Shear stress induces Gα_q/11 activation independently of G protein-coupled receptor activation in endothelial cells

Abstract

Mechanochemical signal transduction occurs when mechanical forces, such as fluid shear stress, are converted into biochemical responses within the cell. The molecular mechanisms by which endothelial cells (ECs) sense/transduce shear stress into biological signals, including the nature of the mechanosensor, are still unclear. G proteins and G protein-coupled receptors (GPCRs) have been postulated independently to mediate mechanotransduction. In this study, we used in situ proximity ligation assay (PLA) to investigate the role of a specific GPCR/Gα_q/11 pair in EC shear stress-induced mechanotransduction. We demonstrated that sphingosine 1-phosphate (S1P) stimulation causes a rapid dissociation at 0.5 min of Gα_q/11 from its receptor S1P₃, followed by an increased association within 2 min of GPCR kinase-2 (GRK2) and β-arrestin-1/2 with S1P₃ in human coronary artery ECs, which are consistent with GPCR/Gα_q/11 activation and receptor desensitization/internalization. The G protein activator AlF₄ resulted in increased dissociation of Gα_q/11 from S1P₃, but no increase in association between S1P₃ and either GRK2 or β-arrestin-1/2. The G protein inhibitor guanosine 5′-(β-thio) diphosphate (GDP-β-S) and the S1P₃ antagonist VPC23019 both prevented S1P-induced activation. Shear stress also caused the rapid activation within 7 s of S1P₃/Gα_q/11. There were no increased associations between S1P₃ and GRK2 or S1P₃ and β-arrestin-1/2 until 5 min. GDP-β-S, but not VPC23019, prevented dissociation of Gα_q/11 from S1P₃ in response to shear stress. Shear stress did not induce rapid dephosphorylation of β-arrestin-1 or rapid internalization of S1P₃, indicating no GPCR activation. These findings suggest that Gα_q/11 participates in the sensing/transducing of shear stress independently of GPCR activation in ECs.

vascular endothelial cells (ECs) are exposed to mechanical forces from blood flow, which are sensed and transduced into intracellular biochemical responses. These signals contribute to the overall phenotype and function of normal ECs, but they can also lead to vascular pathologies (9). The identity of the primary mechanosensor, including its composition and subcellular structure, is still largely unknown. Fluid shear stress, the tangential component of hemodynamic forces, is known to activate heterotrimeric G protein subunits αq and -11 (Gα_q/11) in ECs within seconds of flow onset (16, 18). Gα_q/11 may also be activated independently of cytoskeletal and cytosolic components and occur in the absence of protein receptors (17), suggesting a critical role of the phospholipid bilayer. More recent evidence suggests that Gα_q/11 is part of a mechanosensitive complex together with platelet endothelial cell adhesion molecule-1 (PECAM-1) and GPCRs at the EC junction (11, 30, 46). However, it has also been postulated that GPCRs can be directly stimulated by mechanical forces, including fluid shear stress and stretch (7, 48).

GPCRs are a family of membrane receptors that respond to a diverse set of extracellular physical and chemical stimuli, including light, odor, neurotransmitters, cytokines, growth factors, lipids, and hormones to mediate a wide range of biological processes. Generally, binding of agonists to their respective GPCRs leads to the activation of intracellularly associated heterotrimeric G proteins, composed of α-, β-, and γ-subunits, and the subsequent dissociation of the Gα subunit from the dimeric Gβγ subunit and GPCR. In the continuous presence of agonist, GPCRs are phosphorylated by G protein-coupled receptor kinases (GRKs). Phosphorylation is followed by the recruitment and binding of β-arrestins (typically β-arrestin-1 and -2), desensitization, and internalization, which effectively terminates G protein signaling (31).

Sphingosine 1-phosphate (S1P) is a bioactive lipid that binds to and activates a family of five GPCRs, S1P_1–5, which are differentially coupled to G protein subtypes and are known to be expressed in a wide variety of tissues and cell types. Whereas S1P₁ is known to couple only to Gα_i and S1P₄ and S1P₅ to Gα_i and Gα_12/13, both S1P₂ and S1P₃ can couple to and activate Gα_i, Gα_q/11, and Gα_12/13 (35, 37). Activation of either S1P₁ or S1P₃, both of which are specifically expressed by ECs, can lead to the activation of several downstream signaling pathways, including the extracellular signal-regulated kinase (ERK) pathway (43).

To investigate the specific roles of Gα_q/11 and GPCRs in mechanochemical signal transduction, we compared shear stress- and ligand-induced activation of a specific GPCR/G protein pair by examining the endogenous interactions between S1P₃ and Gα_q/11, S1P₃ and GRK2, and S1P₃ and β-arrestin-1/2 in ECs upon stimulation using in situ proximity ligation assay (PLA). We found that this particular GPCR/G protein pair is mechanosensitive. By targeting Gα_q/11 and S1P₃ separately, we further demonstrated that shear stress-induced activation of Gα_q/11 occurs independently of S1P₃ activation. Our results reveal that shear stress-induced Gα_q/11 activation has a molecular signature distinct from that of ligand-induced Gα_q/11 activation.

