Role of p90 Ribosomal S6 Kinase (p90RSK) in Reactive Oxygen Species and Protein Kinase C β (PKC-β)-mediated Cardiac Troponin I Phosphorylation*

Protein kinase C (PKC)-induced phosphorylation of cardiac troponin I (cTnI) depresses the acto-myosin interaction and may be important during the progression of heart failure. Although both PKCβII and PKCϵ can phosphorylate cTnI, only PKCβ expression and activity are elevated in failing human myocardium during end-stage heart failure. Furthermore, although increased cTnI phosphorylation was observed in mice with cardiac-specific PKCβ II overexpression, no differences were observed in cTnI phosphorylation status between wild type and cardiac-specific PKCϵ overexpression mice. A potentially important downstream effector of PKCs is p90 ribosomal S6 kinase (p90RSK), which plays an important role in cell growth by activating several transcription factors as well as Na+/H+ exchanger. Since both Ser23 and Ser24 of cTnI are contained in putative consensus sequences of p90RSK phosphorylation sites, we hypothesized that p90RSK is downstream from PKCβ II and can be a cTnI (Ser23/24) kinase. p90RSK, but not ERK1/2 activation, was increased in PKCβII overexpression mice but not in PKCϵ overexpression mice. p90RSK could phosphorylate cTnI in vitro with high substrate affinity but not cardiac troponin T (cTnT). To confirm the role of p90RSK in cTnI phosphorylation in vivo, we generated adenovirus containing a dominant negative form of p90RSK (Ad-DN-p90RSK). We found that the inhibition of p90RSK prevented H2O2-mediated cTnI (Ser23/24) phosphorylation but not ERK1/2 and PKCα/βII activation. Next, we generated cardiac-specific p90RSK transgenic mice and observed that cTnI (Ser23/24) phosphorylation was significantly increased. LY333,531, a specific PKCβ inhibitor, inhibited both p90RSK and cTnI (Ser23/24) phosphorylation by H2 O2. Taken together, our data support a new redox-sensitive mechanism regulating cTnI phosphorylation in cardiomyocytes.

There is increasing support for the idea that excessive production of reactive oxygen species (ROS) 1 contributes to the pathogenesis of diabetes. In particular, a strong correlation has been made between increased ROS, activation of specific protein kinase C (PKC) isoforms, and many functional consequences of diabetes (1,2). This correlation has been strengthened by previous transgenic experiments in the heart, where overexpression of the PKC␤ isoform, but not PKC⑀, decreased cardiac function (3)(4)(5). These data also suggest that distinct PKC isoforms may play differential functional roles in cell signaling pathways leading to cardiac dysfunction, but the exact significance of individual isoforms is not yet known.
The family of PKCs includes at least 11 isoforms (␣, ␤I, ␤II, ␥, ␦, ⑀, , , , , ), representing the major downstream targets for lipid second messenger or phorbol esters (6,7). PKC isoforms are classified into classical PKCs (␣, ␤I, ␤II, ␥), which are Ca 2ϩ -dependent and contain two cysteine-rich zinc-finger-like motifs (C1 region) that bind diacylglycerol or phorbol ester and a Ca 2ϩ / phospholipids binding domain (C2 region). Novel PKCs (␦, ⑀, , , ) are diacylglycerol-sensitive but Ca 2ϩ -independent because of the absence of the C2 region, and atypical PKCs (, ) are rather insensitive to diacylglycerol as they lack one cysteine-rich motif in their C1 region but can be activated by phosphatidylserine. Our colleagues have previously reported that in failing human myocardium with end-stage heart failure, the expression and activity of Ca 2ϩ -sensitive PKC␣ and -␤ isoforms are elevated (8). In addition, cardiac-specific overexpression of PKC␤II isoform in mice causes left ventricular hypertrophy, myocardial fibrosis, and decreased in vivo left ventricular performance (4). Of note, these PKC␤II transgenic mice have greater phosphorylation of the myofilament regulatory protein cardiac troponin I (cTnI), which decreases sensitivity of cardiomyocytes to Ca 2ϩ and may lead to cardiac dysfunction (3). In contrast, cardiac-specific overexpression of a constitutively active PKC⑀ mutant results in mild concentric hypertrophy, normal left ventricular performance, no change in cTnI phosphorylation, and an increase in myofilament Ca 2ϩ sensitivity (9 -11). Since both PKC␤ and PKC⑀ could phosphorylate cTnI (12), it remains unclear how PKC␤II and PKC⑀ can differently regulate cTnI phosphorylation.
