A myosin-based mechanism for stretch activation and its possible role revealed by varying phosphate concentration in fast and slow mouse skeletal muscle fibers

Abstract

Stretch activation (SA) is a delayed increase in force following a rapid muscle length increase. SA is best known for its role in asynchronous insect flight muscle, where it has replaced calcium’s typical role of modulating muscle force levels during a contraction cycle. SA also occurs in mammalian skeletal muscle but has previously been thought to be too low in magnitude, relative to calcium-activated (CA) force, to be a significant contributor to force generation during locomotion. To test this supposition, we compared SA and CA force at different P_i concentrations (0–16 mM) in skinned mouse soleus (slow-twitch) and extensor digitorum longus (EDL; fast-twitch) muscle fibers. CA isometric force decreased similarly in both muscles with increasing P_i, as expected. SA force decreased with P_i in EDL (40%), leaving the SA to CA force ratio relatively constant across P_i concentrations (17–25%). In contrast, SA force increased in soleus (42%), causing a quadrupling of the SA to CA force ratio, from 11% at 0 mM P_i to 43% at 16 mM P_i, showing that SA is a significant force modulator in slow-twitch mammalian fibers. This modulation would be most prominent during prolonged muscle use, which increases P_i concentration and impairs calcium cycling. Based upon our previous Drosophila myosin isoform studies and this work, we propose that in slow-twitch fibers a rapid stretch in the presence of P_i reverses myosin’s power stroke, enabling quick rebinding to actin and enhanced force production, while in fast-twitch fibers, stretch and P_i cause myosin to detach from actin.

INTRODUCTION

A primary function of muscles is to generate force for locomotion. The mechanism by which calcium activates and modulates force is well known and its importance for muscles during animal locomotion thoroughly understood (46). However, force needs to be maintained during prolonged muscle use, which produces conditions that cause decreased force production, including P_i accumulation and impaired calcium cycling (2, 7, 8, 11, 30, 36, 43, 69, 71). Other means of modulating muscle force exist that may counteract the decline in calcium-regulated force during extended muscle use and in other situations where calcium cycling is impaired. One way is through a sarcomeric phenomenon known as stretch activation (SA). However, much less is known about SA, its function in some fiber types is unclear, and its molecular mechanism is unknown (31, 50).

SA is a delayed increase in muscle force following a rapid stretch (Fig. 1). The force transient following a very rapid stretch of a muscle has classically been defined as having four phases: an immediate force increase (phase 1), a rapid force decay (phase 2), a delayed, in regard to the fiber length change, secondary force increase (phase 3), and a slow force decay (phase 4) (28). The amplitude of SA, F_SA, is the maximum value of the phase 3 delayed force increase. The rates of these phases have been studied extensively for information about various muscle mechanisms (5, 19, 27).

Fig. 1.

Representative stretch activation (SA) responses from a soleus muscle fiber (A) and an extensor digitorum longus (EDL) muscle fiber (B) following a 1% increase in muscle length over 4 ms. Traces shown are a passive SA trace at pCa 8.0 and two traces at pCa 5.0 with 4 mM and 16 mM P_i. Zero tension is equal to isometric tension immediately before the length step. The phase 3 peak, stretch-activated tension (F_SA), is indicated by vertical dashed arrows.

In contrast, the amplitudes of these phases have received relatively little attention, except in specific muscle types, where a high amplitude of phase 3 is obvious such as asynchronous insect indirect flight muscle (IFM) and in vertebrate heart muscle (57, 59, 61). The greatest phase 3 amplitude occurs in asynchronous IFM, which uses SA instead of calcium to vary muscle force during each contractile cycle (53, 57, 58, 60). These insects have evolved SA to the point where the amplitude of stretch activation (F_SA) is about twofold higher than calcium-activated isometric tension (F₀; a very high F_SA to F₀ ratio) (22). In skeletal muscle, lack of consideration for the importance of phase 3 may be because F_SA is not as prominent as F₀ under experimental conditions typically employed when measuring tension transients, such as high [Ca²⁺] and low [P_i], and thus researchers focus on measuring rates instead of amplitudes (50). However, SA force magnitude (phase 3) has not been measured in vertebrate skeletal muscle under moderate to high [P_i].

Discovering the mechanism behind SA has proven elusive. Many potential mechanisms for SA have been proposed, especially for asynchronous IFM, including the thin filament lattice-matching model (72), the connecting filament (titin) stretch model (21, 24), a filament compliance model involving realignment of myosin-binding sites (12, 45), myosin regulatory light chain phosphorylation (16), a thin filament-based model (6, 54), a model based on the interacting heads motif (IHM) (26), and a myosin isoform kinetic model (67). All of these models have in common a transient increase in the number of cross-bridges bound to actin.

