Scholarly article on topic 'Differential regulation of RhoA-mediated signaling by the TPα and TPβ isoforms of the human thromboxane A2 receptor: Independent modulation of TPα signaling by prostacyclin and nitric oxide'

Differential regulation of RhoA-mediated signaling by the TPα and TPβ isoforms of the human thromboxane A2 receptor: Independent modulation of TPα signaling by prostacyclin and nitric oxide Academic research paper on "Biological sciences"

CC BY
0
0
Share paper
Academic journal
Cellular Signalling
OECD Field of science
Keywords
{"Thromboxane receptor" / RhoA / "Nitric oxide" / Prostacyclin / "Protein kinase" / "Heterologous desensitization"}

Abstract of research paper on Biological sciences, author of scientific article — Katarina Wikström, David J. Kavanagh, Helen M. Reid, B. Therese Kinsella

Abstract In humans, thromboxane (TX) A2 signals through the TPα and TPβ isoforms of the TXA2 receptor that exhibit common and distinct roles. For example, Gq/phospholipase (PL)Cβ signaling by TPα is directly inhibited by the vasodilators prostacyclin and nitric oxide (NO) whereas that signaling by TPβ is unaffected. Herein, we investigated whether TPα and/or TPβ regulate G12/Rho activation and whether that signaling might be differentially regulated by prostacyclin and/or NO. Both TPα and TPβ independently regulated RhoA activation and signaling in clonal cells over-expressing TPα or TPβ and in primary human aortic smooth muscle cells (1° AoSMCs). While RhoA-signaling by TPα was directly impaired by prostacyclin and NO through protein kinase (PK)A- and PKG-dependent phosphorylation, respectively, signaling by TPβ was not directly affected by either agent. Collectively, while TPα and TPβ contribute to RhoA activation, our findings support the hypothesis that TPα is involved in the dynamic regulation of haemostasis and vascular tone, such as in response to prostacyclin and NO. Conversely, the role of TPβ in such processes remains unsolved. Data herein provide essential new insights into the physiologic roles of TPα and TPβ and, through studies in AoSMCs, reveal an additional mode of regulation of VSM contractile responses by TXA2.

Academic research paper on topic "Differential regulation of RhoA-mediated signaling by the TPα and TPβ isoforms of the human thromboxane A2 receptor: Independent modulation of TPα signaling by prostacyclin and nitric oxide"

Cellular Signalling 20 (2008) 1497-1512

ELSEVIER

Contents lists available at ScienceDirect

Cellular Signalling

journal homepage: www.elsevier.com/locate/cellsig

Cellular

Signalling

Differential regulation of RhoA-mediated signaling by the TPa and TPß isoforms of the human thromboxane A2 receptor: Independent modulation of TPa signaling by prostacyclin and nitric oxide

Katarina Wikström, David J. Kavanagh, Helen M. Reid, B. Therese Kinsella *

UCD School of Biomolecular and Biomedical Science, UCD Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland

ARTICLE INFO ABSTRACT

In humans, thromboxane (TX) A2 signals through the TPa and TP| isoforms of the TXA2 receptor that exhibit common and distinct roles. For example, Gq/phospholipase (PL)C| signaling by TPa is directly inhibited by the vasodilators prostacyclin and nitric oxide (NO) whereas that signaling by TP| is unaffected. Herein, we investigated whether TPa and/or TP| regulate G12/Rho activation and whether that signaling might be differentially regulated by prostacyclin and/or NO. Both TPa and TP| independently regulated RhoA activation and signaling in clonal cells over-expressing TPa or TP| and in primary human aortic smooth muscle cells (1° AoSMCs). While RhoA-signaling by TPa was directly impaired by prostacyclin and NO through protein kinase (PK)A- and PKG-dependent phosphorylation, respectively, signaling by TP| was not directly affected by either agent. Collectively, while TPa and TP| contribute to RhoA activation, our findings support the hypothesis that TPa is involved in the dynamic regulation of haemostasis and vascular tone, such as in response to prostacyclin and NO. Conversely, the role of TP| in such processes remains unsolved. Data herein provide essential new insights into the physiologic roles of TPa and TP| and, through studies in AoSMCs, reveal an additional mode of regulation of VSM contractile responses by TXA2.

© 2008 Elsevier Inc. All rights reserved.

Article history:

Received 31 January 2008

Received in revised form 2 April 2008

Accepted 7 April 2008

Available online 23 May 2008

Keywords:

Thromboxane receptor RhoA

Nitric oxide Prostacyclin Protein kinase

Heterologous desensitization

1. Introduction

The phosphorylation status of myosin light chain (MLC) of the actomyosin complex plays a central role in regulating the various types of cytoskeletal reorganizations that widely occur within the vasculature including in platelet shape change and aggregation, tonic-or agonist-induced contraction and relaxation of smooth muscle cells (SMCs), cell migration, cell proliferation and stress fibre formation [1]. Many of the physiologic regulators of platelets and vascular smooth muscle (VSM) contraction, including thromboxane (TX) A2, thrombin, ADP, prostaglandin (PG) I2 or PGD2, act through specific G protein coupled receptor (GPCR) -effector systems [1]. While agents such as TXA2 and thrombin that promote platelet activation or SMC contraction induce Gq-dependent phospholipase (PL)C| activation to evoke calcium (Ca2+) -dependent activation of myosin light chain kinase (MLCK) and MLC20 phosphorylation [1,2], they may also engage the

Abbreviations: GPCR, G protein-coupled receptor; HA, hemagglutinin; HEK, human embryonic kidney; IP, prostacyclin receptor; IP3, inositol 1,4,5-trisphosphate; NO, nitric oxide; eNOS, endothelial nitric oxide synthase; PAGE, polyacrylamide gel electrophor-esis; PG, prostaglandin; PK, protein kinase; PL, phospholipase; sGC, soluble guanylyl cyclase; SIN-1, 3-morpholinosydnonimine, HCl; TP, TXA2 receptor; TX, thromboxane; VSM, vascular smooth muscle.

* Corresponding author. Tel.: +353 1 7166727; fax: +353 1 2837211.

E-mail address: Therese.Kinsella@UCD.IE (B.T. Kinsella).

Ca2+-independent pathway involving receptor co-coupling to G12 and RhoA signalling [1]. G12 members, particularly Ga13, activate RGS (regulators of G protein signaling)-box containing members of the Rho guanine nucleotide exchange factor (RhoGEF) family, such as p115 RhoGEF, PDZ RhoGEF or LARG, to activate RhoA and its key effector in this system Rho kinase 1 (also known as p164 ROKa/ROCK2) and 2 (p160 ROK|/ROCK1), herein termed Rho kinase/ROCK [3-5]. Rho kinases, in turn, phosphorylate, and inactivate, myosin phosphatase, MLC itself and the myosin phosphatase inhibitor CPI-17 resulting in the Ca2+-independent increase in overall levels of phosphorylated MLC through a Rho/Rho kinase-mechanism [2,4,5]. Other targets of Rho kinase include its phosphorylation-dependent activation of LIM kinases which, in turn, phosphorylate and inactivate the actin depolymerizing agent cofilin [4]. The central importance of the Ca2+-independent mechanism of contraction within the vasculature has been highlighted through findings that disorders of the Rho/Rho kinase pathway are major underlying causes of hypertension, vascular spasm and atherosclerosis making Rho kinase an important therapeutic target in the treatment of these diseases [1,2,6].

The prostanoid TXA2 plays an essential role within the vasculature inducing a diversity of cellular responses including platelet shape change, secretion and aggregation, VSMC contraction and migration and is widely implicated in a number of cardiovascular disorders including thrombosis, hypertension, vessel remodelling and atherosclerotic progression [7]. As a predominantly Gq/PLC|-coupled GPCR, the TXA2

0898-6568/$ - see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2008.04.006

receptor or TP can induce both Ca2+-dependent and G12/i3-mediated RhoA/Ca2+ independent responses platelets and VSMCs [1,8]. For example, platelets from Ga13-deficient mice do not undergo RhoA-dependent shape change in response to low levels ofTXA2 but retain the ability to undergo Gq/Ca2+-dependent shape change and aggregation at higher agonist concentrations [9]. Similarly, both Ca2+-dependent/PLC(3 and Ca2+-independent/RhoA mechanisms contribute to TXA2-induced contraction in isolated bovine aortic (Ao) SMCs and in VSM tissue from various other species [10-12]. Notably however, in humans, but not in non-primates, TXA2 actually signals through two distinct TXA2 receptor isoforms termed TPa and TP(3 that arise through alternative splicing and differ exclusively in their carboxyl-terminal (C tail) domains [13-15]. Whilst it is currently unknown whether TPa or TP(3 independently or indeed differentially modulate RhoA activation and downstream signaling, there is substantial evidence that the TPa and TP(3 isoforms can differentially regulate other cellular effectors raising that possibility [16-21].

While both TPa and TP(3 are predominantly coupled to Gq/PLCp activation [22], they can independently regulate other secondary effector systems including opposite regulation of adenylyl cyclase via Gs and Gi, respectively [23]. Additionally, TPa, but not TP(3, couples to PLCS activation via Gh/tissue transglutaminase [24]. Whereas both TPs are expressed in VSMCs, TPa is the predominant isoform expressed in human platelets [25,26]. Consistent with this, in studies investigating intermolecular cross talk between the pro-aggregatory TXA2 and the inhibitory prostanoid prostacyclin (PG12), it was established that Gq/PLCp coupling and signaling by TPa, but not TP(3, undergoes prostacyclin- induced desensitization mediated through direct cAMP-protein kinase (PK) A phosphorylation of TPa at Ser329 within its unique C-tail domain [21,27]. Moreover, Gq/PLCp signaling by TPa, but not TP(3, is also desensitized by the platelet antagonist /vasodilator nitric oxide (NO), involving direct NO/cGMP-dependent PKG phosphorylation of TPa also within its unique C-tail [20] The implication from these studies is that TPa plays a critical role in vascular haemostasis acting as the major TP target for regulation/inhibition by prostacyclin and NO, such as within the anucleate platelet that predominantly expresses TPa. However, the impact of such direct inhibitory effects of prostacyclin and NO mediated by PKA and PKG, respectively, on signaling by TPa and TP(3 through other effector systems, such as through RhoA, is currently unknown but, clearly, any differential regulatory effects by either prostacyclin or NO on such TXA2 signaling may have direct clinical implications, for example within human VSMCs that express both TPa and TP(3 isoforms. Hence, the aim of the current study was to investigate whether TPa and/or TP(3 independently regulate G12/Rho activation and signaling and to establish whether that signaling is differentially regulated by the inhibitory prostacyclin/cAMP/PKA and NO/cGMP/PKG systems. These studies provide essential new insights into the physiologic roles of TPa and TP(3 and, through studies in primary human aortic smooth muscle cells (1° AoSMCs), reveal an additional mode of regulation of VSM contractile responses by the potent autocoid TXA2.