p90RSK has multiple functions. In quiescent cells, inactive p90RSK resides in the cytoplasm and is partially complexed with its upstream regulator, ERK1/2 (13). Stimulation of the cells by growth factors or ROS operating through Ras/Raf-1/ MEK1/2 pathways leads to the activation of ERK1/2, phosphorylation and activation of p90RSK, and the import of these kinases into nucleus. It has been proposed that p90RSK is involved in activation of nuclear factor-B by phosphorylation of I-B (14) or phosphorylates transcription factors, including c-Fos (15), Nur77 (16), and cAMP-response element-binding protein (17). Furthermore, we reported that p90RSK is a serum-stimulated Na ϩ /H ϩ exchanger-1 (NHE-1) kinase and that p90RSK regulates its activity (18). However, the role of p90RSK in the heart remains unclear.
We have previously demonstrated that H 2 O 2 -mediated p90RSK activation is partially dependent on PKC activation in Jurkat cells (19). To determine the role of specific PKC isoforms on cardiac p90RSK activation in vivo, we utilized mice with cardiac-specific PKC␤II and PKC⑀ overexpression. We show here that p90RSK activation is specifically up-regulated in PKC␤II transgenic (Tg) mice but not in PKC⑀-Tg mice. Since cTnI possesses putative phosphorylation sites of p90RSK at Ser 23 and Ser 24 , and cTnI phosphorylation only increases in PKC␤II-Tg mice, we hypothesized that cTnI (Ser 23/24 ) was a p90RSK substrate. We demonstrate here that p90RSK can directly phosphorylate cTnI, and p90RSK activation is required for H 2 O 2 -mediated cTnI (Ser 23/24 ) phosphorylation. To determine the role of p90RSK activation in vivo, we also generated cardiac-specific p90RSK overexpression mice, and we report here that cTnI (Ser 23/24 ) phosphorylation is increased in these mice. Taken together, these data support the importance of p90RSK activation in H 2 O 2 -mediated cTnI (Ser 23/24 ) phosphorylation. The PKC␤-p90RSK-cTnI (Ser 23/24 ) signaling pathway described here may present a new redox-sensitive mechanism to regulate cardiac function via cTnI phosphorylation.

EXPERIMENTAL PROCEDURES
Protein Extract from Heart Tissue-Mouse hearts were washed with 10 ml of cold phosphate-buffered saline. Isolated mice heart tissues were frozen in liquid nitrogen and homogenized with 0.5 ml of lysis buffer (10 mM Tris-HCl pH 7.4, 0.15 M NaCl, 0.05% Triton X-100, FIG. 1. Activation of p90RSK and ERK1/2 in PKC␤II or PKC⑀ transgenic mice. A, immunoblots (IB) of the PKC␤II isoform in whole cardiac homogenates from NLC and cardiac-specific PKC␤II transgenic mice (PKC␤II-Tg) (top panel). p90RSK activity was measured by in vitro kinase assay using GST-NHE-1 (amino acids 625-747) as substrate (second panel from the top). No difference in the amount of p90RSK was observed in lysates from any of the heart samples using Western blot analysis with anti-p90RSK (third panel from the top). ERK1/2 activity was measured by Western blot analysis with a phospho-specific ERK1/2 antibody (pERK1/2, second panel from the bottom). No difference in the amount of ERK1 was observed in lysates from any of the heart samples using Western blot analysis with anti-ERK1 (bottom panel). B, densitometric analysis of p90RSK activation in PKC␤-Tg mice. Results were normalized for all experiments by arbitrarily setting the mean densitometry of control normal heart samples to 1.0 (shown in mean Ϯ S.D., n ϭ 3, **, p Ͻ 0.01). C, immunoblots of the PKC⑀ isoform in whole cardiac homogenates from NLC and cardiac-specific PKC⑀ transgenic mice (PKC⑀-Tg) (top panel). p90RSK activity was measured by in vitro kinase assay using GST-NHE-1 (amino acids 625-747) as substrate (second panel from the top). No difference in the amount of p90RSK was observed in lysates from any of the heart samples using Western blot analysis with anti-p90RSK (third panel from the top). ERK1/2 activity was measured by Western blot analysis with a phospho-specific ERK1/2 antibody (second panel from the bottom). No difference in the amount of ERK1 was observed in lysates from any of the heart samples using Western blot analysis with anti-ERK1 (bottom panel). D, densitometric analysis of p90RSK activation in PKC⑀-Tg mice. Results were normalized for all experiments by arbitrarily setting the mean densitometry of control normal heart samples to 1.0 (shown in mean Ϯ S.D., n ϭ 3; N.S., not significant). 0.05% Nonidet P-40) containing 2 mmol/liter sodium orthovanadate and protease inhibitor mixture (Sigma). Protein concentration was determined with the Bradford protein assay (Bio-Rad). Protein (30 g) was separated on SDS-polyacrylamide gels and transferred to nitrocellulose membranes.