Our previous research with Drosophila muscle types suggests that there exist at least two different SA mechanisms, a myosin-isoform-based mechanism that differentiates muscles possessing minimal and moderate SA force generating ability and a thin filament-based mechanism that enables the even higher SA force-generating ability of IFM (17, 73, 74). The evidence for the myosin isoform mechanism comes from experiments where a myosin isoform [embryonic (EMB)] from a slow, cyclically contracting larval muscle was expressed in the minimally SA jump muscle, which rapidly shortens to power jumping. The EMB isoform exhibited increased F_SA, F_SA/F₀, and power generation when expressed in jump muscle (74), essentially transforming a muscle with minimal SA to one with moderate SA. A critical component of this transformation was elevating [P_i], as increased SA properties were seen only above ∼4 mM P_i. This led us to suggest a phosphate-dependent myosin isoform mechanism for moderate SA muscle types. These findings have inspired us to ask: Does our myosin-isoform-based SA mechanism apply to other species besides Drosophila? If so, is myosin-based SA muscle type specific, and is there a potential benefit to having SA in some muscle types, especially at higher [P_i]?

To answer these questions, we turned to mouse skeletal muscle and examined muscle fiber types that naturally express different myosin isoforms, because the myosin exchange experiment we conducted with Drosophila cannot currently be performed in a mammalian model. We performed skinned fiber mechanics on mouse soleus fibers, which predominantly express the slow-twitch myosin heavy chain (MHC) I and IIA isoforms and extensor digitorum longus (EDL) fibers, which express the fast-twitch MHC IIX and IIB isoforms (34). We found that soleus F_SA increased with [P_i], the same as what we found for Drosophila EMB in jump muscle, suggesting that soleus fibers use the same myosin-isoform-based mechanism for increasing F_SA. However, EDL F_SA decreased with [P_i], leading us to propose a different mechanism for EDL’s response to P_i and stretch. Because high [P_i] and impaired calcium cycling occur during prolonged muscle use, the reason for higher F_SA in soleus than in EDL may be to help maintain force-generating ability during extended slow-twitch muscle use.

MATERIALS AND METHODS

Muscle fiber preparation.

Soleus and EDL muscles were removed from adult (6–20 wk) female CD-1 mice (n = 6; Charles River Breeding Laboratories, Raleigh, NC). The muscles were collected posteuthanasia from animals that were controls for other experiments approved by the University of Massachusetts Institutional Animal Care and Use Committee. All mice were raised under standard housing conditions at room temperature and supplied with breeder’s chow ad libitum. Muscle tissue was immediately placed into cold (4°C) dissecting solution {20 mM N,N-bis[2-hydroxyethyl]-2-aminoethanesulfonic acid (BES), 5 mM ethylene glycol-bis(2-amino-ethylether)-N,N,N’,N’-tetraacetic acid (EGTA), 5 mM MgATP, 1 mM free Mg²⁺, 1 mM dithiothreitol (DTT), and 0.25 mM P_i} with an ionic strength of 175 mEq, pH 7.0, and at pCa 8.0 for isolation of single fiber bundles for mechanical measurements. Muscle bundles of ∼75 fibers were dissected and tied to glass rods and placed in skinning solution, a low-Ca²⁺ solution that begins the removal of the muscle fibers’ external membrane (70) [170 mM potassium propionate, 10 mM imidazole, 5 mM EGTA, 2.5 mM MgCl₂, 2.5 mM Na₂H₂ATP, protease inhibitor (Roche) with an ionic strength of 175 mEq, pH 7.0] for 24 h at 4°C. An osmotic pressure gradient, induced by the addition of glycerol, was used to further permeabilize the muscle fibers’ external membrane. This was accomplished by transferring the fiber bundles to storage solution (identical to skinning solution, but with 1 mM sodium azide and without protease inhibitor) and stepwise adding increasing amounts of glycerol to the solution: 10% vol/vol glycerol for 2 h, 25% glycerol vol/vol for 2 h, and 50% vol/vol glycerol for 2 h. Thereafter, bundles were stored in 50% vol/vol glycerol that included the protease inhibitor at −20°C (glycerol protects the fibers from freezing damage) until single fibers were isolated for mechanical measurements, which occurred within 4 wk of the dissection.

Preparation for mechanical measurements.

