ATP Binding, Not Hydrolysis, at the First Nucleotide-binding Domain of Multidrug Resistance-associated Protein MRP1 Enhances ADP·Vi Trapping at the Second Domain*

Multidrug resistance-associated protein (MRP1) transports solutes in an ATP-dependent manner by utilizing its two nonequivalent nucleotide binding domains (NBDs) to bind and hydrolyze ATP. We found that ATP binding to the first NBD of MRP1 increases binding and trapping of ADP at the second domain (Hou, Y., Cui, L., Riordan, J. R., and Chang, X. (2002) J. Biol. Chem. 277, 5110–5119). These results were interpreted as indicating that the binding of ATP at NBD1 causes a conformational change in the molecule and increases the affinity for ATP at NBD2. However, we did not distinguish between the possibilities that the enhancement of ADP trapping might be caused by either ATP binding alone or hydrolysis. We now report the following. 1) ATP has a much lesser effect at 0 °C than at 37 °C. 2) After hexokinase treatment, the nonhydrolyzable ATP analogue, adenyl 5′-(yl iminodiphosphate), does not enhance ADP trapping. 3) Another nonhydrolyzable ATP analogue, adenosine 5′-(β,γ-methylene)triphosphate, whether hexokinase-treated or not, causes a slight enhancement. 4) In contrast, the hexokinase-treated poorly hydrolyzable ATP analogue, adenosine 5′-O-(thiotriphosphate) (ATPγS), enhances ADP trapping to a similar extent as ATP under conditions in which ATPγS should not be hydrolyzed. We conclude that: 1) ATP hydrolysis is not required to enhance ADP trapping by MRP1 protein; 2) with nucleotides having appropriate structure such as ATP or ATPγS, binding alone can enhance ADP trapping by MRP1; 3) the stimulatory effect on ADP trapping is greatly diminished when the MRP1 protein is in a “frozen state” (0 °C); and 4) the steric structure of the nucleotide γ-phosphate is crucial in determining whether binding of the nucleotide to NBD1 of MRP1 protein can induce the conformational change that influences nucleotide trapping at NBD2.

Multidrug resistance is a major obstacle to successful chemotherapeutic treatment of many types of cancers. Over-expression of P-glycoprotein (P-gp) 1 and/or multidrug resistance-associated protein (MRP1) confers resistance to a broad range of anti-cancer drugs (1,2). Both proteins transport anticancer drugs out of cells in an ATP-dependent manner by utilizing their membrane-spanning domains and two nucleotide binding domains (NBDs) (3)(4)(5), i.e. they couple ATP binding and hydrolysis to transport of solutes (6 -15). However, it is unknown whether they share the same mechanism of this coupling. In the extensively studied P-gp, the two NBDs have been shown to be functionally equivalent with identical ATP hydrolysis steps occurring alternately at each NBD (16 -20) and coupling one transport event with one ATP hydrolysis (21). Ambudkar's group (22) reported that there are two independent ATP hydrolysis events in a single drug transport cycle, one ATP hydrolysis is associated with efflux of drug, whereas the other causes conformational resetting to the original state of the molecule (23). However, in their interpretation the ATP binding/hydrolysis sites of P-gp are recruited in a random manner during hydrolysis (23), meaning that the two NBD sites are functionally equivalent. In other reports, the two NBDs of P-gp were found to be essential for its function but not entirely symmetric (24,25). Vigano et al. (25) proposed recently that ATP binding/hydrolysis at NBD1 is associated with efflux of drug, whereas the event at NBD2 is associated with the "reset" of the molecule. Therefore, how events at NBD1 and NBD2 of P-gp cooperate during drug transport is still not clear. Considerable evidence has accumulated indicating that the two NBDs of some other ATP-binding cassette (ABC) transporters, including the sulfonylurea receptor (SUR1) (26,27), cystic fibrosis transmembrane-conductance regulator (CFTR) (28,29), and MRP1 (30 -33), have very distinctive properties. For example, in the case of