Cell Physiology


We report the novel, heterozygous AE1 mutation R730C associated with dominant, overhydrated, cation leak stomatocytosis and well-compensated anemia. Parallel elevations of red blood cell cation leak and ouabain-sensitive Na+ efflux (pump activity) were apparently unaccompanied by increased erythroid cation channel-like activity, and defined ouabain-insensitive Na+ efflux pathways of nystatin-treated cells were reduced. Epitope-tagged AE1 R730C at the Xenopus laevis oocyte surface exhibited severely reduced Cl transport insensitive to rescue by glycophorin A (GPA) coexpression or by methanethiosulfonate (MTS) treatment. AE1 mutant R730K preserved Cl transport activity, but R730 substitution with I, E, or H inactivated Cl transport. AE1 R730C expression substantially increased endogenous oocyte Na+-K+-ATPase-mediated 86Rb+ influx, but ouabain-insensitive flux was minimally increased and GPA-insensitive. The reduced AE1 R730C-mediated sulfate influx did not exhibit the wild-type pattern of stimulation by acidic extracellular pH (pHo) and, unexpectedly, was partially rescued by exposure to sodium 2-sulfonatoethyl methanethiosulfonate (MTSES) but not to 2-aminoethyl methanethiosulfonate hydrobromide (MTSEA) or 2-(trimethylammonium)ethyl methanethiosulfonate bromide (MTSET). AE1 R730E correspondingly exhibited acid pHo-stimulated sulfate uptake at rates exceeding those of wild-type AE1 and AE1 R730K, whereas mutants R730I and R730H were inactive and pHo insensitive. MTSES-treated oocytes expressing AE1 R730C and untreated oocytes expressing AE1 R730E also exhibited unprecedented stimulation of Cl influx by acid pHo. Thus recombinant cation-leak stomatocytosis mutant AE1 R730C exhibits severely reduced anion transport unaccompanied by increased Rb+ and Li+ influxes. Selective rescue of acid pHo-stimulated sulfate uptake and conferral of acid pHo-stimulated Cl influx, by AE1 R730E and MTSES-treated R730C, define residue R730 as critical to selectivity and regulation of anion transport by AE1.

  • cation leak
  • Xenopus oocyte
  • patch clamp
  • sulfate transport

hereditary stomatocytosis (HSt) is characterized by overt or compensated, autosomal dominant anemia associated with increased red blood cell (RBC) cation leak insensitive to ouabain. The cation leaks of HSt RBCs from different cohorts exhibit widely varying temperature sensitivities and varied red cell membrane abundance of stomatin (Band 7.2) (11, 41), but red cells from all patients share elevated Na content and reduced K content. HSt with red cell cation leak (4) has been associated with multiple missense mutations in the AE1/SLC4A1 gene encoding the erythocyte Cl/HCO3 exchanger Band 3 (5, 22, 40) and with two missense mutations in the erythrocyte NH3/NH4+ transporter Rhesus antigen-associated glycoprotein (RHAG), a component of the Rhesus antigen complex (4). Syndromic cation leak hemolytic anemia has been linked to missense mutations in the erythrocyte glucose transporter GLUT1 (45). However, the specific mechanism(s) or pathways(s) mediating the abnormal cation permeabilities in HSt RBC remains unknown.

In each type of cation leak anemia, the mutation itself has been proposed to mediate a cation leak pathway through the mutant transporter, which usually remains at normal abundance in the red cell membrane. However, none of the cation leaks exhibit the pharmacological profile of the associated wild-type transporter, despite the partial preservation of wild-type transport function in some cases (40). In addition, some investigators have provided evidence that the mutant transporters activate independent, endogenous transport pathways in erythrocytes (2) or in heterologous expression host cells (40).

We report here a case of autosomal dominant hereditary stomatocytosis in association with the novel, heterozygous AE1/SLC4A1/Band 3 mutation R730C. The aim of this study was to define the ion transport properties of the patient's red cells and to characterize the consequences to AE1 ion transport function of the R730C mutation as expressed in Xenopus oocytes. The patient's red cells exhibited greatly elevated Na content and reduced K content without detectably increased cation channel-like activity, and AE1 polypeptide was of normal abundance. Expression of AE1 R730C in Xenopus oocytes revealed normal or supranormal surface abundance and near total loss of Cl transport function unresponsive to attempted functional rescue by the AE1-binding protein glycophorin A (GPA) or by restoration of wild-type side chain charge by sulfhydryl modification. AE1-mediated Cl influx was preserved by substitution at R730 with Lys but not by substitution with Glu, Ile, or His. Wild-type AE1 does not increase cation transport in oocytes. Oocyte expression of AE1 R730C activated endogenous ouabain-sensitive 86Rb+ flux but minimally altered ouabain-insensitive influx of 86Rb+ or Li+. Wild-type AE1 can mediate H+/sulfate cotransport [sulfate transport activated by acidic extracellular pH (pHo)] in place of cis Cl. Acid pHo-activated sulfate influx by AE1 R730C was greatly reduced but was selectively (and unexpectedly) activated by sodium 2-sulfonatoethyl methanethiosulfonate (MTSES) exposure. Consistent with the negative charge modification introduced by MTSES, AE1 mutant R730E exhibited supranormal rates of acid-activated sulfate influx. Moreover, 36Cl influx by AE1 R730C exposed to MTSES was also activated by acidic pHo, in contrast to the inhibition of wild-type AE1-mediated Cl transport by highly acidic pHo. These observations, although not defining the mechanism of the red cell cation leak associated with AE1 R730C, document the importance of residue R730 to the anion translocation pathway of AE1 and its regulation.



Methanethiosulfate (MTS) reagents were from Toronto Research Chemicals (North York, Ontario). 4,4′-Diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS) and charybdotoxin (ChTX) were from Calbiochem (St. Louis, MO). Na36Cl was from GE Healthcare. Na35SO4 and 86RbCl were from Perkin Elmer. Salts, Larcoll, and all other reagents were from Sigma-Aldrich (St. Louis, MO) or Thermo-Fisher Scientific (Waltham, MA).