2. Materials and methods

2.1. Materials

U46619, SQ29,548, BW245C, FK409, FURA2/AM were purchased from Cayman Chemical Company; SIN-1 and Y-27632 from Calbiochem; 3F10 anti-HA, 3F10 anti-HA-horseradish peroxidase (HRP)-conjugated antibody and chemiluminescence detection kit from Roche; anti-RhoA 26C4 (Sc-418), anti-phospho-RhoASer188 (Sc-32954-R), anti-Ga,2;i3 H-300 (Sc-28588), anti-Gaq C15 (SC-392), HRP-conjugated goat anti-mouse (Sc-2005), HRP-conjugated mouse anti-goat (Sc-2354) and HRP-conjugated goat anti-rabbit (Sc-2004) antibodies from Santa Cruz; Glutathione-Sepharose 4B (GE Healthcare) and FITC conjugated goat anti-rabbit antibody from Sigma; anti-cofilin (# 3312) and anti-phospho-cofilin (phosphoSer3, # 3311) were from Cell Signaling; Alexa Fluor® 488 phalloidin (A12379; Excitation / Emission A495/518 nm) from Molecular Probes; anti-HDJ2 antibody from Neomarkers; Opti-MEM® and Oligofecta-mine® were from Invitrogen. All oligonucleotides were synthesised by Genosys

Biotechnologies; small interfering (si) RNAs by Qiagen. Cicaprost was a gift from Schering AG (Berlin, Germany). pcDNA3.1(+):hGaqQ209l-D277N, pCis:GaG?28A and pCis: Gai"3225A were from the UMR cDNA Resource Center (Gaq) or from Dr S. Offermanns, University of Heidelberg, Germany.

2.2. Cell culture and transfections

Human embryonic kidney (HEK) 293 cells were grown in minimal essential medium (MEM), 10% foetal bovine serum (FBS). HEK.TPa, HEK.TPp, HEK.TPaS329A HEK.TPaS331A and HEK.TPaS329,331A cell lines stably over-expressing hemagglutinin (HA) -tagged forms of TPa, TPp, TPaS329A,TPaS331A and TPaS329,331A respectively, have been previously described [20,21]. For transfections, HEK 293 cell lines were routinely plated 48 h previously at -2* 106 cells/10 cm dish in 8 ml media and co-transfected with 10 |og of pADVA and 25 |g of pCMV-based mammalian expression vector using the calcium phosphate/DNA co-precipitation procedure [20].

Primary (1°) human aortic smooth muscle cells (1° h.AoSMCs) were purchased from Cascade Biologics (C-007-5C) and routinely grown in Smooth Muscle Cell Growth Medium 2 (Promocell GMBH, C-22062) supplemented with 0.5 ng/ml epidermal growth factor, 2 ng/ml basic Fibroblast growth factor, 5 |g/ml insulin, 5% FBS.

2.3. Calcium measurements

Measurements of intracellular calcium ([Ca2+]¡) mobilization were carried out in FURA2/AM preloaded HEK 293 cell lines transiently co-transfected with pCMV:Gaq and pADVA some 48 h previously, as described [20]. Cells were stimulated with 1 ^M U46619,1 ^M Cicaprost, 1 |M BW245C, 5 |M SIN-1 or 10 |M FK409, unless otherwise specified. Data (Supplemental Figs. 1 and 2) are representative of 3-4 independent experiments and calculated as changes in intracellular Ca2+ mobilized (A[Ca2+]¡ (nM)) as a function of time (seconds, s) following ligand stimulation.

2.4. Determination ofRhoA activation and cofilin phosphorylation

Activated cellular Rho was determined by interaction with a purified glutathione-S-transferase: rhotekin Rho-binding domain (GST-RBD) fusion protein immobilized on Glutathione-Sepharose 4B resin [28]. Preparation of the GST-RBD protein was carried out as previously reported [28]. For the 'Rho-pulldown assay', in brief, HEK.TPa, HEK. TPp, HEK.TPaS329A HEK.TPaS331A, HEK.TPaS329,331A or 1° h.AoSMCs cells were plated some 48 h previously in complete growth medium onto 10-cm dishes to achieve -70% confluency; cells were then serum starved for 5 h or 20 h (1° h.AoSMCs cells) in growth media containing 0.1% FBS before stimulation for 0-30 min with 0-10 |M U46619, as indicated in specific figure legends. To assess the effect ofprostacyclin, nitric oxide (NO) or PGD2 on TP-mediated Rho signaling, cells were pre-incubated for 10 min with either 0.01-10 ||M Cicaprost; 0.05-50 |M SIN-1; 10 |M FK409 or 1 |M BW245C before stimulation with U46619 (typically 0.1 |M for 10 min). As controls, cells were incubated with an equivalent volume of the drug vehicle, agonist or inhibitor in 0.01% ethanol in HBS (modified Ca2+/Mg2+-free Hank's buffered salt solution) for equivalent incubation times.

Thereafter, cells were lysed in 800 |l Lysis Buffer (125 mM HEPES, pH 7.5, 750 mM NaCl, 5 mM EDTA, 5% NP-40, 10% glycerol, 50 mM MgCl2, and 10 | g/ml each of leupeptin and aprotinin; [29]) and aliquots (600 |l) were subjected to pulldown using Glutathione-Sepharose 4B beads preloaded with 30 |g GST-RBD, essentially as previously described [28]. Following washing, precipitated GTP-bound RhoA was subjected to SDS-PAGE on 12.5% acrylamide gels and immunoblotted with anti-RhoA monoclonal antibody (Sc-418), followed by chemiluminescence detection [21]. In parallel, to confirm equivalent RhoA protein expression in the cell lysates and uniform protein loading, aliquots of whole cell lysates (typically 10 | l, corresponding to 1.25% of total cell lysate) were directly immunoblotted with the anti-RhoA antibody and/or with the anti-HDJ2 antibody. Similarly, to assess U46619-mediated cofilin phosphor-ylation and activation, aliquots ofwhole cells lysates (typically 10 | l, corresponding to 1.25% of total cell lysate) were first immunoblotted with anti-phosphoSer3 cofilin antibody; thereafter, phospho-cofilin blots were stripped and rescreened versus anti-cofilin antibody to normalise for total cofilin protein expression and/or with the anti-HDJ2 antibody to confirm uniform protein loading in each of the assays. All images of RhoA expression/pulldown or cofilin phosphorylation and/or expression were captured using Adobe Photoshop (V6), where band width and intensity was quantified and represented as fold increases relative to basal levels. To account for biological variations in basal activation levels, experiments were normalised to within a comparable range based on measurements from more then 20 individual experiments for each cell type.

2.5. F-actin staining

HEK 293 cell lines or 1° h.AoSMCs, grown on coverslips for 3 days to achieve approximately 50% confluency, were serum-starved for 2 h in growth media containing 0.1% FBS, prior to stimulation with U46619 (0-1 |M; typically 10 nM U46619). To assess the role of prostacyclin or NO, cells were pre-incubated for 10 min with either 0.0110 |M Cicaprost or 0.05-50 |M SIN-1 before stimulation with U46619 (typically 10 nM for 10 min). F-actin polymerization was stained by the addition of Alexa Fluor® 488 phalloidin essentially as described by the supplier (Molecular Probes) and slides were imaged using an Axioplan 2 Imaging Universal Microscope.

2.6. Disruption of TPa and TPfi expression by small interfering (si) RNA

For RNAinterference (RNA,) experiments, HEK 293 cell lines (HEK.TPa, HEK. TP| and, as controls, HEK293 cells) or 1° h.AoSMCs were plated at -2.5* 105 cells /35-mm plate and were allowed to attach for 24 h, achieving -30 % confluency. Thereafter, cells were washed twice with serum-free Opti-MEM® and transfected for 4 h at 37 °C with 0.2 |iM TPa siRNA (siRNATPa; a 50:50 mixture of two individual 19 bp siRNAs duplexes corresponding to nucleotides 2003-2021 and 2380-2398 of GenBank accession D38081, respectively) or 0.2 |iM TP| siRNA (siRNATP|; a 50:50 mixture of two individual 19 bp siRNAs duplexes corresponding to nucleotides 1966-1974 + 2634-2644 and 1970-1974 + 2634-2647 of GenBank accession D38081, respectively) or 0.2 |iM Lamin A/C siRNA (5'-CUGGACUUCCAGAAGAACAtt) using Oligofectamine® (3 il/well) in Opti-MEM® (1 ml /well). Thereafter, 1 ml pre-warmed complete media supplemented with 30% FCS was added per well and cells were harvested following incubation at 0, 24, 48, and 72 h. As additional controls,

HEK.TPa cells were treated according to the latter conditions but using TP| siRNAs and vice versa.

