WO1997031111A2 - A family of organic anion transporters, nucleic acids encoding them, cells comprising them and methods for using them - Google Patents

Abstract

The present invention provides a novel family of organic anion transporters of which until now only one member was known. The family includes multispecific organic anion transporters related to the canalicular multispecific organic anion transporter. The cDNA encoding the latter is also provided. The rat and human cDNA are exemplified. Uses of nucleic acids based on this gene family and of cells comprising such nucleic acids as well as vectors comprising sequences thereof are also disclosed especially in the area of gene therapy.

Description

Title: A family of organic anion transporters, nucleic acids encoding them, cells comprising them and methods for using them

The present invention lies in the field of molecular biology and genetic engineering. It is particularly

concerned with mechanisms of transport for substances across cell membranes. More in particular it is concerned with transport of cytotoxic substances from the inside to the outside of cells.

A group of proteins involved in transport of molecules across membranes is the group of so-called ABC-transporters (ABC: ATP-bindιng cassette). One member of this group called MRP1 (multidrug resistance-associated protein) has been identified as being involved in transporting organic anions across cell membranes. This protein thus transports

different substances than the P-glycoprotem encoded by the MDR-1 gene (MDR: multidrug resistance ) . MDR- 1 is involved in the occurrence of multidrug resistance of for instance tumor cells Multidrug resistance is one of the major problems in chemotherapy of cancer. On the other hand providing cells with multidrug resistance may be very useful in rescuing for instance bone marrow when chemotherapy is applied. Thus on the one hand there is a need for being able to prevent transport of cytotoxic substances out of the cell while on the other hand there is a need to be able to enhance

transport of cytotoxic substances from cells. The P- glycoprotein encoded by MDR-1 is not capable of transport of all cytotoxic substances; its binding specificity is limited to certain groups of molecules. MRP-1 has a different binding specificity in that it transports anionic organic compounds, possibly complexed or conjuged with other

substances. In the liver a protein has been characterized functionally which is an ATP-dependent non-bile salt organic anion transporter called canalicular Multispecific Organic Anion Transporter. This protein has been thought to be identical to MRP-1. The liver plays a major role in the detoxification of many endogenous and xenobiotic, lipophilic compounds.

Detoxification is accomplished by transferase-mediated conjugation with glutathione-, glucuronide-, or sulphate- moieties, resulting in negatively charged, amphiphilic compounds which are efficiently secreted into bile or urine. Hepatobiliary excretion of these conjugates is mediated by an ATP-dependent transport system, the canalicular

Multispecific Organic Anion Transporter (cMOAT), located in the apical (canalicular) membrane of the hepatocyte (1). The identification of a transport-deficient mutant rat strain, the TR- rat (2), has contributed to the functional

characterization of cMOAT (1). These rats have an autosomal recessive defect in the hepatobiliary excretion of bilirubin glucuronides (3) and other multivalent organic anions including, glutathione-S-conjugates (e.g. leukotriene C₄ ), and 3-OH-glucuronidated and -sulphated bile salts (4). Thus far neither a protein nor a complete cDNA encoding cMOAT have been identified. Transport studies in plasma membrane vesicles from cells overexpressing the human Multidrug

Resistance-associated Protein 1 (hMRP1) (5), demonstrated a role for hMRP1 in the ATP-dependent transport of the

glutathione conjugates LTC4 and dinitrophenyl glutathione

(GS-DNP) (6). Because these substrates are also transported by the putative cMOAT protein, MRP1 as stated before, is a possible candidate gene for cMOAT. A recent study suggested lateral as well as canalicular localization of the rmrpl gene product in Wistar liver, but only a lateral

localization in TR- liver (14) and suggested a role for MRP1 in the (defective) hepatobiliary excretion of organic anions in TR- rats. In our view, however, the extremely low MRP1 expression in liver (7,5) renders it unlikely that this gene product is responsible for biliary organic anion secretion. Furthermore, the transport defect in the TR- rat appears to be specific for liver (9), while MRP1 is expressed in all human tissues (7). We have now found that cMOAT is encoded by a different gene and thus that a family of organic anion transporters exists.

The present invention now provides a nucleic acid comprising a sequence encoding at least a part of a member of a family of organic anion transporters, said nucleic acid comprising at least a gene family specific fragment of one of the sequences of fig. la or fig. lb or figs 17, 18 or 19 or the complement thereof, or a sequence having at least 55%, preferably 70%, in particular 90% homology therewith. Of this family sofar only one member, mammalian MRP was known.

We hypothesized that cMOAT might be a liver-specific homologue of MRP1. To obtain a rat mrpl probe, we applied the polymerase chain reaction (PCR) on rat lung cDNA using nested degenerate oligonucleotide primers which were based on the highly conserved first ATP-binding cassette of the hMRP1 sequence (see experimental part). The 213 base pair product obtained shared 83% amino acid sequence identity with the corresponding region of the hMRP1 sequence. When analyzed on Northern (RNA) blot, this PCR fragment

hybridized with a single, 9.5-kb, transcript in all Wistar and TR- rat tissues examined, with high expression in lung and testis, but no detectable expression in liver. Because this expression pattern resembled that of hMRP1 in human tissues (7), we assumed that we had isolated a part of the rat homolog of hMRP1, rat mrpl ( rmrpl ) . In order to find the putative cjnoat gene, two rat liver cDNA libraries were screened, using the rmrpl fragment obtained as a probe (see the experimental part). This resulted in the isolation of a full length cDNA with a single open reading frame of 1541 amino acids (Fig. 1b). Based on similarity searches (10), the protein was identified as a new member of the ABC- transporter family (11), with modest identity to other members of the family. Highest overall identity was found with hMRP1 (47.6%) (5), yeast Cadmium Factor 1 (41.8%) (12), and the human Cystic Fibrosis Transmembrane Conductance Regulator (30.2%) (13). The amino acid sequence identity with hMRP1 ranged from 38-61% outside the ATP-binding domains to 67% and 75% in the first and second ATP-binding domain, respectively. Recently, two different partial rmpl cDNA sequences were disclosed (14). One of these sequences comprised a 347 nt fragment that closely resembles the rat cmoat cDNA sequence found by us, said partial sequence is, however, not identical to cmoat . Moreover, no relation has been made between said partial sequence and the putative cMOAT protein. In contrast, it was suggested that a mutant mrpl gene is responsible for the cMOAT-deflcient phenotype.

Northern (RNA) blot analysis of rat tissues with a 1-kb restriction fragment of our isolated cDNA, revealed three different transcripts, ranginq from approximately 6 5 to 9 5-kb, with high expression onϋy in liver, and low

expression in kidney, duodenum, and lleum (Fig 2A) . These transcripts were strongly decreased (but not absent) in liver (Fig 2B) and other tissues of the TR- rat, which suggests that these transcripts were related to the defect in the TR- rat The three transcripts observed were probably derived from a single gene, because the level of all three transcripts was decreased in the TR- rat. The decrease of this transcript in TR- liver suggests that the isolated cDNA encoded cmoat.

To examine the expression level and the cellular localization of the cMOAT protein in hepatocytes, we

produced a monoclonal antibody (mAb M₂ III- 5) to a bacterial fusion protein containing the 202 -amino acid carboxyl- terminal end of the sequence (see experimental part). On protein blots this antibody detected a protein of

approximately 200-kD in the canalicular, but not the

basolateral plasma membrane fraction of the Wistar rat liver (Fig. 3). This molecular weight was very similar to that of hMRP1 and in good agreement with the predicted molecular weight of the cMOAT protein. The 200-kD protein was

completely absent from the canalicular membrane fraction of the TR- rat (fig.3), which correlated with the decreased mRNA level in TR- rat liver (Fig. 2B). Again, this finding was in good agreement with the defect observed in TR- rats which lack a functional transport system for organic anions in the canalicular membrane .

Thus we have isolated the complete cDNA encoding the cMOAT protein, which is deficient in the TR- rat. Since the cmoa t mRNA was not completely absent in TR- liver it was possible to also amplify the complete TR- cmoa t cDNA by PCR (see the experimental part) using various specific primer sets. To identify the nature of the genetic defect in TR- rats, we sequenced the obtained cDNA. This revealed a 1-bp deletion at ammo acid position 393, which results in a frame-shift and subsequent introduction of a stop-codon at position 401 (Fig. 4) . This deletion results in the

destruction of a N1alll restriction site which provided a means to quickly confirm the mutation in cDNAs from various tissues (see the experimental part) . The very low mRNA expression in TR- rats (Fig 2B) might be due to the fact that the frame shift causes premature termination of translation and subsequent increased degradation of the mRNA.

Our results show a correlation between the cmoa t gene, the absence of the gene product from the canalicular membrane, and the defined congenital transport defect in TR- rats In addition to the exclusive canalicular localization of cMOAT (Fig 3), we have found that hMRP1 is routed only to the lateral domain of the plasma membrane of pig kidney epithelial cells (see experimental part) Our findings thus suggest a differential localization of MRP1 (basolateral) and cMOAT (canalicular) and imply that cMOAT and not MRP1 is involved in biliary organic anion transport This contrasts to suggestions made in the literature (14). It was also suggested (14) that an isoform of MRP1 exists in rat liver which is derived from the same gene by alternative splicing based on the detection of two different sequences for the second ATP-binding domain and only one for the first ATP- binding domain Our complete cDNA data, howevei, show that there are also two different sequences for the first ATP- bindmg domain in mrpl and cmoa t In fact, the two cDNA sequences differ considerably throughout the entire

molecule, thus indicating that MRP1 and cMOAT are encoded by two different genes.

We conclude that the MRP homolog, identified here, encodes the canalicular Multispecific Organic Anion

Transporter, and that a 1-bp deletion, resulting in a truncated, non-viable, protein, is responsible for impaired transport of organic compounds from liver to bile in the TR- rat. TR- rats have the same phenotype as patients with the Dubin-Johnson syndrome, characterized by mild chronic conjugated hyperbilirubinemia (15). Isolation of the human homolog of cmoat is required to elucidate the nature of the defect in humans. Overexpression of hMRP1 confers resistance of human tumor cells to a number of cytostatic drugs (16, 17), and this resistance is dependent on intracellular glutathione levels (18). Apparently, both MRP1 and cmoat are involved in the excretion of organic anions from cells.

Thus, overexpression of cMOAT, like that of MRP1, might also confer resistance to cancer cells against cytostatic drugs or their metabolites.

Using the rat cmoat gene we also found and isolated the cDNA encoding the human cMOAT protein. Now that it is known that these two exist other species of this family of

transporters can be found using the present invention.

These transport mechanisms occur throughout the living world, so family members can be found in bacteria, bacilli, yeasts and fungi, plants, invertebrae, vertebrae, in

particular in mammalians.

In addition to MRP1 and cMOAT (MRP2) , other MRP

homologs encoding GS-X pumps are present in the human genome, considering that there are at least four MRP

homologs expressed in Caenorhabditis elegans (56). We therefore searched the Expressed Sequence Tag (EST) library (57) for putative human MRP homologs, and found three more MRP homologs expressed in humans. We call these new MRP homologs MRP3 , MRP4 , and MRP5 . To investigate a possible role of MRP homologs in drug resistance, we examined a large set of (multι)drug resistant cell lines for the (over) expression of cMOAT, MRP3, MRP4 , and MRP5. We find that especially cMOAT expression is elevated in several cell lines, selected for cisplatin resistance, and also in some sublines of the human non-small lung cancer cell line SW1573/S1, selected for doxorubicm resistance. The expression level of cMOAT correlates with the cisplatin but not the doxorubicin resistance of these cell lines. Although MRP3 and MRP5 were overexpressed in some resistant cell lines, no clear correlation between drug resistance and the expression levels of MRP3 , MRP4 , and MRP5 has emerged from these studies as yet.

Preferred for the purposes of this invention are closely related members of the members identified by the sequences of fig. la and fig.1b, most preferred those members which transport similar or the same compounds when expressed in a cell, or the closely related family members identified herein as MRP 3, 4 and/or 5. Most preferred is the human cmoat gene or its human family members and their products for their usefulness in for instance gene therapy and for their use in preparing blocking agents to che transporting product.

Further embodiments include but are not limited to a vector comprising a nucleic acid according to the invention and suitable means for replication, transduction and/or expression of said nucleic acid.

Preferably such a vector further comprises a gene encoding a therapeutically beneficial protein, which may be any protein having a beneficial effect under certain

circumstances such as giving glutathion elevating activity, which enhances transport of anionic complexes or conjugates by the invented transporters.

Such vectors include vectors wherein the gene encodes at least a functional part of a gamma glutamyl cysteine synthetase or a UDP-glucuronosyltransferase. Other vectors according to the invention include vectors wherein the therapeutically beneficial protein is another multidrugresistance related protein such as MDR1 The invention further provides a cell comprising a nucleic acid or a vector according to the invention Said cells may be any cells, preferred are bone marrow progenitor cells, in particular hematopoietic stem cells.

If said vector thus not encode additional desired functionalities apart from the cMOAT activity as disclosed above, said activity may be present on a separate vector to be introduced into said cell.

The invention also provides a method for providing cells with Canalicular MultispecifIc Organic Anion Transport protein activity, comprising transducing said cell with a nucleic acid or a vector according to tne invention, as well as a method for enhancing Canalicular Multispecific Organic Anion Transport protein activity of cells according to the invention, comprising increasing the intracellulax level of glutathion, glucuronide and/or sulphate This may be done by contacting the cell with for instance glutathion esters, but also by providing additional genetic material as disclosed above. This may be done by cotransducmg UDP- glucosedehydrogenase or sulphotransferase or any other means of enhancing such activity.

Thus the invention also encompasses methods for enhancing Canalicular Multispecific Organic Anion Transport protein activity of cells according to the invention, comprising enhancing the conjugating capacity and/or the complexing activity of said cell for sulphate, glutathion, glucuronide and the like.

