Localization and Genomic Organization of a New Hepatocellular Organic Anion Transporting Polypeptide*

Based on sequence homology to the human organic anion transporting polypeptide 2 (OATP2; SLC21A6), we cloned a new member of the SLC21A superfamily of solute carriers, termed OATP8 (SLC21A8). The protein of 702 amino acids showed an amino acid identity of 80% with human OATP2. Based on Northern blotting, the expression of OATP8 was restricted to human liver. Cosmid clones containing the genes encoding human OATP1 (SLC21A3), OATP2 (SLC21A6), and OATP8 (SLC21A8) served to establish their genomic organization. All three genes contained 14 exons with 13 identical splice sites when transferred to the amino acid sequence. An antibody raised against the carboxyl terminus localized OATP8 to the basolateral membrane of human hepatocytes and the recombinant glycoprotein, expressed in MDCKII cells, to the lateral membrane. Transport properties of OATP8 were studied in stably transfected MDCKII and HEK293 cells. Organic anions transported by human OATP8 included sulfobromophthalein, with aK m of 3.3 μm, and 17β-glucuronosyl estradiol, with a K m of 5.4 μm. Several bile salts were not substrates. Thus, human OATP8 is a new uptake transporter in the basolateral hepatocyte membrane with an overlapping but distinct substrate specificity as compared with OATP2, which is localized to the same membrane domain.

The gene superfamily of solute carriers (SLC) 1 comprises a growing family of different types of transporters. Their substrates comprise a broad range of different structures ranging from inorganic compounds to organic cations and anions as well as amino acids and small peptides. Within this superfamily, the uptake transporters for organic anions are classified into the organic anion/prostaglandin transporter family SLC21A (Human Gene Nomenclature Committee Data Base). Best characterized within this gene family are the organic anion transporting polypeptides (OATPs) (1). In humans, five different organic anion transporting polypeptides have been identified and cloned. Two of them, human OATP1 (SLC21A3) (2) and OATP2 (SLC21A6) (3)(4)(5) have been characterized with respect to their transport characteristics, tissue distribution, and cellular localization. Human OATP1 (originally termed OATP) (2) was cloned as an ortholog of the rat Oatp family member Oatp1 (6). As determined in several expression systems, transported substrates of human OATP1 include sulfobromophthalein (BSP) (2, 6 -8), estrone-3-sulfate (9), dehydroepiandrosterone sulfate (10), and 17␤-glucuronosyl estradiol (E 2 17␤G) (11,12). These substances were also identified as substrates or competitive inhibitors of the recently cloned second member of the human OATPs, OATP2 (SLC21A6) (5). Despite the fact that both proteins share only 44% amino acid identity, they have a similar substrate specificity and a similar predicted topology (5). However, they differ with respect to their tissue-specific expression. Human OATP1 was cloned from liver but exhibits the highest expression in brain (2). In contrast, human OATP2 shows a high expression only in liver as tested by Northern blotting (3)(4)(5). Recently, the sequences of three additional members of the human OATP family designated OATP-B (SLC21A9), OATP-D (SLC21A11), and OATP-E (SLC21A12) were reported (see Table I for accession numbers). The amino acid sequences of these three transporters are quite distinct from human OATP1 and OATP2. Until now, the transporters OATP-B, -D, and -E were not further characterized.
So far, little is known about the genomic organization of the human OATP family members. The human OATP1 gene has been localized to chromosome 12p12 (13), and its promoter region as well as its regulatory properties have been studied (14). In this study (14), the length of the first exon and the first exon/intron splice site were determined, but the complete genomic organization of the OATP1 gene has not been reported. Moreover, no information is currently available on the genomic organization and the chromosomal localization of the other human OATP family members.