To confirm the efficacy of the siRNAs to disrupt TPa or TP| expression, HEK 293 cell lines were harvested and subject to SDS-PAGE (25-50 ig/lane on 12.5% gels) followed by electroblotting onto PVDF membranes (Roche). Membranes were initially screened versus the anti-HA (3F10) antibody and, following stripping, were rescreened versus anti-HDJ antibody to confirm uniform protein loading. Similarly, 1° h.AoSMCs were screened, under permabilising conditions, by indirect immunofluorescence microscopy for TPa and TP| expression using affinity purified isoform specific rabbit anti-TPa (3 |ig/ml) and anti-TP| (3 ig/ml) antibodies [30] incorporating tyramide signal amplification (TSA system; Invitrogen), used as per the manufacturer's instructions. In brief, following incubation with the primary antibodies, 1° h. AoSMCs were incubated with biotinylated goat anti-rabbit (1 in 500 dilution), followed by streptavidin HRP (1 in 2000). Signal amplification was facilitated by incubating the HRP labeled cells with biotinylated tyramide for 10 min at room

Fig. 1. TPa- and TP|-mediated RhoA Activation. HEK.TPa, HEK.TP| and HEK293 cells were serum-starved for 5 h before treatment (Panel A) for 10 min with vehicle or the indicated concentrations of U46619 or (Panel B) with 100 nM U46619 for the specified times, where cells treated with vehicle for 30 min acted as the control. Panel C: Alternatively, HEK.TPa and HEK.TP| cells were transiently transfected with plasmids encoding GaqQ209l,D277N or GaG222SA. Some 48 h post-transfection, cells were serum-starved for 5 h before treatment for 10 min with vehicle or 100 nM U46619. Active Rho was precipitated from the cell lysates using the Rho pulldown assay involving its binding to the GST-RBD (Rho binding domain of rhotekin)-fusion protein, separated by SDS-PAGE and immunoblotted with anti-RhoA antibody (Upper panels). Aliquots of cell lysates (typically 10 il /lane corresponding to 1.25% of total cell lysate) were also analyzed for total RhoA expression with anti-RhoA antibody (Lower panels). The bar charts to the right of the panels signify mean fold increases in Rho activation ±S.E.M. (n=3-6) where basal levels are assigned a value of 1.0.

temperature. Thereafter, 1° h.AoSMCs were incubated with streptavidin FITC (1 in 1000 dilution) and counterstained with propidium iodide (20 |g/ml), prior to mounting and imaging using a Zeiss fluorescence microscope coupled with AxioVision Software (V 4.4).

Thereafter, having optimised the conditions for effective RNAi disruption of TPa or TP(J expression in respective HEK 293 lines and in 1° h.AoSMCs, experiments was scaled up 8.2-fold (2*106 cells on 10-cm dishes) and functional disruption was assessed through Rho pulldown assays or cofilin phosphorylation, as previously outlined herein.

Vehicle

U46619

« ■y 1 . " i JsSi tir^J r

l it r a JÊL * 9 / " i

HEK.TPa

HEK293

Phospho-cofilin

Total cofilin

U46619(mM) -

Phospho-cofilin

Total cofilin

HEK.TPß

U46B19 (fjM) - 0.001 0.01 0.1 -1.0 10.0

HEK293

«10.0 CS

Î 75 — s.o

"Ö 2.5 u.

it- K ö ®

N. K. »S O Q <§> C? <=>' >?'

Phospho-cofilin

Total cofilin

Phospho-cofilin

Total cofilin

U46619 (0.1 nM) Gaq Q209I,D277N G a12 G 228A

CJ 15.0-a> 12.5-

1-10 0-(j

G 7.55 5.0-

0 2.50.015.00)

to 12.5'

at 10.0-g 7.5-

1 5.0. O 2.5-LL

+ - + - +

2.7. Data analyses

Radioligand binding data was analyzed using GraphPad Prism V3.0 to determine the Kd and Bmax values. Statistical analyses were carried out using the unpaired Student's t test using the Statworks Analysis Package. p-values of less than or equal to 0.05 were considered to indicate a statistically significant difference. Throughout the figures, * <0.05, **<0.01, ***<0.001.

3. Results

3.1. TPa and TPß isoforms independently regulate the Gq/PLCß and G12/Rho signaling systems

Whilst a range of studies have investigated Gq/PLCß-mediated signaling by both the TPa and TPß isoforms of the TXA2 receptor (TP) expressed in human tissues, to our knowledge, no such study has investigated the propensities or relative abilities of the individual TPa or TPß isoforms to activate and/or regulate Rho-mediated signaling. Hence, herein we investigated TPa and TPß-mediated Rho signaling in response to the TXA2 mimetic U46619 in clonal HEK 293 cell lines that stably over-express either TPa (HEK.TPa cells) or TPß (HEK.TPß cells). Throughout these studies, TPa/TPß-mediated Gq/PLCß-dependent [Ca2*],- mobilization was monitored as a comparative reference.

Consistent with previous reports [20,21], both TPa and TPß expressed in HEK.TPa cells and HEK.TPß cells, respectively, showed similar concentration-dependent mobilization of [Ca2+]i in response to U46619 stimulation, with maximal responses generated using 1 |jM U46619 (Fig. 1; Supplemental Data). Moreover, both TPa and TPß also mediated rapid RhoA activation in HEK.TPa and HEK.TPß cells in response to U46619 stimulation while no such activation was observed in the vehicle-treated cells or in the control non-transfected HEK 293 cells in the presence of U46619 (Fig. 1A). From concentration-response studies, 10-100 nM U46619 was required for maximal RhoA activation by both TPa and TPß while time-course assays confirmed that this was rapid, occurring within 2 min, and sustained for at least 30 min for both TP isoforms (Fig. 1A and B). RhoA activation through GPCRs predominantly occurs by coupling to G12 (Ga12/Ga13) members but may also occur through Gq coupling, in certain settings at least [31-33]. Herein, over-expression of dominant negative forms of Ga12 (GaG228A), but not of Gaq (GaqQ209 l,D277N), significantly impaired U46619-mediated RhoA activation through both TPa (p=0.0011 and p=0.8011 respectively) and TPß (p=0.0043 and p=0.9235 respectively; Fig. 1C).

To extend these studies, we also examined U46619-mediated stress fibre formation in HEK.TPa and HEK.TPß cells by monitoring F-actin polymerization and Rho-dependent phosphorylation, and inactiva-tion, of the actin depolymerizing agent cofilin using anti-phospho-cofilin antibodies directed to phosphoSer3 [34]. Throughout the latter, assays were normalised for total cofilin expression as presented in the lower panels in each of the figures. Whilst the control HEK 293 cells failed to show any changes in stress-fibre formation in response to U46619 (1 nM-10 |jM), both TPa and TPß induced rapid and profound F-actin polymerization with optimal responses occurring using 10 nM U46619 (Fig. 2A). Moreover, U46619 induced rapid and concentration-dependent cofilin phosphorylation in both HEK.TPa and HEK.TPß cells with optimal responses occurring with 1 ^M U46619 by both TP isoforms (Fig. 2B) while there was no cofilin phosphorylation in either cell line in response to the drug vehicle or in HEK 293 cells in response to U46619 stimulation (Fig. 2B). Furthermore, this U46619-mediated phosphorylation of cofilin was inhibited by over-expression of the

dominant negative form of Ga12 (Ga^ff ), while the dominant negative form of Gaq (GaqQ209l,D277N) had no significant effect (Fig. 2C). The Rho kinase inhibitor Y27632 (10 |jM) effectively abolished U46619-induced cofilin phosphorylation by both TPa and TPß (data not shown).

Collectively, these data confirm that both TPa and TPß can independently couple to both Gq-dependent PLCß activation to mobilize Ca2+ from lP3-operated intracellular stores, for example, and to G12-dependent RhoA activation and effector coupling leading to cofilin phosphorylation and inactivation and to F-actin polymerization.

3.2. The effect ofprostacyclin/cAMP and NO/cGMP on TPa-and TPß-mediated PLCß- and RhoA-signaling

Amongst the many functional differences identified thus far between the individual TPa and TPß isoforms [22], one of the most significant relates to the differential heterologous desensitization of their signaling by the vasodilatory autocoids prostacyclin [21], prostaglandin (PG) D2 [27] and nitric oxide [20]. Hence, in view of those differential sensitivities of Gq/PLCß-mediated signaling by TPa and TPß to both prostacyclin/cAMP and NO/cGMP [20,21], coupled to the well documented inhibitory actions of cAMP and cGMP on Rho-mediated signaling in response to various agents including TXA2 and thrombin, such as within platelets and vascular smooth muscle [35,36], we next investigated the effects of prostacyclin and NO on RhoA-mediated signaling by the individual TPa and TPß isoforms.

Initially the effect of the prostacyclin analogue Cicaprost (1 |jM) or the NO-donor SlN-1 (5 |jM) on U46619-mediated [Ca2+]i mobilization and RhoA activation and signaling by TPa and TPß was examined. Consistent with our previous findings [20,21], pre-incubation with either Cicaprost or SlN-1 significantly impaired U46619-induced [Ca2+]i mobilization by TPa but had no effect on signaling by TPß (Fig. 1; Supplemental Data). While Cicaprost did not induce substantial RhoA activation perse in HEK.TPa, HEK.TPß or HEK 293 cells (data not shown), it significantly impaired U46619-induced RhoA activation by TPa expressed in HEK.TPa cells in a concentration-dependent manner (Fig. 3A). On the other hand, Cicaprost had no effect on RhoA activation by TPß, even at 10 |jM Cicaprost (Fig. 3A). Similarly, SlN-1 also significantly impaired U46619-mediated RhoA activation by TPa in a concentration-dependent manner but had no effect on RhoA activation by TPß, even at 50 |jM SlN-1 (Fig. 3B). While Cicaprost (1-10 |jM) and SlN-1 (5-50 |jM) each significantly impaired U46619-induced F-actin polymerization by both TPa and TPß, consistent with the inhibitory effects of cAMP/PKA and cGMP/PKG on both the Ca2+-dependent and Ca2+-independent paths, it was apparent that at lower concentrations both Cicaprost (100 nM) and SlN-1 (500 nM) impaired F-actin polymerization in HEK.TPa cells but neither agent affected such responses in HEK.TPß cells (Fig. 4A). Moreover, U46619-induced cofilin phosphorylation by TPa was also significantly impaired by either Cicaprost or SlN-1, while neither agent affected such responses in HEK.TPß cells (Fig. 4B), regardless of concentration. Consistent with the latter data, the PGD2 analogue BW245C and the alternative NO donor FK409 also significantly impaired U46619-mediated RhoA activation (Fig. 3C) and cofilin phosphorylation (data not shown) by TPa but had no effect on signaling by TPß (Fig. 3C and data not shown).

