Impact of site-specific phosphorylation of protein kinase A sites Ser²³ and Ser²⁴ of cardiac troponin I in human cardiomyocytes

Abstract

PKA-mediated phosphorylation of contractile proteins upon β-adrenergic stimulation plays an important role in the regulation of cardiac performance. Phosphorylation of the PKA sites (Ser²³/Ser²⁴) of cardiac troponin (cTn)I results in a decrease in myofilament Ca²⁺ sensitivity and an increase in the rate of relaxation. However, the relation between the level of phosphorylation of the sites and the functional effects in the human myocardium is unknown. Therefore, site-directed mutagenesis was used to study the effects of phosphorylation at Ser²³ and Ser²⁴ of cTnI on myofilament function in human cardiac tissue. Serines were replaced by aspartic acid (D) or alanine (A) to mimic phosphorylation and dephosphorylation, respectively. cTnI-DD mimics both sites phosphorylated, cTnI-AD mimics Ser²³ unphosphorylated and Ser²⁴ phosphorylated, cTnI-DA mimics Ser²³ phosphorylated and Ser²⁴ unphosphorylated, and cTnI-AA mimics both sites unphosphorylated. Force development was measured at various Ca²⁺ concentrations in permeabilized cardiomyocytes in which the endogenous troponin complex was exchanged with these recombinant human troponin complexes. In donor cardiomyocytes, myofilament Ca²⁺ sensitivity (pCa₅₀) was significantly lower in cTnI-DD (pCa₅₀: 5.39 ± 0.01) compared with cTnI-AA (pCa₅₀: 5.50 ± 0.01), cTnI-AD (pCa₅₀: 5.48 ± 0.01), and cTnI-DA (pCa₅₀: 5.51 ± 0.01) at ∼70% cTn exchange. No effects were observed on the rate of tension redevelopment. In cardiomyocytes from idiopathic dilated cardiomyopathic tissue, a linear decline in pCa₅₀ with cTnI-DD content was observed, saturating at ∼55% bisphosphorylation. Our data suggest that in the human myocardium, phosphorylation of both PKA sites on cTnI is required to reduce myofilament Ca²⁺ sensitivity, which is maximal at ∼55% bisphosphorylated cTnI. The implications for in vivo cardiac function in health and disease are detailed in the discussion in this article.

during stress and exercise, sympathetic activation of the heart increases heart rate and stroke volume to meet the demands of the body. This is mediated via the stimulation of β₁-adrenergic receptors, which leads to the activation of a downstream kinase, PKA. PKA enhances cardiomyocyte contraction and relaxation by phosphorylation of proteins involved in Ca²⁺ handling and myofilament proteins such as cardiac troponin (cTn)I, cardiac myosin-binding protein-C (cMyBP-C), and titin (for reviews, see Refs. 1 and 37).

PKA-mediated phosphorylation of myofilament proteins is thought to exert a positive lusitropic effect, which enables the heart to relax more rapidly when heart rate increases. This positive lusitropic effect may be induced by a decrease in myofilament Ca²⁺ sensitivity (32, 35, 51) and by enhanced cross-bridge cycling kinetics (11, 20, 33). It is well established (mainly from studies in rodents) that phosphorylation of cTnI at the PKA sites Ser²³ and Ser²⁴ leads to a decrease in myofilament Ca²⁺ sensitivity, through a conformational change of the troponin complex. This structural change reduces the affinity of Ca²⁺ binding to cTnC (16, 30). The role of phosphorylation of cTnI at the PKA sites as a regulator of cross-bridge cycling is less clear. Some studies (11, 20, 36) have reported an increase in cross-bridge kinetics via phosphorylation of cTnI. However, others (7, 33) have attributed an increase in cross-bridge kinetics to phosphorylation of cMyBP-C independent of cTnI phosphorylation, whereas several studies (8, 15, 17, 44) did not find an effect of PKA on cross-bridge kinetics at all. In the present study, we aimed to study the effect of site-specific phosphorylation of cTnI on myofilament Ca²⁺ sensitivity and cross-bridge kinetics in human cardiomyocytes since insights into the functional effects of cTnI phosphorylation and the relation between the level of phosphorylation and the functional effects in the human myocardium are lacking.

The effects of PKA-mediated phosphorylation on the contractile function of cardiomyocytes are challenging to study. First, Ca²⁺-handling and myofilament proteins are simultaneously phosphorylated by PKA. Triton permeabilization of cardiomyocytes eliminates the effects of Ca²⁺-handling proteins, which makes it possible to study the specific effects on myofilament proteins. Second, incubation of triton-permeabilized cardiomyocytes with exogenous PKA does not reveal the functional effects of site-specific cTnI phosphorylation as PKA phosphorylates multiple myofilament proteins. The collective effects are important for in vivo function, but to understand the effects in both healthy and diseased myocardium, information is required of each of the contributing factors. Finally, theoretically, four different PKA-phosphorylated cTnI forms could coexist: one unphosphorylated, two monophosphorylated (either on Ser²³ or Ser²⁴), and one bisphosphorylated (18, 34). A previous study (52) in rodents suggested that bisphosphorylation of cTnI is required for the reduction in myofilament Ca²⁺ sensitivity, but knowledge of the functional consequences of monophosphorylated cTnI also is important since recent studies (48, 50) in human postmortem control hearts and fresh donor transplant hearts with normal cardiac function revealed that ∼40% of cTnI is monophosphorylated at Ser²³. Moreover, differences in the level of monophosphorylated cTnI have been reported between donor and end-stage failing hearts (41, 50).