MATERIALS AND METHODS

Cell culture.

Human coronary artery endothelial cells (HCAECs) from male donors were obtained from either Cell Applications (San Diego, CA) or Lonza (Walkersville, MD) and maintained in complete endothelial growth medium (EGM-2; Lonza) supplemented with 10% heat-inactivated FBS and penicillin-streptomycin. Before all experimental procedures, cells were seeded onto glass microscope slides, grown to confluence, and serum-starved overnight in ATP-free endothelial basal medium (EBM-2; Lonza) supplemented with 0.5% BSA. HCAECs within six passages were used for all experiments.

Reagents.

Sphingosine-1-phosphate (S1P) was purchased from either Tocris Bioscience (Bristol, UK) or Cayman Chemical (Ann Arbor, MI) and resuspended in PBS with 0.4% BSA. Serotonin hydrochloride (5-hydroxytryptamine, 5-HT) was purchased from Tocris Bioscience (Bristol, UK) and solubilized in water. AlF₄^- was prepared by mixing 30 μM AlCl₃ and 10 mM NaF. Guanosine 5′-O-(2-thiodiphosphate), trilithium salt (GDP-β-S) was purchased from EMD Millipore (Billerica, MA). VPC23019 was purchased from Tocris Bioscience and resuspended in dimethyl sulfoxide (DMSO). Anti-phospho-ERK1/2 (T202/Y204) (catalog no. 9102), anti-ERK1/2, anti-phospho-β-arrestin-1 (S412) (catalog no. 2416), and anti-Akt (S473) (catalog no. 9271) antibodies were all purchased from Cell Signaling Technology (Danvers, MA). Anti-β-arrestin-1 (catalog no. 610550) and anti-Akt antibodies were purchased from BD Biosciences (San Jose, CA) and Santa Cruz Biotechnology (Dallas, TX), respectively.

Shear stress.

Glass microscope slides with HCAECs were mounted on a conventional parallel-plate flow chamber (14) and subjected to a steady fluid shear stress of 14 dyn/cm² by perfusion with CO₂-equilibrated EBM-2 containing 0.5% BSA using a PHD 2000 syringe pump (Harvard Apparatus, Holliston, MA). Cells on slides that were mounted but not subjected to shear stress served as “Sham” controls.

In situ proximity ligation assay.

HCAEC monolayers on glass slides were immediately quenched in cold methanol-acetone (−20°C), rehydrated in ice-cold PBS, and then treated according to the manufacturer’s protocol (Olink Biosciences, Uppsala, Sweden). This method of fixing cells with cold methanol is routinely used to quench metabolic activity on a subsecond time scale for quantitative metabolomics (10). Cells were probed using primary antibodies that have been previously shown to be specific for S1P₃, Gα_q/11, and 5-hydroxytryptamine receptor 2A (5-HT_2A) (1, 27, 29). Primary antibodies used were a custom-made rabbit anti-Gα_q/11 (clone no. 47; Epitomics), goat anti-S1P₃/EDG-3 (V-20) (catalog no. sc-16076; Santa Cruz Biotechnology), mouse anti-SR-2A/5-HT_2A (A-4) (catalog no. sc-166775; Santa Cruz Biotechnology), rabbit anti-GRK2 (catalog no. 3982; Cell Signaling Technology), and rabbit anti-pan arrestin (catalog no. ab2914; Abcam, Cambridge, MA). The two fluorescence-labeled PLA probes used were Duolink In Situ PLA Probe Anti-Rabbit PLUS (catalog no. DUO92002) and Duolink In Situ PLA Probe Anti-Goat MINUS (catalog no. DUO92006), purchased from Sigma-Aldrich (Carlsbad, CA). When the two target proteins are bound with both primary and secondary antibodies, the oligonucleotide probes are hybridized to each other and ligated to form a closed circle. Polymerase-driven rolling circle amplification (RCA) generates a product to which the oligonucleotide probes hybridize, thereby forming a visible PLA signal. All primary antibody pairs and PLA probes were tested at different concentrations to ensure that the density of PLA signals were in the linear range for detection of effects and did not reach a saturation point (data not shown). As a negative control to assess background signal for each experiment, one of the primary antibodies for each pair was omitted (13) and showed very few PLA signal (Table 1). Single-sliced images were acquired on a LSM5 PASCAL confocal fluorescence microscope (Carl Zeiss, Germany) equipped with a Plan Apochromatic 63/1.4 numerical aperture oil immersion objective and both the PLA signal (i.e., single dots or pixels) and cell nuclei were quantified using a custom ImageJ image analysis macro. A minimum of ten fields of acquisition were acquired for each of at least three individual experiments.

Table 1.

