Corresponding author R. L. Moss: Department of Physiology, University of Wisconsin Medical School, 1300 University Avenue, Madison, WI 53706, USA. Email: firstname.lastname@example.org
At low levels of activation, unloaded shortening of skinned skeletal muscle fibres takes place in two phases: an initial phase of high-velocity shortening followed by a phase of low-velocity shortening. The basis for Ca2+ dependence of unloaded shortening velocity (Vo) in the low-velocity phase was investigated by varying the level of thin filament activation with Ca2+ and N-ethyl-maleimide myosin subfragment-1 (NEM-S1), a non-tension-generating, strong binding derivative of subfragment-1. Vo was measured with the slack-test method.
Treatment of skinned fibres with 5 μm NEM-S1 eliminated the low-velocity phase of shortening but had no effect on the high-velocity phase of shortening during submaximal activation with Ca2+, or on Vo during maximal activation with Ca2+.
Extensive washout of NEM-S1 from the treated fibres restored the low-velocity phase of shortening and returned low-velocity Vo to pre-treatment values.
The effect of NEM-S1 to increase low-velocity Vo can be explained in terms of a model in which strong binding myosin cross-bridges activate the thin filament to a state in which the rate of ADP release from the actin-myosin-ADP complex and the rate of cross-bridge detachment from actin are accelerated during unloaded shortening.
In mammalian skinned muscle fibres, several investigators have found that unloaded shortening velocity (Vo) decreases when Ca2+ concentration is reduced to levels that induce submaximal isometric tensions (Julian, 1971; Julian et al. 1986; Moss, 1986; Farrow et al. 1988; Martyn et al. 1994; Metzger, 1996). At saturating [Ca2+], Vo is maximal and invariant throughout the time course of shortening. During submaximal activations, unloaded shortening takes place in two distinct phases, an initial rapid phase and a subsequent slow phase (reviewed by Moss, 1992, and Gordon et al. 2000). As [Ca2+] is lowered from maximal, Vo in the high-velocity phase remains high and virtually constant until tension-generating capability is reduced to ≈30 % Po, below which Vo progressively decreases as [Ca2+] is further reduced. On the other hand, Vo in the low-velocity phase continually decreases as [Ca2+] is reduced within the submaximal range.
The activation dependence of the high-velocity phase of shortening has been recapitulated in regulated in vitro motility assays (Homsher et al. 1996; Gordon et al. 1997), where sliding velocities are constant over a wide range of activations but decrease at very low [Ca2+], possibly due to reductions in cross-bridges to less than a critical number. During shortening the number of cross-bridges bound to actin appears to be significantly less than the number during isometric contractions (Ford et al. 1985) and would be even further reduced at low levels of activation. This reduction in cross-bridge number might be expected to slow shortening velocity, since previous studies of the regulation of force development have shown that cross-bridge turnover kinetics are related to numbers of strong binding cross-bridges (Swartz & Moss, 1992; reviewed by Lehrer, 1994, and Gordon et al. 2000).
In contrast, the basis for the low-velocity phase of shortening is not well understood, although it has been suggested that at low [Ca2+] slowed detachment of cross-bridges gives rise to an internal load that slows Vo (Moss, 1986). Earlier work showed that variations in low-velocity Vo are not due to the changes in [Ca2+]per se, since low-velocity Vo scaled similarly with tension-generating capability when activation was changed by varying [Ca2+] or at constant [Ca2+] by partial extraction of the regulatory protein troponin C (Moss, 1986). Involvement of C-protein, a thick filament accessory protein, in mediating the low-velocity phase was suggested by results in which partial extraction of C-protein increased Vo in the low-velocity phase but had no effect on Vo in the high-velocity phase (Hofmann et al. 1991). The effect of C-protein on the low-velocity phase was taken to suggest that C-protein confers an internal load that slows Vo at low levels of activation.
