the journal of chemistry vol. 269, no. issue of … · printed in w.s.a. functional ... coproteins...

THE JOURNAL OF Blomrcu. CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 10, Issue of March 11, pp. 7243-7248, 1994 Printed in W.S.A.

Functional Consequences of Glycine Mutations in the Predicted Cytoplasmic Loops of P-glycoprotein"

(Received for publication, September 28, 1993, and in revised form, November 24, 1993)

Tip W. Loo and David M. Clarke$ From the Medical Research Council Group in Membrane Biology, Departments of Medicine and Biochemistry, University of Toronto, 1 King's College Circle, lbronto, Ontario M5S lA8, Canada

Site-directed mutagenesis was used to investigate whether glycine residues in the predicted cytoplasmic loops play essential roles in the structure and function of human P-glycoprotein. Mutant cDNh in which codons for each of the 20 glycine residues were changed to valine, were expressed in mouse NIH 3T3 cells and analyzed with respect to their ability to confer resistance to various drugs. Mutation of Gly-251, -268, -269, or -781 yielded mutant proteins which were unable to confer drug resistance in transfected cells. Each of these mutant P-glycoproteins had an apparent mass of 150 kDa, compared with 170 kDa for wild-type P-glycoprotein and the apparent mass was altered by endoglycosidase H digestion. These observations suggest that these mutant proteins were improperly processed so that they were located in the endoplasmic reticulum and were not targeted correctly to the plasma membrane. The in vivo processing of mutants Gly-269 to Val and Gly-781 to Val was temperature-sensitive. When cells expressing these mutants were grown at a lower temperature (26 " 0 , the mature 170-kDa form of P-glycoprotein was the major product. Substitution of glycine with alanine at positions 251, 268, 269, or 781 yielded mutants with structural and functional characteristics similar to wild-type enzyme. Mutation of Gly-141,187,288,812, or 830 to Val, altered the drug resistance profile conferred by P-glycoproteins expressed in transfected cells. All five mutations increased relative resistance to colchicine or adriamycin, without altering relative resistance to vinblastine. These results demonstrate that glycines located in the cytoplasmic loops play important roles in structure and function of P-glycoprotein.

P-glycoprotein is a 170-kDa plasma membrane protein that is responsible for the phenomenon of multidrug resistance in mammalian cells. The protein acts as an energy-dependent pump that extrudes a broad range of hydrophobic cytotoxic drugs from the cell (reviewed by Endicott and Ling (19891, Kane et al . (1990), and Pastan and Gottesman (1993)). Human P-glycoprotein, encoded by the MDR1' gene, consists of 1280 amino acids organized in two tandem repeats of 610 amino acids, joined by a linker region of 60 amino acids. Each repeat consists of an NH2-terminal hydrophobic domain containing six potential transmembrane helices, followed by a hydrophilic do-

* This research was supported by a grant from the Medical Research Council of Canada (to D. M. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

.t Scholar of the Medical Research Council of Canada. To whom cor- respondence and reprint requests should be addressed.

The abbreviations used are: MDR, multidmg resistance; PAGE, polyacrylamide gel electrophoresis; CFTR, cystic fibrosis transmembrane conductance regulator.

main containing a nucleotide-binding site. The amino acid sequence and domain organization of the protein is typical of the ABC (ATP-binding cassette) superfamily of transporters (Hyde et al., 1990).

The substrates of P-glycoprotein, such as colchicine, actinomycin D, and the Vinca alkaloids, are generally hydrophobic compounds with a positive charge a t physiological pH (reviewed by Beck and Qian (1992)). These drugs enter cells by passive diffusion through the plasma membrane. P-glycoprotein presumably interacts with drugs embedded in the membrane (Raviv et a l . , 1990) and the extrusion of the drug is coupled to ATP hydrolysis. It has been demonstrated that P- glycoprotein possesses high levels of ATPase activity that is stimulatable in the presence of drug substrates (Ambudkar et al . , 1992; Sarkadi et al . , 1992; Al-Shawi and Senior, 1993). Stimulation of ATPase activity by drug substrates indicates that P-glycoprotein directly binds and transports drug substrates. Further evidence for direct interaction of drug substrates with P-glycoprotein is that photoaffinity analogues of chemotherapeutic agents such as colchicine (Safa et al., 1989), adriamycin (Busche et al., 1989), and vinblastine (Cornwell et al . , 1986) covalently label the transporter.

The segments and amino acid residues of P-glycoprotein in- volved in substrate binding and transport have not been pre- cisely identified. One potentially important region is the cytoplasmic loops connecting the predicted transmembrane segments. Choi et al . (19881, identified a natural mutation within the first intracellular loop of P-glycoprotein which alters the activity and substrate specificity of the transporter. Cells expressing a mutant MDRl gene that encodes a glycine to valine change at position 185 showed increased resistance to colchicine and decreased resistance to vinblastine when compared to cells expressing wild-type P-glycoprotein. The MDRB gene encodes a protein which does not confer drug resistance although 80% of its amino acid sequence is identical to the MDRl gene product. Currier et al . (1992) constructed a human MDR2NDRl chimera in which the MDRl region between amino acids 140 and 229 (containing the first cytoplasmic loop, TM3 and TM41, was replaced with the corresponding region from the MDRB gene. The product was inactive. MDRl and MDRB differs in only 10 amino acids in the first cytoplasmic loop. Changing only 4 of these amino acids (residues 165, 166, 168, and 169) in MDRl to the corresponding MDRB residues abolished function. These results suggest that the first cytoplasmic loop is important for activity and substrate specificity of P-glycoprotein.

