No job name

Biochemistry 2001, 40, 4323-4331
P-Glycoprotein-Mediated Colchicine Resistance in Different Cell Lines Correlates with the Effects of Colchicine on P-Glycoprotein Conformation† Todd E. Druley,‡ Wilfred D. Stein,§ Adam Ruth,‡ and Igor B. Roninson*,‡ Department of Molecular Genetics, UniVersity of Illinois at Chicago, Chicago, Illinois 60607, and Department of Biological Chemistry, Hebrew UniVersity, Jerusalem, Israel 91904 ReceiVed June 15, 2000; ReVised Manuscript ReceiVed NoVember 7, 2000 ABSTRACT: The multidrug transporter P-glycoprotein (Pgp) is an ATPase efflux pump for multiple cytotoxicagents, including vinblastine and colchicine. We have found that resistance to vinblastine but not tocolchicine in cell lines derived from different types of tissues and expressing the wild-type human Pgpcorrelates with the Pgp density. Vinblastine induces a conformational change in Pgp, evidenced by increasedreactivity with a conformation-sensitive monoclonal antibody UIC2, in all the tested cell lines. In contrast,colchicine increases the UIC2 reactivity in only some of the cell lines. In those lines where colchicinealone did not affect UIC2 reactivity, this drug was, however, able to reverse the vinblastine-induced increasein UIC2 reactivity. The magnitude of the increase in UIC2 reactivity in the presence of saturatingconcentrations of colchicine correlates with the relative ability of Pgp to confer colchicine resistance indifferent cell lines, suggesting the existence of some cell-specific factors that have a coordinate effect onthe ability of colchicine to induce conformational transitions and to be transported by Pgp. Colchicine,like vinblastine, reverses the decrease in UIC2 reactivity produced by nonhydrolyzable nucleotides, butunlike vinblastine, it does not reverse the effect of ATP at a high concentration. Colchicine, however,decreases the Hill number for the effect of ATP on the UIC2 reactivity from 2 to 1. Colchicine increasesthe UIC2 reactivity and reverses the effect of ATP in ATPase-deficient Pgp mutants, but not in the wild-type Pgp expressed in the same cellular background, suggesting that ATP hydrolysis counteracts the effectsof colchicine on the Pgp conformation.
The multidrug transporter Pgp1 is an ATPase efflux pump that affect drug transport by Pgp have been identified (8), it for multiple cytotoxic agents, responsible for the best-known is not known whether external factors may also affect the form of multidrug resistance in tumor cells (1, 2). Pgp is a relative efficacy of drug transport by Pgp. Romsicki and 170 kDa glycoprotein consisting of two homologous halves, Sharom (9) have shown that the relative binding affinity of each containing a nucleotide-binding domain with NBS- different drugs for purified Pgp in lipid mixtures was affected carrying consensus Walker A and Walker B sequence motifs, by the lipid composition, suggesting that Pgp-drug interac- characteristic of the ATP-binding cassette (ABC) family of tions may also vary among cell types with different transport proteins (3), and a hydrophobic domain with six transmembrane segments. It has been demonstrated throughchemical means (4) or by amino acid substitutions located We have previously shown that the reactivity of Pgp in either the N-terminal or C-terminal Walker A motif (5, encoded by the human MDR1 gene with a conformation- 6) that both NBS must be intact for Pgp to hydrolyze ATP.
specific monoclonal antibody UIC2 is affected by different The presence of Pgp transport substrates has been shown to Pgp ligands (10, 11). Specifically, the UIC2 reactivity is increase the rate of ATP hydrolysis by Pgp (5, 7). A mutation decreased by different Pgp-binding nucleotides, whereas Pgp that alters the relative ability of Pgp to confer resistance to substrates, such as vinblastine, reverse this effect of nucle- different drugs was also found to change the ability of such otides and increase the UIC2 reactivity. While most of the drugs to stimulate ATP hydrolysis (5). While many mutations Pgp-transported drugs were found to increase the UIC2reactivity in intact Pgp-expressing cells, three of these † This work was supported by NIH Grant R37CA40333 (I.B.R.).
substrates (colchicine, etoposide, and puromycin) failed to W.D.S. was the recipient of a Yamagiwa-Yoshida Memorial Interna- increase the UIC2 reactivity in the original survey (10). In tional Cancer Study grant administered by the International Union the study presented here, we have found that colchicine can * To whom correspondence should be addressed: Department of increase the UIC2 reactivity in some but not all Pgp- Molecular Genetics (M/C 669), University of Illinois at Chicago, 900 expressing cell lines, and that the ability of colchicine to South Ashland Ave., Chicago, IL 60607-7170. E-mail: roninson@uic.edu.
increase the UIC2 reactivity correlates with its relative ability to be transported by Pgp in different cells. We have also 1 Abbreviations: Pgp, P-glycoprotein; FACS, fluorescence-activated used the UIC2 reactivity shift assay to characterize the effects cell sorter; ABC, ATP-binding cassette; NBS, nucleotide-binding site- of colchicine in the presence of vinblastine, ATP, or (s); AMP-PNP, 5′-adenylylimidodiphosphate; MIANS, 2-(4-maleimi-doanilino)naphthalene-6-sulfonic acid.
nonhydrolyzable nucleotides. The results of these assays 4324 Biochemistry, Vol. 40, No. 14, 2001 indicate that ATP hydrolysis by Pgp counteracts the effects sites on 1 × 106 cells. Following addition of the primary of colchicine on the Pgp conformation.
