fsTnC replacement by cTnC: comparison with previous work and implications
Replacement of fsTnC by cTnC has been used to examine the regulation of contraction, but most studies have used partial extraction of fsTnC because exhaustive extraction reduces the maximal isometric force (Moss et al. 1986, 1991; Metzger, 1996). In partial extractions (< 70 % TnC removed), replacement with cTnC produced a variable reduction of maximal isometric force, a reduction of cooperativity (nH fell by 1 unit) and a slight shift in the pCa50 (-0.1 pCa unit; Moss et al. 1986, 1991; Babu et al. 1987). That cTnC replacement produces a reduced cooperativity is paralleled by experiments in transgenic mice expressing fsTnC in their cardiac muscle. Compared to control cardiac myocytes, myocytes from transgenic mice expressing fsTnC/cTnI/cTnT exhibited an increased cooperativity (McDonald et al. 1995).
Recent work in skinned skeletal muscle fibres (Morris et al. 2001) from which > 90 % of the fsTnC was extracted and replaced with cTnC, confirmed these specific observations and showed that both isometric force and ktr at saturating [Ca2+] fell to 65 and 60 %, respectively, of their control values (see also Chase et al. 1994). Identical extraction of fsTnC and replacement with exogenous fsTnC restored isometric force to ≈ 90 % of the control (unextracted) value and ktr to 100 % percent of the control values. Extensive extraction of fsTnC (> 90 %) and its replacement with exogenous fsTnC or cTnC in single myofibrils in the current work produced results similar to those obtained in fibres (Morris et al. 2001). These TnC-extracted myofibrils exhibited no fsTnC (as assessed by SDS-PAGE; see Fig. 1A). After cTnC replacement, maximal isometric tension fell to 55-60 % of the control values, while ktr and kact were both reduced to 45 % (see Table 1). Replacement of myofibrils with the homologous fsTnC restored all parameters to their pre-extraction values. These results imply that the heterologous Tn created by extraction and reconstitution of TnC (mixing of the cardiac and fast skeletal Tn subunits) compromises myofibril activation.
These results appear to be inconsistent with those of Moss et al. (1991) and Metzger (1996), who found that less extensive removal of fsTnC (30-70 % removal) and replacement with cTnC produced no change in maximal isometric force. The loss of function in the more extensive extraction procedures described here is not a consequence of the extraction procedure itself because homologous replacement of TnC restores control values of force, ktr and kact. It could be hypothesized that the significant loss of function in cTnC replacement seen here and in extensively extracted fibres (Chase et al. 1994; Morris et al. 2001, 2003) is a consequence of a failure of the exogenous cTnC to bind to the sites lacking TnC in the extracted fibres. The present results and several observations argue against this hypothesis. First, quantitative SDS-PAGE analysis, although unable to identify the location of specific TnC isoforms in myofibrils, indicates a small surfeit of TnC after reconstitution, not a deficit, regardless of whether cTnC or fsTnC was used (see the first section of Results and Fig. 1A). Second, and more significantly, the diminished kact and ktr observed after cTnC exchange cannot be explained by the presence of regulatory units that are inhibited because they lack TnC. As shown previously in fibres (Metzger & Moss, 1990, 1991), when fsTnC is partially removed from the myofibril, the isometric force at saturating pCa is reduced, but kact and ktr are unchanged (see Fig. 3). In the experiments reported here, isometric force at saturating pCa was reduced from control values when cTnC replaced the missing fsTnC, but the ktr was also reduced (not unchanged as in the case of fsTnC removal). Similarly, in the presence of saturating levels of Ca2+, Morris et al. (2001) have shown in fibres that ktr remains maximal despite 60 % replacement of fsTnC with an inhibitory, mutant cTnC that does not bind Ca2+ at site II. Third, Moss (1986) has shown that partial removal of fsTnC (to a point at which the maximal force is only 50 % of control values) produces two phases of shortening (a high velocity and low velocity phase) at saturating pCa in slack tests. However, when 90 % of the fsTnC has been replaced by cTnC, the unloaded shortening velocity at saturating pCa exhibits only one phase of shortening, which is unchanged from that of controls (Morris et al. 2003). Fourth, using similar fsTnC extraction and cTnC replacement as reported here, it was found that removal of > 90 % of the fsTnC from rabbit psoas fibres and its subsequent replacement with rat cTnC at maximal activation produces only 60-65 % of the control isometric force (Moreno Gonzales et al. 2003 and M. Regnier, personal communication). When these muscle fibres are subsequently incubated in a solution containing fsTnC (1 mg ml−1) to saturate any Tn sites not containing cTnC, maximal isometric force is unchanged from that obtained prior to the exposure to fsTnC. This result argues directly against the presence of sites on the thin filament TnT/TnI binding sites lacking a cTnC molecule. Thus it is unlikely that the reductions of isometric force, kact and ktr seen on cTnC replacement of fsTnC stem from lack of TnC on the Tn sites.
