Article Text

Heterogeneity in the processing defect of SLC26A4 mutants
  1. J S Yoon1,
  2. H-J Park2,
  3. S-Y Yoo3,
  4. W Namkung1,
  5. M J Jo1,
  6. S K Koo4,
  7. H-Y Park4,
  8. W-S Lee3,
  9. K H Kim1,
  10. M G Lee1
  1. 1
    Department of Pharmacology and Brain Korea 21 Project for Medical Science, Yonsei University College of Medicine, Seoul, Korea
  2. 2
    Soree Ear Clinic, Seoul, Korea
  3. 3
    Department of Otorhinolaryngology, Yonsei University College of Medicine, Seoul, Korea
  4. 4
    Department of Biomedical Science, National Institute of Health, Seoul, Korea
  1. Professor Min Goo Lee, Department of Pharmacology, Yonsei University College of Medicine, 134 Sinchon-Dong, Seoul 120-752, Korea; mlee{at}


Background: Mutations in the SLC26A4 gene are responsible for Pendred syndrome and non-syndromic hearing loss (DFNB4). This study analysed non-synonymous SLC26A4 mutations newly identified in East Asians, as well as three common mutations in Caucasians, to characterise their molecular pathogenic mechanisms and to explore the possibility of rescuing their processing defects.

Methods: A total of 11 non-synonymous disease associated mutations were generated and their effects on protein processing and on ion transporting activities were examined.

Results: Most of the mutations caused retention of the SLC26A4 gene product (pendrin) in the intracellular region, while wild-type pendrin reached the plasma membrane. Accordingly, these mutations abolished complex glycosylation and Cl/HCO3 exchange activities of pendrin. However, significant heterogeneity in the processing of mutant pendrin molecules was observed. Each mutant protein exhibited a different cellular localisation, a different degree of N-glycosylation, and a different degree of sensitivity to the treatments that rescue processing defects. For example, H723R-pendrin, the most common mutation in East Asians, was mostly expressed in endoplasmic reticulum (ER), and its defects in protein processing and ion transporting activities were restored considerably by low temperature incubation. On the other hand, L236P-pendrin, the most common mutation in Caucasians, was mainly in the centrosomal region and was temperature insensitive.

Conclusion: These results indicate that the processing of pendrin mutant protein is determined by mutant specific mechanisms, and that a mutant specific method would be required to rescue the conformational defects of each folding mutant.

View Full Text

Statistics from

Mutations in SLC26A4 are considered to be one of the most common causes of congenital deafness and account for up to 10% of all hereditary hearing loss.1 2 Two clinical manifestations, Pendred syndrome (MIM 274600) and non-syndromic hearing loss (DFNB4, MIM 600791), are associated with SLC26A4 mutations.3 4 Pendred syndrome is an autosomal recessive disorder first described in 1896 as the combination of congenital deafness and goitre.5 Non-syndromic recessive hearing loss has been reported to be an allelic disorder of Pendred syndrome.4 The SLC26A4 gene product, commonly referred to as pendrin, is an 82 kDa protein containing 12 putative transmembrane domains that transports monovalent anions, such as Cl, I, HCO3, and formate.6 7

Pendrin is expressed in the non-sensory epithelia of the inner ear, thyroid folliculocytes, and renal cortical collecting ducts.6 8 In thyroid folliculocytes, pendrin mediates Cl/I exchange at the apical membrane. This forms an efficient iodide trapping mechanism in thyroid follicles in cooperation with a sodium iodide symporter at the basolateral membrane that eventually leads to oxidation and organification of iodide by thyroid peroxidase.9 In kidney cells, it has been shown in knock-out animals that pendrin plays a role in the acid–base balance by mediating base secretion in the intercalated cells of renal cortical collecting ducts through its Cl/HCO3 exchange activities.10 The Cl/HCO3 exchange function of pendrin also seems to play an important physiological role in the inner ear.8 Loss of pendrin results in an acidification of endolymph and in a loss of the endocochlear potential via the loss of the K+ channel KCNJ10.11 Most of the hearing loss caused by pendrin mutation is prelingual, severe, and symmetrical. Frequently, enlarged vestibular aqueducts are found in patients with pendrin mutations. Enlargement of vestibular aqueduct (EVA) and its content, the endolymphatic sac and duct, is the most common radiological malformation of inner ear associated with sensorineural hearing loss including Pendred syndrome and DFNB4.12

