RhoA and Rho Kinase-dependent Phosphorylation of Moesin at Thr-558 in Hippocampal Neuronal Cells by Glutamate*

When we were studying phosphorylated proteins in the rat brain after electroconvulsive shock (ECS), we observed the rapid phosphorylation of a 75-kDa protein, which cross-reacted with the anti-phospho-p70 S6 kinase antibody. The phosphorylated protein was purified and identified as moesin, a member of the ezrin/radixin/moesin (ERM) family and a general cross-linker between cortical actin filaments and plasma membranes. The purified moesin from rat brain was phosphorylated at serine and threonine residues. Moesin was rapidly phosphorylated at the threonine 558 residue after ECS in the rat hippocampus, peaked at 1 min, and returned to the basal level by 2 min after ECS. To investigate the mechanism of moesin phosphorylation in neuronal cells, we stimulated a rat hippocampal progenitor cell, H19–7/IGF-IR, with glutamate, and observed the increased phosphorylation of moesin at Thr-558. Glutamate transiently activated RhoA, and constitutively active RhoA increased the basal level phosphorylation of moesin. The inhibition of RhoA and its effector, Rho kinase, abolished increased Thr-558 phosphorylation by glutamate in H19–7/IGF-IR cells, suggesting that the phosphorylation of moesin at Thr-558 in H19–7/IGF-IR cells by glutamate is mediated by RhoA and Rho kinase activation.

the treatment has not been elucidated, recent progress in the cellular and molecular biology of neuronal tissues has enabled neuronal functions to be linked to the regulation of gene expression and intracellular signal transduction (1). Previously, we reported (4) that ECS induces the expressions of various immediate early genes such as c-fos, junB, and TiS-8 and that it suppresses the expression of inositol 1,4,5-triphosphate kinase and inositol 1,4,5-triphosphate receptor genes. We also observed (2)(3)(4)(5)7) that ECS caused the phosphorylation of many signaling molecules, including those in the Pyk2-Ras-Raf-MEK-ERK pathway, and of stress-signaling pathways in various rat brain regions.
Earlier we reported that ECS induces the phosphorylation of CREB in the rat hippocampus (6). Several protein kinases have been suggested to be CREB kinases in neuronal tissues, and we looked for protein kinases potentially responsible for CREB phosphorylation. When we were studying the phosphorylation of one of the suggested CREB kinases, ribosomal S6 kinase (S6K), we observed that a 75-kDa protein in the rat hippocampus was rapidly phosphorylated after ECS. This was detected by an antibody specific to phospho-p70 S6K. The protein was not detected by the anti-p70 S6K antibody, suggesting that this protein was detected nonspecifically by an anti-phospho-p70-S6K antibody. So we purified and identified this protein as moesin (membrane-organizing extension spike protein), a member of the ezrin/radixin/moesin (ERM) family of proteins. ERM proteins consist of three domains: 1) a globular domain at the N-terminal half, which is conserved among the members of the band 4.1 superfamily and is referred to as the FERM (4.1 and ERM) domain, a membrane-binding domain; followed by 2) an extended ␣-helical domain; and 3) a charged C-terminal domain, which includes a consensus sequence motif for actin binding. Thus, ERM proteins have been suggested to function as cross-linkers between actin filaments and plasma membranes and to be involved in the formation of microvilli, cell adhesion sites, ruffling membranes, and cleavage furrows (8 -13).
The ERM proteins are phosphorylated in growth factor-stimulated cells (14,15). In thrombin-activated platelets, moesin is phosphorylated at a specific C-terminal 558 threonine residue (Thr-558), causing filopodia formation (16). In vitro functional analysis suggested that C-terminal threonine phosphorylation maintains ERM proteins in the active state by suppressing intramolecular interaction (17).
RhoA is a member of the Ras-like GTPase superfamily and has been shown to regulate the actin cytoskeleton and mitogenic signaling in response to extracellular signals (18). In fibroblasts, ERM proteins were phosphorylated and relocalized to apical membrane/actin protrusions in a RhoA-dependent manner (19), and in Swiss 3T3 cells at least one of the ERM * This work was supported by Samsung Biomedical Research Institute Grant B-98-022, the Research Fund of Sungkyunkwan University School of Medicine, and Ministry of Health and Welfare Grant HMP-98-N-1-0009. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
proteins was shown to be required for the RhoA-dependent formation of stress fibers and focal contacts (20). RhoA has been reported to activate several serine/threonine kinases such as ROK/ROCK-II/Rho kinase, ROK/ROCK-I, citron kinase, protein kinase N, and protein kinase C1 (20). Rho kinase effectively phosphorylates moesin at Thr-558 in vivo, and this phosphorylation plays a crucial role in the formation of microvilli-like structures (21).
During development and regeneration, neurons grow over large distances in a process controlled by extrinsic factors and cytoskeletal interaction. The suppression of radixin and moesin altered growth cone morphology, motility, and process formation in primary cultured neurons (29). In addition, functional studies have established that in sympathetic neurons, growth cone collapse induced by nerve growth factor deprivation is concomitant with a significant decrease of the radixin staining of growth cones (22). Taken together, these observations suggest that localization of the ERM protein may be essential for the normal expression of growth cone morphology. However, the phosphorylation of ERM proteins in neuron tissues and cells has not been studied, and it is also unknown whether ERM proteins, including moesin, are involved in neuronal cell cytoskeletal changes. We identified moesin as a protein that showed increased phosphorylation after ECS in the rat brain and report that glutamate induces the phosphorylation of moesin through the activation of RhoA and Rho kinase in rat hippocampal neuronal cells.

