dRTA (distal renal tubular acidosis) and HS (hereditary spherocytosis) are two diseases that can be caused by mutations in the gene encoding the AE1 (anion exchanger 1; Band 3). dRTA is characterized by defective urinary acidification, leading to metabolic acidosis, renal stones and failure to thrive. HS results in anaemia, which may require regular blood transfusions and splenectomy. Mutations in the gene encoding AE1 rarely cause both HS and dRTA. In the present paper, we describe a novel AE1 mutation, Band 3 Edmonton I, which causes dominant HS and recessive dRTA. The patient is a compound heterozygote with the new mutation C479W and the previously described mutation G701D. Red blood cells from the patient presented a reduced amount of AE1. Expression in a kidney cell line showed that kAE1 (kidney AE1) C479W is retained intracellularly. As kAE1 is a dimer, we performed co-expression studies and found that, in kidney cells, kAE1 C479W and G701D proteins traffic independently from each other despite their ability to form heterodimers. Therefore the patient carries one kAE1 mutant that is retained in the Golgi (G701D) and another kAE1 mutant (C479W) located in the endoplasmic reticulum of kidney cells, and is thus probably unable to reabsorb bicarbonate into the blood. We conclude that the C479W mutant is a novel trafficking mutant of AE1, which causes HS due to a decreased cell-surface AE1 protein and results in dRTA due to its intracellular retention in kidney.
The human AE1 (anion exchanger 1), encoded by the SLC4A1 gene, is a dimeric or tetrameric membrane glycoprotein which exchanges bicarbonate for chloride in RBCs (red blood cells) and in type A ICCs (intercalated cells) in kidneys (kAE1) . It plays a central role for optimization of respiration by participating in CO2 removal and in the maintenance of acid–base balance in kidneys .
Naturally occurring mutations in the SLC4A1 gene can lead to HS (hereditary spherocytosis), a disease that can cause severe anaemia , or dRTA (distal renal tubular acidosis) , characterized by nephrocalcinosis, metabolic acidosis and failure to thrive. dRTA mutations can be either dominantly or recessively inherited. Patients with a different mutation in each allele are compound heterozygotes. Dominantly inherited dRTA mutants alter the normal trafficking of kAE1 WT (wild-type) in kidney cells, whereas recessively inherited mutants can have their trafficking facilitated by the co-expression with kAE1 WT [5–8].
Mutations that cause both HS and dRTA are extremely rare. RBC-specific proteins such as GPA (glycophorin A) may act as chaperones to improve eAE1 (erythroid AE1) trafficking to the RBC surface, and thus might correct targeting defects caused by dRTA mutations in RBCs. However, these proteins are absent from ICCs, thus patients develop dRTA. Patients with HS and dRTA are either homozygous or compound heterozygous. There are only two case reports of patients homozygous for AE1 mutations that cause both dRTA and HS. Band 3 Coimbra (V488M) was found in a newborn with severe HS, who also developed dRTA . The RBCs of the patient did not contain detectable AE1  and, when expressed in the kidney epithelial MDCK (Madin–Darby canine kidney) cell line, the AE1 mutant was retained intracellularly . Band 3 Courcouronnes was described in a homozygous patient carrying the S667F substitution , resulting in HS and incomplete dRTA. This mutant was retained intracellularly when expressed in MDCK cells, but was partially found at the cell surface when co-expressed with GPA in Xenopus oocytes. Compound heterozygote AE1 gene mutations have been reported previously in Asia. These patients, who usually have dRTA but no HS, carry an SAO (Southeast Asia Ovalocytosis) allele (which causes a condition where RBCs are abnormally rigid) and another mutation . Recently, a compound heterozygous patient carrying the E522K and G701D mutations was described with both HS and dRTA . In kidney epithelial cells, the E522K/G701D heterodimers were retained intracellularly, probably causing complete dRTA in that patient.
