Myosin Light Chain Phosphorylation is Critical for Adaptation to Cardiac Stress

Abstract

Background

Cardiac hypertrophy is a common response to circulatory or neurohumoral stressors as a mechanism to augment contractility. When the heart is under sustained stress, the hypertrophic response can evolve into decompensated heart failure, although the mechanism(s) underlying this transition remain largely unknown. Because phosphorylation of cardiac myosin light chain 2 (MLC2v), bound to myosin at the head-rod junction, facilitates actin-myosin interactions and enhances contractility, we hypothesized that phosphorylation of MLC2v plays a role in adaptation of the heart to stress. We previously identified an enzyme that predominantly phosphorylates MLC2v in cardiomyocytes, cardiac-MLCK (cMLCK); yet the role(s) played by cMLCK in regulating cardiac function in health and disease remain to be determined.

Methods and Results

We found that pressure-overload induced by transaortic constriction in wildtype mice reduced phosphorylated-MLC2v levels by ~40% and cMLCK levels by ~85%. To examine how a reduction in cMLCK and the corresponding reduction in pMLC2v affect function, we generated Mylk3 gene-targeted mice as well as transgenic mice overexpressing cMLCK specifically in cardiomyocytes. Pressure-overload led to severe heart failure in cMLCK knockout mice, but not in mice with cMLCK overexpression in which cMLCK protein synthesis exceeded degradation. The reduction in cMLCK protein during pressure-overload was attenuated by inhibition of ubiquitin-proteasome protein degradation systems.

Conclusions

Our results suggest the novel idea that accelerated cMLCK-protein turnover by the ubiquitin-proteasome system underlie the transition from compensated hypertrophy to decompensated heart failure due to reduced phosphorylation of MLC2v.

Introduction

In 2006, the overall death rate from cardiovascular disease was 262.5 per 100,000, of which 1 in 8.6 deaths was due to heart failure. During failure, the heart is unable to pump sufficient blood to meet circulatory demand due to reduced myocardial force development ¹. Pressure development and ejection are achieved by shortening of the minor- and long-axes of the heart, as well as twisting of the apex (torsion) due to differential orientation of ventricular myofibers in roughly three layers: inner-, mid- and outer ^2–5. The dynamics of cardiac contraction and relaxation are fundamentally related to actin-myosin interactions which are initiated by Ca²⁺ binding to troponin C and relief of inhibition of crossbridge binding to actin. A range of factors regulate this process and contribute to the adaptation of cardiac function to physiological or pathological stressors ^6–8, in turn suggesting potential therapeutic strategies for heart failure ^9–11.

Myosin is a hexamer comprised of two heavy chains and two pairs of light chains (MLC1 and MLC2) bound to the rod and neck regions of the heavy chain, respectively. Phosphorylation of MLC2v has been shown to potentiate the rate and force of cardiac contraction ^12–16. Using mutant MLC2v knock-in mice in which two phosphorylatable serine residues were mutated to alanine (Ser14/15Ala), our recent study with others demonstrated that phosphorylation increases myosin binding and myosin lever arm stiffness, and alters the kinetics of crossbridge cycling to increase crossbridge duty ratio (stroke time/cycle time). Furthermore, in addition to the traditional view of Ca²⁺-dependent activation of thin filaments, phosphorylated MLC2v can facilitate this process by activating neighboring binding sites on actin ¹⁷. Thus, nonphosphorylatable MLC2v in mice reduced contractility, and reduced phosphorylation of MLC2v has been implicated in human heart disease where phosphorylation was reduced to ~18% of total MLC2v in failing hearts from ~30–40% in healthy hearts ^{18, 19}.

The predominant kinase for MLC2v, cardiac-specific MLCK (cMLCK) encoded by Mylk3 , has been identified as a target of transcription factor Nkx2-5 ²⁰, and independently as a gene product that is differentially expressed in failing human hearts ²¹. Knockdown of cMLCK in neonatal cardiomyocytes and in zebrafish embryos resulted in abnormal formation of the sarcomere and depressed contraction ^{20, 21}. A recent study using mice expressing hypomorphic cMLCK confirmed that cMLCK is the predominant MLC2v kinase in the heart and is important for normal cardiac contraction in vivo ²².

