Trafficking defects of a novel autosomal recessive distal renal tubular acidosis mutant (S773P) of the human kidney anion exchanger (kAE1).

Autosomal dominant and recessive distal renal tubular acidosis (dRTA) can be caused by mutations in the anion exchanger 1 (AE1 or SLC4A1) gene, which encodes the erythroid chloride/bicarbonate anion exchanger membrane glycoprotein (eAE1) and a truncated kidney isoform (kAE1). The biosynthesis and trafficking of kAE1 containing a novel recessive missense dRTA mutation (kAE1 S773P) was studied in transiently transfected HEK-293 cells, expressing the mutant alone or in combination with wild-type kAE1 or another recessive mutant, kAE1 G701D. The kAE1 S773P mutant was expressed at a three times lower level than wild-type, had a 2-fold decrease in its half-life, and was targeted for degradation by the proteasome. It could not be detected at the plasma membrane in human embryonic kidney cells and showed predominant endoplasmic reticulum immunolocalization in both human embryonic kidney and LLC-PK1 cells. The oligosaccharide on a kAE1 S773P N-glycosylation mutant (N555) was not processed to the complex form indicating impaired exit from the endoplasmic reticulum. The kAE1 S773P mutant showed decreased binding to an inhibitor affinity resin and increased sensitivity to proteases, suggesting that it was not properly folded. The other recessive dRTA mutant, kAE1 G701D, also exhibited defective trafficking to the plasma membrane. The recessive kAE1 mutants formed dimers like wild-type AE1 and could hetero-oligomerize with wild-type kAE1 or with each other. Hetero-oligomers of wild-type kAE1 with recessive kAE1 S773P or G701D, in contrast to the dominant kAE1 R589H mutant, were delivered to the plasma membrane.

The kidney plays an essential role in maintaining the acidbase balance of the body by reabsorbing bicarbonate in the proximal tubule and secreting acid into the urine. Impairment in acid secretion by ␣-intercalated cells of the distal nephron leads to the development of distal renal tubular acidosis (dRTA), 1 which is inherited in both an autosomal dominant (AD dRTA) (1-7) and autosomal recessive manner (AR dRTA) (4, 8 -10). Patients with AD dRTA usually remain asymptomatic until adulthood, whereas AR dRTA patients are severe cases and early onset (11)(12)(13). Recently, a number of missense, nonsense, and deletion mutations in the anion exchanger 1 (AE1 or SLC4A1) gene resulting in both AD and AR dRTA have been identified and characterized (1)(2)(3)(4)(5)(6)(7). Human AE1 is a red blood cell membrane glycoprotein containing 911 amino acids (14,15) with a large N-terminal cytoplasmic domain of about 400 amino acids and a C-terminal domain consisting of 500 amino acids, which forms 12 transmembrane spans. A short C-terminal cytoplasmic tail (16) provides a binding site for carbonic anhydrase II (17) and is important for the proper trafficking of the protein to plasma membrane (18). Kidney AE1 (kAE1) is transcribed from an alternative promoter (19,20) resulting in the production in human kidney of an isoform that is missing the N-terminal 65 amino acids found in erythroid isoform (eAE1) (21). KAE1 is located in the basolateral membrane of acid-secreting ␣-intercalated cells where it mediates the transport of bicarbonate into the blood (22).
The first reported AD dRTA AE1 mutations were nucleotide substitutions causing an amino acid change of residue 589 from arginine to histidine (R589H) (1)(2)(3), to cysteine (R589C) (1,2), or to serine (R589S) (2). The dominant kAE1 R589H mutant was functional as assessed in transport assay in Xenopus oocyte (3) but was retained in the endoplasmic reticulum in transfected HEK cells (23) and in both nonpolarized and polarized MDCK cells (24). Similarly, a dominant dRTA mutant (kAE1 R901Stop) missing the C-terminal 11 amino acids had normal chloride transport activity in Xenopus oocytes but was retained intracellularly in nonpolarized MDCK cells (25) and HEK cells (26). The kAE1 901Stop mutant was also mis-routed to the apical membrane in polarized MDCK cells (24,27). The R589H and R901Stop mutants could hetero-oligomerize with wild-type kAE1 causing a dominant-negative effect that retained the wild type AE1 protein intracellularly (26). Another recently described dominant mutation at glycine 609 (G609R) had normal transport activity in Xenopus oocyte but was also mis-localized to the apical membrane in polarized cells (7). Taken together, these studies show that dominant AE1 dRTA mutants are functional, impaired in their exit from the ER, and are mislocalized to the apical membrane in polarized cells. The dominant phenotype is because of hetero-oligomer formation and intracellular retention of the wild-type protein.
Recessive dRTA mutations were reported in the studies of patients from Thailand (8 -10), Malaysia, and Papua-New Guinea (4). The first mutation found in two affected siblings was a homozygous substitution of glycine at position 701 of aspartic acid (G701D) (8). AE1 G701D was also found in compound heterozygosity with an in-frame deletion of 9 amino acids (residues 400 -408) of AE1 called Southeast Asian ovalocytosis (SAO) (4,9). AE1 SAO is present in the plasma membrane of red cells but is unable to transport anions (28,29). AE1 G701D could not reach the cell surface in Xenopus oocytes but was rescued when coexpressed with glycophorin A (GPA) (8). GPA is a type I single-span membrane glycoprotein found in the erythrocyte membrane. It associates with AE1 during biosynthesis and promotes AE1 trafficking to the plasma membrane (30). Therefore, AE1 G701D may fold and traffic normally in the presence of GPA in the red cell precursors but not in kidney cells that lack GPA (31). Furthermore, two AE1 mutations, A858D and ⌬V850, in compound heterozygosity with SAO (A858D/SAO or ⌬V850/SAO) or each other (A858D/ ⌬V850) were also found to cause recessive dRTA (4). The AE1 A858D mutant had low Cl Ϫ uptake activity in Xenopus oocytes but had greatly decreased transport activity when coexpressed with AE1 SAO or AE1 ⌬V850. Transport function could only be partially rescued by GPA (4). Similarly, when AE1 ⌬V850 was expressed in Xenopus oocytes by itself or with AE1 SAO, chloride uptake was markedly decreased and was incompletely rescued by GPA (4).
