Selective Loss of Fine Tuning of Gq/11 Signaling by RGS2 Protein Exacerbates Cardiomyocyte Hypertrophy*

Alterations in cardiac G protein-mediated signaling, most prominently Gq/11 signaling, are centrally involved in hypertrophy and heart failure development. Several RGS proteins that can act as negative regulators of G protein signaling are expressed in the heart, but their functional roles are still poorly understood. RGS expression changes have been described in hypertrophic and failing hearts. In this study, we report a marked decrease in RGS2 (but not other major cardiac RGS proteins (RGS3-RGS5)) that occurs prior to hypertrophy development in different models with enhanced Gq/11 signaling (transgenic expression of activated Gαq* and pressure overload due to aortic constriction). To assess functional consequences of selective down-regulation of endogenous RGS2, we identified targeting sequences for effective RGS2 RNA interference and used lipid-based transfection to achieve uptake of fluorescently labeled RGS2 small interfering RNA in >90% of neonatal and adult ventricular myocytes. Endogenous RGS2 expression was dose-dependently suppressed (up to 90%) with no major change in RGS3-RGS5. RGS2 knockdown increased phenylephrine- and endothelin-1-induced phospholipase Cβ stimulation in both cell types and exacerbated the hypertrophic effect (increase in cell size and radiolabeled protein) in neonatal myocytes, with no major change in Gq/11-mediated ERK1/2, p38, or JNK activation. Taken together, this study demonstrates that endogenous RGS2 exerts functionally important inhibitory restraint on Gq/11-mediated phospholipase Cβ activation and hypertrophy in ventricular myocytes. Our findings point toward a potential pathophysiological role of loss of fine tuning due to selective RGS2 down-regulation in Gq/11-mediated remodeling. Furthermore, this study shows the feasibility of effective RNA interference in cardiomyocytes using lipid-based small interfering RNA transfection.

Myocardial hypertrophy accompanies many forms of heart disease and develops in response to hemodynamic stress (such as pressure or volume overload), loss of contractile proteins after cardiac injury, or mutations in sarcomeric proteins. Many different stimuli and signaling pathways have been implicated in hypertrophy development (1). Among them, signaling events that are mediated by heterotrimeric G q/11 proteins are of central importance (2). G q/11 proteins are members of a protein family that transduces signals from many different G protein-coupled receptors (GPCRs) 3 to generate diverse effects on myocyte growth, contractility, and heart rate (3). G proteins are subject to direct regulation by RGS (Regulators of G protein Signaling) proteins, which shorten the duration of the cellular response to external signals and generally cause a reduction in hormone sensitivity (4). Whereas their primary mode of action is acceleration of signal termination by decreasing the lifetime of active, GTP-bound G␣ subunits, some RGS can also block signal generation by antagonizing G␣-mediated effector activation (5). Thus, RGS proteins are believed to play an important role in fine tuning GPCR-induced signal transduction. Several RGS proteins are expressed in the heart, but their exact roles in the normal and diseased heart are still poorly defined (6). Most cardiac RGS proteins belong to the B/R4 family of "small" RGS proteins that primarily consist of the G␣-binding RGS homology domain and short extensions (7). They are mostly nondiscriminatory toward members of the G q/11 and G i/o subfamilies in vitro but tend to display greater specificity in intact cells (8). Myocardial expression studies to date have been restricted to a few RGS proteins and hearts that were already hypertrophic and/or failing (9 -11). At that stage, a multitude of signaling changes occur, many of which are probably secondary to the remodeling process.
The first goal of the present study was to test the hypothesis that changes in the ventricular amount of G q/11 -regulating RGS proteins may precede hypertrophic remodeling and could therefore potentially be of pathophysiological significance. An increase in G q/11 -regulating RGS protein(s) would be expected to attenuate G q/11 signaling and could thereby blunt or delay G q/11 -mediated hypertrophy, whereas conversely a decrease in RGS could exacerbate G q/11 signaling and have a deleterious effect. We focused on RGS proteins that have been consistently detected in cardiac tissue (9,11) and myocytes (12,13), namely RGS2, RGS3, RGS4, and RGS5. Since RGS16 has been reported to be highly inducible in cardiac myocytes (13), it was also included in this analysis. RGS mRNA expression was examined in two different models with enhanced G q/11 signaling: (i) transgenic mice expressing constitutively active G␣ q Q209L (G␣ q *) (14), in which enhanced G␣ q signaling is restricted to myocytes until remodeling eventually affects signaling in nonmyocytes as well, and (ii) pressure overload hypertrophy in response to aortic banding, because its dependence on G q/11 signaling is well recognized (15,16). Importantly, in contrast to existing studies, we examined RGS expression over time, with a special focus on changes that occur early on, and observed a selective reduction in RGS2 protein and mRNA expression in both models prior to hypertrophy initiation.
