The hypertrophic cardiomyopathy myosin variant G256E prolongs cardiac muscle relaxation via altered nucleotide handling Authors: Kerry Y Kao, Matthew Carter Childers, Divya Pathak, Rama Reddy Goluguri, Timothy S McMillen, Joseph D Powers, Kathleen M Ruppel, James A Spudich, Michael Regnier Published: 2026 ABSTRACT AbstractMutations in myosin alter its motor functions in diverse ways by affecting different structural and chemo-mechanical events. Multidisciplinary strategies can be used to understand how varying alterations in motor function converge to common phenotypes like hypercontractility and hypertrophic cardiomyopathy (HCM). We combined molecular dynamics (MD) simulations with protein biochemical and myofibril mechanical analyses to study the HCM-causing myosin variant G256E. MD simulations demonstrated that G256E induces structural changes that increase the work required to displace ADP.Mg2+ from actomyosin. Stopped-flow biochemical analysis demonstrated increased ADP affinity and decreased ADP release rate, and single myofibril mechanics analysis demonstrated increased force generation and reduced ADP sensitivity of the early, slow phase of relaxation. Together, these results demonstrate that slower ADP release from myosin during contraction is a significant contributor to pathological contractile nature of the G256E mutation. This study highlights the importance of detailed chemo-mechanical analysis of mutations associated with hereditary cardiac diseases.Subject terms: Mechanism of action, Molecular modelling, Contractile proteins Introduction Hypertrophic cardiomyopathy (HCM) affects 1 in 500 people globally and is implicated as a contributing cause of other life-threatening heart diseases such as atrial fibrillation, heart failure, and arrhythmia resulting in sudden death1. About 60% of HCM cases are caused by familial inheritance of a single autosomal dominant mutation within a gene encoding a sarcomere protein2. In particular, the MYH7 gene encoding the β-myosin heavy chain (MHC) is one of the two most frequently mutated genes for HCM3. Myosin is the motor protein responsible for muscle contraction, converting chemical energy produced by ATP hydrolysis and release of products (Pi, ADP) into the mechanical work of force production and muscle cell shortening4. This involves changes in myosin’s structure as it interacts with actin to accelerate the release of ATP hydrolysis products (Pi, ADP) in a process known as chemo-mechanical transduction5–7. Structural perturbations resulting from disease-associated mutations can disrupt the kinetics and/or mechanics of the chemo-mechanical cycle and can lead to initiation and progression of disease8. The G256E MYH7 mutation, associated with HCM,9 is located in a hairpin turn that connects two strands in the central β-sheet of transducer region of the myosin head. This region conveys information between the nucleotide binding pocket and actin interface of myosin, both critical to chemo-mechanical transduction during contraction and relaxation10. In a recent study, we employed CRISPR/Cas9 to edit this mutation into human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) to produce a heterozygous (±) isogenic model for studying its impact at multiple scales of cardiac muscle organization8. A biochemical single ATP turnover assay performed on purified human β-cardiac 25-hep heavy meromyosin (HMM) containing the G256E mutation results in a reduced population of myosin in a low basal ATPase state, often described as the super-relaxed (SRX) state, suggesting a greater portion of the myosin population is available to contribute to muscle contraction. Comparison of the actin-activated ATPase activity of two versions of G256E myosin that differed only in the length of their coiled-coil tail confirmed that the mutation disrupted the ability of myosin to fold back into a sequestered, low ATPase state. This supports the hypothesis that the mutation results in more heads being available to interact with actin at the beginning of contractions. Slower motility in an in vitro F-actin motility assay indicated the mutation results in either a smaller step size or higher myosin duty ratio. The latter would affect the actomyosin detachment rate, supported by measures of the contraction and relaxation properties of isolated myofibrils from hiPSC-CMs. The kinetics of the slow, early phase of relaxation, which measures the myosin detachment rate from thin filaments,11 was slower for G256E versus WT myofibrils. Molecular dynamics (MD) simulations of the myosin M.ATP-state suggested that G256E disrupts the shape and stability of the central β-sheet, potentially altering communication between the nucleotide-binding pocket and actin-binding surface. This communication path is critical to open the nucleotide-binding cleft that allows ADP release. This, in turn, is required before a new ATP molecule can bind in the nucleotide pocket and lead to the dissociation of myosin from actin. While these findings are suggestive that myosin-actin chemo-mechanics are altered as part of the hypercontractile phenotype of G256E, it is important to understand which particular step(s) of the crossbridge cycle are affected. The ADP release step is rate limiting for crossbridge detachment in sarcomeres during loaded contractions and is also a determinant of the rate of relaxation12. This is supported by biochemical,13 mechanical,14 single-molecule,15 and structural16 evidence, and holds true across multiple classes of myosins from diverse species. While the precise mechanistic basis for this observation is not definitively established, formation of a strongly-bound complex between actin and myosin plays an important role as binding of myosin to actin alters the conformation of the actin binding cleft and the nucleotide binding pocket16. Myosin ADP bound form binds actin strongly and produces force and the duration of this strongly force-producing state has been shown to have a direct effect on power output. Recent studies have shown that various myosin mutations can alter the affinity of myosin for ADP and consequently affect crossbridge cycling kinetics17. Thus, a detailed study of how the G256E mutation affects ADP release during chemo-mechanical cycling is merited. Prior computational studies have provided important mechanistic insights into conformational changes during the chemo-mechanical cycle. For example, coarse-grained free energy models and Langevin dynamics simulations were used to explore the relationship between ADP release and processive movement in myosin18–20. More recently, umbrella sampling and automated path-searching methods have been applied to chemo-mechanical cycling of myosin V and myosin VI21. Related approaches have been used to study recovery stroke dynamics in myosin II and VI22,23 in myosin II and myosin VI, as well as phosphate release from bovine β-cardiac myosin in the pre-powerstroke state24. These studies of substrate release from myosin have formed an important foundation for computational studies of conformational transitions in myosin; however, many have been performed using coarse-grained approaches and in the absence of actin. In the present work, we explore all-atom dynamics of the human β-cardiac myosin II within the physiologically relevant post-powerstroke actomyosin complex, where ADP release is rate-limiting for crossbridge detachment under mechanical load. Here we report the effects of myosin G256E on ADP release from the nucleotide binding