Under physiological conditions, the level of MLC2v phosphorylation is relatively constant due to countervailing actions of cMLCK and protein phosphatase 1β (PP1β, also termed PP1δ) in association with the myosin-binding phosphatase-target subunit (MYPT)^{23, 24}. The level of MLC2v phosphorylation has been shown to vary through the thickness of the murine ventricular wall, being relatively low in the inner layer and high in the outer layer ¹³. This gradient has been proposed to be critical for non-uniform contractile properties of the myocytes across the wall, in particular for left ventricular (LV) torsion, the counterclockwise twisting of the LV apex, and clockwise twisting of the base during systole, when viewed from the apex ^3–5. It has been thought that the gradient is formed by reduction of phosphatase activity in the outer layer ^{24, 25}. However, the existence of a MLC2v phosphorylation gradient is still controversial ^{13, 22}, and if it is present, the participation of MLC2v kinase as a determinant of the gradient remains to be clarified.

Based on these observations, we sought to investigate the role of cMLCK in the heart by determining its transmural expression and its functional roles through the use of loss- or gain-of-function mutations in mice.

Methods

Mouse models

A conditional null allele of Mylk3 was generated by introducing loxP sites spanning exon 5, which was done through homologous recombination in ES cells. Mice heterozygous for this Mylk3 allele were bred to mice expressing the ACTB1-Cre transgene, resulting in a germline Mylk3^+/− allele, followed by cross-breeding to generate Mylk3^−/− mice having a mixed genetic background mainly with 129/Sv and C57BL/6. Transgenic mice were generated by injection of HA-tagged full-length cMLCK cloned into an α-MHC promoter plasmid (kindly provided by J. Robbins)²⁶. All animal experiments were performed with approval from the University of Florida Institutional Animal Care and Use Committee.

Human heart samples

Human heart samples were obtained from the National Human Tissue Resource Center. All protocols were approved by the University of Florida Institutional Review Board.

Additional experimental procedures are described in the supplement.

Results

Regional expression of cMLCK protein and phosphorylation of MLC2v in mouse and human hearts

We recently identified an enzyme that predominantly phosphorylates MLC2v in cardiomyocytes, cardiac-MLCK (cMLCK) ²⁰. First, we confirmed that the regional expression of cMLCK and the extent of MLC2v phosphorylation (pMLCv) are almost identical in normal mouse hearts, although labeling of each exhibits non-uniform intensity across transverse tissue sections (Figures 1A, S1A). The staining of both was below the level of detection in the absence of cMLCK in Mylk3^−/− mice (described later). Specificity of the pMLC2v antibody against phosphorylated MLC2v was confirmed by Western blotting (Figure S2).

Figure 1.

cMLCK protein and phosphorylated MLC2v expression.(A) Immunostaining of cMLCK and pMLC2v (purple) in transverse sections of Mylk3^+/+ and Mylk3^−/− adult hearts counterstained with 10-fold diluted eosin (pink). Bar = 1 mm. (B) Enlarged images of immunostaining. Bars = 10 µm. (C–L) Distribution of cMLCK (top panels) and pMLC2v (bottom panels) in serial sections from LV and RV of Mylk3^+/+ . Bars = 500 µm (C–F), 50 µm (G–L). (M) Fluorescent immunostaining of serial tissue-sections of Mylk3^+/+ heart including Purkinje fibers using antibodies against cMLCK, pMLC2v and connexin40 (green, all rabbit polyclonal antibodies) with nuclear staining (blue). Bars = 50 µm. (N) Immunostaining of serial tissue-sections of human heart including Purkinje fibers locating in the endocardial layer using antibodies against human cMLCK, pMLC2v, and ANF (all rabbit polyclonal antibodies). Bars = 100 µm.

At higher magnification, cMLCK staining was more diffuse in the cytoplasm compared to the striated staining pattern of pMLC2v (Figure 1B). Globally, expression of cMLCK and pMLC2v were higher in the right ventricle (RV) than in the LV (Figures 1C, D). In the LV, higher levels of expression were observed in the mid to outer epicardial layers than in the inner endocardial layer at interventricular septal wall (Figures 1C–H). A few layers of myofibers at the surface of the endocardium highly expressed cMLCK and pMLC2v (Figures 1I–L). These cells are Purkinje fibers, which form a ventricular conduction system marked by the presence of the gap junction protein connexin40 (Figure 1M). In human hearts, levels of cMLCK and pMLC2v were also high in Purkinje fibers which were co-stained for atrial natriuretic factor (ANF)(Figure 1N). Notably, the antibody against connexin40 was not compatible with the paraffin-sections of the human heart.

These data confirm that cMLCK is the predominant kinase for MLC2v, which is highly expressed in the mid- and epicardial layers compared to the endocardial layer of the LV consistent with a previous report ¹³, and has even greater expression in the RV and Purkinje fibers.