Treatment of Animals and the Preparation of Hippocampal Lysates-
Male Sprague-Dawley rats ranging from 150 to 200 g were used in this study. Animals were treated in accordance with National Institutes of Health guidelines for the care and use of laboratory animals. ECS was administered (130 V, 0.5 s) via an ear electrode. Control animals were treated in the same manner as the ECS-treated animals but without the electric current. The animals were decapitated at the prescribed times (0, 1, 2, 5, 10, 30, and 60 min after ECS), and hippocampi were removed onto ice and homogenized immediately with a glass Teflon homogenizer in 10 volumes of ice-cold homogenization buffer (25 mM Tris, pH 7.5, 1 mM EGTA, 2 mM EDTA, 50 mM NaF, 1 mM Na 3 VO 4 , 10 mM sodium pyrophosphate, 0.2% Nonidet P-40, 1 mM PMSF, 1 mM DTT, 10 g/ml aprotinin, 10 g/ml leupeptin, and 10 g/ml pepstatin) and centrifuged for 10 min at 10,000 ϫ g, 4°C. The protein samples were prepared by boiling the supernatant with 3ϫ Laemmli's sampling buffer. Protein was quantitatively assayed using bicinchoninic acid (Sigma) or Bradford reagent (Bio-Rad).
Purification of the 75-kDa Protein-Thirty rat brains, dissected 1 min after ECS, were homogenized in buffer A (25 mM Tris, pH 7.5, 1 mM EGTA, 2 mM EDTA, 50 mM NaF, 1 mM Na 3 VO 4 , 10 mM sodium pyrophosphate, 1 mM PMSF, 1 mM DTT, 10 g/ml aprotinin, 10 g/ml leupeptin, and 10 g/ml pepstatin) at a ratio of 5 ml/g of brain, wet weight. The homogenate was centrifuged for 30 min at 20,000 rpm in a Beckman rotor, and the resulted supernatant was loaded at the speed of 3 ml/min to 100 ml of packed DEAE-Sepharose (Amersham Biosciences) equilibrated in buffer A. The column was washed with 100 ml of buffer A. The eluted and washed fractions were pooled and applied to a 40-ml SP-Sepharose fast flow column (Amersham Biosciences) equilibrated in buffer A at a speed of 2 ml/min. The column was washed with 250 ml of 50 mM NaCl/buffer A and then eluted with 100 ml of 500 mM NaCl/buffer A. Fractions of 4 ml were collected. The fractions containing enriched 75-kDa protein were pooled and dialyzed against buffer A overnight at 4°C. After centrifugation, the pooled sample was applied to a 20-ml column of heparin-Sepharose (Amersham Biosciences). The column was washed with 100 ml of buffer A and eluted with a NaCl gradient (0 -500 mM). The 75-kDa protein was eluted between 100 and 200 mM NaCl. The eluant was dialyzed against buffer A, filtered through a 0.22-M membrane, and applied at 1 ml/min to a Mono-Q HR 5/5 column (Amersham Biosciences) equilibrated with buffer A. The column was washed with 5 ml of buffer A, and the 75-kDa protein was eluted as flow-through and washed out fractions. The eluant was concentrated with a Centricon 50 (Amicon) for glycerol gradients. Aliquots (200 l) were loaded onto 5 ml of a glycerol gradient mixture (from 15 to 35% glycerol/buffer A and 500 mM NaCl) and centrifuged for 24 h at 47,000 rpm in a Beckman SW 55Ti rotor. Elution was carried out with a Beckman recovery system and 11 drops were collected per fraction. To obtain a single band of the 75-kDa protein, preparative gel electrophoresis (Bio-Rad) was carried out at a flow rate of 0.1 ml/min, and 200 l each were collected per fractions.
The presence of 75-kDa protein was determined by immunoblotting with an anti-phospho-p70 S6 kinase antibody, and the protein amount was estimated by Bradford reagent and gel staining with Coomassie Brilliant Blue R-250.
Immunoprecipitation and Immunoblotting-Samples containing 1 mg of the hippocampal lysates were precleared with 50 l of protein A-Sepharose (Pierce) for 1 h. After preclearing, each lysate was incubated with 20 l of anti-moesin monoclonal antibody (BD Transduction Laboratories) and 50 l of protein A-Sepharose for 1 h, and the resulting immune complexes were pelleted at 10,000 ϫ g and washed three times in homogenization buffer. Sample buffer (3ϫ) was added, and samples were incubated at 100°C for 5 min. 80 mg of the hippocampal lysates and 20 g of the cell lysates or immunoprecipitates were then electrophoresed in 8 or 12% SDS-polyacrylamide gel and transferred to nitrocellulose membranes. The membranes were incubated with antiphospho-p70 S6 kinase (Thr-421/Ser-424) (New England Biolabs), antimoesin polyclonal (Santa Cruz Biotechnology), anti-phosphothreonine and anti-phosphoserine (Zymed Laboratories Inc.), anti-RhoA (Santa Cruz Biotechnology), anti-Myc (provided by Prof. M. Han of Sungkyunkwan University, Suwon, Korea) and anti-phospho-ERM (297S) antibodies (kindly provided by Dr. S. Tsukita of Kyoto University, Kyoto, Japan) for 2 h at room temperature. After washing with Tris-buffered saline/Tween (0.05%), the blots were incubated with horseradish peroxidase-conjugated anti-rabbit, anti-rat, or anti-mouse IgGs, and the bands were visualized using the ECL system (Pierce). Molecular weights were determined by comparing the mobilities of the protein bands with prestained molecular weight markers (Invitrogen).
