RGS3L allows for an M₂ muscarinic receptor-mediated RhoA-dependent inotropy in cardiomyocytes

We used the following primary antibodies: mouse-anti-Rac1 (BD Transd. Laboratories, 610650), mouse-anti-p190A (BD Transd. Laboratories, 610149), mouse-anti-RhoA (26C4, Santa Cruz, sc-418), mouse–anti-RGS3 (CC-Q7, Santa Cruz, sc-100762), rabbit–anti-Tiam1 (C-16, Santa Cruz, sc-872), rabbit–anti-Gβ (T-20, Santa Cruz, sc-378), rabbit–anti-RGS3 (Abcam, ab2564), mouse–anti-c-myc (Klon 9E10, Oncogene), mouse–anti-β-actin (Sigma-Aldrich, A2228), mouse–anti-RhoA-GTP (Biomol). The corresponding horseradish peroxidase-conjugated secondary antibodies were from Sigma-Aldrich (Saint Louis, MO, USA, A-9044, A-9169), the fluorescent labelled secondary antibodies (anti-mouse-Alexa Fluor® 568-antibody, anti-rabbit-Alexa Fluor^® 633-antibody) were from Life Technologies. In this study, the following reagents and inhibitors were used: FGF-2 (Promega), carbamoylcholine chloride (carbachol), 5-bromo-2′-deoxyuridine (5-BrdU) (Sigma-Aldrich), PI3K-inhibitor LY294002 (2-(4-morpholinyl)-8-phenyl-1(4H)-benzopyran-4-one hydrochloride; Alexis), pertussis toxin (PTX; Calbiochem), H1152P ((S)-(+)-2-Methyl-1-[(4-methyl-5-isoquinolinyl)sulfonyl]homopiperazine, 2HCl, Merck).

All the experiments were performed according to the EU animal experiments guidelines and ethical approval by the local German ethics committee (G237/12, Regierungspräsidium Karlsruhe).

NRCM were isolated from hearts of 1–3-day-old male and female neonatal Wistar rats as described previously [92]. Briefly, hearts were minced and subjected to serial digestion in a mixture of collagenase (0.5 mg/ml collagenase type II, Cell systems) and pancreatin (0.6 mg/ml, Sigma-Aldrich) to release single cells. The obtained cell suspension was placed on top of a Percoll gradient (GE Healthcare) to separate cardiomyocytes from other cell types. The cardiomyocyte fraction was seeded on collagen I-coated plates and cultured in DMEM supplemented with 10% (w/v) fetal calf serum, 2 mM l-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin in a humidified atmosphere of 5% CO₂ at 37 °C. 0.1 mM 5-BrdU was used to prevent overgrowth of other, non-cardiomyocyte cell types. The cells were used for experiments between 3 and 5 days after isolation. Serum-reduced conditions (DMEM supplemented with 0.5% FCS) were used for 48 h when indicated.

HEK-293 cells were grown in Dulbecco’s modified Eagle's medium (DMEM, Invitrogen) supplemented with 10% (w/v) fetal bovine serum, 2 mM l-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin in a humidified atmosphere of 5% CO₂ at 37 °C. Transient overnight transfection of the cells with p190RhoGAP, full length RGS3L-N460A, RGS3L or truncated RGS3L constructs were performed using Polyfect (Qiagen) transfection reagent according to the manufacturer's instructions. Cells were incubated under serum-reduced conditions and investigated 24–48 h after transfection. The measurement the RacGAP and RhoGAP activity of p190RhoGAP was performed using a pulldown assay described below.

The HEK-293 cells were from Agilent and routinely monitored for possible mycoplasma contamination.