In this study, we have utilized site-directed mutagenesis to make changes to all of the other glycine residues located in the predicted cytoplasmic loops of P-glycoprotein, thereby allowing us to examine their functional roles in P-glycoprotein. We found that changes to 4 of 20 glycine residues resulted in mutant proteins that were defective in processing. Processing of two of these mutant proteins was temperature-sensitive. Changes to

7243

7244 P-glycoprotein Glycine Mutants

v

FIG. 1. Simplified representation of

dues based on structural models for the locations of muta ted glycine resi-

P-glycoprotein proposed b y Juranka et al. (1989) and Gottesman and Pas- tan (1988). Only residues predicted to reside in transmembrane segments, cytoplasmic loops, and the NH,-terminal sequences are shown. The positions of the mutated glycine residues are indicated as solid circles.

five other glycines altered the substrate specificity conferred by P-glycoprotein in transfected cells.

EXPERIMENTAL PROCEDURES Molecular Cloning-Afull-length MDRl cDNA cloned from a human

kidney cortex cDNA library (Bell et al., 1986) and modified to encode the epitope for monoclonal antibody A52 (Zubrzycka-Gaarn et al., 1984) a t the COOH-terminal end of the protein, was inserted into the mammalian expression vector pMT21, as described previously (Loo and Clarke,

would normally end as TKRQ now became TKRAISLISNSCSPEF- 1993a). The sequence at the COOH terminus of P-glycoprotein that

DDLPLAEQREACRRGDPRQ. Oligonucleotide-directed Mutugenesis-Fragments were removed

from pBSMDRlA52 (Bluescript vector (Stratagene) containing the modified MDRl cDNA) and ligated into the polylinker region of the Bluescript vector for site-specific mutagenesis by the method of Kunkel (1985). The fragments used were the following: EcoRI (position -69) to NsiI (position 9921, Hind111 (position 2109) to KpnI (position 2847). Fragments containing the mutation were subcloned back into their original positions in the pMT21MDRlA52 vector. The integrity of the mutated segment of the cDNA insert was checked by sequencing the entire insert including the cloning sites of the DNA to be used for transfection by the dideoxynucleotide chain termination method (Sanger et al., 1977).

Cell Culture-Procedures for transfection of mouse NIH 3T3 cells, followed by selection in the presence of vinblastine or colchicine have been described previously (Loo and Clarke, 1993a). Drug sensitivity was determined by a tetrazolium-based assay for cell viability (Alley et al., 19881, as described previously (Loo and Clarke, 1993a).

PHJAzidopine Photolabeling-Monolayer cells grown on 6-well plates were washed twice with phosphate-buffered saline, and then incubated with 0.5 p~ [3H]azidopine (48 CVmmol, Amersham) in phosphate-buffered saline in the presence or absence of 100 pg/ml vinblastine for 30 min at room temperature in the dark. The plates were cooled on ice and the labeling medium replaced with cold phosphate-buffered saline. The cells were then W-irradiated on ice for 15 min with a 15-watt G15T8 germicidal lamp at a distance of 5 cm. Cells were col- lected by scraping, suspended in 100 pl of phosphate-buffered saline, and lysed by addition of 9 volumes of buffer I (25 m~ Tris-HC1, pH 7.5, 150 m~ NaCl, 1 m~ EDTA, 1% (v/v) Triton X-100, 0.5% (w/v) sodium deoxycholate, 1% aprotinin, 5 pg/ml leupeptin, and 10 p~ p-aminoeth- ylbenzenesulfonyl fluoride). The lysate was centrifuged at 13,000 x g for 5 min and the pellet was discarded. Five pg of monoclonal antibody A52 was then added to the supernatant. After incubation overnight at 4 "C, the immune complexes were precipitated a t 4 "C for 2 h, using Protein A-Sepharose CL-4B (Pharmacia). The immunoprecipitated proteins were washed five times with buffer I and then subjected to SDS-PAGE, followed by fluorography.

Other Procedures-Immunoblotting was performed on proteins separated by SDS-PAGE (Laemmli, 1970) and transferred onto nitrocellulose (Towbin et al., 1979) using monoclonal antibody A52. The level of P-glycoprotein-A52 expression was measured by using a modified enzyme-linked immunosorbent assay and deoxycholate-purified fast- twitch skeletal muscle Ca2+-ATPase (MacLennan, 1970) as a standard, as described previously (Loo and Clarke, 1993a).

RESULTS

Construction and Expression of Mutants-In the structural models for P-glycoprotein proposed by Jurunka et al. (1989) and Gottesman and Pastan (1988), 4 loops connecting transmem-

brane domains, the nucleotide-binding domains, plus the NH2- terminal sequence are located in the cytoplasm (Fig. 1). Twenty glycine residues are located within the 4 loops connecting the transmembrane domains and in the NHz-terminal sequence, with each sequence containing 3 to 5 glycine residues (Fig. 1). To investigate the role of each of these glycine residues in the structure and function of P-glycoprotein, site-directed mutagenesis was used to change the codon for each cytoplasmic glycine to valine. Valine was chosen as a replacement because a change of Gly to Val at position 185 has been shown to alter the substrate specificity of the transporter (Choi et al., 1988).