On the day of the experiment, a muscle bundle was removed from storage and placed in dissecting solution with 1% vol/vol Triton X-100 for 20 min at 4°C. Triton X-100 removes more of the muscle fiber membrane (sarcolemma) and, importantly, breaks up and removes internal membranes such as the sarcoplasmic reticulum (63). This prevents any internal calcium storage or release and enables the experimenter to control fiber activation level. Afterwards, individual fibers (∼1 mm in length) were isolated from muscle bundles using a pair of fine forceps, and aluminum T-clips were placed on both ends while in dissecting solution. Individual fibers were further demembranated (dissection solution with 1% vol/vol Triton X-100) for 30 min at 4°C to ensure complete removal of sarcolemma and sarcoplasmic reticulum. Side and top fiber diameters were measured in dissecting solution at the midpoint of the fiber to determine its height-to-width ratio. The fiber was mounted on the muscle mechanical apparatus (see Experimental apparatus below) by sliding the holes in the T-clips onto hooks attached to a piezoelectric motor and a strain gauge. The hooks and fiber were initially submerged in relaxing solution (pCa 8.0, 10 mM MgATP, 0 mM phosphate, 45 mM creatine phosphate, 450 U/ml creatine phosphokinase, 1 mM free Mg²⁺, 5 mM EGTA, 20 mM pH 7.0 BES, and 260 mM ionic strength, adjusted with Na methanesulfonate, 1 mM DTT) at 17°C. We calculated the amount of ingredients that needed to be added to our relaxing solution and our other fiber bathing solutions to obtain these concentrations based on the previously reported association constants using a computer program that accounts for the multiple equilibria of ions (3, 23). The 260 mM ionic strength was needed to accommodate 16 mM P_i and 20 mM ATP and was constant at all [P_i]s. P_i competes with MgATP for rigor myosin, especially in fast muscle fiber types, and thus high [ATP] is needed to minimize competition from P_i when working at higher [P_i]s. Additionally, high concentrations of MgATP and regeneration system components help prevent rigor myosins from forming in the interior of the muscle (i.e., rigor core) by increasing diffusion gradients between the solution and interior of the fiber. The fiber bathing solutions for mechanical measurements were the same ones used in our previous Drosophila SA experiments. Sarcomere length was set to 2.65 µm. Cross-sectional area was determined by measuring fiber width on the apparatus’s compound microscope and estimating height by multiplying width by the height-to-width ratio measured during fiber preparation and presuming the fiber cross-sectional area is elliptical.

Experimental apparatus.

A custom-built muscle mechanics apparatus was used to measure isometric tension and to record tension responses to transient length changes. The apparatus was a new, duplicate version of a previously described muscle mechanics apparatus (49). Briefly, a bath plate of 13 wells (∼100 μl) was made to hold experimental solutions and a single large chamber (∼450 μl) for mounting the fiber onto hooks attached to an Akers force gauge (AE-801; SensorOne, Sausalito, CA) and piezo actuator linear motor (P-841.10; Physik Instrumente, Auburn, MA). The plate slides horizontally (x-axis) within a plastic trough, past the fixed motor and force gauge, allowing the fiber to be exchanged between chambers. During solution changes, the time that the fiber was in the air was minimal (∼1 s). The bathing solutions were maintained at a constant temperature by circulating cooling solution through channels milled into the chamber walls. The bathing solution assembly was mounted to an inverted microscope (Zeiss Invertiscope) with a video camera and custom video analysis software that enabled precise measurements of fiber dimensions and sarcomere length.

Stretch activation.

Isometric tension in relaxing solution, pCa 8.0, was measured by slacking the fiber, establishing baseline tension, and returning the fiber to a sarcomere length of 2.65 µm. A passive tension transient was collected by stretching the fiber by 1% of its optimal muscle length over a period of 4 ms. The fiber was held stretched at this length for 2 s and was then returned to its original length over 8 s. The fiber was exchanged into pre-activating solution (same as relaxing, except that EGTA was decreased to 0.5 mM) for 30 s, followed by activating solution (pCa 5.0, 20 mM MgATP, 16 mM phosphate, 25 mM creatine phosphate, 450 U/ml creatine phosphokinase, 1 mM free Mg²⁺, 5 mM EGTA, 20 mM, pH 7.0, BES, 260 mM ionic strength, adjusted with Na methanesulfonate, 1 mM DTT).

After calcium activation, isometric tension was recorded. SA was induced by repeating the same length step (1% over 4 ms, 2-s hold, 8-s return) as that performed at pCa 8.0. The fiber was directly transferred (without returning it to relaxing solution) from 16 to 8, 4, 2, and 0 mM P_i and back to 16 mM P_i, with the length step repeated at each [P_i]. Fibers were exposed to activation solution from high to low [P_i] so that isometric tension started low and increased throughout the protocol. This prolonged fiber stability, which was monitored by performing an isometric tension measurement after each length step. We tested additional fibers (n = 3 for soleus and n = 5 for EDL, data not shown) in the opposite order (starting at 0 mM P_i and increasing to 16 mM P_i), and the same trends were observed. The magnitude of the total stretch-activated force (A_SA) was measured at the peak of the delayed force rise, phase 3 (74), which typically occurred between 30 and 50 ms for EDL fibers and between 60 and 200 ms for soleus fibers following the initiation of the stretch. In fibers where phase 3 was not obvious, which happened primarily at low [P_i], A_SA was determined based upon what point in time this typically occurred in other fibers of the same type and at the same [P_i].

Stretch activation analysis.

The net stretch-activated tension (F_SA), defined as the difference between total phase 3 tension amplitude at pCa 5.0 (A_SA) and passive phase 3 amplitude measured at pCa 8.0 (P_SA), was calculated to help exclude contributions from passive sarcomere elements not found in cross-bridges from our SA tension value. However, we note that titin may become stiffer with increased [calcium] as calcium may bind to titin (38) and increase its stiffness, which is why we also provide total SA tension values (A_SA) in our results (Table 1). To compare the relative tensions generated by the two activation mechanisms, we calculated F_SA/F₀ for each fiber type at each [P_i], where F₀ is calcium-activated isometric tension. F₀ was determined by subtracting passive isometric tension measured at pCa 8.0 (P₀) from total isometric tension at pCa 5.0 (A₀). To determine how much F_SA contributes to overall tension generation, we calculated F_SA/(F_SA+F₀) for each fiber type at each [P_i].