Reductions in cMLCK and pMLC2v after pressure-overload

The transmural expression of cMLCK proteins and pMLC2v was quantified by Western blotting of tissue from hearts subjected to pressure-overload by transverse aortic constriction (TAC) for 3 months. The endocardial layers significantly thicken following TAC, making them easier to dissect (Figure 2A). cMLCK expression and MLC2v phosphorylation were higher in the RV than the LV and were below the level of detection in the endocardial layers (Figures 2B, C, S2B). Our previous study showed that ~30% of total MLC2v was phosphorylated in normal mouse hearts ²⁰. In the remaining experiments described (unless otherwise specified), cMLCK and pMLC2v expression were examined in the apical third of the heart to minimize variations due to regional expression of cMLCK. Reductions in cMLCK and pMLC2 were evident as early as 1 week after TAC, by ~85% and ~40%, respectively, compared to control mice (Figures 2D, E).

Figure 2.

Reduction of cMLCK in mouse hearts due to pressure overload. (A) Immunostaining of cMLCK and pMLC2v (purple) in transverse serial sections of TAC heart counterstained with eosin (pink). (B) Western blotting of cMLCK, total MLC2v and GAPDH in RV, the entire LV and the endocardial layer (endo) dissected from a TAC heart. (C) Unphosphorylated and phosphorylated MLC2v in RV, LV and papillary muscles dissected from TAC hearts examined with 2D-electrophoresis followed by Western blotting with an anti-MLC antibody. (D) Western blotting of cMLCK, pMLC2v, total MLC2v and GAPDH after 1 week of TAC (lanes 6–9) and control (lanes 1–5). (E) Fold difference of cMLCK, pMLC2v and total MLC2v normalized to GAPDH with the value without TAC defined as 1. (F) Representative images of sequential M-mode ultrasound of a Mylk3^+/+ mouse from pre-operation to 4 weeks of TAC. (G) Echocardiographic indices of Mylk3^+/+ mice before and during TAC. Time-dependent effects were not significant in ED dimension and HR using repeated measure ANOVA. *P < 0.05.

One week of TAC resulted in ventricular wall thickening without apparent changes in cardiac function (%FS) or chamber diameters (Figures 2F, G, 1 week). Two weeks of TAC significantly reduced %FS and increased systolic dimensions compared to control values or 1 week of TAC in our experimental conditions (Figures 2F, G). Thus, the functional transition from compensated to decompensated cardiac hypertrophy occurred between 1 and 2 weeks of TAC, shortly after reductions in cMLCK and pMLCv.

Deletion of Mylk3 leads to moderate abnormalities in contraction and conduction under basal conditions

Data from mice with TAC suggest that reduced cMLCK expression and MLC2v phosphorylation may be involved in both chronic and transitional (from compensated to decompensated) cardiac hypertrophy. To examine this possibility mechanistically, we first generated Mylk3 gene-targeted mice and examined whether the absence of cMLCK affects cardiac function. Germline-deletion of floxed-exon 5 resulted in elimination of the first coding exon of the catalytic domain and a frame-shift of the subsequent downstream exons (Figure 3A, Figure S3). cMLCK mRNA expression was below the level of detection in Mylk3^−/− hearts by Northern blotting using a cDNA probe that recognizes exons 1 to 6 (1266 bp)(Figure 3B). This result is likely attributable to a nonsense-mediated mRNA decay ²⁷, with targeted cMLCK mRNA containing a premature termination codon. In Mylk3^−/− hearts, MLC2v phosphorylation was below the level of detection (Figure 3C), consistent with immunostaining in Figure 1A.

Figure 3.

Moderate heart enlargement and reduced contractile function in Mylk3^−/− mice. (A) PCR genotyping demonstrates ~1750 bp in Mylk3^+/+ (lane 1) vs. ~800 bp in Mylk3^−/− (lane 2). (B) Northern blotting shows that cMLCK mRNA (4.3 kb) is below detection levels in Mylk3^−/− mice (lane 1 vs. 2). (C) Unphosphorylated and phosphorylated MLC2v examined with 2D-electrophoresis followed by Western blotting (left panel) and relative amounts of phosphorylated to total MLC2v (right panel). (D) Representative hearts dissected from Mylk3^+/+ and Mylk3^−/− mice at 3 months of age. Bars = 2 mm. (E) HW/BW (mg/g) at 3 and 6 months of age. (F) Representative heart sections with Masson’s trichrome staining and area size of fibrosis (% relative to the total area examined). Bars = 1 mm (top), 100 µm (middle and bottom panels). (G) Cardiac contraction, wall volume, wall thickness and torsion examined using MRI at 3 months of age. Calculation of cardiac torsion is shown. See Supplemental material (video). (H) Representative images of cardiomyocytes, cell area (µm²), long axis (µm) and short axis (µm) at 3 months of age. Bars = 100 µm. (I) Measurements of cardiac contraction and simultaneous Ca²⁺ transients in isolated cardiomyocytes. ED, end-diastolic; ES, end-systolic; %FS, % left ventricular fractional shortening. *P < 0.05.