Mutants were first transiently expressed in HEK 293 cells. This transient expression system allows for rapid (48 h) expression of the mutants, permitting assessment of the structural integrity of the mutant proteins. Fig. 2 shows an immunoblot of wild-type P-glycoprotein-A52 and the glycine mutants expressed in HEK 293 cells. Major structural perturbations were observed for mutants Gly-251 to Val, Gly-268 to Val, Gly-269 to Val, and Gly-781 to Val. Cells transfected with the cDNAs of these mutants expressed an immunoreactive protein of apparent mass 150 kDa as the major product, whereas the major product in cells expressing wild-type P-glycoprotein-A52 had an apparent mass of 170 kDa. A similar 150-kDa product of mutant Pro-709 to Ala, was previously shown to be a partially glycosylated processing intermediate (Loo and Clarke, 1993a). Where processing was perturbed, mutants were constructed to test for the effects of a Gly to Ala change at the same positions since the presence of a smaller side chain may have a less drastic effect on the processing of the enzyme. As shown in Fig. 2, mutants Gly-251 to Ala and Gly-269 to Ala yielded products of apparent mass 170 kDa as the major product, whereas mutants Gly-268 to Ala and Gly-781 to Ala yielded significant levels of both 150- and 170-kDa products. For each of the remaining mutants, the major product in HEK 293 cells was a protein of apparent mass 170 kDa. Each mutant was expressed to approximately the same level as the wild-type P-glycoprotein-A52, suggesting that no major structural perturbation oc- curred by substitution of these glycine residues.

Effect of Mutations on Biological Activity-The ability of wild-type and mutant MDRl cDNAs to confer drug resistance was tested by transfecting the drug-sensitive mouse cell line, NIH 3T3, followed by selection in the presence of colchicine or vinblastine. These two drugs were chosen since there is evidence to suggest that their binding sites do not overlap. Natu- ral mutants (Choi et al., 1988; Gros et al., 1991) and mutations introduced into P-glycoprotein by site-directed mutagenesis (Loo and Clarke, 1993a, 199313; Currier et al., 1992) have been identified which have different effects on the ability of the transporter to confer resistance to colchicine or vinblastine. Accordingly, cells transfected with the mutant cDNAs were selected using both compounds. Control NIH 3T3 cells transfected with vector alone or with the cDNAs of mutants Gly-251 to Val, Gly-268 to Val, Gly-269 to Val, and Gly-781 to Val, did

P-glycoprotein Glycine Mutants 7245

___

FIG. 2. Immunoblots of cell extracts after transfection with mutant constructs. HEK 293 cells were transfected with wild-type and mutant cDNAs. After 48 h, samples of whole cells were separated by 6.5% SDS-PAGE and transferred to nitrocellulose. The samples were probed with monoclonal antibody A52 at a concentration of 2 pg/ml as described under "Experimental Procedures." Bands at positions A and B represent fully glycosylated (170 kDa) and core-glycosylated (150 kDa) P-glycoproteins, respectively.

not yield any colonies in the presence of either colchicine (45 nM) or vinblastine (5 m). The direct transfection and drug selection method used may not have allowed the detection of residual activity that might otherwise have been detected under less stringent drug selection conditions. Therefore, each of the four mutant cDNAs were introduced into NIH 3T3 cells by cotransfection with plasmid pWL-neo (Stratagene), and mass populations of G418-resistant colonies were selected. The drug survival characteristics of these mass populations were tested in medium containing colchicine or vinblastine. These were found to be indistinguishable from that of control NIH 3T3 cells. These inactive mutants were previously found to yield protein products in HEK 293 cells of only 150 kDa. By contrast, cells transfected with either wild-type cDNA or with cDNAs of the remaining mutants yielded drug-resistant colonies in both vinblastine and colchicine. Transfection of NIH 3T3 cells with wild-type P-glycoprotein-A52 cDNA yielded about 5-fold more drug-resistant colonies in the presence of vinblastine compared to colchicine. A similar ratio was obtained when cells were transfected with mutant cDNAs with five exceptions (Table I). Mutants Gly-141, -187, -288, -812, or -830 to Val yielded lower ratios of vinblastine-resistant colonies relative to colchicine- resistant colonies (ratios of 0.8 to 2.1 compared to 5.2 for wild- type enzyme). This was due to an increase in the number of colchicine-resistant colonies rather than to a decrease in the number of vinblastine-resistant colonies. A Gly-185 to Val mutant was also constructed and transfected into NIH 3T3 cells for comparison. A ratio of vinblastinelcolchicine-resistant colonies of only 0.3 was obtained. The large decrease in the ratio of colonies for mutant Gly-185 to Val, was due to a combination of an increased number of colchicine-resistant colonies together with a decreased number of vinblastine-resistant colonies, when compared to the number of colonies obtained with wild- type P-glycoprotein-A52 (ratio of 5.2).

