αB-crystallin Gene Induction and Phosphorylation by MKK6-activated p38

The MAPK kinase MKK6 selectively stimulates p38 MAPK and confers protection against stress-induced apoptosis in cardiac myocytes. However, the events lying downstream of p38 that mediate this protection are unknown. The small heat shock protein, αB-crystallin, which is expressed in only a few cell types, including cardiac myocytes, may participate in MKK6-mediated cytoprotection. In the present study, we showed that, in cultured cardiac myocytes, expression of MKK6(Glu), an active form of MKK6, led to p38-dependent increases in αB-crystallin mRNA, protein, and transcription. MKK6(Glu) also induced p38-dependent activation of the downstream MAPK-activated protein kinase, MAPKAP-K2, and the phosphorylation of αB-crystallin on serine-59. Initially, exposure of cells to the hyperosmotic stressor, sorbitol, stimulated MKK6, p38, and MAPKAP-K2 and increased phosphorylation of αB-crystallin on serine 59. However, after longer times of exposure to sorbitol, the cells began to undergo apoptosis. This sorbitol-induced apoptosis was increased when p38 was inhibited in a manner that would block αB-crystallin induction and phosphorylation. Thus, under these conditions, the activation of MKK6, p38, and MAPKAP-K2 by sorbitol can provide a degree of protection against stress-induced apoptosis. Supporting this view was the finding that sorbitol-induced apoptosis was nearly completely blocked in cells expressing MKK6(Glu). Therefore, the cytoprotective effects of MKK6 in cardiac myocytes are due, in part, to phosphorylation of αB-crystallin on serine 59 and to the induction of αB-crystallin gene expression.

The response of cells to stress can determine the fate of tissues and, ultimately, the fate of the organism. The initial response to most stresses is the induction of intracellular reactions that foster the preservation of cell viability and, thus, organ function. However, if the stress persists, this protective response can be superseded by intracellular events that result in apoptosis or necrosis, leading eventually to deterioration of organ function.
The ability of the myocardium to mount a productive stress response is critical for survival. Examples of stresses to which the heart can be exposed in vivo include transient ischemia, cytokines and chemokines, reactive oxygen species, or toxins (1). In the absence of a fully functional stress response, these insults could rapidly lead to a devastating loss of cardiac myocytes due to apoptosis and to a decline in tissue function, emphasizing the importance of understanding the cellular and molecular events governing the myocardial stress response.
A vital feature of the mechanism by which cardiac myocytes respond to stress is the activation of the stress mitogen-activated protein kinases (MAPKs), 1 such as p38 MAPK (2)(3)(4)(5)(6). For example, p38 is activated in response to ischemia in isolated hearts (7,8). In cultured cardiac myocytes, p38 is activated by agonists such as endothelin and phenylephrine (9 -12), that are known to evoke hypertrophic growth (13), an event that mimics the compensatory phase of the myocardial stress response (14). Additionally, p38 is stimulated in transgenic mouse models of Ras-induced cardiac hypertrophy (15).
Like other MAPKs, p38 is activated after becoming dually phosphorylated on a critical -Thr-Gly-Tyr-motif (16). This phosphorylation is accomplished by an upstream kinase, or a MAPK kinase (MKK). Several MKKs are known to activate p38, as well as some of the other MAPKs. However, among the MAPKs, MKK3 and MKK6 are capable of activating only p38 (17,18). Initial studies of the role of MKK-activated p38 in the heart showed that the expression of a constitutively activated form of MKK6, MKK6(Glu), induced certain features of the stress response in primary neonatal cardiac myocytes, such as sarcomeric stabilization (12) and protection from sphingosineinduced apoptosis (19). Later studies corroborated these results, demonstrating roles for both MKK6 and MKK3 in the cardiac stress response (11,15).
The present study is focused on the mechanism by which MKK6 protects cardiac myocytes from apoptosis, with a particular emphasis on the small heat shock proteins (hsps). Among the best characterized of the small hsps are hsp27 and ␣Bcrystallin. In response to stress, hsp27 is phosphorylated by kinases that lie downstream of p38, notably MAPK-activated protein-K2 (MAPKAP-K2), MAPKAP-K3, and p38-reactive kinase (20,21). In endothelial and smooth muscle cells, the phosphorylation of hsp27, which is believed to contribute to protection against stress-induced cell death, correlates tempo-rally with the translocation of hsp27 to actin filaments, where it contributes to stabilization of the cytoskeleton (22)(23)(24)(25). Furthermore, overexpression of hsp27 has been shown to foster cell survival, although protection from apoptosis was not demonstrated (26).
In contrast to hsp27, which is expressed in nearly all cell types, a structurally related, small hsp, ␣B-crystallin, is expressed mainly in the lens, in certain neurons, in skeletal muscle, and, surprisingly, in very high levels in cardiac myocytes (27)(28)(29)(30). Like hsp27, ␣B-crystallin in lens cells is posttranscriptionally modified by phosphorylation, and in cultured human glioma cells (U373), ␣B-crystallin is phosphorylated in response to stressors such as sorbitol (hyperosmotic) or arsenite (31). Recent studies have shown that when U373 cells are treated with stressors known to activate p38 and MAPKAP-K2, ␣B-crystallin is phosphorylated in a site-specific manner on serines located at positions 19, 45, and 59 (32). The roles and the kinases responsible for each phosphorylation event remain unknown.
