Uncoupling of myofilament Ca²⁺-sensitivity from troponin I phosphorylation by mutations can be reversed by Epigallocatechin-3-Gallate

Abstract

Aims

Heart muscle contraction is regulated via the β-adrenergic response that leads to phosphorylation of Troponin I (TnI) at Ser22/23, which changes the Ca²⁺-sensitivity of the cardiac myofilament. Mutations in thin filament proteins that cause Dilated Cardiomyopathy (DCM) and some mutations that cause Hypertrophic Cardiomyopathy (HCM) abolish the relationship between TnI phosphorylation and Ca²⁺-sensitivity (uncoupling). Small molecule Ca²⁺-sensitisers and Ca²⁺-desensitisers that act upon troponin alter the Ca²⁺-sensitivity of the thin filament but their relationship with TnI phosphorylation has never been studied before.

Methods and Results

Quantitative in vitro motility assay showed that 30 μM EMD57033 and 100 μM Bepridil increase Ca²⁺-sensitivity of phosphorylated cardiac thin filaments by 3.1 and 2.8-fold respectively. Additionally they uncoupled Ca²⁺-sensitivity from TnI phosphorylation, mimicking the effect of HCM mutations. EGCG decreased Ca²⁺-sensitivity of phosphorylated and unphosphorylated wild-type thin filaments equally (by 2.15±0.45 and 2.80±0.48-fold respectively), retaining the coupling. Moreover, EGCG also reduced Ca²⁺-sensitivity of phosphorylated but not unphosphorylated thin filaments containing DCM and HCM-causing mutations, thus the dependence of Ca²⁺-sensitivity upon TnI phosphorylation of uncoupled mutant thin filaments was restored in every case. In single mouse heart myofibrils, EGCG reduced Ca²⁺-sensitivity of force and k_ACT and also preserved coupling. Myofibrils from the ACTC E361G (DCM) mouse were uncoupled; EGCG reduced Ca²⁺-sensitivity more for phosphorylated than unphosphorylated myofibrils, thus restoring coupling.

Conclusion

We conclude that it is possible to both mimic and reverse the pathological defects in troponin caused by cardiomyopathy mutations pharmacologically. Re-coupling by EGCG may be of potential therapeutic significance for treating cardiomyopathies.

Introduction

Hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM) are common inherited diseases. HCM is detected in 1 in 500 of the adult population and the causative mutations are overwhelmingly in the proteins of the cardiac muscle contractile apparatus ^{1, 2}. At least 30% of cases of idiopathic DCM are of genetic origin, and DCM with no additional complications such as conduction disease is usually caused by mutations in the contractile apparatus ^{3, 4}. In many cases, the cardiomyopathies result from missense mutations in one of the proteins of the muscle thin filament (actin, tropomyosin, troponin I, C or T). Investigations of the mechanisms that cause HCM and DCM have generally found that mutations in the muscle thin filament cause abnormalities in the Ca²⁺-regulatory system of the muscle ^5–7.

HCM is closely associated with enhanced myofilament Ca²⁺-sensitivity although the process by which chronically high Ca²⁺-sensitivity leads to the symptoms of HCM: hypertrophy, fibrosis and arrhythmias, is uncertain. In DCM, the situation is more complex. It has been suggested that DCM-causing mutations are associated with a reduced Ca²⁺-sensitivity ⁸ but comprehensive surveys now show that DCM-causing mutations in thin filament proteins can increase or decrease Ca²⁺-sensitivity ^{5, 9}. However, the DCM phenotype is always linked to the absence of modulation of Ca²⁺-sensitivity by troponin I (TnI) phosphorylation (uncoupling) and this has been proposed to be causative of the DCM phenotype ^{9, 10}. In addition, the coupling between TnI phosphorylation and change of Ca²⁺-sensitivity seems to be lost due to HCM mutations in thin filament proteins when studied by in vitro motility assay (IVMA) techniques and in myofibrilar assays ^11–15, but in permeabilized muscle this phenomenon is not observed ¹⁶. Thus, in HCM the relationship between TnI phosphorylation and Ca²⁺-sensitivity is not clear.

Uncoupling appears to be closely associated with mutations in thin filament proteins that cause cardiomyopathies. The decrease in Ca²⁺-sensitivity upon phosphorylation of TnI and the corresponding increase in the rate of Ca²⁺-dissociation from troponin C (TnC), is a key component of the lusitropic response to β1-adrenergic stimulation in the heart. The uncoupling reported in DCM and HCM, is likely to impact on cardiac reserve pathologically, and indeed, most studies of mouse models with cardiomyopathy mutations in thin filament proteins report a blunted response to adrenergic stimulation in vivo (reviewed in ¹⁷).

A number of small molecules have been found that alter myofilament Ca²⁺-sensitivity by binding to troponin and act as either Ca²⁺-sensitisers or Ca²⁺-desensitisers. The effect of these reagents on Ca²⁺-sensitivity is well documented, but the effect of small molecules on the coupling between Ca²⁺-sensitivity and TnI phosphorylation has not previously been considered.

EMD57033 and Bepridil are well established Ca²⁺-sensitisers acting directly upon TnC ^18–20. Epigallocatechin-3-Gallate (EGCG), the principal polyphenol isolated from green tea, is reported to be a Ca²⁺-desensitiser that also acts via a binding site on TnC ^21–23. Structural studies indicate that the regulatory Ca²⁺ binding site (site II in the N-terminal lobe of TnC) is closely coupled both to the binding of the TnI switch peptide (147-163), critical for neutralising the inhibitory action of TnI, and the cardiac specific N-terminal peptide of TnI (1-30) that contains the phosphorylatable serines 22 and 23 ^24–28. We have proposed that DCM-causing mutations in thin filament proteins uncouple phosphorylation from the change in Ca²⁺-sensitivity by disrupting this coupled allosteric system ^{9, 17}. It is likely that Ca²⁺-sensitisers and desensitisers binding to TnC would also have an effect on the coupling between Ca²⁺-sensitivity and TnI phosphorylation by modulating the coupled system.

