Gα₁₃ stimulates gene expression and increases cell size in cultured neonatal rat ventricular myocytes

Abstract

Objectives: Constitutively-active Gα₁₃ causes permissive cell types to proliferate or undergo phenotypic transformation implying a role for G₁₃ in the control of cell growth. Cardiac myocytes are terminally-differentiated cells which respond to growth stimuli by increasing in size rather than by cell division. The objective of this study was to determine whether constitutively-active Gα₁₃ is able to induce a hypertrophic phenotype in cardiac myocytes. Methods: Cultured neonatal rat ventricular myocytes were transiently transfected with an expression vector (pRC/RSV) encoding wild-type Gα₁₃ or constitutively-active Gα₁₃Q226L. Effects on transcription were monitored by co-transfected luciferase (LUX) reporter genes under the control of promoters responsive to hypertrophic stimuli. Cell size was determined by planimetry. Results: Transfection of neonatal myocytes with Gα₁₃Q226L, but not wild-type Gα₁₃, stimulated ANF638LUX and ANF3003LUX expression to 3.0±0.3- and 4.3±0.6-fold of the control, respectively. Likewise, Gα₁₃Q226L stimulated vMLC250LUX and vMLC2700LUX expression to 3.9±1.0- and to 7.7±1.7-fold of controls, respectively, but there was relatively little effect of Gα₁₃Q226L on c-fos-SRE- and β-MHC promoter activity. The effects of Gα₁₃Q226L on ANF3003LUX were inhibited by expression of C3 exoenzyme. Wild-type Gα₁₃ and Gα₁₃Q226L increased myocyte area from 869±43 μm² in control transfections to 1287±64 μm² and 1278±59 μm², respectively. Conclusion: We conclude that Gα₁₃Q226L is able to induce gene expression and morphological changes associated with a hypertrophic response in cardiac myocytes and that the transcriptional effects may be mediated through a Rho-dependent mechanism.

Time for primary review 28 days.

1 Introduction

As a consequence of their terminally differentiated phenotype, cardiac myocytes respond to growth stimuli by increasing their size rather than by undergoing cell division. Thus, whilst the oncogenic counterparts of cellular proto-oncogenes induce proliferation and phenotypic transformation (e.g. anchorage-independent growth) in many cell types, cardiac myocytes initiate a hypertrophic response. This is characterised by an increase in cell size, protein content and myofibrillar organization as well as specific changes in gene transcription (reviewed in Ref. [1]). The latter include a rapid and transient increase in the expression of immediate early genes (c-jun, c-fos and egr-1) [2, 3], the re-expression of atrial natriuretic factor (ANF), β-myosin heavy chain (β-MHC) and skeletal muscle α-actin and the upregulation of contractile protein genes such as ventricular myosin light chain-2 (vMLC2) and cardiac muscle α-actin.

Despite the different phenotypic responses (cell division, transformation or hypertrophy), common intracellular signalling pathways are frequently used to transduce the effects of growth stimuli on different cell types. In cardiac myocytes much recent attention has focused on the role of mitogen-activated protein kinases (MAPKs) as mediators of the hypertrophic response [4–10]. MAPKs are classified into the extracellularly regulated kinases (ERKs) and the stress-activated MAPKs, which themselves are subdivided into the c-Jun N-terminal kinases (JNKs) and the p38MAPKs. Broadly speaking, ERKs are activated by classical hypertrophic agonists such as phenylephrine (PE), endothelin-1 (ET-1) and 12-O-tetradecanoyl phorbol 13-acetate (TPA), whereas the stress-activated MAPKs, are activated more strongly by cell stresses such as ischaemia/reperfusion, osmotic shock and inhibitors of protein synthesis. Recent evidence points to a role for stress-activated MAPK activation in cardiac hypertrophy. Firstly, overexpression of MEKK1 or MKK7, upstream activators of the JNK pathway, induces transcriptional and morphological features of the hypertrophic response [7, 11]. Secondly, the JNK pathway is activated in the hypertrophied hearts of transgenic mice overexpressing oncogenic Ras [8]. Thirdly, overexpression of an upstream activator of p38MAPKβ induces a hypertrophic response [12]and SB 203580, a specific inhibitor of p38MAPK, inhibits the hypertrophic response to PE [10].