Clinical data.

Clinical histories and blood samples were obtained according to protocols approved by the Institutional Review Boards of Stanford University Hospital and Children's Hospital Boston. Whole blood was collected in EDTA tubes and shipped overnight on cold pack from Stanford, CA to Boston, MA.

Hematological parameters.

Aliquots of 300 μl whole blood were analyzed with the ADVIA@120 hematology analyzer (Siemens Medical Solutions Diagnostics, Tarrytown, NY), as previously described (33), for determination of erythrocyte and reticulocyte hematological parameters. Peripheral smears were stained with Wright-Giemsa reagent (47).

Preparation of erythrocytes and measurement of ion content.

Whole blood passed through cotton to preadsorb white cells was centrifuged in a Sorvall RC 28S at 2,500 rpm for 4 min at 4°C as previously described (33). The RBC pellets were washed four times with ice-cold Mg2+-free choline wash solution (CWS) containing (in mM) 150 choline Cl, 20 sucrose, and 10 Tris 3-[N-morpholino]propane sulfonic acid (MOPS), pH 7.4 (4°C). Manual hematocrits were measured with a 50% (vol/vol) cell suspension prepared in Mg2+-free CWS. Aliquots of this suspension were diluted with 0.02% Acationox in double-distilled water to allow measurement of intracellular contents of elemental Na, K, and Mg by atomic absorption spectrometry (Perkin Elmer 800, Wellesley, MA) as previously described (34).

Effects of ion transport inhibitors [0.1 mM ouabain, 10 μM bumetanide, 10 μM 5-(N,N-hexamethylene)amiloride (HMA), or combinations] on erythrocyte ion contents were measured after 3 h incubation at 37°C in plasma-like solution (containing, in mM: 110 NaCl, 5 KCl, 5 glucose, 1 MgCl2, 2 Na biphosphate, pH 7.4, 25 NaHCO3) or in high K medium (containing, in mM: 140 KCl, 10 NaCl, 1 Mg Cl2, 10 glucose, 2.5 K phosphate, pH 7.4). Cells were then washed four times with Mg2+-free CWS, and ion content was measured as described above.

Iso-osmotic discontinuous gradient fractionation of erythrocytes.

Erythrocytes were fractionated on discontinuous Larcoll (arabinogalactan) density gradients as previously described (1). Iso-osmotic Larcoll solutions of densities 1.077, 1.124, and 1.148 g/dl were layered into 12-ml polypropylene tubes on ice. Three milliters of 50% cell suspension were added above the least dense layer, and the tubes were centrifuged at 4°C for 45 min at 50,000 g in a swinging bucket rotor. Three fractions (fF1, fF2, fF3) were collected and washed four times with Mg2+-free CWS to remove residual Larcoll. Isolated fractions resuspended at 50% (vol/vol) were analyzed for hematological parameters as above.

Red cell Gardos channel activity.

Gardos channel activity was measured as described (33). Freshly isolated erythrocytes were suspended at 2% hematocrit in influx medium containing 10 μCi/ml 86Rb+ and (in mM) 145 NaCl, 2 KCl, 0.15 MgCl2, 0.1 ouabain, 10 Tris-MOPS, pH 7.4 (22°C), 0.01 bumetanide, in the presence or absence of 50 nM ChTX. Extracellular [Ca2+] was buffered to 7 μM with 1 mM citrate buffer (46). After addition of 5 μM A23187, aliquots were withdrawn after periods of 2 and 5 min and then immediately centrifuged through 0.8 ml of ice-cold influx medium containing 5 mM EGTA onto an underlying cushion of n-butyl phthalate. Supernatants were aspriated, the tube tip containing the cell pellet was cut off, and erythocyte-associated radioactivity was counted in a gamma counter (41600 HE Isomedic, ICN Biomedicals, Costa Mesa, CA). 86Rb+ uptake was linear up to 5 min. Fluxes were calculated by linear regression (33).

Na-K pump, NKCC, and NHE activities of red cells.

These transporter activities were measured in freshly isolated red cells without or with nystatin pretreatment as previously described (10), with modifications. Cation equilibration was performed in the presence of 40 μg/ml nystatin in a solution containing (in mM) 77 NaCl, 77 KCl, and 55 sucrose, with a final intracellular Na concentration of ∼50 mmol/l cells. Nystatin was then rapidly removed by addition of 1% bovine serum albumin (BSA). Red cell Na-K pump activity measured in a solution containing 155 mM choline chloride and 10 mM KCl was estimated as the fraction of Na+ efflux sensitive to 0.1 mM ouabain. Red cell Na-K-2Cl cotransport (NKCC) activity measured in a solution containing 154 mM choline chloride and 0.1 mM ouabain was estimated as the fraction of Na and K efflux sensitive to 10 μM bumetanide. Red cell Na/H exchange (NHE) activity stimulated by hypertonic shrinkage in a solution containing (in mM) 165 choline chloride, 1 MgCl2, 10 glucose, 0.1 ouabain, 0.01 bumetanide, and 10 Tris-MOPS (pH 7.4 at 37°C) was estimated as Na efflux-sensitive to 10 μM HMA. Leak flux was estimated as ouabain-insensitive Na efflux and reported as the rate constant for Na efflux.


Red cells were washed five times in 140 mM choline chloride, 1 mM MgCl2, 10 mM Tris-MOPS, pH 7.4, and then red cell ghosts were prepared as previously described (40), subjected to SDS-PAGE analysis in Tris-glycine, pH 8 (4–20% gradient gels, NuSep), and stained with Coomassie blue.

On-cell patch recording from red cells.

On-cell patch currents were recorded from patient red cells with the Axopatch 1-D amplifier [Axon Instruments (now Molecular Devices), Sunnyvale, CA] as previously described (43). See legend to Supplemental Fig. S1 at the AJP-Cell Physiol website for details.

Mutation screening.

