Advertisement

Functional divergence of the sarcomeric myosin, MYH7b, supports species-specific biological roles

Open AccessPublished:November 02, 2022DOI:https://doi.org/10.1016/j.jbc.2022.102657

      Abstract

      Myosin heavy chain 7b (MYH7b) is an evolutionarily ancient member of the sarcomeric myosin family, which typically supports striated muscle function. However, in mammals alternative splicing prevents MYH7b protein production in cardiac and most skeletal muscles and limits expression to a subset of specialized muscles and certain non-muscle environments. In contrast, MYH7b protein is abundant in python cardiac and skeletal muscles. Although the MYH7b expression pattern diverges in mammals versus reptiles, MYH7b shares high sequence identity across species. So, it remains unclear how mammalian MYH7b function may differ from that of other sarcomeric myosins and whether human and python MYH7b motor functions diverge as their expression patterns suggest. Thus, we generated recombinant human and python MYH7b protein and measured their motor properties to investigate any species-specific differences in activity. Our results reveal that despite having similar working strokes, the MYH7b isoforms have slower actin-activated ATPase cycles and actin sliding velocities than human cardiac β-MyHC. Furthermore, python MYH7b is tuned to have slower motor activity than human MYH7b due to slower kinetics of the mechanochemical cycle. We found that the MYH7b isoforms adopt a higher proportion of myosin heads in the ultra-slow, super-relaxed state compared to human cardiac β-MyHC. These findings are supported by molecular dynamics simulations that predict MYH7b preferentially occupies myosin active site conformations similar to those observed in the structurally inactive state. Together, these results suggest that MYH7b is specialized for slow, energy-conserving motor activity and that differential tuning of MYH7b orthologs contributes to species-specific biological roles.

      Keywords

      Abbreviations and Nomenclature:

      MyHC (myosin heavy chain), S1 (subfragment-1), Mant-ATP (2'-(or-3')-O-(N-Methylanthraniloyl) Adenosine 5'-Triphosphate), SRX (super-relaxed state), DRX (disordered-relaxed state), EOM (extraocular muscle), IHM (interacting heads motif)

      INTRODUCTION

      Members of the sarcomeric myosin heavy chain (MYH) family function as actin-based motors in striated muscle where they convert chemical energy from ATP hydrolysis into the mechanical force necessary for contraction. Studies have focused almost exclusively on the eight conventional sarcomeric myosin-II genes principally expressed in the major skeletal muscles and the heart (MYH1, -2, -3, -4, -6, -7, -8, and -13). These sarcomeric myosins consist of a globular motor domain with ATP and actin binding sites, a neck domain bound by myosin light chains, and an alpha-helical tail domain that drives assembly into the thick filament of the sarcomere. Despite high sequence conservation across this myosin gene family, the properties of the mechanochemical cycle differ greatly among isoforms allowing for specific tuning of muscle physiology (
      • Resnicow D.I.
      • Deacon J.C.
      • Warrick H.M.
      • Spudich J.A.
      • Leinwand L.A.
      Functional diversity among a family of human skeletal muscle myosin motors.
      ,
      • Deacon J.C.
      • Bloemink M.J.
      • Rezavandi H.
      • Geeves M.A.
      • Leinwand L.A.
      Identification of functional differences between recombinant human α and β cardiac myosin motors.
      ,
      • Bloemink M.J.
      • Deacon J.C.
      • Resnicow D.I.
      • Leinwand L.A.
      • Geeves M.A.
      The Superfast Human Extraocular Myosin Is Kinetically Distinct from the Fast Skeletal IIa, IIb, and IId Isoforms.
      ).
      Given the deep scrutiny of the sarcomeric myosin family due to its essential role in life-critical organs, it was surprising to discover additional members of this family, including MYH7b, after bioinformatic annotation of the human genome (
      • Desjardins P.R.
      • Burkman J.M.
      • Shrager J.B.
      • Allmond L.A.
      • Stedman H.H.
      Evolutionary Implications of Three Novel Members of the Human Sarcomeric Myosin Heavy Chain Gene Family.
      ). This oversight was the result of MYH7b’s atypical expression pattern in mammals. MYH7b transcripts are present in mammalian cardiac and major skeletal muscles, but a regulated alternative splicing event introduces a premature stop codon that prevents MYH7b protein production in these tissues (
      • Bell M.L.
      • Buvoli M.
      • Leinwand L.A.
      Uncoupling of Expression of an Intronic MicroRNA and Its Myosin Host Gene by Exon Skipping.
      ). Moreover, forced cardiac expression of MYH7b protein in a transgenic mouse model resulted in a severe dilated cardiomyopathy, underscoring the importance of evolutionary silencing of MYH7b protein production in the mammalian heart (
      • Peter A.K.
      • Rossi A.C.
      • Buvoli M.
      • Ozeroff C.D.
      • Crocini C.
      • Perry A.R.
      • Buvoli A.E.
      • Lee L.A.
      • Leinwand L.A.
      Expression of Normally Repressed Myosin Heavy Chain 7b in the Mammalian Heart Induces Dilated Cardiomyopathy.
      ). However, MYH7b protein is found in the sarcomeres of certain mammalian specialized muscles such as extraocular muscles, muscle spindles, and the upper esophagus (
      • Rossi A.C.
      • Mammucari C.
      • Argentini C.
      • Reggiani C.
      • Schiaffino S.
      Two novel/ancient myosins in mammalian skeletal muscles: MYH14/7b and MYH15 are expressed in extraocular muscles and muscle spindles: MYH14/7b and MYH15 expression in mammalian skeletal muscles.
      ,
      • Mu L.
      • Wang J.
      • Su H.
      • Sanders I.
      Adult Human Upper Esophageal Sphincter Contains Specialized Muscle Fibers Expressing Unusual Myosin Heavy Chain Isoforms.
      ). Surprisingly, MYH7b protein has also been detected at low abundance in non-muscle tissues, and mutations in MYH7b are linked to hereditary hearing loss (
      • Bell M.L.
      • Buvoli M.
      • Leinwand L.A.
      Uncoupling of Expression of an Intronic MicroRNA and Its Myosin Host Gene by Exon Skipping.
      ,
      • Haraksingh R.R.
      • Jahanbani F.
      • Rodriguez-Paris J.
      • Gelernter J.
      • Nadeau K.C.
      • Oghalai J.S.
      • Schrijver I.
      • Snyder M.P.
      Exome sequencing and genome-wide copy number variant mapping reveal novel associations with sensorineural hereditary hearing loss.
      ,
      • Rubio M.D.
      • Johnson R.
      • Miller C.A.
      • Huganir R.L.
      • Rumbaugh G.
      Regulation of Synapse Structure and Function by Distinct Myosin II Motors.
      ). Given this unique expression pattern, it remains unclear how MYH7b functions in certain mammalian specialized muscles and non-muscle tissue, yet is not well tolerated in major muscles like the heart.
      Phylogenetic analysis of the sarcomeric myosin family classifies MYH7b as an ancient myosin that predates the emergence of the major cardiac and skeletal myosin isoforms (
      • Desjardins P.R.
      • Burkman J.M.
      • Shrager J.B.
      • Allmond L.A.
      • Stedman H.H.
      Evolutionary Implications of Three Novel Members of the Human Sarcomeric Myosin Heavy Chain Gene Family.
      ). Immunostaining has identified MYH7b protein in chicken heart and skeletal muscle (
      • Rossi A.C.
      • Mammucari C.
      • Argentini C.
      • Reggiani C.
      • Schiaffino S.
      Two novel/ancient myosins in mammalian skeletal muscles: MYH14/7b and MYH15 are expressed in extraocular muscles and muscle spindles: MYH14/7b and MYH15 expression in mammalian skeletal muscles.
      ,
      • Machida S.
      • Noda S.
      • Takao A.
      • Nakazawa M.
      • Matsuoka R.
      Expression of slow skeletal myosin heavy chain 2 gene in Purkinje fiber cells in chick heart.
      ), and in this study we report MYH7b protein identification in python heart and skeletal muscles. Given the difference in expression pattern of MYH7b between reptiles/birds and mammals, it appears that the MYH7b gene has evolved to meet different physiological needs of these distinct vertebrate species. Based on this divergent expression pattern across species and the ancient evolutionary roots of MYH7b, we hypothesized that: 1) human MYH7b motor properties diverge from other well-characterized human sarcomeric myosin family members and 2) human and python MYH7b motor functions are differentially tuned to accommodate species-specific roles. We examined these hypotheses using a comprehensive suite of biochemical and biophysical analyses to test MYH7b motor activity. Ultimately, our study of MYH7b defines the activity of a previously uncharacterized sarcomeric myosin motor and provides an evolutionary perspective of motor function within this highly conserved protein family.

