Prolonged myosin binding increases muscle stiffness in Drosophila models of Freeman-Sheldon syndrome

Abstract

Freeman-Sheldon syndrome (FSS) is characterized by congenital contractures resulting from dominant point mutations in the embryonic isoform of muscle myosin. To investigate its disease mechanism, we used Drosophila models expressing FSS myosin mutations Y583S or T178I in their flight and jump muscles. We isolated these muscles from heterozygous mutant Drosophila and performed skinned fiber mechanics. The most striking mechanical alteration was an increase in active muscle stiffness. Y583S/+ and T178I/+ fibers’ elastic moduli increased 70 and 77%, respectively. Increased stiffness contributed to decreased power generation, 49 and 66%, as a result of increased work absorbed during the lengthening portion of the contractile cycle. Slower muscle kinetics also contributed to the mutant phenotype, as shown by 17 and 32% decreases in optimal frequency for power generation, and 27 and 41% slower muscle apparent rate constant 2πb. Combined with previous measurements of slower in vitro actin motility, our results suggest a rate reduction of at least one strongly bound cross-bridge cycle transition that increases the time myosin spends strongly bound to actin, t_on. Increased t_on was further supported by decreased ATP affinity and a 16% slowing of jump muscle relaxation rate in T178I heterozygotes. Impaired muscle function caused diminished flight and jump ability of Y583S/+ and T178I/+ Drosophila. Based on our results, assuming that our model system mimics human skeletal muscle, we propose that one mechanism driving FSS is elevated muscle stiffness arising from prolonged t_on in developing muscle fibers.

Significance

Freeman-Sheldon syndrome is a rare skeletal muscle disease inflicting stiff muscle contractures in the hands, feet, and face. These abnormalities occur during fetal development yet have lasting and debilitating effects on the patient. The mechanism of the disease is largely unknown, and there is no effective treatment besides invasive surgery. Using Drosophila models of Freeman-Sheldon syndrome, we elucidated muscle mechanical and cross-bridge cycle disease defects including increased muscle stiffness and time myosin spends bound to actin. This newfound understanding will increase the probability of discovering new treatments for these patients.

Introduction

Freeman-Sheldon syndrome (FSS) is characterized by debilitating muscle contractures, including facial muscle contractures, extreme camptodactyly and clubfoot, and severely inhibited range of limb motion (1, 2, 3). FSS is the most severe form of a group of diseases with similar symptoms known as distal arthrogryposis (DA), with FSS classified as DA type 2A. Current treatment of FSS is limited but largely involves highly invasive surgical correction (4, 5, 6). Development of better treatments is hindered by a lack of a mechanistic understanding of the disease.

FSS is caused by mutations in the embryonic myosin heavy chain gene (MYH3) (2), but we know very little about the mechanisms by which these mutations impact myosin and muscle structure and function and how these changes contribute to the disease phenotype. A few recent studies have started making progress toward investigating the mechanisms behind FSS mutations, including a study using expressed human embryonic myosin protein with the T178I, R672C, and R672H FSS mutations (7) and an investigation using human FSS patient muscle fiber biopsies (8).

Recently, Drosophila has become a powerful animal model for studying muscle disease mechanisms caused by mutations in myosin (9, 10, 11, 12). In addition to advantages like low cost and rapid life cycle, Drosophila is an outstanding model for analyzing myosin mutations because of its single muscle myosin heavy chain gene (Mhc), which gives rise to all myosin isoforms by alternative RNA splicing (13). Additionally, the myosin-null line Mhc¹⁰ allows for expression of only transgenic myosins in the indirect flight muscle (IFM) and jump muscles (also known as the tergal depressor of the trochanter (TDT)) (14). All model systems have some limitations, and in this case, it is unknown how well the muscles and myosin studied exactly replicate the muscles expressing the human embryonic myosin. The presumption here is that the mutations occur in highly conserved residues (61 and 84% for Y583 and T178 when comparing the specific residue along with its 10 N- and C-terminal amino acids with the equivalent human sequence, respectively (9)), which would result in similar changes in kinetics and mechanics across all muscle myosin isoforms and thus have a similar impact on muscle performance.

We and one other research group have started using Drosophila to analyze the Y583S, T178I, R672C, and R672H FSS mutations (9,10). At the molecular level, both groups found that these mutations caused decreased myosin ATPase activity, with the T178I mutation yielding the largest decrease, which parallels its clinical severity compared with other FSS mutations (15). The Y583S and T178I mutations slowed actin sliding velocity, but only Y583S increased actin affinity (9). At the muscle fiber level, heterozygous animals started to show IFM degradation at 2 days old and degradation increased with age (9). Sarcomere Z-disk distance was reduced in the FSS mutants, which also worsened with age (10). This disruption in structure likely contributed to observed locomotion impairment because FSS mutant organisms showed dramatically reduced larval crawling and adult flight ability.

Here, we aimed to elucidate a fiber and myosin level mechanism for FSS by focusing on the alterations myosin mutations Y583S and T178I cause to Drosophila IFM and TDT muscle functions. These muscles have the advantage of being amenable to a wide range of mechanical tests that give excellent insights into fiber and cross-bridge level mechanisms. Both mutations were studied in the heterozygous state to replicate the FSS patient heterozygous disease condition. We subjected IFM fibers to sinusoidal analysis, variations in ATP concentration, and work loop experiments to measure force, stiffness, work and power output, and rate constants of the cross-bridge cycle. Additional insight was gained by measuring the effects of the mutations on TDT structure, TDT activation and relaxation rates, and Drosophila jumping ability. The main fiber level finding was increased muscle stiffness, which caused decreased power output. Combining these results with our previous study of these two mutant fly lines revealed that a molecular mechanism likely contributing to the FSS phenotype is an increase in time myosin spends strongly bound to actin, with a minor contribution of slowed steps associated with myosin binding to actin and possible alteration of the myosin power stroke.