Protein Sequence Analysis-One microgram of purified 75-kDa protein was electrophoresed in 8% SDS-PAGE and stained with Coomassie Blue. The two stained bands were separately excised from the gel and digested with trypsin. Sequence analysis was performed upon the trypsin-digested peptide by matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry.
Production of Antibodies and Recombinant Proteins-A plasmid (pSK moesin) encoding full-length mouse moesin protein was obtained from Dr. S. Tsukita (Kyoto University). To construct the GST fusion proteins, the moesin DNA sequence was amplified by polymerase chain reaction and subcloned into vector pGEX-4T-1 (Amersham Biosciences). The construct was confirmed by DNA sequencing. GST fusion proteins were expressed in DH5␣ cells and purified on glutathione-agarose beads. A moesin polyclonal antibody was produced in rabbit against purified GST-fused full-length mouse moesin. The antibody reacted only with moesin, not with ezrin or radixin, and we have named the antibody TM2. A polyclonal antibody against Thr-558-phosphorylated ERM proteins was raised in rabbits against a KLH-conjugated synthesized phosphopeptide (KYKpTLRQCCCCC, where pT is phosphothreonine) (peptide from Genemed Synthesis; KLH from Pierce), which corresponded to the mouse moesin sequence from amino acids 555-561. The resulting antibody reacted with only phosphorylated ERM proteins but not unphosphorylated ERM proteins.
Cell Culture and Transfection-H19 -7/IGF-IR cells were obtained from the American Type Culture Collection (Manassas, VA) and maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum, 200 g/ml G418, and 1 g/ml puromycin at 34°C under 5% CO 2 . H19 -7/IGF-IR cells were seeded at 2 ϫ 10 5 in a 6-well plate coated with poly-L-lysine. For glutamate treatment, the cells were cultured in serum-free DMEM for 18 h and then stimulated with or without 100 M glutamate for various periods (2, 5, 10, 30, and 60 min). The transfection of plasmids (3.2 g of pcDNA C3, pCMV-Myc RhoAV14, or empty vectors) into H19 -7/IGF-IR cells was carried out using LipofectAMINE 2000 reagent (Invitrogen) according to the manufacturer's instructions. Transient overexpression of the pCMV-Myc RhoAV14 proteins was verified by immunoblotting cell lysates with an anti-Myc monoclonal antibody. The transient overexpression of C3 transferase was detected by the mobility shift of ADP-ribosylated endogenously expressed RhoA (26). The transfected cells were cultured in serum-free DMEM for 20 h and then stimulated with or without 100 M glutamate for 10 min. To examine the involvement of Rho kinase in glutamate-induced moesin phosphorylation, we used a Rho kinase inhibitor, Y-27632 (Tocris Cookson). Before the glutamate treatments, 30 M Y-27632 was incubated with serum-free DMEM for 1 h in H19-7/ IGF-IR cells Rat1 and MDCK were cultured in DMEM supplemented with 10% fetal bovine serum. For lysophosphatidic acid (LPA, Avanti Polar Lipids) treatment, Rat1 cells were seeded at 4 ϫ 10 5 cells in a 6-well plate. The next day, the cells were serum-starved for 18 h in serum-free DMEM. After adding LPA to a concentration of 1 g/ml, the cells were incubated for various periods (0, 2, 5, 10, and 30 min). For phosphorylation analysis, H19 -7/IGF-IR and Rat1 cells were fixed with 10% trichloroacetic acid, washed with phosphate-buffered saline three times, and then solubilized with 100 l of 1 ϫ Laemmli's sample buffer. RhoA, Rac1, and Cdc42 Activity Assays-RhoA, Rac1, and Cdc42 activities were measured according to the modified method of Ren et al. (23). The H19 -7/IGF-IR cells (2 ϫ 10 7 cells) were seeded in 100-mm culture dishes and serum-starved in serum-free DMEM for 18 h before lysis. The cells were then treated with or without 100 M glutamate. For a RhoA activity assay, the cells were lysed with ice-cold buffer (25 mM Tris, pH 7.5, 250 mM NaCl, 5 mM MgCl 2 , 1 mM sodium orthovanadate, 0.25% sodium deoxycholate, 0.05% SDS, 0.5% Triton X-100, 1 mM PMSF, 1 mM DTT, 1 g/ml aprotinin, and 1 g/ml leupeptin) for 15 min on ice. The cell lysates were then passed 10 times through a 23-gauge needle and centrifuged at 10,000 ϫ g for 10 min at 4°C. The supernatants were incubated with 30 g of glutathione-Sepharose beads bound to the GST-RhoA-binding domain of mouse rhotekin for 60 min at 4°C. For Rac1 or Cdc42 activity assays, the cells were resuspended in 200 l of NS buffer (25 mM Tris, pH 7.5, 1.5 mM MgCl 2, 1 mM sodium orthovanadate, 1 mM PMSF, 1 g/ml aprotinin, and 1 g/ml leupeptin) for 10 min on ice and then passed 10 times through a 25gauge needle. After the addition of 200 l of 2ϫ no-salt lysis buffer (50 mM Tris, pH 7.5, 10 mM MgCl 2, 2% Nonidet P-40, 1 mM sodium orthovanadate, 1 mM PMSF, 1 g/ml aprotinin, and 1 g/ml leupeptin), the cell lysate was passed through a 23G needle 10 times and placed on ice for 5 min, and then 400 l of high salt-binding buffer (25 mM Tris, pH 7.5, 30 mM MgCl 2 , 100 mM NaCl, 0.5% Nonidet P-40, 1 mM sodium orthovanadate, 1 mM PMSF, 1 g/ml aprotinin, and 1 g/ml leupeptin) were added, and the cell lysates were centrifuged at 10,000 ϫ g for 10 min at 4°C. The supernatants were incubated with 30 g of glutathione-Sepharose beads bound to the GST-Rac/Cdc42-binding domain of rat PAK3 (GST-PBD) for 60 min at 4°C. The beads were washed three times with wash buffer (25 mM Tris, pH 7.5, 40 mM NaCl, 30 mM MgCl 2 , 1% Nonidet P-40, and 1 mM DTT). The bound proteins were eluted in Laemmli's sample buffer, separated by 15% SDS-PAGE, and analyzed by immunoblotting with mouse monoclonal anti-RhoA, anti-Rac1, and anti-Cdc42 antibodies.
