| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 13, 12325-12336, March 26, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶¶||**||
From the
Cancer Research Laboratories and the ||Department of Pharmacology and Toxicology, Queen's University, Kingston, Ontario K7L 3N6, Canada
Received for publication, October 17, 2003 , and in revised form, January 8, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-estradiol-17--(D-glucuronide) and methotrexate transport, whereas transport of other organic anions was reduced or unchanged. Significant substrate-specific changes in the ATP dependence of transport and binding by the P1150A mutant were also observed. Our findings demonstrate the importance of TM6, TM8, TM10, TM11, and TM14 in MRP1 transport function and suggest that CL7 may play a differential role in coupling the activity of the nucleotide binding domains to the translocation of different substrates across the membrane. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Increased expression of MRP1 has been detected in drug-resistant tumor cell lines and tumor tissues, and MRP1 has been demonstrated to mediate ATP-dependent cellular efflux of a wide variety of anticancer drugs and other xenobiotics such as antimony and arsenic-centered oxyanions (3, 9). MRP1 has also been shown to transport a broad spectrum of organic anions including oxidized and reduced glutathione (GSSG and GSH) as well as anionic GSH, glucuronide, and sulfate conjugates (3, 10–13). Endogenous organic anion substrates of MRP1 include the cysteinyl leukotriene LTC4 and the conjugated estrogens E217G, E13SO4, and dihydroepiandrosterone sulfate (11, 14–18). Exogenous organic anions such as the folic acid analog, MTX, and conjugated metabolites of the carcinogens, aflatoxin B1 and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol, are also substrates of MRP1 (3, 19–21). MRP1 is expressed in most human tissues, and studies of mice in which the Mrp1 gene has been disrupted have demonstrated that this transporter has a role in inflammation and detoxication (3, 22).
One approach toward understanding the molecular details of the transport mechanism of MRP1 is to identify the amino acid residues critical for its activity by functional analysis of MRP1 mutants created using site-directed mutagenesis. For example, previous studies have identified a number of aromatic, polar, and charged amino acids in or proximal to the TM helices in both MSD2 (amino acids 323–601) and MSD3 (amino acids 972–1248) that play an important role in determining MRP1 transport activity and substrate specificity (23–30). Other studies have shown that Ala substitution of Cys residues in these domains as well as the cytoplasmic and extracytoplasmic regions of MRP1 do not appear to have major effects on its transport function (31).
Sequence analyses have shown that Pro residues are more frequently found in TM 
-helices of ion channels and transporters than in -helices of water-soluble proteins (32). Because their backbone N is not available for intrahelical H-bonding, and because of steric constraints caused by their ring structure, Pro residues may introduce "kinks" in TM -helices, and a role for such Pro residues in substrate binding and/or translocation has been proposed for a number of membrane proteins (32–37). Non-membrane associated Pro residues have also been shown to be important for the function of certain ion channels and transporters (38). In a previous study, we replaced 12 Pro residues in MSD1 and CL3 of MRP1 with Ala and found that while several of the resulting mutants showed reduced levels of MRP1 expression, the substitutions had little or no effect on LTC4 and E217G transport (39). These findings are consistent with an interpretation that MSD1 is not critical for conjugated organic anion transport but probably has a role in the stable expression of MRP1 in mammalian cell plasma membranes. In the present study, we have extended this earlier work by investigating the role of the 18 Pro residues that reside in both the non-membrane and TM regions of MSD2 and MSD3, by examining the consequences of replacing these residues with Ala on the expression and function of MRP1. Many of these Pro residues are highly conserved in other human ABCC family members and MRP1 orthologs in other species (Table I). Thus, the findings of our study may have implications for the structure-function analysis of other ABCC proteins.
|
G, E13SO4, and MTX transport. Unexpectedly, a mutant containing a substitution of Pro1150 predicted to be in the cytoplasmic loop (CL7) linking TM15 to TM16 showed markedly increased levels of E217G and MTX transport, whereas transport of other organic anions was reduced or unchanged. The altered transport activity of P1150A was associated with significant changes in the ATP dependence of E217G but not LTC4 transport, suggesting that CL7 may play a differential role in coupling the activity of the MRP1 NBDs to the translocation of different substrates across the membrane. |