p90RSK in Vitro Kinase Assays-For the autoradiogram studies in Fig. 2A, hearts were crushed in liquid nitrogen, homogenized with 3 volumes of lysis buffer, and centrifuged at 14,000 ϫ g (4°C for 30 min). Protein concentration of supernatants was then determined. p90RSK was immunoprecipitated through the incubation of 1000 g of protein from each sample with 3 l of rabbit polyclonal anti-p90RSK (Santa Cruz Biotechnology) antibody for 3 h followed by the addition of 40 l of a 1:1 slurry of protein A/Sepharose beads to the extract/antibody mixture and then incubation for 1 h at 4°C. This complex was washed, twice each, in cell lysis buffer described above, LiCl buffer (500 mM LiCl, 100 mM Tris-HCl (pH 7.6), 0.1% Triton X-100, 1 mM dithiothreitol) and wash buffer (20 mM Hepes, pH 7.2, 2 mM EGTA, 100 M Na 3 VO 4, 10 mM MgCl 2 , 1 mM dithiothreitol, 0.1% Triton X-100). After the final wash and pelleting, the p90RSK substrates purified human cTnI and cardiac troponin T (cTnT) (Calbiochem) were added with kinase reaction buffer (25 mM Hepes buffer, pH 7.5, 10 mM MgCl 2 , 10 mM MnCl 2 , 15 M cold ATP, 0.5 mCi/ml of [␥-32 P]ATP) in a total volume of 20 l. The kinase reaction was proceed for 20 min at 30°C with agitation and stopped by the addition of 1ϫ gel loading buffer. Samples were separated by 10% SDS-PAGE, transferred overnight at room temperature to nitrocellulose membranes, and then exposed to an x-ray film between two enhancers for 3-6 h. cTnI phosphorylation was determined through densitometry of bands at the correct molecular weights in the linear range of film exposure with the use of a scanner and NIH Image 1.59.
For phosphorylation kinetic studies in Fig. 2B, we determined the K m and V max value for cTnI phosphorylation induced by p90RSK by Hanes-Woolf plot analysis using the Michaelis-Menten equation. Reaction mixtures (200 l) contained 25 mM Hepes buffer, pH 7.5, 10 mM MgCl 2 , 10 mM MnCl 2 , 15 M cold ATP, 5 M [␥-32 P]ATP (2 ϫ 10 6 cpm), and various concentrations of substrate up to 4 M. We utilized recombinant full-length p90RSK (Upstate Biotechnology), which was purified from Sf21 insect cells transfected with baculovirus expression vector. Reactions with p90RSK were carried out for 5 min (or various times up to 90 min) at 30°C and terminated by the addition of 5% trichloroacetic acid-tungstate, and 32 P incorporation was analyzed as described previously (12). Since we found that the phosphorylation rate was linear as a function of a 10 -20-min incubation time under this experimental condition (data not shown), all phosphorylation reactions were carried out for 15 min in the presence of 5 M [␥-32 P]ATP (2 ϫ 10 6 cpm).
We also used S6 kinase substrate peptide to confirm the effect of the dominant negative form of p90RSK on p90RSK kinase activity in Fig. 4 as described previously (20). The in vitro kinase assay was performed according to manufacturer's protocol using a long S6 kinase substrate peptide (Upstate Biotechnology) to determine radiolabeled phosphate incorporation using a scintillation counter. Briefly, washed beads were incubated in 40 l of assay dilution buffer (20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, and 1 mM dithiothreitol), 10 l of 150-m-long S6 kinase substrate peptide, 10 Ci of [␥-32 P]ATP (Amersham Bioscience), 100 M ATP, and 15 mM MgCl for 30 min at 30°C. The reaction was terminated by spotting 40 l of reaction onto P81 phosphocellulose filter paper. The filter was washed five times in 0.75% phosphoric acid and one time in acetone for 5 min, and radioactive incorporation was assayed by Cerenkov counting.