Materials and methods

Fly genetics

Detailed explanation of the generation of these FSS transgenic animals has been reported previously (9). Flies containing the FSS myosin transgene were crossed into a myosin-null (Mhc¹⁰) background so that this myosin is the only expressed copy of myosin in the IFMs and TDT (14). The control line for all experiments was pwMhc2, which is homozygous for full-length wild-type Mhc transgenes in the Mhc¹⁰ background. Experiments using this line have shown that it is indistinguishable from wild-type (16). Heterozygous mutant flies were created by crossing mutant stocks to pwMhc2 flies.

TDT ultrastructure analysis

Transmission electron microscopy was performed as described previously (17). Longitudinal sections of TDT muscles were prepared from 2-day-old female flies expressing Y583S or T178I myosin in the heterozygous form (Y583S/+ or T178I/+). Sections were compared with 2-day-old control samples (pwMhc2/+) to look for any changes in sarcomeric structure. Electron microscopy ultrastructural analysis was performed previously on IFMs from these fly lines (9).

IFM and TDT isolation

Adult female flies were collected just after eclosion and dissected 2 h later for IFMs and 24 h later for TDTs, as described previously (18). Only female flies were used because their larger size makes muscle dissection easier and to avoid potential dosage compensation issues in males because of the control allele being located on the X chromosome. Briefly, the thorax was isolated and split in half using micro scissors. The IFM bundle was removed with a tungsten wire probe and chemically skinned for 1 h in dissection solution (pCa 8.0, 5 mM MgATP, 1 mM free Mg²⁺, 0.25 mM phosphate, 5 mM EGTA, 20 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (pH 7.0), and 175 mM ionic strength, adjusted with Na methane sulfonate, 1 mM dithiothreitol, 50% glycerol, and 0.5% Triton X-100). Individual IFM fibers were separated from the bundle and split in half to a final diameter of ∼125 μm. TDTs were cut out of the thorax cuticle, chemically skinned in dissection solution, and paired down to 8–10 large diameter fibers. Multiple TDT fibers are used in the preparation because each individual fiber is only ∼10–20 μm in diameter (19). Both muscle types were secured to their respective muscle mechanics apparatuses using a pair of aluminum foil T-clips.

IFM mechanics: Sinusoidal analysis

The sinusoidal protocol has been described previously (18). Briefly, the holes of the T-clips holding an IFM fiber were placed on hooks connected to an electric piezo motor and a force transducer. Isometric tension was measured in relaxing solution (pCa 8.0, 12 mM MgATP, 30 mM creatine phosphate, 600 U/mL creatine phosphokinase, 1 mM free Mg²⁺, 5 mM EGTA, 20 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (pH 7.0) and 200 mM ionic strength, adjusted with Na methane sulfonate and 1 mM dithiothreitol) and activating solution (same as relaxing solution, but with calcium adjusted to pCa 5.0), and the net active tension produced by the muscle fiber was calculated as the difference between these measurements (total tension – passive tension = net active tension). Maximal power (P_max), the frequency at which maximal power was generated (f_max), and muscle fiber stiffness (elastic modulus) were measured using sinusoidal analysis. The fiber was oscillated at 0.125% of the muscle length (ML) over a range of 50 frequencies spanning 0.5–650 Hz. Because it is not possible to optically measure length of IFMs because of the individual myofibrils not being lined up with each other in a cross-striated pattern, we found optimal fiber length by sequentially stretching the fiber by 2% ML until power did not increase by more than 3%. As described previously, the viscous versus elastic moduli data (Nyquist plot) were fitted using the complex modulus equation:

$$\mathrm{Y}\left(f\right)=\mathrm{A}{(\mathrm{2}\pi if/\mathrm{a})}^{\mathrm{k}}-\mathrm{B}if/\left(\mathrm{b}+if\right)+\mathrm{C}if/\left(\mathrm{c}+if\right),$$ (1)

where Y is the complex modulus, f is the oscillation frequency, i is the square root of −1, a is 1 Hz, k is a unitless exponent, A represents the viscoelastic elements of passive structures, B and C are representative amplitudes of work-producing and work-absorbing parts of the cross-bridge cycle, and b and c are the relative frequencies of those amplitudes, respectively (18,20,21). The b- and c-values were converted from the time domain to the frequency domain by multiplying by 2π to obtain the apparent rate constants 2πb and 2πc of the work-generating and work-absorbing steps of the cross-bridge cycle, respectively.

IFM mechanics: Work loop analysis

Work loop experiments were used to measure work and power generation of IFMs with length oscillations that are similar to what IFMs experience during in vivo flight (22,23). The muscle fibers were oscillated at 0.25, 0.5, 0.75, or 1.0% ML at frequencies between 50 and 200 Hz. At each amplitude and frequency pair, the fiber underwent a series of 10 oscillation cycles, and work and power values were taken from cycles 7 or 8 because tension values become consistent after cycle 6 (24). ML was plotted against tension for a single contractile cycle, and the area within the trace is the amount of net work produced by the fiber. Power was calculated as net work × ML oscillation frequency. Net work and power values are positive when the work loop trace runs counterclockwise and negative when the trace runs clockwise.