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
G (55 Ci mmol–1), [6,7-3H]E13SO4 (53 Ci mmol–1), and [glycine-2-3H]GSH (40–44.8 Ci mmol–1) were purchased from PerkinElmer Life Sciences. [14,15,19,20-3H]LTC4 (115.3 Ci mmol–1) was from Amersham Biosciences, and [3',5',7-3H]MTX sodium salt (17 Ci mmol–1) and [3',5',7,9-3H]leucovorin diammonium salt (17 Ci mmol–1) were purchased from Moravek Inc. (Brea, CA). 8-Azido-[-32P]ATP (14.1 Ci mmol–1) was from ALT, Inc. (Lexington, KY). LTC4 was purchased from Calbiochem, and nucleotides, GSH, apigenin, 2-mercaptoethanol, acivicin, E217G, leucovorin, E13SO4, NaF, and BeSO4 were purchased from Sigma. Vector Construction and Site-directed Mutagenesis—Pro mutations in MRP1 were generated using the QuikChangeTM site-directed mutagenesis kit (Stratagene, La Jolla, CA). The template for mutagenesis of Pro residues in MSD2 was prepared by cloning a 1.9-kb BamHI/SphI fragment (containing nucleotides 841–2701 encoding amino acids 281–900) from pcDNA3.1(–)-MRP1K into pBluescriptSK(+) (Stratagene, La Jolla, CA) into which a SphI site had been introduced into the multiple cloning site (24). The template for mutagenesis of Pro residues in MSD3 was prepared by cloning a 2-kb XmaI fragment (containing nucleotides 2337–4322 encoding amino acids 780–1440) from pcDNA3.1(–)-MRP1K into pGEM-3Z (Promega, Madison, WI) (23). Proline substitutions were generated in the pBluescriptSK(+) and pGEM-3Z plasmids above according to the manufacturer's instructions with the following mutagenic primers (substituted nucleotides are underlined): MSD2 Pro mutants, P323A, 5'-GTG TTA TAC AAG ACC TTT GGC GCC TAC TTC CTC ATG AGC-3'; P343A, 5'-G ATG ATG TTT TCC GGG GCG CAG ATC TTA AAG TTG C-3'; P359A, 5'-G AAT GAC ACG AAG GCC GCA GAC TGG CAG GG-3'; P448A, 5'-G ATC TGG TCA GCC GCC CTG CAA GTC ATC CTT GC-3'; P464A, 5'-G CTG AAT CTG GGC GCT TCC GTC CTG GCT GG-3'; P478A, 5'-G GTC CTC ATG GTG GCC GTC AAT GCT GTG-3'; P557A, 5'-CC TGG GTC TGC ACG GCC TTT CTG GTG GCC-3'; P595A, 5'-C AAC ATC CTC CGG TTT GCC CTG AAC ATT CTC C-3'; P600A, 5'-CCC CTG AAC ATT CTC GCG ATG GTC ATC AGC AGC-3'; and MSD3 Pro mutants B P1003A, 5'-C TGG ACT GAT GAC GCC ATC GTC AAC GGG-3'; P1060A, 5'-C CTG CGG TCA GCC ATG AGC TTC-3'; P1068A, 5'-C TTT GAG CGG ACC GCG AGT GGG AAC C-3'; P1088A, 5'-C TCC ATG ATC GCG GAG GTC ATC-3'; P1113A, 5'-CTG CTG GCC ACG GCC ATC GCC GCC-3'; P1120A, 5'-GCC ATC ATC ATC GCG CCC CTT GG-3'; P1121A, 5'-C ATC ATC ATC CCG GCG CTT GGC CTC ATC-3'; P1150A, 5'-G GTC AGC CGC TCC GCG GTC TAT TCC C-3'; P1191A, 5'-C CAG AAG GCC TAT TAC GCT AGC ATC GTG GCC AAC-3'; P1120A/P1121A, 5'-C GCC ATC ATC ATC GCA GCG CTT GGC CTC ATC TAC TTC-3'. After confirming all mutations by sequencing or diagnostic restriction enzyme digests, the 1-kb BamHI/Bsu36I fragment (MSD2 Pro mutants) and the 1.3-kb BsmBI/EcoRI fragment (MSD3 Pro mutants) were subcloned back into pcDNA3.1(–)-MRP1K, and the fragments in the full-length constructs were sequenced once again.
Transfections of MRP1 Expression Vectors in Human Embryonic Kidney Cells—Wild-type and mutant pcDNA3.1(–)-MRP1K expression vectors were transfected into SV40-transformed human embryonic kidney cells (HEK293T). For membrane vesicle preparations, 
7 x 106 cells were seeded per 15-cm dish and transfected 24 h later with 16 µg of plasmid DNA using FuGENE 6 (Roche Diagnostics) according to the manufacturer's instructions. After 60 h, the HEK293T cells were harvested, and membrane vesicles were prepared as described previously (40). Empty vector and vector containing the wild-type cDNAs were included as controls in all experiments. Transfections were also carried out with a second independently generated clone of each mutant so that clonal variation could be excluded as an explanation for any changes in expression levels and/or transport activities observed.
Measurement of MRP1 Protein Levels in Transfected Cells—The levels of wild-type and mutant MRP1 proteins were determined by immunoblot analysis of membrane protein fractions from transfected cells essentially as described (41). Proteins were resolved on a 6–7% polyacrylamide gel and electrotransferred to ImmobilonTM P membrane (Millipore, Bedford, MA). Blots were blocked with 4% (w/v) skim milk powder for 1 h followed by incubation with the human MRP1-specific monoclonal antibody QCRL-1 (diluted 1:10,000) (42, 43). After washing, blots were incubated with horseradish peroxidase-conjugated goat anti-mouse antibody (Pierce) followed by application of Renaissance chemiluminescence blotting substrate (PerkinElmer Life Sciences) and exposed to film. Relative levels of MRP1 protein expression were estimated by densitometric analysis using a ChemiImagerTM 4000 (Alpha Innotech, San Leandro, CA). To confirm equal protein loading of the gel lanes, the blots were also stained with Amido Black.