We therefore studied site-specific functional effects of phosphorylation at both PKA sites separately and in combination in human cardiomyocytes. To this end, myofilament force development was measured at various Ca²⁺ concentrations in triton-permeabilized cardiomyocytes in which the endogenous troponin complex was exchanged with exogenous recombinant whole human troponin complexes. These cTn complexes contained unphosphorylated PKA sites, pseudo-monophosphorylated cTnI, and pseudo-bisphosphorylated cTnI. Ser²³ and/or Ser²⁴ were mutated into aspartic acid and into alanine to mimic phosphorylation and dephosphorylation, respectively. This approach allowed us to study the effect of cTnI phosphorylation on active [maximal force (F_max)] and passive isometric force (F_pas), the force-Ca²⁺ relation, and the rate of force redevelopment (k_tr; a measure of cross-bridge kinetics). In addition, cardiomyocytes from idiopathic dilated cardiomyopathic (IDCM) tissue were used for titration of the effect of pseudo-bisphosphorylated cTnI as these cardiomyocytes have a low level of cTnI bisphosphorylation at baseline (13, 41).

The results of our study suggests that phosphorylation of both PKA sites on cTnI is required to reduce Ca²⁺ sensitivity in human cardiomyocytes as no change in Ca²⁺ sensitivity was observed upon exchange with cTn complexes containing pseudo-monophosphorylated cTnI. The maximal reduction in myofilament Ca²⁺ sensitivity was reached at ∼55% bisphosphorylated cTnI. F_max, F_pas, the steepness of the force-Ca²⁺ relation [Hill coefficient (n_Hill)], and cross-bridge kinetics were not significantly altered by pseudo-phosphorylation of cTnI at Ser²³ and/or Ser²⁴. The implications of these findings for in vivo cardiac function in health and disease are detailed in the discussion.

MATERIALS AND METHODS

Expression and purification of recombinant troponin subunits.

Four different cTnI forms were made via site-directed mutations of Ser²³ and/or Ser²⁴ into aspartic acid (D) to mimic phosphorylation or into alanine (A) to mimic dephosphorylation: pseudo-bisphosphorylated cTnI (cTnI-DD), pseudo-phosphorylated cTnI at only Ser²³ (cTnI-DA) or only Ser²⁴ (cTnI-AD), and unphosphorylated cTnI (cTnI-AA). cDNA encoding human cardiac isoforms [cTnC, cTnT, myc tag-labeled cTnT (cTnT-myc), cTnI, cTnI-AA, cTnI-AD, cTnI-DA, and cTnI-DD] were transformed in Esherichia coli Rosetta2 (27) and cultured under carbenicillin/chloramphenicol selection in Overnight Express TB medium (EMD Biosciences). Cultures were harvested by centrifugation, resuspended in PBS, and centrifuged at 10,000 g. Pellets were stored at −80°C until use.

Troponin subunits were purified using fast protein liquid chromatography (AKTA-FPLC System, Amersham Biosciences) essentially as previously described (27).

Reconstitution of troponin complexes.

Fractions containing equal ratios of cTnT, cTnC, and cTnI subunits were pooled and finally dialyzed against 10 mM imidazole, 200 mM KCl, 5 mM MgCl₂, 2.5 mM EGTA, 1 mM DTT, and 0.1 mM PMSF (pH 6.9, 2 times, 1 liter each) before the complexes were concentrated to a final concentration of >2 mg/ml by centrifugation using Centriprep YM-10 centrifugal filters (Millipore).

Exchange of human troponin complex.

Exchange of recombinant cTn in human cardiomyocytes was done as previously described (5, 27) with minor modifications. Cardiomyocytes were isolated from three donor hearts and from an IDCM heart, which were obtained during heart transplantation surgery. Tissue was collected in cardioplegic solution and stored in liquid nitrogen. Samples were obtained after informed consent and with approval of the local Human Research Ethics Committee of The University of Sydney (no. 7326). This investigation conformed with the principles outlined in the Declaration of Helsinki (1997). The human cardiac samples have been extensively characterized (force characteristics and cTnI phosphorylation) in a previous study (13).

Single cardiomyocytes were mechanically isolated with a glass tissue homogenizer as previously described (27). Cardiomyocytes were subsequently incubated overnight at 4°C in exchange solution containing the appropriate concentration of recombinant human cTn complex (1.0 or 2.0 mg/ml in donor cells or a range between 0.0625 to 2.0 mg/ml in IDCM cardiomyocytes) with the addition of 4 mM CaCl₂, 4 mM DTT, 5 μl/ml protease inhibitor cocktail (P8340, Sigma), 10 μl/ml phosphatase inhibitor cocktails 1 and 2 (P2850 and P5726, Sigma), and 50 nM calyculin A (C5552, Sigma) (pH 6.9). The next day, cardiomyocytes were washed twice in rigor solution and finally in relaxing solution (5.95 mM Na₂ATP, 6.04 mM MgCl₂, 2 mM EGTA, 139.6 mM KCl, and 10 mM imidazole, pH 7.0). It has been previously demonstrated that this method results in a homogenous distribution of recombinant cTn complex within the exchanged cardiomyocyte.

Determination of the degree of troponin exchange.

To determine the degree of cTn exchange and to assess the protein phosphorylation status, half of the suspension of cells was treated with a 2D-clean-up kit (GE Healthcare) as described by manufacturer's protocol after overnight cTn exchange. After treatment, tissue pellets were homogenized in sample buffer containing 15% glycerol, 62.5 mM Tris (pH 6.8), 1% (wt/vol) SDS, and 2% (wt/vol) DTT. Protein concentrations were measured with a RCDC Protein Assay Kit II (Bio-Rad) and ranged between 2 and 4 mg/ml.