Total RNA was extracted from whole blood with the RNeasy Mini-kit (Qiagen), and cDNA was reverse transcribed from total RNA (Retroscript Kit, Ambion). The AE1 open reading frame was PCR-amplified from patient cDNA in two overlapping fragments using oligonucleotide primers listed in the online Supplemental Table S1. Gel-purified RT-PCR fragments of AE1 cDNA were sequenced in entirety on both strands. Genomic DNA was isolated from whole blood with the DNeasy Blood and Tissue Kit (Qiagen). A fragment of AE1/SLC4A1 Genomic DNA extending from the 3′-end of intron 16 into the 3′-untranslated region of exon 20 was PCR amplified using oligonucleotides hAE1.I16F and hAE1.3R1 (Supplemental Table S1). Within this genomic fragment, exon 17 was sequenced to confirm the heterozygous mutation detected in patient cDNA.

cDNA preparation, mutagenesis, and cRNA transcription and expression.

The pXT7 oocyte expression plasmids encoding wild-type human AE1/SLC4A1 and AE1(HA), in which the AE1 transmembrane domain third extracellular loop harbors an inserted epitope tag from influenza hemagglutinin (HA) (36), were linearized with SalI. pXT7-AE1 S731P was linearized with StuI (40). Mouse Ae1 mutant E699Q in pBS was linearized with HindIII (7). Human GPA in pXT7 was linearized with XbaI (36). Mutagenesis (38), cRNA transcription, and cRNA expression in Xenopus oocytes were as previously described (38, 40) and detailed in the online Supplemental Methods.

Oocyte isotopic flux measurements.

Unidirectional 36Cl influx studies were conducted for periods of 15–30 min with 0.25 μCi Na36Cl in 150-μl volumes of ND-96 containing (in mM) 96 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, 5 Na HEPES, pH 7.4. 35SO42− influx studies were carried out for 30-min periods in the media of the indicated pH containing 96 mM Na cyclamate and 5 μCi carrier-free Na235SO4 (47 nM) in the presence of 20 mM unlabeled Na2SO4. 86Rb+ influx studies were carried out for 2-h periods in ND-96 containing 1 μCi 86RbCl, in the absence or presence of 10 μM bumetanide, and 0.5 mM ouabain. The above influx measurements and measurement of unidirectional 36Cl efflux were carried out at room temperature as previously described (40). Unidirectional lithium influx experiments were carried out at room temperature for 2-h periods in MBS modifed to contain 85 mM LiNO3 in place of NaCl. Li+ uptake into oocytes was measured as elemental Li by atomic absorption spectrometry (4, 40).

Some oocytes were preincubated as indicated for 15 min at room temperature with the anion transport inhibitor 4,4′-di-isothiocyanato-stilbene-2,2′ disulfonic acid (DIDS, 200 μM), with continued drug exposure during the isotopic influx period. Other oocytes were preincubated 15 min at room temperature with sulfhydryl modifying reagents (24) 2-aminoethyl methanethiosulfonate hydrobromide [MTSEA, 5 mM (25, 50)], sodium 2-sulfonatoethyl methanethiosulfonate (MTSES, 10 mM), 2-(trimethylammonium)ethyl methanethiosulfonate bromide (MTSET, 1 mM) (9), dithiothreitol (DTT, 10 mM), or N-ethylmaleimide (NEM, 1 mM) with continued drug exposure during the 30-min isotopic influx period. Solutions of methanethiosulfonate (MTS) reagents were prepared fresh immediately before use and shielded from light until added to oocytes. Experiments were performed with oocytes harvested from at least two frogs.

Confocal laser scanning immunofluorescence microscopy.

Three to five days after injection with water or with 20 ng cRNA encoding wild-type AE1(HA) or AE1 R730C(HA) (with or without coinjection of 20 ng GPA cRNA), unfixed oocytes were blocked in phosphate-buffered saline (PBS) with 1% bovine serum albumin (BSA) (PBS-BSA) for 1 h at 4°C, then incubated unfixed for 1 h at 4°C with mouse monoclonal anti-HA antibody (dilution 1:100; Sigma), and washed three times for 10 min at 4°C. Antibody-exposed oocytes were fixed at room temperature for 30 min in PBS containing 3% paraformaldehyde, washed three times in PBS supplemented with 0.002% sodium azide, and then again blocked in PBS-BSA for 1 h at 4°C. Oocytes were next incubated 1 h with Cy3-conjugated secondary goat anti-mouse Ig (dilution 1:200; Jackson Immunochemicals) and again thoroughly washed in PBS-BSA. Oocytes expressing HA-tagged constructs were aligned in uniform orientation along a Plexiglas groove and sequentially imaged through the ×10/0.3 numerical aperature objective of a Zeiss LSM510 laser scanning confocal microscope, using the 543-nm laser line with 550 nm long-pass emission filter at 512 × 512 resolution, at a uniform setting of 80% intensity pinhole 188 μm (1.9 Airy units, selected to maximize signal to noise), image depth 12 bits, with detector gain 688, amp gain 1, and 0 amp offset.

Statistical analysis.

Statistical analysis was performed, as indicated, using the unpaired Student's t-test, Dunnett's t-test for one-group comparison with multiple groups, or Tukey Kramer analysis for all groups comparison. Differences were considered significant for P < 0.05.


Clinical data.

The proband is a 41-yr-old male with a lifelong history of being “yellow-tinged.” At age 13 he was hospitalized with jaundice and abdominal pain. At age 17 another hospitalization for jaundice led to a diagnosis of hereditary spherocytosis (HS) with splenomegaly, but no medical records or laboratory data are available from that time. At age 21 abdominal pain prompted ultrasound detection of gallstones, leading 1 yr later to cholecystectomy and splenectomy. The patient has subsequently noted intermittent jaundice without functional limitation. Two of his 3 children (ages 5 and 7, Fig. 1D) and 2 sisters of ages 37 and 52 (not shown) also carry diagnoses of HS.

Fig. 1.

