Phosphorylation of αB-Crystallin in Response to Various Types of Stress*

Phosphorylation of αB-crystallin, a member of the hsp27 family, in human glioma (U373 MG) cells was stimulated by exposure of the cells to various stimuli, which included heat, arsenite, phorbol 12-myristate 13-acetate (PMA), okadaic acid, H2O2, anisomycin, and high concentrations of NaCl or sorbitol, but not in response to agents that elevated intracellular levels of cyclic AMP. Cells exposed to PMA together with okadaic acid yielded three bands of 32P-labeled αB-crystallin when immunoprecipitated samples were subjected to electrophoresis on an isoelectric focusing gel. All of the phosphorylated residues were identified as serine, an indication that three different serine residues can act as sites of phosphorylation in αB-crystallin. Structural analysis by mass spectrometry revealed that phosphorylation of αB-crystallin occurred at serines 19, 45, and 59. Dithiothreitol and staurosporine selectively inhibited the phosphorylation induced by arsenite and the phorbol ester, respectively. SB202190, an inhibitor of p38 mitogen-activated protein (MAP) kinase, suppressed the phosphorylation induced by arsenite, anisomycin, H2O2, sorbitol, NaCl, and heat shock, but not that induced by PMA and okadaic acid. The PMA-induced phosphorylation was selectively suppressed by an inhibitor of p44 MAP kinase kinase, PD98059. Although PMA and arsenite preferentially stimulated the phosphorylation of Ser-45 and Ser-59, respectively, as determined with antibodies that recognized the respective phosphorylated forms of αB-crystallin, all three sites were phosphorylated in response to each stimulus. These results suggest that p38 MAP kinase or p44 MAP kinase might be involved in the signal transduction cascade that leads to the phosphorylation of αB-crystallin. The phosphorylation of αB-crystallin was also enhanced in the heart and diaphragm when rats were exposed to heat stress (42 °C for 20 min).

A major portion in the eye lens of vertebrates is ␣-crystallin, which is found as large aggregates of two closely related subunits, ␣A and ␣B (1). Both ␣A-crystallin and ␣B-crystallin are also present in other tissues (2)(3)(4)(5)(6)(7)(8)(9). Each subunit is highly homologous in terms of amino acid sequence to a small heat shock protein, hsp27 (10,11), and furthermore, the synthesis of ␣B-crystallin is induced by the physiological and nonphysi-ological stimuli that induce the synthesis of hsp27 (12,13). We have shown that hsp27 copurifies with ␣B-crystallin as a large aggregate from skeletal muscle (14).
Phosphorylation is one of the major types of post-transcriptional modification of hsp27, and this phenomenon has been intensively investigated. Phosphorylation of hsp27 is observed when cells are exposed to heat (15,16), arsenite (17), and phorbol ester, okadaic acid, tumor necrosis factor ␣, or interleukin-1␣ (IL-1␣) 1 (17)(18)(19)(20)(21)(22). It has been suggested that phosphorylation of hsp27 is catalyzed by MAP kinase-activated protein kinase-2 (23,24), which itself is activated by a novel protein kinase (p38 MAP kinase) cascade that can be triggered by stress (25,26). However, the physiological relevance of the phosphorylation of hsp27 has not been elucidated. It has been reported that phosphorylation of hsp27 results in the dissociation of an aggregated form of hsp27 to oligomers (27), and phosphorylated hsp27 no longer exhibits the ability to inhibit the polymerization of actin (28).
Reports on the phosphorylation of ␣B-crystallin are mostly related to the phosphorylation of this protein in the lens. Phosphorylated forms of ␣B-crystallin (␣B 1 -crystallin) have been found in bovine lens (29,30), and ␣B-crystallin in extracts of bovine lens can be phosphorylated in a cAMP-dependent manner (31,32) or by the kinase activity of the protein itself (33) in vitro. Recently, Wang et al. (34) reported that H 2 O 2 stimulated the phosphorylation of ␣B-crystallin in rat lens, and Bennardini et al. (35) reported that phorbol 12-myristate 13-acetate induced the phosphorylation of ␣B-crystallin in cultures of bovine articular chondrocytes.
In this report, we show that phosphorylation of ␣B-crystallin in U373 MG human glioma cells and in rat tissues can be induced by various stimuli, and we also identify the sites of phosphorylation in ␣B-crystallin, as determined by use of antibodies that recognized each of the phosphorylation sites specifically.