      RESULTS

      MYH7b is detected at the RNA and protein levels in python muscles

      Most mammalian muscle tissues, including the heart and major skeletal muscles, do not produce MYH7b protein due to an alternative splicing event that shifts a premature stop codon into frame (
      • Bell M.L.
      • Buvoli M.
      • Leinwand L.A.
      Uncoupling of Expression of an Intronic MicroRNA and Its Myosin Host Gene by Exon Skipping.
      ,
      • Lee L.A.
      • Broadwell L.J.
      • Buvoli M.
      • Leinwand L.A.
      Nonproductive Splicing Prevents Expression of MYH7b Protein in the Mammalian Heart.
      ). We investigated whether the MYH7b expression pattern observed in mammals is conserved in reptiles, an evolutionarily distinct class of vertebrates, by examining the cardiac and skeletal muscle myosin composition of two distinct species of python, Burmese python (Python bivittatus) and Ball python (Python regius). We detected MYH7b RNA and unique MYH7b peptides in the skeletal muscle of both python species and in the cardiac muscle of Ball pythons but not Burmese pythons (Figures 1 and S1). Notably, pythons express a broad array of myosins in their muscles where MYH15 (the other recently identified ancient MYH gene) appears to be the predominant cardiac myosin isoform, and MYH1 is the most abundant skeletal myosin isoform (Figure 1A and B and Tables S1 and S2). We also assessed the myosin light chain composition in these tissues to understand which myosin light chains are present (Figure S1 A-C). All python myosin light chain sequences share 69-99% sequence identity with their mouse myosin light chain orthologs. Using an RT-PCR assay that discriminates between the two MYH7b splice forms, we determined that the alternative splicing event first discovered in mammals (
      • Bell M.L.
      • Buvoli M.
      • Leinwand L.A.
      Uncoupling of Expression of an Intronic MicroRNA and Its Myosin Host Gene by Exon Skipping.
      ) also prevents MYH7b protein production in Burmese python hearts (Figure S1E). Ball python cardiac MYH7b RNA also undergoes exon skipping, but a proportion of unskipped RNA that can encode protein is readily detectable (Figure S1E). Based on these results, it appears that, unlike the mammalian ortholog, MYH7b operates in a conventional role in python cardiac and skeletal muscle.
      Figure thumbnail gr1
      Figure 1Mass spectrometry detects MYH7b protein in python muscle. Normalized spectral quantity of python myosin heavy chain proteins detected in cardiac and skeletal samples of two distinct python species, A. Burmese python (n=2) and B. Ball python (n=1). MYH7b is present in skeletal muscle of Burmese pythons and the cardiac and skeletal muscle of Ball pythons. Note that all mass spectrometry values are semi-quantitative as these experiments do not include labeled internal standards. The mass spectrometry data are summarized in . Data are graphed as mean ± SD. C. Percent identity and sequence similarity to human MYH7b calculated from sequence alignments using NCBI BLAST of MYH7b protein sequence across species. NCBI accession codes are as follows: human NP_065935.4, mouse NP_001078847.1, python XP_007419944.1, chicken NP_989918.3, frog XP_031750297.1, zebrafish NP_001311408.1.
      Despite the divergent expression pattern observed in mammals versus reptiles, the MYH7b sequence is highly conserved across species (Figure 1C). Therefore, it is unclear whether inherent molecular function diverges between human and python MYH7b or whether cell type-specific factors within the different tissues that express MYH7b externally tune MYH7b properties to promote species-specific roles. To directly compare human and python MYH7b motor properties, we produced recombinant myosin motor domains consisting of the myosin subfragment-1 (S1) region (Figure S2) and performed a comprehensive analysis of their catalytic motor properties. Our studies use human β myosin heavy chain (β-MyHC, encoded by the MYH7 gene), the myosin isoform expressed in cardiac and slow skeletal muscle, as a comparison because β-MyHC is the closest in sequence identity to human MYH7b and has been extensively characterized in vitro (
      • Deacon J.C.
      • Bloemink M.J.
      • Rezavandi H.
      • Geeves M.A.
      • Leinwand L.A.
      Identification of functional differences between recombinant human α and β cardiac myosin motors.
      ,
      • Sommese R.F.
      • Sung J.
      • Nag S.
      • Sutton S.
      • Deacon J.C.
      • Choe E.
      • Leinwand L.A.
      • Ruppel K.
      • Spudich J.A.
      Molecular consequences of the R453C hypertrophic cardiomyopathy mutation on human -cardiac myosin motor function.
      ).

      Human and python MYH7b have slower actin-activated ATPase activity than human β-MyHC

      We used an NADH-coupled system to assess the steady-state actin-activated ATPase activity of our myosin constructs. We measured the ATPase rates of human β-MyHC, human MYH7b, and python MYH7b at increasing actin concentrations and fit the Michaelis-Menten kinetics equation to the data to calculate the maximum actin-activated ATPase rate (per myosin molecule, kcat) and apparent actin affinity (KM) for each myosin construct (Figure 2A and Table S4). Human MYH7b has a maximal ATPase rate of 0.80 ± 0.08 s-1, strikingly lower by 49% compared to the canonical slow human β-MyHC isoform (1.57 ± 0.18 s-1, p < 0.0001). The maximal ATPase rate for python MYH7b (0.60 ± 0.08 s-1) was substantially lower than both β-MyHC (a 62% decrease, p < 0.0001) and human MYH7b (a 25% decrease, p = 0.0350). The maximum actin-activated ATPase rate is inversely proportional to the total cycle time (tcycle) through the mechanochemical cycle for one myosin (kcat = 1/tcycle ) (
      • Spudich J.A.
      Hypertrophic and dilated cardiomyopathy: four decades of basic research on muscle lead to potential therapeutic approaches to these devastating genetic diseases.
      ). Thus, when comparing total cycle time between isoforms from the values measured, β-MyHC has the shortest cycle time (∼0.6 s) and human and python MYH7b have longer total cycle times comparatively (∼1.3 s and ∼ 1.7 s, respectively). Moreover, the KM of human MYH7b (35.9 ± 9.8 μM) is 62% lower than human β-MyHC (93.4 ± 56.5 μM, p = 0.0096), indicating that MYH7b binds more tightly to actin. The KM of python MYH7b (79.3 ± 16.1 μM) is not significantly different than that of human β-MyHC indicating that altered actin affinity cannot account for the decreased kcat for python MYH7b. Finally, we assessed overall enzymatic efficiency of actin-activated ATPase for each construct determined by kcat/KM. We observed similar measures for human β-MyHC (0.025 ± 0.015 s-1μM-1) and human MYH7b (0.023 ± 0.006 s-1μM-1) whereas comparatively python MYH7b is a less efficient motor (0.008 ± 0.003 s-1μM-1).
      Figure thumbnail gr2
      Figure 2Human and python MYH7b have slow motor properties. A. Actin-activated ATPase curves of human β-MyHC S1, human MYH7b S1, and python MYH7b S1. Each plot shows the average of all technical replicates and error bars represent SD. Data were fit to a Michaelis-Menten kinetics equation to obtain the kcat and KM values summarized in . The curve fit is represented by a solid line with shading to indicate the 95% confidence intervals. The average of ATPase curves run on the same day represent technical replicates (hβ-MyHC n = 9, hMYH7b n = 9, and pMYH7b n = 5) and at least 4 different purifications (biological replicates) are represented for each construct. B. in vitro motility velocities of human β-MyHC S1, human MYH7b S1, and python MYH7b S1. The average velocity of each technical replicate (each motility video) is graphed with error bars representing SD. Data were collected for at least 3 independent protein purifications (biological replicates). Technical replicates: hβ-MyHC n = 11, hMYH7b n = 12, and pMYH7b n = 11. Data are summarized in . **** indicates p < 0.0001.

      Human and python MYH7b actin sliding velocities do not differ from each other but are slower than human β-MyHC

      In order to understand the mechanical properties of MYH7b, we used an unloaded in vitro motility assay to determine the actin sliding velocities of human and python MYH7b compared to human β-MyHC (Figure 2B, Table S5, and Movies S1-S3). For these experiments we used PDZ-based chemistry to immobilize our C-terminally tagged S1 constructs to a coverslip and monitored the velocity at which each myosin moved fluorescently labeled actin. Experiments measuring the actin sliding velocity as a function of myosin concentration were conducted to ensure surface saturation (Figure S3). Human MYH7b and python MYH7b actin sliding velocities (0.597 ± 0.057 μm/s and 0.536 ± 0.076 μm/s, respectively) were lower compared to human β-MyHC (1.017 ± 0.050 μm/s, p < 0.0001 for both). There was no discernable difference between human and python MYH7b actin sliding velocities.

