Major Photoaffinity Drug Binding Sites in Multidrug Resistance Protein 1 (MRP1) Are within Transmembrane Domains 10–11 and 16–17*

MRP1 is an ABC (or ATP binding cassette) membrane transport protein shown to confer resistance to structurally dissimilar drugs. Studies of MRP1 topology suggested the presence of a hydrophobic N-domain with five potential membrane-spanning domains linked to an MDR1-like core (MSD1-NBD1-L1-MSD2-NBD2) by an intracellular linker domain (L0). MRP1-mediated multidrug resistance is thought to be due to enhanced drug efflux. However, little is known about MRP1-drug interaction and its drug binding site(s). We previously developed several photoreactive probes to study MRP1-drug interactions. In this report, we have used eight MRP1-HA variants that were modified to have hemagglutinin A (HA) epitopes inserted at different sites in MRP1 sequence. Exhaustive in-gel digestion of all IAARh123 photoaffinity-labeled MRP1-HA variants revealed the same profile of photolabeled peptides as seen for wild type MRP1. Photolabeling of the different MRP1-HA variants followed by digestion with increasing concentrations of trypsin or Staphylococcus aureus V8 protease (1:800 to 1:5 w/w) and immunoprecipitation with anti-HA mAb identified two small photolabeled peptides (∼6–7 kDa) from MRP1-HA(574) and MRP1-HA(1222). Based on the location of the HA epitopes in the latter variants together with molecular masses of the two peptides, the photolabeled amino acid residues were localized to MRP1 sequences encoding transmembranes 10 and 11 of MSD1 (Ser542-Arg593) and transmembranes 16 and 17 of MSD2 (Cys1205-Glu1253). Interestingly, the same sequences in MRP1 were also photolabeled with a structurally different photoreactive drug, IACI, confirming the significance of transmembranes 10, 11, 16 and 17 in MRP1 drug binding. Taken together, the results in this study provide the first delineation of the drug binding site(s) of MRP1. Furthermore, our findings suggest the presence of common drug binding site(s) for structurally dissimilar drugs.

Increased expression of P-glycoprotein (P-gp1) 1 or the mul-tidrug resistance protein (MRP1) has been associated with the rise of drug resistance in numerous tumor cell lines in vitro (1,2). Gene transfer studies have also provided direct correlation between P-gp1 or MRP1 expression and resistance to dissimilar anti-cancer drugs; hence, the multidrug resistance (MDR) phenotype. In addition, gene disruption studies (3,4) in mice have demonstrate an increased drug accumulation in MRP1 or P-gp1-expressing tissues and organs. Interestingly and despite the low level of sequence identity between MRP1 and P-gp1 (Ͻ20% (5)), the two proteins confer resistance to similar natural product anti-cancer drugs (6). However in contrast to P-gp1, a distinguishing feature of MRP1 endogenous and exogenous substrates is their derivatization by glutathione, glucuronate, or sulfate (7)(8)(9)(10)(11). Notably, the cysteinyl leukotriene LTC 4 is the highest affinity substrate and an important modulator of inflammatory response (9). Other endogenous substrates include glucuronate-and sulfate-conjugated bile salts, glutathione-conjugated prostaglandin A 2 , as well as oxidized glutathione (GSSG) (8,12,13). P-gp1 and MRP1 are members of the ABC (or ATP binding cassette) family of membrane transport proteins (14) that couple the hydrolysis of ATP to drug transport. Members of this family contain one or more blocks of six membrane-spanning domains (MSD) linked to a nucleotide binding domain (NBD) (14). In contrast to P-gp1, MRP1 encodes for an additional hydrophobic N-domain with five potential transmembrane sequences (MSD0) and an extracytoplasmic N terminus (15)(16)(17)(18). Consequently, the linear organization of MRP1 consists of MSD0 linked to an MDR1-like core (MSD1-NBD1-L1-MSD2-NBD2) through an intracytoplasmic linker domain (L0). Functional analyses of truncated MRP1 demonstrate that deletion of the L0 domain abolished MRP1-mediated LTC 4 transport, whereas deletion of the MSD0 while maintaining L0 behaved as the full-length protein (19,20).