Therefore, we screened a genomic data base using the human OATP2 cDNA to identify a cosmid clone containing the genomic sequence of this transporter. We identified one cosmid clone containing cDNA sequences that were highly homologous to but distinct from OATP2 cDNA sequences. By combining these genomic DNA sequences, we identified a new OATP cDNA, which was termed OATP8 (SLC21A8). Based on this partial cDNA sequence, we cloned the full-length cDNA and analyzed the genomic organization of this transporter in comparison with the genomic organization of human OATP1 and human OATP2. Northern blot analysis revealed a liver-specific expression similar to that of OATP2. An OATP8-specific antibody raised against the carboxyl terminus served to detect the protein in the basolateral hepatocyte membrane of human liver. The recombinant protein was localized to the lateral membrane of stably transfected MDCKII cells. In addition to these MDCKII cells, stably transfected HEK293 cells served to determine transport characteristics of this new transporter OATP8.

EXPERIMENTAL PROCEDURES
Materials-Lysozyme and ampicillin were from Roche Molecular Biochemicals, agarose was from Roth (Karlsruhe, Germany), RNase inhibitor (RNAguard), and Stratascript Moloney murine leukemia virus reverse transcriptase, and restriction endonucleases were from Stratagene (Amsterdam, The Netherlands). Leupeptin, pepstatin, aprotinin, fetal calf serum, agar, and the protein standard mixture (relative molecular weight of 26, 600 -180,000) were from Sigma. The human 12lane multiple tissue Northern blot, the ␤-actin primers, the Marathon-Ready cDNA kit, and the Advantage cDNA polymerase mix were from CLONTECH (Heidelberg, Germany).
Antibodies-The carboxyl terminus of the deduced OATP8 sequence (SKTCNLDMQDNAAAN) (see Fig. 1) was used to raise the polyclonal SKT antibody in rabbits. The peptide was synthesized automatically and then coupled to maleimide-activated keyhole limpet hemocyanin. Rabbits were immunized with this conjugate. The mouse monoclonal antibody to desmoplakin and the mouse monoclonal antibody to cytokeratin 19 were purchased from Progen (Heidelberg, Germany). The mouse monoclonal antibody to human dipeptidylpeptidase IV (CD26; anti-DPPIV) was from Pharmingen (San Diego, CA). Cy2-conjugated goat anti-rabbit and Cy3-conjugated goat anti-mouse antibodies were from Jackson ImmunoResearch Laboratories (West Grove, PA).
Immunofluorescence Studies of Human Liver Sections-Normal liver tissue samples were obtained from patients who underwent liver resection because of hepatocellular carcinoma. Only healthy tissue, left after tumor surgery, was used for immunofluorescence studies. Liver tissue was deep-frozen in liquid nitrogen and mounted in Tissue-Tek (Miles, Elkhart, IN). Sections of 5 m were prepared with a cryotome (FrigoCut, 2800E; Leica, Nussloch, Germany), air-dried for 2 h, and fixed in acetone (Ϫ20°C) for 10 min. Sections were then incubated with the primary antibodies for 60 min at room temperature, washed three times with phosphate-buffered saline (PBS) (140 mM NaCl, 10 mM phosphate, pH 7.4), and subsequently incubated with the secondary antibodies. Finally, sections were washed twice in PBS, rinsed with distilled water, and mounted in Moviol (Hoechst, Frankfurt, Germany). Antibodies were prepared in PBS supplemented with 5% fetal calf serum at the following dilutions: SKT and CD26 at 1:100; anti-desmoplakin at 1:20; anti-cytokeratin 19 at 1:5; and Cy2-conjugated antirabbit IgG and Cy3-conjugated anti-mouse IgG at 1:400. Fluorescence microscopy was performed on an Axiovert S100TV microscope (Carl Zeiss, Jena, Germany) as described (5).