PLA signal of HCAECs under basal conditions

Antibodies	Average Number of Fluorescent Dots per Microscope Field	Average Number of Nuclei per Microscope Field
S1P3 only	35.8 ± 2.9	17.2 ± 1.3
Gαq/11 only	21.2 ± 2.7	19.1 ± 0.6
GRK2 only	17.7 ± 1.8	19.7 ± 1.0
β-Arrestins only	25.2 ± 2.2	18.8 ± 0.9
S1P3/Gαq/11 pair	401.9 ± 18.8	22.2 ± 0.9
S1P3/GRK2 pair	87.8 ± 4.5	15.4 ± 0.6
S1P3/β-arrestins pair	83.8 ± 3.2	19.4 ± 0.9

Values are from a representative experiment and depicted as means ± SE. HCAECs, human coronary artery endothelial cells; S1P₃, sphingosine-1-phosphate receptor 3; GRK2, G protein-coupled receptor kinase-2.

Immunofluorescence receptor localization assay.

HCAEC monolayers on glass slides were incubated with 2.5 μM S1P or subjected to shear stress for 0, 5, and 10 min. Cells were then fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.3% Triton X-100 in PBS for 10 min, and blocked in 3% BSA in PBS for 1 h. Cells were incubated with goat anti-S1P₃/EDG-3 (V-20) (catalog no. sc-16076; Santa Cruz Biotechnology) diluted 1:100 in blocking buffer followed by incubation with a 1:200 dilution of Alexa Fluor 488 donkey anti-goat IgG (Molecular Probes, Eugene, OR).

Western blot analysis.

Proteins were separated on NuPAGE 4–12% Bis-Tris gels (Thermo Fisher Scientific) in MOPS SDS running buffer (Thermo Fisher Scientific) and transferred to PVDF membranes (Immobilon-P; Millipore, Temecula, CA). Membranes were blocked for 1 h with 3% BSA in Tris-buffered saline (TBS) and then incubated with a primary antibody for 2 h or overnight in 3% BSA-TBST (TBS with 0.1% Tween 20) at 4°C. After washing and incubating with horseradish peroxidase-conjugated secondary antibodies for 1 h was completed, the membranes were incubated with chemiluminescence substrate (SuperSignal West Pico or West Femto; Thermo Scientific, Rockford, IL). Images were acquired using a C-DiGit Blot Scanner (LI-COR Biosciences, Lincoln, NE).

Statistical analyses.

All experimental data are expressed as means ± SE from at least three independent experiments. Single comparisons between groups were performed using Student’s t-test, whereas multiple group comparisons were analyzed using one-way ANOVA with Bonferroni post hoc tests. P values of <0.05 were considered statistically significant.

RESULTS

S1P₃/Gα_q/11 activation is induced by both S1P and shear stress.

To gain insight into the mechanism(s) by which shear stress induces heterotrimeric G protein activation in endothelial cells, we examined the endogenous interactions between the specific GPCR/G protein pair, S1P₃/Gα_q/11, using in situ proximity ligation assay (PLA). We detected the presence of S1P₃ in close proximity to Gα_q/11 in HCAECs under basal conditions as indicated by the presence of red fluorescent dots, which we refer to as PLA signal and is indicative of the relative number of complexes (Fig. 1A). In cells treated with S1P (2.5 μM) for 0.5 min, we observed a significant decrease (37%) in PLA signal detected by the S1P₃/Gα_q/11 antibody pair compared with that in the untreated control condition, which suggests rapid dissociation of Gα_q/11 from S1P₃ and therefore rapid activation of the complex.

Fig. 1.

Ligand- and shear stress-induced activation of the sphingosine-1-phosphate (S1P₃)-Gα_q/11 complex. In situ proximity ligation assay (PLA) was performed using antibodies directed against S1P₃ and Gα_q/11 (A and B), serotonin (5-HT)2A and Gα_q/11 (C and D), S1P₃ and G protein-coupled receptor kinases (GRK2; E), and S1P₃ and β-arrestin-1/2 (F) on human coronary artery endothelial cells (HCAECs) treated with vehicle alone (Control), S1P; 2 μM for 30 s) (A, E, and F), or 5-HT (100 nM for 30 s) (C) and on HCAECs that were either unstimulated (Sham) or subjected to a step change in shear stress (Flow) for the indicated times (B, D, E, and F). Representative confocal images in a single z-plane of cells probed with the S1P₃/Gα_q/11 antibody pair are shown (A and B). Scale bar, 20 μm. Each bar graph shows quantification of at least 3 independent experiments as PLA signal (red dots) per cell (blue nuclei) relative to the control condition, with error bars indicating SE; n = 6 for C, n = 5 for A, B, E (Flow), and F (Flow), n = 3 for D, E (S1P), and F (S1P). *P < 0.05; **P < 0.01; ***P < 0.001.

To determine whether shear stress can activate the S1P₃/Gα_q/11 complex, we examined the association between S1P₃ and Gα_q/11 in HCAECs by PLA. Interestingly, shear stress induced an 18% decrease in PLA signal as early as 7 s after stimulation (Fig. 1B). The PLA signal is decreased even further at 15 s (39%) and remains decreased through 30 s (32%) and up to 60 s (37%).