In the present paper, we report another compound heterozygote patient carrying the G701D mutation and a novel C479W mutation on the other allele. The patient displayed severe HS and complete dRTA as an infant. We describe the characterization of the molecular and cellular defects associated with the combination of these two mutations.
Part of this work was presented at the Annual American Society of Nephrology meeting held in Philadelphia, PA, U.S.A, on 4–9 November 2008.
The patient is a 19-year-old female of Caucasian Scandinavian origin. The patient was diagnosed with severe HS early after birth requiring splenectomy at 3 years of age, with an improvement in her haematological condition. Failure to thrive and nephrocalcinosis detected at 2.5 years of age led to a diagnosis of complete dRTA, which improved after prescription of daily oral sodium bicarbonate and potassium chloride. The patient did not present with auditory deficits or cognitive impairment. The father has mild HS, but no renal symptoms, and the mother is healthy. No other family member has dRTA (see Figure 1A). Protocols involved in this project have been reviewed and approved by the Health Research Ethics Board (Biomedical panel) at the University of Alberta, and written informed consent was obtained from the patient and her family.
The patient has a RBC defect and is compound heterozygous with a novel mutation in the SLC4A1 gene
Analysis of SLC4A1 mutations
Genomic DNA was isolated from blood samples drawn from the patient, her parents and two siblings. SLC4A1 was analysed by PCR and a standard dye-terminator cycle sequencing method. The segregation of mutations in the family was examined by PCR and RFLP (restriction-fragment-length polymorphism) or derived cleaved amplified polymorphism methods. Analysis indicated that the patient carries nucleotide substitutions in exons 13 (TGC>TGG) and 17 (GGC>GAC), resulting in amino acid changes C479W (Band 3 Edmonton I) and G701D (Band 3 Bangkok I) respectively (see Figure 1B). The Memphis I polymorphism (K56E), which was found previously to be associated with the G701D mutation in Southeast Asia , is not present in this family, suggesting an independent G701D mutation. To our knowledge, this case is the first reported instance of G701D in a non-Asian family.
Analysis of RBC membranes
A total of 40 μg of total RBC membrane protein per lane was loaded on to an 8% (w/v) polyacrylamide gel and separated by SDS/PAGE, followed by Coomassie Blue staining. For Western blot analysis, 10 μg of total protein per lane was loaded on to an 8% (w/v) polyacrylamide gel and separated by SDS/PAGE. After migration, proteins were transferred on to a nitrocellulose membrane, and the membrane was blocked with 3% (v/v) skimmed milk, incubated with a rabbit antibody detecting the last 15 residues of AE1 (generously given by Professor Reinhart Reithmeier, Department of Biochemistry, University of Toronto, Toronto, Canada), followed by a goat anti-rabbit antibody coupled to HRP (horseradish peroxidase) (Jackson Immunoresearch). Relative quantification of the band intensity of proteins was performed using ImageJ software.
Construction of AE1 mutants
Human kAE1 cDNA, containing the sequence encoding an HA (haemagglutinin) epitope at position 557 of the protein, and the kAE1 G701D cDNA cloned into the viral vector pFB-Neo (Stratagene) have been described previously . The kAE1-HA557 WT construct was submitted to site-directed mutagenesis using the QuikChange® site-directed mutagenesis kit (Stratagene). Using the QuikChange® site-directed mutagenesis kit, we also introduced the C479W mutation in kAE1 devoid of the HA epitope, subcloned in the pcDNA3 vector (Invitrogen) . The presence of mutations was verified by automated DNA sequencing.
Cell culture, transfections and viral infections
Viral infections were performed as described previously . Briefly, HEK-293 cells (human embryonic kidney cells) were transfected with p-VPack-GP, p-VPack-VSV-G and pFB Neo kAE1-HA557 WT or mutant plasmids using FuGENE6™ (Roche Applied Science). The cell culture supernatant containing infectious viral particles was added to dividing MDCK cells complemented with 8 μg/ml polybrene. After incubation for 24 h, cells expressing AE1 were selected with 1 mg/ml geneticin. Despite the constant presence of geneticin in the cell culture medium, the level of kAE1 protein expression in MDCK cells progressively decreased within 2–3 weeks. Consequently, all our experiments are performed within 3 weeks post-infection on a heterogeneous population of MDCK cells that have not been cloned.
