INTRODUCTION

Abnormal protein folding or trafficking is associated with a growing number of diseases (1, 2, 3). Diseases such as Alzheimer's and prion-related diseases are characterized by the presence of high levels of insoluble protein aggregates in brain tissue. These plaques appear to be aggregates of misfolded -amyloid protein in Alzheimer's disease or aggregates of misfolded prion protein in the prion-associated diseases such as Creutzfeldt-Jacob disease or Scrapie (mad cow) (4, 5). In cystic fibrosis, the major defect is due to deletion of a single amino acid (F508) in the cystic fibrosis transmembrane conductance regulator (CFTR)¹ resulting in abnormal trafficking to the plasma membrane. The mutant CFTR protein is misfolded, retained in the endoplasmic reticulum, and rapidly degraded (6). Potential therapy for diseases involving folding and/or trafficking defects in the target protein is to prevent misfolding during protein biogenesis.

We have used the human multidrug transporter (P-glycoprotein) as a model system for studying ways to prevent protein misfolding. P-glycoprotein appears to be an excellent model system, because we have identified many misprocessed P-glycoprotein mutants. These temperature-sensitive or -insensitive mutant proteins are misfolded, retained in the endoplasmic reticulum as core-glycosylated biosynthetic intermediates in association with molecular chaperones such as calnexin and Hsc70, and rapidly degraded (7, 8, 9, 10).

P-glycoprotein, the product of the human MDR1 gene, is an energy-dependent pump located at the plasma membrane that interacts with a wide variety of structurally diverse cytotoxic agents that do not have a common intracellular target (11). This protein has clinical importance because it may be one of several mechanisms whereby cancer cells become resistant to chemotherapy.

The protein 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 sequences followed by a hydrophilic domain containing a nucleotide-binding site. The organization of the domains is characteristic of members of the ABC superfamily of (ATP-binding cassette) transporters, the best known member being CFTR.

Our goal was to develop a strategy to specifically rescue the misfolded mutants of P-glycoprotein so that they could exit the endoplasmic reticulum and reach the plasma membrane in a functional form. Nonspecific low molecular weight compounds such as glycerol (12, 13) have been shown to nonspecifically affect protein kinesis. Therefore, we wished to determine whether substrates or modulators of P-glycoprotein could act as specific chemical chaperones and have a more rapid effect on processing of misfolded proteins. We show that biosynthesis of the processing mutants in the presence of substrates or modulators of P-glycoprotein results in the relatively rapid appearance of a fully mature and functional transporter at the cell surface.

EXPERIMENTAL PROCEDURES

Wild-type and mutant MDR1 cDNAs, modified to encode the epitope for monoclonal antibody A52 at the COOH-terminal ends of the proteins, were inserted into the mammalian expression vector pMT21 as described previously (7). Oligonucleotide-directed mutagenesis was carried out as described previously (7). For purification purposes wild-type and mutant MDR1 cDNAs were modified to encode for 10 histidine residues at the COOH ends of the proteins (14). The sequence at the COOH terminus of P-glycoprotein that would normally end as TKRQ became TKRAH₁₀LDPRQ.

HEK 293 cells were transfected with the mutant cDNA constructs. After 24 h, the medium was replaced with fresh medium containing the desired drug concentration. For purification of P-glycoprotein mutants, HEK 293 cells transfected with the cDNA coding for the histidine-tagged P-glycoproteins were solubilized with 1% (w/v) n-dodecyl--D-maltoside, and the mutant P-glycoproteins were purified by nickel-chelate chromatography. Drug-stimulated ATPase activity was determined as described previously (14).

Pulse-chase experiments were done as described previously (8).

Whole cell extracts or purified P-glycoprotein samples were subjected to SDS-PAGE, electroblotted onto nitrocellulose, and developed with monoclonal antibody A52 (15) or with a rabbit polyclonal antibody against P-glycoprotein followed by enhanced chemiluminescence (Amersham Corp.) as described previously (9).

RESULTS

The effect of substrates and modulators of P-glycoprotein on the biosynthesis of two processing mutants were initially studied; G268V in the NH₂-terminal transmembrane domain (16) and Y490 in the NH₂-terminal nucleotide-binding domain (17). Mutant G268V is a temperature-insensitive processing mutant, whereas mutant Y490 contains a deletion at an equivalent position to the F508 mutation in CFTR. The cDNAs coding for these mutant P-glycoproteins were modified to encode for the epitope of monoclonal antibody A52. The presence of the epitope provided us with an important tool for following the synthesis of the misfolded mutants. It allowed us to distinguish the mutant protein from any endogenous P-glycoprotein that may be induced by the presence of drug substrates. These mutants, as well as wild-type enzyme, were expressed transiently in HEK 293 cells and then treated for 24 h with various concentrations of drug substrates. Four structurally different drug substrates or modulators (capsaicin, cyclosporin, verapamil, and vinblastine) of P-glycoprotein were tested. Vinblastine is an antitumor agent that is a substrate of P-glycoprotein, whereas verapamil and cyclosporin A are inhibitors of drug transport. Capsaicin, the pungent ingredient in peppers of the Capsicum family, is a substrate of P-glycoprotein, based on its ability to stimulate ATPase activity (data not shown). All four compounds are hydrophobic and can readily diffuse through the plasma membrane to the site of protein synthesis in the endoplasmic reticulum.