MALDI-TOF Mass Spectrometry Analysis of cTnI Phosphorylation by p90RSK-Tryptic digestion of pooled gel slices containing recombinant human cardiac TnI after incubation with recombinant p90RSK was subjected to enzymatic cleavage for the generation of peptide fragments. Pieces were washed with 100 mM ammonium bicarbonate, reduced (dithiothreitol) and alkylated (iodoacetamide), and then dehydrated via acetonitrile evaporation. The gel pieces were reswollen with 25 mM ammonium bicarbonate containing ϳ0.2 g of enzyme to achieve a substrate/enzyme ratio of ϳ10:1. ZipTip tippets (Millipore, Bedford, MA), packed with C18 matrix, were utilized to clean and concentrate peptide samples prior to analysis. Tips were washed with acetonitrile before peptides were bound and then eluted with either acetonitrile or matrix solution. ZipTip use affords a recovery of 50 -70% in a 1-l volume. Digested protein was mixed with the matrix a-cyano-4-hydroxycinnamic acid, and matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometric analysis was performed in the University of Rochester Protein/Peptide Core Facility as described previously (21). Mass fingerprinting analysis and determination of phosphorylation was performed initially by MS-FIT. The data base search was considered significant if the protein was ranked as the best hit with a sequence coverage of more than 30%. Significance was defined as a MOWSE (Molecular Weight Search) score of at least 1e ϩ003 (MS-FIT) or a difference in probability of 10 Ϫ3 from the first to the second protein candidate (ProFound).
Western Blot Analysis-After treatment with reagents or 24 h after adenovirus transduction, the cells were washed with phosphate-buffered saline and harvested in 0.5 ml of lysis buffer as described previously (22). Western blot analysis was performed as described previously (22). In brief, the blots were incubated for 4 h at room temperature with the anti-phospho-cardiac troponin I (Cell Signaling), which recognizes dual phosphorylation of Ser 23 and Ser 24 , anti-troponin I (Cell Signaling), and PKC⑀ antibody (Santa Cruz Biotechnology) followed by incubation with horseradish peroxidase-conjugated secondary antibody (Amersham Biosciences). Antibodies for assaying ERK1/2, p90RSK, and PKC␣/␤II activation and anti-ERK1 or 2, p90RSK2, and PKC␤ antibody were from Santa Cruz Biotechnology, and the phospho-ERK1/2 (Thr 202 /Tyr 204 ), phospho-p90RSK (Thr 359 /Ser 363 ), and phospho-PKC␣/␤II (Thr 638/641 ) antibodies were from Cell Signaling.
Adenovirus Vector Containing Dominant Negative Form of p90RSK-The dominant negative form of rat p90RSK1 (K93A/K447A) construct was cloned into the AdEasy TM -CMV system (QBIOGene, Carlsbad, CA) using SalI and HindIII restriction enzymes as described previously (23).
Preparation of Rat Neonatal Cardiomyocytes-Primary cultures of cardiac myocytes were prepared from ventricles of 1-3-day-old neonatal Wistar rats (24). Briefly, cells were dissociated by collagenase II (Worthington Biochemical) from the ventricles and plated at a density of 1.5 ϫ 10 5 cells/cm 2 on 35-mm collagen-coated coverslips in culture medium (Dulbecco's modified Eagle's medium with 10% fetal bovine serum and 10% horse serum). After 6 h of plating the isolated cardiomyocytes, 10 M cytosine 1-␤-D-arabinofuranoside was added and cultured 24 h. After that, the culture medium was changed to the medium with 0.5% fetal bovine serum and 10 M Ara C.
Animals-PKC␤ transgenic mice were generated by G. L. King as described previously (4), and PKC⑀ transgenic mice were generated by T. Jalili and R. A. Walsh. Briefly, a 2.7-kb fragment containing mouse PKC⑀ cDNA and endogenous polyadenylation sequence (kindly provided by W. J. van Blitterswijk, Netherlands Cancer Institute, Amsterdam, The Netherlands) was subcloned between the 5.5-kb murine ␣-myosin heavy chain promoter and a 250-bp SV40 polyadenylation sequence (kind gift from J. Robbins, Children's Hospital Research Foundation, Cincinnati, OH). The purified transgene fragment was injected into male pronuclei of fertilized mouse oocytes (University of Cincinnati Transgenic Core, Cincinnati, OH). Genotype of mouse pups was confirmed by PCR analysis of ear clipping using standard procedures.