In this paper we characterized a novel missense AR dRTA mutation in kAE1 resulting in a change of serine at position 773 to proline (S773P), found in a family in the Northeast Thailand. The subject is a compound heterozygote for S773P and G701D and has dRTA as well as rickets and bilateral nephrocalcinosis but does not have hearing loss or anemia. Parents are asymptomatic and normal in acid loading tests. After PCR single-stranded conformation polymorphism screening and sequencing for SLC4A1 mutation, the G701D mutant was reported in the mother, whereas the novel S773P mutant was found in the father (32). In order to understand the molecular basis of dRTA caused by this mutation, we generated the kAE1 S773P mutant and analyzed its biosynthesis and trafficking in transfected HEK and LLC-PK1 cells either alone to mimic the homozygous state, in combination with wild-type kAE1 to mimic the parental state, or with the recessive G701D mutant to mimic the heterozygous state found in the subject.

EXPERIMENTAL PROCEDURES
Construction of Plasmids and Mutations-Human AE1 cDNA (a generous gift of Drs. A. M. Garcia and H. Lodish, Whitehead Institute) was inserted into the HindIII and BamHI sites of pcDNA3 (Invitrogen). The kidney form (kAE1) (23), N555 mutant (33), kAE1 tagged at the C terminus with His 6 (kAE1 His 6 ) (23), and AE1 HA557 containing an extracellular hemagglutinin (HA) epitope inserted in position 557 were constructed as described previously (18). The kidney isoform of Southeast Asian ovalocytosis (kAE1 SAO) was a generous gift from Joanne Cheung (University of Toronto). Both kAE1 S773P and kAE1 S773P tagged at their C termini with His 6 were generated using the Quick-Change TM site-directed mutagenesis kit from Stratagene (La Jolla, CA) and oligonucleotide primers from ACGT Corp. (Toronto, Canada). The mutations were confirmed by sequencing performed by ACGT Corp. (Toronto, Canada). Plasmid DNA for transfections was purified with Qiagen (Valencia, CA) Plasmid Midi columns.
Cell Culture-HEK-293 (human embryonic kidney) cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and 0.5% penicillin and streptomycin (Invitrogen) in 5% CO 2 at 37°C. The day before transfection the cells were trypsinized and replated in either 6-well or 35-mm 2 plates. Either DEAE-dextran method (34) or LipofectAMINE 2000 (Invitrogen) was used to transiently transfect the cells (1 g of DNA per well of a 6-well dish). LLC-PK1 cells were grown in DMEM/F-12 with 10% calf serum and 0.5% penicillin and streptomycin (Invitrogen) in 5% CO 2 at 37°C and transfected as described above.
Western Blotting-SDS-PAGE and Western blotting were used to analyze protein samples (35). KAE1 and mutant proteins were detected using a rabbit antiserum against the C-terminal 15 amino acids of AE1 (anti-Ct AE1). This antibody does not recognize the C-terminal tail of His 6 -tagged kAE1 protein, which could be detected by a rabbit antiserum raised against the 15 N-terminal residues of kAE1 (anti-Nt kAE1). A horseradish peroxidase-conjugated goat anti-rabbit antibody (Cell Signaling Technology, Beverly, MA) was then used to probe for the primary antibody. 1:5000 dilution of mouse anti-GAPDH (Chemicon, Temecula, CA) was used to detect the level of endogenous GAPDH expression in HEK cells. BM Chemiluminescence Blotting Substrate (Roche Applied Science) and exposure to BioMax MR film (Eastman Kodak Co.) were used to detect AE1 protein bands. Densitometric analysis to determine the relative amounts of AE1 was performed by using NIH Image 1.62 software (National Institutes of Health).
Pulse-Chase Assay-The day after transfection, cells (ϳ60 -80% confluent, in 35-mm 2 dishes as described previously) were starved in methionine-free DMEM (Invitrogen) for 30 min and pulsed with 100 Ci/ml of L-[ 35 S]methionine (PerkinElmer Life Sciences) for 20 min to label newly synthesized protein. After labeling, the medium was removed and replaced with DMEM supplemented with 10% calf serum and 0.5% penicillin and streptomycin. In some cases, the proteasome inhibitor, MG262 (Biomol, Hamburg, Germany) (1 M), was added to the cells for 30 min before pulsing. At each chase time point, the cells were lysed in 1 ml of RIPA buffer (1% deoxycholic acid, 1% Triton X-100, 0.1% SDS, 0.15 M NaCl, 1 mM EDTA, 10 mM Tris-HCl, pH 7.5) containing protease inhibitors: 100 M PMSF (Sigma), 1 M aprotinin (Roche Applied Science), 1 M leupeptin (Roche Applied Science), 1 M pepstatin (Roche Applied Science). The wild-type and mutant kAE1 were immunoprecipitated with 4 l of anti-Ct AE1 antibody at 4°C for 1 h, followed by 30 l of protein G-Sepharose (Amersham Biosciences) at 4°C for 1 h. The beads were washed three times with 500 l of RIPA buffer, and proteins were eluted with 30 l of 2ϫ sample buffer. The radiolabeled proteins were analyzed by electrophoresis using 8% SDS-PAGE. The gels were fixed, dried, and exposed to film. The relative amounts of kAE1 were determined using NIH Image 1.62 software as mentioned above.