Cloning of the Human OATP8 cDNA-On the basis of a GenBank TM search against the human OATP2 cDNA sequence (EBI/GenBank accession number AF132573) one cosmid clone was identified (EBI/Gen-Bank accession number AC006582) that exhibited a high degree of base pair identity over several parts of the OATP2 cDNA. This clone contained several unordered partial genomic sequences derived from chimpanzee chromosome 12. An overlay of the OATP2 cDNA sequence and this genomic sequence revealed a partial cDNA sequence of 1726 bp that exhibited 87% identity to the human OATP2 cDNA within the 5Ј-cDNA region. Based on this sequence information two primers were designed located on both ends of this new OATP8 cDNA. Using human liver cDNA from the Marathon-Ready cDNA kit (CLONTECH) and the forward primer oOATP8.for (5Ј-CACGTGGTATCTGTAGTTTAATA-ATG-3Ј) and the reverse primer oOATP8.rev (5Ј-TAAATGTGGTAC-CTCCTGTTGCAG-3Ј), the first human OATP8 cDNA-fragment was amplified. In detail, the polymerase chain reaction was performed in a volume of 50 l containing 5 l of Marathon-Ready cDNA, 0.2 M forward primer, and 0.2 M reverse primer, 5 l of 10ϫ polymerase chain reaction buffer (400 mM Tricine-KOH, pH 9.2, 150 mM potassium acetate, 35 mM magnesium acetate, 750 mg/liter bovine serum albumin), 0.2 mM dNTPs, and 1 l of 50ϫ Advantage cDNA polymerase mix (50% glycerol, 40 mM Tris-HCl, pH 7.5, 50 mM KCl, 25 mM (NH 4 ) 2 SO 4 , 0.1 mM EDTA, 5 mM ␤-mercaptoethanol, 0.25% Thesit, 1.1 mg/ml TaqStart antibody, and KlenTaq-1 DNA polymerase) under the following cycling conditions: 2 min denaturation at 94°C followed by 5 cycles denaturation at 94°C for 10 s and annealing/elongation at 68°C for 2 min, five cycles of denaturation at 94°C for 10 s and annealing/elongation at 66°C for 2 min and subsequently 25 cycles with denaturation at 94°C for 10 s and annealing/elongation at 65°C for 2 min. The polymerase chain reaction was completed by an incubation at 72°C for 10 min. The amplified fragment was analyzed on an agarose gel and subcloned into pCR2.1.TOPO (Invitrogen, BV, Groningen, The Netherlands), resulting in the plasmid pOATP8.5Ј. The lacking 3Ј-end of the human OATP8 cDNA was amplified using a 3Ј-rapid amplification of cDNA reaction with a primer based on the cDNA sequence from the OATP8 partial clone pOATP8.5Ј. In detail, the forward primer oOATP8.3Ј-rapid amplification of cDNA (5Ј-GATGATATCCTTCTT-GTTTCAACTTC-3Ј) and the reverse primer AP1 (supplied with the Marathon-Ready cDNA kit) were subjected to a 3Ј-rapid amplification of cDNA reaction using the Marathon-Ready cDNA kit (CLONTECH) under the same conditions as mentioned above. The amplified fragment of about 1.4 kilobase pairs was subcloned into the vector pCR2.1.TOPO (Invitrogen), resulting in the plasmid pOATP8.3Ј. Based on these sequence informations, the full-length OATP8 cDNA was amplified under the same conditions using the liver Marathon-Ready cDNA (CLON-TECH) and the primer oOATP8.for against the reverse primer oOATP8.3Јrev (5Ј-TATAGATGCATAGACTTATCCAT-3Ј) located within the 3Ј-untranslated region. The amplified fragment of 2241 bp was subcloned into the vector pCR2.1.TOPO (Invitrogen), resulting in the plasmid pOATP8.TOPO. For transfection studies, the OATP8-cDNA was subcloned into the expression vector pcDNA3.1(ϩ) (Invitrogen). Therefore, clone pOATP8.TOPO was restricted with BstXI, and the resulting OATP8 cDNA fragment was subcloned into the BstXI-digested and dephosphorylated vector pcDNA3.1(ϩ), resulting in the plasmid pOATP8.31. The correctness of cloning and orientation were verified by sequencing.
DNA Sequencing-The cDNA clones were sequenced by use of [␣-35 S]dATP and the T7 sequencing kit (Amersham Pharmacia Biotech) according to the dideoxynucleotide chain termination method of Sanger et al. (15). Dried gels were exposed to Kodak BioMax MR-1 films (Sigma).
Sequence Analysis-Throughout this study the computer program Heidelberg Unix Sequence Analysis Resource (16), based on the Wisconsin Genetics Computer group program package (17), was used for restriction mapping, sequence analysis, and sequence alignments.