pocket of myosin and its effects on myosin and myofibril chemo-mechanical function. MD simulations of post-powerstroke myosin containing ADP (with Mg2+) complexed with actin (AM.ADP) suggested that G256E significantly alters nucleotide pocket structure and dynamics. Steered MD simulations suggested that G256E alters the ADP release pathway that more energy is required to displace ADP.Mg2+ from the pocket. These results provide a structural basis for altered ADP-myosin affinity and dynamics. Stopped-flow kinetics experiments confirmed these predictions, showing increased ADP affinity and decreased ADP release rate in G256E sS1. G256E myofibrils generated significantly more force and exhibited a slower, early linear phase kinetics of relaxation (kREL,slow) compared to isogenic controls, similar to our initial study,8 and elevated ADP had considerably less effect on kREL,slow for G256E myofibrils, consistent with impaired ADP release. Combined, these experimental and computational studies support the hypothesis that the G256E mutation alters the structure of the nucleotide binding pocket, resulting in greater ADP affinity and slower ADP release, which results in delayed crossbridge detachment and prolonged contractile relaxation. These results highlight the power of combining experimental measures with in-depth computational dynamics to provide detailed structure-based mechanistic insights at the atomistic scale for protein and contractile organelle (myofibril) function. Results G256E alters myosin nucleotide binding pocket dynamics with ADP To investigate the possible structural changes within myosin that might result in impaired ADP release we employed conventional molecular dynamics (cMD) to simulate the time-dependent changes of the post-powerstroke state of myosin complexed with actin and ADP (AM.ADP, Fig. 1a), and how G256E alters structure and contacts at the atomic level. See Table 1 for MD simulation system details. In general, G256E simulations had a greater change in global Cα RMSD compared to WT simulations (Fig. 1b), presumably as a response to the change from Gly to Glu. These changes tended to stabilize by 300 ns (Fig. 1b). We calculated the change in Cα RMSF per residue and found that the flexibility of the G256E structure remained very similar to WT except for loop 2 (residues 621-646) and the end of the tail (residues 771-810) (Fig. 1c). In the WT stimulations, ADP primarily formed contacts with the P-loop (residues 179-184) and Purine-ring (residues 126-131) of the nucleotide binding pocket (Fig. 2a, c). In G256E simulations, the introduction of the bulky, negatively charged sidechain (E256) formed a novel salt bridge with R169, as was observed previously in simulations of the G256E M.ATP state8. This structural change propagated to the nucleotide pocket. For the majority of G256E simulations, ADP preferred to adopt a conformation that closely associates with Switch-1 (residues 232–244) and the loop that connects the third β-strand of the transducer to the SH1-helix (SH1-loop, residues 673–685). In particular, two non-native contacts formed in the G256E simulations (Fig. 2c S241 and T678) and persisted for greater than 75% of the combined simulation time. This new ADP conformation led to the formation of a salt bridge between D239 of Switch-1 and K679 of the SH1-loop, which persisted longer in G256E simulations (average 28.9% of simulation time) compared to WT (average 3.2% of simulation time, p = 0.051). Overall, this results in a loss of contacts between ADP and the P-loop and the Purine-ring (Fig. 2c) and altered dynamics in the nucleotide pocket in a manner that might affect ADP release. G256E simulations had a greater change in global Cα RMSD compared to WT simulations. a The post-powerstroke model includes 5 actin monomers (light green, space-filling atoms), myosin motor domain, and essential light chain (ELC, dark green). The functional regions of myosin are colored uniquely: upper 50 kDa domain (cyan), lower 50 kDa domain (blue), transducer (seafoam green), N-terminal domain (yellow), converter domain (dark blue), and tail (purple). E256 (red) is located in a β-hairpin turn between beta strands β6 and β7 in the transducer region. b Global Cα RMSD of WT simulations and G256E simulations. On average, G256E simulations had higher global Cα RMSD than WT simulations. c Cα RMSF difference between WT and G256E simulations by residue. Positive values indicate higher flexibility in WT simulations whereas negative values indicate higher flexibility in G256E simulations. Cα RMSF mapped onto (d) WT and (e) G256E structures. Darker colors and thicker ribbons indicate higher Cα RMSF. Table 1Conventional MD system setupNumber of simulations per genotype3Water box dimensionsTruncated octahedron with length 209.8 Å and angle 109.47°Total number of atoms~671 000Total number of water molecules~208 000Salt concentration120 mM NaCl Conventional MD system setup G256E alters ADP nucleotide handling. ADP occupancy maps for (a) WT and (b) G256E simulations show that ADP adopts a novel conformation in G256E simulations. Occupancy maps cover two standard deviations from the mean ADP occupancy. c Changes in ADP contacts with nucleotide pocket residues. In G256E simulations, contacts with the purine-loop and P-loop are lost while new contacts with Switch-1 and the SH-loop are formed. d ADP dihedral angle sampling for (e) WT and (f) G256E simulations. Six dihedral angles were defined: θ1 (O1B-PB-O3A-PA), θ2 (PB-O3A-PA-O5′), θ3 (O3A-PA-O5′-C5′), θ4 (PA-O5′-C5′-C4′), θ5 (O5′-C5′-C4′-O4′), and θ6 (O4′-C1′-N9-C8). These angles were plotted in Ramachandran-style 2-D histogram plots (bin size 5°x5°) to visualize changes in ADP conformation. Each bin is colored according to probability density. The conformation of ADP within the nucleotide binding pocket was assessed by calculation of 6 distinct dihedral angles within ADP: θ1 (O1B-PB-O3A-PA), θ2 (PB-O3A-PA-O5′), θ3 (O3A-PA-O5′-C5′), θ4 (PA-O5′-C5′-C4′), θ5 (O5′-C5′-C4′-O4′), and θ6 (O4′-C1′-N9-C8). The dihedral angles were analyzed via 2D Ramachandran-style plots (Fig. 2d–f). The sampling entropy (Table 2) and percent coverage (Table 3) of the dihedral angle maps indicate that ADP is less flexible in the pocket and adopts a more stable conformation for G256E compared to WT simulations. We calculated the interaction energies between myosin and ADP.Mg2+. Residues with at least 3 kT of interaction energy were considered part of the nucleotide binding pocket (residues 126, 128, 180, 181, 182, 183, 184, 185, 228, 232, 238, 240, 241, 242, 246, 461, 674, 679) (Supplemental Fig. 1). We also observed a statistically different center of mass distance between ADP.Mg2+ and the nucleotide binding pocket residues for G256E. The altered conformation of ADP.Mg2+ in the nucleotide binding pocket led to statistically distinct interaction energies with myosin residues (Supplemental Fig. 1). Based on these data, our simulations predict that the G256E mutation leads to conformational changes in myosin that reshape the nucleotide binding pocket, reorienting ADP.Mg2+ within the binding pocket, and disrupt nucleotide handling during crossbridge cycling. Table 2Conformational entropy of ADP in WT and G256E simulationsDihedral pairADP entropy (cal mol−1 K−1)WTG256Eθ1 vs. θ29.637.62θ3 vs. θ410.008.97θ5 vs. θ611.2510.30 Conformational entropy of ADP in WT and G256E simulations Table 3Percent coverage of dihedral maps sampled by ADP in WT and G256E simulationsDihedral pair% coverageWTG256Eθ1 