Oncogenic src, raf, and rasStimulate a Hypertrophic Pattern of Gene Expression and Increase Cell Size in Neonatal Rat Ventricular Myocytes*

In response to hormones and growth factors, cultured neonatal ventricular myocytes increase in profile, exhibit myofibrillogenesis, and re-express genes whose expression is normally restricted to the fetal stage of ventricular development. These include atrial natriuretic factor (ANF), β-myosin heavy chain (β-MHC), and skeletal muscle (SkM)-α-actin. By using luciferase reporter plasmids, we examined whether oncogenes that activate the extracellular signal-regulated kinase cascade (src F527, Ha-ras V12, and v-raf) increased expression of “fetal” genes. Transfection of myocytes withsrc F527 stimulated expression of ANF, SkM-α-actin, and β-MHC by 62-, 6.7-, and 50-fold, respectively, but did not induce DNA synthesis. Stimulation of ANF expression bysrc F527 was greater than by Ha-ras V12, which in turn was greater than by v-raf. General gene expression was also increased but to a lesser extent. The response to src F527 was inhibited by dominant-negative Ha-ras N17. Myocyte area was increased by src F527, Ha-ras V12, and v-raf, and although it altered myocyte morphology by causing a pseudopodial appearance,src F527 did not detectably increase myofibrillogenesis either alone or in combination with Ha-ras V12. A kinase-dead src mutant increased myocyte size to a much lesser extent thansrc F527 and also did not inhibit ANF-luciferase expression in response to phenylephrine. We conclude that members of the Src family of tyrosine kinases may be important in mediating the transcriptional changes occurring during cardiac myocyte hypertrophy and that Ras and Raf may be downstream effectors.

In response to hormones and growth factors, cultured neonatal ventricular myocytes increase in profile, exhibit myofibrillogenesis, and re-express genes whose expression is normally restricted to the fetal stage of ventricular development. These include atrial natriuretic factor (ANF), ␤-myosin heavy chain (␤-MHC), and skeletal muscle (SkM)-␣-actin. By using luciferase reporter plasmids, we examined whether oncogenes that activate the extracellular signal-regulated kinase cascade (src F527 , Ha-ras V12 , and v-raf) increased expression of "fetal" genes. Transfection of myocytes with src F527 stimulated expression of ANF, SkM-␣-actin, and ␤-MHC by 62-, 6.7-, and 50-fold, respectively, but did not induce DNA synthesis. Stimulation of ANF expression by src F527 was greater than by Ha-ras V12 , which in turn was greater than by v-raf. General gene expression was also increased but to a lesser extent. The response to src F527 was inhibited by dominant-negative Ha-ras N17 . Myocyte area was increased by src F527 , Ha-ras V12 , and v-raf, and although it altered myocyte morphology by causing a pseudopodial appearance, src F527 did not detectably increase myofibrillogenesis either alone or in combination with Ha-ras V12 . A kinase-dead src mutant increased myocyte size to a much lesser extent than src F527 and also did not inhibit ANF-luciferase expression in response to phenylephrine. We conclude that members of the Src family of tyrosine kinases may be important in mediating the transcriptional changes occurring during cardiac myocyte hypertrophy and that Ras and Raf may be downstream effectors.
The cardiac myocyte is a terminally differentiated cell that withdraws from cell division at around birth in mammals. In these cells, adaptive hypertrophic growth in response to an increased requirement for contractile power involves increases in cell size, protein content, and myofibrillar organization (reviewed in Ref. 1). There are also changes in gene transcription that distinguish hypertrophy from maturational growth (eutrophy). In response to hypertrophic stimuli, cultured neonatal rat ventricular myocytes initiate a rapid and transient increase in expression of immediate early genes (c-jun, c-fos, and Egr-1), followed by the re-expression of so-called "fetal" genes including atrial natriuretic factor (ANF), 1 ␤-myosin heavy chain (␤-MHC), and skeletal muscle (SkM) ␣-actin. More chronic exposure to hypertrophic agonists also elicits an increase in the expression of constitutively expressed contractile protein genes such as ventricular myosin light chain-2 and cardiac muscle ␣-actin. Many of these changes are seen during hypertrophy of the rat heart in vivo, providing a justification for the use of this model system (reviewed in Ref. 1).