      Stopped-flow experiments reveal that human and python MYH7b have slow kinetics

      We next performed stopped-flow experiments to better understand the kinetics of key steps of the mechanochemical cycle that may underly the differences in actin-activated ATPase activity observed between human MYH7b and python MYH7b.The dissociation of human and python MYH7b from pyrene-labeled actin was best described by a single exponential, which provides an observed rate constant (kobs). The rates of ATP-induced dissociation of myosin from actin are plotted against ATP concentration in Figure 3A (stopped-flow data are summarized in Table S6). A linear dependence at low ATP concentrations provides the second order rate constant for ATP binding (K1k+2, Scheme 1). The ATP binding rate for human MYH7b was faster (6.7 ± 1.4 μM-1 s-1) than the rate previously measured for human β-MyHC S1 (∼4.4 μM-1 s-1) (
      • Nag S.
      • Sommese R.F.
      • Ujfalusi Z.
      • Combs A.
      • Langer S.
      • Sutton S.
      • Leinwand L.A.
      • Geeves M.A.
      • Ruppel K.M.
      • Spudich J.A.
      Contractility parameters of human β-cardiac myosin with the hypertrophic cardiomyopathy mutation R403Q show loss of motor function.
      ), while python MYH7b had a slower rate than both (1.2 ± 0.0 μM-1 s-1). Over the full range of ATP concentrations, the data are described by a hyperbolic fit (Equation 1), which results in the maximum dissociation rate (k+2) and the ATP binding affinity in the initial step (1/K1). Both human MYH7b (79.8 ± 17.7 μM) and python MYH7b (157.3 ± 11.5 μM) had a much tighter affinity for ATP (smaller 1/K1) compared to values for human β-MyHC (∼365.7 μM) (
      • Nag S.
      • Sommese R.F.
      • Ujfalusi Z.
      • Combs A.
      • Langer S.
      • Sutton S.
      • Leinwand L.A.
      • Geeves M.A.
      • Ruppel K.M.
      • Spudich J.A.
      Contractility parameters of human β-cardiac myosin with the hypertrophic cardiomyopathy mutation R403Q show loss of motor function.
      ). However, human MYH7b ATP affinity was 2-fold tighter than that of python MYH7b (p = 0.0352). Similarly, human MYH7b (520.7 ± 7.0 s-1) and python MYH7b (181.2 ± 6.9 s-1) have a slower maximum actin dissociation rate compared to human β-MyHC (∼991 s-1). It is worth noting that the dissociation rate for python MYH7b is over 2.5-fold slower than human MYH7b (p = 0.0004).
      Figure thumbnail gr3
      Figure 3Human and python MYH7b have distinct kinetic properties. A. ATP-induced dissociation of myosin S1 from actin plotted as observed rate constants against ATP concentration. The python kobs values were lower than the human values at all ATP concentrations. The kobs vs. ATP concentration data were fit to a hyperbolic equation and yield values of k+2 and 1/K1 as described in the methods. B. Myosin S1 affinity for ADP. As ADP concentration increases, the observed rate constants decrease and can be fit to a hyperbolic equation. The reaction rate constants were slower for python MYH7b (see inset) at all ADP concentrations than for human MYH7b, and analysis of the curve indicates that human MYH7b has a weaker affinity for ADP. Final ATP concentrations were 20 μM for human MYH7b and 50 μM for python MYH7b. Data represent mean ± SD, n = 2 for each experiment; values and statistics are summarized in .
      Figure thumbnail sc1
      Scheme 1ATP or ADP binding to actomyosin leading to the dissociation of myosin form actin. A: actin, M: myosin/S1, T: ATP, D: ADP.
      We next determined the ADP affinity for each myosin S1 construct using a competition assay where increasing concentrations of ADP were rapidly mixed alongside a fixed ATP concentration with S1, resulting in a competition between the ATP and ADP for the nucleotide binding site of myosin (Figure 3B). As the concentration of ADP increases, the kobs decreases as more ADP is available to bind myosin. Plotting the kobs against the ADP concentration can be described by a hyperbolic fit (Equation 2), which provides a measure of the ADP affinity (KADP). Human MYH7b had a weaker ADP affinity (120.7 ± 29.5 μM) compared to both python MYH7b (17.0 ± 3.2 μM, p = 0.0386) and human β-MyHC (∼6.1 μM) (
      • Nag S.
      • Sommese R.F.
      • Ujfalusi Z.
      • Combs A.
      • Langer S.
      • Sutton S.
      • Leinwand L.A.
      • Geeves M.A.
      • Ruppel K.M.
      • Spudich J.A.
      Contractility parameters of human β-cardiac myosin with the hypertrophic cardiomyopathy mutation R403Q show loss of motor function.
      ). Taken together, these results indicate that human and python MYH7b have appreciably slower kinetics than human β-MyHC.

      Human and python MYH7b have similar step sizes, but slower detachment kinetics than human β-MyHC

      We next sought to determine the changes in motor function that contribute to the slower actin sliding velocities of the MYH7b constructs compared to β-MyHC. The speed in the motility assay at saturating myosin concentrations is set by the displacement of the myosin working stroke (d) and the amount of time that myosin spends attached to actin at saturating ATP (ton). We used a single-molecule optical trapping assay (Figure 4A) based on the three-bead geometry (
      • Finer J.T.
      • Simmons R.M.
      • Spudich J.A.
      Single myosin molecule mechanics: piconewton forces and nanometre steps.
      ) to define the size of the working stroke for human and python MYH7b (
      • Blackwell T.
      • Stump W.T.
      • Clippinger S.R.
      • Greenberg M.J.
      Computational Tool for Ensemble Averaging of Single-Molecule Data.
      ,
      • Greenberg M.J.
      • Shuman H.
      • Ostap E.M.
      Inherent Force-Dependent Properties of β-Cardiac Myosin Contribute to the Force-Velocity Relationship of Cardiac Muscle.
      ). In this assay, a flow cell is coated with surface beads that are sparsely decorated with myosin S1. Within the flow cell, an actin filament is suspended between two optically trapped beads and brought close to a surface bead to probe for interactions with myosin S1. Actomyosin binding interactions are associated with decreased variance of trapped bead positions and a shift in the mean bead positions, the magnitude of which reports on the step size of myosin S1 undergoing its working stroke (Figure 4B). The distribution of attachment durations can be used to calculate the rate of dissociation of myosin from actin. These experiments are conducted at low ATP concentrations to facilitate the observation of binding interactions.
      Figure thumbnail gr4
      Figure 4Mechanical and kinetic characterization of human and python MYH7b by optical trapping. A. Schematic of three-bead optical trapping assay in which an actin filament is suspended between two optically trapped beads and lowered on to a surface bead that is sparsely coated with myosin. B. Sample data traces for human MYH7b S1 (red) and python MYH7b S1 (green). Binding events, which are associated with reduced variance of the bead position and a shift in the mean bead position, are marked by black bars. C. Cumulative distribution of binding interaction durations observed for human MYH7b S1 (red data trace, black fit line) and python MYH7b S1 (green data trace, gray fit line) at 1 μM ATP. The detachment rate of python MYH7b is faster than the rate for human MYH7b (p < 0.001) D. Cumulative distribution of total step sizes observed for human MYH7b S1 (red data trace, black fit line) and python MYH7b S1 (green data trace, gray fit line). Step sizes are reported as mean ± SEM. The total working strokes of python and human MYH7b are not significantly different (p = 0.27). E. Ensemble averages of individual binding events for human MYH7b S1 (left, red) and python MYH7b S1 (right, green) showing a two-substep working stroke.
      We measured the detachment kinetics of human and python MYH7b S1 at sub-saturating ATP concentrations (1 μM ATP). Under these conditions, the rate of actomyosin dissociation is limited by the rate of ATP binding, not ADP release as seen in the motility assays. We found that the detachment rate for python MYH7b [3.7 (+0.3/-0.2) s-1] was slower than the detachment rate for human MYH7b [9.9 (+0.7/-0.6) s-1; p < 0.001] (Figure 4B and C). This demonstrates that the rate of ATP-induced actomyosin dissociation is slower for python relative to human MYH7b, consistent with the second-order rates of ATP-induced dissociation at low ATP measured in the stopped-flow (K1k+2). Interestingly, the detachment rate for python MYH7b, which is expressed in python hearts, is similar to that previously observed for human β-MyHC myosin (∼3 s-1) (
      • Woody M.S.
      • Greenberg M.J.
      • Barua B.
      • Winkelmann D.A.
      • Goldman Y.E.
      • Ostap E.M.
      Positive cardiac inotrope omecamtiv mecarbil activates muscle despite suppressing the myosin working stroke.
      ). These results demonstrate a biophysical difference between human and python MYH7b.
      We also measured the size of the myosin working stroke. We observed similar total step sizes for human MYH7b (5.3 ± 0.4 nm) and python MYH7b (4.8 ± 0.2 nm; p = 0.27) (Figure 4D). These step sizes are consistent with those previously observed for human and porcine β-cardiac myosin (∼5-7 nm) (
      • Deacon J.C.
      • Bloemink M.J.
      • Rezavandi H.
      • Geeves M.A.
      • Leinwand L.A.
      Identification of functional differences between recombinant human α and β cardiac myosin motors.
      ,
      • Nag S.
      • Sommese R.F.
      • Ujfalusi Z.
      • Combs A.
      • Langer S.
      • Sutton S.
      • Leinwand L.A.
      • Geeves M.A.
      • Ruppel K.M.
      • Spudich J.A.
      Contractility parameters of human β-cardiac myosin with the hypertrophic cardiomyopathy mutation R403Q show loss of motor function.
      ,
      • Greenberg M.J.
      • Shuman H.
      • Ostap E.M.
      Inherent Force-Dependent Properties of β-Cardiac Myosin Contribute to the Force-Velocity Relationship of Cardiac Muscle.
      ,
      • Woody M.S.
      • Greenberg M.J.
      • Barua B.
      • Winkelmann D.A.
      • Goldman Y.E.
      • Ostap E.M.
      Positive cardiac inotrope omecamtiv mecarbil activates muscle despite suppressing the myosin working stroke.
      ,
      • Palmiter K.A.
      • Tyska M.J.
      • Haeberle J.R.
      • Alpert N.R.
      • Fananapazir L.
      • Warshaw D.M.
      R403Q and L908V mutant b-cardiac myosin from patients with familial hypertrophic cardiomyopathy exhibit enhanced mechanical performance at the single molecule level.
      ,
      • Sung J.
      • Nag S.
      • Mortensen K.I.
      • Vestergaard C.L.
      • Sutton S.
      • Ruppel K.
      • Flyvbjerg H.
      • Spudich J.A.
      Harmonic force spectroscopy measures load-dependent kinetics of individual human β-cardiac myosin molecules.
      ). Next, we examined whether there are substeps to the MYH7b working stroke. Many myosin isoforms, including β-MyHC myosin, have a two-substep working stroke (
      • Greenberg M.J.
      • Shuman H.
      • Ostap E.M.
      Inherent Force-Dependent Properties of β-Cardiac Myosin Contribute to the Force-Velocity Relationship of Cardiac Muscle.
      ,
      • Capitanio M.
      • Canepari M.
      • Cacciafesta P.
      • Lombardi V.
      • Cicchi R.
      • Maffei M.
      • Pavone F.S.
      • Bottinelli R.
      Two independent mechanical events in the interaction cycle of skeletal muscle myosin with actin.
      ,
      • Greenberg M.J.
      • Lin T.
      • Goldman Y.E.
      • Shuman H.
      • Ostap E.M.
      Myosin IC generates power over a range of loads via a new tension-sensing mechanism.
      ,
      • Laakso J.M.
      • Lewis J.H.
      • Shuman H.
      • Ostap E.M.
      Myosin I Can Act As a Molecular Force Sensor.
      ,
      • Veigel C.
      • Coluccio L.M.
      • Jontes J.D.
      • Sparrow J.C.
      • Milligan R.A.
      • Molloy J.E.
      The motor protein myosin-I produces its working stroke in two steps.
      ,
      • Veigel C.
      • Wang F.
      • Bartoo M.L.
      • Sellers J.R.
      • Molloy J.E.
      The gated gait of the processive molecular motor, myosin V.
      ), which can be resolved using ensemble averaging of individual binding events (
      • Blackwell T.
      • Stump W.T.
      • Clippinger S.R.
      • Greenberg M.J.
      Computational Tool for Ensemble Averaging of Single-Molecule Data.
      ,
      • Chen C.
      • Greenberg M.J.
      • Laakso J.M.
      • Ostap E.M.
      • Goldman Y.E.
      • Shuman H.
      Kinetic Schemes for Post-Synchronized Single Molecule Dynamics.
      ). Ensemble averaging involves post-synchronization of binding events and averaging forward in time from actomyosin attachment (time-forward average) or backward in time from actomyosin detachment (time-reversed average). If the working stroke consists of multiple substeps, there will be a difference in displacement between the time-forward and time-reversed averages equal to the size of the second displacement. Our data clearly demonstrate a two-substep working stroke for both human and python MYH7b (Figure 4E), demonstrating that the mechanics of the MYH7b working stroke are similar to β-MyHC. Taken together, our results demonstrate that the reduced velocity in the motility assay are due to changes in the myosin kinetics and not mechanics.