vs. θ28.93%3.88%θ3 vs. θ421.06%11.21%θ5 vs. θ633.49%17.48% Percent coverage of dihedral maps sampled by ADP in WT and G256E simulations Changes in ADP release observed in simplistic steered MD simulations We used steered molecular dynamics (sMD) simulations to simulate ADP release from the nucleotide binding pocket of myosin in complex with actin (AM.ADP*). Myosin residues with an average linear interaction energy (lie) of at least 3 kT with either ADP or Mg2+ were selected as nucleotide binding pocket residues for a center of mass (COM) based steering coordinate. We selected 60 structures each from the WT and G256E cMD simulations that had ADP.Mg2+-nucleotide binding pocket residue COM distances within 0.1 Å of the mode values in cMD simulations (WT: 3.6 Å; G256E: 4.4 Å). Then, we performed simplistic steered MD simulations that forced ADP out of the binding pocket over a fast timescale ( ~ 10 ns) (Fig. 3a,b). On average, more work (WT: 58 kcal mol−1, G256E: 89 kcal mol−1 p = 2.4 × 10−18) was needed to displace ADP.Mg2+ from the binding pocket of G256E myosin vs WT myosin suggesting a stronger interaction between G256E myosin and ADP.Mg2+ (Fig. 3c). Snapshots from low work trajectories (Fig. 3d, e) shows that the initial conformation of ADP.Mg2+ was unique for both WT and G256E myosin, but that ADP.Mg2+ tended to interact with E228 at the end of our reaction coordinate. We caution that this approach does not reflect correct absolute energy barriers for ADP release: the simulations do not allow for sufficient relaxation of myosin as ADP.Mg2+ is moved along the reaction coordinate, and sMD does not rigorously calculate a potential of mean force. Indeed, some trajectories sampled much higher work than the average trajectories. Furthermore, these simulations only displace ADP.Mg2+ out of the binding pocket, and some interactions with myosin remain. Under the assumption that both WT and G256E simulations suffer from comparable errors, these simulations predict that it is more difficult for ADP.Mg2+ to leave G256E post-powerstroke myosin. It remains possible for conformational changes that were not captured in these simulations to have large effects on ADP release. To aid in future studies of ADP release from myosin, we analyzed interactions between the ADP.Mg2+ molecules are made with myosin during the release trajectories (Fig. 4a, b). We found that ADP.Mg2+ formed different interactions with WT and G256E myosin as it left the pocket, in large part due to the burial of the adenine ring in the SH1-loop and increased interactions of the phosphates with Switch-1. More work is required to displace ADP in G256E steered MD simulations. Work profiles of 60 sMD simulations of ADP.Mg2+ release from WT (a) and G256E (b) myosin. c The average work done along the PMFs for each genotype. Error bars are ± SD. d Snapshots that depict ADP release from low work trajectories are shown in (d, e) for the WT and G256E simulations, respectively. These snapshots show that movement of the purine ring precedes movement of the phosphate groups in our sMD simulations. At the end of the reaction coordinate, we observe Mg2+ interacting with E228. ADP release trajectory is altered in the presence of the G256E mutation. a, b Depict the contacts made between myosin and ADP.Mg2+ during the sMD simulations. We binned our trajectories into 0.5 ns intervals (x-axis) and calculated myosin-ADP and myosin-Mg2+ contacts within each interval. For each contact (y-axis), we averaged the fraction of simulations for which the contact was present, and these heatmaps are scaled according to the percentage of simulations and time intervals for which a contact was observed. Darker colors indicate that a contact was more likely to be observed. These plots show that ADP.Mg2+ exited WT and G256E myosin via different pathways and formed unique interactions with myosin during release. Contacts are shown only if they appeared in at least 90% of the WT or 90% of the G256E simulations. G256E sS1 has greater affinity for ADP and slower ADP release rate We paired computational results with in vitro experiments to validate findings (Fig. 5a). To experimentally determine the ADP affinity of myosin (KADP), we performed an ATP/ADP competitive binding assay using recombinant human β-cardiac short subfragment-1 (β-sS1), a single-headed myosin construct containing only the human ventricular essential light chain. Experiments were performed by rapidly mixing 0.125 μM pyrene-actin.β-sS1 with 25 μM ATP + varying amounts of ADP (0, 20, 40, 80, 160, 320, and 640 μM) (all reported concentrations are post-mixing, Fig. 6a). In the absence of ADP, experimental curves were fit with a single-exponential which describes ATP-induced actomyosin dissociation. In the presence of ADP, experimental curves were fit with a double-exponential to account for the slow phase, which describes ADP dissociation from actomyosin. We plotted the relative kobs of the fast phase (krel = kobs/kobs,0) vs. ADP concentration to assess ADP affinity (Fig. 6b). G256E β-sS1 demonstrated ~30% increased ADP affinity (KADP = 33 ± 3 µM, Fig. 6b, p = 0.04) compared to WT β-sS1 (KADP = 48 ± 11 µM). To determine the ADP release rate, we pre-incubated pyrene-actin.β-sS1 with a saturating amount of ADP (80 μM) before mixing with a large excess of ATP (1.5 mM, Fig. 6c). Average traces were fit with a double exponential equation with the fast phase describing ADP release and the slow phase describing the population of actin.β-sS1.ADP, which must isomerize from a closed to and open nucleotide pocket conformation before ADP is released. G256E β-sS1 had ~40% slower ADP release (71 ± 4 s−1, Fig. 6d) compared to WT β-sS1 (115 ± 7 s−1, p = 0.0002). Synergy between computational and experimental techniques. a Schematic of experimental and computational workflow synergy. Data from myofibrils derived from CRISPR-edited hiPSCs supported by biochemical experiments with purified β-myosin. Experimental findings are validated by molecular simulations. b SDS-PAGE myosin separation gel showing the conversion of α-MHC of D14 hiPSC-CMs to β-MHC by D30 and D37. Standards are from adult mouse atria (α-MHC), adult rat ventricle (α/β-MHC), and adult rabbit ventricle (β-MHC). c Cardiac directed differentiation of hiPSCs. hiPSC-CMs are matured through lactate purification and line patterning for seven days before being skinned on D37 for myofibril mechanics. Full-sized gel of (b) in Supplemental Fig. 2. Created in BioRender. Kao, K. (2026) https://BioRender.com/nj38w74. G256E sS1 have greater ADP affinity and slower ADP release. a Average traces (8–10 individual traces) of ATP-induced pyrene-actin.WT sS1 dissociation. Traces in the absence of ADP (0 µM) were fit with a single exponential. Traces in the presence of ADP (left to right: 20, 40, 80, 160, 320, 640 µM post-mixing) were fit with a double exponential. b The relative kobs (krel = kobs/k0 where k0 = kobs when ADP = 0 µM) of the fast phase plotted vs. [ADP]. The data was fit with the equation krel = 1/(1 + [ADP]/KADP) to obtain the apparent affinity. One representative trace from WT and G256E is shown in b. Inset: G256E caused a ~ 30% increase in ADP affinity (KADP = 32.82 ± 1.60 μM) compared to WT (KADP = 46.46 ± 3.23 μM). Error bars represent ± SD, n = 4. c Normalized average traces (8–10 individual traces) of ATP-induced pyrene-actin.sS1.ADP dissociation ([ATP] = 1.5 mM). Traces were fit with a double exponential. d ADP release rate (kobs,fast) was significantly slower for