antibody, the reaction mixture was held at 37 °C for anadditional 30 min. Primary antibody reactions were stopped EXPERIMENTAL PROCEDURES
by the addition of 5 mL of ice-cold PBS and the cells thencentrifuged for 5 min at 4 °C and 1500 rpm. Pellets were Materials. ATP, AMP-PNP, vinblastine, colchicine, and resuspended in 100 µL of PBS and 1% BSA buffer propidium iodide were from Sigma. Staphylococcus aureus containing 25 µg/mL goat anti-mouse FITC-conjugated -toxin was purchased from List Biological Labs. The goat secondary antibody. The reaction mixtures were left on ice, anti-mouse IgG2a fluorescein isothiocyanate (FITC)-conju- to prevent the function of Pgp, for 30 min before the reactions gated secondary antibody was obtained from Caltag Labo- were stopped by the addition of 5 mL of ice-cold PBS buffer ratories. MRK16 monoclonal antibody was generously and the mixtures centrifuged for 5 min at 4 °C and 1500 provided by T. Tsuruo (University of Tokyo, Tokyo, Japan).
rpm. Immediately prior to FACS analysis, each pellet was The isolation and preparation of the monoclonal antibody resuspended in 350-500 µL of ice-cold PBS and 1% BSA UIC2 have been previously described (12).
buffer containing 1 µg/mL propidium iodide (PI) and left Cell Lines. LMtk- murine fibroblast cell lines, KK-H, KK- on ice. Two-color cytofluorometric analysis was performed L, KM-H, MK-H, and MM, were derived after transfection by acquiring at least 10 000 individual events using a Becton with either the wild-type (KK) or mutant (KM, MK, and Dickinson FacSort flow cytometer. Flow cytometric data MM) forms of human MDR1 cDNA, followed by vinblastine were analyzed by using the Becton Dickinson Information selection or (in the case of MM) by FACS sorting for the expression of human Pgp (ref 10 and unpublished data). TheMK, KM, and MM mutants contain amino acid substitutions Data Analysis. We determined the kinetic parameters from at either one (KM or MK) or both (MM) conserved lysine the data using the SigmaPlot program. This gave us (1) the residues in the Walker A motifs of the N-terminal or maximum or minimum fluorescence (Fmax or Fmin, respec- C-terminal NBS, K433M and K1076M, respectively.
tively) at asymptotically high or low levels of vinblastine,nucleotide, or vanadate, (2) the concentration of vinblastine, HT1080-MDR1 cells constitute a population of HT1080 nucleotide, or vanadate at which half of the maximal change human fibrosarcoma cells infected with a recombinant retroviral vector carrying human MDR1 cDNA, with Pgp- m), and (3) the Hill number, n. The best-fit regression through the respective data points was positive cells isolated with the FACS (13). 3T3-MDR1 cells determined using the appropriate binding isotherm, as were derived without cytotoxic selection from mouse NIH follows. We used eq 1 below for cases where the fluores- 3T3 fibroblasts as described in ref 14. The KB-GRC1 cell cence signal, F, increases with B, the concentration of ligand, line was derived following transfection of KB-3-1 cells with and eq 2 for those cases in which the fluorescence decreases, human MDR1 cDNA and a single step of colchicine selection The K562/i-S9 cell line was derived from human K562 leukemia cells by infection with a recombinant retrovirus carrying the human MDR1 cDNA followed by subcloning - F )K n]/(K n + Bn) (without cytotoxic selection) and immunostaining for Pgp (16). The multidrug-resistant CEM/VLB-100 cell line wasderived from human T-lymphoblastoid CEM leukemia cells S. aureus R-Toxin Permeabilization. The concentration and by multistep selection with vinblastine (17).
time course of S. aureus R-toxin necessary to yield an All attached cell lines were maintained in 15 cm tissue approximate 50% distribution between PI-positive and PI- culture plates in Dulbecco’s modified Eagle’s medium negative staining cells was determined for each cell line in containing 10% fetal bovine serum, 1% glutamine, and a preliminary experiments. An R-toxin concentration was 1% penicillin/streptomycin mixture. Leukemic cells that grow chosen that was effective in 15-30 min at 37 °C. The in suspension were maintained in 75 cm2 flasks in RPMI approximate percentage of permeabilization was checked 1640 medium containing 10% fetal bovine serum, 1% periodically under a light microscope via trypan blue staining glutamate, and a 1% penicillin/streptomycin mixture.
and counting the percentage of blue cells per high power Fluorescence-ActiVated Cell Sorting Assays. All primary field. Permeabilization was performed in PBS and 1% BSA antibody staining reactions were carried out in a final volume buffer at a final volume of 100 µL containing 1 × 106 cells of 200 µL containing 1 × 106 cells/reaction mixture in at 37 °C. The reaction was stopped by the addition of 30 phosphate-buffered saline (PBS) with 1% bovine serum mL of 37 °C PBS, and the cells were centrifuged for 5 min albumin (BSA). Cells were reacted with their respective at 1500 rpm and room temperature. Pellets were resuspended nucleotides and/or drugs for 10-15 min at 37 °C prior to in PBS containing 10 mM MgCl2 with or without the the addition of the primary antibody. UIC2 was aliquoted, indicated nucleotide at 37 °C for 15 min, and then each heated at 48 °C for 24 h in a thermal cycler, and stored at 4 reaction proceeded through the FACS assay procedure as °C prior to use. This treatment did not affect the reactivity of the antibody to Pgp in the presence of substrate, but Cytotoxicity Assays. For assays involving adherent cell decreased the reactivity of the antibody to Pgp in the absence lines, 250 cells were plated in triplicate in a 3 cm plate of substrate, thus increasing the sensitivity of the assay (our (Falcon) in 2-3 mL of drug-free Dulbecco’s modified unpublished data). The primary antibody concentration that Eagle’s medium containing 10% fetal bovine serum, 1% was used (15 µg/mL) had been previously determined to be glutamine, and a 1% penicillin/streptomycin mixture. After a saturating antibody concentration for all available binding 24 h, the drug-free medium was aspirated and replaced with Colchicine Effects on P-Glycoprotein Conformation Biochemistry, Vol. 40, No. 14, 2001 4325 FIGURE 1: Effects of different concentrations of colchicine on UIC2reactivity in K562/i-S9 and 3T3-MDR1 cells. Intact K562/i-S9 (b)and 3T3-MDR1 (O) cells were incubated at the indicated concen-trations of colchicine for 10 min at 37 °C prior to the addition ofUIC2 for an additional 30 min at 37 °C.
