Effect of complete protein 4.1R deficiency on. ion transport properties of murine erythrocytes

Transcription

1 Page 1 of 36 Articles in PresS. Am J Physiol Cell Physiol (June 14, 2006). doi: /ajpcell Effect of complete protein 4.1R deficiency on ion transport properties of murine erythrocytes Alicia Rivera a, Lucia De Franceschi b, Luanne L. Peters c, Philippe Gascard d, Narla Mohandas e and Carlo Brugnara a a Department of Laboratory Medicine, Children's Hospital Boston, Department of Pathology, Harvard Medical School, Boston, MA USA; b Department of Clinical and Experimental Medicine, Section of Internal Medicine, University of Verona, Verona, Italy; c The Jackson Laboratory, Bar Harbor, ME USA; d Lawrence Berkeley National Laboratory, Berkeley, CA; e New York Blood Center, NY USA Running Title: Altered Cation Transport in Protein 4.1 Deficient Mice Corresponding Author: Alicia Rivera, Ph.D. Children s Hospital Boston, Department of Laboratory Medicine 300 Longwood Avenue, BA 760, Boston, MA USA Phone: ; Fax: alicia.rivera@childrens.harvard.edu Word count: 5,590 Copyright 2006 by the American Physiological Society.

2 Page 2 of 36 2 ABSTRACT Moderate hemolytic anemia, abnormal erythrocyte morphology (spherocytosis), and decreased membrane stability are observed in mice with complete deficiency of all erythroid protein 4.1 protein isoforms (4.1-/-; Shi TS et al., J. Clin. Invest. 103:331,1999). We have examined the effects of erythroid protein 4.1 (4.1R) deficiency on erythrocyte cation transport and volume regulation. 4.1-/- mice exhibited erythrocyte dehydration that was associated with reduced cellular K and increased Na content. Increased Na permeability was observed in these mice, mostly mediated by Na/H exchange with normal Na-K pump and Na-K-2Cl cotransport activities. The Na/H exchange of 4.1-/- erythrocytes was markedly activated by exposure to hypertonic conditions (18.2 ± 3.2 in 4.1 -/- vs. 9.8 ± 1.3 mmol/10 13 cell x h in control mice), with an abnormal dependence on osmolarity, (K 0.5 =417 ± 42 in 4.1 -/- vs. 460 ± 35 mosm in control mice) suggestive of an up-regulated functional state. While the affinity for internal protons was not altered (K 0.5 = ± 0.7 vs ± 0.56 nm in control mice), the V max of the H-induced Na/H exchange activity was markedly elevated in 4.1-/- erythrocytes (V max 91.47±7.2 compared to 46.52±5.4 mmol/10 13 cell x h in control mice). Na/H exchange activation by okadaic acid was absent in 4.1-/- erythrocytes. Altogether, these results suggest that erythroid protein 4.1 plays a major role in volume regulation and physiologically down-regulates Na/H exchange in mouse erythrocytes. Up-regulation of the Na/H exchange is an important contributor to the elevated cell Na content of 4.1 -/- erythrocytes. Keywords: Spherocytosis, Cell Na, and Na/H exchange

3 Page 3 of 36 3 INTRODUCTION The erythrocyte cytoskeleton is a complex arrangement of proteins including spectrin, actin, ankyrin, adducing, protein 4.1 and 4.2. This multi-component structure is responsible for many physiological functions and mechanical properties of the erythrocyte. Deficiency or functional abnormalities of one or more of these cytoskeletal proteins decreases deformability of the erythrocyte plasma membrane leading to hemolysis. Several hemolytic anemias have been associated with defects in the erythrocyte membrane function due to altered cytoskeletal proteins. These anemias have been classified as hereditary spherocytosis (HS), hereditary pyropoikilocytosis and elliptocytosis (30). Chronic hemolysis is the major characteristic of HS. Spherocytic erythrocytes lose mechanical strength and deformability while increasing plasma membrane vesiculation and trapping leading to their destruction in the spleen. Several studies have shown that defects in either band 3, ankyrin, or spectrin account for a large number of HS cases. (24, 29). Spherocytes are also characterized by a lower cellular K, and increased Na content with dehydration (32). It has been suggested that these cation alterations are related to increased Na flux pathways such as Na pump, Na-K 2Cl cotransport (NKCC) and Na/Li exchanger. Recent studies have suggested that the increased permeability of the spherocytes is due to a loss of cellular skeleton integrity (10, 16). An increase in cation permeability due to deficiency of spectrin has been observed in a mouse model of spherocytosis (16). Dehydration and microspherocytosis have been observed in mouse erythrocytes lacking either erythroid band 3 protein (26) or ß adducin cytoskeletal proteins (13). However, there has been no detailed functional characterization of the permeability pathways that mediate changes in cellular Na and K transport in murine sperocytosis.