To verify the general applicability of the in situ PLA technique to monitor the activation of GPCR/G proteins, we examined the endogenous interactions between the serotonin receptor, 5-HT_2A, and Gα_q/11 in response to both its natural ligand, serotonin (Fig. 1C), and shear stress (Fig. 1D). HCAECs stimulated with 5-HT (100 nM) for 30 s showed a decrease in the number of 5-HT_2A·Gα_q/11 complexes (23%). Similarly, the number of complexes was decreased with the rapid onset of flow at 7 s (23%) and reaching a maximum decrease at 30 s (40%) compared with the sham control condition.

S1P, but not shear stress, induces association of GRK2 and β-arrestin1/2 with S1P₃.

We next examined the association of S1P₃ with GRK2 in response to S1P and shear stress stimulation to determine the recruitment of GPCR kinases to activated GPCRs. Stimulation of HCAECs with S1P for 0.5 min resulted in a marked increase (99%) in PLA signal compared with untreated controls (Fig. 1E). However, this relative change in signal was transient as the number of fluorescent dots decreased to baseline levels at 2 and 5 min. In contrast, in cells subjected to shear stress, there was no increase in association between S1P₃ and GRK2 at 0.5 and 2 min and only a modest increase at 5 min (22%).

Since β-arrestins bind to and negatively regulate almost all activated GPCRs, we examined the association of S1P₃ with β-arrestin-1/2 to determine the recruitment of arrestins to activated GPCRs. An increase was observed in the number of dots detected by the S1P₃/ β-arrestin-1/2 antibody pair with a peak (4.5-fold) at 2 min upon stimulation with S1P (Fig. 1F). At 10 min, the number of dots detected was slightly below the baseline level, perhaps reflecting internalization of the receptor. In contrast, shear stress did not increase the PLA signal detected by proximity of S1P₃ and β-arrestin-1/2 at 2 min, but there was a steady increase that began at 5 min after exposure to shear stress and continued through 10 min (51 and 81%, respectively).

S1P, but not shear stress, induces S1P₃ internalization.

Since most canonical GPCRs undergo rapid β-arrestin-dependent internalization upon ligand activation, we investigated whether S1P₃ is internalized in response to either S1P stimulation or shear stress in HCAECs by using immunofluorescence staining. In vehicle-stimulated cells, S1P₃ had a perinuclear Golgi-like staining in addition to a diffuse cell surface localization (Fig. 2A). Stimulation of cells with S1P for 5 min did not appear to change this S1P₃ localization pattern. At 10 min, however, S1P₃ immunostaining became more punctate and localized in the cytoplasm. Shear stress, on the other hand, did not induce cytoplasmic localization of S1P₃ at any of the time points that were evaluated (Fig. 2B). This finding suggests that shear stress, in contrast to S1P, does not induce S1P₃ internalization.

Fig. 2.

Ligand- and shear-induced internalization of S1P₃. HCAECs were treated with S1P at 2.5 μM (A) or exposed to flow at 14 dyn/cm² (B) for indicated time points. Cells were then fixed, permeabilized, blocked, and immunostained with an anti-S1P₃/EDG-3 primary antibody followed by an Alexa fluor 488-conjugated secondary antibody. Representative images from 3 independent experiments are shown.

β-Arrestin-1 is rapidly dephosphorylated in response to S1P but not shear stress.

Cytosolic β-arrestin-1 is known to be constitutively phosphorylated on serine 412 (S412) and is recruited to and rapidly dephosphorylated at the plasma membrane upon agonist stimulation, a process that is required for clathrin-mediated receptor internalization (24). We therefore measured the phosphorylation/dephosphorylation status of β-arrestin-1 in HCAECs that were stimulated with either S1P or shear stress. Western blot analyses of lysates prepared from S1P-activated cells showed rapid dephosphorylation, as reflected by a decrease in phosphorylation of S412 at both 2 and 5 min poststimulation (20 and 15%, respectively) (Fig. 3). In contrast, flow did not induce any β-arrestin-1 dephosphorylation at the same time points. As a positive control, immunoblotting for phosphorylated Akt, which is a known downstream event of both signaling pathways (12, 20), confirmed that cells were indeed activated by each stimulus.

Fig. 3.

Ligand- and shear stress-induced dephosphorylation of β-arrestin-1. HCAECs were treated with S1P (2.5 μM) or exposed to flow (14 dyn/cm²) for indicated time points. Immunoblotting was performed on cell lysates using antibodies against phosphorylated β-arrestin-1 (S412) and total β-arrestin-1 to assess the level of phosphorylation/dephosphorylation. As a positive control for cell stimulation, lysates were also probed for phosphorylated Akt (S473) and total Akt. Representative blots from 5 independent experiments are shown. Bar graph represents the quantification as a ratio of phospho-S412 over total β-arrestin-1, with the mean value of the control condition arbitrarily set to 100% and error bars indicating SE; n = 5. *P < 0.05.

