| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
CARDIOVASCULAR |
1 Department of Bioengineering, University of Washington, Seattle, WA, USA
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
(Received 24 January 2007;
accepted after revision 18 February 2007;
first published online 22 February 2007)
Corresponding author M. Regnier: Department of Bioengineering, University of Washington, Seattle, WA, USA, 98195. Email: mregnier{at}u.washington.edu
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
26 A7TnTm structural regulatory units (RUs) aligned end to end, with one set on each side of the actin helix. This structural arrangement implies that Ca2+ binding to a single Tn should initiate cross-bridge binding along a 7 actin span of the thin filament. Within the geometry of the sarcomere, up to 4 myosins heads could possibly interact with each A7TnTm RU (Gordon et al. 2000). However each Tm interacts with neighbouring Tm molecules via head-to-tail overlapping contacts, thereby allowing the possibility that cross-bridge binding can occur along a greater span of the thin filament than the RU. Structural studies indicate that S1 binding to actin influences the position of Tm on the thin filament (Vibert et al. 1997; Xu et al. 1999). Thus multiple protein–protein interactions are involved in establishing the level of thin filament activation, or number of cross-bridges, during Ca2+-activated force development. Ca2+-dependent activation of striated muscle force development also involves cooperative mechanisms, as clearly demonstrated by the steep slope of the steady-state force–negative log [Ca2+] (pCa) curve of chemically demembranated (skinned) muscle cells. There are a number of mechanisms thought to be responsible for this cooperativity, including events that could occur within an RU or between RUs along thin filaments. These include: (1) coupling between Ca2+ binding at N-terminal sites of an individual TnC (in skeletal, but not cardiac muscle) and/or Ca2+ binding to the N-terminus of TnC influencing the Ca2+ binding of adjacent TnC molecules; (2) strong cross-bridge-mediated enhancement of Ca2+ binding to TnC; (3) cross-bridge-mediated stabilization of Tm during Ca2+-dependent activation allowing for exposure of additional strong myosin binding sites within an RU; and (4) Ca2+ plus cross-bridge binding in one RU influencing the availability of strong cross-bridge binding in adjacent RUs (i.e. an allosteric spread of activation along thin filaments mediated via head-to-tail interactions of Tm) (Gordon et al. 2000).
If activation of one RU influences the activation of an adjacent RU, one possibility is that Ca2+ binding to a single Tn makes more than the 7 actins within the RU available for strong myosin binding. We (Regnier et al. 2002) recently interrupted interactions between neighbouring RUs of thin filaments in fast skeletal muscle and determined that 10–12 actins are made available for strong myosin binding with Ca2+ binding to each Tn. This spread of activation beyond the RU suggests that a functional regulatory unit (FU) is larger than an RU in skeletal muscle, as was also indicated by solution biochemical studies (Greene & Eisenberg, 1980; Geeves & Lehrer, 1994). Our experiments also demonstrated that isolation of individual FUs from one another resulted in loss of most of the slope in the steady-state force–pCa relationship, suggesting that the dominant mechanism of cooperative activation in skeletal muscle involves interactions between adjacent FUs.
Compared to skeletal muscle, the regulation of cardiac muscle contraction appears to be a more graded response to Ca2+-dependent activation. This is demonstrated by a lower Ca2+ sensitivity of both steady-state force generation and the rate of force development in skinned muscle cells (Regnier et al. 1998b, 2002, 2004), as well as cardiac versus skeletal muscle differences in the sarcomere length dependence of force generation (Konhilas et al. 2002). A possible mechanism for these cardiac versus skeletal muscle differences in contractile activation may be the degree to which thin filaments are ‘turned on’ by the Ca2+-dependent activation of Tn. We (Martyn et al. 1999, 2001) and others (Fuchs & Wang, 1991; Wang et al. 2001) have demonstrated that Ca2+ binding to Tn is enhanced by strong cross-bridges in cardiac, but not skeletal muscle. This suggests one mechanism of cooperative thin filament activation that may be available only to cardiac muscle. We have also demonstrated that cardiac (versus skeletal) muscle contractile activation is more dependent on strong cross-bridge binding (Adhikari et al. 2004; Regnier et al. 2004). Additionally, Butters et al. (1997) measured cardiac S1-ATPase activity in the presence of actin reconstituted with various ratios of functional and non-functional cardiac TnC (cTnC), and concluded that less than 7 actins are activated by Ca2+ binding to each RU. These results also suggested that individual Ca2+-bound RUs may be more completely activated by Ca2+ binding to a neighbouring RU, and that the FU size differs between cardiac and skeletal muscle. However, the size and properties of the FUs in cardiac muscle needs to be examined in cardiomyocytes, where steric constraints imposed by the intact lattice structure of sarcomeres can influence thin–thick filament interactions.
In the current study we used the whole cTn replacement and TnC extraction–reconstitution techniques to study the size and properties of FUs in chemically demembranated rat cardiac muscle. We have previously demonstrated that most or all of the native cTn in preparations of trabeculae can be exchanged (Kohler et al. 2003). Native cTn in rat trabeculae was replaced by different ratio mixtures of functional and non-functional cTn. The non-functional cTn (xcTn) contained a cTnC where the single N-terminal Ca2+ binding site (site II) was made non-functional by replacement of Asp65 with Ala (D65A cTnC (xcTnC)). It was assumed that increasing the proportion of xcTn incorporated into thin filaments increased the probability that one or both cTn surrounding a FU would be non-functional.
