Mutation of the aromatic amino acid interacting with adenine moiety of ATP to a polar residue alters the properties of multidrug resistance protein 1.

Structural analyses of several bacterial ATP-binding cassette (ABC) transporters indicate that an aromatic amino acid residue in a nucleotide-binding domain (NBD) interacts with the adenine ring of the bound ATP and contributes to the ATP binding. Substitution of this aromatic residue with a polar serine residue in bacterial histidine transporter completely abolished both ATP binding and ATP-dependent histidine transport. However, substitution of the aromatic amino acid residue in the human cystic fibrosis transmembrane conductance regulator with a polar cysteine residue did not have any effect on the ATP-dependent chloride channel function of the protein. To determine whether the other eucaryotic ABC transporters use the strategy analogous to that in some bacterial ABC transporters, the aromatic Trp653 residue in NBD1 and the Tyr1302 residue in NBD2 of human multidrug resistance-associated protein 1 (MRP1) was mutated to either a different aromatic residue or a polar cysteine residue. Substitution of the aromatic residue with a different aromatic amino acid, such as W653Y or Y1302W, did not affect ATP-dependent leukotriene C4 (LTC4) transport. In contrast, substitution of the aromatic residue with a polar cysteine residue, such as W653C or Y1302C, decreased the affinity for ATP, resulting in greatly increased Kd values for ATP binding or Km values for ATP in ATP-dependent LTC4 transport. Interestingly, although substitution of the aromatic Trp653 in NBD1 of MRP1 with a polar cysteine residue greatly decreases the affinity for ATP, the ATP-dependent LTC4 transport activities are much higher than that of wild-type MRP1, supporting our hypothesis that the increased release rate of the bound ATP from the mutated NBD1 facilitates the protein to start a new cycle of ATP-dependent solute transport.


Structural analyses of several bacterial ATP-binding cassette (ABC) transporters indicate that an aromatic amino acid residue in a nucleotide-binding domain (NBD) interacts with the adenine ring of the bound ATP
and contributes to the ATP binding. Substitution of this aromatic residue with a polar serine residue in bacterial histidine transporter completely abolished both ATP binding and ATP-dependent histidine transport. However, substitution of the aromatic amino acid residue in the human cystic fibrosis transmembrane conductance regulator with a polar cysteine residue did not have any effect on the ATP-dependent chloride channel function of the protein. To determine whether the other eucaryotic ABC transporters use the strategy analogous to that in some bacterial ABC transporters, the aromatic Trp 653 residue in NBD1 and the Tyr 1302 residue in NBD2 of human multidrug resistance-associated protein 1 (MRP1) was mutated to either a different aromatic residue or a polar cysteine residue. Substitution of the aromatic residue with a different aromatic amino acid, such as W653Y or Y1302W, did not affect ATP-dependent leukotriene C4 (LTC4) transport. In contrast, substitution of the aromatic residue with a polar cysteine residue, such as W653C or Y1302C, decreased the affinity for ATP, resulting in greatly increased K d values for ATP binding or K m values for ATP in ATP-dependent LTC4 transport. Interestingly, although substitution of the aromatic Trp 653 in NBD1 of MRP1 with a polar cysteine residue greatly decreases the affinity for ATP, the ATPdependent LTC4 transport activities are much higher than that of wild-type MRP1, supporting our hypothesis that the increased release rate of the bound ATP from the mutated NBD1 facilitates the protein to start a new cycle of ATP-dependent solute transport.