MRP1, the following points clearly indicate that its two nonequivalent NBDs have different properties and functions. First, modifications of the consensus Walker motifs in the two NBDs do not inactivate the protein completely and have different effects on solute transport (30 -32). Second, photoaffinity labeling experiments with 8-azido-ATP also revealed an asymmetry between NBD1 and NBD2, with NBD1 preferentially labeled by 8-N 3 [␥-32 P]ATP (30,31), whereas NBD2 trapped the nucleoside diphosphate hydrolysis product (30,31,33). Third, the ATP binding/hydrolysis sites of MRP1 seem not to be recruited in a random manner because photolabeling by the nonhydrolyzable 8-N 3 [␣-32 P]AMP-PNP occurred predominately at NBD1 (34), and the NBD1 fragment was labeled predominantly with 8-N 3 [␣-32 P]ATP on ice in the dual-expressed N-and C-halves of MRP1 (30). Fourth, ADP trapping at NBD2 enhances ATP binding at NBD1 (31), and ATP binding at NBD1 allosterically enhances ADP trapping or AMP-PNP binding at NBD2 (34), implying that ATP binding at NBD1 or ADP trapping at NBD2 induces conformational change of the MRP1 molecule. It seems likely that the conformational changes of the MRP1 molecule caused by ATP bind-ing/hydrolysis provide mechanical force to pump drug out of the cell. However, experiments to date have not directly determined whether ATP binding alone or hydrolysis at NBD1 can cause conformational change of the MRP1 protein and enhance ADP trapping at NBD2. We have now done this and found that ATP hydrolysis is not required to enhance ADP trapping, and ATP binding alone is sufficient to enhance ADP trapping. In addition, the steric position of the ␥-phosphate of the nucleotide is a crucial factor in determining whether or not the binding of the nucleotide to NBD1 of MRP1 protein can enhance ADP trapping.

EXPERIMENTAL PROCEDURES
Materials-Sodium orthovanadate, EGTA, ATP, ATP␥S, AMP-PNP, AMP-PCP, LiCl, ouabain, and sheep brain lipid were purchased from Sigma. Formic acid and polyethyleneimine-cellulose plates were from Fisher. His-Bind Resin was from Novagen. N-Dodecyl-␤-D-maltoside (DDM) was from Calbiochem. 8 (31,35). These cells were cultured in Dulbecco's modified Eagle's medium/F-12 with 5% fetal bovine serum supplemented with 150 M methotrexate (the original colonies expressing MRP1 were selected in 500 M methotrexate and decreased to 150 M methotrexate because the lesser amount of this drug did not affect the expression of MRP1 protein) at 37°C in 5% CO 2 . Cells for membrane vesicle preparations were grown in roller bottles (Bellco) in Dulbecco's modified Eagle's medium/F-12 media containing 5% fetal bovine serum and 150 M methotrexate at 37°C.
Membrane Vesicle Preparations-MRP1-containing membrane vesicles were prepared according to the procedure described previously (31). Briefly, the cells grown in roller bottles were collected by centrifugation, resuspended in membrane vesicle preparation buffer containing 10 mM Tris-HCl, pH 7.5, 250 mM sucrose, 0.2 mM MgCl 2 and 1ϫ protease inhibitors (2 g/ml aprotinin, 121 g/ml benzamidine, 3.5 g/ml E64, 1 g/ml leupeptin, and 50 g/ml Pefabloc) and equilibrated on ice for 20 min at 800 p.s.i. in a Parr N 2 cavitation bomb. After pressure release, the cell homogenate was adjusted to 1 mM EDTA. The homogenate was diluted 5-fold with 10 mM Tris-HCl, pH 7.5, and 25 mM sucrose and centrifuged at 1000 ϫ g to remove nuclei and unbroken cells. The supernatant was overlaid on a 35% sucrose solution containing 10 mM Tris-HCl, pH 7.5, and 1 mM EDTA and centrifuged at 16,000 ϫ g for 30 min. Membranes at the interface were collected, diluted 5-fold with a solution containing 10 mM Tris-HCl, pH 7.5, and 250 mM sucrose, and then centrifuged at 100,000 ϫ g for 45 min. The pellet was resuspended in a solution containing 10 mM Tris-HCl, pH 7.5, 250 mM sucrose, and 1ϫ protease inhibitors. After passage through a Liposofast™ vesicle extruder (200-nm filter, Avestin, Ottawa, Canada), the membrane vesicles were aliquoted and stored in Ϫ80°C.