4 Page 4 of 36 4 We have previously described an up-regulated response of volume regulatory systems such as NKCC, Na/H exchanger (NHE) and of Na leak pathways due to hypertonic shrinkage in mice lacking protein 4.2 (25). We also have shown that the Ca 2+ -activated K channel (Gardos Channel) is functionally up-regulated in 4.1 knockout mouse erythrocytes (4.1 -/-) and that this channel plays an important role in protecting spherocytes from colloidosmotic lysis (11). Proteins 4.1 stabilize the spectrin-actin complex by anchoring it to the plasma membrane. In erythrocytes, protein 4.1 interacts with spectrin and actin to regulate the mechanical properties of the erythrocyte membrane via a calmodulin-dependent binding site (20). However, little is know about the role that protein 4.1 may play in the modulating volume regulatory increase systems such as the NHE. In the present study, we examine the ion transport pathways and volume regulatory responses of mouse erythrocytes devoid of protein 4.1 and identify a novel functional interaction between protein 4.1 and the erythroid NHE.

5 Page 5 of 36 5 METHODS Mice with targeted deletion of 4.1 (4.1 -/-) and 4.2 (4.2 -/-) have been previously described (13, 25, 30). Control mice (C57Bl/6J and 129/SvEv) were purchased from the Jackson Laboratory, Bar Harbor, ME. Chemicals and inhibitors were purchased from Sigma (ST. Louis, MO). Erythrocyte cation content and transport studies: Erythrocyte sodium and potassium content were determined in fresh mouse erythrocytes by atomic absorption spectrophotometry as previously described in Armsby et al. (1). Briefly, blood was colleted from isofluorene-anesthetized mice into heparinized tubes. After five washes in 172 mm choline chloride washing solution (1 mm MgCl 2, 10 mm TRIS-MOPS ph 7.4 at 4 C), an aliquot of 50% cell suspension in washing solution was used for determinations of hematocrit, cell Na (1:50 dilution in 0.02% acationox), and cell K (1:500 in double-distilled water). Na-K pump, NKCC and NHE : Erythrocytes were loaded with equal amount of Na and K in the presence of nystatin (40mg/ml) and fluxes were measured as described by Armsby et al. (1). Briefly, erythrocytes were loaded with equal amount of Na and K using a nystatin solution (mm): 77 NaCl, 77 KCl, and 55 sucrose. Na-K pump activity was estimated as the ouabainsensitive fraction (1 mm ouabain) of Na efflux into a medium containing 155 mm choline chloride and 10 mm KCl. The Na-K-2Cl cotransport (NKCC) was estimated as bumetanidesensitive Na and K efflux into medium containing 174 mm choline chloride with 1 mm ouabain in the presence and absence of 10 µm bumetanide. NHE activity was estimated as hydroxyl methyl amiloride-sensitive Na efflux in hypertonic shrinkage stimulation with a solution

6 Page 6 of 36 6 containing (mm): 175 mm choline chloride, 1 MgCl 2, 10 glucose, 1 ouabain, 0.01 bumetanide, 10 Tris-MOPS (ph 7.4 at 37 C). Inhibitors were dissolved in DMSO and added to each media as described in the legend. K-Cl cotransport: The K-Cl cotransport (KCC) activity was determined as Cl-dependent K influx (using 86 Rb as a tracer for K) in erythrocytes exposed to hypotonic swelling. The flux was calculated by subtracting total K influx into hypotonic NaCl medium ( mosm; 115 mm NaCl, 5 mm KCl, 1 mm ouabain, 10 NM bumetanide) from K efflux in hypotonic Na sulfamate medium (115 mm Na sulfamate, 5 K sulfamate, 1 mm ouabain, 10 NM bumetanide) and expressed as volume-dependent K influx. Volume regulation experiments: Fresh erythrocytes were incubated at 37 C for 5 h with gentle shaking in an isotonic medium with ionic composition similar to that of normal mouse plasma. The medium contained 160 mm NaCl, 2 mm KCl, 10 mm TRIS-MOPS, ph 7.4 at 37 C, 2 mm Na-biphosphate, 1 mm MgCl 2, and 5 mm glucose. To assess the role of various transport pathways in response to hypertonic shrinkage, fresh erythrocytes were incubated at 37 C for 2 h with gentle shaking in a hypertonic medium containing 220 mm NaCl, 2 mm KCl, 10 mm TRIS-MOPS, ph 7.4 at 37 C, 2 mm Na-biphosphate, 1 mm MgCl 2, and 5 mm glucose. Specific transport inhibitors were added at the specified final concentrations to assess the role of specific ion transport pathways under isotonic and hypertonic conditions. H induced Na influx: Na/H exchange activity was measured by determining the initial rates of net Na influx when an outward H gradients were imposed (27). Briefly, cell Na was removed by