IFM mechanics: ATP response

ATP concentration was decreased in a stepwise fashion from 20 to 0.5 mM using a 0 mM ATP activating solution. Total isometric tension was recorded at each ATP concentration. Myosin affinity for ATP was obtained from K_M- and V_max-values after fitting f_max and 2πc data with the Michaelis-Menten equation at the varying ATP concentrations. After the ATP response experiments, the ATP concentration was brought back to 10 mM to assess fiber degradation. Severely degraded fibers (more than 10% loss of function) were not included in data analysis.

TDT mechanics: Activation and relaxation experiments

To measure muscle activation and relaxation rates, we used a different mechanics apparatus (rig 2) than used for the IFM experiments (rig 1), and we used TDT fibers rather than IFM fibers. Rig 1 has only one fiber bathing well where bathing solutions are exchanged in and out of a solution bubble (under oil) via multiple partial exchanges. Rig 2 has multiple bathing chambers that the fiber can be switched between rapidly and enables recording of activation and relaxation time courses. Details of the rigs have been previously described (18). IFM fibers do not tolerate being switched between chambers, often tearing, but TDT fibers are not affected. Thus, we used TDT fibers, expressing the same mutations as IFM fibers, on rig 2 to collect activation and relaxation time courses. After being attached to rig 2 using T-clips, TDT muscles were stretched to their optimal sarcomere length, 3.6 μm (determined previously (25) in relaxing solution). To activate the muscle, the TDT was first transferred to a bathing chamber with preactivating solution, which contains lower EGTA levels (0.5 mM) than relaxing solution. Isometric tension was set to 0 (baselined) in preactivating solution. Tension was recorded as the muscle was transferred to activating solution, which caused the muscle to reach maximal tension. The muscle was kept in activation solution for 20 s and then transferred back to relaxing solution. Recording ceased when the muscle had fully relaxed, and the muscle was allowed to recover for 2 min before repeating this activation and relaxation procedure. This procedure was repeated between two and four more times for each fiber.

Isometric tension data acquired during the activation, and relaxation events were first processed by a 10-point moving average in Excel to reduce noise, especially the noise generated by the fiber being exchanged between bathing chambers. Activation and relaxation rates were analyzed two ways. The first method fit activation and relaxation traces with a logistic curve:

$$\mathrm{Y}\left(\mathrm{t}\right)=\frac{A}{1+e^{-k\left(t-B\right)}}+C+k_{1}t,$$ (2)

where Y is the tension at any given time point, t. k is the exponential variable describing the steepest part of the curve and is reported as the activation (k_Act) or relaxation (k_Rel) rate. k₁ is the slope of the linear portion of the trace after activation or relaxation, and A, B, and C are constants. This fit was applied to every activation and relaxation event for each fiber, and each parameter was averaged for all fits for each fiber genotype. The second analysis method was to measure the time interval from 10 to 90% of maximal tension activation and from 90 to 10% of maximal tension relaxation.

Locomotion assays

Locomotion assays were executed at 15°C to be comparable with the muscle mechanical measurements performed at 15°C. Free flight performance was tested on 2-day-old female flies by releasing them into a clear plexiglas chamber and scoring their flight pattern; 6 for upward flight (U), 4 for horizontal flight (H), 2 for downward flight (D), and 0 for no flight (N) (26). Flight index (FI) was calculated as FI = 6(U/T) + 4(H/T) + 2(D/T) + 0(N/T), where T is the total number of flies tested. Wing beat frequency was measured after 2-day-old female flies were fitted with a nylon tether and held in front of an optical tachometer (27). Jump distance was calculated by placing a single fly with its wings removed on a blank sheet of paper. A digital camera held above the paper recorded 10–12 jumps. Jump distance in centimeters was calculated using the open-source video analysis and modeling tool Tracker (http://physlets.org/tracker/). The three longest jumps for each fly were averaged.

Results

Muscle ultrastructure

Electron micrographs of TDT longitudinal sections from 2-day-old heterozygous FSS flies are normal compared with control flies (Fig. 1). Z-disks are intact in all three myofibrils pictured for each genotype, and M-lines are distinct. There is no evidence of hypercontraction or myofilament tearing within sarcomeres of the TDTs. IFM structure of FSS heterozygotes was analyzed previously in Fig. 3 of Rao et al. (9). It appeared normal during pupal and 2-h-old adult stages (Rao et al.; Fig. 3, B, F, and J) but started to degrade at 2 days (Rao et al.; Fig. 3, C, G, and K) and progressed with age (Rao et al.; Fig. 3, D, H, and K). Therefore, any degradation in muscle structure did not influence mechanical analysis of 2-h-old IFMs or 1-day-old TDTs.

Figure 1.

TDT ultrastructure. Shown are longitudinal sections of TDT muscles from control (pwMhc2/+) and heterozygous FSS (Y583S/+ and T178I/+) 2-day-old flies. M indicates the M-line and Z indicates the Z-disk, both of which appear normal in control and FSS fibers. The scale bar in the bottom left of the top panel represents 1.1 μm. The IFM electron microscopy data can be found in Fig. 3 of Rao et al. (9).

Figure 3.

