Increased phosphorylation of tropomyosin, troponin I, and myosin light chain-2 after stretch in rabbit ventricular myocardium under physiological conditions

Abstract

After a change in muscle length, there is an immediate intrinsic response in the amount of developed force, followed by a slower response. Although it has been well documented that the slow force response is at least in part generated by modification of calcium handling, it is unclear whether regulation at the myofilament level occurs during the slow force response. We set out to investigate myofilament calcium sensitivity and phosphorylation status of myofilament proteins after a step-wise change in cardiac muscle length. Ultra-thin right ventricular intact trabeculae were isolated from New Zealand White rabbit hearts and iontophoretically loaded with the calcium indicator bis-fura-2. Twitch force-calcium relationships and steady state force-[Ca²⁺]_i relationships were measured at various muscle lengths at 37°C using potassium induced contractures. The EC₅₀ significantly decreased with increase in muscle length and maximal active force development significantly increased, while no significant change in the myofilament cooperativity coefficient was found. Phosphoprotein analysis Pro-Q diamond staining as well as phosphorylation-specific antibodies revealed increased phosphorylation of tropomyosin, troponin I, and myosin light chain-2 at longer muscle lengths. Specifically, TnI phosphorylation at Ser^22/23 was increased. Since the immediate response is seen virtually instantaneously and post-translational modifications are thought not to occur within such a very short timeframe, we hypothesize that these increases in phosphorylation occur during the slow response.

INTRODUCTION

The slow force response after a change in muscle length is primarily thought to result from ionic balancing involving calcium, sodium, and hydrogen [1, 2]. A study by Luers et. al. concluded that NHE1 plays a major role in the slow force response by increasing [Na⁺]_i, which in turn alters NCX flux and increases [Ca²⁺]_i [2]. This increase in calcium could result in increased calcium binding to TnC, resulting in a stronger contraction. However, it still remains unclear to what extent changes in myofilament calcium sensitivity may play a role. The above mentioned study was limited by the use of the fluorescent indicator aequorin, which does not allow for the unambiguous calibration of calcium transients, and also requires that the experiments be performed at 30°C and at 0.2 Hz [2], conditions that are quite remote from the physiological situation. Mostly skinned, or chemically membrane permeabilized muscles have been used in previous studies for the assessment of myofilament sensitivity to calcium. Technical limitations necessitated that these studies were performed at non-physiological temperature, and likely artificial protein phosphorylation status [3]. Also, experimental conditions such as ionic strength were not consistent between research groups [4], and there may have been inaccuracies in free calcium calculations [4]. Recently, a new method has been developed for the assessment of myofilament calcium sensitivity in intact preparations at near physiologic conditions, including temperature and pH [3], and it is this method that we implement in the current study to investigate myofilament calcium sensitivity at various lengths in rabbit myocardium.

The increase in intracellular calcium during the slow force response found by the Pieske group [2] may have additional consequences on myofilament calcium sensitivity other than increased calcium binding to TnC. Increases in intracellular calcium could lead to the activation of various signaling cascades which lead to phosphorylation or dephosphorylation of various myofilament proteins. It is currently not well understood how the phosphorylation status of various myofilament proteins may affect myofilament calcium sensitivity, or how phosphorylation may alter force or twitch kinetics even after calcium has dissociated from the TnC associated with a particular regulatory unit. Likewise, it is incompletely understood which proteins are phosphorylated under stretch conditions, or on which sites are they phosphorylated during these various conditions and by which kinases. For example, TnI phosphorylation has been shown to increase [5, 6], decrease [7], or not change [5, 8] myofilament calcium sensitivity depending upon the phosphorylation site or mechanism.

In a study we showed relaxation rates were prolonged at longer muscle lengths while time for intracellular calcium decline did not change, thus resulting in a dissociation between force relaxation and intracellular calcium decline [9]. We hypothesized that this could involve myofilament proteins undergoing post-translational modifications which modify myofilament calcium sensitivity. In the present study, we set out to investigate myofilament calcium sensitivity and phosphorylation status of myofilament proteins after a length step change. We hypothesized that an increase in muscle length is accompanied by an increase in myofilament calcium sensitivity, and that changes in phosphorylation status of certain myofilament proteins are involved.