G256E sS1. Error bars represent ± SD, n = 3-4. G256E myofibril relaxation is slower and less sensitive to elevated ADP To determine how the increased ADP binding affinity affects contractile properties, particularly the early, linear phase of relaxation that reflects crossbridge detachment, we used myofibrils from CRISPR/Cas9-edited hiPSC lines containing the heterozygous expression of the G256E mutation (MYH7WT/G256E) and an isogenic control (MYH7WT/WT). The cell lines underwent a series of quality control experiments to confirm genotype and validate that the gene-editing process did not have any off-target effects8. We confirmed that our stem cell differentiation and culture conditions (Fig. 5c) provided hiPSC-CMs that express predominantly β-MHC by D30, validating the D37 timepoint was appropriate for experiments (Fig. 5b). The mechanical and kinetic properties of hiPSC-CM myofibrils were measured in a custom-built apparatus using a fast solution switching (~ 5 ms) technique, as previously described25. Myofibrils were activated in a sub-maximal calcium solution (pCa 5.6), relaxed (pCa 9.0), then activated and relaxed with solutions containing 50% ADP (pCa 5.6ADP, 2.5 mM ATP + 2.5 mM ADP) to observe the effects of elevated ADP on contractile properties. Figure 7a shows representative traces for contraction and relaxation, and Fig. 7b shows a higher resolution view of the early, linear phase of relaxation, for measurements of the slope (kREL,slow), which is determined by the rate of crossbridge detachment, and duration (tREL,slow), which assesses the time it takes for thin filament deactivation26,27 Force generated at both pCa 5.6 and pCa 5.6ADP was significantly increased for G256E myofibrils by ~2X (Fig. 7c, p < 0.01) compared to WT myofibrils. Activation kinetics (kACT), which inform on crossbridge attachment and cycling rates, was decreased by ~30% (Fig. 7d, p < 0.01) for both G256E and WT myofibrils. For relaxation, the main, fast phase rate (kREL,fast), which reflects multiple active and passive properties of the contractile elements, was unchanged for G256E myofibrils compared to WT (Fig. 7e). Addition of ADP slowed kREL,fast significantly for both WT and G256E. For the early, linear phase of relaxation, the duration (tREL,slow) was prolonged approximately 2-fold by the addition of ADP for both G256E and WT myofibrils (Fig. 7g, p < 0.01). Additionally, the slow phase rate of relaxation (kREL,slow) was significantly decreased by elevated ADP in G256E myofibrils compared to WT, similar to what we previously reported8. Importantly, the magnitude of change for pCa 5.6 vs. pCa 5.6 + ADP was significantly lower in G256E myofibrils compared to WT (p = 0.019, Wilcoxon rank-sum test). This suggests that ADP release is already significantly impaired in G256E myofibrils, such that the response to ADP product inhibition is blunted. G256E myofibrils isolated from hiPSC-CMs generate more force and have impaired early phase relaxation. a, b Representative activation traces from hiPSC-CM MYH7WT/G256E (red) and isogenic MYH7WT/WT (black) myofibrils with a close-up of the slow, early phase of relaxation with linear fits (kREL,slow). Traces are an average of 2 sequential activations with activating solutions of 5.6 = pCa 5.6 activation paired with pCa 9.0 relaxation or 5.6ADP = pCa 5.6 activation + 50% ADP (2.5 mM ATP, 2.5 mM ADP) paired with pCa 9.0 + 50% ADP (2.5 mM ATP, 2.5 mM ADP) relaxation. c Max force generated is significantly higher in G256E myofibrils in both 5.6 and 5.6ADP activating solutions. ADP does not inhibit force generation. d Activation kinetics (kACT) are unchanged between G256E and WT at both conditions (5.6 and 5.6ADP) and decreased with the addition of ADP. e G256E fast phase relaxation kinetics (kREL,fast) are unchanged compared to isogenic control; ADP inhibition decreases kREL,fast significantly. f The kinetics of the slow, early phase (kREL,slow) are significantly decreased in G256E myofibrils at 5.6 and 5.6ADP. The magnitude of change from pCa 5.6 to pCa 5.6 + ADP was significantly lower in G256E myofibrils compared to WT (p < 0.05, Wilcoxon rank-sum test). g The duration of the slow phase of relaxation (tREL,slow) is not significantly different when comparing G256E and WT. The addition of ADP prolongs the duration of this phase. Data are represented as mean ± SD. Statistical significance was determined by paired Wilcoxon signed-rank tests. ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Individual myofibrils: WT n = 15, G256E n = 13; Biological replicates: WT N = 8, G256E N = 3. Discussion Synergy between computation simulations and experiments Here, we simulated ADP release from myosin using enhanced sampling computational techniques. We validated computational results with data collected from biochemical measurements of ADP affinity for sS1 myosin motors and mechanical measurements of hiPSC-CM-derived myofibrils. Our results highlight the synergy between computational simulation and experimental validation in understanding the structure-function basis of the myosin motor and contractile dysfunction resulting from disease-associated mutations, and this may provide intra-protein targets for more precise therapeutics approaches. In post-powerstroke (AM.ADP) simulations, changing G256 to the E256 mutation forms an enduring salt bridge with R169. This was also observed in the M.ATP state simulations we reported recently8. Even though this contact was consistent in simulations of both chemo-mechanical states, there were different consequences. In the M.ATP simulations, the salt bridge led to strain in sheets β5 and β6 of the transducer region and a reduction in the stability of hydrogen bonds between these sheets. However, in post-powerstroke (AM.ADP) simulations, the instability of the transducer region was not observed. This may be due to intrinsic differences in the nucleotide binding pocket in ‘closed’ (i.e. actin-free, ATP/ADP.Pi-bound forms) vs ‘open’ (i.e. actin-bound, ADP-bound and nucleotide-free forms) conformations. These findings highlight the importance of simulating myosin in different chemo-mechanical states to obtain a more complete assessment of induced changes in structural dynamics, e.g., a single state structure analysis may result in incomplete understanding or incorrect conclusions. In the case of post-powerstroke simulations, it is possible that the presence of actin stabilized the transducer, as the transducer is known to adopt different conformations in the presence and absence of actin28,29. The post-powerstroke simulations ultimately led to the discovery of altered nucleotide pocket dynamics due to the G256E mutation, which was consistent for all replicate simulations of the mutant. In particular, ADP formed novel contacts with the SH1-loop. This loop leads into the SH1 and SH2 helices, which link the converter domain to the rest of the myosin motor domain. In previous experimental studies, SH1-SH2 interactions were found to modulate the affinity of ADP to myosin, and a computational study found electrostatic interactions between the SH1 and SH2 helices alter ADP release30,31. SH1 and the relay helix also connect the converter domain to the actin-binding surface, facilitating lever arm movement31. Although we did not see any differences in either SH1-SH2 interaction or stability (i.e., secondary structure), it is possible that some structural alterations could occur outside of our timescale (i.e., 500 ns). Our cMD simulations predict a structural