By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
Progressive familial intrahepatic cholestasis type 3 (PFIC3) is a rare liver disease characterized by early onset of cholestasis that leads to cirrhosis and liver failure before adulthood. PFIC3 may be improved by chronic administration of ursodeoxycholic acid, although in many cases liver transplantation is the only therapy. The disease is caused by mutations of the adenosine triphosphate (ATP)–binding cassette, sub-family B, member 4 (ABCB4) [multidrug resistance 3 (MDR3)] gene encoding a specific hepatocellular canalicular transporter involved in biliary phosphatidylcholine secretion. Several mutations have been reported; however, the effect of individual mutations has not been investigated. ABCB4 is highly homologous to ATP-binding cassette, sub-family B, member 1 (ABCB1) (MDR1), the multidrug transporter responsible for drug resistance of cancer cells. We have studied the effect of mutation I541F localized to the first nucleotide-binding domain, which is highly conserved between ABCB4 and ABCB1. Plasmids encoding the wild-type human ABCB4 or rat ABCB1–green fluorescing protein (GFP) construct, and corresponding I541F-mutants, were expressed in hepatocellular carcinoma, human (HepG2) and Madin-Darby canine kidney (MDCK) cells. Expression studies showed that ABCB4 was localized at the bile canalicular membrane in HepG2 cells and at the apical surface in MDCK cells, whereas the I541F mutant was intracellular. In MDCK cells, ABCB1-I541F also accumulated intracellularly in compartments, which were identified as the endoplasmic reticulum and cis-Golgi, and remained partially endoH-sensitive. After shifting cells to 27°C, ABCB1-I541F was expressed at the apical cell surface in a mature and active form. Similarly, ABCB4 was significantly trafficked to the membrane of bile canaliculi in HepG2 cells. Conclusion: Mutation I541F causes mislocalization of both ABCB4 and ABCB1. Intracellular retention of ABCB4-I541F can explain the disease in PFIC3 patients bearing this mutation. The observation that plasma membrane expression and activity can be rescued by low temperature opens perspectives to develop novel therapies for the treatment of PFIC3. (HEPATOLOGY 2009.)
Mutations of the adenosine triphosphate (ATP) –binding cassette, sub-family B, member 4 (ABCB4) gene [also called multidrug resistance 3 (MDR3)] are at the origin of progressive familial intrahepatic cholestasis type 3 (PFIC3), a rare lethal autosomal recessive liver disorder, characterized by early onset of persistent cholestasis that progresses to cirrhosis and liver failure before adulthood.1–3 PFIC3 can be distinguished from other PFIC types by a high serum gamma-glutamyltransferase activity and liver histology that shows portal inflammation and ductular proliferation at early stages.2–4 The morbidity is primarily the result of profound cholestasis that often is associated with persistent pruritus. The ABCB4 gene encodes ABCB4, a phospholipid floppase involved in biliary phosphatidylcholine excretion.5 Cholestasis is thought to result from the toxicity of bile in which detergent bile salts are not neutralized by phospholipids, leading to bile canaliculi and biliary epithelium injuries. The absence of phospholipids in bile is expected to destabilize micelles and promote lithogenicity of bile with crystallization of cholesterol, which could favor small bile duct obstruction. There is now strong evidence that in addition to PFIC3, ABCB4 defect can be involved in intrahepatic cholestasis of pregnancy, cholesterol gallstone disease, drug-induced liver injury, and idiopathic biliary fibrosis in the adult.6–13 The patients are mainly heterozygous, contrary to PFIC3 patients who are generally homozygous or compound heterozygous. These cholestases of the adult generally improve or regress under treatment with ursodeoxycholic acid (UDCA), a hydrophilic bile acid. In children suffering from PFIC3, UDCA only improves the symptoms in nearly 30% of the cases, and the only alternative remains liver transplantation.4
ABCB4 belongs to the family of membrane transporters that contain an ATP-binding cassette (ABC-transporters) and hydrolyze ATP during translocation of a wide variety of molecules across membranes. In the human, the ABCB4 gene encodes several splicing variants with a major transcript encoding a protein of 1279 amino acids.14 ABCB4 is a polytopic transmembrane protein composed of two halves, each containing a transmembrane domain with six membrane-spanning segments, and an ATP-binding domain including the ABC signature. ABCB4 is highly homologous to ATP-binding cassette, sub-family B, member 1 (ABCB1) [multidrug resistance 1 (MDR1)], with 77% identity at the amino acid level. Despite their high degree of homology, ABCB1 and ABCB4 are functionally different and have distinct tissue-specific expression patterns. ABCB4 is mainly expressed in the liver15 and is specialized in phosphatidylcholine transport. ABCB1 extrudes a wide variety of drugs and is responsible for multidrug resistance of cancer cells. Since the identification of ABCB4-associated diseases, a growing number of mutations have been reported.16–18 However, it remains unclear how mutations identified from genetics studies affect ABCB4 expression. If it is clear that frame-shift mutations introduce premature stop codons and lead to nonfunctional truncated mutants, missense mutations may affect the protein in different ways. A defined mutation may prevent the protein to be located properly at the canalicular membrane or may impair its function. The precise knowledge of the defect may help to consider new therapeutic strategies, at least for some mutations.