We found that reduction of near-neighbour RU interaction did not greatly reduce either the Ca2+ sensitivity (pCa50) or slope (nH, i.e. the apparent cooperatively) of the force–pCa relationship. Additionally, the relative effect of strong cross-bridge inhibition or augmentation on thin filament activation and force production was not influenced by the reduction of near-neighbour RU interactions in cardiac muscle, but was greatly affected in fast skeletal muscle. As such, our data suggest that the length of thin filament activated by Ca2+ binding to a Tn may be 
7 actins in cardiac muscle, compared with 10–12 actins in skeletal muscle. If so, the spread of activation along thin filaments with Ca2+ binding to a Tn (and subsequent strong cross-bridge binding) may be minimal in cardiac muscle and depend more on concomitant binding of Ca2+ to neighbouring RUs and/or cross-bridge-induced increases in Tn–Ca2+ binding affinity. This would suggest that the cooperativity of the force–pCa relationship in cardiac muscle results primarily from mechanisms that differ from those found in skeletal muscle (Regnier et al. 2002). Preliminary reports of this work have been published previously (Regnier et al. 2001; Gillis et al. 2005a).
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Wild-type (WT) rat cTnC and xcTnC were expressed and purified as previously described (Dong et al. 1996). For cTn replacement studies, cardiac troponin I (cTnI) and cardiac troponin T (cTnT) were purified from bovine heart, obtained at a local abattoir, and purified as previously described (Potter, 1982). For the stopped-flow studies of Ca2+ dissociation, the subunits of cTn were produced using recombinant methods. cTn complex was reconstituted from isolated (recombinant or native) subunits (1: 1: 1 ratio) as previously described (Potter, 1982)
Animal tissue
Male Sprague–Dawley rats (200–250 g) and male New Zealand rabbits were housed in the Department of Comparative Medicine at the University of Washington (UW) and were cared for in accordance with the US National Institutes of Health Policy on Humane Care and Use of Laboratory Animals. All protocols were approved by the UW Animal Care Committee. Rats were killed with sodium pentabarbitol (50 mg kg–1) and hearts were excised and the interior wall of the right ventricle exposed to relaxing solution (see below) containing glycerol (50% v/v) and Triton X-100 (1%) overnight at 5°C. Trabeculae were dissected out of the right ventricular free wall as previously described (Regnier et al. 2000) and stored for up to 5 days at 5°C. For further details and solution recipes see Kohler et al. (2003). Single rabbit psoas fibre segments were prepared as previously described (Regnier et al. 2000). Rabbits were killed with pentobarbital (120 mg kg–1) administered through the marginal ear vein. Isolated fibres were treated with 1% Triton X-100 (v/v) in relaxing solution to remove membranous residue.
Exchange of cTn into permeabilized trabeculae
Endogenous Tn was exchanged with Tn produced from purified proteins as previously described (Kohler et al. 2003). In brief, mounted trabeculae were soaked in a rigor solution for 2 h containing a high concentration (3.0–3.5 mg ml–1) of the recombinant Tn and (mM): 3-(N-morpholino)propanesulphonic acid (Mops) 20, MgCl2 5, EGTA 5, KCl 240, dithiothreitol (DTT) 5, butanedione monoxime (BDM) 5 and pepstatin 0.02; pH 6.5 at 10°C.
Following exposure to exchange buffer, trabeculae were exposed to a solution containing bovine serum albumin (1 mg ml–1). The following mixtures, in percentage, of WT cTn: xcTn were made for exchange into trabeculae: 100: 0, 75: 25, 50: 50, 25: 75 and 0: 100. Exchanges with 100% xcTn were performed on a periodic basis to validate that the exchange with native cTn was nearly complete.
Extraction of TnC and reconstitution of Tn complexes
In some experiments TnC was selectively extracted from muscle cells as previously described using trifluoperazine (TFP) (Regnier et al. 1999, 2002; Moreno-Gonzalez et al. 2005). Cardiac trabeculae or rabbit psoas fibres were placed in extracting solution containing 10 mM Mops, 5 mM EDTA and 0.5 mM TFP at pH 6.6 (Metzger et al. 1989; Hannon et al. 1993). Most muscle preparations were placed in extraction solution for 30 s followed by 10–15 s in relaxing solution (pCa 9.0), and this procedure was repeated five to 10 times. The preparation was then activated at pCa 4.0 to determine the remaining Fmax and additional extraction was performed until Fmax no longer decreased. For trabeculae, the remaining Fmax was 0.26 ± 0.07 of the pre-extracted value. For psoas muscle fibres, extracted values ranged between 0 and 3% of pre-extracted Fmax.
For trabeculae, cTn complexes were reconstituted by incubation with 100% xcTnC (1 mg ml–1) in pCa 9.2 solution without creatine kinase (CK) or dextran for 15–20 min, with no discernable change in Fmax from the extracted value. Reconstitution of skeletal Tn (sTn) complexes was achieved by 1–3 min incubations in 1 mg ml–1 (total) sTnC in solution at pCa 9.2 as previously described (Regnier et al. 2002; Moreno-Gonzalez et al. 2005). For skeletal fibres, mixtures of 10–15% sTnC and 85–90% D28A, D63A sTnC (xxsTnC) were used to produce 20% of the pre-extracted Fmax for comparisons with trabeculae. Reconstitution was considered complete when force at pCa 4.0 no longer increased with subsequent incubations. We have demonstrated relatively equal binding affinities for sTnC and xxsTnC in the absence of Ca2+. Thus, the procedure for reconstitution with sTnC:xxsTnC mixtures is expected to yield a random distribution of these two TnCs along individual thin filaments throughout the entire fibre diameter (Regnier et al. 2002).
