Molecular Regulation of the RhoGAP GRAF3 and Its Capacity to Limit Blood Pressure In Vivo

Abstract

Anti-hypertensive therapies are usually prescribed empirically and are often ineffective. Given the prevalence and deleterious outcomes of hypertension (HTN), improved strategies are needed. We reported that the Rho-GAP GRAF3 is selectively expressed in smooth muscle cells (SMC) and controls blood pressure (BP) by limiting the RhoA-dependent contractility of resistance arterioles. Importantly, genetic variants at the GRAF3 locus controls BP in patients. The goal of this study was to validate GRAF3 as a druggable candidate for future anti-HTN therapies. Importantly, using a novel mouse model, we found that modest induction of GRAF3 in SMC significantly decreased basal and vasoconstrictor-induced BP. Moreover, we found that GRAF3 protein toggles between inactive and active states by processes controlled by the mechano-sensing kinase, focal adhesion kinase (FAK). Using resonance energy transfer methods, we showed that agonist-induced FAK-dependent phosphorylation at ^Y376GRAF3 reverses an auto-inhibitory interaction between the GAP and BAR-PH domains. Y376 is located in a linker between the PH and GAP domains and is invariant in GRAF3 homologues and a phosphomimetic ^E376GRAF3 variant exhibited elevated GAP activity. Collectively, these data provide strong support for the future identification of allosteric activators of GRAF3 for targeted anti-hypertensive therapies.

Keywords: FAK; GRAF3; RhoA; blood pressure; cardiovascular; hypertension; smooth muscle.

Figure 1

Smooth muscle cell (SMC)-specific GRAF3^RQ expression leads to a prolonged decrease in basal systolic blood pressure and limits hypertension (HTN). (A) Schematic of constructs used to develop tamoxifen-inducible SMC-specific GRAF3^RQ expression. (B) Western analysis of Cos cells transfected with the GRAF3^RQ plasmid and infected with Cre or LacZ control virus. (c) Western analysis of bladder lysates from control and GRAF3^RQ SM MHC-CreER^T2 mice treated with tamoxifen (100 mg/kg for 3 consecutive days); n = 3 per group. (D) RT-PCR analysis of (E) GRAF3 and GAPDH or (F) smooth muscle marker gene SM22 and ACTB mRNA levels in thoracic aorta lysates from GRAF3^RQ SM MHC-CreER^T2 and genetic control mice treated with tamoxifen; n = 4 per group, *p < 0.05. (G) Average 24-h systolic blood pressure, measured via radio-telemetry, of unrestrained, conscious GRAF3^RQ and GRAF3^RQ SM MHC-CreER^T2 mice before and after tamoxifen treatment (100 mg/kg for 3 consecutive days) and increasing L-NAME doses (50 mg/L, 150 mg/L or 450 mg/L) given for a week (each) in drinking water. Data are expressed as mean ± SD; n = 4 for GRAF3^RQ mice and n = 3 for GRAF3^RQ SM MHC-CreER^T2; *p < 0.05 vs. GRAF3^RQ (Student’s t-test). Linear regression analysis was performed to compare the slope of the two groups after L-NAME treatment (p = 0.1137).

Figure 2

BAR-PH mediated autoinhibition of GRAF3. (A) Schematic of GRAF3 monomer domain structure. (B) Immunofluorescent staining of RaAoSMCs transfected with Myc-GRAF3 alone or Myc-GRAF3 co-expressed with Flag-BAR-PH (BPH) domain. Yellow arrow and dotted-outline indicate phalloidin-stained cell of interest that is positive for Myc- (green) and/or Flag- (red) staining. Data are representative of over 50 cells/condition from 3 separate experiments (58/59 GRAF3 expressing cells and 5/72 cells co-expressing GRAF3 and BAR-PH exhibited reduced stress fibers). (C) Schematic of GRAF3 in open (active) or auto-inhibited (inactive) conformations. Three-dimensional structures of the GRAF3 BAR-PH-GAP dimer were created using Pymol and the solved, similar structures of Appl1 (BAR-PH) and GRAF1 (GAP). ClusPro docking simulations predicted 2 conformations for GRAF3, (D) open and active or (E) closed and auto-inhibited. Color scheme follows: BAR 1 (dark purple), PH 1 (light purple), BAR 2 (dark green), PH 2 (light green), GAP 1 and 2 (yellow), arginine fingers (active site) (dark blue), RhoA docking sites (pink), C-terminus of PH domain (red), N-terminus of GAP domain (orange); residues in teal aids in visualizing rotation.

Figure 3

GRAF3 conformation is physiologically regulated in a spatial and temporal fashion. (A) Schematic of GRAF3 biosensor when there would be low and high fluorescence resonance energy transfer (FRET) signal. (B) RaAoSMC were transfected with the GRAF3 biosensor and CFP signal observed at baseline and 5 min after treatment with the contractile agonist S1P (10 µM). (C) FRET was monitored in RaAoSMC before and after treatment with 10 µM S1P. Note the dynamic temporal change in FRET and high levels of CFP (but lack of FRET) at the cell periphery 20 min following treatment, which indicates that GRAF3 is most active in these protrusive areas; yellow arrows. (D) Analysis of FRET/CFP ratio over time. Images are representative of 3 independent experiments with n = 7 cells per experiment. Data are expressed as mean ± SD; * p < 0.05 as assessed by one-way ANOVA and Tukey HSD. Each time point is significantly changed except for 0 min vs. 4 min (p = 0.11).

Figure 4

Both Src and FAK kinases phosphorylate GRAF3 at Y376. (A) 3-D (left) and 2-D (right) schematics of GRAF3 indicate location of Y376 in the unstructured, un-modeled linker between the C-terminus of the PH domain (red) and the N-terminus of the GAP domain (orange). (B–E) Cos cells were transfected with the indicated Myc-GRAF3 variant and either (B,D) ^529FSrc (Src) or (C,E) Flag-superFAK cDNAs. Myc-tagged GRAF3 was immunoprecipitated from cell lysate and blots were probed with indicated antibodies. (F,G) Purified GRAF3 BAR-PH-GAP (BPG) domain and ^Y376FGRAF3 BAR-PH-GAP were subjected to a radioactive kinase assay using activated (F) Src or (G) FAK and ATP γ-³²P. Phosphorylated and total GRAF3 are shown by radiograph (top) or Coomassie Blue staining (bottom), respectively. All blots are representative of 3 independent experiments.

Figure 5

Phosphorylation of GRAF3 at Y376 increases GAP activity in vitro and decreases RhoA activity in SMC. TR-FRET GDP assay was performed using (A) 10 nM, (B) 30 nM and (C) 100 nM purified WT or Y376E GRAF3 BAR-PH-GAP variant. A decrease in signal indicates an increase in GTP to GDP conversion; n = 3, * p < 0.05 between WT GRAF3 and ^Y376EGRAF3 at indicated timepoints, as assessed by unpaired t-test. Although not depicted, * p < 0.05 at all points except zero for both WT GRAF3 or ^Y376EGRAF3 compared to either GTP or RhoA-GTP. GTP vs. RhoA GTP is not significant. One-phase decay linear regression revealed significance (‡ p < 0.05) between the slopes of WT GRAF3 and ^Y376EGRAF3 at 100nM. (D) Schematic of GRAF3 BRET activation probe NL, NanoLuc. HT, HaloTag (E) GRAF3 WT vs. GRAF3 Y376E BRET activity in SMC. n = 6. mBU, miliBRET units. (F) Model of GRAF3 activation by Src or FAK phosphorylation of GRAF3 at Y376.