Upon exposure to hypotonic medium, skate red blood cells swell and then reduce their volume by releasing organic osmolytes and associated water. The regulatory volume decrease is inhibited by stilbenes and anion exchange inhibitors, suggesting involvement of the red blood cell anion exchanger skAE1. To determine the role of tyrosine phosphorylation, red blood cells were volume expanded with and without prior treatment with the tyrosine kinase inhibitor piceatannol. At the concentration used, 130 μM, piceatannol nearly completely inhibits p72syk, a tyrosine kinase previously shown to phosphorylate skAE1 (M. W. Musch, E. H. Hubert, and L. Goldstein. J Biol Chem 274: 7923–7928, 1999). Hyposmotic-induced volume expansion stimulated association of p72syk with a light membrane fraction of skate red blood cells. Piceatannol did not inhibit this association but decreased hyposmotically stimulated increased skAE1 tyrosine phophorylation. Movement of skAE1 from an intracellular to a surface detergent-resistant membrane domain and tetramer formation were not inhibited by piceatannol treatment. Two effects of hyposmotic-induced volume expansion, decreased band 4.1 binding and increased ankyrin, were both inhibited by piceatannol. These results suggest that at least one event requiring p72syk activation is pivotal for hyposmotic-induced increased transport; however, steps that do not require tyrosine phosphorylation may also play a role.
- band 4.1
- detergent-resistant membranes
an environmental stress that cells must respond to is osmolarity. Most cells, when exposed to hyposmotic conditions, expand and then accomplish a regulatory volume decrease (RVD) through loss of solutes, including sugar alcohols, poorly metabolized β-amino acids, trimethylamines, and electrolytes (25, 27). Water is obliged to follow the effluxed solute, thus accomplishing the RVD. A number of solute transport systems have been proposed to mediate the solutes required for the RVD. In some cell systems, K+ loss through K+ channels or KCl cotransport contributes to the solute loss. These systems include red blood cells (RBC) of a number of species (10, 11, 13), including cardiac myocytes (33), osteoblasts (37), intestinal and tracheal epithelial cells (35, 36), and Ehrlich ascites cells (24) as limited examples. Organic anions are also pivotal to the solute loss required for the RVD in many cells. One of these organic anions, the β-amino acid taurine, accumulates to high levels in many cells and shows highly stimulated efflux after volume expansion in RBC (6, 7, 9), cultured renal epithelial cells (28), and neuronal cells (1), for example. Taurine efflux and efflux of other organic anions have been speculated to occur through a volume-stimulated organic anion channel. The precise protein nature of this channel remains controversial. A number of Cl channel proteins have been speculated or demonstrated to mediate the anion efflux (26, 30–32).
Elasmobranch and teleost RBC have been extensively used as model systems to investigate the RVD (3–7, 9, 11, 16). In the case of skate RBC, these cells accumulate millimolar amounts of taurine, which appears to be the solute predominantly responsible for the RVD. The hypotonic-stimulated taurine efflux is inhibited by stilbene and other anion exchange inhibitors (9), suggesting the involvement of this protein in the volume-stimulated organic anion efflux. This hypothesis is further supported by experiments in the Xenopus oocyte system where injection with skate AE1 RNA confers volume-stimulated solute efflux (13).
Recent work demonstrated that the skate RBC express at least three mRNA encoding similar anion exchangers (8). The isoform that is most similar to the human and mouse RBC anion exchanger has been termed skate anion exchanger I (skAE1). Previous work using antisera directed to the amino terminus of the human RBC anion exchanger demonstrated that skate RBC express a protein similar to human AE1 and undergo a number of alterations after volume expansion, including tetramer formation, changes in binding of ankyrin and band 4.1, and exocytosis in lipid rafts (18–23). All of these events occur rapidly upon hypotonic-stimulated volume expansion (within 2–10 min), and therefore a temporal sequence of events has been difficult to determine. The dependence of these events on protein tyrosine phosphorylation was investigated using the tyrosine kinase inhibitor piceatannol. This compound is reasonably specific, inhibiting p72syk at concentrations lower than that for other tyrosine kinases. The present results demonstrate that skAE1 movement to the cell membrane surface and oligomerization of skAE1 is not dependent on tyrosine kinase activation, but changes in skAE1 binding to cytoskeletal protein band 4.1 and ankyrin depend on tyrosine kinase activation.
MATERIALS AND METHODS
Isolation and incubation conditions of skate RBC.