MATERIALS AND METHODS

All protocols were in accordance with the guidelines of the Animal Care and Use Committee of The Ohio State University. Male New Zealand White Rabbits (weighing approximately 2 kg and 3 months old) were intravenously injected with 5,000 units/kg Heparin and anesthetized with 50 mg/kg pentobarbital sodium. The chest was opened by bilateral thoracotomy, the heart was rapidly excised, and ultra thin trabeculae (average dimensions 119 ± 13 µm wide, 74 ± 9 µm thick, and 2.35 ± 0.34 mm long, n = 7 for calcium transient measurements) dissected from the right ventricle. Muscles of this size were chosen so as to avoid core hypoxia that will be present in muscles greater than approximately 150 µm thick[10] under the conditions used. The muscles were dissected in a Krebs-Henseleit solution containing (in mM) 137 NaCl, 5 KCl, 1.2 MgSO₄, 1.2 NaH₂PO₄, 20 NaHCO₃, 10 glucose, and 0.25 CaCl₂. 20 mM 2,3-butanedione monoxime (BDM) was added to this solution to minimize cutting damage and to arrest the heart [11]. Muscles were mounted into the setup as previously described [3, 12] and stimulated for a 3 ms duration at 120% threshold (typically about 6 V) at 2 Hz, which is at or slightly below the physiological resting rate, while perfused with an oxygenated Krebs-Henseleit solution, now without BDM and containing 1 mM Ca²⁺. The muscles were stretched until an increase in passive (diastolic) force was no longer accompanied by a substantial increase in developed force. The muscles were allowed time to equilibrate at 37°C (approximately 20 minutes, or until twitches were consistent). Previous studies have shown this length (i.e. optimal length) to correspond to a sarcomere length of about 2.2 µm, which approximates the end-diastolic sarcomere length in the in vivo beating heart [13].

Trabeculae were iontophoretically loaded with the fluorescent calcium indicator bis-fura-2 as previously described [14–16]. Briefly, after allowing the muscle to stabilize at optimal length and body temperature, stimulation was stopped, the temperature was switched to 22.5°C (to minimize dye leakage which is higher at 37°C), the micropipette was placed into a single myocyte, and a current of 2–3 nA over a duration of 15–30 minutes was applied to drive the indicator into the cytosol. After the indicator was loaded, temperature was rapidly switched back to 37 °C, and stimulation was resumed, allowing the bis-fura-2 to spread to neighboring myocytes via gap junctions. The speed of the indicator to quickly associate and dissociate from calcium ions is critical to determine reliable intracellular calcium concentrations in the muscle. We recently showed that the dissociation constant of this dye under these conditions is sufficiently fast to measure the calcium transient within 1% accuracy [9].

After equilibration of the muscle and loading of the calcium indicator at 20°C, temperature was switched back to 37°C. Force-development and intracellular calcium were simultaneously measured at 2 Hz and at an extracellular calcium concentration of 1.5 mM. In order to assess myofilament Ca²⁺ sensitivity, potassium contractures were performed at various muscle lengths (taught, optimally stretched until an increase in developed force is accompanied by a disproportional increase in diastolic force, and mid-length as calculated using a micrometer). To show that changes in myofilament sensitivity are not time dependent, but in fact result from changes in the length of the myofilaments, the order of the muscle lengths (taught, mid-length, and optimally stretched) at which K⁺ contractures were performed was alternated.