fresh medium containing the appropriate type and concentra-tion of cytotoxic drug, in a 3 mL volume. After 8 days at 37 °C, the cells were fixed with methanol and stained withcrystal violet (10% w/v in a 10% methanol solution) and FIGURE 2: Effects of colchicine on UIC2 reactivity in different the number of colonies per plate was counted.
cell types. Each histogram depicts the fluorescence intensity of theindicated cell type, all expressing wild-type human Pgp, stained For assays with suspension cell lines, 3 mL of medium with UIC2 in the absence (s) or presence (‚‚‚) of a saturating containing 50 000 cells/mL and the appropriate concentration concentration (5-10 mM) of colchicine.
of cytotoxic drug were plated in a 3 cm plate, in triplicate.
reactivity up to a concentration of 4% and increased this The cells were left to incubate at 37 °C for 5-7 days. After this incubation, cell clumps were disrupted by repeated therefore account for the colchicine-induced increase in UIC2 pipetting and suspended in Isoton II Electrolyte Solution reactivity in our reaction mixtures, which did not contain (Coulter), and the cell number was determined using a The solid line through the white circles in Figure 1 is the best-fit regression as determined using eq 1 in ExperimentalProcedures. From this regression, the Km value for the effect Differential Effects of Colchicine on the UIC2 ReactiVity of colchicine on UIC2 reactivity in 3T3-MDR1 cells was of the Human MDR1 Pgp in Different Cell Lines. Figure 1 1.11 ( 0.14 mM. The Km values determined via UIC2 depicts the effects of colchicine on the UIC2 reactivity of reactivity assays are apparent affinities of the ligand for Pgp Pgp in two multidrug-resistant cell lines. Due to the large and not the intrinsic affinity of the ligand for the protein.
number of reaction conditions and controls necessary to The Hill number (n) for the colchicine-induced increase in determine accurate parameters of substrate interactions with the UIC2 reactivity was 1.22 ( 0.19, which suggests that Pgp, this figure shows the results of a single experiment, the binding of one molecule of colchicine to Pgp is sufficient but is representative of multiple independent experiments to cause an increase in the UIC2 reactivity of 3T3-MDR1 demonstrating similar results (not shown). In agreement with cells. In contrast to colchicine, the Hill number for the our previous report (10), increasing concentrations of colchi- increase in the UIC2 reactivity brought about by vinblastine cine up to 10 mM had no effect on the UIC2 reactivity of was found to be close to 2 (11). These numbers are in the K562/i-S9 cell line, derived from human K562 leukemia agreement with the report that vinblastine has 2 (18, 20) and cells after retroviral transduction of the human MDR1 gene colchicine only 1 binding site in Pgp (19).
(b). In contrast, 10 µM vinblastine increased the UIC2 Figure 2 demonstrates that a high concentration (5-10 reactivity of this cell line ∼4-fold (11). Colchicine, however, mM) of colchicine has differing effects on UIC2 reactivity induced a 3-fold, dose-dependent increase in UIC2 reactivity of Pgp in various other multidrug-resistant cell lines express- in mouse NIH 3T3 cells transduced with a human MDR1- ing the wild-type human MDR1 Pgp. Colchicine had little expressing retrovirus [3T3-MDR1 (O)]. To ensure that this or no effect on the UIC2 reactivity of HeLa-derived HtTA- effect of colchicine was not due to the ethanol in which MDR1 cells and of two cell lines derived from the KB-3-1 colchicine was solubilized, we analyzed the effect of cell line (also a subclone of HeLa) by colchicine selection increasing ethanol concentrations on the UIC2 reactivity of (KB-8-5) or by transfection with the wild-type MDR1 cDNA 3T3-MDR1 cells. Ethanol had no effect on the UIC2 (KB-GRC1). Colchicine produced a small increase in the 4326 Biochemistry, Vol. 40, No. 14, 2001 Table 1: Drug Resistance and Maximal Antibody Reactivity Values for Different Cell Lines Expressing the Wild-Type P-Glycoproteina a The derivations of all cell lines and their drug-sensitive parental lines are described in Experimental Procedures. The LD50 values were determined from cytotoxicity assays. LD50 values of the drug-resistant cell lines are followed by the LD50 values for the drug-sensitive parental lines in parentheses.
The median fluorescence (in arbitrary units) in the presence of saturating concentrations of MRK16 or UIC2 and in the absence or presence of asaturating concentration of either vinblastine or colchicine is listed.