      Human MYH7b and python MYH7b display differences in SRX/DRX proportions

      Muscle function and energy usage by myosin can be modulated by shifting the equilibrium of myosin into different functional states. Myosin ATPase activity is fastest when actomyosin crossbridge cycling occurs, but ATP turnover decreases when myosin is unavailable to bind actin. Non-actin-bound myosin can adopt two distinct functional states: the disordered-relaxed state (DRX; 100-fold slower than actin-activated rates) or the super-relaxed state (SRX; ultra-low ATPase, 1000-fold slower than actin-activated rates) (
      • Anderson R.L.
      • Trivedi D.V.
      • Sarkar S.S.
      • Henze M.
      • Ma W.
      • Gong H.
      • Rogers C.S.
      • Gorham J.M.
      • Wong F.L.
      • Morck M.M.
      • Seidman J.G.
      • Ruppel K.M.
      • Irving T.C.
      • Cooke R.
      • Green E.M.
      • Spudich J.A.
      Deciphering the super relaxed state of human β-cardiac myosin and the mode of action of mavacamten from myosin molecules to muscle fibers.
      ,
      • Nag S.
      • Trivedi D.V.
      To lie or not to lie: Super-relaxing with myosins.
      ,
      • Rohde J.A.
      • Roopnarine O.
      • Thomas D.D.
      • Muretta J.M.
      Mavacamten stabilizes an autoinhibited state of two-headed cardiac myosin.
      ,
      • Stewart M.A.
      • Franks-Skiba K.
      • Chen S.
      • Cooke R.
      Myosin ATP turnover rate is a mechanism involved in thermogenesis in resting skeletal muscle fibers.
      ). Based on the low actin-activated ATPase activity we observed for MYH7b, we hypothesized that the ratio of SRX/DRX for MYH7b may inherently shift toward SRX compared to β-MyHC and alter the number of myosin motors available to enter the actin-bound state. We used the mant-ATP single turnover kinetics assay to measure the population of myosin motor domains in the SRX and DRX states and the associated enzymatic rates of the myosin S1 constructs in the absence of actin (Figure 5 and Table S7). We observed that human MYH7b has a substantially higher proportion of myosin heads in the SRX state as compared to human β-MyHC (82.9 ± 3.7% vs. 14.4 ± 7.3%, respectively; p < 0.0001). Intriguingly, python MYH7b has a lower proportion of myosin heads in the SRX state compared to human MYH7b (46.4 ± 13.3%, p<0.0001), but still a considerably higher proportion of SRX state myosin compared to human β-MyHC (p<0.0001). The DRX ATP turnover rates were consistent with the actin-activated ATPase rates calculated for each construct where human β-MyHC was faster than human MYH7b and python MYH7b had the slowest rate (Table S7). These assays were performed with single-headed myosin S1 and our measurements are consistent with previous measurements for human β-MyHC short S1 constructs where the slower ATPase rate is thought to arise from specific motor domain and lever arm conformations (
      • Anderson R.L.
      • Trivedi D.V.
      • Sarkar S.S.
      • Henze M.
      • Ma W.
      • Gong H.
      • Rogers C.S.
      • Gorham J.M.
      • Wong F.L.
      • Morck M.M.
      • Seidman J.G.
      • Ruppel K.M.
      • Irving T.C.
      • Cooke R.
      • Green E.M.
      • Spudich J.A.
      Deciphering the super relaxed state of human β-cardiac myosin and the mode of action of mavacamten from myosin molecules to muscle fibers.
      ,
      • Nag S.
      • Trivedi D.V.
      To lie or not to lie: Super-relaxing with myosins.
      ,
      • Rohde J.A.
      • Roopnarine O.
      • Thomas D.D.
      • Muretta J.M.
      Mavacamten stabilizes an autoinhibited state of two-headed cardiac myosin.
      ,
      • Gollapudi S.K.
      • Yu M.
      • Gan Q.-F.
      • Nag S.
      Synthetic thick filaments: A new avenue for better understanding the myosin super-relaxed state in healthy, diseased, and mavacamten-treated cardiac systems.
      ). These data suggest that both MYH7b constructs shift toward a more energy-conserving state.
      Figure thumbnail gr5
      Figure 5Human and python MYH7b have different proportions of SRX populations. A-C. Representative traces for single mant-ATP turnover experiments with human β-MyHC S1 (A), human MYH7b S1 (B), and python MYH7b S1 (C). Each normalized curve was fit to a bi-exponential decay with Y0 = 1 and plateau = 0. The top dark gray dotted line represents data simulated with a single exponential decay with the average slow rate for each construct, and the bottom light gray dotted line represents data simulated with a single exponential decay with the average fast rate for each construct. D. Quantification of the percent SRX and DRX for each construct. Data points represent technical replicates (individual curves) comprised of at least 4 separate purifications (biological replicates) and error bars represent SD. Technical replicates: hβ-MyHC n = 12, hMYH7b n = 10, and pMYH7b n = 10. **** indicates p < 0.0001.