At the cellular level, it has been suggested that endoplasmic reticulum (ER) retention and defective plasma membrane targeting of pendrin evoked by pathogenic mutations play a key role in the pathogenesis of Pendred syndrome.13 14 However, precise intracellular mechanisms responsible for the processing defects of pendrin mutants are not known. More than 150 different pendrin mutations have been reported (, and each ethnic population has its own distinctive mutant allele series with some prevalent founder mutations. Recently, we have identified 16 different mutations in Korean patients with sensorineural hearing loss and EVA.2 15 Eight of them were non-synonymous mutations that resulted in amino acid substitutions. In this study, we investigated the molecular pathogenic mechanisms of the newly identified SLC26A4 mutations and the three most common mutations observed in Caucasians, and explored the possibility of rescuing their processing defects using an integrated molecular and physiologic examination.


Materials, solutions, and cells

Standard HEPES buffered solution A contained (mmol/l) 140 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 10 d-glucose, and 10 HEPES (pH 7.4 with NaOH). HCO3 buffered solution B contained (mmol/l) 120 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 10 d-glucose, 5 HEPES, and 25 NaHCO3 (pH 7.4 with NaOH). Cl-free solutions were prepared by replacing Cl with gluconate. HEK 293 cells were maintained in DMEM-HG supplemented with 10% fetal bovine serum, and HeLa cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum. The fluorescent pH probe, 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF), was purchased from Molecular Probes (Eugene, Oregon, USA). Primary antibodies used for immunostaining and immunoblotting were as follows: anti-Myc monoclonal and polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, California, USA), anti-γ-tubulin (Sigma, St Louis, Missouri, USA), anti-CFTR M3A7 (Upstate Biotechnologies, Lake Placid, New York, USA), anti-KDEL-protein (Stressgen, Victoria, BC, Canada), and anti-His-tagged protein (Calbiochem, Darmstadt, Germany). Fluorescein isothiocyanate (FITC)- or rhodamine-conjugated secondary antibodies were from Zymed (Carksbad, California, USA). Restriction enzymes and peptide-N-glycosidase F (PNGase F) were purchased from New England Biolabs (Beverly, Massachusetts, USA). All other general chemicals were purchased from Sigma.

Pendrin cDNA and site directed mutagenesis

The coding region of pendrin cDNA was amplified by polymerase chain reaction (PCR) from the cDNA library of human thyroid tissue and subcloned into the pCMV-Myc vector (Clontech, Palo Alto, California, USA) using the Xho I and Not I restriction sites. The pendrin cDNA sequence was verified by nucleotide sequencing, and was identical to a registered sequence (Genbank NM_000441). Oligonucleotide directed mutagenesis was performed using the QuikChange II site directed mutagenesis kit (Stratagene, La Jolla, California, USA) according to the manufacturer’s protocol. For addition of polyhistidine (His6) tag, the coding region of pCMV-Myc-pendrin was amplified by PCR and subcloned into pcDNA3.1/His vector (Clonetech) using the BamH I and Not I restriction sites. This resulted in an addition of 35 amino acids to the N-terminus of Myc-tagged pendrin. The cloning and mutagenic primers used in this study are listed in supplemental table 1. The mammalian expressible clones for cystic fibrosis transmembrane conductance regulator (CFTR; pCMV-CFTR and pCMV-ΔF508-CFTR) were described previously.16