Hexokinase Treatment of ATP Analogues-To remove the trace amount of contaminating ATP in ATP analogues, 5 mM AMP-PNP, AMP-PCP, or ATP␥S solutions were treated with hexokinase as described previously (36). Briefly, each nucleotide at 5 mM in a 1-ml solution containing 20 mM Tris-HCl, pH 8.0, 10 mM glucose, 10 mM MgCl 2 , and 20 units of hexokinase was incubated at 30°C for 30 min. Hexokinase in the solution was removed by passing the solution through a Centricon 10 microconcentrator (Amicon, molecular weight cut-off 10,000).
Photoaffinity Labeling of MRP1 Protein-Vanadate preparation and photoaffinity labeling of MRP1 protein were performed according to procedures described previously (31). Briefly, the photolabeling experiments were carried out in a 10-l solution containing 10 g of membrane vesicles from MRP1 expressing cells, 10 The labeled proteins were separated by polyacrylamide gel (7%) electrophoresis and electroblotted to a nitrocellulose membrane.
Purification of MRP1 Protein-The previous procedure (35) used to purify MRP1 protein from BHK cells was modified slightly. Cells collected from roller bottles were washed with a solution containing 10 mM Tris-HCl, pH 7.5, 250 mM sucrose, and 0.2 mM MgCl 2 , and the cell pellet was kept in a Ϫ80°C freezer overnight. The cell pellet was resuspended in a solution containing 10 mM Hepes, pH 7.2, 1 mM EDTA, and 1ϫ protease inhibitors, and then the cells were transferred to a Dounce homogenizer. After seven strokes in this homogenizer, the same volume of a solution containing 10 mM Hepes, pH 7.2, 1 mM EDTA, and 500 mM sucrose was added to the homogenizer, and then another six strokes were performed. Nuclei were removed by centrifugation at 300 ϫ g for 15 min at 4°C. Membranes were collected at 33,000 ϫ g for 45 min at 4°C. The membrane pellet was resuspended in a binding buffer containing 20 mM Tris-HCl, pH 7.9, 500 mM NaCl, 20% glycerol, 25 mM imidazole, 1% DDM, 0.4% sheep brain lipid, and 0.05% ␤-mercaptoethanol and sonicated for a short time on ice. The insoluble material was removed by centrifugation at 10,000 ϫ g for 15 min. The supernatant was applied onto a His-Bind Resin column that had been pre-equilibrated with the binding buffer. The column was washed with 6 column volumes of modified binding buffer containing 0.1% DDM and 25 mM imidazole (first wash), 6 column volumes of modified binding buffer containing 0.1% DDM and 40 mM imidazole (second wash), and 6 column volumes of modified binding buffer containing 20 mM Tris-HCl, pH7.4, 0.1% DDM and 40 mM imidazole (third wash). The bound protein was eluted with 2 column volumes of buffer containing 20 mM Tris-HCl, pH 7.4, 500 mM NaCl, 20% glycerol, 300 mM imidazole, 0.1% DDM, 0.4% sheep brain lipid, and 0.05% ␤-mercaptoethanol. The eluate was dialyzed against a solution containing 20 mM Tris-HCl, pH 7.4, 500 mM NaCl, 20% glycerol, 0.1% DDM, and 0.05% ␤-mercaptoethanol.