7 Page 7 of 36 7 loading the erythrocytes with 145 K solution in the presence of nystatin (10 Ng/ml) at 4 C. The ionophore-free K loaded erythrocytes were incubated at 10% hematocrit with a K loading solution (ph ) containing 1 mm ouabain and 10 NM bumetanide for 20 min at 37 C. To clamp the intracellular ph, DIDS (100 NM) and acetazolamide (200 NM) were added to the cell suspension and incubated for another 30 min at 37 C. Since intracellular acidification induces cell swelling and alkalinization induces shrinkage, the loading medium osmolarities were adjusted between 380 (acid) and 310 (alkaline) by varying sucrose between 0 and 60 mm and KCl between 180 and 160 mm. The cell volume was estimated by comparing the hemoglobin/liter of the loaded cells ( g/l) with the fresh cells ( g/l) for each intracellular ph. Furthermore, all flux media contained 1 mm ouabain and 10 µm bumetanide to avoid contribution of the Na-pump or NKCC co-transport system. Proton loaded erythrocytes were washed 4 times and resuspended at 50% hematocrit with un-buffered washing solution at 4 C, osmotically balanced to prevent volume changes. Erythrocytes (200 Nl) were added to 2ml influx media either at ph 6.0 or ph 8.0 at 37 C. At specified time points, aliquots of 250 NL of the cell suspension were transferred into ice-cold eppendorf tubes containing 0.7 ml of buffer (85 mm choline Cl, 85 mm KCl, 0.25 mm MgCl 2, 10 mm TRIS-MOPS ph 7.4 at 4 C, 0-60 mm sucrose) layered over 0.3 ml of phthalate oil and spun down. Supernatants and oil were removed and the erythrocyte pellets were lysed with 1 ml of 0.02% Acationox detergent. Aliquots of the hemolyzed erythrocytes were used to determine hemoglobin concentration by optical density at 540 nm and Na concentration using atomic absorption spectrophotometry. Linear regression of the Na concentration versus time was used to calculate net Na influx at ph 6.0 and 8.0. The difference between these two conditions is the H induced Na influx expressed as mmol/10 13 cell

8 Page 8 of 36 8 x h. Mean Cellular Volume was measured to adjust flux values to a constant number of erythrocytes. Density Phthalate protocol: Density distribution curves were obtained using phthalate esters in micro-hematocrit tubes as described in detail by Brugnara et al. (3). Briefly, phthalate solutions were prepared to give a range of densities between g/ml. The hematocrit tubes were filled with 30 µl of red cell suspension and 10 µl of different phthalate solutions. Tubes were centrifuged at 12,200 rpm for 10 min at room temperature. The amount of denser cells was calculated from the amount of dense cells (lower layer) over the total amount of cells and expressed as a percent. Statistical analysis: All values are presented as mean ± standard deviation (SD).

9 Page 9 of 36 9 RESULTS Control of cell Na content in plasma-like isotonic conditions To assess the possible determinants of the increased cell Na of 4.1 -/- erythrocytes, freshly collected 4.1-/-, 4.2 -/-, and control erythrocytes were incubated for 5 h in a plasma-like medium in the absence and presence of specific transport inhibitors. In control erythrocytes, Na content is relatively stable in the absence of transport inhibitors; cell Na increases with the use of the Na-K pump inhibitor ouabain (1 mm), while there are no significant changes with either bumetanide (10 NM), an inhibitor of the Na-K-2Cl cotransport, or HOE-642 (10 NM) an amiloride analog inhibitor of the NHE (Fig. 1). Various combinations of these three inhibitors did not produce additional changes to those induced by ouabain alone, suggesting that the Na-K pump is the main regulator of cell Na content in these experimental conditions and that NKCC and NHE play a very minor role in determining the steady state ionic content of normal mouse erythrocytes (Fig. 1, left panel). Results essentially similar to those of control erythrocytes were obtained in 4.2 -/- erythrocytes (Fig. 1, middle panel). This is not surprising, since the transport abnormalities that we had described previously in 4.2 -/- erythrocytes require hypertonic shrinkage (25). In 4.1 -/- erythrocytes, there was a much greater increase in cell Na following treatment with ouabain, an indication that the passive net Na entry in these erythrocytes is greatly enhanced and contributes to the increased steady state cell Na (Fig. 1, right panel). There were no appreciable effects of either bumetanide alone or HOE-642 alone, suggesting that in the presence of an active Na-K pump, blockade of either of these pathways does not result in a measurable change in cell Na. Interestingly, when ouabain and bumetanide were combined, the cell Na gain was greater than with ouabain alone, suggesting that the NKCC in 4.1 -/- erythrocytes performs net Na

10 Page 10 of extrusion probably driven by the increased cell Na. The net gain in cell Na was reduced when erythrocytes were exposed to both HOE-642 and ouabain, suggesting that Na entry via the NHE is an important component of the Na influx into 4.1 -/- erythrocytes under basal conditions. Control of cell Na content in hypertonic conditions Hypertonicity is a known activator of the NHE, which mediates a regulatory volume increase (RVI) upon shrinkage in a variety of erythrocytes. In a first set of experiments, controls and 4.1 -/- erythrocytes were incubated at 37 C in a hypertonic NaCl solution, and erythrocyte densities were measured with the phthalate density technique at baseline, and after two hours incubation in the presence or absence of the NHE inhibitor HOE-642. As shown in Fig. 2, incubation in hypertonic media induced a marked shrinkage of the erythrocytes at baseline compared with isotonic conditions. When control erythrocytes were incubated in hypertonic medium for two hours, there were minimal changes in cell density, with only a small component of erythrocytes showing reduction in cell density, which was slightly inhibited by HOE-642. In 4.1 -/- erythrocytes, a substantial decrease in cell density took place over two hours incubation in hypertonic medium: this reduction in density was completely blocked by HOE-642, which also seemed to induce additional cell shrinkage (Fig. 2). Similar experiments were carried out in control, 4.2 +/-, 4.2 -/- and 4.1 -/- erythrocytes, measuring cell Na at baseline and after two hours incubation in hypertonic medium, with and without HOE-642. As shown in Fig. 3, and in our earlier studies (25), 4.2 -/- erythrocytes exhibit increased NHE activity under hypertonic conditions, which is blocked by HOE-642. A much greater activation of the NHE was found in 4.1 /- erythrocytes, which exhibited a two-fold greater increase in HOE-642-sensitive cell Na compared with 4.2 -/- erythrocytes (Fig. 3).