Mutants Gly-251 to Val, Gly-268 to Val, Gly-269 to Val, and Gly-781 to Val did not yield any drug-resistant colonies. Accord- ingly, we transfected cells with cDNAs in which Gly-251, 268, 269, or 781 was changed to alanine rather than valine. In all four cases, drug-resistant colonies were obtained in the presence of vinblastine or colchicine. The ratio of vinblastine to

TABLE I Ratio of number of vinblastine-resistant colonies to

colchicine-resistant colonies after transfection

Mutant Wnblastindcolchicine"

Wild-type 5.2 Gly-5 + Val 5.5 Gly-9 + Val 5.2 Gly-10 -+ Val 6.0 Gly-141 + Val 1.0 Gly-161 -+ Val 3.7 Gly-169 + Val 4.3 Gly-187 + Val 0.8 Gly-251 + Ala 4.8 Gly-268 + Ala 4.7 Gly-269 + Ala 5.3 Gly-288 + Val 2.1 Gly-778 + Val 4.4 Gly-781 + Ala 4.7 Gly-812 + Val 1.6 Gly-827 + Val 6.1 Gly-830 -+ Val 0.9 Gly-881 + Val 4.9 Gly-892 + Val 5.0 Gly-894 + Val 5.3

Ratio of number of colonies obtained a h r transfection and selection in the presence of 5 nM vinblastine and 45 TIM colchicine. Each value is an average of three separate transfections.

colchicine-resistant colonies was similar to that obtained with cells transfected with cDNA of wild-type P-glycoprotein-A52 (Table I).

Drug Survival Characteristics of Cell Clones Expressing Mu- tant P-glycoproteins-The structural and functional characteristics of the stable mutant cell lines were further investigated by studying individual drug-resistant clones. Three clones of each mutant were isolated and detergent extracts from cells of each clone were immunoblotted to test for expression of the mutant P-glycoprotein-A52 protein. In each case, an immunoreactive protein product of apparent mass 170 kDa was present (data not shown), indicating that drug resistance was due to the expression of the mutant MDRl cDNA.

The drug survival characteristics of each of the mutant clones were analyzed in the presence of vinblastine, colchicine, adriamycin, and actinomycin D. The IDs0 (the drug concentration necessary to inhibit cell growth by 50%) for vinblastine was 0.7 2 0.1 m for control NIH 3T3 cells and 20.8 2 3.3 m for clones expressing wild-type P-glycoprotein-A52, representing a 29.7-fold resistance above that of control cells. Cells expressing wild-type P-glycoprotein-A52 also demonstrated increased relative resistance to colchicine (8.5-fold), adriamycin (9.0-fold), and actinomycin D (24.4-fold). There was some variability in the level of expression of P-glycoprotein-A52 in each of the different cell lines. Accordingly, the relative resistance of each cell line was standardized to the level of P-glycoprotein-A52 expression (Fig. 3). A significant difference in the drug resistance profile was observed with five mutants; Gly-141 to Val, Gly-187 to Val, Gly-288 to Val, Gly-812 to Val, and Gly-830 to Val. Cells expressing mutant Gly-141 to Val exhibited a pro- nounced increase in relative resistance to both colchicine (3.0) and adriamycin (2.60). By contrast, the relative resistance to vinblastine (0.82) and actinomycin D (0.90) were similar to cells expressing wild-type enzyme. A similar drug resistance profile was observed with cells expressing mutant Gly-812 to Val, which had increased resistance to colchicine (2.10) and adriamycin (1.901, while the resistance to vinblastine (0.90) and actinomycin D (0.98) were similar to cells expressing wild-type enzyme. Mutants Gly-187 to Val, Gly-288 to Val, and Gly-830 to Val had similar drug resistance profiles in that they had increased resistance to colchicine (3.20,2.10, and 3.30) and adriamycin (2.70, 1.30, and 3.10) and decreased resistance to acti-


FIG. 3. Comparison of relative resis- b*hblrr tances of cells expressing glycine mutations as a function of level of er- E T pression of P-glycoprotein. Relative resistance was determined by dividing the LDho (the drug concentration that in- hibits plating efficiency by 50%) for each G187-.V

stably transfected cell line by the LDSo of the cell line expressing wild-type P-glycoprotein-A52 and then dividing this value by the amount of P-glycoprotein-A52 per 1 x loK cell. The amount of P-glycoprotein- A52 in each cell line was measured as described under "Experimental Proce- dures." The results are presented relative to the wild-type, which is arbitrarily as-

GOIV GlD,V Q l l l - > V Gl61-rV G1W.rV

GZSl->A G.&%>A O?(IC,A GZ&>V GllI)-,V Gl81-,A G.¶12.,v O(U?-,V w u D > V O M I - r V W192-rV GRc(.>V

signed a value of 1.

nomycin D (0.36, 0.54, and 0.29), but resistance to vinblastine was similar to cells expressing wild-type enzyme. The other 14 glycine mutants yielded drug resistance profiles to all four drugs that were similar to clones expressing wild-type P-glycoprotein-A52.