Table 1.

Effects of P_i on SA of single mouse soleus and EDL muscle fibers

[Pi], mM	PSA, mN/mm2	ASA, mN/mm2	FSA, mN/mm2	FSA/F0, %	FSA/(FSA + F0), %
Soleus
0	0.9 ± 0.1	7.5 ± 0.9#*	6.6 ± 0.9#*	10.5 ± 1.3#*	9.3 ± 1.1#*
2		8.6 ± 1.0#	7.7 ± 1.0#	15.3 ± 1.7#*	13.0 ± 1.4#*
4		9.2 ± 1.0#	8.3 ± 1.0#	20.6 ± 2.1#	16.7 ± 1.6#
8		10.0 ± 1.1abc	9.1 ± 1.1abc	28.6 ± 2.8#	21.5 ± 1.9#
16		10.3 ± 1.1abc*	9.4 ± 1.1abc*	43.3 ± 5.6#*	28.6 ± 2.7#*
EDL
0	0.8 ± 0.1	10.2 ± 0.9e	9.5 ± 0.9e	16.7 ± 1.2de	14.1 ± 0.9d
2		10.2 ± 0.8e	9.4 ± 0.8e	19.7 ± 1.1d	16.3 ± 0.8
4		9.9 ± 0.8e	9.1 ± 0.8e	20.3 ± 1.4	16.6 ± 1.0
8		8.6 ± 0.9e	7.8 ± 0.9e	25.0 ± 2.1ab	19.6 ± 1.3a
16		6.5 ± 0.9#	5.7 ± 0.9#	23.6 ± 2.4a	18.6 ± 1.5

Values are means ± SE; n = 17 for soleus; n = 19 for extensor digitorum longus (EDL). A_SA, total stretch-activated (SA) tension (pCa 5.0); F_SA, net stretch-activated tension (F_SA = A_SA – P_SA); F_SA/(F_SA + F₀), net SA tension as % total tension; F_SA/F₀, net stretch-activated tension as a % net isometric tension; P_i, inorganic phosphate; [P_i], inorganic phosphate concentration; P_SA, passive SA tension (pCa 8.0); SA, stretch activation.

^#Significantly different from all other P_i values within a muscle type. Otherwise, superscript letters indicate statistically significant difference between that concentration of P_i and the indicated concentration:

^a0 mM,

^b2 mM,

^c4 mM,

^d8 mM, and

^e16 mM P_i.

^*P < 0.05 compared with EDL at the same [P_i].

The rates of phases 2–4 were determined by fitting tension responses to the sum of three exponential curves (Eq. 1), with r₂ the rate of rapid tension decrease that occurs during phase 2, r₃ the rate of increase in delayed tension that occurs in phase 3 (SA), and r₄ the rate of tension decay that occurs during phase 4, and the offset accounts for possible differences in starting tension amplitude:

$$T=a_2{e}^{\left(-r_{2}t\right)}+a_3\left[1-e^{\left(-r_{3}t\right)}\right]+a_4{e}^{\left(-r_{4}t\right)}+\text{offset}$$ (1)

Statistics.

A repeated-measures analysis of variance (ANOVA), with [P_i] (0, 2, 4, 8, and 16 mM) as the repeated measure, was performed for each muscle group (soleus and EDL) to examine the effect of [P_i] on all outcomes of interest. A mixed ANOVA, with [P_i] as the within-subjects factor and muscle group as the between-subjects factor, was conducted to determine the presence of an interaction between P_i and muscle group on the same outcomes (e.g., if P_i affected outcomes differently in soleus than in EDL). When the assumption of sphericity was violated for either ANOVA, the Greenhouse-Geisser correction was used. The sample size (number of fibers) for the data in Fig. 4 (i.e., r₂, r₃, and r₄) is lower than the other two tables because we could not fit all tension transients at all [P_i]s, especially 0 mM Pi, with Eq. 1. Using fibers that were missing a [P_i] concentration would have skewed the data and not allowed us to perform a repeated-measures ANOVA. Thus, we used only fibers that could be fit with Eq. 1 at all [P_i]s. Independent-sample t-tests were used to compare values from soleus and EDL fibers (i.e., differences between muscle groups) at each [P_i]. All statistical analyses were conducted using SPSS for Windows version 25.0 (IBM, Armonk, NY), and significance was set at P ≤ 0.05.

Fig. 4.

Kinetics of stretch activation at different P_i concentrations ([P_i]). Individual tension traces following a 1% increase in muscle length over 4 ms were fitted to the sum of three exponentials (see Eq. 1) to obtain the rates of phase 2, r₂ (rate of rapid tension decrease) (A) phase 3, r₃ (rate of delayed tension increase) (B), and phase 4, r₄ (rate of slow tension decrease) (C). All values are means ± SE; n = 8 soleus fibers; n = 13 extensor digitorum longus (EDL) fibers. *P < 0.05 compared with EDL at the same [P_i]; #significantly different from all other P_i values within a muscle type. Otherwise, superscript letters after values indicate a statistically significant difference between that concentration of P_i and the indicated concentration: a, b, c, d, e = 0, 2, 4, 8, and 16 mM Pi, respectively.