Mylk3^−/− mice were born at the expected Mendelian ratios and survived through adulthood beyond 1.5 years of age. Mylk3^−/− hearts were moderately enlarged predominantly along the long-axis with increased heart weight at 3 and 6 months of age (Figures 3D, E). Neither cytoarchitectural disarray nor increased interstitial fibrosis was observed (Figure 3F). Cardiac contraction assessed using MRI (Figure 3G, supplemental movies) and echocardiography (Figure S4) demonstrated a reduction in ejection fraction and increases in the volume of the left ventricular cavity both at end-systole and end-diastole. Cardiac torsion was also reduced in Mylk3^−/− hearts compared to control mice (Figure 3G).

Isolated ventricular cardiomyocytes from adult Mylk3^−/− mice were larger than cardiomyocytes from control mice, and displayed reductions in contractility and speed of relaxation with no changes in amplitude of the intracellular Ca²⁺ transient (Figures 3H, I).

In the ventricular conduction system, cMLCK and pMLC2v were present at higher levels than in the rest of the LV (Figures 1K–N). Ambulatory telemetry ECG recordings demonstrated a small but significant prolongation of the QRS complex in Mylk3^−/− mice compared to age- and sex-matched Mylk3^+/+ controls (Figures S5A–C), indicating that ventricular depolarization was prolonged in Mylk3^−/− mice. This could be due all or in part to elongation of the ventricular conduction system secondary to enlargement of the heart (Figure S5D).

Adaptation to pressure-overload requires cMLCK

In contrast to the finding that Mylk3^−/− mice showed a moderate reduction in contractility under basal conditions, 3 months TAC resulted in profound heart failure in these mice, associated with increased heart size (Figure 4A), decreased survival rate (Figure 4B), and increased heart weight/tibial length (Figure 4C). Hereafter, heart weight is normalized to tibial length in reporting results of TAC studies, since there is an increase in body weight by accumulation of body fluid due to profound heart failure after TAC, particularly in Mylk3^−/− mice.

Figure 4.

Heart failure in Mylk3^−/− mice after pressure overload.(A) Representative images of hearts after 3 months of TAC. Bar = 2 mm. (B) Survival analysis during 3 months of TAC. Several mice died shortly after the operation (1 week) and were eliminated from these studies in consideration of post-operative complications. (C) HW/tibial length (mg/mm) after 3 months of TAC. (D–E) Representative images of M-mode ultrasound and echocardiographic indices after 3 months of TAC. (F–G) Representative tracing of LV pressure, dP/dt and summarized data from hemodynamic measurements after 3 months of TAC. (H) Sequential ehocardiographic indices from pre-operation to 4 weeks TAC. *P < 0.05.

Mylk3^−/− mice had markedly increased LV dimensions at end-diastole and systole with reduced contractility (Figures 4D–E). LV end-systolic pressure (~100 mmHg without TAC) was increased to ~180 mmHg after TAC in Mylk3^+/+ mice, but not in Mylk3^−/− mice (~120 mmHg, Figures 4F–G). Instead, Mylk3^−/− mice exhibited a marked increase in end-diastolic pressure, and reduced rates of contraction (+dP/dt) and relaxation (−dP/dt). These measurements revealed profound systolic and diastolic dysfunction in Mylk3^−/− mice (Figures 4F, G).

One week of TAC did not affect cardiac contractility or chamber dimensions in Mylk3^+/+ mice. In contrast, Mylk3^−/− mice showed a progressive reduction in %FS and increased end-systolic dimension (Figures 2F–G, 4H, S6). Although cardiac function and heart size differed in Mylk3^−/− vs. Mylk3^+/+ mice prior to TAC, Mylk3^−/− mice displayed no compensatory response to TAC (Figure 4H: pre-TAC %FS was 25% reduced compared with Mylk3^+/+ mice, and fell further 1 week after TAC to ~40%).