11 Page 11 of Functional properties of the Na/H exchange in 4.1 -/- erythrocytes To assess the effects of protein 4.1 on the kinetic properties of the NHE, we determined in control and 4.1 -/- erythrocytes the dependence of Na/H exchange activity on osmolarity and internal ph. Dependence of NHE activity on osmolarity: In control erythrocytes, the activity of NHE slowly increased as a function of media osmolarity in a sigmoidal pattern with a nominal saturation at 550 mosm, with a maximal velocity of 9.8 ± 1.3 mmol/10 13 cell x h (Fig. 4). Half of the exchanger activity was reached at 460±35 mosm. In contrast, 4.1 -/- erythrocytes showed a significantly elevated exchanger activity (18.2 ± 3.2 mmol/10 13 cell x h, n=3, p<0.01), which reached nominal V max around 500 mosm in a hyperbolic pattern. Half-maximal activation was achieved at 417± 42 mosm in 4.1 -/- erythrocytes, a value significantly lower (p<0.01) than that of control erythrocytes. Thus, the volume-sensitivity of the Na/H exchange is altered in 4.1 -/- erythrocytes, resulting in functional up-regulation of the system. Dependence of Na/H exchange activity on internal ph: Proton-induced Na influx via NHE was measured in control and 4.1 -/- mouse erythrocytes (Fig. 5). In control erythrocytes, NHE activity increased with intracellular acidification, with a maximal velocity of 46.52±5.4 mmol/10 13 cell x h and proton affinity constant of ± 0.7 nm (ph 6.31). In 4.1 -/- erythrocytes, a significant increase in maximal velocity was observed (91.47±7.2 mmol/10 13 cell x h, n=3, p<0.003), while the affinity constant for intracellular H was unaffected (537.0 ± 0.56 nm, n=3, ph 6.27). These data suggest that the increase in maximal velocity of the exchange is

12 Page 12 of not mediated by alteration on the internal proton allosteric site but rather due to an increase in flux rate. Pharmacological modulation of NHE activity: NHE is regulated by phosphorylationdephosphorylation events, via protein Kinase C (PKC) and protein phosphatases (18). PKC inhibition (with either chelerythrine or calphostin C) did not affect the volume-sensitive NHE activity in either control or 4.1 -/- erythrocytes (Fig. 6). The presence of protein phosphatase inhibitors such as okadaic acid (OKA) and calyculin A (CA) significantly stimulated the volumestimulated NHE activity in control mouse erythrocytes. This was in accord with previous observations of OKA-induced activation of the NHE in human erythrocytes (27). In contrast, in 4.1 -/- erythrocytes NHE was paradoxically inhibited by the presence of either OKA or CA (Fig. 6). Interestingly, both phosphatase blockers inhibited the NHE up to basal values in 4.1 -/- mice suggesting a marked functional dis-regulation of this pathway. NHE activity as a function of media osmolarity was measured in the presence or absence of 100 nm OKA concentration (Fig. 7). In control erythrocytes, the presence of 100 nm OKA resulted in a small shift in the activation of the system by osmolarity (EC 50 changed from 460 ± 35 mosm to 494 ± 34 mosm, p<0.041, n=3). In the 4.1 -/- erythrocytes, exposure to 100 nm OKA resulted not only in inhibition of the system but also in a complete loss of the activation by osmolarity, with a paradoxical inhibition of the system by hypertonic shrinkage. Thus, the absence of protein 4.1 results in a profound alteration of the volume regulatory loop of the NHE. Other transport pathways of 4.1 -/- erythrocytes: Na-K pump, NKCC and KCC

13 Page 13 of Na-K pump and NKCC: The maximal rate of the Na-K pump did not appear to be different in 4.1-/- erythrocytes compared with controls while it was reduced in 4.2 -/- and 4.2 +/- erythrocytes (Fig. 8) (25). Since the % of reticulocytes is significantly greater in 4.1 -/- erythrocytes, the maximal activity of the Na-K pump could be considered to be abnormally low in 4.1 -/- mice when compared with cell-age matched controls. No significant changes were observed in the maximal rate of NKCC compared with normal controls measured as Na or K flux (Fig. 8). NKCC activity was significantly elevated in 4.2 -/- erythrocytes, as previously shown (25). Since the NKCC cotransport is highly dependent on intracellular Na (2), the net Na extrusion carried out by this system will be increased in 4.1 -/- erythrocytes, as shown by the net changes in Na content of 4.1-/- erythrocytes incubated in isotonic conditions (Fig. 1). KCC: Studies of the dependence of KCC on cell swelling were hampered by the extreme fragility of 4.1 -/- erythrocytes, with significant lysis when osmolarity was decreased below isotonic conditions, and associated inability to measure K efflux in a reliable manner. Measurements of K influx using 86 Rb showed better reproducibility, and demonstrated an increase in K efflux, which can be accounted for by the significant reticulocytosis. Interestingly, the passive permeability of 4.1 -/- erythrocytes to K, estimated from the K influx in isotonic chloride-free medium was increased 2-3 fold compared to control erythrocytes (Fig. 9).