Guanine nucleotide-binding proteins (G proteins) play a central role in the signalling response to many extracellular stimuli. The G₁₂ family is the most recently identified family of G proteins and consists of G₁₂ and G₁₃[13]. The signal transduction pathways which G₁₂ and G₁₃ control are largely unknown but their role in cell growth is apparent from their ability to transform cells [14, 15]and to stimulate stress fibre formation [16]. In these Gα₁₃-transformed NIH 3T3 cells, expression of the primary response genes egr-1 and c-fos (both transcription factors) was constitutively-activated [15]. Gα₁₂ and Gα₁₃ robustly activate the JNK pathway but not the ERK pathway in NIH 3T3 cells [17]. The observations that Gα₁₂/Gα₁₃ induce JNK activation, initiate egr-1 and c-fos expression and regulate the actin cytoskeleton suggests a possible role in the development of cardiac hypertrophy, particularly in response to stress. Accordingly, we have determined the effects of Gα₁₃ on transcriptional and morphological markers of the hypertrophic response in cultured cardiac myocytes. We demonstrate that Gα₁₃ is able to induce genes associated with the hypertrophic phenotype and increase myocyte size. We propose that Gα₁₃ has the capacity to participate in the development of cardiac hypertrophy, particularly as a response to cellular stresses.

2 Methods

The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).

Sprague-Dawley rats were bred within the National Heart and Lung Institute. Culture medium and other reagents were from Sigma Chemical (Poole, Dorset, U.K.), Marathon Laboratory Supplies (London, U.K.) or Merck (Lutterworth, Leics., U.K.), as described previously [18]. Endothelin-1 was from Bachem.

2.1 Expression and reporter plasmids

Murine Gα₁₃ was isolated by RT–PCR from liver RNA and subcloned into the expression vector pRC/RSV (Invitrogen). Gα₁₃Q226L harbours a point mutation which disables the intrinsic GTPase activity, resulting in a constitutively active construct. The rat ANF-luciferase (ANF-LUX) reporter constructs pANF(−638)LΔ5′ (ANF638LUX) and pANF(−3003)LΔ5′ (ANF3003LUX) [19], the vMLC2-LUX constructs pMLC(−250 to +18)LΔ5′ (vMLC250LUX) and pMLC(−2700 to +18)LΔ5′ (vMLC2700LUX) [20], and the Rous sarcoma virus (RSV)–LUX and pON249 [21]in which β-gal is expressed from a constitutive cytomegalovirus (CMV) promoter were kindly provided by Dr. K.R. Chien (Dept. of Medicine, University of California San Diego, CA). The AP-1-LUX construct TRE2PRL(-36) was a gift from Drs. J.H. Brown (Dept. of Pharmacology, University of California, San Diego, CA) and M.G. Rosenfeld (Howard Hughes Medical Institute, University of California, San Diego, CA). The LUX reporter constructs for rat β-MHC [22]and the mouse c-fos serum responsive element (c-fos-SRE) [23]were gifts from Dr. M.D. Schneider (Molecular Cardiology Unit, Baylor College of Medicine, Houston, TX) and have been described in detail previously [6]. The dominant-negative N17cdc42 and N17Rac1 constructs in the vector pRK5myc and the Rho inhibitor C3 exoenzyme in pEFmyc were provided by Prof. Alan Hall (MRC Laboratory for Molecular Cell Biology, University College London). Plasmids were purified by polyethylene glycol precipitation [24].