Gel and Western blot analysis
Sarcomere protein stoicheometry of trabeculae prior to and following whole Tn exchange was determined from silver-stained sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE; Fig. 2) according to the methods of Giulian et al. (1983). Western blots were used to determine the phosphorylaton status of bovine cTnI encorporated into cTn complexes and in rat trabeculae prior to exchange (Fig. 3). For cTn complexes containing WT and D65A cTnC, protein kinase A (PKA) and C (PKC) phosphorylation were performed in a typical reaction mixture containing (mM): Hepes 50 (pH 7.5), MgCl2 10, CaCl2 0.5, EGTA 1 and ATP (Tris-salt) 0.1, and 4 µg protein and 3 units of PKA/or 0.13 µg of PKC (Sigma, St Louis, MO, USA). Protein phosphatase 1 (PP1; Sigma) was used to release phosphate groups from phosphorylated serine, threonine and tyrosine residues in buffer containing (mM): Hepes 50, Na2EDTA 0.1, DTT 5 and MnCl2 10, and 0.025% Tween 20 at 30°C for 6–12 h. For rat trabecular, as previously described (Dai et al. 2002), samples were homogenized in a solution of 0.1% Triton X-100, 150 mM NaCl, 10 mM Hepes (pH 7.5), 1 mM EDTA, 0.5 mM 4-(2-aminoethyl)benzenesulphonyl fluoride (AEBSF), 1 µg ml–1 leupeptin, 1 µg ml–1 aprotinin, 10 µg ml–1 soybean trypsin inhibitor and 1 µg ml–1 pepstatin A. The homogenate was incubated on ice for 5 min and then centrifuged at 2000 g for 5 min. The extracts were incubated with goat polyclonal anti-troponin I (Santa Cruz Biotechnology) at 4°C for 1 h, followed by incubation with protein A–sepharose (5 ng ml–1) at 4°C for 30 min, then washed three times with ice-cold lysis solution (see above). All samples were separated on a 12% polyacrylamide gel and transferred to nitrocellulose membranes. The membranes were incubated in Tris-buffered saline (TBS) containing 0.05% Tween 20 and 5% blocking powder for 18 h at 4°C. The membranes were then incubated in a 1: 500 dilution of goat polyclonal anti-phosphoserine (Invitrogen, CA, USA) in TBS with 0.05% Tween 20 and incubated in a 1: 1000 dilution of horseradish peroxidase-conjugated bovine anti-goat IgG (ICN Biochemicals). Detection was by chemiluminescence, and densitometry was performed on the films.
|
|
Solutions contained (mM): phosphocreatine 15, EGTA 15, Mops at least 40, free Mg2+ 1, Na+ + K+ 135 and DTT 1, 250 units ml–1 creatine kinase (Sigma) and 5 mM ATP, 0.5 mM ATP, or 5 mM 2 deoxy-ATP (dATP; Sigma) at pH 7.0 and 15 ± 1°C. Ionic strength was 0.17 M. Affinities of dATP and ATP for Mg2+ were assumed to be similar (Regnier et al. 1998a). For activation solutions, the Ca2+ level (expressed as pCa) was varied between pCa 9.0 and pCa 4.0 by adjusting the concentration of calcium propionate.
Mechanical measurements
The ends of trabeculae or rabbit psoas muscle fibres were wrapped with aluminium foil T-clips for attachment to a force transducer (either Model 400A, 2.2 kHz resonant frequency; Cambridge Technology, Watertown, MA, USA or Model AE801, 5 kHz resonant frequency; SensoNor, Horton, Norway) and a servomotor (model 300, Cambridge Technology) tuned for a 300 µs step response. Skeletal fibre segment ends were chemically fixed by focal application of 1% glutaraldehyde in H2O to minimize compliance (Chase & Kushmerick, 1988). Sarcomere length (Ls) was measured with helium–neon laser diffraction at pCa 9.0. The following measurements were made: steady-state isometric force, the rate of isometric tension redevelopment (ktr) following rapid (< 4 ML s–1) release–restretch (15%
Ls) of the preparation (Brenner, 1986) and stiffness determined by small amplitude (0.1%
Ls) sinusoidal length oscillations (1000 Hz).
Analysis
Force–pCa and ktr–pCa relationships are fitted with the Hill equation:
|
| (1) |
Ca2+ dissociation rates from whole skeletal and cardiac troponin
A model SX.18 MV stopped-flow instrument (Applied Photophysics Ltd, Leatherhead, UK) was used to measure the Ca2+ dissociation rates (koff) from whole Tn complexes using a method modified from that of Tikunova et al. 2002). This method uses Quin-2 (Calbiochem) as a fluorescent chelator. Whole Tn (WT skeletal and WT cardiac) was dialysed into a soltion containing (mM): KCl 250, Mops 20, MgCl2 5 and DTT 1; pH 7.0. The above buffer was used in all stopped-flow experiments. Each complex (6 µM) with 30 µM CaCl2 was rapidly mixed with an equal volume of Quin-2 (150 µM) at 15°C. The samples were excited using a 150 W xenon arc source while emission was monitored through a 510 nm broad band-pass interference filter (Oriel, Stanford, CT, USA). As Quin-2 reports the dissociation of Ca2+ from both the N- and C-terminus, a series of reactions with different durations of time (50 ms–20 s) were accumulated for each protein. These were then fitted with either single or double exponentials as appropriate. Once a time scale had been determined over which the fast N-terminal koff went to completion and allowed for the best fit of the data (determined by the residuals), this was used as the time scale for single exponential fits. The koff values presented for the N-terminus of each protein were calculated by summing and fitting the data from three reactions and then repeating this at least four more times with 12 other reactions. The rate presented for the C-terminus represents the second rate from a double exponential fit of data collected over a 20 s period.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Figure 1A shows an example force record demonstrating the ability to exchange native Tn in a trabecula for complexed Tn containing recombinant rat cTnC. This trabecula was first activated with a pCa 4.5 solution to determine maximal Ca2+-activated steady-state force (Fmax) and rate of force redevelopment (ktr) following the release–restretch protocol (see Methods). The complete loss of force with a 15% Ls release and its redevelopment upon restretch are indicated by the vertical lines in the force record every 5 s. Following maximal activation, the trabecula was relaxed (pCa 9.0) prior to incubation in exchange buffer with 4.2 mg ml–1 xcTn for 2 h (first arrow in Fig. 1A). After return to relaxing solution, Fmax was determined to be 10% of the pre-exchanged level. In this example preparation, we then did a second exchange with whole Tn containing 100% WT cTnC (cTn; second arrow in Fig. 1A), which resulted in recovery of Fmax to 77% of the original pre-exchanged value. Absolute values for Fmax (normalized to cross-sectional area) and ktr are given in the legend to Fig. 1. The exchange protocol was performed periodically with 100% xcTn to verify that complete or near complete replacement of the native cTn was occurring consistently, and a second exchange with 100% cTn was performed a few times to verify that loss of force was due to xcTn (and not preparation degradation). The control experiments with 100% xcTn usually showed more complete loss of Fmax (5.5 ± 1.2%, n = 7, range, 2.1–10.0%). Single exchanges with 100% cTn generally maintained Fmax at a greater level (86 ± 2% of pre-exchanged Fmax, n = 4) than shown in Fig. 1A. This is demonstrated in Fig. 1B where force records of the normal experimental protocol are shown for an example 100% cTn exchange. This trabecula was placed in a series of solutions containing increasing Ca2+ concentration (indicated as pCa values below the force trace), both before and following the exchange protocol. For the example in Fig. 1B, exchange with 100% cTn maintained 86% of pre-exchanged Fmax and pCa50 was 5.38, compared with 5.52 prior to the exchange. Figure 1C shows force records for an exchange mixture containing 50% cTn and 50% xcTn and demonstrates a 75% reduction in Fmax and pCa50 of 5.15, compared with 5.38 prior to the exchange.