HEK 293 cells and HeLa cells were transfected with plasmids for wild-type (WT) or mutant pendrin using the Lipofectamine-Plus reagent (Invitrogen, Carlsbad, California, USA). Immunostaining of HeLa cells was performed using anti-Myc or other appropriate primary antibodies. Briefly, membrane cultured cells were fixed and permeabilised by incubation in cold methanol for 10 min at −20°C, and then stained with the primary antibodies and the fluorescently labelled secondary antibodies. Alternatively, cells were fixed in ice-cold formaldehyde (3.65% in phosphate buffered saline (PBS), 10 min) and permeabilised by saponin (0.1% in PBS, 4 min). Fluorescent images were obtained with a Zeiss LSM 510 confocal microscope. Fluorescence of FITC was excited at 488 nm by argon laser, and emitted fluorescence was detected with 505–530 nm band-pass filter. Fluorescence of rhodamine was excited at 543 nm by He-Ne laser, and emitted fluorescence was detected with 560–615 nm band-pass filter. Pinhole (70 μm), detector gain (850), amplifier offset (0), and amplifier gain (1) were maintained for all images constantly.

Immunoblotting, digestion of N-glycosylation, and surface biotinylation

Immunoblotting was performed using a standard protocol described previously.16 For digestion of glycosylated pendrin by PNGase F, protein samples were first denatured by adding SDS to 0.5% and β-mercaptoethanol to 1%, and incubating for 10 min at 37°C. NP-40 was then added to 1% and N-linked carbohydrates were removed by adding PNGase F (500 U/reaction) and incubating the solution for 2 h at 37°C. For surface biotinylation of pendrins, cells were washed with PBS and then incubated with PBS containing 0.1 mM CaCl2 and 1 mM MgCl2 for 5 min. Cells were then biotinylated using EZ-Link sulfo-NHS-SS-biotin (Pierce, 0.5 mg/ml) for 30 min in the dark. After washing free biotin with bovine serum albumin-containing (1% w/v) PBS, cells were lysed with lysis buffer (1 % NP-40, 150 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, pH 7.4, with complete protease inhibitor cocktails). The lysates were centrifuged for 10 min (13 000 g), and the pellet was discarded. Avidin solution (Streptavidin beads, Pierce, 50 μl) was added to the supernatant (300 μg of protein in 300 μl of lysis buffer), and the mixture was incubated overnight with gentle agitation. Avidin bound complexes were pelleted (13 000 g), washed three times with lysis buffer, and immunoblotted.

Measurement of intracellular pH (pHi) and Cl/HCO3 exchange

Measurements of pHi in HEK 293 cells transiently transfected with pendrin plasmids were performed using a pH sensitive fluorescent probe BCECF with co-transfection of the trans-gene marker pCMV-GFP (Life Technologies) described previously.17 A cluster of cells showing the green fluorescent protein (GFP) fluorescence were loaded with BCECF and pHi was monitored. To minimise the variations from transfection efficiencies, cells showing GFP fluorescence within 80–120% levels of control cells (transfected with pCMV-GFP and mock vector) were chosen for measuring anion exchange in each set of experiments. After dye loading, the cells were perfused with a HCO3 buffered solution B, and BCECF fluorescence was recorded at the excitation wavelengths of 490 and 440 nm at a resolution of 2/s using a recording setup (Delta Ram; PTI Inc, Birmingham, New Jersey, USA). Cli/HCO3o exchange activities were estimated from the initial rate of pHi increase as a result of Clo removal in the HCO3 containing buffer (25 mM HCO3 with a 5% CO2 gassing). The intrinsic buffer capacity (βi) was calculated by measuring ΔpHi in response to 5–40 mM NH4Cl pulses in Na-free solutions.18 The βi of HEK cells was 36.4 (2.2) mM/pH unit (at pHi 7.1), and this value was not significantly different for cells transfected with WT- or mutant-pendrins. Therefore, the results of Cl/HCO3 exchange activity were expressed as ΔpH/min, and this value was directly analysed without compensating for the buffer capacity.

Statistical analysis

The results of multiple experiments are presented as means (SEM), and statistical analysis was carried out using analysis of variance. A value of p<0.05 was considered significant.