2.2 Transient transfection of cultured neonatal rat ventricular myocytes

Ventricular myocytes were isolated from the hearts of 1–2 day old rats as described previously [6]. Fibroblast contamination was reduced during a 30-min pre-plating and myocytes were subsequently plated on gelatin-coated 60-mm tissue culture dishes at a density of 1 million cells per dish (350 cells per mm²) for reporter gene experiments or at a density of 500 000 cells per 60-mm dish for assessment of cell size. Cells were maintained at 37°C in a humidified atmosphere of 5% CO₂. Twenty hours after plating and 4 h before transfection, the medium was changed to 4% horse serum in maintenance medium (Dulbecco’s modified Eagle’s medium and medium 199 in a 4:1 ratio). Plasmids were diluted in 0.25 M CaCl₂ and an equal volume of 50 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulphonic acid (pH 6.9), 280 mM NaCl, 1.5 mM Na₂HPO₄ was added. After 20 min, 0.5 ml of this suspension (containing 15 μg LUX reporter plasmid, 4 μg of pON249 and 0.3–10 μg of test plasmid, as indicated) was added to 4 ml of medium on each plate. After overnight transfection, cells were washed once with 10% horse serum in maintenance medium and twice with maintenance medium. Transfection efficiency assessed by 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-gal) staining of cells and light microscopy is routinely 1–2% by this method [6]but was not assayed here. After 48 h, myocytes were washed three times with ice-cold phosphate-buffered saline (PBS) and extracted on ice into 400 μl 0.1 M potassium phosphate (pH 7.9), 0.5% (v/v) Triton X-100, 1 mM dithiothreitol. Assays for LUX and β-gal were performed as detailed previously [6].

2.3 Determination of myocyte size

To assess the area of transfected myocytes, cells were transfected as above, except that myocytes were plated at a density of 0.5 million cells per dish to facilitate identification of individual cells. Cells were washed three times with ice-cold PBS, fixed with 4% formaldehyde in PBS for 10 min, re-washed three times with PBS and then stained with 0.2 mg/ml X-gal, 5 mM K₄Fe(CN)₆, 5 mM K₃Fe(CN)₆, 2 mM MgCl₂ in PBS. After allowing colour to develop, transfected (blue) cells were randomly selected from all areas of the dishes and video hardcopies taken. Cell area was determined by planimetry.

2.4 Expression of results and statistical analysis

Results are expressed as the ratio of reporter gene expression in the presence of Gα₁₃ relative to empty vector controls. All transfection data was corrected for β-gal expression from the co-transfected pON249 plasmid. Results are presented as means ±S.E.M. for the number of separate myocyte preparations shown in the figure legends. Statistical significance was assessed by using a paired or unpaired Student’s t test, as appropriate, with a significant difference taken as being established at P<0.05.

3 Results

3.1 Stimulation of ANF- and vMLC2 promoter activity by expression of constitutively-active Gα₁₃Q226L in neonatal myocytes