Western blot analysis and enzymatic deglycosylation
MDCK cells expressing kAE1-HA557 WT or mutants were lysed in PBS containing 1% (v/v) Triton X-100 and protease inhibitors (Sigma–Aldrich), and one-tenth of cell lysate supernatant was left untreated, or digested using 1000 units of Endo-H (endoglycosidase H; New England Biolabs) or 500 units of PNGase-F (peptide-N-glycosidase F; New England Biolabs) at 37 °C for 1 h. Samples were then loaded on to an 8% (w/v) polyacrylamide gel, the proteins were separated by SDS/PAGE, transferred on to a nitrocellulose membrane and detected with a mouse anti-HA antibody (Covance), followed by a goat anti-mouse antibody coupled to HRP.
Expression and functional assay in Xenopus oocytes
eAE1 WT cloned in the pSP65 plasmid was used to replace Cys479 with a tryptophan residue by PCR with the QuikChange® site-directed mutagenesis kit (Stratagene). One positive clone was entirely sequenced before further use. A total of 10 ng of eAE1 WT or C479W mutant and 2.5 ng of GPA were co-injected, and oocytes were kept at 19 °C in modified Barth saline [85 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 0.82 mM MgSO4, 0.33 mM Ca(NO3)2, 0.41 mM CaCl2, 10 mM Hepes (adjusted to pH 7.4 with 4.5 mM NaOH) and supplemented with 10 units/ml penicillin and 10 μg/ml streptomycin]. Measurements of pHi (intracellular pH) variations were done by incubating Xenopus oocytes in buffer composed of 63.4 mM NaCl, 1 mM KCl, 24 mM HCO3−, 0.82 mM MgSO4, 0.33 mM Ca(NO3)2, 0.41 mM CaCl2 and 5 mM Hepes/NaOH (pH 7.35), 5% CO2 and 95% O2 until an equilibrium was reached. Then oocytes were bathed in a medium without chloride [63.4 mM sodium gluconate, 1 mM potassium gluconate, 24 mM HCO3−, 0.82 mM MgSO4, 0.74 mM Ca(NO3)2, 5 mM Hepes/NaOH (pH 7.35), 5% CO2 and 95% O2]. Traces are representative of pHi recordings from different oocytes expressing eAE1 WT or the C479W mutant co-expressed with GPA (the hGPA-BSXG plasmid was generously given by Dr Ashley Toye, Department of Biochemistry, University of Bristol, Bristol, U.K.) to ensure maximal expression of the mutant.
MDCK cells expressing kAE1-HA557 WT or mutant were incubated with 0.5 mg/ml EZ-link NHS-SS-biotin [sulfosuccinimidyl-2-(biotinamido)ethyl-1,3-dithiopropionate; MJS Biolynx] for 15 min at 4 °C. Cells were then washed with PBS1 [140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4 and 1.8 mM KH2PO4 (pH 7.35)] containing 0.3% BSA, lysed and, after centrifugation, an aliquot (50 μl) was saved for Western blot analysis. UltraLink Immobilized monomeric avidin (Thermo Scientific) was added to the cell lysates at 4 °C for 1 h. Eluted proteins were loaded on to 8% (w/v) polyacrylamide gels, separated by SDS/PAGE, transferred on to a nitrocellulose membrane and detected using a mouse anti-HA antibody (Covance).