Another plausible explanation for the low-velocity phase of shortening is that activation of the thin filament is reduced during shortening due to reduced numbers of cross-bridges attached to the thin filament (Iwamoto, 1998). As discussed above, the number of cross-bridges bound to the thin filament decreases during shortening, especially at low levels of activation. Furthermore, at low levels of activation, un-overlapped regions of the thin filament would be activated to a lesser degree than overlapped regions, since both Ca2+ and strongly bound cross-bridges are required to fully activate cross-bridge binding (Swartz et al. 1990, 1996) and cross-bridge kinetics (Swartz & Moss, 1992). Thus as shortening continues, cross-bridges would encounter regions of reduced activation, which would slow cross-bridge cycling and thereby reduce unloaded shortening velocity. In this regard, it is interesting to note that the extent of shortening during the high velocity phase, 60-80 nm per half-sarcomere (Moss, 1986), is similar to the distance that the activation signal is thought to spread from the edge of the A-band into the I-band (discussed by Swartz et al. 1996). Most investigators estimate that such activation involves between 7 and 14 actin monomers, i.e. the spread of activation into the I-band could be nearly 80 nm (14 actin monomers × 5.5 nm per monomer diameter).
The present study was undertaken to further investigate the molecular basis for activation dependence of Vo in the low-velocity phase of shortening. The role of strong-binding myosin heads was studied in skinned skeletal muscle fibres by infusing a derivative of myosin subfragment-1, i.e. N-ethyl-maleimide subfragment-1, or NEM-S1, which binds tightly and apparently randomly to the thin filaments (Nagashima & Asakura, 1982; Williams et al. 1988; Swartz & Moss, 1992) and does not develop or bear tension.
Skinned fibre preparation
Skinned single fibres were prepared from rabbit psoas muscle as described previously by Moss (1986). Rabbits were first killed by overdose with pentobarbitone (100 mg (kg body weight)−1, delivered intravenously), an incision was made in the abdomen, and the psoas muscles were then surgically removed. On the day of each experiment, a segment of skinned fibre was mounted to the apparatus between a force transducer (model 403 with resonant frequency of ≈600 Hz and sensitivity of 20 mV mg−1 or model 400 with frequency of ≈2 kHz and sensitivity of 2 mV mg−1, Cambridge Technology) and a DC torque motor (model 300H, Cambridge Technology). After mounting the fibre, the experimental chamber was placed on the modified stage of a microscope (Zeiss model WL) and fibre length was measured by translating the fibre horizontally through the field of the microscope. Depth was measured by focusing from top to bottom of the fibre, which was viewed with a ×10 eye-piece, a ×1.6 intermediate lens, and an ultra-long working distance objective (Zeiss UD 40). Sarcomere length of the relaxed fibre was adjusted to 2.3-2.5 μm by changing overall fibre length. Sarcomere lengths in relaxed and activated fibres were measured from photomicrographs of the central region of each fibre.
Measurement of Vo
Unloaded shortening velocity was measured at 15 °C using the slack-test method (Edman, 1979) while the fibres were steadily activated in solutions of either maximal (pCa 4.5, where pCa = -log[Ca2+]) or submaximal [Ca2+] (pCa > 4.5). Slack-test data obtained at pCa 4.5 were well fitted by a single straight line, while data obtained at submaximal [Ca2+] were well fitted by two straight lines corresponding to the high- and low-velocity phases of shortening (Moss, 1986).
Solutions of variable Ca2+ concentration contained 4 mm MgATP, 1 mm free Mg2+, 20 mm imidazole (pH 7.0), 14.5 mm creatine phosphate, 7 mm EGTA, and different total CaCl2. Ionic strength was adjusted to 180 mm by adding KCl. Ca2+ concentrations are expressed as pCa, i.e. -log[Ca2+], which were calculated using stability constants listed by Godt & Lindley (1982) and the computer program of Fabiato (1988).
