A dynamic pathway for calcium-independent activation of CaMKII by methionine oxidation

Abstract

Calcium/calmodulin (Ca2+/CaM)-dependent protein kinase II (CaMKII) couples increases in cellular Ca2+ to fundamental responses in excitable cells. CaMKII was identified over 20 years ago by activation dependence on Ca2+/CaM, but recent evidence shows that CaMKII activity is also enhanced by pro-oxidant conditions. Here we show that oxidation of paired regulatory domain methionine residues sustains CaMKII activity in the absence of Ca2+/CaM. CaMKII is activated by angiotensin II (AngII)-induced oxidation, leading to apoptosis in cardiomyocytes both in vitro and in vivo. CaMKII oxidation is reversed by methionine sulfoxide reductase A (MsrA), and MsrA-/- mice show exaggerated CaMKII oxidation and myocardial apoptosis, impaired cardiac function, and increased mortality after myocardial infarction. Our data demonstrate a dynamic mechanism for CaMKII activation by oxidation and highlight the critical importance of oxidation-dependent CaMKII activation to AngII and ischemic myocardial apoptosis.

Figure 1. CaMKII is activated by ROS

(A) General structure of a subunit from the multimeric holoenzyme CaMKII and mechanism of CaMKII activation by autophosphorylation. The amino acid sequence of the regulatory domain is highlighted to show the autoinhibitory (AI) and calmodulin-binding (CaM-B) regions. Yellow symbols represent CaM. Pretreatment with Ca²⁺/CaM (1°) followed by phosphorylation at T287 (2°) yields persistent activity even after the removal of Ca²⁺/CaM (3°). (B) Kinase assays were performed after three distinct treatment steps: (1°) ± Ca²⁺/CaM, (2°) ± H₂O₂ or ATP, and (3°) ± EGTA. (n = 6 assays/group, * p < 0.05 vs. WT no treatment). (C) CaMKII is activated by H₂O₂ in a dose-dependent manner after pre-treatment with Ca²⁺/CaM. Oxidation-dependent CaMKII activity is ablated in M281/282V mutants (n = 6 assays/group, * p < 0.05 vs. WT no treatment). (D) M281/282V mutants have normal Ca²⁺/CaM-dependent and T287-autophosphorylation-dependent activation (n = 6 assays/group, * p < 0.05 vs. WT no treatment). (E) Proposed mechanism for activation of CaMKII by oxidation. After initial activation of the holoenzyme by Ca²⁺/CaM (1°), oxidation at M281/282 (2°) blocks reassociation of the catalytic domain, yielding persistent CaMKII activity (3°).

Figure 2. AngII induces oxidation of CaMKII in vivo

(A) Immunoblot of WT CaMKII and M281/282V mutant after no treatment, oxidation, or autophosphorylation probed with antibodies against total, autophosphorylated (p-T287), or oxidized CaMKII. Summary data shows relative band intensity using the oxidized CaMKII antibody (n = 3 trials/group, * p < 0.05 vs. band intensity of WT CaMKII treated with H₂O₂). (B) Immunoblot and summary data of oxidized WT CaMKII probed with antiserum against oxidized M281/282 with increasing ratios of oxidized antigen peptide. (n = 3 trials/group, * p < 0.05 vs. band intensity with no peptide). (C) Immunofluorescent staining of heart sections from mice treated with saline, AngII, or Iso and probed for oxidized or total CaMKII. Red staining is positive for oxidized or total CaMKII and blue staining is for nuclei. Calibration bars are 100 microns. (D) Immunoblot and summary data of heart lysates from mice treated with saline (Sal), Iso, or AngII probed with antibodies against total CaMKII, oxidized CaMKII, or actin (n = 3 hearts/group, * p < 0.05 vs. band intensity of saline treatment).