Immunoblot analysis was used to determine the degree of exchange of endogenous cTn by the recombinant cTn complex. Therefore, recombinant cTnT was labeled with a myc tag to allow discrimination between endogenous and recombinant cTn complexes. Proteins were separated on a one-dimensional 13% SDS-polyacrylamide gel and blotted onto a nitrocellulose membrane (Hybond) using the protocol supplied by the manufacturer in 1 h at 75 V. A specific monoclonal antibody against cTnT (clone JLT-12, Sigma, dilution: 1:1,250) was used to detect endogenous and recombinant cTnT by enhanced chemiluminescence (Amersham Biosciences).

Myofilament protein phosphorylation.

Phosphorylation levels of sarcomeric proteins were determined before and after cTn exchange using ProQ diamond-stained one-dimensional gels, as previously described (49). Phosphorylation signals were normalized to the intensities of SYPRO ruby-stained myosin light chain (MLC)2 bands to correct for small differences in protein loading. To correct for differences in staining between gels, the PeppermintStick phosphoprotein marker (Molecular Probes) was used. The ratio of the intensities of the ProQ diamond- and SYPRO ruby-stained ovalbumin band was used to correct for intergel differences.

The distribution of endogenous phosphorylated species of cTnI in the donor hearts and in IDCM cardiomyocytes was analyzed using Phos-tag acrylamide gels (FMS Laboratory, Hiroshima University, Japan) as previously described (13, 24).

Isometric force measurements.

Force measurements in cardiomyocytes exchanged with pseudo-phosphorylated cTn complexes were performed as previously described (27). Isometric force was measured at 15°C and at a sarcomere length of 2.2 μm (40). The following parameters were determined: F_max at pCa 4.5, F_pas at pCa 9.0, the Ca²⁺ sensitivity of force development (pCa₅₀), n_Hill, and k_tr at maximal and submaximal Ca²⁺ concentrations. k_tr was determined using a slack-restretch test. Thereafter, after the cell had reached steady-state force in activating solution, it was shortened within 1 ms to 70% of its original length and restretched after 30 ms. As a result of this intervention, force first dropped to zero and then quickly redeveloped to the original steady-state level. A single exponential was fitted to force redevelopment to determine k_tr.

Data analysis.

Data analysis was performed as previously described using the following Hill equation to fit force-Ca²⁺ relations: F(Ca²⁺)/F₀ = [Ca²⁺]^n_Hill/(Ca₅₀^n_Hill + [Ca²⁺]^n_Hill), where F is steady-state force, F₀ is the steady-state force at saturating Ca²⁺ concentration, n_Hill is the steepness of the relationship, and Ca₅₀ (or pCa₅₀) is the midpoint of the relation. Data in donor and failing cardiomyocytes were compared using one-way ANOVA followed by a Bonferroni post hoc test. Values are means ± SE; n is the number of myocytes.

RESULTS

Quantification of troponin exchange in human cardiomyocytes.

Figure 1A shows a representative immunoblot loaded with samples of cardiomyocytes incubated overnight with increasing concentrations (range: 0–2 mg/ml) of recombinant cTn. cTnT-myc migrated more slowly through the gel compared with endogenous cTnT, and, therefore, two cTnT bands were found. The percentage of cTn exchange was calculated from the ratio of cTnT-myc and the total amount of cTnT. In Fig. 1B, the percentage of exchanged cTn was plotted against the cTn concentration in the exchange solution. The results shown in Fig. 1, C and D, indicate that the affinity of the cTnT antibody was the same for cTnT compared with cTnT-myc and that cTnT loading (ranging from 60 to 240 ng cTn/lane) was within the linear range.

Fig. 1.

Quantification of troponin exchange in cardiomyocytes by immunoblotting. A: immunoblot stained with an antibody against cardiac troponin (cTn)T that recognizes both endogenous cTnT (bottom band) and recombinant myc tag-labeled cTnT (cTnT-myc; top band). The example shows a suspension of cardiomyocytes exchanged with increasing concentrations of recombinant cTn. B: average percentages of cTn exchange obtained during all exchange experiments plotted against the cTn concentration in the exchange solution. The following double-exponential curve was fitted to the data points: y = A[1 − exp(−x/k₁)] + (100 − A)[1 − exp(−x/k₂)], yielded an A value of 68.9%, a k₁ value of 0.13 mg/ml, and a k₂ value of 1.29 mg/ml. Error bars are shown when larger than symbol size. C: immunoblot of recombinant cTn containing cTnT without (lanes 1, 3, 5, and 7) or with (lanes 2, 4, 6, and 8) the Myc tag label. D: average cTnT intensities of four blots plotted against the cTn amount (in ng) loaded per lane for cTnT- and cTnT-myc-containing cTn complexes. Linear regression analysis indicated that the slopes and intercepts did not significantly differ. E: average percentages of cTn exchange in cardioymyocytes after an overnight incubation in exchange solution with 1 mg/ml cTn containing the different pseudo-phosphorylated cTnI species (average values represent cTn exchange experiments in cardiomyocytes isolated from 3 donor hearts). No significant differences were found in exchange percentages between the various cTnI complexes. cTnI-AA, pseudo-dephosphorylated cTnI; cTnI-AD, pseudo-monophosphorylation of Ser²⁴; cTnI-DA, pseudo-monophosphorylation of Ser²³; cTnI-DD, pseudo-bisphosphorylated cTnI.

No significant differences in the percentages of cTn exchange were found between cardiomyocytes with the different recombinant cTn complexes. This indicates that the different pseudo-phosphorylated cTn complexes incorporated similarly in the myofilaments. The percentage of exchange with the four pseudo-phosphorylated cTn complexes at a concentration of 1 mg/ml in three donor samples was, on average, 68.7 ± 1.6% (Fig. 1E). At a concentration of 2 mg/ml cTn during overnight incubation, the average percentage of exchange amounted to 86.2 ± 3.5% in cardiomyocytes isolated from three donor samples.

cTnI phosphorylation after exchange.