Photoafinity Labeling with Azidopim-Azidopine is a substrate of P-glycoprotein (Tamai and Safa, 1991) and will also label the protein. I t was previously reported by Safa et al. (1990) that mutant Gly-185 to Val exhibited about a 4-fold increase in azidopine labeling relative to wild-type P-glycoprotein. Accordingly, we carried out photoaffinity labeling with t3H]azidopine on each of the glycine mutants which conferred altered drug resistance profiles. Stable mutant cell lines were incubated with [3H]azidopine. Following UV cross-linking, cells were solubilized and the labeled P-glycoprotein were immunoprecipitated with monoclonal antibody A52. Samples containing equivalent amounts of P-glycoprotein-A52 were subjected to SDS-PAGE and fluorography. As shown in Fig. 4, mutants Gly-141 to Val, Gly-187 to Val, Gly-812 to Val, and Gly-830 to Val were all labeled by azidopine in a manner indistinguishable from wild-type enzyme. In all cases, labeling of the mutant proteins was inhibited by vinblastine (100 p ~ ) , indicating that azidopine labeling was specific.

Endoglycosidase H Digestion-Four mutants, Gly-251 to Val, Gly-268 to Val, Gly-269 to Val, and Gly-781 to Val, did not yield any drug-resistant colonies. To obtain stable cell lines expressing these mutant proteins, NIH 3T3 cells were co-transfected with the mutant cDNA with pWLneo (Stratagene). Colonies were selected in the presence of Geneticin (G418) and individual clones expressing mutant P-glycoprotein-A52 were isolated. In each case, a protein product of apparent mass 150 kDa was detected, a result similar to that obtained with transient expression of the mutant cDNAs in HEK 293 cells. As shown in Fig. 5, the 150-kDa products are sensitive to endoglycosidase H, indicating that these products are partially glycosylated intermediates and are probably localized to the endoplasmic reticulum. Drug-resistant cells expressing mutants Gly-268 to Ala and Gly-781 to Ala contained significant amounts of the 150-kDa product, in addition to the 170-kDa mature P-glycoprotein-A52. Only the 150-kDa product was sensitive to endoglycosidase H treatment. Since the oligosaccharide processing which results in endoglycosidase resistance occurs in the Golgi apparatus, the acquisition of endoglycosidase H resistance can be used to identify products that have left the endoplasmic reticulum and have been targeted to the plasma membrane (Hubbard and Ivatt, 1981).

Zhnperature Sensitivity of Processing-The most common mutation in cystic fibrosis is the deletion of Phe-508 (Kerem et al., 1989) in the cystic fibrosis transmembrane conductance regulator (CFTR), which yields a partially glycosylated product that is not delivered to the plasma membrane. It was recently found that the mutant APhe-508 yields a fully glycosylated

1

. -

> 0 c-3

3

2 7 - +

KUB

FIG. 4. Immunoprecipitation of [sHlazidopine-labeled cells. Stable cell lines expressing the mutant P-glycoprotein-AS2 were labeled with I3HIazidopine in the absence ( - ) or presence (+ ) of 1 0 0 pdml vinblastine and immunoprecipitated with monoclonal antibody A52. as described under "Experimental Procedures." The immunoprecipitated proteins were separated by SDS-PACE, followed by fluorography.

I__.. F1c.3. Endoglycosidane H digestion. Whole cells expressing wild-

type P-glycoprotein-A52 or mutant P-glycoprotein-A52 were Rolubilized with buffer containing SO mM sodium citrate, pH 5.5.0.5% (w/v) SDS. 10 mM EDTA, and 1% (v/v) 2-mercaptwthanol. The sample was divided into two equal portions, which were then incubated either with (+ ) or without (-) endoglycosidase Hf (1,000 units. New England Biolabs) for 15 min at room temperature. The digestion waR stopped by addition of an equal volume of buffer containing 0.25 M "is-HCI. pH 6.8, 4'7 (w/vJ SDS. 4% (v/v) 2-mercaptwthanol. and 20% (v/v) glyceml. The reaction mixtures were subjected to SDS-PAGE and transferred to a nitrocellulose membrane for Western blot analysis with monoclonal antibodyA52 as described under "Experimental Procedures." Rands a t ponitions A. R . and C represent fully glycosylated. core-glycosylated. and unglym- sylated forms of the mutant P-glycoprotein, reapectively.

product when the incubation temperature is reduced (Denning et al., 1992). Since CFTR is similar in structure to P-glycopro- k in , we tested whether the defect in processing of mutants Gly-251 to Val, Gly-268 to Val, Gly-269 to Val. and Gly-781 to Val could be corrected by growing the stable cell line expressing these mutant proteins at reduced temperature. We found that lowering the incubation temperature had a dramatic effect on the processing of mutants Gly-269 to Val and Gly-781 to Val. As shown in Fig. 6, the 150-kDa protein (hand B ) is the major product a t 37 "C. When the cells were grown at 26 "C, however, there was a shift in the glycosylation pattern. The major product was now 170 kDa (band A ), corresponding to that of fully glycosylated P-glycoprotein. By contrast, incubation of cells expressing processing defective mutants Gly-251 to Val and Gly- 268 to Val at 26 "C resulted in only a small increase in the level

P-glycoprotein Glycine Mutants 7247

3 % - o b & 5 % zi 3 c r r z c r r c 7 b

nnnnn 37 26 37 26 37 26 37 26 37 26

251 to Val, Gly-268 to Val, Gly-269 to Val, and Gly-781 to Val. NIH FIG. 6. Effect of temperature on the processing of mutants Gly-

3T3 cells expressing the mutant P-glycoprotein-A52 were grown for 48 h a t 26 or 37 "C, as indicated. Samples of whole cells were subjected to SDS-PAGE. transfected to nitrocellulose, and probed with monoclonal antibody A52 as described under 'Experimental Procedures." Rands a t positions A and R represent fully glycosylated and core-glycosylated forms of P-glycoprotein, respectively.

of mature P-glycoprotein. Processing of P-glycoprotein in cells expressing wild-type enzyme was not affected by temperature.