|
The steady-state force–pCa relationship was measured for different ratio mixtures of cTn:xcTn in the exchange solution. Butters et al. (1997) demonstrated that the affinity of cTn and xcTn for thin filaments was similar in the low [Ca2+] exchange solutions, as determined using the method of Huynh et al. (1996). This suggests that the amount of each Tn exchanged into the thin filament of the trabeculae was similar to the proportions in the exchange solutions. The data are summarized in Figs 2 and 3 and the Hill fit parameters Fmax, pCa50 and apparent cooperativity of activation (nH) are summarized in Table 1. Data are normalized to Fmax following exchange with 100% cTn for ease of comparison. Figure 4 shows that decreasing the fraction of functional cTn in the exchange mixture reduced Fmax in a more than proportional manner. The concave (up) nature of the Fmax versus fraction of function cTn relationship is similar to that found by Butters et al. (1997) for the relationship myosin S1-thin filament MgATPase versus fraction of function cTn. Our data regarding the organized structure of the sarcomere supports their conclusions from protein solution studies that positive cooperativity between RUs must occur to achieve the activation state found with saturating levels of Ca2+. It also implies that < 7 actins may become available for strong cross-bridge formation with Ca2+ binding an individual cTn. These data contrast with our findings in rabbit (psoas) skeletal muscle fibres (Regnier et al. 2002) where Fmax was reduced by a less than proportional manner with decreasing functional Tn. The best fit line for the skeletal muscle fibre data (Fig. 4 of Regnier et al. 2002) is redrawn in Fig. 4 for comparison (dotted line). From these experiments we concluded that 10–12 actins were activated by Ca2+ binding the Tn of each FU. This comparison demonstrates a distinct difference between cardiac and skeletal muscle in the extent that activation spreads along thin filaments with Ca2+ binding.
|
|
0.15 of pre-extracted Fmax), a 0.45 pCa unit shift to decreased Ca2+ sensitivity and significant reduction of nH (1.7 versus 3.8 for control). This comparison demonstrates that the loss of near-neighbour FUs has much less effect on the pCa50 and nH of force development in cardiac versus skeletal muscle.
|
11–24% remaining force (with saturating [Ca2+]) following the extraction procedure. To determine how our results with the whole Tn exchange protocol compare with this method, the TnC extraction protocol (see Methods) was used on a small group of trabeculae from normal rats (containing mostly 
-myosin) or PTU-treated rats (containing 
-myosin), followed by incubation with 100% xcTnC. The results are summarized in Table 2. For these trabeculae the reconstituted Fmax was 32 ± 3% and 20 ± 2% of pre-extracted values for normal and PTU-treated rats, respectively, comparing most closely to the values obtained for cTn exchanges with 50: 50 and 25: 75 mixtures of cTn:xcTn (Table 1), respectively. The decrease in Ca2+ sensitivity was 0.3 pCa units for each group of trabeculae, and nH was not significantly reduced (similar to the results of the cTn exchange protocol). However, in trabeculae reconstituted with 100% purified rat cTnC, we find an approximate 0.1–0.15 pCa unit decrease in Ca2+ sensitivity with similar or slightly reduced nH compared with pre-extracted values (data not shown), suggesting that changes in these parameters are similar to those using the whole cTn exchange protocol. Thus it appears that either method of reducing the level of functional cTn has little or no effect on the apparent cooperativity (nH) of cardiac thin filament activation and force development.
|
50% with addition of 10 mM BDM to the pCa 4.0 activation solution under control (pre-extraction) conditions. Following the reduction of FUs to obtain 0.20 Fmax the relative level of inhibition by BDM was not changed. Similar experiments were done in rabbit psoas skeletal fibres for comparison (Fig. 6). Under control conditions 10 mM BDM reduced Fmax by 50%, as in cardiac muscle. However, in contrast to cardiac muscle, the relative level of force inhibition following FU reduction to produce 0.2 Fmax was much greater (85% reduction). Figure 7A shows the converse experiment, where augmenting the number of strong cross-bridges (via replacement of ATP with dATP) increased Fmax both in pre-extraction (control) conditions and following the exchange protocol to produce 0.2 Fmax with isolated FUs in cardiac trabeculae (Regnier et al. 2002, 2004). By contrast, when this experiment was repeated using either rabbit psoas (fast) or soleus (slow) skeletal muscle, dATP did not increase Fmax in control measurements but increased Fmax by 40% when FUs were isolated in fast skeletal muscle (Fig. 7B). Together these experiments with BDM and dATP demonstrate that maximal activation of skeletal muscle fibres becomes more dependent on strong cross-bridge binding when the number of FUs is greatly reduced, but there is no increased dependence of maximal activation on strong cross-bridge formation in cardiac muscle with reduced numbers of FUs. The combined data suggest that near-neighbour FU interactions are more important for activation in skeletal (versus cardiac) muscle, and that the predominant form of apparent cooperativity in thin filament activation probably differs between the two striated muscle types.