In the second part of the study, we tested the hypothesis that endogenous RGS2 inhibits G q/11 -mediated signaling and hypertrophy in cardiac myocytes and determined whether or not selective RGS2 knockdown has functional implications despite the presence of other RGS proteins with potentially overlapping function. In order to achieve that goal, we identified targeting sequences that allowed for effective and selective RNA interference (RNAi) of endogenous RGS2 and established experimental conditions for RNAi in cultured cardiac myocytes. RNAi has evolved in recent years as a powerful approach to decode gene function of endogenous proteins. It describes a phenomenon by which double-stranded RNA induces potent and specific inhibition of eukaryotic gene expression via the degradation of complementary mRNA (17,18).

EXPERIMENTAL PROCEDURES
In Vivo Models of Hypertrophy-The generation and respective phenotypes of two independent transgenic mouse lines with heart-specific expression of constitutively active G␣ q * have been described elsewhere (14,19,20). Pressure overload was induced by ascending aortic constriction in 2.5-month-old FVB mice for 8 weeks as described (21). All experimental protocols were reviewed and approved by the institutional animal care and use committee. Additional information on RGS expression in response to pressure overload hypertrophy due to ascending aortic constriction was obtained. 4 Northern Blot Analysis-Total ventricular RNA (RNAzol B; TelTest) was extracted from mouse ventricular tissue, size-fractionated (10 g/lane), and transferred to nylon membranes that were incubated with 32 P-labeled cRNA probes (Strip-EZTM SP6/T7; Ambion) for RGS2, RGS4, and RGS5, random-primed 32 P-labeled (Prime-ItII; Stratagene) ANF and GAPDH cDNA probes, and an 18 S oligonucleotide probe that had been labeled with [␥-32 P]ATP by T 4 polynucleotide kinase (Gibco). The RGS mRNA transcripts sizes were 1.5/1.8 kb (RGS2), 3.3 kb (RGS4), and 4.4 kb (RGS5) with no other bands detectable. Blots were exposed to x-ray film followed by quantitative densitometry. Data were normalized to GAPDH or 18 S rRNA.
Real Time PCR-Reverse transcribed (TaqMan Reverse Transcription Reagents; Applied Biosystems) RNA samples from neonatal and adult rat ventricular myocytes (Trizol; Invitrogen) were subjected to PCR using FAM TM -labeled Taqman MGB probes for RGS2-RGS5 and 18 S and TaqMan Universal PCR master mix (Applied Biosystems) according to the manufacturer's instructions. Each sample was assayed in two independent RT reactions and duplicate reactions each and normalized to 18 S expression. Negative controls included the absence of enzyme in the RT reaction and absence of template during PCR. The PCR cycling was performed at 95°C for 10 min and then 95°C for 15 s and 60°C for 1 min for a total of 45 cycles. The cycle threshold (C T ) values corresponding to the PCR cycle number at which fluorescence emission in real time reaches a threshold above the base-line emission were determined using ABI Prism 7700 software. For quality control, serial dilutions of cDNA plasmids for rat RGS2-RGS5 (0.3-3 ϫ 10 6 copies) were tested, and the linearity of the resulting C T values was confirmed for RGS2 and RGS5 but not RGS3 or RGS4. Relative mRNA expression levels were therefore assessed by real time PCR for RGS2 and RGS5 only.
CardioGenomics Microarray Data Base-The data base 4 was mined for information on RGS mRNA expression in response to aortic constriction. Signals from each chip (Affymetrix MG-U74) were subjected to linear normalization and analyzed with the Affymetrix Microarray Suite 5 algorithm. Similar results were obtained upon additional nonlinear normalization at the probe level (data not shown).
Detection of endogenous RGS proteins by Western blot analysis is notoriously difficult due to their low abundance. To increase the protein amount loaded, particulate and soluble proteins from ventricular tissue were extracted by ultracentrifugation (50,000 ϫ g for 30 min at 4°C), separated on 15% SDS-PAGE (60 g/lane), and immunoblotted as described above. Only one of five RGS2 antibodies tested recognized endogenous ventricular RGS2 protein by enhanced chemiluminescence (Fig. 1D). It recognized a double band in the appropriate molecular weight range that had a slightly faster mobility than His 6 -tagged RGS2 (provided by Dr. C. W. Dessauer (University of Texas Health Science Center at Houston); data not shown) and was absent in crude homogenates of brain cortex (provided by Dr. D. P. Siderovski (University of North Carolina)) from RGS2 knock-out (KO) mice. It is conceivable that the two bands reflect the presence of alternative splice forms or posttranslational modifications. Importantly, RGS2 was reduced in the particulate fraction from ␣ q *52 ventricles compared with wild type (see Fig. 1D). Due to low signal intensity, expression of soluble RGS2 protein could not be reliably assessed (data not shown). Only two of the other four RGS2 antibodies tested also recognized protein in the appropriate molecular weight range. However, the fact that the corresponding bands were still present in samples lacking RGS2, suggested cross-reaction with other proteins.