RESULTS

Stretch and calcium-activated tension at 0 mM P_i.

Stretching soleus and EDL fibers produced the four expected transient tension phases at [P_i]s ≥2 mM (Fig. 1). However, at 0 mM, phase 3 was not always visually apparent. A definitive phase 3 peak was observed in 41% of soleus fibers (7 of 17) but not in any EDL fibers. There was usually at least a discernible change in slope of the rapid decaying portion of the transient after phase 1, indicating the presence of phase 3 at 0 mM P_i. This occurred in ∼50% of EDL fibers and ∼80% (including the 41% that showed the definitive phase 3 peak) of soleus fibers. This lack of a visible phase 3 peak under certain experimental conditions has been reported previously (1). To obtain the SA amplitude values for these cases, slope changes and times from fibers with definitive phase 3 peaks were used to help estimate at what time point to measure A_SA.

At 0 mM P_i, the passive (pCa 8.0) phase 3 amplitude values (P_SA) were not different between fiber types. Total phase 3 amplitude values (A_SA) and active (pCa 5.0) phase 3 amplitude values (F_SA) were greater in EDL than in soleus fibers (Table 1 and Fig. 2A). For calcium-activated tension values, there was no difference in total isometric tension (A₀) and net calcium-activated tension (F₀), but there was a twofold higher amount of passive isometric tension (P₀) in soleus than in EDL fibers (Table 2 and Fig. 2B). To put the amount of F_SA in perspective, we divided F_SA by F₀ (normalizing SA tension to calcium-activated isometric tension), generating a value of 17% for EDL that was greater than the 11% soleus value (Table 1). Normalizing to total tension [F_SA/(F_SA + F₀)] resulted in values of 14 and 9%, again with EDL slightly greater (Table 1).

Fig. 2.

A: stretch-activated tension, F_SA, increased for soleus fibers, but decreased for extensor digitorum longus (EDL) fibers with increasing P_i concentration ([P_i]). B: calcium-activated isometric tension, F₀, decreased with increasing [P_i] for both soleus and EDL fibers. All values are means ± SE; n = 17 soleus fibers; n = 19 EDL fibers. *P < 0.05 compared with EDL at the same [P_i]. For statistical analysis of [P_i] effects, see Tables 1 and 2.

Table 2.

Effects of P_i on calcium-activated isometric tension of single mouse soleus and EDL muscle fibers

[Pi], mM	P0, mN/mm2	A0, mN/mm2	F0, mN/mm2
Soleus
0	6.3 ± 0.8	71.1 ± 3.9#	64.8 ± 3.9#
2		57.2 ± 3.1#	50.8 ± 3.0#
4		47.0 ± 2.9#	40.7 ± 3.0#
8		38.1 ± 2.4#	31.8 ± 2.5#
16		29.5 ± 2.1#	23.2 ± 2.2#
EDL
0	2.8 ± 0.3	62.1 ± 4.9#	59.3 ± 4.7#
2		51.8 ± 3.9#	49.0 ± 3.7#
4		49.2 ± 3.9#	46.4 ± 3.7#
8		35.2 ± 3.1#	32.4 ± 3.0#
16		27.3 ± 2.5#	24.5 ± 2.3#

Values are means ± SE; n = 17 for soleus; n = 19 for extensor digitorum longus (EDL). A₀, total isometric tension (pCa 5.0); F₀, net Ca²⁺-activated isometric tension (F₀ = A₀ – P₀); [P_i], inorganic phosphate concentration; P₀, passive isometric tension (pCa 8.0).

^#Significantly different from all other P_i values within a muscle type. No differences were found between soleus and EDL at the same [P_i].

Stretch and calcium-activated tension at higher [P_i].

At higher [P_i], the phases of the tension transients were better defined for both fiber types (Fig. 1). Both fiber types also showed very similar decreases in calcium-activated tension with increasing [P_i] (Fig. 2B). By 16 mM Pi, both muscle types had lost 59–64% of their calcium-activated tension, i.e., F₀ (Table 2). This isometric tension loss is a well-known muscle fiber phenomenon (11). The amount of isometric tension loss we observed was similar in magnitude to that reported previously, with investigator’s results differing (higher, lower, or the same) for whether fast- and slow-twitch fiber types’ isometric tensions are equally affected by P_i (20, 48, 51, 56, 64). However, we found a response of F_SA to [P_i] that, to our knowledge, has not previously been observed in mammalian skeletal muscle fibers. Specifically, soleus and EDL showed opposite F_SA trends in response to [P_i], with soleus increasing and EDL decreasing (Fig. 2A). In soleus, F_SA increased 1.4-fold (Table 1), and its F_SA/F₀ ratio increased more than fourfold to 43% by 16 mM P_i because F_SA increased and F₀ decreased with greater [P_i] (Table 1 and Fig. 3A). In contrast, for fast EDL fibers, F_SA decreased by 40%, and the F_SA/F₀ ratio increased by only 41% by 16 mM P_i. As a percentage of total force [F_SA/(F_SA + F₀)], SA tension increased (9 to 29%) with [P_i] in soleus fibers but remained relatively unchanged (14–19%) in EDL fibers (Table 1 and Fig. 3B). In other words, soleus force generation from SA is enhanced under conditions where calcium-activated tension is declining. This contrasts with EDL, where both stretch and calcium-activated force generation decreased with increasing [P_i] (Fig. 2, A and B).