UIC2 reactivity of human MDR1-transfected mouse LMtk- the absence of vinblastine, was used in place of MRK16 cells (KK-L), a stronger increase in a CEM leukemia cell reactivity to represent the Pgp levels (data not shown). The line isolated by multistep selection with vinblastine (CEM/ results depicted in Figure 3A indicate that the Pgp density VLB-100), and an even stronger increase in the 3T3-FACS14 is the primary determinant of vinblastine resistance in cell line, derived from NIH 3T3 cells by transduction with different Pgp-expressing cell lines, in agreement with the the human MDR1 retrovirus, and in human HT1080 fibro- previous study, which was limited to the derivatives of a sarcoma cells transduced with MDR1 (HT1080-MDR1). In single cell line (20). In contrast to vinblastine resistance, contrast to the variable effects of colchicine, saturating however, colchicine resistance showed no correlation with concentrations of vinblastine consistently increased UIC2 Pgp density (r ) 0.12; Figure 3B), indicating that some reactivity to a maximal level that was similar to the reactivity factors in the cellular environment other than the Pgp density of a conformation-insensitive antibody MRK16 (Table 1) in affect the ability of Pgp to transport colchicine.
almost all of these cell lines, with the exception of NIH 3T3 We asked if the ability of Pgp to confer colchicine derivatives where the UIC2 reactivity in the presence of resistance could be determined by its ability to undergo a vinblastine remained lower than that of MRK16 (Table 1 conformational transition in the presence of colchicine. If this were the case, we would expect that relative colchicine The Ability of Pgp To Confer Colchicine Resistance resistance, normalized by relative vinblastine resistance (a Depends on the Cellular EnVironment and Correlates with function of the Pgp density), would show a Michaelis- the Effect of Colchicine on UIC2 ReactiVity. We have Menten dependence on the colchicine-induced increase in previously shown that the levels of vinblastine resistance in UIC2 reactivity. Figure 3C shows a type of Lineweaver- MDR1-transduced NIH 3T3 cell lines correlate with the Burke plot of the inverse values for the ratio of colchicine density of Pgp in the cell membrane (20). We asked if similar resistance to vinblastine resistance relative to the fold increase correlations could be established for vinblastine and colchi- in UIC2 reactivity in the presence of saturating amounts of cine in a comparison of different Pgp-expressing cell types, colchicine. This plot provides a linear regression with r ) and if the ability of Pgp to confer resistance to different drugs 0.97, indicating a highly significant correlation between the could be related to the ability of such drugs to increase UIC2 ability of colchicine to induce a UIC2 reactivity shift and reactivity. For this analysis, we measured the levels of the relative colchicine resistance conferred by Pgp in different vinblastine and colchicine resistance in five different drug- sensitive cell lines and in their multidrug-resistant derivatives, Colchicine Counteracts the Vinblastine-Induced Increase expressing the wild-type human MDR1 Pgp. Table 1 shows in UIC2 ReactiVity. We investigated whether colchicine, the LD50 values for vinblastine and colchicine resistance for which does not alter UIC2 reactivity in many of the tested these multidrug-resistant cell lines and their drug-sensitive cell lines, would affect the reactivity of such cell lines in parents (in parentheses). Table 1 also includes the results of the presence of a saturating concentration of vinblastine.
concurrent FACS assays for immunoreactivity (expressed as Figure 4A shows the results of such an experiment with the the mean fluorescence value) of the multidrug-resistant cell KK-H cell line, which was derived from KK-L cells shown lines with a conformation-insensitive antibody MRK16 in Figure 2 by selection for increased resistance to vinblas- specific for the human MDR1 Pgp, with UIC2 in the absence tine. Cells were reacted with increasing concentrations of or presence of vinblastine or colchicine, and with the UPC10 colchicine, in the presence or absence of 10 µM vinblastine.
isotype control, the latter reflecting autofluorescence and While KK-H cells show a strong increase in UIC2 reactivity nonspecific antibody binding (proportional to the cell in the presence of vinblastine alone [compare the UIC2 reactivity for the starting point of the KK-H-labeled curve Using the data in Table 1, we have plotted the relative with the data point for KK-H cells at the same colchicine resistance to vinblastine (Figure 3A) or colchicine (Figure concentration (4) in the absence of vinblastine], their UIC2 3B) against the relative density of Pgp in the individual cell reactivity is not significantly affected by colchicine alone lines. This latter parameter was determined by dividing the (compare the data points in the absence of vinblastine). In mean MRK16 immunofluorescence by the fluorescence of the presence of 10 µM vinblastine, however, increasing cells reacted with UPC10. The data show a strong correlation concentrations of colchicine steadily reduced the UIC2 between vinblastine resistance and Pgp density, with a reactivity of KK-H cells, bringing it close to the low level correlation coefficient (r) of 0.96 (Figure 3A). Similar results of reactivity observed in the presence of colchicine alone were obtained when UIC2 reactivity, in the presence or in (Figure 4A). The best-fit regression using eq 2 in Experi- Colchicine Effects on P-Glycoprotein Conformation Biochemistry, Vol. 40, No. 14, 2001 4327 FIGURE 3: Correlations of vinblastine and colchicine resistance in different Pgp-expressing cell lines with Pgp density and UIC2 reactivityshift. Resistance and median fluorescence values for each of the five labeled cell lines are listed in Table 1. The solid line in each panelis the best-fit linear regression, and the error bars represent the standard error of the LD50 (n ) 3) as determined using SigmaPlot. In panelA, relative vinblastine (VLB) resistance (fold increase in LD50 relative to that of the parental cell line) for each Pgp-expressing cell line isplotted on the Y axis vs the ratio of median MRK16 reactivity to median fluorescence of cells stained with the UPC10 control, a measureof Pgp density (X axis). In panel B, the relative colchicine resistance is plotted vs the same measure of Pgp density. Panel C is a type ofLineweaver-Burke plot for the ratio of median UIC2 reactivity to median UIC2 reactivity in the presence of a saturating concentration ofCOL (UIC2+COL) vs the ratio of relative VLB resistance to relative COL resistance.
mental Procedures shows that the Km of colchicine in the cine (Figure 4B). Surprisingly, both cell lines carrying single presence of vinblastine in KK-H cells was 6.68 ( 3.01 mM.