Subcelluar Fractionation-To obtain cytosolic and membranous fractions, Rat1 cells were rinsed once in cold phosphate-buffered saline and then lysed in detergent-free buffer (25 mM Tris, pH 7.5, 150 mM NaCl, 5 mM MgCl 2 , 1 mM sodium orthovanadate, 1 mM PMSF, 1 mM DTT, 1 g/ml aprotinin, and 1 g/ml leupeptin) by passing 10 times through a 25 gauge needle. The lysed cells were pelleted at 100,000 ϫ g for 1 h. The supernatants were used as cytosol, and the pellets as crude membrane. The samples were added with sample buffer (3ϫ) and incubated at 100°C for 5 min.

RESULTS
Purification and Identification of the 75-kDa Phosphoprotein in the Rat Hippocampus-Previously we reported that ECS induces CREB phosphorylation in the rat hippocampus (6). To identify the kinases responsible for the phosphorylation of CREB in the rat hippocampus after ECS, we studied the phosphorylation of Rsk, a cytosolic protein kinase for ribosomal S6, which was reported as a CREB kinase in glial cell progenitors (24). Initially, we examined the mobility shifts of Rsks after ECS in the rat hippocampus and could observe the mobility shifts (data not shown). We then examined the phosphorylation status of S6K in the rat brain after ECS using an anti-phospho-p70 S6K antibody. S6K was highly phosphorylated in the rat hippocampus before stimulation, and we were unable to find any further S6K phosphorylation (Fig. 1, arrowhead). However, we found that this antibody cross-reacted with five more bands (Fig. 1), and three of these showed density changes after ECS (upper, middle, and lower arrows). The antibody against S6K itself did not detect any protein bands coincident with these three proteins, indicating that these proteins might not be directly related to S6K. Because the phospho-specific antibody was generated against the serine-and threonine-phosphorylated peptide of S6K, we presumed that the increase in band intensities suggested the increased phosphorylation of serine and/or threonine residues in these proteins. Of the three bands, we focused on the middle protein, which was a doublet of 75 and 80 kDa. The phosphorylation of the doublet was so rapid that it only lasted 1 min after ECS in the rat hippocampus ( Fig. 1).
To identify the protein, we purified it from the brain lysates of ECS-treated rats (1 min after ECS) using the six-step procedure described under "Experimental Procedures." During the purification steps, we lost the 80-kDa protein and, therefore, were only able to purify the 75-kDa protein to homogeneity ( Fig. 2A, left panel). Because we traced the 75-kDa protein during the purification procedures by immunoblotting, we were unable to estimate purification fold and yield. Coomassie staining of the final preparation revealed two protein bands, a major band with molecular size of 75 kDa and a minor smaller molecular size band. These two proteins both reacted with the anti-phospho-p70 S6K antibody, and their band intensities coincided with the protein amounts ( Fig. 2A, right panel), suggesting that the smaller protein might be a proteolytic product of the major protein. The major band was eluted from the gel and digested with trypsin. The masses of the tryptic peptides were measured by MALDI-TOF. The masses of 10 peptides were analyzed using a data base search program (NCBInr 02.15.00), and it was found that seven of these peptides matched those of rat moesin (data not shown). To confirm whether the purified 75-kDa protein was really a moesin, the purified protein was immunoblotted with a commercially available monoclonal antibody against human moesin (Figs. 2 and 4, ␣-ME). This antibody reacted with the lower band as well as the upper band, and the intensities corresponded well with the amount of purified protein (Fig. 2B, left panel). Although we did not identify the lower band of purified proteins by peptide analysis, these results imply that the lower band may have been caused by a cleavage of moesin during the purification.