A pair of complementary oligonucleotides encoding a short hairpin (shRNA) sequence (sense and antisense target sequence separated by a 9-nucleotide loop sequence) plus BamHI and HindIII restriction sites at the ends were synthesized, annealed, and subcloned into a plasmid (pSilencer 3.0-H1 vector; Ambion) at the BamHI and HindIII restriction sites between a human H1 RNA polymerase III promoter and a RNA polymerase III terminator sequence. pSilencer plasmids containing siRNA hairpin sequences of GAPDH or without significant homology to any known rat gene (Ambion) were used as positive and negative controls, respectively. Inserted sequences were confirmed by restriction enzyme digestion and DNA sequencing. In addition, siRNAs corresponding to the shRNA sequences were synthesized and high pressure liquid chromatography-purified by Ambion.
Cell Culture and Transfection-HEK-293 cells were maintained in DMEM containing 10% fetal bovine serum and antibiotics and transfected using Lipofectamine 2000 following the instructions of the manufacturer (Invitrogen). Lipofectamine 2000 was first diluted in OPTI-MEM I (Invitrogen) and then mixed with plasmid DNA or siRNA at the indicated amounts. The ratio (w/w) between Lipofectamine 2000 and DNA or RNA was 4:1 or 3.7:1, respectively. One and 2 ml of medium/ well were used for 12-and 6-well plates, respectively. Following a 20-min incubation at room temperature, the DNA-or RNA-Lipofectamine 2000 mixture was added to antibiotics-free transfection medium (DMEM containing glutamine plus 10% fetal bovine serum) and replaced cell culture medium. The cells were maintained in a 37°C incubator with 95% O 2 , 5% CO 2 until analysis. The medium was changed to medium with no DNA or siRNA 24 h after transfection.
Cardiomyocyte siRNA Transfection-AVM and NVM were transfected with siRNA 1 or 2 days after isolation, respectively, in the same way as HEK-293 cells (see above), albeit in different media. For NVM, the transfection medium was composed of DMEM, 2 mM glutamine, and 2% fetal bovine serum. The transfection medium was replaced with serum-free DMEM containing 2 mM glutamine, 1ϫ ITS Liquid Media Supplement (Sigma), and penicillin/streptomycin 24 h after transfection. For AVM, the transfection medium was identical to the ACCTI culture medium and was replaced with fresh ACCTI culture medium 24 h after transfection.
Fluorescent Labeling of RGS2 siRNA-RGS2 siRNA (5 g) was labeled with fluorescin (FAM)-labeling reagent (Silencer TM labeling kit; Ambion) for 1 h at 37°C in the dark according to the manufacturer's instructions and then precipitated by ethanol. Increasing concentrations (0 -50 nM) of FAM-labeled RGS2 siRNA were transfected into NVM as described above, and cellular siRNA uptake was documented by fluorescent microscopy.
Immunofluorescent Staining-Immunofluorescent staining in HEK293 cells and cardiomyocytes was performed as previously described (27). In brief, cells were fixed in a solution containing 4% paraformaldehyde, 0.9% NaCl, and 0.1 M NaH 2 PO 4 (pH 7.4) for 15 min at room temperature. After permeabilization with 0.1% Triton X-100 in PBS for 15 min at room temperature, nonspecific binding sites were blocked with 3% bovine serum albumin in PBS for 2 h at room temperature or overnight at 4°C. They were then incubated in 1% bovine serum albumin/PBS containing mouse monoclonal antibodies against myosin (MF-20, 1:100 dilution; Developmental Studies Hybridoma Bank) or the FLAG epitope (M2, 1:1000; Sigma), followed by incubation with rhodamine-or fluorescein-conjugated secondary antibodies (goat anti-mouse from Pierce (1:200) and goat anti-rabbit from Amersham Biosciences (1:50), respectively), each for 1.5 h at room temperature.
Phospholipase C␤ Activity-Total inositol phosphate formation was measured as a surrogate parameter for PLC␤ activity as described (19). Briefly, 24 h after siRNA infection, NVM and AVM were labeled in inositol-free medium supplemented with myo-[ 3 H]inositol (2 Ci/well, 15 Ci/mmol; Amersham Biosciences) overnight. The next day, LiCl (10 mM final) was added before the addition of phenylephrine (PE) plus timolol (Tim) or endothelin-1 (ET-1) at increasing concentrations. After 30 min at 37°C, inositol phosphates were extracted in 20 mM formic acid, neutralized, separated by anion exchange chromatography (Dowex AG1-X8), and quantitated by liquid scintillation counting. Cell density was monitored for each condition. Normalization to protein amount/well yielded similar results (data not shown).
Assessment of Mitogen-activated Protein Kinase (MAPK) Activation-After siRNA transfection of NVM for 48 h, PE/Tim (1/0.1 M) or ET-1 (10 nM) was added to the medium for the indicated incubation times (5 or 20 min). Cells were washed with PBS and lysed in 1ϫ lysis buffer (Cell Signaling) with protease inhibitor mixture (Roche Applied Sciences) for 30 min at 4°C and sonicated for 3 ϫ 5 s. Equal quantities of protein lysates were separated on 12% SDS-PAGE and transferred to nitrocellulose. Nitrocellulose strips were incubated with antibodies against ERK1/2, p38, and JNK (all 1:1000; Cellular Signaling) that either recognized their respective phosphorylated forms only (9101, 9211, and 9255) or both phosphorylated and unphosphorylated MAPKs (9102, 9212, and 9252). After three washes with Tris-buffered saline containing 0.1% Tween 20 and incubation with appropriate peroxidase-coupled secondary antibodies, proteins of interest were visualized by chemiluminescence (SuperSignal West Pico Substrate; Pierce).