14 Page 14 of DISCUSSION Erythroid protein 4.1-deficient mice exhibit a unique set of membrane transport abnormalities, some of them shared by other mouse models of cytoskeletal disorders and others unique to this mouse model. Cellular dehydration due to K loss, and marked increases in cellular Na content are features seen in both human and mouse spherocytosis. The increased cell Na is due to an increased Na entry via ouabain and bumetanide-resistant pathways. Our studies extend these observations and indicate that a major mediator of this abnormal Na entry in 4.1 -/- erythrocytes is the NHE. The NHE isoform present in erythroid cells is NHE1, which is the prototype for all other members of this family (4, 6, 19). In erythrocytes and other cell types, NHE activity is activated by phosphorylation (23), and is modulated by insulin and osmotic stress (5, 9). NHE-function seems to be markedly altered in 4.1 -/- erythrocytes. Our studies suggest that NHE is constitutively activated in 4.1 -/- erythrocytes (see Fig. 3), with an increased sensitivity to activation by hyperosmotic shrinkage (Fig. 4), and normal regulation by intracellular H (Fig. 5). Deleted: In addition, the up-regulated NHE of 4.1 -/- erythrocyte is paradoxically inhibited by OKA, a known stimulator of this system in normal mouse erythrocytes (Fig. 7). These alterations suggest that there must be a functional interaction between 4.1 protein and NHE in normal mouse erythrocytes, which contributes to modulation and regulation of this transporter. These alterations are different from previously described functional abnormalities of the erythroid NHE: alterations of the internal ph NHE regulatory sites have been described in patients with

15 Page 15 of essential hypertension and insulin-resistant diabetes (7, 8), while the affinity of the internal H site was not altered in 4.1 -/- erythrocytes. Pharmacological blockade of phosphatase activity by OKA has been shown to stimulate NHE activity. While the absence of additional stimulation by OKA in 4.1 -/- can be rationalized by the up-regulated state of NHE, the paradoxical inhibition of the system by OKA remains unexplained (Fig. 7). It remains to be determined if this paradoxical effect of OKA in 4.1 -/- erythrocytes may indicate the presence of a previously unknown modulator which is functionally silent in normal erythrocytes, and becomes active in the absence of 4.1 protein. The Ca-activated phosphatase, Calcineurin Homologous Protein (CHP), has been shown to down-regulate NHE activity (17, 21). Perhaps in the absence of erythroid protein 4.1 protein, this interaction is altered, resulting in constitutive activation of the exchanger, as shown for CHP2, an isoform of CHP (22). Previous reports on interactions of protein 4.1 with other non-cytoskeletal membrane proteins have shown an interaction between the 30 kd FERM (Four4.1/Ezrin/Radixin/Moesin) domain of erythroid protein 4.1 that mediates cytoskeletalmembrane interactions and picln, a cytosolic protein believed to regulate volume-sensitive anion channels (28, 31). The crystal structure of this region of protein 4.1 has been recently solved, and the specific sites of interaction with band 3, glycophorin C/D and the PDZ (PSD95/Dlg/ZO-1) domain containing the protein p55 binding site, have been mapped. Erythroid protein 4.1 interaction with its binding partners is markedly affected by the binding of calmodulin to two separate binding regions of FERM domain in the presence of calcium (14). Calmodulin also regulates NHE-1 function and response to hypertonicity and acidification (12, 15, 34).

16 Page 16 of The Na/H exchanger regulatory factor (NHERF), is an essential element in the protein kinase A-mediated inhibition of the predominantly renal NHE3 isoform (35-37). The C-terminus of NHERF specifically interacts with protein of the family of membrane-cytoskeletal adapters ERM (ezrin-radixin-moesin) (33). This interaction is critical for the inhibition on NHE3 activity mediated by camp via PKA-dependent phosphorylation (38). The occurrence of such an interaction in mouse erythrocytes and its involvement in the abnormal regulation of NHE1 in 4.1-/- erythrocytes remains to be investigated. The data presented in this manuscript do not provide conclusive evidence on the mechanisms leading to the reduced K content and dehydration of mouse 4.1 -/- erythrocytes. We have reported increased activities of K-Cl cotransport (see Fig. 9) and Ca-activated Gardos channel (11), but the role of these two pathways in generating K loss and dehydration is still unclear. Human studies have also provided inconclusive evidence for the pathophysiology of K loss, with little evidence supporting a role for the K-Cl cotransport in this process (10). However, we have recently demonstrated that the K loss mediated by the Gardos channel is an important compensatory mechanism for the reduction in surface membrane area of HS erythrocytes, which protects erythrocytes from premature destruction due to colloid-osmotic lysis (11). In conclusion, 4.1 -/- erythrocytes exhibit distinct functional abnormalities, indicative of an important functional interaction between this cytoskeletal protein and the transmembrane ion transport protein NHE. The constitutive up-regulation of NHE in 4.1 -/- erythrocytes provides an explanation for their elevated baseline cell Na content.

