Nuclear translocation of cardiac G protein-Coupled Receptor kinase 5 downstream of select Gq-activating hypertrophic ligands is a calmodulin-dependent process

Abstract

G protein-Coupled Receptors (GPCRs) kinases (GRKs) play a crucial role in regulating cardiac hypertrophy. Recent data from our lab has shown that, following ventricular pressure overload, GRK5, a primary cardiac GRK, facilitates maladaptive myocyte growth via novel nuclear localization. In the nucleus, GRK5's newly discovered kinase activity on histone deacetylase 5 induces hypertrophic gene transcription. The mechanisms governing the nuclear targeting of GRK5 are unknown. We report here that GRK5 nuclear accumulation is dependent on Ca(2+)/calmodulin (CaM) binding to a specific site within the amino terminus of GRK5 and this interaction occurs after selective activation of hypertrophic Gq-coupled receptors. Stimulation of myocytes with phenylephrine or angiotensinII causes GRK5 to leave the sarcolemmal membrane and accumulate in the nucleus, while the endothelin-1 does not cause nuclear GRK5 localization. A mutation within the amino-terminus of GRK5 negating CaM binding attenuates GRK5 movement from the sarcolemma to the nucleus and, importantly, overexpression of this mutant does not facilitate cardiac hypertrophy and related gene transcription in vitro and in vivo. Our data reveal that CaM binding to GRK5 is a physiologically relevant event that is absolutely required for nuclear GRK5 localization downstream of hypertrophic stimuli, thus facilitating GRK5-dependent regulation of maladaptive hypertrophy.

Figure 1. PE and AngII induce translocation of GRK5 from the membrane to the nucleus.

(A) Representative immunofluorescence staining of endogenous GRK5 in AdRbM shows increased nuclear GRK5 following PE (50 µM) and AngII (10 µM) treatment, but not ET-1 (100 nM). (B) NRVM were infected with Ad-GRK5 (50 MOI). After 48 hr, cells were treated with 50 µM PE for 5 different time points, harvested by subcellular fractionation. Nuclear were fractions immunoblotted for GRK5 and fibrillarin. (C) The amount of GRK5 in the nucleus was calculated by denistometry, normalized to fibrillarin, and reported as Fold Change over baseline. *p<0.05, one-way ANOVA with a Bonferroni correction, n = 4. (D) Rabbit myocytes were infected with an adenovirus expressing GRK5-GFP and cultured overnight. Using TIRFM cells were imaged at 10 sec intervals for 700 sec. Baseline myocytes were untreated while stimulated myocytes were treated with either PE (50 µM), AngII (10 µM), or Et-1 (100 nM) at 120 sec. Fluorescence was normalized and reported as fold change versus baseline. n = 4. (E) Representative TIRF images for each agonist at the beginning and end of imaging.

Figure 2. Mice with cardiac-overexpression of GRK5 (Tg-GRK5) show increased nuclear accumulation of GRK5 following 3 days of continuous infusion of a subpressor dose of PE or AngII.

(A) Osmotic minipumps containing either a subpressor dose of PE (30 µM/kg/day) or phospho-buffered saline (PBS) were implanted subcutaneously in Tg-GRK5 mice. After 72 hr, hearts were isolated and subjected to subcellular fractionation and immunoblotted for GRK5 and fibrillarin. (B) The amount of GRK5 in the nucleus was calculated by denistometry, normalized to fibrillarin, and reported as the fold change increase with PE. *p<0.001 v. PBS treated, student’s t-test, n = 8. (C) Osmotic minipumps containing either a subpressor dose of AngII (200 nM/kg/min) or PBS were implanted subcutaneously in Tg-GRK5 mice. After 72 hr, hearts were isolated and subjected to subcellular fractionation and immunoblotted for GRK5 and fibrillarin. (D) The amount of GRK5 in the nucleus was calculated by denistometry, normalized to fibrillarin, and reported as the fold change increase due to AngII. *p<0.01 v. PBS treated, student’s t test, n = 9. (E) HW/BW ratio following 3 days of continuous PBS or AngII infusion in NLC and Tg-GRK5. *p<0.01 v. Tg PBS and NLC AngII, one-way ANOVA with a Bonferroni correction, n = 5–9 (F) Systolic LV Posterior Wall thickness (LVPWT) measured in mm by echocardiogram following 3 days of continuous PBS or AngII infusion in NLC or Tg-GRK5 mice. *p<0.01 v. Tg PBS and NLC AngII, one-way ANOVA with a Bonferroni correction, n = 5–9.

Figure 3. GRK5 nuclear accumulation is diminished after treatment with a CaM inhibitor.

(A) NRVM were infected with Ad-GRK5 and either Ad-LacZ or Ad-Gq-CAM. 48 hr after infection, cells were treated with DMSO or inhibitor: BIM1 (10 µM), Gö6976 (10 µM), CDZ (10 µM) and KN93 (10 µM) for 1 hr. The cells were harvested using subcellular fractionation and immunoblotted for GRK5. (B) Immunoblots were quantitated by densitometry, normalized to fibrillarin, and reported as fold change over baseline. * p<0.05 v. untreated baseline, # p<0.05 v. CDZ, one-way ANOVA with a Bonferroni correction, n = 4. (C) NRVM were infected with Ad-LacZ, Ad-GRK5 and Ad-Gq-CAM. 48 hr after infection, cells were treated with DMSO or CDZ (10 µM) for 1 hr. The cells were harvested using subcellular fractionation, and immunoblotted for GRK5. (D) Densitometric analysis for (C) with GRK5 normalized to fibrillarin and calculated as fold change over baseline. *p<0.01 v. DMSO GRK5, #p<0.01 v. DMSO GRK5+ Gq, one-way ANOVA with a Bonferroni correction, n = 4. (E) NRVM were infected with either Ad-LacZ or Ad-Gq-CAM. 48 hr after infection, cells were treated with DMSO or CDZ (10 µM). Immunofluorescence was detected using a polyclonal GRK5 antibody. (F) TIRF analysis of AdRbM infected with an adenovirus expressing GRK5-GFP and cultured overnight. Cells were imaged at 10 sec intervals for 700 sec. Cells were pre-treated with CDZ or DMSO for 30 min at 37°C prior to imaging. Baseline myocytes were untreated while stimulated myocytes were treated with PE (50 µM) at 120 sec. Fluorescence was normalized and reported to fold change versus baseline. n = 4. (G) Same experimental design as (F) except cells were stimulated with AngII (10 µM) at 120 s. n = 4.