The question then is why, in previous work by Moss et al. (1991) and Metzger (1996), partial replacement of fsTnC by cTnC had so little effect on isometric force. In those studies, fsTnC extraction and cTnC replacement occurred at sarcomere lengths of > 2.5-2.6 μm, the TnC removal from the region of the thin filament lacking overlap was much more extensive than that from the region of overlap (Yates et al. 1993; Swartz et al. 1996). In muscle fibres at a sarcomere length of 2.5 μm, partial TnC extraction preferentially removes TnC from a 0.4 μm length of the 1.1-μm-long thin filament region adjacent to the Z-line and lacking overlap (i.e. 37 % of the total sarcomeric TnC is removed and would have little impact on isometric force). If a total of 70 % of the TnC is removed from the fibre (Metzger, 1996), 37 % would come from the non-overlap region and 33 % from the overlap region, which accounts for only 52 % of the TnC in the overlap region. Thus only 50 % of the fsTnC is removed from the overlap region, producing the Ca2+-sensitive force. With low extraction of fsTnC in the overlap region, the cooperative effects of adjacent active regions may obscure a reduction of communication between cTnC and fsTnI/fsTnT.
The effects of cTnC replacement on force transients following abrupt changes in [Pi] have not been previously measured. Values of kPi(+) were the same in cTnC- and fsTnC-replaced myofibrils (12.6 and 12.5 s−1 respectively at 5°C and 2 mm final [Pi], see Table 2). These values are consistent with those we reported previously for untreated rabbit psoas myofibrils (7.4 and 18.2 s−1 at 1 and 5 mm final [Pi], respectively; Tesi et al. 2000). They are also in reasonable agreement with results from caged Pi experiments in skinned fibres (Millar & Homsher, 1990; Dantzig et al. 1992; Walker et al. 1992; Regnier & Homsher, 1998) when differences in experimental temperature are taken into account. As reported previously (Tesi et al. 2000), kPi(+) values were two to four times faster (see Fig. 4) than the apparent rate of force generation measured at the same [Pi] by ktr (as well as by kact, data not shown). kPi(+) is thought to probe the crossbridge transitions associated with force generation and, unlike kact and ktr, it is either unaffected (Millar & Homsher, 1990; Tesi et al. 2000) or only marginally affected (Walker et al. 1992) by thin filament activation level. The facts that activation-independent indicators of power stroke kinetics (such as kPi(+) (present data) and unloaded shortening velocity; Morris et al. 2003) are unchanged by cTnC replacement, while kact and ktr are reduced support the hypothesis that cTnC substitution compromises myofibril activation. Thus cTnC replacement alters a step associated with a weakly to strongly bound crossbridge transition preceding the power stroke, while subsequent crossbridge steps are unaffected (Morris et al. 2001, 2003).
The rate constant of the rise in force initiated by a sudden decrease in [Pi] (kPi(-)) was not significantly different from kact and ktr and, like kact and ktr, was significantly reduced following cTnC replacement in fast skeletal myofibrils (see Table 2 and Fig. 4). It has been found that kPi(-) shares the same activation dependence shown by kact and ktr (Tesi et al. 2000). Although the striking differences between kPi(-) and kPi(+) are inconsistent with the predictions of simple one- or two-step crossbridge kinetic models of force generation (e.g. Pate & Cooke, 1989; Millar & Homsher, 1990; Kawai & Halvorson, 1991; Dantzig et al. 1992; Regnier & Homsher, 1998), the reduction of kPi(-) observed after cTnC replacement confirms the failure of the chimeric Tn complex to maximally activate crossbridge interactions even in the presence of saturating concentrations of Ca2+.