The K⁺ contracture has been used by others [17], and modified by ourselves [3], to further develop this technique by combining it with Ca²⁺ measurements. Since bis-fura-2 to some extent leaked and/or was photobleached during the experiments, especially at body temperature, it was not possible to obtain an in vivo K_d because this procedure would extend the experimental time by about a full hour, within which the dye concentration would have fallen too much to obtain unambiguous calibration parameters. However, in order to calibrate each experiment, we obtained an R_min and R_max using solutions containing either a calcium chelating agent (EGTA) or high calcium respectively, and used those values together with the published K_d value for bis-fura-2 (370 nM Molecular Probes) to construct a curve of Ca²⁺ concentration vs. the ratio of 340/380. By using the equation [Ca²⁺]_i=K’[R-R_min]/[R_max-R] where K’=380(no Ca²⁺)/380(max Ca²⁺)*K_d, we calculated the amount of [Ca²⁺]_i that corresponds with our fluorescent ratios [14]. Even if the in situ K_d would somewhat differ from the assumed K_d, this variable would impact the data only in a quantitative, but not directional qualitative manner. Moreover, since each muscle served as its own control, little or no impact on the data of a possible different in situ K_d is assumed.

Twitch contractions were continuously recorded throughout the experiment. Force development was normalized to the cross sectional area of the trabeculae to allow for comparison between muscles of different diameters, including those from previously published work. Twitches were recorded at each experimental condition upon stabilization of developed tension. Data were collected and analyzed using custom-designed software (in LabView, National Instruments).

An additional 60 trabeculae twitching at 1 Hz and at either no preload or optimal preload were rapidly frozen with liquid nitrogen, and then quickly removed from the setup. For Pro-Q diamond staining, the tissue from a subset of these trabeculae was homogenized in an SDS protein lysis buffer and loaded on a 13% 8 × 10 cm SDS-PAGE gel (0.75 mm thickness, 4% stacking gel, 15 wells). The gel was then run for 45 min at 175 V, fixed overnight, and stained using the Pro-Q Diamond phosphoprotein basic staining protocol (Invitrogen). The gel was imaged in a Typhoon variable mode scanner (GE Healthcare) using an excitation wavelength of 532 nm and a 610 nm (BP30) emission filter at a photomultiplier tube setting of 450. The gel was then stained using Sypro Ruby and imaged. Densitometric analysis was performed on each band, and the ratio of Pro-Q stain intensity to total protein intensity was calculated. In a different subset of muscles the site-specific phosphorylation of TnI was determined by Western blot similar to that previously described with slight modification [18]. Briefly, single trabeculae were solubilized in sample buffer (2% SDS, 0.1% bromophenol blue, 10% glycerol and 50 mM Tris-HCl, pH 6.8) by heating at 80°C for 5 minutes with vortexing. Following clarification by centrifugation for 5 minutes samples were aliquoted and stored at −80°C. Due to the variable size of the trabeculae, total protein amount was first approximated by the Lowry based RC DC protein assay (Bio-Rad) and similar protein amounts then separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on a 12% Laemmli gel with an acrylamide-to-bisacrylamide ratio of 29:1. The resultant gel was Coomassie stained, scanned and the TnI band quantified (ImageQuant TL, GE) to generate a TnI normalized loading for each individual trabeculae. Total protein of TnI normalized trabeculae was then separated by SDS-PAGE as above and transferred to PVDF. Determination of TnI phosphorylation at Ser^22/23 was conducted by Western blot employing a rabbit anti-pTnI^22/23 primary antibody (Phospho-Troponin I (cardiac) (Ser^23/24); Cell Signaling), anti-rabbit Horseradish Peroxidase linked secondary antibody and autoradiography by chemiluminescence (ECL Plus, GE). Following development the blot was striped with Western Restore Plus (Pierce), re-probed for total TnI using a mouse anti-TnI primary antibody (C5, Fitzgerald) and anti-mouse Alkaline Phosphatase conjugated secondary antibody detected by development with NBT/BCIP substrate. Quantification of pTnI^22/23 and total TnI using different species primary antibodies and development methods is critical to ensure no carryover of the pTnI signal into the total TnI detection. Resultant blots were scanned, quantified by ImageQuant TL and the density of the pTnI^22/23 signal normalized to the ratio of total TnI density for each trabecula. Data were statistically analyzed using t-tests (paired or unpaired) where applicable. A two-tailed value of P < 0.05 was considered significant. Data are represented as mean ± SEM.

