The Raf-MEK-ERK Cascade Represents a Common Pathway for Alteration of Intracellular Calcium by Ras and Protein Kinase C in Cardiac Myocytes*

Ras and protein kinase C (PKC), which regulate the Raf-MEK-ERK cascade, may participate in the development of cardiac hypertrophy, a condition characterized by diminished and prolonged contractile calcium transients. To directly examine the influence of this pathway on intracellular calcium ([Ca2+] i ), cardiac myocytes were cotransfected with effectors of this pathway and with green fluorescent protein, which allowed the living transfected myocytes to be identified and examined for [Ca2+] i via indo-1. Transfection with constitutively active Ras (Ha-RasV12) increased cell size, decreased expression of the myofibrils and the calcium-regulatory enzyme SERCA2, and reduced the magnitude and prolonged the decay phase of the contractile [Ca2+] i transients. Similar effects on [Ca2+] i were obtained with Ha-RasV12S35, a Ras mutant that selectively couples to Raf, and with constitutively active Raf. In contrast, Ha-RasV12C40, a Ras mutant that activates the phosphatidylinositol 3-kinase pathway, had a lesser effect. The PKC-activating phorbol ester, phorbol 12-myristate 13-acetate, also prolonged the contractile [Ca2+] i transients. Cotransfection with dnMEK inhibited the effects of Ha-RasV12, Raf, and phorbol 12-myristate 13-acetate on [Ca2+] i . The effects of Ha-RasV12 and Raf on [Ca2+] i were also counteracted by SERCA2 overexpression. Both Ras and PKC may thus regulate cardiac [Ca2+] i via the Raf-MEK-ERK cascade, and this pathway may represent a critical determinant of cardiac physiological function.

quence SERCA2 (Ϫ13/3726). Next, the plasmid was linearized with DraIII at nucleotide 3076 of SERCA2 (located in the 3Ј noncoding region of the SERCA2 cDNA), and the overhang was filled with T4 polymerase, generating a blunt end. The plasmid was subsequently digested with KpnI, and the SERCA2 encoding fragment was inserted into the adenoviral shuttle vector, pACCMV.pLpA, at the KpnI/blunted HindIII sites. The resulting vector (pACCMV.pLpA-SERCA2) contains SERCA2(Ϫ13/ 3076) inserted between the CMV promoter and the SV40 polyadenylation sequence.
Immunocytofluorescence-Cardiac myocytes were fixed and stained for ␤-galactosidase, actin, and SERCA2 as published (7,11). An fluorescein isothiocyanate-conjugated goat anti-rabbit secondary antibody was utilized to visualize the anti-SERCA2 primary antibody. Actin was visualized using rhodamine phalloidin (Molecular Probes), according to the manufacturer's instructions.
Indo-1 Measurements-Cells were loaded with indo-1 as described previously (13). Indo-1 measurements were made in air-compatible Dulbecco's modified Eagle's medium at room temperature at 20 Hz. GFP-expressing myocytes, corresponding to myocytes that had taken up the GFP expression vector during the transfection procedure, were identified by monitoring fluorescence with an excitation wavelength of 475 nm. Excitation was then switched to 355 nm (optimal for indo-1), and indo-1 fluorescence was recorded from the cytoplasmic region of the myocyte. In the absence of indo-1, GFP-expressing myocytes were nonfluorescent at the excitation wavelength of 355 nm, indicating that GFP does not interfere with measurement of indo-1. Although indo-1 fluorescence is related to free calcium concentration via well established equilibrium relationships, exact calibration of the signal in vivo is problematic (14); the results are thus reported as the ratio between the fluorescence emission at 405 and 485, or as "normalized" transients, where the ratios at diastole and at the peak systole of the transient equal 0 and 100%, respectively.