We have therefore investigated how EMD57033, Bepridil and EGCG affect Ca²⁺-regulation and its modulation by phosphorylation in native human heart thin filaments and how these reagents interact with mutations in thin filament proteins associated with HCM or DCM. Using IVMA and single myofibril contractility we have confirmed the Ca²⁺-sensitising effects of EMD57033 and Bepridil and demonstrated that they also uncouple Ca²⁺-sensitivity from the TnI phosphorylation level, thus mimicking the effects of HCM-causing mutations. In contrast, EGCG decreases Ca²⁺-sensitivity in native thin filaments whilst retaining the modulation of Ca²⁺-sensitivity by TnI phosphorylation. Moreover, EGCG has the unique ability to restore the coupling to uncoupled HCM and DCM mutant thin filaments and myofibrils, thus antagonising the disease-causing defect. This property of EGCG may be of therapeutic significance for treating some cardiomyopathies.

Methods

Sources of contractile proteins

Donor heart tissue, used as control, and end-stage failing heart tissue from explanted hearts were obtained from the Sydney Tissue Bank Sydney, Australia ^1,2. Ethical approval for collection and use of tissue samples was obtained from the St Vincent’s Hospital, Sydney and Brompton, Harefield and NHLI Research Ethics Committees.

Troponin was isolated from 100 mg of human heart muscle using an anti-cardiac troponin I (TnI) monoclonal antibody affinity column as described by Messer ²⁹. Troponin containing the TNNC1 G159D mutation was purified from explanted heart samples as previously described ^{30,9, 31} . Recombinant TNNT2 K280N and TNNI3 K36Q was introduced into donor heart troponin by exchange as described ^{29, 32}. Wild-type α-tropomyosin (Tpm1.1) and the mutants E40K, E54K and E180G were expressed in a baculovirus/Sf9 system with a protocol based on that of Akkari et al. ^{9, 33, 34}. Native, E361G and E99K mutant mouse cardiac actins were extracted from transgenic mouse hearts as described by Song et al. ³⁵.

Manipulation and measurement of TnI phosphorylation level

Troponin isolated from human heart samples has a high level of phosphorylation, which was reduced by treatment with shrimp alkaline phosphatase (Sigma P9088). Recombinant TnI was phosphorylated by treatment with protein kinase A (PKA) catalytic subunit (Sigma, P2645-400) as previously described. ^{9, 32}. To dephosphorylate mouse heart troponin, mice were treated with Propranolol as described ¹⁰. TnI phosphorylation levels in myofibrils and isolated troponin was measured by Phosphate affinity SDS-PAGE as described ^{9, 36}, results are shown in Supplement C.

Quantitative in vitro motility assay

Thin filaments were reconstituted with 10 nM rabbit skeletal or mouse cardiac muscle α-actin (labelled with TRITC phalloidin), tropomyosin (40-60 nM) and troponin (60 nM) to study Ca²⁺-regulation of filament motility by the quantitative in vitro motility assay ^{29, 37}. Thin filament movement over a bed of immobilised rabbit fast skeletal muscle heavy meromyosin (100 μg/ml) was compared in paired motility chambers in which troponin varied by a single factor (mutation, phosphorylation state or treatment with drug). Filament movement was recorded and analysed as previously described ³⁸, yielding two parameters, the fraction of filaments moving and the speed of moving filaments.

Contraction of isolated myofibrils

Contraction and relaxation of isolated mouse myofibrils was initiated using a fast-solution change system and sensitive force transducer system recently described ¹⁰. Further details are in Supplement A.

Results

Relationship between phosphorylation and Ca²⁺-sensitivity

We measured the Ca²⁺-dependence of thin filaments containing human heart troponin by the in vitro motility assay (IVMA). The assay yields two parameters, fraction of filaments motile and sliding speed of the motile thin filaments. Both parameters are Ca²⁺-dependent and Ca²⁺-activation curves were plotted (Figure 1A). 3.9 μM Ca²⁺ increased the fraction of filaments motile from 0.03 to 0.83 and the sliding speed from 1.5 μm/sec to 3.0 μm/sec. The EC₅₀ and Hill co-efficient (n_H) determined by fitting the Hill equation to the data were similar for both parameters under all conditions; therefore, in this manuscript we have only shown the fraction motile parameter. Full data for both parameters is shown in Supplement B.

Figure 1. Relationship between EC₅₀ and TnI phosphorylation in native human thin filaments and thin filaments containing HCM-causing mutation TPM1 E180G.

Sliding speed and fraction of filaments motile, measured in the same experiment by in vitro motility assay (IVMA) is plotted against [Ca²⁺] for representative experiments. Raw data from one experiment is shown here, the mean values of EC₅₀ from replicate experiments is shown in Table 2. Solid lines and points, phosphorylated troponin (P); dotted lines and open points, unphosphorylated troponin (unP). Error bars represent SEM of 4 measurements of motility in the same motility chamber. Blue, native thin filaments; purple, HCM-causing mutation TPM1 E180G present.

A. Native thin filaments: phosphorylation increased EC₅₀ (decreased Ca²⁺-sensitivity) but had no effect on the maximum sliding speed or fraction of filaments motile at saturating Ca²⁺.

B. Thin filaments containing TPM1 E180G HCM-causing mutation. The relationship of Ca²⁺-sensitivity to TnI phosphorylation is uncoupled.

