Bioscience Reports, Vol. 17, No. 2, 1997

REVIEW

The GS-X Pump in Plant, Yeast, and AnimalCells: Structure, Function, and GeneExpression

Toshihisa Ishikawa,1'3 Ze-Sheng Li,2 Yu-Ping Lu,2 andPhilip A. Rea2

Received January 2, 1997

This review addresses the recent molecular identification of several members of the glutathioneS-conjugate (GS-X) pump family, a new class of ATP-binding cassette (ABC) transporters responsiblefor the elimination and/or sequestration of pharmacologically and agronomically important com-pounds in mammalian, yeast and plant cells. The molecular structure and function of GS-X pumpsencoded by MRP, cMOAT, YCF1, and AtMRP genes, have been conserved throughout molecularevolution. The physiologic function of GS-X pumps is closely related with cellular detoxification,oxidative stress, inflammation, and cancer drug resistance. Coordinated expression of GS-X pumpgenes, e.g., MRP1 and YCF1, and -y-glutamylcystaine synthetase, a rate-limiting enzyme of cellularglulathione (GSH) biosynthesis, has been frequently observed.

KEY WORDS: ABC-transporter; GS-X pump; multidrug resistance associated protein (MRP);canalicular membrane multispecific organic anion transporter (cMOAT); yeast cadmium factor 1(YCF1); •y-glutamylcystaine synthetase; glutathione.

INTRODUCTION

Plant and animal cells eliminate a broad range of lipophilic toxins from thecytosol after their conjugation with glutathione (GSH) (1-3). This transportprocess is mediated by a GS-X pump, a novel organic anion-transportingMg2+-ATPase. The term of "Gs-X pump" has been originally proposed based onits transport activity and high affinity toward glutathione S-conjugates (GS-conjugates), glutathione disulfide (GSSG), and cysteinyl leukotrienes (1). Thekinetic properties and substrate specificity of the GS-X pump have beenintensively studied using plasma membrane vesicles from different biologicalsources. Accumulating evidence suggests that the GS-X pump has a broad

1 Department of Medicinal Biology, Central Research, Pfizer Inc., 5-2 Taketoyo, Aichi 470-23, Japan.2 Plant Science Institute, Department of Biology, University of Pennsylvania, Philadelphia, Penn-

sylvania 19104-6018.3 To whom correspondence should be addressed.

189OI44-X4ft.V97/(l4(M)-(}|K<)$]2.50/00 I'M7 Plenum Publ ishing Corporation

190 Ishikawa, Li, Lu, and Rea

substrate specificity toward different types of organic anions and thereby plays aphysiologically important role in inflammation, oxidative stress, xenobioticsmetabolism, and tumor drug resistance. In plants, the GS-X pump is competent inthe transport of GS-conjugates of the chloroacetanilide and triazine herbicides,metolachlor and simetryn, respectively, as well as GSSG and GS-conjugates ofl-chloro-2,4-dinitrobenzene (CDNB) and monochlorobimane (2,3). It has re-cently been suggested that the GS-X pump is involved in the anthocyaninbiosynthetic pathway (4).

During the past three years, there was a remarkable progress in ourunderstanding of the molecular nature of the GS-X pumps. Overexpression of themultidrug resistance-associated protein (MRP1) gene (5,6) in human cancer cellshas been first reported to result in increased ATP-dependent GS-conjugatetransport, thus demonstrating that the MRP1 gene encodes a human GS-X pump(7, 8). A liver specific GS-X pump, named cMOAT, has also been cloned from ratliver cDNA libraries and it exhibited an extensive homology with human MRP1(9-11). Mutation of the cMOAT gene is implicated to be a cause of hereditaryhyperbilirubinemia associated with defective function of the hepatic GS-X pump.The yeast cadmium factor (YCF1) gene from Saccharomyces cerevisiae has beenidentified on the basis of its ability to confer cadmium resistance (12). The YCFIgene encodes an ATP-binding cassette (ABC) protein homologous to humanMRP1, and the protein has recently been identified as vacuolar GS-X pump inyeast cells (13). Furthermore, two cDNAs, i.e. AtMRPl and AtMRP2, encodingGS-X pump in a plant Arabidopsis thaliana have been cloned and they also showextensive sequence homologies to MRP1, cMOAT and YCFI (Lu, Y.-P., Li,Z.-S., Rea, P. A., unpublished). Thus, evidence is accumulating to support theidea that the GS-X pumps comprise a multigene family in both the animal andplant kingdoms.

GS-X PUMP AND GSH METABOLISM

One of the major functions of the GS-X pump is excretion and/orsequestration of toxic compounds as a cellular protection system. The metabolismand detoxification of xenobiotics comprises three main stages: phases I, II and III(1). In phase I, xenobiotics and endogenous substances are oxidized, reduced orhydrolyzed to expose or introduce a functional group of the appropriatereactivity. Cytochrome P450s and mixed function oxidases are examples of phaseI enzymes that confer the requisite electrophilicity on otherwise unreactivecompounds for their subsequent metabolism by phase II enzymes. In phase II, theactivated derivative is conjugated with GSH, glucuronic acid or sulfate by theaction of GSH transferases (GSTs), UDP-glucuronyl transferases, and sulfo-transferases. In the phase III, the resulting conjugates are transported out of thecytosol to the extracellular space or into intracellular compartments. Theunderlying principle is that the conjugation reaction confers negative charges onand increases the water solubility of the compound to promote its extrusion bythe GS-X pump and other unidentified transporters.

The GS-X Pump 191

The function of the GS-X pump is closely linked with cellular GSHmetabolism. Glutathionation provides negative charges to compounds andthereby enables the substrate recognition by the GS-X pump. GSH bears twoimportant features in its structure, namely, the y-glutamate linkage and theSH-group, both of which are intimately linked to its intracellular stability andbiological functions. The intracellular concentration of GSH in mammalian cells is1-10 mM (14), which is even higher than the intracellular concentration of ATP.In many cells, GSH accounts for more than 90% of the total nonprotein sulfur.Such high intracellular concentration of GSH is made possible by the -y-Glulinkage structure, which protects the GSH molecule from protease cleavage. TheSH-group' of GSH is strongly nucleophilic and confers to the molecule the uniqueability to react with a wide variety of agents including free radicals, reactiveoxygen species, heavy metals, and cytotoxic electrophilic compounds, therebyplaying a critical role of detoxification in living cells.

The biosynthesis of GSH takes place in the cytosol. The reaction consists oftwo steps catalyzed by -y-glutamylcysteine synthetase and by GSH synthetase,where each step requires one molecule by ATP (Fig. 1). y-Glutamylcysteinesynthetase (y-GCS) catalyzes the first step (reaction 1) in which an amide linkageis formed between the amino group of cysteine and the y-carboxyl group of

Fig. 1. Schematic illustration for the biosynthesis of GSH, conjugation of GSH with electrophiliccompounds or heavy metals (X), and ATP-dependent transport of GSH conjugates (GS-X) via theGS-X pump. In the animal cell, GS-X is exported across the plasma membrane or compartmental-ized into intracellular vesicles and subsequently released from the cell by exocytosis. In the yeastand plant cell, GS-X is transported by the GS-X pump into the vacuole.