|
|
To determine how loss of near-neighbour RU interactions affects the maximal rate of force development, ktr was measured in solution of pCa 4.0 in all trabeculae for each of the different Tn mixtures. The data, summarized in Fig. 8A, clearly demonstrate that maximal ktr is not dependent on either the number of FUs or the absolute level of isometric force production in cardiac muscle. We (Moreno-Gonzalez et al. 2003, 2007) and others (Morris et al. 2001) have demonstrated that this is also the case for skeletal muscle fibres reconstituted with mixtures of sTnC:xxsTnC or cTnC:xcTnC, respectively.
|
Ca2+ dissociation rates (koff) from cTn and sTn
So why would cardiac muscle thin filament activation be more dependent on strong cross-bridge binding than skeletal muscle? One possible reason could be a less effective transmission of the Ca2+ signal for contractile activation, perhaps due to the different isoforms of Tn and/or Tm. For example, if the time Ca2+ stays bound to Tn is shorter in cardiac than in skeletal muscle, it may decrease the probability of strong cross-bridge binding in cardiac muscle. Therefore, to characterize kinetic differences between the Ca2+-dependent activation of cardiac Tn and fast skeletal Tn, the koff values of these two protein complexes were measured. The results illustrate that the rate of Ca2+ dissociation from the N-terminus of cTn (66.0 ± 1.0 s–1) is 2.5-fold greater than that of sTn (25.8 ± 0.1 s–1) at 15°C (Fig. 9). Ca2+ dissociation from the C-terminus was at least two orders of magnitude slower for cTn and sTn (0.67 ± 0.01 s–1 and 0.15 ± 0.01 s–1, respectively) and did not influence the N-terminal measurements. If it is assumed that Ca2+ binding affinity of the TnC Ca2+ trigger sites (N-terminus) is determined primarily by koff (Johnson et al. 1994), our results support a hypothesis that shorter interaction time of cTnC with cTnI, compared with the interaction of sTnC with sTnI, could effectively provide a lower level of thin filament activation at any given [Ca2+].
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Our interpretation of the results in this study relies on the assumption that the recombinant cTn exchanged into the trabeculae binds along the entire length of the thin filament and that the binding affinity of cTn and xcTn complexes are similar. We have previously demonstrated through colocalization studies of phalloidin green-labelled actin and rhodamine-labelled exchanged into trabeculae that there is uniform distribution of the exchanged xcTn along the thin filament (Kohler et al. 2003). Comparison of the fluorescence intensity scans of the labelled xcTn along the trabeculae with that of actin illustrated that the two labelled molecules are present in the same location throughout the preparation. Additionally, xcTn and cTn bind to the thin filament with similar affinity in low Ca2+
solutions, as determined by Butters et al. (1997). These results, together with our demonstration that exchange with xcTn can almost eliminate (up to 98%) Ca2+-activated force production, suggest that cTn is uniformly and effectively replaced throughout the thin filament. Another potential concern is that cTn complexes used for exchange were composed of cTnI and cTnT purified from bovine heart and recombinant rat cTnC (WT or xcTnC). The amino acid sequence of rat cTnI is 89% similar to bovine cTnI and rat cTnT is 87% similar to bovine cTnT. The use of ‘chimeric’ cTn in this replacement could result in a difference in how the subunits interact (compared with all rat proteins) as a result of the sequence differences. However, the control measurements were made using 100% chimeric cTn (or xcTn). There was a small decrease in the Ca2+ sensitivity of force following exchange of native cTn with WT cTnC-cTn. This decrease does not appear to result from non-specific extraction of other sarcomere contractile proteins, as demonstrated in Fig. 2. It could result from a different phosphorylation state of the exchanged WT cTnC-cTn, as Fig. 3 shows the bovine cTnI used was phosphorylated. However, Fig. 3 also demonstrates cTnI phosphorylation of the native rat cTnI. Thus, we believe our results were not due to loss of sarcomeric proteins and are at least relatively independent of phosphorylation status.
Cooperative thin filament activation in cardiac versus skeletal muscle
A primary purpose of the present study was to compare experiments in cardiac muscle with similar data we have obtained in fast skeletal muscle here and in previous work (Regnier et al. 2002; Moreno-Gonzalez et al. 2005). The results of the current study suggest that while cardiac and skeletal muscle myofilament structure is similar and they contain the same compliment of thin and thick filament proteins, the underlying mechanisms that regulate thin filament activation and steady-state force development in chemically demembranated preparations differ. Here we characterized the role of individual FUs compared with near-neighbour FU interactions along the thin filament during contractile activation by replacing endogenous Tn with different mixtures of ‘functional’ and ‘non-functional’ Tn to increase isolation of FUs in cardiac trabeculae. In skeletal muscle, we were able to accomplish this by extraction of endogenous sTnC and replacement with mixtures of ‘functional’ and ‘non-functional’ (D28A and D63A) sTnC. Interestingly, our Fmax versus the fraction of functional cTn data suggest that the size of the FU in cardiac muscle (7 actins) is less than that found with similar measurements in skeletal muscle (10–12 actins; Regnier et al. 2002) when Ca2+ binds to an individual Tn (Fig. 4). This result implies that the activation of a single Tn complex by Ca2+ binding in cardiac muscle ‘turns on’ a smaller length of the thin filament than in skeletal muscle. A potential implication of this is that smaller FU size may result in less near-neighbour FU interaction in cardiac (versus skeletal) muscle and, consequently, a greater reliance on local (within each FU) cooperative feedback mechanisms for contractile activation.