EXPERIMENTAL PROCEDURES
Reagents-Phorbol 12-myristate 13-acetate (PMA), 4-␣-phorbol 12myristate 13-acetate (4␣-PMA), okadaic acid, and staurosporine were obtained from Wako Pure Chemicals (Osaka, Japan). Dithiothreitol was obtained from Nacalai Tesque Inc. (Kyoto, Japan). SB202190 and PD98059 was obtained from Calbiochem-Novabiochem Corp. (La Jolla, CA). Recombinant human IL-1␣ and tumor necrosis factor ␣ (with specific activities of 2 ϫ 10 7 and 2.55 ϫ 10 6 units/mg of protein, respectively) were obtained from Dainippon Pharmaceuticals (Osaka). Forskolin, cholera toxin, anisomycin, and protein A were obtained from Sigma. Dibutyryl cAMP was obtained from Boehringer (Mannheim, * This work was supported in part by a grant-in-aid for scientific research on priority areas from the Ministry of Education, Science, and Culture of Japan. 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.
Culture and Treatment of Cells-U373 MG cells (obtained from American Type Culture Collection, Rockville, MD) were grown in Eagle's minimal essential medium (Nissui Pharmaceutical Co., Tokyo) supplemented with 10% fetal calf serum (Life Technologies, Inc., Tokyo) at 37°C in a humidified atmosphere of 95% air and 5% CO 2 . The cells were seeded on 90-mm dishes, and the medium was changed every 2 or 3 days. When the cells had reached confluence, various chemicals, dissolved in dimethyl sulfoxide or in H 2 O, were added to the culture medium. After a 90-min incubation in a CO 2 incubator, unless otherwise specified, cells were washed three times with phosphate-buffered saline (8 g of NaCl, 0.2 g of KCl, 1.15 g of Na 2 HPO 4 , and 0.2 g of KH 2 PO 4 in 1000 ml of H 2 O) and frozen at Ϫ80°C for a few days prior to analysis. The frozen cells on each dish were collected and suspended in 0.5 ml of 50 mM Tris-HCl, pH 7.5, containing 0.1 M NaF, 0.1 M okadaic acid, and 5 mM EDTA. Each suspension was sonicated and centrifuged at 125,000 ϫ g for 20 min at 4°C. The supernatant was used for analysis by isoelectric focusing (IEF) of ␣B-crystallin. In some experiments, pellets were washed once with the Tris buffer by sonication as described above and then solubilized with 8 M urea for analysis by IEF.
IEF and Western Blot Analysis-Isoelectric focusing was performed as described by O'Farrell (36) using the Protean II system from Bio-Rad (Tokyo). In brief, an aliquot of cell extract that contained 10 -40 g of protein was mixed with 4 volumes of sample buffer (2% Ampholine mixture (4 parts Ampholine pH 6 -8 and 1 part Ampholine pH 3.5-10), 9.5 M urea, 2% Nonidet P-40, and 5% 2-mercaptoethanol). Fifty-microliter aliquots of the mixture were then applied to a gel composed of 9.2 M urea, 2% Ampholine mixture, 4% acrylamide, and 2% Nonidet P-40, and electrophoresis was performed at 400 V for 16 h at 16.5°C. For Western blot analysis, proteins on a gel were transferred electrophoretically to a polyvinylidene difluoride membrane (type GV; Nihon Millipore Ltd., Yonezawa, Japan), and the membrane was immunostained, as described previously (8), with antibodies against the carboxyl-terminal decapeptide (8) or against specifically phosphorylated forms of ␣B-crystallin that had been prepared as described below.
Immunoprecipitation of 32 P-Labeled ␣B-Crystallin-Cells were labeled with [ 32 P]orthophosphate as described previously (27), and then they were incubated for 90 min in a CO 2 incubator with 1 M PMA and 0.2 M okadaic acid. The cells were washed three times with phosphatebuffered saline and then frozen at Ϫ80°C for 2 h. The frozen cells were scraped off the plates, disrupted by passage through a 24-gauge needle, and then centrifuged as described above. The soluble fraction of the cells (100 g of protein) was incubated at 4°C for 5 h with 5 g of affinity-purified rabbit IgG against ␣B-crystallin (8), and then 50 l of a suspension of protein A-Sepharose were added to the mixture with further incubation at 4°C overnight. The Sepharose beads were washed three times with 0.5 ml of 50 mM Tris-HCl, pH 8.0, containing 1 M NaCl and 0.1% Nonidet P-40 and then once with 50 mM Tris-HCl, pH 7.0. The beads were subjected to rapid agitation on a mixer for 5 min with 50 l of the sample buffer for IEF, and then the mixture was centrifuged at 10,000 ϫ g for 5 min. The supernatants were subjected to IEF or analysis of phosphoamino acids.