Transfection of neonatal myocytes with Gα₁₃Q226L caused a dose-dependent increase in LUX expression from a co-transfected ANF638LUX transgene (Fig. 1). Maximal expression was elicited with 3 μg Gα₁₃Q226L and this concentration was used in all subsequent studies. In contrast, there was little change in ANF-LUX expression with increasing concentrations of wild-type Gα₁₃ or backbone vector (Fig. 1). The proximal 638 bp of the rat ANF promoter is sufficient to confer selective upregulation in response to hypertrophic agonists in cardiac myocytes [19]. To determine if this promoter fragment was sufficient to transduce the full effect of Gα₁₃Q226L on ANF-LUX expression, the responses of ANF638LUX and ANF3003LUX were compared (Fig. 2). In this series of experiments Gα₁₃Q226L stimulated ANF638LUX expression to 3.0±0.3-fold of control and ANF3003LUX to 4.3±0.6-fold of control, but there was no significant difference in the extent of activation of these two promoter fragments. Transfection with wild-type Gα₁₃ had no effect on either promoter construct. A similar pattern was observed for vMLC2-LUX constructs (Fig. 3). Again, two promoter fragments were tested because, although the proximal 250 bp of the promoter is sufficient for tissue-specific and inducible expression of vMLC2-LUX transgenes [19], it is still possible that sequences in the promoter between −250 and −2700 harbour cis elements responsive to Gα₁₃Q226L. Gα₁₃Q226L stimulated vMLC250LUX and vMLC2700LUX expression to 3.9±1.0− and to 7.7±1.7-fold of controls, respectively, but these were not statistically different. There was no effect of wild-type Gα₁₃ on the 250 bp vMLC2 construct but expression of vMLC2700LUX was enhanced to 2.3±0.5-fold of control. Two additional control experiments were performed to determine that the effects of Gα₁₃Q226L on ANF- and vMLC2-LUX expression were specific (Fig. 4). Neither Gα₁₃Q226L nor wild-type Gα₁₃ had any effect on basal expression from the promoterless LUX vector pSVOALΔ5′ or on LUX expression driven by the constitutively-active Rous sarcoma virus (RSV) promoter. For comparison, it should be noted that basal LUX/β-gal expression from the RSV construct was 40.0±12.9-fold greater than that from ANF3003LUX (P<0.025) and 35.1±7.2-fold greater than that from vMLC250LUX (P<0.005). The effects of Gα₁₃Q226L were therefore specifically attributable to activation of the ANF and vMLC2 promoters.

3.2 Stimulation of c-fos-SRE- and β-MHC promoter activity by Gα₁₃Q226L

In addition to ANF and vMLC2, several other genes are specifically activated in response to hypertrophic agonists including skeletal muscle α-actin, β-MHC and immediate-early genes such as c-fos, c-jun and Egr-1[1]. Transfection with Gα₁₃Q226L had much smaller effects on LUX expression under the control of a c-fos-SRE promoter element or the β-MHC promoter (Fig. 5). Expression of the c-fos-SRE- and β-MHC-LUX constructs were only increased by 26% and 70%, respectively, though both effects were statistically significant. Treatment of myocytes transfected with the c-fos-SRE- and β-MHC-LUX constructs with TPA demonstrated that these transgenes were permissive to much stronger activation (4 to 5-fold) than that elicited by Gα₁₃Q226L (Fig. 5).

3.3 Gα₁₃Q226L does not stimulate AP-1-dependent promoter activity

As an extension of these studies we examined the effect of Gα₁₃Q226L on a LUX reporter construct under the control of two consensus TPA response elements (2×TRE) which bind the transcription factor AP-1, most usually composed of a c-Fos/c-Jun heterodimer. This construct (2×TRE) was completely unresponsive to Gα₁₃Q226L but was strongly activated by TPA (LUX/β-gal fold-induction for Gα₁₃Q226L and TPA were 0.85±0.14 (n=10) and 19.2±4.3 (n=5) relative to controls, respectively). Thus, Gα₁₃Q226L does not transactivate genes through an AP-1-dependent mechanism in cardiac myocytes.

3.4 Gα₁₃Q226L stimulation of transcription requires Rho but not Rac1 or cdc42

Gα₁₃ activates Rho in a wide variety of cell types and, since ANF-LUX expression in myocytes in response to PE is dependent on Rho [25, 26], we investigated whether Rho may mediate Gα₁₃Q226L-induced transcription of ANF-LUX (Fig. 6). Expression of C3 exoenzyme, which specifically ADP-ribosylates and inactivates Rho, prevented ANF-LUX induction in response to Gα₁₃Q226L. In contrast, no such inhibition was observed when dominant-negative constructs of two other members of the Rho family of GTPases, cdc42 and Rac1, were co-transfected with Gα₁₃Q226L, suggesting a specific role for Rho itself.