ATP Enhancement of ADP Trapping in MRP1 Protein Is
Greatly Diminished on Ice-We found that photolabeling of MRP1 protein with 8-N 3 [␣-32 P]ADP was enhanced 4-fold mainly at NBD2 (34), suggesting that ATP binding or hydrolysis at NBD1 caused conformational change of the protein and increased affinity for ATP at NBD2. However, these experiments were performed at 37°C and did not distinguish whether ATP binding or hydrolysis caused the conformational change of the protein. To distinguish these two possibilities, the same photolabeling experiments were performed on ice. Fig. 1, A and B, showed that the trapping of ADP to the protein was enhanced ϳ30 -40% in the presence of 5 to 20 M ATP. The trapping was inhibited almost 70% in the presence of 640 M ATP, simply because of competition between ATP binding and ADP trapping. These results were interpreted in the following two ways: 1) ATP hydrolysis at NBD1 may be required to induce the conformational change of the molecule to enhance ADP trapping at NBD2 and limited ATP hydrolysis on ice greatly diminishes the enhancement effect; 2) ATP binding alone can cause a conformational change that increases ADP trapping at NBD2 at higher temperature, such as 37°C, and the smaller augmentation of trapping by ATP on ice may reflect a membrane structure and/or MRP1 protein that are in a "frozen state." Therefore these results cannot be used to distinguish the two possibilities mentioned above.
Hexokinase-treated AMP-PNP Does Not Enhance ADP Trapping by MRP1-The nonhydrolyzable ATP analogue, AMP-PNP, enhanced ADP trapping by MRP1 protein (34), support-ing the hypothesis that nucleotide binding alone can cause the conformational change of the protein. However, this could be misleading if there is a trace amount of hydrolyzable nucleotide contaminant. This can be removed efficiently by treating the solution with hexokinase (36). The results in Fig. 2 show that AMP-PNP, after hexokinase treatment, did not enhance ADP trapping by MRP1 protein, but instead it inhibited ADP trapping. Previous results clearly indicated that 8-N 3 [␣-32 P]AMP-PNP bound to NBD1 of MRP1 (34). Therefore it seems likely that nucleotide binding alone, at least in the case of AMP-PNP, cannot enhance ADP trapping by MRP1.
The Nonhydrolyzable ATP Analogue, AMP-PCP, Enhances ADP Trapping-If nucleotide binding alone cannot cause the conformational change leading to increased trapping, then the binding of other nonhydrolyzable ATP analogues, such as AMP-PCP, should also not enhance ADP trapping. To test this hypothesis, both hexokinase-treated and untreated AMP-PCP were utilized in the ADP trapping experiments. Fig. 3, A and B, shows that AMP-PCP, which was not treated with hexokinase, enhanced ADP trapping ϳ50%, whereas the hexokinasetreated AMP-PCP did so by ϳ20% (Fig. 3, C and D). The enhancing effect of the untreated AMP-PCP is slightly greater than the treated compound, presumably reflecting the removal of a trace amount of ATP. These results may mean that nucleotide binding at NBD1 alone can induce the conformational change required to enhance ADP trapping at NBD2. However, this interpretation is not consistent with the conclusion derived from Fig. 2. Therefore we postulated that in some nucleotides, such as ATP or AMP-PCP, binding alone without hydrolysis can induce this conformational change, whereas in other nucleotides, such as AMP-PNP, binding cannot induce this change.