Analysis of Phosphoamino Acids-The immunopurified extract was subjected to SDS-polyacrylamide gel electrophoresis and then transferred electrophoretically from the gel to a polyvinylidene difluoride membrane. The membrane was subjected to autoradiography. Each band containing phosphorylated ␣B-crystallin was cut out and subjected to hydrolysis in 6 N HCl at 110°C for 1 h. The hydrolysate was then evaporated in dryness under nitrogen gas. The residue was dissolved in 5 l of formate/acetate buffer (formic acid, acetic acid, and water (26:78:900, v/v)), pH 1.9, supplemented with 67 g each of phosphoserine, phosphothreonine, and phosphotyrosine (Sigma) and applied to a silica gel plate (Merck, Darmstadt, Germany). Electrophoresis was performed in the above buffer at pH 1.9 and 30 mA for 40 min. Electrophoresis in the second dimension was performed in a different buffer (acetic acid, pyridine, and water (10:1:200, v/v)) at pH 3.5 and 35 mA for 30 min. The phosphoamino acids on the dried plate were visualized by staining with ninhydrin, and then the plate was subjected to autoradiography at Ϫ80°C.
Peptide Mapping and Identification of Phosphoamino Acids by Mass Spectrometry-The sites of phosphorylation in ␣B-crystallin were determined by high-performance liquid chromatography-electrospray ionization/triple stage quadrupole mass spectrometry (HPLC-ESI/TSQMS) (37). The system consisted of a Model 140A liquid chromatograph (Perkin-Elmer Applied Biosystems, Foster City, CA) that was connected to a Model TSQ-700 mass spectrometer equipped with an ESI interface (Finnigan MAT, San Jose, CA). ␣B-Crystallin (2 g), purified from U373 MG cells that had been exposed to 1 M PMA and 0.2 M okadaic acid for 90 min as described below, was digested with tosylphenylalanyl chloromethyl ketone-treated trypsin at 37°C for 5 h in 25 mM Tris-HCl, pH 8.0, at a enzyme/substrate ratio of 1:50 (w/w). The digest was fractionated by reversed-phase chromatography on a capillary column (0.5 mm, inner diameter, ϫ 150 mm) packed with Aquapore C18 (particle size, 5 mm; Perkin-Elmer Applied Biosystems) with a gradient of 0 -80% acetonitrile in 0.1% formic acid over 20 min at a flow rate of 8.5 ml/min. The eluate was introduced directly to the ESI interface of the mass spectrometer. After mapping of tryptic peptides, the phosphopeptides were analyzed by tandem mass spectrometry (MS/MS) for identification of the sites of phosphorylation. The spectrometer was operated under the following conditions: electrospray voltage, 4.5 kV; temperature, 200°C; and electron multiplier voltage, 1000 V for peptide mapping and 1600 V for sequencing.
Purification of ␣B-Crystallin from Extracts of U373 MG Cells-For analysis of phosphorylation sites by mass spectrometry, ␣B-crystallin was purified from extracts of U373 MG cells that had been exposed to 1 M PMA plus 0.2 M okadaic acid for 90 min. The cells were scraped from ϳ100 dishes (90 mm in diameter), sonicated, and centrifuged as described above. The supernatant (ϳ80 ml) was incubated at room temperature with 0.5 mg of affinity-purified antibodies against the carboxyl-terminal decapeptide of ␣B-crystallin. After 2 h of a gentle shaking, 1 ml of a 50% (v/v) suspension of protein A-coupled Sepharose beads (Sigma) was added to the mixture with additional incubation at 4°C overnight. The Sepharose beads were washed with 50 mM Tris-HCl, pH 8.0, containing 1 M NaCl and 0.1% Nonidet P-40 and then with 50 mM Tris-HCl, pH 7.0. ␣B-Crystallin trapped on the beads was eluted with 0.1 M sodium acetate buffer, pH 4.5, containing 7 M urea, 1 mM glycol ether diaminetetraacetate, and 1 mM dithiothreitol. The eluate was then applied to a column (0.8 cm, inner diameter, ϫ 7.5 cm) of TSK-SP-5PW (Tosoh, Tokyo), and ␣B-crystallin was eluted with a linear gradient of NaCl (0 -0.4 M) in the above buffer as described previously (38).