Cells overexpressing P-glycoprotein, multidrug resistanceassociated protein 1 (MRP1), 1 breast cancer resistant protein, and/or other unidentified drug transporters become resistant to a broad range of anticancer drugs (1)(2)(3)(4)(5)(6). These proteins couple ATP binding/hydrolysis to anti-cancer drug transport. Because of the decreased concentration of anti-cancer drugs inside of the cells overexpressing these proteins, the cells become multidrug-resistant. The most conserved features of these proteins are the two cytoplasmic nucleotide-binding domains (NBDs) containing Walker A and B motifs (7); therefore, they belong to the ATP-binding cassette (ABC) transporter superfamily. Recent crystal structural analyses of bacterial ABC transporters, such as HisP (8), MJ0796 (9), MalK (10), and GlcV (11), indicate that: 1) the residues from Walker A motif of one NBD interact with the ␣-, ␤-, and ␥-phosphate of the bound ATP, providing the majority of the interactions that stabilize the nucleotide in the NBD; 2) the Mg 2ϩ in the GlcV-ADP-Mg 2ϩ complex (11) is bound through an octahedral coordination involving one oxygen atom from ␤-phosphate, one oxygen atom from ␥-phosphate, one oxygen atom of Thr 45 (a residue from Walker A motif), one oxygen atom of Gln 89 (a residue from Q-loop), two water molecules coordinated with phosphate and Asp 165 (a residue from Walker B motif), and Glu 166 (the putative catalytic base); 3) residues from the ABC signature motif of another NBD interact with ␣-, ␤-, and ␥-phosphate, ribose, and adenine base of the bound ATP, further stabilizing the bound nucleotide in the NBD; and 4) additional stabilizing force comes from the interactions between an aromatic amino acid residue and the adenine base of the bound ATP.
Because eucaryotic ABC transporters, such as P-glycoprotein, MRP1, and cystic fibrosis transmembrane conductance regulator (CFTR), also contain Walker A and B motifs, ABC signature motif, and an aromatic amino acid residue; therefore, the way they bind ATP should be analogous to those procaryotic ABC transporters. Mutations of the lysine residue in Walker A motif or aspartic acid residue in Walker B motif in P-glycoprotein (12)(13)(14)(15)(16) or in MRP1 (17)(18)(19)(20) greatly reduced the protein-mediated ATP-dependent drug transport and drug resistance, supporting the hypothesis that the way to bind ATP by these eucaryotic ABC transporters should be analogous to those procaryotic ABC transporters. Additional evidence comes from the naturally occurring CFTR mutant G551D, a mutation of the Gly 551 residue in the ABC signature motif of NBD1 causing cystic fibrosis (21)(22)(23). Presumably the mutation of this critical Gly residue in the LSGGQ ABC signature motif to an acidic amino acid affects ATP binding and the chloride channel function of the protein (24). The corresponding mutations in MRP1, G771D in NBD1 and G1433D in NBD2, almost completely abolished ATP-dependent LTC4 transport (20). However, substitution of the aromatic amino acid with a polar cysteine residue in human CFTR did not disrupt the chloride channel function (25), which was in contrast with the Y16S mutation in bacterial ABC transporter HisP that abolished both ATP binding and histidine transport (26). These authors suggested that ATP binding in CFTR may not involve the coordination of the adenine base with an aromatic amino acid residue analogous to that in some bacterial ABC transporters (25). Whether this is due to the edge-to-face interactions between the adenine ring and Phe 430 as shown in the crystal structure of mouse CFTR NBD1 (27) is not yet known. However, the close interactions between Trp 653 in NBD1 of human MRP1 and the adenine ring of the bound ATP were identified by the two-dimensional NMR spectra (28). To test whether this aromatic residue is involved in ATP binding to the NBD1 of human MRP1, Trp 653 was mutated to either Tyr or Cys. In the meantime, the corresponding aromatic residue in NBD2, Tyr 1302 , was mutated to either Trp or Cys. The substitution of Trp 653 or Tyr 1302 with a different aromatic amino acid, such as W653Y or Y1302W, has little effect on the K m values for ATP in ATP-dependent LTC4 transport. In contrast, the substitution of Trp 653 or Tyr 1302 with a polar cysteine residue greatly increased the K d values for ATP binding and the K m values for ATP in ATP-dependent LTC4 transport, implying that these aromatic residues in NBD1 and NBD2 of MRP1 do play an important role in nucleotide binding. Cell Culture and Expression of MRP1s-Spodoptera frugiperda 21 (Sf21) cells were cultured in Grace's insect cell medium supplemented with heat-inactivated 5% fetal bovine serum at 27°C. Viral infection was performed according to Invitrogen's recommendation.