      MYH7b is more likely to adopt closed switch-2 conformations

      We next used molecular dynamics to investigate the structural differences between human and python MYH7b and human β-MyHC that may give rise to the large shift in SRX/DRX proportions seen in the ATP turnover assay. At present there are no available experimental structures of MYH7b. Therefore, we homology modeled the MYH7b sequence into a β-MyHC S1 structural template (PDB: 5N6A) (
      • Planelles-Herrero V.J.
      • Hartman J.J.
      • Robert-Paganin J.
      • Malik F.I.
      • Houdusse A.
      Mechanistic and structural basis for activation of cardiac myosin force production by omecamtiv mecarbil.
      ) to perform simulations (Figure 6A). Though human MYH7b, python MYH7b, and human β-MyHC simulations were initiated from the same starting structure, previous molecular dynamics simulations revealed conformational heterogeneity missing in ground state structural models (
      • Porter J.R.
      • Meller A.
      • Zimmerman M.I.
      • Greenberg M.J.
      • Bowman G.R.
      Conformational distributions of isolated myosin motor domains encode their mechanochemical properties.
      ). Based on our SRX measurements, we hypothesized that human MYH7b, and to a lesser extent python MYH7b, would shift the distribution of conformations adopted in the active site towards inactive states with slow basal ATPase activity. A recently published cryo-electron microscopy structure of the Interacting Heads Motif (IHM) reveals a novel active site geometry that may be crucial for achieving the SRX biochemical state (
      • Heissler S.M.
      • Arora A.S.
      • Billington N.
      • Sellers J.R.
      • Chinthalapudi K.
      Cryo-EM structure of the autoinhibited state of myosin-2.
      ). Importantly, in the IHM structure, switch-2, an active site loop that coordinates phosphate, is shifted away from the relay helix into a putative phosphate release pathway (Figure 6C). We reasoned that MYH7b would be more likely to adopt similar active site conformations to those observed in the IHM.
      Figure thumbnail gr6
      Figure 6Molecular dynamics simulations predict that human MYH7b is more likely to adopt closed switch-2 conformations similar to those observed in the Interacting Heads Motif (IHM) active site. A. Homology model of human MYH7b S1 in the pre-powerstroke state. Red spheres indicate key active site residues whose structural fluctuations differ between MYH7b and β-MyHC. The active site ligands (ADP*Pi*Mg) are aligned and superimposed as reference. B. Zoom-in of the human MYH7b active site. The three residues (D464, E469, and I481) depicted in A are shown in red sticks. The relay helix is shown in blue while switch-2 is shown in cyan. The active site ligands (ADP*Pi*Mg) are aligned and superimposed as reference. C. Free energy landscapes of switch-2 E469’s position show that E469 is more likely to occupy positions farther away from the relay helix and transducer in MYH7b, more similar to the IHM free head structure. Black circle indicates the position of E469 in the pre-powerstroke crystal structure that was used as the starting point for simulations (PDB: 5N6A) while the black triangles indicate the position of E469 in the Interacting Heads Motif (upward-facing for the blocked head and downward-facing for the free head).
      To find differences between the structural ensembles of MYH7b and β-MyHC, we used a self-supervised autoencoder called DiffNets (
      • Ward M.D.
      • Zimmerman M.I.
      • Meller A.
      • Chung M.
      • Swamidass S.J.
      • Bowman G.R.
      Deep learning the structural determinants of protein biochemical properties by comparing structural ensembles with DiffNets.
      ). Subsequently, we measured the probabilities of these structural changes occurring between isoforms by combining hundreds of microseconds of simulation data using Markov State Models (MSMs) (
      • Pande V.S.
      • Beauchamp K.
      • Bowman G.R.
      Everything you wanted to know about Markov State Models but were afraid to ask.
      ). Briefly, DiffNets learns a low-dimensional representation of each simulation frame (i.e. a latent vector) and outputs a classification label for each frame in a self-supervised manner based on the expectation maximization algorithm. The DiffNet assigns each frame a probability from 0 to 1 that represents the likelihood of an active site conformation being unique to the MYH7b structural ensemble. By finding the correlation between structural features and this output label, we can identify structural differences that discriminate MYH7b from β-MyHC (Figure S4). To quantify these differences, we constructed MSMs of the entire S1 construct (see experimental methods) and compared the equilibrium probability-weighted distributions of these features between MSMs.
      We find that switch-2 is more likely to adopt closed positions in MYH7b than in β-MyHC (Figure 6). Crystal soaking experiments have suggested that switch-2 must open in order to allow phosphate release (
      • Llinas P.
      • Isabet T.
      • Song L.
      • Ropars V.
      • Zong B.
      • Benisty H.
      • Sirigu S.
      • Morris C.
      • Kikuti C.
      • Safer D.
      • Sweeney H.L.
      • Houdusse A.
      How Actin Initiates the Motor Activity of Myosin.
      ). Interestingly, the distances from E469 in switch-2 (human MYH7b numbering) to D464 at the phosphate binding site and I481 in the relay helix are strongly correlated with the DiffNet output label, suggesting that the position of E469 discriminates MYH7b’s structural ensemble from β-MyHC’s (Figure S4). In the human MYH7b MSM, E469 is more likely to be shifted away from the relay helix, positioning it into the putative phosphate release tunnel, likely blocking phosphate release (Figure S5). Furthermore, E469 is more likely to adopt positions where its relative position to D464 and I481 is similar to that seen in the IHM free head (Figure 6C). We also note that simulations of MYH7b were more likely to adopt structures where the purine-binding A-loop was in more extended conformations, like that observed in the IHM free head (Figure S6). Overall, these simulations suggest a structural rationale for the experimental differences in SRX/DRX proportions between MYH7b and β-MYHC.