The coding sequences of RGS3L and RGS3L-N460A were subcloned into the adenoviral shuttle vector pAdTrack-CMV (a gift from Dr. B. Vogelstein, Baltimore, MD). The cDNA oligonucleotide 5′-ATCCCGGAAGGAATCCTTTTCAGGTTCAAGAGAACTGAAAAGGAT-TCCTTCCTTTTTGGAAA-3′ encoding the RGS3L-shRNA sequence was subcloned into the BglII/HindIII sites of the pShuttle H1 vector (Stratagene). For knockdown of Tiam1 the cDNA oligonucleotide 5′GATCCCGCGAGCTTTAAGAAGAAACTTCAAGAGAGTTTCTTCTTAA-AGCTCGCTTTT-GGAAA-3′ encoding Tiam1-shRNA was used. The complete H1 promoter driven expression cassette was subcloned into pAdTrack [12]. After recombination of the created shuttle vectors with pAdEasy-1, the Pac I linearized recombinant adenoviral genome was transfected and amplified in HEK-293 cells.

The cDNA sequences of the variant/mutant RGS3L-N460A was cloned into a single-stranded AAV genome plasmid (pSSV9-CMV-MLC1500-luc) via XbaI restriction replacing the luciferase (luc) gene. The correct fragment size and orientation was controlled by agarose gel-electrophoresis and sequencing resulting in pSSV9-CMV-MLC1500-RGS3L-N460A. AAV9 vectors were generated by double-transfection of helper plasmid pDP9rs [79] and either pSSV9-CMV-MLC1500-RGS3L-N460A or -luciferase (as control). Vectors were then purified using iodixanol step gradient ultracentrifugation followed by genomic titer determination using quantitative real-time PCR as described previously [31].

Cell lysates (about 50 µg protein) were mixed with 70 µl rehydration buffer and applied to 7 cm IPG gel strips (GE Healthcare) containing a linear 3–10 pH gradient. Isoelectric focusing was carried out using an Ettan IPGphor unit (GE Healthcare). The subsequent SDS–PAGE was performed on 8% polyacrylamide gels.

For immunoblot analysis the protein samples were separated by SDS–PAGE using 8–15% denaturating acrylamide gels and transferred onto nitrocellulose membranes. Membranes were blocked with Roti-Block (Carl Roth) for 1 h at room temperature, and incubated with specific primary antibodies (see above) overnight at 4 °C and according to the manufacturer’s recommendations. After incubation with appropriate secondary antibodies for 1 h, proteins were visualized by enhanced chemoluminescence using an imaging system (Alpha Innotech). Image J Software was used for the analysis of the blots.

The cellular Rac1-GTP and RhoA-GTP levels were measured with a pull-down assay using GST fusion proteins containing the Rho-binding domain of rhotekin (GST–RBD) or Rac-binding domain of p21-activated kinase (PAK1) (GST–PBD). GST–RBD and GST–PBD were expressed and purified from E. coli. NRCM were stimulated with FGF2 or transduced with adenoviruses and pretreated with the different inhibitors described above. After activation of the NRCM with carbachol (1 mM, 5 min), cells were lysed in ice-cold GST-Fish buffer [11, 87] and pelleted by centrifugation (12,000 rpm, 3 min at 4 °C). The GTP-bound RhoGTPases contained in the supernatant was incubated for 1 h at 4 °C with either GST–PBD bound to magnetic glutathione–sepharose beads or GST–RBD bound to nonmagnetic glutathione–sepharose beads. The beads were separated using a magnetic plate and centrifugation. After twice washing of the beads, bound proteins were eluted with sample buffer and separated by SDS–PAGE. The amounts of activated GTPases and total GTPases from lysates as loading control were then determined by immunoblot analysis as described above.

NRCM were cultured on 96-well plates (Sarstedt). The cells were transduced with shEGFP, shTiam or RGS3L–N460A encoding adenoviruses, and treated with FGF2 or PTX as described above. Thereafter, the cells were stimulated with 1 mM carbachol or solvent for 5 min at 37 °C. Myocytes were then loaded for 30 min with 5 µM 5(6)-carboxy-2′,7′-dichlorofluorescein (DCF-DA, Sigma-Aldrich) at 37 °C. After washing the cells with HEPES buffered salt solution (HBSS; 25 mM HEPES pH 7,4, 120 mM NaCl, 5,4 mM KCl, 1,8 mM CaCl₂, 25 mM NaHCO₃, 15 mM Glucose), DCF fluorescence was measured using the Envision 2102 Multilabel Reader with a set of FITC filters (excitation 485 ± 10 nM and emission at 535 ± 20 nM, PerkinElmer).