Treatment of Rats and Preparation of Tissue Extracts-Male
Wistar rats (body weight, ϳ250 g) were treated in accordance with the guidelines of the Animal Care and Use Committee of the Institute for Developmental Research. The rats were subjected to heat stress as described previously (40). After 20 min of heat stress at 42°C, rats were killed under ether anesthesia, and tissues were dissected out on dry ice and kept frozen at Ϫ80°C for a few days. The frozen tissues were homogenized in 10 volumes of 50 mM Tris-HCl, pH 7.5, containing 0.1 M NaF, 5 mM EDTA, and 0.1 M okadaic acid, and each suspension was sonicated and centrifuged at 125,000 ϫ g for 20 min at 4°C. Supernatants were subjected to IEF followed by Western blot analysis. The insoluble fraction was also analyzed as described above.
Other Methods-Levels of cyclic AMP in cells were determined using a cAMP enzyme immunoassay system (EIA, Amersham International, Buckinghamshire, United Kingdom). Concentrations of protein in soluble fractions of cells and tissues and in fractions solubilized with urea from pellets after centrifugation were determined with a protein assay kit (Bio-Rad), with bovine serum albumin as the standard (8). Bovine ␣B 1 -crystallin and ␣B 2 -crystallin and rat ␣-crystallin, used as standards for IEF, were purified from lenses as described previously (8).

Various Stimuli Induce the Formation of Acidic Forms of
␣B-Crystallin in U373 MG Cells-U373 MG human glioma cells were exposed to NaAsO 2 at various concentrations for 60 min, and then cell extracts were subjected to IEF with subsequent Western blot analysis. As shown in Fig. 1A, arsenite acted in a dose-dependent manner to induce the generation of a form of ␣B-crystallin that had a lower isoelectric point than the control. Since the newly induced form of ␣B-crystallin had an isoelectric point similar to that of ␣B 1 -crystallin, a phosphorylated form of ␣B-crystallin, it appeared that the observed modification of ␣B-crystallin under the above conditions was due to phosphorylation. The arsenite-induced phosphorylated form of ␣B-crystallin was detected after 20 min of exposure, and its level increased in a time-dependent manner for 120 min (Fig. 1B). When cells that had been exposed to arsenite for 120 min were incubated for 2 h in the standard medium, the phosphorylated form of ␣B-crystallin disappeared (Fig. 1B). These results suggest that considerable dephosphorylating activity was present in the cells.
We next examined whether or not okadaic acid, an inhibitor of phosphoserine/phosphothreonine protein phosphatases, could induce the accumulation of the phosphorylated form of ␣B-crystallin. As shown in Fig. 2A, exposure of cells to 0.2 M okadaic acid stimulated the accumulation of ␣B-crystallin in a time-dependent manner. These results suggest that the phosphorylation and dephosphorylation of ␣B-crystallin in cells are under dynamic equilibrium. The band that migrated between the unphosphorylated (p0) and phosphorylated (p1) forms, which was also detected in the preparation of ␣-crystallin purified from lens, seemed to be ␣B-crystallin that had received other modification. Generation of the phosphorylated form of ␣B-crystallin was also observed when cells were exposed to IL-1␣ and PMA, a potent activator of protein kinase C, although the extent of the phosphorylation induced by IL-1␣ was smaller than that induced by arsenite or PMA (Fig. 2B). 4␣-PMA, an inactive analog of PMA, and tumor necrosis factor ␣ each barely induced any phosphorylation of ␣B-crystallin (Fig.  2B).
It has been reported that ␣B-crystallin in extracts of bovine lens can be phosphorylated in a cAMP-dependent manner in vitro (31,32). However, activators of protein kinase A, such as dibutyryl cAMP, forskolin, and cholera toxin, barely induced any phosphorylation of ␣B-crystallin in U373 MG cells (Fig.  2C).
Other compounds, such as H 2 O 2 (oxidative stress), sorbitol and NaCl (hypertonic stress), and anisomycin (an activator of p38 MAP kinase) (41), also stimulated the phosphorylation of ␣B-crystallin in the soluble fraction of U373 MG cells (Fig. 3A). The phosphorylated form of ␣B-crystallin was also detected in the insoluble fraction of cells that had been treated with H 2 O 2 ( Fig. 3A) or with arsenite or PMA (data not shown). However, except in the case of heat stress, most of the phosphorylated form of ␣B-crystallin was detected in the soluble fraction of the cells.
The phosphorylated form of ␣B-crystallin was not detected in the soluble fraction of cells that had been heated at 43-47°C for 20 min (Fig. 3B). ␣B-Crystallin in cells is converted from a soluble form to an insoluble form soon after heat treatment (13). Therefore, we solubilized the insoluble fractions of cell extracts using 8 M urea and then subjected the resultant mixtures to IEF and subsequent Western blot analysis. As shown in Fig. 3B, we detected the heat stress-induced and temperaturedependent phosphorylation of ␣B-crystallin in the insoluble fraction of U373 MG cells.