Little skate (Raja erinacea) were caught in Frenchman's Bay, ME, and kept in running seawater tanks before bleeding. Blood was removed in a heparinized syringe from the midtail vein and washed in isotonic (940 mosmol/l) elasmobranch incubation medium (940EIM; composition in mmol/l: 300 NaCl, 5.2 KCl, 2.7 MgSO4, 5 CaCl2, 370 urea, and 15 HEPES, pH 7.4) two times before use. The small buffy coat was removed each time. RBC were resuspended in 940EIM at 50% hematocrit and, when appropriate, treated with 130 μM piceatannol (Biomol, Plymouth Meeting, PA) for 15 min before use. Nonincubated cells were always treated with appropriate dilution of DMSO used to introduce piceatannol. Samples were diluted to 10% hematocrit in either 940EIM or hypotonic 460EIM (460 mosmol/l, NaCl reduced to 100 mM and urea to 250 mM).
Skate RBC were labeled for 2 h with 2 μCi/ml [3H]taurine (Perkin-Elmer Radiochemicals, Boston, MA) in 940EIM. Cells were washed three times with 940EIM with 10 mM, then 1 mM, and then 0.1 mM taurine. RBC were resuspended to 20% hematocrit in 940EIM and diluted 1:10 into either 940- or 460EIM. Fluxes were terminated at 15 min by centrifugation (14,000 g at room temperature for 20 s). The supernatant was removed and counted, the pellet was extracted with perchloric acid (10% wt/vol), and radioactivity was measured by liquid scintillation spectroscopy.
Light membrane isolation.
Light membranes were isolated by a modification of Luria et al. (17). RBC were incubated in varying media for 5 min, pelleted, and snap-frozen in an alcohol-dry ice bath. Cell pellets were resuspended in 10 ml lysis buffer (10 mM HEPES, pH 7.4, 5 mM EDTA, with the Complete protease inhibitor cocktail; Roche Molecular Biochemical, Indianapolis, IN) with 0.1 mM sodium orthovanadate and 1 mM NaF to inhibit phosphatases. Ghosts were prepared by three cycles of lysis in this buffer (centrifugation at 10,000 g for 15 s at 4°C). Pelleted ghosts were resuspended in 2 ml lysis buffer and sonicated gently for a total of 30 s. Unbroken cells, nuclei, and mitochondria were pelleted by centrifugation (10,000 g for 30 s at 4°C), and the postmitochondrial membranes were isolated by centrifugation (100,000 g for 15 min at 4°C). Membranes were solubilized in lysis buffer (10 mM HEPES, pH 7.4, 1 mM EDTA containing 1% vol/vol Triton X-100) for 15 min on ice. Detergent-insoluble material was pelleted (14,000 g for 10 min at 4°C) and solubilized in Laemmli stop solution for Western blots.
For analysis of movement of tyrosine kinases into light membranes upon volume expansion, a fraction consisting of light membranes was isolated to preserve protein-protein interactions. Ghosts and postmitochondrial membranes were prepared as above and resuspended in a buffer of 150 mM NaCl, 10 mM Tris, pH 7.4, and 2 mM EDTA with the protease inhibitor cocktail. The membranes were sonicated at a setting of three using a microtip probe on a Branson sonicator (Danbury, CT) for 30 s and resuspended in 1 ml buffer with 80% (wt/vol) sucrose in the above buffer. The sonicated membranes (2-ml volume) were overlaid with 3 ml each of buffer with 35, 22.5, and 10% sucrose and centrifuged at 100,000 g at 4°C for 3 h. Fractions were collected at the interfaces between the sucrose solutions and concentrated on 10 kDa molecular mass cutoff filters (Millipore, Medford, MA). Aliquots of each fraction were saved for protein determination, and the remainder was mixed with Laemmli stop solution as preparation for SDS-PAGE.
For immunoprecipitation studies, microsomal membranes were solubilized in triple-detergent immunoprecipitation buffer (composition in mmol/l: 150 NaCl, 10 HEPES, pH 7.4, and 1 EDTA with 0.1% wt/vol SDS, 0.5% wt/vol sodium deoxycholate, and 1% vol/vol Triton X-100). Samples were cleared of nonspecific binding with 40 μl of a 50% wt/vol slurry of Pansorbin Staphylococcus aureus cells (EMD Biosciences, San Diego, CA) for 15 min by rotation, and cells were removed by centrifugation (500 g for 30 s at room temperature). skAE1 antiserum was added (2 μl) and rotated overnight at 4°C. Protein G-agarose (40 μl; Pierce Chemical, Rockford, IL) was added and rotated for an additional 120 min. Beads were pelleted and washed three times with immunoprecipitation buffer, samples were eluted with Laemmli stop solution at 65°C for 10 min, and Western blots were generated.