Figure 4. A mutant GRK5 (W30AK31Q) unable to bind CaM at its N-terminal CaM binding site displays less nuclear accumulation following Gq or PE stimulation.

(A) Cartoon of GRK5’s structure illustrating pertinent domains and regulatory sites. (B) NRVM infected with Ad-GRK5 or Ad-GRK5W30A were stimulated with Ad-Gq-CAM (48 hr) or PE (1hr). Cells were then harvested by subcellular fractionation and immunoblotted for GRK5. (C) Quantitative analysis of (B) normalized to fibrillarin and reported as fold change over baseline. *p<0.001 v. WT GRK5, one-way ANOVA with a Bonferroni correction, n = 4. (D) AdRbM were co-infected with an adenovirus expressing either WT GRK5 tagged with GFP or GRK5 W30A tagged with GFP and Ad-Gq-CAM. Following an overnight culture, cells are imaged by confocal microscopy. Fluorescence within the nucleus was measured and normalized to cytoplasmic fluorescence. *p<0.001 v. WT GRK5+ Gq, one-way ANOVA with a Bonferroni correction, n = 4 (E) Images of representative myocytes showing WT GRK5-GFP (left) and GRK5W30A-GFP (right). (F) MEF2 activity in NRVM was measured using a luciferase assay system. Cells were co-infected with an adenovirus expressing a MEF2-luciferase reporter construct, Ad-LacZ, Ad-GRK5 or Ad-GRK5W30A and stimulated for 48 hr with the Ad-Gq-CAM virus. *p<0.001 v. WT GRK5, one-way ANOVA with a Bonferroni correction, n = 4, done in triplicate. Inset shows whole cell lysate of NRVM used in this experiment.

Figure 5. GRK5W30A displays increased plasma membrane association following agonist treatment and differential ability to desensitize GPCRs compared to WT.

(A) AdRbM were infected with an adenovirus expressing GRK5-GFP or GRK5W30A-GFP and cultured overnight. Using TIRFM cells were imaged at 10 sec intervals for 700 sec. Baseline myocytes were untreated while stimulated myocytes were treated with either AngII (10 µM) (A) or PE (50µM) (B) at 120 sec. Fluorescence was normalized and reported to fold change versus baseline. n = 4 (C) Changes in GRK5 activity at the membrane was measured using an IP₁ ELISA to determine changes in desensitization. NRVM were infected with Ad-LacZ, Ad-GRK5 or Ad-GRK5W30A. After 48 hours, cells were stimulated with PE or ET-1 for 2 hr, then assayed for IP₃ generation via IP₁ ELISA. *p<0.01 v. LacZ PE and WT PE, #p<0.01 v. LacZ ET-1, one-way ANOVA with a Bonferroni correction, n = 3, done in duplicate.

Figure 6. GRK5W30A demonstrates altered nuclear translocation in vivo.

(A) Total cell lysates from GRK5KO injected with Ad-GRK5W30A into their LV free wall taken 10 days post-injection. (B) Nuclear lysates from mice with cardiac expression of only GRK5W30A that had received 72 hr of chronic PBS or AngII infusion were immunoblotted for GRK5. (C) Quantitative analysis of the nuclear lysates for nuclear GRK5 accumulation normalized to fibrillarin and reported as fold change. n = 8. (D) HW/BW ratio following 3 days of continuous PBS or AngII infusion for mice expressing GRK5W30A. (E) Total cell lysates from GRK5KO mice injected with Ad-GRK5 CTPB into their LV free wall taken 10 days post-injection. (F) Nuclear lysates from mice cardiac specific expression of only GRK5 CTPB that had received 72 hr of chronic PBS or AngII infusion were immunoblotted for GRK5. (G) Quantitative analysis of the nuclear lysates for nuclear GRK5 accumulation normalized to fibrillarin and reported as fold change. *p<0.05, student’s t test, n = 6 (H) HW/BW ratio following 3 days of continuous PBS or AngII infusion for mice expressing GRK5CTPB. *p<0.05, student’s t test, n = 6.

Figure 7. Cartoon depicting the select Gq-coupled receptor CaM-mediated translocation of GRK5 into the nucleus of cardiomyocytes.

Gq activation due to catecholamines or AngII binding at the α₁AR or AT-1R, respectively, causes CaM to bind GRK5 at its N-terminus, dislodging GRK5 from the plasma membrane. Via its NLS, GRK5 is directed to the nucleus where its interaction with CaM is stabilized by IP₃R-regulated Ca²⁺ release. Once in the nucleus, GRK5 can act as an HDAC5 kinase, relieving repression of MEF2 and inducing hypertrophic gene transcription. In contrast, endothelin-1 binding leads to a selective interaction between the ET-1R substrate and the desensitizing GRK5. CaM cannot bind the kinase in this state, thus keeping GRK5 at the plasma membrane.