MRP1-mediated Transport by Membrane Vesicles—Membrane vesicles were prepared from transiently transfected HEK293T cells, and ATP-dependent transport of 3H-labeled substrates by the membrane vesicles was measured using a rapid filtration technique as described previously (40). Briefly, LTC4 transport assays were performed at 23 °C in a 50-µl reaction containing 50 nM LTC4 (40 nCi per reaction), 4 mM ATP, 10 mM MgCl2, an ATP-regenerating system (10 mM creatine phosphate, 100 µg ml–1 creatine kinase), and 2 µg of vesicle protein in transport buffer (50 mM Tris-HCl, 250 mM sucrose, pH 7.4). Control uptake assays contained 4 mM AMP, and the regenerating system was omitted. Uptake was stopped at the desired time by rapid dilution in ice-cold buffer, and then the reaction was filtered through glass-fiber filters presoaked in transport buffer. Radioactivity was quantitated by liquid scintillation counting. All data were corrected for the amount of [3H]LTC4 that remained bound to the filter, which was usually <10% of the total radioactivity. Transport in the presence of AMP was subtracted from transport in the presence of ATP to determine ATP-dependent LTC4 uptake. All transport assays were carried out in triplicate, and results are expressed as means (±S.D.).
Uptake of [3H]E217G was measured in a similar fashion, except that membrane vesicles (2 µg of protein) were incubated at 37 °C in a total reaction volume of 50 µl containing 400 nM E217G (40 nCi per reaction) and the components as described for LTC4 transport. [3H]E13SO4 uptake was performed at 37 °C in a 50-µl total reaction volume containing membrane vesicles (2 µg of protein), 300 nM E13SO4 (100 nCi per reaction), in the presence of 3 mM GSH and 10 mM dithiothreitol and the same components described above (16, 44). Apigenin-stimulated [3H]GSH uptake was also measured by rapid filtration with membrane vesicles (20 µg of protein) incubated at 37 °C for 20 min in a 60-µl reaction volume with 100 µM [3H]GSH (120 nCi per reaction) and 30 µM apigenin (13). To minimize GSH catabolism by 
-glutamyltranspeptidase during transport, membranes were preincubated with 0.5 mM acivicin for 10 min at 37 °C prior to measuring [3H]GSH uptake. [3H]MTX uptake was performed at 37 °C for 20 min in a 50-µl total reaction volume containing membrane vesicles (10 µg of protein), 100 µM MTX (200 nCi per reaction), and the same components described above (26, 45), whereas [3H]leucovorin uptake was performed at 37 °C for 20 min in a 50-µl total reaction volume containing membrane vesicles (10 µg of protein) and 250 µM leucovorin (500 nCi per reaction).
Kinetic Analysis of ATP-dependent [3H]LTC4 and [3H]E217G Transport—Km and Vmax values of ATP-dependent LTC4 transport by membrane vesicles (4 µg of protein) were determined by measuring uptake at eight different [3H]LTC4 concentrations (10–1000 nM) for 1 min at 23 °C in 50 µl of transport buffer containing components as described above. Kinetic parameters for [3H]E217G uptake were determined by measuring uptake at eight different [3H]E217G concentrations (0.25–25 µM) for 1 min at 37 °C as before. The apparent Km value for ATP was determined by measuring initial rates of [3H]LTC4 and [3H]E217G uptake in the presence of 10 different concentrations of nucleotide (2–4000 µM). Data were analyzed using Graph Pad PrismTM software and kinetic parameters determined by non-linear regression and Michaelis-Menten analyses.
Northern Blot Analysis—Total RNA was isolated from transfected HEK cells using a PureScript RNA isolation kit (Gentra Systems, Minneapolis, MN) according to the manufacturer's instructions. RNA (5 µg) was resolved on a 1% agarose gel containing 0.66 M formaldehyde and transferred onto a Magna charge nylon membrane (Micron Separations, Westborough, MA) by capillary action and cross-linked to the membrane by exposure to UV light. The membrane was probed with a human MRP1 1.8-kb EcoRI cDNA fragment that had been radiolabeled with [-32P]dCTP (specific activity, 3000 Ci mmol–1) (PerkinElmer Life Sciences) by random priming (Invitrogen, Random Primers DNA labeling system) (9).
Confocal Microscopy—HEK293T cells transfected with the MRP1 constructs were seeded at 3.5 x 105 cells per well in a 6-well plate on coverslips coated with 0.1% gelatin in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. Forty-eight hours later, the coverslips were washed with PBS, and cells were fixed with 4% paraformaldehyde and permeabilized by adding 0.2% Triton X-100 in PBS. The coverslips were incubated with MRP1-specific mAb QCRL-3 (diluted 1:2,500 in 0.1% Triton X-100/PBS containing 10 µgml–1 RNase A and 1% bovine serum albumin) for 60 min (41, 42). The coverslips were then incubated with Alexa Fluor 488 anti-mouse IgG (H+L) (Fab')2 fragment in 0.1% Triton X-100/PBS containing 2 µg ml–1 propidium iodide and then placed on slides containing 1 drop of Antifade Solution (Molecular Probes, Inc., Eugene, OR). Cells were examined using a Leica TCS SP2 MS multiphoton system confocal microscope (Leica Microsystems, Heidelberg, Germany) as before (39).
Photolabeling of MRP1 with [3H]LTC4—Membrane proteins were photolabeled with [3H]LTC4 essentially as described (40). Briefly, membrane vesicles prepared from transfected cells (70 µg of protein in a final volume of 50 µl) were incubated with [3H]LTC4 (0.08–0.1 µCi; 200 nM) and 10 mM MgCl2 for 30 min at room temperature and then frozen in liquid nitrogen. Samples were then alternately irradiated at 302 nm for 1 min using a CL-1000 UV cross-linker (DiaMed, Mississauga, Ontario, Canada) and snap-frozen in liquid N2 10 times. Radiolabeled proteins were resolved by SDS-PAGE and processed for autoradiography. After drying, the gel was exposed to X-Omat film for 2 days at –70 °C. Relative levels of photolabeling were estimated by densitometric analysis as before.
