The effects of regulatory proteins on crossbridge-actin interaction
The increased isometric force and unloaded thin filament sliding speed observed in the presence of regulatory proteins were unexpected. The maximal rate of ATP hydrolysis by actin and myosin S1 heads,
Vmax, is unaffected by the presence or absence of regulatory proteins at pCa 5 at roughly equivalent molar equivalents of actin and myosin heads (
Williams et al. 1988). However, when the myosin head concentration is reduced to less than 0.2 of the actin concentration in the presence of regulatory proteins,
Vmax falls to ~20 % of its maximal value, despite the fact that the F-actin concentration necessary to reach 50 % of the maximal ATPase (
KATP) rate is unchanged (
Williams et al. 1988). Measurements of actomyosin ATPase rates in solution are made at low ionic strength (13 mM) and the interaction of the myosin heads with the actin filament are mechanically unconstrained. In motility assays and muscle fibres, interaction between the myosin heads and thin filament are subjected to marked mechanical constraints which demonstrably affect the rate of ATP hydrolysis (
Siemankowski et al. 1985;
Woledge et al. 1985;
Ma & Taylor, 1994). Although the unloaded shortening velocity is roughly proportional to the actomyosin (AM) ATPase rate over four orders of magnitude (
Barany, 1967), the ATPase rate is not limited by the rate of crossbridge detachment from the thin filament (
Ma & Taylor, 1994). At physiological [ADP] and [ATP], the rate of ADP release from the AM.ADP complex, which does limit the unloaded thin filament sliding speed, is 10–100 times faster than the actomyosin ATPase
Vmax (
Siemankowski et al. 1985).
Might the binding of the regulatory proteins to the thin filament stiffen it and cause the crossbridge to more efficiently and rapidly push the thin filament in the motility assay? If this were the case the persistence length (Lp, a measure of the thin filament stiffness) of actin should be markedly increased by the presence of regulatory proteins. However, the persistence length of F-actin is 9 μm and increases to just 12 μm in the presence of regulatory proteins and calcium (Isambert et al. 1995). Further, Tm addition to the thin filament increases Lp to 21 μm, but the thin filament sliding speed is not significantly changed (Fig. 3). These results suggest that a change in thin filament stiffness cannot account for the increased sliding speed and force observed in the presence of regulatory proteins.
Which regulatory proteins produce the changes in force and unloaded thin filament sliding speed? The experiments shown in Fig. 3 show that the presence of Tn produces a significant increase in Vf, but Tm does not. These results agree with those of Bing et al. (1997a), who found that Tm or Tm + TnI + TnC binding to the thin filament does not increase thin filament sliding speed beyond that of actin alone. However, when skeletal Tm and skeletal Tn, skeletal Tm + skeletal TnI + TnC + TnT, or cardiac Tm and cardiac Tn are bound to the thin filament, the regulated thin filament sliding speed is increased over that of actin alone by ~30 %. These results suggest that TnT acts in concert with TnI and TnC to increase Vf. The extended amino terminus of TnT binds to both the overlapping regions of adjacent tropomyosin molecules and to actin. TnT interaction with TnC, TnI and Tm both increases inhibition of actomyosin ATPase in the absence of Ca2+ and increases stimulation in the presence of Ca2+ (Potter et al. 1995). These results are consistent with the increased thin filament sliding speed seen upon replacement of native TnT with the I79N FHC TnT (Lin et al. 1996; Tobacman et al. 1999). Finally, other regulatory protein mutations have been linked to an increased thin filament sliding speed in motility assays. These include FHC TnT mutations ΔExon-15,16 and E244D (Tobacman et al. 1999), and FHC Tm mutations D175R and Q180G (Bing et al. 1997b). These mutant regulatory proteins increase regulated thin filament unloaded sliding speed by 6–21 % over thin filaments containing native or wild-type regulatory proteins. Studies in which cultured cardiac myotubes were transfected with and expressed I79N or R92Q TnT produced a 2-fold increase in unloaded shortening velocity and the I79N reduced the maximal isometric force to 75 % of its control value (Sweeney et al. 1998). This behaviour of the I79N TnT containing myotubes is consistent with the observations made in the force and sliding speed reported here.