Assessment of Myocyte Hypertrophy-After siRNA transfection of NVM for 48 h, PE/Tim or ET-1 was added to the medium at the indicated concentrations. After 72 h, myocyte size was assessed in MF-20stained myocytes using ImageProPlus software (MediaCybernetics). Protein synthesis was measured in quadruplicates for each condition by continuous labeling of cell protein with [ 14 C]phenylalanine (0.05 Ci/ ml, Ͼ450 mCi/mmol; ICN) during treatment with G q/11 -coupled receptor agonists (23). On the day of the assay, cells were washed with PBS, followed by precipitation of cellular protein with 10% trichloroacetic acid for at least 1 h at 4°C. Cellular proteins were washed once with 10% trichloroacetic acid, solubilized in 1% SDS for 1 h at 37°C, and then quantified by liquid scintillation counting.
Statistical Analyses-Data are reported as mean Ϯ S.E. for n determinations or mean Ϯ range of duplicate determinations, as indicated. Each assay is representative of at least two additional independent assays. Statistical differences were assessed where appropriate by unpaired, two-tailed Student's t test or 2-way analysis of variance followed by Bonferroni post-tests for comparison of individual means (GraphPad Prism 4). p values smaller than 0.05 were considered statistically significant.

Expression of Endogenous RGS Proteins in Hearts with Enhanced G q/11
Signaling-Using Northern blot analysis and RT-PCR, we examined mRNA expression of the major cardiac RGS proteins in two different models of G q/11 -mediated hypertrophy, both prior to and during hypertrophic remodeling.
Transgenic Mice with Heart-restricted G␣ q * Expression-We previously reported that G␣ q * expression in vivo constitutively elevates cardiac PLC␤ activity (14,19) and provided a detailed phenotypic characterization (ventricular weights and hypertrophic marker gene expression, echocardiography, and signaling changes) of two independent lines (␣ q *52 and ␣ q *44h (20)). We reported that mice from both lines develop the same end stage phenotype (cardiac dilation with premature death in heart failure) with a similar rate of disease progression (within 2-3 months) but a very different time of disease onset (ϳ9 months later in ␣ q *44h). For the present study, time points were chosen that span the entire time frame from normal morphology to hypertrophy, dilation, and failure in each line (Fig. 1). Mice from line ␣ q *52 were examined at 0.5 months of age (when cardiac PLC␤ activity is increased but ventricular weight and ANF (see Fig. 1B) or ␤-myosin heavy chain expression are not yet significantly increased and the histology is still indistinguishable from age-matched wild types (indicated by gray bars in Fig. 1A)) and at 2.5 months of age (when ␣ q *52 mice re-express embryonic marker genes (see Fig. 1B) and have a markedly increased ventricular weight and chamber diameter and significantly impaired fractional shortening (indicated by the black bars in Fig. 1A)). Mice from line ␣ q *44 were examined between 0.5 and 14 months of age. Because of the late disease onset (at ϳ10 -12 months of age) but rapid disease progression toward the end-stage phenotype within 2-3 months (see Ref. 20 for details), 12-14-month-old ␣ q *44h mice were grouped into mice that were primarily still in their "predilated" stage (gray bars) and those that had already reached "end stage" (black bars).
The most striking change in the transgenic hearts was a reduction in RGS2 mRNA by over 60% that was already detectable in the respective "predilated" stages (0.5 months in ␣ q *52 and 6 months in ␣ q *44; Fig. 1A). RGS2 continued to be reduced (or progressed even further) until the end stage phenotype was reached at 2.5 and 12-14 months, respectively. RGS4 and RGS5 mRNA amounts were also decreased but to a lesser degree and, importantly, only at end stage. Details on the time course of RGS2, RGS4, and RGS5 mRNA expression in wild-type ventricles (between 2 days and 18 months of age) are also shown in Fig. 1A. RGS4 and RGS5 increased between 3-and 4-fold in the first 2 weeks of life and then remained at that elevated level (RGS5) or declined after 2 months (RGS4). In contrast, RGS2 mRNA levels remained largely constant over time, suggesting a relative shift in the overall composition of cardiac RGS isoforms during development and maturation.
RGS3 and RGS16 mRNA were difficult to detect/quantify by Northern blot analysis and were therefore assessed by RT-PCR. Overall, the expression of RGS3 was quite variable among different samples but was not consistently altered in both lines (data not shown). Consistent with previous findings (13), RGS16 mRNA was barely detectable in ventricular tissue from wild-type mice and appeared not to be altered in G␣ q * mice (data not shown).