Role of cMLCK in response to 4 weeks of swimming exercise

Besides the hypertrophic response to pressure-overload, the heart can also exhibit hypertrophic growth as an adaptation to exercise. This adaptation results in an increase in absolute force production by the ventricular wall due at least in part to increase in the total number of actin-myosin motor units working in parallel ²⁸. To understand the role of cMLCK in the response of the heart to exercise, 3 month-old Mylk3^−/− and Mylk3^+/+ mice were subjected to 4 weeks of swim training (Figure 5A).

Figure 5.

Effects of cMLCK on cardiac adaptations to physiological stress.(A) Representative images of hearts dissected after 4 weeks of swimming exercise. Bars = 2 mm. (B) Survival analysis during 4 weeks of swimming exercise. (C) HW/BW with or without 4 weeks of swimming exercise at ~4 months of age. (D) Echocardiographic indices after swimming exercise. (E) Unphosphorylated and phosphorylated MLC2v examined with 2D-electrophoresis followed by Western blotting and relative amounts of phosphorylated to total MLC2v. (F) Western blotting of cMLCK, pMLC2v, total MLC2v and GAPDH with or without swimming exercise. Fold difference of cMLCK and pMLC2v normalized to total MLC2v with the value in Mylk3^+/+ without swimming exercise defined as 1. (G) Real-time RT PCR demonstrates relative expression of ANF and BNP mRNA normalized to β-actin mRNA with the value in Mylk3^+/+ without swimming exercise defined as 1. *P < 0.05.

Wildtype mice tolerated 4 weeks of swimming, while 25% of Mylk3^−/− mice died during or shortly after exercise (Figure 5B). Both Mylk3^+/+ and surviving Mylk3^−/− mice demonstrated an increased heart weight/body weight (Figure 5C); however, exercise failed to improve cardiac contraction evident as a reduction in %FS in Mylk3^−/− mice compared to control Mylk3^+/+ (Figure 5D) and to sedentary Mylk3^−/− mice at a similar age from 26% to 19% (Figure S4, 3 months of age). In Mylk3^+/+ but not in Mylk3^−/− mice, swimming increased pMLC2v to 39% (33% without swimming) (Figure 5E), and increased expression of cMLCK protein (Figures 5E, F). These results indicate that cMLCK is the predominant kinase mediating increased MLC2v phosphorylation in response to swim training.

Increased expression of fetal genes, such as ANF and brain natriuretic peptide (BNP), is often observed in failing hearts but was not seen in sedentary Mylk3^−/− mice (no swimming), consistent with moderate cardiac dysfunction without an increase in interstitial fibrosis (Figure 3F). Exercise-induced hypertrophy normally does not increase expression of ANF and BNP (Figure 5G, +/+ no swim vs. swim)^{29, 30}. Despite the reduction in %FS in Mylk3^−/− mice after exercise, there was no substantial increase in ANF or BNP before vs. after exercise in Mylk3^−/− mice.

Overexpression of cMLCK protects against pressure-overload induced cardiac dysfunction

Profound heart failure in pressure-overloaded Mylk3^−/− mice supports our hypothesis that reduced cMLCK expression and MLC2v phosphorylation are involved in decompensated TAC-induced cardiac hypertrophy. We further examined whether excess cMLCK protects the heart during short-and long-term pressure-overloading using a second mouse model that over-expresses cMLCK under the control of α-myosin heavy chain (α-MHC) promoter/enhancers. Among five transgenic lines, three lines expressed cMLCK relatively homogenously in the LV (data not shown) and were further examined. The level of transgenic protein expression relative to the endogenous cMLCK in non-TG (NTG) mice ranged from 6.5-fold in line TG1 to 21-fold in TG2 and TG3 (Figure 6A), leading to a moderate increase in pMLC2v from 34% of total MLC2v in NTG to 42–50% in the transgenics (Figure 6B). Basal cardiac function in the three transgenic lines was not statistically different from NTG, with the exception of a slight reduction in end-diastolic dimensions in TG3 (Figure S7A).

Figure 6.

Overexpression of cMLCK attenuates pressure overload-induced pathological hypertrophic responses. (A) Schematic of the transgenic construct. Western blotting demonstrates transgenic protein expression in three lines of transgenics using antibodies against cMLCK (top panel) and HA (lower panel). Relative expression compared to the endogenous cMLCK proteins is shown. (B) Unphosphorylated and phosphorylated MLC2v examined with 2D-electrophoresis followed by Western blotting and relative amounts of phosphorylated to total MLC2v (n=2-4 hearts from each line). (C) Western blotting of heart lysates with or without 1 week of TAC using antibodies against cMLCK, HA, pMLC2v, MLC2v and GAPDH. (D) Summarized data of LV systolic pressure and HW/tibial length with or without 1 week of TAC. Representative M-mode ultrasound images (E) and echocardiographic indices (F) following 3 months of TAC. (G) Masson’s trichrome staining of heart sections and area size of fibrosis (% relative to the total area size examined) following 3 months of TAC. Bar = 50 µm. (H) Representative co-staining for TUNEL, DAPI and troponin T of LV following 3 months of TAC, and summary for TUNEL positive nuclei (% of total nuclei). Bar= 20 µm. Total number of nuclei counted is indicated. MHC, myosin heavy chain; NTG, non-transgenic; TG, transgenic. *P < 0.05.