NRCM were transduced with EGFP (control) or RGS3L–N460A encoding adenovirus for 24 h, treated with 100 nM H1152P for 1 h as indicated and stimulated with 1 mM carbachol or solvent for 90 s at 37 °C. Thereafter, the cells were lysed in a buffer containing phosphatase inhibitor PhosStop (Roche Applied Science). The lysates were cleared by centrifugation and subjected to SDS–PAGE and immunodetection. Finally, densitometric band intensities of pMLC-2 were quantitated using total MLC-2a as control.

NRCM were transduced with RGS3L–N460A adenovirus for 24 h and stimulated with 1 mM carbachol or solvent for 5 min at 37 °C. Co-immunoprecipitation was performed as described previously [5]. Briefly cells were lysed with immunoprecipitation buffer (50 mM Tris–HCl, pH 7.4, 2 mM EDTA, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 10 mM NaF) containing 1 mM sodium orthovanadate, 1 mM Pefablock, 10 µg/ml aprotinin, 10 µg/ml leupeptin). After centrifugation the cleared lysates were incubated with the indicated antibodies (2 μg) under agitation for 1 h at 4 °C. After addition of 40 μl 1:1 (v/v) protein-A–sepharose-beads (Amersham Biosciences), the mixture was gently shaken for an additional 3–4 h at 4 °C. Beads were washed three times with immunoprecipitation buffer and eluted in SDS-containing buffer for 5 min at 95 °C. After SDS–PAGE and transfer to nitrocellulose membranes, immunoprecipitated proteins were detected by Western blot analysis using the indicated antibodies according to standard protocols. Final detection was done with an ECL system (Amersham), band intensity was quantified with ImageJ-software.

Proximity ligation assay was performed by following the manufacturer’s protocol (Olink Bioscience, Uppsala, Sweden). Briefly, NRCM seeded on collagen-coated coverslips in 12-well dishes were stimulated with or without carbachol for 5 min, washed 3 times with PBS, fixed in 4% formaldehyde in PBS for 10 min, permeabilized in 0.02% Triton-X-100 for 10 min, and blocked with Blocking Solution (Olink) for 30 min in 37 °C. After incubation with the indicated primary antibodies (mouse–anti-p190RhoGAP, rabbit–anti-RGS3) over night at 4 °C, wells were washed with Wash Buffer (Olink), incubated with PLA Probe anti-mouse plus and PLA Probe anti-rabbit minus for 1 h at 37 °C. After washing, a ligation and amplification step followed using the manufacturer’s protocol and reagents. Cells were mounted with Duolink In Situ Mounting Medium with DAPI, dried in room temperature and visualized by confocal microscopy (Leica, Germany). The LAS-X software was used for image processing. Quantification of fluorescent signals was performed using Image J software.