The definition of FU size that we have presented here is the length of the thin filament (i.e. the number of actins) activated by Ca2+ binding to a single Tn. However, if the FU is defined as the conditions that are required to most completely activate an RU (A7TmTn), then it may be that an FU contains two or three RUs if Ca2+ binding to neighbouring Tn complexes can more completely activate an RU. This conclusion was reached from the solution studies of Butters et al. (1997), who used similar methods to vary the fraction of FUs in reconstituted cardiac thin filaments. They found a less than proportional increase in maximal Ca2+-activated cardiac S1-thin filament ATPase as the fraction of functional RUs was increased, with the shape of the curve being quite similar to what we found for Fmax versus the fraction of functional Tn (Fig. 4). Their interpretation of the data was that individual RUs were fully activated only when neighbouring RUs are simultaneously activated by Ca2+. As such, both solution and cellular mechanical data indicate that the FU size (with Ca2+ binding to a single Tn) is smaller in cardiac muscle. Additionally, both definitions of FU size could explain the moderate shift in the force–pCa curve (Table 1) as cTn content in trabeculae is increased. However, whether Ca2+ binding to neighbouring Tn complexes on cardiac thin filaments contributes significantly to cooperative activation remains an open question. This is because, if we achieved isolation of individual RUs (i.e. each functional Tn is surrounded by non-functional Tn on each side) it is more difficult to determine how loss of near-neighbour RU interactions has little influence on nH unless a primary form of cooperativity in thin filament activation resides within the individual RU. Alternatively, if Ca2+-dependent activation of an RU requires activation of neighbouring RUs, these completely isolated RUs may not be functional. It may be that the only RUs that can produce force are ones with neighbouring RUs that can bind Ca2+ and activate, as suggested by Butters et al. (1997). If this is the case, nH may be less dependent on the fraction of functional cTn in thin filaments. The present work does not allow us to confidently distinguish between these two possibilities.
Additional support for the hypothesis of minimal interaction between FUs comes from the force–pCa relationship. In skeletal muscle we found that isolating FUs greatly reduced nH and pCa50 of the force–pCa relationship (Regnier et al. 2002), suggesting that spread of activation beyond the boundaries of an RU with Ca2+ binding into near-neighbour RUs, appear to be the primary determinant of cooperative activation in skeletal muscle. In agreement with our study, Moss et al. (1985) showed that partial extraction of sTnC resulted in a reduced Fmax and a right-shift of the relative force–pCa relationship. These authors suggested that the Ca2+ sensitivity of an RU within the thin filament may vary with the state of activation of adjacent RUs. Our finding that isolation of FUs in cardiac muscle has little effect on nH (Table 1) indicates that the steepness of the force–pCa curve probably involves cooperative mechanisms within individual RUs and that interactions between neighbouring RUs has less influence (at least at submaximal levels of Ca2+) in cardiac muscle. The decrease in pCa50 was also much less than in skeletal muscle, but may reflect at least some near-neighbour interactions. Additionally, in experiments where native cTnC was extracted until the remaining Fmax was 20–30% of pre-extracted values, we found similar results to when the whole cTn exchange protocol was used. Whether or not the cTnC of these preparations was randomly extracted from thin filaments, it supports our conclusion that reducing near-neighbour FU interactions has little effect on nH in cardiac muscle. Therefore, our current results in cardiac muscle illustrate a difference in the mechanisms that determine the highly cooperative Ca2+ dependence of force development in the two muscle types (see below).
Cooperative events within individual FUs
In a recent study we (Moreno-Gonzalez et al. 2005) found that when skeletal fibre near-neighbour FU interactions are reduced, the level of thin filament activation at lower [Ca2+] is determined primarily by the direct effects of Ca2+ on regulatory protein mobility, while at higher [Ca2+] the final level of thin filament activation is primarily determined by strong cycling cross-bridges. Work by Fitzsimons et al. (2001) also supports the role of strong cycling cross-bridges in fully activating the skeletal thin filament. In cardiac muscle, the cross-bridge-dependent component controlling the force–pCa curve may predominate under normal conditions as well as when near-neighbour FU interactions are minimized. This would explain the relative insensitivity of cardiac pCa50 to loss of near-neighbour FU interactions (0.1–0.2 pCa units; Table 1) compared with skeletal muscle (0.5 pCa units; Regnier et al. 2002). It could also explain the interesting difference between these two muscle types, namely that loss of near-neighbour FU interactions greatly reduced nH in skeletal (Regnier et al. 2002) but not cardiac muscle (Table 1). As mentioned above, this suggests that cooperative activation mechanisms are controlled locally in cardiac muscle, at the level of individual FUs.
One ‘local’ form of positive cooperative events is strong crossbridge feedback to increase Ca2+ binding to Tn. Both 45Ca2+ binding (Fuchs & Wang, 1991; Wang et al. 2001) and TnC structural changes associated with increased TnC–TnI interaction (Martyn et al. 1999, 2001) are enhanced by strong cross-bridges in cardiac, but not skeletal muscle. As such, cross-bridge-induced increases in cTn Ca2+ binding may be a significant mechanism of cooperative activation in cardiac muscle (see below).