Northern Blot Analysis-The Northern blot analyses were performed using the commercial human 12-lane multiple tissue Northern blot from CLONTECH. For OATP8, the 589-bp EcoRI restriction fragment (bp 2059 -2647) from clone pOATP8.3Ј and for the control a human ␤-actin cDNA fragment supplied with the kit were used as probes. The multiple tissue Northern blot membrane was prehybridized at 42°C for 2 h in 10 ml of hybridization buffer (6ϫ SSC, 0.5% SDS, 5ϫ Denhardt's solution, 50% formamide, 100 g/ml denatured salmon sperm DNA). DNA fragment labeling was performed using the Rediprime DNA labeling system (Amersham Pharmacia Biotech) according to the manufacturer's instructions. The labeled fragments were purified using NucTrap purification columns from Stratagene. After hybridization, the membrane was washed once in 2ϫ SSC, 0.1% SDS at 42°C for 20 min followed by two washing steps at 60°C for 20 min, one step with 1ϫ SSC, 0.1% SDS, the following by 0.5ϫ SSC, 0.1% SDS. The blot was air dried, and autoradiography was performed at Ϫ80°C with an intensifying screen for 36 h (OATP8) and 18 h (␤-actin).
Cell Culture and Transfection Studies-Human embryonic kidney (HEK293) and Madin-Darby canine kidney (MDCKII) cells were cultured in minimum essential medium (Sigma), supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C, 95% humidity, and 5% CO 2 . Cells were transfected with the polybrene (hexamethrin bromide) method as described earlier (5). After 3 weeks of G418 selection (600 g/ml), single colonies were screened for OATP8 expression by immunoblot analysis and immunofluorescence microscopy. Expression of recombinant OATP8 was further enhanced by culturing cells with 10 mM sodium butyrate (18).
Preparation of Membrane Vesicles-Basolateral membrane vesicles from human liver (19) and crude membrane vesicles from cultured cells were prepared as described earlier (18).
Immunoblot Analysis-Membrane proteins were diluted with sample buffer and incubated at 37°C for 30 min prior to separation on 4% stacking and 10% resolving SDS-polyacrylamide gels. Immunoblotting was performed using a tank blotting system from Bio-Rad and an enhanced chemiluminescence detection kit (NEN Life Science Products). Primary antibody was diluted 1: 5000 in TTBS (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween 20). The secondary antibody was a horseradish peroxidase-conjugated goat anti-rabbit IgG (Bio-Rad) used at a 1: 2000 dilution.
Deglycosylation and Tunicamycin Treatment-Basolateral membrane proteins from human liver were deglycosylated using peptide N-glycosidase F (EC 3.5.1.52) as described earlier (5). To study the glycosylation of the recombinant protein, OATP8-transfected HEK293 cells were cultured with 1 g/ml tunicamycin for 48 h. Crude membrane vesicles were then prepared from the treated cells and studied by immunoblot analysis.
Immunofluorescence Microscopy of Transfected Cells-Transfected HEK293 or MDCK cells were grown on Transwell membrane inserts (pore size, 3 m; Costar, Cambridge, MA). Sodium butyrate was added to the culture medium 24 h before the experiment. HEK293 cells were fixed and permeabilized for 5 min in methanol (Ϫ20°C), MDCK II cells were fixed for 10 min with 4% paraformaldehyde in PBS and permeabilized for 10 min in 1% Triton X-100 in PBS. Cells were thereafter incubated with the primary antibody SKT (diluted 1:50 in PBS) for 30 min at room temperature. After three washes with PBS, cells were incubated with Cy2-conjugated goat anti-rabbit IgG (diluted 1:200 in PBS) for 30 min at room temperature. Nuclei were stained with propidium iodide (0.2 g/ml) added into the dilution of the secondary antibody. Membranes were cut from the inserts and mounted onto slides with 50% glycerol in PBS. Confocal laser scanning immunofluorescence microscopy was performed using a LSM-510 apparatus from Carl Zeiss (Oberkochen, Germany).