On the other hand the invention provides a method for reducing Canalicular Multispecific Organic Anion Transport protein activity and/or the multidrug resistance of a cell comprising providing said cell with an antisense construct of a nucleic acid or a vector according to the invention, which antisense constructs are thus also part of the present invention. These methods can be used to block or at least reduce transport of substances by the transporter protein according to the invention thus reducing resistance of for instance tumor cells to certain chemotherapeutic substances Other ways of blocking the invented transporter are also part of the invention. These include methods of reducing the level of the conjugating or complexing molecules that enhance transport by the invented transporter Antibodies to (in particular the extracellular domain of) the transporter For even further reducing multidrug resistance of for instance tumor cells said cells can be further provided witn an antisense construct derived from another multidrug resistance related protein such as MDR1.

Proteins encoded by a nucleic acid according to the invention or obtainable by expression of a vector according to the invention are of course also part of the present invention, in particular proteins having Canalicular

Multispecific Organic Anion Transport protein activity or Canalicular Multispecific Organic Anion Transport protein specific antigenicity comprising at least part of the sequence of fig 4 or being encoded by at least a part of the sequences of MRP 2, MRP 3 or MRP 4 (as given in the

accompanying figures) or derivatives thereof having the same or similar function.

In the following a number of uses of the molecules, cells and methods of the invention are disclosed.

The invention enables the use of a nucleic acid

according to the invention or a protein according to the invention in the diagnosis of Dubm-Johnson disease, Rotor disease or another disease involving Canalicular

Multispecific Organic Anion Transport protein, as well as the use of a nucleic acid according to the invention or a protein according to the invention in the treatment of

Dubin-Johnson disease, Rotor disease or another disease involving Canalicular Multispecific Organic Anion Transport protein.

Furthermore the nucleic acids according to the

invention can be used as a selectable marker gene. The members of the gene family disclosed herein have several useful applications in the context of gene therapy.

The concept of gene therapy has a very broad range of applications with one common denominator and that is the transfer of additional, new or corrected genetic information into cells which have a genetic or acquired defect. Examples of genetic disorders eligible for gene therapy are cystic fibrosis, Duchenne's Muscular Dystropy, cancer, Gaucher disease, Cπgler Najjar and Dubin-Johnson syndrome Examples of acquired diseases are cancer, viral and parasitic diseases . In addition , gene transfer can augment the efficacy of conventional therapies. Vehicles for the

transfer of genes into target cells and tissues include vectors of viral and non -viral origin. Among the viral vectors munne based retroviruses and human based

adenoviruses are the preferred embodiments .

Retroviruses are RNA viruses which efficiently

integrate their genetic information into the genomic DNA of infected cells via a reverse-transcribed DNA intermediate as a proviral copy. Integration into the host's genome and the tact that parts of their genetic material can be replaced by foreign DNA sequences make retroviruses one of the more lucrative vectors for gene delivery in human gene therapy procedures, most notably for gene therapies which rely on gene transfer into dividing tissues. Recombinant murine retroviruses have been the vectors of choice since the start of gene therapy and several clinical trials using

recombinant retroviruses are ongoing. In order to generate a recombinant retrovirus which carries the cDNA sequence of a particular gene one needs to introduce the retroviral construct into an appropriate packaging cell line. The retroviral construct carries the cDNA of interest and the cis acting elements for packaging and transcription of the viral RNA genome. The packaging cell line provides the trans acting factors needed for packaging: the gag, pol and env genes. Expression of the retroviral construct into the packaging cell line results in the production of recombinant retroviral particles capable of transducing susceptible target cells and transferring a particular therapeutic gene. The recombinant retrovirus is stably integrated into the target cell genome and conferred to its daughter cells upon cell division.

Adenoviruses are non-enveloped DNA viruses. The genome consists of a linear, double stranded DNA molecule of about 36 kb. Recombinant adenovirus vectors have been generated for gene transfer purposes. Recombinant adenoviruses can be generated by co-transfection of two E1 -deleted recombinant adenoviral DNA constructs, one of which comprising the sequences of interest, into an El-expressing cell line. In contrast to retroviruses, adenoviruses do not integrate into the host cell genome, are able to infect non-dividing cells and are able to efficiently transfer recombinant genes in vivo. These features make adenoviruses attractive candidates for in vivo gene transfer into target cells which are difficult or impossible to treat ex vivo, such as cells of lung and liver.

Although the skilled artisan will be α.ble to employ other vector systems than those exemplified here, such as Adeno Associated Virus (AAV) , adenoviruses and retroviruses are preferred embodiments, because of the extensive

experience with these viruses in gene therapy concepts.

Vectors comprising nucleic acids encoding and

expressing functional members of the family of organic anion transporters disclosed in the present invention are of particular importance for the treatment of diseases caused by defects in these transporters. Examples of such diseases include Dubin-Johnson syndrome, Rotor syndrome and other cholestatic disorders. The human Dubin-Johnson syndrome

The earliest evidence for distinct canalicular

transport systems for bile acids and non-bile acid organic anions came with the recognition of the human Dubin-Johnson syndrome; this is a rare congenital chronic conjugated hyperbilirubinemia. The hepatic clearance of bilirubin and other cholephilic organic anions, like BSP and indocyanine green, is impaired in these patients, whereas bile acid clearance is normal The urinary excretion of

coproporphyrins, metabolic by-products of heme synthesis, is normal but the proportion of coproporphyπn isomer I is increased The liver histology of this syndrome is

characterized by lysosomal pigment accumulation.

Preferred target tissues for the genetic treatment of these diseases include the liver, gut and kidney

Retroviral vectors comprising the nucleic acid

sequences disclosed in this invention are constructed as exemplified in EP/95 201211.0 incorporated herein by

reference Recombinant retrovirus supernatant stocks are produced by introduction of the retroviral constructs in appropriate retroviral packaging cell lines. Adenoviral vectors comprising the nucleic acid sequences disclosed in this invention are constructed as exemplified in

EP/95 . 202213 5 incorporated herein by reference Adenovirus stocks are produced by transfectmg the adenoviral construct into appropriate E1 complementing cell lines.

Hematopoietic stem cells (HSC) are the source for life- long production of all mature blood cell types. Therefore, genetic correction of HSC is expected to result in permanent mitigation of the clinical manifestation of inherited and aquired hematopoietic diseases . Of particular interest are Gaucher disease, thalassemia, sickle-cell anemia, AIDS, and others exemplified in WO93/07281. This makes HSC attractive targets for gene therapy. Currently, the established

procedure for gene transfer into long-term repopulating HSC relies on the use of recombinant retroviral vectors.

However, the gene transfer efficiency into human HSC is insufficient for the treatment of most hematopoietic

diseases This forms the bottle-neck for a broader

application of bone marrow gene therapy Therefore, it is preferred to provide the recombinant retroviral vector with a marker sequence for positive selection of transduced cells. Selection for the presence of this sequence can be performed in vitro by culturing the transduced cells in the presence of a selective drug. Another approach is to select for transduced cells in vivo, following transplantation of transduced HSC. Both approaches can be taken by inclusion in the recombinant retroviral construct genes encoding

transporter proteins conferring resistance to cytostatic drugs.

The members of the family of organic anion transporters disclosed in the present invention are important examples of genes that can be used for this purpose. Another important embodiment of the present invention is the use of the disclosed members of the family of organic anion transporters to provide the hematopoietic system of cancer patients with resistance to chemotherapeutic drugs . This makes increased dose- intensity in the chemotherapeutic treatment of cancer possible. For most anticancer drugs increasing the dose-intensity results in increased response rates and a higher proportion of cures. Dose- intensity is the amount of drug administered per unit time, and can be augmented either by increasing the chemotherapy dose or by reducing intervals between cycles. Dose- intensive

chemotherapy can produce complete regressions and improve survival in patients with historically refractory solid tumors and non-Hodgkin's lymphomas .Dose-response

relationships have been demonstrated for many anticancer drugs. The major dose-limiting toxicity of many anticancer drugs is myelosuppression, which thus prevents optimum dose- intensity administration. Severe myelosuppression makes the patient particularly prone to opportunistic infections and is a frequent reason for curtailing chemotherapy before an adequate therapeutic response has been obtained. With a few exeptions (e.g., hormones) most anticancer drugs used in the clinic today, cause moderate to severe bone marrow toxicity (e.g., vinblastine, epipodophyllotoxins, cisplatin,

carboplatin, melphalan, methotrexate, alkylating agents, nitrosoureas, anthracyclines and anthraquinones).

Increasing the therapeutic index of myelosuppressive anticancer drugs by retrovirus-mediated transfer of genes encoding proteins conferring drug resistance is an

attractive prospect. A recombinant retrovirus encoding a mutated dihydrofolate reductase (DHFR) that is highly resistant to the anticancer drug methotrexate has been constructed. Infection of murine bone marrow cells with this retroviral vector and subsequent reconstitution of lethally irradiated mice conferred protection from methotrexate- induced marrow toxicity. Furthermore, it has been

demonstrated that transfer of the MDR1 cDNA into drug- sensitive cells can introduce drug resistance, in vitro as well as in vivo. Members of the family of drugs extruded from the cell by the MDR1 drug pump are e.g. anthracyclines, vinca alkaloids, podophyllotoxins, and colchicine .

Etoposide, a commonly used podophyllotoxin of which the dose-limiting toxicity is restricted to the hematopoietic system, is also pumped by the MλRl encoded drug pump albeit only poorly. The MDR related drugs have in common that they are lipophilic compounds derived from various natural products. In general, MDR cells are not cross-resistant to alkylating agents (e.g., chlorambucil and cyclophosphamide), antimetabolites (e.g., cytarabine, methotrexate, and 5- fluorouracil), cisplatin, carboplatin or melphalan. The members of the family of organic anion transporters

disclosed in the present invention efficiently extrude organic anion compounds from the cell, including GS-DNP and chemotherapeutic agents such as the conjugated forms of cisplatin, carboplatin, etoposide, chlorambucil, and

melphalan. This is shown in a nonlimiting example for cisplatin below. Therefore, introduction of the disclosed transporters into the hematopoietic system of cancer

patients allows dose- intensification of frequently used chemotherapeutic drugs such as etoposide and are different from those protected against by MDR1 or mutant DHFR .

Therefore introduction of members of the disclosed invention into the hematopoietic system could lead to increased efficacy of cancer treatment.

The invention will be described in greater detail in the following experimental part.

Experimental

Example 1. Identification and isolation of the rat cmoat (mrp2 ).

A 213-bp PCR product was obtained from rat lung cDNA after first round amplification with degenerate primers

corresponding to ammo acid residues 678-648 (forward) and 770-776 (reverse), and subsequent second round amplification v/ith nested primers corresponding to ammo acid residues 694-700 (forward) and 760-766 (reveise) of the M4RP1 sequence (5) .

Partial cDNA clones were isolated from a rat hepatocyte cDNA library (23) which was screened with the 213-bp probe according to standard procedures (24) From a 4 5-kb positive clone a 5'-located, 0 6-kb HphI restriction fragment was used to screen a gtlO 5'-stretch rat liver cDNA library (Clontech, Palo Alto) A 0 8-kb overlapping clone was obtained from which a 0 6-kb Avail probe was isolated to rescreen the same library, resulting in the isolation of another overlapping clone. The 5' end of the cDNA was obtained using the anchored PCR procedure [M.A Frohman, M.K. Dush, G.R Martin, Proc Natl Acad Sci .

U.S. A 85, 8998 (1988)] cDNA synthesis was carried out with 5 μg of total RNA isolated from Wistar rat liver and random hexamer primers using Superscript Reverse Transcriptase II. After purification the cDNAs were tailed with a synthetic oligonucleotide anchor sequence using a 5 ' -RACE kit (Life Technologies, Gaithersburg). Two rounds of nested PCR (96 C, 30sec; 60°C, 30sec; 72°C, 45sec) using an anchor specific primer and two c/noat-specific primers (5'- tgtccagtatcttctgtgagcg-3' (first round), 5'- aacacgacgaacacctgcttggc-3 ' (nested) ) resulted m the

isolation of the missing 5'-sequence. Probes were labeled with [a-³²P]dCTP using random primers. Hybridization of the filters was performed at 65°C in 0.5 M NaPO₄ (pH 7 0), 2 mM EDTA, and 7% SDS (hybridization solution), for 20 hours. Filters were washed four times in 2x SSC, 1% SDS for 30 min at 65°C, and autoradiographed. Nucleotide sequences were determined by the dideoxy-nucleotide chain method [F.Sanger, S. Nicklen, A.R. Coulson, Proc. Natl. Acad. Sci. U.S.A. 74, 463 (1977)]. The cmoat sequence is being submitted to the Genbank database and is available under accession number L49379.

A fusion gene, consisting of the gene for the

Escherichia Coli maltose-binding protein, and the 3' part of the cmoa t cDNA corresponding to amino acid residues 1340-

1541, was constructed in pMal-c [C.V. Maina et al . , Gene 74, 365 (1977)]. The fusion protein was produced in E.coli strain JM101 and purified by amylose resin affinity

chromatography. Mice were injected three times over six weeks with 200 μg of the purified protein. The first

injection was in the presence of Freund's complete adjuvant, and the subsequent boosts in Freund's incomplete adjuvant. Two weeks after the final boost with a glutathione-5- transferase-cMOAT fusion protein, splenocytes were isolated and fused with myeloma cells. Hybridomas were screened by ELISA with the glutathione-S-transferase-cMOAT fusion protein and subsequently tested for positivity in Western blots.

cmoat cDNA was amplified from liver, kidney, ileum and duodenum from both Wistar and TR- rats using primers

corresponding to amino acid residues 366-375 (forward) and 451-458 (reverse) of the cmoat sequence. The obtained PCR product was digested with Nialll . In all PCR products from TR- rat digestion produced two bands of 206 and 66 bp whereas in the Wistar three bands of 83, 122 and 67 bp were observed.