The arsenite-induced phosphorylation of ␣B-crystallin was suppressed in the presence of 2 mM dithiothreitol, whereas the PMA-induced phosphorylation was selectively inhibited by 100 nM staurosporine, an inhibitor of protein kinase C (Fig. 4A). By contrast, the phosphorylation of ␣B-crystallin that was stimulated by anisomycin (Fig. 4B) or by other chemicals (data not shown) was unaffected by the presence of dithiothreitol or staurosporine. However, an inhibitor of p38 MAP kinase, SB202190 (42), inhibited the phosphorylation that was induced by arsenite, anisomycin, H 2 O 2 , sorbitol, NaCl, and heat, but not the phosphorylation that was induced by PMA and okadaic acid (Fig. 5). On the other hand, an inhibitor of p44 MAP kinase kinase, PD98059 (43), selectively suppressed the phosphorylation of ␣B-crystallin inducible by PMA (Fig. 5). The phosphorylation that was induced by anisomycin (Fig. 5) or by each of the other stimuli described above (data not shown) was barely inhibited by the presence of PD98059. These results indicate that the phosphorylation of ␣B-crystallin by the various agents involves different signal transduction cascades.
Three Different Serine Residues Serve as Sites of Phosphorylation in ␣B-Crystallin-When U373 MG cells were exposed to PMA plus okadaic acid, generation of three acidic forms of ␣B-crystallin was induced in a time-dependent manner (Fig. 6,  A and B, p1, p2, and p3). To confirm that the modifications were due to phosphorylation, we labeled cells with [ 32 P]orthophosphate and then exposed them to PMA plus okadaic acid for 90 min. The labeled ␣B-crystallin in cell extracts was collected by immunoprecipitation with antibodies against ␣B-crystallin and then subjected to IEF with subsequent Western blotting (Fig. 7A) and autoradiography (Fig. 7B). As expected, all three forms of ␣B-crystallin induced by the treatment were identified as phosphorylated forms (Fig. 7, A and B).
The immunoprecipitates were subjected to analysis of phosphoamino acids. As shown in Fig. 7C, radioactivity was detected only in the spot that comigrated with phosphoserine. These results indicate that three different serine residues in ␣B-crystallin can be phosphorylated in U373 MG cells, as reported also for the phosphorylation of ␣B-crystallin in the lens (29,30).
Identification of the Sites of Phosphorylation in ␣B-Crystallin by Mass Spectrometry-A peptide map of the tryptic digest of ␣B-crystallin obtained by capillary HPLC-ESI/TSQMS is shown in Fig. 8A. Most of the fragments predicted from the amino acid sequence of ␣B-crystallin were identified by comparison of the observed molecular mass with the theoretical molecular mass of each fragment. In addition to these fragments, three additional fragments were detected with molecular masses ϳ80 Da larger than the theoretical masses of T2 (residues 12-22; average mass, 1374), T3 (residues 23-56; average mass, 4006), and T4 (residues 57-69; average mass, 1497), respectively (designated T2*, T3*, and T4*). These results suggest that each of these fragments includes a single phosphorylated residue. The mass spectra deconvoluted from the multiply charged ions are shown in Fig. 8 (B, T2*; D, T3*; and F, T4*). These fragments were analyzed by MS/MS for identification of the sites of phosphorylation. Fig. 8C shows part of the MS/MS spectrum of the triply charged ion of T2* (m/z ϭ 485). In the spectrum, the b and y series ions were hardly detected, probably because this peptide contained three proline residues and a single histidine residue, all of which are susceptible to collision-activated cleavage. However, internal sequence ions corresponding to His-Ser*, Pro-Phe-His, and Pro-Phe-His-Ser* (HS*, PFH, and PFHS* in Fig. 8C) suggested that the site of phosphorylation in T2 was Ser-19. In this spectrum, some of the oligopeptide ions derived from unphosphorylated T2 were also detected. Fig. 8E shows a spectrum for the quadruply charged ion of T3* (m/z ϭ 1021). The presence of a series of doubly charged ions, namely, y 11 , y* 12 (y 12 ϩ 80 Da), y* 14 , y* 16 , y* 18 , and y* 19 , indicated that the site of phosphorylation in T3* was Ser-45. In the case of T4*, the MS/MS analysis of the doubly charged ion (m/z ϭ 789)  Polyvinylidene difluoride membranes were stained as described in the legend to Fig. 1. p0, unphosphorylated ␣B-crystallin; p1 and p2, phosphorylated ␣B-crystallin.
clearly revealed a series of b ions, b 2 , b* 3 (b 3 ϩ 80 Da), and b* 4 (Fig. 8G), indicating that Ser-59 was the site of phosphorylation. Thus, the structural analysis showed that the phosphorylation of ␣B-crystallin occurred at three sites: Ser-19, Ser-45, and Ser-59.