192 Ishikawa, Li, Lu, and Rea

glutamate. In the second step, GSH synthetase catalyzes the reaction betweenglycine and the cysteine carboxyl group of y-glutamylcysteine:

L-Glu + L-Cys + ATP -> y-glutamylcysteine + ADP + Pi (Reaction 1)

y-Glutamylcysteine + L-Gly + ATP -» GSH + ADP + Pi (Reaction 2)

The reaction catalyzed by y-GCS is the rate-limiting step of GSH biosynthesisand is controlled by negative feedback from its end product, GSH, vianon-allosteric competitive inhibition (15).

The pathway of GSH-mediated drug inactivation is a biologically "expen-sive" mechanism. The synthesis of GSH and the export of GS-conjugates fromthe cytosol requires at least three molecules of ATP in order to metabolize 1 molof the drug molecule. In the case of the detoxification of cadmium, at least fivemolecules of ATP are required, namely, four molecules of ATP for the synthesisof two molecules of GSH and at least one molecule of ATP for the export ofbis-glutathionato cadmium(II), a glutathione-cadmium chelate complex. Thiscompares unfavorably with the glycolytic pathway, in which only two molecules ofATP are gained from one molecule of glucose. Nevertheless, the fact that theGS-X pump is ubiquitously distributed in the plant and animal kingdoms stronglysuggests that the GSH-associated metabolism and transport pathway is fun-damentally important for the survival of living cells.

It is estimated that GSH biosynthesis originated about 3.5 billion years ago.GSH is found in the vast majority of eukaryotes, whereas in eubacteria GSHbiosynthesis is limited to only two groups, i.e., cyanobacteria and purple bacteria(16). The former appeared on the earth about 35 billion years ago and wascapable of oxygenic photosynthesis. The cyanobacteria is the group considered tohave given rise to plant chloroplasts, whereas the purple bacteria is considered tohave introduced the ancestor responsible for eukaryotic mitochondria. GSHproduction appears to be closely associated with those prokaryotes responsiblefor the oxygen-producing and oxygen-utilizing pathways of eukaryotes, suggestingthat the ability of GSH biosynthesis may have been acquired by eukaryotes inthose endosymbiotic process that give rise to chloroplast and mitochondria (16).In fact, GSH plays a pivotal role in protection of living cells under oxidative stress(17). Intracellular glutathione disulfide (GSSG) is maintained at low levels (lessthan 3% of total GSH pool) by the action of GSSG reductase, whereas enhancedGSH peroxidase reaction (e.g., via redox-cycling of quinone compounds) leads toan increase in the cellular GSSG level and GSSG efflux from cells. GSSG effluxhas been observed in many different organs and cell types, including erthyrotytes,liver, lung, and heart. It is important to note that studies on the GS-X pumpstarted with the discovery of GSSG efflux from human erythrocytes.

In 1969, the frontier work was made by Srivastava and Beutler who reportedthat GSSG from human erythrocytes is an unidirectional and energy-dependentprocess (18). GSSG transport occurred even against a concentration gradient ofGSSG and the transport was halted almost entirely by exhausting endogenousATP by preliminary incubation of erythrocytes in a glucose-free medium for

The GS-X Pump 193

eight hours or by the presence of fluoride in the incubation medium. Their reportprovided the first evidence that GSSG efflux is mediated not by simple diffusionbut by active membrane transport. Although they suggested energy-dependenceof GSSG transport, it was not elucidated whether ATP is directly required.Eleven years later, 1980, the transport of GSSG across the plasma membrane wasproven to be an ATP-dependent "primary" active process in inside-out mem-brane vesicles from human erythrocytes (19). Subsequently, ATP-dependentGSH conjugate transport was demonstrated by Board (20), Kondo et al. (21) andLabelle et al. (22).

In 1984, we have reported that of GSSG and GSH conjugates are releasedfrom the isolated perfused heart (23). The heart is the organ continuouslyexposed to highly oxygenated blood from the lung. Cardiomyopathy resultingfrom oxidative damage inflicted by hyperoxia or administration of certainanticancer drugs, e.g., doxorubicin, has been described. GSSG efflux wassuggested to be an important defense mechanism against oxidative stress (24).The relationship of GSSG efflux rate vs. cytosolic free ATP/ADP ratio shows thatGSSG efflux rate is half-maximal at (ATP/AD)frei. = 10(25), suggesting anATP-dependent transport process. GSSG efflux from the heart was not affectedby epinephrine, nor-epinephrine or dibutyryl cyclic AMP, suggesting the GSSGtransport is independent of a- or /3-adrenergic hormonal regulations (25). Usingplasma membrane vesicles prepared from rat hearts, we have demonstratedATP-dependent primary active transport of GSSG and GS-conjugates (26).Cardiac GS-X pump was shown to have high affinities toward GS-conjugates witha long aliphatic carbon chain (25). We first provided evidence that leukotriene C4

(LTC4), a pro-inflammatory mediator, is an endogenous substrate for the GS-Xpump (26,27).

In the liver, GSSG and GS-conjugates are predominantly excreted into bile.Akerboom et al. reported that hepatobiliary transport of GSSG was inhibited byGS-conjugates, suggesting the existence of a common transport system forhepatobiliary transport of both GSSG and GS-conjugates (28). Although themembrane potential had been speculated to be a potential driving force for thetransport of GSSG and GSH conjugates (29), Kobayashi and his coworkersprovided evidence for ATP-dependent primary active transport of S-(2,4-dinitrophenyl)-glutathione (GS-DNP) using rat hepatocyte canalicular membranevesicles (30). Moreover, Koabayashi et al. reported that transport of p-nitrophenyl-glucuronide across rat liver canalicular membrane was an ATP-dependent process (31). Inhibition of the transport of the glucuronide conjugateby GS-DNP suggested that the hepatobiliary elimination of glucuronide conjug-ates is mediated by the hepatic GS-X pump (31).

The discovery of two strains of jaundiced mutant Wistar and Sprague-Dawleyrats, i.e., TR and EHBR, greatly enhanced the study of the hepatic GS-X pump.The mutant rats manifest predominantly conjugated hyperbilirubinemia and aredefective in the biliary secretion of GSH, GSSG, GS-conjugates (32-34),GSH-metal complexes (35), bilirubin-glucuronide conjugates (36) and cysteinylleukotrienes (37, 38), as well as organic anions, e.g., bromosulphthalein, indo-

194 Ishikawa, Li, Lu, and Rea

cyanine green and dibromosulfophthalein (39,40). These findings stimulated theidea that the hepatobiliary GS-X pump has a broad substrate specificity towarddifferent organic anions. Thus, the hepatic GS-X pump is also called "canalicularmultispecific organic anion transporter (cMOAT)". The functional defect of thehepatic GS-X pump or cMOAT is inherited with characteristics of an autosomalrecessive abnormality, and those mutant rats are regarded as animal models ofthe Dubin-Johnson syndrome in humans. Interestingly, erythrocyte membranesfrom patients of the Dubin-Johnson syndrome and from the TRT mutant ratsexhibited ATP-dependent transport of GS-DNP and GSSG as normal (41). Thus,it has been suggested that erythrocyte and hepatic GS-X pumps are encoded bydistinct genes. This hypothesis has recently been verified by molecular identifica-tion of hepatic and extrahepatic GS-X pump genes, i.e. cMOAT and MRP1 (seebelow).