MDCK cells expressing kAE1-HA557 WT or mutants were grown on glass coverslips or on semi-permeable Transwell polycarbonate filters (Corning), fixed, permeabilized or kept intact and blocked with 1% (w/v) BSA. Cells were then incubated with a mouse (Covance) or rat (Roche) anti-HA antibody, a rabbit anti-giantin antibody (Covance), a rabbit anti-calnexin antibody (generously given by Dr David Williams, Department of Biochemistry, University of Toronto, Toronto, Canada) or a rat anti-E-cadherin antibody (Sigma–Aldrich). Secondary antibodies were a goat anti-rabbit antibody coupled to Alexa Fluor® 488 (Molecular Probes), a donkey anti-rat antibody coupled to Alexa Fluor® 488 (Molecular Probes), a donkey anti-rat antibody coupled to Cy5 (indodicarbocyanine; Jackson Immunoresearch), and a goat anti-mouse antibody coupled to Cy3 (indocarbocyanine; Jackson Immunoresearch) or to Alexa Fluor® 488 (Molecular Probes). HEK-293 or MDCK cells transiently transfected with kAE1 C479W (with no HA epitope) were fixed, permeabilized and blocked with 1% (w/v) BSA. Cells were then incubated with either a mouse Bric155 antibody (IBGRL) or a rabbit antibody detecting the last 15 residues of AE1, followed by anti-mouse or anti-rabbit antibodies coupled to Cy3 (Jackson Immunoresearch). Samples were examined using an Olympus IX81 microscope equipped with a Nipkow spinning-disk optimized by Quorum Technologies.
Oocytes used for immunofluorescence detection of AE1 were from the same batches as the ones used for the pHi measurements. Oocytes were fixed for 20 min in 3% (w/v) paraformaldehyde in PBS2 [2.7 mM KCl, 100 mM Na2HPO4 and 1.8 mM KH2PO4 (pH 7.4)], then rinsed in PBS2 and incubated overnight in PBS2 containing 15% (w/v) sucrose at 4 °C. Four oocytes for each experimental condition were frozen simultaneously in tissue freezing medium (Leica Microsystems) by immersion in isopentane at −180 °C and cryosectioned (7-μm-thick; Leica CM3050).
Cryosections were placed on silane-coated glass slides (Electron Microscopy Sciences) and heated for 1 h at 50 °C. Samples were then incubated in PBS2 containing 1% (w/v) BSA before addition of antibodies against the AE1 N-terminal domain (anti-CDB3; 1:100 dilution), at room temperature (23 °C) for 1 h. Following washing with PBS2 containing 1% (w/v) BSA, a 1:1000 dilution of FITC-labelled anti-rabbit antibody was added for 1 h at room temperature. The samples were washed in PBS2 and mounted. Immunofluorescence was carried out at room temperature, and samples were viewed using an Axioplan2 Zeiss microscope.
MDCK cells expressing kAE1-HA557 WT or mutants were lysed in PBS1 containing 1% (v/v) Triton X-100 and protease inhibitors. Aliquots of the cell lysates were saved as the total fraction, and the remaining cell lysates were incubated with a rabbit anti-myc antibody (Santa Cruz Biotechnology), followed by Protein G–Sepharose. The bound AE1 proteins were eluted with Laemmli buffer before detection by Western blot analysis using a mouse anti-HA antibody (Covance).
RBCs from the patient carry a reduced amount of plasma-membrane AE1
Sequencing of the SLC4A1 gene indicated that the patient is compound heterozygous with the previously described G701D substitution (Band 3 Bangkok I)  that she inherited from her healthy mother, and a novel C479W mutation, designated as ‘Band 3 Edmonton I’. As the father who is heterozygous for the C479W novel mutation has HS only (Figure 1A), this new mutation is dominantly inherited for HS. Since the C479W mutation only causes dRTA when combined with the recessive dRTA mutation G701D , we conclude that the novel C479W mutation is recessive for dRTA. HS typically occurs when α- or β-spectrin, ankyrin, protein 4.2 or AE1 proteins are defective and reduced in RBCs . To measure the amount of plasma-membrane proteins in the RBCs of the patient, we performed SDS/PAGE, followed by staining using Coomassie Blue on ghost membranes (Figure 2A). Spectrin, and protein 4.1 and 4.2 amounts did not display an obvious reduction (Figure 2A); however, AE1 was decreased to approx. 65 and 80% of the normal amount in the RBCs of the patient and father respectively (Figure 2B). The intensity of two lower-molecular-mass bands (approx. 70 and 35 kDa as indicated by arrowheads in Figure 2A) that may correspond to AE1 degradation fragments was also reduced in the RBCs of the patient. The amount of ankyrin was also slightly decreased. The decreased amount of AE1 in RBCs of the patient and father is probably the reason for the appearance of HS symptoms.