Red cell properties in proband's family. A: Wright-Giemsa stain of peripheral blood smear of proband, patient I:1, showing occasional stomatocytes (arrows) and acanthocytes (*). B: Wright-Giemsa stain of peripheral blood smear of proband's unaffected wife I:2. C: Wright-Giemsa stain of peripheral blood smear of proband's affected child II:1. D: pedigree showing red cell mean corpuscular volume (MCV, fL) and Na content (mmol/kg Hb) for each family member. E: SDS-PAGE profile of Coomassie blue-stained proteins from red cell ghosts of family members as indicated in the pedigree (D).

A complete blood count (CBC) from an overnight shipment of blood from patient I:1 revealed (Table 1) macrocytosis (MCV 112 fL) without anemia or elevated reticulocyte count, normal corpuscular hemoglobin concentration mean (CHCM) and red blood cell distribution width (RDW), and (not shown) normal white cell and platelet counts. Similar results were observed for red cells from the patient's two affected children (Table 1, child II:1 and child II:2). The patient's peripheral blood smear (Fig. 1A) revealed mostly normal red cells, with occasional stomatocytes, acanthocytes, and Howell-Jolly bodies consistent with postsplenectomy status. However, the peripheral smear of the affected child II:1 was normal (Fig. 1C). The incubated osmotic fragility test (not shown) revealed increased susceptibility to hypotonic lysis in a large proportion of patient I:1 cells. The SDS-PAGE profile of erythrocyte membrane proteins revealed a normal protein profile, with grossly normal abundance of Band 3/AE1 and grossly normal spectrin/Band 3 ratio (Fig. 1E).

View this table:
Table 1.

Hematological indices and red cell ion contents

The elevated mean corpuscular volume (MCV) and periodically elevated reticulocyte count, atypical for postsplenectomy HS, prompted measurement of glucose-6-phosphate dehydrogenase (G6PD) activity in least dense (nominally young) and most dense (nominally oldest) red cell fractions. Control red cell G6PD activity was 8.7 (in EU/g Hb) in the top fraction and 5.4 in the bottom fraction, consistent with the loss of G6PD activity as cells age (21). In contrast, the patient's G6PD activity was 6.2 in the top fraction and 8.8 in the bottom fraction, suggesting an altered red cell density distribution. Indeed, density fractionation of patient I:1 red cells revealed that most reticulocytes sedimented in the more dense fF2 and fF3 fractions, despite values of MCVr and CHCMr indistinguishable from normal controls (Table 2). The elevated G6PD activity in the novel, reticulocyte-rich dense fraction of patient RBC in the patient was consistent with a membrane permeability abnormality.

View this table:
Table 2.

Hematological parameters and ion content in mature erythrocytes and reticulocytes from whole blood and density-fractionated cells of patient I:1

As shown in Table 2, the patient's red cells exhibited elevated Na content and reduced K content. The patient's red cells also exhibited decreased Na/H exchange and Na-K-2Cl cotransport activities when compared with control values under the maximal velocity conditions of nystatin treatment (Table 3). Na+-K+-ATPase pump activity in nystatin-treated patient red cells was comparable with the control values. Similar patterns of activity were evident in density-fractionated red cells (not shown). Intact (nystatin-untreated) patient I;1 RBC had eightfold-elevated intracellular Na content (>200 mmol/kg; Table 4). These cells also exhibited a ouabain-resistant “Na leak” efflux rate constant sixfold higher than control, consistent with the 50-fold increased leak fluxes [0.2 ± 0.04 mmol/1013 cells × h for control cells (n = 12) vs. 10.3 mmol/1013 cells × h for patient cells]. Patient red cells also exhibited a ouabain-sensitive Na flux elevated nearly sixfold compared with control cells (Table 4), as previously reported (5, 18, 30). These findings are all consistent with the elevated intracellular Na content in red cells from patient I:1 and characteristic of HSt. Note also that in nystatin-untreated patient red cells, Na-K-2Cl cotransport is normal and Na/H exchange is elevated (Table 4), suggesting no intrinsic reduction in function of these transporters.

View this table:
Table 3.

Ion transport in nystatin-treated control erythrocytes and mutant erythrocytes from patient I:1 (mmol/1013 cell × h)

View this table:
Table 4.

Ion transport in fresh mutant erythrocytes from patient I:1 compared with normal control erythrocytes

Red cell channel-like activity.

The elevated Na content and increased cation leak of the patient red cells suggested the possibility of a conductive cation leak. Supplemental Fig. S1 presents a trace from an on-cell patch record of the patient's red cell. The patient's red cells displayed occasional short bursts in activity suggesting multiple channel openings, with NPo of 0.019 in the 32 s sweep shown. The linear current-voltage relationship from this patch was consistent with a unitary conductance of 27 pS. The positive reversal potential close to zero measured in the absence of pipette chloride identified the inward current measured at negative values of negative intracellular potential (−Vp) as carried largely by cations (likely inward Na+ current and outward K+ current). Comparable spontaneous activity was observed in a second tight patch, but a third tight patch was quiescent. This activity level resembled and was no greater than that recorded in on-cell patches of normal HbAA red cells (43). Thus on-cell patch clamp recording from three patient red cells provided no evidence that the increased cation leak measured in the whole cell population was conductive. As additional blood samples were unavailable to increase the recording sample size, these preliminary results do not rule out the possibility that some patient red cells do exhibit increased membrane conductance.

Mutation detection.

The presence of overhydrated cation-leak stomatocytosis prompted evaluation of the patient's RHAG and AE1 genes. The RHAG cDNA sequence was free of mutations. The AE1/SLC4A1 cDNA sequence revealed a novel, heterozygous C-to-T substitution in position 1 of codon 730, giving rise to the missense substitution R730C (Fig. 2B). This substitution was confirmed in the AE1/SLC4A1 genomic sequence as heterozygous mutation c.16161.C>T (Fig. 2C). No additional sequence variants were noted in the AE1 open reading frame.

Fig. 2.