Moesin Phosphorylation at Serine and Threonine in the Rat Hippocampus after ECS-After the purified protein had been identified as moesin, we examined the phosphorylation of the purified moesin. Because the anti-phospho-p70 S6K antibody we used was produced against a synthetic peptide dually phosphorylated at the Thr-421 and Ser-424 residues of S6K, we examined the phosphorylation of the serine and threonine residues of the purified moesin with the anti-phosphoserine and anti-phosphothreonine antibodies. Both the anti-phosphoserine and anti-phosphothreonine antibodies reacted with the purified moesin, indicating that the purified moesin was phosphorylated at both the serine and threonine residues (Fig. 2B,  middle and right panels). Because the brain lysate used for the purification of the 75-kDa protein was prepared from rats 1 min after ECS, the results suggested that ECS might increase the phosphorylation of moesin in the rat hippocampus. To   FIG. 1. Immunoblot analysis with the anti-phospho-p70 S6 kinase antibody of rat hippocampal lysate after ECS. After ECS, rats were sacrificed at prescribed times. The hippocampi were separated on ice and homogenized immediately in 10 volumes of homogenization buffer. Tissue lysates were fractionated in 8% SDS-PAGE, transferred to a nitrocellulose membrane, and then immunoblotted with an anti-phospho-p70 S6 kinase antibody. Molecular size standards are shown on the left. S, sham-treated animal. determine whether this was the case, the moesin in the hippocampal lysates of sham and ECS-treated (1 min) rats were immunoprecipitated with anti-moesin monoclonal antibody and immunoblotted with polyclonal anti-phosphoserine and anti-phosphothreonine antibodies. We confirmed that the phosphorylation of moesin at both the serine and threonine residues indeed increased in the rat hippocampus 1 min after ECS (Fig. 2C).
Phosphorylation of Moesin at Thr-558 in the Rat Hippocampus after ECS-Having confirmed that the purified moesin from rat brain was phosphorylated at its serine and threonine residues, we examined the phosphorylation of moesin at Thr-558 in the rat hippocampus after ECS. Thr-558 is a well known phosphorylation site of moesin and is shared by other ERM protein family members (Thr-567 in ezrin and Thr-564 in radixin). The rat hippocampal lysates after ECS were examined with the phospho-Thr-558-specific antibody 297S (kindly provided by Professor Tsukita), which can detect all three phosphorylated ERM proteins. The antibody 297S detected at least five proteins in the rat hippocampal lysate. Three protein bands had molecular sizes bigger than 100 kDa, suggesting that these proteins were not ERM proteins and were detected nonspecifically. The molecular sizes of the other two protein bands, which showed density increases after ECS in accordance with those of the phosphoprotein bands in rat hippocampus detected by the anti-phospho-S6K antibody after ECS, were estimated to be 75 and 82 kDa (Fig. 3A), suggesting that these phosphoproteins were moesin and radixin, respectively. To confirm this, we immunoblotted the deprobed membrane with a polyclonal anti-moesin antibody, which was capable of reacting with moesin and radixin (Figs. 2 and 3, ␣-MR). The polyclonal anti-moesin antibody detected two bands in the hippocampal lysate, which corresponded to the major and minor bands detected by the phospho-specific antibody, 297S. These results showed that ECS increased the phosphorylation of two of the ERM proteins in the rat hippocampus, namely moesin and radixin (but primarily moesin) and that the phosphorylation of these proteins was rapid and transient, peaking at 1 min (4-fold that of the basal level) and returning to the base line 2 min after ECS (Fig. 3B). Because moesin was the major phosphorylated ERM protein in the rat hippocampus and the phosphoprotein we had purified from rat brain after ECS, we decided to focus on moesin phosphorylation.
Production of Anti-moesin and Thr-558-phosphorylated Moesin-specific Antibodies-To obtain the moesin-specific antibody, the purified full length of mouse moesin fused to GST was injected subcutaneously into two rabbits. As shown in Fig. 4A, the polyclonal antibody raised against whole moesin (designated as TM2) reacted only with moesin even in MDCK cells (which expressed large amounts of ezrin) and did not crossreact with ezrin or radixin, whereas the monoclonal antibody moesin (Figs. 2 and 4, ␣-ME) detected two protein bands with molecular sizes of 75 and 85 kDa, which were moesin and ezrin, respectively (Fig. 4A).
In addition, we produced a phospho-ThrERM (p-ThrERM)specific antibody. The polyclonal antibody p-ThrERM was raised in rabbits against the KLH-conjugated synthesized phosphopeptide (KYKpTLRQCCCCC) corresponding to the mouse moesin sequence from amino acids 555-561, which is shared by all three ERM proteins.
To test the reactivity and specificity of this antibody, the lysates of Rat1 cells were immunoblotted after LPA stimulation. A protein band detected by the anti-p-ThrERM antibody increased in intensity 2 min after LPA stimulation, and this was sustained up to 30 min (Fig. 4C). The molecular size of the detected band was 75 kDa, and this coincided well with moesin when the blot was reprobed with the moesin antibody TM2. The temporal pattern of increased moesin phosphorylation was similar to that reported previously (17,19), indicating that the phospho-specific antibody detected the Thr-558-phosphorylated form of moesin. To ensure that this phospho-Thr-558specific antibody was specific for Thr-558-phosphorylated moesin and did not cross-react with unphosphorylated moesin, we immunoblotted the membrane and the cytosolic fractions of Rat1 cell lysates with p-ThrERM and TM2 antibodies. It was found that the phosphorylated moesin localized only to the

FIG. 2. Purification and identification of the 75-kDa protein as moesin.