The existence of the GS-X pumps in yeast and plants has also been reportedby St-Pierre et al. (42), Martinoia et al. (2) and Li et al. (3). In yeast and plantcells, instead of excretion, conjugates of xenobiotics are stored in the vacuole.The GS-conjugates of l-chloro-2,4-dinitrobenzene (CDNB), N-ethylmaleimideand metalachor were barely taken up by isolated vacuoles in the absence of Mg2+

and ATP. In the presence of Mg2+ and ATP, however, uptake was markedlyenhanced. ATP-dependent accumulation of GS-conjugates was inhibited byvanadate, but not by bafilmoycin, a specific inhibitor of the vacuolar H+-ATPases(3). Thus, ATP-dependent accumulation of GS-conjugates in the vacuole isindependent of H+ gradient across the membrane. Furthermore, GSSG was also,actively transported into isolated vacuoles, while uptake of GSH was onlymarginal (43). The uptake of GSSG was completely inhibited by the GS-conjugate of the herbicide metolachor, indicating that transport of GSSG ismediated by a vacuolar GS-X pump. In many functional respects, vacuolar GS-Xpumps of yeast and plants show a striking resemblance to the GS-X pump foundin rats and humans.

MOLECULAR IDENTIFICATION OF GS-X PUMPS

Identification of the GS-X pump in human, rat, yeast, and plants at themolecular level has served to confirm its wide distribution and demonstrate thatGS-conjugate transporters constitute a multigene family within the ABC trans-porter superfamily. In 1994, Muller et al. and Leier et al. provided evidence thatoverexpression of MRP1 gene, first isolated from human small cell lung cancercell lines (5,6), conferred increased Mg-ATP-dependent GSH conjugate transport(7,8), thus demonstrating that MRP1 gnee encodes a human GS-X pump. Sincethen, other closely related genes have been characterized. The liver-specific GS-Xpump, i.e., cMOAT, has been cloned from cDNA libraries of TR and EBHR ratliver (9-11). The yeast cadmium factor (YCF1) gene, which was originallyisolated on the basis of its ability to confer cadmium resistance (12), was found tobe a vacuolar GS-X pump in Saccharomyces cerevisiae (13). Most recently, twogenes (AtMRPl, AtMRP2) encoding plant GS-X pumps have been isolated from

The GS-X Pump 195

Fig. 2. Phylogeny tree of ABC transporters. Human MRP1,human cMOAT, rat cMOAT and yeast YCF1 belong to theGS-X pump family.

a vascular plant, Arabidopsis thaliana, and they exhibit extensive homology withMRP1, YCF1 and cMOAT (Lu, Y.-P., Li, Z.-S., Rea, P. A., unpublished data).

Sequence comparisons of MRP1, cMOAT, YCF1, AtMRPl and AtMRP2with other members of the ABC transporter superfamily reveal two majorsubgroups (Fig. 2). One consists of MRP1 (5,6,44), cMOAT (9-11,45), YCF1(12), AtMRPl, AtMRP2 (Lu, Y.-P. Li, Z.-S., Rea, P. A., unpublished data), SUR(46), EBCR (47), Yrs/Yorl (48), the Leishmania P-glycoprotein-related molecule(Lei/PgpA) (49-51) and the cystic fibrosis transmembrane conductance regulator(CFTR) chloride channel (52). The other consists of the P-glycoprotein (MDR)(53, 54), major histocompatibility complex transporters (TAP1 and TAP2) (55)and STE6 (56). Of all the ABC transporters defined to date, MRP1, cMOAT,

196 Ishikawa, Li, Lu, and Rea

YCF1, AtMRPl and AtMRPZ are known to have GS-X pump activity. Unlikethe similarities between the GS-X pump subgroup, Lei/PgpA and CFTR, whichcenter on the nucleotide-binding folds (NBFs), those within the GS-X pumpfamily are found throughout the sequence. GS-X pump family members are40-45% identical (60-65% similar) at the amino acid level, possess NBFs with anequivalent spacing of conserved residues and are collinear with respect to thelocation, extent and alternation of putative transmembrane spans and ex-tramembrane domains. Two features of members of the GS-X pump family thatdistinguish them from other ABC transporters are their possession of a centraltruncated CFTR-like "regulatory" domain, rich in charged amino acid residues,and an approximately 200 amino acid residue N-terminal extension.

The conclusion that MRP1 gene encodes a human GS-X pump transportingmultivalent organic anions is deduced from the following observations: (a)ATP-dependent transport of LTC4 and GS-conjugates observed in plasmamembrane vesicles from MRPl-overexpressing, doxorubicin-resistance HL60/ADR cells was greater than that observed in drug-sensitive cells (57). (b) A190-kDA membrane protein was detected by [3H]LTC4 photoaffinity labeling inthe plasma membrane from HL60/ADR cells (57). (c) The [3H]LTC4-labeled190 kDa membrane protein was immunoprecipitated by MRP-specific monoclonalantibody (57): (d) Plasma membrane vesicles prepared from MRP-transfectedcells exhibited ATP-dependent Transport of LTC4, GS-DNP, GSSG, 17-/3-estradiol-17-(0-D-glucuronate, and sulfate conjugates (7,8,58-61). (e) MRP-Mediated transport of [3H]LTC4 was strongly inhibited by GSH conjugatescarrying a long aliphatic carbon chain, but not inhibited by doxorubicin (60),being consistent with our previous observations on GS-X pump function(1,24,26,27). (/) MRP is expressed in a variety of normal tissues and cell types,including heart, skeletal muscle, lung, adrenal gland, erythrocytes, and macro-phages (62), however its expression level in the liver is very low (6,62).

cMOAT, the liver-specific GS-X pump, was cloned from cDNA libraries of ratliver (9-11), and from a cisplatin-resistant human epidermoid cancer cell line,KB/KCP4 (44). While MRP1 gene is localized on 16pl3.12-13 (5), humancMOAT gene was found on 10q24 (44), thus demonstrating distinct localization ofhepatic and extrahepatic GS-X pump genes. Rat cMOAT consists of 1541 aminoacids, exhibiting highest overall identity to human MRP1 (47.6%), YCF1(41.8%), and CFTR (30.2%) (9-11). cMOAT is almost exclusively expressed inthe liver and to a lesser extent in the duodenum, jejunum and kidney, whereasnone is detectable in the brain, heart, lung, testis, and skeletal muscle (9,11).Double immunofluorescence and confocal laser scanning microscopy revealedexclusive localization of cMOAT at the canalicular membrane domain ofhepatocytes in the normal rat, but its loss in TR" and EHBR rats (10). NocMOAT mRNA was detectable in these mutant rats (9-11). In TR" rat, asingle-nucleotide deletion at amino acid 393 resulted in the introduction of thestop codon at amino acid 401 (9). In EHBR rat, one base pair replacement fromG to A at nucleotide 2564 resulted in the introduction of the premature stopcodon at the corresponding amino acid 855 in all tissues examined (11). SinceEHBR and TR" are allelic mutants and both strains exhibit the autosomal

The GS-X Pump 197

recessive inheritance in hepatobiliary excretion of organic anions, it is concludedthat the impaired expression of this hepatic GS-X pump, cMOAT, is related tothe pathogenesis of hyperbilirubinemia in the mutant rats. Genetic alterationsassociated with Dubin-Johnson syndrome in humans remain to be elucidated.