A wide variety of agonists induce the hypertrophic phenotype in cultured ventricular myocytes. These include agonists acting through G protein-coupled receptors such as endothelin-1 (2, 3), angiotensin II (4,5), and ␣ 1 -adrenergic agonists (4,6,7). Other hypertrophic stimuli include passive stretch (8 -10), phorbol esters that activate protein kinase C isoforms (11,12), and growth factors acting through receptor protein tyrosine kinases such as fibroblast growth factors (13) and insulin-like growth factor 1 (14,15). A feature common to each of these stimuli is their ability to activate the extracellular-regulated kinase (ERK) members of the mitogen-activated protein kinase (MAPK) superfamily (16 -22). This has prompted the suggestion that activation of ERKs may be a central component of the intracellular signaling mechanism through which hypertrophic agonists exert their effects (16,17,23), although it is still not clear whether ERK activation is obligatory or sufficient (19, 24 -26).
Activation of ERK1 (p44 MAPK) and ERK2 (p42 MAPK) is brought about by phosphorylation by MAPK kinases (MEKs, for MAPK (or ERK) kinases) (reviewed in Refs. [27][28][29][30][31]. MEKs are themselves activated by phosphorylation by MAPK kinase kinases, the best-characterized being the Raf family (c-Raf-1, A-Raf, and B-Raf in higher organisms) (32)(33)(34)(35). For agonists acting through receptor protein tyrosine kinases, the intervening steps between receptor occupancy and activation of the Raf 3 MEK 3 ERK cascade are relatively well established and involve SH2/SH3-containing adaptor proteins such as Grb2 and Shc, guanine-nucleotide exchange factors such as Sos, and activation through GDP/GTP exchange of the small G proteins of the Ras family (reviewed in Refs. 36 -38). Ras⅐GTP mediates the activation of c-Raf-1 by recruiting it to the plasma membrane (39 -42), where other reactions that possibly include the phosphorylation of Tyr-340/Tyr-341 by Src family protein tyro-sine kinases lead to its full activation (43,44). Recent evidence shows that A-Raf is also activated synergistically by oncogenic ras and oncogenic src, whereas oncogenic ras alone is sufficient for maximal activation of B-Raf (44). In terms of the hypertrophic response, there is good evidence for a role for activated Ras. Thus, angiotensin II activates Ras in myocytes (45); injection of oncogenic Ras V12 initiates hypertrophic changes in myocytes (46), and PE-induced ANF expression is inhibited by a dominant-negative ras mutant (46).
Recent work has suggested that protein tyrosine kinases of the Src family (Src, Fyn, Lyn, etc.) may participate in the coupling of both receptor protein tyrosine kinases and G protein-coupled receptors to ERK activation (reviewed in Refs. [47][48][49][50][51]. Indeed, a role for Fyn has been proposed in angiotensin II-induced hypertrophy of the cardiac myocyte (45). Here we have examined the effect of overexpression of src F527 , an oncogenic variant of the c-src protooncogene, on the transcriptional and morphological characteristics associated with the hypertrophic phenotype in cardiac myocytes. We show that src F527 expression potently activates genes that are up-regulated in response to hypertrophic stimuli, a response in which it may act in concert with Ras and Raf. Our results suggest that non-receptor protein tyrosine kinases such as c-Src may play a role in some of the altered growth responses of myocytes exposed to hypertrophic stimuli.