3.5 Na⁺/H⁺ exchanger activity is not required for Gα₁₃Q226L-induced ANF-LUX activity

Gα₁₃ activates Na⁺/H⁺ exchange in many cell types, an effect which may or may not be dependent on Rho [27]. Accordingly, to determine whether Na⁺/H⁺ activity may be required for Gα₁₃Q226L-induced transcription, myocytes were incubated with the Na⁺/H⁺ exchanger inhibitor 5-(N-ethyl-N-isopropyl) amiloride (EIPA). As shown in Fig. 7, EIPA (which inhibits the Na⁺/H⁺ exchanger with a K_i of 0.4 μM) failed to inhibit Gα₁₃Q226L-induced ANF-LUX expression at concentrations ranging from 1–16 μM. Indeed, at a concentration of 4 μM, EIPA actually increased ANF-LUX/β-gal in response to Gα₁₃Q226L, though basal ANF-LUX activity was also increased to a similar extent by EIPA in this series of experiments. In contrast, 4 μM EIPA almost completely inhibited ANF-LUX expression in response to PE.

3.6 Gα₁₃Q226L induces morphological features characteristic of the hypertrophic phenotype in neonatal cardiac myocytes

Another characteristic feature of neonatal myocytes exposed to hypertrophic stimuli is their increased size [3, 6, 28]. Myocytes transfected with either wild-type Gα₁₃ (Fig. 8A, top right panel) or Gα₁₃Q226L (bottom left panel) were larger in cross-sectional area than the surrounding non-transfected cells. In contrast, myocytes transfected with the backbone vector were no different in size from the untransfected cells (top left panel). For comparison, cells in another dish were exposed to the powerful hypertrophic agonist ET-1, which increased the area of all the cells (bottom right panel). Quantitative analysis of 100 transfected cells determined that myocyte area was increased from 869±43 μm² in control transfections to 1287±64 μm² and 1278±59 μm² in transfections with wild-type Gα₁₃ and Gα₁₃Q226L, respectively (Fig. 8B). The increase in area of Gα₁₃-transfected cells (47–48%) was similar to that in vector-transfected cells treated with ET-1 (54%, results not shown). Wild-type Gα₁₃ and Gα₁₃Q226L also increased the area/perimeter ratio, indicative of a more regular profile [18].

4 Discussion

Heterotrimeric G proteins act as intermediates between receptors and second messenger systems. They are characterised by the distinct properties of their α-subunits and are classified into four families (G_s, G_i, G_q and G₁₂) based on amino acid sequence homology (reviewed in [29, 30]). The G₁₂ family is the most recently identified and consists of G₁₂ and G₁₃, which have been detected at both the mRNA and protein levels in a large number of tissues including the heart [13, 31, 32]. Their widespread distribution suggests that they may be involved in signal transduction pathways common to most tissues and their ability to transform cells [14, 15]points to a role in the regulation of cell growth. In this study we have demonstrated that Gα₁₃ can initiate a hypertrophic response in neonatal cardiac myocytes, characterised by upregulation of ANF-, vMLC2- and c-fos-SRE promoter activity and an increase in cell size.

The intracellular signalling pathways through which Gα₁₃ transduces its growth-promoting effects are still poorly understood. In many cell types Na⁺/H⁺ exchanger activity is a target for both Gα₁₂ and Gα₁₃[33], though the isoforms involved and the consequences of their action may differ. Thus, whereas Gα₁₂Q229L inhibits NHE1 and activates NHE2 and NHE3, Gα₁₃Q226L activates all three isoforms [34]. Since NHE1 is the predominant isoform in the heart, Gα₁₃ may be expected to activate Na⁺/H⁺ exchanger activity more strongly than Gα₁₂ in this tissue. Activation of Na⁺/H⁺ exchanger activity leads to intracellular alkalinization which is a potent anabolic stimulus in the heart and in cardiac myocytes [35, 36]. However, stimulation of Na⁺/H⁺ exchange activity does not appear to be the mechanism through which Gα₁₃Q226L induces gene expression since this effect was not inhibited by EIPA but was somewhat augmented by it. In contrast, EIPA was able to inhibit PE-induced ANF expression, implying that this α₁-adrenergic effect is dependent on Na⁺/H⁺ exchanger activity.