We have previously established that while both prostacyclin analogues, such as Cicaprost, and NO-donors, such as SlN-1, were

Fig. 2. TPa and TP|-mediated F-Actin Polymerization and Cofilin Phosphorylation. Panel A: HEK.TPa, HEK.TP| and HEK293 cells were serum-starved for 5 h before treatment with the vehicle (MEM) or 10 nM U46619 for 10 min. Following fixation and permeabilization, F-actin polymerization was detected with Alexa Fluor® 488 phalloidin followed by fluorescence microscopy. Panel B: HEK.TPa, HEK.TP| and HEK293 cells were serum-starved for 5 h before treatment for 10 min with 0-10 iM U46619. Panel C: Alternatively, HEK.TPa and HEK.TP| cells were transiently transfected with plasmids encoding GaqQ209l,D277N or Gaif2SA. Some 48 h post-transfection, cells were serum-starved for 5 h before treatment for 10 min with vehicle or 100 nM U46619. Aliquots of the cell lysates (typically 10 il/lane corresponding to 1.25% of total cell lysate) in B and C were separated by SDS-PAGE and immunoblotted with anti-Phospho cofilin (Upper panels) or anti-cofilin (Lower panels) antibodies to detect phosphorylated and total cofilin expression. The bar charts to the right of the panels signify mean fold increases in cofilin phosphorylation ±S.E.M. (n=3-6) where basal levels were assigned a value of 1.0. (The reader is referred to the web version of this article to see color images of this figure, where relevant.)

indeed capable of cross-desensitizing or impairing Gq/PLCß signaling by TPa, they did so by entirely independent mechanisms and at different, though adjacent, sites. Specifically, prostacyclin-desensitization occurs by direct PKA phosphorylation of Ser329 while NO-desensitization occurs through PKG phosphorylation of Ser331, both within the unique C-tail domain of TPa [20,21]. Hence, to further investigate the mechanism by which SlN-1 and Cicaprost impair signaling by TPa, we examined their effects on U46619-induced [Ca2+],- mobilization, Rho activation, F-actin polymerization and cofilin phosphorylation by TPa and its specific site directed variants TPaS329A, TPaS331A,

TPaS329,331A defective in the Cicaprost-sensitive PKA (at Ser329), NO-sensitive PKG (at Ser331) or both (at Ser329,331) phosphorylation sites, as previously described by us [20,21]. Consistent with those previous studies, pre-incubation with SlN-1 specifically impaired U46619-induced [Ca2+]i mobilization in HEK.TPa and HEK.TPaS329A cells while having no affect on such signaling in HEK.TPaS331A and HEK. TPaS329,331A cells (Fig. 2; Supplemental Data). Moreover, both SlN-1 and the alternative NO-donor FK409 also specifically impaired U46619-induced RhoA activation by TPa and TPaS329A cells but

had no affect on signaling by TPaS331A and TPaS329,331A (Fig. 5A).

Fig. 3. Cicaprost- and SlN-1-induced desensitization of TP-mediated signaling. Panels A-C: HEK.TPa and HEK.TPß cells were serum starved for 5 h before treatment for 10 min with vehicle (Panels A and B), 0.01 -10 |jM Cicaprost (Panel A), 0.05-50 |jM sIN-1 (Panel B), 1 |jM BW345C or 10 |jM FK409 (Panel C). Thereafter, cells were incubated with 100 nM U46619 for 10 min (Panels A-C). Active Rho was precipitated from the cell lysates using GST-RBD fusion protein, separated by SDS-PAGE and immunoblotted with an anti-RhoA antibody (Upper panels). Cell lysates were analyzed by western blotting for total RhoA expression (Lower panels). The bar charts to the right of the panels signify mean fold increases in Rho activation ±S.E.M. (n=3-6) where basal levels were assigned a value of 1.0. The asterisks indicates that the level of U46619-mediated RhoA activation was significantly reduced in the presence of Cicaprost, Sin-1, BW245C and FK409 where * and ** indicates p<0.05 and p< 0.01, respectively.

Additionally, S1N-1 and FK409 also impaired U46619-induced F-actin polymerization, at low agonist concentration, and cofilin phosphorylation by TPa (Fig. 4 and data not shown) and TPaS329A cells, but had no affect on signaling by TPaS331A and TPaS329 331A (Fig. 5B and data not shown). On the other hand, pre-stimulation with Cicaprost impaired U46619-induced [Ca2"1"],- mobilization and RhoA activation by TPa and TPaS331A while having no affect on signaling by TPaS329A and TpaS329331A (Fig. 2; Supplemental Data and Fig. 5A). Consistent with this, the PGD2 receptor (DP) agonist BW245C also impaired RhoA activation by TPa (Fig. 3C) and TPaS331A without affecting signaling by TPaS329A and TPaS329 331A (Fig. 5A). Additionally, Cicaprost specifically impaired U46619-induced F-actin polymerization and cofilin phos-phorylation by TPa (Fig. 4) and TPaS331A cells but had no affect on signaling by TPaS329A and TPaS329,331A (Fig. 5B and data not shown).

Taken together these data clearly indicate that both Gq/PLCß-mediated [Ca2+],- mobilization and the G12/RhoGEF-dependent RhoA activation and cofilin phosphorylation by TPa are specifically impaired by the potent vasodilators S1N-1 and Cicaprost. On the other hand, neither agonist-induced Gq/PLCß nor G12/RhoA signaling by TPß is directly affected by either vasodilator. Moreover, consistent

with our previous findings [20,21], our data herein further suggest that the mechanisms whereby Cicaprost and SIN-1 impair both the Gq/PLCp-mediated [Ca2+],- pathway and the Rho-dependent pathway are similar but entirely independent where NO/SIN-1-mediated desensitization occurs through a PKG-dependent mechanism involving direct phosphorylation of TPa at Ser331 while that of prostacy-clin/Cicaprost involves a PKA-dependent mechanism where Ser329 is the phospho-target.

3.3. TPa- and TPfi-mediated RhoA signaling in primary human AoSMCs

We next investigated TP-mediated Rho activation and cytoskele-tal signaling in a physiologically relevant, vaso-responsive model by investigating U46619-induced signaling in 1° h.AoSMCs, cells that express both TPa and TPp [30]. Consistent with our findings herein in HEK.TPa and HEK.TPp cells, stimulation of 1° h.AoSMCs with U46619 led to rapid RhoA activation with maximal responses observed with 100-1000 nM U46619 (Fig. 6A). Stimulation of 1° h. AoSMCs also led to rapid F-actin polymerization (Fig. 6B) and cofilin phosphorylation with optimal responses generated using 1 |iM

Fig. 4. Cicaprost- and SIN-1-induced desensitization of TP-mediated signaling. Panel A: HEK.TPa and HEK.TPp cells were serum starved for 5 h before treatment for 10 min with vehicle (Vehicle), 10 nM U46619 (U46619), 100 nM Cicaprost followed by 10 nM U46619 (U46619, Cicaprost) or 500 nM SIN-1 followed by 10 nM U46619 (U46619, SIN-1). F-actin formation was detected with Alexa Fluor® 488 phalloidin followed by fluorescence microscopy. Images presented are representative of the majority of cells examined and of 3/4 independent experiments. Panel B. HEK.TPa and HEKTPp cells were serum starved for 5 h before treatment for 10 min with vehicle (-), 1 |iM Cicaprost or 10 |jM SIN-1 (Panel B). Thereafter, cells were incubated with vehicle (-) or 100 nM U46619 for 10 min. Cell lysates were separated by SDS-PAGE and immunoblotted with anti-Phospho cofilin (Upper panels) or anti-cofilin (Lower panels) antibodies. The bar charts signify mean fold increases in cofilin phosphorylation±S.E.M. (n=3-6) where basal levels were assigned a value of 1.0. The asterisks indicates that the level of U46619-mediated cofilin phosphorylation was significantly reduced in the presence of Cicaprost or Sin-1 where * indicates p < 0.05. (The reader is referred to the web version of this article to see color images of this figure, where relevant.)

U46619 (Fig. 6C). While the IP agonist Cicaprost did not lead to substantial RhoA activation and cofilin phosphorylation relative to the drug vehicle per se, it significantly impaired such U46619-mediated signaling in 1° h.AoSMCs (Fig. 7A). Consistent with this, the specific PGD2 receptor (DP) agonist BW245C also significantly impaired RhoA activation (Fig. 7C) and cofilin phosphorylation in 1° h.AoSMCs. Similarly, while the NO donors SlN-1 and FK409 alone

did not induce substantial RhoA signaling relative to the drug vehicle per se, they each significantly impaired U46619-induced RhoA activation and cofilin phosphorylation following their pre-incubation in 1° h.AoSMCs (Fig. 7B and D). Moreover, while Cicaprost, Sin-1, BW245C or FK409 did not induce F-actin polymerization per se, they each significantly impaired U46619-induced F-actin polymerization (data not shown).

Fig. 5. Cicaprost- and SlN-1-induced desensitization of TP Signaling in HEK 293 cells. Panel A: HEK.TPaS329A, HEK.TPaS331A and HEK.TPaS331,329A cells were serum-starved for 5 h before treatment for 10 min with vehicle (-), 5 |M SlN-1,10 |M FK409,1 |M Cicaprost or 1 |M BW245C as indicated. Thereafter, cells were incubated for 10 min with vehicle (-) or 100 nM U46619 for 10 min. Active Rho was precipitated from the cell lysates using GST-RBD fusion protein, separated by SDS-PAGE and immunoblotted with an anti-RhoA antibody (Upper panels). Cell lysates were analyzed by western blotting for total RhoA expression (Lower panels). Panel B: HEK.TPaS329A, HEK.TPaS331A and HEK.TPaS331,329A cells were serum starved for 5 h before treatment for 10 min with vehicle, 500 nM SlN-1 or 100 nM Cicaprost. Thereafter, cells were incubated for 10 min with vehicle (-) or 10 nM U46619 for 10 min. F-actin formation was detected with Alexa Fluor® 488 phalloidin followed by fluorescence microscopy. lmages presented are representative of the majority of cells examined in 8 independent fields and of 3/4 independent experiments. (The reader is referred to the web version of this article to see color images of this figure, where relevant.)