A Northern blot in Fig. 1B illustrates RGS2 mRNA down-regulation in ␣ q *52 ventricles prior to and after the development of the cardiac phenotype (0.5 and 2.5 months, respectively). Importantly, it occurred in ventricular myocytes (Fig. 1C). Fig. 1D suggests that RGS2 mRNA down-regulation was accompanied by a reduction in RGS2 protein as early as 0.5 months of age (see "Experimental Procedures" concerning the difficulty in detecting endogenous RGS2 protein).
Pressure Overload Due to Aortic Constriction-RGS2 down-regulation was also observed in ventricles from mice 8 weeks after ascending aortic constriction that caused an increase in heart/body weight ratio compared with sham-operated controls (8.6 Ϯ 1.2 versus 4.5 Ϯ 0.1 mg/g (n ϭ 3 each), p Ͻ 0.05). Decreased RGS2 mRNA expression levels after banding were observed by Northern blot (Fig. 2A) and RT-PCR (56 Ϯ 5% of sham, n ϭ 6). The other RGS proteins were not significantly altered (RGS4, 116 Ϯ 12% of sham; RGS5, 133 Ϯ 23% of sham (n ϭ 6 each)) or increased by 52 Ϯ 28% (RGS3; n ϭ 6). Additional information on changes in RGS expression in response to pressure overload, particularly a detailed time course of events, was obtained from the Car-dioGenomics microarray data base. 4 In that study, ascending aortic constriction led to a 40% increase in heart body weight ratio compared with control after 1 week (Fig. 2B, top) (i.e. when marked ANF induction became detectable). 4 Importantly, a marked decrease in RGS2 mRNA was observed within 48 h that persisted for up to 8 weeks (Fig. 2B,  middle). In contrast, RGS5 mRNA was not significantly altered in ventricles from banded mice at any time point (Fig. 2B, bottom). Probes for RGS3 and RGS4 mRNA were represented at the arrays, but their levels were either below or too close to detection limit, respectively.
RGS2 RNAi Target Sequence Identification-To determine the functional significance of selective down-regulation of endogenous RGS2 protein, we scanned the entire RGS2 coding sequence for potential RNAi target sequences (Fig. 3A) and subcloned them between a human H1 RNA polymerase III promoter and a RNA polymerase III terminator sequence for in vivo shRNA synthesis (see "Experimental Procedures" for details).
Because of low expression of endogenous RGS2 in most cell types, potential target sequences were first screened for their effectiveness in RGS2 silencing in HEK-293 cells that were co-transfected with FLAG-tagged rat RGS2 (F-RGS2). Two of them (sequences 3 (Fig. 3B) and 4 (data not shown)) reduced exogenous F-RGS2 below the detection limit by Western blot analysis. F-RGS2 levels were largely comparable in the presence or absence of the other four constructs but significantly reduced when compared with cells that had been co-transfected with plasmids encoding negative control and GAPDH shRNAs. For reasons still unknown, RGS2 expression was increased in the presence of either control plasmids, whereas GAPDH was not affected by the negative control plasmid. Consistent with a ϳ50% transfection efficiency in HEK-293 cells under the conditions used (Fig. 3C), GAPDH pSilencer reduced endogenous GAPDH expression by approximately half (Fig. 3,  B and D). RGS2 silencing was dose-dependent and detectable by immu-nostaining (Fig. 3C), indicating that RGS2 shRNA was efficiently expressed and correctly processed intracellularly to generate mature and functional siRNA. Effective RGS2 suppression could also be achieved through direct introduction of synthetic siRNAs for sequences 3 and 4 (data not shown), eliminating the need for intracellular shRNA processing. Importantly, the expression of RGS3-RGS5 was largely unchanged (Fig. 3D).
RNAi to Endogenous RGS2 in Myocytes-After optimizing myocyte/ siRNA/lipid carrier ratios, RGS2 RNAi was also achieved in cardiomyocytes using traditional cationic lipid-based transfection. Fig. 4A illustrates dose-dependent uptake of fluorescently labeled RGS2 siRNA in NVM 24 h after transfection that was still detectable after 4 days (data , and RGS5 mRNA expression in wild type mice. *, p Ͻ 0.05 versus 2 day-old (set at 1) for each RGS. B, Northern blot of ventricular RNA from 0.5-and 2.5-monthold ␣ q *52 and wild type mice, probed with 32 Plabeled RGS2 cRNA, ANF cDNA, and 18 S oligonucleotide probes. C, RT-PCR on ventricular myocyte RNA from 2.5-month-old ␣ q *52 and wild type. D, Western blot of ventricular particulate protein (60 g/lane) from 0.5-month-old ␣ q *52 and wild type mice, probed with antibodies against RGS2 and muscle-specific caveolin-3 (Cav-3). Crude brain cortex homogenates from wild type and RGS2 knock-out mice (KO) served as controls.
not shown). RGS2 siRNA uptake was uniform and detectable in more than 90% of cells at 30 nM. In AVM, cellular uptake of FAM-labeled siRNA was also observed but difficult to ascertain due to greater autofluorescence (data not shown). Neither neonatal (Fig. 4A) nor adult (Fig. 4B) myocytes showed apparent changes in cell viability or morphology upon siRNA transfection.