Human heart troponin has a high level of TnI phosphorylation in the 1.4-1.8 mols Pi/mol range ³⁶. This is reduced to less than 0.3 mols Pi/mol by treatment with phosphatase, see Supplement C. The Ca²⁺-sensitivity of unphosphorylated thin filaments is 1.88±0.10 (p<0.0001, n=16) times greater than the native phosphorylated thin filaments as reported previously ²⁹. n_H was significantly reduced from 2.11±0.18 for phosphorylated to 1.72±0.11 for unphosphorylated thin filaments, p=0.017, n=16. The maximum sliding speed was not significantly affected by phosphorylation level (Table 1, Figure 1A).

Table 1. Effect of EMD57033, Bepridil and EGCG on Ca²⁺-regulation of motility.

TREATMENT	EC50	nH	ΔVmax	EC50, TPM1 E54K DCM mutation
Native    P	0.16±0.03(4)	2.08±0.14(4)	1.40±0.11(4)	0.12±0.02(3)
unP	0.091±0.014(4)†	1.51±0.25(4)†	1.30±0.04(4)	0.11±0.008(3)
+EMD57033  P	0.057±0.010(4)**	1.26±0.07(4)*	1.55±0.08(4)	0.041±0.002(3)*
+EMD57033   unP	0.053±0.010(4)	1.34±0.14(4)	1.54±0.08(4)*	0.041±0.002(3)*
Native   P	0.13±0.007(8)	2.89±0.48(8)	2.09±0.14(8)	0.13±0.014(3)
unP	0.072±0.002(6)††	1.96±0.08(6)	2.13±0.20(6)	0.13±0.012(3)
+Bepridil   P	0.045±0.001(8)**	1.94±0.08(8)	2.40±0.15(8)**	0.060±0.009(3)*
+Bepridil   unP	0.045±0.001(6)**	2.03±0.13(6)	2.24±0.16(6)	0.058±0.006(3)**
Native   P	0.14±0.03(7)	1.61±0.19(7)	1.91±0.18(7)	0.11±0.01(5)
unP	0.059±0.011(6)††	1.50±0.21(6)	1.92±0.22(6)	0.11±0.02(5)
+EGCG   P	0.26±0.02(7)**	1.96±0.24(7)	1.38±0.12(7)**	0.23±0.03(5)*
+EGCG   unP	0.15±0.02(6)**††	1.79±0.30(6)	1.52±0.13(6)*	0.071±0.005(5)††

^*p<0.05

^**p<0.01; for presence and absence of reagent

^†p<0.05

^††p<0.01; unphosphorylated compared with phosphorylated. Paired t test was used for EC₅₀, n_H and ΔV_max (the change in velocity due to Ca²⁺). EC₅₀ values rounded to 2 significant figures. In brackets is the number of experiments.

We have investigated thin filaments incorporating 6 DCM-causing mutations that have been shown to abolish this relationship (uncouple) in previous studies ⁹ and 3 HCM mutations that uncouple in vitro (Figure 1B ^{15, 39}).

EMD57033 and Bepridil increase Ca²⁺-sensitivity and uncouple TnI phosphorylation from changes in Ca²⁺-sensitivity

EMD57033 and Bepridil substantially increased the Ca²⁺-sensitivity (Figure 2) and slightly increased the maximum sliding speed of native human thin filaments as expected.

Figure 2. Effect of Ca²⁺-sensitisers on Ca²⁺ control of motility.

A, B, D, E. Fraction of filaments motile, measured by IVMA is plotted against [Ca²⁺] for representative experiments, details as for Figure 1.

A, D. Effect of Ca²⁺-sensitisers on Ca²⁺-regulation of native thin filaments. Blue, native thin filaments; orange, presence of 30 μM EMD57033; brown, presence of 100 μM Bepridil.

B, E. Effect of Ca²⁺-sensitisers on Ca²⁺-regulation of thin filaments containing the TPM1 E54K DCM-causing mutation. Red, E54K-containing thin filaments; orange, presence of 30 μM EMD57033; brown, presence of 100 μM Bepridil.

C. EMD57033 dose response curve: EC₅₀ was determined at a range of EMD57033 concentrations. The change in EC₅₀ (± SEM of 4 measurements of motility in the same motility chamber) is plotted for three separate experiments.

F. Bepridil dose response curve: motility was measured at a constant 0.037 μM Ca²⁺ with increasing concentrations of Bepridil. The increase in fraction of filaments motile (± SEM of 4 measurements of motility in the same motility chamber) is plotted for three separate experiments and the curve represents the fit of the pooled data to the Hill equation. Values of parameters obtained are shown.

The effect of EMD57033 on EC₅₀ for Ca²⁺-activation of motility was biphasic with a maximum at 30 μM and an EC₅₀ in the range 15-20 μM (Figure 2C). Interestingly, we observed that 30 μM EMD57033 increased the Ca²⁺-sensitivity of thin filaments containing phosphorylated troponin more than unphosphorylated TnI (mean ratio EC₅₀ P/EC₅₀ P + EMD57033 = 3.07±0.71, p=0.0001, n=4 in phosphorylated compared with a mean ratio of EC₅₀ unP/EC₅₀ unP + EMD57033 = 1.83±0.36, p=0.07, n=4 in unphosphorylated, see Table 1, Figure 2A, Supplement B).

Consequently, in the presence of EMD57033, Ca²⁺-sensitivity of thin filaments was the same, independent of the level of phosphorylation (EC₅₀ P + EMD57033/EC₅₀ unP + EMD57033 = 1.08±0.03, p=0.06, n=4). EMD57033 increased sliding speed slightly at saturating Ca²⁺ concentration in phosphorylated thin filaments and in unphosphorylated thin filaments by 18.4±4.1% (p=0.02, n=4). n_H was slightly decreased in phosphorylated but not unphosphorylated thin filaments. EMD57033 also increased the Ca²⁺-sensitivity of thin filaments containing the DCM-causing mutations, TPM1 E40K and E54K. In both these cases, Ca²⁺-sensitivity was independent of troponin phosphorylation (uncoupled) due to the mutation (Table 1, Figure 2B).