We have studied the effect of mutation I541F, located in the first nucleotide-binding domain (NBD), which has been described in a homozygous patient with PFIC3.4 This female patient experienced chronic cholestasis since 1 year of age and developed cirrhosis. Chronic administration of oral UDCA had no beneficial effect, and she received a liver graft at age 5 years. MDR3 liver immunostaining showed absence of canalicular protein. We have reproduced the mutation in both ABCB4 and ABCB1 and show that this mutation is not specifically linked to ABCB4 function but causes a folding defect of the NBD domain that affects normal processing and trafficking of both transporters to the plasma membrane. When folding is improved at reduced temperature, both traffic and activity are rescued.
The monoclonal P3II-26 anti-ABCB4 antibody was obtained from Alexis Biochemicals (Lausen, Switzerland). Monoclonal C219 was from Abcam (Cambridge, UK). Rabbit polyclonal anti-calnexin and mouse monoclonal anti-disulphide isomerase) were from Stressgen (Ann Arbor, MI). Monoclonal anti-GFP was from Roche Diagnostics (Meylan, France). Monoclonal AC17 anti–lysosomal-associated membrane protein 2 was a kind gift of A. Le Bivic (IBDM, Marseille, France). Mouse monoclonal antibody to Golgi matrix protein was from Sigma (St-Quentin-Fallavier, France). Goat polyclonal antibody anti-early endosome antigen and monoclonal anti-giantin were from Santa Cruz Biotechnology (Santa Cruz, CA). The polyclonal antibody against GFP was raised in the laboratory by J.L.D.. Cy3-conjugated secondary antibodies were purchased from Jackson Immunoresearch (Montluçon, France). Alexa Fluor 488 secondary antibodies and culture media were from Invitrogen (Cergy-Pontoise, France). Peroxidase-conjugated secondary antibodies were from Rockland Immunochemicals (Gilbertsville, PA). Endo-β-N-acetylglucosaminidase H (endoH) and peptide N-glycosidase F (PNGase-F) were from New England Biolabs (Ipswich, MA). Calcein-AM was from AnaSpec (San Jose, CA). Verapamil was from Calbiochem (Merck Chemicals Ltd., Nottingham, UK). The ECL-Plus detection kit was from GE Healthcare France (Orsay, France). All other reagents were from Sigma.
Constructs and Mutagenesis.