Rat p90RSK1 cDNA was subcloned into a pBluescript-based Tg vector between the 5.5-kb murine-␣-MHC (␣-myosin heavy chain) promoter and the 250-bp SV40 polyadenylation sequences (a kind gift from J. Robbins, Children's Hospital Research Foundation, Cincinnati, OH). The purified transgene fragment was injected into male pronuclei of fertilized mouse oocytes (University of Rochester Transgenic Core). The genotype of mouse pups was confirmed by PCR analysis of tail clipping using standard procedures.
Statistical Analysis-Values presented are mean Ϯ S.D. Statistical analysis was performed with the StatView 4.0 package (Abacus Concepts). Differences were analyzed with one-or two-way repeated measures analysis of variance as appropriate followed by Scheffé's correction.

RESULTS
p90RSK Is Activated in PKC␤II Transgenic Mice but Not in PKC⑀ Transgenic Mice-It has been reported that cardiacspecific overexpression of the PKC␤II isoform in transgenic mice causes a cardiomyopathy, which is characterized by left ventricular hypertrophy, myocardial fibrosis, and decreased in vivo left ventricular performance (4). In these mice, PKC␤IIinduced phosphorylation of the myofilament regulatory protein cTnI decreases cardiomyocyte calcium sensitivity and may cause depressed cardiomyocyte function (5). On the other hand, the calcium-independent PKC⑀ isoform has been implicated in cardiomyopathy and ischemic preconditioning (25). Of note, unlike PKC␤II overexpression mice, transgenic mice with cardiac-specific overexpression of wild type PKC⑀ demonstrated concentric hypertrophy with normal in vivo cardiac function. We evaluated p90RSK activity in both of these transgenic mice and found that p90RSK activation is significantly increased in PKC␤II transgenic mice in comparison with control non-transgenic littermate control (NLC) mice (Fig. 1, A and B). However, in PKC⑀ transgenic mice, we could not find any differences in p90RSK activity between NLC and transgenic mice (Fig. 1, C  and D). These data are consistent with our hypothesis that p90RSK could contribute to PKC␤-mediated depressed cardiac function by phosphorylation of cTnI.
p90RSK Phosphorylates TnI but Not Troponin T with High Substrate Affinity-We asked whether recombinant p90RSK could phosphorylate cTnI and cTnT in vitro. Recombinant p90RSK phosphorylated cTnI, but we could not detect any phosphorylation of cTnT ( Fig. 2A). We also performed a kinetic study to determine the K m and V max value for cTnI phosphorylation induced by p90RSK using a Hanes-Woolf plot analysis with the Michaelis-Menten equation based on data from three experiments. The initial rate of phosphorylation of human cTnI wild type was studied as shown in Fig.  2B. Since the phosphorylation rate was linear as a function of a 10 -20-min incubation time under this experimental condition (data not shown), all phosphorylation reactions were carried out for 15 min in the presence of 5 M [␥-32 P]ATP (2 ϫ 10 6 cpm). It was found that K m (M) ϭ 0.6 Ϯ 0.1, V max (pmol/min) ϭ 4.6 Ϯ 0.1, and V max /K m ϭ 7.5, which were similar to TnI phosphorylation induced by PKC␣ and PKC␦ as reported previously (26). These data support the hypothesis that cTnI is a good substrate for p90RSK.

FIG. 3. p90RSK activation and cTnI (Ser 23/24 ) phosphorylation were increased by oxidative stress in cardiomyocytes.
A, cardiomyocytes were stimulated for the indicated times with 100 M H 2 O 2 . Cells were harvested in lysis buffer, and Western blot analysis was performed with anti-phosphospecific p90RSK antibody (top panel). No difference in the amount of p90RSK was observed in lysates from any of the heart samples using Western blot analysis with anti-p90RSK (second panel from the top). Bottom panel, densitometric analysis of p90RSK activation. Results were normalized to control (time ϭ 0), which was arbitrarily set to 1.0 (shown is mean Ϯ S.D., n ϭ 3). IB, immunoblots. B, cardiomyocytes were stimulated for the indicated times with 100 M H 2 O 2 . Cells were harvested in lysis buffer, and Western blot analysis was performed with anti-phosphospecific cTnI (Ser 23/24 ) antibody (p-TnI, top panel). No difference in the amount of cTnI was observed in lysates from any of the heart samples using Western blot analysis with anti-cTnI (second panel from the top). Bottom panel, densitometric analysis of cTnI (Ser 23/24 ) phosphorylation. Results were normalized to the mean densitometry of control (time ϭ 0), which was arbitrarily set to 1.0 (shown is mean Ϯ S.D., n ϭ 3).