YCF1 gene encodes a yeast vacuolar GS-X pump consisting of 1,515 aminoacids with extensive homology to MRP1 (42,6% identity, 63.3% similarity) andCFTR (31% identity, 56.7% similarity). The protein encoded by YCF1 gene isrequired for cadmium resistance in S. cerevisiae. Another type of cadmiumresistance gene designated HMT1 has been identified in Schizosaccharomycespombe, and this protein contains a single transmembrane domain and nucleotidebinding fold (63,64). Sequence comparison reveals that HMTI belong to asubclass of ABC transporters which is distinct from YCF1, MRP1, cMOAT andCFTR. S. cerevisiae, cells harboring a deletion of the YCF1 gene are hypersensi-tive to cadmium compared to wild type cells. Mutagenesis experiments havedemonstrated that conserved amino acid residues, functionally critical in MRP1and CFTR, play a vital role in YCFl-mediated cadmium resistance (12).Mutagenesis of phenylalanine 713 in the NBF1 of YCF1 completely abolished incadmium detoxification. Furthermore, substitution of a serine to alanine residuein a potential protein kinase a phosphorylation site in a central region of YCF1,which displays sequence similarity to the central regulatory domain of MRP1 andCFTR, also rendered YCF1 nonfunctional, suggesting that phosphorylation of theserine residue is critical for GS-X pump activity of YCF1.

TRANSPORT OF GSH-METAL COMPLEXES BY GS-X PUMP

The GS-X pump plays an important role in excretion or sequestration ofheavy metals and electrophilic compounds after conjugation with cellular GSH. Liet al. (13,65) have recently provided evidence that YCF1 sequestrates aGSH-cadmium complex, i.e. bis-glutathionato cadmium(II), as well as GS-conjugates from the cytosol into the vacuole in the yeast. Disruption of the YCF1gene resulted in yeast strains (DTY167 and DTY168) hypersensitive to cadmium(12). While wild type cells mediate vacuolar accumulation of the fluorescentGS-conjugate of monochlorobimane, the mutant DTY167 cells lack the activity ofvacuolar accumulation (13). Introduction of plasmid borne, epitope-tagged geneencoding functional YCF1 and DTY167 cells restored cellular resistance tocadmium concomitant with the activity for ATP-dependent transport of DNP-SGand bis-glutathionato cadmium(II) (GS2-Cd) in vacuolar membrane vesicles(13, 65). ATP-dependent transport of GS2-Cd in vacuolar membrane vesicles wasinhibited by GS-DNP, suggesting a common pathway for the transport of thosecompounds (65). YCF1 mRNA level as well as GS-X pump activity in DTY165cells (wild type) was dramatically enhanced by pretreatment with 200 /u,M CdSO4

or 150 /iM CDNB for 24 h (65).We previously demonstrated that human GS-X pump encoded by the

MRP1 gene eliminates a GSH-platinum (GS2-Pt) complex, bis-glutathionato

198 Ishikawa, Li, Lu, and Rea

platinum(II), from cancer cells (66). The Km value for the GS2-Pt complex wasestimated to be 100 ̂ M, being similar to that for GSSG (67). The MRP1 gene isoverexpressed in cisplatin-resistant human leukemia HL-60/R-CP cells, whereasno gene amplification was detected (67,68). Instead, MRP1 mRNA level inHL-60/R-CP cells was significantly induced within 48 h by incubation of cells withcisplatin, cadmium, zinc or arsenite (68). Tommasini et al. has recently reportedthat transfection of the human MRP cDNA in yeast mutant DTY168 cellsrestored cadmium resistance to the wild-type level (69), suggesting that MRP1functionally complements YCF1. MRPl-transfected cancer cells are reportedlyresistant to arsenite and antimony (70). MRPl-mediated efflux of arseniteaccompanies GSH efflux, supporting the hypothesis that MRP1 transports theGSH-arsenite complex (7f). Through the reaction with cellular GSH, heavymetals are converted to organic anions. The multivalent negative charge is crucialfor the recognition of substrates by GS-X pumps, where one molecule of heavymetal must be ligated by at least two molecules of GSH. Mono(glutathionato)-cadmium is not substrate for YCF1 (65) or MRP1 (Ishikawa, T. unpublishedobservation). Furthermore, it is important to note that the overexpression of theMRP1 gene per se does not necessarily result in cellular resistance to heavymetals. In the case of cisplatin, formation of the GS2-Pt complex in the cells isslow and this is the rate limiting step of the over-all reaction including the exportprocess (66), whereas the reaction of cadmium with GSH is much faster (65).High cellular GSH levels are important to propel the interaction of heavy metalswith GSH and to shift the thermodynamic equilibrium toward the formation ofGSH-heavy metal complexes. y-Glutamylcysteine synthetase (y-GCS) plays animportant role in the regulation of cellular GSH levels, thereby modulating theGS-X pump function.

In cisplatin-resistant human epidermoid carcinoma KCP-4 cells, Taniguchi etal. (45) and Chuman et al. (72) have recently shown that cMOAT may function asa transporter for GS2-Pt complex to reduce cellular accumulate of cisplatin.MRP1 gene was not expressed in their cisplatin-resistant cell line (45,72). To ourknowledge, the paper by Taniguchi et al. (45) is the first report demonstrating theexpression of cMOAT gene in human cancer cells. In conjunction with ourprevious studies (66-68), these reports suggest that GS-X pumps encoded byMRP1 and cMOAT genes have an extensive overlapping substrate specificitytoward organic anions, including GSH-metal complexes.

Hepatobiliary transport of GSH is critically involved in the biliary excretionof the physiologically important copper and zinc, as well as the toxic metals, suchas arsenic, mercury, cadmium, lead, antimony and bismuth (73-80). It has beenshown that excretion of these metals in bile is increased when biliary excretion ofendogenous GSH is enhanced (73, 75). In contrast, biliary excretion is diminishedwhen hepatobiliary transport of GSH is decreased by agents that depletes hepaticGSH (e.g., diethylmalate) or inhibit of GSH transport from liver to bile (e.g.,sulfobromophthalein and indocyanine green) (77-79). In transport mutant (GYor TR ) rats, hepatobilary transport of GSH and intravenously injected metals,e.g., copper, zinc, and selenium, was severely impaired (35,80). Th underlyingmechanism for the GSH-dependent biliary excretion of those metals may involve

The GS-X Pump 199

(a) complexation of metals with hepatic GSH, and (b) subsequent hepatobiliarytransport of GSH-metal complexes mediated by cMOAT.