The mechanism of MRP1-mediated transport is not well understood and appears to differ from that of P-gp1. Drug transport studies using P-gp1-expressing plasma membranes clearly demonstrate an energy-dependent transport through direct binding to P-gp1 (21). Likewise, MRP1-mediated transport of LTC 4 is energy-dependent and likely to occur by direct binding to MRP1 (22). However, transport of unmodified natural product drugs by MRP1 has been shown to occur in association with glutathione (23)(24)(25)(26). Using several MRP1-specific photoreactive drugs, we have demonstrated direct binding between MRP1 and unmodified natural product drugs (27)(28)(29)(30)(31).
Moreover, glutathione was not required for MRP1-drug interaction. In addition, proteolysis of IACI-or IAARh123-photolabeled MRP1 produced three small peptides (27,28). In this report, we have mapped the drug binding site(s) of MRP1. To achieve this, we made use of eight MRP1 variants containing hemagglutinin A (HA) epitopes inserted at different positions in MRP1 (17,32). Our results show for the first time the presence of two major drug binding domains, confined to ϳ50 amino acids each and encoding TMs 10 -11 and TMs 16 -17.

EXPERIMENTAL PROCEDURES
Materials-Iodine-125 (100.7 mCi/ml) and the Protein A-coupled Sepharose were purchased from Amersham Pharmacia Biotech. The monoclonal anti-hemagglutinin A antibody 16B12 (anti-HA) was from the Berkeley Antibody Co. (Richmond, California). All chemicals were of the highest commercial grade available.
Cell Culture and Plasma Membrane Preparation-HeLa cells transfected with vector alone or MRP1-HA cDNA variants (17,32) were used in this study. HeLa-MRP1-HA variants describe cells expressing MRP1 containing one or more copies of hemagglutinin A epitope (YPYDVPD-YAS) inserted at amino acids 4, 163, 271, 574, 653, 938, 1001, or 1222 from the N terminus of MRP1 (see Fig. 1). Briefly, cells were grown in ␣-minimal essential medium (␣MEM) containing 10% fetal bovine serum (Hyclone). Cells were detached with trypsin-EDTA and then mixed immediately with cell culture media containing 20% horse serum followed by several washes in phosphate-buffered saline solution, pH 7.4. The cell pellets were resuspended in hypotonic Tris-Mg 2ϩ buffer (1 mM MgCl 2 and 10 mM Tris-HCl, pH 7.0) containing protease inhibitors (2 g/ml leupeptin, 2 g/ml aprotinin, 1 g/ml pepstatin A) and homogenized in a glass Dounce homogenizer. The resulting cell homogenate was spun for 10 min at 400 ϫ g, and the resulting supernatant was spun for an additional 60 min at 100,000 ϫ g. Plasma membrane-enriched pellets were resuspended in 5 mM Tris buffer, pH 7.4, and washed once to remove residual protease inhibitors. The final membrane pellets were resuspended in Tris-sucrose buffer (5 mM Tris, 250 mM sucrose, pH 7.4) using a 27-gauge needle and separated into aliquots for storage at Ϫ70°C. Proteins were quantified using the Lowry assay (33).
Photoaffinity Labeling and Protease Digestion-Plasma membranes from HeLa or HeLa-MRP1-HA variants were photoaffinity-labeled with IAARh123 or IACI photoreactive drugs as previously described (27,28). Photolabeled MRP1-HA variants were excised from SDS-PAGE and processed for in-gel digestion with Staphylococcus aureus V8 protease (20 g/gel slice) according to the method of Cleveland et al. (34). Alternatively, 100-g aliquots of photolabeled plasma membranes were digested with increasing concentrations of trypsin (sequencing grade and TPKC (L-1-tosylamido-2-phenylethyl chloromethyl ketone)-treated from Roche Molecular Biochemicals) at 37°C for 40 min. After digestion, samples were transferred to ice, and the digestion was stopped by the addition of 80 l of buffer A (1% SDS, 0.05 M Tris, pH 7.4) containing protease inhibitors (10 g/ml leupeptin, pepstatin A, aprotinin, and 1 mM phenylmethylsulfonyl fluoride). Samples were left on ice for 15 min and then diluted with 320 l of buffer B (1.25% Triton X-100, 190 mM NaCl, 0.05 M Tris, pH 7.4). IAARh123-or IACI-photolabeled MRP1-HA was immunoprecipitated with anti-HA mAb as previously described (35).