ATP␥S Enhances ADP Trapping-To further test the above hypothesis, ATP␥S was utilized in the ADP trapping experiments. Fig. 4, A and B, show that ATP␥S enhanced ADP trapping to MRP1 protein almost 3-fold. Similarly, hexokinasetreated ATP␥S also increased the trapping ϳ3-fold (Fig. 4, C and D). Because the enhancing effect of ATP␥S is much greater than that of AMP-PCP (Fig. 3) and slightly less than that of ATP (34), ATP␥S may have an effect on ADP trapping similar to that of ATP under other conditions, such as at 0°C. The experiments in Fig. 4E were performed on ice and showed that ATP␥S stimulated ADP trapping by MRP1 protein ϳ20 -30% (Fig. 4F). These results imply that ATP␥S binding alone can cause the underlying conformational change.
ATP␥S Binding without Hydrolysis Can Enhance ADP Trapping-Thus far, we have not ruled out the possibility that the poorly hydrolyzable ATP analogue, ATP␥S, might be hydrolyzed to some extent during the 10-min incubation period at 37°C. Therefore we estimated how long it would take to hydrolyze one ATP␥S molecule. Table I shows that it took ϳ20 min for one MRP1 molecule at 37°C to hydrolyze one ATP␥S. Therefore, ϳ50% of the ATP␥S bound to MRP1 protein should be hydrolyzed during the 10-min of incubation at 37°C. Hence, some hydrolysis did occur.
Because it takes ϳ20 min for one MRP1 molecule to hydrolyze one ATP␥S, at shorter times hydrolysis is essentially negligible. The experiments in Fig. 5A were performed as follows. The pelleted membrane proteins containing MRP1 were resuspended on ice with a 10-l ice-cold reaction mixture, transferred to a 37°C water bath, and incubated for 1, 2, 4, 8, 16, or 32 min. The samples were brought back to ice after incubation for the indicated times at 37°C and diluted with 500 l of ice-cold Tris-EGTA buffer immediately, and then the membrane was pelleted by centrifugation in a cold room (4°C). Therefore, the temperature inside of the tubes incubated at 37°C for only 1 min should not be 37°C at the beginning of the incubation, and the incubation time at 37°C must be less than 1 min. The amount of ATP␥S hydrolyzed during this short period must be much less than during the 32-min period (Fig.  5A, lane 32Ј). Yet, this sample (1-min incubation at 37°C) had the greatest enhancing effect on ADP trapping (Fig. 5, A and  C), indicating that ATP␥S hydrolysis was not responsible for the conformation change and that ATP␥S binding alone was sufficient. Fig. 5C shows that the enhancing effects gradually decreased with incubation time. The mechanism of this diminution is not yet clear. However, one of the possible reasons is that the unstimulated ADP trapping in the absence of ATP␥S gradually increased with incubation time (Fig. 5, A and B), implying that occlusion or trapping of ADP by vanadate takes time or requires conformational change induced by ATP binding at NBD1. ATP␥S did not significantly enhance ADP trapping if the samples were incubated on ice (Fig. 5A, lanes 0Ј, Ϫ  and ϩ). This result is consistent with the data in Figs. 1 and 4E showing that no matter if ATP or ATP␥S was employed, the nucleotides did not significantly enhance the ADP trapping if the experiments were performed on ice. DISCUSSION Over-expression of MRP1 protein confers resistance to a broad range of anti-cancer drugs (2). Solutes (for example, anti-cancer drugs) are extruded out of cells by MRP1 protein in an ATP-dependent manner (7,14,37) by utilizing its two NBDs to bind and hydrolyze ATP. Both NBDs of MRP1 protein can bind nucleotides (30,31,33). However, the properties