The plasmid pJ3omega- MDR3 encoding human ABCB4 isoform A was obtained from the American Type Culture Collection (LGC Standards, Molsheim, France). The ABCB4 complementary DNA (cDNA) was extracted with HindIII and XbaI restriction enzymes and cloned into the pcDNA3 expression vector (Invitrogen) using the same restriction sites as described.19 The pEGFP-MDR1 plasmid encoding rat ABCB1 fused with GFP at the C-terminus,20 was a gift of Pr. I. M. Arias (Tufts University School of Medicine, Boston, MA). For construction of the ABCB4-I541F and ABCB1-I541F-GFP mutants, site-directed mutagenesis was performed using the QuikChange XL mutagenesis kit (Stratagene Europe, Amsterdam Zvidoost, The Netherlands). The primers used for mutagenesis were GCA GAA GCA GAG GTT CGC CAT TGC ACG TGC (sense) and GCA CGT GCA ATG GCG AAC CTC TGC TTC TGC (antisense) for ABCB4 and GGA CAG AAA CAG AGG TTC GCC ATT GCC CGG GCC (sense) and GGC CCG GGC AAT GGC GAA CCT CTG TTT CTG TCC (antisense) for ABCB1. Constructs were verified by automated sequencing.
Cell Culture and Transfection.
Hepatocellular carcinoma, human (HepG2) cells and Madin-Darby canine kidney (MDCK) II cells were grown at 37°C in Dulbecco's modified Eagle medium with 10% heat-inactivated fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin, in a humidified incubator under a 5% CO2/air atmosphere. HepG2 and MDCK cells were transfected using nucleofector II (Amaxa, Cologne, Germany) using program H22 (solution V) and K29 (solution T), respectively. Stable transfectants were obtained by selection with 800 μg/mL G-418 sulfate for 3 weeks and were subsequently grown in the presence of 100 μg/mL G-418. Enrichment of the cell population with ABCB1-I541F-GFP–expressing cells was performed by fluorescence-activated cell sorting. Trypsinized cells (usually 106 cells) were harvested and washed in Leibovitz's medium with 20% fetal bovine serum. Fluorescent cells were sorted using an Influx 500 Cytopeia cytofluorimeter (Seattle, WA). After sorting, cells were immediately placed in standard culture conditions and expanded. For 27°C experiments, cells were fed with Leibovitz's medium and placed in an incubator without CO2. Control cells were placed in the same conditions, but at 37°C. Before the experiments, cells were treated overnight with 10 mM sodium butyrate, to increase the cytomegalovirus promoter transcriptional activity.
Immunofluorescence and Confocal Microscopy.
For immunofluorescence, HepG2 cells were grown on glass coverslips. MDCK cells were grown either on coverslips or on Transwell polycarbonate filter units (Costar Corp., Cambridge, MA). For the detection of ABCB4, cells were fixed with methanol/acetone (4/1, vol/vol) for 3 minutes at −20°C, and processed for immunostaining with the mouse monoclonal anti-ABCB4 P3II-26 as described.21 For the detection of ABCB1 with C219, cells were fixed with 4% paraformaldehyde for 1 minute, then with methanol for 10 minutes. In all other cases, immunofluorescence was performed on cells fixed with 4% paraformaldehyde and subsequently permeabilized with 0.075% saponin. Incubations with primary and secondary antibodies were performed as described.22 Nuclei were stained with propidium iodide. Confocal imaging was acquired with a Leica TCS SP2 Laser Scanning Spectral system attached to a Leica DMR inverted microscope with a 63/1.4 immersion objective. Digital images were analyzed using the on-line ScanWare software and processed with Image J software (NIH, MA).
Immunoprecipitation, Deglycosylation and Western Blot.
Transfected cells were washed with phosphate-buffered saline containing 1mM CaCl2 and 0.5 mM MgCl2 and lysed on ice for 30 minutes in 20 mM Tris HCl, 150 mM NaCl, 1 mM ethylenediaminetetra-acetic acid, pH 7.4, containing 1 % (wt/vol) Triton X-100 in the presence of a protease inhibitor cocktail (Roche Diagnostics, Basel, Switzerland). Lysates were passed through a 23-gauge needle to shear the DNA, and then centrifuged at 16,000 × g for 10 minutes to remove insoluble materials. Protein content was determined by Uptima bicinchoninic acid protein assay (Interchim, Montluçon, France). Equal amounts of protein were directly processed for sodium dodecyl sulfate polyacrylamide gel electrophoresis or immunoprecipitated. Immunoprecipitation of ABCB1-GFP was performed using polyclonal anti-GFP (1.5 μL) adsorbed onto 70 μL protein A–Sepharose beads (1 mg protein A/mL) for 4 hours at 4°C. Beads were then washed and resuspended in appropriate buffers for digestion with endoH (2000 U) or PNGase-F (2000 U) (New England Biolabs) according to the instructions of the manufacturer. The reaction was stopped with 5× Laemmli buffer, and samples were processed for sodium dodecyl sulfate polyacrylamide gel electrophoresis on 6% polyacrylamide gels. Immunoblotting was performed using the mouse monoclonal anti-GFP followed by horseradish peroxidase–conjugated mouse-specific secondary antibody. Development of peroxidase activity was performed with the ECL Plus detection kit.