Hence, taken together, both NO-donors and prostanoid vasodila-tory agents, such as prostacyclin and PGD2 signaling through the prostacyclin (IP) and PGD2 (DP) receptors, respectively, can impair U46619-mediated Rho activation and cytoskeletal signaling in 1° h. AoSMCs. Moreover, our data generated in the HEK.TPa and HEK.TPp cell lines over-expressing the individual TPa and TPp isoforms, respectively, suggest that such inhibitory responses of prostacyclin and NO are mediated, at least in part, at the interface of the stimulatory GPCR (i.e. the TP). More specifically, by directly targeting TPa, prostacyclin and NO may impair its RhoA-signaling both at the level of TPa itself in addition to at the level of the well documented targets ofcAMP/PKAand cGMP/PKG [35,36]. On the other hand, as TPp is not as such a direct target of prostacyclin- or NO-mediated phosphorylation and inhibition, their effect on TPp-mediated RhoA signaling may be solely manifest at a later point in the cascade [36]. Clarity on this issue in 1° h.AoSMCs is, however, confounded by the fact that h.AoSMCs express both TPa and TPp isoforms [30] and, therefore, it is possible that the vasodilatory agents NO and Cicaprost may target TPa, TPp or both in addition to other downstream targets.

To address this issue, we generated small interfering RNA (siRNA)-targeting agents to selectively disrupt or knock-down expression of TPa and TPp in 1° h.AoSMCs. To begin with, the siRNA agents were

validated by examining their ability to affect TPa and TPß expression and RhoA signaling in HEK.TPa and HEK.TPß cells. Under optimized experimental conditions, we observed effective isoform-specific knock-down of both TPa and TPß expression following 72 h incubation of HEK.TPa and HEK.TPß cells with siRNATPa and siRNATPß, respectively (Fig. 8A), with ~50-60% specific knock-down achieved as assessed by densitometry and radioligand binding assay in each case (Fig. 8A and data not shown). On the other hand, the siRNA^ did not affect TPß expression in HEK.TPß cells and siRNAipß did not affect TPa expression in HEK.TPa cells (Fig. 8A) thereby confirming the specificity of the TPa and TPß isoform-directed siRNAs. Additionally, RNA!' directed to Lamin A/C, acting as a control, had no effect on either TPa or TPß expression in either cell line (Fig. 8A). Moreover, preincubation of HEK.TPa cells with siRNATPa significantly impaired U46619-induced RhoA activation but had no significant effect on such signaling in HEK.TPß cells (Fig. 8B). Conversely the anti-TPß siRNATPß significantly impaired U46619-induced RhoA activation in HEK.TPß cells but had no effect on such signaling in HEK.TPa cells (Fig. 8B). RNA,- directed to Lamin A/C had no effect on either TPa- or TPß-mediated RhoA activation (Fig. 8B). Consistent with these findings, siRNAs directed to TPa and TPß also impaired U46619-mediated F-actin polymerization and cofilin phosphorylation in HEK.TPa and

Fig. 6. TP-mediated Rho signaling in 1° human Aortic Smooth Muscle Cells. Panel A: 1° AoSMCs were serum starved for 20 h before treatment for 10 min with vehicle (-) or 0.00110 |iM U46619 as indicated. Active Rho was precipitated from the cell lysates using GST-RBD fusion protein, separated by SDS-PAGE and immunoblotted with an anti-RhoA antibody (Upper panels). Cell lysates were analyzed by western blotting for total RhoA expression (Lower panels). Panel B: 1° AoSMCs were serum starved for 5 h before treatment for 10 min with vehicle or 1 |iM U46619 for 10 min. F-actin formation was detected with Alexa Fluor® 488 phalloidin followed by fluorescence microscopy. Images presented are representative of the majority of cells examined and of 3/4 independent experiments. Panel C: 1° AoSMCs were serum starved for 20 h before treatment for 10 min with vehicle (-) or 0.001 -10 |iM U46619 as indicated. Cell lysates were separated by SDS-PAGE and immunoblotted with anti-Phospho cofilin (Upper panels) or anti-cofilin (Lower panels) antibodies to detect phosphorylated and total cofilin expression. The bar charts to the right of the panels signify mean fold increases in Rho activation or cofilin phosphorylation ±S.E.M. (n=3-6) where levels of basal levels are assigned a value of 1.0. (The reader is referred to the web version of this article to see color images of this figure, where relevant.)

HEK.TPß cells, respectively, and in an entirely isoform specific manner (data not shown).

Having established the specificity of the siRNA reagents to impair expression and RhoA-dependent signaling by both TPa and TPß in HEK 293 cells, we next examined their ability to affect expression and signaling by the individual TP isoforms in h.AoSMCs. The effective delivery and utility of the latter siRNAs in 1° h.AoSMCs was initially confirmed whereby the siRNATPa impaired expression of TPa but not of TPß while siRNATPß reduced expression of TPß but not of TPa (Fig. 9A). Consistent with the reduced expression of TPa and TPß following incubation of the 1° h.AoSMCs with the isoform-specific siRNAs, there were reductions in U46619-induced Rho activation and cofilin phosphorylation in the presence of RNAi directed to either TP but not directed to Lamin A/C (Fig. 9B). Moreover, incubation of the 1° h.AoSMCs with both siRNATPa and siRNATPß led to a further significant reduction in U46619-induced Rho activation and cofilin phosphoryla-tion. Hence, these data clearly suggest that both TPa and TPß contribute to the Rho activation in h.AoSMCs.

We next examined the effect of the inhibitory vasodilatory agents SlN-1 and Cicaprost on U46619-mediated RhoA activation and signaling in 1° h.AoSMCs in the presence of the respective TP-isoform specific siRNA reagents. ln the absence of siRNA, SlN-1 and Cicaprost significantly impaired U46619-mediated RhoA activation (Fig. 9C) consistent with our earlier findings in both 1° h.AoSMCs (Fig. 7A and B)

and in HEK.TPa cells (Fig. 3B and C). Following 72 h incubation with siRNATPp, the NO donor SIN-1 specifically impaired U46619-mediated RhoA activation in h.AoSMCs to levels greater than that observed in vehicle treated cells. On the other hand, the inhibitory effect of SIN-1 on U46619-mediated RhoA activation in 1° h.AoSMCs was significantly less in the presence of siRNATPa at 72 h (Fig. 9C). Similarly, SIN-1 impaired U46619-mediated cofilin phosphorylation in the presence of siRNATPß to levels similar to those observed in vehicle-treated cells but its ability to impair U46619-signaling in the presence of siRNATPa was almost fully abolished (Fig. 9D). Moreover, the prostacyclin analogue Cicaprost significantly impaired Rho activation (Fig. 9C) and cofilin phosphorylation (Fig. 9D) in 1° h.AoSMCs pre-treated with siRNATPß to levels similar to that observed in the control, vehicle-treated cells while its ability to impair that signaling in cells pre-treated with the siRNATPa was almost completely abolished.

Hence, we conclude that, similar to that which occurs for TP-mediated Gq/PLCß activation, both the NO and prostacyclin analogues SIN-1 and Cicaprost impair TP-mediated cytoskeletal changes involving RhoA activation and cofilin phosphorylation in 1° h.AoSMCs and that they do so, at least in part, by specifically and directly targeting TPa, impairing its downstream signaling. On the other hand, neither vasodilatory agent directly target TPß. Hence, TPa- and TPß-mediated RhoA signaling functionally diverge at the point at which prostacyclin and NO affect the RhoA signaling cascade.

Fig. 7. Desensitization of TP Signaling in 1° h.AoSMCs. 1° AoSMCs were serum starved for 20 h before treatment for 10 min with vehicle (-), 1 |iM Cicaprost (Panel A), 5 |jM SIN-1 (Panel B), 1 |iM BW245C (Panel C) or 10 |iM FK409 (Panel D). Thereafter, cells were incubated for 10 min with vehicle (-) or 1 |iM U46619, as indicated. Active Rho was precipitated from the cell lysates using GST-RBD fusion protein, separated by SDS-PAGE and immunoblotted with an anti-RhoA antibody (GST-RBD pulldown) while aliquots of cell lysates were analyzed by western blotting for total RhoA expression, phospho-cofilin and total cofilin expression as indicated. The bar charts to the right of the panels signify mean fold increases in Rho activation or cofilin phosphorylation ±S.E.M. (n=3-6) where basal levels are assigned a value of 1.0. The asterisks indicates that the level of U46619-mediated RhoA activation and cofilin phosphorylation was significantly reduced in the presence of Cicaprost, Sin-1, BW245C or FK409 where * indicates p<0.05.

4. Discussion

Much of the molecular events that underlie the complex physiologic processes of platelet aggregation and thrombosis, VSM contraction, migration, proliferation involve the fundamental reorganization of the cellular cytoskeleton [5]. A key step in this cytoskeletal reorganization involves alterations in MLC phosphorylation that occurs either through Ca2+-dependent activation of MLCK or through the alternative Ca2+-independent pathway, involving RhoA [1,2,5,6]. For example, the processes that contribute to platelet activation and

secretion and shape change are under the dual control of the Gq/Ca2+-dependent and G12/Ca2+-independent pathways, respectively [1,8].