These differences in the molecular association patterns, receptor internalization, and β-arrestin-1 phosphorylation/dephosphorylation status between the S1P-induced and shear stress-induced conditions led us to hypothesize that shear stress may cause Gα_q/11 activation independently of GPCR activation. We therefore sought to determine the patterns if Gα_q/11 was directly activated independently of S1P₃ activation. To this end, we examined cells treated with the G protein activator AlF₄⁻ in the absence of S1P stimulation. Our results showed that Gα_q/11 is transiently dissociated from S1P₃ at 2 min (Fig. 4A). In contrast to S1P-induced cells, in which S1P₃ was directly activated by ligand binding, there was no observable increase in association of S1P₃ with either GRK2 (Fig. 4B) or β-arrestin-1/2 (Fig. 4C) in AlF₄^--treated ECs. In fact, GRK2 appeared to be dissociated from S1P₃ at 2 min, as suggested by the decrease in PLA signal. To verify that AlF₄⁻ was used at an active concentration, we examined ERK1/2 activation, a pathway known to be mediated by Gα_q/11. Western blot analysis showed that AlF₄^- is a potent activator of ERK1/2 activation, as indicated by its marked increase in phosphorylation at 5 min (Fig. 4D).

Fig. 4.

AlF4 (30 μM AlCl₃ + 10 mM NaF)-induced activation of Gα_q/11. HCAECs were treated with vehicle or AlF₄^- for indicated time points. In situ PLA was performed using antibodies against S1P₃ and Gα_q/11 (A), S1P₃ and GRK2 (B), and S1P₃ and β-arrestin-1/2 (C). Each bar graph shows quantification of 3 independent experiments as PLA signal per cell relative to the vehicle-treated control condition, with error bars indicating SE; n = 3. D: immunoblotting for phospho-ERK1/2 was performed on lysates prepared from cells stimulated with AlF₄^- for 5 min. A representative blot from 3 independent experiments is shown. Bar graph represents the quantification as a ratio of phospho-ERK1/2 over total ERK1/2 levels, with the mean value of the control condition set to 1 and error bars indicating SE; n = 3. *P < 0.05; **P < 0.01; ***P < 0.001.

GDP-β-S prevents both S1P- and shear stress-induced Gα_q/11 activation.

We next investigated whether the observed S1P-induced dissociation of the S1P₃/Gα_q/11 complex represents Gα_q/11 activation, S1P₃ activation, or both. For the first set of studies, cells were treated with GDP-β-S, a metabolically stable analog of GDP that binds to Gα proteins and inhibits binding and activation by GTP. GDP-β-S (300 μM, 4 h), but not vehicle, prevented the decrease in association between S1P₃ and Gα_q/11 in S1P-induced cells at 30 s (Fig. 5A). GDP-β-S also abrogated S1P-induced association of GRK2 with S1P₃ at 30 s (Fig. 5B). Interestingly, GDP-β-S did not completely block the S1P-induced increase in association of β-arrestin-1/2 with S1P₃, but caused a significant decrease in their S1P-induced association relative to vehicle control at 2 min (Fig. 5C). These findings suggest that the dissociation of Gα_q/11 from S1P₃ in response to S1P treatment is indicative of its activation.

Fig. 5.

Effects of G protein inhibition on ligand- and shear stress-induced Gα_q/11 activation. HCAECs were pretreated with vehicle or guanosine 5′-(β-thio)diphosphate (GDP-β-S; 300 μM, 4 h), an inhibitor of G protein activation, before stimulation with S1P (2 μM). In situ PLA was performed using antibodies directed against the pairs S1P₃ and Gα_q/11 (A) and S1P₃ and GRK2 (B) at 0.5 min, and the pair S1P₃ and β-arrestin-1/2 (C) at 2 min. Each bar graph shows quantification of at least 3 independent experiments as PLA signal per cell relative to the vehicle-treated control condition, with error bars indicating SE; n = 6 for A, n = 4 for B, n = 3 for C. D: HCAECs were pretreated with vehicle or with GDP-β-S (300 μM, 4 h) before exposure to shear stress (Flow; 15 s). In situ PLA was performed with antibodies against S1P₃ and Gα_q/11. Bar graph shows quantification of 3 independent experiments as PLA signal per cell relative to the vehicle-treated control condition with error bars indicating SE; n = 3. *P < 0.05; **P < 0.01; ***P < 0.001.

To determine whether activation of the S1P₃/Gα_q/11 complex by shear stress is due to direct activation of Gα_q/11 or is mediated specifically through S1P₃, we sought to target each molecule separately. GDP-β-S blocked the shear stress-induced rapid dissociation of S1P₃/Gα_q/11 at 15 s (Fig. 5D), while vehicle alone did not.

VPC23019 blocks Gα_q/11 activation by S1P but not by shear stress.

To determine whether the dissociation that we observed between S1P₃ and Gα_q/11 within 0.5 min of HCAEC stimulation with S1P is also indicative of S1P₃ activation, we performed PLA experiments using cells that were pretreated for 2 min with either vehicle or the S1P₃-selective antagonist VPC23019 at 10 μM (Fig. 6A). In vehicle control-treated cells induced with S1P for 0.5 min, we observed a significant decrease in PLA signal detected by the S1P₃/Gα_q/11 antibody pair. VPC23019 prevented this decrease in PLA signal, as indicated by the lack of a significant change in the number of S1P₃/Gα_q/11 complexes detected at 0.5 min.