Total RNA was extracted using the acid-phenol single step method [P. Chomczynski and N. Sacchi, Anal . Biochem . 8, 148 (1987)]. Poly(A)⁺ RNA was isolated using the polyAtract mRNA system III ( Promega, Madison). RNA was fractionated on a 0.8% denaturating agarose gel, transferred to Hybond N⁺ nitrocellulose membrane filters and hybridized with a [a- ³²P]dCTP-labeled 213 -bp rat lung mrpl probe and a 1-kb HindiII/AvalI fragment of cmoat m hybridization solution (11) for 20 hours at 65ºC Filters were washed 4x30 mm m 0 2x SSC/0 1% SDS at 65°C and autoradiographed A ³²P- labeled 1 2-kb PstI fragment of the rat glyceraldehyde-3- phosphate dehydrogenase cDNA [Ph Forth et al , Nucleic Acid Res 13, 1431 (1985)] was used to estimate variations in RNA loading.

Canalicular and basolateral membranes were isolated as described by [P J Meier, E S Sztul A Reuben J L

Bover J Cell Biol 98, 991 (1984)] Membranes, containing 50 μg of protein were fractionated by 7 5% SDS

polyacrylamide gel electrophoresis, electrophoretically transterred to nitrocellulose filters blocked for at least 2h in PBS/M/T (phosphate-buffered saline containing 1% BSA and 1% milk powder and 0 05% Tween-20), and incubated with the monoclonal antibody (M₂ III -5 hybπdoma culture medium diluted eightfold with PBS/M/T) for 2h Immunoreactivity was visualized with peroxidase-conjugated rabbit anti-mouse lmmunoglobulms and subsequent staining with 3,3'- diaminobenzidine and 4-chloro-1-naphthol substrate P- glycoprotems were detected using the monoclonal antibody C219 and peroxidase-conjugated rabbit anti-mouse IgG Immune complexes were visualized by enhanced chemiluminescence detection.

Example 2. Isolation and characterization of the human cMOAT

The human homolog of the rat cMOAT cDNA was isolated using a 4 kb fragment of the rat cMOAT cDNA The fragment was labelled as described for the rat cMOAT cDNA . The labelled probe was then used to screen a human lambda gtll liver cDNA library. Three clones with inserts hybridizing with the rat cMOAT cDNA sequence were isolated and

designated clone 12,7 and 20. Clone 12 contained an insert of 2716 nucleotides comprising coding sequence 130-2846

Clone 7 contained an insert of 2000 nucleotides comprising coding sequence 2517-3185. Clone 20 contained an insert of 2231 nucleotides comprising the coding sequence 3069-5300 Missing nucleotides 1-130 encompassing the translation initiation site were obtained from the Washϋ-Merck EST library, clone 1243479. Furthermore noncoding 3' sequences were found to be present in additional EST clones and were used to complete the full coding sequence of the human cMOAT cDNA. Clones 193244 and 199655 were used for this purpose and completed the full length sequence from 5300 to 5582 nucleotides.

Example 3. Transport experiments with rat cMOAT transfectants.

The rat cMOAT cDNA was cloned into the mammalian expression vector pSVK3 (Pharmacia). pSVK3-rat-cMOAT and ρSVK3 with rat-cMOAT in the reverse orientation (pSVK3 -rat-cMOAT/Rev) relative to the promotor were transfected into COS-7 cells grown in 75 cm² tissue culture flasks. Three days after transfection, the cells were used for GS-DNP transport experiments and analyzed for cMOAT protein expression using anti-cMOAT antibodies. For transport measurements the cells were washed with Hanks buffer and loaded with Hanks/¹⁴C-CDNB at 15'C. Samples were taken after various time points.

Input CDNB and cell mediated formation and transport of GS- DNP were separated by extraction of the samples with water- saturated ethylacetate. The water phase which contains the excreted ¹⁴C labelled GS-DNP was counted in a scintillation counter. Total protein was determined using the Lowry assay GS-DNP efflux from the transfected cells was measured m a scintillation counter . Relative to COS - 7 cells

transfected with pSVK3 -rat-cMOAT/Rev, cells transfected with pSVK3-rat - cMOAT excreted two fold more GS-DNP (See figure 5). This suggests that rat cMOAT transfected cells express a functional organic anion transporter protein in line with the expression of a protein reactive with anti-cMOAT antibodies.

COS- 7 cells transfected as described above were also used to isolate membrane vesicles and perform transport experiments. For this purpose cell homogenates were prepared from

transfected COS-7 cells and were centrifuged over a

discontinous gradient of 19, 38 and 56 % sucrose The 38-19 % interface was collected and revesiculated and total protein content was determined using the Lowry method The vesicle suspensions were incubated wuth ³H-GS-DNP at 37 "C in the presence of an ATP regenerating system

After the indicated time points the vescile suspensions were filtered Then the filters were washed with ice-cold stop buffer (250 mM sucrose, 20 mM HEPES/Tris pH = 7 4 ) and counted in a scintillation counter In agreement with the cell transport experiments, vesicles isolated from cells expressing rat-cMOAT exhibited GS-DNP transport above the level of transport observed with vesicles isolated from pSVK3-rat-cMOAT/Rev transfected COS-7 cells. This transport was completely dependent on the presence of ATP

characteristic for a member of the ABC transporter

superfamily (figure 6).

Example 4 : Transport experiments with human cMOAT trans fectants

We have attempted to express the human cMOAT protein in several mammalian cell lines and in the yeast Saccharomyces cerevisiae with various expression constructs carrying the human cMOAT cDNA . From 80 independent human cMOAT

transfected and cloned human SI cells only 2 clones

expressed the human cMOAT protein at low levels These expressing cells were tested for functional cMOAT by plating the cells in varying concentrations of doxorubicin,

vincristin and cisplatinum.

The human LLC-PK1 cell line was also transfected with the same human cMOAT DNA construct and 90 clones were screened for expression None of these clones expressed the human cMOAT protein as detected with antibody M₂III-6. In

contrast, expression of human cMOAT in yeast was also studied and was high after the translational core sequences of the human cMOAT cDNA were converted to yeast consensus sequences.

However in both cases no active cMOAT mediated transport of GS-DNP could be observed. To investigate the role of cellular polarity in determining functional expression of the human cMOAT protein we have infected MDCK cells with an amphotropic retrovirus carrying the human cMOAT cDNA . For this purpose a Hindlll-Ncol DNA fragment containing the complete predicted human cMOAT open reading frame was cloned into the retroviral vector pCMV-nec (Benαer et al., 1987) resulting in a construct designated pCMV-neo-human - cMOAT . The retroviral amphotropic packaging cell line Phenix kindly provided by G.P. Nolan, Stanford University Medical Center, Stanford, USA) was cultured in Iscove's with 10% fetal calf serum. Phenix cells were transfected with pCMV- neo-human-cMOAT DNA using a commercially available calcium phosphate transfection kit (Gibco/BRL). After 16h at 5% CO₂

, 37 ºC, medium was changed and growth was continued for 48 h followed by collection of recombinant retrovirus

containing medium. Storage of virus supernatants was at -20 °C. The MDCK cell line strain II (MDCKII; Louvard) was used for transduction experiments. For this purpose 2 x 10⁵

MDCKII cells were seeded and incubated with a 5 ml 1/10 diluted virus stock in medium containing 30 mg Transfection Reagent (DOTAP; Boehringer Mannheim, Germany). After 10 h medium was replaced with fresh medium. Thirty six hours after infection cells were trypsinized and seeded at dilutions varying between 1/12-1/64. Stably infected cells were selected for 2-3 weeks in medium with G418 at 200 mg/ml . Thirty clones were picked and analyzed for the presence of hcMOAT protein. Western blot analysis of crude membrane fractions of these clones revealed that several clones contained a substantial, but between individual clones variable, amount of human cMOAT . Two of these clones are shown in Figure 7. A weak signal was observed in wild- type MDCKII cells with a slightly higher molecular weight than hcMOAT after prolonged exposure. This might either represent canine cMOAT or another protein to which this mAt cross-reacts. The subcellular distribution of human cMOAT protein was determined in MDCKII -217 hcMOAT transfectants, which showed the strongest signal in Western blot analysis (See figure 7). Cells were grown to confluency on

microporous polycarbonate membrane filters (3 mm pore size, 24.5 mm diameter, Transwell™ 3414; Costar Corp., Cambridge, MA) at a density of 2 x 10^ cells per well as previously described (8). For confocal laser scanning microscopy, cells were washed in PBS and fixed for 10 min in aceton at rt. Filters were incubated with mAb M2-III-6 (undiluted) for 60 min. Antibody binding was detected with a FITC-labeled sheep anti-mouse IgG (1:50; Boehringer Mannheim, Germany). Filters were mounted with Vectashield (Vector Laboratories,

Burlingame, CA) containing propidiumiodide (1 mg/ml) for counterstaining of nucleic acids. Cells were examined with a MRC-600 confocal microscope (Bio Rad, Hertfordshire, UK). Expression of hcMOAT protein was visualized by indirect immunolocalization using confocal laser scanning microscopy (CLSM) . Clear staining was observed in approximately 50% of the cells (Figure 8 A), whereas in MDCKII wild-type only a very weak signal was detected (data not shown). Examination of the cells at the plane perpendicular to the membrane filter revealed that almost all human cMOAT immunostaining was confined to the apical plasma membrane, although intracellular staining was observed in some cells (Figure 8 B).

To investigate whether human cMOAT expression in MDCK cells allows functional transport of GS-DNP the following experiment was performed Export of [¹⁴C]GS-DNP (GSH dmitrophenyl) from cells was determined by incubating cells with [¹⁴C]CDNB ([¹⁴C]-1-chloro-2,4-dιnιtrobenzene 10 mCi/mmol) as described (8) . The resulting hydrophilic GS-DNP only leaves the cell by active transport. Transport of GS- DNP across the apical and the basolateral membrane can be distinguished by growing cells as a monolayer on microporous membrane filters Briefly, cells were grown on polycarbonate filters (see under immunocytochemistry) for 3 4 days Two ml of medium (at room temperature) containing 2 mM [¹⁴C]CDNB was applied to both the apical and basal compartment of the monolayer and 200 ml aliquots were taken at various time points. After extraction with 200 ml of ethylacetate

radioactivity in 160 ml of the water phase was determined by liquid scintillation counting. The amount of radioactivity was corrected for the decrease in volume of culture medium. To determine intracellular radioactivity, cells were washed with cold PBS, filters were cut from the plate and counted directly in liquid scintillation fluid. The resulting pattern of GS-DNP export after exposing MDCKII, and the human cMOAT transfected clones is shown in Figure 9. MDCKII wild-type cells showed a substantial endogenous level of GS- DNP transport Comparable transport to both the apical and basal compartment was measured in parental cells.

Remarkably, apical GS-DNP export was substantially higher in both human cMOAT transfected clones, demonstrating that human cMOAT is active as a glutathione conjugate pump in these cells (figure 9). Comparing the Western blot data (Figure 7) with transport data suggests that there is a correlation between the amount of human cMOAT detectable and the level of apical GS-DNP transport. To exclude that differences in transport capacity between individual clones were due to differences in GST activity, and therefore differences in conjugation capacity, the total amount of GS- DNP retrieved in cells plus medium after 20 min was

calculated. These data revealed that all clones had

comparable levels of GS-DNP formed (data not shown).

Example 5: A mutation in the human cMOAT gene causes the Dubin-Johnson syndrome. The human Dubm-Johnson syndrome is an autosomal recessive liver disorder characterized by chronic conjugated

hyperbilirubmemia Patients have impaired hepatobiliary transport of non-bile salt organic anions a phenotype similar as has been described for the TR- rat, which has a defective cmoat. In view of the identical phenotypes of TR- rats and Dubm-Johnson patients, we have tested whether a mutation in the human cMOAT gene also underlies the

transport defect in Dubm-Johnson syndrome. In this example we demonstrate that the human homologue of rat cMOAT, human cMOAT also subject of this invention, is deficient in a patient with Dubm-Johnson syndrome. Furthermore we show that we have used the DNA sequence of the human cMOAT cDNA to develop two diagnostic assays for Dubm Johnson syndrome.

We have studied a female, Caucasian patient (age 54) who was diagnosed for Dubm-Johnson at the age of 20. She frequently complained of pains in the upper abdomen . General liver function was normal except for elevated conjugated (38 to 70 μM) and unconjugated (12 to 25 μM) serum bilirubin levels. It was not possible to visualize the gallbladder after administration of oral contrast reagent, a

characteristic feature of Dubin-Johnson. The patient

demonstrated a delayed plasma clearance of i.v. injected BSP, followed by a secondary rise in plasma BSP levels. At the age of 32 the patient underwent cholecystectomy. A characteristic black liver was observed and microscopic analysis of a liver section revealed mild fibrosis and the pigment accumulation indicative of Dubm-Johnson. Liver from this patient was obtained by a needle biopsy. Normal control liver was obtained from surgical pathology specimens .

Biopsies were fixed for histology in 4 % formaldehyde and embedded in paraffin. Skin fibroblasts from the patient and a normal control were obtained by skin biopsy and cultered in Ham F-10 (Life Technologies), supplemented with 10 % fetal bovine serum and antibiotics, at 37 ºC.