An important question raised by the cTnC replacement studies (present study and Morris et al. 2001, 2003) is why cTnC interaction with fsTnI/fsTnT produces a small reduction in pCa50 and cooperativity, no significant change in unloaded shortening velocity and kPi(+), but does produce significant reductions in isometric force, ktr, kact and kPi(-). Comparison of the structures of fsTnC and cTnC yields some clues. First, cTnC lacks a Ca2+-binding region in the N-terminal domain of cTnC (in the loop linking helix A and helix B) because negatively charged residues chelating Ca2+ in fsTnC (residues 28, 30 and 34) are uncharged in cTnC (Tobacman, 1996). This change alone should reduce the cooperativity. Second, when Ca2+ binds to the N-terminal domain of fsTnC, helices B and C move away from a structural unit composed of helices N, A and D in transition from a ‘closed’ to an ‘open’ structure. This motion exposes a ‘hydrophobic’ patch to which a regulatory region of fsTnI (fsTnI residues 115-131 or cTnI residues 147-163) binds (Herzberg & James, 1988; Gagne et al. 1995; Vassylyev et al. 1998). This binding is thought to detach the inhibitory peptide of fsTnI from a binding site on actin and allows movement of Tn and tropomyosin structures to expose S-1 binding sites on actin (Tobacman, 1996; Gagne et al. 1995; Herzberg et al. 1986). In cardiac muscle, Ca2+ binding to the N-terminal regions does not create an ‘open’ structure (Spyracopoulos et al. 1998). Rather, residues cTnI150-158 of the regulatory region of cTnI147-163 bind to hydrophobic side chains (particularly the alanine residues 22 and 23 of cTnC) in helices A, B and D, which detach the cTnI inhibitory peptide from its binding site on actin resulting in thin filament activation (Vassylyev et al. 1998; Li et al. 1999). The binding of cTnI147-163 to the cTnC N-terminal domain is only about 15 % as strong as the binding of the analogous regions of fsTnI (fsTnI 115-131) to fsTnC (Li et al. 1999). If the strength of this interaction affects the distribution among thin filament states in the presence of Ca2+ and the absence of bound crossbridges, then the rate of activation of the thin filament (and therefore kact, ktr and kPi(-)) might be correspondingly slowed by substitution of cTnC for fsTnC. More generally, any of the structural differences between the cardiac and fast skeletal Tn subunits could compromise the efficiency of the signalling between the cTnC and fsTnI.
Regardless of which set of interactions within Tn are perturbed, we suggest that the mechanical consequence is a slowing of the rate of the weak-to-strong crossbridge binding, which limits the rate of force development and the steady-state number of strongly bound crossbridges. Earlier we showed that such a kinetic interpretation is consistent with the reduction in isometric force and ktr seen with cTnC substitution in single skinned skeletal muscle fibres (Morris et al. 2001). The Geeves model (McKillop & Geeves, 1993) and the structural interpretations of Lehman and co-workers (Xu et al. 1999) of thin filament regulation suggest an explanation of these results. We suggest that the rate of the weak-to-strong (non-force-bearing) crossbridge transition is determined by the equilibrium between two of the structural states of the thin filament: the B state, and the C state (Lehman et al. 2000; Vibert et al. 1997). In this model of regulation, strong crossbridge formation, even though non-force-bearing, requires that the local tropomyosin position be that of the C (or M state), but not that of the B state. In thin filaments containing chimeric Tn, weak TnC-TnI interactions cause one-third of the thin filament to remain in the B state in the presence of saturating Ca2+. Such an altered distribution might decrease kact, ktr and kPi(-), as well as reducing maximal isometric tension.
Tn complex exchange and Ca2+ regulation
The failure of the replacement of fsTn by cTn to alter crossbridge mechanics (isometric force, ktr, kact) per se suggests that the crossbridges themselves determine these effects (providing that regulatory proteins do not limit the rate of crossbridge attachment). These findings agree with the effects on nucleotide-free myosin S1-thin filament binding resulting from substitution of cardiac for skeletal muscle Tn (Maytum et al. 2003). In the presence of Ca2+, neither the kinetics nor the equilibrium of skeletal myosin S1-actin binding was altered by the Tn source. The current data show that this is also true in the sarcomere for cycling crossbridges in the presence of ATP, as assessed for a broad panel of crossbridge kinetic parameters. In the presence of Ca2+, myosin-thin filament interactions are unaffected by Tn isoform.
On the other hand, significant changes in the pCa50 and Hill coefficient seen in pCa-tension experiments (see Fig. 6) indicate that Tn isoforms cause important changes in Ca2+ regulation. The observed effects in myofibrils are thoroughly consistent with those reported previously for cTn-exchange in rabbit psoas muscle fibres (Brenner et al. 1999; see their Fig. 2). As shown here by control experiments in which homologous fsTn was used in the exchange protocol, the large Ca2+-sensitizing effect of cTn-replacement in rabbit psoas myofibrils was not a consequence of the exchange procedure itself.