Western Blot Analysis with Antibodies against Phosphorylated Forms of ␣B-Crystallin-To clarify the sites of phosphorylation induced by each stimulus, we prepared antibodies that recognized each of the phosphorylated sites in ␣B-crystallin. The antibodies against peptides that contained phosphorylated Ser-19 or Ser-45 reacted with ␣B 1 -crystallin purified from bovine lens and from rat lens (data not shown) because the amino acid sequences of peptides p19S and p45S, used as antigens, were identical to those of bovine and rat ␣B-crystallins. How-ever, the sequence of peptide p59S included a residue different from that in bovine and rat ␣B-crystallins. Residue 61 in human ␣B-crystallin is phenylalanine, whereas that in bovine and rat ␣B-crystallins is isoleucine (44,45). Therefore, antibodies against p59S did not react with bovine and rat ␣B 1 -crystallins (data not shown), but they did react with the phosphorylated form of ␣B-crystallin in U373 MG cells. As shown in Fig.  9, most of the phosphorylated forms of ␣B-crystallin induced by various stimuli, which were detectable with antibodies against the carboxyl-terminal peptide (Fig. 9, ␣B cry), also reacted with antibodies specific to each of the three phosphopeptides. The antibodies against p19S often detected the phosphorylated ␣Bcrystallin as doublet bands for unknown reasons. Although PMA and arsenite preferentially stimulated the phosphorylation of Ser-45 and Ser-59, respectively, all three sites were   7. Phosphorylation and analysis of phosphoamino acids in ␣B-crystallin using 32 P-labeled cells. U373 MG cells were incubated at 37°C for 1 h with 0.5 mCi/dish of carrier-free [ 32 P]orthophosphate in phosphate-free Eagle's minimal essential medium without serum. Then they were exposed to 1 M PMA plus 0.2 M okadaic acid at 37°C for 90 min. ␣B-Crystallin in extracts of cells was immunoprecipitated and subjected to IEF as described under "Experimental Procedures." A, shown are the results from Western blot analysis of samples from cells exposed to PMA plus okadaic acid (lane 3) or from unexposed cells (lane 2) using antibodies against the carboxyl-terminal peptide of ␣B-crystallin. Lane 1, rat ␣-crystallin. B, shown is an autoradiogram of a gel after IEF. Lane 2, unexposed control cells; lane 3, cells exposed to PMA plus okadaic acid. C, the immunoprecipitate of extracts of cells that had been exposed to PMA plus okadaic acid was subjected to SDS-polyacrylamide gel electrophoresis, and proteins on the gel were transferred to a nitrocellulose membrane. The membrane was subjected to autoradiography, and the band of 32 P-labeled ␣Bcrystallin was excised from the membrane for analysis of phosphoamino acids by thin-layer chromatography. The locations of unlabeled phosphoserine (PS), phosphothreonine (PT), and phosphotyrosine (PY) are marked by circles. p0, unphosphorylated ␣B-crystallin; p1, p2, and p3, phosphorylated forms of ␣B-crystallin.
found to be phosphorylated in response to each stimulus.