and the functions of the two NBDs do not seem to be equal as discussed in the Introduction. How are the events at NBD1 and NBD2 related during solute transport? We have found that the binding of ATP at NBD1 enhances ADP trapping or AMP-PNP binding to NBD2, implying that ATP binding at NBD1 causes conformational change of the MRP1 molecule (34) and the trapping of ADP⅐Vi, mimicking the ATP hydrolysis intermediate ADP⅐Pi, enhances intact ATP binding at NBD1 (31), implying that ATP hydrolysis at NBD2 causes conformational change in the MRP1 molecule. These conformational changes may contribute to active solute transport by the molecule. However, it is not clear whether the enhancing effect of nucleotide binding at NBD2 requires ATP hydrolysis at NBD1. The nonhydrolyzable ATP analogue, AMP-PNP, slightly enhances ADP trapping to MRP1 protein (34), implying that nucleotide binding alone can induce the conformational change. However, why was the enhancing effect of AMP-PNP much lower than that of ATP (34)? Interestingly, a poorly hydrolyzable ATP analogue, ATP␥S, induced the conformational change to the same extent as ATP as determined by fluorescence quenching (38), implying that nucleotide binding alone may be responsible for induction of the conformational change. However, because ATP␥S is a poorly hydrolyzable ATP analogue, the nucleotide used in the experiments might be hydrolyzed by MRP1. Therefore, the question of whether ATP binding or hydrolysis is required to induce the conformational change was still unanswered. Our present results clearly indicate that ATP binding alone is sufficient to induce the conformational change of MRP1  change. However, in contrast to AMP-PNP, another nonhydrolyzable ATP analogue, AMP-PCP, can enhance ADP trapping (Fig. 3), implying that nucleotide binding alone can induce the conformational change of the protein. Interestingly, the enhancing effect of AMP-PCP on ADP trapping was much less than that of ATP, perhaps because of the structural difference between AMP-PCP and ATP. If the structure difference between nucleotides is a major factor determining the enhancing effects, then a nucleotide with a structure similar to that of ATP, such as ATP␥S, should have a similar effect. Indeed, the hexokinase-treated ATP␥S had an effect similar to that of ATP (Fig. 4), although we had not ruled out the possibility that ATP␥S might be hydrolyzed during the 10-min incubation at 37°C. The results shown in Table I (Figs. 1 and 4), implying that ATP binding alone can induce the conformational change only under proper conditions such as those of temperature. In conclusion: 1) nucleotide hydrolysis at NBD1 is not required to induce the conformational change stimulating ADP trapping at NBD2; 2) nucleotide binding alone under proper conditions is sufficient to induce the conformational change; 3) the proper steric structure of the ␥-phosphate of the nucleotide is a crucial factor affecting the ability of the nucleotide to induce the conformational change; when nitrogen replaces oxygen between the ␤ and ␥ phosphates in AMP-PNP, this eliminates the enhancing effect, whereas when carbon replaces oxygen in AMP-PCP this greatly reduces the enhancing effect; 4) the replacement of oxygen with sulfur on the ␥-phosphate in ATP␥S also slightly reduces the enhancing effect.
How does ATP binding to NBD1 induce the conformational change? The original "unexcited" structure of MRP1 protein should be the same no matter whether ATP, ATP␥S, AMP-PNP, or AMP-PCP has been utilized to excite MRP1 protein.