RBCs from the patient carry reduced amount of plasma-membrane AE1
The kAE1 C479W mutant is not functional at the surface of Xenopus oocytes
To test whether the new C479W mutation affects the function of AE1, we expressed and tested the presence of functional eAE1 WT and C479W at the cell surface of Xenopus oocytes by immunostaining (Figure 3A). GPA is a protein acting as a chaperone that facilitates the surface trafficking of eAE1 in RBCs . In the absence or presence of GPA, the eAE1 C479W mutant protein was detected at the cell surface, in contrast with oocytes where no cRNA was injected or in the absence of primary antibody. We next compared the functional activity of eAE1 WT or C479W mutant expressed at the cell surface of Xenopus oocytes by measuring variations of pHi over time, in the absence or presence of extracellular chloride. If the AE1 protein is functional, switching the chloride-containing extracellular solution to one containing gluconate will result in intracellular alkalinization caused by the exchange of intracellular chloride with extracellular bicarbonate . Figure 3(B) shows that, in contrast with the eAE1 WT, there was no alkalinization in oocytes expressing eAE1 C479W, indicating that this novel mutant is inactive in Xenopus oocytes.
The C479W mutant is inactive at the surface of Xenopus oocytes
In MDCK cells, the kAE1 C479W mutant does not carry complex glycosylation
To evaluate the effect of the new C479W mutation on kAE1 biosynthesis in kidney cells, we introduced the C479W mutation into the cDNA encoding human kAE1 (kAE1 C479W) and expressed it in MDCK cells that do not express endogenous AE1 [19,20]. This construct as well as kAE1 WT contains an extracellular HA epitope at position 557, a modification that does not affect folding or trafficking of AE1 [5,20,21]. kAE1-HA557 WT (subsequently referred to as kAE1 WT) is targeted to the basolateral membrane of polarized MDCK cells [5,20], a similar location to that in type A ICCs. Immunoblots from MDCK cells indicated that kAE1 WT migrates as two major bands (Figure 4A). kAE1 WT carries a single N-glycosylation site at position 642. Figure 4(B) shows that the upper band is Endo-H-resistant (lane H), but PNGase-F-sensitive (lane F), indicating that, in MDCK cells, kAE1 WT was processed to complex oligosaccharide and had moved from the ER (endoplasmic reticulum) to the medial Golgi. Processing of kAE1 protein was efficient in MDCK cells, as 73% of kAE1 WT carried complex oligosaccharide. The lower-molecular-mass band, which was Endo-H- and PNGase-F-sensitive, corresponds to kAE1 WT in the ER carrying high mannose oligosaccharide .
In MDCK cells, the C479W mutant does not carry complex glycosylation
kAE1 G701D also carries complex oligosaccharide, but processing of kAE1 G701D in the medial Golgi was less efficient than in MDCK cells expressing kAE1 WT, as seen by the slight predominance of the core glycosylated fraction (53% of high mannose glycosylated kAE1 G701D) . In contrast, the kAE1 C479W mutant runs as a low-molecular-mass band corresponding to high mannose glycosylated kAE1 protein. These experiments suggest that the intracellular trafficking of the kAE1 C479W mutant from the ER is impaired in MDCK cells.