Protein expression of pendrin mutants

A total of 11 non-synonymous SLC26A4 mutations were analysed in this study. Eight of these were identified in a recent study on Korean patients with sensorineural hearing loss and EVA.15 In addition, the three most common disease-causing mutations in Caucasians, L236P, E384G, and T416P,19 were included in this study. Locations of individual mutations are depicted in supplemental fig 1 using a 12-transmembrane domain topology of pendrin predicted by the MEMSAT3 package (

Figure 1 Immunostaining of wild-type (WT) and mutant pendrins. After a 48 h transfection with plasmids for WT or mutant pendrins, HeLa cells were incubated with anti-Myc (A–L) or anti-CFTR M3A7 (M and N) antibodies, and stained with the fluorescein isothiocyanate (FITC) conjugated secondary antibodies. Arrowheads indicate membrane expressions of WT-pendrin, S166N-pendrin, and WT-CFTR.

In the initial step, intracellular localisation of pendrin was examined in cells heterologously expressing WT- or mutant pendrins using antibodies against an NH2 terminal Myc-epitope. Immunostaining of each pendrin mutant protein showed similar results in both HeLa and HEK 293 cells. However, because of frequent cell detachment problems during the fixation of HEK 293 cells, images collected from HeLa cells were more reproducible (fig 1). Cells were fixed and immunostained 48 h after transfection with each plasmid. All the mutations except S166N caused retention of pendrin in the intracellular region, while WT- and S166N-pendrin reached the plasma membrane. Most mutant pendrin proteins were retained intracellularly, though their exact localisations varied. For example, L236P-pendrin was expressed in a punctate pattern with a small hot spot in the perinuclear region (fig 1F). On the other hand, H723R-pendrin was localised in ER tube-like structures (fig 1L). The control set of immunostaining was performed with wild-type and mutant (Δ508) CFTR, which are representative epithelial transporters expressed in the plasma membrane and retained in intracellular regions, respectively (fig 1M and 1N).16 20

Expression of pendrin mutants was also determined by immunoblotting in HEK 293 cells using anti-Myc antibodies (fig 2A). In all pendrins except for E625X, where the 155 COOH-terminal amino acids were deleted, an 85 KDa protein band was detected. In addition, WT- and S166N-pendrin showed higher molecular weight bands around 110 KDa. It has been speculated that pendrin is a glycoprotein because higher molecular weight proteins are frequently observed in immunoblotting rather than the calculated weight of 82 KDa.7 21 In the present study, we successfully analysed pendrin glycosylation patterns using NH2-terminal Myc-epitope tagged constructs and peptide N-glycosidase (PNGase) F. As seen in fig 2B, digestion of WT- and S166N-pendrin with PNGase F cleared both the 85 KDa and 110 KDa bands and produced a single 82 KDa band. These results indicate that the 85 KDa band is an ER core-glycosylated pendrin (band B, fig 2A and B) and that the 110 KDa protein of WT- and S166N-pendrin is a complex glycosylated pendrin that can be expressed on the plasma membrane (band C, fig 2A and B). In fact, the surface biotinylation experiment revealed the membrane expression of WT- and S166N-pendrins (fig 2C). This correlates well with the findings in immunostaining (fig 1), where only WT- and S166N-pendrins are expressed on the plasma membrane. In some immunoblots, band A-like protein was observed below the band B protein in the absence of PNGase F treatments, such as L236P-pendrin in fig 2C. These unglycosylated species are probably ER associated degradation (ERAD) retro-translocation precursors, of which deglycosylation occurs upon targeting to degradation.22

Figure 2 Immunoblotting and digestion of N-glycosylation. (A) HEK 293 cells transfected with plasmids for wild-type (WT) or mutant pendrins were blotted with anti-Myc antibodies. (B) Samples from cells transfected with WT- or S166N-pendrin were digested with peptide N-glycosidase F (PNGase F). An aliquot of samples was treated with digestion buffer (Bf) and bovine serum albumin (BSA) to examine the specificity of PNGase F digestion (middle lane). (C) Surface biotinylation of membrane proteins was performed in HEK 293 cells transfected with WT-, S166N-, L236P-, and H723R-pendrins.