Measurement of the functional RhoGAP activity of p190RhoGAP was performed by a pulldown assay as described previously [5, 58]. This assay is based on the principle that the functional activation of p190RhoGAP is indicated by an increased ability to associate with its substrate, the active, GTP-bound form of monomeric GTPases. The ability of p190RhoGAP to act as a RhoAGAP was detected using constitutively active RhoAQ63L-GST coated beads. The more p190RhoGAP protein binds to the beads, the higher is its RhoAGAP activity (corresponding to a decreased level of RhoA-GTP in the cell [58]). Using RacQ61L-GST-coated beads instead the Rac1GAP-activity of p190RhoGAP can be measured. Briefly, cells (NRCM or HEK-293 cells) were washed carefully with PBS and lysed in a buffer containing 50 mM Tris–HCl, pH 7.4, 10 mM MgCl₂, 150 mM NaCl, 1 mM DTT, 1% Triton X-100, 10 g/ml each of aprotinin and leupeptin, 1 mM Pefabloc and 1 × Phosstop (Roche). After centrifugation for 3 min at 12,000×g, the supernatants were incubated for 60 min at 4 °C with glutathione–sepharose magnetic beads coated with RhoQ63L or RacQ61L conjugated with GST and purified previously from Rosetta E. coli bacteria. The beads were separated using a magnetic plate. After three times washing of the beads, bound proteins were eluted with sample buffer and separated by SDS–PAGE. p190RhoGAP was detected by immunoblotting.

Subconfluent NRMC were cultured on 12 well plates (Sarstedt), washed three times with PBS and fixed with 3% paraformaldehyde/PBS for 15 min at room temperature. After the cells were treated with 0.05% Triton-X-100 for 3 min at room temperature and 0.5% bovine serum albumine (BSA) for 45 min at 4 °C, they were incubated with the appropriate antibodies for 16 h at 4 °C. After washing with PBS, the cells were incubated with the indicated secondary antibodies for 1 h at room temperature. Images of the cells mounted at room temperature in PBS were acquired using fluorescence microscopy (Olympus IX 81).

Measurement of the RhoA-activity in AVMCM was performed by immunofluorescence staining using anti-RhoA–GTP-antibody, microscope image acquisition, and quantification of fluorescence. Cells isolated from mouse hearts were seeded on collagen-coated coverslips in 12-well dishes in Modified Eagle Medium (MEM) supplemented with 1% BSA, 100 units/ml penicillin and 100 μg/ml streptomycin, 1% l-glutamine, 1% (−) blebbistatin, 1% ITS (insulin–transferrin–selenium) in a humidified atmosphere of 5% CO2 at 37 °C. Cells were transduced with Myc-tagged RGS3L–N460A–adenovirus or control EGFP–adenovirus for 48 h. After stimulation with or without 1 mM carbachol for 5 min, cells were fixed in 4% formaldehyde in PBS for 10 min, permeabilized and blocked for 10 min in a blocking buffer containing 10% FCS, 0.2% Triton-X 100 in Ca²⁺- and Mg²⁺-free PBS. Incubation with primary antibodies (mouse–anti-RhoA–GTP-antibody 1:500, rabbit–anti-c-myc-antibody 1:100) was performed in blocking buffer over night at 4 °C. After washing with PBS, cells were incubated with the corresponding fluorochrome-labeled secondary antibodies (anti-mouse-Alexa Fluor 568-antibody 1:1000, anti-rabbit-Alexa Fluor 633-antibody 1:1000) over night at 4 °C, and washed thereafter with PBS. DAPI (1:1000 in PBS for 1 h) was used to detect nuclei. After washing with PBS and mounting, coverslips were dried at room temperature. Cells were visualized by confocal microscopy (Leica, Germany), software LAS-X was used for image processing. Analysis of the fluorescence intensity of the Alexa Fluor 568 conjugate was performed using Image J (Version 1.47v).

Neonatal rat engineered heart tissues (EHT) were generated and analyzed as previously reported [24]. In brief, agarose casting molds were prepared in 24 well plates (NUNC) with liquid agarose (2% (w/v), PBS) and polytetrafluorethylen (PTFE) spacer. After agarose solidification PTFE spacer were removed and polydimethylsiloxan (PDMS) were placed on the 24 well plate so that pairs of flexible PDMS posts reach into each agarose casting mold. Dissociated neonatal rat heart cells were re-suspended in a fibrinogen solution (5 mg/ml). 97 µl were mixed with 3 µl thrombin aliquot (100 U/ml) and pipetted into a casting mold. This pipetting step was repeated for each EHT. Fibrin polymerization took place in an incubator (2 h) and PDMS racks with fibrin gels attached to the PDMS posts were transferred to new 24 well plates. EHT were cultivated in maintenance media (DMEM, Horse serum 10% (v/v), insulin 10 μg/ml, aprotinin 33 μg/ml) with media replacement on Mondays, Wednesdays and Fridays. After development (days 25–30) EHT were transferred to modified Tyrode’s solution containing 0.1 mM free Ca²⁺ (120 mM NaCl, 22.6 mM NaHCO₃, 5.4 mM KCl, 5 mM glucose, 1 mM MgCl₂, 0.4 mM NaH₂PO4, 0.1 mM CaCl₂, 0.05 mM Na₂EDTA, 25 mM HEPES) and force analysis was performed by video-optical recording of EHT shortening as recently described [24].