DISCUSSION

It was previously shown that a natural mutant of P-glycoprotein which had a Gly-185 to Val change in the first cytoplasmic loop, resulted in a mutant protein which conferred altered drug specificities in transfected cells (Choi et al., 1988). In light of these observations, the goal of this study was to determine whether other glycine residues located within the predicted cytoplasmic loops are also important for function. We found that changes to 10 of the 20 glycine residues predicted to be in the cytoplasmic sequences had profound effects on protein structure or function. Mutation to valine of 4 glycine residues located within cytoplasmic segments affected the processing of P-glycoprotein. Mutants Gly-251 to Val, Gly-268 to Val, Gly-269 to Val, and Gly-781 to Val yielded products of 150 kDa in both HEK 293 and NIH 3T3 cells, while the wild-type P-glycoprotein-A52 and the remaining glycine mutants yielded a product of 170 kDa. The smaller proteins were sensitive to endoglycosidase H digestion, indicating that they probably reside in the endoplasmic reticulum and, therefore, are unable to participate in the efflux of drugs from the cell. The presence of elevated amounts of partially glycosylated intermediate is likely due to a slowing of a step in the biosynthetic process which could be due to misfolding of the protein. Introduction of a Val in place of a Gly has the potential to seriously disrupt localized folding in the protein. Subtle alterations of the native conformation could be sufficient to disrupt transport out of the endoplasmic reticulum. It is unlikely, however, that these changes yield a protein that is totally unfolded or denatured since degradation products were not observed. I t has been well established that the endoplasmic reticulum possess a mechanism that prevents the transport of misfolded proteins that would otherwise be destined for further processing (Lodish, 1988; Pelham, 1989; Klausner and Sitia, 1990). In previous studies (Loo and Clarke, 1993a, 1993b), we found that mutation of Pro-709 to Ala and Phe-777 to Ala also affected processing of P-glycoprotein. Mu- tant Pro-709 to Ala expressed the 150-kDa protein as the major product, whereas mutant Phe-777 to Ala contained approximately equivalent amounts of the 150- and 170-kDa proteins.

Changes to 2 glycine residues resulted in temperature-sensitive mutants. Processing of mutants Gly-269 to Val and Gly- 781 to Val to the fully mature, 170-kDa form of the enzyme was dramatically increased when cells were incubated at 26 "C. These results suggest that lowering the growth temperature allows the mutant protein to leave the endoplasmic reticulum and proceed to the Golgi complex, where more extensive glycosylation is known to occur. Temperature-sensitive processing of proteins has been observed with other proteins. For example,

studies on temperature-sensitive mutants of the vesicular sto- matitis virus G protein have shown that oligomerization and movement in the cell can be promoted by reducing the incubation temperature (Machamer and Rose, 1988). Similarly, processing of a mutant form of CFTR has also been found to be temperature-sensitive. The most common mutation in CFTR is a deletion of Phe-508 (CFTR F508) (Kerem et al . , 1989). Studies on the localization of CFTR F508 indicate that the mutant protein is only core-glycosylated and does not reach the plasma membrane (Cheng et al., 1990). Denning et al . (1992) showed that lowering the growth temperature for NIH 3T3 cells expressing the mutant resulted in a shift in the glycosylation pattern such that about 30% of the protein corresponded to mature CFTR. Therefore, the defect in processing and temperature sensitivity of CFTR Phe-508 is similar to mutants Gly-269 to Val and Gly-781 to Val.

There may be a mechanism in the endoplasmic reticulum that prevents transport of misfolded proteins otherwise destined for further processing (Lodish, 1988; Pelham, 1989). If such a control mechanism acts on P-glycoprotein, then i t prevents or slows the movement of mutants Gly-269 to Val and Gly-781 to Val a t 37 "C, but to a lesser extent at 26 "C. It is not clear how such a control mechanism would recognize the differences between wild-type and mutant P-glycoproteins. One possibility is that they fold differently. It is quite possible that mutation of glycine to valine affects folding, since an amino acid with no side chain is being replaced with one containing a bulky side chain. The temperature-sensitive mutants may be folding mutants that are in a more native conformation a t the permissive temperature. I t is also possible that P-glycoprotein interacts with another protein, such as a chaperonin, during the folding process. This interaction may be increased by glycine to valine changes in critical positions of P-glycoprotein. Another possibility is that P-glycoprotein is processed as an oligomer and the presence of certain mutations may interrupt the interaction of monomers a t 37 "C. P-glycoprotein may be particularly sensitive to slowing of processing since maturation of the wild-type enzyme is already a slow process relative to other glycosylated proteins (Green- berger et al., 1987; Richert et al., 1988).