NBS mutants of Pgp, KM-H (O) and MK-H (b), demon- The Hill number was 1.15 ( 0.34, suggesting that 1 molecule strated a colchicine-dependent increase in UIC2 reactivity.
of colchicine was sufficient to decrease the UIC2 reactivity.
The magnitude of this increase was higher for KM-H than Essentially the same effects of colchicine in the presence of for MK-H cells. The solid line through the data in Figure a saturating concentration of vinblastine were obtained with 4B is the best-fit regression using eq 1 in Experimental K562/i-S9 cells, except that the Km was 1.96 ( 0.88 mM Procedures. The Hill numbers for the effect of colchicine (data not shown). These results indicate that colchicine binds on KM-H and MK-H cells were close to 1 (0.83 ( 0.50 and to Pgp and affects its conformation, even in the cells where 1.44 ( 0.25, respectively). Since neither regression line colchicine alone does not bring about a change in the UIC2 reaches a saturation plateau for colchicine concentration, the reactivity. Furthermore, we have found that vinblastine, when apparent affinity of colchicine could not be measured from tested up to a concentration of 72 µM, is unable to produce these data. These results indicate that colchicine can increase a significant increase in the UIC2 reactivity of K562/i-S9 the UIC2 reactivity of Pgp mutants deficient in ATP cells in the presence of 5 mM colchicine (data not shown).
hydrolysis, even in the cells where it does not alter the Mutations of the Nucleotide-Binding Sites of Pgp Enhance reactivity of the functional Pgp. We have previously shown the Effects of Colchicine on UIC2 ReactiVity. As described that Pgp mutated in both NBS (MM) has a high UIC2 in the accompanying paper (11), the UIC2 reactivity of Pgp reactivity which is unchanged in the presence of vinblastine is affected by nucleotide binding and debinding. We asked (10). As expected, colchicine also had no effect on the UIC2 if the ability of colchicine to change the UIC2 reactivity reactivity of the MM mutant (not shown).
would be affected by mutations in the NBS of Pgp. Figure As with the wild-type KK-H cell line, increasing concen- 4B compares the effects of increasing concentrations of trations of colchicine decreased the UIC2 reactivity of KM-H colchicine on UIC2 reactivity of LMtk- cell lines transfected and MK-H cells in the presence of 10 µM vinblastine to the with either the wild-type human Pgp (KK-H) or Pgp mutants levels approaching those that are seen with colchicine alone carrying K433M or K1076M substitutions of the essential (Figure 4A). The Hill numbers for this effect of colchicine lysine residues in the Walker A motifs of the N-terminal were close to 1 in all three cell lines (1.15 ( 0.34 for KK- (MK-H) or C-terminal (KM-H) NBS, respectively. These H, 0.99 ( 0.46 for KM-H, and 1.00 ( 0.15 for MK-H cells).
mutants are capable of binding nucleotides but are devoid The apparent affinity for this effect of colchicine, however, of ATPase activity (5), and their UIC2 reactivity is increased was more than 1 order of magnitude higher for the single by vinblastine with the same Hill number (2) as the wild- type Pgp (11). As mentioned above, the UIC2 reactivity of ( 0.04 mM for MK-H cells) than in the wild-type Pgp of the wild-type Pgp in KK-H cells was unaffected by colchi- KK-H cells (6.68 ( 3.01 mM). These results provide 4328 Biochemistry, Vol. 40, No. 14, 2001 FIGURE 4: Effects of colchicine on UIC2 reactivity of the wild-type Pgp (A) and of ATPase-deficient Pgp mutants (B) in the presence ofvinblastine. Both panels depict data for LMtk- cells expressing either wild-type Pgp, KK-H, or single-NBS mutants of Pgp, KM-H, andMK-H. In panel A, cells were incubated in the absence [KK-H (3), KM-H (]), and MK-H (0)] or presence of 10 µM vinblastine (b),individually labeled) for 10 min at 37 °C prior to the addition of the indicated concentration of colchicine. The cells were then stained withUIC2 for an additional 30 min at 37 °C. The solid line through each set of data is the best-fit regression as determined using eq 2. In panelB, KK-H (2), KM-H (O), and MK-H (b) cells were stained at the indicated concentration of colchicine for 10 min at 37 °C prior to theaddition of UIC2 and subsequent incubation for an additional 30 min at 37 °C.