The endogenous cTnI phosphorylation as assessed by Phos tag analysis (Fig. 2, A and B) showed an average of 67.2 ± 4.6% bisphosphorylated cTnI, 31.0 ± 4.6% monophosphorylated cTnI, and 1.8 ± 0.5% unphosphorylated cTnI in the three donor hearts used in the exchange experiments (Fig. 2B), whereas unphosphorylated (60.3%) and monophosphorylated (33.1%) cTnI were the most abundant forms in the IDCM sample used (Fig. 2A). Figure 2, C (1 mg/ml) and D (2 mg/ml), shows the percentages of unphosphorylated (endogenous unphosphorylated cTnI plus recombinant cTnI-AA), monophosphorylated (endogenous monophosphorylated cTnI plus recombinant cTnI-AD or DA), and bisphosphorylated (endogenous bisphosphorylated cTnI plus recombinant cTnI-DD) cTnI species after exchange calculated on the basis of the cTn exchange percentage determined with immunoblot analyses (Fig. 1E). ProQ Diamond-stained gels demonstrated that no significant differences in the level of phosphorylation of other myofilament proteins (cMyBP-C, desmin, and MLC2) were induced upon exchange with the various cTn complexes (Table 1).

Fig. 2.

cTnI (pseudo-)phosphorylation levels after exchange. A: Phos tag-stained gel showing cTn phosphorylation at baseline (i.e., before exchange). B: distribution of phosphospecies of the three donor samples used in the exchange experiments. C (1 mg/ml) and D (2 mg/ml): percentages of unphosphorylated (0P; endogenous unphosphorylated cTnI plus recombinant cTnI-AA), monophosphorylated (1P; endogenous monophosphorylated cTnI plus recombinant cTnI-AD/DA), and bisphosphorylated (2P; endogenous bisphosphorylated cTnI plus recombinant cTnI-DD) cTnI species after exchange calculated with the cTn exchange percentages determined with immunoblot analyses (as shown in Fig. 1E). IDCM, ideopathic dilated cardiomyopathic sample.

Table 1.

Overview of myofilament protein phosphorylation before and after exchange

	Cardiac Myosin-Binding Protein-C	Desmin	Myosin Light Chain 2
Before exchange	0.60 ± 0.07	0.27 ± 0.02	0.25 ± 0.04
Control exchange	0.58 ± 0.07	0.26 ± 0.03	0.31 ± 0.01
cTnI-AA	0.56 ± 0.04	0.24 ± 0.06	0.31 ± 0.03
cTnI-AD	0.54 ± 0.05	0.24 ± 0.05	0.28 ± 0.02
cTnI-DA	0.58 ± 0.11	0.29 ± 0.04	0.27 ± 0.02
cTnI-DD	0.55 ± 0.03	0.23 ± 0.06	0.25 ± 0.02

Values are means ± SE of the ProQ/SYPRO intensity ratio. Shown is myofilament protein phosphorylation determined before and after exchange in three donor samples via ProQ diamond staining. Control exchange cardiomyocytes were kept overnight in exchange solution without recombinant cardiac troponin (cTn)I added. No significant effect was found of the control exchange protocol on myofilament protein phosphorylation compared with before exchange (by Student's t-test). cTn exchange at 1 mg/ml also did not affect myofilament protein phosphorylation (by one-way ANOVA comparing control exchange and the four cTnI exchange groups). A, alanine substitution; D, aspartic acid substitution.

Comparison of functional parameters after exchange with cTn with or without the myc tag.

To test whether the myc tag, added to cTnT as a tool to quantify cTn exchange, interfered with myofilament function, recombinant cTn complex containing cTnT or cTnT-myc was exchanged in donor cardiomyocytes. No significant differences were found between cells (2 donors, 6–8 cells) exchanged with these complexes (Table 2), demonstrating that the myc tag did not influence any of the functional parameters studied.

Table 2.

Force measurements in cardiomyocytes after exchange with recombinant cTnI

	Fmax	Fpas	pCa50	nHill	ktr-max
1 mg/ml cTn
cTnI-Wt	20.6 ± 1.7	3.3 ± 0.5	5.54 ± 0.02	2.6 ± 0.2	0.59 ± 0.04
cTnI-Wt (myc)	20.0 ± 1.8	3.3 ± 0.4	5.51 ± 0.02	3.0 ± 0.1	0.63 ± 0.03
cTnI-AA (myc)	24.6 ± 1.5	3.2 ± 0.2	5.50 ± 0.01	3.2 ± 0.1	0.59 ± 0.03
cTnI-AD (myc)	23.3 ± 1.8	3.3 ± 0.3	5.48 ± 0.01	3.1 ± 0.1	0.50 ± 0.04
cTnI-DA (myc)	22.3 ± 1.4	3.2 ± 0.3	5.51 ± 0.01	3.3 ± 0.1	0.55 ± 0.02
cTnI-DD (myc)	20.9 ± 1.3	3.2 ± 0.4	5.39 ± 0.01***	3.6 ± 0.3	0.60 ± 0.06
2 mg/ml cTn
cTnI-AA (myc)	24.4 ± 1.0	2.6 ± 0.2	5.48 ± 0.01	3.3 ± 0.2	0.55 ± 0.02
cTnI-AD (myc)	22.7 ± 1.2	3.2 ± 0.2	5.47 ± 0.02	3.0 ± 0.1	0.52 ± 0.04