The prostanoid TXA2 plays an essential role within the vasculature inducing a range of cellular responses including platelet shape change and aggregation; contraction of vascular and bronchial smooth muscle (SM) cells; mitogenic and hypertrophic growth of VSM cells; inhibition of angiogenesis/vascularization [37-39]. Elevations in the levels of TXA2, its synthase or its receptor have been implicated in various cardiovascular disorders including thrombosis, myocardial infarction, unstable angina, atherosclerosis, systemic- and pregnancy-

Fig. 8. Effect of siRNA-mediated down-regulation of TPa and TPß Expression on Rho-signaling in HEK 293 cells. Panel A: HEKTPa and HEKTPß cells were transfected with siRNA-directed to TPa (siRNATPa) and TPß (siRNATPß), respectively, for 0-72 h. As controls, HEKTPa cells were transfected with siRNATPß or HEK.TPß cells were transfected with siRNATPa for 72 h. Alternatively, as additional controls, HEK.TPa cells or HEK.TPß cells were transfected with siRNALaminA/C for 72 h (Right panels). HA-tagged TPa or TPß expression was detected by immunoblotting using anti-HA 3F10 antibody (Upper panels) or equal protein loading was verified by secondary screening of blots with an anti-HDJ2 antibody (Lower panels). Panel B: HEK. TPa and HEK.TPß cells were transfected with siRNATPa, siRNATPß or siRNALaminA/C for 72 h, as indicated. Thereafter, cells were incubated for 10 min with vehicle (-) or 100 nM U46619 for 10 min. Active Rho was precipitated from the cell lysates using GST-RBD fusion protein and immunoblotted with an anti-RhoA antibody (Upper panels). Cell lysates were analyzed by western blotting for total RhoA expression (Lower panels). The bar charts to the right of the panels signify mean fold changes in TP isoform expression (A) and Rho activation (B) ±S.E.M. (n=3-6) where basal levels are assigned a value of 1.0. The asterisks indicates that TPa (siRNATPa) andTPß (siRNATPß) expression (Panel A) or U46619-mediated RhoA activation (Panel B) was significantly reduced in the presence of their respective siRNAs where *, ** and *** indicates p<0.05, p<0.01 and p<0.001 respectively.

induced hypertension and ischemic heart disease, processes in which RhoA dysfunction is widely implicated [7]. In humans, TXA2 signals through 2 distinct isoforms referred to as TPa and TP| [13,15,22]. While the functional requirement for two types of receptor for TXA2 in humans is unknown there is substantial evidence that they may have distinct physiologic/pathophysiologic roles [16,17,19,22].

Bearing this in mind and the growing appreciation of the critical role of the RhoA-mediated Ca2+-independent pathways to both normal and disease-processes within the vasculature [2,5,6], the central aim of the current study was to investigate the ability of the individual TPa and TP| isoforms to regulate RhoA signaling. Moreover, in view of the critical involvement of inhibitory agents including prostacyclin and NO, that largely signal through cAMP and cGMP second messengers, in regulating RhoA-dependent mechanisms

[1,6,35,36] coupled to their role in differentially regulating TPa and TPß-mediated Gq/PLCß signaling [20,21,27], we also sought to investigate the impact of both vasodilators on RhoA signaling through the individual TP isoforms.

To this end, we investigated the ability of TPa and TPß to mediate RhoA signaling in established clonal HEK 293 cell lines that overexpress the individual TP isoforms [20,21] and in cultured 1° AoSMCs, a physiologically relevant cell type that expresses both TPa and TPß [26,30]. Throughout these studies, TP-mediated RhoA signaling was determined by monitoring its activation-dependent interaction with the Rho-binding domain (RBD) of its effector rhotekin in GST-RBD pulldown assays in response to the TXA2 mimetic U46619. Moreover, we also investigated the ability of TPa and TPß to regulate events downstream of RhoA: effector coupling by monitoring U46619-

anti-TPa artti-TPß

Fig. 9. Effect of siRNA-mediated down-regulation of TPa and TPß Expression on Rho-signaling in 1° h.AoSMCs. Panels A-D: 1° AoSMCs were transfected with siRNA-directed to TPa (siRNATPa), TPß (siRNATPß) or Lamin A/C (siRNALamin A/C) for 72 h where non-transfected cells served as controls, as indicated. ln Panel A, following fixation and permeabilization, cells were screened by indirect immunoflourescence microscopy with anti-TPa or anti-TPß isoform specific 1 ° antibody and stained using FlTC-labelled goat anti-rabbit lgG, where cell nuclei were counter stained with propidium iodide. ln Panel A, data are representative of 3 independent experiments where a total of 16 independent fields of cells were analysed for each TP isoform. lt was estimated that there was greater than 70% reduction TPa/TPß expression in 80% of cells analysed. ln Panel B, cells were serum-starved for 20 h before treatment for 10 min with vehicle or 1 |M U46619 as indicated. Active Rho was precipitated from the cell lysates using GST-RBD fusion protein, separated by SDS-PAGE and immunoblotted with an anti-RhoA antibody (Active RhoA) while cell lysates were analyzed for total RhoA expression, phospho-cofilin and total cofilin expression as indicated. Panels C and D, the cells were serum-starved for 20 h before treatment for 10 min with vehicle, 1.0 |M Cicaprost or 5.0 |M SlN-1. Thereafter, cells were incubated for 10 min with vehicle (-) or1 |M U46619. ln Panel C, active Rho was precipitated from the cell lysates using GST-RBD fusion protein, separated by SDS-PAGE and immunoblotted with an anti-RhoA antibody (Upper panels) while cell lysates were analyzed for total RhoA expression (Lower panels). ln Panel D, cell lysates were immunoblotted with anti-Phospho cofilin (Upper panels) or anti-cofilin (Lower panels) antibodies. ln Panel B, the bar charts to the right signify mean fold changes in Rho activation and cofilin phosphorylation ± S.E.M. (n=3-6) where basal levels are assigned a value of 1.0. The asterisks indicates that U46619-mediated RhoA activation and cofilin phosphorylation was significantly reduced in the presence of their respective TPa (siRNATPa) and TPß (siRNATPß) siRNAs where * indicates p<0.05 and p<0.01, respectively. The bar charts below Panels C and D depict mean reductions (±S.E.M., n=3-6) in U46619-mediated Rho activation (C) and cofilin phosphorylation (D) in response to pre-treatment with Sin-1 and Cicaprost, and the asterisks indicate that the level of desensitization was significantly altered in the presence of the siRNATPa or siRNATPß where * and ** indicates p < 0.05 and p < 0.01, respectively. (The reader is referred to the web version of this article to see color images of this figure, where relevant.)

Fig. 9 (continued).

induced F-actin polymerization and cofilin phosphorylation. The ubiquitously expressed actin-depolymerising factor cofilin readily undergoes Rho/Rho kinase-dependent phosphorylation at Ser3 either by the LIM kinase 1/2 [3] and was used herein as a monitor of events downstream of Rho kinase in the Rho signaling cascade. Our conclusions are several-fold. Both TPa and TPß expressed in HEK 293 cells readily induced RhoA activation, F-actin polymerization and cofilin phosphorylation in response to U46619. In general, GPCR-mediated RhoA activation largely occurs through a G12, mainly Ga13, -dependent mechanism but in certain settings, particularly at higher agonist concentrations, may also occur through a Gq-mechanism through the specific involvement of the LARG (Leukemia-associated Rho guanine-nucleotide exchange factor), but not the p115- or PDZ-, member of the RGS-containing RhoGEF family [1,31-33]. Hence, herein, we sought to clarify the involvement of G12 and Gq on TP-mediated RhoA signaling and found that dominant negative forms of

Ga12 (GaGPA), not of Gaq (GaqQ209l,D277N), significantly impaired U46619-mediated RhoA activation and cofilin phosphorylation. Collectively, these data confirmed that both TPa and TPp can independently couple to Gq/PLCp activation and Ga12/Ga13/Rho-GEF-RhoA activation and are in agreement with a host of studies in mouse platelets whereby the single TP in that species couples to Gq/ PLCp and to G12/RhoA activation to independently regulate platelet activation (aggregation and secretion) and platelet shape change responses, respectively [1,8].

Thereafter, we investigated the effect of the selective prostacyclin analogue Cicaprost and the NO donor SIN-1 on TP-mediated RhoA activation and signaling. Consistent with our previous reports [20,21,27], Gq/PLCp-mediated [Ca2+],- mobilization by TPa, but not by TPp, was desensitized in response to both Cicaprost and NO stimulation. In keeping with this, TPa-mediated RhoA activation, F-actin polymerization and cofilin phosphorylation was also specifically

impaired by Cicaprost and SIN-1 while neither agent affected Rho-mediated signaling by TPß. As stated, while both prostacyclin and NO desensitize TPa-mediated Gq/PLCß signaling, they do so by entirely independent mechanisms involving direct PKA- and PKG- mediated phosphorylation of TPa at Ser329 and Ser331, respectively, within its unique C-tail domain [20,21]. Hence, we next compared the effect of SIN-1 and Cicaprost on Rho-signaling by TPaS329A, TPaS331A, TPaS329,331A, variants of TPa defective in the prostacyclin-sensitive PKA (at Ser329), NO-sensitive PKG (at Ser331) or both (at Ser329,331) phospho-target sites [20,21]. While SIN-1 and the alternative NO-donor FK409 impaired U46619-induced RhoA activation, cofilin phosphorylation, F-actin polymerization as well as [Ca2+],- mobilization by TPa and TPaS329A, they had no affect on signaling by TPaS331A and TPaS329,331A. Conversely, both Cicaprost and the PGD2 receptor agonist BW245C impaired RhoA activation, cofilin phosphorylation, F-actin polymerization and [Ca2+],- mobilization by TPa and TPaS331A but had no affect on signaling by TPaS329A and TPaS329,331A. Collectively, these data suggest that both Gq/PLCß-mediated [Ca2+],- mobilization and G12/RhoGEF -dependent RhoA activation of its effector rhotekin and cofilin phosphorylation by TPa, but not by TPß, are specifically impaired by the potent vasodilators SIN-1 and Cicaprost in this cellular context, at least. Of course the inhibitory effects of both prostacyclin and NO, and other agents that signal through cAMP and cGMP, on RhoA signaling are widely documented and form an essential component of the homeostatic regulatory mechanism that determines the balance between activation and inhibition, particularly within the vasculature [1,6,36]. Hence, it is arguable that the effects of Cicaprost and SIN-1 on TPa-mediated Rho signaling in HEK 293 cells are perhaps somewhat predictable. However, the fact that RhoA-mediated signaling by TPaS331A and TPaS329,331A is unaffected

by SIN-1 while that signaling by TPaS329A and TPaS329,331A is

unaffected by Cicaprost clearly suggest that the observed effects of SIN-1 and Cicaprost on TPa, in the HEK 293 over-expression system at least, are due to direct effects on TPa itself, namely through site specific prostacyclin-induced PKA (at Ser329) and NO-induced PKG (at Ser331) phosphorylation rather than at some other intermediary in the RhoA signaling cascade. Moreover, in keeping with that hypothesis, the finding that agonist-induced G12/RhoA signaling by TPß is not affected by either prostacyclin or NO again suggests that the effects of both vasodilators are due to direct effects on TPa itself and is entirely consistent with previous findings involving both prostacyclin- and NO-mediated desensitization of TPa and TPß signaling through the Gq/PLCß effector system [20,21]. The fact that we do not observe any measurable inhibitory effects on TPß-mediated RhoA signaling by either Cicaprost or SIN-1, such as might be expected to occur at a later point in the signaling cascade [36], could in theory be due to the fact that the level of TP receptor expression in the HEK 293 stable cell lines produces an overriding forward signal, overwhelming any inhibitory effects of prostacyclin or NO.