SDS-PAGE and Western
Blotting-Protein samples were resolved on SDS-PAGE using the Fairbanks gel system (36). Radiolabeled proteins were visualized on dried gels after exposure to Kodak X-AR film at Ϫ80°C. For immuno-detection of recombinant MRP1-HA variants, 20-g samples of enriched plasma membranes were resolved on SDS-PAGE and transferred to nitrocellulose membrane by using a wet electroblotting technique essentially as outlined by Towbin et al. (37). Nitrocellulose membranes were blocked in 5% skim milk in phosphatebuffered saline solution (pH 7.4) and incubated with anti-HA mAb overnight at 4°C. Membranes were washed and incubated with 1:3000 (v/v) of goat anti-mouse antibody conjugated to horseradish peroxidase. Immunoreactive proteins were visualized by chemiluminescence using Pierce SuperSignal substrate.

RESULTS
Tumor cells overexpressing MRP1 show reduced sensitivity to natural product toxins and certain anti-cancer drugs likely due to enhanced drug clearance (2). However, little is known about MRP1-drug interaction and binding site(s). Using photoreactive drug analogues, we have recently demonstrated the photoaffinity labeling of MRP1 at multiple sites (27,28). In this study, eight different MRP1-HA variants (see Fig. 1) were used to identify the sequences of MRP1 drug binding site(s). The expression and characterization of all eight MRP1-HA variants have been previously described (17,32), and their function was determined to be similar to wild type or unmodified MRP1.  variants, plasma membranes were labeled with IAARh123, and proteins were either resolved on SDS-PAGE ( Fig. 2B) or used for immunoprecipitation with anti-HA mAb (Fig. 2C). Fig. 2B shows an IAARh123-photolabeled ϳ190-kDa polypeptide in all 8 membrane samples. Moreover, Fig. 2C confirms the identity of the labeled protein as MRP1. The lower photoaffinity labeling signal associated with MRP1-HA (574) is consistent with the low level of expression seen in plasma membranes from cells expressing this variant ( Fig. 2A). To determine whether the insertion of HA epitopes at different positions in MRP1 has affected the binding domains relative to unmodified MRP1, immunopurified IAARh123-photolabeled MRP1-HA variants were excised from SDS-PAGE gels and digested with S. aureus V8 protease according to the method of Cleveland et al. (34). The results in Fig. 2D show the same digestion profile for the labeled MRP1-HA variants; hence, the insertion of HA epitopes in MRP1 at the specified positions ( Fig. 1) did not affect its drug interaction. Moreover, the lack of effect of HA epitope insertion on photolabeling characteristics of MRP1 is in agreement with the lack of effect of HA epitope insertion on the drug resistance profile in transfectant HeLa cells (17,32). In addition, similar results were also obtained using the quinoline-derived photoreactive drug, IACI (results not shown).
We have recently shown that MRP1 is photoaffinity-labeled at multiple sites with two structurally dissimilar drugs (27,28). Mild proteolytic cleavage of MRP1 with trypsin generated 111-and 85-kDa photoaffinity-labeled polypeptides that, respectively, produced two and one small photolabeled peptides upon further digestion (27,28). To achieve a higher resolution map of MRP1 photoaffinity-labeled sites, plasma membranes from each of the eight MRP1-HA variants were photoaffinitylabeled, digested with increasing concentrations of trypsin (1: 800 -1:5 w/w), and then immunoprecipitated with anti-HA mAb. Only those polypeptides having an HA epitope and a cross-linked IAARh123 were utilized in the analyses of MRP1 binding domains. The 50-kDa photolabeled protein is nonspecifically immunoprecipitated and appears in all the lanes from HeLa and HeLa-HA variants (Fig. 3). Although identity of this 50-kDa photolabeled protein is not known, comparison of photolabeled MRP1 and 50-kDa proteins by Cleveland digestion (34) did not reveal similar photolabeled peptides or peptide maps (results not shown). Interestingly, however, the 50-kDa protein was much less apparent when the IAARh123-photolabeled membranes immunoprecipitated with hybridoma supernatant containing anti-HA mAb was isolated from cells grown in absence of 10% fetal bovine serum. Thus, the presence of serum proteins is likely to be responsible for the nonspecific immunoprecipitation of the 50-kDa protein (results not shown).