Values are means ± SE. Recombinant wild-type (Wt) cTn complexes with or without the myc tag were exchanged at 1 mg/ml in two donor samples (6–8 myocytes/group). No significant differences were found in the force parameters between cTn complexes with or without the myc tag label. Myc tag-labeled cTnI mutated at Ser²³ and Ser²⁴ into alanine or aspartic acid was exchanged at 1 mg/ml (13 myocytes/cTnI mutant) in three donor samples. The midpoint of the Ca²⁺ sensitivity of force development (pCa₅₀) was significantly lower in cTnI-DD compared with cTnI-AA, cTnI-AD, or cTnI-DA. Myc tag-labeled cTn-AA and cTn-AD were exchanged at 2 mg/ml (11 myocytes/cTnI mutant) in three donor samples. No significant differences were found in the maximal rate of force development (k_tr-max; in s⁻¹) at saturating Ca²⁺ concentration (pCa 4.5) between cTn-AA and cTn-AD (as analyzed by a Student's t-test). F_max, maximal force (in kN/m²); F_pas, passive force (in kN/m²); n_Hill, Hill coefficient (steepness of the force-pCa curve).

^***P < 0.0001 (by one-way ANOVA followed by a Bonferroni post hoc test).

Effects of pseudo-phosphorylated cTn on myofilament Ca²⁺ sensitivity.

The effect of site-specific pseudo-phosphorylation of PKA sites Ser²³ and Ser²⁴ on pCa₅₀ was measured at various Ca²⁺ concentrations in permeabilized donor cardiomyocytes (each complex: 3 donor hearts, 13 cells) in which the endogenous cTn complex was partially exchanged with recombinant cTn complexes (1 mg/ml cTn complex; Figs. 1E and 2C). Endogenous cTnI phosphorylation was largely removed by exchange with nonphosphorylated recombinant cTn complex (Fig. 2C) and enabled us to compare the functional effects of the four different (un)phosphorylated cTn complexes in donor myocardium. Compared with unphosphorylated cTnI (cTnI-AA), myofilament Ca²⁺ sensitivity was significantly lower after exchange with pseudo-phosphorylated cTnI at both sites (cTnI-DD), as evident from the rightward shift of the force-pCa curve for cTnI-DD compared with cTnI-AA (Fig. 3, A and B). It can also be seen in Fig. 3 that the curves after cTn-DA, cTn-AD, and cTn-AA exchange were indistinguishable. This indicates that pseudo-monophosphorylation of the PKA sites did not influence Ca²⁺ sensitivity (Fig. 3, A and B, and Table 2) and suggests that bisphosphorylation is required to cause a change in Ca²⁺ sensitivity. The difference in pCa₅₀ in donor cardiomyocytes upon exchange with cTnI-DD compared with the averaged pCa values of cells exchanged with cTnI-AA, cTnI-AD, or cTnI-DA was 0.11 ± 0.01 pCa units. In a previous study from our group (27), it has been shown that PKA treatment of human donor cardiomyocytes after exchange with wild-type unphosphorylated cTn decreased Ca²⁺ sensitivity to a similar extent (0.08 pCa units). The similar decrease in myofilament Ca²⁺ sensitivity observed in the present study of 0.11 pCa units reflects a change in EC₅₀ of 0.91 μM between cTn-DD (pCa₅₀: 5.39 and EC₅₀: 4.07 μM) and nonphosphorylated cTnI-AA (pCa₅₀: 5.50 and EC₅₀: 3.16 μM; Fig. 3B), which is likely to be of functional significance. Note that myofilament Ca²⁺ sensitivity after exchange with unphosphorylated cTnI-AA and unphosphorylated wild-type cTn complex were the same (Table 2). This indicates that cTnI-AA exchange mimics the effects of exchange using the unphosphorylated Ser²³- and Ser²⁴-containing cTn complex. In addition, experiments were performed using 2 mg/ml cTn-AA and cTn-AD (3 donors, 11 cells) to increase the extent of cTn exchange (86.2 ± 4%) and thereby increase the level of (pseudo-)monophosphorylated cTnI to 88.1% (Fig. 2D). Also, at higher cTn exchange levels, pseudo-monophosphorylation of Ser²⁴ (cTnI-AD) did not lower myofilament Ca²⁺ sensitivity compared with unphosphorylated cTnI (cTnI-AA; Table 2). This strengthens the observations at 1 mg/ml cTn that pseudo-monophosphorylation has no effect on Ca²⁺ sensitivity.

Fig. 3.

Ca²⁺ sensitivity of force development (pCa) is reduced only after phosphorylation of both Ser²³ and Ser²⁴ on cTnI. A: myofilament force development measured at various Ca²⁺ concentrations in permeabilized donor cardiomyocytes in which the endogenous cTn complex was partially exchanged (68.7 ± 2%) with 1 mg/ml of recombinant myc tag-labeled cTn complexes (13 cardiomyocytes from 3 donor hearts in all groups). B: compared with unphosphorylated cTnI (cTnI-AA), pCa derived from the midpoint of the force-pCa relationship (pCa₅₀) was significantly reduced after exchange with pseudo-phosphorylated cTnI at both PKA sites (cTnI-DD). Exchange with cTnI-DA or cTnI-AD did not alter Ca²⁺ sensitivity compared with unphosphorylated cTnI. The reduced Ca²⁺ sensitivity upon exchange with cTnI-DD is evident from the rightward shift of the force-pCa curve in cTnI-DD-exchanged cardiomyocytes compared with cells exchanged with cTnI-AA, cTnI-AD, or cTnI-DA. ***P < 0.0001, cTnI-DD vs. all other complexes (by posttest Bonferroni analyses of one-way ANOVA).