EXPERIMENTAL PROCEDURES
Materials-Sprague-Dawley rats were bred within the National Heart and Lung Institute. Culture medium and other reagents were from Sigma, Life Technologies, Inc., or Merck. The plasmids encoding constitutively active v-Raf and Ha-Ras V12 and dominant-negative (DN) Ha-Ras N17 and DNRaf were from Professor C. J. Marshall, and those encoding chicken Src F527 and kinase-dead mutants of c-Src (srcKϪ) and Src F527 (src F527 KϪ) were from Dr. R. M. Marais (both at the Chester Beatty Laboratories, Institute of Cancer Research, London, UK). In Src F527 the tyrosine residue at position 527 is mutated to phenylalanine rendering it insensitive to phosphorylation and inactivation by the tyrosine kinase Csk. The ANF-luciferase (ANF-LUX) reporter construct pANF(Ϫ638)L⌬5Ј (7) and pON249 (52) in which ␤-gal is expressed from a constitutive cytomegalovirus promoter were kindly provided by Dr. K. R. Chien (Department of Medicine, University of California, San Diego). The AP-1-LUX construct TRE2PRL(Ϫ36) was a gift from Drs. J. H. Brown (Department of Pharmacology, University of California, San Diego) and M. G. Rosenfeld (Howard Hughes Medical Institute, University of California, San Diego). The LUX reporter constructs for ␤-MHC (53), SkM-␣-actin (53)(54)(55), and the c-fos serum responsive element (SRE) (56) were gifts from Dr. M. D. Schneider (Molecular Cardiology Unit, Baylor College of Medicine, Houston, TX). The LUX reporter plasmids have been described in detail previously (57). Plasmid pNG1 was constructed by subcloning a BamHI-NotI fragment of pPD46.21 (58) (a gift from Dr. A. Fire, Carnegie Institute, Baltimore, MD) encoding a ␤-gal gene containing a nuclear localization signal into the BamHI-NotI sites in pBK-cytomegalovirus (Stratagene). Plasmids were purified by polyethylene glycol precipitation (59).
Transient Transfection of Cultured Neonatal Ventricular Myocytes-Ventricular myocytes were isolated from the hearts of 1-2-day-old rats by a modification of the method of Iwaki et al. (60) as detailed previously (57). Transfections (by the calcium phosphate method) involved the addition of 5 or 15 g of LUX reporter plasmid, 2 or 4 g of pON249, and a total of up to 10 g of test plasmids to 60-mm cell culture dishes containing 1 million cells. For sarcomeric staining, transfections involved the addition of 6 g of ANF-LUX reporter plasmid, 1.6 g of pNG1, and 2 g each of test plasmids to 100,000 cells per well of a 2-well chamber slide. Phenylephrine (PE), when present, was added to a concentration of 10 -100 M (as stated in individual experiments) approximately 20 h after the transfection. After a further 48 h, myocytes were extracted and assayed for LUX and ␤-gal as detailed previously (57). Results are presented as the mean Ϯ S.E. of experiments on at least four separate preparations of myocytes.
Determination of Myocyte Size-To assess the area of transfected myocytes, cells were washed three times with ice-cold Dulbecco's Ca 2ϩ / Mg 2ϩ -free phosphate-buffered saline (PBS), fixed with 4% formaldehyde in PBS for 10 min, re-washed three times with PBS, and then stained with 0.2 mg/ml 5-bromo-4-chloro-3-indolyl-␤-D-galactopyrano-side, 5 mM K 4 Fe(CN) 6 , 5 mM K 3 Fe(CN) 6 , 2 mM MgCl 2 in PBS. Transfected (blue) cells were randomly selected from all areas of the dishes and imaged using a video hardcopier from which the cell area was determined by planimetry.
Expression of Results and Statistical Analysis-The absolute values of chemiluminescence differs considerably between the various reporter constructs. Thus, in Figs. 1-3, results are expressed as the ratio of reporter gene expression in the presence of src F527 , v-raf, and Ha-ras V12 expression plasmids relative to empty vector controls. This facilitates comparison of the relative potencies of the expression plasmids. Results are presented as means Ϯ S.E. For statistical analysis, absolute values of chemiluminescence were used, and significance was assessed by using an unpaired or paired Student's t test as appropriate with a significant difference taken as being established at p Ͻ 0.05. V12 , and v-raf-The proximal 638 base pairs of the ANF promoter are sufficient to