Photolabeling of MRP1 by 8-Azido-[-32P]ATP—MRP1-enriched membrane vesicle proteins were photolabeled with 8-azido-[-32P]ATP essentially as described (46, 47). Briefly, membrane vesicles (20 µg) were dispersed in 20 µl of transport buffer containing 5 mM MgCl2 and 5 µM 8-azido-[-32P]ATP (1 µCi). After 5 min on ice, the samples were UV cross-linked at 302 nm in a 96-well plate for a further 8 min, washed twice, and then solubilized in Laemmli buffer and subjected to SDS-PAGE. After drying, the gel was exposed to film for 1 h.
Orthovanadate-induced Trapping of 8-Azido-[-32P]ADP by MRP1— MRP1-enriched membrane vesicles (20 µg of protein) were incubated in transport buffer (20 µl) containing MgCl2 (5 mM), sodium orthovanadate (1 mM), 8-azido-[-32P]ATP (5 µM, 1 µCi) for 15 min at 37 °C. In some cases, BeFx (mixture of 1 mM NaF and 200 µM BeSO4) was added instead of sodium orthovanadate (48). Membrane vesicles prepared from untransfected HEK cells were included as a negative control. Reactions were terminated by the addition of ice-cold Tris-EGTA buffer (50 mM Tris-HCl, pH 7.4, 0.1 mM EGTA, 5 mM MgCl2) and then centrifuged at 21,000 x g for 15 min at 4 °C. Resuspended membrane proteins were transferred to a 96-well plate and exposed to UV light at 302 nm on ice for 8 min. Membrane vesicles were then solubilized in Laemmli buffer and subjected to SDS-PAGE. After drying, the gel was exposed to film for 12–24 h.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
Ala mutants, and levels of ATP-dependent uptake (after normalization to take into account differences in MRP1 protein expression levels) are shown in Fig. 2A. Thus, relative to wild-type MRP1, [3H]LTC4 uptake by the MSD2 mutants P323A and P359A was moderately reduced (by 25%), whereas uptake by the P343A, P448A, P478A, P557A, and P595A TM mutants was substantially reduced (by 55–70%). [3H]LTC4 uptake by the P464A and P600A mutants was comparable with wild-type MRP1 (Fig. 2A).
|
G uptake assays using membrane vesicles prepared from cells expressing the MSD2 Pro mutants were also performed. As shown in Fig. 2B, ATP-dependent uptake levels of this substrate by mutants P323A and P359A relative to wild-type MRP1 were moderately reduced (by 30%), whereas uptake by mutants P343A, P448A, P557A, and P595A was substantially reduced (by 75–90%), and uptake by the P478A mutant was increased 1.5-fold. In contrast, as observed for [3H]LTC4 uptake, [3H]E217G uptake by the P464A and P600A mutants was comparable with wild-type MRP1.
To determine whether the Pro to Ala substitutions in MSD2 affected the transport of other MRP1 substrates, levels of MTX, GSH, and E13SO4 uptake in inside-out membrane vesicles were determined. Relative to wild-type MRP1, [3H]MTX uptake by the P323A and P359A mutants was moderately reduced (by 30–40%), whereas uptake levels of the P343A, P448A, P557A, and P595A mutants were substantially reduced (by 75–90%). In contrast, [3H]MTX uptake by the P464A and P478A mutants was moderately increased by 30% relative to wild-type MRP1, whereas uptake by the P600A mutant was comparable with wild-type MRP1 (Fig. 2C).
GSH transport was measured in the presence of apigenin (30 µM), which previous studies have shown markedly enhances the uptake of this tripeptide by MRP1, thereby allowing more accurate quantitation (13, 47). Relative to wild-type MRP1, GSH uptake by six of the nine MSD2 Pro mutants (P323A, P343A, P448A, P478A, P557A, and P595A) was substantially reduced (60–93%), whereas GSH uptake levels by the P359A, P464A, and P600A mutants were comparable with wild-type MRP1 (Table II).
|
G, and MTX uptake data shown in Fig. 2, are summarized in Table II. Taken together, these results indicate that the MSD2 TM-associated residues Pro323 (TM6), Pro343 (TM6), Pro448 (TM8), Pro478 (TM9), Pro557 (TM10), and Pro595 (TM11) as well as the non-membrane-associated Pro359 (ECL3) play some role in the organic anion transport function of MRP1, whereas Pro464 (ECL4) and Pro600 (TM11) do not appear to be involved.