A potential consequence of smaller FU size may be that Ca2+ binding to cTn is less effective in activating cardiac thin filaments. The formation of strong cross-bridges in the cardiac thin filament may act to stabilize the interaction of cTnC with cTnI and therefore might be required to fully activate the thin filaments (or maintain activation). Several lines of evidence support this hypothesis. Using preparations with isolated FUs, we have shown that both inhibition (Fig. 6) and augmentation (Fig. 7) of strong cross-bridge binding greatly influences Fmax in skeletal, but not cardiac muscle. BDM, which inhibits cross-bridge transition from weak to force-generating states (Regnier et al. 1995) had a much larger effect on Fmax of rabbit psoas skeletal muscle fibres after reconstitution with only small fractions of functional sTnC. We concluded that this was due to the loss of near-neighbour FU interactions (spread of activation along thin filaments) in psoas fibres. Conversely, strong cross-bridge inhibition by BDM had no greater (relative) effect on Fmax with the loss of near-neighbour FU interactions in rat trabeculae (Fig. 6), suggesting that Ca2+-mediated strong cross-bridge binding and force development are controlled mainly at the level of individual FUs. Additionally, in the present study (Fig. 7) and in previous work, we have provided evidence that thin filaments of normal cardiac muscle may not be completely activated even with saturating levels of Ca2+. This is suggested by the fact that augmenting strong cross-bridge binding (as determined from stiffness measurements) with dATP substantially increases both Fmax and maximal ktr (Regnier et al. 2000, 2004) in cardiac muscle. Indeed, the increase in Fmax for control trabeculae is similar to that seen in preparations with isolated FUs (Fig. 8A). This is not due to a dATP-induced increase in the rate of cross-bridge cycling per se, because the enhancement of steady-state force and force development kinetics was independent of cardiac myosin isoform ( versus myosin heavy chain). Also shown in Fig. 7, the augmentation is particular to cardiac muscle, as dATP does not increase Fmax of either fast or slow twitch rabbit skeletal muscle fibres. As soleus muscle contains cardiac TnC, it suggests that the augmentation of Fmax in cardiac muscle is not due specifically to TnC isoform either, but could be due to TnC interaction with other Tn subunits (see below). It is interesting that in skeletal muscle fibres with isolated FUs, dATP increased Fmax by approximately 40% (Fig. 7B), similar to what is seen in cardiac muscle. Taken together the data suggest that strong interaction between near-neighbour RUs and/or the extent of spread in activation along thin filaments are/is an important determinant of the amount of cross-bridge binding and force generation, and that this is more limited in cardiac (versus skeletal) muscle.
Maximal and submaximal ktr are determined by properties of individual FUs
Similar to steady-state force, ktr appears to be controlled at the level of individual FUs in cardiac muscle. We found no influence of near-neighbour FU interactions on maximal ktr (Fig. 8A). Similarly, we recently reported that loss of near-neighbour FU interactions has minimal effect on maximal ktr in skeletal fibres (Moreno-Gonzalez et al. 2003, 2007). Additionally, Morris et al. (2001) found similar results in skeletal muscle fibres when endogenous TnC was extracted and Tn complexes were reconstituted with varying mixtures of functional and non-functional cardiac TnC. We also found that decreasing the number of FUs has little or no effect on the Ca2+ dependence of ktr in cardiac muscle (Fig. 8B). We previously reported minimal effect of decreasing FUs on the Ca2+ dependence of ktr in skeletal fibres (Moreno-Gonzalez et al. 2003, 2007). However, when native TnC of skeletal muscle fibres was replaced with cTnC or an sTnC mutant, both of which had lower Ca2+ binding affinity, maximal ktr and the Ca2+ dependence of ktr were greatly affected independent of force. Thus, it appears that ktr is controlled by the properties of regulatory proteins and by mechanisms that work at the level of individual FUs in both cardiac and skeletal muscle.
The role of troponin in cooperative activation
One possible explanation for why strong cross-bridges enhance Ca2+ binding to cTn in cardiac muscle is that they may strengthen the interaction between cTnI and the N-terminus of cTnC (cNTnC) during activation. Solution NMR studies by Li et al. (1999) demonstrated that the strength of interaction between the hydrophobic patch of cNTnC (exposed upon Ca2+-dependent activation) and the cTnI peptide 147–163 is six times less than that between sNTnC and the corresponding sTnI peptide. It is this interaction that is thought to pull cTnI away from an inhibitory position on the actin monomers of the thin filament. A likely reason for this lower strength of interaction between cardiac proteins is that the size of the hydrophobic patch exposed on cNTnC with Ca2+-dependent activation is approximately one-third the size of that exposed on sNTnC (162 Å2 versus 500 Å2) (Gagne et al. 1995; Li et al. 1999). There are two possible consequences of the lower strength of interaction between cTnC and cTnI during Ca2+-dependent activation. The first is that it is harder for cTnC to ‘pull’ cTnI away from interacting with actin and assume the ‘activated’ conformation. The second is that it may be easier for activated cTn to become deactivated by the cTnI re-establishing a strong interaction with actin. There is also a measurable delay between Ca2+ binding or dissociation and the associated confirmation changes in isolated cTnC or as part of the Tn complex, as measured by extrinsic fluorescent probes in solution (Dong & Cheung, 1996; Dong et al. 1997; Hazard et al. 1998). Furthermore, the conformational changes for saturated Ca2+ binding are not complete, indicating that not all cTnC is in the activated conformation (Dong & Cheung, 1996; Dong et al. 1997). This could have great significance in terms of thin filament activation in cardiac muscle.
In the current study, the koff of Ca2+ from the N-terminus of cTn was 2.5-fold faster than that from sTn. This measurement represents Ca2+ dissociation from cTnC site II while the protein complex is in the activated state. As a result, the protein complex would become deactivated as the hydrophobic patch on the N-terminus of cTnC loses contact with the switch region of cTnI (Li et al. 2004). The difference in koff between cTn and sTn suggests that, under identical conditions, cTn may deactivate at a faster rate and would therefore be in the activated state for a shorter period of time than sTn. This could be due, at least in part, to site II of cTnC having a lower Ca2+ affinity than the N-terminus of sTnC (sites I and II). An additional factor could be a lower affinity of cNTnC for the C-terminus of cTnI during activation. The koff data (Fig. 9) suggest there may be fundamental differences in the functional characteristics of sTn and cTn, such that the kinetics of Tn subunit interactions can influence initial strong cross-bridge binding and subsequent cTn-mediated thin filament activation.