Protein purification and modification
All procedures were done at 4 °C unless specified otherwise. Myosin was purified from rabbit fast skeletal muscle as described previously (Swartz et al. 1990), and myosin subfragment-1 (myosin S1) was prepared by chymotryptic digestion as described by Weeds & Pope (1977). The A1 isoform of S1, i.e. S1 with light chain 1 bound, was then purified by cation-exchange chromatography (Swartz & Moss, 1992). Prior to modification with N-ethyl-maleimide (NEM), the S1 was first incubated with 25 mm Tris (pH 7.6), 5 mm EDTA, and 10 mm dithiothreitol (DTT) for 30 min at 25 °C, and subsequently dialysed against 25 mm Tris (pH 7.6) and 5 mm EDTA. After dialysis, the protein was reacted with 15 mol NEM per mol S1 for 30 min at 25 °C. The reaction was quenched with excess DTT, and the protein was precipitated with ammonium sulphate at 75 % saturation. The precipitate was collected by centrifugation and stored at 4 °C until use. When used for an experiment, the NEM-S1 was dialysed against 20 mm imidazole (pH 7.0) and 1 mm DTT and was then filtered through a 0.45 μm filter. To assess the effects of NEM-S1 on Vo, fibres were treated for 20 min with relaxing solution containing 5 μm NEM-S1, which was prepared by mixing one volume of ×2 concentrated relaxing solution (see below) and one volume of 20 mm imidazole (pH 7.0) containing NEM-S1. This concentration of NEM-S1 was selected as the maximum concentration that did not competitively inhibit maximum tension development in activating solution of pCa 4.5 (Swartz & Moss, 1992). As described previously (Swartz & Moss, 1992), 4-8 μm NEM-S1 had no significant effects on the Ca2+ sensitivity of force in skinned skeletal muscle fibres, i.e. there was no change in the pCa at which force was half-maximal, but NEM-S1 significantly increased forces at pCa values for which force was less than half-maximal. Resting forces in solutions of pCa 9.0 were greater in the presence of 5 μm NEM-S1 (≈4 % of maximum Ca2+-activated force) than in its absence (< 1 % of maximum force).
Tension and velocity in control fibres
Skinned fibre segments used in this study developed maximal isometric tensions of 132 ± 20 kN m−2 (n = 6) and had average Vo values of 3.36 muscle lengths per second (ML s−1) during maximal activations at pCa 4.5. These values agree well with earlier data from this (e.g. Moss, 1986) and other laboratories (e.g. Farrow et al. 1988). As reported previously, slack-test plots obtained at submaximal concentrations of Ca2+ were biphasic (see control data in Fig. 2), i.e. shortening up to ≈80 nm per half-sarcomere proceeded at relatively high velocity and subsequent shortening was slower (Moss, 1986; Farrow et al. 1988; Metzger, 1996).
Effects of NEM-S1 on shortening velocity
Figure 1 presents motor position and force traces from a fibre subjected to identical changes in length both before (b) and after (a) treatment with 5 μm NEM-S1. To allow direct comparisons of these records, the concentration of Ca2+ was reduced in the presence of NEM-S1 to yield similar forces in the two cases, since treatment with NEM-S1 increases the Ca2+ sensitivity of tension, as shown previously (Swartz & Moss, 1992). Typical of such comparisons, treatment with NEM-S1 substantially reduced the duration of unloaded shortening following the slack step and increased the initial rate of rise of tension re-development following the period of unloaded shortening, even though steady isometric force prior to the length step was virtually identical in the presence and absence of NEM-S1. This result is consistent with previous findings that increased numbers of strongly bound cross-bridges speed the kinetics of force development in skinned skeletal muscle fibres (Swartz & Moss, 1992). NEM-S1 had minimal effects on force records obtained during slack-test measurements performed at maximal Ca2+ concentrations: Vo decreased from 3.29 ± 0.11 ML s−1 in control fibres to 2.95 ± 0.05 ML s−1 in the same fibres following treatment with NEM-S1. This difference was significant (P < 0.02) but was probably due to run-down in fibre performance between measurements (Moss, 1992) and not a specific effect of NEM-S1 on maximum cross-bridge cycling rate.