The procedure was essentially performed as described.23 Cells were seeded at 70,000 cells/well in 96-well plates with clear bottoms (Greiner Bio-One, Les Ulis, France). Cells were fed 24 hours after seeding, and the assay was performed 48 hours later. Medium was removed and monolayers washed three times with Leibowitz's medium. Calcein-AM was added at a final concentration of 5 μM and 0.1% dimethylsulfoxide in 100 μL Leibowitz's medium. Maximum calcein fluorescence was measured in cells treated with 40 μM verapamil, which was added 10 minutes before and during the assay. Plates were immediately placed in a Tecan SpectraFluor cytofluorimeter (MTX Lab Systems, Vienna, VA) for 60 minutes and read at 15-minute intervals at 485-nm excitation and 530-nm emission.
Expression of ABCB4 and ABCB4-I541F in HepG2 and MDCK Cells.
HepG2 cells derive from a human hepatocellular carcinoma and are able to polarize and develop bile canaliculi in culture. To examine the cellular localization of ABCB4 and ABCB4-I541F, HepG2 cells were transiently transfected with the pcDNA3 plasmid encoding wild-type ABCB4 or ABCB4-I541F or with the empty plasmid. Immunofluorescence was performed 60 hours later with the P3II-26 monoclonal antibody. Control cells transfected with the empty plasmid did not express detectable amounts of ABCB4 (not shown). Cells transfected with the plasmid encoding ABCB4 expressed the protein exclusively on the membranes of bile canaliculi, which correspond to the apical surface (Fig. 1A), consistent with the expected canalicular localization of ABCB4 in hepatocytes.15 By contrast, cells transfected with the plasmid encoding ABCB4-I541F did not express the protein at the membrane of bile canaliculi. The mutant was exclusively detected intracellularly (Fig. 1B).
The MDCK cell line is a well-characterized polarized kidney cell line, which forms a fully polarized monolayer when grown on filters. MDCK cells are widely used to study polarized expression and traffic of membrane proteins. MDCK cells were transfected with the plasmids encoding ABCB4 and ABCB4-I541F or the empty plasmid. Transfection was more efficient than in HepG2 cells, and stably transfected cell populations were established. Again no ABCB4 protein was detected in control cells transfected with the empty plasmid (not shown). In cells transfected with wild-type ABCB4 cDNA, the protein was exclusively detected at the apical surface, over the nuclei (Fig. 1C). Transverse confocal sections confirmed that ABCB4 was apically localized (Fig. 1E). By contrast, ABCB4-I541F was detected around the nuclei (Fig. 1D). Transverse confocal sections showed a roughly uniform distribution in all the height of the cells, indicating that the mutation causes intracellular retention (Fig. 1F). These expression studies show that ABCB4 is expressed at the apical membrane, whereas ABCB4-I541F leads to a major trafficking defect and accumulates within the cell cytoplasm, in both HepG2 and MDCK cells.
ABCB1-I541F-GFP Is Also Intracellular.