      DISCUSSION

      For decades, the properties of the eight myosin-II isoforms found in mammalian skeletal and cardiac muscle have been a primary focus of the myosin field. Phylogenetic analysis suggests that MYH7b is an ancient sarcomeric myosin that predates the emergence of the skeletal and cardiac-specific myosin isoforms (
      • Desjardins P.R.
      • Burkman J.M.
      • Shrager J.B.
      • Allmond L.A.
      • Stedman H.H.
      Evolutionary Implications of Three Novel Members of the Human Sarcomeric Myosin Heavy Chain Gene Family.
      ). Furthermore, MYH7b orthologs have been found in fish, chicken, snakes, and frogs, suggesting that MYH7b was present in a common vertebrate ancestor (
      • Rossi A.C.
      • Mammucari C.
      • Argentini C.
      • Reggiani C.
      • Schiaffino S.
      Two novel/ancient myosins in mammalian skeletal muscles: MYH14/7b and MYH15 are expressed in extraocular muscles and muscle spindles: MYH14/7b and MYH15 expression in mammalian skeletal muscles.
      ,
      • Ikeda D.
      • Ono Y.
      • Snell P.
      • Edwards Y.J.K.
      • Elgar G.
      • Watabe S.
      Divergent evolution of the myosin heavy chain gene family in fish and tetrapods: evidence from comparative genomic analysis.
      ,
      • McGuigan K.
      Evolution of Sarcomeric Myosin Heavy Chain Genes: Evidence from Fish.
      ). Thus, understanding the molecular properties of MYH7b is important and can inform an evolutionary perspective of the sarcomeric myosin family. This myosin gene escaped detection for years due to its unique expression pattern and regulation in mammals that restricts MYH7b protein production to certain striated muscles with specialized functions (
      • Bell M.L.
      • Buvoli M.
      • Leinwand L.A.
      Uncoupling of Expression of an Intronic MicroRNA and Its Myosin Host Gene by Exon Skipping.
      ,
      • Rossi A.C.
      • Mammucari C.
      • Argentini C.
      • Reggiani C.
      • Schiaffino S.
      Two novel/ancient myosins in mammalian skeletal muscles: MYH14/7b and MYH15 are expressed in extraocular muscles and muscle spindles: MYH14/7b and MYH15 expression in mammalian skeletal muscles.
      ). The surprising discovery of low abundance MYH7b protein in non-muscle tissue has also sparked recent interest in this myosin’s function (
      • Rubio M.D.
      • Johnson R.
      • Miller C.A.
      • Huganir R.L.
      • Rumbaugh G.
      Regulation of Synapse Structure and Function by Distinct Myosin II Motors.
      ,
      • Carlyle B.C.
      • Kitchen R.R.
      • Kanyo J.E.
      • Voss E.Z.
      • Pletikos M.
      • Sousa A.M.M.
      • Lam T.T.
      • Gerstein M.B.
      • Sestan N.
      • Nairn A.C.
      A multiregional proteomic survey of the postnatal human brain.
      ,
      • Guo Z.
      • Shao C.
      • Zhang Y.
      • Qiu W.
      • Li W.
      • Zhu W.
      • Yang Q.
      • Huang Y.
      • Pan L.
      • Dong Y.
      • Sun H.
      • Xiao X.
      • Sun W.
      • Ma C.
      • Zhang L.
      A Global Multiregional Proteomic Map of the Human Cerebral Cortex.
      ,
      • Xu J.
      • Patassini S.
      • Rustogi N.
      • Riba-Garcia I.
      • Hale B.D.
      • Phillips A.M.
      • Waldvogel H.
      • Haines R.
      • Bradbury P.
      • Stevens A.
      • Faull R.L.M.
      • Dowsey A.W.
      • Cooper G.J.S.
      • Unwin R.D.
      Regional protein expression in human Alzheimer’s brain correlates with disease severity.
      ). In contrast to mammals, MYH7b appears to have a role in cardiac and skeletal muscle of birds (
      • Rossi A.C.
      • Mammucari C.
      • Argentini C.
      • Reggiani C.
      • Schiaffino S.
      Two novel/ancient myosins in mammalian skeletal muscles: MYH14/7b and MYH15 are expressed in extraocular muscles and muscle spindles: MYH14/7b and MYH15 expression in mammalian skeletal muscles.
      ,
      • Machida S.
      • Noda S.
      • Takao A.
      • Nakazawa M.
      • Matsuoka R.
      Expression of slow skeletal myosin heavy chain 2 gene in Purkinje fiber cells in chick heart.
      ), and as described in this study, pythons. Here, we present the first detailed characterization of the ancient sarcomeric myosin MYH7b’s motor properties. We found that although MYH7b has overall slower kinetics and working stroke mechanics than other sarcomeric myosins, MYH7b’s enzymatic properties are consistent with a role in force generation. Furthermore, we identified differential tuning between the motor activities of human and python MYH7b, suggesting evolutionary adaptations based on species-specific physiological demands.
      Our results show that several key motor properties of human and python MYH7b are slower than the canonical slow human β-MyHC isoform. Both human and python MYH7b have a reduced actin-activated ATPase activity compared to β-MyHC that is due, in part, to slower kinetics of steps in the mechanochemical cycle. Furthermore, we observe that the sliding velocities of both human and python MYH7b are slower than β-MyHC. In the motility assay, the speed is proportional to the displacement of the working stroke (d) divided by the attachment duration at saturating ATP (ton). The optical trapping experiments show that the step sizes of MYH7b and β-MyHC are similar. This is consistent with the fact that both β-MyHC and MYH7b have 2 bound light chains, and thus have similar length lever arms. Furthermore, both MYH7b and β-MyHC have two-substep working strokes, and thus similar mechanics. Therefore, the differences in the sliding speeds are due to changes in the actomyosin detachment rate, which is limited by the rate of ADP release. Overall, our results suggest that evolutionary pressures have preserved MYH7b activity for slower contractile functions than β-MyHC.
      Additionally, we observed striking changes in the SRX/DRX ratio of the MYH7b isoforms compared to human β-MyHC. Human MYH7b showed a distinct stabilization of the SRX state, an almost complete reversal of the ratio of SRX/DRX in human β-MyHC. These assays were performed using myosin S1, and our measurements for β-MyHC are consistent with values previously reported for SRX proportions using β-MyHC short S1 and S1 fragments (
      • Anderson R.L.
      • Trivedi D.V.
      • Sarkar S.S.
      • Henze M.
      • Ma W.
      • Gong H.
      • Rogers C.S.
      • Gorham J.M.
      • Wong F.L.
      • Morck M.M.
      • Seidman J.G.
      • Ruppel K.M.
      • Irving T.C.
      • Cooke R.
      • Green E.M.
      • Spudich J.A.
      Deciphering the super relaxed state of human β-cardiac myosin and the mode of action of mavacamten from myosin molecules to muscle fibers.
      ,
      • Rohde J.A.
      • Roopnarine O.
      • Thomas D.D.
      • Muretta J.M.
      Mavacamten stabilizes an autoinhibited state of two-headed cardiac myosin.
      ,
      • Gollapudi S.K.
      • Yu M.
      • Gan Q.-F.
      • Nag S.
      Synthetic thick filaments: A new avenue for better understanding the myosin super-relaxed state in healthy, diseased, and mavacamten-treated cardiac systems.
      ). Our experimental results are supported by molecular dynamics simulations, which predict that MYH7b adopts motor domain conformations more similar to those observed in the interacting heads motif (IHM). Based on these results, we hypothesize that MYH7b may be more likely to stabilize the auto-inhibited IHM state. Future work using two-headed constructs with a portion of the myosin rod will be necessary to experimentally confirm this prediction. Stabilization of this ultra-slow ATP turnover state could be an evolutionary advantage and energy-saving mechanism as it is known that skeletal muscle myosin can be recruited out of the IHM in situations of increased demand (
      • Hooijman P.
      • Stewart M.A.
      • Cooke R.
      A New State of Cardiac Myosin with Very Slow ATP Turnover: A Potential Cardioprotective Mechanism in the Heart.
      ,
      • McNamara J.W.
      • Li A.
      • dos Remedios C.G.
      • Cooke R.
      The role of super-relaxed myosin in skeletal and cardiac muscle.
      ).
      Given the absence of MYH7b from most mammalian skeletal muscles and cardiac muscle, it appears that these tissues have evolved to use myosin isoforms with faster motor properties. The slow kinetics and high proportion of MYH7b in the SRX state are likely key contributors to the cardiac dilation and dysfunction observed in a transgenic mouse model with forced cardiac expression of MYH7b protein (
      • Peter A.K.
      • Rossi A.C.
      • Buvoli M.
      • Ozeroff C.D.
      • Crocini C.
      • Perry A.R.
      • Buvoli A.E.
      • Lee L.A.
      • Leinwand L.A.
      Expression of Normally Repressed Myosin Heavy Chain 7b in the Mammalian Heart Induces Dilated Cardiomyopathy.
      ). Notably, these results are consistent with a recent study reporting a dilated cardiomyopathy-causing myosin mutation that slows kinetics and stabilizes the SRX state (

      Rasicci, D. V., Tiwari, P., Desetty, R., Sadler, F. W., Sivaramakrishnan, S., Craig, R., and Yengo, C. M. (2022) Dilated cardiomyopathy mutation E525K in human beta-cardiac myosin stabilizes the interacting heads motif and super-relaxed state of myosin. 10.1101/2022.02.18.480995

      ). However, specialized mammalian muscles appear to tolerate inclusion of MYH7b; these muscles typically express a broad array of sarcomeric myosin isoforms with different motor properties allowing for functional adaptability (
      • Lee L.A.
      • Karabina A.
      • Broadwell L.J.
      • Leinwand L.A.
      The ancient sarcomeric myosins found in specialized muscles.
      ). For example, the extraocular muscle performs both ultra-fast movements and slower pursuit and vergence motions (
      • Porter J.D.
      • Baker R.S.
      • Ragusa R.J.
      • Brueckner J.K.
      Extraocular muscles: Basic and clinical aspects of structure and function.
      ), which require a repertoire of functionally diverse myosins. The specialized myosin isoform MyHC-extraocular encoded by MYH13 has the fastest contractile properties of the sarcomeric myosin family and is found exclusively in EOM and laryngeal muscles (
      • Bloemink M.J.
      • Deacon J.C.
      • Resnicow D.I.
      • Leinwand L.A.
      • Geeves M.A.
      The Superfast Human Extraocular Myosin Is Kinetically Distinct from the Fast Skeletal IIa, IIb, and IId Isoforms.
      ,
      • Schachat F.
      • Briggs M.M.
      Phylogenetic implications of the superfast myosin in extraocular muscles.
      ). Likewise, it is possible that MYH7b performs a specialized role in slow contractility in these heterogeneous tissues that show a diverse range of contractility. Finally, MYH7b could co-polymerize with other myosins to modulate contractile filaments, specifically in the context of non-muscle tissues, as has been reported for a distinct class-18 myosin and non-muscle myosin II (
      • Billington N.
      • Beach J.R.
      • Heissler S.M.
      • Remmert K.
      • Guzik-Lendrum S.
      • Nagy A.
      • Takagi Y.
      • Shao L.
      • Li D.
      • Yang Y.
      • Zhang Y.
      • Barzik M.
      • Betzig E.
      • Hammer J.A.
      • Sellers J.R.
      Myosin 18A Coassembles with Nonmuscle Myosin 2 to Form Mixed Bipolar Filaments.
      ). Future studies testing the ability to co-assemble with non-sarcomeric myosins will be essential for understanding MYH7b’s role in non-muscle cells and tissues.
      In contrast to mammals, which express two cardiac myosin isoforms, pythons express a broad array of myosin isoforms in their hearts, and even more in their skeletal muscles. We observed that python MYH7b has slower steady-state actin-activated ATPase activity and slower kinetic rates of key individual steps within the mechanochemical cycle compared to human MYH7b. Intriguingly, despite the lower maximal actin-activated ATPase rate, python MYH7b demonstrated an increase in DRX myosin population compared to human MYH7b. These results suggest that although python MYH7b has overall slower enzyme kinetics, it has been adapted to modulate SRX/DRX such that an increased population of myosin heads are readily available for interactions with actin. The higher proportion of DRX myosin for python MYH7b compared to human MYH7b is potentially the result of evolutionary pressures on body wall skeletal muscles and the heart, which require a constant population of myosin heads in the DRX state to support normal function. This idea is consistent with the requirement for these high-demand muscles to quickly activate in response to environmental cues like predation or threats. Additionally, pythons are infrequent feeders that shift between resting, energy-reserving states and periods of high metabolic demand and post-prandial physiological adaptation (
      • Andersen J.B.
      • Rourke B.C.
      • Caiozzo V.J.
      • Bennett A.F.
      • Hicks J.W.
      Postprandial cardiac hypertrophy in pythons.
      ,
      • Riquelme C.A.
      • Magida J.A.
      • Harrison B.C.
      • Wall C.E.
      • Marr T.G.
      • Secor S.M.
      • Leinwand L.A.
      Fatty Acids Identified in the Burmese Python Promote Beneficial Cardiac Growth.
      ,
      • Secor S.M.
      • Diamond J.
      A vertebrate model of extreme physiological regulation.
      ). The slower motor properties of MYH7b and overall significant proportion of myosin in the SRX state could be an evolutionary adaptation to minimize ATP usage during prolonged periods of rest.
      In this study, we present the first comprehensive functional characterization of the ancient sarcomeric myosin MYH7b. We have defined MYH7b’s activity as even slower than that of β-MyHC, which was previously the slowest human myosin isoform characterized (
      • Deacon J.C.
      • Bloemink M.J.
      • Rezavandi H.
      • Geeves M.A.
      • Leinwand L.A.
      Identification of functional differences between recombinant human α and β cardiac myosin motors.
      ,
      • Sommese R.F.
      • Sung J.
      • Nag S.
      • Sutton S.
      • Deacon J.C.
      • Choe E.
      • Leinwand L.A.
      • Ruppel K.
      • Spudich J.A.
      Molecular consequences of the R453C hypertrophic cardiomyopathy mutation on human -cardiac myosin motor function.
      ). Our results suggest that MYH7b has been adapted for specialized roles in mammals who do not tolerate the slow properties of MYH7b in conventional skeletal and cardiac muscle. In contrast, the slow properties of MYH7b are likely advantageous to different species like pythons that have different physiological needs and demands on cardiac and skeletal muscle. Questions remain as to whether MYH7b carries out an essential role in mammalian specialized muscle or whether this myosin is gradually being silenced as is the case in mammalian heart and skeletal muscle. Furthermore, MYH7b’s specific role and mechanism of action in mammalian non-muscle environments remains unresolved. Ultimately, this study provides a perspective of how highly conserved myosin motors can be differentially tuned for specialized activities across species and tissues.