SEQUESTRATION BY GS-X PUMP

Plant GS-X pump plays a role in vacuolar sequestration of pigments, naturalherbicides, allelochemicals and pathogen-related compounds (2,3) (Fig. 1).Recent functional analyses of the maize gene, Bronze-2, which participates inanthocyanin pigment biosynthesis, have provided compelling evidence thatendogenous compounds are metabolized and sequestered into the vacuole in amanner similar to xenobiotics (4). Anthocyanins share a common biosyntheticorigin and core structure based on cyanidin-3-glucoside (81). It is through thespecies-specific modification of cyanidin-3-glucoside by hydroxylation, methyla-tion, glucosylation and acylation that the wide spectrum of red, blue, and purplecolors in the vacuoles of flowers, fruits, and leaves is produced. Walbot andcolleagues (4) have recently shown that the Bronze-2 (GZ-2) gene in maizeencodes a GST gene responsible for conjugating anthocyanin with GSH. bz-2mutants, on the other hand, are unable to pump pigments from the cytosol intothe vacuole lumen, being due to defective in the glutathionation of anthocyanins(4). Thus, GS-conjugation of anthocyanin appears to be a prerequisite for thevacuolar accumulation of anthocyanin pigments and the vacuolar GS-X pumpmay transport the GS-conjugation of anthocyanin into the lumen.

Yeast and animal cells also sequestrate GS-conjugates from the cytosol intosubcellular compartments. Using monochlorobimane, its fluorescent GS-conjugate was demonstrated to be sequestrated into endosomal intracellularvesicles in the primary culture of rat hepatocytes (82) and cisplatin-resistantHL-60/R-CP cells (67). In yeast cells, the fluorescent conjugate was accumulatedin the central vacuole. The involvement of cMOAT, MRP1, and YCF1 insequestration of GS-conjugates is supported by the following observations: (a)Monochlorobimane, a nonfluorescent compound, is specifically conjugated withGSH in the cell by the action of GSTs, and the resulting fluorescent GS-conjugates is a substrate for the GS-X pump, (b) When cultured normal and TR~rat hepatocytes were loaded with the GS-conjugate of bimane, both cell typesdisplayed a strong cytosolic fluorescence. Normal rat cells completely lost thecytosolic fluorescence during incubation in monochlorobimane-free mediumbecause of excretion of the GS-conjugate. However, fluorescent vesicles wereobserved in the perinuclear region and around a canaliculus. In contrast,TR-cells lost their cytosolic fluorescence more slowly and completely lackedvesicular fluorescence (82). (c) Likewise, the accumulation of fluorescence inintracellular vesicles was more prominent in HL-69/R-CP cells that in HL-60cells. The cisplatin-resistant HL-60/R-CP cells expressed MRP1 gene at highlevels. The vesicular accumulation of the fluorescent GS-conjugate in HL-60/R-CP cells was inhibited by ATP depletion (67). (d) In cultured hepatocytes andcisplatin-resistant HL-60/R-CP cells, vesicular fluorescence was significantlyincreased by preincubation with monensin or methylamine that interfere with

200 Ishikawa, Li, Lu, and Rea

vesicular trafficking out of the trans-Golgi complex (82,67), suggesting thatGS-conjugates accumulated in intracellular vesicles are excreted by exocytosis. (e)DTY165 (wild type) cells exhibited the vacuolar accumulation of the fluorescentGS-conjugate, whereas such vacuolar accumulation was abolished in DTY167(ycflA strain) cells (13). The capacity for ATP-dependent, uncoupler-insensitiveDNP-SG uptake was strictly associated with vacuolar membrane vesicles purifiedfrom DTY165 cells (13). These results support the idea that intracellularsequestration of GS-conjugate occurs in plant, yeast, and animal cells, and thatthe GS-X pump mediates "storage export" of GS-conjugates, and probably othermultivalent organic anions (2).

REGULATION OF GS-X PUMP GENE EXPRESSION

At present, information on molecular mechanisms involved in the regulationof GS-X pump gene expression is substantially limited. GS-X pump activity inplant cells was shown to be significantly enhanced by CDNB, a substrate forGSTs (83). This increase of GS-X pump activity could be ascribed to enhanced denovo synthesis of GS-X pump protein and/or increased recruitment of GS-Xpump molecules in the vacuolar membrane.

YCF1 in the yeast was shown to be induced by pretreatment with cadmiumor CDNB (65). RNase protection assays of YCF1 expression in DTY165 cellsrevealed that YCF1 mRNA levels were increased by 1.9- and 2.5-fold by 200 ju,MCdSO4 and 150/iM CDNB, respectively. Concomitantly, the GS-X pump activityassessed by ATP-dependent GS2-CD and GS-DNP was increased to similarextents. Wu and Moye-Rowley have revealed that the YCF1 gene and the GSH1gene encoding y-GCS are coordinately regulated by yAP-1 which encodes yeasttranscription factor AP-1 (84). Transcriptional activation of these genes mediatedby yAP-1 is essential for cadmium resistance in yeast cells (85). Both organicelectrophiles and heavy metals induce YCF1 expression and yAP-1 protein, amember of the basic-leucin zipper (BZIP) family of transcription factors,transcriptionally activates YCF1 and GSH1 genes (84,86).

In cisplatin-resistant HL-60/R-CP cells, our group has demonstrated thatMRP1 and -y-GCS genes are coordinately induced by cisplatin and heavy metals,such as arsenite, cadmium, and zinc (68). In addition, human glioma A172 cellspretreated with ACNU for 24 hours enhanced mRNA levels of both MRP1 andy-GCS by 3- to 5-fold (87). These results strongly suggest that expression ofMRP1 gene is closely related with cellular GSH biosynthesis and that certaincommon factor(s) may regulate the expression of both MRP and -y-GCS genes.Since c/.y-regulatory elements including AP-1 binding sites have been identified inthe promoter regions of human y-GCS and MRP genes (88-90) (Fig. 3), it isextremely interesting to examine whether AP-1 is responsible to transcriptionallyregulate the expression of MRP1 and y-GCS genes in mammalian cells.

The fact that MRP1 and y-GCS can be coordinately induced by cisplatin,l-(4-amino-2-methyl-5-pyridinyl)methyl-3-(2-chloroethyl)3-nitrosourea (ACNU),and heavy metals provide important information to our understanding of how

The CS-X Pump 201

Fig. 3. Potential regulation sites in the promoter regions of human MRP1 and y-GCS genes. CRE,cyclic AMP response element; ERE, estrogen response element; GRE glucocorticoide responseelement, XRE, xenobiotic response elements, EpRE, Electrophile response element; CAT box,CCAAT box. Data from refs. 88-90.

drug resistance genes are acutely induced upon drug treatments. Unlike cellculture studies where drug-resistant variants are usually obtained throughcontinuous drug exposure, such transient induction of drug-resistance geneexpression is more directly related to cancer chemotherapeutic protocols.Interestingly, human colorectal cancers frequently overexpress MRP1 mRNA(91) (Fig. 4). Because the patients involved in the study had not been treated withchemotherapeutic agents, it is suggested that up-regulation of different drug genesare associated with different human cancers and, more importantly, developmentof drug resistance in these cancers is an intrinsic mechanism. Because intrinsicdrug resistance is the major factor controlling efficacy of chemotherapy in cancertreatment, a better understanding of how different drug-resistance genes areregulated in different tumor systems is of great importance for development ofeffective strategies to circumvent drug resistance in cancer chemotherapy.