In all experiments, samples were resolved on 10% SDS-PAGE, except for those experiments on oligomer formation when 5% SDS-PAGE was used. Resolved proteins were immediately transferred to polyvinylidene difluoride membranes (Polyscreen; Perkin-Elmer) using 1× Towbin's buffer (25 mM Tris, 192 mM glycine, pH 8.8 with 10% vol/vol methanol). Blots were blocked in 5% wt/vol nonfat dry milk in Tris-buffered saline containing Tween 20 (T-TBS, composition in mmol/l: 140 NaCl, 5 KCl, 10 Tris, pH 7.4 with 0.05% vol/vol Tween 20) for 60 min. Blots were incubated overnight at 4°C with a polyclonal antiserum directed against 17 amino acids of the amino terminus of skAE1 characterized previously (14). After primary antibody, blots were washed five times with T-TBS, incubated with peroxidase-conjugated secondary antibody for 60 min, washed four times with T-TBS, one time with TBS, and then developed using an enhanced chemiluminescent system (Supersignal; Pierce). When blots were analyzed for phosphotyrosine, BSA (3% wt/vol) replaced the milk in the blocking solution. The 4G10 monoclonal anti-phosphotyrosine antibody and rabbit polyclonal anti-p72syk (both from Upstate Biotechnology, Lake Placid, NY) were used.
To determine if oligomers of skAE1 formed, cross-linking studies with the homobifunctional cross-linker BS3 were performed as previously described (18). Postmitochondrial microsomal membranes (100,000 g membrane pellet after removal of mitochondria at 10,000 g) were isolated as above, and an amount containing 200 μg protein was reacted with BS3 (5 mM for 60 min; Pierce Chemical). Reactions were stopped by the addition of 0.1 volume of 1 M Tris, pH 8.0. Samples were mixed with Laemmli stop solution and heated, and Western blots were performed.
Band 4.1 and ankyrin binding.
Binding of the cytoskeletal proteins band 4.1 and ankyrin was performed as previously described (20, 21). Briefly, plasma membranes were isolated as described above, and then peripheral membrane proteins were removed by incubation with spectrin extraction buffer (0.2 mM EDTA, pH 7.4, with protease and phosphatase inhibitors described above). The removal of spectrin by this buffer results in vesicles that are >80% inside-out. The vesicles were pelleted (100,000 g for 10 min at 4°C) and resuspended in 5 ml KI extraction buffer (composition in mmol/l: 1,000 KI, 7.5 NaH2PO4, and 1 EDTA with protease and phosphatase inhibitors as above). The vesicles were pelleted (100,000 g for 10 min at 4°C) and resuspended in binding buffer (composition in mmol/l: 5 NaH2PO4, 1 EDTA, pH 7.4, with protease and phosphatase inhibitors). Ankyrin and band 4.1 were purified from human RBC and iodinated, and binding reactions were performed as described previously (20, 21).
Effect of piceatannol on taurine efflux.
Hyposmotic exposure stimulates a rapid efflux of a number of solutes, including taurine, from skate RBC. To determine whether p72syk inhibition blocked this efflux, cells were prelabeled with [3H]taurine, and taurine efflux was measured. Aliquots of the cells, when appropriate, were treated with piceatannol for the last 30 min of the 120-min taurine labeling period. Piceatannol treatment did not alter the small efflux of taurine measured in cells in isosmotic conditions (Fig. 1). Hyposmotic medium stimulated a large increase in taurine efflux that was inhibited by piceatannol treatment (Fig. 1). Therefore, similar to results on taurine uptake, piceatannol treatment inhibits the taurine efflux (22) stimulated by hyposmotic exposure.
Effect of piceatannol on hyposmotic-induced skAE1 tyrosine phosphorylation.
To confirm that piceatannol inhibited hyposmotic-induced increases in skAE1 tyrosine phosphorylation, skAE1 was immunoprecipitated from cells in isosmotic or hyposmotic conditions, either untreated or treated with 130 μM piceatannol. Basal tyrosine phosphorylation of skAE1 was only slightly decreased by piceatannol treatment (Fig. 2, bottom). After hyposmotic-induced volume expansion (10 min), the level of phosphotyrosine increased ∼100%, and this increase was nearly entirely inhibited by piceatannol treatment (Fig. 2, bottom). No differences in the total skAE1 were observed (Fig. 2, top).