1 week TAC reduced transgenic proteins expression, similar to endogenous cMLCK proteins (Figure 6C), despite endogenous and transgenic cMLCK having different promoter/enhancers, exon-intron structures, and 3’ poly-A tails. These results suggest that the reduction in cMLCK protein by pressure-overload is regulated by post-transcriptional mechanisms.

While NTG and TGs were subjected to similar degrees of pressure-overload (Figure 6D), the increase in heart weight (determined by heart weight/tibial length) was attenuated in transgenics, including the TG3 line with higher cMLCK expression (Figure 6D). Attenuation was accompanied by a parallel reduction in cardiomyocyte cell size and preservation of contractility at 2 and 4 weeks after TAC (Figures S7B, S7C). After chronic pressure-overload (3 months; degree similar in all groups; Figure S8A), and reductions in both endogenous and transgenic cMLCK proteins (Figure S8B), cardiac function was significantly preserved in the high-expressor TG3 line vs. NTG with respect to contractility and chamber size (Figures 6E, F), interstitial fibrosis (Figure 6G), and apoptosis (Figures 6H, S9).

These results further support our hypothesis that elevated expression of cMLCK protein (6.5 to 21-fold higher than endogenous) protects the heart against cardiac dysfunction during short- and long-term pressure-overload.

Involvement of the ubiquitin-proteasome system in cMLCK protein degradation

To understand potential mechanisms underlying the reduction in cMLCK protein during pressure-overload, we explored the possible role of the ubiquitin-proteasome system, the major non-lysosomal pathway for intracellular protein degradation. Recently, altered activity of the ubiquitin-proteasome system has been emphasized in studies of cardiac hypertrophy ^31–33. Target proteins are polyubiquitinated by concerted action of three enzymes: E1 (ubiquitin-activating enzyme), E2 (ubiquitin-conjugating enzyme) and E3 (ubiquitin-ligase) prior to degradation by the 26S proteasome composed of the 20S catalytic and 19S regulatory cores ^31–33. Both synthetic and natural proteasome inhibitors block proteasome activity: synthetic peptide aldehyde, such as MG132, mimics proteasome substrates, binds to the proteasome active site, and transiently disrupts its activity. A natural inhibitor, lactacystin, covalently binds to the β-subunit of 20S unit and irreversibly and more specifically block its activity ³⁴.

We first examined whether endogenous cMLCK expression was upregulated by two different proteasome inhibitors, MG132 (Figure S10) and lactacystin (Figure 7A). Both inhibitors increased cMLCK expression in cardiomyocytes. Next, mice were treated with lactacystin, MG132, or both MG132 and lactacystin 3 hrs prior to the TAC operation, followed by daily injection of the inhibitors for 1 week. In comparison to the 80% reduction in cMLCK protein after 1 week of TAC without proteasome inhibitors (Figure 7B, lane 1 vs. 2), addition of proteasome inhibitors attenuated the reduction of cMLCK proteins by 40-60% in all 3 groups and was most effective when the two inhibitors were used in combination (Figure 7B). In contrast to reduced cMLCK expression after TAC, the expression of MLC2v and myosin binding protein C (MyBP-C) assessed on the same membrane was unchanged with or without TAC (Figure 7C). Proteasome inhibitors increased the expression of MyBP-C but not MLC2v. Thus, TAC-induced reductions and partial rescue by proteasome inhibitors were observed for cMLCK but not MLC2v or MyBP-C.

Figure 7.