Saturation of the effect of cTnI bisphosphorylation on myofilament Ca²⁺ sensitivity.

To test how much PKA-mediated cTnI bisphosphorylation is necessary to maximally reduce myofilament Ca²⁺ sensitivity, cTnI-DD was exchanged in IDCM cardiomyocytes. This was done at cTn concentrations between 0 and 2 mg/ml to vary the degree of exchange (4–5 cells/concentration). IDCM cardiomyocytes were used in these experiments because the endogenous levels of cTnI Ser²³/Ser²⁴ bisphosphorylation in the end-stage failing human myocardium are low compared with donor tissue. The results shown in Fig. 2A demonstrate, in agreement with previous studies (13, 41), that cTnI Ser²³/Ser²⁴ bisphosphorylation was low (∼7%) in the IDCM cardiomyocytes used. Comparison of the phosphorylation levels of other myofilament proteins in IDCM tissue (cMyBP-C: 0.39 ± 0.03, desmin: 0.23 ± 0.04, and MLC2: 0.21 ± 0.03) with the values in donor tissue (Table 1) also indicates that, in addition, cMyBP-C phosphorylation was lower in IDCM tissue compared with donor tissue. Cardiomyocytes incubated overnight in exchange solution without complex served as controls. cTnI-DD exchange caused a gradual decline in pCa₅₀ saturating at 51 ± 5% of cTnI-DD (Fig. 4). The maximal decrease in pCa₅₀ was 0.13 ± 0.02 units. This value corresponds well with the difference in pCa₅₀ between cTnI-AA and cTnI-DD in donor cells. Assuming a linear decline in Ca²⁺ sensitivity (slope: 0.0028) between 0% and 51% of cTnI-DD exchange, it can be calculated that 3.6% of cTnI-DD is sufficient to reduce Ca²⁺ sensitivity by 0.01 pCa units.

Fig. 4.

Ca²⁺ sensitivity decreased maximally at ∼55% of bisphosphorylated cTnI Ser²³/Ser²⁴ in end-stage failing cardiomyocytes. Shown is the relation between bisphosphorylated cTnI and myofilament Ca²⁺ sensitivity in IDCM cardiomyocytes (4–5 cells) exchanged with different concentrations of recombinant cTn-DD (concentrations of 0.06125, 0.125, 0.25, 0.5, 1.0, and 2.0 mg/ml of cTnI-DD were used). Cardiomyocytes immersed in exchange solution without complex during the overnight exchange served as controls (0% cTn exchange). cTnI-DD exchange caused a gradual decline in pCa₅₀ of 0.13 ± 0.03, saturating at 51 ± 5% of troponin exchange. *P < 0.05, **P < 0.01, and ***P < 0.001, control vs. cTnI-DD (by posttest Bonferroni analyses of one-way ANOVA). Solid line: pCa₅₀ = 5.608 − 0.0028 × exchange percentage(cTnI-DD).

Effects of pseudo-phosphorylated cTn on F_max, F_pas, and n_Hill.

F_max, F_pas, and n_Hill were determined in permeabilized donor cardiomyocytes (3 donors, 13 cells at 1 mg/ml and 11 cells at 2 mg/ml) in which the endogenous cTn complex was exchanged with recombinant human cTn complexes. F_max, F_pas, and n_Hill (Table 2) were not significantly different after exchange with cTn containing the mono- and bisphosphorylated cTnI species.

Effects of pseudo-phosphorylated cTn on k_tr.

Maximal k_tr at pCa 4.5 (k_tr-max) after exchange with pseudo-unphosphorylated cTn did not differ from k_tr-max after exchange with pseudo-monophosphorylated or pseudo-bisphosphorylated cTn (Table 2). However, because k_tr-max showed a tendency to decrease in cTnI-AD, additional experiments were done at even higher exchange percentages (cTn concentration: 2 mg/ml) where cTnI-AA was compared with cTnI-AD (Table 2). In agreement with the results at 1 mg/ml exchange, k_tr-max did not significantly differ between cTnI-AA and cTnI-AD. In addition, no significant differences in k_tr values measured at submaximal Ca²⁺ concentrations were found after exchange with cTn containing mono- and bisphosphorylated cTnI species (data not shown). Note that k_tr-max after exchange using cTnI-AA was the same as in the exchange using unphosphorylated wild-type cTnI (Table 2). The effect of pseudo-phosphorylation at the PKA sites on force redevelopment was also analyzed with a double-exponential equation (Fig. 5), as previously described by Caremani et al. (6). A fivefold difference in the rate constants of the fast phase (1.12 ± 0.07 s⁻¹, n = 52) and slow phase (0.22 ± 0.03 s⁻¹, n = 52) were observed. In agreement with the results of the monoexponential fit, no differences in the parameter values were observed between groups. On the basis of these data, it can be concluded that pseudo-phosphorylation at Ser²³ and/or Ser²⁴ did not affect k_tr.

Fig. 5.

Determination of the rate of tension redevelopment (k_tr). Maximal k_tr at pCa 4.5 (0.57 s⁻¹) was determined from a single exponential curve fit of force redevelopment after a slack-restretch test in activating solution. This example includes the results of a second-order curve fit (slow component of 0.19 s⁻¹ and fast component of 1.50 s⁻¹). The first-order and second-order curve fit overlapped, and, therefore, only the second-order fit is shown.

DISCUSSION

Both Ser²³ and Ser²⁴ need to be phosphorylated to reduce Ca²⁺ sensitivity.