Stresses such as ischemia, sorbitol, and arsenite activate p38 and MAPKAP-K2 in the heart and in cultured cardiac myocytes (7,8,11,33). In comparison with stress-induced hsp27 translocation to actin filaments in other cell types, in the isolated perfused heart, ischemic stress induces ␣B-crystallin translocation from a diffuse, cytosolic localization to sarcomeres (34). This translocation was shown to be temporally coordinated with increased ␣B-crystallin phosphorylation, although the identities of phosphorylated residues have not been established (35). Since hsp27 translocates to and stabilizes actin filaments in endothelial and smooth muscle cells (4,23), the sarcomeric localization of ␣B-crystallin could foster myofilament stabilization in stressed cardiac myocytes. Since myofilament disintegration takes place in cardiac myocytes during hypoxia, which leads to apoptosis (36), the sarcomeric translocation of ␣B-crystallin could be important for protecting myocytes against apoptosis. Consistent with this possibility is the finding that overexpression of ␣B-crystallin increased cardiac myocyte survival during ischemia (26), although apoptosis was not assessed in that study.
The mechanisms responsible for the tissue-restricted expression of ␣B-crystallin are largely unknown. However, the ␣Bcrystallin promoter possesses a serum response element (SRE) (38) that is homologous to SREs in other stress-induced cardiac genes (e.g. ANF (39)). Thus, it is plausible that ␣B-crystallin expression is induced in concert with other SRE-containing cardiac genes in response to stress and that the increased levels of ␣B-crystallin contribute to myocardial protection. Moreover, it is possible that stress-mediated activation of the p38 pathway can induce the selective, site-specific phosphorylation of ␣B-crystallin and that this phosphorylation may be critical for optimal ␣B-crystallin-mediated cytoprotection. These hypotheses were tested in the present study using a well characterized cultured cell model system for evaluating the response of cardiac myocytes to stress and activated MKK6.

Cell Culture
Primary ventricular myocytes were prepared from 1-4-day-old Harlan Sprague-Dawley rats as described (12,39). Briefly, hearts were dissected in DMEM/air; the apical two-thirds of the ventricles were dissected away from the atria. After mincing and washing the ventricles twice with air-compatible DMEM, cells were isolated by multiple rounds of 10-min-long tissue dissociation with 0.001% trypsin. After each incubation with trypsin, the supernatant was added to an equal volume of DMEM/F-12 (1:1) containing 20% fetal bovine serum, and all of the supernatants were combined. Plastic wells were treated for at least 1 h with 5 ng of fibronectin/ml of DMEM/F-12 (1:1). Myocytes, which, in some experiments, were derived from the nonadherent cells in a 1-h preplating on untreated plastic dishes, were plated finally in DMEM/F-12 (1:1) containing 10% fetal bovine serum for approximately 16 h. After this time, cultures were washed with DMEM/F-12 (1:1) and then incubated in DMEM/F-12 (1:1) containing bovine serum albumin at 1 g/liter (minimal medium) with or without any test agents for the indicated times.

Northern Analysis
Northern analyses were performed as described previously (39), with minor modifications. RNA was isolated from 3 ϫ 10 6 myocytes using RNAzol B (Tel-Test, Inc., Friendswood, TX) as described by the manufacturer. Five g of each sample were fractionated on a 1% agarose/ formaldehyde gel and then transferred to Hybond Nylon membrane (Amersham Pharmacia Biotech). Membranes were cross-linked and then incubated with radioactive probes in QuikHyb (Stratagene, Inc., Madison, WI). The ␣B-crystallin probe was prepared by digesting a rat ␣B-crystallin cDNA (cLens/pBluescript(Ϫ) from Dr. J. Piatigorsky, National Institutes of Health) with EcoRI/HindIII to generate a 730-bp fragment of the ␣B-crystallin cDNA. The glyceraldehyde-3-phosphate dehydrogenase probe was prepared by digesting a glyceraldehyde-3phosphate dehydrogenase cDNA (ATCC 78105) with EcoRI to generate a 1.2-kilobase pair fragment. cDNA fragments were then labeled with [␣-32 P]dCTP, using Klenow fragment and standard protocols.

Western Analyses
␣B-crystallin-Cultures composed of approximately 10 6 myocytes were lysed in 100 l of Laemmli sample buffer supplemented with 0.1 mM sodium o-vanadate, 10 g/ml aprotinin, 2 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 1 mM p-nitrophenyl phosphate, and 10 g/ml leupeptin and boiled for 5 min. Eighty l of each lysate were analyzed for phosphospecific ␣B-crystallin, and 5 l of each lysate were analyzed for total ␣B-crystallin. Proteins were resolved on a 15% SDS-PAGE and transferred to a nitrocellulose membrane in methanol transfer buffer at 60 V for 5 h or 30 V overnight. Membranes were blocked for 30 min in 5% nonfat milk dissolved in TBS-Tween (0.01%) at room temperature and then probed with rabbit affinity-purified ␣B-crystallin antisera specific for phosphorylation at serine 19, 45, or 59. All ␣B-crystallin antisera were generously provided by Dr. Kanefusa Kato (Institute for Developmental Research, Japan). The ␣B-crystallin phosphoserine-specific antisera were used at 1 g/ml for phosphoserine 19 and at 0.5 g/ml for phosphoserines 45 and 59; the antisera were diluted into 5% nonfat milk TBS-Tween. Blots were incubated with this mixture for 2 h at room temperature. Some blots were incubated with an affinitypurified ␣B-crystallin antiserum raised against the C terminus of ␣Bcrystallin, which was used at a final concentration of 0.1 g/ml. This antiserum recognizes all phosphorylated and nonphosphorylated forms of ␣B-crystallin, thus serving as a measure of the total amount of ␣B-crystallin in a sample. Following incubation with the ␣B-crystallin antisera, blots were washed for 30 min in TBS-Tween and then incubated in a 1:2000 dilution of anti-rabbit IgG horseradish peroxidase (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) followed by washing for 30 min in TBS-Tween. Visualization of immune complexes were carried out by an enhanced chemiluminescence method using enhanced chemiluminescence Western blotting detection reagents (NEN Life Science Products) according to the manufacturer's instructions.
p38 and Phospho-p38 -Cultures composed of approximately 2 ϫ 10 6 myocytes were lysed in 100 l of supplemented Laemmli sample buffer (see above), boiled for 5 min, and then submitted to 10% SDS-PAGE and transferred to a nitrocellulose membrane (see above). Western analyses were performed as described above with 1:1000 dilutions of antisera specific for either phospho-p38 (New England BioLabs, Beverly, MA; catalog no. 9211S). Blots were subsequently stripped with 6.25 mM Tris, 2% SDS, and 100 mM 2-mercaptoethanol for 30 min at 50°C, washed for 1 h in TBS-Tween, and reprobed with 1:1000 dilution of p38 antibody (Stressgen Biotechnologies Corp., Victoria, Canada; catalog no. KAP-MAOO9E) for normalization.