The I541F mutation is localized next to the ABC signature (LSSGGQ) in the first nucleotide-binding domain. This domain is involved in binding and hydrolysis of ATP. It is highly conserved between members of the same ABC transporter family, as shown by comparison with the multidrug transporter ABCB1 (Fig. 2). To determine whether the mutation affected ABCB4 specifically or not, the mutation was introduced into ABCB1. We used the vector pEGFP-ABCB1, which encodes ABCB1 fused with GFP at the C-terminus.20 It has been shown that the presence of GFP does not affect ABCB1 trafficking or activity.20 In MDCK cells, ABCB1-GFP was exclusively expressed at the apical surface (Fig. 3A), as already reported.24 In contrast, ABCB1-I541F-GFP was intracellular, as observed for ABCB4-I541F. In these experiments, the green fluorescent signal of GFP was taken as a reporter for the localization of the protein. To check that it was indeed representative of the fusion protein, cells were also immunostained using the C219 monoclonal antibody, which is directed against MDR proteins. Cells were grown to subconfluence to visualize more precisely the intracellular localization. Figure 3B shows that the two fluorescent signals colocalized perfectly. ABCB1-I541F-GFP accumulated around the nuclei, in membrane compartments that appeared sometimes granular (Fig. 3B). These expression experiments show that mutation I541F causes a trafficking defect both in ABCB1 and ABCB4, and that ABCB1-GFP chimera provides an interesting model to further investigate the effect of I541F mutation.
ABCB1-I541F-GFP Is Retained in an Endoplasmic Reticulum/Golgi Compartment.
To precisely determine the sites of intracellular accumulation, different cellular compartments were identified with specific antibodies in MDCK/ABCB1-I541F-GFP cells grown to subconfluence. Detection of the mutant by the green fluorescence of GFP prevented any misinterpretation attributable to possible cross-reaction between antibodies. In all of these colocalization experiments, ABCB1-I541F-GFP was observed exclusively in the cytoplasm, especially around the nuclei, whereas little ABCB1-GFP was intracellular (Fig. 4A). ABCB1-I541F-GFP colocalized strongly with calnexin and to a lesser extent with protein disulfide isomerase, two endoplasmic reticulum (ER) markers (Fig. 4B). Alteration of the ER was evident, especially when the pattern of calnexin was examined. In nontransfected cells, the distribution of calnexin was finely granular, whereas in ABCB1-I541F cells, bundles strongly labeled for calnexin colocalized with the fluorescent mutant (Fig. 4B). This pattern was strongly suggestive of an ER accumulation with local enlargements of ER cisternae. A large colocalization was also observed with the cis-Golgi marker giantin (Fig. 4B), but only little colocalization with Golgi matrix protein, a middle-Golgi marker. No colocalization was observed with the early endosomal antigen or with the lysosomal protein lysosomal-associated membrane protein 2 (Fig. 4B). These colocalization experiments show that the mutant is retained and accumulates in an ER/Golgi compartment during its biosynthetic process.
ABCB1-I541F-GFP Is Not Processed Correctly.
A biochemical means to investigate the state of processing of proteins is to check their glycosylation. ABCB4 has two N-glycosylation consensus sites in the first extracellular loop, and ABCB1 has three. To assess more precisely the processing defect caused by the mutation, the glycosylation status of ABCB1 and ABCB1-I541F was studied with endoglycosidases. On western blots, ABCB1-GFP migrated essentially with an apparent molecular mass of 190 kDa. A very faint band migrating at 180 kDa was also detected. The molecular mass of the 190 kDa band decreased to slightly less than 180 kDa after deglycosylation with PNGase-F (Fig. 5), but did not change after endoH treatment, indicating that the wild-type protein is fully matured with complex N-glycosylated chains. Only the faint 180-kDa band was endoH-sensitive and shifted to a band migrating slightly faster. In contrast, ABCB1-I541F-GFP migrated as two bands of roughly equal intensity of 180 and 190 kDa (Fig. 5). Treatment with endoH shifted the 180-kDa band to a slightly lower molecular mass (Fig. 5), indicating that a large part of the mutant was not complex glycosylated. Thus, biochemical studies confirmed that the I541F-mutant does not mature as the wild-type protein during the biosynthetic process, and remains largely in a high-mannose endoH-sensitive form.
Lowering the Temperature Allows the Mutant to Reach the Plasma Membrane.