confer tissue specificity and inducible expression on the LUX reporter (7). Transfection with src F527 , a constitutively active oncogenic mutant of c-src, stimulated expression of this construct by 62.3 Ϯ 10.3-fold ( Fig. 1, panel A). This was significantly greater (p Ͻ 0.01) than the stimulation of ANF-LUX by oncogenic Ha-ras V12 (18.0 Ϯ 2.7-fold). Co-transfection with src F527 and Ha-ras V12 stimulated ANF-LUX expression by 106.1 Ϯ 21.0-fold which was significantly greater than transfection with Ha-ras V12 alone (p Ͻ 0.01) but was not significantly greater than transfection with src F527 alone ( Fig. 1, panel A). When LUX expression was normalized for ␤-gal expression ( Fig. 1, panel B), the effect of src F527 (7.2 Ϯ 0.8-fold) was still significantly greater (p Ͻ 0.02) than activation by Ha-ras V12 (3.7 Ϯ 0.7-fold). Again, the combined effect of src F527 and Ha-ras V12 (8.3 Ϯ 1.3-fold) was significantly greater than that of Ha-ras V12 alone (p Ͻ 0.025) but was not significantly greater than that of src F527 alone. Since transfection efficiency using the calcium phosphate method is routinely about 2% and independent of the plasmids transfected (57), the reduction in the magnitude of the responses when normalized to ␤-gal is a consequence of a global stimulation of gene expression. This is manifest by the activation of ␤-gal expression despite it being driven by the constitutive cytomegalovirus promoter. Thus, the fold induction of ␤-gal was 4.7 Ϯ 1.1, 7.7 Ϯ 1.7, and 11.6 Ϯ 3.0 for Ha-ras V12 , src F527 , and Ha-ras V12 plus src F527 -transfected cells, respectively. src F527 is therefore a more potent activator of ANF expression than is Ha-ras V12 and, like oncogenic ras (53), src F527 also has a powerful effect on gene expression in general. Similar conclusions can be drawn from experiments using a LUX reporter responsive to the activation of the AP-1 transcription factor complex ( Fig. 1, panels C and D). This reporter has very low basal activity in cardiac myocytes compared with the ANF-LUX transgene, accounting for the greater variability in stimulation when expressed as fold induction by src F527 and Ha-ras V12 .

Stimulation of Promoter Activities by src F527 , Ha-ras
The specific activation of the ANF and AP-1 reporter genes by src F527 and Ha-ras V12 shown here contrasts with the reported (53) lack of a specific effect of ras R12 on other reporter genes (SkM-␣-actin, cardiac-␣-actin, and ␤-MHC) whose upregulation is also associated with the hypertrophic phenotype. To determine whether this was a peculiarity of the ANF and AP-1 transgenes, the effects of src F527 and Ha-ras V12 were extended to include some of these alternative marker genes of hypertrophy (Fig. 2). The SkM-␣-actin-LUX transgene was upregulated 7.8-fold by Ha-ras V12 and 6.7-fold by src F527 (Fig. 2,  panel A). However, there was only a slight (1.9-fold) activation of the SkM-␣-actin-LUX/␤-gal ratio (Fig. 2, panel B) indicating that SkM-␣-actin transcription was not dramatically activated above the general increase in transcription. By using a c-fos-SRE transgene, a similar conclusion was reached for the effect of Ha-ras V12 (2.3-fold activation), but this transgene was specifically activated by src F527 (9.0-fold activation, see Fig. 2, panels C and D). In contrast, the ␤-MHC transgene was specifically activated by both Ha-ras V12 (6.5-fold) and src F527 (15.9fold, see Fig. 2, panels E and F). Of possible relevance to these findings, we found that the sensitivity of the SkM-␣-actin, c-fos-SRE and ␤-MHC transgenes to activation by src F527 (and by Ha-ras V12 ) was inversely related to their basal activities. Thus, although the absolute LUX activity (arbitrary light units in thousands) in the presence of src F527 differed by less than 2-fold (142 Ϯ 33, 201 Ϯ 58, and 115 Ϯ 25 for transfections with SkM-␣-actin, c-fos-SRE and ␤-MHC, respectively), the basal activities were 21.4 Ϯ 1.0, 4.6 Ϯ 0.6, and 2.3 Ϯ 0.3 arbitrary light units in thousands, respectively. For ANF-LUX and AP-1-LUX, the basal activities (in thousands of arbitrary light units) were 3.6 Ϯ 1.4 and 0.9 Ϯ 0.5, respectively, and following src F527 stimulation were 178 Ϯ 34 and 308 Ϯ 113, respectively.