Determination of the Phosphorylation Sites by MALDI-TOF
Mass Spectrometry-To determine p90RSK phosphorylation sites on cTnI, we performed tryptic peptide mapping using MALDI-TOF mass spectrometry analysis after recombinant p90RSK was immunoprecipitated and incubated with cTnI in an in vitro kinase reaction. Computer-assisted proteomic analysis revealed two phosphorylated tryptic ions with mass/charge ratios (m/z) of 786.2225 and 942.3236, which correspond to cTnI residues 23-27 ϩ 2PO 4 and cTnI residues 22-27 ϩ 2PO 4 , respectively. Since p90RSK is a Ser/Thr kinase, and the cTnI peptide residues 22-27 contains two Ser at positions 23 and 24, it is most likely that both Ser 23 and Ser 24 are phosphorylated by p90RSK activation. In addition, we utilized an anti-phosphospecific TnI antibody (Cell Signaling), which recognized dual phosphorylation of Ser 23 and Ser 24 sites. As shown below (see Fig. 5), we found dual phosphorylation of Ser 23/24 in cardiac TnI in p90RSK-Tg mice in vivo and also detected Ser 23/24 dual phosphorylation by p90RSK in vitro kinase assay using cTnI as a substrate (data not shown). Taken together, these data suggest that p90RSK phosphorylates cTnI at both Ser 23 and Ser 24 . Interestingly, we also found that the phosphopeptides with m/z 693.2374, 821.3324, and 908.3644 correspond to cTnI residues 41-45 ϩ 2PO 4 , 41-46 ϩ 2PO 4 , and 39 -45 ϩ 2PO 4 , respectively. Of note, cTnI 41-45 residues contain reported PKC phosphorylation sites of Ser 42 and Ser 44 in human cTnI (12). These data suggest that p90RSK can phosphorylate cTnI at multiple sites.

H 2 O 2 Induced cTnI (Ser 23/24 ) Phosphorylation via p90RSK
Activation in Cardiomyocytes-ROS and growth factors stimulate similar intracellular signal transduction events including activation of Src kinase family members and extracellular signal-regulated kinase (ERK1/2) (27)(28)(29). To determine whether ROS activates both p90RSK and cTnI phosphorylation, we used H 2 O 2 (100 M) to stimulate p90RSK activity and cTnI phosphorylation. Of note, we investigated here the Ser 23/24 ( 20 RRRSS 24 ) phosphorylation of cTnI using anti-dual phosphospecific (Ser 23/24 ) TnI antibody, which have been reported as PKA phosphorylation sites because these sites have a significant role in regulating TnI function (12) and contain putative consensus sequence of p90RSK (RRXS). As shown in Fig.  3A, we found that H 2 O 2 stimulated p90RSK activation after 10 min and was sustained up to 60 min. Using an anti-dual phosphospecific Ser 23/24 cTnI antibody, we found that cTnI phosphorylation was increased after 10 min and also sustained after 60 min of stimulation, paralleling the time course of H 2 O 2 -mediated p90RSK activation.
To determine the role of p90RSK activation in H 2 O 2 -mediated cTnI phosphorylation, we generated an adenovirus vector containing the dominant negative form of p90RSK (Ad-DN-p90RSK) and transfected it into cardiomyocytes. As shown in Fig. 4, A (second panel from the top) and B (upper panel), we found that H 2 O 2 -mediated cTnI phosphorylation was significantly inhibited in Ad-DN-p90RSK transfected cells. In contrast, we did not observe any inhibition on ERK1/2 activation (Fig. 4, A, second panel from the bottom, and B, lower panel). We also confirmed that p90RSK activity was significantly decreased in Ad-DN-90RSK transfected cells (Fig. 4C). These data suggest that p90RSK activation is critical for H 2 O 2 -mediated cTnI phosphorylation.
cTnI Phosphorylation in Cardiac-specific p90RSK Transgenic Mice (p90RSK-Tg)-To examine the effect of activation of p90RSK on cTnI phosphorylation at the whole organ level, we created Tg mice with cardiac-specific expression of the wild type of p90RSK. The level of Tg protein expression in three different lines of Tg mice was determined by Western blot using an anti-p90RSK antibody. We found an 8 -9-fold increase in the level of p90RSK expression in all Tg lines relative to wild type mice, and there was a concordant 10 -12-fold increase in p90RSK phosphorylation. However, there was no significant ERK1/2 and PKC␣/␤II activation in p90RSK-Tg mouse hearts, confirming a selective activation of p90RSK (Fig. 5). cTnI (Ser 23/24 ) phosphorylation was significantly increased in p90RSK-Tg mice (Fig. 5, A, third panel from the top, and B,  upper panel), which was also observed in other two lines of p90RSK-Tg mice (data not shown). These data support a role for p90RSK activation in cTnI (Ser 23/24 ) phosphorylation in vivo.