A similar pattern of results was observed with Bepridil. The dose-response curve was cooperative with a calculated EC₅₀ of 52 μM (Figure 2F). At 100 μm, which is the saturated concentration, the increase in Ca²⁺-sensitivity of phosphorylated donor thin filaments (measured by the mean ratio EC₅₀ P/EC₅₀ P + Bepridil) was 2.8±0.14-fold for fraction motile (p<0.0001, n=6) and 3.09±0.59-fold for sliding speed (p=0.016, n=6), see Supplement B.

Bepridil increased the Ca²⁺-sensitivity of thin filaments and also uncoupled Ca²⁺-sensitivity from TnI phosphorylation, as shown by the EC₅₀ ratios of phosphorylated to dephosphorylated troponin (EC₅₀ P/EC₅₀ unP = 1.73±0.03, p<0.0001; EC₅₀ P + Bepridil/EC₅₀ unP + Bepridil = 1.03±0.01, p=0.04, n=6) (Figure 2D, Supplement B). Bepridil increased sliding speed at saturating Ca²⁺ concentrations significantly in phosphorylated thin filaments (14.8±3.2%, p=0.006, n=6) but did not have a significant effect in unphosphorylated thin filaments (7.2±4.6%, p=0.18, n=6); it had no significant effect on n_H. Bepridil also reversibly increased the Ca²⁺-sensitivity of both phosphorylated and unphosphorylated thin filaments containing the DCM-causing TPM1 E54K mutation and these mutant thin filaments remained uncoupled (Table 1, Figure 2E, Supplement E).

EGCG decreases Ca²⁺-sensitivity whilst retaining coupling

EGCG decreased Ca²⁺-sensitivity of human cardiac thin filaments measured by IVMA and at saturating concentrations (100 μM) reduced the maximum sliding speed up to 20% but had no significant effect on n_H (Table 1, Figure 3, Supplement B). Titration of EGCG at 0.074 μM Ca²⁺ yielded a cooperative dose-response curve with a calculated EC₅₀ of 70±7 μM for phosphorylated and 60±1 for unphosphorylated (Figure 3C). Unlike EMD57033 and Bepridil, 100 μM EGCG had a similar effect on the EC₅₀ of thin filaments with phosphorylated and unphosphorylated troponin, (mean ratio EC₅₀ + EGCG/EC₅₀ control = 2.15±0.45-fold for phosphorylated and 2.80±0.48-fold for unphosphorylated thin filaments, see Supplement B). Therefore, the coupling of TnI phosphorylation level to changes in Ca²⁺-sensitivity was retained (EC₅₀ P/EC₅₀ unP = 2.24±0.10, p<0.0001, n=6; EC₅₀ P + EGCG/EC₅₀ unP + EGCG = 1.73±0.16, p=0.006, n=6) (Table 1, Figure 3A, Supplement B).

Figure 3. Effect of EGCG on Ca²⁺ control of motility.

A, B. Fraction of filaments motile, measured by IVMA, is plotted against [Ca²⁺] for representative experiments, details as for Figure 1. Raw data from one experiment is shown here, the mean values of EC₅₀ from replicate experiments are shown in Tables 1 and 2.

A. Effect of EGCG on Ca²⁺-regulation of native thin filaments. Blue, native thin filaments; green, presence of 100 μM EGCG.

B. Effect of EGCG on Ca²⁺-regulation of thin filaments containing TPM1 E54K DCM-causing mutation. Red, E54K-containing thin filaments; green, presence of 100 μM EGCG.

C, D. Dose response curves, the fraction motile was measured at a constant [Ca²⁺] with increasing concentrations of EGCG for a representative experiment.

C. Effect of EGCG on donor thin filaments at 0.074 μM Ca²⁺. The inset plots the change in fraction motility with increasing EGCG concentration and the curve represents the fit of the pooled data to the Hill equation. Values of parameters obtained are shown.

D. Effect of EGCG on thin filaments containing TPM1 E180G tropomyosin at 0.037 μM Ca²⁺. Initially, motility is the same (see Figure 1); the addition of EGCG reduced motility of phosphorylated and increased motility of unphosphorylated thin filaments, indicating recoupling or the restoration of the phosphorylation-dependent Ca²⁺-sensitivity difference (see Figure 3B). Details as for Figure 2F.

EGCG restores coupling to HCM and DCM mutant thin filaments and myofibrils

EGCG also affected Ca²⁺-sensitivity in thin filaments containing mutations associated with HCM or DCM in a phosphorylation-dependent way. This was tested with five DCM-causing mutations and three HCM-causing mutations. EGCG decreased the Ca²⁺-sensitivity of phosphorylated mutant thin filaments in a similar way to wild-type troponin (Ratio EC₅₀ with/without EGCG was 1.36-2.85). However, with unphosphorylated troponin EGCG tended to have no effect or, in the case of the three tropomyosin mutations, increased Ca²⁺-sensitivity (Table 2). As a result, the dependence of Ca²⁺-sensitivity on TnI phosphorylation level was restored. For example, thin filaments containing the DCM-causing TPM1 mutation E54K were uncoupled with EC₅₀ of 0.106±0.013 μM for phosphorylated and 0.114±0.019 μM for unphosphorylated giving a ratio of 0.95±0.03 compared with a ratio of 2.24±0.10 for wild-type thin filaments (Table 1). In the presence of 100 μM EGCG, the EC₅₀ were 0.23±0.03 μM for phosphorylated and 0.071±0.005 μM for unphosphorylated giving a ratio of 3.17±0.37 (p=0.004, n=5) (Table 1, Figure 3B). A dose-response assay for recoupling yielded an EC₅₀ of 58.8±13.3 μM (Figure 3D). Similarly, EGCG was able to reversibly restore coupling of all the thin filament HCM and DCM-causing mutations in this study (Table 2, Figures 4 and 5, Supplement F).