Cl/HCO3 exchange activities of pendrin mutants

It has been suggested that the defect in Cl/HCO3 exchange activities at the apical membrane of inner ear epithelial cells is the key factor that causes deafness.8 11 Therefore, we evaluated the Cl/HCO3 exchange activity of each pendrin mutant as compared with that of WT-pendrin. Although pendrin secretes HCO3in most of epithelial cells by Clo/HCO3i exchange, we measured HCO3influx upon Cl removal (Cli/HCO3o exchange) because of experimental advantages. Representative traces are shown in fig 3A–E, and summarised results are presented in fig 3F. Transfection with WT-pendrin evoked an 8.4-fold increase in Cl/HCO3 exchange activity from 0.033 (0.004) to 0.276 (0.031) ΔpH unit/min in HEK 293 cells. Pendrin-mediated Cl/HCO3 exchange activities were notably decreased in most of the mutants, confirming the pathogenicity of these mutations. On the other hand, the S166N mutant showed a decrease of approximately 25% in average activity compared with WT, but the difference was insignificant (p = 0.09).

Figure 3 Measurements of Cl/HCO3 exchange activity. A batch of HEK 293 cells showing a minimal intrinsic Cl/HCO3 exchange activity was monoclonally selected and used for the experiments. Twenty-four hours after the plating of HEK 293 cells to glass coverslips, WT or mutant pendrin plasmids and GFP expressing plasmids (5:1 in μg) were transiently transfected into the cells. Two days after the transfection, a cluster of cells showing the GFP fluorescence was loaded with BCECF and intracellular pH was monitored. Cli/HCO3o exchange activities were estimated from the initial rate of pHi increase as a result of Clo removal. (A–E) Representative traces of cells transfected with mock vector, WT-pendrin, S166N-pendrin, L236P-pendrin, and H723R-pendrin are shown. (F) A summary of Cl/HCO3 exchange activity. All measurements were taken a minimum of three times on cells transfected with each of two separate batches of mutant plasmids (n⩾6).

Heterogeneity in the processing defect of mutant pendrins

The above results indicate that aberrant membrane transport caused by a processing defect is the major pathogenic mechanism of pendrin mutations examined in this study. Many disease-causing folding mutants such as ΔF508-CFTR are temperature sensitive, of which folding defects are significantly corrected by low temperature incubation (fig 4B).22 23 Therefore, the possibility of folding rescue by low temperature incubation was initially explored by investigating the N-glycosylation of pendrins, a marker of protein processing. Each mutant pendrin was transfected into HEK 293 cells incubated at the normal temperature (37°C) or at a low temperature (27°C for 36 h), and its expression pattern was analysed (fig 4). As expected, WT- and S166N-pendrin showed the highest levels of complex glycosylated band C proteins at both normal and low temperature incubations. Interestingly, all mutants except E625X showed an increase in the total amount of protein by the low temperature incubation. However, the most noticeable finding is that each mutant showed a different degree of processing correction in response to the low temperature incubation. For example, significant fractions of M147V- and H723R-pendrins showed band C proteins in response to low temperature incubation, while virtually no L236P- and E384G-pendrins were complexly glycosylated in response to the same treatment (fig 4). Based on these results, the L236P and H723R mutants were chosen for further study as representative mutants that have temperature insensitive and temperature sensitive processing defects, respectively. In addition, L236P and H723R are the most common disease associated mutations in Caucasians and in East Asians, respectively.2 19

Figure 4 Response to low temperature incubation. (A) WT- or mutant pendrin-transfected HEK 293 cells were incubated at 37°C (−) or at 27°C (+) for 36 h, and protein samples were blotted with anti-Myc antibodies. (B) A control immunoblot was performed using anti-CFTR antibodies and samples from HEK 293 cells transfected with WT- and ΔF508-CFTR, a representative epithelial transporter bearing the temperature sensitive processing defect. Arrowheads indicate the complex glycosylated band C form of pendrin and CFTR.