      EXPERIMENTAL PROCEDURES

      RNA isolation, cDNA synthesis, and qPCR

      Total RNA was purified from Burmese and Ball python ventricle and skeletal muscle by homogenization in Tri Reagent (Molecular Research Center, TR118). Chloroform was added and incubated at room temperature for 15 minutes followed by a centrifugation spin at 12,000 x g for 15 minutes at 4 °C. The aqueous layer was removed and RNA was precipitated with isopropanol and washed with 75% ethanol. RNA pellets were dissolved in HPLC grade water.
      RNA was transcribed to cDNA using the SuperScript III Reverse Transcription kit (ThermoFisher/Invitrogen, 18080044) and random hexamer primers. Real time quantitative polymerase chain reaction (qPCR) was performed on a CFX96-Real-Time PCR Detection System (BioRad) using SYBR Green PCR Master Mix (ThermoFisher/Applied Biosystems, 4312704) and gene specific primer sets (Table S3). Relative RNA expression was measured using the Pfaffl standard curve method and all genes were normalized to 18S expression. Primers were designed against the Burmese python genome (accessed on NCBI) and many primer sets were unable to successfully amplify from RNA produced using Ball python tissues, presumably due to species sequence differences.

      Exon Skipping PCR

      Amplification of the alternative splicing products were carried out on Burmese and Ball python cardiac and skeletal muscle cDNA by PCR using a forward primer specific to Burmese python exon 5, 5’- TAAGGGTAAGCGGAGGTCT and a reverse primer specific to exon 9, 5’- TTCCATGGCAGGGTTAGC. These primers amplify a 271 base pair product corresponding to the unskipped transcript and a 174 base pair product corresponding to the exon-7 skipped transcript. Primer sequences were checked against unpublished Ball python RNA sequences and designed to match both species.

      Protein lysate preparation and Mass spectrometry

      Protein lysates were prepared by homogenizing tissue in ice cold RIPA buffer (50 mM Tris pH 8, 1 mM EDTA, 0.5 mM EGTA, 0.5% sodium deoxycholate, 1% Triton X-100, 0.1% SDS, 140 mM NaCl with protease inhibitor cocktail (Millipore Sigma/Roche, 11873580001). 2 samples (animals) per tissue type from Burmese pythons and 1 sample per tissue type for Ball python (due to availability of tissue) were prepared. 20 μg total protein was run on a NuPAGE 4-12% Bis-Tris mini protein gel (ThermoFisher/Invitrogen, NP0323BOX). The gel was stained with Coomassie Blue (Research Products International, B43000) and destained in 50% methanol, 10% acetic acid, 40% water followed by 10% methanol, 10% acetic acid, 80% water. Bands at molecular weight marker 250 kDa (myosin heavy chain) and between ∼15 and ∼20 kDa (myosin light chains) were excised and sent to the University of Colorado Anschutz Mass Spectrometry Proteomics Shared Resource Facility. Excised protein bands were subjected to tandem mass spectrometry using an Orbitrap Fusion Lumos with Easy nLC 1200 UPLC system (ThermoFisher).
      All MS/MS samples were analyzed using Mascot (Matrix Science, London, UK; version 2.7.0). Mascot was set to search custom python-specific myosin databases (Burmese python myosin heavy chains and myosin light chains, accession codes are included in Tables S1 and S2) with assumed trypsin digestion. Scaffold (Proteome Software Inc, version Scaffold_5.1.0) was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95% probability by the Peptide Prophet algorithm (
      • Keller A.
      • Nesvizhskii A.I.
      • Kolker E.
      • Aebersold R.
      Empirical Statistical Model To Estimate the Accuracy of Peptide Identifications Made by MS/MS and Database Search.
      ) with Scaffold delta-mass correction. Protein identifications were accepted if they could be established at greater than 99% probability and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm (
      • Nesvizhskii A.I.
      • Keller A.
      • Kolker E.
      • Aebersold R.
      A Statistical Model for Identifying Proteins by Tandem Mass Spectrometry.
      ). Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Spectrum count normalization is calculated by the Scaffold software by multiplying each spectrum count in each sample by the average count over the biosample’s total spectrum count. A full mass spectrometry sample report including values for percent coverage and number of unique peptides is included in the supporting information (Tables S1 and S2).

      Recombinant myosin expression and purification

      Construct design

      Recombinant human and python myosin was generated by cloning myosin genes into an adenovirus using the pAdEasy Vector system (Qbiogene). All experiments used the myosin heavy chain motor domain known as subfragment 1 (S1), which is sufficient to produce the catalytic activity that drives actin-based contractility (

      Toyoshima, Y. Y., Kron, S. J., McNally, E. M., Niebling, K. R., Toyoshima, C., and Spudich, J. A. (1987) Myosin subfragment-l is sufficient to move actin filaments in vitro

      ). Myosin S1 constructs correspond to amino acids 1-842 for human β-MyHC S1, 1-850 for human MYH7b S1, 1-850 for python MYH7b S1 followed by a flexible Gly-Ser-Gly or Gly-Ser-Ser linker and a C-terminal 8 amino acid PDZ binding peptide (Arg-Gly-Ser-Ile-Asp-Thr-Trp-Val). All myosin S1 constructs were expressed in C2C12 myotubes and were purified using a C-terminal tag that binds PDZ domains. The constructs used in these experiments are bound by endogenous mouse C2C12 light chains (Figure S2).
      The human β-MyHC S1 sequence was obtained by amplifying the S1 region from human heart MYH7 cDNA as previously described (
      • Deacon J.C.
      • Bloemink M.J.
      • Rezavandi H.
      • Geeves M.A.
      • Leinwand L.A.
      Identification of functional differences between recombinant human α and β cardiac myosin motors.
      ). The human MYH7b S1 sequence was obtained by amplifying the S1 region from the human cDNA clone pF1KA1512 (Kazusa DNA Research Institute, product ID FXC06072). The python MYH7b S1 sequence was obtained by amplifying the gene from Burmese skeletal muscle cDNA using restriction enzyme carrying primers designed for the 5’ and 3’ regions. All myosin S1 inserts were sub-cloned into a pUC19 vector containing the PDZ binding C-tag. Myosin S1 C-tag inserts were sub-cloned into the pShuttle-CMV plasmid (Addgene, plasmid #16403), which was transformed into dh5ɑ competent cells. Colonies were miniprepped and subject to analytical digest to screen for positive clones. The pShuttle-CMV plasmid containing myosin S1 C-tag was transformed into BJ5183 competent cells (RecA+, pAdEasy) and miniprepped. Miniprepped DNA was transformed into dh5ɑ competent cells and midi-prepped to produce pure plasmid sufficient for transfection.
      Original pShuttle-CMV constructs were later updated to include the translation enhancing WPRE sequence cloned using a synthetic construct (Genewiz) directly after the C-tag and stop codon using the Gibson HiFi DNA Assembly Cloning Kit (NEB, E5520S). Adenovirus was produced and used to infect C2C12 cells.

      Virus production

      Replication deficient recombinant adenoviruses containing myosin S1 were produced by transfecting HEK293 cells and amplified through HEK293 cells to produce adequate quantities of virus. Infected HEK293 cells were collected and lysed by 3 freeze/thaw steps. Cell debris was pelleted by spinning at 15,000 x g 30 minutes and supernatant was overlayed onto a CsCl step gradient (1.5 mL of 1.25 g/mL and 1 mL of 1.4 g/mL). Viral particles were separated by spinning at 160,000 x g for 1 hour in a Beckman SW 41 Ti rotor. Virus particles were collected, pooled, and purified further using a second gradient of 1.35 g/mL CsCl. Virus was collected and stored at -20 °C in a glycerol containing buffer.