The kAE1 C479W mutant does not traffic to the cell surface in kidney cells
We determined the level of cell-surface expression of kAE1 WT, G701D and C479W proteins by cell-surface biotinylation (Figure 5A) and immunostaining on MDCK cells (Figures 5B and 5C). In contrast with kAE1 WT, neither the kAE1 C479W nor the G701D mutant were biotinylated and, thus, neither were detectable at the cell surface (Figure 5A). The absence of biotinylated GAPDH (glyceraldehyde-3-phosphate dehydrogenase) indicated that no biotinylation reagent had leaked inside cells during the experiment. As kAE1 G701D or C479W did not seem to reach the plasma membrane, we determined their respective intracellular location by immunostaining in MDCK cells (Figures 5B and 5C). kAE1 WT was predominantly located at the plasma membrane and displayed minimal co-localization with the ER marker calnexin (Figure 5B). In polarized MDCK cells, kAE1 WT co-localized with the basolateral marker E-cadherin (Figure 5C). The kAE1 G701D mutant showed partial co-localization with calnexin, but was predominantly retained in the Golgi, as seen by co-localization with giantin . However, the kAE1 C479W mutant co-localized with calnexin, indicating that this new mutant is retained in the ER. When expressed in HEK-293 or MDCK cells, the kAE1 C479W mutant protein devoid of the HA epitope was also retained intracellularly (results not shown), indicating that the HA epitope did not cause the intracellular retention of the kAE1 C479W mutant. Neither of the mutants co-localized with the basolateral membrane marker E-cadherin in polarized MDCK cells. Interestingly, the kAE1 C479W mutant was not present at the cell surface of kidney MDCK cells, but the eAE1 C479W mutant trafficked to the plasma membrane of Xenopus oocytes. This suggests that the cellular environment plays a role in trafficking of the C479W mutant, as was observed previously for other dRTA mutants such as eAE1 or kAE1 R589H and S613F [5,18,20], or kAE1 R901X (Walton) [5,18–20,22]. These results indicate that trafficking of the novel kAE1 C479W mutant is altered in kidney cells.
The C479W AE1 mutant does not traffic to the cell surface in MDCK cells
The kAE1 C479W mutant can physically interact with the kAE1 G701D mutant in MDCK cells
AE1 forms dimers , and individuals carrying normal and mutant AE1 or two different AE1 mutants may express AE1 heterodimers in their kidney cells. To determine the molecular mechanisms resulting in dRTA in the kidney of the patient, we co-expressed kAE1 G701D and C479W in MDCK cells, as well as kAE1 WT with either kAE1 G701D or C479W to mimic the situation in the kidney cells of the parent. To distinguish kAE1 mutant proteins from kAE1 WT, we introduced different epitopes at position 557: kAE1 G701D or kAE1 WT carried a myc epitope, whereas kAE1 C479W protein carried an HA epitope. We determined whether kAE1 WT–myc or G701D–myc protein could physically interact with kAE1 C479W–HA mutant in MDCK cells using co-immunoprecipitation. Figure 6(A) shows that kAE1 C479W mutant can physically interact with the kAE1 WT or with kAE1 G701D mutant.
The C479W mutant does not overlap with the G701D mutant, and co-expression with kAE1 WT does not rescue its trafficking to the cell surface
When co-expressed, the kAE1 C479W and G701D mutants do not overlap
We then examined the respective location of kAE1 C479W and G701D within MDCK cells by immunostaining (Figure 6B). When co-expressed in polarized MDCK cells, we observed that the kAE1 G701D and C479W mutants remained located in the Golgi and ER respectively. Neither protein was localized to the cell surface. Despite their ability to physically interact, no obvious change in the cellular location of the individual mutants was detectable in cells co-expressing both proteins (compare Figures 5B and 6B). This suggests that the kAE1 G701D and C479W mutants do not substantially overlap at sites where either protein is present in high concentrations in kidney cells, and the patient probably does not have functional kAE1 proteins at the basolateral membrane of ICCs.