Stimulation of the Phosphorylation of ␣B-Crystallin in Tissues of Heat-stressed Rats-We examined whether or not the phosphorylation of ␣B-crystallin was stimulated in tissues of heat-stressed rats. Phosphorylated ␣B-crystallin was detected in soluble and insoluble fractions of hearts and diaphragms from untreated control rats (Fig. 10, A and B). After heat stress (42°C for 20 min), the level of phosphorylated ␣B-crystallin was markedly elevated in both the soluble and insoluble fractions of hearts and diaphragms (Fig. 10, A and B). When diaphragms excised from untreated control rats were subjected to heat stress at 45°C as described previously (38), enhanced phosphorylation of ␣B-crystallin was also detected in both the soluble and insoluble fractions of tissue extracts (Fig. 10C). DISCUSSION This is the first systematic study, to our knowledge, that demonstrates the stimulation of the phosphorylation of ␣Bcrystallin in cultured cells during their exposure to various stimuli and in rat tissues after heat stress. The phosphorylation of ␣B-crystallin in response to the stimuli was also observed in 3Y-1 rat fibroblasts and heat-stressed C6 rat glioma cells, which contain high levels of ␣B-crystallin (data not shown). Two forms of ␣B-crystallin are present in the lens: unphosphorylated ␣B 2 -crystallin and phosphorylated ␣B 1 -crystallin (1,31). Wang et al. (34) reported recently that the phosphorylation of ␣B-crystallin in cultured rat lens is stimulated upon exposure to H 2 O 2 . Moreover, Kantorow and Piatigorsky (33) reported that ␣B-crystallin and ␣A-crystallin have autokinase activity. It has been reported that ␣B-crystallin in extracts of bovine lens can be phosphorylated in vitro in a cAMPdependent manner (31,32). However, in our experiments, exposure of cells to activators of cAMP-dependent protein kinase, which included forskolin, cholera toxin, and dibutyryl cAMP, barely stimulated the phosphorylation of ␣B-crystallin in U373 MG cells (Fig. 2C). This discrepancy might be due to a difference between conditions in vitro and in vivo. We estimated the concentrations of cyclic AMP in U373 MG cells that had been exposed for 45 min to various stimuli. The concentration of cyclic AMP in untreated control cells was 7.8 pmol/mg of protein (a mean of results from duplicate dishes), and it was markedly enhanced by 20 M forskolin (72.0 pmol/mg of protein), but barely by 200 M arsenite (7.4 pmol/mg of protein), 1 M PMA (7.0 pmol/mg of protein), 10 g/ml anisomycin (8.3 pmol/mg of protein), 0.4 M sorbitol (10.1 pmol/mg of protein), and 4 mM H 2 O 2 (12.2 pmol/mg of protein), These results also suggest that the phosphorylation of ␣B-crystallin in U373 MG cells is induced in a cAMP-independent manner.
Klemenz et al. (12) reported that, after heat stress, only a single spot of ␣B-crystallin was visible after two-dimensional gel electrophoresis of an extract of NIH 3T3 cells. We also failed to detect the phosphorylated form of ␣B-crystallin in soluble extracts of heat-stressed U373 MG cells (Fig. 3B). However, phosphorylated ␣B-crystallin was detected in the insoluble fraction of cells (Fig. 3B). It has been reported that ␣B-crystallin relocalizes rapidly to the insoluble nuclear/cytoskeletal fraction after heat stress (13). Thus, phosphorylated ␣B-crystallin might be preferentially redistributed to the insoluble fraction, as suggested also for hsp27 (27).
The arsenite-induced phosphorylation of ␣B-crystallin was inhibited by dithiothreitol, and the PMA-induced phosphorylation was inhibited by staurosporine (Fig. 4A). Moreover, phosphorylation of ␣B-crystallin in response to anisomycin, arsen- ite, sorbitol, H 2 O 2 , NaCl, and heat stress was suppressed by SB202190, an inhibitor of p38 MAP kinase, and the PMAinduced phosphorylation was selectively inhibited by PD98059, a inhibitor of p44 MAP kinase kinase (Fig. 5). These results indicate that several kinase cascades might be associated with induction of the phosphorylation of ␣B-crystallin. The agents that induced phosphorylation of ␣B-crystallin in the present study were also effective stimulators of the phosphorylation of hsp27 (15)(16)(17)(18)(19)(20)(21)(22). It has been reported that hsp27 can be phosphorylated by MAP kinase-activated protein kinase-2, which is under the control of a novel protein kinase cascade (23)(24)(25)(26). However, it is unknown whether or not ␣B-crystallin can serve as a substrate for MAP kinase-activated protein kinase-2, and the kinase(s) responsible for the phosphorylation of ␣B-crystallin remains to be identified.
In a cell-free system, we could not detect the increased activity for phosphorylation of ␣B-crystallin even in extracts of cells that had been exposed to 1 M PMA plus 0.2 M okadaic acid (data not shown). Therefore, there might be other mechanisms that account for the phosphorylation of ␣B-crystallin under the experimental conditions described. For example, these conditions may alter the cellular distribution or the aggregation properties of ␣B-crystallin, which may increase the accessibility of ␣B-crystallin to protein kinases. Previously, we (27) reported that a polymeric form of hsp27 in cells was dissociated in response to various stimuli employed in the present study. When the aggregation properties of ␣B-crystallin were analyzed by centrifugation on sucrose density gradients, ␣B-crystallin in extracts of U373 MG cells sedimented as a polymeric form without any dissociated forms even in extracts of cells that had been exposed to 1 M PMA plus 0.2 M okadaic acid for 90 min. However, hsp27 in the same extracts of cells treated with PMA plus okadaic acid was completely dissociated to a dimeric form (data not shown). Dissociation of hsp27 from a heteropolymeric complex with ␣B-crystallin might increase the accessibility of ␣B-crystallin to protein kinases.