basis for altered nucleotide handling by G256E myosin. The G256E mutation induced a series of conformational changes that altered the structure and dynamics of the nucleotide-binding pocket as well as the conformation of ADP within the pocket. Our sMD simulations, though simplistic, predict that G256E myosin has a relatively stronger interaction with ADP.Mg2+ than does WT myosin. We propose that altered interactions between ADP.Mg2+, Switch-1, and the SH1-loop underlie the altered ADP release rates that we observed experimentally. These simulations also suggest residues that are important in the rate of ADP dissociation. K184, a positively charged residue in the P-loop lining the bottom of the pocket, anchors ADP in the binding pocket, and the loss of a salt bridge between this residue and Pβ is necessary to initiate ADP release. Mutation of this residue to a glutamine (K184Q) has been seen in patients with HCM or left ventricular noncompaction cardiomyopathy and is currently a variant of unknown significance (VUS)32. The next residue identified on the reaction coordinate is D239. In WT simulations, this residue forms a salt bridge with K679 to keep the nucleotide pocket closed. This residue is of particular interest because it has been tied to multiple disease-related mutations. D239N is a likely pathogenic mutation implicated in early-onset HCM33,34. Based on our observations, we hypothesize that the change from Asp, a negatively charged side chain, to the uncharged polar residue Asn, will disrupt the salt bridge between 239 and 679. This would likely result in faster ADP release. This may explain results from a study demonstrating faster actin gliding velocities and actin-activated ATPase rates in D239N myosin35. At the end of our reaction coordinates, ADP. Mg2+ was strongly coordinated by E228, which may act as a “loading zone” to shuttle ADP out and a new ATP in to the nucleotide binding pocket. Finally, K246 is another residue that is important for Pβ stabilization. The mutation K246I is currently classified as likely pathogenic by Walsh et al36. A change from a positive residue to a non-polar one would likely affect coordination of the nucleotide in the pocket through destabilization of Pβ36. In our previous study we found that, for isolated myofibrils from hiPSC-CMs, the kinetics of the slow, early phase of relaxation, kREL,slow following maximal (pCa 4.0) calcium activation were reduced8. This could result from either slowed ADP release from myosin or slowed ATP binding to myosin, as both steps are required for crossbridge detachment. Here, we utilized elevated ADP as a product inhibition method to probe if ADP release from myosin is specifically altered with the G256E mutation. We chose to perform these experiments at a sub-maximal calcium level to make conditions more physiologically relevant, as pCa 5.6 is much closer to the rise in intracellular calcium that occurs during a cardiac twitch. Importantly, kREL,slow was significantly slower for G256E myofibrils at pCa 5.6 (Fig. 7f), a similar result as in our previous study at pCa 4.08. The addition of 2.5 mM ADP to WT myofibrils slowed kREL,slow to a rate that was similar to that for G256E myofibrils in the absence of ADP, and addition of ADP for G256E myofibrils resulted in a significant but smaller further reduction of kREL,slow. The addition of ADP should inhibit ADP release and prolong the kinetics of this phase, which is seen in WT myofibrils but not G256E myofibrils. This suggests that the release of ADP is significantly compromised within G256E myofibrils, explaining the minimal response to ADP product inhibition. These data agree with the conclusions from our computational simulations and myosin biochemical measures regarding ADP.Mg2+ release and greater ADP.Mg2+ affinity for myosin, respectively. However, we should not rule out other possible contributions to the relaxation rate, e.g., the rates of ATP binding and actomyosin dissociation. Our preliminary stopped-flow measures of mant-ATP binding to S1 or HMM indicate that the rate of ATP binding is multi-fold faster than kREL,slow for both WT and G256E myofibrils, suggesting ADP release may control myosin detachment and relaxation kinetics. This is supported by measurements of kREL,slow, which was decreased for G256E compared to WT myofibrils at pCa 5.6 (Fig. 7f), and decreased by elevation of ADP for WT, but much less for G256E myofibrils. A note on force development At pCa 5.6, steady-state force (Fig. 7c) was approximately doubled for G256E myofibrils compared with WT myofibrils, indicated increased calcium sensitivity of force, while activation kinetics, kACT (Fig. 7d), were not affected. This contrasts with measurements we reported for maximal activation (pCa 4.0), where both force and kACT were approximately doubled for G256E myofibrils8. This suggests that the rate-limiting chemo-mechanical process for the rate of force development at the sub-maximal calcium used (pCa 5.6) is new strong crossbridge formation and the force-generating isomerisation of myosin associated with inorganic phosphate release, and not ADP release or crossbridge detachment. The elevation of ADP did not affect steady-state force but decreased kACT by a similar amount for both WT and G256E myofibrils. Together, this suggests the mutation affects the number of myosins that bind, but not the binding and force development kinetics. Limitations of our study Here, we report MD simulations of post-powerstroke β-cardiac myosin in complex with actin and have used this model to evaluate the effects of the G256E mutation on the structure of post-powerstroke actomyosin and the dynamics and pathway of ADP release. While these simulations provide molecular mechanistic insight, they are subject to several limitations and do not fully reflect the full complexity of ADP release within sarcomeres. First, our model includes only an actin pentamer and myosin residues 1–810, and there is no effective load in our system; thus effects of tension (force) on ADP release or interaction between more complex constituents of the sarcomere are not included. Some studies suggest Mg2+ may be released from myosin prior to ADP release; however, here we have simulated the release of ADP.Mg2+ as a complex. We have not yet simulated the rigor actomyosin complex for this mutation, nor have we explored differences in ATP binding or actomyosin attachment/detachment for G256E myosin. Additional impacts of thin filament regulatory proteins were also not considered in our current simulation studies. Finally, our models do not account for additional, more complex behavior of myosin, which includes destabilization of the interacting heads motif (IHM) that occurs with myofibril contractile activation and contributes to the hypercontractility of the G256E mutation8. Methodologically, the 1D COM-based reaction coordinate used in the sMD does not capture coupled conformational changes in the Switch-1, Switch-2, and P-loop loops or the potential for separate release of ADP and Mg²⁺ dissociation. The sMD approach that we employed here is simplistic and was performed at very fast pulling speeds, resulting in unrealistic energy barriers for ADP dissociation. Despite these limitations, we propose that the combination of our cMD and sMD simulations points to structural changes that give rise to a major change in crossbridge cycling and that meaningful changes due to an HCM mutation are detectable with such approaches. In future studies, more complex