A, purified 75-kDa protein was electrophoresed in 8% SDS-PAGE and stained with Coomassie Blue (left panel) or immunoblotted with anti-phospho-p70 S6 kinase antibody (right panel). B, purified protein was electrophoresed, transferred to a membrane, and immunoblotted with anti-moesin monoclonal (␣-ME), antiphosphoserine (␣-p-Ser) or anti-phosphothreonine (␣-p-Thr) antibodies. C, 1 min after ECS or sham treatment the rats were sacrificed, and moesin was immunoprecipitated from the hippocampal lysates with the anti-moesin monoclonal antibody. The immunoprecipitates were fractionated by SDS-PAGE, immunoblotted with anti-phosphoserine and antiphosphothreonine antibodies, and then stripped and reprobed with an anti-moesin polyclonal antibody (␣-MR) as a protein content control. S, sham-treated animal; E1, 1 min after ECS. The results are representative of three independent experiments. membrane fraction. Because the antibody did not detect any protein in the cytosolic fraction of Rat1 cells despite the existence of a substantial amount of moesin (Fig. 4B), we concluded that the antibody specifically reacted with the Thr-558-phosphorylated moesin. Even though all the ERM proteins share this peptide sequence containing phosphorylated threonine, the polyclonal p-ThrERM antibody detected only one band in the Rat1 cell lysate. This result suggests that moesin might be the predominant ERM protein in the Rat1 cell. Another possibility was that the phosphorylated statuses of radixin and ezrin in the Rat1 cell might be comparatively low. To address this point, we determined the amounts of ERM proteins in Rat1 cells. Indeed, moesin was found to predominate in Rat1 cells (Fig. 4A). To confirm that the anti-p-ThrERM antibody could detect other phosphorylated ERM proteins, we immunoblotted MDCK cell lysate and observed phosphorylated ezrin and moesin (data not shown).
Glutamate Induces Thr-558 Phosphorylation of Moesin in H19 -7/IGF-IR Cells-Previously, it was reported (24) that tonic-clonic seizure after ECS induced the release of neurotransmitters in the rat hippocampus. Therefore, there is a possibility that ECS may induce the phosphorylation of moesin through the release of neurotransmitters, so we studied whether glutamate could induce the phosphorylation of moesin in H19 -7/IGF-IR-immortalized rat hippocampal progenitor cells. The anti-p-ThrERM antibody detected only one band in the lysate of the H19 -7/IGF-IR cells, and this band corresponded to moesin. Like Rat1 cells, moesin was the predominant ERM protein in H19 -7/IGF-IR cells, and the levels of ezrin and radixin were very low (Fig. 4A). The Thr-558 of moesin was phosphorylated in H19 -7/IGF-IR cells even without glutamate stimulation, as has been described previously (16), but glutamate treatment rapidly increased moesin phosphorylation at Thr-558. This phosphorylation level was doubled in 2 min, was sustained to 30 min, and then declined slowly to reach the base line 60 min after glutamate treatment (Fig. 5, A and B).
RhoA and Rho kinase Are Involved in the Phosphorylation of Moesin by Glutamate-The Rho family proteins RhoA, Rac1, and Cdc42 have been implicated in the phosphorylation of moesin in various cell lines. To determine whether the activation of RhoA, Rac1, or Cdc42 is involved in the phosphorylation of moesin at Thr-558, we examined the activities of RhoA, Rac1, or Cdc42 in H19 -7/IGF-IR cells after glutamate treatment by measuring the amounts of RhoA bound to the GSTfused RhoA-binding domain of rhotekin (GST-RBD) or Rac1/ Cdc42 bound to the GST-fused Rac/Cdc42-binding domain of PAK3 (GST-PBD). Glutamate induced a rapid increase in the amount of cellular GTP-bound RhoA in H19 -7/IGF-IR cells but did not increase in the amount of cellular GTP-bound Rac1 (Fig. 6, A and B). The amount of GTP-bound Cdc42 was barely detectable in H19 -7/IGFR-IR cells even after glutamate treatment, suggesting that the activity of Cdc42 in this hippocampal progenitor cell line was very low and not increased by gluta- mate treatment (data not shown). These results indicate that glutamate activates RhoA, not Rac1 or Cdc42, in the rat hippocampal neuronal cell H19 -7/IGF-IR. The activity of RhoA peaked at 2 min (265% average) and was sustained 10 min after glutamate treatment, which preceded the phosphorylation of moesin at Thr-558, suggesting that RhoA might be involved in the phosphorylation of moesin at Thr-558. To determine whether the inactivation of RhoA might inhibit glutamate-induced moesin phosphorylation, we transiently transfected expression plasmids carrying the C3 cDNA, which encodes a potent inhibitor of RhoA (C3 transferase) (25) or an empty vector into H19 -7/IGF-IR cells. The glutamate-induced moesin phosphorylation was significantly suppressed in the presence of C3 toxin. Because C3 transferase catalyzes the ADP-ribosylation of RhoA on asparagine 41, we examined the mobility of RhoA in C3-transfected hippocampal neuronal cells to confirm that the C3 transferase acted properly. As shown in Fig. 7B, the mobility of RhoA was retarded in C3transfected H19 -7/IGF-IR cells, indicating that RhoA was ADP-ribosylated. Because the suppression of the phosphorylation of moesin by RhoA C3 inhibition indicated that RhoA was involved in the phosphorylation of moesin in H19 -7/IGF-IR cells after glutamate stimulation, as a next step we investigated whether activated RhoA might be sufficient for the phosphorylation of moesin in neuronal cells. Expression plasmids encoding Myc epitope-tagged activated RhoA (RhoAV14) were transfected into H19 -7/IGF-IR cells. The cells were then serum-starved and lysed for immunoblotting. The basal level of moesin phosphorylation at Thr-558 was increased 2-fold in RhoAV14-transfected H19 -7/IGF-IR cells versus that of nontransfected cells, indicating that activated RhoA can increase the phosphorylation of moesin at Thr-558. The expression of RhoAV14 was confirmed by the detection of tagged Myc by using an anti-Myc antibody (Fig. 7C).