Processing of the N-linked oligosaccharide was determined using kAE1 N555 construct (33). After immunoprecipitation using anti-Ct AE1 antibody, the high mannose and complex oligosaccharides were distinguished by endo H digestion (33).
Cell Surface Biotinylation-The cells were washed with borate buffer (10 mM boric acid, 154 mM NaCl, 7.2 mM KCl, 1.8 mM CaCl 2 , pH 9.0) and treated twice with 1 ml of 0.5 mg/ml EZ-Link NHS-SS-Biotin (Pierce) in borate buffer for 15 min at room temperature (36). Unreacted reagent was then quenched by rinsing the cells with 0.192 M glycine, 25 mM Tris, pH 8.3. Cells were lysed with 1 ml of RIPA buffer with protease inhibitors, and an aliquot of the lysate was taken for Western blotting. ImmunoPure (Pierce) immobilized streptavidin (100 l) was added to the lysate for 1 h at 4°C to bind the biotinylated proteins. The supernatant was removed, and an aliquot (unbound fraction) was taken for Western blotting. The streptavidin beads were washed three times with RIPA buffer. SDS sample buffer (containing 5% ␤-mercaptoethanol) was added to the beads (bound fraction), and the samples were incubated for 1 h at room temperature to cleave the disulfide bond in the biotinylating reagent and release the captured proteins. Samples were analyzed by SDS-PAGE and Western blotting using the anti-Ct AE1 antibody. The blot was also incubated with mouse anti-tubulin antibody to confirm that biotinylation was solely restricted to the cell surface.
Immunofluorescence-HEK and LLC-PK1 were grown on glass coverslips. The transiently transfected HEK or LLC-PK1 cells were rinsed twice in PBS containing 1 mM CaCl 2 and 1 mM MgCl 2 . The cells were fixed with 3.7% paraformaldehyde for 10 min, washed with 100 mM glycine, and permeabilized in 0.1% Triton X-100 for 15 min. After blocking nonspecific binding with 1% BSA, the cells were incubated either with a 1:1,000 dilution of a rabbit anti-Ct AE1 antibody or with a 1:1,000 dilution of a mouse anti-HA antibody (Covance Inc, Princeton, NJ) in PBS containing 1% BSA for 30 min at room temperature. After several washes, the coverslips were incubated with a 1:1,000 dilution of Cy3-conjugated anti-rabbit IgG (Jackson ImmunoResearch) and goat anti-mouse Alexa488 (1:1,000) (Molecular Probes, Eugene, OR) for 30 min at room temperature. The coverslips were washed with PBS and mounted before observation using a laser scanning confocal Zeiss LSM 510 microscope.
SITS-Affi-Gel Binding Assay-Transfected cells were lysed in 500 l of PBS containing 1% C 12 E 8 (Nikkol, Tokyo, Japan) and protease inhibitors (PMSF, leupeptin, pepstatin A, and aprotinin) at 4°C for 30 min. The lysates were centrifuged (13,000 ϫ g) to remove insoluble material. 25 l of SITS-Affi-Gel 102 prepared as described previously (37) and 100 l of lysate were combined in 1% C 12 E 8 , 253 mM sodium citrate buffer, pH 7.1, with or without 1 mM free anion transport inhibitor, H 2 DIDS (Sigma), in a 1-ml total volume. The mixture was incubated at 4°C for 30 min. Resin was collected by centrifugation (8000 ϫ g) for 5 s and washed three times with 0.1% C 12 E 8 , 228 mM sodium citrate buffer, pH 7.1. To elute bound proteins from the SITS-Affi-Gel, 75 l of 2ϫ sample buffer was added to the SITS-Affi-Gel beads. The total and bound fractions were analyzed for kAE1 content by Western blotting using anti-Ct AE1 antibody.
Proteolysis Assay-Transfected HEK cells expressing either kAE1, kAE1 S773P, kAE1 R589H, or kAE1 SAO were lysed in 1% C 12 E 8 without any protease inhibitors at 4°C for 30 min. The 2 l of 1, 10, or 100 g/ml dilution of trypsin or chymotrypsin were added to 20 l of protein samples. All samples were digested at 4°C for 1 h, and the enzyme was inhibited by addition of 2 l of 40 mM PMSF protease inhibitor. To visualize all fragments generated from the enzyme digestion, SDS-PAGE and Western blot were used. Molecular weight markers were ␤-galactosidase (116.0 kDa), bovine serum albumin (66. Transport Assay-KAE1 WT and kAE1 S773P were tested for the chloride/bicarbonate exchange activity in transfected HEK cells (38). Briefly, HEK-293 cells were grown on poly-L-lysine-coated coverslips. Two days after transfection, the coverslips were incubated with serumfree DMEM containing 2 M 2Ј,7Ј-bis(carboxyethyl)-5,6-carboxyfluorescein-AM at 37°C for 30 min. Coverslips were mounted in the fluorescence cuvette, which was perfused with Ringer buffer (5 mM glucose, 5 mM potassium gluconate, 1 mM calcium gluconate, 1 mM MgSO 4 , 2.5 mM NaH 2 PO 4 , 25 mM NaHCO 3 , 10 mM Hepes, pH 7.4) with 140 mM NaCl or 140 mM sodium gluconate continuously bubbled with air/CO 2 . The fluorescence changes were monitored using a fluorimeter Photon technology international (London, Ontario, Canada) at excitation at 440 and 502.5 nm and emission at 528.7 nm. The fluorescence data were converted to pH values by using the nigericin/high potassium calibration. The transport rates were measured by linear regression of the initial H ϩ equivalent flux using Kaleidagraph software (Synergy Software, Reading, PA) (38).