The effectiveness of both RGS2 siRNAs in suppressing endogenous RGS2 was examined by RT-PCR (Fig. 5A) and real time PCR (Fig. 5B). RGS2 sequence 4 caused a concentration-dependent knockdown of RGS2 mRNA (by up to 90% of control at 50 nM) in both NVM and AVM, without affecting GAPDH expression. The effect of RGS2 sequence 3 was comparable (as shown for NVM) (Fig. 5A). RGS2 knock-down was observed for up to 6 days (data not shown).
Importantly, RGS3, RGS4, and RGS5 mRNA expression were largely unchanged compared with control in both NVM and AVM transfected with RGS2 sequence 4 (Fig. 5C) and sequence 3 (data not shown). Real time PCR confirmed continued expression of RGS5 in both NVM (0.9 Ϯ 0.2-fold negative control; n ϭ 3) and AVM (1.4 Ϯ 0.3-fold; n ϭ 3) upon RGS2 RNAi. FAM TM -labeled Taqman MGB probes (Applied Biosystems) for rat RGS3 and RGS4 have not yet worked in our hands (see "Experimental Procedures"). Taken together, these results suggest that endogenous RGS2 expression could be effectively and specifically suppressed by transient siRNA transfection in cultured cardiomyocytes.
Effect of RNAi to Endogenous RGS2 on G q/11 -mediated Signaling-To minimize the risk for off target effects, the lowest siRNA concentration causing homogenous and widespread RGS siRNA uptake and marked RGS2 suppression (30 nM) was chosen for all subsequent functional experiments. First, we examined the response of PLC␤, the main direct downstream target for G q/11 , to increasing concentrations of two different G q/11 -coupled receptor agonists in myocytes that had been transfected with RGS2 sequence 4 or negative control siRNA for 48 h. Silencing of endogenous RGS2 increased the maximal effect of agonistinduced PLC␤ activation in both NVM (Fig. 6A) and AVM (Fig. 6B), whereas basal PLC␤ activity was not altered. The PLC␤ activity-enhancing effects of RGS2 siRNA sequences 3 and 4 were comparable (Fig.  6C) and could not be attributed to changes in the expression of endogenous G␣ q/11 , PLC␤ 1 , or PLC␤ 3 (Fig. 6D).
We then assessed whether or not silencing of endogenous RGS2 also leads to enhanced G q/11 -mediated activation of MAPKs using phospho-  rylation-specific antibodies. As expected, short term stimulation of NVM (5 min for ERK1/2 and p38 or 20 min for JNK) with PE/Tim or ET-1 resulted in increased phosphorylation of ERK1/2, p38, and JNK(p54), suggesting increased activation of each of the three MAPK branches in response to GPCR-induced G q/11 activation (Fig. 7). Importantly, the extent of MAPK phosphorylation was largely comparable in cells that had been transfected with RGS2 sequence 4 siRNA compared with negative control siRNA. Only in PE/Tim-treated cells, phosphorylated JNK(p54) appeared to be slightly decreased (similar results were obtained for JNK(p46), although total expression of this isoform was difficult to detect (data not shown)). Total ERK1/2, p38, and JNK(p54) expression levels were comparable at each experimental condition (Fig.  7). Silencing of endogenous RGS2 mRNA to 0.16 Ϯ 0.02-fold negative control (n ϭ 4, p Ͻ 0.05) was confirmed by real time PCR in transfection experiments conducted in parallel. Taken together, these findings suggest that, in contrast to PLC␤, G q/11 -mediated MAPK activation was not enhanced upon RGS2 silencing.
Effect of RNAi to Endogenous RGS2 on G q/11 -mediated Hypertrophy-We next assessed the hypertrophic response that NVM typically assume upon G q/11 activation (28) in cells with targeted knockdown of endoge-   nous RGS2. In these experiments, PE/Tim and ET-1 doses were used that led to a modest increase in cell size (Fig. 8, A and B) and radiolabeled protein content (Fig. 8C) in control cells. Although there was no significant difference under base-line conditions in RGS2 siRNA sequence 4-transfected NVM compared with control cells, both G q/11 -coupled receptor agonists caused a markedly enhanced increase in cell size and protein content upon selective RGS2 down-regulation, indicating an enhanced hypertrophic response.

DISCUSSION
This study revealed that endogenous ventricular RGS2 expression is selectively reduced in two different hypertrophy models with enhanced G q/11 signaling (transgenic G␣ q * expression and pressure overload). Importantly, RGS2 down-regulation was pronounced (by Ͼ50%) and detectable prior to the onset of hypertrophy. We identified RNAi targeting sequences that are effective and selective in suppressing RGS2 expression in NVM and AVM and demonstrate that endogenous RGS2 exerts functionally important inhibitory restraint on G q/11 -mediated PLC␤ (but not MAPK) activation and hypertrophy. Together, these data suggest that loss of cardiac fine tuning of PLC␤ signaling by RGS2 down-regulation could potentially play a pathophysiological role in the development of G q/11 -mediated hypertrophy.