It is unclear how exchanging fsTn for cTn produces the combination of increased Ca2+ sensitivity and decreased cooperativity. The most straightforward explanation involves two separate mechanisms. After cTn for fsTn exchange, the thin filament regulatory sites have: (1) a decreased cooperativity due to the presence of one rather than two Ca2+ binding sites in each cTnC regulatory domain and (2) a higher intrinsic Ca2+ affinity. A reduction of the number of regulatory Ca2+-binding sites could reduce the cooperativity by about 1 Hill unit (Grabarek & Gergely, 1983). However, an increased intrinsic Ca2+ affinity is not supported by comparisons among various chimeric and non-chimeric thin filaments (Rosenfeld & Taylor, 1985; Walsh et al. 1985; Tobacman, 1987; Tobacman & Sawyer, 1990; Korman & Tobacman, 1999). An alternative explanation for the increased pCa50 following exchange is a greater myosin-induced activation of the thin filament. A weakness of this idea is that this same mechanism should promote cooperativity in the pCa-tension curve, compensating for the decrease in cooperativity resulting from one Ca2+ binding site per cTn. Indeed, substitution of cTn for fsTn in the presence of EGTA does cause the myosin S1-thin filament binding curve to become more cooperative (Maytum et al. 2003). However, in the present experiments the Hill coefficient after cTn exchange drops to 1.4, not much greater than 1. It is not clear how myosin-mediated effects might greatly shift the pCa50 and yet not preserve or increase cooperativity. Despite this uncertainty, the most likely explanation for the increased pCa50 after Tn exchange is enhanced thin-filament activation by crossbridges, and the decreased cooperativity presumably reflects the lone regulatory site of cTnC.
To account in more detail for the Ca2+-sensitizing effect of cTn in fast skeletal myofibrils, we suggest that cTn is less effective than fsTn in suppressing thin-filament activation. Several observations support this view. First it is known that the addition of cardiac regulatory proteins does not alter the Vmax of the actin activated S1 ATPase (Tobacman & Adelstein, 1986; Williams et al. 1988; Tobacman et al. 2002), but addition of skeletal muscle Tn-tropomyosin decreases the Vmax (Williams et al. 1988). In addition, the same data show that the cardiac regulatory proteins decrease the thin filament Km, but the skeletal muscle proteins do not. In filaments regulated by fsTn, addition of Ca2+ does not produce a fully active ATPase unless strongly bound crossbridges (e.g. NEMS-1) are present or the ratio of actin to S1 is less than 2-4 (Williams et al. 1988). However, thin filaments controlled by cTn are almost fully activated in the presence of Ca2+, independent of the ratio of actin to S1 (Butters et al. 1997; Tobacman et al. 2002). Furthermore, unlike the equal results found in the presence of Ca2+, in the presence of EGTA myosin S1-thin filament binding is more inhibited by skeletal than by cardiac Tn (Maytum et al. 2003). Maytum and co-workers suggested this is due to a larger occupancy of the fully active or M state of the thin filament when cardiac rather than skeletal Tn is present, in EGTA. However, other interpretations are possible because their analysis used a model that accounts for some but not all major aspects of thin filament behaviour (Tobacman & Butters, 2000) and fails to explain the profound inhibition of actin-S1 phosphate release in the absence of Ca2+ (Heeley et al. 2002). Finally, in the case of skeletal muscle myofibrils containing exchanged cTn, there is also the added complication that the regulation by cTn occurs on thin filaments controlled by fast tropomyosin (fsTm). In that case, the effectiveness of the cooperativity may be altered by differences in the structures of fsTm and cardiac tropomyosin (i.e. different ratios of alpha and beta tropomyosins). It is possible that the cooperativity in the presence of fsTm/fsTn is therefore greater than that of fsTm/cTn present in the exchange experiments described above.
The ease with which Tn exchange occurs in myofibrils means that many of these hypotheses can be tested by using chimeras of Tn in which those areas hypothesized to control the behaviour of the muscle fibre have been altered. The relative ease of exchange of whole Tn also means that such exchanges can be made using hypertrophic cardiomyopathy mutant Tn (TnI and TnT forms) to test for effects of these mutations on Ca2+ sensitivity and cooperativity. However, additional work is needed using cardiac skinned myocytes or myofibrils to learn whether whole Tn exchange in these preparations is as effective as it is in skeletal muscle myofibrils. The ease with which such exchange can be made in skeletal muscle preparations opens the way to perform tests of cooperativity using different ratios of fsTn or cTn and permanently inhibited fsTn or cTn (Morris et al. 2001; Regnier et al. 2002).