The effect of v-raf on promoter activity is also shown in Fig.  2. Expression of v-raf significantly stimulated promoter activity for all three promoters tested, and the effect of v-raf was similar to but slightly lower than that of Ha-ras V12 , in keeping with its role as a downstream effector of Ras (reviewed in Ref. 42). To confirm that the relative order of potency of the three oncogenes (src F527 Ͼ Ha-ras V12 Ն v-raf) was not an artifact of the amount of plasmid transfected, the response of ANF-LUX expression was examined over a range of oncogene concentrations. Results from four separate preparations expressed as fold induction of ANF-LUX/␤-gal in response to 0.3, 1, 3, and 10 g of plasmid, respectively, were as follows: for src F527 , 8.0 Ϯ 1.1, 8.0 Ϯ 0.4, 10.0 Ϯ 0.9, and 10.4 Ϯ 1.5; for Ha-ras V12 , 3.7 Ϯ 1.2, 3.5 Ϯ 1.5, 4.5 Ϯ 1.6, and 3.4 Ϯ 0.9; and for v-raf, 2.0 Ϯ 0.5, 2.0 Ϯ 0.5, 2.7 Ϯ 0.7, and 2.5 Ϯ 0.7. Thus, increasing the amounts of oncogene transfected over the range of 0.3 to 10 g of DNA had no effect on the extent of ANF-LUX induction. The absolute values of LUX observed for any one of the oncogenes also differed by less than 2-fold over this concentration range. This suggests that the amount of oncogene expressed in each case was sufficient to give a maximum response and that the relative potencies of the oncogenes relates to the efficacy with which they stimulate transcription.

Oncogenic src F527 Does Not Induce DNA Synthesis in Cardiac
Myocytes-To determine whether the ability of src F527 to induce strong activation of gene expression (Figs. 1 and 2) might reflect a capacity to re-initiate DNA synthesis, myocytes were transfected with src F527 and subsequently analyzed for BrdUrd incorporation into nuclei. Myocytes transfected with src F527 were identified by staining for co-transfected ␤-gal and did not show any nuclear staining for BrdUrd (results not shown). Thus, the strong effects of src F527 on gene expression are transcriptional effects and are not the result of changes in DNA synthesis.
Inhibition of src F527 -stimulated Gene Expression by Dominant-negative ras-The relative order of potency of src F527 , Ha-ras V12 , and v-raf is in accord with observations in other cell types that the transforming activity of Src is dependent upon Ras (61) and that of Ras on c-Raf-1 (62). To investigate whether the increase in cardiac gene expression in response to src F527 requires functional Ras and/or Raf, the effects of DN Ha-ras N17 and DNraf were determined. In this series of experiments ANF-LUX expression was increased about 4-fold by src F527 , and this was inhibited by co-transfection with Ha-ras N17 (Fig.  3). DNraf had a modest but non-significant inhibitory effect (35.9 Ϯ 8.3%) on src F527 -induced ANF-LUX/␤-gal expression and did not significantly potentiate the inhibition by Ha-ras N17 (Fig. 3). However, the results with the DNraf construct were complicated by its effect on ␤-gal expression. Transfection with DNraf inhibited src F527 -induced ANF-LUX expression by 50.2 Ϯ 5.2% but also inhibited ␤-gal expression in the presence of src F527 by 36.7 Ϯ 5.6% in these same transfections. Similarly, the DNraf construct reduced ␤-gal expression in the presence of Ha-ras V12 and in the presence of 0.1 mM PE by 20.2 Ϯ 2.5 and 71.0 Ϯ 1.3%, respectively (n ϭ 4 separate preparations). Thus, the DNraf construct has a predominantly general effect to depress gene expression and which masks the effects of DNraf on PE-, Ha-ras V12 -, and src F527 -induced ANF-LUX expression. These results suggest that transmission of the signal from src F527 to ANF-LUX expression is mediated through Ras but that Raf may not be the only downstream Ras effector through which the signal is propagated.
The Effect of Genistein on ANF-LUX Expression-It has previously been reported that activation of Ras and induction of ANF expression in cardiac myocytes by PE can be blocked by the tyrosine kinase inhibitor genistein (63). Surprisingly, we found that genistein (20 M) enhanced ANF-LUX/␤-gal expression in response to src F527 (Fig. 4, panel A). However, genistein also increased basal ANF-LUX activity to a similar extent in these experiments such that fold induction of ANF-LUX/␤-gal by src F527 was unaltered by genistein when expressed relative to the genistein control (Fig. 4, panel B). In the light of these unexpected results, we re-examined the effects of genistein on PE-and TPA-induced ANF-LUX expression. Genistein had no effect on the fold activation of ANF-LUX/␤-gal by PE or TPA when expressed relative to the Me 2 SO control (Fig. 4, panel A). Again, because of its effect on basal ANF-LUX/␤-gal expression, when the results were expressed relative to the matched control, genistein appeared to inhibit PE-and TPA-induced ANF-LUX/␤-gal expression (Fig. 4, panel B). Thus, interpretation of these results is dependent upon the manner in which they are expressed. A possible explanation for the altered responses to TPA and PE in the presence of genistein is that under basal conditions c-Src is kept in an inactive conformation through phosphorylation of tyrosine 527 (in chicken c-Src) in the C-terminal tail by Csk (47,64). If the tyrosine kinase activity of Csk is more sensitive to inhibition by genistein than is c-Src itself, the net effect of genistein could result in an increase in c-Src activity.