ERK Assays-Myocytes were transfected with the test constructs and a hemaglutinin-tagged ERK2 expression vector (Cobb, University of Texas Southwestern Medical Center, Dallas, TX). Cultures were extracted, immunoprecipitated for the hemaglutinin epitope tag, and assayed for ERK activity in vitro as described (11).  1A) and averaged 841 Ϯ 115 m 2 (mean Ϯ S.E., n ϭ 13) in surface area. Ha-Ras V12 transfection (16 g/electroporation) resulted in an approximate 6-fold increase in cell area (to 4984 Ϯ 412, n ϭ 25) and the cells exhibited a highly irregular, sprawling morphology (Fig. 1B). Although Ha-Ras V12 -transfected myocytes sometimes stained strongly for actin, the actin filaments were not striated (Fig. 1C). The absence of striations in the Ha-Ras V12transfected myocytes does not represent a limitation of the detection technique, because striated myofibrils were easily identified in neighboring, nontransfected myocytes (Fig. 1C). Also, this staining procedure visualizes striations in myocytes subjected to pacing of contractions, phenylephrine, or transfection with the p38 activator MKK6 (11,15).

Effects of Ha-Ras
Because Ha-Ras V12 had such a profound effect on morphology, it was of interest to determine whether gene expression relating to regulation of [Ca 2ϩ ] i, such as SERCA2, would also be altered. Myocytes were therefore cotransfected with Ha-Ras V12 along with CMV-␤-galactosidase; the cells were then fixed and stained for ␤-galactosidase and endogenous SERCA2, which were visualized with Texas Red-or fluorescein isothiocyanate-conjugated secondary antibodies, respectively. Control myocytes exhibited SERCA2 staining that was equivalent to neighboring, nontransfected myocytes (Fig. 1, E and F). In contrast, myocytes transfected with Ha-Ras V12 exhibited weaker staining for SERCA2 (Fig. 1, G and H). Photomicrographs representing SERCA2 staining were scanned, and the intensity of the images were quantified via NIH Image. For nontransfected, control-transfected, and Ha-Ras V12 -transfected myocytes, pixel intensities averaged 18.4 Ϯ 0.7 (n ϭ 50), 19.3 Ϯ 0.8 (n ϭ 6), and 8.8 Ϯ 0.7 (n ϭ 13) (mean Ϯ S.E., arbitrary pixel intensity units), respectively, and the values for Ha-Ras V12 transfection were significantly different (p Ͻ 0.05, Student-Newman Keuls). Although pixel intensity and SERCA2 expression may not be strictly linearly related because of factors such as photobleaching, these data indicate that Ha-Ras V12 transfection diminishes SERCA2 expression.
The above results suggested that the regulation of [Ca 2ϩ ] i might be modified by Ha-Ras V12 . To test this hypothesis, myocytes were cotransfected with the GFP expression plasmid, along with the Ha-Ras V12 expression plasmid, then loaded with indo-1, and monitored for fluorescence during the electrical pacing of contractions. Myocytes transfected with 1 g of the Ha-Ras V12 expression plasmid exhibited statistically significant reductions in systolic indo-1 ratios (R sys ) and a trend toward prolongation of decay ( Fig. 2A and Table I). Transfection with 16 g of Ha-Ras V12 prolonged decay even more ( Fig.  2A and Table I) and, additionally, increased t peak , the time required for the calcium transients to reach their maximal values; transfection with 16 g of Ha-Ras V12 also significantly increased the diastolic (R dia ) indo-1 ratios. Virtually every myocyte transfected with Ha-Ras V12 (e.g. 19 of the 20 myocytes transfected with 16 g of Ha-Ras V12 in Fig. 2A) exhibited abnormal calcium transients, suggesting a high "cotransfection efficiency" for GFP and Ha-Ras V12 . These results are consistent with Ha-Ras V12 -mediated reduction of SERCA2, because reduced calcium removal from the cytoplasm might elevate R dia and prolong the decay phase of the calcium transient.