We next investigated whether the co-expression of kAE1 WT and kAE1 G701D or C479W mutant could restore cell-surface trafficking of the mutant in MDCK cells. As the patient's father and mother do not display kidney symptoms, we anticipated that a sufficient amount of functional kAE1 is present at the basolateral membrane of type A ICCs in these individuals. We thus co-expressed kAE1 WT and C479W or G701D mutant in MDCK cells, and investigated whether the co-expression of kAE1 WT could restore the trafficking of the mutant to the cell surface by immunofluorescence (Figure 6C). We used the extracellular HA epitope to determine whether co-expressing kAE1 WT–myc could restore trafficking of kAE1 C479W–HA or kAE1 G701D–HA mutants to the plasma membrane of MDCK cells: blue staining corresponds to kAE1 mutant protein, whose trafficking to the cell surface has been restored by the presence of kAE1 WT, stained in red. Although kAE1 WT–myc could partially restore trafficking of kAE1 G701D–HA in MDCK cells, there was no plasma-membrane kAE1 C479W–HA mutant in MDCK cells co-expressing kAE1 WT–myc (Figure 6C), indicating that kAE1 WT cannot restore the cell-surface trafficking of the kAE1 C479W mutant.
In the present study, we report the case of a new patient with both HS and complete dRTA, who is compound heterozygous for two SLC4A1 mutations: the previously described G701D mutation and a novel C479W mutation (Figure 1C). Our data indicate that the RBCs of the patient display a 35% reduction in plasma-membrane AE1 (Figure 2), which probably results in decreased anion transport function in RBCs and in destabilizing interactions between the cytoskeleton and plasma membrane .
The G701D mutation was described previously in compound heterozygous dRTA patients from South east Asia, in combination with either SAO or other dRTA mutations [13,15,24,25]. Our patient is the second compound heterozygous individual with both HS and dRTA, as G701D was recently found in a compound heterozygous Taiwanese patient in combination with the novel E522K mutation . Interestingly, our patient has a Caucasian origin and is, therefore, the first one reported with the G701D mutation in a non-Asian background. When expressed in Xenopus oocytes, the G701D mutant was retained intracellularly unless co-expressed with GPA [13,15]. These results indicate that GPA may act as a chaperone on the trafficking of AE1 in RBCs . In contrast, our present results indicate that, in the absence or presence of GPA, the C479W mutant is targeted to the surface of Xenopus oocytes (Figure 3). However, given that the RBCs of the patient display approximately a 35% reduction in plasma-membrane AE1 (Figure 2), we hypothesize either that the cell-surface location of the mutants in Xenopus oocytes does not reflect trafficking of the mutant proteins in RBCs, or that recovery of the trafficking of the mutant protein by GPA is incomplete in RBCs. Further investigations are required to fully understand the molecular basis of HS in this patient.
The patient also developed dRTA and nephrocalcinosis at an early age. To investigate the molecular mechanisms in the development of dRTA, we expressed kAE1 WT, G701D or C479W proteins in MDCK cells. As observed previously [5,7], the kAE1 G701D mutant was predominantly located in the Golgi in MDCK cells (Figures 5B and 5C). In MDCK cells, we determined that the kAE1 C479W mutant did not carry complex oligosaccharide (Figure 4), did not traffic to the plasma membrane (Figure 5A) and predominantly co-localized with the ER marker calnexin (Figures 5C and 5D). The intracellular retention of kAE1 G701D and C479W in MDCK cells indicate that these two mutant proteins have a trafficking defect in kidney cells. As kAE1 protein is a dimer, we next mimicked the situation found in the kidneys of the patient's mother and father by co-expressing kAE1 WT with either kAE1 G701D or C479W in MDCK cells. We found that kAE1 WT could rescue the trafficking of kAE1 G701D to the cell surface as was observed previously (Figure 6C, and [5,7,27]). However, no cell-surface kAE1 C479W protein was detectable when co-expressed with kAE1 WT in MDCK cells. This result is in contrast with previous findings on other recessive dRTA mutants whose trafficking to the plasma membrane was restored by co-expression with kAE1 WT [5,7], including the recently described case of a novel compound heterozygous patient carrying the E522K and G701D mutations .