Fig. 6.

Effects of S1P₃ inhibition on ligand- and shear stress-induced Gα_q/11 activation. A and C: HCAECs were pretreated with vehicle or the S1P receptor antagonist VPC23019 (10 μM), for 2 min before stimulation with S1P (2.5 μM). In situ PLA was performed using antibodies directed against S1P₃ and Gα_q/11 on cells stimulated for 0.5 min (A), S1P₃ and GRK2 on cells stimulated for 0.5 min (B), and S1P₃ and β-arrestin-1/2 on cells stimulated for 2 min (C). Each bar graph shows quantification of 3 independent experiments as PLA signal per cell relative to the vehicle-treated control condition with error bars indicating SE; n = 3. D: immunoblotting for phospho-ERK1/2 was performed on cell lysates at 5 min poststimulation with S1P in the absence or presence of VPC23019 pretreatment. A representative blot from 3 independent experiments is shown. Bar graph represents the quantification as a ratio of phospho-ERK1/2 over total ERK1/2 levels, with the mean value of the control condition set to 1 and error bars indicating SE. E: HCAECs were pretreated with vehicle or VPC23019 (10 μM, 2 min) before flow. In situ PLA was performed using antibodies against S1P₃ and Gα_q/11. Bar graph shows quantification of 3 independent experiments as PLA signal per cell relative to the vehicle-treated control condition, with error bars indicating SE; n = 3. F: immunoblotting for phospho-ERK1/2 was performed on lysates from cells subjected to flow for 5 min in the absence or presence of VPC23019 pretreatment. A representative blot from 4 independent experiments is shown. Bar graph represents the quantification as a ratio of phospho-ERK1/2 over total ERK1/2 levels, with the mean value of the control condition set to 1 and error bars indicating SE; n = 4. *P < 0.05; ***P < 0.001.

The effects of VPC23019 on the S1P-induced associations of S1P₃ with GRK2 and β-arrestin-1/2 were also investigated. The dramatic increase in association of S1P₃ with GRK2 by S1P at 0.5 min in the presence of vehicle alone was blocked by the pretreatment with VPC23019 (Fig. 6B). Likewise, there was an increase in the number of S1P₃-β-arrestin-1/2 complexes in vehicle-treated cells after 2 min of S1P stimulation, but no apparent increase in S1P₃-β-arrestin-1/2 complexes with VPC23019 (Fig. 6C).

Western blot analysis was also performed to confirm that VPC23019 under the selected conditions inhibits downstream signaling induced by S1P stimulation (Fig. 6D). As expected, we observed a significant increase in ERK1/2 phosphorylation (2.9-fold increase) with vehicle control at 5 min poststimulation with S1P, which was completely abrogated by VPC23019.

For cells subjected to shear stress in the presence of VPC23019, there was still a significant decrease in the number of S1P₃/Gα_q/11 complexes (Fig. 6E), just as there had been in the absence of the S1P₃ inhibitor, indicating that S1P₃ inhibition does not have an effect on shear stress-induced S1P₃/Gα_q/11 activation. Additionally, shear stress-induced ERK1/2 phosphorylation, which is known to be mediated by Gα_q/11 activation (15), was unchanged in the presence of VPC23019 at 5 min (Fig. 6F). It should be noted here that ERK1/2 phosphorylation was examined at 5 min as opposed to 10 min poststimulation to avoid autocrine effects due to shear stress-induced S1P secretion (40). Collectively, these data demonstrate that shear stress-induced activation of the S1P₃/Gα_q/11 complex is not dependent on activation of S1P₃.

DISCUSSION

GPCRs are activated by a wide array of external stimuli, which can be either chemical or physical in nature. Here, we show that the specific GPCR/G protein pair S1P₃/Gα_q/11 is mechanosensitive and is activated within 7 s of the onset of fluid shear stress. Although mechanical stimulation, in the form of either membrane stretch or fluid shear stress, has been previously shown to activate a number of different GPCRs in different cell types (38), the majority of the studies either examined events that were distal from receptor activation (i.e., Ca²⁺ measurements and ERK phosphorylation), or utilized approaches [i.e., fluorescence resonance energy transfer (FRET) and bioluminescence resonance energy transfer (BRET)], which required the transfection and overexpression of constructs. Furthermore, these expression constructs contain fluorescence proteins on the third intracellular loop and COOH terminus of GPCRs, which are known to decrease ligand-binding affinity and coupling to G proteins (41). Our study presents a novel approach to investigate a more proximal event in the activation of endogenous GPCRs and measure it in the context of their native cellular and subcellular environment, without the need for overexpression.

