Uncovering G protein-coupled receptor kinase-5 as a histone deacetylase kinase in the nucleus of cardiomyocytes

Abstract

G protein-coupled receptor (GPCR) kinases (GRKs) are critical regulators of cellular signaling and function. In cardiomyocytes, GRK2 and GRK5 are two GRKs important for myocardial regulation, and both have been shown to be up-regulated in the dysfunctional heart. We report that increased levels and activity of GRK5 in failing myocardium may have unique significance due to its nuclear localization, a property not shared by GRK2. We find that transgenic mice with elevated cardiac GRK5 levels have exaggerated hypertrophy and early heart failure compared with control mice after pressure overload. This pathology is not present in cardiac GRK2-overexpressing mice or in mice with overexpression of a mutant GRK5 that is excluded from the nucleus. Nuclear accumulation of GRK5 is enhanced in myocytes after aortic banding in vivo and in vitro in myocytes after increased G alpha q activity, the trigger for pressure-overload hypertrophy. GRK5 enhances activation of MEF2 in concert with Gq signals, demonstrating that nuclear localized GRK5 regulates gene transcription via a pathway critically linked to myocardial hypertrophy. Mechanistically, we show that this is due to GRK5 acting, in a non-GPCR manner, as a class II histone deacetylase (HDAC) kinase because it can associate with and phosphorylate the myocyte enhancer factor-2 repressor, HDAC5. Moreover, significant HDAC activity can be found with GRK5 in the heart. Our data show that GRK5 is a nuclear HDAC kinase that plays a key role in maladaptive cardiac hypertrophy apparently independent of any action directly on GPCRs.

Fig. 1.

GRK5 potentiates pressure-overload cardiac hypertrophy in vivo. (A) Cardiac hypertrophy as determined by HW/BW ratios 4 weeks post-TAC in NLC mice (sham and TAC, n = 8 each) and TgGRK5 mice (sham and TAC, n = 9 each). (B) Ejection fraction 4 weeks post-TAC in these mice. (C and D) Real-time quantitative RT-PCR data from mRNA from the left ventricles of NLC (n = 4) and TgGRK5 mice (n = 5) 2 weeks post-TAC. Atrial natriuretic factor (ANF) (C) and β-myosin heavy chain (β-MHC) (D) are normalized to 28S RNA. (E and F) Myocyte size is increased in TgGRK5 mice post-TAC compared with post-TAC NLC control mice. All values are calculated from the average of 50 cells from n = 3 animals. All data presented above are the mean ± SEM. *, P < 0.05 versus NLC sham; ‡, P < 0.05 versus GRK5 sham; #, P < 0.05 versus NLC post-TAC (one-way ANOVA, Bonferroni's multiple comparison test).

Fig. 2.

Nuclear deficient GRK5 transgenic mice display normal pathology after TAC. (A) Western blot for GRK5 from whole heart lysates isolated from TgGRK5 and TgGRK5ΔNLS mice. (B) Nuclear and nonnuclear fractions were prepared from whole hearts isolated from TgGRK5, NLC, and TgGRK5ΔNLS mice, and GRK5 was visualized by Western blot analysis. Endogenous GRK5 in NLC hearts was observed on longer exposure (data not shown). (C and D) TgGRK5ΔNLS mice (sham, n = 14; TAC, n = 5) or NLC mice (sham, n = 19; TAC, n = 7) were subjected to pressure overload. Heart weight-to-body weight ratio (HW/BW) (C) and left ventricular function (ejection fraction) (D). *, P < 0.05 versus NLC sham; ‡, P < 0.05 versus ΔNLS sham (one-way ANOVA, Bonferroni's multiple comparison test); †, P < 0.05 versus sham (two-tailed t test).

Fig. 3.

GRK5 accumulates in the nucleus of myocytes after hypertrophic stimuli and enhances Gq-mediated cellular growth. (A) Adult cardiomyocytes were isolated, and immunofluorescent images are shown on representative myocytes (n = 4 preparations each) from TgGRK5 and NLC hearts. The GRK5 is shown in green and is observed at the plasma membrane and in the nucleus. Propidium iodide was used as a nuclear control (shown in red). (B and C) Nuclear and nonnuclear fractions (data not shown) were prepared from adult myocytes, and GRK5 localization was visualized by Western blot analysis, by using GAPDH and histone as loading controls for nonnuclear and nuclear fractions, respectively. (D and E) Western blot for GRK5, GAPDH, and histone of nonnuclear and nuclear fractions from NRVMs infected with Ad-GRK5 along with Ad-LacZ or Ad-CAM-Gαq. (F) [³H]Phenylalanine incorporation was measured in NRVMs from either Ad-LacZ- or Ad-GRK5-treated cells. All data presented above are the mean ± SEM. *, P < 0.05 versus LacZ veh; ‡, P < 0.05 versus GRK5 veh; #, P < 0.05 versus NLC plus CAM-Gq (one-way ANOVA, Bonferroni's multiple comparison test); †, P < 0.05 versus sham (two-tailed t test) (n = 3 separate experiments).