Membranes are heterogeneous in structure, and microdomains of specialization may exist. One region of interest is domains high in cholesterol with adjacent sphingo- and phospholipids with highly unsaturated fatty acids, regions known as lipid rafts. These lipid rafts are stable in nonionic detergents, such as Triton X-100. Because of the high cholesterol and tight packing with the unstaturated fatty acids, these regions are less dense than most membranes and can be isolated by their differential density in nonionic detergent-insoluble material. Attempts to detect p72syk in the detergent-resistant membranes, under isosmotic or hyposmotic conditions, were not successful (data not shown). It is possible that the Triton extraction removes many peripheral-associated proteins. Therefore, an alternative technology analyzing light membranes was used instead of buoyant membranes derived from the detergent-insoluble membranes (17). Postmitochondrial membranes were sonicated and placed in 40% sucrose and a discontinuous sucrose gradient placed above. Membranes accumulated at the interfaces between the 40 and 35%, 35 and 22.5%, and 22.5 and 10% regions of the gradient. Because the cholesterol/lipid packing makes the lipid rafts less dense, these regions should be at the 22.5–10% interface. As confirmation that the membrane fraction at the interface of the 10 and 22.5% sucrose solutions was enriched in lipid rafts, the marker flotillin-2 was used. This member of the band 7 family is highly enriched in lipid rafts in skate RBC (23). The light membranes possessed abundant flotillin-2, whereas lesser amounts were detected in membranes from the two other interfaces of the sucrose gradients (Fig. 3, top). These membranes were then analyzed for p72syk. Under isosmotic conditions, with or without piceatannol, very little p72syk could be found in the light membranes. Analyzing membranes at 5 min after volume expansion, a time that precedes the maximal increase in tyrosine phosphorylation (20), increased p72syk was detected in the light membranes. Treatment with piceatannol did not, however, alter the hyposmotic-induced redistribution of p72syk (Fig. 3, middle). As a control, aliquots of the total p72syk were also measured, and no differences were observed (Fig. 3, bottom).
Piceatannol does not inhibit skAE1 movement to the surface or oligomerization.
Hyposmotic conditions induce appearance of skAE1 on the surface (23). However, the dependence on p72syk is not known. It should be noted that the skAE1 that appears on the surface resides in detergent-resistant regions (DRM). This surface-expressed skAE1 arises from skAE1 within DRM in membranes inside the cell, possibly in an endosomal compartment. In isosmotic conditions, although one-third of the cell skAE1 resides in DRM, little of this can be surface biotinylated, suggesting an intracellular compartment. It should be noted that the form of biotin used, Sulfo-N-hydroxysuccinimido biotin, is membrane impermeant, labeling only proteins that face outside the cell. Piceatannol treatment under isosmotic conditions did not alter surface expression or the distribution of skAE1 in DRM (Fig. 4). Upon hyposmotic-induced expansion, the amount of skAE1 in DRM increases. The most notable change is that skAE1 can be surface biotinylated after volume expansion and therefore is on the outside surface membrane (Fig. 4). Piceatannol treatment did not inhibit the hyposmotic-induced appearance of skAE1 on the surface, thus suggesting that activation of p72syk is not required for this event.
In cells under isosmotic conditions, skAE1 is nearly evenly distributed in monomers and dimers, as assessed by the ability to be cross-linked by the homobifunctional cross-linker BS3. This distribution was not affected by piceatannol treatment (Fig. 5). Hyposmotic-induced volume expansion stimulates association of skAE1, and tetramers comprise over one-third of the skAE1 (Fig. 5). The formation of tetramers was not altered by piceatannol treatment, demonstrating that p72syk activation was also not required for this step. Whether the syk is bound to the tetramers, but is not required for their formation, is a question that is under investigation.
Tyrosine kinase inhibition blocks changes in band 4.1 and ankyrin binding.
Similar to its mammalian counterparts, skAE1 has regions in the cytoplasmic amino terminus that would be predicted to bind ankyrin or band 4.1. The site for band 4.1 binding to human AE1 has been identified as a five-amino-acid section, IRRRY. Skate AE1 possesses a IKRRY. Thus this site appears to be conserved (8). Band 4.1 may bind to other proteins in the skate RBC membrane, and only a portion of the band 4.1 membrane binding is inhibited by inclusion of the peptide IRRRY or purified cytoplasmic domain of band 3 (21). Hypotonic-stimulated volume expansion decreases the IRRRY-sensitive binding of band 4.1 by 40% (Fig. 6).