Table 2. Effect of EGCG and TnI phosphorylation on the Ca²⁺-sensitivity of thin filaments containing HCM and DCM mutations.

Mutation	EC50 of thin filaments containing phosphorylated troponin μM ± SEM	EC50 of thin filaments containing unphosphorylated troponin μM ± SEM	Ratio EC50 P / unP ± SEM	EC50 of thin filaments treated with EGCG containing phosphorylated troponin μM ± SEM	EC50 of thin filaments treated with EGCG containing unphosphorylated troponin μM ± SEM	Ratio EC50 P + EGCG / unP + EGCG ± SEM	Ratio EC50 P + EGCG / P ± SEM	Ratio EC50 unP + EGCG / unP ± SEM
WT	0.14 ± 0.03(7)	0.059 ± 0.011(6)††	2.24 ± 0.10(6)††	0.26 ± 0.02(7)**	0.15 ± 0.02(6)**††	1.73 ± 0.16(6)††	2.15 ± 0.45(7)	2.80 ± 0.48(6)
DCM
TPM1 E54K	0.11 ± 0.013(5)	0.11 ± 0.02(5)	0.95 ± 0.03(5)	0.23 ± 0.03(5)*	0.071 ± 0.005(5)††	3.17 ± 0.37(5)††	2.33 ± 0.53(5)	0.68 ± 0.10(5)
TPM1 E40K	0.18 ± 0.03(8)	0.17 ± 0.03(8)	1.02 ± 0.007(8)	0.22 ± 0.04(5)	0.052± 0.01(5)*†	4.63 ± 0.96(5)†	1.70 ± 0.43(3)	0.34 ± 0.05(3)
TNNC1 G159D	0.092 ± 0.004(5)	0.095 ± 0.005(5)	0.97 ± 0.03(5)	0.19 ± 0.03(5)*	0.088 ± 0.005(5)*†	2.25 ± 0.41(5)†	2.13 ± 0.36(5)	0.93 ± 0.02(5)
TNNI3 K36Q	0.18 ± 0.03(10)	0.18 ± 0.04(10)	1.02 ± 0.05(10)	0.19 ± 0.02(5)	0.10 ± 0.009(5)*†	1.97 ± 0.12(5)†	2.32 ± 0.29(3)	1.28 ± 0.04(3)
ACTC E361G	0.087 ± 0.002(5)	0.080 ± 0.002(5)	1.08 ± 0.02(5)	0.20 ± 0.02(5)**	0.081 ± 0.002(5)††	2.41 ± 0.28(5)††	2.27 ± 0.28(5)	1.01 ± 0.02(5)
HCM
TPM1 E180G	0.086 ± 0.013(3)	0.087 ± 0.011(3)	0.98 ± 0.02(3)	0.11 ± 0.01(5)*	0.048 ± 0.003(5)*†	2.41 ± 0.32(5)†	1.36 ± 0.10(3)	0.51 ± 0.06(3)
TNNT2 K280N	0.11 ± 0.007(5)	0.10 ± 0.004(5)	1.08 ± 0.04(5)	0.20 ± 0.03(5)*	0.10 ± 0.004(5)†	1.96 ± 0.24(5)†	2.15 ± 0.18(3)	1.09 ± 0.006(3)
ACTC E99K	0.074 ± 0.005(5)	0.074 ± 0.005(5)	0.99 ± 0.02(5)	0.21 ± 0.03(5)**	0.10 ± 0.007(5)*††	2.03 ± 0.21(5)††	2.85 ± 0.28(5)	1.43 ± 0.13(5)

^*p<0.05

^**p<0.01; for presence and absence of EGCG

^†p<0.05

^††p<0.01; unphosphorylated compared with phosphorylated using paired t test. Ratios: single value t test compared with 1. ANOVA analysis of this dataset is shown in Supplement E. EC₅₀ values rounded to 2 significant figures. In brackets is the number of experiments.

Figure 4. Coupling is restored to HCM and DCM mutations by EGCG.

Fraction of filaments motile, measured by IVMA in the presence of 100 μM EGCG is plotted against [Ca²⁺] for representative single experiments. Details as for Figure 1.

Red lines, thin filaments containing DCM-causing mutations, Purple lines, thin filaments containing HCM-causing mutations.

The mean values of EC₅₀ from replicate experiments are plotted in Figure 5 and summarised in Table 2.

Figure 5. The re-coupling effect of EGCG.

The ratio EC₅₀ phosphorylated/EC₅₀ unphosphorylated (coupling constant) is plotted in the absence (blue) and presence (green, red and purple) of 100 μM EGCG. Green, control troponin; red, DCM-causing mutants; purple, HCM-causing mutants, error bars are ± SEM. All the mutant samples are uncoupled but the coupling constant is restored to control levels by EGCG. Data from Table 2.

The effect of EGCG on contraction of myofibrils

We studied the effect of EGCG on mouse myofibril contractility in basally phosphorylated and dephosphorylated states, obtained by treating mice with propranolol prior to sacrifice (Table 3, Figure 6).

Table 3. Effect of phosphorylation, ACTC E361G mutation and EGCG on Ca²⁺-regulation of myofibril contraction.

In brackets is the number of experiments. Statistical analysis carried out by un-paired t test (equal variance).