FSS fibers’ response to changing ATP concentration. The f_max (A) and 2πc (B) data were fitted with a Michaelis-Menten curve (solid lines) to obtain V_max- and K_M-values. Both FSS mutants had a significantly lower V_max-value for the f_max data, but no significant changes in 2πc V_max (see Table 3 for values and statistical analysis). T178I/+ fibers had lower K_M-values (vertical dashed red lines) in both data sets compared with control line fibers (vertical dashed blue line). Y583S/+ K_M-values (vertical dashed yellow line) were not significantly different from the control lines. (C) Total isometric tension did not significantly change for Y583S/+ and control fibers as ATP concentration increased, but T178I/+ total isometric tension increased. n = 6 for all fibers except control 2πc data, where n = 5. Values are mean ± SEM. To see this figure in color, go online.

IFM muscle stiffness

The most striking mechanical alteration in this study was the dramatic increase in IFM muscle stiffness (i.e., elastic modulus) caused by both Y583S and T178I FSS mutations. Under relaxed conditions (pCa 8.0), the average values suggest a possible increase in passive stiffness for the Y583S/+ and T178I/+ mutants of 22 and 27%, respectively, compared with the control, but it was not significantly different at 500 Hz or any one specific frequency when analyzed using a standard one-way analysis of variance (ANOVA) (for example, p = 0.1 and p = 0.06 for Y583S and T178I, respectively, at 500 Hz) (Fig. 2 A; Table 1). However, using a repeated measure ANOVA or a nested one-way ANOVA resulted in an overall statistical difference in passive stiffness between both mutants and the control fiber (p < 0.0001). Once activated (pCa 5.0), FSS muscle fibers showed a larger increase in stiffness, especially in the high frequency range, compared with control fibers. Y583S/+ and T178I/+ muscle stiffness increased 70 and 82%, respectively, above control values at 500 Hz (Fig. 2 B; Table 1). These increases in elastic modulus values suggest that both mutant fibers either have more cross-bridges bound to the actin at any given time compared with the control or that each myosin cross-bridge is stiffer. At saturating ATP concentrations during the initial sinusoidal analysis portion of the protocol, the T178I/+ fibers exhibited a significant increase (29%) in total isometric tension production (4.17 ± 0.34 mN/mm²) compared with control fibers (3.22 ± 0.18 mN/mm², one-way ANOVA, p < 0.05). The average value of the net active isometric tension production was also higher for T178I/+ (0.90 ± 0.27 mN/mm²) compared with control (0.77 ± 0.21 mN/mm²) fibers, but the values were not significantly different. Passive tension of T178I/+ fibers (3.27 ± 0.26 mN/mm²) was not significantly different from control fibers (2.45 ± 0.30 mN/mm²). No significant differences in passive (2.79 ± 0.62 mN/mm²), total (3.69 ± 0.42 mN/mm²), or active (0.90 ± 0.36 mN/mm²) isometric tension values were observed for Y583S/+ fibers compared with control values.

Figure 2.

Muscle stiffness (elastic modulus), power, and work production of control, Y583S/+, and T178I/+ IFM fibers. (A) Passive stiffness, pCa 8.0, between mutant and control fibers is not different at 500 Hz (vertical dashed lines, one-way ANOVA, p < 0.05), but the ∼25% increase in FSS fibers is significant using a repeated measures and nested ANOVA looking at the effect of the mutations at all frequencies collectively. (B) Active stiffness (pCa 5.0) in FSS mutant fibers is dramatically different from control fibers at 500 Hz (one-way ANOVA, p < 0.05). (C) Positive power, measured by sinusoidal analysis, was generated between 30 and 230 Hz. T178I/+ maximal power is about half of the control, whereas Y583S/+ maximal power is not statistically different from control. However, Y583S/+ power is statistically reduced compared with the control at ML oscillation frequencies above 160 Hz, which is also the frequency at which control fibers generate maximal power (f_max). Vertical dashed lines indicate f_max. (D) Shown are representative work loop traces under conditions in which the control produced maximal power, 0.5% ML and 150 Hz (Table 2). Arrows indicate the direction of the work loop. The majority of the T178I/+ loop is in the clockwise direction, indicating net work absorption. In (A)–(C), n = 12 for both FSS mutants, and n = 13 for control fibers. In (D), n = 11 for Y583S/+ and control fibers, and n = 7 for T178I/+. Values are mean ± SEM.

Table 1.

Sinusoidal analysis parameters of IFM fibers

	Pmax (W/m3)	fmax (Hz)	2πb (s−1)	2πc (s−1)	Ee (kN/m2)	Eep (kN/m2)
Control	71 ± 7	158 ± 3	1261 ± 79	5046 ± 272	214 ± 15	288 ± 26
Y583S/+	54 ± 11	130 ± 5a	924 ± 64a	6696 ± 971a	364 ± 29a	366 ± 38
T178I/+	32 ± 6a	108 ± 3a	747 ± 28a	5230 ± 370	389 ± 23a	352 ± 17

P_max is maximal power, f_max is the frequency at which maximal power was generated, Ee is elastic modulus (active stiffness, pCa 5.0) measured at 500 Hz, E_ep is elastic modulus (passive stiffness, pCa 8.0) measured at 500 Hz. 2πb and 2πc are apparent rate constants of the cross-bridge cycle. Values are mean ± SEM. n = 12 for FSS fibers, and n = 13 for control.

^aIndicates significant difference from the control (one-way ANOVA, p < 0.05).