Copurification of kAE1 Mutants with His-tagged kAE1-In order to determine whether mutant kAE1 can oligomerize with wild-type or mutant kAE1 proteins, a C-terminal His 6 -tagged version of kAE1 (kAE1 His) was used. Cells were either transfected with kAE1 His alone or cotransfected with kAE1 His in combination with kAE1, the dominant kAE1 R589H mutant, or the recessive kAE1 S773P and kAE1 G701D mutants. Transfected cells were lysed in 500 l of PBS, pH 7.4, containing 1% C 12 E 8 , 5 mM imidazole, and protease inhibitors at 4°C. After removing the insoluble material by centrifugation, 400 l of cell lysate was incubated with 40 l of ProBond TM nickel beads (Invitrogen) for 1 h at 4°C. Beads were washed three times with 0.5 ml of PBS, pH 7.4, 0.1% C 12 E 8 , 40 mM imidazole. Bound proteins were eluted with the same buffer containing 500 mM imidazole. KAE1 in the eluant was detected by Western blotting as described above. KAE1 His is not recognized by the anti-C-terminal antibody but can be detected by an anti-N-terminal (anti-Nt AE1) antibody. Also, kAE1 S773P His was constructed to test for oligomerization with itself and with kAE1 WT and kAE1 G701D. No interaction was detected by mixing samples prepared from two batches of transfected cells. Control experiments showed no binding of untagged constructs to the Ni 2ϩ resin.
Biotinylated Histidine-tagged Copurification-HEK cells were cotransfected with eAE1 His and either kAE1, recessive S773P, G701D, or dominant R589H. To biotinylate the cell surface proteins, the cells were treated twice with 1 ml of 0.5 mg/ml EZ-Link NHS-LC-Biotin (Pierce) in borate buffer for 15 min at 4°C following by quenching buffer as described above (36). Cells were lysed in 1% C 12 E 8 in PBS containing 5 mM imidazole, centrifuged, and incubated with ProBond TM nickel beads (Invitrogen) for 1 h at 4°C to capture eAE1 His and any associated proteins as described previously. Erythroid AE1 and any associated kAE1 wild-type or mutant protein was eluted in 500 mM imidazole and detected by Western blotting. Anti-Nt eAE1 and kAE1 antibodies were used to detect AE1 in both erythroid and kidney isoforms in total cell lysate, whereas anti-biotin horseradish peroxidase (Cell Signaling Technology, Beverly, MA) was used to visualize biotinylated AE1. The anti-Ct AE1 antibody detected kAE1 or mutant kAE1 that was associated with eAE1 His. The erythroid version of AE1 His was used in these experiments to allow detection of both eAE1 and kAE1 in the immunoblot.
Fluorescence Activated Cell Sorting-Transfected HEK cells expressing either kAE1 HA557 or kAE1 S773P HA557 alone and in combination with wild-type or S773P, G701D, and dominant R589H mutants were trypsinized for 5 min at 4°C, centrifuged to collect the cells, and resuspended in Hanks' solution with 1% BSA. The cells were incubated with 1:1,000 dilution of mouse anti-HA antibody in Hanks' solution with 1% BSA for 15 min, and then the cells were stained with goat anti-mouse Alexa488 (1:1,000) (Molecular Probes, Eugene, OR) for 15 min on ice. The samples were washed and analyzed using Beckman-Coulter EPICS Elite (BD Biosciences), and the percent of positive cells was determined to quantify the level of the cell surface expression. Fig. 1A shows the location of mutations that are characterized as causing autosomal dominant (filled star) or recessive (open star) distal renal tubular acidosis (AD and AR dRTA) in a folding model of human AE1. The dRTA mutations are distributed throughout the membrane domain of kAE1 but are absent from the N-terminal cytosolic domain (1-10). Here we characterized a novel recessive kAE1 S773P mutant found as a compound heterozygote with kAE1 G701D in a patient from northeastern Thailand. This mutation is located within the tenth transmembrane segment in the C-terminal region of the protein. Most interestingly, serine 773 is highly conserved among the anion exchanger family members, and introduction of a proline residue within a transmembrane segment may be expected to alter the structure of this region of the protein.

RESULTS
Expression of kAE1 S773P in Transfected HEK-293 Cells-To determine the effect of the S773P mutation on the biosynthesis of kAE1, we expressed kAE1 and kAE1 S773P in transiently transfected HEK cells. KAE1 was expressed as a 96-kDa band as shown in Fig. 1B. The kAE1 S773P mutant was expressed at a much lower level (34 Ϯ 5%, n ϭ 3 Ϯ S.D.) relative to normal kAE1 (Fig. 1B, lane 2). The level of expression was normalized to endogenous GAPDH expressed in HEK cells (Fig. 1B, lower panel).
Biosynthesis and Turnover of kAE1 S773P-In order to investigate the rates of biosynthesis and turnover of kAE1 S773P, pulse-chase experiments using [ 35 S]methionine were performed (Fig. 2). The kAE1 protein persisted over 24 h and exhibited a half-life of ϳ11 h, whereas mutant kAE1 S773P demonstrated a shorter half-life of about 5 h. Hence, the reduction in kAE1 S773P expression as shown by immunoblot (Fig.  1B) was likely caused by the more rapid turnover of kAE1 S773P. This is in contrast to the AD dRTA mutant R589H, which exhibited a normal half-life and expression level in transfected HEK cells (23).