Cellular localisations of these pendrin mutant proteins were compared with those of the ER resident KDEL proteins and the centrosomal marker γ-tubulin (fig 5). Forty-eight hours after the transfection, L236P-pendrin was highly concentrated at the centrosome, with a small amount scattered in other intracellular regions including the ER (fig 5 G–L). However, when cells were stained 16 h after transfection, L236P-pendrin was mostly localised to the ER (fig 5 S–U). These findings imply that L236P-pendrin is rapidly censored by the ERAD system and transported to the centrosomal region to be degraded.24 On the other hand, H723R-pendrin remained mainly in the ER with only a small amount in the centrosome 48 h after transfection (fig 5 M–R).

Figure 5 Cellular localisation of WT-, L236P-, and H723R-pendrin. Each pendrin protein (green) was co-stained with ER or centrosomal markers (red). (A–R) After a 48 h transfection with plasmids for WT- or mutant pendrins, HeLa cells were stained with anti-Myc (pendrin), anti-KDEL protein (ER), and anti-γ-tubulin antibodies (centrosome). Arrowheads indicate the localisation of centrosome (γ-tubulin). Note that L236P-pendrin was highly concentrated at the centrosome, while H723R-pendrin accumulated in the ER. (S–U) After a 16 h transfection with plasmids for L236P-pendrin, HeLa cells were stained with anti-Myc (pendrin) and anti-KDEL protein (ER) antibodies.

Next, we carefully examined whether common measures that rescue folding defects could structurally and functionally correct the processing defects of L236P- and H723R-pendrins. Twenty-four hours after transfection with each construct, HEK 293 cells were further incubated at 27°C for 36 h. Alternatively, cells were treated with Na-butyrate, which has been shown to induce the membrane expression of ΔF508-CFTR.25 As seen in fig 4, H723R-pendrin, but not L236P-pendrin, exhibited complex glycosylation in response to low temperature incubation and Na-butyrate treatment (fig 6A). Rescue of H723R-pendrin was also demonstrated by immunolocalisation, with the membrane expression of H723R-pendrin induced by low temperature incubation (fig 6B). Next, Cl/HCO3 exchange activities of L236P- and H723R-pendrins were measured after low temperature incubation or Na-butyrate treatment for 36 h. As illustrated in fig 6C and 6D, these treatments significantly increased pendrin mediated Cl/HCO3 exchange activity in H723R-pendrin but not in L236P-pendrin. For example, Cl/HCO3 exchange activity of H723R-pendrin reached 0.101 (0.009) ΔpH unit/min through low temperature incubation, which corresponds to approximately 37% of the activity of WT-pendrin.

Figure 6 Temperature sensitivity of L236P- and H723R-pendrin. In panels A–D and F, after a 24 h transfection with each construct at 37°C, HEK 293 cells were further incubated at 27°C for 36 h or treated with Na-butyrate for 36 h at 37°C. Control cells were continuously kept at 37°C for 60 h after transfection. (A) HEK 293 cells were incubated at 27°C or treated with 5 mM Na-butyrate (NaB), and protein samples were blotted with anti-Myc antibodies. Arrowhead indicates the band C form of pendrin. (B) HeLa cells were stained with anti-Myc antibodies and FITC conjugated secondary antibodies. The arrowhead indicates the membrane expression of H723R-pendrin after a 36 h incubation at 27°C. (C) Representative traces of Cl/HCO3 exchange measured in HEK 293 cells incubated at 27°C (blue) or treated with Na-butyrate (green) after transfection with H723R-pendrin. (D) A summary of Cl/HCO3 exchange activity after the incubation at 27°C or treatment with Na-butyrate in cells transfected with mock vector, L236P-pendrin, and H723R-pendrin. (E) To minimise the protein overload and ER stress, HEK 293 cells were continuously kept at 27°C immediately after transfection and protein samples were harvested at earlier time points. The band C form of pendrin is observed in cells transfected with H723R-pendrin, but not L236P-pendrin. (F) Immunoblot was performed with anti-His monoclonal antibody and the His6-tagged pendrin clones, of which gene products express additional 35 amino acids at the N-terminus of the original Myc-tagged pendrin. Arrowhead indicates the band C glycosylated forms of pendrin.