additional evidence that mutations that abolish the ATPase it had in the presence of AMP-PNP. In panel B, colchicine activity of Pgp increase the effect of colchicine on UIC2 reverses the effect of the nonhydrolyzable analogue almost as efficiently as vinblastine. Colchicine was similarly effec- Effects of Different Nucleotides on the UIC2 ReactiVity tive in reversing the effect of ADP (data not shown). In in the Presence of Colchicine. To investigate further the contrast to its effect with nonhydrolyzable nucleotides, relationship between the conformational effects of colchicine colchicine provides a moderate increase in UIC2 reactivity and the nucleotide binding and hydrolysis by Pgp, we in the presence of lower concentrations of ATP, but at the analyzed the effects of different nucleotides on the UIC2 highest ATP concentration (20 mM), the UIC2 reactivity is reactivity in the presence of colchicine, using R-toxin- unaffected by the presence of colchicine (Figure 5A). An permeabilized cells. As described in the accompanying paper important distinction between the decrease in UIC2 reactivity (11), cell permeabilization depletes cells of endogenous provided by ATP alone versus ATP and colchicine is found nucleotide, thereby increasing the UIC2 reactivity to the when comparing the Hill number for each regression line.
maximal level. The addition of ATP, ADP, or nonhydro- While the Hill number for ATP alone was approximately 2 lyzable ATP analogues decreases the UIC2 reactivity of (2.08 ( 0.42), the Hill number for ATP in the presence of permeabilized cells, but this effect is reversed by the addition colchicine was 0.60 ( 0.11 (which was not significantly different from 1). This result suggests that the number of Figure 5 shows the effects of ATP (panel A) or its ATP molecules required to decrease UIC2 reactivity is nonhydrolyzable analogue AMP-PNP (panel B) on the UIC2 decreased in the presence of colchicine.
reactivity of permeabilized KK-H cells in the presence of Similar experiments were carried out to analyze the UIC2 nucleotide alone (b), nucleotide and 20 µM vinblastine (1), reactivity of single-mutant Pgps in permeabilized KM-H and or nucleotide and 10 mM colchicine (9). In agreement with MK-H cells. As shown in Figure 6, not only vinblastine but our previous findings (11), both nucleotides provide a dose- also colchicine was able to reverse the decrease in UIC2 dependent decrease in UIC2 reactivity in the absence of reactivity provided by ATP. This effect of colchicine in single drugs, with the Hill number close to 2 for ATP (2.08 ( 0.42; NBS mutants parallels its ability to increase the UIC2 Figure 5A) and closer to 1 (1.41 ( 0.21; Figure 5B) for reactivity of intact MK-H and KM-H cells (Figure 4B). As AMP-PNP. Also as observed in the previous study (11), the in intact cells, colchicine caused a smaller change in the UIC2 presence of vinblastine (1) reverses the decrease in UIC2 reactivity of MK-H than of KM-H cell line. Colchicine and reactivity provided by either nucleotide alone (Figure 5A,B).
vinblastine were also effective in reversing the effects of The addition of the saturating concentration of colchicine, ADP and AMP-PNP in permeabilized MK-H and KM-H however, had a different effect in the presence of ATP than Colchicine Effects on P-Glycoprotein Conformation Biochemistry, Vol. 40, No. 14, 2001 4329 FIGURE 6: Effects of colchicine on the ATP-induced decrease inUIC2 reactivity in R-toxin-permeabilized MK-H (A) and KM-H FIGURE 5: Effects of colchicine on the decrease in UIC2 reactivity of R-toxin-permeabilized cells brought about by ATP (A) or AMP- PNP (B). Both panels depict KK-H cells that were permeabilized incubated in the presence of ATP and 10 mM colchicine, and with the R-toxin prior to the addition of nucleotide and/or drug. In 1) depict data for cells incubated in the presence of ATP b) depict data for cells incubated in the presence and 20 µM vinblastine. The solid line through each set of data is of nucleotide alone (ATP in panel A or AMP-PNP in panel B), the best-fit regression as determined using eq 2.
squares (9) depict data for cells incubated in the presence ofnucleotide and 10 mM colchicine, and triangles (1) depict datafor cells incubated in the presence of nucleotide and 20 µM ground depends on the specific Pgp isoforms (21) and on vinblastine. Cells treated with the R-toxin were incubated in the the presence of mutations that alter the substrate specificity presence of nucleotide for 10 min at 37 °C followed by incubation of Pgp (15). In the study presented here, we asked if relative for an additional 10 min at 37 °C after the addition of the resistance to different drugs, conferred by the same wild- appropriate drug. The solid line through each set of data is the best- type human MDR1 Pgp, would be affected by the cell type fit regression as determined using eq 2.
where this Pgp is expressed. We have shown earlier that the DISCUSSION
resistance to vinblastine, one of the best transport substratesof Pgp, correlated with the cell-surface density of Pgp in In the study presented here, we compared the ability of different multidrug-resistant derivatives of the same cell line, Pgp expressed in different cell types to provide resistance suggesting that the Pgp level was the principal determinant to one of its relatively poor transport substrates, colchicine, of vinblastine resistance in such cells (20). We now extended and we have correlated this resistance with the ability of this analysis to compare Pgp-expressing cell lines of different colchicine to induce conformational transitions of Pgp, which tissue types, and we still observed the same correlation can be detected by altered reactivity with a conformation- between vinblastine resistance and Pgp density. In contrast, sensitive antibody UIC2. We also compared how colchicine the relative levels of colchicine resistance varied widely and affects the Pgp conformation in the presence of another did not correlate with the Pgp density in different cell types.