In situ PLA has recently been utilized to demonstrate GPCR heterodimerization without the need for heterologous expression (5) and has even been used as a stand-alone method in detecting and quantifying dimerization between transcription factors (2) as well as interactions between receptors and their effector molecules (39). This technique has also been used to confirm junctional associations between Gα_q/11 and PECAM-1 in quiescent endothelial cells and to detect dissociations between these two proteins in response to a sudden temporal change in shear stress (11). The association between Gα_q/11 and PECAM-1 under basal conditions and their dissociation from one another in response to shear stress had previously been shown independently using coimmunoprecipitation (co-IP) (30). It should be noted that endogenous levels of PECAM-1 in ECs are typically high, which eliminates the need for its heterologous expression. There have also been several other studies that have shown interactions between two proteins by in situ PLA and verified by co-IP or vice versa (4, 33, 34, 36, 45). In the majority of the aforementioned studies, co-IP was performed using lysates from cells that were transfected with constructs to overexpress their proteins of interest. In situations in which the endogenous levels of two proteins of interest are relatively low, the likelihood of detecting an interaction by conventional co-IP is also predictably low and therefore other alternative methods, such as in situ PLA, must be utilized. The solubility of a target protein, especially a seven-transmembrane receptor such as S1P₃, in low-stringency conditions that would favor protein-protein interactions, may also be low. We attribute the lack of detection of an interaction between endogenous S1P₃ and Gα_q/11 in ECs by co-IP (data not shown) in our studies to these two aforementioned reasons. The strengths of in situ PLA are that it allows for detection of associations between endogenous proteins without the need for expression constructs, the resolution is on the order of individual interactions between two proteins, and it is highly sensitive due to rolling circle amplification (8). Through the use of in situ PLA, we demonstrated that S1P stimulation and exposure of endothelial cells to shear stress both cause rapid dissociation of Gα_q/11 from S1P₃, which suggests the activation and release of Gα_q/11 from the GPCR/G protein complex in both instances.

In the case of S1P stimulation, there was also a rapid and transient increase in the association of S1P₃, first with GRK2, and then with β-arrestin-1/2. These events were then followed by a return to their respective baseline complex levels. Together, these findings are consistent with the recruitment of GRK2, which functions to phosphorylate ligand-bound and activated GPCRs, and the subsequent binding of β-arrestin-1/2, which either serves to initiate GPCR inactivation through the canonical pathway of desensitization and internalization or to mediate G protein-independent signaling (42). S1P stimulation indeed caused a change in S1P₃ localization from the cell surface to the cytoplasm after 10 min, which is consistent with two previous studies, one that showed that S1P induces internalization of S1P₃ within 45 min in human umbilical vein endothelial cells and another that demonstrates epitope-tagged S1P₃ labeling to be punctate and accumulate in the cytoplasm in cells treated with S1P for 15 min (19, 22). This result strongly suggests that the β-arrestin-1/2 recruitment observed by in situ PLA at 2 min may be initiating S1P₃ internalization in response to S1P stimulation.

In addition to demonstrating for the first time that shear stress rapidly induces the dissociation of Gα_q/11 from a specific GPCR, we also provide evidence that it was not a result of direct activation of the GPCR but rather direct activation of Gα_q/11, independent of GPCR activation. The timing of endogenous Gα_q/11 activation in the present study is consistent with previous work showing Gα_q/11 to be rapidly activated within 1 s of flow onset (18). Gα_q/11 mediates both shear stress-induced Ras activation (16) and intracellular Ca²⁺ release (27) within 5–20 s, which implies that Gα_q/11 activation must occur before these events. Indeed, detection of shear-induced Gα_q/11 activation by in situ PLA places it right at or before these two relatively early signal transduction events. There is evidence that GPCRs can be directly activated by shear stress. For example, it has been shown that the activities of the Gα_q/11-coupled GPCRs, bradykinin receptor B2 (B2R) and parathyroid hormone type 1 receptor are increased by shear stress (7, 47). However, in these particular studies, each GPCR was overexpressed in cells as a FRET sensor, and activation was inferred from conformational changes. Additionally, responses to shear stress were detected only after long times (e.g., 80 s) (7), which are far longer than it takes for shear stress to activate Gα_q/11.

Direct activation of Gα_q/11 with the G protein activator AlF₄^- induced Gα_q/11 dissociation from S1P₃. However, GRK2 and β-arrestin-1/2 were not recruited to S1P₃, as AlF₄^- activates G proteins independently of GPCR activation. Unexpectedly, a decreased association of GRK2 with S1P₃ was observed, which coincided with the dissociation of Gα_q/11 from S1P₃. This is likely due to the higher binding affinity that activated Gα_q/11 has for GRK2 (6). Activated Gα_q/11 may function to sequester GRK2, preventing it from triggering desensitization/internalization of S1P₃ in the absence of its ligand.