The present study is the first to examine the effects of site-specific phosphorylation of cTnI- Ser²³ and/or Ser²⁴ in human cardiomyocytes. Our study suggests that (mono)phosphorylation of either Ser²³ or Ser²⁴ does not alter myofilament Ca²⁺ sensitivity and that both serines need to be phosphorylated to reduce the Ca²⁺ sensitivity of myofilament force. These results are compatible with a model proposed on the basis of biochemical studies (29) indicating that monophosphorylation and consecutive bisphosphorylation exert a stepwise effect on the affinity of binding between cTnI and cTnT as well as cTnC, whereas bisphosphorylation of cTnI causes a reduction in the affinity of cTnC for Ca²⁺. A more recent study (16) has provided insights on the conformational transitions within the cTn complex upon bisphosphorylation, but alternative models have been proposed as well (for a recent review, see Ref. 31).

In a previous study by Zhang et al. (52), a chimera of a mouse-bovine cTn complex with alanine mutations was exchanged in skinned porcine cardiac myocytes to mimic dephosphorylation at Ser²³ and/or Ser²⁴. PKA treatment of porcine cells containing alanine-mutated cTn complex did not reduce Ca²⁺ sensitivity, whereas PKA did reduce Ca²⁺ sensitivity in cells exchanged with the wild-type complex. This indicated that Ser²³ and Ser²⁴ were required for the PKA-mediated decrease in Ca²⁺ sensitivity but did not rule out the possibility that the decrease depended on the phosphorylation of other target proteins, such as cMyBP-C and titin. To circumvent this problem, Dohet et al. (10) mutated both Ser²³ and Ser²⁴ of cTnI to aspartic acid or alanine and exchanged human cTn mutants in porcine cardiac muscle fibers. This study showed that bisphosphorylated cTn complex (cTnI-DD) lowered Ca²⁺ sensitivity compared with the nonphosphorylated cTn complex (cTnI-AA). Our study extends these previous studies in that human recombinant proteins were exchanged in human tissue and demonstrates that phosphorylation of both Ser²³ and Ser²⁴ of cTnI is required and necessary to reduce Ca²⁺ sensitivity in the human myocardium.

This conclusion relies on the assumption that alanine and aspartic acid mimic dephosphorylated and phosphorylated serine, respectively. Previous studies have indicated that this is indeed the case with respect to interactions between the cTn subunits (12, 29), Ca²⁺ dissociation rates from cTnC (45), and pCa₅₀ (10). Moreover, it can be noted that we observed similar functional parameters in cells exchanged with dephosphorylated cTnI mimicked with alanine substitutions at Ser²³ and Ser²⁴ and unphosphorylated wild-type cTnI (Table 2). Therefore, it appears safe to assume that alanine and aspartic acid substitutions at Ser²³ and Ser²⁴ of cTnI indeed reflect the structural and functional consequences of Ser²³ and Ser²⁴ (de)phosphorylation.

Maximal reduction in myofilament Ca²⁺ sensitivity at ∼55% of bisphosphorylated cTnI.

To study how much Ser²³/Ser²⁴ bisphosphorylation is required to maximally decrease Ca²⁺ sensitivity, exchange experiments with cTnI-DD were done at different concentrations in cardiomyocytes from IDCM tissue, in which Ser²³/Ser²⁴ bisphosphorylation is low (∼7%; Fig. 2A) (13, 41). The maximal decrease in Ca²⁺ sensitivity was reached at 51 ± 5% of cTnI-DD exchange (Fig. 4). Because of the low percentage of endogenous bisphosphorylation in IDCM cardiomyocytes remaining after exchange (∼3.5%), we conclude that ∼55% bisphosphorylated cTnI Ser²³/Ser²⁴ is required to maximally decrease myofilament Ca²⁺ sensitivity.

The fact that approximately half of the cTn units along the myofilament need to be phosphorylated to cause a maximal decrease in Ca²⁺ sensitivity implies that the "phosphorylation signal' is spread along the thin filament. However, we did not find a significant change in cooperativity (n_Hill) of myofilament force development upon cTnI-DD exchange, and pseudo-monophosphorylation also did not affect the cooperativity of myofilament force development (Table 2). Dohet et al. (10) also did not find a difference in n_Hill after exchange with cTnI-DD compared with cTnI-AA in porcine cardiac muscle fibers. Moreover, n_Hill was not significantly different in myofibrils from wild-type mice compared with transgenic mice with both serine residues mutated to aspartic acid (22, 35). Thus, although n_Hill itself does not appear to be changed by phosphorylation of the NH₂-terminal extension of cTnI, some form of communication along the thin filament needs to be present to explain why partial cTnI phosphorylation maximally reduces the activation state of all participating troponin units.

No effect of exchange of pseudo-phosphorylated cTnI at Ser²³ and Ser²⁴ on maximal and passive force.

Our results indicated that pseudo-phosphorylation of Ser²³ and/or Ser²⁴ of cTnI did not affect F_max at saturating Ca²⁺ levels of human cardiomyocytes. This is in line with previous studies (4, 39) showing no effect of PKA incubation on F_max at saturating Ca²⁺ in human skinned cardiomyocytes. In addition, no change was found in F_pas upon exchange with either pseudo-monophosphorylated or pseudo-bisphosphorylated cTnI. These results are consistent with a previous study (47) indicating that PKA reduces passive tension in cardiomyocytes via phosphorylation of titin's cardiac-specific N2B domain.

k_tr is not affected by cTnI phosphorylation at Ser²³ and Ser²⁴.