Novel stomatocytosis mutation in the AE1 gene. A: control DNA sequence trace centered around codon R730 (CGT) in AE1 cDNA. B: patient cDNA sequence trace revealing heterozygous missense mutation in codon 730 from CGT to TGT, encoding C730 in one allele. C: patient genomic DNA sequence trace of the corresponding region of exon 17 in the AE1 gene. D: schematic of the AE1 polypeptide transmembrane domain locating the R730C mutation (dark square) in the context of previously reported stomatocytosis-associated AE1 missense mutations (circles).

Amino acid residue 730 of AE1 is located within the region previously delineated to encompass other AE1 missense mutations associated with cation leak stomatocytosis (Fig. 2D). The mutation R730C shares its location within the proposed reentrant loop 1 (RL1) of the AE1 polypeptide with a subset of previously reported stomatocytosis mutations, in particular S731P, H734R, R760Q (5), and E758K (40). RL1 has been proposed as an important determinant of anion selectivity and transport regulation (38, 40).

AE1 R730C at the Xenopus oocyte surface is inactive as a Cl transporter.

Figure 3 demonstrates that the fluorescent signal of ecto-epitope-tagged stomatocytosis mutant AE1 R730C(HA) at the Xenopus oocyte surface was approximately twice that of wild-type AE1(HA). These results suggest that the mutant polypeptide might traffic more efficiently to the plasma membrane or might be impaired in physiological internalization. Alternatively, HA epitope accessibility to antibody was increased in the mutant polypeptide. Coexpression of GPA did not increase surface expression with 20 ng of each AE1 cRNA injected. Neither AE1 R730C nor AE1 R730C(HA) mediated detectable unidirectional 36Cl influx in conditions of robust wild-type AE1 Cl transport activity, although the HA-tagged mutant was abundantly expressed at the oocyte surface. GPA coexpression had little or no stimulatory effect on Cl influx (Fig. 4A). The presence of the novel Cys residue at AE1 codon 730, a site exposed to the extracellular medium in both AE1 (13), suggested its likely accessibility to modification by the poorly permeant, selective, sulfydryl-reactive MTS reagents. We therefore tested the hypothesis that alkylation or reduction of AE1 R730C might rescue Cl transport activity. As shown in Fig. 4B, the greatly reduced Cl uptake by AE1 R730C was not increased by prolonged exposure to high concentrations of MTSEA (5 mM), MTSES (10 mM), or MTSET (1 mM). Treatment with MTSEA partially inhibited Cl influx by wild-type AE1, whereas MTSES and MTSET were without effect. NEM treatment (1 mM) of oocytes from Ambystoma mexicanum also failed to rescue 36Cl transport by AE1 R730C (n = 10, data not shown; background Cl transport prevented study in X. laevis). Since addition of a single Cys residue might plausibly predispose to disulfide crosslinking between promoters of dimeric AE1 (42), we tested the ability of 10 mM DTT to rescue Cl transport. However, this maneuver, too, failed to correct loss of function (not shown).

Fig. 3.

Epitope-tagged AE1 R730C polypeptide is expressed at the Xenopus oocyte surface. Whole mount confocal immunofluorescence micrographs through the equatorial plane of oocytes previously injected with water (A) or with 20 ng cRNA encoding AE1 R730C(HA) (B), AE1 R730C(HA) and GPA [20 ng of each cRNA; (C)], wild-type AE1(HA) (D), or wild-type AE1(HA) and GPA (E). Unfixed, unpermeabilized oocytes were incubated with anti-HA antibody, then postfixed and reacted with secondary antibody. Shown are median intensity, representative images from 4–7 oocytes studied in each group. Bottom left: mean fluorescence intensity values (± SE) of (n) oocytes in AE. *P < 0.05 for AE1 R730C vs. wild-type AE1 in the absence of GPA, or the same comparison in the presence of GPA, each by unpaired Student's t-test.

Fig. 4.

Positive charge at codon 730 is necessary but not sufficient for Cl transport by AE1. A: AE1 R730C and epitope-tagged AE1 R730C(HA) exhibit loss of function in a 36Cl influx assay, with minimal or no rescue by glycoprotein A (GPA). B: pretreatment with methanethiosulfate reagents (MTSEA, 5 mM; MTSES, 10 mM; MTSET, 1 mM) did not rescue Cl transport by AE1 R730C. However, MTSEA partially inhibited Cl transport by wild-type AE1 (*P < 0.05; Dunnett's t-test). C: substitution at R730 of AE1 with Lys (R730K) largely preserved Cl transport, but Cl transport is severely reduced upon substitution with Ileu (R730I), Glu (R730E), or His (R730H). Loss of function is not rescued by coexpression of GPA. Gray bars indicate coexpression of GPA. Values are means ± SE for (n) oocytes.

We next tested the consequences of side chain substitution at position R730 of AE1. As shown in Fig. 4C, preservation of side chain positive charge by Lys substitution retained most Cl uptake activity, although with loss of enhancement by GPA. In contrast, replacement of R730 with either the uncharged Ile, the negatively charged Glu, or the polar His residue resulted in loss of Cl uptake that was not rescued by coexpression of GPA.

Cation transport in oocytes expressing AE1 R730C.

Most AE1 mutations previously associated with cation leak stomatocytosis (5) have elicited cation leak upon overexpression in Xenopus oocytes (17, 44). However, some reports have suggested that cation leak represented indirect activation of endogenous cation transport pathways in oocytes (40) or in erythrocytes (2). In this context we examined the effect of AE1 R730C expression on cation transport in Xenopus oocytes. As shown in Fig. 5A, ouabain-insensitive 86Rb+ influx associated with AE1 R730C expression was not elevated in the absence or presence of coexpressed GPA. However, expression of AE1 R730C was associated with a substantial increase in ouabain-sensitive 86Rb+ influx, consistent with secondary activation of endogenous oocyte Na+-K+-ATPase. Li+ uptake has also been measured as an index of cation leak in oocytes expressing stomatocytosis-associated AE1 mutations (17). Figure 5C shows that 2 h of Li+ uptake into oocytes expressing AE1 R730C was GPA insensitive and not significantly higher than for wild-type AE1, in contrast to Li+ uptake by oocytes expressing AE1 stomatocytosis mutant S731. Thus, although AE1 R730C expression does not significantly increase oocyte influx of 86Rb+ or Li+, it increases Na+-K+-ATPase transport activity in a manner consistent with intra-oocyte Na accumulation in the steady state.