Fig. 4.

Gq-mediated MEF2 activity depends on nuclear GRK5 activity. (A) NRVMs were infected with a MEF2-luciferase adenovirus and Ad-GRK5 with either Ad-LacZ (control) or Ad-CAM-Gαq. MEF2 activity is normalized to total cellular protein (n = 4 separate experiments). *, P < 0.05 versus LacZ veh; ‡, P < 0.05 versus GRK5 veh; #, P < 0.05 versus NLC + CAM-Gq (one-way ANOVA, Bonferroni's multiple comparison test). (B) HeLa cells were transfected with pcDNA3.1 (control) and either pcDNA3.1 containing GRK5 or GRK5-K215R (catalytically inactive GRK5) and then infected with Ad-MEF2-luciferase and Ad-CAM-Gαq (n = 4 separate experiments). *, P < 0.05 versus pcDNA; ‡, P < 0.05 versus GRK5-K215R (one-way ANOVA, Bonferroni MCT). (C) GRK5 nuclear localization is needed for increased MEF2 activity. NRVMs were infected with Ad-LacZ, Ad-GRK5, or Ad-GRK5ΔNLS (nonnuclear GRK5) and then infected with Ad-MEF2-luciferase and Ad-CAM-Gαq. (n = 5 separate experiments). *, P < 0.05 versus LacZ; ‡, P < 0.05 versus GRK5ΔNLS (one-way ANOVA, Bonferroni MCT). (D) Gq-mediated MEF2 activity in cardiac fibroblasts isolated from WT mice and GRK5 KO mice. †P < 0.05 Gq-mediated response over basal versus WT cells (Student's t test, n = 5).

Fig. 5.

GRK5 interacts with and phosphorylates HDAC5. (A and B) NRVMs were infected with adenovirus containing GFP-HDAC5 and GRK5. GFP-HDAC5 was then immunoprecipitated from either whole-cell lysates (A) or the nuclear fraction (B) and probed for either GFP or GRK5. Cells infected with GRK5 only were used as a negative control. Shown are representative blots from three individual experiments. (C) GRK5 but not GRK2 acts as an HDAC kinase on HDAC5. Twenty picomoles of purified GRK5 and GRK2 was mixed with purified GST-HDAC5 (1 μM). Phosphorylation was detected by autoradiography and quantified with ImageQuant (n = 3 separate experiments, †P < 0.05 two-tailed t test). The GRK2 and GRK5 used in these assays were also added to rhodopsin-enriched rod outer segment membranes, and both GRKs were found to be equally active against this GPCR substrate. (D) GRK5 is able to phosphorylate GST-HDAC5 in a dose-dependent manner. Increasing picomoles (0.1–20 ρmol) of purified GRK5 resulted in increased HDAC5 phosphorylation. (E) Cos-7 cells transfected with plasmids containing GFP-HDAC5 and either (Upper Left) empty vector (EV) (Upper Right) nuclear-excluded GRK5 and GRK5ΔNLS (in red), or (Lower Left) nuclear-trapped GRK5 and GRK5ΔNES (in red). Only cells transfected with the nuclear form of GRK5 (GRK5ΔNES) resulted in specific HDAC5 export. (Lower Right) When serines 259 and 498 are mutated to alanines, this nonphosphorylated form of HDAC5 is restricted to the nucleus with coexpression of GRK5ΔNES. (F) Endogenous GRK5 is needed for maximal Gq-mediated phosphorylation of HDAC5. Cardiac fibroblasts from either WT or GRK5 KO cells were infected with GFP-HDAC5 and either LacZ (used as basal levels) or CAM-Gαq. After immunoprecipitation of GFP-HDAC5, the phosphorylation status of HDAC5 was detected by using a S498 phosphospecific antibody and normalized to total HDAC5 levels (P < 0.05 (t test), n = three independent experiments). IB; immunoblot IP; immunoprecipitation.

Fig. 6.

GRK5 interacts with HDACs in vivo. GRK5 was immunoprecipitated from TgGRK5 or NLC hearts under both sham and 2 weeks post-TAC. After IP, HDAC activity was measured by using the Caymen HDAC activity assay (see Methods). (n = 3–5 mice per group). *, P < 0.05 versus NLC sham; ‡, P < 0.05 versus GRK5 sham, #, P < 0.05 versus NLC post-TAC (one-way ANOVA, Bonferroni's multiple comparison test). IB, immunoblot.