Transport Assays-Transfected HEK293 cells were seeded in 6-well plates (coated with 0.1 mg/ml poly-D-lysine) at a density of 2 ϫ 10 6 cells/well and cultured with 10 mM sodium butyrate for 24 h. For uptake studies cells were first washed with uptake buffer (142 mM NaCl, 5 mM KCl, 1 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 1.5 mM CaCl 2 , 5 mM glucose, and 12.5 mM 4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid, pH 7.3) and then incubated with 1 ml of uptake buffer containing tritiumlabeled substrate at 37°C. 3 H-labeled substrates were obtained from NEN Life Science Products. 3 H-labeled BSP was obtained through custom synthesis from Hartmann Andlytik (Koln, Germany). For cisinhibition studies reduced glutathione (5 mM) was included into the uptake buffer. At different time points uptake was stopped by adding ice-cold buffer. After three washes with cold buffer, cells were lysed in 1 ml of 0.2% SDS in water. The cell-associated radioactivity was determined by transferring 250-l aliquots of the lysate to scintillation vials and counting the radioactivity with a Beckman scintillation counter (LS 6000IC, Beckman Instruments GmbH, Munich, Germany). Protein content was determined according to Lowry using 100 l of cell lysate.

RESULTS
Cloning of the Human OATP8 cDNA-A GenBank search using OATP2 (SLC21A6) cDNA as reference sequence revealed a chimpanzee genomic sequence (GenBank/EBI accession number AC006582) that exhibited a high percentage of identity over several parts of the cDNA. An overlay of the OATP2 cDNA sequence (GenBank/EBI accession number AJ132573) with this genomic DNA sequence, taking into consideration the conserved exon/intron splice junction sites, resulted in a new partial cDNA sequence of 1726 bp in length. This new cDNA sequence was located in the 5Ј-part of the human OATP2 cDNA and exhibited 87% base pair identity. Based on this information, the full-length human cDNA of this new member of the organic anion transporter family was cloned and further analyzed. The coding region covers 2106 bp and shows the highest degree of identity to human OATP2 (SLC21A6) of about 87% as calculated under the default parameters of the Bestfit program Comparison of the Genomic Organization of the Human OATP8 (SLC21A8) Gene with the Human OATP2 (SLC21A6) and OATP1 (SLC21A3) Genes-After completion of cloning and sequencing of the OATP8 cDNA, a data base search using the full-length cDNA revealed another genomic clone (GenBank/ EBI accession number AC011604), that covers the complete gene of OATP8. The genomic locus of the gene is chromosome 12p12-31.7 to 12p12-37.2. In similar approaches, data base searches using the human OATP1 cDNA (GenBank/EBI accession number NM_005075) and the human OATP2 cDNA (Gen-Bank/EBI accession number AJ132573) revealed two genomic clones, with the GenBank/EBI accession numbers AC022224 and AC022335, covering the human OATP1 (SLC21A3) and OATP2 (SLC21A6) gene, respectively. Like OATP8, the OATP1 gene is located on chromosome 12p12 (13). Using the OATP cDNAs and the identified genomic sequences, the organization of these OATP genes was determined (Fig. 1). All exons were flanked by the dinucleotide GT on the 5Ј-donor site and AG on the 3Ј-acceptor site, which is consistent with the consensus sequences for splice junctions in eukaryotic genes (20). Despite the fact that OATP1, when compared with OATP2 and OATP8, had a different amino acid composition, the genomic organization was very similar (Fig. 1). All three genes contained 14 exons with an average exon size of about 150 bp. For OATP2 and OATP8, 13 exons were identical in length. A comparison of the genomic organization of human OATP8 and OATP1 revealed 9 exons identical in length. For all three OATP family members, the longest exon is exon 14 (OATP1, 220 bp plus 3Ј-untranslated region; OATP2, 211 bp plus 3Ј-untranslated region; and OATP8, 244 bp plus 3Ј-untranslated region). The shortest exon in all three transporter genes is exon 12 with 65 bp. A comparison of the splice junction sites of the OATP1, OATP2, and OATP8 genes on the amino acid level shows that all 13 splice sites are at identical positions (Fig. 1). Because the DNA sequences used for the analysis of the genomic organization comprised several pieces released in an unordered fashion, the length of the introns could not be determined.