Paraffin-embedded liver sections of Dubin-Johnson and control liver were examined for the presence and

localization of the cMOAT protein, using monoclonal antibody M₂III-6. For this purpose formaldehyde-fixed paraffin- embedded liver sections were deparaffinized in xylene and rehydrated. Endogenous peroxidase activicy was blocked with 0.3 % (v/v) H2O₂ in methanol for 30 min. Before staining, the sections were pretreated with 0.01 M citric acid (pH

6.0) for 3 x 5 min at 100 C. The sections were blocked with normal rabbit serum for 10 min and incubated with monoclonal antibody M2IH-6 for 1 h. Immunoreactivity was visualized with biotinylated rabbit anti -mouse Fab2 (Dako Copenhagen, Denmark), followed by streptavidin-conjugated horseradish peroxidase (Dako) in PBS/1% BSA, and subsequent staining with 3 , 3 ' -diaminobenzidine tetrahydrochloriαe and 0.02 % (v/v) H₂O₂ in PBS. P-glycoproteins were detected with monoclonal antibody JSB-1. All sections were

(counter)stained with hematoxylin and mounted. The M2IH-6 antibody was produced against a bacterial fusion protein containing the 202-amino acid COOH-terminus of rat cMOAT; it cross-reacts with human cMOAT, but not with human MRP1. In human control liver, like in rat control liver, the antibody stained the canalicular membrane of the hepatocyte . In

Dubin-Johnson liver , as in TR- rat liver, no canalicular staining was observed, indicating that this patient lacks the cMOAT protein (figure 11). A positive canalicular staining was observed in both Dubin Johnson syndrome and control liver with JSB-1, an antibody against P- glycoprotein, used as a positive control. To investigate the nature of the genetic defect, total RNA was isolated from cultured fibroblasts obtained from a skin biopsy of both the patient and a normal control according to the acid-phenol single step method. cDNA synthesis was carried out with 6 μg of total RNA and random hexamer primers with Moloney Murine Leukaemia Virus Reverse Transcriptase (M-MLV-RT, Life

Technologies, Gaithersburg, MD), at 37°C for 1 h, followed by 10 min 65°C to inactivate the M-MLV-RT. The complete cMOAT cDNA was amplified by the "touch down" PCR protocol from both patient and control fibroblast cDNA using five sets of cMOAT-specific primers:

5'-TAGAAGAGTCTTCGTTCCAGACGCAG- 3 ' ( forwardl ) and

5'-GCAATTTCAGCAGCTGAGGACTCAC-3 ' (reversel),

5'-AAATCCTGGTTGATGAAGGCTCTG - 3 ' ( forwardlI) and

5'-TCCAGGTTCACATCTCGGACTCTGGC-3 ' (reverselI),

5 -ACATCTGCCATTCGAGATGACTGC-3 ' ( forwardlII) and

5'-CAACTCTCATGTCCCTCTGAGATGC-3 ' (reverselll),

5'-TGAAGTTCTCCATCTACCTGGAGTACC-3 ' (forwardlV) and

5'-GATGATGGTCAGCTTCTCTCGGAGG-3 ' (reverseIV), and

5'-GTCATCCCTCACAAACTGCCTCTTCAGAATCTTAG-3' ( forwardV) and 5'-CTGCTAGAATTTTGTGCTGTTCACATTC-3 ' (reverseV). PCR reactions were carried out in a Perkin Elmer GeneAmp PCR system 2400, in Ix Taq polymerase buffer (Life Technologies), 1.5 mM of MgCl2, 0.5 mM of dNTPs, 400 nM of each primer, and 0.5 units of Taq polymerase. The PCR products were obtained after application of the "touch down" PCR protocol; the reactions were denatured at 96 C for 5 min, followed by five times 2 cycles with annealing temperatures of 72, 70, 68, 66, and 65 C respectively, and subsequent 30 cycles with an

annealing temperature of 64°C. Each cycle started with 20 s at 94°C, 30s at the indicated annealing temperature, and 90s at 72°C. The PCR reaction was terminated after an extension step at 72°C for 10 min. PCR fragments obtained from

fibroblasts were excised from agarose gel, purified, ligated into the TA-cloning plasmid pCR™II (Invitrogen, Leek, The Netherlands), and transformed into INVaF' competent cells (Invitrogen). White colonies were picked, grown overnight, and plasmid DNA was isolated using the alkaline lysis method. Nucleotide sequences of 5-8 pooled clones were determined by the dideoxynucleotide chain method. Sequence analysis of multiple independent clones revealed a mutation in the patient at codon 1066 (CGA to TGA; arginine to stop- codon) (figure 12), which leads to premature termination of cMOAT protein synthesis, the normal protein being 1545 amino acids long (see also figure 10). The mutation results in the loss of a TagI restriction site, and we have confirmed the absence of this site in the patient by TagI digestion of her cMOAT cDNA (figure 13). From this observation we conclude that this patient is either homozygous for the mutation m codon 1066, or that the second allele does not give rise to a mRNA.

In conclusion, the identification of a mutation in human cMOAT in a patient with the DJS confirms the hypothesis that the TR- rat is an animal model for Dubin Johnson syndrome and provides additional evidence that the cMOAT gene encodes the major transporter for organic anions in the liver canalicular membrane . Our demonstration of a low, but

detectable expression of the cMOAT gene in fibroblasts in addition to a nucleic acid based diagnostic assay for Dubm- Johnson syndrome, allows a simple identification of this inherited disorder, without the need for liver biopsy. Example 6: Identification and isolation of other members of the family of anorganic anion

transporters.

6.1 Materials and Methods

6.1.1 Cell lines

All cell lines used in this study have been described in the literature before: the drug-sensitive and doxorubicin- selected MDR sublines of the non-small-cell lung cancer cell lines SW1573/S1 and COR-L23 (58-61); the small cell lung cancer cell line GLC4 (62); the lung adeno carcinoma cell line MOR/P (61); and the leukemia cancer cell line HL60 (63); the T24 bladder carcinoma cell line and three CDDP- resistant sublines (45); the 2008 ovarium carcinoma cell line, two CDDP-resistant sublines and a Cd²⁺-resistant subiine (40, 41); the A2780 ovarium carcinoma and the HCT8 colon carcinoma cell lines and CDDP-resistant sublines of both (39, 47); the PXN94 ovarium carcinoma and the

tetraplatin-resistant subiine PXN94tetR (42); the GCT27 testicular carcinoma cell line and the CDDP-resistant subiine GCT27cisR (44); the KB-3-1 epidermoid carcinoma cell line and a CDDP-resistant subiine KCP-4 (37, 38). All cells were grown in DMEM or RPMI medium (Gibco, BRL), supplemented with 10% fetal calf serum, 2 mM glutamine, penicillin (50 units/ml) and streptomycin (50 μg/ml). All cells were free of Mycoplasma as tested by the use of the Gene- Probe rapid Mycoplasma detection system (Gen-Probe, San Diego, USA).

6.1.2 Clonogenic survival assays

The drug sensitivity of cells was determined in clonogenic survival assays in the continuous presence of drugs. Five hundred cells per well were seeded in 24 -well plates and incubated for 24 hrs at 37ºC. Drugs, of which concentrations were varied in 2- fold steps, were added and cells were incubated for 5-6 days at 37ºC. After this the cells were stained with 0.2% crystal violet in 3.7% glutaraldehyde and colonies containing more then 50 cells were counted. The relative resistance was calculated as the ratio of IC50

(Inhibitory Concentration where 50% of the cells survives) of the resistant cell line to the IC₅₀ of the parental cell line.

6.1.3 Cloning and sequencing of MRP3 , MRP4 , and MRP 5 cDNA

For the isolation of MRP3 , MRP4 , and MRP5 cDNA, human cDNA clones were obtained from the I.M.A.G.E. consortium (64). Additional MRP3 cDNA clones were isolated by screening a human liver 5' stretch plus cDNA library, oligo(dT) and random primed (Clontech, Palo Alto, USA), using a 1 kb EcoRI-SacI fragment of a human cDNA clone (no. 84966,

Stratagene liver cDNA library #937224) as probe. Several overlapping cDNAs were isolated and sequenced. For MRP4 the insert of a human cDNA clone (no. 38089, Soares infant brain 1NIB cDNA library) was sequenced, containing the 3'-terminal end of the gene. MRP5 cDNA clones were isolated by screening a fetal brain cDNA library (Clontech, Palo Alto, USA), using the insert of human cDNA clone (no. 50857, Soares infant brain 1NIB cDNA library) as probe (J. Wijnholds, C. Mol, and P. B., unpublished results). Several overlapping cDNAs were isolated and sequenced. For sequencing the ABI 377 Automatic Sequencer was used. Sequence analysis was done using the GCG package of the Wisconsin University (20). All the sequences have been deposited with GenBank (MRP3 accession number U83659; MRP4 accession number U83660; MRP5 accession number U83661).

6.1.4 RNA

Cytoplasmic RNA from cell lines was isolated by a Nonidet P- 40 lysis procedure (24). Total cellular RNA from tissue samples obtained during surgery or at autopsy was isolated by acid guanidium isothiocyanate-phenol-chloroform

extraction (65). 6.1.5 RNase protections

By PCR amplification of human cMOAT cDNA a 241 bp fragment corresponding to nucleotides (nts) 4136-4376 (49; GenBank accession number U49248) was generated. The primers used for amplification were 5'-CTGCCTCTTCAGAATCTTAG-3' (forward primer) and 5'-CCCAAGTTGCAGGCTGGCC-3 ' (reverse primer). For MRP3, MRP4, and MRP5 RNA detection the following fragments were generated by PCR amplification: (i) for MRP3 a 262 bp fragment was generated using the primers

5'-GATACGCTCGCCACAGTCC-3 ' (forward primer) and

5'-CAGTTGGCCGTGATGTGGCTG-3 ' (reverse primer); (ii) for MRP4 a 239 bp fragment was generated using the primers 5'- CCATTGAAGATCTTCCTGG-3' (forward primer) and 5'- GGTGTTCAATCTGTGTGC-3' (reverse primer); (iii) for MRP5 a 381 bp fragment was generated using the primers

5'-GGATAACTTCTCAGTGGG-3' (forward primer) and

5'-GGAATGGCAATGCTCTAAAG-3' (reverse primer). All the

fragments were cloned into pGEM-T (Promega, Madison, USA), resulting in the plasmids hcMOAT- 241, MRP3-262, MRP4-239, and MRP5-381, and the sequences were confirmed. For RNase protections, a-³²P-labeled RNA transcripts were transcribed from Wotl-linearized DNA of hcMOAT-241 and MRP3-262, using T7 RNA polymerase, or from Ncol - linearized DNA from MRP4-239 and MRP5-381, using Sp6 RNA polymerase. For MDRl RNA

detection, a 301 bp MDR1 cDNA fragment was used (nt

positions 3500-3801 (66)), and for MRPl RNA detection a 244 bp MRPi cDNA fragment was used ( nt positions 239-483 (7)). RNase protections were carried out according to Zinn et al . (67), modified by Baas et al . (58). Protected probes were visualized by electrophoresis through a denaturing 6% acrylamide gel, followed by autoradiography. In all

experiments a probe for actin (68) was included as control for RNA input. The amount of MDRl , MRPl , cMOAT, MRP3 , MRP4 , or MRP5 RNA relative to the amount of actin was calculated using a phosphorimager (Fuji BAS 2000, TINA 2.08b).

6.1.6 Protein analysis

Total cell lysates were made by lysing harvested cells in 10 mM C1 /1.5 mM MgCl₂/10mM Tris-HCl, pH 7.4/0.5% (wt/vol) SDS supplemented with 1 mM phenylmethylsulfonyl fluoride, leupeptin (2 μg/ml ), pepstatin (1 μg/ml), and aprotinin (2 μg/ml). DNA was sheared by sonication and samples containing 40 μg of protein were fractionated by SDS/7.5% PAGE and then transferred onto a nitrocellulose filter by electroblotting. After blotting the filters were blocked for at least 2 hours in Blotto (Phosphate-buffered saline containing 1% bovine serum albumin, 1% milk powder, and 0.05% Tween-20), followed by incubation for 2 hours with the primary antibody in

Blotto. cMOAT protein was detected with mouse monoclonal antibodies M2III-5 or M2III-6, generated against a bacterial fusion protein containing the 202 amino acid COOH-terminus of rat cmoat (48). Immunoreactivity was visualized with peroxidase-conjugated rabbit anti-mouse immunoglobulins (Dako, Denmark) followed by enhanced chemiluminescence detection (Amersham, U.K.).

6.1.7 Fusion proteins of cMOAT, MRP3, and MRP5.

To test the cross-reactivity of the cmoat monoclonal antibodies with human cMOAT and other MRP homologs, fusion proteins were made of the Escherichia coli maltose-binding protein with COOH-terminal ends of human cMOAT, MRP3, and MRP5, respectively, using the plasmid vector pMal-c (69). The expression plasmids encoded, respectively, for cMOAT the 202-ammo acid COOH-terminal end, for MRP3 the 190-amιno acid COOH-terminal end, and for MRP5 the 169-ammo acid COOH-terminal end. The fusion proteins were produced in E.coli DH5a and purified by amylose resin affinity

chromatography (69). 6.1.8 Glutathione assay

Cells (1-2 x10⁶ per well) were plated in triplicate in 6 wells plates in medium with or without drugs. 48 hrs after plating the cells were washed with phosphate-buffered saline and scraped in 10% perchloric acid. Precipitated protein was removed by centrifugation and the supernatant was

neutralized by adding 0.5 M MOPS/5 M KOH. The concentration of total glutathione (GSH and glutathione disulfide (GSSG) ) was determined according to the recycling method of Tietze (70).