Expression Levels and Transport Activities of MRP1 Mutants Containing Ala Substitutions of MSD3 Pro Residues—We next replaced the nine Pro residues that are predicted to be in MSD3 of MRP1 with Ala as before (Fig. 3A). Eight of these mutants were expressed in HEK cells at levels comparable with those of wild-type MRP1. The exception was P1113A which was expressed at levels that were 20% those of wild-type MRP1. This finding suggests that Ala substitution of Pro1113 (which is predicted to be located in ECL7 connecting TM14 to TM15) in some way impairs the expression or stability of MRP1 (Fig. 3B). The electrophoretic mobility of the small amounts of Pro1113 mutant protein present was comparable with wild-type MRP1, indicating that processing had occurred.
|
40–50% less than those of wild-type MRP1, whereas LTC4 uptake by the other five MSD3 Pro mutants was similar to wild-type MRP1.
|
G uptake by six of the nine MSD3 mutants (P1003A, P1060A, P1068A, P1088A, P1120A, and P1121A) were moderately reduced (by 30–55%), whereas uptake of this substrate by the P1113A and P1191A mutants was comparable with wild-type MRP1. In marked contrast, uptake of [3H]E217G by the P1150A mutant was 2.2-fold higher than wild-type MRP1 (Fig. 4B).
[3H]MTX uptake by the P1150A mutant was also dramatically increased (6-fold) whereas uptake of this antifolate by the remaining eight MSD3 Pro mutants was either moderately reduced (by 50–60% for P1068A, P1088A, and P1120A) or comparable with wild-type MRP1 (P1003A, P1060A, P1113A, P1121A, and P1191A) (Fig. 4C). To determine whether Ala substitution of Pro1150 also affected the transport of other folic acid derivatives, [3H]leucovorin uptake by the P1150A mutant was assessed. As shown in Fig. 4D, levels of [3H]leucovorin uptake by wild-type and P1150A mutant MRP1 were similar, indicating that the enhanced MTX transport activity caused by the Pro1150 substitution is quite substrate-specific.
The levels of [3H]GSH and [3H]E13SO4 uptake were also determined for the nine MSD3 Pro mutants, and the results from these experiments are summarized together with the LTC4, E217G, and MTX transport data in Table III. Relative to wild-type MRP1, apigenin-stimulated GSH uptake by the P1068A, P1088A, and P1150A mutants was significantly reduced (by 50–70%), whereas uptake by the other six MSD3 Pro mutants was similar to wild-type MRP1 except for P1120A, which was 1.4-fold higher. GSH-stimulated E13SO4 uptake by the P1003A, P1060A, P1088A, P1150A, and P1191A mutants was moderately reduced (by 25–50%), whereas uptake of this sulfated estrogen by the other four MSD3 Pro mutants (P1068A, P1113A, P1120A, and P1121A) was comparable with wild-type MRP1.
|
Expression of MRP1 Mutant Proteins P1113A, P1113A/P1120A, and P1113A/P1121A Is Reduced, but Membrane Localization and MRP1 mRNA Levels Are Comparable with Wild-type MRP1—To explore further the reduced levels of expression caused by Ala substitution of Pro1113, two double mutants were created to determine whether Ala substitution of an additional Pro residue in relatively close proximity to Pro1113 (Pro1120 and Pro1121) might restore MRP1 expression by a compensatory change in structure (P1113A/P1120A and P1113A/P1121A); a third mutant, P1120A/P1121A, was also generated and included in these experiments. Expression levels of these double Pro mutants were determined in membrane proteins isolated from the transfected HEK cells by immunoblotting with mAb QCRL-1 as before (Fig. 5A). The results show that the expression levels of the two Pro1113 double mutants were reduced by more than 80% as was observed for the single P1113A mutant. On the other hand, the third mutant, P1120A/P1121A, was expressed at levels that were at least 60% those of wild-type MRP1. These results indicate that mutation of Pro1113 to Ala is consistently associated with a marked reduction in MRP1 protein expression levels that cannot be compensated for by additional Pro substitutions nearby. Despite the decreased expression levels of the three P1113A containing MRP1 mutants, indirect immunofluorescence confocal microscopy of intact transfected HEK293T cells showed that all of the Pro1113 mutants were correctly routed to the plasma membrane as was the P1120A/P1121A mutant (Fig. 6).
|
|
Ala substitution is caused by some event that occurs after transcription.
Kinetic Parameters of [3H]LTC4 and [3H]E217G Uptake by MRP1 Pro Mutants—MSD2 TM mutants P343A, P448A, P478A, P557A, and P595A, MSD3 CL7 mutant P1150A, and TM14 mutant P1088A whose [3H]LTC4 or [3H]E217G transport properties were substantially altered relative to wild-type MRP1 were further characterized by kinetic analyses (Table IV). It was not technically feasible to measure kinetic parameters of LTC4 uptake of all mutants in a single experiment; therefore, results from two separate experiments are reported. The differences in the apparent Km(LTC4) values obtained for wild-type MRP1 in the two sets of experiments are within the range reported in the literature. Vmax values that were normalized to take into account the relatively small differences in expression levels of the various MRP1 mutants compared with the wild-type protein are also shown.
|
The apparent Km(LTC4) values for the MSD3 mutants P1088A and P1150A were also somewhat decreased (55 and 40 nM, respectively) compared with wild-type MRP1 (72 nM), as were their Vmax(LTC4) values (by 50%), yielding overall LTC4 transport efficiencies (Vmax/Km) for these two mutants of 4.15 and 4.55, respectively, compared with a value of 5.86 for wild-type MRP1.