Mutant membrane proteins are often misfolded and targeted to the proteasome where they are degraded (39). To investigate the degradation pathway of kAE1 S773P, pulse-chase experiments were performed in the presence and absence of the proteasome inhibitor MG262 (1 M) (Fig. 2C). At the 5-h chase time point, about 81% of kAE1 WT remained. There was a small effect of the proteasome inhibitor on the amount of kAE1 WT remaining at this time point (85%) compared with the untreated sample. In contrast, only 60% of kAE1 S773P remained at the 5-h chase point, and this protein was protected from degradation by treatment of cells with MG262 since 79% of the protein remained at the 5-h chase time. This indicates that degradation of the kAE1 S773P mutant is mediated by the proteasome.
Oligosaccharide Processing of the kAE1 S773P-We then studied the trafficking of the protein from the ER through the Golgi by following oligosaccharide processing. AE1 maintains a high mannose oligosaccharide in transfected HEK cells although it traffics to the plasma membrane. Li et al. (33) found that by moving the acceptor site of N-linked glycosylation from the endogenous site at Asn-642 to position 555 in the extracellular loop 3, AE1 is then processed from high mannose to complex oligosaccharide. This processing can be used to conveniently monitor protein trafficking as the glycoprotein moves from the ER to the medial Golgi. Although kAE1 N555 was processed from high mannose (open circle) to complex oligosaccharide (filled circle) (Fig. 1B, lane 3), kAE1 S773P N555 did not show the presence of a higher molecular weight band containing a complex oligosaccharide (Fig. 1B, lane 4), indicating impaired trafficking from the ER to the Golgi. Again, the level of expression of the S773P N555 mutant was lower (26 Ϯ 4%, n ϭ 3 Ϯ S.D.) than the normal protein. This suggests that the kAE1 S773P mutant is less stable than the normal protein, and its oligosaccharide processing is different from the wild-type kAE1.
As mentioned above, kAE1 N555 can be used to monitor the rate of glycoprotein processing and trafficking by pulse-chase experiments. At the beginning of the chase period, both kAE1 N555 and kAE1 N555 with the S773P mutation showed a single endo H-sensitive band indicating a high mannose oligosaccharide (Fig. 3A, open circle). After a 2-h chase, kAE1 N555 conversion of the high mannose to a higher molecular weight complex form that was resistant to endo H digestion was apparent (Fig. 3A, filled circle). The time course of conversion of the high mannose to the complex form (Fig. 3B) indicates that ϳ50% of the radioactivity in pulse-labeled kAE1 was present in the upper complex band by 5 h of chase time. This means that kAE1 acquired complex oligosaccharide and had moved from the ER to the medial Golgi. On the other hand, the mutant kAE1 S773P N555 did not acquire complex oligosaccharide over the same chase period (Fig. 3, A and B). This indicates that kAE1 S773P N555 did not traffic from the ER to the medial Golgi. The kAE1 S773P N555 mutant was more rapidly degraded than kAE1 N555, with no detectable protein remaining after the 24-h chase period.
Cell Surface Biotinylation of kAE1 S773P-To determine the cell surface expression of kAE1 in transfected HEK cells, biotinylation of intact cells was performed. KAE1 expressed in transfected cells could be biotinylated by NHS-SS-biotin (Fig.  4, lane B). Moreover, the diminished unbiotinylated fraction (Fig. 4, lane U) compared with total fraction (lane T) indicated the presence of kAE1 at the cell surface, determined to be 21 Ϯ 2% (n ϭ 5 Ϯ S.D.) of the total kAE1 expression. We could not detect either the recessive kAE1 S773P or G701D mutants in the biotinylated fraction (2.5 Ϯ 2 and 0.5 Ϯ 0.4%, respectively, n ϭ 5 Ϯ S.D.) (Fig. 4, lane B). There was no protein in the bound fraction of control cells expressing kAE1 that were not treated with biotinylating reagent, indicating no binding of the unbiotinylated protein to the resin. No biotinylation of intracellular tubulin was detected indicating that the labeling was restricted to the cell exterior (data not shown).
Immunofluorescence Localization-We next determined the subcellular location of the recessive kAE1 S773P and G701D mutants in transfected HEK and LLC-PK1 cells by immunolocalization. Confocal images of transiently transfected HEK cells showed that kAE1 was present at the cell membrane, as well as in intracellular compartments (Fig. 5A) as described previously (23). In contrast, kAE1 S773P did not show plasma membrane staining and was localized intracellularly (Fig. 5C). The intracellular protein colocalized with calnexin, an ER marker (data not shown). The intracellular retention of kAE1 S773P in the ER was confirmed in transiently transfected nonpolarized LLC-PK1 cells, which exhibited a better defined cell morphology than HEK cells (Fig. 5D). The kAE1-transfected LLC-PK1 cells showed a clear plasma membrane staining (Fig. 5B) that was absent in cells transfected with kAE1 S773P (Fig. 5D). The kAE1 G701D recessive mutant was also localized intracellularly in both HEK (Fig. 5E) and nonpolarized LLC-PK1 cells, with a perinuclear staining pattern suggesting Golgi localization (Fig. 5F).
Transport Activity and Inhibitor Binding-Measurements of the anion transport activity were carried out in the transfected HEK cells to determine the Cl Ϫ /HCO 3 Ϫ transport rate of kAE1 at the plasma membrane (38). Whereas normal kAE1 showed The bound fraction of biotinylated kAE1 was 15 times overloaded with respect to the total and supernatant to be in the linear range of the chemiluminescence exposure. The position of the 96-kDa kAE1 is indicated by an arrow. The percentage of the cell surface expression was quantitated by scanning the total and supernatant samples. The pixel density was analyzed using NIH Image 1.62f (n ϭ 5). robust transport activity (31 Ϯ 6 mM H ϩ /min, n ϭ 3 Ϯ S.D.), the recessive S773P mutant had no significant transport activity (5 Ϯ 2 mM H ϩ /min, n ϭ 3 S.D.) over pcDNA3-transfected cells (4 Ϯ 2 mM H ϩ /min, n ϭ 3 Ϯ S.D.), which is consistent with a lack of expression at the cell surface. We also found that the recessive G701D mutant showed no transport activity (5 Ϯ 3 mM H ϩ /min, n ϭ 3 Ϯ S.D.) in transfected HEK cells. These results confirm the lack of functional expression of the recessive dRTA mutants at the level of the plasma membrane in transfected HEK cells.