		Fmax, kN/m2	EC50, μM	nH	kACT, s-1	tLIN, ms	kREL, S-1
WT¶	P	100.9±6.3(14)	0.93±0.06(11)	10.43±1.84(10)	4.16±0.43(11)	50.8±3.5(12)	35.0±4.0(10)
	unP	87.1±6.0(16)	0.51±0.06(14)†††	4.74±0.66(13)††	4.50±0.24(15)	67.0±4.2(16)†	23.2±2.8(11)†††
WT+EGCG	P	92.5±5.0(15)	1.85±0.17(9)***	4.30±0.85(8)**	4.44±0.24(15)	46.7±3.9(12)	41.0±4.0(12)
	unP	92.6±8.9(8)	1.13±0.13(5)††***	5.08±1.24(5)	3.47±0.55(6)	88.9±9.9(7)†††*	22.3±1.2(7)††
E361G¶	P	93.5±8.9(11)	0.38±0.05(12)§§§	5.39±1.26(9)§	4.51±0.32(13)	75.4±6.5(10)§§	21.6±2.8(10)§§§
	unP	89.3±8.6(10)	0.43±0.06 (10)	4.48±0.70(8)	4.12±0.32(11)	70.8±9.7(11)	21.6±2.1(10)
E361G+EGCG	P	89.1±4.5(12)	2.07±0.20(5)*** ‡‡‡	7.18±1.54(5)	3.43±0.20(13)*§ ‡‡	49.7±3.3(12)**	31.0±2.6(12)*§
	unP	96.3±11.1(11)	1.28±0.10(8)††*** ‡‡‡	4.14±0.42(7)†	3.25±0.26(11)*** ‡‡	63.0±4.7(12)†§	31.0±2.0(12)**§§ ‡‡

^*p<0.05

^**p<0.01

^***p<0.001; EGCG treated compared with no EGCG.

^†p<0.05

^††p<0.01

^†††p<0.001; unphosphorylated compared with phosphorylated.

^§p<0.05

^§§p<0.01

^§§§p<0.001; ACTC E361G compared with wild-type.

^‡p<0.05

^‡‡p<0.01

^‡‡‡p<0.001; ACTC E361G EGCG treated compared with no EGCG wild-type.

^¶Vikhorev et al. 2014 ¹¹.

Figure 6. EGCG restores the relationship between Ca²⁺-sensitivity of force production and TnI phosphorylation in wild-type and ACTC E361G mouse myofibrils.

Phosphorylation level was reduced by propranolol treatment and measured by phosphate affinity SDS-PAGE. Basally phosphorylated wild-type was 1.02±0.03 and ACTC E361G was 1.08±0.01 mols Pi/mol TnI. For propranolol-treated (dephosphorylated) muscle the phosphorylation level was 0.30±0.04 in wild-type and 0.34±0.07 mols Pi/mol TnI for ACTC E361G.

A. Ca²⁺-activation curves for ACTC E361G myofibrils in the presence (green) and absence (red) of 10 μM EGCG. Solid lines and points, phosphorylated troponin (P); dotted lines and open points, unphosphorylated troponin (unP). The plot shows averaged isometric tension ± SEM from 11-14 myofibrils for experiments performed at SL=2.17 μm. EGCG shifts the activation curve to the right and restores the difference between phosphorylated and unphosphorylated myofibrils.

B. Effects of the ACTC E361G mutation, phosphorylation and EGCG on the mean EC₅₀ (± SEM) for myofibril isometric contraction. Significant differences, calculated by t-test, between phosphorylated and unphosphorylated myofibrils are indicated: ** p < 0.01; *** p < 0.001. Data from Table 3.

We found that 10 μM EGCG decreased Ca²⁺-sensitivity of isometric force for both phosphorylated and unphosphorylated mouse myofibrils equally (EC₅₀ P/EC₅₀ P + EGCG = 0.50±0.06, p<0.001, n=9, EC₅₀ unP/EC₅₀ unP + EGCG = 0.45±0.07, p<0.00, n=5). Consequently, the effect of phosphorylation on Ca²⁺-sensitivity in wild-type myofibrils was preserved in the presence of EGCG (EC₅₀ P + EGCG/EC₅₀ unP + EGCG = 1.64±0.24, p<0.001, n=5).

As previously shown ¹⁰, in myofibrils from ACTC E361G DCM mice, the Ca²⁺-sensitivity of force is uncoupled (Table 3, Figure 6). EGCG decreased Ca²⁺-sensitivity in myofibrils from ACTC E361G mice; however, this effect was greater in phosphorylated myofibrils than unphosphorylated myofibrils. Thus, EGCG restored the phosphorylation-dependent shift in Ca²⁺-sensitivity for ACTC E361G myofibrils to the same level as in wild-type myofibrils (EC₅₀ P + EGCG/EC₅₀ unP + EGCG = 1.62±0.20, p<0.01, n=5 compared with 1.82±0.24 for wild-type in the absence of EGCG). n_H values were similar to those found in wild-type myofibrils (Table 3). The duration of the initial, nearly isotonic, phase of relaxation, t_LIN, in phosphorylated and unphosphorylated ACTC E361G myofibrils was restored to wild-type values (Table 3). EGCG decreased the rate of force development (k_ACT) but did not affect maximum force. k_ACT was decreased by EGCG in unphosphorylated wild-type myofibrils (decreased by 22.8%, p=0.059, n=15) as well as in phosphorylated and unphosphorylated ACTC E361G (decreased by 23%, p<0.05, n=10 and by 21%, p<0.001, respectively, n=11).