Voorter et al. (30) reported that ␣B-crystallin from bovine lens has two sites of phosphorylation: Ser-19 and Ser-45. Chiesa et al. (29) also identified two sites of phosphorylation: Ser-59 and Ser-43 and/or Ser-45. Smith et al. (46) reported that the sites of phosphorylation of ␣B-crystallin were Ser-45, Ser-59, and Ser-19 or Ser-21. In the present study, exposure of U373 MG cells to PMA plus okadaic acid resulted in phosphorylation of three serine residues (Fig. 6). Using mass spectrometry, we identified the sites of phosphorylation as Ser-19, Ser-45, and Ser-59 (Fig. 8).
The phosphorylated form of ␣B-crystallin was detected in tissues of untreated rats, including the heart and the diaphragm (Fig. 10). Thus, phosphorylation of ␣B-crystallin seems to occur even under normal physiological conditions in vivo, and phosphorylation is stimulated during exposure to stress (heat).
We raised antibodies that recognized each phosphorylated FIG. 9. Western blot analysis of phosphorylated ␣B-crystallin with antibodies that recognized each specific phosphoserine residue. Soluble extracts of cells that had been exposed for 90 min to 200 M arsenite, 1 M PMA, 4 mM H 2 O 2 , 0.4 M sorbitol, 0.15 M NaCl, or 10 g/ml anisomycin, as well as the urea-solubilized fraction of pellets of cells that had been heated for 30 min at 45°C (Heat shock), were subjected to IEF with subsequent Western blot analysis using antibodies (1 g of IgG/ml) against p19S, p45S, or p59S or using antibodies (0.1 g/ml) against the carboxyl-terminal peptide of ␣B-crystallin (␣B cry). Control, untreated cells. The location of the phosphorylated ␣B-crystallin in each blot is shown by an arrowhead.
FIG. 10. Phosphorylation of ␣B-crystallin in rat tissues. Rats were subjected to heat treatment at 42°C for 20 min, and hearts (A) and diaphragms (B) were dissected out. Soluble (upper panels) and insoluble (lower panels) fractions of each tissue were subjected to IEF with subsequent Western blot analysis with antibodies against the carboxylterminal peptide of ␣B-crystallin. Lanes 1-3, untreated control rats; Lanes 4 -6, heat-stressed rats. In C, the diaphragm from an untreated rat was divided into three pieces, and the fragments were incubated at 45°C for 0 min (lane 1), 10 min (lane 2), and 20 min (lane 3), respectively, in serum-free Eagle's minimal essential medium equilibrated with 5% CO 2 and 95% O 2 . Soluble (upper panel) and insoluble (lower panel) fractions of the tissues were analyzed as described above. Arrowheads indicate the locations of the phosphorylated form of ␣B-crystallin. residue in ␣B-crystallin, and we used them to determine the selectivity of each stress with respect to the site of phosphorylation. Although PMA and arsenite preferentially stimulated the phosphorylation of Ser-45 and Ser-59, respectively, all three serine residues were phosphorylated to some extent during exposure to each stress. These results, together with the effects of SB202190, suggest that p38 MAP kinase might be involved in the signal transduction cascade that leads to phosphorylation of ␣B-crystallin in response to stress.
The physiological significance of the phosphorylation of ␣Bcrystallin is unclear. Wang et al. (34) reported that phosphorylation of ␣B-crystallin in rat lens has no effect on its chaperone activity. Nicholl and Quinlan (47) reported that ␣crystallin inhibits the assembly in vitro of glial fibrillary acidic protein and vimentin in an ATP-independent manner and, moreover, that both phosphorylated and unphosphorylated forms of ␣B-crystallin are equally effective inhibitors. Wang and Spector (48) reported recently that ␣B-crystallin stabilizes actin filaments and prevents their cytochalasin-induced depolymerization, whereas phosphorylated ␣B-crystallin lacks such an activity in vitro. When the proportion of phosphorylated forms of ␣B-crystallin was estimated by the NIH Image program, it was ϳ10% in cells exposed to heat and NaCl and 15-20% in cells treated with each of the other stimuli (arsenite, PMA, anisomycin, H 2 O 2 , and sorbitol). Since only a small fraction of ␣B-crystallin was phosphorylated under our various conditions, phosphorylated ␣B-crystallin might act as a second messenger in stress responses in cells rather than as a molecular chaperone. Further investigations are required to clarify the role of the phosphorylation of ␣B-crystallin in cells.