transport substrate, vinblastine, in the presence of ATP or One possible interpretation of this variability was that some nonhydrolyzable nucleotides, and in the wild-type Pgp cell-specific mechanisms of colchicine resistance, unrelated relative to Pgp mutants deficient in ATP hydrolysis. We have to Pgp, determine the final level of colchicine resistance in found that the effects of colchicine on Pgp conformation Pgp-expressing cells. Alternatively, such cell-specific factors depend on the cellular environment and on the ATP could still act through Pgp, by influencing its interactions hydrolysis by Pgp and that these effects correlate with the with colchicine. Analysis of the effects of colchicine on the ability of Pgp to confer colchicine resistance in different cell UIC2 reactivity shift in different cell lines supports the second interpretation. The levels of colchicine resistance Previous studies indicated that the relative resistance to relative to vinblastine resistance in different cell types showed different Pgp-transported drugs in the same cellular back- an excellent correlation in a Lineweaver-Burke-type plot 4330 Biochemistry, Vol. 40, No. 14, 2001 with the ability of colchicine to induce UIC2-detectable also effective in reversing the effect of nonhydrolyzable conformational transitions. The differences in relative colchi- nucleotides (AMP-PNP and ADP), suggesting that this cine resistance or in the effects of colchicine on the Pgp substrate could also promote nucleotide dissociation. With conformation cannot be ascribed to the selection history of ATP, however, colchicine was able to provide partial reversal the cell lines analyzed in Figure 3. Three of the five lines of the UIC2 reactivity only at lower ATP concentrations.
were isolated without drug selection after retroviral trans- Since ATP, but not the nonhydrolyzable nucleotides, was duction with wild-type human MDR1; one (KB-GRC1) was able to overcome the conformational effect of colchicine, derived by transfection with a wild-type MDR1-expressing this result suggested that ATP hydrolysis could play a role plasmid vector and a single step of small-scale colchicine in limiting the ability of colchicine to increase UIC2 selection to isolate the transfectants, and only one line (CEM/ reactivity. This hypothesis was confirmed in the assays using VLB-100) was derived by multistep selection with vinblas- the KM-H and MK-H cell lines, which carry mutant forms tine, which has a potential for selecting additional mutations.
of Pgp that bind, but do not hydrolyze, nucleotide (5). In The KB-GRC1 and CEM/VLB-100 lines, however, show no contrast to cells carrying wild-type Pgp, colchicine increased increase in the resistance to their selective agent relative to the UIC2 reactivity of intact MK-H and KM-H cells and the Pgp density, which argues against selection-associated overcame the effects of the highest ATP concentrations in mutations in these cell lines. The differences in the colchicine response in different cell lines can therefore be attributed to Another important effect of colchicine on the Pgp-ATP the cellular environment rather than the effects of selection.
interaction is indicated by the finding that the saturating To the best of our knowledge, this is the first demonstration concentration of colchicine (10 mM) altered the Hill number that the cellular environment can alter the relative ability of for the ability of ATP to decrease UIC2 reactivity from 2 to Pgp to confer resistance to different drugs.
1. This effect of colchicine may be interpreted in light of a The finding that colchicine resistance is associated with recent hypothesis by Sauna and Ambudkar (24). They colchicine-induced changes in the UIC2 reactivity was suggest that hydrolysis of 1 ATP molecule is required to surprising, since colchicine (in contrast to vinblastine and transport the Pgp-bound substrate, with an associated change most other Pgp substrates) was previously shown to be in the Pgp conformation [from E1 to E2, as depicted in the unable to bring about a change in UIC2 reactivity in K562/ formal scheme given in Figure 6 of the accompanying paper i-S9 cells (10). Similarly, colchicine was reported to induce (11)], that lowers substrate affinity. The hydrolysis of a no change in the proteolytic profile of Pgp, which is another second ATP molecule would change the Pgp conformation assay for Pgp conformational transitions (22). By examining back to E1 to allow the transporter to bind a new substrate a variety of Pgp-expressing cell lines, we have now found molecule. Colchicine binding appears to shift the equilibrium that colchicine increases the UIC2 reactivity in some but not of these conformational transitions, in the direction from E2 all cell lines. Furthermore, in those lines where colchicine to E1. In the absence of colchicine, such a shift in equilibrium alone did not affect UIC2 reactivity, this drug was able to would otherwise require the hydrolysis of a second ATP reverse the vinblastine-induced increase in UIC2 reactivity.
molecule. Therefore, in the case of colchicine, only one ATP This result is consistent with the ability of colchicine to would be required for the overall conformational transition, inhibit vinblastine-induced changes in the proteolytic profile resulting in the decrease in the Hill number for ATP.
of Pgp (22) and with the finding that high concentrations of What is the nature of the cellular factors that determine colchicine inhibit vinblastine transport by Pgp (23). Our the effects of colchicine on the Pgp conformation and the results suggest that the UIC2 reactivity shift may be used as ability of Pgp to efflux colchicine? It seems likely that Pgp- an assay to identify Pgp-interacting agents by their ability colchicine interactions may be affected by the makeup of to decrease the UIC2 reactivity in the presence of an the lipid bilayer in the different cell lines. Several studies upshifting substrate, such as vinblastine, even if such agents have reported that changes in the lipid composition of the by themselves do not alter the UIC2 reactivity.