Effects of Ha-Ras V12S35 and Ha-Ras V12C40 on ERK Activation and [Ca 2ϩ ] i -Raf and phosphatidylinositol 3-OH kinase (PI 3-kinase) are among the best known Ras effectors in mammalian cells. To distinguish between these pathways, Ha-Ras V12 mutants have been identified that selectively couple to Raf (Ha-Ras V12S35 ) or PI 3-kinase (Ha-Ras V12C40 ) (16,17). As a first step toward evaluating the role of these effector pathways in the growth and [Ca 2ϩ ] i responses to Ha-Ras V12 , the differential coupling of Ha-Ras V12S35 and Ha-Ras V12C40 to the Raf-MEK-ERK pathway was verified; to do this, myocytes were cotrans-fected with hemaglutinin-ERK2 along with Ha-Ras V12 , Ha-Ras V12S35 , or Ha-Ras V12C40 , and in vitro kinase assays were performed on hemaglutinin-ERK2-specific immunoprecipitations. Cotransfection with Ha-Ras V12 resulted in a strong activation of ERK (17-fold), whereas transfection with Ha-Ras V12S35 yielded a moderate activation (2.5-fold). In contrast, transfection with Ha-Ras V12C40 did not activate ERK, at all (0.8-fold). Thus, although Ha-Ras V12S35 specifically activates the Raf-MEK-ERK pathway, it is a less effective stimulator of ERK activation than Ha-Ras V12 ; this result is consistent with other reports (17), and should be considered when interpreting the effects of these Ras mutants on cell growth and [Ca 2ϩ ] i .
Myocytes were transfected with equimolar amounts of Ha-Ras V12S35 or Ha-Ras V12C40 and analyzed for cell size and [Ca 2ϩ ] i . Ha-Ras V12S35 -transfected myocytes were larger than controls (1751 Ϯ 206 m, n ϭ 15) but were smaller than myocytes transfected with Ha-Ras V12 (see above). Ha-Ras V12C40 -transfected myocytes were similar in size to controls (1010 Ϯ 96 m, n ϭ 16). Transfection with Ha-Ras V12S35 modified [Ca 2ϩ ] i in a manner similar to the effects of Ha-Ras V12 , with statistically significant, pronounced prolongation of t decay and t peak (Fig. 2, C and D, and Table I). Transfection with Ha-Ras V12C40 also modified [Ca 2ϩ ] i , but the effect of Ha-Ras V12C40 was limited to the prolongation of transient decay, and the prolongation of decay was more modest than that elicited by Ha-Ras V12S35 . Thus, the relative effects of the Ras mutants on [Ca 2ϩ ] i (Ha-Ras V12 Ͼ Ha-Ras V12S35 Ͼ Ha-Ras V12C40 ) are consistent with the rank order of these mutants on ERK activation, suggesting that the Raf-MEK-ERK pathway may serve as a major route via which [Ca 2ϩ ] i may be altered by Ras activity. The data with Ha-Ras V12S35 further suggest that the PI 3-kinase pathway might also participate in the Ha-Ras V12 -mediated disregulation of [Ca 2ϩ ] i but most likely to a lesser degree than the Raf-MEK-ERK pathway.
Effects of BXB-Raf and MEKK1 COOH on [Ca 2ϩ ] i -Because Raf links Ras to ERK activation, it was of interest to test the effects of constitutively active Raf (BXB-Raf) on [Ca 2ϩ ] i . BXB-Raf transfection results in substantial activation of ERK and cell enlargement but does not increase myofilament expression or organization (11). BXB-Raf-transfected myocytes exhibited normal R dia but displayed reduced R sys , increased t peak , and prolonged decay (p Ͻ 0.05 for all parameters, data pooled from three separate cell preparations; Fig. 2, E and F, and Table I). The alterations of transient kinetics by BXB-Raf are consistent with the results obtained with Ha-Ras V12 and Ha-Ras V12S35 and show striking similarities to the contractile defects induced by pressure overload hypertrophy (5,6).