To test whether kAE1 S773P and G701D were properly folded, we employed an inhibitor affinity resin, 4-acetamido-4Јisothiocyanostilbene-2,2Ј-disulfonate (SITS)-Affi-Gel (37,40). This binding assay provides a measure of the integrity of the inhibitor binding site of the expressed kAE1 protein, including the intracellular pool present in solubilized cell extracts. The immunoblot showed that the majority of wild-type kAE1 bound to SITS-Affi-Gel (85 Ϯ 3%, n ϭ 5 Ϯ S.D.), under the conditions employed, and that this binding was efficiently blocked by the free inhibitor H 2 DIDS (Fig. 6A, lane D). On the contrary, the kAE1 SAO mutant, which does not bind inhibitors (40), bound very poorly to SITS-Affi-Gel relative to kAE1 (15 Ϯ 2%, n ϭ 5 Ϯ S.D.). The amount of mutant kAE1 S773P bound to SITS-Affi-Gel was lower than wild type (35 Ϯ 8%, n ϭ 5 Ϯ S.D.). The kAE1 G701D also bound to SITS-Affi-Gel less well (63 Ϯ 11%, n ϭ 4 Ϯ S.D.) than the wild-type protein. The results suggest that the recessive mutations induce a change in the inhibitorbinding site of the protein, resulting in poorer binding to the affinity resin, although binding is not completely eliminated.
In order to further examine the folding of the mutant proteins, we determined their sensitivity to proteolysis by trypsin and chymotrypsin in detergent extracts of transfected cells. Both kAE1 S773P and G701D were more sensitive to proteolytic cleavage than the wild type and generated a different fragment pattern (data not shown). The impaired binding to an inhibitor affinity matrix and increased protease sensitivity suggest that the recessive mutants kAE1 S773P and G701D were not properly folded. This is in contrast to dominant dRTA mutant (kAE1 R589H), which bound to SITS Affi-Gel at the same level as wild-type kAE1 (23) and had an unaltered protease sensitivity compared with wild-type kAE1 (data not shown).
Native Gel Electrophoresis-AE1 exists as dimers in erythroid cells (41) and in transfected HEK cells (23,42). Native gel electrophoresis using the detergent PFO was used to determine the oligomeric state of kAE1 and kAE1 mutants in transfected cells. It was shown previously that PFO preserves the oligomeric state of native AE1 (43). Cell extracts were prepared by solubilizing transfected HEK cells with PFO. The protein was resolved on native PFO gels, and the position of kAE1 relative to native protein markers was determined by immunoblotting. KAE1, S773P, and R589H mutants all ran as dimers (180 kDa) with no detectable monomer in the PFO system (data not shown). Thus, the recessive mutations do not prevent kAE1 from forming a dimer.
Copurification of kAE1 Mutant with Histidine-tagged kAE1-Because kAE1 and the mutant can form dimers in transfected cells, coexpression studies were carried out to determine whether kAE1 and the recessive kAE1 mutants could heterodimerize. In order to carry out this experiment, a Cterminal 6-histidine-tagged kAE1 (kAE1 His) was created (23). The 6-histidine tag interferes with the recognition epitope of anti-Ct AE1; therefore kAE1 His, in contrast to kAE1, is unable to be detected by the antibody against the C terminus. All mutants could be detected by using an anti-N-terminal antibody. KAE1 His was expressed alone or in combination with kAE1 WT, dominant kAE1 R589H, or recessive kAE1 S773P or G701D (Fig. 7A). As expected, kAE1 His could be detected by the anti-Nt AE1 antibody (Fig. 7A, lane 2, top panel) but not by the anti-Ct AE1 antibody (Fig. 7A, lane 2, lower panel). Thus, any band detected in the eluant by the anti-Ct AE1 indicates the presence of the non-His-tagged kAE1 protein. Coexpression of kAE1 His with kAE1 and mutant proteins revealed that the wild-type protein could form hetero-oligomers with all of the mutant proteins. KAE1 alone does not bind to the Ni 2ϩ beads; thus, no band was detected by antibody against the N terminus or against the C terminus (Fig. 7A, lane 1, upper and lower  panels).
To understand the mechanism of dRTA in this patient, who is compound heterozygous for recessive S773P and G701D, we also studied the interaction between kAE1 S773P and G701D by using histidine-tagged copurification (Fig. 7B). KAE1 S773P with a C-terminal 6-histidine tag was created. We found that kAE1 S773P His can form a hetero-oligomer not only with kAE1, confirming the data presented above, but also with kAE1 G701D. Thus, the mutant proteins were not only capable of interacting with the wild-type protein but also with each other, indicating that heterodimers of kAE1 S773P and G701D can form.