Another group of intracellular signalling pathways which are modulated by Gα₁₂ and Gα₁₃[17, 37, 38]and which are important regulators of growth in a wide variety of cell types [39, 40]are the MAPK cascades. Transient transfection of NIH 3T3 [17]or COS-7 [38]cells with Gα₁₂Q229L or Gα₁₃Q226L activates the JNK pathway but not the ERK pathway. On the other hand, the ERK pathway can be activated by Gα₁₂ and Gα₁₃ in Rat-1 fibroblasts and both ERK and JNK activation appear to be required for mitogenesis in response to Gα₁₂ activation in NIH 3T3 fibroblasts [37, 41]. In cardiac myocytes there is a growing consensus that one or more MAPK pathways may be important in mediating the transcriptional and morphological responses to hypertrophic agonists [4–7, 9, 10, 18, 42–47]. Although activation of the ERK family of MAPKs was initially thought to be of paramount importance [4–6, 18, 44, 45], other more recent data suggests that activation the JNKs and p38MAPKs may be more closely associated with the hypertrophic response [7, 9, 10, 47]. Activation of the JNKs but not the ERKs [17, 38]by Gα₁₃Q226L would be consistent with a role for the stress-regulated MAPKs in the hypertrophic response.

The Rho family of GTPases includes Rho, Rac and cdc42 family members and there is increasing evidence that Rho plays an important role in the response of cardiac myocytes to hypertrophic agonists, perhaps in concert with MAPK pathway activation [25, 26, 48]. Myocytes infected with a recombinant adenovirus expressing constitutively activated Rho displayed enhanced ANF expression and increased assembly and organization of sarcomeric units, whereas expression of the Rho inhibitor C3 exoenzyme inhibited ANF expression and myofibrillar organization in response to PE [48]. The studies reported herein show that C3 exoenzyme is also able to inhibit Gα₁₃-induced ANF expression, suggesting that Rho is required for this response. In contrast, transfection with dominant-negative cdc42 and Rac1 constructs failed to inhibit Gα₁₃-induced ANF expression but this data should be viewed with caution since a dominant-negative Rho construct also failed to inhibit ANF expression significantly (data not shown). One possible explanation is that the dominant-negative constructs may be less effective since they act by mass action in binding the upstream Rho, cdc42 and Rac1 guanine nucleotide exchange factors, whereas the C3 exoenzyme will act catalytically to inhibit Rho directly. Hence, the current data support a role for Rho in Gα₁₃Q226L-mediated ANF expression but do not definitively rule out a potential role for the other Rho family members.

Little is known about the transcription factors which mediate the effects of the G₁₂ family on gene expression. In NIH 3T3 cells transformed with Gα₁₃Q226L, early growth response gene-1 (egr-1) and c-fos are constitutively-active [15]. In the same cell type a constitutively-active Gα₁₂ mutant stimulated expression from an SRE-reporter gene [49], suggesting that activation of c-fos expression is most likely effected through the SRE in its promoter. Curiously, we observed only a small activation of a c-fos-SRE reporter gene by Gα₁₃Q226L, despite the ability of TPA to strongly activate it (Fig. 5). The reason for this difference is not clear but might involve tissue-specific differences in transcription factors. The β-MHC-LUX transgene was also only modestly upregulated by Gα₁₃Q226L, whereas the ANF- and vMLC2-LUX transgenes were strongly activated.