Fig. 5.

Cation influx in oocytes expressing AE1 R730C. A: AE1 R730C coexpressed without or with GPA does not exhibit significantly increased ouabain- and bumetanide-insensitive 86Rb+ influx. B: expression of AE1 R730C substantially increases ouabain-sensitive 86Rb+ influx. Bumetanide (5 μM) was also present or absent as indicated. C: AE1 R730C exhibited slightly increased Li+ influx in a GPA-insensitive manner and at rates lower than those mediated by AE1 stomatocytosis mutant S731P. Values are means ± SE for (n) oocytes (*P < 0.05 compared with wild-type AE1 in absence of GPA, Dunnett's t-test; **P < 0.01 compared with AE1 R730C with coexpressed GPA; Student's unpaired t-test). Gray columns indicate coexpression of GPA.

Negative charge modification of AE1 R730 rescues sulfate influx.

Since AE1 can mediate H+/sulfate cotransport (7, 8), we assessed sulfate uptake in oocytes expressing AE1 R730C. As shown in Fig. 6A, sulfate uptake at pHo 7.4 by wild-type human AE1 was low and by mouse Ae1 was even lower, but both were stimulated at pHo 5.5. As previously demonstrated for sulfate efflux (8), conditionally electrogenic mouse Ae1 mutant E699Q (7) exhibited anomalous inhibition of sulfate influx by acidic pHo. AE1 R730C exhibited reduced sulfate transport at pH 7.4 as well as lack of stimulation by acidic pHo.

Fig. 6.

Negative charge at codon 730 enhances sulfate transport by AE1. A: human AE1 (hAE1) and mouse Ae1 (mAe1) both exhibit acid pH-stimulated sulfate influx. In contrast, sulfate influx by hAE1 R730C and mAe1 E699Q is inhibited rather than activated by acidic bath pH (*P < 0.05 vs. pHo 7.4, Student's unpaired t-test). B: acidic pH-stimulated sulfate uptake by AE1 R730C is rescued by treatment with MTSES but not by treatment with MTSEA or MTSET. *P < 0.05 vs. absence of MTSES and vs. pHo 7.4 (Tukey-Kramer all pairs analysis). C: acidic pH-stimulated sulfate uptake is at wild-type levels in oocytes expressing AE1 R730K, enhanced in oocytes expressing AE1 R730E, and abrogated in oocytes expressing AE1 R730I. AE1 R730H exhibits greatly reduced sulfate transport activity but with preservation of acid pH-induced stimulation (*P < 0.05 vs. pHo 7.4, Tukey-Kramer). D: sulfate influx at pH 5.5 in oocytes expressing AE1 mutant R730E is inhibited by 200 μM DIDS. *P < 0.05 compared with the absence of DIDS. Values are means ± SE for (n) oocytes.

Modification of the Cys residue at position R730 with a positive charge through reaction with MTSET (1 mM) or MTSEA (5 mM) did not increase sulfate uptake at either pHo 7.4 or 5.5. In contrast, charge reversal at position R730 by treatment with MTSES (10 mM) partially restored the ability of acid pHo to stimulate sulfate uptake (Fig. 6B). This unexpected observation prompted evaluation of pH-sensitive sulfate uptake by AE1 containing additional missense substitutions at position R730 (Fig. 6C). Preservation of the sidechain positive charge in AE1 R730K was associated with wild-type sulfate transport stimulated by acidic pHo (compare with Fig. 6A). However, AE1 R730E exhibited enhanced, acid-stimulated sulfate uptake, at rates twice those exhibited by wild-type AE1 and consistent with the MTSES result. In contrast, AE1 mutants R730I and R730H were essentially inactive (Fig. 6C). Acid pHo-stimulated sulfate uptake into oocytes expressing AE1 R730E or wild-type AE1 was fully inhibited by 200 μM DIDS (Fig. 6D).

Remarkably, AE1 R730C treated with MTSES exhibited acid pHo-stimulated 36Cl influx, as well (Fig. 7A). However, the apparent increase in AE1 R730E-mediated 36Cl influx by acid pHo (Fig. 7B) fell short of statistical significance (P = 0.07). Oocyte pretreatment with MTSES at pH 7.4 before influx measurement at pH 5.5 showed indistinguishable results (n = 10, data not shown), establishing that pH dependence reflected a direct effect on anion transport rather than pH-dependent MTSES reactivity. These results contrast markedly with the inhibitory effect of acid pHo on wild-type AE1-mediated Cl transport (39). However, the low rates of 36Cl influx into oocytes expressing AE1 R730H (Fig. 4C) were not stimulated by acid pHo (n = 10, not shown).

Fig. 7.

Negative charge at codon 730 enhances AE1-mediated Cl transport. A: oocytes expressing AE1 R730C were exposed to 10 mM MTSES 15 min before and during the 30 min influx assay at the indicated pHo. MTSES rescued loss of Cl transport function only at pHo 5.5. *P < 0.01 for presence vs. absence of MTSES at pH 5.5; **P < 0.01 for pH 7.4 vs. 5.5 in the presence of MTSES (Tukey-Kramer analysis). B: acid pHo modestly stimulated 36Cl influx into oocytes expressing AE1 R730E (P = 0.07, Student's unpaired t-test). Values are means ± SE for (n) oocytes.