The care and experimental use of all animals in this study were in accordance with institutional guidelines and approved by the local ethics committee (Regierungspraesidium Karlsruhe).

We injected 6-week-old male Wistar rats with AAV9–CMV–MLC1500–RGS3L–N460A-virus or with an AAV9–CMV–MLC1500-luc control virus (10¹² vg/rat) intravenously. After 2 months, the rats were anaesthetized (2–3% isoflurane flow) and subsequently euthanized by cervical dislocation. The hearts were harvested and mounted on a Langendorff rig perfused with a relaxing solution containing 118.3 mM NaCl, 3 mM KCl, 0.2 mM CaCl₂, 4 mM MgSO₄, 2.4 mM KH₂PO₄, 24.9 mM NaHCO₃, 10 mM glucose and 2.2 mM mannitol [77]. The left ventricle was exposed and posterior left ventricular papillary muscles and strips of left ventricles (approx. 1 mm diameter) were excised and mounted in organ baths (Radnoti organ bath System, ADInstruments) containing the relaxing solution (32 °C) as described above and oxygenated (95% O₂, 5% CO₂).

After mounting, the strips were stretched to ~ 3 mN diastolic tension, allowed to relax ~ 5 min and then the relaxing solution was replaced with a solution of identical composition with the exception of 1.8 mM CaCl₂ and 1.2 mM MgSO₄ concentrations. The muscle strips were field stimulated at a frequency of 1 Hz with impulses of 5 ms duration and current about 10–20 mA. The isometrically contracting muscles were stretched to the maximum of their length–tension curve. Contraction–relaxation cycles were recorded and analyzed as previously described [75, 76]. Basal contractility was expressed as maximal developed force (F_max, mN). The descriptive parameters at the end of the equilibration period were used as basal (control) values. Inotropic responses were expressed as changes in the maximal development of force ((dF/dt)_max). The measurements were based on averaging 10–20 contraction–relaxation cycles. Antagonists of α1-adrenoceptors (prazosin 1 μM) and β-adrenoceptors (timolol 1 μM) were added 90 min before the muscarinic agonist carbachol which was added to the organ bath as a bolus (100 μM). The contractility measurements were conducted at increasing extracellular Ca²⁺-concentrations (ranging from 1.8 to 3.0 mM).

FGF2 increases RGS3 expression in neonatal rat cardiomyocytes (NRCM) [96] and the level of RGS3L determines the M₂R coupling to Rac1 or RhoA in H10 cells [87]. Therefore, we analyzed RGS3 expression with and without FGF2 treatment and its effect on RhoGTPase activation in NRCM. As established before [87], we used 2D gel electrophoresis with subsequent immunoblotting to identify specific isoforms of RGS3. RGS3L was detected at the predicted isoelectric point (pI) and molecular mass of 4.79 and 61 kDa, respectively (Fig. 1A). In accordance with the published data [91], a marked increase in the amount of RGS3L protein was detected in NRCM treated with FGF2 for 24 h. Next, we examined the activation of Rac1 and RhoA in FGF2-treated NRCM in response to carbachol. Under control condition carbachol caused a prominent Rac1 and a weak RhoA activation, whereas in FGF2-treated cells the opposite was found (Fig. 1B, C). Both the carbachol-induced Rac1 and RhoA activations were reduced by treatment with PTX or the phosphoinositide-3-kinase (PI3K) inhibitor LY294002 independent of FGF2 (Supplemental Fig S1A). These data are in accordance with the G_i- and PI3K-dependent RhoA and Rac1 activation observed in carbachol-treated NRCM-derived H10 cells and thus indicate the existence of a similar signaling pathway in primary NRCM [87].