Early and Selective Down-regulation of RGS2 in Hearts with Enhanced G q/11 Signaling-Studies on the role of RGS protein expression changes in the pathophysiology of cardiac hypertrophy and failure are still scarce. Cardiac RGS expression has been investigated in patients with end-stage failing hearts and organ donors with acute failure (9,10) and animal models of pressure overload and heart failure (11,29). The reported findings were not uniform, presumably due to species-and/or model-specific differences, and were limited to a few RGS proteins (mainly RGS3 and RGS4) and already hypertrophied and failing hearts.
The present study identified changes in RGS2 expression that preceded the development of hypertrophy and may therefore be of pathophysiological importance for the initial events triggering hypertrophy development. Of the major endogenous cardiac RGS proteins, only RGS2 was markedly altered in its expression prior to the development of hypertrophy. RGS2 down-regulation was observed at both mRNA and protein levels and in cardiac myocytes. It was found in two independent models (G␣ q * expression and pressure overload), suggesting that it is a common feature in hearts with enhanced G q/11 signaling. In contrast, changes in RGS3, RGS4, and RGS5 expression (at end stage) were found to be not uniform in the two models.
Delineating Endogenous RGS Protein Function Using RNAi-To date, the function of RGS proteins has mainly been assessed with recombinant proteins and by heterologous overexpression, and very little is known about their role in cardiomyocyte biology. Negative regulation of cardiac G protein signaling by RGS proteins has been described (9,30), but these studies focused primarily on RGS4, the G q/11 signaling pathway, and neonatal stages of cardiac development. Using a similar type of overexpression studies, we recently demonstrated that RGS2 is a potent and selective inhibitor of G q/11 -mediated PLC␤ activation and hypertrophy in cardiac myocytes, whereas the other major cardiac RGS proteins (RGS3-RGS5) can regulate both G q/11 and G i/o signaling (31). Whereas gain-of-function studies provide valuable information about the capacity of RGS proteins to regulate cardiac G protein signaling, they do not necessarily reflect the functions of endogenous RGS. This is of particular importance in light of generally low expression levels of RGS proteins and the simultaneous presence of multiple RGS proteins with often redundant G␣ specificity. In this study, we used a novel approach to delineate the function of endogenous RGS proteins in living cells, which represents a significant extension to previously used antibodies (32) or inhibitory RGS peptides (33) targeting the RGS-G␣ interface, antisense oligonucleotides (34), or ribozymes (35).
Feasibility and Significance of RNAi in Cardiac Myocytes Using siRNA Transfection-Functional gene characterization via RNAi requires efficient siRNA introduction into cells in the absence of nonspecific or cytotoxic effects. Generally, only low transfection efficiencies can be achieved in cardiomyocytes using cationic lipids. However, we and others have observed much higher transfection efficiencies for siRNAs compared with plasmid DNAs in some cell types, which prompted us to  explore this approach in myocytes. In this study, we demonstrate homogenous cellular siRNA uptake in more than 90% of NVM and effective and selective RGS2 RNAi in both NVM and AVM without any apparent compromise in cell viability or morphology. The gene silencing effect was transient but sufficiently long to conduct experiments requiring several days of culture.
Our findings suggest that transfection of siRNA into neonatal and adult cardiomyocytes could become a useful complement to gene targeting in whole animals. It offers a less labor-and cost-intensive alternative for effective gene silencing than previously reported adenoviral approaches (such as adenoviral infection of U6 promoter-driven shRNA (36,37) and siRNA introduction via hemagglutinating virus of Japan envelope vector (38)). The transfection efficiency that can be achieved with this novel approach is superior to pSilencer electroporation (39). Recently, another group reported gene silencing via siRNA transfection (40), although this study (like the others cited above) was limited to neonatal myocytes and provided few experimental details.
Endogenous RGS2 Negatively Regulates Cardiomyocyte G q/11 -mediated PLC␤ Activation and Hypertrophy-Total inositol phosphate formation and hypertrophy in response to G q/11 -coupled receptor stimulation were markedly increased upon gene silencing of endogenous RGS2 in NVM and AVM, suggesting that RGS2 exerts functionally important inhibitory restraint on G q/11 -mediated PLC␤ activation in cardiac myocytes of different differentiation states.
In contrast to PLC␤, the immediate downstream target of G q/11 , none of the MAPK signaling branches (ERK1/2, p38, and JNK) appeared to be enhanced in response to G q/11 activation upon RGS2 silencing. ERK1/2, p38, and JNK have all been implicated to have prohypertrophic effects in cell culture-based systems (such as NVM), although more recent evidence suggests that p38 and JNK may not positively regulate hypertrophy in vivo (reviewed in Refs. [41][42][43]. Activation of MAPK signaling by G q/11 -coupled GPCRs involves a number of intermediaries (44), which could in part contribute to the apparent lack of regulation of G q/11mediated MAPK activation by endogenous RGS2. Furthermore, it is conceivable that one or more of the other cardiac RGS proteins play a significant role. These possibilities will be explored in future studies.