Additionally, transfection with MEKK1 COOH , a truncated version of MEKK1 that very strongly activates ERK in this cell type (11), also modified the calcium transients in a manner similar to Ha-Ras V12 and BXB-Raf, eliciting a trend toward increasing R dia , statistically significant reductions in R sys , and prolongations of t peak and t decay (transients not shown, but see Table I (Fig. 3, A and B, and Table I). The ability of dnMEK to antagonize the effects of Ha-Ras V12 is consistent with the hypothesis that Ha-Ras V12 alters [Ca 2ϩ ] i via the Raf-MEK-ERK pathway. Consistent with the relative position of Raf and MEK in the ERK activation cascade, cotransfection of dnMEK also strongly inhib- ited the effects of BXB-Raf on [Ca 2ϩ ] i (Fig. 3, C and D), returning the transients to normal kinetics.
Because incubation of cultured cardiac myocytes with protein kinase C-activating phorbol esters activates ERK (18,19), down-regulates SERCA2, and prolongs transient decay (7,8), it was of interest to test the ability of dnMEK to inhibit the effects of PMA on [Ca 2ϩ ] i . Consistent with previous studies, PMA treatment prolonged t peak and t decay (Fig. 3, E and F, and Table  I). Transfection with dnMEK significantly inhibited the effect of PMA on t peak and t decay (Fig. 3, E and F, and Table I). Thus, PMA, which activates the ␣-, ⑀-, and ␦-PKC isoforms in this cell type (20), may modify [Ca 2ϩ ] i primarily via activation of the Raf-MEK-ERK pathway.
Transfection with PKAC, a constitutively active PKC-␣ isoform (21), also strongly prolonged decay of the transients (transients not shown, but see Table I for descriptive statistics), and this effect was inhibited by the dominant-negative C4BRaf (22). PKAC also reduced R dia in a C4BRaf-sensitive manner. These results further suggest that PKC may modify cardiac [Ca 2ϩ ] i via the Raf pathway.

Expression of SERCA2 Inhibits the Effects of Ha-Ras V12 and BXB-Raf Transfection-Because transfection with
Ha-Ras V12 leads to down-regulation of SERCA2 expression and alteration of [Ca 2ϩ ] i , it was of interest to monitor [Ca 2ϩ ] i handling in myocytes cotransfected with expression vectors for both Ha-Ras V12 and for SERCA2. Coexpression of SERCA2 with Ha-Ras V12 resulted in partial but not complete restoration of the calcium transients (Fig. 4, A and B, and Table I). For example, cotransfection of SERCA2 did not counteract the effect of Ha-Ras V12 on R dia but significantly counteracted the effects of Ha-Ras V12 on R sys and t peak ; SERCA2 also exhibited a trend toward correcting the effect of Ha-Ras V12 on t decay . SERCA2 transfection, by itself, had no effect on cell morphology and did not alter the effect of Ha-Ras V12 on cell morphology. Cotransfection with SERCA2 also strongly corrected the effect of BXB-Raf on [Ca 2ϩ ] i (Fig. 4, C and D, and Table I). These results are consistent with the hypothesis that the Ras-Raf-MEK-ERK pathway may alter [Ca 2ϩ ] i , in part, by regulating SERCA2 expression. However, the inability of SERCA2 cotransfection to completely restore the alterations in [Ca 2ϩ ] i induced by Ha- ] i in cultured neonatal ventricular myocytes R dia represents the indo-1 ratio (405/485) monitored prior to the electrical stimulation for myocytes paced at 0.3 Hz. R sys represents the maximal indo-1 ratio attained by each cell. t peak represents the time between the start and the peak of the transient. t decay represents the half-time of decay, which was estimated by fitting the decay phase of each transient to a monoexponential function. t peak and t decay were further normalized to the corresponding mean control value for each experiment. Control t peak and t decay averaged 0.146 and 0.275 s, respectively, over nine cell preparations. For the data below, each value represents the average Ϯ S.E. for the number of cells given in parentheses. Statistical significance was evaluated using the Student-Newman-Keuls test.