While slack-test plots obtained during submaximal activation were biphasic in untreated control fibres (Fig. 2), application of 5 μm NEM-S1 eliminated the low-velocity phase, so that shortening proceeded at the high-velocity Vo characteristic of that level of activation. In Fig. 2, slack-test data from a fibre before treatment with NEM-S1 shows that shortening velocity was constant during maximal activation. Slack-test plots obtained during submaximal activation were biphasic, and Vo in the low-velocity phase decreased as the level of activation was reduced. After application of 5 μm NEM-S1, the low-velocity phase of shortening was eliminated so that shortening velocity was constant at each level of activation. Thus NEM-S1 eliminated Ca2+ dependence of Vo at least in the range of activations studied here.
Summary data of the effects of NEM-S1 on Vo are shown in Fig. 3 in which Vo, as a percentage of the control value at pCa 4.5, was plotted as a function of the level of thin filament activation, expressed as percentage maximum tension measured in the same fibres at pCa 4.5. Data were obtained only during Ca2+ activations yielding forces up to 0.5 Po, since it is in this range that reduced activation has its greatest effects on Vo, as shown previously (Julian et al. 1986; Moss, 1986; Farrow et al. 1988; Metzger, 1996). The data show that in submaximally activated control fibres Vo values in the high-velocity phase of shortening were somewhat less than Vo obtained during maximal activation, while Vo in the low-velocity phase was substantially less. Once fibres were treated with NEM-S1, slack-test plots became linear and monophasic, yielding Vo values that were ≈75 % of the values measured during maximal activation.
Reversibility of mechanical effects due to NEM-S1
These effects on Vo were readily reversed by wash-out of NEM-S1 from the treated skinned fibres. As an example, Fig. 4 shows results of experiments in which Vo was measured at submaximal pCa both before and after treatment with NEM-S1 and following wash-out of NEM-S1. Slack-test data were first obtained during a Ca2+ activation yielding 0.31 Po (open squares); the fibres were then bathed for 20 min in the presence of NEM-S1 and subsequently activated at a reduced [Ca2+], which yielded 0.34 Po (filled squares). The fibre was then bathed in relaxing solution with no NEM-S1 for a total of 45 min, with changes to fresh relaxing solution at 15 min intervals. Following the wash-out, slack-test data were obtained during activation with a [Ca2+] yielding 0.29 Po (half-filled squares). Treatment with NEM-S1 eliminated the low-velocity phase of shortening, which was restored by wash-out of NEM-S1. This result is consistent with earlier results in which wash-out of NEM-S1 from skinned fibres reversed the effect of NEM-S1 to speed the rate of submaximal tension redevelopment following release and re-stretch (Swartz & Moss, 1992).
Low-velocity phase of shortening is eliminated by strong binding cross-bridges
The main finding of this study is that N-ethyl-maleimide-modified myosin subfragment-1 (NEM-S1) speeds Vo during the low-velocity phase of shortening, but has no effect on Vo in the high-velocity phase or during maximal activation with Ca2+. Since NEM-S1 binds strongly and nearly irreversibly to actin (Nagashima & Asakura, 1982; Swartz & Moss, 1992), these results imply that strong binding cross-bridges, corresponding to states in which one or both products of nucleotide hydrolysis have dissociated, accelerate detachment of endogenous cross-bridges in partially activated fibres. Conversely, the fact that the low-velocity phase of shortening appears only at submaximal levels of activation suggests that this phase arises when there is less than a critical number of cross-bridges strongly bound to the thin filament. The lack of effect of NEM-S1 on Vo in maximally Ca2+-activated fibres implies that under these conditions the turnover kinetics of myosin are already fully activated by Ca2+ and/or strong binding cross-bridges.
Determinants of unloaded shortening velocity in the low-velocity phase
In Huxley's 1957 model of muscle contraction, Vo is determined by the rate of cross-bridge detachment from actin, since cross-bridges that are slow to detach at the end of the working stroke would give rise to an internal load that opposes further shortening. To account for the higher rate of cross-bridge turnover in shortening muscle relative to the rate of ADP dissociation from actomyosin in solution, Cooke et al. (1994) proposed that significant numbers of cross-bridges are mechanically disrupted following their working stroke and thus do not release ADP during that cycle of interaction with actin. Thus Vo does not seem to be simply related to the rate of ADP release from actomyosin, although it is likely that Vo increases as the rate of ADP release increases (Siemankowski et al. 1985).