Role of PKC␤ in H 2 O 2 -mediated p90RSK Activation and cTnI (Ser 23/24 ) Phosphorylation in
Cardiomyocytes-It has been reported that Ser 43/45 and Thr 144 of cTnI are phosphorylation sites of PKC, but it is unclear whether PKC activation can indirectly result in greater Ser 23/24 phosphorylation because only minor phosphorylation of Ser 23/24 has been observed by PKC (9). Since we found that p90RSK activation was increased in PKC␤II-Tg mice and that p90RSK could directly phosphorylate cTnI in vitro, we investigated whether PKC␤ can regulate H 2 O 2 -mediated p90RSK activation and subsequent dual TnI (Ser 23/24 ) phosphorylation in cardiomyocytes. To determine the role of PKC␤ activation in H 2 O 2 -mediated p90RSK activation and cTnI phosphorylation, we used a myristoylated PKC␤ C2-4 inhibitor that inhibits the translocation and function of all classical PKC isoforms (30,31) and PKC␤-specific inhibitor, LY333,531. We confirmed that LY333,531 did not directly inhibit p90RSK activation by directly adding LY333,531 (20, 200 nM) in p90RSK in vitro kinase assay mixture (data not shown). As shown in Fig. 6, pretreatment with PKC␤ C2-4 inhibitor and LY333,531 abolished H 2 O 2 -mediated p90RSK activation (Fig.  6, A and C, top panel, and B and D, upper panel) and cTnI (Ser 23/24 ) phosphorylation (Fig. 6, A and C, third panel from the  top, and B and D, lower panel). These data indicate that PKC␤ activation is important for p90RSK activation as well as downstream cTnI (Ser 23/24 ) phosphorylation. DISCUSSION In the present study, we found that p90RSK activation regulates H 2 O 2 -mediated cTnI (Ser 23/24 ) phosphorylation. To our knowledge, this is the first report to document cTnI (Ser 23/24 ) phosphorylation by p90RSK. In addition, we found that p90RSK activation and cTnI (Ser 23/24 ) phosphorylation were increased in PKC␤II-Tg but not in PKC⑀-Tg mice and that the PKC␤ inhibitor prevented H 2 O 2 -mediated p90RSK activation and cTnI (Ser 23/24 ) phosphorylation. These data support a critical role for PKC␤II in regulating p90RSK activation and subsequent cTnI (Ser 23/24 ) phosphorylation. It has been reported that specific PKC-mediated phosphorylation of Ser 43/45 of cTnI plays an important role in regulating force development in cardiac muscle (9). Since p90RSK kinase, which is downstream of PKC␤II, could phosphorylate Ser 23/24 in N-terminal of cTnI, and PKC phosphorylation sites of Ser 43/45 already exist on cTnI, there appears to be a novel complexity of PKC␤II activation on cTnI function in vivo.
p90RSK consists of three isoforms that show the same overall structure consisting of two kinase domains, a linker region, and short N-terminal and C-terminal tails. The C-terminal kinase belongs to the calcium/calmodulin-dependent kinase group of kinases, but the only known function of the C-terminal kinase is the regulation of the activity of N-terminal kinases. The N-terminal kinase belongs to the AGC group of kinases, which include PKA and PKC. The N-terminal kinase is responsible for phosphorylating the RSK substrates and recognizes the basic consensus motif: (R/K)XRXX(S/T), or RRX(S/T), which fits to both Ser 23 and Ser 24 of cTnI ( 20 RRRSSN 25 ) (32,33).
In the current study, we also determined that human cardiac TnI residues 41-45 can be phosphorylated by p90RSK using MALDI-TOF mass spectrometry. TnI residues 41-45 contains two Ser at positions 42 and 44, which have been reported as major PKC-dependent phosphorylation sites in cardiac TnI (12). Although these sites do not show a consensus sequence of p90RSK phosphorylation sites, it is still possible that p90RSK can phosphorylate these Ser 42 and Ser 44 sites because the RSK amino-terminal kinase domain is homologous to the kinase domains in other members of the AGC family, including PKA, various PKCs, and p70RSK (34).