plasma membrane alter drug and/or nucleotide binding to In those cell lines where colchicine induces an increase Pgp. Binding of the substrate [3H]azidopine to reconstituted in UIC2 reactivity, the increase follows a conventional Pgp in liposomes was improved when the lipid composition ligand-binding curve, which can be used to determine the of the liposomes was increased for cholesterol, stigmasterol, apparent Km and the Hill number parameters that reflect or ergosterol, in descending order (25). Also, [3H]azidopine Pgp-colchicine interactions. In all the cell lines analyzed photolabeling of Pgp was abolished in the presence of in the study presented here, the Hill number for colchicine, nonionic detergents, but not in the presence of urea or a determined in some lines by the increase in the UIC2 zwitterionic detergent, due presumably to disruption of the reactivity and in other lines by the decrease in UIC2 reactivity lipid bilayer (26). Similarly, alterations in the lipid headgroup in the presence of vinblastine, was approximately 1, sug- and the acyl chain composition or the lipid bilayer alter the gesting that the binding of only 1 molecule of colchicine is apparent affinities of vinblastine, verapamil, and daunorubicin required to alter the UIC2 reactivity. This finding is for Pgp, and also affect ATP binding and hydrolysis (9).
consistent with other work reporting that colchicine has a It is also possible that Pgp interactions may be affected single binding site on Pgp (19).
by some cytoplasmic factors other than lipid composition.
We have analyzed how the effects of colchicine on UIC2 Thus, Zhang and Ling (27) found that cytoplasmic compo- reactivity are affected by nucleotide binding and ATP nents modulate the membrane topology of Pgp molecules hydrolysis by Pgp. As we have previously shown (11), both produced in cell-free translation systems, and suggested that ATP and nonhydrolyzable adenine nucleotides decrease the Pgp expressed in various cell types may have different UIC2 reactivity in permeabilized cells, but vinblastine topological structures. Such topological changes could ac- efficiently reverses this effect of nucleotides. Colchicine was count for the differences in proteolytic profiles of Pgp Colchicine Effects on P-Glycoprotein Conformation Biochemistry, Vol. 40, No. 14, 2001 4331 observed in the presence of different ligands (22, 28) and may provide a plausible explanation for altered UIC2 14. Ruth, A. C., Stein, W. D., Rose, E., and Roninson, I. B. (2001) reactivity. The exact nature of substrate-induced conforma- tional transitions of Pgp and the cellular factors that affect 15. Choi, K., Chen, C., Kriegler, M., and Roninson, I. (1988) Cell these transitions remain a subject for future investigation.
16. Chaudhary, P. M., and Roninson, I. B. (1991) Cell 66, 85- REFERENCES
17. Beck, W. T., Mueller, T. J., and Tanzer, L. R. (1979) Cancer 1. Gottesman, M. M., and Pastan, I. (1993) Annu. ReV. Biochem. 18. van Veen, H. W., Margolles, A., Muller, M., Higgins, C. F., 2. Ambudkar, S. V., Dey, S., Hrycyna, C. A., Ramachandra, M., and Konings, W. N. (2000) EMBO J. 19, 2503-2514.
Pastan, I., and Gottesman, M. M. (1999) Annu. ReV. Phar- 19. Shapiro, A. B., and Ling, V. (1997) Eur. J. Biochem. 250, 3. Higgins, C. F. (1992) Annu. ReV. Cell Biol. 8, 67-113.
20. Choi, K., Frommel, T., Stern, R., Perez, C., Kriegler, M., 4. Urbatsch, I. L., Sankaran, B., Weber, J., and Senior, A. E.
Tsuruo, T., and Roninson, I. (1991) Proc. Natl. Acad. Sci. (1995) J. Biol. Chem. 270, 19383-19390.
5. Muller, M., Bakos, E., Welker, E., Varadi, A., Germann, U.
21. Tang-Wai, D. F., Kajiji, S., DiCapua, F., de Graaf, D., A., Gottesman, M. M., Morse, B. S., Roninson, I. B., and Roninson, I. B., and Gros, P. (1995) Biochemistry 34, 32- Sarkadi, B. (1996) J. Biol. Chem. 271, 1877-1883.
6. Loo, T. W., and Clarke, D. M. (1995) J. Biol. Chem. 270, 22. Wang, G., Pincheira, R., and Zhang, J. T. (1998) Eur. J. 7. Sarkadi, B., Price, E. M., Boucher, R. C., Germann, U. A., and Scarborough, G. A. (1992) J. Biol. Chem. 267, 4854- 23. Horio, M., Gottesman, M. M., and Pastan, I. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 3580-3584.
8. Germann, U. A. (1996) Eur. J. Cancer 32A, 927-944.
24. Sauna, Z. E., and Ambudkar, S. V. (2000) Proc. Natl. Acad. 9. Romsicki, Y., and Sharom, F. J. (1999) Biochemistry 38, 25. Saeki, T., Shimabuku, A. M., Ueda, K., and Komano, T. (1992) 10. Mechetner, E. B., Schott, B., Morse, B. S., Stein, W. D., Biochim. Biophys. Acta 1107, 105-110.
Druley, T., Davis, K. A., Tsuruo, T., and Roninson, I. B.
26. Zordan-Nudo, T., Ling, V., Liu, Z., and Georges, E. (1993) (1997) Proc. Natl. Acad. Sci. U.S.A. 94, 12908-12913.
11. Druley, T. E., Stein, W. D., and Roninson, I. B. (2001) 27. Zhang, J. T., and Ling, V. (1995) Biochemistry 34, 9159- 12. Mechetner, E., and Roninson, I. (1992) Proc. Natl. Acad. Sci. 28. Wang, G., Pincheira, R., Zhang, M., and Zhang, J. T. (1997) 13. Kandel, E. S., Chang, B. D., Schott, B., Shtil, A. A., Gudkov, A. V., and Roninson, I. B. (1997) Somatic Cell Mol. Genet.

Source: http://research.peds.wustl.edu/Portals/Druley_T/Biochem2001b.pdf