Another cytoskeletal protein in which binding is altered by volume expansion is ankyrin. Under isoosomotic conditions, ankyrin binds to KI-stripped inside out vesicles, and over 80% of this binding can be inhibited with the purified cytoplasmic domain of band 3 (20). Exposure to hypotonic medium increases the number of binding sites, and much of this binding is to sites of higher affinity (Fig. 7). Pretreatment with piceatannol did not alter the binding under isoosomotic conditions. However, the formation of the high-affinity sites after hyposmotically stimulated volume expansion was essentially abolished after piceatannol (Fig. 7).
The regulation of the activity of a number of transport proteins has been the focus of many investigations. A variety of regulatory events has been hypothesized, including changes in activity by direct phosphorylation events, regulation of activity by interaction with cytoskeletal proteins, and as altered distribution among membrane microdomains (2, 34). Hyposmotic-induced increases in taurine efflux in skate RBC appear to depend on a number of these regulatory factors. To help elucidate the dependence on protein tyrosine phosphorylation, we have used the p72syk tyrosine kinase inhibitor piceatannol and investigated changes in the distribution and associations of the skate RBC anion exchanger skAE1. The present results demonstrate that, although p72syk inhibition inhibits taurine efflux stimulated by hyposmotic medium and increased tyrosine phosphorylation of skAE1, skAE1 movement to the surface and oligomerization are not affected. Piceatannol does, however, inhibit decreased band 4.1 binding and high-affinity ankyrin binding that occur after hyposmotic-induced volume expansion.
We hypothesize that an early step in the conversion of skAE1 to mediate taurine efflux is movement to the plasma membrane. This step does not require p72syk activation. We cannot yet address the question whether tetramers are formed before the movement of skAE1 to the surface membrane or whether this occurs afterward. We have attempted to cross-link after surface biotinylation but were unsuccessful, perhaps because of the fact that both reagents label epsilon amino groups of lysine residues, and they may share the same sites in skAE1. It is clear, however, that oligomerization does not require p72syk activation. Potentially, the oligomerized skAE1 recruits p72syk because of conformational changes and results in phosphorylation of tyrosine residues. Preliminary results performing proteolytic digests suggest that more than one tyrosine residue occurs, and we are currently investigating the identity and potential roles of these individual sites. One tyrosine in particular might be predicted to play a role in the band 4.1 binding. The motif IRRRY mediates an interaction of human band 4.1 with human band 3, the human AE1 (12). In the skate, this motif is replaced by IKKRY, and this peptide or IRRRY will displace band 4.1 binding from skAE1. We do not know if the state of this specific tyrosine residue phosphorylation changes upon hyposmotic-induced volume expansion, but it is under investigation. Whether decreased band 4.1 binding is permissive for high ankyrin binding to occur has not been determined. However, the formation of high-affinity ankyrin binding to skAE1 requires a piceatannol-sensitive tyrosine kinase. It appears that tyrosine phosphorylation of skAE1 may play a pivotal role, since neither ankyrin nor band 4.1 tyrosine phosphorylation is increased upon hyposmotic-induced volume expansion (data not shown). We hypothesize that the IKKRY motif may mediate band 4.1 binding, which is potentially altered through phosphorylation. However, less is known about the ankryin interactions with skAE1. There may be ankyrin binding sites of low and high affinity, or there may be one population that has greater access under volume-expanded conditions and that binds with a higher affinity under these conditions. This question may be difficult to address using a heterologous expression system, since interactions between different cell-specific proteins may be required.
A number of other protein motifs of skAE1 may participate in recruitment of not only kinases but also adaptor/accessory proteins to skAE1. When the protein sequence of skAE1 (TRMBL Q7T1P6) is analyzed using the Scansite 1.0 software (scansite.mit.edu), an SH2 site at tyrosine residue 891 is predicted. SH2 domains are known to mediate the interactions of a wide variety of proteins and could mediate skAE1 self interaction or interaction with p72syk.
The role of altered binding of ankryin and band 4.1, and potentially other interacting proteins, as well as tetramer formation, remains to be precisely determined. Increased skAE1 tyrosine phosphorylation may be important in regulation of activity through interactions with other proteins or might directly alter activity of the exchanger. Tetramer formation, and altered protein interactions (e.g., ankyrin and band 4.1), might regulate removal of skAE1 from the surface membrane during the RVD.
This work was supported by National Science Foundation Grants LG IBN-0412702 (to L. Goldstein) and National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-47722 (to M. W. Musch).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society