17 Page 17 of ACKNOWLEDGMENTS We thank Lin-Chie Pong, Michelle Langlois, Michelle Rotter, Maria Argos, and Sarah Sheldon for technical assistance.

18 Page 18 of GRANTS Supported by NIH grants DK and HL (AR), DK50422 (CB), HL64885 (LLP), HL31579 (NM), and by FIRB-grant RBNE01XHME-003 (LDF). Funding was also provided for AR from the Center of Excellence in Minority Health and Health Disparities at Harvard Medical School.

19 Page 19 of REFERENCES 1. Armsby CC, Stuart-Tilley AK, Alper SL, and Brugnara C. Resistance to osmotic lysis in BXD-31 mouse erythrocytes: association with upregulated K-Cl cotransport. Am J Physiol 270: C , Brugnara C, Canessa M, Cusi D, and Tosteson DC. Furosemide-sensitive Na and K fluxes in human red cells. Net uphill Na extrusion and equilibrium properties. J Gen Physiol 87: , Brugnara C, Gee B, Armsby CC, Kurth S, Sakamoto M, Rifai N, Alper SL, and Platt OS. Therapy with oral clotrimazole induces inhibition of the Gardos channel and reduction of erythrocyte dehydration in patients with sickle cell disease. J Clin Invest 97: , Cala PM. Volume regulation by Amphiuma red blood cells. The membrane potential and its implications regarding the nature of the ion-flux pathways. J Gen Physiol 76: , Cala PM, Maldonado H, and Anderson SE. Cell volume and ph regulation by the Amphiuma red blood cell: a model for hypoxia-induced cell injury. Comp Biochem Physiol Comp Physiol 102: , Cala PM and Maldonado HM. ph regulatory Na/H exchange by Amphiuma red blood cells. J Gen Physiol 103: , Canessa M, Morgan K, Goldszer R, Moore TJ, and Spalvins A. Kinetic abnormalities of the red blood cell sodium-proton exchange in hypertensive patients. Hypertension 17: , Canessa M, Romero JR, Ruiz-Opazo N, and Herrera VL. The alpha 1 Na(+)-K+ pump of the Dahl salt-sensitive rat exhibits altered Na+ modulation of K+ transport in red blood cells. J Membr Biol 134: , 1993.

20 Page 20 of Ceolotto G, Conlin P, Clari G, Semplicini A, and Canessa M. Protein kinase C and insulin regulation of red blood cell Na+/H+ exchange. Am J Physiol 272: C , De Franceschi L, Olivieri O, Miraglia del Giudice E, Perrotta S, Sabato V, Corrocher R, and Iolascon A. Membrane cation and anion transport activities in erythrocytes of hereditary spherocytosis: effects of different membrane protein defects. Am J Hematol 55: , De Franceschi L, Rivera A, Fleming MD, Honczarenko M, Peters LL, Gascard P, Mohandas N, and Brugnara C. Evidence for a protective role of the Gardos channel against hemolysis in murine spherocytosis. Blood 106: , Garnovskaya MN, Mukhin YV, Vlasova TM, and Raymond JR. Hypertonicity activates Na+/H+ exchange through Janus kinase 2 and calmodulin. J Biol Chem 278: , Gilligan DM, Lozovatsky L, Gwynn B, Brugnara C, Mohandas N, and Peters LL. Targeted disruption of the beta adducin gene (Add2) causes red blood cell spherocytosis in mice. Proc Natl Acad Sci U S A 96: , Han BG, Nunomura W, Takakuwa Y, Mohandas N, and Jap BK. Protein 4.1R core domain structure and insights into regulation of cytoskeletal organization. Nat Struct Biol 7: , Ikeda T, Schmitt B, Pouyssegur J, Wakabayashi S, and Shigekawa M. Identification of cytoplasmic subdomains that control ph-sensing of the Na+/H+ exchanger (NHE1): phmaintenance, ATP-sensitive, and flexible loop domains. J Biochem (Tokyo) 121: , 1997.

21 Page 21 of Joiner CH, Franco RS, Jiang M, Franco MS, Barker JE, and Lux SE. Increased cation permeability in mutant mouse red blood cells with defective membrane skeletons. Blood 86: , Lin X and Barber DL. A calcineurin homologous protein inhibits GTPase-stimulated Na-H exchange. Proc Natl Acad Sci U S A 93: , Livne AA, Aharonovitz O, Fridman H, Tsukitani Y, and Markus S. Modulation of Na+/H+ exchange and intracellular ph by protein kinase C and protein phosphatase in blood platelets. Biochim Biophys Acta 1068: , McLean LA, Zia S, Gorin FA, and Cala PM. Cloning and expression of the Na+/H+ exchanger from Amphiuma RBCs: resemblance to mammalian NHE1. Am J Physiol 276: C , Nunomura W, Takakuwa Y, Parra M, Conboy JG, and Mohandas N. Ca(2+)- dependent and Ca(2+)-independent calmodulin binding sites in erythrocyte protein 4.1. Implications for regulation of protein 4.1 interactions with transmembrane proteins. J Biol Chem 275: , Pang T, Hisamitsu T, Mori H, Shigekawa M, and Wakabayashi S. Role of calcineurin B homologous protein in ph regulation by the Na+/H+ exchanger 1: tightly bound Ca2+ ions as important structural elements. Biochemistry 43: , Pang T, Wakabayashi S, and Shigekawa M. Expression of calcineurin B homologous protein 2 protects serum deprivation-induced cell death by serum-independent activation of Na+/H+ exchanger. J Biol Chem 277: , 2002.