Materials-[␣-
Generation of Constructs-The pDual construct expressing the Nhalf (1-932) and C-half (933-1531) simultaneously 2 was made from the pNUT/MRP1/His (30) and used as a template for the mutagenesis. The tryptophan residue at position 653 was mutated to either tyrosine or cysteine (see Fig. 1A, W653Y or W653C) by using the forward/reverse primers and the QuikChange site-directed mutagenesis kit from Stratagene (19). The forward and reverse primers for W653Y and W653C are: W653Y/forward, 5Ј-AGG AAT GCC ACA TTC ACC TAT GCC AGG AGC GAC CCT CCC-3Ј; W653Y/reverse, 5Ј-GGG AGG GTC GCT CCT GGC ATA GGT GAA TGT GGC ATT CCT-3Ј; W653C/forward, 5Ј-AGG AAT GCC ACA TTC ACC TGT GCC AGG AGC GAC CCT CCC-3Ј; and W653C/reverse: 5Ј-GGG AGG GTC GCT CCT GGC ACA GGT GAA TGT GGC ATT CCT-3Ј. The underlined sequences are codons for mutated residues. Y1302W and Y1302C mutations were generated by using the same strategy as for W653Y. The forward and reverse primers for Y1302W and Y1302C are: Y1302W/forward, 5Ј-CGG AAC TAC TGC CTG CGC TGG CGA GAG GAC CTG GAC TTC-3Ј; Y1302W/reverse, 5Ј-GAA GTC CAG GTC CTC TCG CCA GCG CAG GCA GTA GTT CCG-3Ј; Y1302C/forward, 5Ј-CGG AAC TAC TGC CTG CGC TGC CGA GAG GAC CTG GAC TTC-3Ј; and Y1302C/reverse, 5Ј-GAA GTC CAG GTC CTC TCG GCA GCG CAG GCA GTA GTT CCG-3Ј. The N-half and C-half containing these mutations were sequenced completely and used to make other combinations (see Fig. 1A). To make these com-  (19,31). The samples are: The molecular weight markers are indicated on the left. N-Half and C-Half on the right indicate the N-proximal half (1-932) and C-proximal half (933-1531) of MRP1 proteins. Relative expression levels of MRP1 proteins were analyzed by densitometry. The average ratios of N-half and C-half expression levels are indicated below each blot. Because some of the bands in 0.2 g of membrane vesicles were barely detected, only 0.4-and 0.8-g blots were analyzed. The ratios of the band intensities with the same amount of total membrane proteins, for example, 0.4 g of wild-type N-half (co-expressed with wild-type C-half) versus 0.4 g of the W653Y-mutated N-half (co-expressed with wild-type C-half) were determined. Because the ratio of N-half, for example, W653Y-mutated N-half, is similar to that of the wild-type C-half co-expressed with W653Y-mutated N-half, the mean ratio of the protein expression includes N-half and C-half.
binations, a KpnI-RsrII fragment containing C-half, for example, with a Y1302W mutation, was cloned into a KpnI-RsrII fragment containing pDual vector DNA and N-half, for example, with a W653Y mutation, to generate W653Y/Y1302W. The regions containing these mutations were sequenced again to confirm that the correct clones were obtained.
Recombinant Viral DNA Preparation and Viral Particle Production-Generation of recombinant viral DNA was performed according to Invitrogen's recommendation. PDual/MRP1 donor plasmid DNA was transformed into DH10Bac-competent cells harboring the parent Bacmid DNA with a mini-attTn7 target site and the helper plasmid. Bacteria transformed with the donor plasmid DNA were selected on LB plates containing 50 g/ml kanamycin, 7 g/ml gentamicin, and 10 g/ml tetracyclin. In addition, colonies containing recombinant Bacmids were identified by disruption of the lacZ␣ gene (white colonies) on the LB plates containing 100 g/ml X-gal and 40 g/ml isopropyl thio-␤-D-galactoside. A single colony confirmed as having a white phenotype on the LB plate with X-gal and isopropyl thio-␤-D-galactoside was inoculated in medium containing 50 g/ml kanamycin, 7 g/ml gentamicin, and 10 g/ml tetracyclin. The purified recombinant Bacmids were confirmed by polymerase chain reaction with MRP1-specific primers and then used to transfect Sf21 cells with CellFECTIN reagent (Invitrogen). After 3-4 days of incubation at 27°C, the supernatants containing viral particles were collected, and the cell lysates (with 2% SDS) were used to do a Western blot, probed with MRP1-specific monoclonal antibodies 42.4 and 897.2 (19).