Among the downstream effectors of RhoA, Rho kinase has been reported to be involved in the phosphorylation of moesin in various non-neuronal cell lines. Therefore, we examined glutamate-induced moesin phosphorylation in H19 -7/IGF-IR cells after pretreating with the Rho kinase inhibitor Y-27632. The moesin phosphorylation induced by glutamate treatment was completely suppressed in the presence of 30 M Y-27632 (Fig. 8, A and B), suggesting that Rho kinase was involved in the phosphorylation of moesin at Thr-558. Although the inhibition of RhoA and Rho kinase abolished the phosphorylation of moesin at Thr-558 in H19 -7/IGF-IR cells induced by glutamate, its inhibitory effect did not change the basal level of moesin phosphorylation, suggesting that there might be pathways of moesin phosphorylation not involving RhoA and Rho kinase. This was the same picture that emerged in the brain. ECS rapidly induced the phosphorylation of moesin in the rat hippocampus, but we also observed a basal phosphorylation of moesin. Further studies will be needed to elucidate the pathways involved in the basal phosphorylation of moesin at Thr-558 in H19 -7/IGF-IR cells.

DISCUSSION
In the present study, we demonstrate that ECS induced the phosphorylation of moesin at Thr-558 in the rat hippocampus.
The phosphorylation and activation of signaling molecules in vivo is not easy to study because it is difficult to find effective stimuli for tissues or organs. We have used ECS as a unique stimulus for the study of signaling molecule activation and the induction of genes in brain tissues. We have reported previously the activation of the Pyk2-Ras-Raf-MEK-ERK cascade and the activation of other signaling molecules such as CREB and stress kinases in the rat hippocampus after ECS (2, 3, 5-7). However, the mechanism of activation of these signaling molecules remains unknown. We assumed that glutamate, known to be released after ECS and/or depolarization induced by electricity, may be involved in the activation of these signaling molecules. Based on this assumption, we used a hippocampal progenitor cell H19 -7/IGF-IR to study the effects of glutamate on the phosphorylation of moesin. Indeed, the glutamate treatment did induce the phosphorylation of moesin at Thr-558 in H19 -7/IGF-IR cells, suggesting the possibility that ECS may induce the phosphorylation of moesin by the release of glutamate in the rat hippocampus. However, the temporal pattern of moesin phosphorylation by ECS in the rat hippocampus was quite different from that found for glutamate in H19 -7/IGF-IR cells. Moesin phosphorylation was very rapid and transient in the rat hippocampus; it immediately increased after ECS and returned to the basal level after 2 min, whereas the phosphorylation lasted for 30 min after glutamate treatment in H19 -7/IGF-IR cells. This kind of discrepancy was observed even in the activation of other signaling molecules such as ERKs. The phosphorylation of ERKs in the rat hippocampus by ECS was rapid and transient and returned to a basal level 5 min after ECS, whereas the activation of ERKs in rat hippocampal slice cultures lasted longer (3,27). The reasons for these discrepancies are unknown, but they indicate that the activation and inactivation of signal transduction systems in the live brain are more dynamic than that in in vitro cultured cells.
ERM proteins are activated by C-terminal threonine phosphorylation, which maintains ERM proteins in an active state by suppressing intramolecular interactions (28). These proteins play a crucial role in the formation of microvilli, cell-tocell adhesion, the maintenance of cell shape, cell motility, and membrane trafficking. Recent analyses have shown that ERM proteins are not only involved in the cytoskeletal organization but are also involved in signaling pathways. In neuronal cells, moesin and radixin are reported to be important for growth cone development and maintenance (29). However, the functional role of ERM proteins as well as their phosphorylation in mature neurons has not been elucidated. In this study, for the first time, we showed that the ERM protein moesin was phosphorylated in the rat hippocampus by ECS and that glutamate can induce the phosphorylation of moesin at Thr-558 by activating RhoA and Rho kinase. However, we have not studied the functional role of phosphorylated moesin. Thus, it remains to be elucidated whether phosphorylated moesin is also involved in the cytoskeletal organization of neuronal cells and/or other signaling pathways in neuronal tissues.
Although many cultured cells express all three ERM proteins, the amounts of ERM proteins widely vary from cell to cell. In epithelial cells, ezrin expression is much higher than moesin or radixin, and moesin is reported to predominate among ERM proteins in endothelial cells. In neuronal cells, moesin is highly expressed in PC12 cells and glioma cells, whereas the three ERM proteins are equally expressed in astrocytes (30 -32). In the rat hippocampus and hippocampal progenitor cells, moesin predominated ( Figs. 2A and 4A). Thus, moesin is the major phosphorylated ERM protein in the rat hippocampus after ECS and in H19 -7/IGF-IR cells after glutamate treatment. In primary cultured neurons, moesin and radixin but not ezrin regulate growth cone development and cell motility (29). Thus, it seems likely that moesin and radixin are needed for neuronal development and that moesin may be the major ERM protein in terms of neuronal development.