Transfection by Electroporation
Immediately following the final dissociation step, myocardial cells were resuspended in serum-free DMEM:F-12 (1:1). Between 5 and 12 ϫ 10 6 cells were combined with the indicated amounts of plasmids in a total volume of 300 l of DMEM/F-12 (1:1). The total amount of plasmid used for each electroporation in one experiment was equalized using pCMV6. Cells were electroporated at 500 V, 25 microfarads, and 100 ohms in a 0.2-cm gap electroporation cuvette (Bio-Rad) using a Gene Pulser II (Bio-Rad). Under these conditions, only cardiac myocytes are transfected (39,40). For reporter assays, 1.5 ϫ 10 6 cells were plated per 24-mm well, whereas 4.5 ϫ 10 6 myocytes per 35-mm well were plated to perform kinase assays.

MKK6 Kinase Assays
Myocytes were transfected by electroporation with 30 g of pcDNA3 FLAG-MKK6(wt), as described above, using 9 ϫ 10 6 cells per transfection. Transfected myocytes were plated, and after 48 h in minimal media, cultures were treated with 400 mM sorbitol or 0.5 mM arsenite for 5 min and then extracted in 1 ml of lysis buffer containing 20 mM Tris, pH 7.6, 20 mM ␤-glycerophosphate, 250 mM NaCl, 3 mM EGTA, 3 mM EDTA, 0.5% Nonidet P-40, 0.1 mM sodium o-vanadate, 10 g/ml aprotinin, 2 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 1 mM pnitrophenyl phosphate, and 10 g/ml leupeptin. The cell debris was removed by brief centrifugation. Lysates were incubated overnight at 4°C with 4 g of anti-FLAG M2 antibody (Sigma) per sample. Forty l of a protein G-Sepharose slurry were then added to each sample and incubated for 60 min at 4°C. The Sepharose pellet was washed twice with lysis buffer and once with preincubation buffer (30 mM HEPES, pH 7.4, 10 mM MgCl 2 , 20 mM ␤-glycerophosphate, 1 mM DTT). Kinase reactions were performed using 2 g of GST-3XHA p38␤ 2 (see below) per reaction as substrate in a final volume of 30 l of kinase buffer (30 mM HEPES, pH 7.4, 10 mM MgCl 2 , 20 mM ␤-glycerophosphate, 1 mM DTT, 20 M ATP, and 6 M [␥-32 P]ATP (5000 Ci/mmol). After 20 min at 30°C, the reactions were terminated by the addition of Laemmli sample buffer, and the phosphorylation level of substrate proteins was evaluated by SDS-PAGE followed by autoradiography and phosphor image analyses.

Recombinant Protein Expression and Purification
A kinase-inactive form of human p38␤ 2 (p38␤ 2 (K53R)), which retains the native sites that are phosphorylated by MKK6 (Thr 181 and Tyr 183 ), was used as the substrate for MKK6 kinase assays. To prepare this protein, the plasmid, pGEX-6P-1/3XHA p38␤ 2 (K53R), was transformed into BL-21-competent cells (Stratagene, Inc., Madison, WI; catalog no. 200133). Cultures were grown in 2ϫ YTA medium with 100 g/ml ampicillin for 12-15 h until A 600 ϭ 2.0. Protein expression was induced with 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside (Sigma; catalog no. I-5502), at 30°C and with shaking at 150 rpm for 6 h. Cells were the collected by centrifugation at 6000 ϫ g for 10 min and then resuspended in 50 ml of buffer G (1ϫ phosphate-buffered saline, 50 mM EDTA, 5 mM benzamidine, 100 M phenylmethylsulfonyl fluoride) per liter of culture. The resuspended cells were sonicated using a Branson 450 sonifier, at setting 5 for 15-30 s. Triton X-100 was added to 0.1% after sonication, and the sonicate was then centrifuged at 12,000 ϫ g for 20 min. Recombinant GST-3XHA p38␤ 2 (K53R) was then purified from the crude cell extract using a Bulk GST Purification Module (Amersham Pharmacia Biotech.; catalog no. 27-4570-01), essentially as described in the accompanying protocol. Briefly, 1 ml of 50% glutathione-Sepharose was added per 100 ml of sonicate and incubated on ice for 1-2 h. The Sepharose was pelleted at 1000 ϫ g for 5 min, washed three times with cold 1ϫ phosphate-buffered saline, and eluted with 10 mM reduced glutathione (in 50 mM Tris-HCl, pH 8) in fractions of 1 ml/500 l of Sepharose. The eluted protein was dialyzed overnight at 4°C in 2 liters of dialysis buffer (100 mM Tris, pH 7; 200 mM NaCl; 200 M EDTA; 2 mM DTT). The purification of p38␤ 2 (K53R) was verified by SDS-PAGE followed by staining with Coomasie Blue, coupled with Western analyses using an anti-HA antiserum. Final protein concentration was determined using the DC Protein Assay (Bio-Rad; catalog no. 500-0116).