To ascertain the positions of the IAARh123-photolabeled peptides of MRP1-HA (574) and -HA (1222) more closely, photolabeled membrane samples were digested with higher concentrations of trypsin or S. aureus V8 protease (up to 1:5 w/w), and the total digest was subjected to immunoprecipitation with anti-HA mAb. Fig. 6A shows a photolabeled peptide derived from the exhaustive digestion of MRP1-HA (574) with trypsin, migrating with an apparent molecular mass of a ϳ6.5 kDa. Careful examination of MRP1 amino acid sequence for the smallest possible tryptic peptide that include the HA epitope at position 574 revealed the following peptide sequence 542 SAYL-SAVGTFTWVCTPFLVALCTFAVYVTIDEN[HA]NILDAQTAFV SLALFNILR 593 , where the underlined sequences represent the predicted TM10 and TM11. The calculated molecular mass of the photoaffinity-labeled peptide, including the amino acid sequence of one HA epitopes, is 6.8 kDa. Exhaustive digestion of IAARh123-photolabeled MRP1-HA (574) variants with V8 protease followed by immunoprecipitation with anti-HA mAb did not show any other photolabeled small peptides (data not shown). Similarly, exhaustive digestion of IAARh123-photolabeled membranes of MRP1-HA (1222) with V8 protease revealed the presence of a ϳ7-kDa photolabeled peptide (Fig. 6B). Analysis of MRP1 sequence for all possible V8 peptides that include the HA epitope at position 1222, revealed the following amino acid sequence 1205 CVGNCIVLFAALFAVISR[HA-HA]HSLSAGLVGLS-VSYSLQVTTYLNWLVRMSSE 1253 , where the underlined sequences represent the predicted TM16 and TM17. The calculated molecular mass of this peptide including the amino acid sequence of the two HA epitopes is ϳ7.6 kDa. As indicated earlier, we have recently demonstrated that both IAARh123 and IACI, two structurally dissimilar photoreactive drugs, labeled similar tryptic peptides that migrate with apparent molecular masses of 4 -6 kDa on SDS-PAGE (27,28). Given the localization of IAARh23 drug binding sites in MRP1, it was of interest to determine the IACI-photolabeled sequences using the MRP1-HA variants. Fig. 7 shows immunopurified peptides from IACI-labeled MRP1-HA (574) and -HA (1222) after digestion with trypsin and V8 protease, respectively. In-terestingly, Fig. 7A shows a ϳ6.5-kDa IACI-photolabeled peptide immunoprecipitated from MRP1-HA (574) tryptic digest. Likewise, MRP1-HA (1222) showed an ϳ7-kDa IACI-photolabeled peptide (Fig. 7B). Digestion of IACI-photolabeled plasma membranes prepared from the other MRP1-HA variants led to identical photolabeled peptides as IAARh123 (data not shown). Taken together these results demonstrate the binding of two structurally dissimilar drugs to the same site(s) of MRP1. DISCUSSION Similar to P-gp1, MRP1 is believed to cause resistance to anti-cancer drugs through an enhanced drug clearance mechanism (39). Indeed, there is considerable overlap in substrate specificity to anti-cancer drugs between MRP1 and P-gp1 (39). However, MRP1 also mediates the transport of other normal cell metabolites such as glutathione-, sulfate-, and glucuronatemodified ligands. Furthermore, given the structural differences between MRP1 and P-gp1 and the low sequence identity (Ͻ20% (5)), it has been difficult to speculate about the mechanism of MRP1 based on the current understanding of P-gp1-mediated drug transport. For example, unlike P-gp1, it has not been possible to demonstrate direct binding between MRP1 and unmodified natural product drugs, likely due to low binding affinities. Moreover, it remains unclear how reduced glutathione mediates MRP1 drug transport. Using a quinoline-based photoreactive drug, we have previously shown direct interaction between MRP1 and natural product drugs (29 -31). Consistent with our findings, Leier et al. Analyses of the first three MRP1-HA(4, 163, and 271) variants using IACI and IAARh123 revealed the presence of only one photolabeled polypeptide with an apparent molecular mass of 111 kDa. Further digestion of the 111-kDa fragment, which contained three HA epitopes at positions 4, 163, or 271 from the N terminus did not produce smaller photolabeled peptides. The smallest immunoprecipitable IAARh123-photolabeled peptides attainable after trypsin digestion of MRP1-HA(574) or V8 protease digestion of MRP1-HA(1222) were 6.5 and 7 kDa, respectively (A and B). The schematic to the right of panels A and B represents the equivalent domains of MRP1. Each circle represents one amino acid residue, with the white circles for those outside the membrane and gray circles representing the amino acids that form TM domains. Black circles are the trypsin-sensitive arginines and lysines (A) or V8-sensitive aspartic and glutamic acids (B). Therefore, the MRP1 sequence within the first 274 