6.1.9 Chromosome localizations

For the chromosome localization of MRP3, MRP4 , and MRP5, radiation hybrid mapping was performed with MRP3, MRP4 , and MRP5 specific primers and two different cell panels,

Stanford G3 (StG3; 71) and Genebridge 4RH (Gb4RH; 72). The primers used for amplification were: (i) for MRP3

5'-CTCAATGTGGCAGACATCGG-3 ' and 5'-GGGAGCTCACAAACGTGTGC-3'; (ii) for MRP 4 5'-CCATTGAAGATCTTCCTGG- 3 ' and 5'- GGTGTTCAATCTGTGTGC-3'; (iii) for MRP5 55'-

CCTGTTTGGGAAGGAATATGA-3' and 5'-GGGTCGTCCAGGATGTAGAT-3'. For the PCR reactions 25 ng DNA, 2 ng/μl of each specific primer, 0.8 units Goldstar polymerase (Eurogentec, Seraing, Belgium) (MRP3 and MRP4 ) or 1.5 units Amplitaq Gold

polymerase (MRP5 ) were used in a total volume of 25 μl with 1.5 mM MgCl2 and 100 μM of each dNTP at final

concentrations. The PCR conditions were: initial

denaturation 5 min 94°C ( MRP3 and MRP4 ) or 12 min 95°C

( MRP5), followed by 42 cycles of 15 sec 94°C, 30 sec 58°C, and 45 sec 72ºC. Final extension was for 10 min at 72ºC. PCR products were resolved by 1 percent agarose gel

electrophoresis and the cell line scored positive, negative or ambiguous for presence of the gene. Datafiles were submitted to the Stanford Human Genome Center or Whitehead Institute radiation hybrid mapping databases for placing of the MRP genes in context of the respective radiation hybrid map framework markers.

6.1.10 Microsatellite repeat analysis

To confirm identity of cell lines and subclones 9 highly polymorphic microsatellite markers were used (D1S1649,

D2S434, D2S1384, D3S2427, D9S301, D9S934, D12S2070, D14S611, and D17S969). PCR conditions were as described in the Genome Database (GDB). One primer of each set was labelled with a fluorescent dye and PCR products were visualized by

electrophoresis on a ABI 377 automatic sequencer. Data were analyzed with Genetyper software version 1.1.1 (Perkin

Elmer, Foster City). Allele sizes were within expected range.

6.1.11 deposited clones

The sequences for human MRP- 3, MRP-4, MRP-5 and for human and rat cMOAT are deposited at ECACC (European Collection of Cell Cultures, Salisbury, Wiltshire SP4 OJG, UK) under provisional numbers. 96010801-hu-MRP3 #96A

96010802-hu-MRP5 #97

96010803-hu-MRP4 #38089

96010804 -hu-cMOAT #33A

96010805-hu-MRP3 #20.11

96010806-hu-MRP5 #101

96010807-hu-MRP5 #104

96010808-hu-MRP5 #105

96010809-hu-MRP3 #20.1

96010810-rat-cMOAT

96010811-hu-MRP3 #97F

6.2 Results 6.2.1 Database search for MRP homologs

We searched human EST databases (dbEST, TIGR) for MRP homologs other than MRPl and cMOAT. Alignment and comparison of EST sequences with homology specific to the 3'-terminal ends of MRPl and cMOAT, including the coding sequence for the second ATP-binding domain, revealed that there are at least 4 more MRP homologs expressed in humans. One of these homologs is the human sulfonylurea receptor ( SUR) gene (73). The other three MRP homologs had not been identified before, and were designated MRP3, MRP4, and MRP5.

6.2.2 Cloning and sequencing of MRP3 , MRP4 , and MRP 5 cDNA

Additional cDNA clones for MRP3 and MRP5 were isolated from a human liver and a fetal brain cDNA library, respectively. MRP3 and MRP5 cDNA clones were sequenced as well as the MRP4 cDNA clone obtained from the I.M.A.G.E. consortium. Both MRP3 and MRP5 encode four domain proteins, i.e. proteins with two ATP-binding domains and two domains with

transmembrane regions (M.K. and J. Wijnholds, unpublished results). More sequence data will determine whether this is also the case for MRP4 . Figure 14 shows the protein

alignment for the COOH-terminal ends of the various members of the human MRP family and human SUR. The alignment

includes the Walker A and B motifs and the signature

sequence of the second ATP-binding domain. The percentages of homology for the COOH-terminal 124 amino acids are shown in table 1. The highest homology is found between MRP1 and MRP3 (86% similarity) and the lowest between SUR and any of the MRPs (< 69% similarity).

6.2.3 Chromosome localization of MRP3, MRP4 , and MRP5 The MRPl gene has been mapped to chromosome 16 at band pl3.13-13.12 (5) and recently the cMOAT gene to chromosome 10, band q24 (52, 74). We mapped the other MRP homologs on the Gb4RH and StG3 radiation hybrid mapping panels, using MRP3 , MRP4 , or MRP5 specific primers. MRP3 , MRP4 , and MRP5 , are located on chromosomes 17, 13, and 3, respectively. The most closely linked markers were D17S797 (Gb4RH) and

D17S1989 (StG3) for MRP3, WI-9265 (Gb4RH) and D13S281 (StG3) for MRP4 , and WI-6365 (Gb4RH) and D3S4205 (StG3) for MRP5. These results are consistent between the radiation hybrid mapping panels and demonstrate that the new MRP homologs are indeed new genes, and not splice variants of MRPl or cMOAT.

6.2.4 Human tissue distribution of cMOAT, MRP 3 , MRP4 , and MRP 5 RNA

RNase protections were performed to determine the tissues that express cMOAT and MRP3 , MRP4 , and MRP5. The results are summarized in table 2. Both cMOAT and MRP3 are highly expressed in liver, and to a lower extent also in duodenum. Low expression of cMOAT was found in kidney and peripheral nerve. For MRP3, substantial expression, similar to

expression in duodenum, was also detected in colon and adrenal gland. MRP4 is expressed at a low level in only a few tissues tested. MRP5 RNA was detected in substantial amounts in every tissue tested, with relatively high

expression in skeletal muscle and brain. 6.2.5 Expression of MRP homologs in resistant cell lines

In view of their homology with MRPl , cMOAT and the three new MRP homologs are believed to encode transporter proteins involved in drug resistance. We therefore screened a large set of human cell lines derived from various tissues and their resistant sublines selected with either doxorubicin, cisplatin, tetraplatin, or CdCl₂- Only resistant lines showing decreased cellular accumulation of drugs were analyzed. All cell lines were analyzed by RNase protection for levels of MDR1, MRPl , cMOAT, MRP3 , MRP4 , MRP5, and actin RNA. The results are summarized in tables 3 and 4, and an example of each probe is shown in Fig. 15.

High MDRl overexpression was detected only in two sublines of the human non-small-cell lung cancer cell line SW1573/S1, both selected for high level doxorubicin

resistance (2R160 and 1R500). The low level of MDR1 RNA in the other cell lines is not remarkable as most of the cell lines selected for our panel were known to have a non-Pgp MDR phenotype . Low MDR1 overexpression was found in the

2R120, a subiine of the SW1573/S1, and in three cisplatin selected sublines of the bladder carcinoma cell line T24. Interestingly, a decrease rather than an increase in MDRl RNA was seen in two cisplatin selected sublines of the ovarium carcinoma cell line 2008 (table 4). This phenomenon has been reported earlier in the SW1573/S1 sublines lR50b, 2R50, and 3R80, selected for low level doxorubicin

resistance (58, 59; Table 3).

MRPl RNA is highly overexpressed in the four non-Pgp MDR cell lines GLC₄/ADR, MOR/R, COR-L23/R, and HL60/ADR, all selected for high level doxorubicin resistance (7, 75, 76). The doxorubicin selected cell lines, derived from the

SW1573/S1 cell line, showed no or only a minor increase in MRP1 RNA, as reported before (7, 77). In the cell lines, selected for cisplatin resistance, we detected no major changes in MRPl RNA. Only in two sublines of the T24 cell line, T24/DDP7 and T24/DDP10, and in HCT8/DDP, a subiine of the colon carcinoma HCT8 cell line, a slight (less than 2 fold) increase in MRPl RNA was found .

Expression of cMOAT varied greatly between the cell lines Most parental cell lines did not express cMOAT or at very low levels Only the MOR/P and the KB-3-1 parental cell lines showed substantial cMOAT RNA levels Overexpression of cMOAT was found in several doxorubicin-resistane sublines of SW1573/S1 (30 3M, lR50b, 2R120, 2R160, and 1R500), and some cisplatin selected cell lines (2008/C13*5 25, 2008/A,

A2780/DDP, and HCT8/DDP) .

Similar to cMOAT, most parental cell lines either did not express MRP3 or only at very low levels The only two parental cell lines which shov, high expression of MRP3 the MOR/P and the KB-3-1 also show high expression of cMOAT Overexpression of MRP3 in resistant lines was only found in several doxorubicin-resistant sublines of the SW1573/S1 cell line and the cisplatin resistant HCT8/DDP cell line

MRP4 is expressed only at low or very low levels in the cell lines we analyzed and no overexpression of MRP4 was detected in resistant sublines .

MRP5 is expressed in every cell line we analyzed, with the highest levels in MOR/P and 2008, but in none of the resistant sublines MRP5 is highly overexpressed Only in three cisplatin resistant cell lines, T24/DDP10, HCT8/DDP, and in the KCP-4(-), a minor increase in MRP5 RNA was detected.

6.2.6 cMOAT protein in resistant cell lines

To investigate whether the increased cMOAT RNA levels in the resistant cell lines were accompanied by increased cMOAT protein levels, total cell lysates were tested on Western blot with the monoclonal antibodies M2III-5 and M2III-6, generated against amino acids 1340 to 1541 of the rat cmoat protein (48) To test the specificity for human proteins of the Mabs generated against rat cmoat, fusion proteins containing COOH-terminal ends of human cMOAT, MRP3, and MRP5, were made Both cMOAT Mabs, M2IH-5 and M₂III-6, recognize human cMOAT. M2lII-5 also reacts with the MRP5 fusion protein, and M2lII-6 also reacts with the MRP3 fusion protein. No cross-reaction was detected for both Mabs with MRP1 (data not shown) .

Protein analysis of the cell lines with the cMOAT Mabs showed the presence of a 190-200 kDa protein in several lines (Fig. 16). Similar results were obtained witn M2III-5 and with M2lII-6 (not shown), indicating that the protein detected is cMOAT. The level of cMOAT protein in each cell line correlated very well with the level of cMOAT RNA, even for the cell lines with only a marαmal increase in cMOAT RNA, such as the 2008/C13*5.25 and the 2008/A. The only exception was the cisplatin resistant subiine of KB-3-1 KCP-4(-) . The Western blot shows that the cMOAl protein level was aoout 2-3 fold higher in the KCP-4(-) ceil line than in the KB- 3-1, whereas the RNA levels were the same in parental and resistant cells Mab M2III-5 also reacts with

MRP5 and MRP5 RNA is raised in the KCP-4(-) cells, but a similar result was obtained with Mab M2III-6 which does not cross-react with MRP5.

All cell lines with no or only very low levels of cMOAT RNA also contained no detectable cMOAT protein (Fig. 16). The small amount of cMOAT detected in the parental A2780 cell line migrated faster in the gel than the cMOAT protein present in the cisplatin resistant A2780/DDP cell line, or the protein detected in the HCT8, HCT8/DDP, KB-3-1, and KCP- 4 ( - ) cells. The varying mobility of cMOAT in the gel could be caused by different degrees of post-translational

modification of cMOAT protein in each cell line, as we have observed for MRP1 (77, M.K. unpublished results), but this needs to be verified.

6.2.7 Glutathione assays

In view of the proposed role of cMOAT as a GS-X pump, intracellular GSH levels were measured for the cell lines in table 4 GSH levels were elevated in all resistant cell lines (Table 4), and were not detectably different in cells cultured with or without drugs (data not shown).

6.2.8 Drug resistance of the cell lines analyzed To determine whether there is a correlation between the elevation of expression of transporters and resistance pattern, we have extended the existing information on these cell lines with a more complete survey of resistance against either cisplatin or doxorubicin (Table 5). Interestingly, all the doxorubicin selected SW1573 cell lines with

overexpression of cMOAT are also cross-resistant against cisplatin, and the level of cMOAT expression correlates quite well with the level of cisplatin resistance (Tables 3 and 5). Cytotoxicity analysis of the KCP-4(-) cell line showed that the IC50 for cisplatin for this cell line was much lower than reported (700 nM, RF 1.8 [table 5] instead of 25.000 nM, RF 62.5 [37]), suggesting that this cell line was a revertant or contaminated with another low-level- cisplatin resistant cell line. When these KCP-4(-) cells were cultured in the presence of 6.7 μM cisplatin, more than 99% of the cells died. The surviving population, KCP-4(+), was highly cisplatin resistant again (IC₅₀ 22.400, RF 59

[table 5]), but did not express cMOAT anymore (Fig. 16). Microsatellite repeat analysis showed that both cell lines, KCP-4(-) and KCP-4(+), were derived from the parental KB-3- 1, indicating that the KCP-4(-) is most likely a revertant.

All cell lines selected for resistance against

cisplatin, tetraplatin or CdCl₂ are not cross-resistant against doxorubicin (Table 5), with two exceptions: the KCP-4(-) cell line and the PXN94/tetR cell line. Cross- resistance did not correlate with cMOAT expression.

6.3. Discussion. 6.3.1. The MRP gene family

Our database search of expressed sequence tags has revealed that at least five homologs of MRP1 are expressed in man. cMOAT or MRP2 encodes the major organic anion transporter in the canalicular membrane of hepatocytes (48-52, 55). The product of another homolog, SUR, plays a role in the regulation of insulin secretion (73) The other three homologs, MRP3 -5, are all more related to MRP1 than SUR (Table 1). Identity is highest between MRP1 and MRP3 (75%). Since the region taken for comparison is small and one of the most conserved parts of the protein, the overall

identity between the MRP homologs will probably be lower than the percentages in Table 1.