Involvement of the ubiquitin-proteasome system in cMLCK protein degradation. (A) Representative Western blotting of ubiquitin and cMLCK as well as Coomassie stained actin using cultured neonatal mouse cardiomyocytes with or without lactacystin treatments (5 and 10 µM) for 24 or 48 hrs. (B) Representative Western blotting of cMLCK using heart lysate obtained from mice treated with the proteasome inhibitors, lactacystin and/or MG132 s(1 mg/kg/day). GAPDH expression is also shown. Expression of cMLCK normalized to GAPDH is shown with the value for TAC(-) defined as 1. (C) Western blotting of MLC2v and MyBP-C probed using the same membrane used as in Figure 7B. (D) Western blotting of cMLCK in heart lysates from the pressure-overload Mylk3^+/+ hearts in the presence of proteasome inhibitors (lanes 1, 2) or Mylk3^−/− (lane 3) enriched by TUBE-conjugated beads (left panel), and cMLCK and GAPDH in the input heart lysates (right panel). Representative Western blotting of cMLCK after infection of adenovirus encoding MuRF1 and atrogin1 at ~5 moi and (E) upto 20 moi (F). Expression of cMLCK normalized to GAPDH is shown with the value for control (5 moi adenovirus-GFP) defined as 1 (E).

Third, in the fraction of ubiquitylated proteins enriched by TUBE (tandem ubiquitin binding entity)-conjugated beads from the same amount of heart lysate, cMLCK antibodies preferentially recognized proteins with several different molecular weights in the heart lysate from pressure-overloaded Mylk3^+/+ hearts in the presence of proteasome inhibitors (Figure 7C).

Substrate specificity of ubiquitin-ligation is determined by E3 ligases. Among ~10 muscle-specific E3 ligases, muscle-specific RING finger protein 1 (MuRF1) and atrogin1/MAFbx, are the most intensively studied muscle specific E3 ligases ³², and are upregulated in rodent hypertrophic models ³⁵. Overexpression of E3 ligases, including MuRF1 and atrogin 1, has been shown to promote degradation of target proteins in cultured cardiomyocytes ^{36, 37} but not cMLCK by adenovirus infection into mouse neonatal cardiomyocytes from ~5 to 20 moi (Figures 7E, F).

Collectively, these results suggest that ubiquitin-proteasome-dependent protein degradation and/or processing participates in reduction of cMLCK protein due to TAC, which was not seen for two other proteins and was independent of MuRF1 or atrogin1 in cultured cardiomyocytes.

Discussion

Potential mechanism(s) underlying the transition from compensated to decompensated heart failure are largely unknown. In this study, we demonstrated that functional transition from compensated to decompensated hypertrophy occurred between 1 and 2 weeks of TAC. Reduced cMLCK expression and MLC2v phosphorylation occurred as early as 1 week and continuously during TAC, suggesting a role for MLC2v phosphorylation in this transition and under chronic pressure-overload. Second, this possibility was examined mechanistically in loss- and gain-of function cMLCK mutant mouse models. TAC led to severe heart failure in Mylk3^−/− mice but did not affect cardiac contractility in cMLCK-overexpressing mice when the level of cMLCK expression exceeded that of the normal heart. Third, this process involved ubiquitin-proteasome dependent protein degradation. Fourth, increased cMLCK expression and MLC2v phosphorylation were observed in exercise-induced cardiac hypertrophy after 4 weeks of swim training in Mylk3^+/+ mice with unchanged contractility; while induction of MLC2v phosphorylation was not observed in Mylk3^−/− mice, which exhibited 25% mortality and reduced contractility. Collectively, these data demonstrate that reduced cMLCK expression and consequent reduction of pMLC2 play a pivotal role in mediating the transition from compensated to decompensated hypertrophy. Also, induction of cMLCK is critical for the cardiac adaptive response to exercise stress, presumably due to MLC2v phosphorylation-dependent facilitation of crossbridge formation as discussed in our earlier study ¹⁷.

Transgenic mice overexpressing cMLCK did not display an obvious cardiac phenotype (based on size or function) under basal conditions even with an ~21-fold increase in expression. Following TAC, however, beneficial effects were evident, including maintenance of contractility without fibrosis or apoptotic cell death. Levels of exogenous, transgenically expressed cMLCK protein were also substantially decreased by TAC. Thus, a greater than 6.5-fold induction of cMLCK under basal conditions causes protein synthesis to exceed degradation during TAC, and the increased levels of cMLCK are sufficient to have a beneficial effect on cardiac function after TAC.

Proteasome inhibitors could attenuate, at least in part, the fall in cMLCK protein after TAC: however, inhibition of proteasomal processing does not necessarily improve cardiac function in rats ³⁸. The proteasome inhibitor, Bortezomib, is the first FDA-approved for clinical application. Patients with multiple myeloma treated with Bortezomib demonstrated cardiotoxicity including heart failure, as did normal rats ^{39, 40}. Thus, a more specific methodology to attenuate cMLCK protein degradation during overloading must be identified. Because substrate specificity is defined by E3 ligases, we reasoned that identification of an E3 ubiquitin-ligase specific for cMLCK might be a step in this direction. The two best characterized muscle-specific E3 ligases, MuRF1 and atrogin1, however, failed to reduce cMLCK expression in cardiomyocytes. Thus, identification of cMLCK-specific E3 ligases as well as how proteasomal processing works on cMLCK, whether it be by degradation and elimination or a process yielding biological active polypeptide fragments of cMLCK, remain to be elucidated ⁴¹.