Conditions
R dia R svs t peak t decay H-Ras V12 dose response (Fig 2, A and B (Fig. 2, C and D) Control (15) (Fig. 3, A and B) Control ( (Fig. 4, A and B) Control ( Ras V12 suggest that Ha-Ras V12 activity might also alter [Ca 2ϩ ] i by additional pathways that are independent of SERCA2.

DISCUSSION
In the present study cultured neonatal rat ventricular myocytes were transiently transfected with effectors of the Ras-Raf-MEK-ERK and related pathways, along with "transfection marker" constructs allowing the transfected myocytes to be identified and analyzed for cell size and morphology. Additionally, a novel technique was developed in which GFP was used as a transfection marker for living myocytes; this provided the opportunity to analyze the effects of the second messenger effectors on [Ca 2ϩ ] i regulation, a physiological parameter of extreme importance to cardiac function in health and disease. The data obtained are generally consistent with earlier reports demonstrating that microinjection of Ha-Ras V12 into cardiac myocytes elicits a strong growth response and hypertrophic gene expression (2), and the cardiac-targeted expression of Ha-Ras V12 in transgenic mice increases heart size concomitant with a decrease in myofibrillar organization (3,23). Indeed, in the present study, transfection with Ha-Ras V12 elicited an extreme increase in cell size along with loss of the contractile apparatus, a morphological phenotype resembling hypertrophic cardiomyopathy (24,25).
Transfection with Ha-Ras V12 also altered expression of SERCA2, a calcium-regulatory enzyme involved in calcium reuptake during the decay phase of the contractile calcium transients, which is also commonly down-regulated in whole animal models of pressure overload hypertrophy and failure (1,5,6). Indo-1 measurements of [Ca 2ϩ ] i further revealed that Ha-Ras V12 transfection severely compromised the magnitude of the contractile calcium transients and prolonged the decay phase of the transients, a calcium mobilization phenotype that is consistent with the effect of Ha-Ras V12 to diminish SERCA2. To investigate the potential effector pathways via which Ha-Ras V12 alters [Ca 2ϩ ] i , expression plasmids encoding Ras mutants that selectively couple to Raf-MEK-ERK (Ha-Ras V12S35 ) and PI 3-kinase (Ha-Ras V12C40 ) pathways were employed. Ha-Ras V12S35 altered [Ca 2ϩ ] i and cell size in a manner more consistent with Ha-Ras V12 than did Ha-Ras V12C40 , suggesting that the Raf-MEK-ERK pathway is more involved in the growth and calcium disregulation responses to Ha-Ras V12 than PI 3-kinase. The effect of Ha-Ras V12S35 on [Ca 2ϩ ] i and cell size was less than that of Ha-Ras V12 , however, indicating that Ras effectors other than Raf may also play a role in these responses. Another factor to be considered, though, is the degree to which Ha-Ras V12 and Ha-Ras V12S35 activate ERK. Although Ha-Ras V12S35 is selective for the ERK pathway, it is not as effective a stimulator of ERK as Ha-Ras V12 . Therefore the lesser effect of Ha-Ras V12S35 on [Ca 2ϩ ] i compared with Ha-Ras V12 might relate to the lesser stimulatory effect of Ha-Ras V12S35 on ERK activity.
Transfection with a constitutively active Raf construct (BXB-Raf) also strongly modified [Ca 2ϩ ] i , generally reducing the magnitude of the transients, slowing the time to reach the peak magnitude, and strongly prolonging transient decay, which are alterations in [Ca 2ϩ ] i handling commonly observed in myocytes isolated from hypertrophic and failing hearts (5,6). Notably, the effects of both Ha-Ras V12 and BXB-Raf were strongly antagonized by cotransfection with dnMEK, further implicating the Raf-MEK-ERK pathway as a possible route for the disregulation of cardiac [Ca 2ϩ ] i . On the other hand, cotransfection with dnMEK did not fully correct the effects of Ha-Ras V12 on [Ca 2ϩ ] i ; this suggests that Ras effectors other than Raf may also modify [Ca 2ϩ ] i , a hypothesis that will be the subject of further investigation.