The MRP homologs MRP3-5 are all located on other chromosomes than MRPl and cMOAT . This confirms that MRP3 , MRP4 , and MRP5 are not alternative splice products of MRP1 or cMOAT Klugbauei and Hofmann (78) recently cloned another ABC transporter ( ABC-C), located in the same chromosomal band as MRP1 , but this is not a MRP homolog, because the identity between these two proteins is only 18%. After our work was completed Allikmets et al . (79) reported the identification of 21 new ABC genes also based on a search of the human EST database and they mapped the identified partial sequences.

6.3.2 Physiological functions of the MRP family members

The physiological role of these new MRP proteins is probably a role in cellular detoxification processes by exporting GSH 5-conjugates or other organic anions GSH S-conjugate carriers have been described in many mammalian cells, including liver, heart lung, and mast cells and

erythrocytes (1B, 80) Kinetic studies indicate that both liver canaliculi and erythrocytes contain two different ATP- dependent transport activities for organic anions (81-841 cMOAT is localized in the canalicular membranes of

hepatocytes and the absence of this protein in the TR- rats as well as in a patient with the Dubin-Johnson svndrome shows a role for the cmoat/cMOAT proteins in the transport of non-bile acidic organic compounds from liver to bile (48 49, 55). The other ATP-dependent transport activity in liver canaliculi responsible for transport of bile acids from liver to bile is not attributable to cMOAT, because studies with TR- rats and Dubin-Johnson patients showed that bile acid transport was not affected (1B) .

Two other congenital liver diseases characterized by a conjugated hyperbilirubmemia, like the Dubin-Johnson syndrome, are Benign Recurrent Intrahepatic Chclestasis (BRIC) and Progressive Familial Intrahepatic Cholestasis (PFIC or Byler disease) (85, 86). The clinical and

biochemical features of BRIC and PFIC are suggestive of a defect in primary bile acid secretion (87, 88). BRIC and PFIC have both been mapped to the same region on chromosome 18, 18q21-q22 (89, 90) .

In view of the high expression of MRP3 in the liver

(Table 2), MRP3 may be the bile salt transporter. Since none of the human ABC transporter genes identified thus far maps to chromosome 18 (79; this study) it is unlikely that

BRIC/PFIC is caused by a defect in a readily recognizable ABC transporter gene.

GS-X activity has also been found in erythrocytes

Several studies have shown that human and rat erythrocytes contain a low- and a high-affinity S-(2,4-dinitrophenyl)- glutathione (DNP-SG) transporter (84, 91, 92). The high- affinity DNP-SG transporter is most likely MRP1, since the presence of this protein and its binding to LTC4 have been shown for erythrocytes (93, 94). The other transporter with low affinity for DNP-SG but high affinity for glucuronides and mercapturates (84) is not cMOAT or the bile salt transporter, because (i) no major alterations in DNP-SG transport in erythrocytes from TR- rats and Dubin-Johnson patients were detected (IB), and (ii) erythrocytes transport DNP-SG and GSSG but no bile salts (83). This second

transporter may be encoded by one of the other MRP homologs.

6.3.3 Expression of MRP homologs in resistant cell lines

We screened a large set of cell lines and their resistant sublines to see whether MRPl, cMOAT or one of the other MRP homologs is overexpressed. MRP4 was not overexpressed in any of the lines. MRP3 RNA was only found to be elevated in the cisplatin resistant HCT8/DDP cell line and several SW1573/S1 sublines selected for doxorubicin resistance. However, overexpression did not correlate with the level of

doxorubicin resistance. For MRP5 low overexpression was found in three cell lines selected for cisplatin resistance (T24/DDP10, HCT8/DDP, and KCP-4(-); Table 4), but many other cisplatin selected cell lines showed no overexpression.

Table 3 shows that the classical non-Pgp cell lines selected for high doxorubicin resistance and known to highly overexpress the MRP1 gene, do not significantly overexpress other members of the MRP family. This is compatible with the interpretation that MRP1 is the transporter responsible for MDR in these cell lines. In the non-Pgp derivatives of the SW1573/S1 cell line presented in Table 3 a more complex situation is found and the contribution of MRP1, cMOAT, MRP3, and the major vault protein, also present at increased levels in some of these cell lines (95), remains to be sorted out. 6.3.4 The involvement of organic anion transporters in cisplatin resistance

Whereas P-glycoproteins do not transport small or highly charged molecules, organic anion transporters, such as MRP1 and cMOAT have been speculatively linked to resistance to oxyanions (arsenite, antimonite) and cisplatin. These compounds can form complexes with GSH and there is now considerable evidence that these complexes are substrates for organic anion transporters. Resistance caused by increased export of these complexes is bound to be complex as pointed out by Ishikawa (80) and by us (6b, 19, 95) Increased levels of pump or GSH, increased GSH synthesis, or a combination may be required depending on the rate limiting step in drug export.

In the protozoal parasite Leishmania, resistance to arsenite and antimonite can be associated with both a 40- fold increase in the Leishmania GSH homolog trypanothione (97) and an increase in the MRP-related ABC-transporter PgpA (98). Cancer cells selected for high levels of cisplatin may sometimes also contain extremely high concentrations of GSH (99) and the GSH synthesis in these cells is upregulated (99-101). All of the cisplatin resistant cell lines studied by us have elevated GSH levels as well, albeit not as high as the cell lines isolated by Godwin et al . (99). In contrast to published data, we also find raised GSH levels in the T24 sublines, the GCT27cιsR, and the PXN94/tetR cell lines (42, 44, 45). We find no clear correlation, however, between the degree of cisplatin resistance and GSH levels, as observed by Godwin et al . (99). Moreover, all the cell lines studied by us show a decreased accumulation of cisplatin and an organic ion pump may therefore be involved in resistance.

Ishikawa et al (36) showed that MRPl is overexpressed in the cisplatin resistant human leukemia cell line HL60/R- CP. They concluded that an increased GSH synthesis in combination with raised MRP1 levels can cause cisplatin resistance. Active cisplatin efflux has been described in three of the cell lines in Table 4: KCP-4, A2780/DDP, and HCT8/DDP (37-39, 102). The ATP-dependent efflux was

inhibited by DNP-SG, indicating that it was catalyzed by a GS-X pump. In addition, the membrane vesicles of the KCP-4 cell line were shown to catalyze an increased uptake of LTC₄

(37, 38), known to be the substrate with the highest affinity for MRP1 However, data from these papers and our study show that MRPl is not overexpressed in these cisplatin resistant cell lines, suggesting that MRP1 is not the major pump responsible for cisplatin resistance This is supported by transfection studies with MRPl , which showed no cisplatir resistance of the transfected cells (28, 17). Nevertheless it remains possible that transport of cisplatin conjugates by MRP1 is efficient and that the lovv levels of MRP1 present in parental cells suffice for resistance, if formation of cisplatin conjugates in resistant cells is increased, e g. by an increase in GSH synthesis.

An organic anion pump that is important in cisplatin resistance is cMOAT Especially striking is the correlation between cisplatin resistance and cMOAT expression in the non-Pgp MDR cell lines derived from the SV/1573/S1 cell line (Table 5). These lines were selected for doxorubicin

resistance and it is therefore unlikely that other

mechanisms of cisplatin resistance are activated in these lines. It should be noted, that these non-Pgp MDR lines, selected for low level doxorubicin resistance, contain multiple alterations in the expression of ABC-transporters Besides upregulation of MRPl , cMOAT, and MRP 3 (Table 3), down-regulation of MDR1 has occurred in these lines (59;

this study, Table 3).

Some other cisplatin-resistant lines contain increased levels of cMOAT as well, notably 2008/C13*5.25, 2008/A, A2780/DDP, and HCT8/DDP (Fig. 3, Table 4 and 5).

The combination of cisplatin with doxorubicin resistance in resistant cell lines has been reported before (100, 103) and is also present in twj other platin selected lines, studied here, PXN94/tetR and KCP-4(-) (Table 5). All other platin-selected lines in Table 5 are doxorubicin sensitive, however. The substrate specificity of the organic anion pumps in the liver canalicular membrane (cMOAT) and in erythrocytes (presumably mainly if not exclusively MRP1) is very similar (91) We therefore expect both pumps to confer similar resistance spectra. We have recently succeeded in obtaining stably transfected kidney cells in which cMOAT is properly routed to the plasma membrane (R Evers, M.K., and P B., unpublished). These cells should allow a direct test of the druq resistance spectrum that can be associated with cMOAT overexpression.

Overexpression of cMOAT in cisplatin resistant cell lines was recently also reported by Taniguchi et al. (52). However, in contrast to our results (Table 4) they detected raised cMOAT RNA levels in the KCP-4 and T24/DDP10 cell lines. We do not find this. The level of cMOAT RNA was even decreased in the highly cisplatin resistant KCP-4(+) cells, and in the T24/DDP10 cell line cMOAT RNA is hardly

detectable by RNase protection. We also detect no cMOAT protein in these cell lines (Fig. 16). Cross-hybridization of the cMOAT probe used by Taniguchi et al. (52), which contains the coding sequence of the first ATP-bmding domain, with RNA transcribed from the other MRP homologs might explain the discrepancy. This underlines the

importance of the use of gene specific probes to determine expression of MRP homologs.

In conclusion, our data and those recently published by Ishikawa et al . (35, 36), Fujii et al. (37, 38), Goto et al. (39), Chuman et al. (102), and Taniguchi et al. (52) provide evidence that an organic anion pump, notably cMOAT,

contributes to cisplatin resistance by exporting the

cisplatin-GSH complex. Elevated GSH levels and synthesis may be required to drive formation of the complex if contact with cisplatm is extended, as is usually the case for cell lines selected for resistance in vitro. LEGENDS TO THE FIGURES.

Figure 1a. cDNA sequence of human cMOAT

Figure 1b. cDNA sequence of rat cMOAT

Figure 2. (A) Northern blot analysis of 2 μq poly(A)⁺ RNA from Wistar rat tissues hybridized to a 1-kb Hmdl ll/Avall cDNA fragment of cmoa t. RNA was analyzed as described in (the experimental part. Prolonged exposure of the film revealed no detectable expression in other tissues then kidney, duodenum, and ileum. (B) Northern blot analysis of 2 μg of poly(A)⁺ RNA from Wistar and TR- rat liver and hepatocytes hybridized with the same probe as described in (A) The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) signal is shown at the bottom Molecular size standards are indicated at the right.

Figure 3. Immunoblot analysis of cmoat and P-glycoprotem in canalicular and basolateral membrane fractions of Wistar and TR- rat hepatocytes Lane 1, Wistar basolateral

membranes, Lane 2, Wistar canalicular membranes, Lane 3 TR- basolateral membranes, Lane 4, TR- canalicular membranes Upper panel the blot was incubated with the monoclonal antibody M2 III -5 directed to cmoat (27). This antibody did not crossreact with the hMRP1 protein as tested in total lysates from the MRP-overexpressmg cell line GLC4/ADR (20). Lower panel: Immunodetection of P-glycoprotems with the Mab C219 in the same membrane preparations. The 150-kD P- glycoproteins are exclusively expressed in canalicular membranes (22) Differential staining of the two fractions demonstrates the separation of the two membrane domains with slight contamination of the basolateral fraction by

canalicular membranes. Molecular weight markers are

indicated.

Figure 4 Deduced amino acid sequence of the rat cMOAT and alignment with the deduced 70 ammo acid sequence of the translated 213 -bp putative rat mrpl cDNA.

Figure 5. Transport of GS-DNP in COS-7 cells transiently transfected with rat cMOAT expression constructs. Closed circles represent cells transfected with pSVK3-rat cMOAT. Open circles represent cells that have been transfected with a pSVK3 construct with the rat cMOAT cDNA in the reverse orientation and serves as a negative control. The results depicted are the mean of three measurements.

Figure 6 Transport of GS-DNP in membrane vesicles prepared from COS -7 cells transiently transfected with rat cMOAT expression constructs in the presence or absence of an ATP regenerating system Closed squares represent cells

transfected with pSVK3-rat cMOAT Open squares represent cells that have been transfected with a pSVK3 construct with the rat cMOAT cDNA in the reverse orientation and serves as a negative control. The results depicted are the mean of three measurements.

Figure 7. Human cMOAT expression in crude lysates from MDCKII derived transfectants. 2 or 20 mg of total protein was size fractionated in a 7.5% polyacrylamide gel

containing 0.1% (wt/vol) SDS. After electroblotting, human cMOAT protein was visualized by staining with mAB M₂-III-6. Protein antibody interaction was detected using the Amersham enhanced chemilummescence kit (ECL). Lane 1,2 MDCKII cells, lane 3,4 human cMOAT expressing clone MDCKII-216, lane 5,6 human cMOAT expressing clone MDCKII -217. In lanes 1,3 and 5 two micrograms of total protein were loaded and in lanes 2,4 and 6 20 micrograms were loaded.

Figure 8. Detection of human cMOAT in MDCKII monolayers by confocal laser-scanning microscopy. A. Indirect

immunofluorescence (FITC) picture with mAb M2-III-6 on

MDCKII -217 cells Nucleic acids were detected using

propidiumiodide (red signal). Top view of the cell layer is shown. B. Optical section perpendicular to the plane of the cell layer.

Figure 9. Export of GS-DNP from MDCK-II, MDCKII-216, and

MDCKII-217 cells. Cells were incubated with [¹⁴C]CDNB (2 mM) in both the apical and basal compartments. Samples were taken at t = 1, 3, 6, 12, and 20 min from both compartments and extracted with ethylacetate. The amount of [¹⁴C]DNP-GS excreted (in pmol per 2 ml) was measured and plotted. All experiments were done in duplicate and repeated at least twice. Variation between measurements was below 10%. Dotted line: transport to the basal compartment. Continuous line: transport to the apical compartment.

Figure 10. Deduced amino acid sequence of human cMOAT .

Predicted transmembrane regions are underlined. Walker A, B, and signature sequence are doubly underlined. Predicted N- glycosylation sites conserved in other cMOAT proteins (rat, rabbit) and MRP1 proteins (human, mouse) are indicated with triple asterisks. The triangle indicates the location (amino acid 1066) at which a stop codon is introduced by a C to T transition in DJS cMOAT.