There were no significant differences in the degree of activation of the ANF638- and ANF3003-LUX constructs by Gα₁₃Q226L, indicating that the proximal 638 base pairs of the rat ANF promoter contains the cis elements necessary for induction by Gα₁₃Q226L. This region of the rat ANF promoter has many consensus sequence elements for binding transcription factors, including, egr-1, AP-1 and cAMP response element (CRE) motifs [19]. Different ANF promoter elements, including more specific cis sequences [50–52], appear to be responsible for ANF expression under a variety of experimental conditions [53], so any or several of these cis elements may be important. Since egr-1 is activated in Gα₁₃Q226L-transformed cells [15], this is one possibility. It is unlikely that the AP-1 site is important, however, because Gα₁₃Q226L failed to stimulate expression from a LUX reporter construct under the control of consensus AP-1 binding site. Furthermore, C3 exoenzyme does not inhibit PE-induced activation of this same AP-1 construct, whereas it does inhibit PE-induced ANF promoter activity [26]. Though there are several consensus cis elements within the vMLC250 and vMLC2700 promoter fragments [54], none are obvious candidates for activation by Gα₁₃Q226L.

As well as inducing the selective activation of marker genes for hypertrophy, overexpression of Gα₁₃Q226L increased myocyte size (Fig. 8). Interestingly, although wild-type Gα₁₃ was largely ineffective at inducing transcriptional responses, it was equally as effective as Gα₁₃Q226L in increasing myocyte size. This presumably reflects a low basal level of activity of Gα₁₃ which is sufficient to induce a response when overexpressed and suggests that cell size is more sensitive to Gα₁₃ than is transcription. It is not altogether surprising that wild-type Gα₁₃ can initiate a growth response in myocytes because even wild-type Gα₁₂ can transform NIH 3T3 cells when overexpressed, albeit at a much lower frequency than with Gα₁₂Q229L [55]. Overexpression of the α subunit of G₁₃ will most likely disrupt the equilibrium between the α and βγ subunits of G₁₃ and in doing so may prolong the activity of Gα₁₃GTP, or may simply increase its concentration by a mass action effect. The mechanism for the effect of Gα₁₃ on cell size is not known, but, like its effect on transcription, may also be dependent on Rho family GTPases. Overexpression of constitutively-active Rho in myocytes by using a recombinant adenoviral vector increased sarcomeric organization [48], though a different study failed to detect any effect when recombinant constitutively active Rho protein was microinjected or expression plasmids were transfected [26]. Other processes such as phospholipase D (PLD) activation might also be important since hypertrophic agonists activate PLD in myocytes [56]and PLD activation by Gα₁₃Q226L is dependent on Rho GTPases [57].

Finally, the extracellular stimuli which may be coupled to a hypertrophic response through the G₁₂ family of G proteins in cardiac myocytes are not known. Little information is available on what the upstream activators of Gα₁₂ and Gα₁₃ are in any cell type, but the best characterised system is in platelets where thrombin activates both Gα₁₂ and Gα₁₃[58]. Thrombin has also been shown to activate ANF expression in myocytes [59], though it is not a classical hypertrophic agent. Another potential agonist is angiotensin II which initiates characteristic hypertrophic changes in cardiac myocytes [60]and which has been proposed to couple to increases in cytoplasmic Ca²⁺ in rat portal vein myocytes via G₁₃ activation [61]. However, it is the βγ subunits of G₁₃ which are responsible for the Ca²⁺-elevating effect in the latter case.

In summary, we have shown that a GTPase-deficient mutant of Gα₁₃ is able to induce transcriptional and morphological features of the hypertrophic phenotype in cardiac myocytes. The mechanisms involved in transducing these effects of Gα₁₃ have not been unequivocally defined but most likely involve Rho, though activation of a number of other components including stress-activated MAPK pathways, PLD and other Rho family members may also be important. Upstream effectors which may elicit their hypertrophic effects through Gα₁₃ have still to be determined.

Acknowledgements

This work was supported by grants BS1 and FS/96057 from the British Heart Foundation (to S.J.F.). We thank Ken Chien, Michael Schneider, Joan Heller Brown and Alan Hall for providing plasmids and Nicola Haward and Richard Taylor for preparation of the myocytes.

References