22 Page 22 of Pedersen SF, King SA, Rigor RR, Zhuang Z, Warren JM, and Cala PM. Molecular cloning of NHE1 from winter flounder RBCs: activation by osmotic shrinkage, camp, and calyculin A. Am J Physiol Cell Physiol 284: C , Pekrun A, Eber SW, Kuhlmey A, and Schroter W. Combined ankyrin and spectrin deficiency in hereditary spherocytosis. Ann Hematol 67: 89-93, Peters LL, Jindel HK, Gwynn B, Korsgren C, John KM, Lux SE, Mohandas N, Cohen CM, Cho MR, Golan DE, and Brugnara C. Mild spherocytosis and altered red cell ion transport in protein 4. 2-null mice. J Clin Invest 103: , Peters LL, Shivdasani RA, Liu SC, Hanspal M, John KM, Gonzalez JM, Brugnara C, Gwynn B, Mohandas N, Alper SL, Orkin SH, and Lux SE. Anion exchanger 1 (band 3) is required to prevent erythrocyte membrane surface loss but not to form the membrane skeleton. Cell 86: , Pontremoli R, Zerbini G, Rivera A, and Canessa M. Insulin activation of red blood cell Na+/H+ exchange decreases the affinity of sodium sites. Kidney Int 46: , Ritter M, Ravasio A, Jakab M, Chwatal S, Furst J, Laich A, Gschwentner M, Signorelli S, Burtscher C, Eichmuller S, and Paulmichl M. Cell swelling stimulates cytosol to membrane transposition of ICln. J Biol Chem 278: , Savvides P, Shalev O, John KM, and Lux SE. Combined spectrin and ankyrin deficiency is common in autosomal dominant hereditary spherocytosis. Blood 82: , Shi ZT, Afzal V, Coller B, Patel D, Chasis JA, Parra M, Lee G, Paszty C, Stevens M, Walensky L, Peters LL, Mohandas N, Rubin E, and Conboy JG. Protein 4.1R-deficient mice

23 Page 23 of are viable but have erythroid membrane skeleton abnormalities. J Clin Invest 103: , Tang CJ and Tang TK. The 30-kD domain of protein 4.1 mediates its binding to the carboxyl terminus of picln, a protein involved in cellular volume regulation. Blood 92: , Vives Corrons JL and Besson I. Red cell membrane Na+ transport systems in hereditary spherocytosis: relevance to understanding the increased Na+ permeability. Ann Hematol 80: , Voltz JW, Weinman EJ, and Shenolikar S. Expanding the role of NHERF, a PDZdomain containing protein adapter, to growth regulation. Oncogene 20: , Wakabayashi S, Ikeda T, Iwamoto T, Pouyssegur J, and Shigekawa M. Calmodulinbinding autoinhibitory domain controls "ph-sensing" in the Na+/H+ exchanger NHE1 through sequence-specific interaction. Biochemistry 36: , Weinman EJ, Evangelista CM, Steplock D, Liu MZ, Shenolikar S, and Bernardo A. Essential role for NHERF in camp-mediated inhibition of the Na+-HCO3- co-transporter in BSC-1 cells. J Biol Chem 276: , Weinman EJ, Minkoff C, and Shenolikar S. Signal complex regulation of renal transport proteins: NHERF and regulation of NHE3 by PKA. Am J Physiol Renal Physiol 279: F , Weinman EJ, Steplock D, Donowitz M, and Shenolikar S. NHERF associations with sodium-hydrogen exchanger isoform 3 (NHE3) and ezrin are essential for camp-mediated phosphorylation and inhibition of NHE3. Biochemistry 39: , 2000.

24 Page 24 of Weinman EJ, Steplock D, Wade JB, and Shenolikar S. Ezrin binding domaindeficient NHERF attenuates camp-mediated inhibition of Na(+)/H(+) exchange in OK cells. Am J Physiol Renal Physiol 281: F , 2001.