Mutants Gly-141 to Val, Gly-187 to Val, Gly-288 to Val, Gly- 812 to Val, and Gly-830 to Val altered the substrate specificity of P-glycoprotein. These mutants, however, showed no detect- able change in structure or biosynthesis, since the protein products corresponded in size to the mature form of P-glycoprotein when expressed in either HEK 293 or NIH 3T3 cells. All five mutants exhibited enhanced abilities to confer resistance to colchicine and adriamycin relative to wild-type enzyme, whereas the ability to confer resistance to vinblastine was es- sentially unchanged. Some decrease in the ability to confer resistance to actinomycin D was also observed in some cases. These characteristics are in some ways similar to mutant Gly- 185 to Val (Choi et al., 1988; Safa et al . , 1990) which also showed an increased ability to confer resistance to colchicine and adriamycin. Mutant Gly-185 to Val, however, differs from the glycine mutants in this study in that there was also a significant decrease in the ability to confer resistance to vinblastine. In addition, mutant Gly-185 to Val showed about a 4-fold increase in binding of azidopine relative to wild-type P-glycoprotein, while azidopine binding by glycine mutants in this study did not show any increased labeling.

The majority of glycine mutations which affect substrate specificity are found within the first loop of the NH,-terminal half, as well as in the first loop of the COOH-terminal half, of the tandemly duplicated molecule. It is not clear how changes in these loops alter drug specificity. Safa et al . (1990) have shown that mutant P-glycoprotein (Val-185) binds less of a colchicine photoactive analog and more of a vinblastine photo-


active analog, despite the fact that the mutant is more efficient in transporting colchicine. Furthermore, both colchicine and vinblastine equally inhibited the labeling of mutant (Val-185) and wild-type (G185) P-glycoprotein by their respective photoactive analogues. On the basis of these results, the authors proposed that mutation of Gly-185 to Val affected release of the drug substrate from the transporter. By contrast, Bruggemann et al. (1992) showed that vinblastine inhibition of labeling by azidopine differed between the mutant and wild-type P-glycoprotein. Their results suggest that the mutation also affects initial drug-protein interactions. Therefore, changes to amino acids in these loops may have complex effects on both initial drug-protein interactions and on their subsequent release from the protein.

The results of this study add to a growing list of point mutations of P-glycoprotein which affect substrate specificity. In recent studies, we have used site-directed mutagenesis to examine the effects of changes to proline and phenylalanine residues within the predicted transmembrane domain of P-glycoprotein ( L o o and Clarke, 1993a, 1993b). We found that mutation of Pro-223, Pro-866, Phe-335, and Phe-978, located within putative transmembrane segments TM4, TM10, TM6, and TM12, respectively, dramatically altered the drug resistance profiles of cells expressing the mutant proteins. Devine et al. (1992) reported that mutations of Gly-338 to Ala and Ala- 339 to Pro in TM6 of hamster P-glycoprotein resulted in pref- erential resistance to actinomycin D in Chinese hamster ovary cells. In another mutant, a Ser-941 to Phe change in TMll of mouse P-glycoprotein resulted in decreased transport of colchicine and adriamycin, while transport of vinblastine was main- tained (Hsuet aE., 1990; Gros et al., 1991). The ability of various mutations to alter substrate specificity of P-glycoprotein sug- gests that there may be multiple overlapping drug-binding sites. All of these changes are within transmembrane segments or within cytoplasmic loops connecting these segments. Fur- ther mutagenesis studies of residues in these domains will aid in the identification of those residues which affect transport function. Such studies should lead to a better understanding of how P-glycoprotein can transport a wide variety of substrates and to better ways to approach the design of new inhibitors of drug transport for clinical applications.

Acknowledgments-We are grateful to Dr. David H. MacLennan for the use of the A52 epitope and monoclonal antibody used in this study and for critically reading the manuscript. We thank Dr. Randal Kauf- man for pMT21 and Dr. David Hampson (Faculty of Pharmacy, Univer- sity of Toronto) for HEK 293 cells.

REFERENCES

Alley, M. C., Scudiero, D. A,, Monks, A,, Hursey, M. L., Czerwinski, M. J., Fine, D. L.,Abbott, B. J., Mayo, J. G., Shoemaker, R. H., and Boyd, M. R. (1988) Cancer Res. 48.589401

Al-Shawi, M. K, and Senior, A. E. (1993) J. Biol. Chem. 266,4197-4206 Ambudkar, S . V., Lelong, I. H., Zhang, J., Cardarelli, C. O., Gottesman, M. M., and

Beck, W. T., and Qian, X.-D. (1992) Biochem. Phnrmacol. 43'89-93 Pastan, I. (1992) Proc. Natl. Acad. Sci. U. S. A 89,8472-8476

Bell, G. I., Fong, N. M., Stempien, M. H., Wormsted, M. A,, Caput, D.. Ku, L., Urdes, M. S., Rall, L. B., and Sanchez-Pescador, R. (1986) Nuckie Acids Res. 14, 84274446

Bruggemann, E. P., Currier, S . J., Gottesman, M. M., and Paatan, I. (1992) J. Biol. Chem. 267,21020-21026

Busche, R., Tummler, B., Cano-Gauci, D. F., and Riordan, J. R. (1989) Eur J.