In the absence of drug substrates (Fig. 1A), the major product of the mutant P-glycoproteins had an apparent mass of 150 kDa compared with 170 kDa for the wild-type enzyme. The 170- and 150-kDa forms of the enzyme represent mature and core-glycosylated biosynthetic intermediate, respectively. In the presence of capsaicin, cyclosporin A, verapamil, or vinblastine, however, the amount of mature enzyme (170 kDa) for both mutants increased in a dose-dependent manner. The ability to "rescue" misprocessed mutants appeared to be independent of the location of the mutation. In addition to the mutants G268V and Y490, we were able to facilitate processing of P-glycoproteins with mutations in the predicted transmembrane segments (TM1, G54V; TM5, G300V; TM7, A718L; and TM9, A841L), in the extracellular loops between transmembrane segments (G854V), in the cytoplasmic loops (G251V and W803A), in the nucleotide-binding domains (G427C and S434C), and in the linker region connecting the two halves of the molecule (E707A) (data not shown).

Expression of wild type and mutants of P-glycoprotein-A52 (A) or wild-type and mutant (F508) CFTR (B) in the absence or the presence of drug substrates.

 C

 B

The most potent of the four compounds was cyclosporin A. More than 50% of the mutant protein was present as the mature form of the enzyme in the presence of 2-10 µM of cyclosporin A, 5-20 µM vinblastine, 12.5-50 µM verapamil, or 75-150 µM capsaicin. The highest concentration of capsaicin (300 µM) appeared to be quite toxic to the cells. Except for vinblastine, the cells continued to proliferate in the presence of the various drug substrates. Vinblastine, which is an inhibitor of microtubule assembly, did not cause immediate cell death but resulted in the detachment of the cells from the dish. Other hydrophobic compounds that are not substrates of P-glycoprotein, such as 3-methoxy-tyramine or 3-hydroxy-4-methoxyphenethyl amine, had no effect on the processing of the misfolded mutants (data not shown). The effects of drug substrates on folding appear to be specific to P-glycoprotein because the drug substrates of P-glycoprotein could not rescue the temperature-sensitive CFTR F508 processing mutant (Fig. 1B).

Restoration of processing of the misfolded mutants appeared to be quite rapid and occurred within a few hours after the addition of drug substrates to the medium. Fig. 2A shows that after 4 h in the presence of 15 µM cyclosporin A, about 50% of mutant G268V was present as the fully mature (170-kDa) form of the enzyme and that after 24 h, more than 80% of the mutant protein was present in the fully mature form. The total amount of P-glycoprotein also increased dramatically in the presence of cyclosporin A. These results suggested that the drug stabilized the mutant protein resulting in decreased turnover. This was confirmed by pulse-chase studies. Fig. 2B shows that in the absence of cyclosporin A, the 150-kDa P-glycoprotein of mutant G268V was not processed to the mature enzyme. The core-glycosylated protein was rapidly degraded (half-life about 2 h), and there was little product remaining after 8 h. In the presence of cyclosporin A, however, the kinetics of maturation of the P-glycoprotein of mutant G268V was similar to that of wild-type enzyme. By 4 h post-labeling, the majority of the 150-kDa protein was processed to the mature enzyme (170 kDa). The processed P-glycoprotein was stable for at least 24 h. Cyclosporin A had no detectable effect on the processing of the wild-type enzyme.

Cell surface labeling was performed (8) to determine whether the mutant proteins reached the plasma membrane. HEK 293 cells were transfected with the histidine-tagged mutant cDNAs and then incubated in the presence or the absence of 15 µM cyclosporin A. The transfected cells were then treated with periodate to convert the carbohydrate moieties to aldehydes, followed by addition of biotin-LC-hydrazide (Pierce). Biotin-LC-hydrazide is a nonpermeable compound that forms covalent attachments to extracellular glycoproteins after periodate oxidation (18). The histidine tagged P-glycoprotein mutants were then purified by nickel-chelate chromatography and immunoblotted with streptavidin-conjugated horseradish peroxidase. Fig. 3 shows that wild-type but not the mutant P-glycoproteins, was present at the cell surface when expression was done without drug substrate. When the transfected cells were incubated in the presence of 15 µM cyclosporin A, however, both mutant proteins were detected at the cell surface.

Cell surface labeling of histidine-tagged wild-type and mutant forms of P-glycoprotein after expression in the absence () or the presence (+) of cyclosporin A.