Insight as to how regulatory proteins affect the crossbridge cycle may be gained by considering the steps in the crossbridge cycle.
Scheme 1 represents a minimalist series of reaction steps for the crossbridge ATPase mechanism. Strain-dependent values for each of the rate constants for the steps have been modelled in a way which accounts for the observed force-velocity curve, isometric force, unloaded shortening velocity, and their dependence on ATP, ADP and Pi (Pate & Cooke, 1989). In this model M.ADP and M.ADP.Pi are detached crossbridge states, AM.ADP.Pi is a weakly attached crossbridge, and AM.ADP and AM are strongly attached force exerting states. It is also very likely that steps 3, 4 and 5 are composed of two steps each (an isomerization and a dissociation step). Because regulatory proteins are unlikely to affect the behaviour of detached states, only steps 2–5 should be affected by the addition of regulatory thin filament proteins. Examination of the Pate & Cooke (1989) model reveals that: (1) increases in the forward (clockwise) rate constants of steps 2 and 3 significantly increase isometric force, Pmax, but have little effect on unloaded shortening velocity, Vu; and (2) increases in the forward rate constants of steps 3, 4, or 5 increase Vu but only 4 and 5 have a slight negative or no effect on Pmax (for a summary of the effects of changes in these rate constants see Table 8, Regnier et al. 1998). Increasing the forward rate of step 5 markedly increases the unloaded shortening velocity but has no effect on isometric force or ATPase rate. It may be that the binding of regulatory proteins to the thin filament changes the myosin interaction surface of actin so that the a combination of changes in the rate constants of steps 2–5 occur to produce an increased force and unloaded sliding speed. Alternatively, once bound to the thin filament, the crossbridge head might itself interact with the regulatory proteins, thereby altering the rate constants of one or more of steps 2–5. Because the rates of these steps are strain dependent (there is not simply one fixed rate), it is possible that changes occur primarily at negative or positive strains. Nevertheless, changing the interaction between the thin filament and its substrate (M.ADP.Pi) by effects at a separate site interacting with the regulatory proteins is, by definition, an allosteric effect.
Relation to studies in single muscle fibres
The microneedle force-pCa curve (
Figs 4 and
5) (
nH, 1.7; p
K, 6.2) is comparable to those of skinned cardiac cells (
Hofmann et al. 1991) and skinned skeletal fibres whose endogenous TnC was replaced with cardiac TnC (
Moss, 1992). Thus the single regulated thin filament behaves very similarly to skinned muscle fibres. These results suggest that at reduced [Ca
2+], either fewer crossbridges are attached and pulling on the thin filament or the force exerted per crossbridge is reduced. The constancy of the force/stiffness ratio as pCa is reduced in skinned muscle fibres (
Brenner, 1988) argues for the former interpretation. In motility assays, however, the Hill coefficient is 15–35 % less than that for fibres. Two factors may contribute to the reduced co-operativity of the force-pCa curve for
in vitro motility assays. First, the random array of S1 heads on the coverslip and their less productive attachment to the filaments may reduce the co-operative effects of strongly bound crossbridges (
VanBuren et al. 1994). Second, the force measurements were made at 50 mM ionic strength, compared with the more physiological value of 200 mM used in fibre experiments. Reductions in
nH are observed at lower ionic strength in both skeletal and cardiac muscle (
Fink et al. 1986).