Analysis of the Deduced Amino Acid Sequence of OATP8 -
The open reading frame of the OATP8 cDNA of 2106 bp encodes a protein of deduced 702 amino acids in length with a calculated molecular mass of 77,402 Da. The amino acid identity of OATP8 to other members of the organic anion transporter family were calculated (16) using the Heidelberg Unix Sequence Analysis Resource tree alignment program (Table I). This analysis demonstrated that the OATP8 transporter ( Fig.  1) shares the highest degree of amino acid identity with human OATP2 (SLC21A6) (3)(4)(5). The identity with other OATPs varied between 28% for human OATP-E and 40% for human OATP1 (Table I). Based on a CLUSTAL alignment a computeraided transmembrane analysis using the TopPred II program (21,22) demonstrates that OATP8 consists of 12 predicted transmembrane domains. This transmembrane organization is identical to the one predicted for the other known OATPs (1, 5), underlining the relationship of human OATP8 with the other members of the OATP family. A comparison of OATP8 and OATP2 with respect to potential N-glycosylation sites revealed 11 glycosylation sites for both proteins with four of them located outside in predicted extracellular loops for OATP8 but with six predicted outside glycosylation sites for OATP2. Both proteins share predicted N-glycosylation of the fifth extracellular loop.
Tissue Distribution of Human OATP8 -Tissue distribution of human OATP8 mRNA was studied by taking into account the high nucleotide identity between human OATP2 and OATP8. Therefore, an OATP8 cDNA restriction fragment was used as probe that shared only 78% identity to the human OATP2 cDNA. To prevent cross-hybridization of this OATP8 cDNA fragment with OATP2 mRNA, the Northern blot was performed under high stringency hybridization and washing conditions. As for OATP2 mRNA, a strong signal for OATP8 mRNA was only detected in liver (Fig. 2). Prolonged exposure of the blot for up to 84 h revealed no additional signals in other tissues, suggesting that human OATP8 (Fig. 2), like OATP2 (3)(4)(5), is highly and apparently exclusively expressed in human liver. The length of the detected mRNAs was 2.8 kb, likely corresponding to the fully spliced mRNA, and 4.5 kb, probably corresponding to unspliced or partially spliced mRNA.
Immunofluorescence Localization of OATP8 in Hepatocytes-Incubation of cryosections from human liver tissue with the SKT antibody yielded fluorescent staining of the basal and lateral plasma membrane domains of hepatocytes (Fig. 3, A, D,  G, and H). Hepatocytes near the central vein showed stronger   FIG. 2. Analysis of OATP8 mRNA expression in different human tissues. A human 12-lane multiple-tissue Northern blot was hybridized using a 589-bp EcoRI restriction fragment as probe. This fragment was located at the 3Ј-end of the OATP8 mRNA and shares only 79% identity to the human OATP2 mRNA. Using stringent hybridization and washing conditions, a cross-hybridization of the labeled probe with OATP2 mRNA was minimized. with related human transporters of the OATP (SLC21A) family Identities are given in relation to human OATP8 (black box). Analysis was performed using the tree alignment from the Heidelberg Unix Sequence Analysis Resource program package (16). Gaps in the consecutive numbering of the gene symbols have been reserved for yet unknown human orthologs of rat OATPs. The SLC21A gene nomenclature is based on the Human Gene Nomenclature Committee Data Base. staining than hepatocytes close to the portal vein (Fig. 3, G and  H), indicating a lobular zonation of protein expression. Absence of OATP8 from the canalicular membrane was confirmed by double labeling of cryosections with the SKT antibody and an antibody against the apical plasma membrane marker dipeptidylpeptidase IV (Fig. 3, A-C). Expression of OATP8 in the lateral membrane was demonstrated by colocalization of OATP8 with desmoplakin (Fig. 3, D-F). OATP8 was not detected in cholangiocytes (Fig. 3H), which were identified by staining of a consecutive cryosection with an antibody to cytokeratin 19 (Fig. 3I) as a marker for bile ductular cells (23).