Cell Surface Expression of Hetero-oligomer by Cell Surface Biotinylation and Histidine-tagged Copurification-The dominant R589H mutant retained wild-type kAE1 proteins intracellularly because of hetero-oligomerization (23). In order to elucidate whether the recessive S773P and G701D mutants had the same effect on trafficking of kAE1 to the plasma membrane, we used a biotinylation histidine-tagged copurification assay. Wild-type eAE1 His and mutant kAE1 were coexpressed, and cell surface protein was biotinylated. KAE1 His and associated proteins were then purified by nickel affinity chromatography. In these experiments, eAE1 His was used in order to resolve it from associated kAE1 mutants during SDSgel electrophoresis. Any kAE1 protein associated with eAE1 FIG. 7. Histidine-tagged kAE1 copurification of kAE1 and recessive mutant S773P. A, cell lysates from HEK cells cotransfected with 0.5 g of kAE1 His cDNA per well with the same amount of kAE1, recessive kAE1 S773P, kAE1 G701D, or dominant kAE1 R589H (0.5 g of DNA). KAE1-His and associated proteins were then copurified by nickel-affinity resin and eluted in 500 mM imidazole. The upper panel shows all kAE1 proteins expressed in the eluted fraction (kAE1 His and coexpressed protein) probed by anti-Nt AE1 antibody. Anti-Ct AE1 detects kAE1 and mutant kAE1 but not kAE1-His in the bound fraction as shown in the lower panel of the immunoblot. This is representative of five independent experiments. B, cell lysates from HEK cells cotransfected with 0.5 g of kAE1 S773P His cDNA per well with the same amount of kAE1, recessive kAE1 S773P, or kAE1 G701D (0.5 g of DNA). As described previously, the upper panel shows all kAE1 proteins expressed in the bound fraction (kAE1 S773P His and coexpressed protein) probed by anti-Nt AE1 antibody. The associated kAE1 and mutant kAE1 but not kAE1 S773P-His in the bound fraction can be detected by anti-Ct AE1 antibody as shown in the lower panel of the immunoblot. This is representative of five independent experiments.
FIG. 8. Detection of recessive kAE1 S773P with wild-type AE1 at the plasma membrane by cell surface biotinylation. Cell lysates from HEK cells cotransfected with 0.5 g of eAE1 His cDNA per well with the same amount of kAE1, recessive kAE1 S773P, kAE1 G701D, or dominant kAE1 R589H (0.5 g of DNA). After surface biotinylation with NHS-LC-biotin reagent, eAE1-His and associated proteins were then copurified by nickel-affinity resin and eluted in 500 mM imidazole. Immunoblots probed by both anti-Nt eAE1 and anti-Nt kAE1 antibodies revealed the AE1 proteins coexpressed in the bound fraction. The anti-biotin-horseradish peroxidase was used to detect the surfacebiotinylated proteins in the bound fraction as shown in the lower panel of the immunoblot. This is representative of three independent experiments.
His was detected by immunoblot, and the biotinylated protein was detected by streptavidin blotting. Most interestingly, we found that kAE1 and the recessive kAE1 S773P and G701D mutants that copurified with eAE1 His were biotinylated but dominant kAE1 R589H was not (Fig. 8). This suggested that the wild-type protein could rescue the trafficking of recessive mutants, allowing the movement of the mutant protein from the ER to the plasma membrane, because of hetero-oligomerization. In contrast, the dominant dRTA mutant exhibited a dominant-negative effect, retaining the wild-type protein intracellularly.
Confocal images (Fig. 9) of LLC-PK1 cells cotransfected with eAE1 containing an external HA epitope (eAE1 HA557) with either dominant or recessive kAE1 confirmed that the dominant mutant could retain eAE1 HA557 intracellularly. The anti-HA antibody detected eAE1 at the cell surface in nonpermeabilized cells cotransfected with kAE1, recessive kAE1 S773P, and recessive kAE1 G701D. In contrast, cell surface staining of eAE1 was decreased in cells cotransfected with the dominant mutant kAE1 R589H. The recessive kAE1 S773P or G701D proteins were detected at the plasma membrane when coexpressed with the wild-type protein, whereas the dominant mutant kAE1 R589H was localized intracellularly.
FACS Analysis-We also examined the expression of kAE1 at the cell surface by fluorescence-activated cell sorting (FACS) using kAE1 with an hemagglutinin (HA) epitope inserted in the third extracellular loop (position 557). The effect of coexpression of dominant and recessive dRTA mutations on the cell surface expression of kAE1 HA557 was determined. For this experiment, cells were transfected with 0.25 g of kAE1 HA557 plus an equal amount of mutant DNA. At 0.25 g of kAE1 HA557 DNA alone, 19 Ϯ 4% (n ϭ 3 Ϯ S.D.) of cells showed higher than background staining (2 Ϯ 1%, n ϭ 3 Ϯ S.D.), indicating the population of cells that expressed kAE1 at the plasma membrane. Coexpression of wild-type kAE1 with kAE1 HA557 did not affect the percentage of cells (18 Ϯ 4%) expressing the tagged protein at the plasma membrane. There was little effect on the surface expression of kAE1 HA557 by coexpression with either recessive kAE1 S773P (17 Ϯ 8%, n ϭ 3 Ϯ S.D.) or G701D (14 Ϯ 5, n ϭ 3 Ϯ S.D.) mutants compared with coexpression with wild-type kAE1 (Fig. 9). In contrast, coexpression with dominant R589H caused a decrease (8 Ϯ 4%, n ϭ 3 Ϯ S.D.) in the percentage of cells showing cell surface expression.