We have presented a family exhibiting autosomal dominant cation leak stomatocytosis with well-compensated, intermittent anemia, associated with the novel heterozygous AE1 mutation R730C. Affected red cells exhibited increased Na content, decreased K content, increased cation leak, and increased Na+-K+-ATPase activity, without evidence of increased red cell cation channel activity. Overexpression of the mutant AE1 polypeptide at the Xenopus oocyte surface was accompanied by a GPA-insensitive loss-of-function phenotype with respect to transport of Cl and sulfate, in parallel with substantially increased Na+-K+-ATPase activity. However, ouabain plus bumetanide-insensitive influxes of 86Rb+ and Li+ were not significantly increased. Substitution of R730 with Lys largely preserved anion transport function at neutral bath pH, but substitutions with Ile or His led to loss of function. Unexpectedly, chemical addition or genetic substitution of a negative charge at AE1 residue 730 partially or fully rescued acid pHo-stimulated transport of sulfate. These maneuvers rescued Cl transport only at acid pHo, exhibiting a pH dependence of Cl transport opposite that of wild-type AE1. Our study introduces a new stomatocytosis mutation within a restricted region of the AE1 polypeptide harboring previously reported stomatocytosis mutations and additionally uncovers new modes of regulation of pH-dependent anion transport by AE1 sequence variants.

Significance of clustered stomatocytosis mutations in the AE1 polypeptide.

The novel stomatocytosis mutation R730C is located at the NH2-terminal nominally extracellular portion of the RL1 reentrant loop structure (38), immediately adjacent to previously reported HSt mutations S731P and H734R (5) (Fig. 2D). In the Cys-less AE1 background, S731C was labeled by both biotin maleimide and by lucifer yellow iodoacetamide, marking it as exofacially oriented. In contrast, nearby S725C was inaccessible to labeling (13). AE1 S731C was not susceptible to short- or long-linker disulfide crosslinking to S731C in the adjacent protomer within the AE1 homodimer (42).

R730C is also near previously reported HSt mutations E758K (40) and R760Q (5), located at the COOH-terminus of RL1 (Fig. 2D). The HSt mutation G796R (22) has been modeled to reside at the adjacent exofacial surface of reentrant loop 2 (RL2). These residues mutated in stomatocytosis constitute a potential surface that might contribute to the outer vestibule of the anion translocation pathway through AE1 or an important regulatory region for that pathway. Two additional stomatocytosis mutations L687P and D705Y (5) are situated at the cytoplasmic face of the AE1 polypeptide. These may contribute to an inner vestibule for ion translocation. However, the wide variety of side chain substitutions associated with cation-leak stomatocytosis does not allow a straightforward electrostatic explanation for the proposal (5) that each mutation generates a pathological cation translocation pathway through the AE1 anion exchanger. It remains possible that each mutation alters interaction with native cation permeability pathway(s) of the erythrocyte or the heterologous host cell. Such pathways could be either uniform for all AE1 stomatocytosis mutations or unique for any single mutation.

The nature of the cation leak in stomatocytosis.

The selectivity of the nonspecific cation leak in stomatocytosis remains incompletely described. Despite the substantially elevated cation leak and increased Na+-K+-ATPase activity in heterozygous AE1 R730C erythrocytes, on-cell patch recording revealed no evident increase in cation channel-like activity (Supplemental Fig. S1). Similarly, AE1 R730C expression in Xenopus oocytes did not detectably increase ouabain-insensitive influxes of 86Rb+ or of Li+, but ouabain-sensitive 86Rb+ influx was significantly increased (Fig. 5), consistent with activation by elevated intra-oocyte [Na+]. The greatly increased ouabain-sensitive Na+ efflux rate measured in fresh patient red cells (Table 4) was normalized or reduced after nystatin treatment (Table 3), also consistent with pump activation secondary to elevated intracellular [Na+] as a result of increased cation leak (Table 4). Thus the cation leak in both affected red cells and in oocytes expressing AE1 R730C exceeded the corrective capacity of endogenous pump activity activated by elevated intracellular Na+. In oocytes, however, this occurred in the absence of statistically increased ouabain and bumetanide-insensitive influx of 86Rb+ or of Li+.

The presumed polypeptide pathway(s) of the stomatocytosis cation leak remain undefined. Multiple missense mutations associated with cation-leak stomatocytosis have been found not only in the AE1 gene but also in the gene encoding ammonia channel RhAG (4) and that encoding glucose/ascorbate transporter GLUT1 (45) (the latter in association with paroxysmal, exertion-induced dyskinesia and seizure disorder). These findings have led to the proposal that certain types of transporter mutations, whether in anion transporters such as AE1, polar solute transporters such as GLUT1, or a transporter of neutral solute NH3 (or, some propose, the cation NH4+), share the common phenotype of cation leak through the varied transporter polypeptides themselves (45).

Cation channel activity has indeed been recorded in GLUT1 mutant stomatocytic red cells (45). Missense mutations can alter cation selectivity of ion channels, as in mutant KV channels with Arg-to-His substitutions in the voltage sensor that exhibit voltage- and pH-regulated proton conductance via a proton wire mechanism (6). Single missense CLC mutations can convert coupled 2Cl/H+ exchange activity to uncoupled Cl conductance (26). These transitions can also be brought about in wild-type transporters by ligand toxins. Thus binding of palytoxin converts coupled 3Na+/2K+ exchange of α1-Na+-K+-ATPase into constitutive, nonspecific cation channel activity (14), and maitotoxin can produce an apparently similar effect on Ca2+-ATPase (35). Charge reversal mutations in the outer vestibule of palytoxin-uncoupled Na+-K+-ATPase suffice to convert selectivity of the conductance from cations to anions (32).