reaction coordinates and sophisticated methods (e.g., umbrella sampling21) should refine these results, and our current study suggests several factors for consideration when formulating reaction coordinates. We observed three major conformational effects: the Mg2+ ion had strong and distinct interaction energies with WT and G256E myosin; ADP formed unique interactions with WT and G256E, and ADP interacted with residues in the P-loop, central β-sheet, Switch-1, and the SH1-loop along the release coordinate. All of these factors should be accounted for in enhanced sampling simulations. Stopped-flow biochemical experiments involve isolated actomyosin constructs in the absence of load, which may over-report the effects of the G256E mutation on ADP affinity, as load reduces the rate of ADP release from myosin37,38. Additionally, shortened myosin constructs are unable to form the IHM found in native thick filaments. As such, we are unable to assess the combined effects or interplay between alterations of myosin recruitment (ON and OFF states) and crossbridge cycling. Myofibrils derived from hiPSC-CMs may contain fetal isoforms of sarcomeric proteins; however, our maturation protocol results in myofibrils that contain predominantly β-myosin, which is the adult ventricular isoform affected by the G256E mutation, and the crossbridge cycling rate is dominated by myosin isoform. To mitigate the limitations of any one technique, our interdisciplinary research strategy used different models to explore the effects of a mutation at different scales to identify the common feature (ADP release) that G256E affects. Conclusions In summary, we have demonstrated that the MYH7 G256E impairs ADP release as well as the previously reported increase in myosin recruitment from the OFF state8. Together, these changes contribute to the clinically observed HCM phenotype of hypercontractility and impaired relaxation through delayed crossbridge detachment and a greater population of myosin in force bearing states. Our conventional MD results suggest that mutations in the transducer region can allosterically affect the nucleotide binding pocket. Using sMD we developed a predictive method to computationally assess a specific transition step in the crossbridge cycle–ADP release. Our results were corroborated with experimental data obtained from myofibrils obtained from hiPSC-CMs and recombinant sS1 protein, emphasizing the necessity for both computational approaches and validation against experimental data. Similar studies could help elucidate the mechanisms of other specific mutations in myosin implicated in disease and contribute to developing additional sarcomere targeted therapies. Materials and methods Computational model preparation The initial actin+myosin+ADP+Mg2+ (AM.ADP-Mg2+) structure model was generated based on cryo-EM structures of post-powerstroke actin-bound human cardiac myosin solved by Doran et al. (PDB ID: 8EFH, 3.3 Å resolution; PDB ID: 8EFE, 3.8 Å resolution)16,39. Our model includes the myosin motor and actin 5-mer from 8EFH; tropomyosin was excluded (chains O, P). Then we grafted the converter domain, tail and essential light chain (ELC) from 8EFE onto the actomyosin complex. Missing disordered heavy atoms were constructed via Modeller40. The disordered N-terminal extension of the ELC was not modeled. The final WT model includes actin residues (2–377), myosin residues (1–810), and ELC residues (44–195). A model for the mutant G256E actomyosin complex was generated via in silico mutation of the WT structure with UCSF Chimera41,42. The resulting two systems (WT and G256E) were prepared for molecular dynamics (MD) simulation using the AMBER20 simulation package43 and the ff14SB force field44. All-atom explicit solvent simulations were chosen to model the effects of a single amino acid substitution as accurately as possible. Water molecules were treated with the TIP3P force field45. Metal ions were modeled using the Li and Merz parameter set46–48 ADP molecules were treated with the GAFF2 force field49 Partial charges for ADP were derived from a restrained electrostatic potential (resp) fit to quantum mechanics calculations performed with ORCA50 The SHAKE algorithm was used to constrain the motion of hydrogen-containing bonds51 Long-range electrostatic interactions were calculated using the particle mesh Ewald (PME) method. Hydrogen atoms were modeled onto the initial structure using the leap module of AMBER, and each protein was solvated with explicit water molecules in a truncated octahedral box, and then neutralizing + 120 mM counterions (Na+, Cl-) were added. Each system was minimized in three stages. First, hydrogen atoms were minimized for 1000 steps in the presence of 100 kcal mol−¹ Å−2 restraints on all heavy atoms. Second, all solvent atoms were minimized for 1000 steps in the presence of 25 kcal mol−¹ Å−2 restraints on all protein atoms. Third, all atoms were minimized for 8000 steps in the presence of 25 kcal/mol restraints on all backbone heavy atoms (N, O, C, and Cα atoms). After minimization, systems were heated to 310 K using the NVT (constant number of particles, volume, and temperature) ensemble and in the presence of 25 kcal mol−¹ Å−2 restraints on backbone heavy atoms. Next, the systems were equilibrated over five successive stages using the NPT (constant number of particles, pressure, and temperature) ensemble. During the first four stages, the systems were equilibrated for 5 ns in the presence of 25 to 1 kcal mol−¹ Å−2 restraints on backbone heavy atoms. During the final equilibration stage, the systems were equilibrated for 5 ns in the absence of restraints. A 10 Å nonbonded cutoff was used for all preparation and production simulations. Conventional MD production and analysis The equilibrated systems were then simulated using conventional molecular dynamics protocols in the NVT ensemble in triplicate for 500 ns each (2 systems, three replicate simulations per system, 500 ns sampling per replicate simulation = 3 µs net sampling), and coordinates were saved every 10 ps. Molecular analyses were performed using cpptraj52 Cα RMSD values for each simulation were calculated for all residues and relative to the minimized starting structure. Cα RMSF values were calculated after aligning the trajectory to the average structure. Residue-residue contacts were counted if at least one pair of heavy atoms between two residues were within 5 Å of one another, and contacts are presented as the fraction of simulation time that two residues were interacting. Nucleotide dynamics were evaluated by monitoring 6 dihedral angles that describe ADP structure using internal coordinates, as previously described53 The 6 calculated dihedral angles were defined as: θ1 (atoms: O1B-PB-O3A-PA), θ2 (PB-O3A-PA-O5′), θ3 (O3A-PA-O5′-C5′), θ4 (PA-O5′-C5′-C4′), θ5 (O5′-C5′-C4′-O4′), and θ6 (O4′-C1′-N9-C8). Ramachandran-style dihedral angle maps (θ1 vs. θ2, θ3 vs. θ4, and θ5 vs. θ6) describe the relative orientation of the phosphate groups, the sugar ring, and the purine ring. We used these plots to calculate the Shannon entropy and sampling of dihedral angle space as described before53. The orientation of the myosin tail relative to the actin filament was calculated using the “vector” module of cpptraj as previously described54 We also calculated interaction energies between myosin residues and either ADP or Mg2+ using the "linear interaction energy" module. Electrostatic and van der Waals interaction energies were combined. Steered