Fig. 3.

A: stretch-activated tension, F_SA, normalized to calcium-activated isometric tension, F₀. B: F_SA, normalized to total tension production, F_SA + F₀. These normalizations show that stretch activation becomes a major contributor to soleus fiber force production as P_i concentration ([P_i]) increases. All values are means ± SE; n = 17 soleus fibers; n = 19 EDL fibers. *P < 0.05 compared with extensor digitorum longus (EDL) at the same [P_i]. For statistical analysis of [P_i] effects, see Tables 1 and 2.

Kinetics of stretch activation.

Fitting the tension transients with Eq. 1 revealed that the rates of phases 2 and 3, r₂ and r₃, were faster for EDL than for soleus, confirming that these muscles contain fast and slow fiber types, respectively. In both fiber types, r₂ and r₃ increased with [P_i] (Fig. 4). By 16 mM P_i, r₂ and r₃ had increased 3.6-fold and 5.5-fold for EDL, respectively, while for soleus the increases were 2.0-fold and 2.5-fold, respectively. Values for r₄ did not change with P_i and did not differ between fiber types (Fig. 4).

DISCUSSION

We observed that both soleus and EDL fibers exhibit SA force generation. However, as [P_i] increased, the two fiber types showed opposite [P_i] dependencies. Soleus F_SA increased with [P_i], while EDL F_SA decreased. The F_SA results in soleus are particularly intriguing because they are in stark contrast with P_i’s well-known negative influence on calcium-activated force generation. This contrast was especially apparent in the quadrupling of the soleus SA to calcium-activated tension ratio (e.g., F_SA/F₀), from 11% at 0 mM P_i to 43% at 16 mM P_i. Furthermore, at 16 mM P_i soleus F_SA accounted for almost 30% of total force production. Based upon these results, we propose a physiological reason for SA in skeletal muscle fibers and, incorporating our previous Drosophila work (73, 74), possible molecular mechanisms behind this fiber type-specific SA response.

Physiological relevance of skeletal muscle SA.

Does the increase in F_SA with [P_i] in slow-twitch fibers have physiological relevance to muscle and locomotory performance? For SA to be potentially helpful, a muscle must be going through repetitive cyclical contractions such as insect IFM powering flight (32) or, as we propose here, antagonistic leg muscles during running. The antagonistic muscle stretches the agonist muscle immediately before the agonist shortens. The delayed tension increase (F_SA), brought about by stretch, must be timed to occur as the agonist muscle is shortening. This results in higher force during shortening, producing greater useful work and power for locomotion. To roughly test whether the timing of our in vitro SA kinetics match the sarcomere length change rates used by mice during locomotion, we overlaid the soleus muscle length change pattern while running with fits of our 2 and 16 mM P_i SA traces from a soleus fiber (Fig. 5). The soleus sarcomere length change patterns for a mouse running with stride frequencies of 5 and 8 Hz are based on the work of others (29). Synchronizing the start of the lengthening portions of the running pattern and our soleus SA tension transient traces shows that the increased force due to the phase 3 peaks (SA peaks) occurs when the muscle powering running is shortening. This looks highly promising for skeletal F_SA to be physiologically relevant because augmenting force during shortening is the most useful time during a contraction cycle to increase work and power output from the muscle (31). Furthermore, at higher running speeds, [P_i] is proportionally greater in muscle (14, 18). Our tension trace fits show that higher [P_i] shifts the increased force from the SA peak to the left, which better matches the timing of the faster running speed than if the effect of increasing P_i had instead decreased or not changed SA kinetics (compare Fig. 5, A and B). Thus, the kinetic response of SA to [P_i] is also beneficial for adjusting to different running speeds. In other words, the effect of [P_i] on the time interval between stretch and phase 3 positively correlates with mouse stride rate and in vivo [P_i].

Fig. 5.