However, generation of nominally uniform cation leak via multiple types of missense mutation in multiple types of transporter supports consideration of an alternate possibility. Thus stomatocytosis mutations in AE1, RhAG, and GLUT1 genes might each result in activation of nonspecific cation permeabilities endogenous to the erythrocyte [perhaps with similar activation in heterologous expression systems (40)]. The incompletely penetrant association of stomatocytosis with deficiency of the cholesterol-binding red cell membrane protein stomatin (Band 7.2b) suggests a possible link with alternate cation pathways. The Caenorhabditis elegans stomatin homolog Mec2 is part of a mechanosensitive ENaC/ASIC-related degenerin cation channel complex of sensory neurons (19), and mammalian stomatin indeed modulates ASIC (31). The stomatin homolog podocin binds to and activates Ca2+-permeable cation channel Trpc6 in renal glomerular podocytes (20). Interestingly, interaction of stomatin with Glut1 also promotes Glut1-mediated transport of dehydroascorbate by red cells (27), secondarily altering transport of electrons and protons. The recent report of kidney AE1's interaction with the podocyte slit-diaphragm protein nephrin (48), another podocin-binding protein, suggests an indirect link between AE1 and Trpc6 [reportedly expressed also in red cells (12)]. However, the conductive nature of stomatocyte red cell cation leak remains controversial. In the case of AE1 stomatocytosis mutation H734R, the cation leak has been attributed to electroneutral cation/H+ exchange and to dysregulated electroneutral K-Cl cotransport, without evidence of macroscopic changes in red cell conductance (2). The preliminary patch clamp data of Supplemental Fig. S1 similarly suggest that the increased cation permeability of AE1 R730C stomatocytosis erythrocytes is unaccompanied by detectably increased membrane conductance. The explanation for decreased Gardos channel activity in AE1 R730C erythrocytes (Table 4) is unclear.

Anion transport by AE1 R730C.

Despite the apparently normal abundance of AE1 R730C in red cells (Fig. 1) and its near-normal surface expression in Xenopus oocytes (Fig. 3), Cl transport by the mutant was nearly abrogated and was not susceptible to rescue by coexpressed GPA (Fig. 4A). Similarly, Ala substitution at the corresponding R1056 residue in RL1 of mouse AE2a greatly reduced Cl transport, without evident reduction of oocyte surface abundance (37). Human AE1 residue R730 corresponds to mouse Ae1 residue R748. Passow and colleagues (15, 23) and Zaki (3) suggested mouse R748 as the principal and perhaps sole target for transport inhibition by phenylglyoxal at high concentrations, but the responsible residue(s) remain undefined. Expression in Xenopus oocytes of engineered mAe1 mutants R748K and R748Q revealed >90% loss of 36Cl influx activity in the setting of wild-type levels of total oocyte polypeptide content, accompanied by apparently wild-type mutant suface expression as detected by proteolytic susceptibility to extracellular chymotrypsin (23). The severe loss of Cl transport in oocytes expressing mAe1 R748K (and R748Q) differed remarkably from the minimal loss of Cl transport exhibited by those expressing hAE1 R730K, in the absence or presence of coexpressed GPA (Fig. 4C). The reason for this difference is unknown, but the mouse and human Arg residues reside within large regions of near-identical amino acid sequence. Within human AE1 amino acid 698–871 (mouse amino acids 716–889), a region extending from the cytoplasmic loop before the putative ninth transmembrane span until the middle of the COOH-terminal-most transmembrane span, the only amino acid sequence differences are at human A767/mouse S785 in the middle of the putative transmembrane span believed to follow RL1 and at human Y824/mouse F842 in distal RL2. One of these residues, or the more divergent sequences of the third, or adjacent fourth extracellular loops of AE1 may be the major determinant of this functional difference between human AE1 R730K and mouse Ae1 R748K.

Although human AE1 mutant R730K exhibited near-wild-type Cl transport activity (Fig. 4C), nominal restoration of side chain positive charge to mutant residue R730C by prolonged treatment with saturating concentrations of MTSEA or MTSET failed to rescue Cl transport activity (Fig. 4B). Surprisingly, however, nominal conferral of negative charge to R730C with MTSES partially rescued acid pHo-stimulated sulfate transport (Fig. 6B). Confirming the importance of side chain negative charge, the R730E mutant exhibited greatly enhanced DIDS-sensitive sulfate transport at acid pHo (Fig. 6, C and D). Moreover, MTSES-treated AE1 R730C further showed acid pHo-stimulated 36Cl influx. Thus partial rescue of Cl transport function by either method was accompanied by an unprecedented pH sensitivity opposite that of wild-type AE1.

These data suggest that R730E and MTSES-derivatized R730C might contribute a novel protonation site (or unmask a cryptic site) with unique regulatory consequences to transport of both monovalent and divalent anions. Alternatively, conformational changes produced by the charge modification might outweigh local electrostatics in determining the consequent gain-of-transport function. The RL1 region of mouse Ae2 has been shown to play a crucial role in regulating the physiological pattern of pH-dependent Cl transport, a pattern of proton-mediated inhibition of transport (38). The reversed pH dependence of AE1 R730C-mediated Cl transport recalls the reversed pH dependence for mouse Ae1 E699Q-mediated sulfate transport (8), but AE1 R730C-mediated sulfate transport exhibits wild-type pH dependence. Interestingly, phenylglyoxal treatment of intact red cells was reported to acid-shift the acid pHo-stimulated sulfate uptake into the cells (15). Possible contributions to the altered pH sensitivity of AE1 R730C of the human AE1 residues corresponding to pH-sensitive residue mouse Ae1 residue H752 and its modifiers, pH-sensitive E699 and the H2DIDS-reactive K558 (8, 28, 29), remain unknown. The currently available resolution of the AE1 transmembrane domain structure (49) neither reveals clearly the region of RL1 nor allows proposal of a structural basis for the reversed pH dependence of AE1-mediated Cl transport or for the role of negative charge at residue 730 in transport rescue. However, the functional modifications of AE1-mediated anion transport introduced by MTSES treatment of R730C and by the R730E substitution add to the properties that may be more completely explained by progressively higher resolution AE1 structures (49). These structures will, in turn, allow more detailed evaluation of the proposed cation translocation pathway through AE1 stomatocytosis mutant polypeptides.


This work was supported by National Institutes of Health Grants DK-43495 to S. L. Alper, HL-077765 to S. L. Alper and C. Brugnara, and HL-090632 to A. Rivera. A. K. Stewart was supported by a Pilot Feasibility award of the Harvard Digestive Diseases Center (DK-34854). P. S. Kedar was supported by an International Fellowship for Young Biomedical Scientists from the Indian Council of Medical Research.


We thank Drs. R. Colah and K. Ghosh (National Institute of Immunohematology, Mumbai) for their encouragement.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
View Abstract