IFM power production

IFM power production was negatively impacted by the FSS mutations. T178I/+ IFMs generated less than half the amount of maximal power during sinusoidal analysis as control fibers (Fig. 2 C; Table 1). Y583S/+ fibers did not experience a decrease in maximal power production in these experiments (Table 1) but did generate significantly lower power than the control at higher frequencies (Fig. 2 C). Y583S/+ and T178I/+ fibers displayed decreased frequency at which maximal power is produced (f_max) values (reductions of 17 and 32%, respectively), which suggests an overall slowing of cross-bridge cycling (Table 1). Slowed myosin kinetics was supported by 27 and 41% decreases in the apparent rate constant 2πb for Y583S/+ and T178I/+, respectively, which suggests slowing of at least one step associated with work production (i.e., actin binding, P_i release and/or the power stroke) of the cross-bridge cycle (Table 1; (20,21)). A 33% increase in the 2πc-value was observed for the Y583S/+ fibers, but there was no significant difference in 2πc for the T178I/+ fibers.

We also used work loop experiments to measure power generation (22). The work loop technique allows for longer ML oscillations, which are more similar to those that occur for IFMs during Drosophila flight (23). The Y583S/+ and T178I/+ mutants generated 46 and 56% less work and 49 and 66% less power (P_Wmax) than control fibers, respectively (Table 2), when each muscle fiber was optimized for power production, i.e., at the ML change oscillation frequency and amplitude conditions that produced maximal power. The optimal frequency for work loop power production, f_Wmax, was 10% lower for Y583S/+ and 30% lower for T178I/+ (Table 2), which was similar to the observed decreases in sinusoidal analysis f_max (Table 1). The mutant fibers also performed better at ∼25% shorter ML change amplitudes than control fibers (Table 2).

Table 2.

Work loop analysis

	PWmax (W/m3)	fWmax (Hz)	ML (%)	Net work (J/m3)	Work generated (J/m3)	Work absorbed (J/m3)
Control	158 ± 19	141 ± 8	0.59 ± 0.04	1.15 ± 0.15	20.0 ± 2.0	−18.8 ± 1.9
Y583S/+	80 ± 11a	127 ± 2a	0.42 ± 0.04a	0.62 ± 0.08a	17.3 ± 1.8	−16.7 ± 1.7
T178I/+	53 ± 9a	100 ± 3a	0.48 ± 0.02a	0.52 ± 0.09a	22.2 ± 1.9	−21.6 ± 1.8
Control	134 ± 16	150	0.5	0.89 ± 0.11	16.9 ± 1.2	−16.0 ± 1.1
Y583S/+	56 ± 14a	150	0.5	0.40 ± 0.1a	19.2 ± 1.4	−18.8 ± 1.3
T178I/+	−16 ± 14a	150	0.5	−0.13 ± 0.09a	21.7 ± 2.4	−21.8 ± 2.3a

The top three rows, mean ± SEM, were taken from the %ML amplitude and frequency conditions in which each fiber produced maximal power (P_Wmax). n = 12 for FSS fibers, and n = 11 for the control. The bottom three rows, mean ± SEM, were taken from when FSS fibers were subjected to %ML amplitude and frequency conditions at which control fibers produced maximal power (thus, %ML amplitude and frequency values are constants rather than variables). Note that P_Wmax and net work values for T178I/+ fibers are negative. n = 11 for Y583S/+ and control fibers, and n = 7 for T178I/+.

^aIndicate significant difference from the control (one-way ANOVA, p < 0.05).

When all fibers were subjected to the ML change conditions at which the control fibers generated maximal power (0.5% ML and 150 Hz), work and power generation were greatly compromised for both FSS mutants (Fig. 2 D; Table 2). This was especially true for the T178I/+ fibers, where the representative work loop trace ran clockwise, meaning net work and power production were negative, and more work was absorbed by the fiber than it generated. Although the mutations increased both work absorbed and work produced, net work was lower than for the control fibers because the increase in the amount of work absorbed caused by the FSS mutant fibers was greater than the increase in work produced (Table 2). Y583S/+ fibers generated 14% more work than the control but work absorption increased even more, 18%. Similarly, T178I/+ fibers generated 28% more work, but work absorption increased 36%.

IFM ATP response

To elucidate more details about the mutations’ effects on the cross-bridge cycle, we performed sinusoidal analysis at varying concentrations of ATP and found that the T178I mutation decreased myosin ATP affinity but Y583S did not. The decrease in T178I myosin ATP affinity was shown by ∼3-fold and ∼7-fold higher K_M-values that were obtained by fitting the Michaelis-Menten equation to f_max and 2πc data, respectively (Fig. 3, A and B; Table 3). Decreased ATP affinity was also supported by 40% higher isometric tension values in the T178I/+ fibers than in control fibers at saturating ATP (20 mM), which increased with decreasing ATP concentration, gaining an additional 42% by 0.75 mM ATP (Fig. 3 C). Under the same conditions, control isometric tension did not change significantly as ATP concentration decreased. This shows that the increase in total tension is due to myosin-based tension (net active tension) rather than passive sarcomere elements because only net active, myosin-based tension is expected to increase with decreasing ATP concentration. Y583S/+ fibers showed no changes in K_M-values or a significant change in tension as ATP concentration decreased, suggesting that Y583S does not alter myosin’s affinity for ATP and that there is no difference in net active tension generation compared with control fibers.

Table 3.