Viral Plaque Assay, Viral Infection, and Membrane Vesicle Preparation-Viral plaque assay was performed according to Invitrogen's recommendation. The expression levels of the dually expressed Nhalves and C-halves with varying multiplicity of infection were determined by Western blot. The multiplicities of infection producing similar amounts of N-halves (comparing the wild-type N-half with mutated N-halves) and C-halves (comparing wild-type C-half with mutated C-halves) were used to infect Sf21 cells for membrane vesicle preparations.

SDS-PAGE and Western Blot-SDS-PAGE
and Western blot were performed as described previously (19,31). The primary antibodies used were mouse anti-human MRP1 monoclonal antibodies 42.4 and 897.2 (19,31), and the secondary antibody was anti-mouse Ig conjugated with horseradish peroxidase. Chemiluminescent film detection was performed according to the manufacturer's recommendations (Pierce).
Photoaffinity Labeling of MRP1 Protein-Photoaffinity labeling experiments were carried out in a 10 l of solution containing 5 g of membrane vesicles, 40 mM Tris-HCl, pH 7.5, 2 mM ouabain, 0.1 mM EGTA, 10 mM MgCl 2 , and varying concentrations of [␣-32 P]8-N 3 ATP for 5 min on ice. The samples were UV-irradiated on ice for 2 min (19, 31), separated on a polyacrylamide gel (7%), and electroblotted to a nitrocellulose membrane.
Membrane Vesicle Transport-ATP-dependent transport of 3 H-labeled LTC4 into the membrane vesicles was assayed by a rapid filtration technique (32,33). The assays were carried out in a 30-l solution containing 3 g of membrane vesicles, 50 mM Tris-HCl, pH 7.5, 250 mM sucrose, 10 mM MgCl 2 , 200 (see Fig. 2) or 400 nM (see Fig. 3) LTC4, and varying concentrations of ATP as indicated in the figure legends. AMP was used as a negative control. After incubation at 37°C for the time indicated in the figure legends, the samples were brought back to ice, diluted with 1 ml of ice-cold 1ϫ transport buffer (50 mM Tris-HCl, pH 7.5, 250 mM sucrose, and 10 mM MgCl 2 ) and filtered through a nitrocellulose membrane (19). This nitrocellulose membrane was then washed with 10 ml of ice-cold 1ϫ transport buffer, air-dried, and placed in a 10 ml of biodegradable counting scintillant (Amersham Biosciences). The radioactivity bound to the nitrocellulose membrane was determined by liquid scintillation counting (Beckman LS 6000SC).  (11), indicate that there is an aromatic amino acid residue upstream of the Walker A motif interacting with the adenine ring of the bound ATP and contributing to ATP binding. The aromatic residue Trp 653 in NBD1 of human MRP1 protein interacts with the adenine moiety of the bound ATP (28). The corresponding aromatic amino acid residue in NBD2 of human MRP1 is Tyr 1302 , based on the alignment of the amino acid sequences in that region. To test whether these aromatic amino acid residues are involved in ATP binding, these aromatic residues were mutated to either another aromatic residue, such as Trp or Tyr, or a polar residue, such as Cys, as shown in Fig. 1A. The recombinant viruses expressing wild-type or mutated N-halves (1-932) and C-halves (933-1531) were prepared and expressed in Sf21 cells. 2 Fig. 1B shows the Western blot results of the proteins expressed in Sf21 cells, probed with antibody 42.4 against NBD1 and 897.2 against NBD2 of MRP1 (19). 42.4 antibody detected a 93-kDa N-half band, whereas 897.2 antibody detected 57-and 52-kDa C-half bands (Fig. 1B). We have found that the 57-kDa band is the glycosylated C-half, whereas the 52-kDa band is the unglycosylated C-half. 2 Fig. 1B also shows the relative expression levels of the MRP1 proteins in the membrane vesicles. Overall the amounts of N-and C-halves in these membrane vesicles are similar (Fig. 1B), indicating that mutations of the aromatic residues in NBD1 and/or NBD2 did not significantly affect the protein processing.