A small but statistically insignificant trend for a decrease in k_tr-max at 1 mg/ml cTn exchange in cTnI-AD compared with cTnI-AA was found (Table 2). Additional experiments with higher cTn concentrations (2 mg/ml) did not show any difference in k_tr-max (Table 2). In accordance with our findings, it has been demonstrated that PKA treatment did not affect cross-bridge cycling in skinned trabeculae (8, 17) and cardiomyocytes (15) from rat hearts and in human myofibrils (44). Other studies (11, 20, 36) in rodents showed that PKA-mediated phosphorylation of cTnI did increase cross-bridge kinetics. More recently, it has been shown that PKA-mediated phosphorylation of cMyBP-C increases cross-bridge kinetics in transgenic mice, independent of cTnI phosphorylation (7, 33). The reasons for these differences are unclear, but our data indicate that in the human heart, PKA-mediated phosphorylation of cTnI may induce positive lusitropic effects by affecting Ca²⁺ sensitivity, whereas no effect was found on k_tr, which represents the sum of the apparent rates of cross-bridge attachment and detachment under isometric conditions. However, it cannot be excluded that cross-bridge kinetics are affected by phosphorylation of cTnI in the in vivo situation because Layland and Kentish (21) observed, in line with our findings, in isolated sarcoplasmic reticulum-inhibited cardiac trabeculae no relaxant effect of isoprenaline during isometric contractions, but, on the other hand, an increase in the rate of relaxation during work-loop contractions (which mimic the in vivo situation).

Residual force after the shortening restretch protocol used to determine k_tr ranged between 30% and 50%. This residual force is most likely mainly caused by cross-bridge reattachment during restretch (duration: 2 ms). However, the k_tr values obtained did not depend on the level of residual force, and the average residual force within the experimental groups after exchange did not differ. Moreover, the k_tr values after the exchange protocol were comparable with values reported in previous studies (40, 46) in cardiomyocytes without prior treatment. Therefore, we are confident that the comparison of k_tr in the different experimental groups was valid.

Implications of cTnI phosphorylation at Ser²³ and Ser²⁴ in health and disease.

In the present study, we provided evidence that bisphosphorylation of PKA sites Ser²³ and Ser²⁴ of human cTnI reduces myofilament Ca²⁺ sensitivity and that the functional range lies between 0% and 55% of bisphosphorylation. The level of phosphorylation of these two serines is determined by the balance between kinase and phosphatase activity at the myofilaments. PKA phosphorylates Ser²³ and Ser²⁴ upon stimulation of the β-adrenergic receptor pathway with different affinities (18). In addition, PKC (34), PKD (14), and PKG (2) are all known to phosphorylate cTnI at Ser²³ and Ser²⁴. Although in vitro studies (14, 18, 26, 28, 52) have demonstrated that PKA preferentially phosphorylates Ser²⁴ over Ser²³, it has been recently demonstrated using top-down mass spectrometry that cTnI in human cardiac tissue is only monophosphorylated at Ser²³ (48, 50). This surprising finding may be explained by preferential dephosphorylation of Ser²⁴ by phosphatases (9, 19, 46). Indeed, it has been demonstrated that protein phosphatase 2A has a preference for Ser²⁴ (18). It cannot be excluded that the level of bisphosphorylation needed to maximally reduce Ca²⁺ sensitivity depends on the phosphorylation status of other myofilament proteins and thus might vary with disease state.

In the donor samples used in this study, the level of endogenous bisphosphorylation amounted to 67.2%, which is above the level of saturation of Ser²³/Ser²⁴ bisphosphorylation. Recent studies (48, 50) have reported relatively low levels of bisphosphorylation of cTnI Ser²³/Ser²⁴ (∼15%) and a relatively large fraction of monophosphorylated cTnI at Ser²³ (∼40%) in human postmortem control hearts and in fresh transplant donor hearts. This suggests that the healthy human heart has a reserve to decrease myofilament Ca²⁺ sensitivity during β-adrenergic receptor stimulation. In addition, several studies have demonstrated lower cTnI phosphorylation of the PKA sites (3, 23–25, 41, 50) and higher myofilament Ca²⁺ sensitivity in end-stage failing hearts relative to donor hearts (25, 41), which has been ascribed to downregulation and desensitization of the β-adrenergic receptor pathway. Comparison of human cardiac samples from heart failure patients with different disease severity [ranging from New York Heart Association (NYHA) class I to IV] showed increased Ca²⁺ sensitivity only in the end stage (NYHA class IV) of cardiac disease (38), which suggests that the detrimental effects of reduced cTnI phosphorylation may only become evident at the end stage of heart failure. However, recent studies in patients with obstructive familial hypertrophic cardiomyopathy and normal systolic but impaired diastolic function (NYHA class III) showed increased myofilament Ca²⁺ sensitivity (43) and lower cTnI phosphorylation (24, 42) in familial hypertrophic cardiomyopathic myocarium compared with nonfailing myocardium. In addition, a recent study (50) showed that the level of cTnI bisphosphorylation in postmortem hearts with mild hypertrophy was significantly lower (4.1%) compared with control levels (18.4%). Collectively, these studies indicate that cTnI bisphosphorylation and the associated impact on Ca²⁺ sensitivity depends on the stage of heart failure (NYHA class) as well as on etiology.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant R01-HL-063038.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: P.J.M.W., A.M.M., G.J.M.S., and J.v.d.V. conception and design of research; P.J.M.W., D.B.F., A.L.T., and A.H.F. performed experiments; P.J.M.W., G.J.M.S., and J.v.d.V. analyzed data; P.J.M.W., G.J.M.S., and J.v.d.V. interpreted results of experiments; P.J.M.W. prepared figures; P.J.M.W. drafted manuscript; P.J.M.W., G.J.M.S., and J.v.d.V. edited and revised manuscript; P.J.M.W., D.B.F., A.L.T., A.H.F., C.d.R., A.M.M., G.J.M.S., and J.v.d.V. approved final version of manuscript.