Michaelis-Menten parameters derived from fitting the response of fibers to changing [ATP]

	fmax	2πc
Control Vmax	185 ± 4 Hz	5739 ± 300 s−1
Y583S Vmax	169 ± 4 Hza	6411 ± 1074 s−1
T178I Vmax	128 ± 4 Hza	6370 ± 988 s−1
Control KM	1.2 ± 0.1 mM	0.5 ± 0.1 mM
Y583S KM	1.2 ± 0.1 mM	1.3 ± 0.3 mM
T178I KM	3.3 ± 0.7 mMa	3.6 ± 0.8 mMa

V_max- and K_M-values were calculated for f_max and 2πc data versus ATP concentration. n = 6 except control 2πc data, for which n = 5. Values are mean ± SEM

^aIndicates significant difference from the control (one-way ANOVA, p < 0.05).

TDT activation and relaxation rates

To test the impact of the mutations on activation and relaxation rates, we used Drosophila TDTs expressing the same mutations as the IFMs. Although activation and relaxation rates are typically limited by the diffusion rate of calcium into and out of the skinned fibers, given the extremely prolonged relaxation rates observed by Racca et al. for FSS R672C, we thought we might observe changes if the cross-bridge release rate had slowed enough to be of a similar or slower rate as calcium diffusion (8). Activation k_Act-values decreased 22 and 19%, and the time from 10 to 90% activated was 45 and 20% longer than the control for Y583S/+ and T178I/+ muscles, respectively (Fig. 4 A; Table 4). Only the T178I mutation slowed TDT relaxation rate, as k_Rel decreased by 9% (Fig. 4 B; Table 4).

Figure 4.

Rates of activation and relaxation of heterozygous FSS jump muscles. Equation 2 was fitted to all activation and relaxation events for control and FSS TDTs, and the resultant fit parameters were averaged for each genotype. Activation and relaxation traces derived from the average fit parameters are shown. The tension level at the start of the experiment for activation traces (A) was normalized to 0% of maximal tension, and tension values for relaxation traces were normalized to 100% of maximal tension. FSS activation rates (k_Act) were significantly slower (one-way ANOVA, p < 0.05), as shown by rightward shifts in the curves, but only the T178I/+ relaxation rate (B) (k_Rel) was significantly slower than the control. The k_Act- and k_Rel-values were generated when Eq. 2 was fitted to the activation and relaxation curves and are listed in Table 4. n = 55 activation/relaxation runs from 14 control fibers, 48 runs from 10 Y583S/+ fibers, and 50 runs from 10 T178I/+ fibers. To see this figure in color, go online.

Table 4.

Activation and relaxation rates of TDT muscles

	kAct	kRel	Time to Act (s)	Time to Rel (s)
Control	1.64 ± 0.08	6.36 ± 0.19	4.09 ± 0.27	0.89 ± 0.02
Y583S/+	1.29 ± 0.06a	5.90 ± 0.34	5.93 ± 0.26a	0.91 ± 0.03
T178I/+	1.39 ± 0.07a	5.36 ± 0.2a	4.89 ± 0.24a	0.95 ± 0.03

Activation and relaxation rates were determined from k-values when Eq. 2 was fitted to activation (k_Act) and relaxation (k_Rel) curves. Time to activation (Act) and time to relaxation (Rel) are the times (in seconds) it takes for the fiber to change from 10 to 90% activated or 90 to 10% relaxed, respectively. Values are mean ± SEM n = 55 activation or relaxation runs from 14 control fibers, 48 runs from 10 Y583S/+ fibers, and 50 runs from 10 T178I/+ fibers.

^aIndicates significant difference from the control (one-way ANOVA, p < 0.05).

Drosophila locomotion

Diminished power production and slowed muscle kinetics due to the FSS mutations negatively influenced Drosophila flight ability. FI (1.5 ± 0.1) and wing beat frequency (146 ± 1 Hz) of Y583S/+ 2-day-old flies diminished 35 and 6% compared with controls (2.3 ± 0.1 and 156 ± 3 Hz; Student’s t-test, p < 0.05) at 15°C. 2-day-old heterozygous T178I mutant Drosophila could not fly or beat their wings. Y583S/+ Drosophila showed no change in jump distance (1.86 ± 0.04 cm) compared with control Drosophila (1.72 ± 0.07 cm), but the T178I/+ organisms displayed an 88% reduction in jump distance (0.19 ± 0.11 cm; Student’s t-test, p < 0.05).

Discussion

FSS Y583S and T178I mechanisms

Our study sought to determine the effect of two FSS-causing myosin mutations, Y583S and T178I, with particular focus on alterations to muscle mechanical properties and the myosin cross-bridge cycle. Using Drosophila IFM, we found that both FSS mutations increased active muscle stiffness, likely increased passive stiffness, decreased power, increased tension, and slowed f_max. TDT-muscle-skinned preparations enabled us to discern additional mechanical changes: decreased activation and relaxation rates. The IFM mechanical findings are similar but more pronounced than our recent IFM mechanical study of a milder form of DA (28).

Our previous study of isolated myosin from the Y583S and T178I FSS lines showed decreased actin-activated ATPase rate and slowed in vitro motility (9). Based on these changes, our muscle mechanics experiments, and our knowledge of what sarcomeric mechanisms contribute to these parameters, we can deduce cross-bridge level mechanisms for these two mutations. We will start by considering what our three findings of increased muscle stiffness (elastic modulus), muscle tension, and slower actin motility from IFM studies are suggesting.