A high amount protein loading by heterologous expression may cause ER stress and aberrant protein trafficking.21 Therefore, the difference between L236P- and H723R-pendrins was examined at earlier time points with low temperature incubation immediately after transfection in order to minimise the possible effects from protein overloading. As shown in fig 6E, the temperature sensitive processing correction of H723R-pendrin, but not L236P-pendrin, was also observed at the minimum protein loading states. Although the 13 amino acid Myc-tag at the N-terminus is relatively small, it may affect the protein processing and trafficking. Therefore, we also performed a control experiment using the pendrin clones with a different tag, of which gene products express additional 35 amino acids including His6-tag at the N-terminus of the original Myc-tagged pendrin (fig 6F). The results from new clones again clearly showed the differences in the temperature sensitivity between L236P- and H723R-pendrins. This result implies that: (1) the N-terminus of pendrin is a relatively unaffected region from ERAD; and (2) the difference between L236P- and H723R-pendrins is more critical for protein processing than that of the variations at N-terminus.


Over the past decade, the 10 member SLC26 gene family has emerged as a new group of anion exchangers structurally distinct from the classical SLC4 family of anion exchangers.6 26 27 Special interest has been focused on four members of the SLC26 family that are associated with human diseases: chondrodysplasias (SLC26A2), congenital chloride diarrhoea (SLC26A3), Pendred syndrome and non-syndromic deafness (SLC26A4), and non-syndromic hearing impairment (SLC26A5).6 26 The present study analysed non-synonymous SLC26A4 mutations newly identified in East Asians, as well as three common mutations in Caucasians, to characterise and compare their molecular pathogenic mechanisms. Results from this study revealed that most of the disease associated mutants did not reach the plasma membrane, which in turn abolished pendrin mediated anion transport. The most notable finding is significant heterogeneity in the processing defects and intracellular localisations among the pendrin mutants. Previously, molecular studies on SLC26A4 mutation dealt with L236P-pendrin, the most common mutation in Caucasians, and suggested that pendrin mutation generally causes temperature insensitive processing defects, and that glycosylation patterns are not associated with the degradation of pendrin.13 21 However, the present study clearly demonstrates that these phenotypes are not representative of all SLC26A4 mutations. The H723R pendrin mutant, the most common mutation in East Asians, showed a temperature sensitive processing defect that can be significantly rescued by the low temperature incubation (fig 6). In addition, defects in N-glycan processing were highly associated with the intracellular retention of disease-causing mutant protein (fig 1, fig 2, and supplemental table 2).

One of the key remaining questions is the underlying molecular mechanism behind the differences in the ERAD of L236P- and H723R-pendrins. This may result from either quantitative or qualitative differences in the ERAD processes. For example, L236P may bear a quantitatively more severe defect in protein folding than that of H723R, and undergo much faster degradation that could result in small remnants of protein in the ER regions and a large accumulation of protein in the centrosomal region where the proteasomal machinery assembles.24 The finding that L236P-pendrin is abundant in the ER 16 h after transfection but moves to the centrosomal region 48 h after transfection partially supports this hypothesis (fig 5). On the other hand, differences between the ERAD of L236P- and H723R-pendrins may stem from a qualitative reason. For example, L236P-pendrin could be censored by an ER quality control system that presides in an earlier step of the ERAD pathway, whereas H723R-pendrin could evoke a defect in a later step of protein processing, such as the COP II mediated ER exit pathways. Therefore, L236P-pendrin might be sent to the degradation machinery immediately after protein synthesis, while H723R-pendrins could be packed within the ER tubular structures. These potential underlying mechanisms are currently under investigation. Since defects in protein processing cause many loss-of-function type conformational diseases, such as cystic fibrosis and α1 antitrypsin deficiency, an extensive effort has been made to reverse the processing defects. For example, low temperature incubation, chemical chaperones including glycerol, Na-butyrate treatment, and Golgi alkalinisation have been shown to have some degree of effect, although there are no clinically available therapies for conformational diseases at present.16 20 25 The results in this study imply that individual mutant specific methods will be required for rescuing the conformational defects of folding mutants. It also appears that the processing defects of mutants rapidly censored by an early ERAD pathway are more difficult to rescue.

