.

27. Le Gall, A.H., Powell, S.K., Yeaman, C.A. & Rodriguez-Boulan, E. (1997) The neural cell adhesion molecule expresses a tyrosine- independent basolateral sorting signal. J. Biol. Chem. 272, 4559– 4567.

44. Bakos, E., Hegedus, T., Hollo, Z., Welker, E., Tusnady, G.E., Zaman, G.J., Flens, M.J., Varadi, A. & Sarkadi, B. (1996) Membrane topology and glycosylation of the human multidrug resistance-associated protein. J. Biol. Chem. 271, 12322–12326. 45. Sonnhammer, E.L., von Heijne, G. & Krogh, A. (1998) A hidden Markov model for predicting transmembrane helices in protein sequences. Proc. Int. Conf. Intell. Syst. Mol. Biol. 6, 175–182. 46. Jedlitschky, G., Leier, I., Buchholz, U., Hummel-Eisenbeiss, J., Burchell, B. & Keppler, D. (1997) ATP-dependent transport of bilirubin glucuronides by the multidrug resistance protein MRP1 and its hepatocyte canalicular isoform MRP2. Biochem. J. 327, 305–310. 28. Distel, B., Bauer, U., Le Borgne, R. & Hoflack, B. (1998) Baso- lateral sorting of the cation-dependent mannose 6-phosphate receptor in Madin–Darby canine kidney cells. Identification of a

1876 A. T. Nies et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

sion of type IIb NaPi cotransporters. Proc. Natl Acad. Sci. USA 97, 2916–2921.

47. Coloma, M.J., Hastings, A., Wims, L.A. & Morrison, S.L. (1992) Novel vectors for the expression of antibody molecules using variable regions generated by polymerase chain reaction. J. Immunol. Methods 152, 89–104.

54. Blum, R., Stephens, D.J. & Schulz, I. (2000) Lumenal targeted GFP, used as a marker of soluble cargo, visualises rapid ERGIC to Golgi tra(cid:129)c by a tubulo-vesicular network. J. Cell Sci. 113, 3151–3159.

55. McCarthy, J.B., Lim, S.T., Elkind, N.B., Trimmer, J.S., Duvoisin, R.M., Rodriguez-Boulan, E. & Caplan, M.J. (2001) The C-terminal tail of the metabotropic glutamate receptor subtype 7 is necessary but not su(cid:129)cient for cell surface delivery and polarized targeting in neurons and epithelia. J. Biol. Chem. 276, 9133–9140.

48. Gronwald, R.G., Grant, F.J., Haldeman, B.A., Hart, C.E., O’Hara, P.J., Hagen, F.S., Ross, R., Bowen-Pope, D.F. & Mur- ray, M.J. (1988) Cloning and expression of a cDNA coding for the human platelet-derived growth factor receptor: evidence for more than one receptor class. Proc. Natl Acad. Sci. USA 85, 3435–3439. 49. Sormunen, R., Eskelinen, S. & Lehto, V. (1993) Bile canaliculus formation in cultured HepG2 cells. Lab. Invest. 68, 652–662. 50. Zegers, M.M. & Hoekstra, D. (1997) Sphingolipid transport to the apical plasma membrane domain in human hepatoma cells is controlled by PKC and PKA activity: a correlation with cell polarity in HepG2 cells. J. Cell Biol. 138, 307–321.

56. Perego, C., Vanoni, C., Villa, A., Longhi, R., Kaech, S.M., Frohli, E., Hajnal, A., Kim, S.K. & Pietrini, G. (1999) PDZ–mediated interactions retain the epithelial GABA transporter on the baso- lateral surface of polarized epithelial cells. EMBO J. 18, 2384– 2393. 57. Bezprozvanny, I. & Maximov, A. (2001) PDZ domains: more than just a glue. Proc. Natl Acad. Sci. USA 98, 787–790. 51. Moyer, B.D., Denton, J., Karlson, K.H., Reynolds, D., Wang, S., Mickle, J.E., Milewski, M., Cutting, G.R., Guggino, W.B., Li, M. & Stanton, B.A. (1999) A PDZ-interacting domain in CFTR is an apical membrane polarization signal. J. Clin. Invest. 104, 1353– 1361.

58. Cuppen, E., Gerrits, H., Pepers, B., Wieringa, B. & Hendriks, W. (1998) PDZ motifs in PTP-BL and RIL bind to internal protein segments in the LIM domain protein RIL. Mol. Biol. Cell 9, 671–683.

52. Roush, D.L., Gottardi, C.J., Naim, H.Y., Roth, M.G. & Caplan, M.J. (1998) Tyrosine-based membrane protein sorting signals are differentially interpreted by polarized Madin–Darby canine kid- ney and LLC-PK1 epithelial cells. J. Biol. Chem. 273, 26862– 26869.

53. Karim-Jimenez, Z., Hernando, N., Biber, J. & Murer, H. (2000) Requirement of a leucine residue for (apical) membrane expres- 59. Hillier, B.J., Christopherson, K.S., Prehoda, K.E., Bredt, D.S. & Lim, W.A. (1999) Unexpected modes of PDZ domain scaffolding revealed by structure of nNOS-syntrophin complex. Science 284, 812–815.