25 Page 25 of FIGURE LEGENDS Figure 1. Changes in erythrocyte Na content in normal control, 4.2-/-, and 4.1-/- mice after a 5-hour incubation in isotonic plasma-like medium. Changes in cell Na content from baseline are presented in the absence or presence of either of the transport inhibitors ouabain, bumetanide, or HOE-642, or various combinations of the three. Data are expressed as mean ± SD of three separate experiments. Statistical analysis compared to 0 drugs (ANOVA): * p values < 0.05, n=3. Figure 2. Effects of a two-hour incubation in hypertonic media on phthalate density profile. Phthalate density profiles for control and 4.1-/- erythrocytes are presented at baseline for isotonic conditions, and hypertonic conditions and following two hours incubation in the absence or presence of the NHE inhibitor HOE-642. Figure 3. Effects of a two-hour incubation in hypertonic media on erythrocyte Na content. Data for cell Na content of control, 4.2+/-, 4.2-/-, and 4.1-/- erythrocytes are presented at baseline and following a two-hour incubation in the absence or presence of the NHE inhibitor HOE-642. Statistical analysis between control and in the presence of HOE642: n=3,* p<0.05 Figure 4. Osmotic activation curve of NHE in control and 4.1 -/- erythrocytes. Na efflux was measured in the presence or the absence of 10 NM HMA in Na-loaded erythrocytes from control () and 4.1 -/- () erythrocytes as described in the Methods section. Kinetic analysis of the curves gave a V max of 9.8 ± 1.3 mmol/l cell x h, and EC 50 of 460 ± 35 mosm for control

26 Page 26 of data. For 4.1 -/- erythrocytes, corresponding values for V max of 18.2 ± 3.2 mmol/l cell x h, and for EC 50 of 417±42 mosm were found. Data are expressed as mean ± SE of triplicate experiments (n=3). Figure 5. H- induced NHE in control and 4.1 -/- erythrocytes. Na influx was measured in acid loaded erythrocytes as described in the Methods section. Kinetic analysis of the curve gave a V max of ± 5.4 mmol/10 13 cell x h and a K 0.5 of ± 0.06 nm (ph=6.31) for control () and a V max of 91.5 ± 7.2 mmol/l cell x h, and a K 0.5 of ± 0.56 nm (ph=6.27) for 4.1 -/- () erythrocytes. The data represent the mean ± SE of triplicate experiments (n=3). Figure 6. Effects of phosphatase and kinase inhibitors on NHE activity. HMA-sensitive Na efflux was measured in control () and 4.1 -/- () erythrocytes. The inhibitors (100 nm) were added to the flux media at time 0. OKA=okadaic acid, CH= chelerythrine, CC= calphostin C, CA= calyculin A. *= p<0.04, **= p<0.03. Data are expressed as mean ± SE of three experiments in duplicate determinations (n=3). Figure 7. Effect of okadaic acid (OKA) on the volume stimulated NHE activity. HMAsensitive Na efflux was measured in control and 4.1 -/- erythrocytes. A) control erythrocytes in the presence () or absence () of 100 nm OKA. B) 4.1 -/- erythrocytes in the presence () or absence () of 100 nm OKA. The values are expressed as mean ± SE of triplicate experiments (n=3).

27 Page 27 of Figure 8. Ion transport via the ouabain-sensitive Na-K pump and the bumetanide-sensitive NKCC in control, 4.2+/-, 4.2 -/-, and 4.1 -/- mouse erythrocytes. For Na-K pump and NKCC, erythrocytes were treated with the nystatin technique to obtain similar intracellular concentrations of Na and K as described in the Methods section. Data are expressed as mean ± SD of three separate experiments. Statistical analysis compared to control animal: *p value <0.05, n=3. Figure 9. Ion transport via KCC in control and 4.1 -/- mouse erythrocytes. K influx was measured in isotonic Cl, hypotonic Cl, and isotonic sulfamate (SFA) media. Volume stimulated and Cl-dependent influxes were calculated as hypotonic Cl minus isotonic Cl, and isotonic Cl minus isotonic SFA, respectively. Data are expressed as mean ± SD of three separate experiments. Statistical analysis between isotonic medium vs. different conditions for control and 4.1 -/-: *p value<0.05, n=3.

28 Page 28 of Figure 1 Change in Cell Na + (mmol/kg Hb) Control * * * * Ouabain Bumetanide HOE * 4.2 -/ * * * * 4.1 -/ * * *

29 Page 29 of Figure 2 Percent Denser Cells Control KO Density Isotonic, baseline Hypertonic, baseline Hypertonic, 2 hrs. Hypertonic, 2 hrs. + HOE642 Percent Denser Cells Density

30 Page 30 of Figure 3 Erythrocyte Na + mmol/kg Hb Baseline 2 hours 2 hours + 50 µm HOE642 * * * * * * 0 +/+ 4.2+/- 4.2-/- 4.1-/-

31 Page 31 of Figure 4

32 Page 32 of Figure 5

33 Page 33 of Figure * * HMA-sensitive Na + efflux mmol/10 13 cell x h ** ** 0 control CH CC OKA CA 100 nm

34 Page 34 of Figure 7

35 Figure /+ 4.2+/- 4.2-/- 4.1-/- 35 +/+ 4.2+/- 4.2-/- 4.1-/- Ouabain-sensitive Na + flux mmol/10 13 cells x h Bumetanide-sensitive efflux mmol/10 13 cells x h Page 35 of 36 A * * Na + efflux K + efflux * * B

36 Page 36 of Figure 9 Deleted: 86 Rb influx mmol/10 13 cell x h 2.0 Isotonic Cl Hypotonic Cl Isotonic SFA Volume-stimulated Flux 1.5 Chloride-dependent Flux * * * 0.0 Control 4.1 -/-