The finding that widely scattered pendrin mutations induced an altered glycosylation and processing defect suggests that pendrin is a very sensitive substrate of ERAD. Among the pendrin mutations examined in this study, those in the transmembrane domains and intracellular regions evoked processing defects, whereas S166N, located in the extracellular domain, did not. This result implies that ERAD-cytosolic (ERAD-C) might play a larger role than ERAD-luminal (ERAD-L) in the degradation of mutant pendrin proteins, as in the case of CFTR.28 29 In particular, ERAD of H723R-pendrin resembles that of ΔF508-CFTR in many ways: (1) both are temperature sensitive; (2) both function as membrane transporters when they reach the plasma membrane; and (3) both exhibit increased membrane expression in response to the non-specific histone deacetylase inhibitor Na-butyrate.23 25 It has been suggested that Na-butyrate may increase the membrane trafficking of ΔF508-CFTR by inhibiting the association with the cytoplasmic molecular chaperone Hsc70.25

All of the East Asian mutations examined in this study, except S166N, were on the conserved sites in rats (Genbank NP035997) and mice (Genbank NP062087), which indicates that these sites are important for either the folding or the function of pendrin. Some of the mutations are located in regions highly homologous among mammalian SLC26 transporters and even among some bacterial and plant sulfate transporters. For example, E384G is located on the Saier motif, which includes the triplet –NQE- residues at the COOH-terminus of a hydrophobic transmembrane domain and is conserved in the plant homologue of Sporobolus hamata.30 Three mutations, E625X, L676Q, and H723R, are located on the sulfate transporter and anti-sigma (STAS) domain, which is preserved in bacterial sulfate transporters.6 The physiological roles of the Saier motif and the STAS domain in SLC26 transporters are unknown; however, results from this study suggest that these sites are important for the structural integrity of SLC26A4 and that alterations at these sites evoke folding defects. S166N was found in a deaf patient as a compound heterozygote with the IVS4+4A>G mutation without any phase information.15 However, based on current data, S166N does not seem to be a severe pathogenic mutation. S166N-pendrin folds properly, reaches the plasma membrane, and retains approximately 75% of the Cl/HCO3 exchange activity of WT-pendrin (figs 13). In addition, the S166N mutant retains 70–100% function of other pendrin anion transport activities, such as Cl/OH exchange, Cl/formate exchange, and I/HCO3 exchange (data not shown).

This study is the first to characterise clearly the glycosylation patterns of WT- and mutant pendrins. Previously, N-glycosylation of pendrin was estimated based on the molecular weight differences between the observed mass of 110 KDa by immunoblot and the calculated mass of 82 KDa deduced from amino acid sequences.7 21 However, the lack of specific antibodies that can sensitively detect both glycosylated and non-glycosylated pendrins has hampered precise determination of pendrin glycosylation patterns. In this study, using the short NH2 terminal Myc-tag, which would not greatly affect the overall tertiary structure of pendrin, both ER core glycosylated and complex glycosylated pendrins were identified. Furthermore, it was also found that the expression of complex glycosylated pendrins is highly correlated with their membrane presentation.

In conclusion, we have identified a significant heterogeneity in the processing of mutant pendrin molecules that are associated with Pendred syndrome and non-syndromic deafness. Each mutant protein exhibited a different cellular localisation, a different degree of N-glycosylation, and a different degree of sensitivity to the treatments that rescue processing defects. These results indicate that the processing of folding mutants is determined by individual mutant specific mechanisms, and suggest that a mutant specific method would be required to rescue the conformational defects of each folding mutant.


We thank Sangmi Em and Sanghee Ko for technical assistance and WonSun Han for editorial assistance. This work was supported by grants R11-2007-040-01001-0 from the Korea Science and Engineering Foundation, Ministry of Science and Technology, Korea, and AO30001 from the Korea Health 21 R&D Project, Ministry of Health & Welfare, Korea.


View Abstract

Supplementary materials


  • Competing interests: None declared.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.