Figure 3. AngII increases ROS production and apoptosis by a CaMKII-dependent pathway in cardiomyocytes

(A) Percent of total isolated cardiomyocytes positive for TUNEL staining after treatment with saline, AngII, Iso, or H₂O₂ (n = 6 hearts/group, * p < 0.05 vs. WT with saline). (B) Caspase-3 activity induced by saline, AngII, or Iso normalized to WT cells treated with saline (n = 3 hearts/group, * p < 0.05 vs. WT with saline). (C) DHE stained cardiomyocytes after treatment with 100nM AngII or Iso. Red coloration indicates presence of ROS above control cells. Scale bars equal 50 μm. (D) Percent of total cells positive for DHE staining above control (n = 3 assays/group, * p < 0.05 vs. WT saline). (E) Example traces of intracellular calcium concentration of cultured WT cardiomyocytes treated with 100nM AngII (red symbols) or Iso (blue symbols) measured by real-time calcium imaging. The arrow indicates addition of AngII or Iso. (F) Peak intracellular Ca²⁺ concentration in response to either AngII or Iso for WT or p47^−/− cells (n = 3 trials/group, NS = not statistically different).

Figure 4. AngII-induced apoptosis is blocked by CaMKII silencing

(A) Representative immunoblot with anti-CaMKII to measure protein expression after treatment with shRNA and shRNA-resistant rescue constructs. Immunoblot against actin was used as a loading control (not shown). Middle panel shows summary data of CaMKII expression relative to untreated cells (n = 3 experiments/group, * p < 0.05 vs. no treatment). Bottom panel shows summary data for CaMKII activity assays of lysates (n = 3 experiments/group, * p < 0.05 vs. total activity with no treatment, † p < 0.05 vs. ROS-dependent activity with no treatment). Only the WT CaMKII construct was able to reconstitute both Ca²⁺/CaM- and ROS-dependent activity observed in untreated cells. (B) Immunostaining and (C) summary data from isolated rat cardiomyocytes transduced with shRNA against CaMKII followed by rescue with WT CaMKII, M281/282V, or GFP control. Immunostaining shows total nuclei (DAPI) and DNA nicking (TUNEL) consistent with apoptosis. Scale bars equal 100 μm. Summary data show percent of total nuclei with positive TUNEL staining (n = 6 hearts/group, * p < 0.05 vs. GFP with AngII).

Figure 5. AngII causes cardiac apoptosis in vivo via a ROS and CaMKII-mediated pathway

(A) Immunostaining of mouse heart sections for total nuclei (DAPI) and nuclear damage (TUNEL) consistent with apoptosis. WT, p47^−/−, and AC3-I mice were treated with Ang II (3mg/kg/day) or Iso (30mg/kg/day) for seven days. Scale bars equal 100 μm. (B) Percent of total nuclei that showed positive TUNEL staining (n = 3 hearts/group, * p < 0.05 vs. WT with saline).

Figure 6. MsrA^−/− mice have increased susceptibility to AngII-mediated apoptosis

(A) Immunofluorescent staining of heart sections from WT and MsrA^−/− mice treated with AngII and probed for oxidized or total CaMKII. Red staining is positive for oxidized or total CaMKII and blue staining is for nuclei. Calibration bars are 100 microns. (B) Quantification of average staining intensity for AngII treated hearts, relative to WT (n = 3 hearts/group, * p < 0.05 vs. WT with AngII). (C) Summary data for TUNEL staining of heart sections from WT and MsrA^−/− mice treated with saline or AngII (n = 5 hearts/group, * p < 0.05 vs. WT with saline).

Figure 7. Mice lacking MsrA have increased CaMKII oxidation, apoptosis, reduced survival and impaired heart function after myocardial infarction

(A) Immunostaining and (B) stain intensity quantification of oxidized CaMKII in heart sections from WT, p47^−/−, and MsrA^−/− mice post-MI (n = 3 hearts/group, * p < 0.05 vs. WT). (C) Summary data for TUNEL staining of heart sections from WT, p47^−/−, and MsrA^−/− mice post-MI (n = 3 hearts/group, * p < 0.05 vs. WT). (D) Mortality is significantly increased post-MI in MsrA^−/− mice compared to WT controls. Numbers in bars represent post-MI deaths/total number of mice receiving MI. Post-MI left ventricular dilation (E) and function (F) were compromised in surviving MsrA^−/− mice compared to WT controls three weeks after surgery (n = 17 hearts/group for WT, n = 9 hearts/group for MsrA^−/−).