The Effect of src F527 , Ha-ras V12 , and v-raf on Myocyte Size-As well as characteristic alterations in transcription, hypertrophied cardiomyocytes are larger and more regular in shape than control cells (2,6,7,25,60,65). In separate experiments, transfection of myocytes with src F527 , Ha-ras V12 , or v-raf significantly increased the size of transfected cells by 68, 129, and 62%, respectively, but there was no additional stimulation in the presence of both Ha-ras V12 and src F527 (Fig. 5). The increases in cell size induced by these oncogenes were as great as that elicited by maximally effective concentrations of PE (57). To determine whether the increase in cell size by src F527 is dependent on its kinase activity, the effect of a kinase dead src F527 mutant (src F527 KϪ) on myocyte size was determined (Fig. 5, Expt. 3). Whereas src F527 increased myocyte area by 74.8% (p Ͻ 0.001), src F527 KϪ only increased myocyte area by 22.5% (p Ͻ 0.001 versus src F527 ), although this was still significantly greater than the controls (p Ͻ 0.02). It is unlikely that this is due to any residual kinase activity as src F527 KϪ was completely ineffective in stimulating ANF-LUX activity (0.8 Ϯ 0.13-fold of control compared with 11.8 Ϯ 0.48-fold of control for src F527 , n ϭ 4). Thus, the ability of src F527 to induce an increase in cell size is largely dependent on its kinase activity, although there is a small kinase-independent component that may reflect an ability of src F527 (and src F527 KϪ) to induce cell spreading.
The Effect of src F527 on Sarcomeric Organization-Another feature characteristic of myocytes exposed to hypertrophic agonists is increased assembly and organization of contractile proteins into sarcomeric units (2,25,60). To examine the effects of src F527 on sarcomerogenesis, myocytes were co-transfected with pNG1, a vector that targets ␤-gal to the nucleus. This helps to prevent problems with cross-reactivity which can arise when staining proteins in the same intracellular compartment with two primary mouse antibodies. Myocytes were subsequently stained for ␤-gal to identify transfected cells (results not shown) and with ␤-MHC to highlight the contractile apparatus (Fig. 6). Transfection with empty vector had no effect on sarcomeric organization (Fig. 6, panel A; compare the transfected cell (see arrow) to the surrounding non-transfected cells). Likewise, cells transfected with src F527 did not display increased organization of the contractile apparatus into sarcomeric structures (Fig. 6, panel B), even though the transfected cell is morphologically distinct with a more clearly defined outline and outgrowth of extended processes/pseudopodia. These probably represent points of contact with the substratum. A combination of src F527 and Ha-ras V12 enhanced this definition of shape (Fig. 6, panel C), but despite the clear change in size and appearance, there was little evidence of increased sarcomerogenesis when compared with the organization of the myofibrils seen in response to PE (Fig. 6, panel D).
Myocytes treated with PE also displayed more defined points of anchorage to the dish.