Wild type and
The Substitutions of the Aromatic Residues with a Polar Amino Acid in NBD1 and NBD2 Decreased LTC4 Transport Activities at Low ATP Concentration-If these aromatic residues are involved in ATP binding by providing van der Waals' interactions with the adenine ring of the bound nucleotide, then mutation of these aromatic residues to a moderately polar residue should reduce the binding force and decrease the affinity for ATP. To test this possibility, membrane vesicles containing wild-type and mutated MRP1s (Fig. 1B) were utilized to do ATP-dependent LTC4 transport in the presence of 50 M ATP, which is within the range of the K m values for ATP (K m (ATP)) of wild-type N-half ϩ wild-type C-half mediated LTC4 transport (34). 2 Fig. 2 shows that the transport activity of W653Ymutated N-half ϩ wild-type C-half is ϳ2-fold of wild type, whereas the transport activity of wild-type N-half ϩ Y1302W-  C-half or W653Y-mutated N-half ϩ Y1302W-mutated C-half is similar to that of wild type, indicating that substitution of an aromatic residue with a different aromatic amino acid, such as Trp to Tyr or Tyr to Trp, does not have a significant negative effect on the ATP binding and hydrolysis. In contrast, the substitution of the aromatic residue, no matter whether it is in NBD1 or NBD2, with a nucleophilic cysteine residue, such as W653C-mutated N-half ϩ wild-type C-half, W653C-mutated N-half ϩ Y1302W-mutated C-half, wild-type N-half ϩ Y1302C-mutated C-half, W653Y-mutated N-half ϩ Y1302C-mutated C-half, or W653C-mutated N-half ϩ Y1302Cmutated C-half, greatly decreased the ATP-dependent LTC4 transport activities (Fig. 2), implying that both aromatic residues, Trp 653 in NBD1 and Tyr 1302 in NBD2, are involved in ATP binding.
The Substitutions of the Aromatic Residues with a Polar Amino Acid in NBD1 and/or NBD2 Increased K m Values for ATP in ATP-dependent LTC4 Transport-If the two aromatic residues, Trp 653 and Tyr 1302 , are involved in ATP binding, then mutation of one or both of these residues to a polar amino acid should decrease the hydrophobic interaction force between the aromatic residue and the adenine ring of the bound ATP, i.e. the affinity for ATP, and affect K m (ATP) values in ATP-dependent LTC4 transport. To test this possibility, membrane vesicles containing wild-type and mutated MRP1s (Fig. 1B) were utilized to do ATP-dependent LTC4 transport in the presence of varying concentrations of ATP. Panels A and B of Fig. 3 show that the transport activities of those N-halves and Chalves containing an aromatic residue reach plateau within low concentrations of ATP, implying low K m (ATP) values. The transport activity of W653Y-mutated N-half ϩ wild-type C-half is much higher than that of wild type, whereas the activities of wild-type N-half ϩ Y1302W-mutated C-half and W653Y-mutated N-half ϩ Y1302W-mutated C-half are similar to that of the wild type (Fig. 3A), which are consistent with the results in Fig. 2. Table I shows their K m (ATP) and V max for LTC4 (V max (LTC4)) values in ATP-dependent LTC4 transport. Substitution of the aromatic Trp 653 with a different aromatic amino acid (Tyr), i.e. W653Y, slightly decreased the K m (ATP) value but increased V max (LTC4) value 2-fold ( Table I). Substitution of the aromatic residue Tyr 1302 in NBD2 with a different aromatic Trp residue, Y1302W, also slightly decreased K m (ATP) value; however, the V max (LTC4) value was not changed (Table I). Interestingly, the switch of those two aromatic residues in NBD1 and NBD2, W653Y-mutated N-half ϩ Y1302W-mutated C-half, did not alter the K m (ATP) value but slightly increased the V max (LTC4) value (Table I). Panels C and D of Fig. 3 show that the transport activities of those N-half mutants containing a substitution with a polar cysteine residue reach plateau within high concentrations of ATP, implying high K m (ATP) values. Table I shows that the K m (ATP) values for W653C, W653C/Y1302W, and W653C/Y1302C are 4.5-, 4.2-, and 22.8fold higher than that of wild-type MRP1. Interestingly, the transport activities of these N-half mutants are also higher than that of wild-type MRP1 (Table I). The transport activities of those C-half mutants containing a substitution with a polar cysteine residue also reach plateau within high concentrations of ATP (Fig. 3, E and F). Their K m (ATP) values are even higher than those N-half mutants containing only one cysteine replacement, such as W653C and W653C/Y1302W (Table I). However, their V max (LTC4) values are less than those of the N-half mutants containing only one cysteine replacement, such as W653C and W653C/Y1302W (Table I).