The proposed interdependence of Ca2+ and cross-bridge binding in thin filament activation is illustrated in Scheme 1, which describes the relative binding affinity of cTnI for cTnC versus actin, in the absence and presence of Ca2+ and cross-bridges. In the absence of Ca2+ and cross-bridges, the strength of cTnI interaction with cNTnC is weak whereas cTnI interaction with actin monomers is strong (*). When Ca2+ binds to cTnC, the interaction with cTnI strengthens and cTnI interaction with actin weakens (**). Subsequent strong cross-bridges further weaken cTnI interaction with actin and stabilize the interaction with cTnC, as demonstrated by Robinson et al. (2004).
The role of Tm in cooperative activation
Above, we discussed possible mechanistic differences between cardiac and skeletal muscle Tn in cooperative activation of thin filaments and force generation. These mechanistic differences are likely to occur ‘locally’ (i.e. within the confines of an FU or between two neighbouring RUs). However, the biophysical nature of Tm may be an important determinant of the size of FUs. In other words, one significant mechanistic difference between cardiac and skeletal muscle may be the role that Tm plays in transmitting the activation signal along the thin filament via end-to-end Tm interactions in neighbouring RUs. Chandy et al. (1999) demonstrated that cardiac Tm is more mobile than skeletal Tm. This greater mobility could result in greater hysteresis of cardiac Tm movement across the surface of actin with Ca2+-dependent activation, compared with skeletal Tm. A ‘stiffer’ Tm could facilitate exposure of additional myosin binding sites beyond the boundaries of a Ca2+-activated RU, into neighbouring RUs along the thin filament. In cardiac muscle the more flexible or ‘floppy’ Tm would limit the number of actins made available for myosin binding with the activation of a single FU. The importance of strong cross-bridge binding and facilitating exposure of a maximal number of myosin binding sites may therefore be more critical in allowing cardiac (versus skeletal) muscle to be fully activated. Studies by Lehman et al. (2000) using electron microscopy and image reconstruction have demonstrated that cardiac Tm is in a more inhibitory position on the thin filament than skeletal Tm at low [Ca2+]. This implies that cardiac Tm would have to move further to expose the same number of myosin binding sites during the Ca2+-dependent activation compared with skeletal Tm. Additionally, if the Ca2+–Tn interaction time is shorter in cardiac muscle (as indicated from a faster koff; Fig. 9), TnC–TnI interaction time may be less as well, potentially reducing the time that Tm is mobile. This could result in fewer cross-bridge binding sites exposed and/or a reduced time of exposure at the same level of Ca2+. Finally, if TnC–TnI interaction is weaker/shorter in cardiac muscle, the resulting stronger TnI–actin interaction could further reduce Tm mobility and/or the time that Tm is mobile.
Significance
An important consequence of cardiac thin filaments not being fully activated by Ca2+ alone may be that the level of activation can be more affected by increases/decreases in strong cross-bridge binding. In other words, Ca2+ binding to cTn enables initial cross-bridge binding, but the activation level is determined and maintained more directly by the number of strong cross-bridges that are bound. This would provide cardiac muscle with a way to finely control activation and the level of force generation in individual sarcomeres. Such fine control at the cellular (sarcomere) level is important because all cardiomyocytes are activated during systole, in contrast to skeletal muscle where a main mechanism of varying force is via the number of fibres recruited. The greater dependence of cardiac muscle activation on strong cross-bridges may also explain the steeper length dependence of force in cardiac versus skeletal muscle and provide a mechanism for rapid deactivation at the end of systole at shorter sarcomere lengths.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Brenner B (1986). The cross-bridge cycle in muscle. Mechanical, biochemical, and structural studies on single skinned rabbit psoas fibers to characterize cross-bridge kinetics in muscle for correlation with the actomyosin-ATPase in solution. Basic Res Cardiol 81 (Suppl. 1), 1–15.[CrossRef][Medline]
Butters CA, Tobacman JB & Tobacman LS (1997). Cooperative effect of Ca2+ binding to adjacent troponin molecules on the thin filament-myosin subfragment 1 MgATPase rate. J Biol Chem 272, 13196–13202.
Chandy IK, Lo JC & Ludescher RD (1999). Differential mobility of skeletal and cardiac tropomyosin on the surface of F-actin. Biochemistry 38, 9286–9294.[CrossRef][Medline]
Chase PB & Kushmerick MJ (1988). Effects of pH on contraction of rabbit fast and slow skeletal muscle fibers. Biophys J 53, 935–946.
Dai Y, Iwata K, Fukuoka T, Kondo E, Tokunaga A, Yamanaka H, Tachibana T, Liu Y & Noguchi K (2002). Phosphorylation of extracellular signal-regulated kinase in primary afferent neurons by noxious stimuli and its involvement in peripheral sensitization. J Neurosci 22, 7737–7745.
Dong WJ & Cheung HC (1996). Ca2+-induced conformational change in cardiac troponin C studied by fluorescence probes attached to Cys-84. Biochim Biophys Acta 1295, 139–146.[CrossRef][Medline]
Dong W, Rosenfeld SS, Wang CK, Gordon AM & Cheung HC (1996). Kinetic studies of Ca2+ binding to the regulatory site of troponin C from cardiac muscle. J Biol Chem 271, 688–694.
Dong WJ, Wang CK, Gordon AM, Rosenfeld SS & Cheung HC (1997). A kinetic model for the binding of Ca2+ to the regulatory site of troponin from cardiac muscle. J Biol Chem 272, 19229–19235.
Farah CS & Reinach FC (1995). The troponin complex and regulation of muscle contraction. FASEB J 9, 755–767.[Abstract]
) 
); 50: 50 cTn:xcTn (
); 75: 25, cTn:xcTn (
); and 100: 0, cTn:xcTn (
) are added to the figure for comparison in A.