The only difference between ATP, AMP-PNP, and AMP-PCP is the atom between the ␤and ␥-phosphates, which determines the distance and angle between the ␤and ␥-phosphates (39). Therefore the steric structure of the ␥-phosphate is the crucial determinant of the structural perturbation. Which residues of MRP1 protein interact with the ␥-phosphate when the nucleotide binds to NBD1? Although the three-dimensional structure of MRP1 protein has not been solved, the structures of other ATP-binding cassette transporters, such as periplasmic histidine permease of Salmonella typhimurium (40) and MJ0796 from Methanococcus jannaschii (41), may provide a clue. By analogy with these, Ser-685 (a residue in Walker A of NBD1), Gln-713 (␥-phosphate linker), and Asp-792 (a residue in Walker B of NBD1) may interact with the ␥-phosphate via Mg 2ϩ . Lys-684 (a residue in Walker A of NBD1) may interact with the ␥-phosphate directly. Asp-793 may function as a catalytic base to attack the water molecule to hydrolyze the bound ATP. Val-680 and Gly-681 (residues in the Walker A motif of NBD1) and Ser-1431, Val-1432, and Gly-1433 (residues in the ATP-binding cassette signature sequence of NBD2) may interact with the ␥-phosphate. Upon ATP binding to NBD1 the ␥-phosphate of the bound nucleotide either "pushes" or "pulls" (by electrostatic interactions between the ␥-phosphate and the charged residues) some of these residues, leading to the conformational change and increasing the affinity for ATP at NBD2 (34). Changing of the atom between the ␤and ␥-phosphates changes the spatial orientation of the ␥-phosphate, either eliminating (such as AMP-PNP) or diminishing (such as AMP-PCP) the pushing or pulling force. If that is the case, mutations of these residues may affect ATP binding at NBD1 and decrease the ATP enhancing effect on ADP trapping. Indeed, mutations of K684L and D792A greatly diminish the ATP enhancing effect on ADP trapping (34). Interestingly, mutation of D792E, which did not significantly change the negative charge at that position, also greatly diminished the ATP enhancement effect on ADP trapping, 2 indicating that the distance between the negative charged residue, Asp-792, and the ␥-phosphate of ATP is also very important. The samples were resuspended in a 10-l ice-cold reaction mixture on ice, transferred immediately to a 37°C water bath, and then incubated at 37°C for the indicated time. Lanes 0Ј, Ϫ and ϩ, the samples were incubated on ice for 1 min without transferring to 37°C water bath. The other samples were incubated at 37°C for 1 min (1Ј, Ϫ and ϩ), 2 min (2Ј, Ϫ and ϩ), 4 min (4Ј, Ϫ and ϩ), 8 min (8Ј, Ϫ and ϩ), 16 min (16Ј, Ϫ and ϩ), and 32 min (32Ј, Ϫ and ϩ). The samples were washed with 500 l of ice-cold Tris-EGTA buffer immediately after the 37°C incubation and UV irradiated on ice for 2 min. Because the samples were transferred from 0 to 37°C directly, the temperature inside of the tubes was not 37°C at the beginning of the 37°C incubation. Although these data provide evidence that ATP binding, not hydrolysis, at NBD1 is sufficient to induce the conformational change and enhance nucleotide binding at NBD2, they do not speak directly to how the protein couples ATP hydrolysis to solute transport. Upon binding of ATP to NBD2 there should be a transient state in which both NBDs bound ATP. Whether the bindings of ATP to both NBD1 and NBD2 will lead to the formation of a transient ATP sandwich (41)(42)(43) between NBD1 and NBD2 is not known. However, the efficient hydrolysis of the bound ATP at NBD2 (30, 31) may cause conformational change of the MRP1 protein (34) and generate the negatively charged products, ADP and phosphate. The electrostatic repelling force between them may facilitate the release of the hydrolysis products from NBD2. What is the fate of the ATP bound at NBD1 during or after release of the ATP hydrolysis products, phosphate and ADP, from NBD2? When the dually expressed N-and C-halves were labeled with 8-N 3 [␣-32 P]ATP on ice, the NBD1 fragment was predominantly labeled (30). However, when the experiments were performed at 37°C, NBD2 was predominantly labeled (30). Consistent with the above finding, when full-length wild-type MRP1 was labeled with 8-N 3 [␣-32 P]ATP on ice and digested with trypsin, the labeling at NBD1 was greater than at NBD2. 2 In contrast, when the experiments were performed at 37°C, NBD2 was predominantly labeled (31). We interpreted these results to mean that the ATP bound at NBD1 was released during the incubation period at 37°C. If that is the case, whether the ATP bound at NBD1 is released as an intact ATP or hydrolyzed first and then released is not known. No matter how the ATP bound at NBD1 is released, the releasing of this bound nucleotide may bring the MRP1 molecule back to the original unexcited state so that the MRP1 molecule can start another cycle of solute transport.