Preparation of Recombinant Adenovirus
AdV-p38AGF and AdV-MKK6(Glu)-The AdEasy system was used for preparing recombinant adenoviral strains using previously described methods (41,42). Briefly, human MKK6(Glu) and human p38␣(AGF) were polymerase chain reaction-amplified from the parent templates (see above) such as to create restriction sites that would facilitate cloning into pAdTrack-CMV, an adenoviral shuttle vector that harbors CMV-driven green fluorescent protein, and a CMV-flanked multiple cloning site for the insertion of the gene of interest. Polymerase chain reaction-amplified p38␣(AGF) and MKK6(Glu) were cloned into the EcoRI and NotI sites of pGEX-6P-1 (Amersham Pharmacia Biotech), which served as a shuttle cloning vector. pGEX-6P-1/ p38␣(AGF) and pGEX-6P-1/MKK6(Glu) were then digested with BamHI and NotI, and the resulting products of interest were cloned into the BglII and NotI sites of pAdTrack-CMV to create pAdTrack-CMV-p38␣(AGF) and pAdTrack-CMV-MKK6(Glu). pAdTrack-CMV-p38␣(AGF) or pAdTrack-CMV-MKK6(Glu) was linearized and then cotransformed with the adenoviral vector, pAdEasy-1, into Escherichia coli strain BJ5183. This strain of E. coli allows for homologous recombination of pAdEasy-1 and the pAdTrack-CMV shuttle vector containing the gene of interest. Recombinants were selected on kanamycin and screened by restriction digestion with PacI. Recombinant plasmids were then retransformed into E. coli DH5␣ for propagation purposes. Recombinant adenoviral plasmids were linearized with PacI and then transfected into 293 human embryonic kidney cells, using Lipo-fectAMINE® (Life Technologies, Inc.). Transfection efficiency was determined by observing green fluorescent protein fluorescence, as described previously (41). The recombinant viruses were then harvested 7-10 days postinfection. Viral titers were determined by observing green fluorescent protein fluorescence of primary neonatal cardiac myocytes; the minimum quantity of viral stock that afforded 100% transfection efficiency was selected for the experiments in this study.

Reporter Assays
␤-Galactosidase-Following the appropriate time in culture, each well of myocytes was washed twice with phosphate-buffered saline and then lysed in 500 l of ice-cold lysis buffer (25 mM Gly-Gly, pH 7.8, 15 mM MgSO 4 , 4 mM EDTA, 0.25% Triton X-100) containing 1 mM DTT. The cell debris was removed by centrifugation. To measure ␤-galactosidase activity, 200 l of the cell extract was added to 400 l of ␤-galactosidase buffer (60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgSO 4 ) containing 1 mg/ml chlorophenol red-␤-D-galactopyranoside and 50 mM ␤-mercaptoethanol. The reaction was incubated for 2 h at 37°C, after which absorbance was measured at 570 nm.
Luciferase-To measure luciferase activity, 100 l of buffer (25 mM Gly-Gly, pH 7.8, 15 mM MgSO 4 , 4 mM EGTA, 45 mM KPO 4 pH 7.8, 1 mM DTT, 0.3 mM D-luciferin, 3 mM ATP) were added to 100 l of cell lysate. Light emission of each sample was measured by an Optocomp II luminometer for 10 s. The relative luciferase activities were determined by dividing the relative luciferase activity by the relative ␤-galactosidase activity.
Apoptosis DNA Laddering-Assessment of apoptosis was performed by DNA laddering as described previously (19). Briefly, each sample began with a 100-mm culture containing approximately 8 ϫ 10 6 cardiac myocytes. The cells from each culture dish were lysed in a digestion buffer containing 100 mM NaCl, 10 mM Tris (pH 8.0), 25 mM EDTA, 0.5% SDS, 0.4 mg/ml proteinase K (added fresh); scraped; pipetted several times; and incubated at 55°C for 5-14 h. Samples were then extracted twice with phenol and once with chloroform/isoamyl alcohol, and then DNA was isopropyl alcohol-precipitated and washed once with 80% ethanol and allowed to air-dry. The DNA pellet was then dissolved in a buffer containing 10 mM Tris (pH 7.9), 10 mM MgCl 2 , 50 mM NaCl, 1 mM DTT and then digested for 2 h at 37°C with RNase A (3 mg/ml). These samples were then extracted once with phenol and once with chloroform/isoamyl alcohol and then ethanol-precipitated. The DNA pellets were then washed with 80% ethanol, air-dried and dissolved in a small volume of Tris/EDTA. The absorbance of the sample was determined and identical amounts of each sample were then fractionated on a 2% agarose gel.

MKK6(Glu) Induces ␣B-Crystallin Expression in Cardiac
Myocytes-To determine whether the ␣B-crystallin gene is induced upon stimulation of p38 by MKK6, an activated form of MKK6, MKK6(Glu), was expressed in primary neonatal rat ventricular myocytes (NRVMCs) using adenovirus (AdV)-mediated gene transfer, followed by Northern and Western analyses of ␣B-crystallin mRNA and protein levels. Previous studies have demonstrated that using AdV-mediated gene transfer, a transfection efficiency of 100% can be achieved in NRVMCs (41). Compared with control cells, cultures expressing MKK6(Glu) displayed a 6.3-fold increase in the quantity of ␣B-crystallin mRNA, which was reduced to 2.8-fold upon treatment with the membrane-permeable, p38-specific inhibitor SB203580 (Fig. 1A). Western analysis showed that cells expressing MKK6(Glu) displayed a 2.2-fold increase over control in the quantity of ␣B-crystallin protein, which was reduced to control levels upon treatment of the cells with SB203580 (Fig.  1A).