The cell surface expression of the recessive S773P mutant containing an external HA epitope was also determined directly by FACS analysis of transfected HEK cells (Fig. 10). KAE1 S773P HA557 transfected alone showed a low signal (8 Ϯ 5%, n ϭ 3 Ϯ S.D.) revealing poor cell surface expression. Furthermore, HA epitope-tagged kAE1 S773P showed a 2-fold FIG. 9. Colocalization of recessive kAE1 S773P with wild-type AE1 at the plasma membrane. Confocal images of nonpolarized transiently transfected LLC-PK1 cells coexpressing eAE1 HA557 with either wild-type, recessive kAE1 S773P, G701D, or dominant R589H. After 2 days of transfection, the cells were fixed in 3.7% paraformaldehyde and permeabilized by incubation in 0.2% Triton, blocked with 3% BSA, and then stained with a mouse anti-HA and rabbit anti-Nt kAE1 antibodies recognizing eAE1 HA557 and kAE1, respectively. After washing, goat anti-mouse Alexa488 (1:1000) and Cy3-conjugated goat anti-rabbit antibody were used to visualize eAE1 HA557 and kAE1, respectively, using a Zeiss LSM 510 confocal microscope. The results indicate that the dominant dRTA mutant retains the wild-type protein intracellularly, whereas the wild-type protein can rescue the recessive mutant proteins and allow their expression at the cell surface. Coexpression of the two recessive mutants resulted in no cell surface expression of these mutants.

DISCUSSION
SLC4A1 mutations can lead to the molecular pathogenesis of both AD (1-7) and AR dRTA (4, 8 -10). In this paper, the effect of a novel missense recessive mutation (S773P) in the SLC4A1 gene on the biosynthesis of kAE1 was characterized in transfected cells. KAE1 S773P was localized to the endoplasmic reticulum in both HEK cells and in nonpolarized LLC-PK1 cells. Whereas kAE1 N555 acquired complex oligosaccharide, kAE1 S773P contained solely high mannose oligosaccharide through 24 h of chase time. This indicates that the recessive kAE1 S773P mutant is impaired in its trafficking from the ER to the medial Golgi. As well, kAE1 S773P could not be biotinylated at the cell surface, was not detected by FACS analysis of nonpermeabilized cells, and showed no transport activity confirming its intracellular localization in HEK cells. The kAE1 S773P mutant is found as a compound heterozygous in the patient with another recessive mutant, kAE1 G701D. KAE1 G701D was also impaired in its trafficking to the plasma membrane in transfected HEK cells. Intracellular retention of recessive mutants was observed previously in other membrane proteins, such as cystic fibrosis transmembrane conductance regulator, aquaporin, and pendrin proteins causing human diseases (44 -47). KAE1 S773P showed a reduced level of expression compared with wild-type kAE1, because of a decrease in its stability. An inhibitor binding assay and protease sensitivity showed that the kAE1 S773P was improperly folded. The effect of this mutation in the folding of the protein might be expected since the mutation leads to the replacement of a highly conserved serine to proline within a putative transmembrane segment. Studies of proline substitution in the transmembrane domain concluded that the introduction of proline in transmembrane segment tends to have a strong helix-breaking property (48). Moreover, the proline residue may induce the kink formation in transmembrane segment and subsequently be the reason for misfolding. Misfolded membrane proteins are often retained in the ER by cellular quality control involving interactions with chaperones such as calnexin (49). Although only one amino acid is changed in the protein structure, kAE1 S773P is recognized by the quality control system of the cell, which leads the defective protein to be degraded by the proteasome (50 -52). Protein misfolding because of the presence of a missense mutation is a common mechanism in the pathogenesis of human diseases such as cystic fibrosis, nephrogenic diabetes insipidus and long QT syndrome (44,45,53).
Mutations in the SLC4A1 gene causing the autosomal dominant form of dRTA such as R589H, S613F, G609R, and R901X are impaired in trafficking to the cell surface and are mistargeted in polarized epithelial cells (7,27). R589H and R901X exert a dominant effect because of hetero-oligomerization with the normal protein, resulting in intracellular retention of the normal protein (23,26). In contrast, kAE1 could traffic to the plasma membrane after forming hetero-oligomers with the recessive kAE1 S773P and G701D mutants. This is similar to AE1 SAO, a mutant that forms heterodimers with the normal protein that are present at the red cell surface (54). The normal FIG. 10. FACS analysis revealing the surface-expressed kAE1. The bar graph showed the percentage of cell surface expression of kAE1 HA557 wild-type or kAE1 S773P HA557 from transfected HEK-293 cells. After staining with mouse anti-HA antibody, the anti-mouse conjugated with Alexa 488 was then visualized, and the signal was analyzed by using Beckman-Coulter EPICS Elite flow cytometry. The error bar shows the S.D. with n ϭ 3. AE1 protein retains transport function, although it is associated with an inactive SAO subunit (28).
In conclusion, the defective trafficking of kAE1 S773P in a compound heterozygous state with kAE1 G701D provides an explanation for the dysfunction found in dRTA (Fig. 11). Both kidney AE1 S773P and G701D recessive mutants are impaired in trafficking to the plasma membrane in transfected HEK cells. The recessive S773P mutant was misfolded and targeted to the proteasome for degradation. The kAE1 S773P mutant can form a homo-oligomer, as well as a hetero-oligomer with kAE1 G701D. With the lack of the functional expression of either mutant protein at the basolateral membrane of the ␣-intercalated cells, the cells would not be capable of transporting bicarbonate into the blood. In the heterozygous state with wild-type AE1, the mutant can form hetero-oligomers, but this does not prevent trafficking of the wild-type protein to the plasma membrane. This is in contrast to the dominant dRTA mutant (kAE1 R589H) that retains the wild-type protein in the ER. Also, the more rapid degradation observed for the mutant protein may limit the amount of hetero-oligomer formation leaving the wild-type protein relatively free to homo-dimerize and traffic normally to the cell surface. Thus, in the heterozygous state, sufficient kAE1 would be present in the basolateral membrane of ␣-intercalated cells to retain adequate bicarbonate transport into the blood. The expression of dominant and recessive dRTA mutants in MDCK cells is currently under investigation in order to determine the effects of the mutations on kAE1 trafficking to the basolateral membrane in polarized epithelial cells.