From this discussion, it is plausible that the low-velocity phase of shortening at submaximal activation might involve greater cross-bridge-derived internal loads, due possibly to activation-dependent slowing of ADP release and cross-bridge detachment. Metzger's (1996) observation that increased MgADP gives rise to a low-velocity phase of shortening in maximally activated muscle fibres provides support for this idea.
The basis for the low-velocity phase of shortening has previously been proposed to involve a population of long-lived cross-bridges that initially aid contraction, but because they are slow to detach, these cross-bridges ultimately give rise to an internal load that opposes contraction and thereby slows Vo (Moss, 1986). To account for the progressive slowing of Vo in the low-velocity phase as [Ca2+] was reduced, it was hypothesized that the long-lived, rigor-like cross-bridges were bound in zones of transition between active and inactive regions of the thin filament. Evidence for such cross-bridges was obtained by Allen & Moss (1987), who showed that stiffness increased relative to tension when submaximally activated muscle fibres actively shortened. In this model, the number of zones would increase as activation was lowered and the proportion of cross-bridges of the long-lived type would increase. Application of NEM-S1 would presumably speed the rate of cross-bridge detachment in the transition zones by allosteric effects on the thin filament. The present results neither support nor refute this idea.
Another possible explanation for a biphasic time course of shortening is that the degree of Ca2+ saturation of the thin filament decreases during shortening, similar to the decrease in Ca2+ binding observed when intact barnacle muscle fibres are rapidly released to a new length (Ridgway et al. 1983). In such a case, Vo might decrease as a result of a reduction in the number of cross-bridges interacting with actin. However, it seems unlikely that shortening-dependent changes in Ca2+ binding are responsible for biphasic shortening, since biphasic shortening was observed in fibres activated in the absence of Ca2+ by partial removal of the regulatory protein troponin (Moss, 1986). Also, Vo in both phases of shortening scaled to the level of activation of the thin filaments when activation was varied at constant [Ca2+] by partial extraction of troponin C (Moss, 1986).
Roles of strongly bound cross-bridges in modulating low-velocity Vo
Recent evidence suggests that biphasic shortening arises because the number of cross-bridges bound to the thin filament progressively decreases as shortening proceeds, and once a threshold number is reached, cross-bridge cycling kinetics and Vo become slower. For example, Iwamoto (1998) showed that Vo in the low-velocity phase recovered when fibres that had shortened into the low-velocity phase were held isometric and allowed to redevelop force - the recovery of Vo was directly related to the extent of recovery of force and presumably to the increased number of cross-bridges bound to the thin filament. Consistent with this idea, we have previously shown that NEM-S1 facilitates activation of skinned fibres at submaximal [Ca2+] (Swartz & Moss, 1992), which was evident in increased force and rate of force development. Other experiments showed that in the presence of Ca2+ endogenous cross-bridges in the zone of overlap between thick and thin filaments induce greater binding of fluorescently labelled myosin subfragment-1 than is observed in the adjacent I-band which contains only thin filaments and is thus devoid of cross-bridges (Swartz et al. 1990, 1996). During muscle shortening there is a substantial decrease in numbers of cross-bridges bound to the thin filament (Julian & Sollins, 1975; Ford et al. 1985), and at submaximal activations, such a decrease might be sufficient to slow the kinetics of cross-bridge interaction and Vo by co-operative inactivation of the thin filament.