Comparison of in vivo soleus sarcomere length change timing with in vitro soleus stretch activation kinetics. A: fit of Eq. 1 to a soleus fiber’s tension response at 2 mM P_i (purple) is shown with its y-axis on the left side of the graph. Horizontal black arrows above the stretch activation (SA) trace indicate areas of increased force production due to SA (peak of phase 3). In vivo sarcomere length change pattern (y-axis on right side of the graph) of a soleus muscle when a mouse is running at a stride frequency of 5 Hz (trotting, red), based on video analysis of running mice and their associated sarcomere length measurements (29). The start of in vivo and in vitro lengthening is set to 0 s to enable comparisons. The area of increased force generation due to SA coincides with muscle shortening, suggesting that SA could be highly beneficial for increasing soleus work and power output while the mouse is running. B: comparison of the effects of increased P_i concentration ([P_i]) with faster running speed. Fit of Eq. 1 to the same soleus fiber’s tension response at a higher P_i concentration, 16 mM P_i (blue). Horizontal black arrows above the SA trace indicate areas of increased force production due to SA (peak of phase 3). In vivo sarcomere length change pattern (y-axis on right side of the graph) of a soleus muscle when a mouse is running at a stride frequency of 8 Hz (galloping, green). The majority of increased force at 16 mM P_i overlays the shortening portion of the faster 8 Hz length change. That the timing of the faster running speed better matches the timing of enhanced force at higher [P_i] (than vice versa) is beneficial because faster running speeds increase muscle [P_i].

The exact time alignment of the peak of phase 3 and the in vivo sarcomere patterns in Fig. 5 will be influenced by a number of factors that we will need to investigate in future studies. For example, the higher in vivo temperature, 37°C, would shift the SA peak to the left, while factors such as lower [calcium] and slower lengthening (closer to in vivo conditions) would likely shift the SA peaks to the right. Thus, these current alignments are only to show plausibility, and more studies are required to test our hypothesis that SA helps modulate force in slow-twitch fibers during locomotion.

While slow-twitch fibers are known to have higher [P_i] than fast-twitch fibers, ∼6 mM P_i compared with 1 mM P_i, respectively, under resting conditions (37), and thus could potentially benefit from F_SA as soon as locomotion starts, the most benefit from F_SA for mouse soleus muscles is likely during longer-duration running (trotting and possibly galloping). As running starts, calcium fluxes are likely sufficient to modulate muscle force levels as needed for locomotion. However, as a muscle continues to be used, studies have shown that calcium activation ability decreases from both impaired nerve firing and decreased calcium release and re-uptake (2, 30, 43). This is likely exacerbated at faster speeds, where there is less time for calcium cycling. Additionally, [P_i] increases with extended use from myosin, calcium, and Na/K ATPases (11, 14, 18) to further impair calcium-activated force production.

In contrast to impaired calcium-activated force, our study shows P_i increases soleus F_SA and that SA contributed nearly 30% of total force production at 16 mM P_i. There are reasons to think that this percentage is even higher during extended locomotion. Impaired calcium cycling likely causes subsaturating fluxes in calcium concentration, rather than the fixed saturating concentration used here, and some studies in heart and insect flight muscle suggest that SA is most prominent at subsaturating levels (40, 62). Furthermore, shortening deactivation (SD), the opposite of SA (31), is also likely occurring in soleus during running. SD causes a transient decrease in force following rapid muscle shortening, with the same but opposite four phases. If we presume an equal drop in force from SD as the force increase we observed for SA, then working together, SA and SD would modulate ∼60% of total force production at 16 mM P_i. An expected benefit from high SA would be recovery of work done to stretch the muscle during subsequent shortening. A study of work generated during a lengthening and shortening protocol on live muscle fibers (presumably with [P_i] between the resting value of 6 mM and the 30–40 mM range that occurs during extreme exercise; see Ref. 69)) found that mouse soleus fibers returned a greater percentage of work than EDL fibers (65). Although several possible mechanisms were proposed, one was recruitment of additional cross-bridges by stretch (i.e., SA) in soleus muscle. Thus, SA may have evolved to offset the decline in calcium-activated force and work production, enabling slow-twitch muscle types to power extended locomotion.

Mechanism of skeletal muscle SA.

Our mammalian skeletal muscle measurements of SA characteristics are very similar to what we observed for Drosophila jump muscle fibers expressing different myosin isoforms (73, 74). Jump muscle expressing EMB myosin displayed the same F_SA response to [P_i] as soleus, while wild-type jump muscle was similar to EDL. Combined with our current mammalian results, we now propose two possible P_i-dependent mechanisms for how different myosin isoforms influence skeletal muscle F_SA. The first mechanism is for myosin isoforms, such as those found in soleus, that enable muscles to produce moderate F_SA levels at high P_i (Fig. 6A). When a moderate SA muscle fiber is stretched (phase 1), some of the strongly bound post-power stroke cross-bridges rebind P_i and are forced backwards into the weakly bound pre-power stroke state, causing a transient increase in the population of weakly bound cross-bridges and the phase 2 decline in force. The higher the concentration of P_i present, the greater the probability of this backward step occurring. The temporary increase in the number of cross-bridges in weakly bound states increases the number of cross-bridges available to rapidly bind and contribute to force generation during phase 3. An important aspect of myosin from moderate SA muscle fibers would be their ability to go through the power stroke again without having to release P_i and ADP and hydrolyze ATP. This is the same mechanism we proposed to explain our Drosophila EMB myosin F_SA results (74).

Fig. 6.