Substitution of an Aromatic Residue with a Polar Amino Acid in NBD1 or NBD2 Increases its K d Value for ATP Binding-The greatly increased K m (ATP) values of those mutants containing a polar residue imply that the substitution of the aromatic residue with a nucleophilic cysteine residue decreases the affinity for ATP. However, because the K m (ATP) value reflects the total events of ATP binding/hydrolysis at both NBD1 and NBD2 and the ATP-dependent LTC4 transport, it might be difficult to make that conclusion based solely on the K m (ATP) values. To test whether these substitutions really alter their affinities for ATP, membrane vesicles containing wild-type N-half (Trp 653 ) ϩ wild-type C-half (Tyr 1302 ), W653Cmutated N-half ϩ Y1302W-mutated C-half, W653Y-mutated N-half ϩ Y1302C-mutated C-half, and W653C-mutated N-half ϩ Y1302C-mutated C-half were labeled with [␣-32 P]8-N 3 ATP on ice to determine their K d values (Fig. 4). Because the labeling reactions were performed on ice, ATP hydrolysis reaction should be greatly reduced, and the labeling level should mainly reflect the equilibrium of forward and backward reactions: Panels A, D, G, and J of Fig. 4 show the autoradiograms reflecting [␣-32 P]8-N 3 ATP labeling of the wild-type and mutated N-and C-halves on ice. Labeling was determined by Packard Instant Imager and plotted out against [␣-32 P]8-N 3 ATP concentrations (Fig. 4, B, C, E, F, H, I, K, and L). The K d for wild-type NBD1 co-expressed with wild-type NBD2 is ϳ9 M ATP, whereas the K d for wild-type NBD2 co-expressed with wild-type NBD1 is ϳ33 M ATP (Table II), which are similar to the K d values of NBD1 and NBD2 determined from the fulllength MRP1 protein (35). In contrast, the K d value for W653Cmutated NBD1, co-expressed with Y1302W-mutated NBD2, could not be determined because of very weak labeling of this mutated fragment (Fig. 4D), presumably with a very high K d value. Interestingly, the K d value for Y1302W-mutated NBD2, co-expressed with W653C-mutated NBD1, increased from 33 (the K d of wild-type NBD2) to 139 M ATP (Table II), presumably because of the negative effect of W653C-mutated NBD1 on the Y1302W-mutated NBD2. Therefore there might be an allosteric interaction between the two NBDs even though the experiments were performed on ice. The very weak labeling of W653C-mutated NBD1, including the labeling of W653C-mutated NBD1 co-expressed with Y1302W-mutated (Fig. 4D) and Y1302C-mutated (Fig. 4J) NBD2, indicates that substitution of the aromatic residue with a polar amino acid greatly decreases the affinity for ATP at this mutated NBD1. The K d value of W653Y-mutated NBD1, co-expressed with Y1302C-mutated NBD2, increased from 9 (the K d of wild-type NBD1) to 29 (Table II), presumably because of the negative effect of Y1302C-mutated NBD2. The K d values of the Y1302C-mutated NBD2 (Table II) increased from 33 (the K d of wild-type NBD2) to 122 (the K d of Y1302C-mutated NBD2 co-expressed with W653Y-mutated NBD1) and 160 M ATP (the K d of Y1302Cmutated NBD2 co-expressed with W653C-mutated NBD1), indicating that substitution of this aromatic residue with a polar amino acid also decreased the affinity for ATP at the mutated NBD2. The combinations of the decreased affinities for ATP at the W653C-mutated NBD1 and Y1302C-mutated NBD2 lead to a very high K m (ATP) value (1573 M ATP in Table I) of the double mutated MRP1 in ATP-dependent LTC4 transport. DISCUSSION As discussed in the Introduction, the objective of the project described here is to determine whether the aromatic Trp 653 residue in NBD1 and/or Tyr 1302 in NBD2 of human MRP1 protein are/is involved in nucleotide binding. The approach utilized is to replace the aromatic residues at those two positions with either a different aromatic residue or a moderately polar amino acid. Fig. 1B shows that all of these mutants, no matter whether they were mutated to a different aromatic residue or a polar cysteine residue, produced similar amounts of MRP1 protein as wild type, implying that the substitution with a moderately polar amino acid does not have a significant effect on protein processing and stability. Figs. 2 and 3 show that all of these mutants have the ability to transport LTC4, indicating that these mutations, including the one with a polar residue replacement, do not induce a significant global conformational change and have the ability to bind and hydrolyze ATP.