Retention in the ER/Golgi is suggestive of a folding defect that may be recognized by the cell quality control machinery and prevent the mutant from maturing. There are examples of folding defects that can be overcome by lowering the temperature.25 MDCK-ABCB1-I541F-GFP cells were filter grown to confluence at 37°C, then the cells were switched or not to 27°C for 24 hours. Figure 6A shows that in control cells grown at 37°C, ABCB1-I541F-GFP was intracellular, whereas in cells grown at 27°C, ABCB1-I541F-GFP was expressed at the apical surface. In the same conditions, ABCB1-GFP remained apically localized (not shown). Western blotting confirmed that, after switching to 27°C, the mutant acquired full maturation and migrated with a higher apparent molecular mass on western blots (Fig. 6B). These results show that I541F is a temperature-sensitive mutation and that processing and trafficking of the mutant reverts toward that of wild-type on lowering temperature.
ABCB1-I541F-GFP Is Functional When Expressed at the Apical Membrane.
We took advantage of the drug transporting activity of ABCB1 to monitor the activity of the mutant. The calcein assay uses calcein-AM as an ABCB1 substrate.23 Calcein-AM is the nonfluorescent acetoxymethyl ester derivative of the fluorescent probe calcein and can penetrate cell membranes. Once inside the cell, the lipophilic blocking groups are cleaved by nonspecific esterases, resulting in a charged fluorescent form that leaks out of cells far more slowly than its parent compound. Control MDCK cells accumulated calcein in both the presence and absence of the competitive inhibitor verapamil. In contrast, MDCK/ABCB1-GFP cells showed verapamil-inhibited extrusion of calcein (Fig. 6C). Only very weak activity was measured in MDCK/ABCB1-I541F-GFP cells grown at 37°C; however, when cells were grown at 27°C for 24 hours before the assay, they accumulated significantly less fluorescent calcein (Fig. 6C). These results show that mutation I541F does not impair activity, provided that the transporter can reach the plasma membrane.
ABCB4 Is Trafficked to the Bile Canaliculi of HepG2 Cells at Low Temperature.
To check whether the temperature rescue of ABCB1-I541F in MDCK cells might be relevant for PFIC3 therapy, we studied the effect of low temperature on ABCB4 expression in HepG2 cells. Cells were transiently transfected with either ABCB4 or ABCB4-I541F cDNA. Forty-eight hours after transfection, cells were either left at 37°C or shifted to 27°C for 17 hours before studying ABCB4 expression. Figure 7A shows that ABCB4-I541F was detected in significant amount at the membrane of bile canaliculi in cells grown at 27°C. By contrast, no labeled bile canaliculi were observed in cells grown at 37°C. The bile canalicular localization of the wild-type protein was not changed by the temperature shift (Fig. 7A). By immunoblotting, ABCB4 was detected as a major band migrating in a region of 160 kDa and a minor band at 140 kDa, both at 37°C and 27°C (Fig. 7B). ABCB4-I541F migrated as a single 140-kDa polypeptide in cells grown at 37°C. After shifting the cells to 27°C, a significant proportion of the mutant was detected with an apparent molecular mass of approximately 160 kDa (Fig. 7B). These results show that, for ABCB1-I541F in MDCK cells, low temperature is effective on processing and bile canalicular delivery of the ABCB4 mutant in hepatic cells.
We show that a mutation identified in a PFIC3 patient causes a maturation defect that can be rescued by low temperature. Our observations not only provide a rational explanation for linking the mutation to the disease but also bring perspectives to find novel therapies for the targeted treatment of PFIC3.
Homozygous I541F ABCB4 mutation was described in a patient with PFIC3 who presented with persistent cholestasis from the age of 1 year and developed severe portal hypertension.4 In the patient's liver, immunohistochemistry did not show any canalicular staining for ABCB4.4 This suggested that in the liver in vivo, the mutant protein could not reach the canalicular membrane because of a targeting defect or protein degradation. In both HepG2 and MDCK cells, we found that ABCB4-I541F accumulated intracellularly and was not detected at the canalicular membrane, as in the previously reported mutated patient. The site of accumulation was identified as the ER and cis-Golgi, in accordance with the endoH sensitivity of the mutant. It must be noted, however, that the mutant was partly endoH resistant. Whether this resistant form is at the plasma membrane or in a medium/distal Golgi compartment is difficult to assess. We cannot exclude that some molecules may be fully processed and reach the plasma membrane in overexpressing systems, such as transfected cell line models.