To evaluate whether increased transcription contributed to MKK6-inducible ␣B-crystallin expression, a reporter construct comprised of 426 nucleotides of the mouse ␣B-crystallin 5Јflanking sequence and promoter driving firefly luciferase (␣BC(426)/Luc) was prepared. When NRVMCs were co-transfected by electroporation with ␣BC(426)/Luc, along with a construct encoding MKK6(Glu), there was an increase in the amount of luciferase activity, which correlated with the quantity of MKK6(Glu) (Fig. 1B). At the highest dose of MKK6(Glu) tested, there was an approximate 5.5-fold increase in luciferase expression. MKK6(Glu)-inducible ␣BC(426)/Luc was significantly inhibited by SB203580 or by co-transfecting a dominantinterfering form of p38 (Fig. 1C). Taken together, these results demonstrated that in cardiac myocytes the expression of MKK6(Glu), which among the MAPK kinases has been shown to activate only p38 (12), increased the levels of ␣B-crystallin transcription, mRNA, and protein in a p38-dependent manner. This is consistent with roles for ␣B-crystallin gene induction in MKK6/p38-mediated protection from apoptosis.
Site-selective Phosphorylation of ␣B-crystallin in Response to MKK6 -Antisera that bind selectively to ␣B-crystallin that is phosphorylated on either serine 19, 45, or 59 (32) were used to determine the extent of phosphorylation at these sites in cells expressing MKK6(Glu). In control NRVMCs, ␣B-crystallin was poorly phosphorylated on serine 19; however, phosphorylation on serines 45 and 59 was clearly detectable (Fig. 2A, top). The extent of phosphorylation of all three of these ␣B-crystallin serine residues increased in cells expressing MKK6(Glu). However, relative to serine 45 and serine 59, phosphorylation at serine 19 was nearly undetectable, even in MKK6(Glu)-expressing cells. Accordingly, phosphorylation of ␣B-crystallin at serine 19 was not pursued in further detail in this study. Expression of MKK6(Glu) increased phosphorylation at serines 45 and 59 by about 2-and 6.5-fold, respectively (Fig. 2A). The phosphorylation of ␣B-crystallin at both serine 45 and 59 observed in MKK6(Glu)-expressing cells was reduced to control levels or below control levels upon incubation with SB203580, indicating a requirement for p38. The activation of both p38 and MAPKAP-K2 in MKK6(Glu)-expressing cells (Fig. 2B) was consistent with the hypothesis that p38-activated kinases, such as MAPKAP-K2, are responsible for ␣B-crystallin phosphorylation in response to MKK6. SB203580, which interacts with the ATP binding site of p38 and, thus, inhibits its kinase activity, did not block the ability of MKK6(Glu) to phosphorylate p38, as expected. However, the kinase activity of endogenous MAPKAP-K2, which was stimulated by about 9-fold in MKK6(Glu)-expressing cells, was inhibited completely by SB203580 (Fig. 2B). Thus, the coordinate reduction of MKK6(Glu)-activated MAPKAP-K2 and ␣B-crystallin phosphorylation at serines 45 and 59 by SB203580 is consistent with the premise that in cardiac myocytes, ␣B-crystallin serves as a target for p38-activated MAPKAP-K2 and that the phosphorylation of ␣B-crystallin may contribute to the antiapoptotic effects of MKK6.
Stressors Activate MKK6, p38, MAPKAP-K2, and ␣B-crystallin Phosphorylation in Cardiac Myocytes-In most cell types, the MAPK pathways are activated in response to various stresses and are thus thought to serve important roles in the cellular stress response (43). Accordingly, we determined whether treating myocardial cells with two well characterized stressors, sorbitol and arsenite, activated the MKK6/p38/MAP-KAP-K2 pathway in a manner that could lead to ␣B-crystallin phosphorylation. Either sorbitol or arsenite coordinately induced the activities of MKK6, p38, and MAPKAP-K2, from approximately 8-to 15-fold (Fig. 3, A-C). Importantly, the activation of MAPKAP-K2 was blocked by SB203580 (Fig. 3C), indicating that under these conditions, p38 serves as the primary activator of MAPKAP-K2 in response to sorbitol or arsenite.
Sorbitol and arsenite also induced the site-specific phosphorylation of ␣B-crystallin on serines 45 and 59 by 2.5-and 10-fold, respectively (Fig. 4). Interestingly, ␣B-crystallin phosphorylation on serine 45 in response to sorbitol was very sensitive to the cell-permeable inhibitor of the ERK MAPK pathway, PD980590, but not to SB203580 (Fig. 5A). In contrast, ␣B-crystallin phosphorylation at serine 59 was not inhibited by PD098059 but was completely blocked by SB203580 (Fig. 5B). Thus, the enhancement of the phosphorylation of ␣B-crystallin on serine 45 was relatively small and appeared to require ERK, while sorbitol-mediated phosphorylation of ␣B-crystallin on serine 59 was relatively high and appeared to require p38. This site-selective phosphorylation was consistent with the relatively low ERK activation and high p38 activation observed in response to sorbitol or arsenite (Fig. 5C).
The Role of p38 in Sorbitol-induced Apoptosis-Long term exposure to sorbitol or arsenite leads to apoptosis in most cell types (43)(44)(45). Thus, it is possible that the initial activation of the p38 pathway, and, ultimately, ␣B-crystallin phosphorylation, which occur within minutes of exposure to these stressors, represent an attempt by cells to mount a protective response. To test this possibility, we evaluated the effects of inhibiting the p38 pathway on sorbitol-induced apoptosis in cardiac myocytes. As measured by DNA laddering, the initiation of apoptosis in cultured cardiac myocytes required approximately 5 h of exposure to 400 mM sorbitol (Fig. 6A). Compared with control (0 h of sorbitol), 5 or 8 h of sorbitol increased band intensity in the DNA ladders by about 2-and 6-fold, respectively (Fig. 6B). Consistent with the hypothesis that MKK6-activated p38 can afford protection from apoptosis was the finding that SB203580 increased DNA band intensity in sorbitol-treated cultures by as much as 2-fold (Fig. 6, A and B).