Importantly, transfection with dnMEK also inhibited the effect of the PKC-activating phorbol ester, PMA, on [Ca 2ϩ ] i . Additionally, cotransfection with a dominant-negative Raf construct (C4BRaf) inhibited the ability of a constitutively active protein kinase C construct (PKAC) to alter [Ca 2ϩ ] i . These data indicate that the Raf-MEK-ERK cascade represents a point of convergence via which hypertrophic stimuli that activate Ras and/or PKC lead to diminishment and prolongation of the cardiac contractile calcium transients. PKC might activate Raf by directly phosphorylating it (10,26), or, alternatively, it has been proposed that PKC can activate Ras (27), which would also subsequently lead to Raf and ERK activation. PMA is also known to have hypertrophic effects on cardiac myocytes (28) and to stimulate the disorganization and breakdown of the contractile apparatus in cultured adult rat myocytes (29) resulting in a morphological phenotype similar to the effects of Ha-Ras V12 transfection.
Previous studies have demonstrated two mechanisms via which ERK activation might regulate SERCA2 expression. Ha-Ras V12 transfection leads to inhibition of the rat SERCA2 promoter (30), suggesting a control of SERCA2 expression at the transcriptional level. Additionally, SERCA2 mRNA stability is reduced by phorbol esters (8). PMA, in conjunction with ionomycin, also down-regulates other SERCA isoforms in lymphocytes (31), suggesting that the regulation of SERCA by PKC might be a general strategy for [Ca 2ϩ ] i regulation in both cardiac and noncardiac cell types. The observation in the present study that cotransfection with SERCA2 tends to compensate for the effects of Ha-Ras V12 and BXBRaf on [Ca 2ϩ ] i is consistent with previous studies, demonstrating that infection of cultured neonatal ventricular myocytes with a SERCA2expressing adenovirus acts to correct the effects of PMA on [Ca 2ϩ ] i (32).
SERCA2 down-regulation may not be the only means by which the [Ca 2ϩ ] i is altered in hypertrophic myocytes. For example, prolongation of relaxation occurs in pressure overload rats prior to the reduction in SERCA2 levels (33), and calcium transients may be reduced in cardiac myocytes obtained from hypertrophic hearts under conditions in which the SR is well loaded with calcium (34). It has been suggested that impaired calcium-induced calcium release occurs in hypertrophic myocytes because of increased distance between voltage-dependent calcium channels and ryanodine receptors at the T-tubule-SR dyads (34). Because Ha-Ras V12 transfection down-regulates myofibrillar expression and the SR is closely associated with myofibrils (35), there is a strong possibility Ha-Ras V12 transfection might alter SR structure and function. Further analysis of cellular ultrastructure will be needed to resolve this question.
Because excessive ERK activation may prove deleterious to the maintenance of the cardiac phenotype, it is interesting to consider the role of the related MAPKs JNK and p38 in cardiac gene expression. Recent results suggest that JNK activity in cultured neonatal ventricular myocytes may be regulated by [Ca 2ϩ ] i and contractile activity (19) and that contractile activity and JNK activation represent positive regulators of cell growth and the contractile apparatus (15,36). Transfection with constitutively active upstream regulators of p38 also dramatically increases cell growth, myofibril expression, and organization (11,37) in a manner consistent with the compensatory phase of cardiac hypertrophy. Thus, MAPK family members may participate in the development and maintenance of the cardiac phenotype, and altered activity of these enzymes may lead to the development of both compensatory or potentially pathological changes in morphology and physiology. The striking ability of dnMEK to inhibit the effects of both Ha-Ras V12 and PMA on cardiac [Ca 2ϩ ] i suggests that blockade of the Raf-MEK-ERK cascade via gene transfer or other techniques could conceivably be an effective therapeutic strategy in treating the contractile defects associated with cardiac hypertrophy and failure.