Exercise is a stressor of the heart resulting in regional adaptive hypertrophy and alterations in contractility and biochemical properties of myocardial proteins ⁴². Typically, greater responses to exercise are observed in endocardial than epicardial layers ⁴². In this study, swimming increased the expression of cMLCK and phosphorylation of MLC2v. Mylk3^−/− mice exhibited cardiac dysfunction in response to chronic exercise and also reduced survival compared to control Mylk3^+/+ mice. These findings are in contradistinction to the beneficial effects of exercise reported in other genetic mouse models with cardiac dysfunction, in which exercise improved cardiac function and/or lifespan ^{43, 44}. Collectively, this suggests cMLCK plays an essential role by inducing a beneficial adaptation of the heart to exercise, which is consistent with pMLC2v playing a critical role in cardiac adaptation to increased loading of the heart, such as occurs with regular exercise ⁴⁵.

We also found that spatial distributions of pMLC2v and cMLCK expression were nearly identical, and were neither homogenous nor simple gradients between the inner and outer layers of the rodent heart ¹³. Non-uniform distributions of cMLCK and pMLC2v are consistent with previous observations that the ventricular walls are not homogenous in terms of size, electro-physiological coupling, or active wall stress under basal conditions ^46–48. It is possible that expression of cMLCK is related to spatial orientation of fibers and corresponding variations in fiber stress, which may confer to myocytes across the wall variable contractile properties that contribute to LV torsion ^{3–5, 13}. In this regard, torsion was substantially reduced in Mylk3^−/− and MLC2v (Ser14/15Ala) knock-in mutant mice ¹⁷.

cMLCK expression is higher in RV than in LV myocardium, exhibiting an inverse relationship to systolic pressure (~17 mmHg in RV⁴⁹ vs. ~100 mmHg in LV). cMLCK expression in the LV was markedly downregulated due to TAC. Furthermore, cMLCK is highly expressed in the specialized cardiomyocytes of the ventricular conduction system, even though the contractile protein content in these cells is substantially less than in contractile myocytes. These results could be interpreted as showing that increased mechanical stress reduces net expression of cMLCK by increasing ubiquitin-proteasome activity. It is also possible that cMLCK expression and pMLC2v levels are related to diastolic pressure, decreasing with chronic increases in diastolic pressure as proposed previously ⁵⁰. Of course, both active and passive stresses vary within the wall as a function of the radius of curvature, which in turn varies with depth within the wall and also position along the base-apex axis. Viewed in this way, an inverse relationship between wall stress and cMLCK and pMLC2v expression might provide a regulatory mechanism by which variations in stress are normalized.

With respect to the ventricular conduction system, our observation of high levels of cMLCK expression and pMLC2v in normal mouse and human hearts was unexpected. It seems unlikely that increased cMLCK and pMLC2v would influence the activity of the conduction system, although our experiments do not rule out this possibility.

Our finding of depressed cardiac contraction in near complete absence of MLC2v phosphorylation in our Mylk3^−/− mice is consistent with results of a recent study using hypomorphic cMLCK mutant mice ²². However, our Mylk3^−/− mice demonstrated milder cardiac dysfunction without an increase in interstitial fibrosis, and only a marginal increase in fetal gene expression including ANF and BNP under basal conditions. The milder phenotype displayed in our Mylk3^−/− mice might be attributed to the absence of Neo^R cassettes, or differences in mouse genetic backgrounds and other experimental conditions ⁵¹.

A limitation of this study is that cardiac function was different between the mouse models (mutants vs. wildtype) under basal conditions prior to being subjected to a cardiac stress. This was particularly true for Mylk3^−/− mice. However, since we observed reciprocal phenotypes in mice with constitutively less or more cMLCK, our conclusions appear to be sound.

In summary, this study demonstrates that cMLCK plays essential roles in cardiac adaptations to exercise and pressure overload. Development of heart failure under pressure overload is attenuated by greater expression of cMLCK without apparent adverse effects on normal cardiac function in the absence of increased loading. Thus, an increase in expression of cMLCK or enhancement of its function is a potentially important therapeutic strategy for the treatment of heart failure resulting from pressure overload (i.e., hypertensive heart failure).