In a second experiment, the effect of PD098059 on apoptosis Cultured cardiac myocytes were infected with either AdV-Con or AdV-MKK6(Glu) and treated with or without 5 M SB203580, as indicated and as described under "Materials and Methods." Culture extracts were then analyzed for ␣B-crystallin phospho-Ser 19 , ␣B-crystallin phospho-Ser 45 , ␣B-crystallin phospho-Ser 59 , or total ␣B-crystallin by Western blotting with the appropriate antisera, as described under "Materials and Methods." Each treatment was carried out on two identical cultures. The phosphor image results are shown at the top, and the quantitated results of each blot, as determined using ImageQuant software (Molecular Dynamics), are shown at the bottom as the mean densities of phospho-Ser 19 , phospho-Ser 45 , or phospho-Ser 59 /␣B-crystallin total and are compared with the control (AdV-Con without SB203580). Data are presented as the mean of two identically treated cultures, Ϯ S.D. This experiment was replicated at least three times; the results shown are from one representative experiment. *, p Ͻ 0.05 different from control (no MKK6(Glu)) as determined by ANOVA followed by Neuman Keuls post hoc analysis of variance. B, p38 and MAPKAP-K2 activities. Cultured cardiac myocytes were infected with either AdV-Con or AdV-MKK6(Glu) and treated with or without 5 M SB203580, as indicated and as described under "Materials and Methods." Culture extracts were then analyzed for phospho-p38 and total p38 by Western blotting and for MAPKAP-K2 kinase activity as described under "Materials and Methods." Each treatment was carried out on three identical cultures. The phosphor image results are shown at the top, and the quantitated results of each blot, as determined using ImageQuant software (Molecular Dynamics), are shown as the mean density Ϯ S.E. compared with the control (AdV-Con without SB203580). The phospho-p38 band intensities were normalized to total p38. The data are presented as the mean of three identically treated cultures. This experiment was replicated at least three times; the results shown are from one representative experiment. *, p Ͻ 0.05 different from no MKK6(Glu), as determined by ANOVA followed by Neuman Keuls post hoc analysis of variance.
induced by 5 h of sorbitol treatment was compared with SB203580. As expected, sorbitol-mediated apoptosis was increased by inhibiting p38 with SB203580; however, PD098059 had no effect (Fig. 6, C and D). Thus, treating cells in a manner known to block ERK-mediated increases in phosphorylation of ␣B-crystallin on serine 45 had no effect on apoptosis, while treating cells in a manner known to block only p38/MAPKAP-K2-mediated phosphorylation of ␣B-crystallin on serine 59 enhanced apoptosis.
To further demonstrate a protective role for p38 under these conditions, sorbitol-induced apoptosis was assessed in cells expressing either MKK6(Glu) or dominant negative p38. As expected, sorbitol strongly induced DNA laddering in control cells, amounting to nearly 8-fold over control; however, there was no visible laddering in MKK6(Glu)-expressing cells, even after 8 h of sorbitol treatment (Fig. 6, E and F). In fact, cells expressing MKK6(Glu) were so refractory to apoptosis that after 24 h of sorbitol treatment they displayed about 5-6-fold less DNA laddering, as determined by band intensity, than control cells (Fig. 6, G and H). As expected, overexpression of a dominant-negative p38, p38AGF, resulted in no protection from apoptosis (Fig. 6, E and F). DISCUSSION The reduction of heart function due to cardiac myocyte apoptosis contributes to the progression of heart failure (47,48), emphasizing the importance of identifying signaling pathways that regulate myocardial cell apoptosis. In this regard, recent attention has focused on p38 MAPK. In cardiac myocytes, it has been shown that the activation of p38␣ by MKK3 leads to apoptosis, while activation of p38␤ by MKK6 leads to protection from apoptosis (15,19). The mechanism of MKK6/p38␤mediated cytoprotection remains largely unknown. However, one hypothesis is that by activating p38␤, MKK6 induces the expression of cytoprotective proteins. In support of this possibility are recent findings from our laboratory showing that MKK6-activated p38␤ can stimulate NFB, leading to the induction of the gene for the cytokine interleukin-6, which is secreted from and has cytoprotective effects on myocardial cells (41). This finding suggests that some other cytokines and/or trophic factors, such as tumor necrosis factor ␣, cardiotrophin-1, and leukocyte inhibitory factor, which are likely to be induced upon NF-B activation, also participate in cytoprotec- FIG. 4. Effects of sorbitol and arsenite on ␣B-crystallin phosphorylation. Cultured cardiac myocytes were treated for 10 min with or without 400 mM sorbitol or 0.5 mM arsenite, as indicated and as in Fig. 3. Culture extracts were then analyzed for ␣B-crystallin phospho-Ser 45 , ␣B-crystallin phospho-Ser 59 , or total ␣B-crystallin by Western blotting, as described under "Materials and Methods." Phospho-Ser 45 and phospho-Ser 59 analyses were carried out on separate cultures; the ␣B-crystallin total was determined for each. The phosphor image of the region of the blot where ␣B-crystallin migrates is shown at the top. The mean intensity of each phospho-␣B-crystallin band/total ␣B-crystallin Ϯ S.D. for each treatment is shown in the bar graph at the bottom. Data are presented as the mean of three identically treated cultures. *, p Ͻ 0.05 different from control as determined by ANOVA followed by Neuman Keuls post hoc analysis of variance. FIG. 3. Effects of sorbitol and arsenite on MKK6, p38, and MAPKAP-K2. A, MKK6. Cultured cardiac myocytes were transfected by electroporation with pcDNA3 FLAG-MKK6(wt). Forty-eight h later, they were treated for 5 min with or without 400 mM sorbitol or 0.5 mM arsenite. Following immunoprecipitation of FLAG-MKK6 from culture extracts, MKK6 kinase assays were carried out using p38␤ 2 (K53R) as the substrate and SDS-PAGE separation of products, as described under "Materials and Methods." The phosphor image of the region of the gel containing p38␤ 2 is shown at the top, and the digital analyses of those images are shown in the bar graph at the bottom as the mean density Ϯ S.E. The data are presented as the mean of three identically treated cultures. B, p38. Cardiac myocytes were cultured identically to those used in A, except they were not transfected. After 48 h, they were treated for 5 min with or without 400 mM sorbitol or 0.5 mM arsenite, and the levels of phospho-p38 (P-p38) and total p38 were determined by Western blotting, as described under "Materials and Methods" and as in Fig. 2. The image of the p38 region of the blot is shown at the top. The relative intensities of each phospho-p38 band were divided by the relative intensities of each total p38 band to normalize for slight differences in total p38 levels between the cultures. The means Ϯ S.E. of the relative intensities of phospho-p38/p38 for each treatment are shown in the bar graph at the bottom. The data are presented as the mean of three identically treated cultures. C, MAPKAP-K2. Cardiac myocytes were cultured identically to those used in A, except they were not transfected. After 48 h, they were treated for 10 min with or without 400 mM sorbitol or 0.5 mM arsenite with or without SB203580, and then endogenous MAPKAP-K2 was immunoprecipitated from culture extracts, and MAPKAP-K2 kinase assays were carried out using hsp27 as the substrate, as described under "Materials and Methods." The image of the region of the blot where hsp27 migrates is shown at the top of the panel. The mean intensity of each hsp27 band Ϯ S.E. for each treatment is shown in the bar graph at the bottom. The data are presented as the mean of three identically treated cultures. *, p Ͻ 0.05 different from control as determined by ANOVA followed by Neuman Keuls post hoc analysis of variance. tion in response to stress-mediated MKK6 and p38␤ activation.