The apparent Km(E217G) for the MSD3 P1150A mutant, which showed significantly enhanced E217G transport, was decreased 4-fold (0.24 versus 1.02 µM for wild-type MRP1). In contrast, the apparent Km(E217G) value for P1088A was increased 1.4-fold (1.39 µM). On the other hand, the Vmax(E217G) values of P1150A and P1088A (160 and 165 pmol mg–1 min–1, respectively) were comparable with wild-type MRP1 (176 pmol mg–1 min–1). Thus, the overall E217G transport efficiency of P1150A was enhanced 4-fold whereas that of P1088A was reduced by 35%, largely because of changes in the apparent uptake affinity of the mutant proteins for this conjugated estrogen.
Photolabeling of Wild-type and Pro Mutant MRP1 Proteins with [3H]LTC4—To investigate further whether the reduced [3H]LTC4 transport activity of the TM mutants P343A, P448A, P478A, P557A, P595A, and P1088A and the CL7 mutant P1150A was associated with a decrease in substrate binding, photolabeling experiments were carried out with this intrinsically photoactivable arachidonic acid derivative. As shown in Fig. 7, a radiolabeled protein band of 190 kDa was readily detectable in [3H]LTC4 photolabeled membrane vesicles prepared from cells expressing all of the mutants except P595A. Thus the P595A mutation completely abrogates photolabeling by LTC4, whereas the other mutations caused no or only a moderate reduction in photolabeling compared with wild-type MRP1.
|
G Transport by MRP1-P1150A—To explore further the mechanism of enhanced E217G transport by the P1150A mutant, the ATP dependence of E217G uptake by membrane vesicles expressing this mutant MRP1 was investigated by measuring initial rates of uptake of this conjugated organic anion at different ATP concentrations (Fig. 8A). The ATP dependence of LTC4 transport by P1150A was also measured for comparison (Fig. 8B). In the case of E217G transport (Fig. 8A), the Km(ATP) and Vmax values were 108 µM and 295 pmol mg–1 min–1 and 483 µM and 142 pmol mg–1 min–1 for P1150A and wild-type MRP1, respectively. In the case of LTC4 transport (Fig. 8B), the Km(ATP) and Vmax values were 97 µM and 90 pmol mg–1 min–1 and 109 µM and 170 pmol mg–1 min–1 for P1150A and wild-type MRP1, respectively. Thus the MRP1-P1150A mutant showed a significantly lower (nearly 5-fold) Km (higher apparent affinity) for ATP than wild-type MRP1 when supporting transport of E217G. In contrast, when supporting LTC4 transport, the Km(ATP) of P1150A and wild-type MRP1 were similar.
|
-32P]ATP Binding and Orthovanadate-induced Trapping of 8-Azido-[-32P]ADP by MRP1-P1150A—To investigate further the ATP binding and hydrolysis properties of the MRP1-P1150A mutant, we used a 32P-labeled photoactivable nucleotide analog to evaluate its relative ability to bind ATP. As shown in Fig. 9A, at 4 °C 8-azido-[-32P]ATP photolabeled a 190-kDa protein corresponding to MRP1 in membrane proteins prepared from both wild-type and P1150A mutant MRP1-transfected cells. The intensity of labeling of the wild-type and mutant MRP1 proteins was similar, indicating a comparable level of ATP binding by the two proteins. In contrast, orthovanadate-induced trapping of 8-azido-[-32P]ADP at 37 °C by the P1150A mutant was markedly reduced compared with wild-type MRP1 (Fig. 9B). This indicates that although mutation of Pro1150 does not appear to affect binding of ATP, it may reduce the ability of MRP1 to hydrolyze ATP. This finding was confirmed by using an alternate nucleotide trapping agent, BeFx (48).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
280 amino acids that comprise MSD1 and the CL connecting it to MSD2. Several of the MSD1 Pro mutations caused a substantial reduction in MRP1 expression, suggesting that MSD1 is important for the stable expression of MRP1 in mammalian cell membranes (39). In the present study, we found that the MRP1 mutant containing an Ala substitution of Pro1113 in MSD3 was also poorly expressed. According to most topological models of MRP1, Pro1113 is located in the relatively short ECL that connects TM14 to TM15. Consequently, as suggested previously for Pro104 (39), Pro1113 may facilitate insertion of TM14 and TM15 into the membrane during biosynthesis. The absence of Pro104 or Pro1113 may cause misfolding and destabilization of MRP1, leading to its enhanced degradation and low expression levels. Consistent with this conclusion is the observation that incubation of P1113A-transfected HEK293T cells at a reduced temperature (30 °C) significantly enhances the levels of plasma membrane expression of the P1113A mutant protein (results not shown). With relatively few amino acids predicted to be in the ECL between TM14 and TM15, it may be that these two helices form a closely interacting structure. However, the structural change introduced by substitution of Pro1113 in MRP1 could not be compensated for by introducing another structural change by replacing nearby Pro residues in TM15 (Pro1120 and Pro1121) because the double mutants P1113A/P1120A and P1113A/P1121A were also poorly expressed. Nevertheless, despite their low levels of expression, the Pro1113 mutants could all be detected at the plasma membrane of the transfected cells. Given the very substantial structural change that would be introduced by simultaneous substitution of two adjacent Pro residues (33, 49), it is somewhat surprising that the TM15 double Pro mutant P1120A/P1121A was expressed at levels comparable with wild-type MRP1, and its transport activity was also similar to the wild-type protein (results not shown). However, these Pro residues are predicted to be located at the NH2-proximal end of TM15 in the outer leaflet of the membrane bilayer which may make it less likely for any structural changes introduced into TM15 by their substitution to disrupt the interhelical interactions that are presumed to occur between TM15 and other TM helices of MRP1. In addition, according to our current model of MSD2 and MSD3 of MRP1 (which is based on the crystal structures of two bacterial MsbA transporters and an energy minimized P-glycoprotein simulation (50)), amino acids in this region of TM15 are not predicted to be directly in the substrate translocation pathway.