RESULTS

First, we set out to quantify myofilament calcium sensitivity in intact, stimulated, trabecular preparations. Figure 1 shows representative K⁺ contracture tracings in the same muscle at different preloads. The contracture developed when the K⁺ solution flowed over the muscle. When the contracture had reached the peak, the solution was switched back to the Krebs-Henseleit solution, thus allowing the muscle to relax. The muscles had been allowed to stabilize at steady-state before each of the potassium contractures were performed. The potassium contractures were performed while the muscles were continuously stimulated at a frequency of 2 Hz. Figure 2 shows a different representative K⁺ contracture experiment plotted and fit with a curve using the Hill equation. An increase in preload is accompanied by an increase in developed tension and increase in myofilament calcium sensitivity. Maximal force increased significantly from no preload to the intermediate length, and again from the intermediate length to optimal preload.

Figure 1.

A. Representative K⁺ contracture tracing at optimal length. B. Representative K⁺ contracture tracing at taut length in same muscle as panel A. A decrease in preload is accompanied by a decrease in developed tension and decrease in myofilament calcium sensitivity. Data collected at a stimulation rate of 2 Hz, 1.5 mM [Ca²⁺]_o, 37 °C, n = 1 trabecula.

Figure 2.

Representative K⁺ contractures fit to Hill equation in same muscle at different preloads. An increase in preload is accompanied by an increase in developed tension and increase in myofilament calcium sensitivity. Data collected at a stimulation rate of 2 Hz, 1.5 mM [Ca²⁺]_o, 37 °C, n = 1 trabecula. Experiment date: 070628.

Next, we observed the intracellular calcium concentration at which force is half maximal (EC₅₀). Figure 3 shows EC₅₀ significantly decreases with increase in preload (P < 0.05), showing an increase in myofilament calcium sensitivity at higher preloads. Figure 4 shows EC₅₀ significantly decreases with an increase in developed tension (P < 0.001). As the muscles are stretched toward optimal preload and developed tension increases, myofilament sensitivity to calcium also increases.

Figure 3.

EC₅₀ decreases with increase in preload, showing an increase in myofilament calcium sensitivity at higher preloads. Data collected at a stimulation rate of 2 Hz, 1.5 mM [Ca²⁺]_o, 37 °C, n = 7 trabeculae. *denotes P < 0.05.

Figure 4.

EC₅₀ decreases with increase in maximal tension. As the muscles are stretched toward optimal preload and maximal tension increases, myofilament sensitivity to calcium increases. Data collected at a stimulation rate of 2 Hz, 1.5 mM [Ca²⁺]_o, 37 °C, n = 7 trabeculae.

Since changes in myofilament calcium sensitivity could result from crossbridge cooperativity, we set out to measure n_Hill. Figure 5 shows no significant difference in the hill coefficient with changes in preload (P = 0.76 from no to optimal preload).

Figure 5.

The hill coefficient does not significantly change with preload (P = 0.76 from no preload to optimal preload). Data collected at a stimulation rate of 2 Hz, 1.5 mM [Ca²⁺]_o, 37 °C, n = 7 trabeculae.

The assessment of myofilament calcium sensitivity in intact muscles now makes it possible to study physiological regulation of this calcium sensitivity that involves post-translational modification of myofilaments. Since we observed that at steady state myofilament calcium sensitivity is increased when the muscle is kept at a longer length, we set out to investigate whether post-translational modifications of myofilaments occur. First, we performed a Pro-Q Diamond staining to globally elucidate potential myofilament proteins which experience a change in phosphorylation status (Figure 6). A Sypro Ruby stain in panel A shows total protein, followed by the phosphoprotein stain in panel B. Panels C–E show significant changes in phosphorylation status of TnI, tropomyosin, and myosin light chain–2 with increasing muscle length (P < 0.05). “Slack” denotes no preload, and “Optimal” denotes optimal muscle preload, or optimal muscle length (see methods for determination of optimal muscle length).

Figure 6.