molecular dynamics (sMD) Pre-production protocols were identical to those for the conventional MD simulations except that we employed the hydrogen mass partitioning scheme55 to enable a 4 fs time step, the total time for the heating and equilibration phases was 1 ns, and 25 kcal mol−¹ Å−2 restraints were present on all protein heavy backbone atoms, ADP, and Mg2+ during the entire equilibration. During ~10 ns long production sMD simulations, we steered ADP.Mg2+ from ~3 Å to ~15 Å away from myosin over along the reaction coordinate using a center of mass distance-based reaction coordinate and a force constant of 7.2 kcal mol−¹ Å−2. Specifically, Mg2+, Pα, and Pβ were steered away from the center of mass position of the myosin residues with strong linear interaction energies in the cMD simulations (residues 126, 128, 180, 181, 182, 183, 184, 185, 228, 232, 238, 240, 241, 242, 246, 461, 674, 679). sS1 protein purification Recombinant WT and G256E human β-cardiac sS1 myosins were purified as described previously56. Briefly, human β-cardiac myosin heavy chain (MYH7) was co-expressed with human ELC (MYL3) in the differentiated murine myoblast C2C12 cell line (ATCC) using adenovirus generated in HEK293T cells (ATCC) using the AdEasy Vector System (Qbiogene Inc, USA). The sS1 construct is a truncated version of MYH7 (residues 1-808) followed by a flexible GSG (Gly-Ser-Gly) linker and a carboxy-terminal 8-residue (RGSIDTWV) PDZ binding peptide. The MYL3 construct contains an N-terminal FLAG tag followed by a TEV protease site for purification. C2C12 cells were infected with adenovirus constructs 48 h after differentiation and harvested 4 days after infection in a lysis buffer containing 50 mM NaCl, 20 mM MgCl2, 20 mM imidazole pH 7.5, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 3 mM ATP, 1 mM PMSF, 5% sucrose, and Roche protease inhibitors. The harvested cells were immediately flash frozen in liquid nitrogen. The cell pellets were stored up to 2 months at –80 °C and thawed at room temperature for purification. Cells were lysed in a Dounce homogeniser with 50 strokes, and the lysate was clarified by spinning at 90,377 × g for 30 min at 4 °C. Supernatant was transferred to a fresh tube and bound to anti-FLAG resin for 1.5 h at 4 °C on a nutator. The resin was washed with a buffer (150 mM NaCl, 5 mM MgCl2, 20 mM imidazole pH 7.5, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 3 mM ATP, 1 mM PMSF, 10% sucrose, and Roche protease inhibitors) and then nutated overnight with TEV protease at 4 °C to cleave the ELC-myosin complex from the resin. The next day, the supernatant was purified using a HiTrap Q HP column on a fast protein liquid chromatography (FPLC) apparatus with a gradient of 0–600 mM NaCl over 20 column volumes in a buffer containing 10 mM imidazole pH 7.5, 4 mM MgCl2, 10% sucrose, 1 mM DTT, and 2 mM ATP. Fractions containing human β-cardiac myosin and devoid of mouse skeletal myosin (Coomassie staining on 12.5% SDS PAGE) were pooled and flash frozen in liquid nitrogen. Purified sS1 was buffer exchanged into stopped-flow buffer (120 mM KCl, 5 mM MgCl2, 20 mM MOPS, 1 mM DTT, pH 7.0) overnight prior to experiments. Concentration was measured through a Bradford protein assay (5000201, Bio-Rad). Pyrene-actin labeling Pyrene-actin was prepared fresh and used within ~1 week. First, G-actin was prepared from lyophilized bovine cardiac muscle as described before57. G-actin was diluted to 1 mg/mL before polymerizing with 1 mM ATP, 50 mM KCl, and 2 mM MgCl2 at 4 °C for 2 h. To label, 1.28% w/w N-(1-Pyrene)iodoacetamide (AA00G316, aablocks) was added to F-actin and stirred for 18 h at room temperature. Excess N-(1-Pyrene)iodoacetamide was removed by centrifuging at 2000 × g for 5 min at 4 °C. Labeled F-actin was pelleted by centrifugation at 92,000 × g for 3 h at 4 °C and resuspended in stopped-flow buffer plus minimal ATP (120 mM KCl, 5 mM MgCl2, 20 mM MOPS, 1 mM DTT, 200 μM ATP, pH 7.0) in a glass Dounce homogenizer. Pyrene-actin was stabilized with 1:1 molar phalloidin (5301, AAT Bioquest) for at least 4 h before dialyzing into stopped-flow buffer overnight to remove residual ATP. Stopped-flow kinetics Stopped-flow kinetics experiments were performed at 20 °C using a Hi-Tech “KinetAsyst” stopped-flow (SF-61DX2, TgK Scientific, UK). Measurements were made in 120 mM KCl, 5 mM MgCl2, 20 mM MOPS, 1 mM DTT, pH 7.0 solution. Pyrene fluorescence was measured by excitation at 365 nm and emission through a GG 400 nm cutoff filter. ADP affinity (KADP) was assessed by competitive binding of ATP and ADP to actomyosin. Briefly, 0.125 μM actomyosin (sS1 + pyrene-actin) was rapidly mixed with 25 µM ATP plus various amounts of ADP (0–640 µM). An average trace of at least eight individual traces was generated for each condition and fit with a double exponential. The relative kobs of the fast phase (krel) was plotted vs. ADP concentration and fitted with an equation as described previously58. ADP release rate was measured as previously described59. 0.33 μM sS1 + 0.5 μM pyrene-actin was pre-incubated with 80 μM ADP before rapidly mixing with a large excess (1.5–2.5 mM) of MgATP. Average traces (8-10 individual traces) were fit with a double exponential. The kinetics of the fast phase (kobs,fast) describes the ADP release rate. All concentrations reported are post-mixing. Human induced pluripotent stem cell (hiPSC) culture The parental WTC hiPSC cell line was generated by the Bruce R. Conklin Laboratory at the Gladstone Institutes and University of California, San Francisco (UCSF)60. Cells were maintained using described methods61,62. A detailed cell culture protocol from the Allen Institute for Cell Science can be found on its webpage (http://allencell.org). AICS-0075-085 ACTN2-mEGFP was developed at the Allen Institute for Cell Science (http://allencell.org/cell-catalog) and is available through Coriell. AICS-0097-113 ACTN2-mEGFP MYH7WT/WT and AICS-0097-141 ACTN2-mEGFP MYH7WT/G256E were developed at the Allen Institute for Cell Science and are available on request. hiPSC-cardiomyocyte (hiPSC-CM) differentiation and maintenance hiPSCs were differentiated into cardiomyocytes using an established small-molecule protocol63 with modifications (Fig. 5c)64. Briefly, hiPSCs were seeded onto Matrigel-coated 12-well tissue culture treated plates at densities ranging from 50 × 103 – 125 × 103 cells per well and maintained for 2 days with mTeSR1 media (85850, Stemcell Technologies). On day -1, the day prior to differentiation, mTeSR1 media was supplemented with 1 μM CHIR99021 (13122, Cayman Chemical). On day 0, hiPSCs were induced with 5 μM CHIR99021 in a chemically defined media composed of RPMI 1640 (11875-093, Gibco) + 500 μg/mL bovine serum albumin (A9418, Sigma-Aldrich) + 213 μg/mL L-ascorbic acid (A8960, Sigma-Aldrich), hereafter abbreviated RBA. After 48 hours, the media was replaced with 2 μM Wnt-C59 (S7037, Selleck Chem) in RBA media. On day 4, the media was replaced with RBA media. From day 6 onwards, differentiated cells were maintained in RPMI 1640 + 1X B27 plus insulin (A1895601, Gibco) with media changes every 2–3 days. Solutions for myofibril mechanics Solutions for isolated myofibril mechanics experiments were prepared as previously described65. Experimental solutions were composed of the following: 15 mM total EGTA, 80 mM total MOPS, 5 mM MgATP, 15 mM total creatine phosphate, 83 mM free K, and 52 mM free Na, all free ion concentrations were determined using MAXCHELATOR66). Solutions were adjusted to pH 