Figure 11. Immunohistochemical detection of the cMOAT protein in human and rat liver using monoclonal antibody

M2III-6. Sections of a normal human liver (A) and normal rat liver (B), which demonstrate the exclusive canalicular localization of the protein. In liver sections of the DJS patient (C) and the TR- rat (D), no canalicular staining is observed. Magnifications are 20 x (A, C), and 100 x (B, D). Figure 12. Part of the cMOAT cDNA sequence encompassing the mutation which results in the absence of the functional protein in the patient. The normal sequence is depicted on the right. The arrow indicates the site of the mutation at codon 1066. This codon normally encodes an arginine residue (CGA), but is changed into a stop-codon (TGA) in the

patient. The mutation of C to T eliminates the recognition site for the restriction enzyme TagI ( 5 ' -TC.GA-3 ' ) .

Figure 13. Taql digest of a part of the cMOAT cDNA that was obtained with primer combination forwardlV/reverselV. Lane 1 represents healthy control-, lane 2 the patient cDNA digest. Molecular size markers are indicated on the left in kilo base pairs.

Figure 14

Protein alignment of COOH-terminal ends of the five human

MRP homologs and human SUR. The alignment was performed with the PILEUP program of GCG (48). The GenBank accession numbers for the proteins used in this comparison are the following MRP1 - L05628, CM0AT/MRP2 - U49248, MRP3 - U83659, MRP4 -U83660, MRP5 - U83661, SUR - L78207 The nucleotide binding domain specific signature sequence and the Walker A and B motifs are shown in bold. Asterisks above the alignment indicate identical amino acids in at least four of the five MRP proteins.

Figure 15

RNase protection assays of RNA transcript levels of MDR1 MRPl cMOAT (MRP2 ) MRP3, MRP4, and MRP5 m the human non- small-cell lung cancer cell line SW1573/S1 and its

doxorubicin selected subiine 30 3M 10 μg total cytoplasmic RNA from each cell line was used per probe. The positions of the protected fragments of MDR1 MRPl - 5, and t-actm are indicated.

Figure 16

Immunoblot detection of cMOAT protein in the cell lines analyzed in this paper. Total cell lysates were size

fractionated (40 μg per lane) in a 7 5% polyacrylamide gel containing 0 5% SDS The fractionated proteins were

transferred to a nitrocellulose membrane, and cMOAT protein was detected by incubation with monoclonal antibody M₂III-5.

The size (kDa) and position of molecular weight markers are indicated.

Figure 17 A/B.

Nucleotide sequence and ammo acid sequence of MRP3

Figure 18.

Partial MRP4 sequence

Figure 19 A/B

MRP 5 sequences.

Table 1

Homology between the COOH-terminal 124 ammo acids of the five human MRP homologs and human SUR Percentages of identity and similarity were determined using the BESTFIT program of GCG (48) . Table 2

Levels of RNA transcripts of MRP1 , cMOAT (MRP2) , MRP3 , MRP4 , and MRP5 in human tissues. RNA expression levels were determined by RNase protection assays with 10 μg total RNA from various human tissues per probe Expression of t-actin was taken as control for total RNA input Data for MRPl RNA levels are from Zaman et al . (52). The relative expression level is indicated by filled circles, very low or

undetectable RNA levels by open circles nd = not

determined.

Table 3

Characteristics of the doxorubicin-selected cell lines analyzed in this paper. Resistant cell lines were selected by chronically exposing them to the concentrations of doxorubicin as shown. RNA levels were determined as m

Figure 2. The relative expression level is indicated by filled circles, very low expression by _, and undetectable RNA levels by open circles.

Table 4

Characteristics of cell lines selected for resistance to cisplatin, tetraplatin or CdCl₂. Resistant cell lines were selected by chronically exposing them to the concentrations of drugs as shown Only A2780/DDP and HCT8/DDP were selected by challenging them 1 h weekly with 50 μM cisplatm RNA levels were determined as in Figure 2. The relative

expression level is indicated by filled circles, very low expression by _, and undetectable RNA levels by open

circles. Data for total intracellular glutathione

concentrations were obtained from three independently isolated cell extracts assayed in three independent

experiments using the recycling method of Tietze (56) and presented as the mean GSH ± SD.

Table 5

IC₅₀ values and relative resistance factors (RF) of the cell lines analyzed for cisplatin and doxorubicin. IC₅₀ data were obtained from clonogenic survival assays with continuous exposure to drugs. The relative resistance factor was determined by dividing the IC₅₀ of each resistant cell line by the IC₅₀ of the corresponding parental cell line. Also shown are the levels of RNA transcripts of MRPl and cMOAT , taken from Table 3 and 4.

REFERENCES AND NOTES

1a. R.P.J. Oude Elferink et al., Am. J. Physiol. 258, G699 (1990); T. Kitamura et al., Proc. Natl. Acad. Sci. U.S.A. 87, 3557 (1990);

1B R.P.J. Oude Elferink et al. , Biochim. Biophys. Acta.

1241, 215 (1995).

2. P.L.M. Jansen et al., Heriditary chronic conjugated hyperbilirubmemia in mutant rats caused by defective hepatic anion transport. Hepatology 5: 573-579, 1985 3a. P.L.M. Jansen, W.H.M. Peters, D.K.F. Meijer,

Gastroenterol. 93, 1094 (1987);

3b 1. Nishida et al. , Clm Invest 90, 2130 (1992)

4a M. Huber, et al , Hepatology 1, 224 (1987);

4b T. Ishikawa et al., J. Biol. Chem. 265, 19279 (1990), 4c. R.P.J. Oude Elferink et al Hepatology 7, 1109 (1987), 4d. F. Kuipers et al., J. Lipid Res. 30, 1835 (1989).

5. S.P.C. Cole et al., Science 258, 1650 (1992).

6a. G. Jedlitschky et al, ATP-dependent transport of

glutathione S-conjugates by the multidrug resistance- associated protein. Cancer Res. 54: 4833-4846, 1994 6b. M. Miiller et al., Overexpression of the gene encoding the multidrug resistance-associated protein results in increased ATP-dependent glutathione S-conjugate transport. Proc. Natl Acad. Sci., USA 91: 13033-13037, 1994

7. G.J.R. Zaman et al., Analysis of the expression of MRP, the gene for a new putative transmembrane drug

transporter, in human multidrug resistant lung cancer cell lines. Cancer Res. 53: 1747-1750, 1993

8. R. Evers et al., Basolateral localization and export activity of the human multidrug resistance-associated protein in polarized pig kidney cells. J. Clin. Invest. 97: 1211-1218, 1996

9. M.H. de Vries et al., Naunyn-Schmiedeberg's Arch.

Pharmacol 340, 588 (1989). 10. W.R. Pearson and D.J. Lipman, Proc. Natl. Acad. Sci. U.S.A. 85, 2444 (1988).

11. C.F. Higgins, ABC-transporters : from micro organisms to man. Ann. Rev. Cell Biol. 8: 67-113, 1992

12. M.S. Szczypka et al., J. Biol. Chem. 269, 22853 (1994).

13. J.R. Riordan et al., Science 245, 1066 (1989).

14. R. Mayer, et al., J. Cell Biol. 131, 137 (1995).

15. J.M. Crawford and J.L. Gollan, in Disease of the liver, L. Schiff and E.R. Schiff, Eds. (Philadelphia, 1993), vol .1, pp. 42-84 [seventh edition].

16. G.J.R. Zaman et al., The human multidrug resistance- associated protein MRP is a plasma membrane drug-efflux pump. Proc. Natl. Acad. Sci., USA 91: 8882-8826, 1994.

17. S.P.C. Cole et al., Cancer Res. 54, 5902 (1994).

18. G.J.R. Zaman et al., Role of glutathione in the dxport of compounds from cells by the multidrug-resistance- associated protein. Proc. Natl Acad. Sci., USA 92:

7690-7694, 1995.

19. Abbrevations for the amino acid residues are as

follows: A, Ala; C, Cys ; D, Asp; E, Glu; F, Phe; G,

Gly; H, His; I, He; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gin; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr.

20. J. Devereux et al., A comprehensive set of sequence

analysis programs for the VAX. Nucl. Acids Res. 12:

387-395, 1984

21. P. Klein, M. Kanehisa, C. Delisi, Biochim. Biophys.

Acta 815, 468, 1985.

22. J.J.M. Smit et al., Lab. Inv. 71, 638, 1994.

23. M. Otter, J. Kuiper, D. Rijken, A.J. van Zonneveld,

Biochem. Mol. Biol. Int. 37, 563 (1995).

24. J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular

Cloning; A Laboratory Manual. (Cold Spring Harbour

Laboratory, Cold Spring Harbour, New York, 1989).

25. M.M. Gottesman et al., Biochemistry of multidrug

resistance mediated by the multidrug transporter. Ann.

Rev. Biochem. 62: 385-427, 1993. 26 U.A. Germann, P-glycoprotein - A mediator of multidrug resistance in tumour cells. Eur. J. Cancer 32A: 927- 944, 1996.

27. D.W. Loe et al., Biology of the multidrug resistance- associated protein, MRP. Eur. J. Cancer 32A: 945-957,

1996.

28. G.J R. Zaman et al., MRP, mode of action and role m MDR. In: "Multidrug Resistance in Cancer Cells, edited by S. Gupta and T Tsuruo John Wiley & Sons, England, 1996.

29 I. Leier et al., The MRP gene encodes an ATP-dependent export pump for leukotriene C4 and structurally related conjugates. J. Biol Chem. 45- 27807-27810, 1994.

30 C H M Versantvoort et al. , Regulation by glutathione of drug transport in multidrug-resistant human lung tumour cell lines overexpressing multidrug resistance- associated protein Br . J. Cancer 72: 82-89, 1995.

31. R. Tommasini et al., The human multidrug resistance- associated protein functionally complements tne yeast cadmium resistance factor 1. Proc. Natl Acad. Sci., USA 93: 6743-6748, 1996.

32 M.S. Szczypca et al., A yeast metal resistance protein similar to human cystic fibrosis transmembrane

conductance regulator (CFTR) and multidrug resistance- associated protein. J. Biol. Chem. 269: 22853-22857, 1994.

33. Z. Li et al., The yeast cadmium factor protein (YCF1) is a vacuolar glutathione 5-conjugate pump. J. Biol. Chem. 271: 6509-6517, 1996.

34. T. Ishikawa and F. Ali-Osman, Glutathione-associated cis-diammιnedichloroplatinum(II) metabolism and ATP- dependent efflux from leukemia cells. J. Biol. Chem. 268: 20116-20125, 1993.

35. T. Ishikawa et al., GS-X pump is functionally

overexpressed m cis-diammιnedichloroplatmum(II)- resistant human leukemia HL60 cells and down-regulated by cell differentiation. J. Biol. Chem. 269: 29085- 29093, 1994.

36. T. Ishikawa et al., Coordinated induction of MRP/GS-X pump and gamma-glutamylcysteine synthetase by heavy metals in human leukemia cells. J. Biol. Chem. 271: 14981-14988, 1996.

37. R. Fujii et al., Active efflux sytem for cisplatin in cisplatin-resistant human KB cells. Jap. J. Cancer Res. 85, 426-433, 1994a.

38. R. Fujii et al., Adenosine triphosphate -dependent

transport of leukotriene C₄ by membrane vesicles prepared from cisplatin-resistant human epidermoid carcinoma tumor cells. J. Natl Cancer Inst. 86: 1781- 1784, 1994b.

39. S. Goto et al., Augmentation of transport for

cisplatin-glutathione adduct in cisplatin-resistant cancer cells. Cancer Res. 55: 4297-4301, 1995.

40. P. Naredi et al., Cross-resistance between cisplatin and antimony in a human ovarian carcinoma cell line. Cancer Res. 54: 6464-6468, 1994.

41. P. Naredi et al., Cross-resistance between cisplatin, antimony potassium tartrate, and arsenite in human tumor cells. J. Clin. Invest. 95: 1193-1198, 1995.

42. K.J. Mellish et al., Circumvention of acquired

tetraplatin resistance in a human ovarian carcinoma cell line by a novel trans platinum complex, JM335

[trans ammine (cyclohexylamine) dichloro dihydroxo platinum (IV)]. Int. J. Cancer 59: 65-70, 1994.

43. C.W. Taylor et al., Different mechanisms of decreased drug accumulation in doxorubicin and mitoxantrone resitant variants of the MCF7 human breast cancer cell line. Br. J. Cancer 63: 923-929, 1991.

44. L.R. Kelland et al., Establishment and characterization of an in vitro model of acquired resistance to

cisplatin in a human testicular nonseminomatous germ cell line. Cancer Res. 52: 1710-1716, 1992. 45. S. Kotoh et al., Increased expression of DNA

topoisomerase I gene and collateral sensitivity to camptothecin in human cisplatin-resistant bladder cancer cells. Cancer Res. 54: 3248-3252, 1994.

46. T. Misawa et al., Establishment and characterization of acquired resistance to platinum anticancer drugs in human ovarian carcinoma cells. Jap. J. Cancer Res. 86: 88-94, 1995.

47. W. Schmidt and S.G. Chaney, Role of carrier ligand in platinum resistance of human carcinoma cell lines.

Cancer Res. 53: 799-805, 1993.

48. C.C. Paulusma et al., Congenital j?ιndιce in rats with a mutation in a multidrug resistance-associated protein gene. Science 271: 1126-1128 1996.

49. C.C. Paulusma et al., A mutation in the human cMOAT

gene causes the Dubin-Johnson syndrome. Submitted.

50. M. Buchler et al., cDNA cloning of the hepatocyte

canalicular isoform of the multidrug resistance

protein, cMrp, reveals a novel conjugate export pump deficient in hyperbilirubinemic rats. J. Biol. Chem. 271: 15091-15098, 1996.