Another possible mechanism by which MKK6 could protect cells from apoptosis involves the small hsps, such as ␣B-crystallin. The results of the present study show that in cardiac myocytes, constitutively active MKK6 can augment ␣B-crystallin gene expression, and it can induce the site-selective phosphorylation of ␣B-crystallin on serine 59. Moreover, the blockade of ␣B-crystallin gene induction and serine 59 phosphorylation by inhibitors of p38 was shown to augment myocyte apoptosis. These results indicate that the activation of MKK6, p38, and MAPKAP-K2 by stress probably represents a cellular response that is mounted in an effort to provide protection from apoptosis.
The observation that MKK6-mediated increases in ␣B-crystallin gene expression contribute to cytoprotection is in accord with a previous report that showed that overexpression of ␣B-crystallin protects cultured cardiac myocytes from ischemic injury, as measured by cytosolic enzyme release (49). The mechanism of ␣B-crystallin induction in the heart is unknown. However, since the 5Ј-flanking sequence of ␣B-crystallin possesses an SRE, transcriptional induction is probably coordinate with that of other stress-activated cardiac genes that possess regulatory SREs, such as ANF. Two SREs, located at Ϫ406 and Ϫ134 in the promoter-proximal region of the ANF 5Ј-flanking sequence, have been shown to bind serum response factor and to be required for both basal ANF expression and ANF transcriptional induction in response to MKK6-activated p38 (39,50). Moreover, it is through the phosphorylation of a recently discovered serum response factor binding partner, ATF6, that p38 is thought to increase ANF transcription (50). Thus, it is feasible that the SRE located at nucleotide Ϫ384 in the ␣Bcrystallin 5Ј-flanking sequence, which, like the SREs in the ANF gene, binds serum response factor (38), serves a role in cardiac expression and p38 inducibility of ␣B-crystallin. In support of this hypothesis is a preliminary experiment where we showed that truncating Ϫ426(␣BC)/Luc to Ϫ164(␣BC)/Luc resulted in an approximately 200-fold reduction in MKK6(Glu)inducible promoter activity. Therefore, it seems likely that ␣B-crystallin induction depends on the ability of p38 to phosphorylate ATF6 and may not rely directly upon p38-mediated activation of MAPKAP-K2.
The precise mechanism by which ␣B-crystallin contributes to protecting cardiac myocytes against apoptosis is not known; however, there are likely to be many pathways in which ␣Bcrystallin participates. One observation that has captured a great deal of attention is that during stresses that typically lead to apoptosis in either isolated hearts or cultured cardiac myocytes, ␣B-crystallin translocates from a soluble, cytosolic location to the Z-bands of the sarcomeres (34,35,51). It is possible that this translocation fosters sarcomeric stabilization, and then for 10 min with or without 400 mM sorbitol or 0.5 mM arsenite with or without SB203580 or PD098059, as indicated. A, phospho-Ser 45 . Culture extracts were analyzed for ␣B-crystallin phospho-Ser 45 or ␣Bcrystallin total (not shown) by Western blotting, as described under "Materials and Methods." B, phospho-Ser 59 . Culture extracts were analyzed for ␣B-crystallin phospho-Ser 59 or total ␣B-crystallin (not shown) by Western blotting, as described under "Materials and Methods." The phosphor image of the region of the blot where ␣B-crystallin migrates is shown at the top of each panel. The mean intensity of each phospho-␣B-crystallin band/total ␣B-crystallin Ϯ S.D. for each treatment is shown in the bar graph at the bottom. The data are presented as the mean Ϯ S.D. of three identically treated cultures. C, ERK and p38. The abilities of sorbitol and arsenite to increase levels of phospho-p38 were determined as described in the legend to Fig. 3. Phospho-ERK assays were carried out as described under "Materials and Methods." *, p Ͻ 0.05 different from control as determined by ANOVA followed by Neuman Keuls post hoc analysis of variance. which provides a degree of resistance to apoptosis. This concept is supported by studies showing that a collection of proteins, including ␣-actinin, ␣-actin, titin, nebulin, and Nsp11, localize to the Z-band, where they are believed to be critical for proper sarcomere formation during early myofibrilogenesis (52,53). Moreover, the recent finding that a naturally occurring point mutation in human ␣B-crystallin, R120G, is the cause of desmin-related myopathy, a disease where patients present with symptoms similar to cardiomyopathy, supports a role for ␣B-crystallin in sarcomeric stabilization as well as protection from apoptosis (54). Interestingly, this point mutation results in a form of ␣B-crystallin that is less stable to heat denaturation and acts as a poor chaperone and may thus contribute to the cardiomyopathic phenotype (55).