To verify the dependence between the expression level of RGS3L and the coordinated activation of either RhoA or Rac1 in NRCM, we altered RGS3L expression using different experimental approaches. First, we transduced NRCM with an adenovirus encoding the GAP-deficient RGS3L mutant RGS3L–N460A, which mediated in M₂R expressing cells a prolonged carbachol-induced RhoA activation compared to wild-type RGS3L [87]. Increasing expression of RGS3L–N460A in NRCM induced a pronounced RhoA activation in the absence and presence of carbachol (Fig. 1D). In addition, the increasing expression of RGS3L–N460A preferentially suppressed the carbachol-stimulated Rac1 activation (Fig. 1E). Second, loss of function experiments were performed by transducing NRCM with an adenovirus encoding a shRNA directed specifically against RGS3L (Ad-shRGS3L) [15]. As shown in Fig. 1F, the expression of RGS3L was nearly abolished in the Ad-shRGS3L-transduced NRCM compared to cells transduced with a control adenovirus encoding for a shRNA against EGFP. Gβ, which served as a loading control, remained unaffected. Most importantly, RGS3L depletion reduced the carbachol-induced RhoA activation to control level (Fig. 1G). This effect was dependent on the virus load used for NRCM transduction (Supplemental Fig S1B). In contrast, the carbachol-induced Rac1 activity remained unchanged. In summary, these data demonstrate that the expression level of RGS3L dynamically regulates RhoA activation in NRCM and determines whether Rac1 or RhoA activation prevails after carbachol application.

In the neonatal rat heart as well as in an experimental heart failure model, the M₂R-induced inotropic response was sensitive to ROCK inhibition and accompanied by increased MLC-2 phosphorylation [30, 40]. In NRCM, both the ventricular isoform MLC-2v and the atrial isoform MLC-2a are present and both localize along the myofilaments (Fig. 2A). To study the influence of RGS3L–N460A on the carbachol-induced MLC-2 phosphorylation, we used a polyclonal antibody against MLC-2a-P, which recognizes mono- and diphosphorylated (Ser22 and/orSer23) MLC-2a. Treatment with carbachol in control-transduced NRCM slightly increased MLC-2 phosphorylation, which was, however, enhanced by the expression of RGS3L–N460A (Fig. 2B). The quantification of the carbachol-induced MLC-2 phosphorylation revealed a 33% increase in the presence of RGS3L–N460A compared to the non-stimulated cells (Fig. 2C). Treatment of the cells with the selective ROCK inhibitor H1152P completely blocked the carbachol-induced MLC-2 phosphorylation in the presence of RGS3L–N460A (Fig. 2B, C).

To study whether the increase in ROCK-mediated MLC phosphorylation is associated with an increase in contractility, rat engineered heart tissues (EHT) were prepared from neonatal rat heart cells. A shown in Fig. 2D, the EHT express detectable levels of RGS3L. In contractility measurements a moderate but statistically significant increase in contractile force was observed after carbachol stimulation (Fig. 2E). This inotropic response was blunted by adding the ROCK inhibitor H1152P (Fig. 2E), which is in accordance with the described carbachol-induced inotropy of the neonatal rat heart [40].