51. K. Ito et al., Expression of the putative ATP-binding cassette region, homologous to that in multidrug resistance associated protein (MRP), is hereditarily defective in Eisai hyperbilirubinemic rats (EHBR). Int. Hepatol. Comm. 292: 292-299, 1996.

52. K. Taniguchi et al., A human canalicular multispecific organic anion transporter (cMOAT) gene is overexpressed in cisplatin-resistant human cancer cell lines with decreased drug accumulation. Cancer Res. 56: 4124-4129, 1996.

53. H. Takikawa et al., Biliary excretion of bile acid

conjugate in a hyperbilirubinemic mutant Sprague-Dawley rat. Hepatology 14: 352-360, 1991.

54. P. Zimniak, Dubin-Johnson and Rotor syndromes:

Molecular basis and pathogenesis. Sem. Liver Dis . 13: 248-260, 1993. 55 J. Kartenbeck et al., Absence of the canalicular isoform of the MRP gene-encoded conjugate export pump from the hepatocytes in Dubm-Johnson syndrome

Hepatology 23: 1061-1066, 1996.

56 A. Broeks et al. , Homologues of the human multidrug resistance genes MRP and MDR contribute to heavy metal resistance in the soil nematode Caenorhabditis elegans EMBO J. 15: 6132-6143 1996.

57 M.S. Boguski et al., dbEST-database for 'expressed

sequence tags' Nature Genet 4 332-333 1993

58 F. Baas et al ., Non- P-glycoprotein mediated mechanism for multidrug resistance precedes P-glycoprotein expression during in vitro selectio I for doxorubicin resistance in a human lung cancel cell line Cancel Res 50 : 5392-5398, 1990.

59 E. W . H. M. Eijdems et al. , Reduced topoisomerase II

activity in multidrug-resistant human non-small cell lung cancer cell lines British J. Cancer 71: 40-47, 1995a.

60 C. M. Kuiper et al., Drug transport variants without P- glycoprotem overexpression from a human squamous lung cancer cell line after selection with doxorubicin. J. Cell Pharmacol. 1: 35-41, 1990

61 J. G. Reeve et al., Non-P-glycoprotem-mediated

multidrug resistance with reduced EGF receptor

expression in a human large cell lung cancer cell line Br. J. Cancer 61: 851-855, 1990.

62 J .G. Zijlstra et al., Multifactorial drug resistance in an adriamycin-resistant human small cell lung carcinoma cell line Cancer Res. 47: 1780-1784, 1987

63 W. Marsh and M.S Center Adramycin resistance in HL60 cells and accompanying modification of a surface membrane protein contained in drug-sensitive cells Cancer Res. 47: 5080-5087, 1987.

64 G. G. Lennon et al., The I .M.A.G.E. consortium: An

integrated molecular analysis of genomes and their expression Genomics 33 151-152, 1996. 65. P. Chomczynski and N. Sacchi Single step method of RNA isolation by acid guadinium thiocyanate-phenol- chloroform extraction. Anal. Biochem. 162 : 156-159, 1987.

66. C. Chen et al . Internal duplication and homology with bacterial transport proteins in the mdrl (P- glycoprotein) gene from multidrug-resistant human cells. Cell 4 7 : 381-389, 1986.

67. K. Zinn et al ., Identification of two distinct

regulatory regions adjacent to the human β-interferon gene. Cell 34 : 865-879, 1983.

68. T. Enoch et al . , Activation of the human β-interferon gene requires an interferon inducible factor. Mol .

Cell. Biol. 6 : 801-810, 1986.

69. C.V. Maina et al . , An Escherichia col i vector to

express and purify foreign proteins by fusion to and separation from maltose-binding protein. Gene 74 : 365- 373, 1988.

70. F. Tietze, Enzymic method for quantitative

determination of nanogram amounts of total and oxidized glutathione. Anal. Biochem. 27 : 502-522, 1969.

71. D.R. Cox et al., Radiation hybrid mapping: a somatic cell genetic method for constructing high-resolution maps of mammalian chromosomes. Science 250 : 245-250, 1990.

72. M.A. Walter et al . , A method for constructing radiation hybrid maps of whole genomes. Nature Genet. 7: 22-28, 1994.

73. L. Aguilar-Bryan et al ., Cloning of the β-cell high

affinity sulfonyl-urea receptor: a regulator of insulin secretion. Science 268 : 423-426, 1995.

74. M.A. van Kuijck et al ., Assignment of the canalicular multiorganic anion transporter ( cMOAT) gene to human chromosome 10q24 and mouse chromosome 19D2 by in situ hybridization. Submitted.

75. N. Krishnamachary and M.S. Center The MRP gene

associated with a non-P-Glycoprotein multidrug resistance encodes a 190-kDa membrane bound

glycoprotein. Cancer Res. 53: 3658-3661, 1993.

76. M.A. Barrand et al., A 190-kilodalton protein

overexpressed in non-P-Glycoprotein-containing

multidrug-resistant cells and its relationship to the MRP gene. J. Natl Cancer Inst. 86: 110-117, 1994.

77. E W.H.M. Eijdems et al., Altered MRP is associated with multidrug resistance and reduced drug accumulation m human SW-1573 cells. British J. Cancer 72: 298-306, 1995b.

78. N. Klugbauer and F. Hofmann Primary structure of a

novel ABC transporter with a chromosomal localization on the band encoding the multidrug resistance- associated protein FEBS Lett. 391 61-65, 1996.

79 R. Allikmets et al., Characterization of the human ABC superfamily: isolation and mapping of 21 new genes using the Expressed Sequence Tags database. Human Mol Genet. 5: 1649-1655, 1996.

80. T. Ishikawa The ATP-dependent glutathione S-conjugate export pump. Trends Biochem. Sci. 17: 463-468, 1992.

81. T. Kondo et al., Studies on glutathione transport

utilizing inside-out vesicles prepared from human erythrocytes. Biochim. Biophys. Acta 645: 132-136, 1981.

82. P. Zimniak and Y.C. Awashti ATP-dependent transport

systems for organic anions. Hepatology 17: 330-339, 1993.

83. M. Heijn Characterization of ATP-dependent organic

anion transport. PhD thesis, Amsterdam, the

Netherlands, 1994.

84. M. Saxena and G.B. Henderson MOAT4 , a novel

multispecific organic-anion transporter for

glucuronides and mercapturates in mouse L1210 cells and human erythrocytes. Biochem. J. 320: 273-281, 1996. 85. W.H.J. Summerskill and J.M. Walshe Benign recurrent

intrahepatic "obstructive" jaundice. Lancet 2: 686-690, 1959. 86 R J. Clayton et al., Byler disease: fatal familial intrahepatic cholestasis in an Amish kindred. Am. J. Dis Child 117: 112, 1969.

87 E. Jaquemm et al., Evidence for defective primary bile acid secretion in children with progressive familial intrahepatic cholestasis. Eur. J. Pediatrics 153: 424- 428, 1994.

88 G.P Van Berge Henegouwen et al., Is an acute

disturbance m hepatic transport of bile acids the primary cause of cholestasis in benign recurrent intrahepatic cholestasis? Lancet J 1249-1251, 1974

89 V.E H Carlton et al., Mapping of a locus for

progressive familial intrahepatic cholestasis (Byler disease) to 18q21-q2? the benign recurrent

intrahepatic cholestasis region Human Mol Genet 4 1049-1053, 1995.

90 R.H Houwen et al., Genome screening by searching for shared segments: mapping a gene for benign recurrent intrahepatic cholestasis. Nature Genet. 8: 380-386, 1994.

91. M. Heijn et al. ATP-dependent multispecific anion

transport system in rat erythrocyte n embrane vesicles Am J. Physiol 262- C104-C110, 1992.

92. G. Bartosz et al., Organic anions exhibit distinct

inhibition patterns on the low-K_m and hιgh-K_m transport of S-(2,4-dinitrophenyl)glutathιone through the human erythrocyte membrane. Biochem. J. 292: 171-174, 1993.

93. M.J. Flens et al., Tissue distribution of the multidrug resistance protein. Am. J. Pathol. 148 1237-1247, 1996.

94. L. Pulaski et al., Identification of the multidrug- resistance protein (MRP) as the glutathione-S-conjugate export pump of erythrocytes. Eur. J. Biochem 241: 644-

648.

95 R.J. Scheper et al., Overexpression of a M_r 110,000

vesicular protein in non-P-glycoprotem-mediated multidrug-resistance Cancer Res. 53: 1475-1479, 1993. 96. P. Borst and M. Ouelette New mechanisms of drug resistance in parasite protozoa Ann Rev. Microbiol 49: 427-460, 1995.

97. R. Mukhopadhyay et al., Trypanothione overproduction and resistance to antomonials and arsenicals in

Leishmania. Proc. Natl Acad. Sci., USA 93: 10383-10387, 1996.

98. B. Papadopoulou Gene disruption of the P-glycoprotem related gene pgpa of Leishmania tarentolae. Bichem. Bioph ys. Res. Comm. 224: 112-11%, 1996

99. A. K. Godwin et al ., High resistance to cisplatm in

human ovarian cancer cell lines is associated with marked increase of glutathione synthesis. Proc. Natl Acad. Sci. 89: 3070-3074, 1992.

100. X. Y . Hao et al., Acquired resistance to cisplatin and doxorubicin in a small cell lung cancer cell line is correlated to elevated expression of glutathione- linked detoxification enzymes Carcinogenesis 15: 1167-1173, 1994.

101. K. S. Yao et al., Evidence for altered regulation of t- glutamylcysteme synthetase gene expression among cisplatm-sensitive and cisplatin-resistant human ovarian cancer cell lines Cancer Res. 55: 4367-4374, 1995.

102. Y. Chuman et al., Characterization of the ATP-dependent LTC4 transporter in cisplatm resistant human KB cells

Biochem. Biophys Res. Comm. 226 158-165, 1996.

103. K. Hamaguchi et al., Cross-resistance to diverse drugs is associated with primary cisplatm resistance in ovarian cancer cell lines. Cancer Res. 53: 5225-5232, 1993.

Claims

1 A nucleic acid comprising a sequence encoding at least a part of a member of a family of organic anion

transporters, with the exclusion of mammalian Multidrug Resistance Associated Protein, said nucleic acid comprising at least a gene family specific fragment of one of the sequences of fig 1a or fig 1b or the complement thereof or a sequence having at least 55%, pieferably 70%, in

particular 90% homology therewith.

2 A nucleic acid according to claim 1 encoding at least a part of a mammalian member of said family.

3 A nucleic acid according to claim 2 encoding at least a part of a human member of said family.

4 A nucleic acid and/or its complement naving at least part of the sequence of fig.1A and encoding a protein having human Canalicular Multispecific Organic Anion Transport protein or similar activity or antigenicity.

5 A nucleic acid and/or its complement having at least part of the sequence of fig.1b and encoding a protein having rat Canalicular Multispecific Organic Anion Transport protein activity or antigenicity.

6 A nucleic acid encoding rat Canalicular Multispecific Organic Anion Transport protein or human Canalicular

Multispecific Organic Anion Transport protein.

7 A vector comprising a nucleic acid according to anyone of the foregoing claims and suitable means for replication, transduction and/or expression of said nucleic acid.

8 A vector according to claim 7 further comprising a gene encoding a therapeutically beneficial protein.

9 A vector according to claim 7 or 8 further comprising a gene encoding glutathion elevating activity.

10 A vector according to claim 9 wherein the gene encodes at least a functional part of a gamma glutamyl cysteme synthetase or a UDP-glucuronosyltransferase.

11. A vector according to claim 7 wherein the

therapeutically beneficial protein is another

multidrugresistance related protein such as MDR1.

12. A cell comprising a nucleic acid or a vector according to anyone of the foregoing claims.

13. A cell according to claim 12 further comprising a vector encoding glutathion elevating activity.

14. A cell according to claim 12 or 13 provided with other resistance related proteins such as MDR1.

15 A method for providing ceils with Canalicular

Multispecific Organic Anion Transport protein activity, comprising transducing said cell with a nucleic acid or a vector according to anyone of claims 1-11.

16. A method for enhancing Canalicular Multispecific

Organic Anion Transport protein activity of cells according to claim 15, comprising increasing the intracellular level of glutathion, glucuronide and/or sulphate.

17. A method for enhancing Canalicular Multispecific Organic Anion Transport protein activity of cells according to claim 15, comprising enhancing the conjugating capacity and/or the compelxing activity of said cell for sulphate, glutathion, glucuronide and the like.

18. A method for reducing Canalicular Multispecific Organic Anion Transport protein activity and/or the multidrug resistance of a cell comprising providing said cell with an antisense construct of a nucleic acid or a vector according to anyone of claims 1-11.

19. A method according to claim 18, further comprising providing the cell with an antisense construct derived from another multidrug resistance related protein such as MDRl.

20. A protein encoded by a nucleic acid according to anyone of claims 1-6 or obtainable by expression of a vector according to anyone of claims 7-11.

21. A protein having Canalicular Multispecific Organic Anion Transport protein activity or Canalicular

Multispecific Organic Anion Transport protein specific antigenicity comprising at least part of the sequence of fig.4 or another mammalian equivalent thereof.

22. Use of a nucleic acid according to anyone of claims 1-6 or a protein according to claim 20 or 21 in the diagnosis of Dubin-Johnson disease, Rotor disease or another disease involving Canalicular Multispecific Organic Anion Transport protein.

23. Use of a nucleic acid according to anyone of claims 1-6 or a protein according to claim 20 or 21 in the treatment of Dubin-Johnson disease, Rotor disease or another disease involving Canalicular Multispecific Organic Anion Transport protein.

24. Use of a nucleic acid according to anyone of claims 1-6 as a selectable marker gene.