Cheng, S . H., Gregory, R. J., Marshall, J., Paul, S., Souza, D. W., White, G. A,, Biochem. 183, 189-197

Choi, K, Chen, C.J., Kriegler, M., and Roninson, I. B. (1988) Cell M, 519529 O'Riordan, C. R., and Smith, A. E. (1990) Cell 69,827-834

Cornwell, M. M., Safa, A. R., Felsted, R. L., Gottesman, M. M., and Pastan, I. (1986)

Currier, S . J. , h e , S . E., Willingham, M. C., Cardarelli, C. O., Pastan, I., and

Denning, G. M., Anderson, M. P., Amara, J. F., Marshall, J., Smith, A. E., amd

Devine. S . E., Ling, V., and Melera, P. W. (1992) Proc. Natl. Acad. Sci. U. S. A 89,

Endicott, J. A., and Ling, V. A. (1989) Annu. Rev. Biochem. 58,137-171 Gottesman, M. M., and Pastan, I. (1988) J. Biol. Chem. 263, 12163-12166 Greenberger, L. M., Williams, S. S . . and Honvitz, S . B. (1987) J. Biol. Chem. 262,

Gros, P., Dhir, R., Crmp, J., and Talbot, F. (1991) Proc. Natl. Acad. Sci. U. S. A 88,

Hubbard, S . C., and Ivatt, R. J. (1981)Annu. Reu. Biochem. 60,553-583 Hsu, S . I. H., Cohen, D., Kirschner, L. S., Lothstein, L., Hartatein, M., and Honvitz,

Hyde, S . C., Emsley, P., Hartshorn, M. J., Mimmack, M. L., Gileadi, U., Pearce, S . S . B. (1990) Mol. Cell. Bwl. 10, 3596-3606

R., Gallagher, M. P., Gill, D. R., Hubbard, R. E., and Higgins, C. F. (1990) Nature 346,362-365

Proc. Natl. Acad. Sci. U. S. A 83,3847-3850

Gottesman, M. M. (1992) J. Biol. Chem. 267,25153-25159

Welsh, M. J. (1992) Nature 358,761-764

45644568

13685-13689

7289-7293

Jurunka, P. F., Zastawny, R. L., and Ling, V. (1989) FMEB J. 3,2583-2592 &ne, S . E., Pastan, I., and Gottesman, M. M. (1990) J. Bioenerg. Biomembr 22,

59%618 Kerem, B. S . , Rommens, J. M., Buchanan, J. A, Markiewin, D., Cox, T. K,

Klausner, R. D., and Sitia, R. (1990) Cell 62,611-614 Chakravarti, A., Buchwald. M., and Taui, L.-C. (1989) Science 246,1073-1080

Kunkel, T. A. (1985) Proc. Natl. Acad. Sei. U. S. A. 82,488492 Laemmli, U. K. 11970) Nature 227,68Mi85 Lodish, H. E (1988) J. Biol. Chem. 265,2107-2110 Loo, T. W.. and Clarke, D. M. (1993a) J. Biol. Chem. 268,3143-3149 Loo, T. W., and Clarke, D. M. (1993b) J. Biol. Chem. 268, 19965-19972

Machamer, C. E., and Rose, J. K (1988) J. Biol. Chem. 269,5955-5960 MacLennan, D. H. (1970) J. Biol. Chem. 245,45084518

Pastan, I., and Gottesman, M. M. (1993) Annu. Reu. Biochem. 82,385427 Pelham, H. R. B. (1989) Annu. Reu. Cell Biol. 5, 1-23 Raviv, Y., Pollard, H. P., Bruggernann, E. P., Pastan, I., and Gottesman, M. M.

Richert, N. D., Aldwin, L., Nitecki, D., Gottesman, M. M., and Pastan, I. (1988)

Safa, A. R., Mehta, N. D., and Agresti, M. (1989) Biochem. Biophys. Res. Commun.

Safa, A. R., Stern. R. K, Choi, K, Agresti, M., Tamai. I., Mehta, N. D.. and

Sanger, F., Nicklen, S., and Coulson, A. R. (1977) ?'roc. Natl. Acad. Sci. U. S. A. 74,

Sarkadi, B., Price, E. M., Boucher, R. C., Germann, U. A,, and Scarborough, G. A.

Tamai, I., and Safa, A. R. (1991) J. Biol. Chem. 266,16796-16800 Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. U. S. A 76,

Zubnycka-Gaarn, E., MacDonald, G., Phillips, L., Jorgensen, A. 0.. and MacLen-

(1990) J. Biol. Chem. 26S, 39753980

Biochemistry 27,760G7613

162, 1402-1408

RoNnson, I. B. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,7225-7229

54635467

(1992) J. Biol. Chem. 267, 48544858

4350-4354

nan, D. H. (1984) J. Bioenerg. Biomembr. 1 8 , 4 4 1 4 6

the journal of chemistry vol. 269, no. issue of … · printed in w.s.a. functional ... coproteins...

Documents