Active muscle stiffness = N × S × duty ratio, where N is the number of heads that are functionally accessible for interaction with actin in the sarcomere, S is the average stiffness of an individual myosin head, and duty ratio is the fraction of the accessible myosin heads strongly bound to actin at the same time during contraction (29,30). Duty ratio = t_on/(t_on + t_off), where t_off is the average time myosin spends detached from actin. We will assume N is unchanged by the mutations, which is supported by our observations of at least no gross structural changes in the mutant sarcomere at the age used for mechanical experiments (Fig. 1; (9)). Therefore, to account for our observed increase in active elastic modulus, either S increased and/or t_on was greater and/or t_off decreased. Because actin-activated ATPase rates decreased (9), and ATPase = 1/(t_on + t_off); t_off more likely increased than decreased if it changed at all. Thus, we are left with greater S and/or increased t_on to account for higher active muscle stiffness.

Similarly, isometric tension = N × F × duty ratio, where F is the mean force generated by each head during contraction and duty ratio, is again =t_on/(t_on + t_off). F = d_uni × S, where d_uni is the average step size of the myosin head during contraction (29,30). We previously argued that t_off is unlikely to change, based on ATPase results. We cannot rule out increased step size, but again, greater S and/or increased t_on are left as strong candidates to account for the increase in isometric tension caused by at least one of the FSS mutations.

Our previously observed slower in vitro actin motility also supports an increase in t_on (9). Unloaded actin motility is not influenced by S (although a recent study by Brizendine et al., might suggest otherwise (31)) but would be slowed by increased t_on because velocity = d_uni/t_on (31). The only common factor that influences motility, active stiffness, and isometric tension is t_on. Thus, the simplest cross-bridge-based mechanism to account for all three results would be increased t_on.

Further support for increased t_on is the decreased f_max of the mutants compared with control fibers. Longer t_on would slow the rate myosin is moving actin and would therefore decrease f_max. Additionally, a slower relaxation rate for the T178I mutant TDT fibers suggests slower detachment of myosin from actin, which could be caused by a longer t_on, although we do not have the breadth of evidence for increased t_on in the TDT as we do in the IFM. For one of the mutations, T178I, we have evidence for a change to a specific step of the cross-bridge cycle that would increase t_on. The decrease in affinity for ATP that we found for T178I would slow ATP-induced detachment of myosin from actin. However, the change seems unlikely to be large enough to be the only step decreased, and because the other mutant did not show decreased ATP affinity, another strongly bound step is likely slowed for both.

The slower activation rates for the mutant TDT fibers (Fig. 4, A and B) and slower 2πb rate (Table 1) compared with control rates may provide some insight into an additional altered cross-bridge rate. Slower activation and 2πb both suggest a decrease in the rate of force generation, which given that both activation rate and 2πb are slower, suggest a slowing of myosin binding rate to actin, Pi release, and/or the power stroke (20,21). Given that Y583S interfaces with the G699 residue in the SH1-SH2 helix, which has been proposed to be a pivot point for the lever arm during myosin force generation (9,32), it is possible that the Y583S mutation influences the conformational change taking place during the force generating portion of the cross-bridge cycle. Similarly, the T178I mutation sits on the base of the p-loop, which binds P_i when ATP enters the active site, and could influence nucleotide release rate (9).

Increased t_on also could explain the observed decreases in net work and power during cyclical contractions. We observed that during the shortening portion of our work loops, the increased force generation ability of the mutant fibers (likely due to increased t_on) caused increased work generation because work = force × length change (see bottom three rows of Table 2). However, increased t_on likely also caused the greater increase in work absorption during muscle lengthening because an increased number of bound cross-bridges resist lengthening. This greater increase in work absorbed compared with work produced resulted in decreased net work per contraction cycle compared with control fibers. Because power per contraction cycle = net work × frequency of ML change (f_Wmax), the lower net work decreases power generation. Further, we also observed slower f_Wmax, which as we previously explained, is also slowed by longer t_on.

Longer t_on, and perhaps increased passive stiffness, likely also contributes to the progressive sarcomere structural defects we previously observed with age in Drosophila and similar muscle ultrastructure abnormalities found by other investigators (9,10,33). The higher t_on causes the muscle to experience increased force and stiffness, for example during the lengthening portion of our work loop experiments. The increased stress on these myofibrils likely causes sarcomere degradation, as we have seen in several other mutants that increase muscle contractility (12,28,34,35).

Differences between the two mutants

Although both FSS mutations caused many similar changes in muscle mechanical properties, there were a few differences between the two mutants. In general, the impact of the T178I mutation on mechanics was more severe than Y583S. T178I caused more power loss, a lower 2πb-value, and slower f_max than Y583S. These results fall in line with what is observed clinically because patients expressing the T178I mutation show more severe symptoms compared with patients expressing the Y583S mutation (15). Additionally, the severity in phenotype observed for the T178I mutation is reflected in its recent pathological reclassification. Previously, it had been reported that the T178I mutation could cause FSS or Sheldon-Hall syndrome, a less severe form of DA referred to as DA2B. Beck et al. reevaluated the prior evidence that T178I caused DA2B and determined those DA2B patients were misdiagnosed and had characteristics of the more severe disease, FSS (15).