MECHANISMS OF ANTITUMOR DRUG RESISTANCE MEDIATED BYMRP1/GS-X PUMP

Human MRP1 cDNA was originally isolated from a doxorubicin-selected,multidrug resistant lung cancer cell line (5). Overexpression of MRP1 has beensubsequently observed in many doxorubicin-resistant cell lines. Furthermore,transfection of expression vectors harboring human MRP1 cDNA reportedlyconfers resistance to doxorubicin and many other antitumor drugs, e.g., vincris-tine, vinblastine, and VP-16, in otherwise drug-sensitive cell lines (70). These

202 Ishikawa, Li, Lu, and Rea

Fig. 4. Coordinated expression of MRP1 and•y-GCS in human colorectal cancers. Relativelevels of MRP1 and y-GCS mRNA (in arbit-rary units) are plotted. The correlationcoefficient is r = 0.779. Data from Kuo et al.(ref. 91),

results suggest the important role of MRP1 expression and drug resistance incultured mammalian cells. To date, however, there is little knowledge about themechanism by which GSH participates in MRP-mediated drug resistance. Theseantitumor agents are poor substrates in membrane vesicles transport assayprepared from MRPl-overexpressing cells and also poor inhibitors for LTC4

transport mediated by MRP1. On the other hand, in the presence of GSH,inhibitions of ATP-dependent transport of LTC4 by certain antitumor drug (e.g.,vincristine, VP-16 and Taxol) can be enhanced (60). These findings are consistant,in part, with the notion that the presence of GSH can facilitate MRP-mediateddrug transport.

It was reported that overexpression of MRP1 by transfection of its cDNAresulted in increased resistance to doxorubicin in the transfected cells. However,plasma membrane vesicles prepared from the transfected cells failed to displaytransport activity of doxorubicin (59, 60; cf. 92). In our study using membranevesicles, doxorubicin per se was not the direct substrate for the GS-X pumpencoded by the MRP1 gene. ATP-dependent transport of GS-conjugates andLTC4 in plasma membrane vesicles prepared from MRP1 cDNA-transfected cellsare not inhibited by doxorubicin (93). Moreover, the 190-kDa protein inMRPl-overexpressing H69AR cells was not labeled with a photoaffinity analog ofdoxorubicin (93). On the other hand, cellular GSH has been shown to be acritical factor for the export of daunorubicin by MRP1 (71). Non-P-glycoprotein-

The GS-X Pump 203

mediated resistance to doxorubicin is related to cellular GSH levels andGSH-metabolizing enzyme systems (94). Thus, it is conceivable that GSH mayform conjugates with the antitumor drug prior to transport.

We previously proposed a hypothetical scheme to explain that glutathionateddoxorubicin may be recognized by MRP1/GS-X pump as shown in Fig. 5. In thisscheme, two-electron reduction of anthracyclines yields active intermediates.Reductive activation of anthracycline compounds is known to lead to covalentbinding to proteins and nucleic acids (96). Mitochondrial sulfhydryl groups aremodified by doxorubicin aglycones (97). In mammalian cells, two-electronreduction of quinone compounds is catalyzed by DT-diaphorase[NAD(P)H:quinone oxidoreductase], and this enzyme activity is greatly

Fig. S. A pu ta t i ve metabolic pa thway of the reductive bioactivation and GSHconjugation of doxorubicin as well as the subsequent transport of the GS-conjugate ofdoxorubicin aglycon.

204 Ishikawa, Li, Lu, and Rea

increased in the preneoplastic state (98). The reduced doxorubicin readilyundergoes deglycosylation. The C-7 of the aglycone intermediate is electrophilicand can react with nucleophiles, such as GSH, cysteine, N-acetyl cysteine, andprotein thiol groups. The reaction of the aglycone with GSH may be catalyzed byGSTs. The GS-conjugates thus formed are recognized by the MRP1 as itssubstrates and transported either immediately or after oxidation by molecularoxygen and/or by autocatalysis with the original drug. This scheme indicates thatboth the MRP1 and the metabolic enzyme system including DT-diaphorase andGSTs, as well as intracellular GSH, may be critically involved in the mechanismunderlying doxorubicin resistance.

To test our hypothesis, we have prepared doxorubicin- and daunorubicin-conjugates by linking the cysteine residue of GSH to aglycon of these anthracycl-ines at C-7 or C-14 position. The conjugates were purified by HPLC and thestructures of the conjugates were confirmed by NMR and MS spectra. Beingconsistent with our previous findings, doxorubicin and daunorubicin did notinhibit ATP-dependent LTC4 transport in plasma membrane vesicles preparedfrom MRPl-overexpressing SR3A cells. In contrast, GSH-conjugates ofdoxorubicin- and daunorubicin-aglycons competitively inhibited LTC4 transport(95), suggesting that glutathionation of anthracyclines, or addition of negativecharges on them, may facilitate the transport of these antitumor drugs inMRPl-overexpressing cancer cells.

While glutathonation of doxorubicin and daunomycin may be a straightfor-ward mechanism of MRP1-mediated resistance to these antitumor drugs. Itremains possible that GSH may be transiently associated with these antitumordrugs and/or MRP thereby creating favorable conformational conditions for drugtransport, in viewing the fact that GSH-doxorubicin complex has not beendetected in the culture medium of MRPl-overexpressing cells. In any event, themolecular bases involved in MRP1-associated overexpressing cells. In any event,the molecular bases involved in MRP1-associated drug resistance remain to becritically investigated.

CONCLUDING REMARKS

GS-X pumps in plants and animals exhibits a striking resemblance in thefunction and molecular structure, thereby comprising a novel ABC transporterfamily. It is important to note that the GS-X pumps are expressed in normaltissues and cell types and their function is closely associated with the cellularprotection mechanism. The glutathione moiety is an important property formaximal affinity to the active site of the GS-X pumps, whereas the moiety itself isnot considered to be a structural determinant. Thus, GS-X pumps have a broadsubstrate specificity toward different types of substrates that contain a largehydrophobic moiety and two negative charges. Like cMOAT, human MRP1transported not only GS-conjugates, glucuronide and sulfate conjugates, such as17/3-glucuronosyl estradiol, glucuronosyl etoposide, 3a-sulfatolithocholyl taurine.

The GS-X Pump 205

These findings are in support of the idea that the GS-X pump is a member of the"phase III" detoxification system.

Recent molecular identification of the GS-X pump in human, rat, yeast andplants has shed light on the genetic diversity of the GS-X pump family.Regulation of GS-X pump gene expression is coming into the focus. Site-directedmutation techniques would provide powerful tool to study the gene regulationmachinery, as well as the structure-activity relationship of the GS-X pumpmolecule. The next important goal would be to understand the molecularmechanism involved in the energy transfer within the pump molecule from ATPhydrolysis to the transport of organic anions. Furthermore, knock-out ortransgenic animals and plants may uncover unknown physiological functions ofthe GS-X pump. Our recent study suggests that the GS-X pump may modulatecell cycle arrest of human cancer cells induced by anticancer prostaglandins(99,100). The GS-X pump is considered to be a novel and important target ofdrug discovery. In this context, we have just entered the new era of GS-X pumpresearch that propels both basic and applied sciences.

ACKNOWLEDGMENT

The authors thank Professor M. T. Kuo (Department of Molecular Pathol-ogy, University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030)for his helpful discussion and critical reading of this article.