Therefore, we extended our studies by investigating TP-mediated Rho activation and cytoskeletal signaling in the more physiologically relevant primary human aortic smooth muscle cells. As expected, stimulation of cultured 1° h.AoSMCs with U46619 led to a concentration-dependent RhoA activation, F-actin polymerization and cofilin phosphorylation. Moreover, Cicaprost (IP agonist) and BW245C (DP agonist) and the NO donors SIN-1 and FK409 each significantly impaired such U46619-induced RhoA activation, F-actin polymerization and cofilin phosphorylation in 1° h.AoSMCs. Hence, collectively, both NO-donors and the vasodilatory prostanoids prostacyclin and PGD2 readily desensitize TP-mediated Rho activation and cytoskeletal signaling in 1° h.AoSMCs, findings entirely predicted from and in keeping with outcomes from other systems [6,8,35,36]. However, as stated, our data generated in the HEK 293 cell lines over-expressing the individual TPa and TPß isoforms suggest that such inhibitory responses of prostacyclin and NO are mediated, at least in part, directly at the level of TPa itself rather than at the level of other well

documented targets of cAMP/PKA and cGMP/PKG on Rho signaling [35,36]. Human AoSMCs express both TPa and TPß isoforms [26,30]. Hence, through the use of TP isoform-specific siRNA, we sought to determine whether TPa and TPß independently contribute to U46619-induced RhoA activation and signaling in 1° h.AoSMCs and to ascertain whether the inhibitory effects of NO and/or Cicaprost may directly target TPa, or indeed TPß, at the level of the receptor itself and/or in addition to other downstream targets [36]. Under optimized experimental conditions, the specificity and utility of the siRNATPa and siRNATPß reagents were validated whereby we observed effective isoform-specific knock-down of both TPa and TPß expression and RhoA-mediated signaling in their respective HEK 293 cell lines. Moreover, the effective delivery and utility of the latter siRNAs in 1° h. AoSMCs was confirmed whereby the siRNATPa reduced expression of TPa but not of TPß, while siRNATPß reduced expression of TPß but not of TPa. It was notable that the level of siRNA-mediated impairment of TPa and TPß expression in the 1° h.AoSMCs was significantly higher than observed in HEK.TPa or HEK.TPß cells. The reason for this apparent discrepancy is simply owing to the fact that the stably transfected HEK 293 cell lines express TPa and TPß in abundance (~2 pmol/mg protein) relative to that expressed in 1° h.AoSMCs (20-50 fmol/mg protein) and hence, the inability of the siRNA to completely konckdown TPa or TPß expression in HEK.TPa or HEK.TPß cells was not surprising. Consistent with their reduced expression in the 1° h.AoSMCs, there was a significant reduction in U46619-mediated Rho activation and cofilin phosphorylation in the presence ofRNAi directed to either TPa or TPß but not to Lamin A/C confirming that both TPa and TPß contribute to the RhoA activation in h.AoSMCs. While SIN-1 and Cicaprost significantly impaired U46619-mediated RhoA activation in the presence of the siRNA directed to Lamin A/C to levels similar to that in vehicle-treated cells, the inhibitory action of both agents on RhoA activation and cofilin phosphorylation in 1° h. AoSMCs exposed to the siRNATPa was substantially impaired. On the other hand, in the presence of siRNATPß both SIN-1 and Cicaprost reduced U46619-mediated RhoA, F-actin polymerization (data not shown) and cofilin phosphorylation to levels not significantly different to those observed in vehicle-treated cells. Hence, both the NO and prostacyclin impair TP-mediated cytoskeletal changes involving RhoA activation, F-actin polymerization and cofilin phosphorylation in 1° h. AoSMCs and they do so, at least in part, by specifically and directly targeting TPa impairing its downstream signaling. On the other hand, neither vasodilatory agent directly target TPß.

As stated, it is widely held that agents that signal through either cAMP- or cGMP-second messenger systems play a critical counterbalancing/inhibitory affect on RhoA-mediated signaling cascades [35,36] as well as regulating Rho-mediated transcriptional responses through the serum response factor [40]. In fact within the vasculature, there is a critical reciprocal relationship between RhoA signaling and expression and that of NO-signaling and expression of endothelial nitric oxide synthase (eNOS) [6,36,41,42]. Moreover, in platelets there is a differential effect whereby cAMP/PKA inhibits both the Gq/PLCß-mediated aggregation and secretion and the G12/Rho-mediated shape change while cGMP/PKG signaling inhibits the former Gq/Ca2+ dependent mechanism but does not affect the latter RhoA/Ca2" independent mechanism [35]. Clearly many of the actions of cAMP and cGMP on RhoA signaling are mediated through their respective second messenger kinases PKA and PKG, respectively [6,35,36,43] and more recently it has been established that this may largely occur through their direct phosphorylation of RhoA itself at an identical site, namely Ser188 within its hypervariable region [36,42,44,45]. Whilst phosphorylation of RhoA at Ser188 does not apparently alter its association with either RhoGEFs or RhoGAPs (GTPase activating proteins), it significantly increases its interaction with RhoGDI (GDP dissociation inhibitor) thereby reducing the level of membrane bound RhoA and impairing its ability to activate its key effectors including Rho Kinases [36,46]. Moreover, in a recent study investigating NGF-

mediated RhoA responses in neuronal PC12 cells, Nusser et al. provided in vitro and in vivo evidence to suggest that Ser188 phosphorylation of RhoA impairs activation of Rho kinase (ROCK 1/ 2), but does not affect its ability to activate other Rho effectors including rhotekin, mDia-1 and PKN [47]. From their studies, they proposed that Ser188 phosphorylation of RhoA may act as a 'secondary molecular switch' capable of overriding GTP-elicited activation of certain RhoA effectors, such as ROCK, but directing it to signal with (an)other subset of Rho effectors, perhaps in a cell specific manner. Returning to studies herein on TPa- and TP(3-mediated RhoA signalling, both NO and prostacyclin directly target RhoA phosphor-ylation at Ser188 through their regulation of PKG and PKA signaling, respectively (data not shown). Hence, RhoA-mediated signaling by TPa is subject to regulation by both direct prostacyclin/PKA and NO/ PKG-inhibition mediated through their respective phosphorylation of Ser329 and Ser331 within the unique C-tail domain of TPa in addition to the more general type of regulation through Ser188 phosphorylation of RhoA. On the other hand, TP(3 is not a direct target for either PKA or PKG phosphorylation or inhibition, but its RhoA-mediated signaling would be sensitive to RhoA phosphorylation by either second messenger kinase. Whilst it has not as yet been established whether the "molecular switch mechanism" resulting from RhoAS188 phosphorylation proposed by Nusser et al. [47] to exist in neuronal cells can be extended to other cell/tissue types, such as smooth muscle, it is tempting to speculate.

Hence, as presented in our model (Fig. 10), we propose that as TPa is directly targeted for inhibition by prostacyclin and NO, its signaling

would be fully impaired by either vasodilator beginning at the level of the receptor itself. On the other hand, as TP(3 is not subject to direct PKA or PKG phosphorylation, its signaling by prostacyclin or NO may only be regulated at downstream intermediary level(s), such as at the level of RhoA phosphorylation. In the event that the 'phospho-RhoASer188 switch mechanism' exists within TXA2-responsive VSM or indeed in platelets, RhoA signaling through TP(3 may be directed away from one effector system, such as ROCK signaling, in the direction of another effector(s), such as rhotekin, mDia-1 and PKN (Fig. 10) as proposed by Nusser et al. [47] in the neuronal system, or indeed toward other subset(s) of the many diverse RhoA effectors, perhaps in a cell specific manner. Final clarification as to whether such a mechanism exists will require further detailed investigation.