Furthermore, despite their ability to interact with each other (Figure 6A), the two kAE1 G701D and C479W mutants were both retained intracellularly and did not substantially overlap when co-expressed in polarized MDCK cells (Figure 6B). These results suggest that kAE1 G701D/C479W heterodimers may represent a minority of kAE1 dimers in cells co-expressing both proteins or that they may be prematurely degraded. Together, these data suggest that the patient, who is compound heterozygous for the G701D and C479W mutations, does not have functional kAE1 at the basolateral membrane of ICCs and, thus, developed dRTA.
The highly conserved Cys479 is one of the three cysteine residues located in the TM (transmembrane domain) of AE1 (Figure 1B) . AE1 in which all five cysteine residues were replaced by serine was functional, suggesting that a cysteine residue at position 479 is not absolutely required for the function of the protein . However, it is possible that replacing a cysteine with a bulky tryptophan residue at the membrane interface alters the structure of the region around TM3 and forces TM3 to adopt a conformation unfavourable for the function of the protein. Indeed, tryptophan residues display a striking preference for the extracellular boundary of TMs in membrane proteins [30,31]. Alternatively, the C479W substitution may stabilize TM3 and reduce its conformational stability for substrate transport as Cys479 was proposed to be part of the pore-lining helix . Consistent with this, our results indicate that, in Xenopus oocytes, the C479W mutant located at the cell surface does not alkalinize the oocyte after an acid load, in contrast with AE1 WT, suggesting that the kAE1 C479W mutant is misfolded. As for other protein-misfolding diseases, such as nephrogenic diabetes insipidus , we hypothesize that protein synthesis quality control systems recognize and retain the C479W mutant intracellularly in kidney cells, which results in dRTA.
Cys479 is located 11 amino acids away from the Arg490, which dominantly causes HS when replaced with cysteine (Band 3 Bicetre I) . In COS-7 and the erythroleukaemia cell line K562, the eAE1 R490C mutant does not traffic to the plasma membrane and is retained in the ER. Furthermore, eAE1 R490C has a dominant-negative effect on the trafficking of the eAE1 WT protein .
In conclusion, the clinical recognition that this patient was different from patients described thus far with dRTA and HS led to a multidisciplinary collaborative investigation. Our results suggest that the new C479W mutation causes major trafficking defects in kAE1 protein in both RBCs and kidney cells.
anion exchanger 1
distal renal tubular acidosis
- HEK-293 cell
human embryonic kidney cell
- MDCK cell
Madin–Darby canine kidney cell
red blood cell
Southeast Asia ovalocytosis
Carmen Chu and Naomi Woods performed the majority of the experiments under the supervision of Emmanuelle Cordat. Helene Guizouarn, Franck Borgese and Bernard Pellissier performed the Xenopus oocytes experiments. Nunghathai Sawasdee sequenced the DNA of the patient and the family under the supervision of Pa-thai Yenchitsomanus. Manjula Gowrishankar is the clinician who is treating the patient.
We thank the patient and her family for their invaluable contribution to the present study. We are also grateful to Sandra Ungarian for excellent technical assistance, and to Professor Reinhart Reithmeier and Dr Ashley Toye for providing the constructs and antibodies.
This work was supported by the Stollery Children's Hospital, the SickKids Foundation/the Canadian Institutes of Health Research [grant number XG09-016], and the Banting Research Foundation. P.Y. is supported by a Senior Research Scholar Grant from the Thailand Research Fund (TRF) and by a Pre-Clinic Staff Development Fund from the Faculty of Medicine Siriraj Hospital, Mahidol University. E.C. is supported by an AMGEN Western Canadian Kidney Research Senior Fellowship.
These authors contributed equally to this work.