NRCMs are a widely used model to study cardiomyocyte signaling, nevertheless, they largely differ from fully differentiated adult cardiomyocytes with regard to myofilament and subcellular organization. Therefore, we studied whether RGS3L–N460A overexpression could also evoke RhoA activation in isolated adult mouse ventricular cardiomyocytes (AMVCM) in which RhoA activation was visualized by confocal microscopy using a specific anti-RhoA–GTP antibody. In addition, the expression of RGS3L–N460A was monitored by immunofluorescence (Fig. 3A, B). Interestingly, we observed that RhoA–GTP and RGS3L–N460A co-localized at the sarcolemma of AMVCM. Moreover, carbachol treatment did not alter the RhoA–GTP level in control transduced AMVCM, but significantly increased it in RGS3L–N460A expressing AMVCM (Fig. 3C, D). These data demonstrate that higher RGS3L expression levels allow for a M₂R-induced RhoA activation also in adult cardiomyocytes.

We, therefore, aimed to verify that high expression levels of RGS3L in cardiomyocytes allow for a carbachol-induced inotropic response in the healthy ventricular tissue. We injected 6-week-old male Wistar rats either with AAV9–CMV–MLC1500–RGS3L–N460A virus (AAV–RGS3L–N460A) or AAV9–ssCMV–MLC1500-luc control virus (AAV-luc). The animals were housed for a further 2 months prior to experimentation. They did not show any signs of distress and none died. After sacrifice, the hearts were explanted and cardiac contractility was measured ex vivo in ventricular muscle strips (Fig. 4). In line with the data reported before from failing rat hearts [30], carbachol initially decreased the contractility of control as well as RGS3L–N460A expressing muscle strips in a transient manner (Fig. 4C, D). In the RGS3L–N460A-expressing ventricular muscle strips, the return to baseline was accelerated, followed by an inotropic response significantly exceeding the baseline (Fig. 4D). This inotropy was more prominent at higher Ca²⁺ concentrations and later timepoints (Fig. 4D–F). Plotting of the expression level of RGS3L–N460A vs. the measured force of contraction of the individual muscle strip revealed a statistically significant positive correlation between both parameters (Fig. 4G). These results suggest that the expression of RGS3L is indeed linked to the contractile response.

RGS3L displays no GAP activity towards monomeric GTPases and thus likely influences the activity of RhoGTPases by interacting with either GEFs or GAPs. However, it was shown that its interference with G_i-mediated Rac1 activation depends on its capacity to bind Gβγ dimers [87]. Previous studies reported that in NRCM the G_iβγ/PI3K-induced activation of Rac1 via G_i-protein-coupled receptors is mediated by the guanine nucleotide exchange factor Tiam1 [21, 85]. To investigate the role of Tiam1 in the carbachol-induced Rac1 as well as RhoA activation, we transduced NRCM with an adenovirus encoding a Tiam1-specific shRNA and studied the activation of RhoA and Rac1. As shown in Fig. 5, the expression of Tiam1 was largely reduced compared to cells transduced with a control shRNA encoding virus. The downregulation of Tiam1 abolished the carbachol-induced Rac1 activation (Fig. 5B) as well as the carbachol-induced ROS production (Fig. 5C), which likely originates from the Rac1-dependent activation of NADPH oxidase in NRCM [1]. Consistent with this interpretation, the carbachol-induced production of ROS was sensitive to PTX treatment, RGS3L–N460 overexpression, and FGF2, all suppressing the carbachol-induced Rac1 activation (Fig. 5D). In line with these data, the phenylephrine (PE)-induced protein synthesis, which requires a Giβγ- and PI3K-dependent, Tiam1-mediated Rac1 activation [85], was suppressed by the overexpression of RGS3L–N460A (Fig. 5E). These data suggest that RGS3L could additionally protect the heart not only by reducing the Rac1 activity induced by carbachol, but also by interfering with other Rac1-dependent pathways involved in the development of cardiac hypertrophy. This interpretation is further supported by data showing that the hypertrophy induced by the inflammatory stimulus PGE2 through a Tiam1/Rac1-dependent activation of the transcription factor MEF2 could also be suppressed by RGS3L [83]. Interestingly, the depletion of Tiam1 also abolished the carbachol-induced RhoA activation in the presence of RGS3L–N460A (Fig. 5B), indicating that Tiam1 is also essential in this pathway. Nevertheless, Tiam1 is a bona fide RacGEF. An exchange activity against RhoA has never been reported in a cellular context but it is a well-known protein–protein interaction partner integrating a variety of signaling proteins [7, 50]. As GAPs are as important for Rac1 and RhoA signaling as GEFs, we postulated that p190RhoGAP, which was reported to alter its substrate specificity from RhoA to Rac1 under certain conditions [39, 41, 46], is part of a putative complex and might be responsible for the observed changes in the presence of RGS3L. We, therefore, performed reciprocal co-immunoprecipitation experiments in NRCM transduced with the RGS3L–N460A encoding adenovirus. In line with our hypothesis, p190RhoGAP could be co-immunoprecipitated from NRCM lysates with an anti-RGS3 antibody. Similarly, RGS3L–N460A was co-immunoprecipitated with the anti-p190RhoGAP antibody. Treatment of the NRCM with carbachol before cell lysis increased the amount of the co-precipitated interaction partner by two-to-threefold (Fig. 6A–D). As a weak carbachol-induced RhoA activation also occurred in NRCM in a RGS3L sensitive manner (see Fig. 1E), we studied the p190RhoGAP/RGS3L interaction also in non-transduced NRCM. As these conditions are below the detection level of the co-immunoprecipitation assay, we performed a more sensitive proximity ligation assay (PLA) in NRCM. Thereby, we detected a complex formation already under basal conditions. In accordance with the co-immunoprecipitation assays, carbachol significantly increased the interaction between both proteins, indicating that the carbachol-induced p190RhoGAP/RGS3L complex formation occurs also in native conditions (Fig. 6E, F).