Hence, in summary TPa- and TPp-mediated RhoA signaling functionally diverge at the point at which prostacyclin and NO affect the RhoA signaling cascade. These data further support the hypothesis that TPa is the major regulatory TP isoform involved in vascular hemostasis being a direct target for inhibition of both its Gq/PLCp/Ca2+-dependent and G12/RhoA/Ca2+-independent signaling by prostacyclin and NO within the vasculature. On the other hand, as TP(3 remains unaffected by either agent, at the interface of the receptor at least, the functional role of TPp remains to be further clearly defined. The data herein highlight further critical differences between the TPa and TP(3 receptor isoforms in terms of their regulation of Rho signaling that are likely to be physiologically relevant in human tissues such as SM and suggest that selective targeting and impairment of TPa-mediated signaling may offer a useful therapeutic approach in the treatment of

Fig. 10. Model of TPa and TPp-mediated RhoA activation and Signaling. Panels A and B: Agonist (TXA2/U46619)-activated TPa and TP(i couples to Gaq/PLC(i, yielding increases in IP3, mobilization of [Ca2+]¡ leading to sequential Ca2+/calmodulin-dependent activation of MLCK, MLC20 phosphorylation and actomyosin formation resulting in Ca2+-dependent F actin polymerization. Agonist-activated TPa and TP(i can simultaneously co-couple to Gi2/RhoGEF to activate RhoA and a host of its effectors including Rho kinase (ROCK), mammalian diaphanous protein (mDIA), Rhotekin, protein kinase (PK)N, amongst many others. ROCK phosphorylates: (i) the myosin-binding subunit (MBS) of myosin phosphatase (MPTase), inhibiting its activity; (ii) and activates CPI-17, a phosphorylation-dependent inhibitor of MPTase; (iii) MLC20 itself; (iv) LIM kinase (LIMK) which, in turn, phosphorylates and inactivates the actin depolymerizing agent Cofilin. These combined actions of ROCK contributes to the Rho A/Ca2+-independent mechanism for regulating stress fibre formation in non-muscle cells, smooth muscle contraction and platelet shape change. Panel A: The second messenger kinases cAMP-dependent PKA and cGMP-dependent PKG, activated in response to prostacyclin and NO signalling, respectively, cannot only impair TPa-mediated Gq-PLCp signaling and G12-RhoGEF signaling through direct phosphorylation of TPa at Ser329 and Ser331, respectively, but may also impair that TPa-mediated signalling, such as at the level of RhoA itself through Ser188 phosphorylation. Panel B: On the other hand, as TPp is not subject to direct PKA or PKG phosphorylation, its signaling by prostacyclin or NO may only be regulated at downstream intermediary level(s), such as at the level of RhoA (Ser188) phosphorylation. Y27632 is a selective inhibitor of Rho kinase. (The reader is referred to the web version of this article to see color images of this figure, where relevant.)

certain vascular diseases such as systemic- and pulmonary-hypertension in which both TXA2 and RhoA dysfunction has been implicated [2,48,49]. Moreover, the data also suggests that for effective impairment of TPß-mediated RhoA activation and signaling in such clinical settings, it may be necessary to fully antagonize it at the level of the TPß receptor itself rather than at a later downstream step, such as most typically at the level Rho kinase/ROCK inhibition [2].

Acknowledgements

This work was supported by grants from The Wellcome Trust, The Health Research Board, Ireland and Science Foundation Ireland.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.cellsig.2008.04.006.

References

[1] S. Offermanns, Circ. Res. 99 (12) (2006) 1293-1304.

[2] Y. Fukata, M. Amano, K. Kaibuchi, Trends Pharmacol. Sci. 22 (1) (2001) 32-39.

[3] A.L. Bishop, A. Hall, Biochem. J. 348 (Pt 2) (2000) 241-255.

[4] M. Schwartz, J. Cell Sci. 117 (Pt 23) (2004) 5457-5458.

[5] A.P. Somlyo, A.V. Somlyo, J. Physiol. 522 (Pt 2) (2000) 177-185.

[6] T.M. Seasholtz, J.H. Brown, Mol. Interv. 4 (6) (2004) 348-357.

[7] Y. Kaneko, T. Nakayama, K. Saito, A. Morita, I. Sato, A. Maruyama, M. Soma, T. Takahashi, N. Sato, Hypertens. Res. 29 (9) (2006) 665-671.

[8] B. Klages, U. Brandt, M.I. Simon, G. Schultz, S. Offermanns, J. Cell Biol. 144 (4) (1999) 745-754.

[9] A. Moers, B. Nieswandt, S. Massberg, N. Wettschureck, S. Gruner, I. Konrad, V. Schulte, B. Aktas, M.P. Gratacap, M.I. Simon, M. Gawaz, S. Offermanns, Nat. Med. 9 (11) (2003) 1418-1422.

[10] X. Ding, P.A. Murray, Am. J. Physiol. Lung Cell Mol. Physiol. 289 (5) (2005) L825-L833.

[11] A. Gohla, G. Schultz, S. Offermanns, Circ. Res. 87 (3) (2000) 221-227.

[12] M.H. Tsai, M.J. Jiang, Pflugers Arch. 453 (2) (2006) 223-232.

[13] M. Hirata, Y. Hayashi, F. Ushikubi, Y. Yokota, R Kageyama, S. Nakanishi, S. Narumiya, Nature 349 (6310) (1991) 617-620.

[14] R.M. Nusing, M. Hirata, A. Kakizuka, T. Eki, K. Ozawa, S. Narumiya, J. Biol. Chem. 268 (33) (1993) 25253-25259.

[15] M.K. Raychowdhury, M. Yukawa, L.J. Collins, S.H. McGrail, K.C. Kent, J.A. Ware, J. Biol. Chem. 269 (30) (1994) 19256-19261.

[16] A.W. Ashton, Y. Cheng, A. Helisch, J.A. Ware, Circ. Res. 94 (6) (2004) 735-742.

[17] A.W. Ashton, J.A. Ware, Circ. Res. 95 (4) (2004) 372-379.

[18] A.T. Coyle, S.M. Miggin, B.T. Kinsella, Eur. J. Biochem. 269 (16) (2002) 4058-4073.

[19] J.L. Parent, P. Labrecque, M.J. Orsini, J.L. Benovic, J. Biol. Chem. 274 (13) (1999) 8941-8948.

[20] H.M. Reid, B.T. Kinsella, J. Biol. Chem. 278 (51) (2003) 51190-51202.

[21] M.T. Walsh, J.F. Foley, B.T. Kinsella, J. Biol. Chem. 275 (27) (2000) 20412-20423.

[22] B.T. Kinsella, Biochem. Soc. Trans. 29 (Pt 6) (2001) 641-654.

[23] T. Hirata, F. Ushikubi, A. Kakizuka, M. Okuma, S. Narumiya, J. Clin. Invest. 97 (4) (1996) 949-956.

[24] R Vezza, A. Habib, G.A. FitzGerald, J. Biol. Chem. 274 (18) (1999) 12774-12779.

[25] A. Habib, G.A FitzGerald, J. Maclouf, J. Biol. Chem. 274 (5) (1999) 2645-2651.

[26] S.M. Miggin, B.T. Kinsella, Biochim. Biophys. Acta 1425 (3) (1998) 543-559.

[27] J.F. Foley, L.P. Kelley, B.T. Kinsella, Biochem. Pharmacol. 62 (2) (2001) 229-239.

[28] X.D. Ren, M.A. Schwartz, Methods Enzymol. 325 (2000) 264-272.

[29] W.G. Barnes, E. Reiter,J.D. Violin, X.R. Ren, G. Milligan, R.J. Lefkowitz,J. Biol. Chem. 280 (9) (2005) 8041-8050.

[30] S.M. Miggin, B.T. Kinsella, Biochim. Biophys. Acta 1539 (1-2) (2001) 147-162.

[31] M.A. Booden, D.P. Siderovski, C.J. Der, Mol. Cell Biol. 22 (12) (2002) 4053-4061.

[32] H. Chikumi, J. Vazquez-Prado, J.M. Servitja, H. Miyazaki, J.S. Gutkind, J. Biol. Chem. 277 (30) (2002) 27130-27134.

[33] S. Vogt, R Grosse, G. Schultz, S. Offermanns, J. Biol. Chem. 278 (31) (2003) 28743-28749.

[34] T. Sumi, K. Matsumoto, Y. Takai, T. Nakamura, J. Cell Biol. 147 (7) (1999) 1519-1532.

[35] M.P. Gratacap, B. Payrastre, B. Nieswandt, S. Offermanns, J. Biol. Chem. 276 ( 51) (2001)47906-47913.

[36] G. Loirand, C. Guilluy, P. Pacaud, Trends Cardiovasc. Med. 16 (6) (2006) 199-204.

[37] S. Ali, M.G. Davis, M.W. Becker, G.W. Dorn II, J. Biol. Chem. 268 (23) (1993) 17397-17403.

[38] A.W.Ashton, RYokota, G.John, S. Zhao, S.O. Suadicani, D.C. Spray,J.A.Ware,J. Biol. Chem. 274 (50) (1999) 35562-35570.

[39] S. Narumiya, Y. Sugimoto, F. Ushikubi, Physiol. Rev. 79 (4) (1999) 1193-1226.

[40] S. Zhuang, G.T. Nguyen, Y. Chen, T. Gudi, M. Eigenthaler, T. Jarchau, U. Walter, G.R Boss, R.B. Pilz, J. Biol. Chem. 279 (11) (2004) 10397-10407.

[41] M. Rolli-Derkinderen, V. Sauzeau, L. Boyer, E. Lemichez, C. Baron, D. Henrion, G. Loirand, P. Pacaud, Circ. Res. 96 (11) (2005) 1152-1160.

[42] V. Sauzeau, H. Le Jeune, C. Cario-Toumaniantz, A. Smolenski, S.M. Lohmann, J. Bertoglio, P. Chardin, P. Pacaud, G. Loirand, J. Biol. Chem. 275 (28) (2000) 21722-21729.

[43] J.M. Manganello, J.S. Huang, T. Kozasa, T.A Voyno-Yasenetskaya, G.C. Le Breton, J. Biol. Chem. 278 (1) (2003) 124-130.

[44] V. Sauzeau, M. Rolli-Derkinderen, C. Marionneau, G. Loirand, P. Pacaud, J. Biol. Chem. 278 (11) (2003) 9472-9480.

[45] N. Sawada, H. Itoh, J. Yamashita, K. Doi, M. Inoue, K. Masatsugu, Y. Fukunaga, S. Sakaguchi, M. Sone, K. Yamahara, T. Yurugi, K. Nakao, Biochem. Biophys. Res. Commun. 280 (3) (2001) 798-805.

[46] S.M. Ellerbroek, K. Wennerberg, K. Burridge, J. Biol. Chem. 278 (21) (2003) 19023-19031.

[47] N. Nusser, E. Gosmanova, N. Makarova, Y. Fujiwara, L. Yang, F. Guo, Y. Luo, Y. Zheng, G. Tigyi, Cell. Signal. 18 (5) (2006) 704-714.

[48] E.A. Meagher, G.A. FitzGerald, Circulation 88 (3) (1993) 1324-1333.

[49] W.G. Zhao, J.S. Richardson, Clin. Invest. Med. 13 (6) (1990) 343-352.