PE-induced ANF-LUX Is Not Inhibited by Kinase-dead src-In an attempt to determine whether c-Src may be a component of the intracellular signaling pathway through which PE exerts its hypertrophic transcriptional responses, myocytes were transfected with kinase-dead wild type c-src (srcKϪ) or src F527 KϪ, and its effect on PE-induced ANF-LUX was determined. Transfection with srcKϪ did not affect ANF-LUX/␤-gal expression in response to 10 -100 M PE (fold induction by PE FIG . 5. Effects of src F527 , v-raf, and Ha-ras V12 expression on myocyte size. In experiments 1 and 2, neonatal cardiac myocytes were transfected with ANF-LUX (15 g/dish), pON249 (4 g/dish), and 5 g of v-raf, Ha-ras V12 , src F527 , or their backbone vector control as described under "Experimental Procedures." In experiment 3 ANF-LUX was omitted, and pON249 and src F527 (or pEF or src F527 KϪ) were used at 3 g/dish each. After a further 48 h the cells were fixed and stained with 5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside, and cell area was in the absence and presence of 3-20 g of srcKϪ was 14.9 Ϯ 3.5 and 20.5 Ϯ 7.1, respectively, n ϭ 7). Similar results were obtained using the src F527 KϪ construct; fold induction of ANF-LUX/␤-gal by 10 M PE was 11.6 Ϯ 3.3 and 18.8 Ϯ 4.9 in the absence and presence of 3 g of src F527 KϪ, respectively (n ϭ 4). In these experiments ␤-gal expression in the presence of src F527 KϪ was reduced to 50.5 Ϯ 11.2% which accounted for the apparent stimulation of ANF-LUX/␤-gal expression. These results suggest that c-Src may not be an essential component of the pathway through which PE transduces its transcriptional responses. However, these conclusions remain to be substantiated because the efficacy with which these kinase-dead Src mutants can act to inhibit the function of endogenous c-Src is not known. 2

DISCUSSION
The Src family of non-receptor protein tyrosine kinases is activated by a wide variety of agents (growth factors, cytokines, G-protein-coupled receptor agonists, UV light, etc.) and is intimately connected with cell growth (reviewed in Refs. 47-51 and 66). This is reflected by the ability of oncogenic src to transform cells (61), for which there is a functional dependence on Ras (61). Similarly, the transforming ability of oncogenic ras is dependent on c-Raf-1 (62), and unregulated c-Raf-1 protein kinase activity is itself oncogenic (67). The function of all three protooncogenes is closely interconnected because, although Src and Ras can independently activate c-Raf-1, both are required for full c-Raf-1 activation (43,44,68). The same is true for activation of A-Raf but not B-Raf which can be fully activated by oncogenic Ras alone (44). One view is that Ras is required to recruit c-Raf-1 to the plasma membrane where it can be activated by phosphorylation on tyrosine residues 340 and 341 by membrane-bound Src (42)(43)(44)69). B-Raf lacks the tyrosine residues in c-Raf-1 and A-Raf which are phosphorylated by Src, which may explain its lack of sensitivity to activation by Src (44). However, since the activation of c-Raf-1 autokinase activity by v-src is not inhibited by expression of dominant-negative Ha-ras N17 , v-src may at least partially activate c-Raf-1 via a Ras-independent pathway (68).
The importance of Ras activation in the hypertrophic response has been demonstrated by the ability of oncogenic Ha-Ras V12 protein to induce the characteristic morphological and transcriptional changes in myocytes and for dominant-negative Ha-ras A15 to inhibit the hypertrophic response to PE (46). Oncogenic raf has also been shown to induce the characteristic transcriptional changes in myocytes, although not morphological changes associated with hypertrophy (70). Here we have compared the effects of src F527 , Ha-ras V12 , and v-raf with respect to their ability to initiate a gene program associated with hypertrophy ( Figs. 1 and 2). The order of potency observed (src F527 Ͼ Ha-ras V12 Ͼ v-raf) is in accord with their functional hierarchy as described above and which is reinforced by the observation that Ha-ras N17 inhibits activation of ANF-LUX expression by src F527 (Fig. 3). DNraf had a lesser specific effect on the response of ANF-LUX to src F527 , suggesting that c-Raf-1 may not be the only downstream effector that can transduce the effects of activated Ras. Other possible Ras effectors include MEK kinase-1, phosphatidylinositol-3-kinase, p120 GAP , Ral-GDS and protein kinase C (reviewed in Ref. 71). Of these other potential candidates, MEK kinase-1 and protein kinase C have been shown to induce ANF-LUX when constitutively active forms are overexpressed in cardiac myocytes (72)(73)(74), and Ras-GAP (a GTPase-activating protein for Ras) has also recently been proposed to have an effector-like function in these cells (75). Observations in fibroblasts that oncogenic src associates with and phosphorylates Ras-GAP suggests an ad-ditional level of complexity to the functional interaction of Src with Ras (76).