In our previous study, we found that single Ala substitutions of MSD1 Pro residues had little effect on the organic anion transport activity of MRP1 (39). In contrast, Ala substitutions of the Pro residues of MSD2 had variable, and in some cases profound, effects on MRP1 transport activity. For example, Pro Ala substitutions at positions 343 (TM6), 448 (TM8), 557 (TM10), and 595 (TM11) caused a marked decrease in, or eliminated, transport of at least four of the five organic anions tested, whereas Ala substitution of MSD2 Pro residues at positions 359 (in ECL3), 464 (in ECL4), and 600 (at the TM11-cytosol interface) had little or no effect. Between these two extremes of transport activity were mutants P323A and P478A. Replacement of Pro323 at the TM6 membrane-cytosol interface with Ala caused a significant and relatively specific loss of GSH transport, whereas Ala substitution of Pro478 (TM9) was associated with a substantial loss of LTC4 and GSH transport on the one hand, and a moderate increase in E217G, MTX, and E13SO4 transport on the other. The partial and substrate-selective alterations in transport activity observed for the P323A and P478A mutants provide evidence that although these Pro residues are not essential for function, they may be involved in some substrate-specific interactions with the transporter.
The almost total loss of MRP1 transport activity of the MSD2 TM mutants P343A, P448A, P557A, and P595A does not distinguish between a structural or dynamic role for these Pro residues but does confirm that these highly conserved residues are required for transport. However, all but one of these mutants (P595A) could still be labeled with [3H]LTC4, indicating that if the Pro substitutions caused any structural changes, they did not disrupt the binding site for this substrate. Predicted to be located well within the membrane bilayer, Pro343 (TM6), Pro448 (TM8), Pro557 (TM10), and Pro595 (TM11) may introduce a kink into the -helix between the segments preceding and following the Pro residue. Such so-called Pro-kink motifs can occur because of the inability of the Pro residue to form intrahelical hydrogen bonds. This lack of hydrogen bond donor activity in turn can result in a greater flexibility of a TM helix, which may allow it to respond conformationally to specific interactions with substrate or to function in the propagation of conformational changes from one protein domain to another (33). Thus, a Pro residue could act as a "switch" between the different conformations adopted by the protein in different steps of its catalytic or transport cycle. Additionally or alternatively, Pro-kinks may contribute to a narrowing of the ends of a substrate translocation pathway and, in this way, contribute to the architecture of the "chamber" through which substrates are effluxed out of the cell. However, it has been established that the number of energetically favored conformations available to a kinked -helix and its preferred bend angle can be altered by the residues surrounding the Pro (either within the same helix and/or in neighboring helices), and this alteration can be reflected in the functional properties of the mutants (37, 51, 52). Whether or not this is the case with any or all of the Pro-kink motifs in the MSD2 TM helices of MRP1 requires further investigation. According to our current MRP1 model based on the crystal structures of the bacterial lipid transporter MsbA (50), Pro448 (TM8), Pro557 (TM10), and Pro595 (TM11) are located in the substrate translocation pathway of MRP1. Pro448 is predicted to introduce a modest kink in TM8, whereas Pro557 creates a very significant kink in TM10, likely influenced by the Thr residue that precedes it (37). It may be that the loss of these Pro residues modifies the architecture of the MRP1 substrate translocation pathway directly by changing the geometry of the TMs in which they are located and/or indirectly by causing a shift in overall TM helix packing.
Finally, the functional importance of the four MSD2 TM Pro residues at positions 343, 448, 557, and 595 is consistent with their location in the same TM helices (TM6, TM8, TM10, and TM11, respectively) that contain other amino acids shown in previous studies to be important for MRP1 drug binding and/or transport activity (24, 26, 50, 53, 54). However, of these four MSD2 TM Pro mutants that display a global and substantial loss of transport activity, only one of them (P595A) showed a complete loss of photolabeling by LTC4 and a major decrease in uptake affinity for this substrate. Similarly, non-conservative substitutions of the nearby Phe594 (50) and Arg593 (54)2 residues in TM11, which also caused a global loss of MRP1 transport activity, eliminated LTC4 photolabeling. These observations point to a particularly critical role for this region of TM11, together with charged residues in TM6 (26), in the high affinity binding of LTC4 and likely other substrates as well.
In contrast to the MSD2 mutants, none of the nine MSD3 Pro mutants examined showed a comparable global decrease in transport activity, although Ala substitution of Pro1088 (at the membrane cytosol interface of TM14) and to a lesser extent Pro1003 (in ECL6) did cause a significant reduction in the transport of all five organic anion substrates tested. The significantly reduced transport activity of the P1088A mutant, together with our previous findings that non-conservative mutations of polar residues in the same region of TM14 alter MRP1 transport activity, confirm that the TM14-cytosol interface is a particularly important region for MRP1 activity (25, 29).3 Other MSD3 Pro mutants showed more substrate-selective changes in transport activity. For example, Ala substitution of Pro1120 (in TM15) caused a moderate and selective loss of E217G and MTX uptake. In contrast, the relative transport activities of the single Pro1113 (in ECL7) and Pro1121 (in TM15) mutants were comparable with those of wild-type MRP1 except for a moderate and selective loss of E217G transport activity observed for the latter mutant.