Our present finding that NEM-S1 eliminated the low-velocity phase of shortening provides further support for the idea that the low-velocity phase is the result of shortening-induced inactivation of the thin filament. When a muscle contracts isometrically, activation of the thin filament presumably includes binding of both Ca2+ and strongly bound cross-bridges (Swartz et al. 1990, 1996; reviewed by Lehrer, 1994, and Gordon et al. 2000). When shortening begins, the number of strongly bound cross-bridges in the zone of overlap decreases and the ends of the thick filament slide into the I-band and overlap thin filaments that are activated to a lesser degree due to the initial absence of cross-bridges. Both factors would be expected to slow cross-bridge interaction kinetics and Vo. Attainment of a steady velocity in the low-velocity phase would then be a manifestation of reaching a reduced, steady-state number of strongly bound cross-bridges. Co-operative inactivation of the thin filament can also explain the gradation of low-velocity Vo with [Ca2+], since the number of cross-bridges strongly bound to the thin filament should decrease as [Ca2+] is reduced, thereby slowing cross-bridge cycling kinetics and Vo. Consistent with this interpretation, addition of NEM-S1 to the skinned fibres eliminated the low-velocity phase of shortening, presumably by maintaining a high number of strongly bound cross-bridges.
Other studies (Homsher et al. 1996, 2000; Gordon et al. 1997) have shown that the sliding velocity of regulated thin filaments in an in vitro motility assay decreases as [Ca2+] is reduced, an effect which Homsher et al. (1996) attributed to a reduced number of cross-bridges bound to the thin filament. In this case, however, the mechanism envisaged by the authors was not a slowing in cycling kinetics but instead a reduction in overall sliding velocity due to momentary periods (because of the reduced number of interacting cross-bridges) when no cross-bridges are actually bound to the thin filament. Gordon et al. (1997) further suggested that a critical number of force-generating cross-bridges must be bound to the thin filaments in order to overcome the resistance to movement due to the presence of weakly bound cross-bridges at low [Ca2+]. It seems unlikely that these in vitro motility results involve the low-velocity phase of shortening that we see in skinned fibres, since in each of these studies it was necessary to substantially reduce the [Ca2+] in order to observe a slowing of sliding velocity.
Biphasic shortening at low levels of activation might also be explained on the basis of the time required for co-operative inactivation of the thin filament once shortening has begun. Recent studies of the kinetics of state transitions in tropomyosin (Ishii & Lehrer, 1993) have shown that switching of tropomyosin from ‘on’ to ‘off’ lags behind the dissociation of myosin S1 from regulated thin filaments by tens of milliseconds. Thus during the transition from isometric to unloaded contractions, there is probably a delay in the change in state of tropomyosin following the onset of shortening and the concomitant reduction in numbers of cross-bridges bound to the thin filaments. The duration of high-velocity shortening (20-40 ms) in this and previous studies (Moss, 1986; Farrow et al. 1988) was relatively constant at all levels of activation and agrees well with the delay in change of activation state of tropomyosin measured by Ishii & Lehrer (1993).
Possible mechanisms of effects of strong binding cross-bridges on low-velocity Vo
The molecular mechanism by which strong binding cross-bridges accelerate Vo during submaximal activations probably involves several processes. For example, biochemical studies have shown that strong binding of myosin cross-bridges induces changes in the position of tropomyosin within the thin filament (Lehrer & Ishii, 1988). Additional work has shown that Ca2+ regulation of actomyosin ATPase activity involves acceleration of product release (Rosenfeld & Taylor, 1987), which may be induced by Ca2+ binding to troponin C and/or strong binding of cross-bridges to the thin filaments. While ADP release is an obvious candidate for activation-dependent regulation of Vo, some mechanical phenomena are not easily explained by this mechanism. When the concentration of Pi is increased during submaximal activation, the low-velocity phase of shortening is eliminated (Metzger, 1996) but there is no effect on Vo during maximal activation (Cooke & Pate, 1985; Metzger, 1996), similar to the present results with NEM-S1. This is a surprising result since increased Pi would be expected to reduce the number of strongly bound cross-bridges. The fact that increased [Pi] increases the rate of tension development (Millar & Homsher, 1990) suggests that Pi release is associated with tension-generating steps early in the cross-bridge interaction cycle. Thus the finding that Pi eliminates the low-velocity phase might indicate that Vo in this phase is determined primarily by the rate of force-generating transitions rather than the ADP release steps. Alternatively, as suggested by Metzger (1996), Pi may bind preferentially to strained cross-bridges, thereby reducing the shortening-induced internal load.
This work was supported by National Institutes of Health grant P01 HL47053 to R.L.M.