REFERENCES

1. Ishikawa, T. (1992) Trends Biochem. Sci. 17:463-468.2. Martinoia, E., Grill, E., Tommaini, R., Kreuz, K. and Amrhein, N. (1993) Nature 364:247-249.3. Li, Z.-S., Zhao, Y. and Rea, P. A. (1995) Plant Physiol. 107:1257-1268.4. Marrs, K., Alfenito, M. R., Lloyd, A. M., and Walbot, V. (1995) Nature 375:397-400.5. Cole, S. P. C, et al. (1992) Science 258:1650-1654.6. Zaman, G. J. R., et al. (1994) Proc. Nad. Acad. Sci. USA 91:8822-8826.7. MUller, M., et al. (1994) Proc. Nad. Acad. Sci. USA 91:13033-13037.8. Leier, I., Jedlitschky, G., Buchholz, U., Cole, S. P. C., Deeley, R. G. and Keppler, D. (1994) J.

Biol. Chem. 269:27807-27810.9. Paulusma, C. C., et al. (1996) Science 271:1126-1128.

10. BUchler, M., Konig, J., et al. (1996) J. Biol. Chem. 271:15091-15098.11. Ito, K., Suzuki, H., Horohashi, T., Kume, K., Shimizu, T. and Sugiyama, Y. (1996) Am. J.

Physiol (in press).12. Szczypka, M. S., Wemmie, J., Moye-Rowley, W. S. and Thiele, D. J. (1994) J. Biol. Chem.

269:22853-22857.13. Li, Z.-S., Szczypka, M., Lu, Y.-P., Thiele, D. J. and Rea, P. A. (1996) J. Biol. Chem.

271:6509-6517.14. Meister, A. (1989) In: Glutathione: Chemical, Biochemical, and Medical Aspects—Part A

(Dolphin, D., Poulson, R. and Avramovic, O. eds), John Wiley and Sons, New York, pp.367-474.

15. Meister, A. and Anderson, M. (1983) Ann. Rev. Biochem. 52:711-760.16. Fahey, R. C. and Sundquist, A. R. (1991) Adv. Enzymol. Relat. Mol. Biol. 64:1-53.17. Sies, H. (1985) Oxudative Stress, Academic Press, London.18. Srivastava, S. K. and Beutler, E. (1969) J. Biol. Chem. 244:9-16.19. Kondo, T., Dale, G. and Beutler, E. (1980) Proc. Natl. Acad. Sci. USA 77:6359-6362.

206 Ishikawa, Li, Lu, and Rea

20. Board, P. (1981) FEBS Lett. 124:163-165.21. Kondo, T., Murao, M. and Taniguchi, N. (1982) Eur. J. Biochem. 125:551-554.22. Labelle, E. F., Singh, S. V., Srivastava. S. K., and Awasthi, Y. C. (1986) Biochem. J.

238:443-449.23. Ishikawa, T. and Sies, H. (1984) J. Biol. Chem. 259:3838-3843.24. Ishikawa, T., Esterbauer, H. and Sies, H. (1986) J. Biol. Chem. 261:1576-1581.25. Ishikawa, T., Zimmer, M. and Sies, H. (1986) FEBS Lett. 200:128-132.26. Ishikawa, T. (1989)J. Biol. Chem. 264:17343-17348.27. Ishikawa, T., Kobayashi, K., Sogame, Y. and Hayashi, K. (1989) FEBS Lett. 259:95-98.28. Akerboom, T. P. M., Bilzer, M. and Sies, H. (1982) FEBS Lett. 140:73-76.29. Inoue, M., Akerboom, T. P. M., Sies, H., Thao, T. and Arias, I. M. (1984) J. Biol. Chem.

259:4998-5002.30. Kobayashi, K., Sogame, Y., Hara, H. and Hayashi, K. (1990) J. Biol. Chem. 265:7737-7741.31. Kobayashi, K., Komatsu, S., Nishi, T., Hara, H. and Hayashi, K. (1991) Biochem. Biophys. Res.

Commun. 176:622-626.32. Oude Elferink, R. P. J., Ottenhoff, R., Liefting, W., de Haan, J. and Jansen, P. L. M. (1989) J.

Clin. Invest. 84:476-483.33. Oude Elferink, R. P. J., Ottenhoff, R., Lielting, W. G. M., Schoemaker, B., Groen, A. K. and

Jansen, P. L. M. (1990) Am. J. Physiol. 258:G699-G706.34. Takikawa, H. et al. (1991) Hepatology 14:352-360.35. Houven, R., Dijkstra, M., Kuipers, F., Smit, E. P., Havinga. R. and Vonk, R. J. (1990) Biochem.

Pharmacol. 6:1039-1044.36. Jansen, P. L. M., Peters, W. H. M., Meijer, D. K. F. (1987) Gastroenterology 93:1094-1103.37. Huber, M., Guhlmann, A., Jansen, P. L. M. and Kepplcr. D. (1987) Hepatology 7:224-228.38. Ishikawa, T., Miiller, M., Klunemann, C., Shaub, T. and Keppler, D. (1990) J. Biol. Chem.

265:19279-19286.39. Jansen, P. L. M., Groothuis, G. M. M., Peters, W. H. M. and Meijer, D. F. M. (1987) Hepatology

7:71-76.40. Sathirakul, K., Suzuki, H., Yamada, T., Hanano, M., Sugiyama, Y. (1993) J. Pharmacol. Exp.

Therap. 268:65-73.41. Board, P., Nishida, T., Gatmaitan, Z., Che, M. and Arias, I. M. (1992) Hepatohgv 15:722-725.42. St-Piere, M., Ruetz, S., Epstein, L. F., Gros, P. and Arias, 1. M. (1994) Proc. Natl. Acad. Sci.

USA 91:9476-9479.48.43. Tommasini, R., Martinoia, E., Grill, E., Dietz. K.-J. and Amrhein, N. (1993) Z. Naturforsch.

48c: 867-871.44. Paul, S., Belinsky, M. G., Shen, H. and Kruh, G. D. (1996) Biochemislrv 35:13647-13655.45. Taniguchi, K. et al. (1996) Cancer Res. 56:4124-4129.46. Aguilar-Bryan, L. (1995) Science 268:423-426.47. van Kuick, M. A. (1996) Proc. Natl. Acad. Sci. USA 93:5401-5406.48. Cui, Z., Hirata, D., Tsuchiya, E., Osada, H. and Miyakawa, Y. (1996) J. Biol. Chem.

271:14712-14716.49. Ouellette, M., Fase-Fowler, F. and Borst, P. (1990) EMBO J. 9:1027-1033.50. Callahan, H. L. and Beverley, S. M. (1991)J. Biol. Chem. 266:18427-18430.51. Papadopoulou, B., Roy, G., Dey, S., Rosen, B. P. and Ouellette, M. (1994) 7. Biol. Chem.