Immunolocalization of Recombinant OATP8 in Transfectants-OATP8 showed a similar localization as OATP2 in transfected cells. OATP8 was sorted to the lateral membrane in the polarized MDCKII cells (Fig. 4, B and D). In the nonpolarized HEK293 cells, OATP8 was localized over the entire plasma membrane (Fig. 4, A and C).
Immunoblot Analysis of Endogenous and Recombinant OATP8 -The specificity of the polyclonal antibodies SKT (Fig.  1) and ESL (5) was first established in immunoblot analyses using crude membranes from HEK293 cells transfected either with OATP2 (HEK-OATP2) or with OATP8 (HEK-OATP8). As shown in Fig. 5, antibody SKT was specific for OATP8 and antibody ESL was specific for OATP2. The recombinant OATP8 in HEK293 cells has an apparent molecular mass of about 90 kDa. When the cells were treated with tunicamycin to inhibit N-glycosylation, the molecular mass of the recombinant OATP8 was reduced to about 65 kDa (Fig. 5). In the basolateral membrane from human liver, however, the polyclonal antibody SKT detected a signal with an apparent molecular mass of about 120 kDa. Deglycosylation with peptide N-glycosidase F reduced the molecular mass to about 65 kDa (Fig. 5). Apparently, OATP8 is heavily glycosylated in the hepatocytes. Interestingly, when OATP8 was expressed in the hepatocyte-derived cell line HepG2, it showed a similar molecular mass of about 90 kDa as in transfected HEK293 or MDCKII cells (not shown). Thus, the glycosylation of OATP8 in the liver differs from the glycosylation in the transfected cultured cells.
Transport Activity of Recombinant OATP8 -Several organic anions were tested as substrates for uptake by OATP8-transfected HEK293 cells. Uptake of BSP by OATP8 was time-dependent (Fig. 6). Transport activity of OATP8 was not sodiumdependent because replacement of sodium chloride by choline chloride did not significantly affect the transport activity of OATP8. The apparent K m values for BSP and E 2 17␤G were 3.3 and 5.4 M, respectively. Leukotriene C 4 and dehydroepiandrosterone sulfate (Table II) were also transported by recombinant human OATP8. The bile salts cholate, cholyl taurine, and cholyl glycine tested under the same experimental conditions at a substrate concentration of 5 M were not taken up into OATP8-transfected cells at an enhanced rate.
Reduced glutathione, an important endogenous antioxidant, was tested both in cis-inhibition studies and in direct uptake experiments. At a concentration of 5 mM, reduced glutathione did not cis-inhibit OATP8-mediated uptake of 1 M BSP. Moreover, no significant difference in uptake rates of [ 3 H]reduced glutathione, at a concentration of 100 M, was detected between HEK293 cells expressing OATP8 and HEK293 cells transfected with control vector alone, suggesting that reduced glutathione is not a substrate for OATP8.  (5), and OATP8 was detected by the polyclonal antibody SKT (Fig. 1). A, ESL detected only OATP2 and SKT detected only OATP8. B, glycosylation of OATP8 in human liver. OATP8 has an apparent molecular mass of about 120 kDa in comparison to 90 kDa of OATP2. The high molecular mass of OATP8 in hepatocytes is a result of different glycosylation as deglycosylation with N-peptidyl glycosidase F reduced the molecular mass of OATP8 to about 60 kDa. C, glycosylation of OATP8 in transfected HEK293 cells. Recombinant OATP8 in the transfected HEK293 cells is glycosylated differently as compared with liver. Both OATP2 and OATP8 have an apparent molecular mass of about 90 kDa. After treatment of the cells with tunicamycin (10 g/ml), the molecular mass of OATP8 was reduced to about 65 kDa.

DISCUSSION
We have cloned and characterized a new member of the human OATP family belonging to the solute carrier superfamily. The newly identified transporter is the eighth member of the SLC21A family, and the gene symbol accordingly is SLC21A8. By following the numbering of the gene nomenclature, the translation product of this new gene was termed OATP8. All known OATPs within the SLC family exhibit common features. They all are glycosylated and have a similar secondary structure with 12 predicted transmembrane helices (1,5). Within the OATP family, OATP8 shows the highest degree of identity, both on nucleotide and amino acid level, to human OATP2 (4,5,24). Both are preferentially, if not exclusively, expressed in human liver (Fig. 2) (4,5,24), whereas other OATPs are expressed in a variety of tissues as shown for human OATP1 with its expression in brain, kidney, liver, and testis (2).