As well as these specific transcriptional effects associated with a hypertrophic response, src F527 and Ha-ras V12 also stimulated general or global gene expression, as previously reported for ras R12 (53). In partial agreement with that study (53), we only detected a 1.6-fold induction of SkM-␣-actin by Ha-ras V12 when corrected for ␤-gal expression (Fig. 2, panel B), although we observed a 6.5-fold activation of ␤-MHC under these conditions (Fig. 2, panel F). This discrepancy may be due in part to the different oncogenic ras constructs used or to differences in culture conditions because the stimulation of SkM-␣-actin-LUX and ␤-MHC-LUX reported here was much greater than the effect reported for the Arg-12 Ras construct using the same LUX reporters (53). The ability of both oncogenic src and ras to stimulate general gene transcription after serum withdrawal probably reflects their ability to substitute for growth factors, the responses to which are dependent on the activation of the cellular counterparts of these oncogenes.
The most striking features of myocyte morphology evoked by powerful hypertrophic agonists such as endothelin-1 and PE are an increase in cell size and the organization of the contractile apparatus into sarcomeric structures (Fig. 6, panel D, and Refs. 2,25,60). In common with oncogenic ras or raf, src F527 increased myocyte size, but the effects of Ha-ras V12 and src F527 were not additive (Fig. 5). This may reflect a limit on the size that these cells can attain under the conditions in which they are cultured. Despite its potent effects on gene transcription and cell size, src F527 , either alone or in combination with Haras V12 , did not detectably induce sarcomeric organization (Fig.  6, panels B and C). This result differs from the reported similarity in the pattern of ventricular myosin light chain-2 organization in myocytes microinjected with Ras oncoprotein and myocytes treated with PE (46). However, this could be due to a difference between injection of the Ras oncoprotein and transfection with a Ha-ras V12 expression plasmid because we have consistently failed to observe effects of transfected Ha-ras V12 on sarcomeric structural organization. 3 One clear morphological effect of transfection of myocytes with src F527 or with src F527 plus Ha-ras V12 is the increased definition of the cell periphery and in particular the contact points with the substratum. A possible explanation for this phenomenon is that c-Src is known to associate with the plasma membrane and especially with focal adhesions (77). Here, it may phosphorylate cytoskeletal proteins, several of which have been described as Src substrates including the focal adhesion kinase p125 FAK (78,79), paxillin (80), and p130 Cas (81). In accord with this view, the effects of src F527 on myocyte morphology were greatly reduced in the absence of Src kinase activity. Indeed, interaction and phosphorylation of Src and p125 FAK have been proposed as critical early steps in the process by which cell adhesion initiates intracellular signaling (reviewed in Ref. 49). Another recent report suggests that Src family tyrosine kinases are essential for ERK activation in myocytes in response to hydrogen peroxide (82). It is likely that Src family tyrosine kinases will also serve to transduce the effects of other stimuli in cardiac myocytes, although our initial studies do not support a role for Src in the response to PE.
The lack of effect of srcKϪ on PE-induced transcription might be interpreted in one of several ways. First, and most obviously, c-Src may not be involved in the hypertrophic response to PE. Second, the hypertrophic response to PE could require c-Src but not its kinase activity, such that transfected srcKϪ may be equally as effective as c-Src in transducing the effect of PE. However, overexpression of SrcKϪ or Src F527 KϪ did not stimulate ANF-LUX expression making this explanation unlikely (results not shown). Third, although PE may signal through c-Src, when this pathway is inhibited by SrcKϪ other Src family members may be able to substitute for it. Alternatively, the response to PE may require another member of the Src family that is not inhibited by SrcKϪ but whose effects on activation by PE can be reproduced by src F527 . Another point for consideration is that it is also possible that src F527 does not stimulate the same pathways as activated endogenous c-Src. Finally, SrcKϪ and Src F527 KϪ may not be effective competitors of c-Src. Thus, whether c-Src or one of its family members is involved in the hypertrophic response to PE remains equivocal.
In summary, src F527 initiates changes in transcription associated with the hypertrophic response, an effect for which it is dependent on Ras. Transfection with src F527 also increases myocyte size, although it does not increase the organization of the contractile apparatus into sarcomeric units. We conclude that members of the Src family of non-receptor protein tyrosine kinases may be important in mediating the transcriptional responses that occur during the development of cardiac hypertrophy, although the hypertrophic stimuli which may utilize Src as part of their signaling pathway have yet to be defined.