Two siRNAs targeting different sequences in RGS2 (sequences 3 and 4) caused a similar degree of RGS2 suppression in cardiomyocytes. They caused comparable functional effects, making nonspecific off-target effects (45) unlikely. Importantly, enhancement of G q/11 -mediated PLC␤ activation was observed despite the continued presence of RGS3-RGS5 with a similar capacity to regulate G q/11 signaling (31). Our findings do not exclude potential contributions by other proteins (such as RGS3, RGS4, or RGS5) in regulating G q/11 -mediated PLC␤ activation, but they indicate for the first time a functionally important negative regulatory effect of endogenous RGS2.
Endogenous RGS2 negatively regulated PLC␤ activation in response to two different G q/11 -coupled receptors (␣ 1 -adrenergic and endothelin-1). Recent evidence suggests that RGS proteins may not be able to freely interact with available G␣ proteins but may be selectively sorted by GPCRs at the plasma membrane (46). Binding of RGS2 to G q/11coupled GPCR sequences has been described via direct interaction (47) or a scaffolding protein (48), but neither we (Figs. 6 and 8) nor others (49,50) have observed receptor-specific regulation of G q/11 signaling by RGS2 to date. It is important to note that cellular interactions between a full-length GPCR and RGS protein have yet to be described. Screening of a larger number of receptor agonists will be required to determine if endogenous RGS2 displays receptor selectivity in cardiomyocytes and to delineate its profile.
Taken together, this study shows that endogenous RGS2 proteins play a central role in fine tuning cardiomyocyte G q/11 signaling and suggests that RGS2 down-regulation in hearts with enhanced G q/11 signaling might further exacerbate G q/11 -mediated PLC␤ activation. We previously demonstrated that RGS2 can effectively diminish G q/11 -mediated PLC␤ activation in the absence of GTPase activity (5). In G␣ q * transgenic hearts, RGS2 down-regulation therefore probably acts in concert with a previously reported up-regulation in endogenous G␣ q/11 and PLC␤ (19,20) to further enhance G q/11 signaling. In pressure overload, the crucial role of G q/11 signaling was demonstrated by a marked reduction or absence of the hypertrophic response upon heart-specific disruption of the receptor-G q/11 interface (15) or deletion of G q/11 ␣ subunits (16). It is likely that RGS2 down-regulation contributes to further exacerbation of G q/11 signaling in this setting as well.
Further work is required to address the functional significance of RGS2 down-regulation in vivo. Targeted deletion of RGS2 in mice leads to a hypertensive phenotype, presumably due to prolonged G q/11 signaling and vasoconstriction in the vasculature (51,52). Central nervous mechanisms (an enhanced sympathetic tone that could lead to a resetting of the baroreceptor reflex) appear to also contribute to blood pressure elevation in this model (53). To date, very little is known about the cardiac phenotype of RGS2KO. Only echocardiographic results at 6 months of age have been reported, which showed no signs of hypertrophy and normal systolic function (51). It is possible that the increase in blood pressure in RGS2KO is not sufficient to induce hypertrophic changes or that hypertrophic changes might develop at a later point in time. Furthermore, whether or not long term deletion of RGS2 may lead to compensatory changes in the expression of other RGS proteins has not yet been addressed in any of the studies published to date. The present study suggests that endogenous RGS2 fine tunes cardiomyocyte G q/11 signaling and function in response to receptor activation but not under base-line conditions. Therefore, the response to hypertrophic stimuli in mice lacking RGS2 warrants further investigation.
Implications-G protein-mediated signaling pathways are of fundamental importance for the regulation of cell size, differentiation, and function. As regulators of G protein signaling, RGS proteins provide an alternative approach to modulate the GPCR-G protein system. A better understanding of this important protein family and the biological function of endogenous RGS proteins under normal and pathophysiological conditions will probably provide new opportunities in therapeutic design.
Our findings implicate endogenous RGS2 as an important negative regulator of G q/11 -mediated PLC␤ activation and hypertrophy in cardiomyocytes. We propose that early and selective down-regulation of RGS2 might play a critical role in G q/11 -mediated hypertrophy via positive feedback. The question of whether changes in RGS2 expression also play a role in other forms of cardiac hypertrophy (e.g. as a result of ischemic injury or sarcomeric mutations) will be the subject of future investigation.
Enhancing the expression and/or function of RGS2 could potentially be a useful approach to modulate one of the central pathways in hypertrophy development. It is increasingly recognized that modulating negative regulators of hypertrophy could become an important strategy for heart failure treatment, because it attempts to mimic negative feedback mechanisms that are often central for maintaining cellular homeostasis (54). Cardiac hypertrophy is generally believed to be initially an adaptive response that helps normalize ventricular wall stress and maintain cardiac output but then often progresses to overt heart failure. However, increasing evidence suggests that hypertrophy may not necessarily be a compensatory event and that its inhibition may be a promising therapeutic approach (22). MARCH 3, 2006 • VOLUME 281 • NUMBER 9 JOURNAL OF BIOLOGICAL CHEMISTRY 5819