In contrast with T178I, the Y583S mutation did not cause any changes in ATP affinity or relaxation rate. Therefore, its t_on-value may not be increased as much as T178I, and its altered t_on is likely influenced by a different strongly bound cross-bridge cycle rate than ATP binding. Myosin structure modeling performed by Rao et al. (9) showed that Y583 is located on the periphery of the nucleotide-binding pocket and also interacts with the SH1-SH2 helix. Before ATP binding and myosin release from actin, ADP, and P_i must first be released from the myosin active site. Therefore, it is possible that one of these release steps is slowed in Y583S myosin.

Drosophila mechanisms and human experimental results

In general, our findings corroborate and expand upon the findings of the two human-based studies that have investigated causes of FSS. In Racca et al., the only previous muscle mechanics study of FSS, a decrease in relaxation rate, incomplete relaxation, and a 2-fold increase in passive stiffness of muscle fibers and/or myofibrils from human FSS R672C patient biopsies were observed (8). We also measured a reduced relaxation rate for one of our mutants, T178I, a likely significant ∼25% increase in passive stiffness, and our proposed mechanism of increased t_on is similar to their proposed explanation for their reduced relaxation rate being a subpopulation of long-lasting cross-bridges. Additionally, Racca and colleagues measured a decreased activation rate, which we also observed for both of our mutants. Their interpretation of slower force generation causing this decrease matches our slower muscle apparent rate constant 2πb of work production, suggesting slower cross-bridge cycle steps associated with myosin binding to actin, Pi release, and/or the power stroke.

Most of Walklate et al.’s findings that used expressed human myosins are compatible with our proposed mechanism of increased t_on. They studied three FSS mutations R672H, R672C, and T178I using the C2C12 myosin expression system (7). Their results showed that at least two of the mutations decreased ATP affinity for actomyosin, increased affinity for actin in the presence of ADP, and yielded evidence for increased lifetime of the actin-myosin complex. These data all suggest a longer time bound to actin, i.e., increased t_on, which agrees with our main proposed mechanism. In contrast to our conclusion, however, they interpreted their decreased actin-activated ATPase V_max and decreased rate of the ATP hydrolysis step of the cycle to suggest that the FSS mutations greatly increase the time myosin spends detached from actin and that this would decrease the myosin duty ratio. A decreased duty ratio would cause less force production in a muscle, which contrasts with our results for T178I/+ fibers in which total tension was significantly increased. Based upon our findings, we suggest that both on and off times of myosin binding with actin are increased by the FSS mutations but that the time spent attached (i.e., t_on) has increased more than the time spent detached, causing an increase in duty ratio. This would cause the increased stiffness and force that we observed in our Drosophila fibers. The different conclusions about duty ratio are likely due to our measurements being performed while myosin is constrained in thick filaments where specific motor orientation and resistive forces are present, as opposed to the solution-based S1 studies presented by Walklate et al. (7).

It is encouraging that our understanding of the mechanisms behind the FSS phenotype is growing, but how can we make use of this knowledge in the clinical setting, and what aspects of the disease still need to be studied? One of the more interesting aspects about FSS in humans is that the mutations are only expressed in the MYH3 embryonic myosin heavy chain. Therefore, after birth, the congenital abnormalities would be expected to be nonprogressive because MYH3 expression is downregulated in adult tissue (36, 37, 38, 39, 40). The mutation presumably only obstructs muscle development in utero, yet its effects last throughout the patient’s lifetime. Apparently, the distal muscles (i.e., hands and feet) develop sooner in utero and, therefore, are expressing MYH3 for a longer period of time, so this is likely why FSS exerts a greater effect on these areas (41).

However, MYH3 expression does not disappear completely in adulthood. MYH3 is still expressed in skeletal muscle in combination with an adult form of myosin (MYH2), meaning that FSS mutations could still influence contraction postnatally (8,42,43). This could explain the lack of mitigation of symptoms with the expression of adult myosin isoforms. In the future, a drug that specifically decreases t_on may help reduce the debilitating FSS phenotype if used early enough. There are currently a number of well-known skeletal myosin inhibitors such as 2,3-butanedione monoxime (BDM), blebbistatin, and N-benzyl-p-toluene sulphonamide (BTS), which are powerful tools for studying muscle contraction (44, 45, 46). In fact, a very recent study that developed a zebrafish model of FSS showed that treatment of embryos with para-aminoblebbistatin dramatically improved phenotypic outcomes (33). Drugs that have similar effects as these examples, particularly if they reduced t_on, could be particularly helpful, but care would be required to avoid off-target effects, particularly effects on cardiac myosin.

In conclusion, our study suggests a major component of the FSS molecular mechanism is an increase in the time myosin spends attached to actin, i.e., t_on. For the T178I mutation, we determined that a decrease in ATP affinity contributed to longer t_on. Increased t_on drastically increased IFM muscle stiffness, slowed muscle kinetics, and inhibited muscle power generation. These molecular and fiber mechanisms are likely contributing to human FSS by inhibiting normal muscle and limb movements during fetal development, causing anatomical abnormalities and a limited range of motion in FSS patients.

Author contributions

K.M.B. contributed to the investigation, data curation, formal analysis, preparation of the original draft, and review and editing. A.H. contributed to the investigation and formal analysis. W.A.K. contributed to the investigation and methodology. S.I.B. contributed to the conceptualization, methodology, resources, review and editing of the manuscript, supervision, and funding acquisition. D.M.S. contributed to the conceptualization, methodology, resources, review and editing of the manuscript, supervision, and funding acquisition.

Editor: Samantha Harris.