To test if the mutant P-glycoproteins were active when expressed in the presence of drug substrates, we attempted to purify the mutant P-glycoproteins by nickel chelate chromatography for measurement of drug-stimulated ATPase activity. We have previously found that wild-type but not misprocessed P-glycoprotein containing a histidine tag can readily be recovered by nickel-chelate chromatography (10). Apparently the processing mutants are misfolded, resulting in masking of the histidine tag. We modified the cDNAs of the mutant P-glycoproteins to code for 10 tandem histidine residues at the COOH end of the molecule to facilitate purification by nickel-chelate chromatography (14). The cDNAs of the mutant P-glycoproteins were transiently expressed in HEK 293 cells and then incubated for 24 h in the presence or the absence of 15 µM cyclosporin A. Histidine-tagged P-glycoprotein was isolated by nickel-chelate chromatography. The majority of the wild-type P-glycoprotein was bound to the nickel column and was eluted with 0.3 M imidazole regardless of whether expression was carried out in the presence or the absence of cyclosporin A. This was determined by loading equivalent amounts of the flow-through and eluted fractions on SDS-PAGE followed by immunoblotting with a polyclonal antibody against P-glycoprotein (data not shown). By contrast, most of the P-glycoprotein of mutants G268V and Y490 grown without drug substrate were recovered in the flow-through fractions during nickel-chelate chromatography. In the presence of 15 µM cyclosporin A, however, the majority of the mutant P-glycoproteins were recovered by nickel-chelate chromatography and had an apparent mass of 170 kDa. Similar results were obtained when the transfected cells were incubated in the presence of capsaicin, verapamil, or vinblastine (data not shown).

Drug-stimulated ATPase activity of the wild-type and mutant P-glycoproteins was measured in the presence of verapamil, vinblastine, or colchicine after the addition of lipid. Fig. 4 shows that the purified P-glycoprotein of mutant Y490 after expression in the presence of cyclosporin resulted in a functional molecule that exhibited a similar pattern of drug-stimulated ATPase activity as the wild-type enzyme. Similarly, drug-stimulated ATPase activity was detected in the mutant G268V after expression in the presence of cyclosporin A. The observation that mutant G268V exhibits reduced activity is consistent with previous observations that several glycine to valine mutations in the cytoplasmic loops of P-glycoprotein also alter the substrate specificity of the enzyme (16). These results show that the presence of drug substrates during biosynthesis of the misfolded proteins results in the appearance of a functional transporter at the cell surface.

DISCUSSION

Drug substrates or modulators of P-glycoprotein appear to be acting as powerful "chaperones" for processing misfolded P-glycoproteins. All misprocessed mutants of P-glycoprotein that we have tested could be converted to the fully mature form of the enzyme, even when the mutations were located in different domains of the molecule. The effects of these substrates or modulators were also independent of the cell lines because the misfolded mutants could be rescued when expressed either transiently in HEK 293 cells (this study) or stably in NIH 3T3 cells (data not shown).

The ability of drug substrates to facilitate folding appears to be specific for P-glycoprotein because these substrates were ineffective on the CFTR mutant (F508). Reversal of the misfolding phenotype of the F508 CFTR mutant can be accomplished by incubation at lower temperatures (19) or by exposure to nonspecific low molecular weight compounds such as glycerol (12, 13), trimethylamine N-oxide, or deuterated water (13). In our hands, the effectiveness of glycerol treatment or lowering the growth temperature to facilitate processing was cell line-dependent. For example, some misfolded P-glycoprotein mutants such as G714A are temperature- and glycerol-sensitive only when stably expressed in NIH 3T3 cells but not when expressed in HEK 293 cells (data not shown). In addition, maturation of the misfolded P-glycoproteins in the presence of glycerol or by lowering the incubation temperatures is a slow process that usually requires 24-72 h. By contrast, the appearance of the mature form of any of the misprocessed mutants of P-glycoprotein occurs within 2-4 h after addition of any drug substrate (Fig. 2). Another interesting observation is that misfolded mutants that are temperature- and glycerol-insensitive, such as G251V, G268V, and E707A could also be rescued by these drug substrates when expressed in either HEK 293 or NIH 3T3 cells (data not shown).

The exact mechanism of how these specific drug substrates or modulators facilitate processing of misfolded P-glycoproteins mutants is not known. A possible explanation is that the drug-binding sites(s) in P-glycoprotein are formed early in the folding intermediates during biosynthesis. Occupation of the drug-binding site(s) stabilized the folding intermediates in a "near native" conformation, thus escaping the cell's quality control mechanism (Fig. 2B). Recently, Qu and Thomas (20) studied the effects of the CFTR F508 on the thermodynamic stability and folding yield of the nucleotide-binding domain 1 and concluded that the major deleterious effect of the mutation was to allow accumulation of a folding intermediate that was prone to self-association. The F508 mutation had little effect on the thermodynamic stability of the folded nucleotide-binding domain 1. The mutation also did not appear to enhance in vivo proteolysis because inhibition of proteasomes (21, 22) did not enhance the efficiency of maturation of the full-length F508 CFTR. These results suggest that mutations that cause misprocessing slow one or more folding steps, resulting in an increased concentration of the intermediate that is prone to self-aggregation. Therefore, in P-glycoprotein, it is possible that the occupation of the drug-binding site(s) in the early stages of folding may reduce the concentration of the intermediate that is prone to self-aggregation.

In summary, the results of this study demonstrate that a potential strategy in the treatment of diseases involving trafficking/misfolding of proteins would be to identify specific synthetic and natural substrates or modulators and to include these during biosynthesis.