Discussion

In previous studies it was found that mutations in proteins of the cardiac muscle thin filament that are associated with inherited cardiomyopathies (HCM and DCM) alter myofibrillar Ca²⁺-sensitivity. They also cause the modulation of myofilament Ca²⁺-sensitivity to become independent of the PKA-dependent phosphorylation of TnI. We have named this phenomenon uncoupling. Moreover, this uncoupling effect may be sufficient to generate the disease phenotype of familial DCM ^{9, 17}. We have investigated whether small molecules might be able to mimic or reverse the molecular effects of mutations.

The Ca²⁺-sensitisers EMD57033 and Bepridil, known to bind to TnC, induce uncoupling in wildtype thin filaments, thus mimicking the effects of HCM mutations. The Ca²⁺-desensitiser EGCG has an opposite effect. It preserves coupling in wild-type troponin and restores coupling to thin filaments with HCM and DCM-causing mutations in TnI, TnC, TnT, tropomyosin and actin, thus antagonising the effect of the HCM or DCM mutation. These findings suggest the potential of EGCG for treating the symptoms of inherited cardiomyopathies.

Ca²⁺-sensitisers mimic the effects of HCM-causing mutations in thin filaments

Both EMD57033 and Bepridil increase Ca²⁺-sensitivity of phosphorylated thin filaments by 2.8 and 2.9-fold, measured by quantitative IVMA, in common with previous measurements ^{40, 41}. However, it has not been shown before that they increase Ca²⁺-sensitivity of unphosphorylated thin filaments considerably less than phosphorylated thin filaments. As a consequence, Ca²⁺-sensitivity becomes independent of the TnI phosphorylation level (uncoupled).

The effect of Ca²⁺-sensitisers is analogous to the effect of HCM-causing mutations in vitro (see Table 2) where both an increase in the Ca²⁺-sensitivity and uncoupling of the Ca²⁺-sensitivity from the TnI phosphorylation are observed. Uncoupling due to HCM mutations was first reported in 2001 ¹³ and several more thin filament mutations have subsequently been demonstrated by in vitro assays ^{5, 17}, as the TPM1 E180G mutation, shown in Figure 1B, demonstrates.

Mechanistic considerations of Ca²⁺-sensitisers

Since both EMD57033 and Bepridil appear to modulate Ca²⁺-sensitivity with a minimal effect on the sliding speed in IVMA or maximum isometric force in muscle fibres ^{10, 42} it is likely that they are acting on troponin rather than the cross-bridge cycling mechanism in our systems.

The Ca²⁺-sensitivity of cardiac troponin is modulated by the unique N-terminal peptide of TnI (1-30) that contains the PKA phosphorylation sites, Ser22 and 23. When unphosphorylated the peptide interacts with the N-terminal domain of TnC, close to the regulatory Ca²⁺-binding site and this interaction is lost or reduced when the two serines are phosphorylated ^{43, 44}. The change in Ca²⁺-sensitivity with TnI phosphorylation is a 2-3 fold reduction.

Bepridil is suggested to displace TnI 1-30 and/or the linker helix TnI 31-70 ^{20, 45}. EMD57033 is believed to bind to the C-terminal domain of TnC and also displaces the TnI 31-70 peptide ^{18, 19, 46}. Thus both these compounds could act allosterically by interfering with the modulation of Ca²⁺-sensitivity due to Ser22/23 phosphorylation, resulting in uncoupling.

Interestingly, in experiments measuring isometric force in cardiac muscle myofibrils, EMD57033 increased Ca²⁺-sensitivity but only caused uncoupling at short sarcomere lengths. Thus the uncoupling phenomenon may be graded rather than all-or-none. One possibility is that uncoupling may be related to Ca²⁺-sensitivity if the Ca²⁺-sensitiser shifts the conformational equilibrium so far towards the N-terminal bound conformation that it cannot be significantly influenced by phosphorylation of the N-terminal peptide of TnI (see Supplement D). However the Ca²⁺-sensitising property of these compounds does not necessarily involve the same molecular mechanism as uncoupling.

The actions of small molecules that uncouple highlights the crucial role of the allosteric coupling between the TnI 1-30 peptide and ligand binding sites that could be remotely located. Likewise, it is notable that DCM-causing mutations that uncouple are distributed in all the proteins of the thin filament, also demonstrating long-range allosteric interactions ^{9, 47}. We have proposed that these mutations destabilize the unphosphorylated state of the 1-30 peptide ⁹.

The Ca²⁺-desensitiser EGCG restores coupling to thin filaments with HCM and DCM mutations

Our observation that EGCG reduces Ca²⁺-sensitivity 2-3 fold both in human thin filaments and in mouse myofibrils with only small effects on maximum sliding speed, isometric force or Hill coefficient, is in accord with previous measurements in skinned cardiac muscle fibres ^{21, 23}. Moreover we have demonstrated that the phosphorylation-dependent shift in myofilament Ca²⁺-sensitivity is unaffected by EGCG in both isolated filaments and intact myofibrils, in contrast to the Ca²⁺-sensitisers (Figures 2, 3 and 6).

EGCG was also able to decrease the Ca²⁺-sensitivity of phosphorylated thin filaments containing HCM and DCM-causing mutations, however it had different effects on phosphorylated and unphosphorylated thin filaments. With phosphorylated troponin, EGCG decreased Ca²⁺-sensitivity similarly to wild-type troponin, but EGCG either had no effect or increased Ca²⁺-sensitivity with unphosphorylated troponin (Table 2). This resulted in the restoration of the coupling of Ca²⁺-sensitivity change to TnI phosphorylation (Figures 3, 4 and 5).

The ability of EGCG to re-couple all the uncoupled mutations in any of the thin filament proteins equally (actin, tropomyosin, TnI, TnT and TnC; Figures 4 and 5 ^{9, 17}) is compatible with a common mechanism for uncoupling and implies that the primary effect of EGCG would be to re-stabilize the unphosphorylated state of the TnI 1-30 peptide. We have previously demonstrated that uncoupling and absolute Ca²⁺-sensitivity are not related when comparing a series of DCM-causing mutations ⁹ so the Ca²⁺-desensitising and re-coupling activities of EGCG may involve separate activities. Indeed, we have found analogues of EGCG that do not affect Ca²⁺-sensitivity but can still re-couple.