It should also be mentioned that during the preparation of this study, Kobayashi et al. (35) have published that Ser 23 or Ser 24 of cTnI is the most permissive site for PKC-dependent FIG. 5. cTnI (Ser 23/24 ) phosphorylation, but not ERK1/2 and PKC␣/␤II activation, was increased in cardiac-specific p90RSK transgenic mice. A, immunoblots (IB) of the p90RSK in whole cardiac homogenates from wild-type and cardiac-specific p90RSK transgenic mice (p90RSK-Tg) (second panel from the top). p90RSK phosphorylation was detected by Western blotting with anti-phosphospecific p90RSK antibody (top panel). cTnI (Ser 23/24 ) phosphorylation was measured by Western blot analysis with a phospho-specific cTnI (Ser 23/24 ) antibody (p-TnI, third panel from the top). No difference in the amount of cTnI was observed in lysates from any of the heart samples using Western blot analysis with anti-TnI (fourth panel from the top). ERK1/2 and PKC␣/␤II activity was measured by Western blot analysis with a phospho-specific ERK1/2 antibody (fourth panel from the bottom) and PKC␣/␤II (second panel from the bottom). No difference in the amount of ERK1 and PKC␤ was observed in lysates from any of the heart samples using Western blot analysis with anti-ERK1 (third panel from the bottom) and PKC␤ (the bottom panel). B, densitometric analysis of cTnI (Ser 23/24 ) phosphorylation and ERK1/2 activation in p90RSK-Tg mice. Results were normalized for all experiments by arbitrarily setting the mean densitometry of control normal heart samples to 1.0 (shown in mean Ϯ S.D., n ϭ 4, **, p Ͻ 0.01). phosphorylation, and the phosphorylated cTnI with Ser 42 and/or Ser 44 , previously reported as the phosphorylation sites of PKC (12), could not be detected. They explained the discrepancy of their data with a previous report (12) by the different usage of their recombinant PKC (Kobayashi et al. (35)) instead of a mixture of PKC isoforms from brain (Noland et al. (12)). Since we found that endogenous p90RSK activation was regulated by PKC␤ (Fig. 6) and also detected coimmunoprecipitation of endogenous p90RSK with PKC␤ (data not shown), it may be possible that activation of p90RSK is responsible for Ser 42/44 phosphorylation in a mixture of PKC isoforms from brain. Further investigation is necessary to clarify this issue.
The activation of PKC and increased diacylglycerol levels initiated by hyperglycemia are associated with many vascular abnormalities in retinal, renal, and cardiovascular disease (36,37). It has been reported that increased PKC activity in diabetic rat hearts could be due to ␤ and isoforms (36). Although we could not detect ERK1/2 activation in PKC␤-Tg mouse hearts, we observed significant activation of p90RSK activation. We have previously reported the presence of another ROS-stimulated pathway that activates p90RSK independently of MEK1/2 and ERK1/2 (19). It is possible that PKC␤II directly phosphorylates p90RSK and activates p90RSK activation, but this is beyond the scope of this study.
Desensitization of myofibrils to Ca 2ϩ appears to oppose the inotropic effect produced by phospholamban and sarcolemmal L-type calcium channel phosphorylation, but several studies also indicate that cTnI phosphorylation by PKA plays an essential role in the rate of force production (10,11,38). PKA phosphorylation of cTnI increases the rates of cross-bridge cycling, and transgenic mouse hearts expressing constitutively phosphorylated cTnI at PKA phosphorylation sites exhibit augmented force production and faster relaxation (39). We generated cardiac-specific p90RSK transgenic mice and found that TnI (Ser 23/24 ) phosphorylation was increased as shown in Fig.  5. In the current study, we focused on the role of p90RSK to phosphorylate cTnI (Ser 23/24 ), as a downstream event to PKC␤II activation. Since we also found that p90RSK can phosphorylate and regulate NHE-1 activity, it would be intriguing to determine the role of p90RSK activation on cardiac function via regulating cTnI (Ser 23/24 ) phosphorylation and NHE-1 activation, especially after ischemia/reperfusion injury. Since the cell localization of PKA and p90RSK is apparently different and compartmentation of kinase may be critical to determine its downstream effect (20,40), the effect of cTnI (Ser 23/24 ) phosphorylation on cardiac function may be also different between p90RSK and PKA, although they could phosphorylate the same sites of cTnI. Further studies will be required to clarify the role of p90RSK activation on cardiac function and Ca 2ϩ sensitivity via cTnI (Ser 23/24 ) phosphorylation.