The genomic organization of human OATP1 (SLC21A3), OATP2 (SLC21A6), and OATP8 is very similar. All three genes contain 14 exons with an average exon size of about 150 bp. All 13 introns share identical phases with a high proportion of class 1 phases, where the codon is interrupted between the first and the second nucleotide (Fig. 1). The genomic organization of human OATP2 (SLC21A6) and OATP8 (SLC21A8) is closely related with 13 exons identical in length. The genomic organization of OATP8 compared with OATP1 revealed 9 exons identical in length. Fig. 1 demonstrates that the 5 exons differing in length exhibit gaps in the amino acid alignment leading to the different number of base pairs. OATP1 has been mapped to chromosome 12p12 (13), and for OATP8 the genomic clone allows for localization to chromosome 12p12-31.7 to 12p12-37.2. Thus, human OATP1 (SLC21A3), OATP2 (SLC21A6), and OATP8 (SLC21A8) differ with respect to their amino acid composition, but their genomic organization indicates that they are closely related and possibly originate from a gene amplification.
The cellular localization of OATP8 was examined by the antibody SKT directed against the carboxyl terminus of the protein (Fig. 1). This antibody did not cross-react with human OATP2 (Fig. 5). The OATP8 glycoprotein was detected in the basolateral membrane of human hepatocytes (Fig. 3), which is the same membrane domain to which human OATP2 has been localized (5). Furthermore, this is the same localization as demonstrated for human OATP1 (1) and for several rat OATP family members (25)(26)(27) except for rat OATP1, which is apically localized in kidney (28) and brain (29). Interestingly, OATP8 was not detected in bile ductular cells (Fig. 3, H and I). These cells also lack expression of human OATP2 (5).
To investigate the transport properties of the recombinant glycoprotein, we used transfected MDCKII and HEK293 cells overexpressing human OATP8. In polarized MDCKII cells transfected with OATP8 cDNA, the cellular localization was restricted to the lateral membrane, whereas in the nonpolarized HEK293 cells the recombinant protein was localized to the whole plasma membrane (Fig. 4). Using the HEK293 cells stably expressing the recombinant protein, several typical organic anions including BSP, E 2 17␤G, and dehydroepiandrosterone sulfate were transported by OATP8 (Table II and Fig.  6). The apparent K m values for BSP and E 2 17␤G were 3.3 and 5.4 M, respectively. These K m values were in the same range as the ones for transport by recombinant rat Oatp1 (25). The K m value for E 2 17␤G was comparable with the one determined with human OATP2 (5). A major difference between the hepatic basolateral uptake transporters OATP2 and OATP8 concerns their ability to transport certain bile salts. In contrast to human OATP2, which transports cholyl taurine and whose 17␤glucuronosyl estradiol transport was inhibited by cholate (5), OATP8 did neither transport cholyl taurine nor cholate nor cholyl glycine. Both transporters thus have overlapping substrate specificities as well as clear differences that require further detailed characterization. An additional marked difference between OATP2 and OATP8 became apparent in the deglycosylation experiments indicating that OATP8 in human liver is much more glycosylated than OATP2 (Fig. 5). Deglycosylation yields a similar molecular mass for both proteins that corresponds to the mass predicted on the basis of the number of amino acids. It will be of interest to examine in more detail a role of the different oligosaccharide moieties of OATP8 and OATP2 in human hepatocytes.
In summary, human OATP8 represents the third member of the human OATP family of transporters that was cloned, localized, and functionally examined. OATP8 is the second uptake transporter, in addition to OATP2 (3)(4)(5), which is predominantly, if not exclusively, expressed in human liver and mediates the sodium-independent, high affinity uptake of important physiological amphiphilic organic anions into hepatocytes. It will be of interest to identify additional mutually exclusive endogenous substrates for human OATP2 and OATP8, in addition to cholate and cholyl taurine.