A. Sypro Ruby stain showing total protein. B. Pro-Q Diamond stain showing phosphorylated protein. C–E. TnI, Tm, and MLC-2 phosphorylation is significantly increased at optimal preload compared to no preload (P < 0.05). Trabeculae flash-frozen at a stimulation rate of 1 Hz, 1.5 mM [Ca²⁺]_o, 37 °C.

Although a comprehensive analysis of all possible proteins and specific amino acid targets was deemed well beyond the initial scope, we did further investigate whether these changes were sufficiently large to be unraveled and detected using antibodies specifically targeted against defined amino-acid residues on a protein. Using an antibody that specifically recognizes TnI when phosphorylated at Ser residues 22 and 23 (anti-pTnI^22/23), figure 7 demonstrates the level of TnI phosphorylation at Ser^22/23 was significantly increased when the muscle contracted in steady-state at the longer length (Optimal) compared to the shorter length (Slack) (TnI normalized pTnI^22/23 phosphorylation in arbitrary units: Slack 489 ± 217; Optimal 966 ± 239; P < 0.05).

Figure 7.

Quantification of TnI Ser^22–23 phosphorylation in single trabecula contracted at either no preload (Slack) or optimal preload (Optimal) was conducted by Western blot analysis. A. Western blot of optimally stretched single trabeculae (lanes 5–6) probed with the anti-pTnI^22/23 antibody shows increased TnI phosphorylation at Ser^22/23 compared to slack trabeculae (lanes 1–4) at relatively similar loading of total TnI detected by the anti-TnI antibody. B. Quantification of the pTnI^22/23 signal normalized to total TnI demonstrates the phosphorylation of optimally stretched trabeculae was significantly increased compared to slack (TnI normalized pTnI^22/23 phosphorylation in arbitrary units, * P < 0.05, n = 4 individual trabeculae/group).

DISCUSSION

It is a well-known and generally accepted phenomenon that when cardiac muscle is stretched within its physiological range, the maximal isometric and twitch developed force increase. In addition it is known that myofilament calcium sensitivity increases when cardiac muscle is stretched [19], but this latter finding has exclusively been shown at low temperature, and/or in permeabilized preparations. We now show, using intact muscles contracting at physiological temperature, that myofilament calcium sensitivity indeed significantly increases when cardiac muscle is stretched under physiologically relevant conditions, while maximal developed force is increased.

Using intact muscle preparations, we found an EC₅₀ which is lower than previously reported for skinned preparations [20, 21]. Likewise, Gao et al. have previously reported that intact preparations exhibit an increase in myofilament calcium sensitivity compared to skinned when assessed in the very same preparation [22]. Our observed EC₅₀ is very close to the EC₅₀ of 620 nM observed by this previous study [22]. However, in the latter study, rat rather than rabbit was used, and the temperature was much lower. In addition, the steady state force calcium relationship was induced by high frequency stimulation, and frequency itself may modulate the EC₅₀ at steady state [12], and does not allow for a direct unambiguous comparison with our work. A direct comparison of our data with a previous study, also on rabbit at body temperature using this protocol, indicates the reproducibility of this method; at 2 Hz at optimal preload, our current value for the EC₅₀ of 653 ± 120 nM is nearly identical to the 679 ± 69 nM observed by our previous study [12]. The potassium-contracture/iontophoresis based assessment is much more technically challenging and has a lower experimental success rate than the more straightforward assessment of myofilament calcium sensitivity in permeabilized preparations at cold temperatures. However, our protocol for intact muscle at body-temperature offers the advantage of exploring regulatory systems that are otherwise either washed away because of the permeabilization (such as kinases and phosphatases), or are inactive/modified by cold temperature.