Mutation of Pro residues in the predicted CLs of MSD3 also resulted in substrate-selective changes in transport activity. These intracellular domains (and the comparable domains in MSD2) are extensively -helical but are presumably not part of the substrate translocation pathway through the membrane bilayer. Ala substitution of Pro1060 (in CL6) caused a moderate and selective loss of E217G transport, and Ala substitution of Pro1068 (in CL6) caused a moderate overall loss of organic anion transport except for E13SO4 transport, which remained comparable with wild-type MRP1. In current models of MRP1 (29, 50), CL6 contains three -helices, with the central helix flanked by Pro1060 and Pro1068. The analogous helix in the bacterial transporter MsbA appears to be in contact with the NBD, and it has been proposed that it may be important for transduction of conformational changes that occur between the NBD and MSD following nucleotide binding and hydrolysis (55). That structural changes are introduced into CL6 by replacement of Pro1060 or Pro1068 with Ala is evident from the loss of immunoreactivity of the P1060A and P1068A mutants with mAb MRPm5 which maps to amino acids 1063–1072.4 However, the moderate changes in the transport properties of the P1060A and P1068A mutants, and the fact that mAb MRPm5 does not inhibit MRP1 transport activity, suggest that considerable mobility of this loop can be accommodated without loss of MRP1 activity.4
CL7 is also predicted to be extensively -helical and, like CL6, contains two Pro residues. One of these, Pro1191, is located at the COOH-proximal end of CL7, and Ala substitution of this residue caused only a modest and selective loss of E13SO4 uptake. This suggests that the structural change introduced into this region of CL7 by mutation of Pro1191 does not significantly alter binding or transport of organic anions by MRP1.
The second, more NH2-proximal CL7 Pro residue at position 1150, is located at the end of a putative -helix that begins with TM15 and is predicted to extend well into the cytoplasm. The P1150A mutant showed a moderate but significant decrease in LTC4, GSH, and E13SO4 transport but, remarkably, also exhibited a substantial increase in E217G and MTX transport. The decrease in LTC4 transport caused by the P1150A mutation was associated with a 2.5-fold decrease in Vmax, whereas the increase in E217G transport was associated with a 4-fold increase in uptake affinity (decreased Km). These observations indicate that the structural change introduced by Ala substitution of Pro1150 in some way selectively enhances the apparent uptake affinity of MRP1 for the glucuronidated estrogen while at the same time diminishes the efficiency of LTC4 transport.
Intriguingly, the P1150A mutation also affected the apparent ATP dependence of organic anion transport in a substrate-selective way. Thus, the P1150A mutation caused a 5-fold increase in apparent affinity (decreased Km) for ATP and 2-fold increase in Vmax when supporting E217G transport. However, for LTC4 transport, the Km(ATP) of P1150A was unchanged, and a 2-fold decrease in Vmax was observed. Furthermore, whereas the P1150A mutant appeared to bind azido-ATP as well as wild-type MRP1, it showed substantially reduced vanadate- and BeF-induced trapping of azido-ADP. We have shown previously that at 4 °C, NBD1 is predominantly photolabeled by azido-ATP, whereas under trapping conditions at 37 °C, the NBD2 is much more strongly labeled by azido-ADP than NBD1. Thus, our present observations suggest that the CL7 Pro1150 mutation somehow compromises the ability of NBD2 to hydrolyze ATP and/or trap ADP in the presence of vanadate and BeF. This could contribute, at least in part, to the moderately reduced ability of the P1150A mutant to transport LTC4, because previous studies have indicated that ATP binding/hydrolysis at NBD2 plays a dominant role in the transport of this substrate (46, 56). However, it is not immediately obvious how reduced ATP binding/hydrolysis activity at NBD2 can be reconciled with the enhanced ability of the P1150A mutant to transport E217G and MTX. One possibility is that ADP release from NBD2 of the P1150A mutant after hydrolysis is so rapid that the efficiency of vanadate-induced trapping of ADP is reduced, but this cannot explain the substrate-specific changes observed. A more likely explanation is that the Pro1150 mutation has had a differential effect on the coupling of ATP binding and hydrolysis to the conformational changes involved in translocating LTC4 compared with E217G. Further studies are needed to distinguish among these and other possibilities.
MTX transport by the P1150A mutant was increased to an even greater extent than E217G transport. Interestingly, however, transport of the related folate analog, leucovorin, remained comparable with that of wild-type MRP1. The chemical structures of leucovorin and MTX are very similar, differing only in the nature of three functional groups located in different regions of the molecules (Fig. 4D). These structural differences are certain to affect the H-bonding capabilities of the two drugs, but whether or not they are related to the differing ability of the P1150A mutant to transport these two substrates awaits the availability of a more refined structure of MRP1.
Pro residues are significantly more
) and the P1150A mutant (
) were measured for 1 min in the presence of 10 different concentrations of ATP (2 µM to 4 mM). The results shown are the means (±S.D.) of triplicate determinations in a single experiment. Similar results were obtained in a second experiment with an independently prepared batch of vesicles from a second transfection.