It has been reported that the expression of constitutively active mutants of members of the small GTPase family, Rho, Rac, and Cdc42, induce phosphorylation on the C-terminal threonine residue of ERM proteins (33). In the present study, we found that glutamate transiently activates only RhoA and not Rac1 or Cdc42 in H19 -7/IGF-IR cells and that the inhibition of RhoA resulted in the suppression of moesin phosphorylation induced by glutamate. Our results suggest that RhoA activation is involved in the phosphorylation of moesin at Thr-558 in hippocampal progenitor cells. Previously, LPA-induced RhoA activation was reported to induce the threonine phosphorylation of moesin and changes in the actin-based cytoskeleton of fibroblast cells (17). However, we did not observe any changes in either the morphology or the actin-based cytoskeleton in H19 -7/IGF-IR cells after glutamate treatment (data not shown), which suggests that the activated RhoA in H19 -7/IGF-IR cells may not induce cytoskeletal changes. In neuronal cells, the Rho signaling pathway has classically been implicated in axonal outgrowth, dendrogenesis, cell migration during neural development, exocytosis, and endocytosis (34). In addition, the Rho effector protein citron forms a heteromeric complex with PSD-95 and N-methyl-D-aspartate receptors and is concentrated at postsynaptic sites in hippocampal neurons (35). Although the functional roles of citron have not been defined, previous reports suggest its involvement in the synaptic function. Taken together, it is possible that glutamateinduced RhoA activation and activated RhoA-mediated phosphorylation of moesin in the rat hippocampus and neuronal cells may affect cellular functions other than cytoskeleton changes, such as synaptic vesicle recycling or glutamate receptor endocytosis. Thus, studies on the exact intracellular localizations of moesin and its phosphorylated form and on the proteins bound to them are needed to elucidate the function of moesin in neuronal cells.
Thr-558 of moesin was reported to be effectively phosphorylated in vitro by Rho kinase, phosphatidylinositol-4-phosphate 5-kinase, PKC-, and myotonic dystrophy kinase-related Cdc42-binding kinase (17,21,33,36). However, the kinases for moesin phosphorylation in vivo remain controversial. Originally, Rho kinase was suggested to be a downstream kinase for the phosphorylation of ERM proteins by a constitutively active form of RhoA in Swiss 3T3 cells (37), and this was supported by the phosphorylation of moesin by the dominant Rho kinase in transfected COS cells (21). However, Rho kinase phosphorylated myosin light chain and not ERM proteins in glioma cells (38), and Matsui et al. (28) reported that the inhibition of phosphatidylinositol-4-phosphate 5-kinase, rather than Rho kinase, eliminated the threonine phosphorylation of ERM proteins in NIH 3T3 cells by LPA stimulation. It is not clear whether this discrepancy was caused by differences in the cell lines used for these studies. In the present study, the phosphorylation of moesin at Thr-558 by glutamate treatment in hippocampal progenitor cells was completely suppressed by the inhibitory effect of Rho kinase, which indicates that the phosphorylation is mediated by Rho kinase in this neuronal cell line.
Taken together, our results show that the phosphorylation of moesin at Thr-558 in hippocampal progenitor cells (H19 -7/ IGF-IR) by glutamate treatment is mediated by RhoA and Rho kinase. However, we do not exclude the possibility that kinases other than RhoA and Rho kinase may be involved in the phosphorylation of moesin in hippocampal neuronal cells. We observed a basal level of moesin phosphorylation in H19 -7/ IGF-IR cells and in the rat hippocampus. Moreover, this basal level of phosphorylation in H19 -7/IGF-IR cells was not affected by the inhibition of RhoA or Rho kinase, suggesting that kinases other than Rho kinase are involved in the phosphorylation of moesin and that the activities of these other kinases are not affected by glutamate treatment. We will continue studying to identify those kinases responsible for the basal level of moesin phosphorylation in hippocampal neuronal cells.
In the present study, we identified the phosphorylation of moesin in the rat hippocampus after ECS with the anti-phospho-S6K antibody, which suggests a possible sequence homology between the amino acid sequences around the Thr-558 of moesin and those around the dual phosphorylation site, Thr-421 and Ser-424 of S6K. However, the examination of amino acid sequences revealed no homology between these two sequences, indicating that the detection of moesin phosphorylation by anti-phospho-S6K antibody was nonspecific and not amino acid sequence related.
We also observed the phosphorylation of serine residue(s) in the purified moesin from the rat brain. The phosphorylation of ERM proteins at the serine residue(s) has been reported but not extensively studied. As yet, the phosphorylated serine residue(s) of ERM proteins and the signaling system involved in the phosphorylation of the serine residue(s) remain unknown. Although ECS increased the phosphorylation at the serine residue(s) of moesin, we did not observe serine phosphorylation in moesin in H19 -7/IGF-IR cells after glutamate treatment (data not shown). The intensities of the stimuli may explain the difference. The increase of moesin phosphorylation in the rat hippocampus by ECS was at least twice that observed in H19 -7/IGF-IR cells by glutamate. Therefore, glutamate stimulation may not be strong enough to induce the phosphorylation of moesin at serine residue(s), and a second stimulus such as depolarization may be needed to activate the signaling system for the phosphorylation of moesin at serine residue(s) in the rat brain. Further studies on the phosphorylation of ERM proteins at serine residue(s) are needed to elucidate the mechanism of serine phosphorylation in moesin.
In conclusion, we demonstrate the phosphorylation of moesin at Thr-558 by ECS in the rat hippocampus and by glutamate treatment in the hippocampal progenitor cells H19 -7/IGF-IR. Glutamate also activated RhoA in H19 -7/IGF-IR cells, and this moesin phosphorylation was mediated by RhoA and Rho kinase. To our knowledge, this is the first report on the phosphorylation of moesin in neuronal tissues. However, the signaling pathways involved in the activation of RhoA by glutamate remain to be elucidated.