Although S1P ligand stimulation and shear stress both appear to rapidly activate the S1P₃/Gα_q/11 complex in endothelial cells with similar temporal kinetics (within 0.5 min), our collective data support the notion that shear stress activates the complex through a mechanism that is distinct from the classical model of ligand-induced GPCR signaling. First, VPC23019, a selective S1P₃ antagonist, dramatically inhibited S1P-induced S1P₃/Gα_q/11 dissociation, S1P₃/GRK2 and S1P₃/β-arrestin-1/2 associations, and ERK1/2 phosphorylation but had no effect on the shear stress-induced dissociation of Gα_q/11 from S1P₃. This indicates that shear stress activates S1P₃-associated Gα_q/11 without activation of S1P₃. Second, S1P stimulation, but not shear stress exposure, induces a change in S1P₃ localization from the cell surface to the cytoplasm, indicating that shear stress does not initiate S1P₃ internalization upon EC activation. Finally, β-arrestin-1 is not dephosphorylated at S412 in response to shear stress as it is upon stimulation with S1P. Dephosphorylation of β-arrestin-1 occurs upon activation of a GPCR, as has been shown specifically for the follicle-stimulating hormone (FSH) receptor in Sertoli cells upon FSH stimulation (26). This suggests that shear stress does not activate GPCRs which utilize β-arrestin-1 for clathrin-mediated receptor internalization. Since β-arrestin-2 can also be dephosphorylated upon GPCR activation (23), one cannot exclude the possibility that shear stress activates GPCRs that utilize β-arrestin-2 for this purpose.

GRK2 and β-arrestin-1/2 showed increased associations with S1P₃ with shear stress stimulation, but these associations were significantly delayed compared with those for ligand-induced stimulation, and there was no rapid return to baseline levels. This delayed phenomenon may in fact be due to autocrine signaling, as it has been previously shown that ECs can synthesize and secrete S1P in response to shear stress (40). Thus it is likely that the increased and more sustained associations of S1P₃ with GRK2 and β-arrestin-1/2 that we observed at the later time points (5 and 10 min) in response to shear stress are due to autocrine S1P-induced activation of S1P₃.

While our data demonstrate that shear stress does not directly activate S1P₃, one cannot exclude the possibility that another unidentified Gα_q/11-coupled GPCR, which is heterodimerized with S1P₃, is directly activated. Recent reports describe cross-signaling in GPCR heteromers whereby ligand activation of one GPCR constituent leads to the G protein activation of another GPCR (28). The S1P₁ receptor, which can also be induced by S1P binding, has previously been shown to be essential for both acute and chronic shear stress signaling in ECs (21). However, the “acute” signaling events referred to in that study, i.e., ERK, Akt, and eNOS activation, were all examined 10 min after flow onset, which suggests that the delayed GRK2 and β-arrestin-1/2 that we observed in the present study, are mediated by autocrine signaling rather than direct activation of S1P₁.

It has been previously reported that cyclic mechanical stretch induces the activation of the GPCR angiotensin II type I receptor (AT1R), which requires β-arrestins and GRKs but is both ligand and G protein activation independent (32). While this may appear to contradict our findings, the differences may be attributed to the fact that cyclic stretch and shear stress are distinct mechanical forces with opposing biophysical mechanisms of mechanotransduction (44). In addition, each study was performed using a different cell system, with the present study looking specifically at endogenous activity in human coronary artery ECs as opposed to cells stably expressing the GPCR of interest.

Despite our findings here that shear stress induces activation of Gα_q/11 independently of S1P₃ activation, S1P₃ may still play an important role in endothelial mechanosensing in response to temporal changes in shear stress. However, it is unlikely that S1P₃ is the only GPCR involved in mechanotransduction. In fact, there are an increasing number of reports that claim the importance of GPCRs in shear stress-induced signaling, including B2R (7, 46), S1P₁ (21), formyl peptide receptor (25), and AT1R (3). We speculate that a mechanosensitive complex that contains heterooligomerized GPCRs, which function as a scaffold and reservoir for G proteins, is localized at cell-cell junctions with PECAM-1 through interactions with HSPGs (11, 30, 46). By not functioning as receptors per se, these GPCRs need not be desensitized and recycled back to the cell surface by a β-arrestin-dependent mechanism in response to shear stress. The absence of any changes in S1P₃ localization in response to shear stress implies that shear stress does not trigger S1P₃ activation, desensitization, and/or internalization.

Our data not only demonstrate that shear stress causes the rapid dissociation of Gα_q/11 from S1P₃, but that it does so without the rapid association of GRK2 and β-arrestin-1/2, without the internalization of S1P₃, without the global dephosphorylation of β-arrestin-1, and despite the presence of a S1P₃ antagonist. Together, these findings indicate that shear stress can activate Gα_q/11 and its downstream signaling in a manner that is distinct from its activation via agonist stimulation. We also show that in situ PLA can be a powerful tool for the general detection of endogenous GPCR/G protein activation, as supported by our findings on two different GPCR/G protein pairs, which may have future applications for the screening of lead compounds for the identification of novel agonists and/or antagonists targeting GPCR/G protein activation.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute MERIT Award R37 HL040696.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

N.G.d.P., B.M., and J.A.F. conceived and designed research; N.G.d.P. and B.M. performed experiments; N.G.d.P. and B.M. analyzed data; N.G.d.P., B.M., and J.A.F. interpreted results of experiments; N.G.d.P. prepared figures; N.G.d.P. drafted manuscript; N.G.d.P. and J.A.F. edited and revised manuscript; N.G.d.P., B.M., and J.A.F. approved final version of manuscript.