We used this protocol to investigate whether at different lengths, in steady-state, phosphorylation of the myofilaments is altered by regulatory mechanisms. It is well known that, for instance, when TnI is phosphorylated during β-adrenergic stimulation, myofilament calcium sensitivity is decreased, aiding the heart in relaxation. Thus, myofilament protein phosphorylation could very well have a functionally significant impact on length-dependent myofilament function in steady-state conditions. From our Pro-Q diamond studies we conclude that phosphorylation of tropomyosin, TnI, and MLC-2 is increased after stretch. This data was observed in an animal model that closely resembles human EC-coupling, as well as obtained at physiological temperature. In our previous study [9], we observed an increase in systolic calcium concentration at longer muscle lengths, and this increase in intracellular calcium transients could lead to increased protein phosphorylation and/or dephosphorylation by calcium-dependent mechanisms in the slow force response. Also, these pathways could perhaps remain active even after the decline of the intracellular calcium. Additionally, since the myofilaments were likely to have buffered a greater amount of calcium at higher muscle lengths [4], the actual increase in systolic calcium was probably greater than actually measured. Therefore this substantial increase in systolic calcium could lead to the activation of calcium-dependent kinases and phosphatases, which could in turn post-translationally modify the myofilament proteins. This would be in agreement with a previous study that suggested an increase in intracellular calcium after stretch may activate Ca²⁺/calmodulin, which in turn could activate MLCK, which would phosphorylate MLC-2 [23]. Increases in intracellular calcium could also activate calcineurin, or PP3, a protein phosphatase [24, 25].

In order to demonstrate that these small preparations in the current study can be used to identify specific amino acid involvement, we measured phosphorylation at Ser^22/23 on TnI using a site specific antibody rather than a capture-all stain like Pro-Q diamond for proof-of-principle. While there was a significant increase in phosphorylation at these sites, we can currently not determine if there is an even more pronounced effect at other sites, such at TnI Thr¹⁴⁴ or Ser^43/45. Although phosphorylation of TnI is typically thought to decrease myofilament calcium sensitivity, TnI phosphorylation has been shown to increase [5, 6], decrease [7], or not change [5, 8] myofilament calcium sensitivity, depending upon the phosphorylation site or mechanism. The effect of phosphorylation at certain sites is even uncertain. For example one study showed that replacement of Thr¹⁴⁴ with Glu, mimicking phosphorylation, showed no significant difference in Ca²⁺ sensitivity [5, 8], while another study showed that myofilament calcium sensitivity was decreased by substitution of cTnI Thr¹⁴⁴ with Pro and increased by ssTnI Pro¹²² substitution with Thr [26], suggesting an increase in myofilament calcium sensitivity with TnI Thr¹⁴⁴ phosphorylation. Yet another study showed that Thr¹⁴⁴ phosphorylation results in decreased myofilament calcium sensitivity [27], and another showed that p21-activated kinase may increase myofilament calcium sensitivity by phosphorylation of TnI [28]. A study by Wang et al. showed increased myofilament calcium sensitivity by PKCβII phosphorylation of TnI Thr¹⁴⁴ [6]. Therefore, the mechanisms of TnI phosphorylation that develop after a muscle length change require further study and a more precise determination of the specific phosphorylation sites involved.

Although we also found Tm and MLC-2 to be involved, unambiguously determining the impact of the phosphorylation of these proteins or the underlying signaling cascade is well beyond the scope of the current study. Tropomyosin is phosphorylatable near the C-terminus at Ser²⁸³, possibly by PKC [29, 30]. The phosphorylation of MLC-2 is in agreement with a previous study, which found MLC phosphorylation and increased myofilament responsiveness to contribute to the slow force response [23]. Currently, it is not well understood how myosin light chain–2 influences cardiac contractility or kinetics in this context, and this may be the subject of future investigations.

In conclusion, we show that in intact muscles at physiological temperature, myofilament calcium sensitivity is significantly and greatly enhanced. As a force response to a change in length is seen virtually instantaneously, post-translational modifications observed in this study at steady state when switching from a short to long muscle length are likely occurring as a regulatory mechanism during the slow response. Our initial studies into involvement of the myofilaments in this slow force response have identified MLC-2, Tm, and TnI to be involved to some extent, and have further identified Ser^22/23 on TnI as one of the specific TnI sites. Future studies will have to determine the relative contribution of these proteins and specific sites, and to what extent they contribute to the slow force response, versus the change in ion handling that modifies intracellular calcium handling during this response.