7.0 with NaOH. Elevated ADP solutions (5.6ADP) were made by replacing 50% of the nucleotide pool with ADP for final concentrations of 2.5 mM ATP and 2.5 mM ADP. Isolated myofibril assay hiPSCs were differentiated into cardiomyocytes as described above. On day 14 post-differentiation, hiPSC-CMs were replated at 100k cells/cm2. On day 16, hiPSC-CMs were subjected to glucose starvation and lactate enrichment for 4 days in DMEM minus glucose (11966-025, Gibco) media supplemented with 4 mM sodium L-lactate (L7022, Sigma Aldrich). hiPSC-CMs were maintained in RPMI 1640 + 1X B27 plus insulin. On day 30, hiPSC-CMs were replated onto 10 kPa polyacrylamide gels in 6-well glass bottom plates (P06-20-1.5-N, Cellvis) patterned with 15 μm width Matrigel lines prepared by microcontact printing as described previously67, and matured for 7 days to promote elongation and formation of cell-to-cell contacts. On day 37, isolated myofibrils for mechanical and kinetic analysis hiPSC-CMs were demembranated in a 1% triton-X pCa 9.0 relaxing solution supplemented with 1:100 protease inhibitor cocktail (P8340, Sigma-Aldrich) for 30 minutes on ice. A cell scraper was then used to gently free isolated myofibrils from permeabilized cells. The supernatant was collected and gently homogenized in a glass Dounce homogenizer ~5–7X. Myofibrils were mounted into a custom-built apparatus with rapid Ca2+ solution switching (15 °C) as previously described25,65. Myofibrils were initially suspended in pCa 9.0 relaxing solution before rapidly switching to pCa 4.0 maximal activation solution. Myofibrils were also activated in a pCa 5.6 sub-maximal activation solution and a pCa 5.6ADP solution containing 50% ADP (2.5 mM ATP, 2.5 mM ADP) solution to challenge myofibrils with ADP. Myosin separation gel Cell pellets were flash frozen in liquid nitrogen and stored at −80 °C until ready to use. Cell pellets were resuspended in ~3x volume myofibril lysis buffer (5% SDS, 50 mM Tris (pH 8.5), 0.75% sodium deoxycholate, and 1x protease inhibitor cocktail). Sarcomeres and proteins were denatured by boiling samples for 10 min. Concentration was measured by blotting for total protein and adjusting loading. Myosin separation gel was made by combining a stacking gel (2.95% Acrylamide with 10% glycerol and pH 8.8) and resolving gel (6% Acrylamide with 10% glycerol and pH 8.8). Separation gels were run for 2 h and 15 min at 32 mA in a cold room (4 °C). Gels were stained for total protein using One-Step Lumitein™ Protein Gel Stain, 1X following manufacturer instructions. Gels were imaged on a BioRad ChemiDoc using SYPRO Ruby protein stain setting. Statistics Conventional MD (cMD) simulations were run in triplicate (n = 3) for each genotype. Statistics on cMD analyses were performed with a Student’s t test. Steered MD (sMD) simulations were run 60 times (n = 60) for each genotype. A Student’s t test was used to compare work required for ADP dissociation for WT vs. G256E. For myofibrils, within each genotype, differences between pCa 5.6 and pCa 5.6ADP were assessed using paired Wilcoxon signed-rank tests. To evaluate genotype-dependent effects of ADP, the change (Δ = pCa 5.6–pCa 5.6ADP) was calculated for each myofibril and compared between genotypes using a Wilcoxon rank-sum test. (Individual myofibrils: WT n = 15, G256E n = 13; Biological replicates: WT N = 8, G256E N = 3). Stopped-flow experiments were replicated 3–4 times, and a Student’s t test was used to determine significance. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplemental Data Reporting Summary Acknowledgements This work used the Extreme Science and Engineering Discovery Environment (XSEDE) resource COMET through allocation TG-MCB200100 to M.R. XSEDE was supported by the National Science Foundation grant no. ACI-1548562. Partial funding for M.C.C. was provided by awards T32HL007828 and K99HL173646 from the National Heart, Lung, and Blood Institute. Partial support for K.Y.K. was provided by NIBIB T32EB032787. Partial support for K.Y.K. was provided by NIH-NHLBI R00HL159224 awarded to J.D.P. This research was supported by the University of Washington Center for Translational Muscle Research (CTMR) via the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health award no. P30AR074990 to M.R. This work was initiated and supported by the NIH/NIGMS grant 1RM1 GM131981-03 to J.A.S., K.M.R. and M.R. and supported by NHLBI R01HL128368 to M.R. This work was made possible by the Allen Institute for Cell Science team, which generated the stem cell lines. The parental WT unedited hiPSC line, WTC, was provided by the Bruce R. Conklin Laboratory at the Gladstone Institutes and UCSF. The Allen Institute for Cell Science wishes to thank the Allen Institute for Cell Science founder, Paul G. Allen, for his vision, encouragement, and support. The content is solely the responsibility of the authors and does not necessarily represent the official view of the NHLBI or the NIH. Author contributions Designed research, M.C.C. and M.R. Provided purified protein, D.P., R.R.G., and K.M.R. Performed research, M.C.C., K.Y.K., T.S.M., D.P., and R.R.G. Analyzed data, M.C.C. and K.Y.K. Manuscript writing, K.Y.K., M.C.C., and M.R. Resources and Funding, M.R., M.C.C, J.D.P., and J.A.S. Peer review Peer review information Communications Chemistry thanks the anonymous reviewers for their contribution to the peer review of this work. Data availability Data generated for this study is available in the Supplementary Data file. Raw traces for biochemical and myofibril data are available upon request. Structures used for generating starting structures for MD simulations can be accessed through the following PDB codes: 8EFH and 8EFE. Trajectory files for molecular dynamics simulations are available upon request. Code availability Custom R code used to analyze myofibril traces and Python code used to analyze molecular dynamics simulations are available upon request. Competing interests M.R. is a consultant for Kardigan Bio and Bristol Myers Squibb; serves on the scientific advisory board for FilamenTech; and has equity in StemCardia, Inc and KineaBio, Inc; none of the current work is in conflict with these associations. J.A.S. is cofounder and on the Scientific Advisory Board of Cytokinetics, Inc., a company developing small molecule therapeutics for treatment of hypertrophic cardiomyopathy, but this work was done independently. J.A.S. is cofounder and Executive Chair, and K.M.R. is cofounder and Research and Clinical Advisor, of Kainomyx, Inc., a company developing small molecule therapeutics targeting cytoskeletal proteins for multiple diseases, but this work was done independently. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Supplementary information The online version contains supplementary material available at 10.1038/s42004-026-02048-w. References Associated Data Supplementary Materials Supplementary Information Description of Additional Supplementary Files Supplemental Data Reporting Summary Data Availability Statement Data generated for this study is available in the Supplementary Data file. Raw traces for biochemical and myofibril data are available upon request. Structures used for generating starting structures for MD simulations can be accessed through the following PDB codes: 8EFH and 8EFE. Trajectory files for molecular dynamics simulations are available upon request. Custom R code used to analyze myofibril traces and Python code used to analyze molecular dynamics simulations are available upon request.