Several studies have located EGCG binding in the C-terminal domain of TnC in the region of the hydrophobic cleft ^{21, 22, 46}. Molecular Dynamics (MD) simulation suggests that it can bind in several interchangeable orientations ⁴⁶ and unlike EMD57033, EGCG can bind to the C-terminal domain of cTnC even in the presence of TnI 34-71 helix ⁴⁶. We suggest that it is possible that EGCG can restabilise the TnI – TnC interaction, restoring the Ca²⁺ response to phosphorylation found in DCM and HCM. MD calculations indicate that the uncoupling mutation cTnC G159D strengthens TnI 34-71 helix binding to the hydrophobic cleft whilst EGCG weakens this interaction ⁴⁶, compatible with their opposite effects on coupling.

Effects of EGCG on myofibrillar contractility

EGCG decreased Ca²⁺-sensitivity in phosphorylated and unphosphorylated wild-type myofibrils by ~2 fold but did not change the relaxation parameters t_LIN and k_REL in phosphorylated myofibrils (see Table 3). The rate of force development (k_ACT), measured at high Ca²⁺, was unchanged in myofibrils with phosphorylated TnI and slightly decreased (p=0.059) in unphosphorylated myofibrils. As k_ACT depends strongly on the Ca²⁺ concentration ¹⁰, we conclude that EGCG shifts the [Ca²⁺]-k_ACT relationship towards higher Ca²⁺ concentration in agreement with IVMA data (Table 1). Thus, unlike EMD57033, EGCG decreases cross-bridge activation kinetics ¹⁰.

EGCG reduced Ca²⁺-sensitivity and k_ACT in both phosphorylated and unphosphorylated ACTC E361G myofibrils. In addition it restored the lost difference in Ca²⁺-sensitivity between phosphorylated and unphosphorylated myofibrils and also the difference in the relaxation parameter t_LIN (Table 3). The observation that EGCG does not affect the EC₅₀ P/ EC₅₀ unP or t_LIN in wild-type myofibrils but changes them in ACTC E361G, suggests that EGCG can restore coupling in ACTC E361G myofibrils independently of its Ca²⁺-desensitising function.

Although uncoupling can be clearly demonstrated at the level of skinned muscle fibres or myocytes with thin filament mutations causing DCM ^{35, 48, 49} this does not appear to be true for HCM-causing mutations, despite being indistinguishable at the single filament level. A near-normal response of Ca²⁺-sensitivity to TnI phosphorylation has been reported for TNNC1 L29Q and TPM1 E180G in transgenic mice ^{50, 51} and TNNI3 R145W and TNNT2 K280N in human heart myectomy samples ¹⁶. Despite this, both DCM and HCM linked mutations in transgenic mice models are associated with an impaired response to β1-adrenergic stimulation and reduced cardiac reserve as would be expected if uncoupled ^{15, 35} (reviewed by Messer ¹⁷). The physiological manifestations of the uncoupling seen in unloaded filaments requires further investigation.

Clinical relevance of re-coupling by EGCG

Ca²⁺-antagonists have been suggested as being potentially useful for treatment of HCM ⁵². EGCG represents a new class of Ca²⁺-antagonists with a very favourable functional profile. EGCG acts directly on the Ca²⁺-regulatory system of the thin filament that is also the main target of HCM-causing mutations in sarcomeric proteins ⁵. By binding to troponin it decreases the enhanced Ca²⁺-sensitivity characteristic of HCM whilst also reversing the uncoupling effect we observed in HCM mutations. The ability to restore coupling to DCM mutant myofibrils could be beneficial, since this was proposed to be the primary cause of DCM, but the reduction in Ca²⁺-sensitivity in cases where the mutation reduces Ca²⁺-sensitivity may be deleterious.

EGCG has a wide range of pharmacological properties; indeed it has been cited as an example of a promiscuous molecule ⁵³. Moreover, low doses of EGCG have a positive inotropic effect in intact heart muscle, due to effects on the ryanodine receptor type 2, Na⁺-K⁺ ATPase and Na⁺-Ca²⁺ Exchanger, that may override the Ca²⁺-desensitising effect in the contractile apparatus ^{54, 55}.

Although EGCG is too non-specific to be a viable drug for treating HCM, this study has demonstrated a significant proof of principle: it is possible to directly reverse the molecular mechanism of HCM-causing mutations pharmacologically. The coupling of Ca²⁺-sensitivity to cTnI phosphorylation has been demonstrated to be a labile property of troponin. HCM or DCM mutations and Ca²⁺-sensitisers can induce uncoupling, indicating that small perturbations can destabilise troponin ^{9, 17, 32}. It is very unusual to find a reagent that will give a gain of function by apparently restoring the native conformational state in the presence of mutations that destabilised this state.

Our findings provide a starting point for investigating molecules related to EGCG that may be more efficacious and act specifically on troponin ⁵⁶.

Abbreviations

DCM

Dilated Cardiomyopathy

EGCG

Epigallocatechin-3-Gallate

F_MAX

Maximum force

HCM

Hypertrophic Cardiomyopathy

IVMA

In vitro Motility Assay

k_ACT

Rate of force development

k_REL

Rate of fast relaxation phase

MD

Molecular Dynamics

n_H

Hill Coefficient

P

Phosphorylated

PKA

Protein Kinase A

t_LIN

Duration of slow relaxation phase

TnC

Troponin C

TnI

Troponin I

TnT

Troponin T

unP

Unphosphorylated