The location of the I541F mutation, in the first NBD domain, next to the LSSGQ signature, predicted that ATP-binding or hydrolysis could be affected. This is not apparently the case, because the primary defect was failure of the mutant to traffic properly to the apical plasma membrane. Dixon et al.26 have shown that the A546D mutation described in a patient with intrahepatic cholestasis of pregnancy affected the traffic of an ABCB1 mutant to the plasma membrane, but not its activity. In the work of Dixon et al.,26 ABCB1 was used as a model to reproduce the A546D ABCB4 mutation. Here, we also used ABCB1 and showed that the I541F mutation caused similar intracellular retention in both ABCB1 and ABCB4. ABCB1 represents an interesting model, especially when considering mutations in the NBD domains. These domains are highly conserved between ABCB1 and ABCB4 and have been shown to be interchangeable.27 Furthermore, ABCB1 can be expressed as a GFP-fusion protein that retains drug-transporting activity. Because many ABCB4 missense mutations are found in the first NBD domain, ABCB1-GFP may be a good model to understand the effect of these mutations, to screen for means of rescuing the mutants, and to test whether they are active.
Many inherited diseases are known to arise because of point mutations within a gene that result in the production of proteins unable to assume a stable conformation within the cell.28, 29 Mutations lead to non-native protein folding intermediates that are recognized by specialized chaperones and eventually targeted for destruction by the quality control machinery of the cell.30 The fact that trafficking of I541F-mutant can be rescued by lowering the temperature suggests that this mutation causes a folding defect. One of the principles of protein folding is to ensure that hydrophobic residues are buried inside the folded molecule. In the case of I541F, both isoleucine and phenylalanine are hydrophobic amino acids, but isoleucine has a branched chain, whereas phenylalanine has a bulky aromatic chain. This bulky chain may be difficult to hide inside the polypeptide. At reduced temperature, intermediate states would last longer, thus giving more chance to allow burying the bulky hydrophobic side chain of phenylalanine. Alternatively, quality control by chaperones may be less stringent at lower temperature, thus allowing the mutant to be released and to reach the plasma membrane. A positive effect of reduced temperature has already been reported in the case of other ABC-transporter mutations. At 30°C, canalicular expression and stability of the D482G-mutant bile salt export pump was improved in HepG2 cells.31 In 3T3 fibroblasts, the cystic fibrosis transmembrane regulator ΔF508-mutant partially matured after 2 days at 26°C.25 However, rescue of trafficking mutants is interesting only if they are functional. Our observation that folding of ABCB1-I541F at 27°C allows translocation of calcein indicates that, despite the mutation, the NBD domain is able to adopt a native transport-competent conformation.
The only therapy for PFIC3 is UDCA treatment. Response to UDCA therapy is variable among PFIC3 patients.4 The patient harboring the I541F mutation was not improved by UDCA therapy and required liver transplantation. In accordance with the lack of effect in vivo, treatment of MDCK-I541 cells with UDCA had no effect on the localization of the mutant (our unpublished results). Conversely, our results bring new perspectives to develop alternative therapies. The observation that the I541F-mutant is transport-competent if it successfully transits to the plasma membrane raises the possibility that strategies to influence protein folding inside cells might prove to have therapeutic value. It would therefore be interesting to identify those mutations that affect folding of the NBD-domain but not binding and hydrolysis of ATP, and research into pharmaceutical agents designed to improve folding could prove beneficial. These data open perspectives to develop novel therapeutic tools for PFIC3.
The authors thank Philippe Fontanges and Romain Morichon for help with the confocal microscope (INSERM IFR65), and Marie-Claude Gendron for cell-sorting (Institut Jacques Monod). We also thank Julie Gabillet and Sarah Gora for help in the initial phase of this work.