The specificity of a GAP can be monitored by binding to constitutively active monomeric GTPases [58]. Therefore, we studied whether RGS3L expression can alter the substrate specificity of overexpressed p190RhoGAP in HEK cells, the system in which the RGS3L-dependent switch from Rac1 to RhoA was originally characterized [87]. Co-transfection of RGS3L–N460A with p190RhoGAP reduced the binding of p190RhoGAP to the constitutively active RhoAQ63L mutant (Fig. 7A, B) and reciprocally increased the binding to the constitutively active Rac1Q61L mutant (Fig. 7C, D), supporting our hypothesis of a substrate switch of p190RhoGAP. Like the RGS3L–N460A mutant, wild-type RGS3L, but not the N-terminally truncated isoform RGS3S is able to confer RhoA activation [87]. To test whether the switch in GAP activity of p190RhoGAP follows a similar pattern, we co-transfected HEK-293 cells with p190RhoGAP together with RGS3–N460A or RGS3S encoding constructs (Fig. 7E). In accordance with its ability to confer RhoA activation, the change in the GAP activity of p190RhoGAP was only observed in the presence of wild-type RGS3L and its N460A mutant, but not when RGS3S was expressed. This clearly demonstrates that the extended N-terminus of RGS3L is not only required for the Gβγ binding [87], but is also a requisite for the regulation of the GAP activity of p190RhoGAP (Fig. 7F). Next, we determined whether carbachol could change the substrate specificity of endogenously expressed p190RhoGAP for RhoA and Rac1 in the presence of RGS3L in NRCM. Without RGS3L–N460A expression, p190RhoGAP bound predominantly to constitutively active RhoA (Fig. 7G). In contrast, when RGS3L–N460A was expressed, carbachol reduced the binding of p190RhoGAP to RhoAQ63L and increased the binding to Rac1Q61L (Fig. 7G– - J). The switch was detectable after 3 min and more prominent at 5 min (Supplemental Fig S2). Taken together these data clearly indicate that the interaction of RGS3L and p190RhoGAP is essential for re-balancing the activity of Rac1 and RhoA in response to carbachol in cardiomyocytes.