Proposed stretch activation (SA) mechanisms for mammalian skeletal muscle fibers. The standard, accepted cross-bridge cycle pathway (42) is shown outlined in blue, while our proposed pathways for cross-bridges subjected to stretch and P_i are shown outlined in green for soleus (A) and purple for extensor digitorum longus (EDL) (B). The green-colored lever arm of myosin indicates a pre-power stroke state, while the red lever arm indicates a post-power stroke state. For both fiber types, the detachment of myosins (with ADP and P_i bound) from actin due to stretch and P_i binding is responsible for the phase 2 decrease in tension (see Fig. 1A). However, for soleus, the stretch and P_i reverse the myosin power stroke (green arrow), making this population of weakly bound myosins, along with the other myosins in this same state, primed to reattach and generate greater force than if stretch had not occurred. This reattachment is responsible for the phase 3 force regeneration and phase 3 peak (stretch activation). Because we do not know whether P_i binds to myosin and then stretch alters the myosin conformation, or vice versa, we do not indicate any intermediaries in the green pathway. For EDL, the pathway differs from the soleus in that P_i and stretch do not reverse the power stroke (purple arrow pathway); instead, myosin detaches in response to stretch when in the presence of P_i. We do not mean to imply that the green pathway is exclusively available to soleus myosin and the purple only to EDL myosins. A small fraction of EDL myosins likely take the green pathway (shown in A), and this may explain why some SA occurs in EDL fibers, but their probability of using the purple pathway becomes higher with increasing P_i concentration ([P_i]).

The hypothesis that P_i allows cross-bridges to move back into the weakly bound pre-power stroke state from the strongly bound post-power stroke state is well supported by many studies and is one explanation for how increasing [P_i] reduces calcium-activated isometric force (10, 13, 25, 35, 47, 52, 66). Several studies also support the idea that stretch can strain the strongly bound myosin and cause its transition into the weakly bound, pre-power stroke position (9, 41, 44, 55). Our addition to these hypotheses is that only some myosin isoforms (e.g., from slow-twitch skeletal muscle) are able to store the energy from stretch, which is used to generate a second power stroke, without the need for hydrolyzing another ATP molecule. The free energy for the reversal, and possibly another power stroke, could come from the stretch as put forth by theoretical models (4).

In contrast, for minimal SA muscle types such as EDL, we hypothesize that the power stroke is not as easily reversed. In these fibers, stretch might occasionally drive strongly bound, post-power stroke myosin heads backwards into a weakly bound, pre-power stroke state at low [P_i] (which may be why similar F_SA values were measured for soleus and EDL at low [P_i]). However, as [P_i] is increased, stretch is more likely to cause detachment of strongly bound myosin heads from actin without a reversal of their power stroke. We propose that the EDL myosin heads then take an off-actin pathway to release P_i and ADP (Fig. 6B). Even if they do rebind to actin, they are not able to contribute to SA because these EDL heads (with ADP and P_i bound) are at a lower energy state than the soleus heads (with ADP and P_i bound), because they did not store energy from stretch. This mechanism is very similar to those proposed by Linari et al. (39) to explain the effects of P_i on fast skeletal muscle isometric force production and by Debold et al. (15) to explain the influence of P_i on fast skeletal myosin when propelling actin in the in vitro motility assay. However, we note that the effects of P_i, as well as the underlying mechanism(s) for those effects, on isometric force, unloaded filament sliding, and F_SA are not necessarily expected to be the same.

Other possible mechanisms that may be contributing to SA in one or both muscle types include stretch-slowing rates of some cross-bridge steps and briefly synchronizing a subgroup of myosin heads at force-generating steps of the cycle. There could also be some contribution from stretch delaying detachment of strongly bound heads, thus temporarily increasing force levels. This is suggested by experiments that imposed negative strain on myosin, which caused decreases in ADP release rate (33, 68). Another possibility is that strain on the thin filament realigns myosin-binding sites, leading to a transient increase in myosin binding (12, 45). However, none of these would likely be affected by increasing [P_i] and, therefore, are not likely to account for the SA differences we observed between soleus and EDL fibers as [P_i] increased.

In conclusion, we propose that slow-twitch fibers have evolved myosin isoforms that enhance SA. In contrast, fast-twitch fibers, which are not used for extended time periods, faced less selective pressure to evolve SA enhancing myosin isoforms. Our results suggest that mammalian skeletal muscle SA deserves more investigation and that in some skeletal muscle types SA should not be considered negligible compared with IFM and heart muscle SA.

GRANTS

This work was supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant R01 AR064274 to D. M. Swank.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.R.S., M.S.M., and D.M.S. conceived and designed research; C.R.S., K.M.B., M.S.M., and D.M.S. performed experiments; C.R.S., K.M.B., J.N.S., M.S.M., and D.M.S. analyzed data; C.R.S., K.M.B., J.N.S., M.S.M., and D.M.S. interpreted results of experiments; C.R.S., K.M.B., J.N.S., M.S.M., and D.M.S. prepared figures; C.R.S. and D.M.S. drafted manuscript; C.R.S., K.M.B., J.N.S., M.S.M., and D.M.S. edited and revised manuscript; C.R.S., K.M.B., J.N.S., M.S.M., and D.M.S. approved final version of manuscript.