The effects of pCa on fibre unloaded shortening velocity are different from its effects on thin filament sliding speed. In fibres, Moss (1986) found that at submaximal pCa unloaded shortening occurred in two phases. An initial fast phase of shortening, whose velocity, Vu, exhibited little dependence on calcium (occurring for shortening of less than 8–10 % the muscle length), was followed by a second phase of shortening whose velocity declined in proportion to the reduction in [Ca2+] (Moss, 1986). In the intact muscle fibre whose maximal activation has been reduced by treatment with the E-C coupling inhibitor, dantrolene, no significant change in Vu was observed (Edman, 1979). This discrepancy may stem from the fact that in Edman's (1979) experiments, the distance shortened at reduced pCa was always less than 10 % of the muscle length. Further dantrolene treatment may produce a non-uniform reduction in pCa throughout the fibre diameter, i.e. it may inhibit calcium release deep in the fibre but not at the fibre surface. The effects of pCa observed on sliding speed in the motility assays more closely resembles the effects of calcium on the initial phase of the unloaded shortening velocity. The lack of correspondence between the motility sliding speed and the second phase of unloaded shortening in skinned muscle fibres suggests that the second phase is related to constraints present in the skinned fibre but absent in the isolated thin filament.
Unlike the Vf-pCa data reported here and elsewhere (Homsher et al. 1996; Gordon et al. 1997), Fraser & Marston (1995) reported only a 30 % reduction in Vf at very low calcium concentrations using silanized motility surfaces. When they used nitrocellulose-coated surfaces (as in the current study) they found a marked drop in Vf (Bing et al. 1997a). They attributed the different behaviour between the two surfaces to the presence of a drag present on nitrocellulose-treated surfaces which is absent on silanized motility surfaces. Unlike Bing et al. (1997a), we observed no filament sliding at pCa 8 or 9 (Fig. 4) over either nitrocellulose-coated or silanized surfaces (E. Homsher & N. Back, unpublished results). Our interpretation of the significant sliding speeds Fraser & Marston (1995) observed at pCa 8 and 9 is that the [Tn] in their motility assay was too low to completely regulate the thin filament. The thin filament sliding seen at pCa 8 and 9 was due to thin filaments with Tm but no Tn bound. This interpretation is supported by similarity of the Vf at pCa 8 and 9 to that of unregulated thin filaments.
The role of calcium in regulating force and sliding speed
Two hypotheses have been proposed to explain the [Ca
2+] regulation of muscle contraction. The first was a ‘recruitment’ model derived from the steric blocking model of contraction (H. E.
Huxley, 1972). In this model calcium binding to TnC is an ‘on/off’ switch, which controls the number of crossbridges bound to the thin filament but does not alter the rate of crossbridge attachment to or detachment from the thin filament or the rate or extent of the power stroke (
Podolsky & Teichholz, 1970). The second hypothesis is a ‘kinetic’ model, in which calcium binding to TnC modulates the rate of crossbridge attachment to or detachment from the thin filament or the rate or extent of the power stroke (
Julian, 1971;
Brenner, 1988). The recruitment model as originally proposed cannot be correct because crossbridges bind (albeit weakly) to thin filaments in the absence of calcium (
Brenner et al. 1982). Further the increase in steady state ATPase rate as [Ca
2+] rises is far out of proportion to the change in crossbridge binding to actin (
Chalovich et al. 1981). The rate of force redevelopment (
ktr) in fibres varies with pCa (
Brenner, 1988;
Millar & Homsher, 1990;
Walker et al. 1992) while the simplest recruitment model predicts that
ktr will be independent of the pCa. The kinetic model of regulation originally suggested that calcium regulated the rate of P
i release from the actomyosin crossbridge (
Chalovich & Eisenberg, 1982) which corresponds to a regulation of the power stroke. The apparent independence of the rate of P
i release in isometrically contracting muscles at various pCa (
Millar & Homsher, 1990;
Walker et al. 1992) implies, however, that calcium regulation occurs at a step prior to P
i release, but after weak binding to the thin filament. This result implies, according to
Scheme 1, that calcium regulates a crossbridge step after step 2 but before step 3, i.e. a transition from a weakly bound crossbridge to a strongly bound crossbridge that has not produced any force. The data in
Figs 4–
6 support this view, because thin filament sliding speed (which is most affected by changes in steps 3, 4 and 5) fell by less than 20 % when reductions in [Ca
2+] reduced the isometric force by greater than 75 % (which are strongly affected by changes in the rates of steps prior to 3, 4 and 5). The similarity of alterations in sliding speed due to changes in pCa and changes in number of available actin binding sites at pCa 5 (in CBMII-Tn) (
Fig. 6) indicates that the change in speed is less related to the pCa than to the number of attached crossbridges.