amino acid residues (relative to the position of the first lysine located right after the HA epitope insertion at position 271) are not photolabeled with IAARh123 or IACI. Alternatively, it is possible that the first 274 amino acid residues of MRP1 are photoaffinity-labeled; however, after the addition of higher trypsin concentrations, the photolabeled site was separated from the HA epitope. Although this latter possibility is conceivable with one HA epitope, it is unlikely to occur with all HA epitopes inserted at three different positions in MRP1 (at amino acids 4, 163, and 271 from the N terminus). Moreover, we and others previously demonstrated the presence of two trypsin hypersensitive sites, one in each of the linker domains (L1 and L0) of MRP1, with cleavage at L1 proceeding that at L0 (27,28,38). Consequently, cleavage at L1 would generate the 111-and 85-kDa polypeptides followed by a cleavage of the 111-kDa at L0 producing two polypeptides with apparent molecular masses of 40 -60 and 60 kDa (16). The 40 -60-kDa fragment would include MSD0, whereas the 60-kDa fragment would include MSD1 (16). The absence of photolabeled 40 -60-kDa polypeptides with either one of the three MRP1-HA (4, 163, 271) variants is consistent with the lack of drug interactions within the first 274 Nterminal amino acids. These results are consistent with earlier findings whereby deletion of the first N-terminal 203 amino acids behaved like wild-type MRP1 in vesicle uptake of LTC 4 (19). Having previously demonstrated that IACI and IAARh123 interacts with MRP1 at the same or an overlapping site(s) as that of LTC 4 , we speculate that MSD0 does not interact with either drug.
Earlier in-gel digestion studies of IACI-or IAARh123-photolabeled 111-kDa fragments, which encodes MSD0-L0-MSD1-NBD1 and parts of L1, showed two photolabeled peptides with apparent molecular masses of ϳ6 and ϳ4 kDa (27,28). Based on the observed results with MRP1-HA (4,163,271,574,653), which yielded only one photolabeled peptide containing the HA epitope at position 574, the previously observed photolabeled peptides (4 and 6 kDa) may represent incomplete digestion of the photolabeled fragment (Leu 536 -Glu 623 ) at internal V8 protease cuts at Asp 572 or Asp 578 . Thus, V8 protease cuts at Asp 572 or Asp 578 would produce a fragment with an estimated molecular mass of ϳ4 kDa (Leu 536 to Asp 572 /Asp 578 or Asp 572 /Asp 578 to Glu 623 ) in addition to an ϳ6-kDa fragment representing Leu 536 -Glu 623 .
Tryptic digestion of MRP1-HA (1222), by contrast to MRP1-HA(938 and 1001), produced two photolabeled peptides (41 and 7 kDa) in addition to the 85-kDa polypeptide (Fig. 5). The 41-kDa photolabeled peptide did not contain the first 1013 N-terminal amino acid residues of MRP1 (relative to the position of the first lysine located right after the HA epitope insertion at position 1001) and is likely to encode part of the MSD2 and NBD2. The second smallest photolabeled peptide produced from the total cleavage of the MRP1-HA (1001) is likely to encode the last two transmembrane domains of MSD2. Analysis of MRP1 amino acid sequence for potential trypsin cuts near the 1222 HA epitope showed three potential sites, Arg 1202 , Arg 1222 , and Arg 1249 . Cleavage at Arg 1202 and Arg 1249 would generate a 67-amino acid peptide, including the two 10-residue repeats for the HA epitope, with a calculated molecular mass of 7.6 kDa versus an apparent molecular mass of 7 kDa. Interestingly however, exhaustive trypsin digestion of MRP1-HA (1222) variant did not produce a smaller peptide that could represent cleavage at Arg 1222 . One possibility is that the Arg 1222 site was not accessible to trypsin. Alternatively, cleavage at Arg 1222 did occur, but the peptide generated from this cleavage was not photolabeled. Consequently, the photolabel is cross-linked to sequences between Arg 1202 and Arg 1222 or within TM16. The localization of the 7-kDa photolabeled peptide in MRP1-HA (1222) was further confirmed after exhaustive proteolytic digestion with V8 protease, which produced a 7-kDa peptide. Scanning MRP1 sequences for potential V8 sites revealed two glutamate residues (Glu 1204 and Glu 1253 ) near HA(1222). Hence the resulting polypeptide of 69 amino acid residues has a calculated molecular mass of 7.5 kDa versus an apparent mass of 7 kDa on SDS-PAGE. The localization of the 7-kDa IAARh123-photolabeled peptide to MRP1 sequences, which include TMs 16 and 17, is consistent with proteolytic cleavage of MRP1 with two different proteases. The latter findings are in agreement with an earlier study suggesting a possible role of the C-terminal domain (MSD2 and NBD2 or residues 959 -1531) in determining MRP1 resistance to and transport of anthracyclines (40).