The nature of the binding of ␣B-crystallin to Z-bands remains to be elucidated. However, it has been observed that ␣B-crystallin binds directly to portions of titin that are located near the Z-band (35). It has also been shown that desmin can bind to ␣B-crystallin in vitro, prompting some speculation that ␣B-crystallin may bind to a portion of desmin that resides at or near the Z-bands (56). Another potential ␣B-crystallin binding partner is nebulette, a cardiac-specific isoform of nebulin that is localized at the actin-Z-band junction (57).
The molecular mechanism underlying the translocation of ␣B-crystallin to Z-bands is unknown; however, there is evidence supporting a role for phosphorylation. For example, the phosphorylation of ␣B-crystallin following myocardial stress is temporally correlated with its translocation to the Z-bands (35). In fact, some of the stresses that lead to this phosphorylation and translocation, such as ischemia/reperfusion, are well known to activate p38 and MAPKAP-K2 in the heart (7,8,22). These findings support the hypothesis that the phosphorylation of ␣B-crystallin on serine 59 by p38-activated MAP-KAP-K2 might play an important role in the mechanism by FIG. 6. Effects of p38 on sorbitol-induced apoptosis. A, cultured cardiac myocytes were treated with 400 mM sorbitol for the times shown with or without SB203580. DNA was then extracted and fractionated on an agarose gel, stained with ethidium bromide, and photographed. L, 100-bp DNA ladder; 100 bp is the fastest migrating band. B, the intensity of the 400-bp band (lowest) from each sample shown in A was assessed using ImageQuant software (Molecular Dynamics). Each intensity relative to the control (no sorbitol or SB203580) is shown in the bar graph. C, triplicate cultures of cardiac myocytes were treated for 5 h with 400 mM sorbitol with or without SB203580 or PD098059, as shown. DNA fragmentation was assessed as described for A. D, the intensities of the 400-bp bands from C were determined and plotted as means Ϯ S.E. *, p Ͻ 0.05 different from control; †, p Ͻ 0.05 different from sorbitol or sorbitol plus PD098059, as determined by ANOVA followed by Neuman Keuls post hoc analysis of variance. E, cultured cardiac myocytes were infected with or without AdV-Con, AdV-MKK6(Glu), or AdV-p38AGF and, after 48 h, treated with or without 400 mM sorbitol for 8 h. DNA fragmentation was assessed as described for A. F, the intensity of the 400-bp band (lowest) from each sample shown in E was assessed using ImageQuant software (Molecular Dynamics). Each intensity relative to the control (no sorbitol) is shown in the bar graph. G, duplicate cultures of cardiac myocytes were infected with or without AdV-Con or AdV-MKK6(Glu) and, after 48 h, treated with or without 400 mM sorbitol for 0, 12, or 24 h, as shown. DNA fragmentation was assessed as described for A. H, the intensities of the 400-bp bands from G were determined and plotted as means Ϯ S.D. *, p Ͻ 0.05 different from control (0 h sorbitol; AdV-Con), as determined by ANOVA followed by Neuman Keuls post hoc analysis of variance.
which ␣B-crystallin translocates from a diffuse cytosolic localization to Z-bands. Moreover, the finding in the present study that inhibiting the phosphorylation of ␣B-crystallin on serine 59 leads to the potentiation of apoptosis in response to stress, coupled with other ␣B-crystallin localization studies, suggests that it is the movement of ␣B-crystallin to the Z-bands that participates in protecting the integrity of myofibrils during the cardiac stress response.
The importance of ␣B-crystallin in the cellular stress response extends beyond the heart. In addition to the lens, where ␣B-crystallin was first discovered and where it is believed to protect from cataract formation (58), ␣B-crystallin is also expressed in skeletal muscle cells as well as some CNS cells, including oligodendrocytes, astrocytes, and neurons (58). Interestingly, the same mutation that results in desmin-related myopathy in humans also causes premature cataracts, which are believed to be the result of the inability of lens cells to respond to stresses, such as UV exposure (54). ␣B-crystallin has also been implicated in several neurodegenerative diseases. For example, the expression of ␣B-crystallin in astrocytes is considerably increased in Parkinson's patients suffering from dementia and in Alzheimer's disease, where it is believed to contribute to cellular mechanisms of adaptation to the stress caused by the disease (60). ␣B-crystallin has also been implicated in multiple sclerosis, where it is thought to serve as the autoantigen against which antibodies are expressed that lead to eventual neurodegeneration (37). Thus, although ␣B-crystallin displays a tissue-restricted expression pattern, it appears to play potentially important roles in providing protection against potentially harmful stresses in a variety of cell types.
The quantity of ␣B-crystallin expression in the heart is strikingly high, amounting to as much as 5% of the total protein in cardiac myocytes. This level of expression is equal only to the major sarcomeric proteins that comprise the basic functional units of all cardiac myocytes. Thus, it is apparent that in comparison with the sarcomeric proteins, ␣B-crystallin plays an important structural role in the heart. However, in contrast to the sarcomeric proteins, the localization of ␣B-crystallin is clearly dynamically regulated in a manner that allows it to transiently localize to myocyte structures that require a molecular chaperone to maintain structure and cell viability. It is probable that ␣B-crystallin localization is regulated in response to numerous signaling pathways in addition to the p38/MAPKAP-K2 pathway studied in the present paper. Future studies focused on determining the precise biochemical mechanisms governing the localization of ␣B-crystallin to the cellular machinery that is critical for mounting a protective stress response will reveal new information about this important molecular chaperone. Moreover, such studies will provide a better understanding of how to devise molecular approaches aimed at exploiting the cytoprotective roles of ␣B-crystallin to treat myocardial pathologies.