269:11980-11986.52. Riodan, J. R. et al. (1989) Science 245:1066-1073.53. Wndieott, J. A. and Ling, V. (1989) Annu. Rev. Biochem. 58:137-171.54. Gottesman, M. M. and Paslan, I. (1988)J. Biol. Chem. 263:12163-12166.55. Monaco, J. J., Cho, S. and Attaya, M. (1990) Science 250:1723-1726.56. McGralh, J. P. and Varshavsky, A. (1989) Nature 340:400-404.57. Jedlitschky, G., Leier. I., Buccholz, U., Center, M. and Keppler, D. (1994) Cancer Res.

54:4833-4836.58. Leier, I., et al. (1996) Biochem. J. 314:433-437.59. Jedlitschky, G., Leier, I., Buchholz, U., Barnouin, K., Kurz, G. and Keppler, D. (1996) Cancer

Res. 56:988-994.60. Loe, D. W., Almquist, K. C, Deeley, R. G. and Cole, S. P. C. (1996) 7. Biol. Chem.

271:9675-9682.61. Loe, D. W., Almquist, K. C., Cole, S. P. C. and Deeley, R. G. (1996) 7. Biol. Chem.

271:9683-9689.62. Flens, M. J., et al. (1996) 4m. 7. Pathol. 148:1237-1247.

The GS-X Pump 207

63. Ortiz, D. F., Kreppel, L., Speiser, D. M., Scheel, G., McDonald, G. and Ow, D. W. (1992)EMBO J. 11:3491-3499.

64. Ortiz, D. F., Ruscitti, T., McCue, K. F. and Ow, D. W. (1995) J. Biol. Chem. 270:4721-4728.65. Li, Z.-S., Lu, Y.-P., Zhen, R.-G., Szczypka, M., Thiele, D. J. and Rea, P. A. (1996) Proc. Natl.

Acad. Sci. USA (in press).66. Ishikawa, T. and Ali-Osman, F. (1993)J. Biol. Chem. 268:20116-20125.67. Ishikawa, T., Wright, C. D. and Ishizuka, H. (1994) J. Biol. Chem. 269:29085-29093.68. Ishikawa, T., Bao, J.-J., Yamane, Y., Akimaru, K., Frindrich, K., Wright, C. D. and Kuo, M. T.

(1996) J. Biol. Chem. 271:14981-14988.69. Tommasini, R. et al. (1996) Proc. Natl. Acad. Sci. USA 93:6743-6748.70. Cole, S., Sparks, K. E., Fraser, K., Loe, D. W., Grant, C. E., Wilson, G. M. and Deeley, R. G.

(1994) Cancer Res. 54:5902-5910.71. Zaman, G. J. R., et al. (1995) Proc. Natl. Acad. Sci. USA 92:7690-7694.72. Chuman, Y., Chen, Z. S., Sumizawa, T., Furukawa, T., Haraguchi, M., Takebayashi, Y., Niwa,

K., Yamada, K., Aikou, T. and Akiyama, S. (1996) Biochem. Biophys. Res. Commun.226:158-165.

73. Alexander, I. and Aaseth, J. (1980) Biochem. Pharmacol. 29:2129-2133.74. Graf, P. and Sies, H. (1984) Biochem. J. 33:639-643.75. Ballatori, N. and Clarkson, T. W. (1985) Fundam. Appl. Toxicol. 5:398-403.76. Alexander, J., Aaseth, J. and Mikalsen, A. (1986) Ada Pharmacol. Toxicol. 59 (suppl VII):

486-489.77. Gyurasics, A., Varga, F. and Gregs, Z. (1991) Biochem. Pharmacol. 41:937-944.78. Gyurasics, A., Varga, F. and Gregs, Z. (1991) Biochem. Pharmacol. 42:465-468.79. Gyurasics, A., Koszorus, R., Varga, F. and Gregs, Z. (1992) Biochem. Pharmacl. 44:1275-1281.80. Dijkstra, M., Kuipers, F., Havinga, R., Smit, E. P. and Vonk, R. J. (1990) Peditric Res.

28:339-343.81. Kondo, T., Yoshida, K., Nakagawa, A., Kawai. T., Tamura, H. and Goto, T. (1992) Nature

359:515-518.82. Oude Elferink, R., Bakker, C. K. K., Roelofsen, H., Middelkoop, E., Ottenhoff, R., Heijin, M.

and Jansen, P. L. M. (1993) Hepatology 17:434-444.83. Li, Z.-S., Zhen, R.-G. and Rea, P. A. (1995) Plant Physiol. 109:177-185.84. Wu, A. and Moyc-Rowley, W. S. (1994) Mol. Cell Biol. 14:5832-5839.85. Wu, A., Wemmie, J. A., Szcypka, M. S., Thiele, D. J. and Moye-Rowley, W. S. (1994) J. Biol.

Chem. 269:32592-32597.86. Wemmie, J. A., Szczypka, M. S., Thiele. D. J., Moye-Rowley, W. S. (1994) J. Biol. Chem.

269;32592-32597.87. Gomi, A., Masuzawa, T., Ishikawa, T. and Kuo, M. T. (1997) Proc. Am. Assoc. Cancer. Res. (in

press.)88. Zhu, Q. and Center, M. (1994) Cancer Res. 54:4488-44902.89. Mulcahy, R. T. and Gipp, J. J. (1995) Biochem. Biophys. Res. Commun. 209:227-233.90. Yao, K.-S., Godwin, A. K., Johnson, S. W., Ozols, R. F., O'Dwyer, P. J. and Hamilton, T. C.

(1995) Cancer Res. 55:4367-4374.91. Kuo, M. T., Bao, J.-J., Curley, S. A., Ikeguchi, M., Johnston, D. A. and Ishikawa, T. (1996)

Cancer Res. 56:3642-3644.92. Paul, S., Breuninger, L. M., Tew, K. D., Shcn, H. and Kruh, G. D. (1996) Proc. Natl. Acad. Sci.

USA 93:6929-6934.93. Cole, S, P. C., Sparks, K., Fraser, K., Loe, D. W., Grant, C. E., Wilson, G. M. and Deeley. R. G.

(1994) Cancer Res. 54:5902-5910.94. Lutzky, J., et al. (1994) Cancer Res. 49:4120-4125.95. Priebe, W., Krawczyk, M., Kuo, M. T., Yamane. Y., Savaraj, N. and Ishikawa, T. (1997) Proc.

Am. Assoc. Cancer Res. (in press.)96. Gianni, L., Corden, B. J. and Myers, C. E. (1983) In Reviews in Biochemical Toxicology vol. 5,

Hodgson, E., Bent, J. R. and Philpot, R. M. (cds.). Elsevier, New York, pp. 1-82.97. Sokolove, P. M. (1988) FEBS Lett. 234:199-202.98. Benson, A. M., Hunkeler, M. J. and Talalay, P. (1980) Proc. Natl. Acad. Sci. USA 77:5216-5220.99. Akimaru, K., Nakanishi, M., Furuta, M., Suzuki, R., Noyori, R. and Ishikawa, T. (1997) In:

Eicosanoids and Other Bioactive Lipids in Cancer, Inflammation and Radiation Injury, Honn, K.V., Marnett, L. J. and Wong, P. Y.. Eds.), Plenum Press, New York (in press).

100. Ishikawa, T., Akimaru, K., Kuo, M. T., Nakanishi, M., Furuta. M., Suzuki. R., Noyori, R. (1997)Proc. Am. Assoc. Cancer Res. (in press.)