The photolabeling of the same peptides of MRP1 by two structurally dissimilar photoreactive drugs (IAARh123 and IACI) suggests common binding domains. However, given the size of the photolabeled peptides with a combined sequence of 100 amino acid residues, it is conceivable that the photolabeled peptides can encode several binding sites for different groups of compounds. Alternatively, the photolabeled peptides in MRP1 represent domains that are distant from the drug transport site(s). Although the latter possibility is plausible, earlier mapping studies of P-gp1 drug binding site(s) using photoreactive drugs have shown that photoaffinity-labeled sequences represent functionally important drug binding sites, and mutations of amino acids in such domains modulate P-gp1 drug transport (41)(42)(43). However, several questions relating to P-gp1 broad substrate specificity and the ability of P-gp1 to accommodate the binding of drugs remain unanswered. The photoaffinitylabeled domains of MRP1, which encode TMs 10 -11 and TMs 16 -17, are separated by ϳ400 amino acid residues in its linear sequence (5). As such, photolabeling of MSD1 and MSD2 sequences in MRP1 suggest one of two possibilities; (a) the two photolabeled sequences from MSD1 and MSD2 are brought together by protein folding to form one large binding cavity that can accommodate interactions with many structurally dissimilar drugs or (b) the two photolabeled peptides represent two separate drug binding sites in MSD1 and MSD2 that are allosterically regulated. Although it is not possible to deter mine which model best fits the observed results in the current study, earlier reports with P-gp1 suggested two or more spatially separate sites linked allosterically through protein conformation (44,45).
Secondary structure predictions of MRP1 have suggested the presence of a hydrophobic domain with five transmembrane domains (MSD0) and an extracytoplasmic N terminus linked to an MDR1-like core consisting of two MSDs and two NBDs in tandem (MSD1-NBD1-MSD2-NBD2). In contrast to MRP1, secondary structure predictions of P-gp1 have suggested two tandemly repeated MSD and NBD or an MDR1-like core. Alignment of the MRP1 and P-gp1 predicted secondary structures revealed TM 5-6 and TM 11-12 of P-gp1 to correspond to TM 10 -11 and TM 16 -17 of MRP1, respectively. Thus, the drug binding sites of both P-gp1 and MRP1 are localized to parallel domains in both proteins. This finding is unexpected in light of the low sequence identify between the two proteins (Ͻ20% (5)). Comparison of the amino acid sequences of the four major transmembrane domains in MRP1 and P-gp1 showed 14,9,4, and 14% sequence identity between TM 5, 6, 11, 12 of P-gp1 and TM 10, 11, 16, 17 of MRP1, respectively. In addition, helical wheel presentation of the transmembrane domains did not reveal any obvious similarities between MRP1 and P-gp1 transmembrane sequences.
The findings that the photolabeled sequences of MRP1 encode transmembrane domains is interesting, since similar results have been obtained with P-gp1 whereby TMs 5-6 and 11-12 were also shown to mediate P-gp1-drug interactions (42)(43)(44)(45). However, unlike P-gp1, which mediates the transport of hydrophobic drugs mainly from within the lipid bilayer, MRP1 appears to interact with and mediate the transport of drugs from the cytoplasm and the hydrophobic environment of the lipid bilayer (46 -48). Consequently, the P-gp1-drug interaction model may be an oversimplified version of MRP1-drug interactions.
In summary, the results of this study have identified two domains in MRP1 to a resolution of ϳ50 amino acid residues that are photoaffinity-labeled with two structurally dissimilar drugs. Moreover, based on positions of the photolabeled peptides, the results of this study show that the drug binding domains are localized to TMs 10 -11 of MSD1 and TMs 16 -17 of MSD2. Efforts to identify the precise amino acid residues that are cross-linked to IAARh123 and IACI are ongoing. In addition, it will be of interest to know if the photolabeled sites identified in this study are the same or different from the binding site(s) of normal MRP1 substrates (e.g. LTC 4 and other cell metabolites).