The above conclusions make it possible to determine whether these aromatic residues in NBD1 and NBD2 are involved in ATP binding. Substitution of the aromatic residue with a different aromatic amino acid do not have a significant negative effect on their ATP-dependent LTC4 transport (Figs. 2 and 3). Their Michaelis constant K m (ATP) values in ATP-dependent LTC4 transport are similar to that of wild type (Table I), implying that these substitutions with a different aromatic amino acid do not significantly change their affinities for ATP. In contrast, the substitutions of the aromatic residues with a polar cysteine residue greatly increased their Michaelis constant K m (ATP) values (Table I), implying that these mutations do affect their ability to bind and hydrolyze ATP, probably because of decreased affinity for ATP. Consistent with this conclusion, substitution of the aromatic Trp 653 in NBD1 and Tyr 1302 in NBD2 with a polar cysteine residue greatly increased their K d (ATP) values (Table II), indicating that the substitution of the aromatic residues with a polar cysteine residue greatly decreased affinity for ATP in the mutated NBD1 and NBD2. These results clearly indicate that the aromatic Trp 653 residue is involved in ATP binding in NBD1, and the Tyr 1302 residue is involved in ATP binding in NBD2. However, mutation of either one or both did not abolish the ATPdependent LTC4 transport, which is in contrast with the Y16S mutation in bacterial ABC transporter HisP that abolished both ATP binding and ATP-dependent histidine transport (26), implying that the interactions between the aromatic residue and the adenine ring of the bound ATP in MRP1 may not be as critical as in bacterial ABC transporter HisP. These results also imply that the stereo orientation of the aromatic residue in MRP1 may be different from that in bacterial ABC transporter HisP (8). In addition, these results are also in contrast with the mutation in human CFTR (25), implying that the stereo orientation of the aromatic residue in human MRP1 may also be different from that in human (25) or mouse CFTR (27).
Interestingly, substitution of the aromatic residue Trp 653 with a polar cysteine residue, such as W653C, W653C/Y1302W, and W653C/Y1302C, greatly decreased their affinity for ATP but did not abolish ATP binding completely and lead to very high K m (ATP) and V max (LTC4) values in ATP-dependent LTC4 transport (Table I). We have found that release of bound ATP, no matter whether it is hydrolyzed or not, from the NBD1 of MRP1 facilitates the protein to start a new cycle of ATP-dependent solute transport. 2 The increased V max values of the W653C-mutated NBD1s, including W653C, W653C/Y1302W, and W653C/Y1302C, can also be explained by this hypothesis. The rationales are: 1) although the binding of ATP to the W653C-mutated NBD1 requires higher nucleotide concentration than that of the wild type because of lower affinity of W653C-mutated NBD1 for that nucleotide, ATP still can easily bind to the W653C-mutated NBD1 because of the ATP concentration in mM range; 2) the much higher K d value of the W653C-mutated NBD1 also indicates that the release rate of the bound ATP from the W653C-mutated NBD1 is much higher than that of wild type; and 3) release of the bound ATP from the W653C-mutated NBD1 resets the MRP1 protein back to its original conformational state so that the molecule can start a new cycle of ATP-dependent solute transport, leading to a higher V max (LTC4) value than that of wild-type MRP1. The substitution of the aromatic residue Tyr 1302 in NBD2 with a polar cysteine residue also decreased its affinity for ATP, resulting in increased K m (ATP) values in ATP-dependent LTC4 transport (Table I) and K d values in ATP binding (Table II). Whether this decreased affinity for nucleotide in NBD2 also facilitates the molecule to start a new cycle of ATP-dependent solute transport is not clear, because the Y1302C-mutated NBD2 co-expressed with wild-type NBD1 increased its V max (LTC4) 1.8-fold (Table I), whereas the Y1302C-mutated NBD2 co-expressed with W653Y-mutated NBD1 did not have a significant effect on its V max (LTC4) value (Table I). In summary, these results clearly indicate that the release of the bound nucleotide from NBD1 plays an important role during ATP-dependent solute transport.  Fig. 4. b Not determined because of the very weak labeling of this fragment.