The behaviour of unloaded thin filament sliding speed is similar to that reported earlier in that reduction of [Ca2+] slows regulated thin filament sliding speed (Fraser & Marston, 1995; Homsher et al. 1996; Gordon et al. 1997). At first glance this result seems contradictory to A. F. Huxley's (1957) postulate that unloaded shortening velocity is limited by the rate of crossbridge detachment from the thin filament. However, as Uyeda et al. (1990) have shown, this is true only if there are large numbers of crossbridges (>75) interacting with a filament. In motility assays thin filament length ranges from 1–15 μm and given the crossbridge density this means that 130–1950 crossbridges can interact with the thin filament. However, if the number of accessible actin sites is reduced to 10 % of maximal, the number of crossbridges pulling on the thin filament must fall to 13–195 for the thin filaments. Given a duty cycle of 2.17 ms and a maximal turnover rate of 25 s−1, this would yield sliding speeds of 0.52–1.0 of maximal (see legend to Fig. 6 for details of calculation). In the fibre thin filaments are arranged in parallel and 1000 thin filaments (Woledge et al. 1985) are attached to each side of the Z line per myofibril. Thus if only one crossbridge could attach per thin filament, there would be 1000 crossbridges pulling on each side of the Z line in a myofibril.
In both isometric and isotonic conditions the power stroke rapidly follows strong binding, succeeded by product release, ATP binding, and S1-ATP dissociation from actin. The rate at which weakly bound crossbridges enter into the strongly bound force-generating state will be calcium dependent, if there is a Ca2+-dependent equilibrium step (weak to strongly bound but non-force exerting crossbridge attachment) prior to the power stroke itself. This mechanism would then account for the well-documented effect of [Ca2+] on the rate of rise of force (Brenner, 1988). Because a Ca2+-regulated transition from a weakly to strongly bound non-force generating state would not contribute to thin filament motion, unloaded thin filament sliding speed would be independent of the extent of thin filament activation as long as there are enough crossbridges attached to constantly and uniformly propel the thin filament. When insufficient numbers of actin molecules are exposed to assure constant movement of the thin filament, sliding speed will fall (Uyeda et al. 1990).
Relation to FHC mutant regulatory proteins
Isometric force is reduced and unloaded thin filament sliding speed is increased by the presence of the I79N TnT mutant, which agrees well with corresponding measurements in quail myotubes expressing this mutant (
Sweeney et al. 1998). The TnT mutation in this report and several of those studied by
Sweeney et al. (1998) are located in a region of troponin forming an extended tail stretching along and binding to the tropomyosin strand (
Tobacman, 1996). The similar effects occasioned by the presence of the I79N and R92Q TnT mutations (reduced isometric force and increased
Vu) together with the fact that large changes in Tm affinity for actin occur in the presence of S1, Tn and Ca
2+ suggest these mutations change the actin-Tm interacting surface. Alterations in the actin interaction surface may allosterically produce changes in acto-S1 binding and crossbridge kinetics. If the changes in acto-S1 interaction increase the rate of crossbridge dissociation from actin (R
51 in the model of
Pate & Cooke, 1989), isometric force will be reduced and
Vu will rise. The mechanical changes would stem from an altered position of Tm over the actin-Tm interface mediated by alterations in the TnT tail flexibility. The reduced isometric force and increased
Vu may then initiate the compensatory hypertrophy seen in familial hypertrophic cardiomyopathy.