      Protein expression

      C2C12 mouse skeletal myoblasts were cultured in Dulbecco's Modified Eagle Medium (DMEM) containing 10% Fetal Bovine Serum (FBS), 1% L-Glutamine, 1% Penicillin-Streptomycin and differentiated into myotubes using DMEM containing 2% Horse Serum, 1% L-Glutamine, 1% Penicillin-Streptomycin. Myotubes were then infected with myosin S1 C-tag adenovirus 3 days after differentiation. Cells were collected 4 days post infection using trypsin, centrifuged 30 minutes at 2200 x g at 4 °C, scraped into liquid nitrogen, and stored at -80 °C.

      Protein purification

      Cell pellets were lysed in 50 mM Tris pH 8.0, 200 mM NaCl, 4 mM MgCl2, 0.5% Tween-20, 5 mM DTT, 1 mM ATP, 0.2 mM PMSF, and 1X protease inhibitor cocktail (Millipore Sigma/Roche, 11873580001) and mechanically dounced preceding a 25 minute centrifugation at 39,000 x g. The supernatant was collected and filtered through 5 μM and 1.2 μM filters. The filtered supernatant was applied to a column containing SulfoLink resin (ThermoFisher, 20402) coupled to PDZ. The flow through was collected and the column washed with 30 mM Tris pH 7.5, 50 mM KCl, 5 mM MgCl2, 1 mM DTT, and 1 mM ATP. Myosin S1 was eluted using a peptide with higher PDZ specificity (Genescript, Trp-Glu-Thr-Trp-Val). Recombinant myosin-S1 was dialyzed against storage buffer containing 20 mM MOPS pH 7.0, 25 mM KCl, 5 mM MgCl2 and 10% sucrose. Proteins were flash frozen after the addition of 1 mM DTT and 1 mM ATP and stored at -80 °C. The resulting purified protein consists of recombinant myosin-S1 with endogenous C2C12 light chains.

      Actin-activated ATPase assay

      Actin-activated ATPase rates were measured using the NADH-coupled ATPase assay. Myosin was thawed on ice and diluted to 0.8 μM in 20 mM MOPS pH 7.0, 25 mM KCl, and 5 mM MgCl2 with 5 mM DTT. Rabbit skeletal actin was purified as previously described ((

      Pardee, J. D., and Spudich, J. (1982) [18] Purification of muscle actin. in Methods in Enzymology, pp. 164–181, Structural and Contractile Proteins Part B: The Contractile Apparatus and the Cytoskeleton, Academic Press, 85, 164–181

      ) and diluted accordingly to reach concentrations ranging from 10 μM to 100 μM and was mixed with 4 μM Gelsolin (bacterially expressed and prepared from E.coli (
      • Sommese R.F.
      • Sung J.
      • Nag S.
      • Sutton S.
      • Deacon J.C.
      • Choe E.
      • Leinwand L.A.
      • Ruppel K.
      • Spudich J.A.
      Molecular consequences of the R453C hypertrophic cardiomyopathy mutation on human -cardiac myosin motor function.
      )) and incubated for at least 30 minutes. Myosin was added to actin/gelsolin in a clear 384-well plate to achieve a final myosin concentration of 0.4 μM. One well with Myosin alone (0.4 μM) plus buffer was measured to determine the basal ATPase rate in the absence of actin. Right before measuring, 10X buffer consisting of 20 mM ATP, 30 mM phospho(enol)pyruvate (Sigma, 860077), 10 mM NADH (Sigma, N8129), and 8 mM pyruvate kinase/lactate dehydrogenase (Sigma, P0294) was added to the each well to achieve a final 1X concentration. Experiments were performed at 25 °C using a SpectraMax iD3 Multi-Mode Microplate Reader (Molecular Devices) absorbance was monitored at 340 nm every 30 seconds for 1 hour. A standard curve of NADH from 1 mM to 0.008 mM was run on each plate to determine the conversion factor of nmol NADH/absorbance value.
      Individual rates for each myosin and actin well and myosin alone wells were calculated by taking the linear range of the absorbance vs. time and converting to nmol ATP/second using the NADH conversion factor. Rates were divided by the myosin concentration in order to obtain the per second rate and plotted against actin concentration. The basal (actin-free) myosin ATPase activity was subtracted from each actin rate and technical replicates comprised of one or two curves per construct on a given day were averaged. The data were fit to a Michaelis-Menten kinetics equation using the nonlinear fit feature in Graphpad Prism, which calculates the maximum actin-activated ATPase activity rate (kcat) and Michaelis-Menten constant (KM). Ambiguous fits were discarded across all data sets. Statistical significance was determined using a one-way ANOVA with Tukey’s multiple comparison test.

      In vitro Motility assay

      Labeling actin

      1 μM F-actin (prepared as described for the actin-activated ATPase assay) diluted in 1X assay buffer (20 mM MOPS pH 7.0, 25 mM KCl, 5 MgCl2) was incubated with 2 μM Rhodamine-phalloidin (ThermoFisher, R415) and incubated 1 hour at room temp. Labeled actin was stored at 4 °C protected from light.

      Deadheading myosin

      A “deadheading” spin down was performed prior to flow cell loading to eliminate inactive myosin bound to actin. A 3:1 molar ratio of actin:myosin in 1X assay buffer was incubated on ice for 5 minutes. 1 mM ATP was added to myosin and actin and mixture was centrifuged at 90,000 rpm in a TLA-100 rotor (Beckman) 25 minutes at 4 °C. The supernatant was collected and diluted in 1X assay buffer to a final concentration of 0.4 μM.

      Motility Experiments

      Flow cells were constructed using a microscope slide, double sided tape, and coverslips coated in 0.2% nitrocellulose (LADD Research Industries, #53152) diluted in amyl acetate. SNAP-PDZ (plasmid provided by the Spudich Lab (
      • Aksel T.
      • Choe Yu E.
      • Sutton S.
      • Ruppel K.M.
      • Spudich J.A.
      Ensemble force changes that result from human cardiac myosin mutations and a small-molecule effector.
      ), purified from E. coli) at 3 μM diluted in 1X assay buffer (20 mM MOPS pH 7.0, 25 mM KCl, 5 mM MgCl2, 10 mM DTT) was flowed into each chamber of the flow cell and incubated at room temperature for 2 minutes. 1 mg/mL BSA diluted in 1X assay buffer was flowed through each chamber followed by myosin S1 diluted to 0.4 μM (∼54 μg/mL) in 1X assay buffer and incubated at room temperature for 3 minutes. Each chamber was blocked again with 1 mg/mL BSA followed by labeled actin and incubated at room temperature for 1 minute. Finally, motility buffer consisting of 1X assay buffer, 3 mM ATP, 1 mg/mL BSA, 1 mM EGTA, and 0.5% methylcellulose and an oxygen scavenging solution of 4 mg/mL glucose, 0.135 mg/mL glucose oxidase (Sigma, G2133), 0.0215 mg/mL catalase (Sigma, C30). Motility videos were obtained at 25 °C with a frame rate of 1 frame per second for 30 seconds using a 100X oil objective on a Nikon Ti-E Eclipse Inverted Fluorescence Microscope equipped with a Hamamatsu ORCA-Flash 4.0 V3 Digital CMOS Camera.

      Analysis of in vitro Motility data

      In vitro motility videos were analyzed using MATLAB code developed in-house. Briefly, the code identifies fibers between 2 and 6 pixels in width using a Hessian transform (

      Frangi, A. F., Niessen, W. J., Vincken, K. L., and Viergever, M. A. (1998) Multiscale vessel enhancement filtering. in Medical Image Computing and Computer-Assisted Intervention - MICCAI’98 (Wells, W. M., Colchester, A., and Delp, S. eds), pp. 130–137, Lecture Notes in Computer Science, Springer Berlin Heidelberg, Berlin, Heidelberg, 1496, 130–137

      ). A threshold value was used to generate a binary mask of the identified fibers in the image. The resulting mask was then skeletonized and branches in the skeleton were removed to convert the mask of each fiber into a line. The midpoint coordinate of the line was then identified and used to track the position of the migrating fibers. Each individual fiber was then tracked using a previously developed custom toolbox which implemented the linear assignment framework (
      • Jaqaman K.
      • Loerke D.
      • Mettlen M.
      • Kuwata H.
      • Grinstein S.
      • Schmid S.L.
      • Danuser G.
      Robust single particle tracking in live cell time-lapse sequences.
      ,

      Tay, J. W., and Cameron, J. C. (2020) CyAn: A MATLAB toolbox for image and data analysis of cyanobacteria. 10.1101/2020.07.28.225219

      ). The code for this analysis can be accessed at https://github.com/Biofrontiers-ALMC/actin-tracking-toolbox. The mean velocity across technical replicates was determined from the tracked positions, and statistical significance was determined using a one-way ANOVA with Tukey’s multiple comparison test.

      Stopped-flow kinetic experiments

      All stopped-flow kinetic experiments were conducted in 25 mM KCl, 20 mM MOPS, 5 mM MgCl2, and 1 mM DTT, pH 7.0, at 20 °C. A High-Tech Scientific SF-61 DX2 stopped-flow system was used to perform the measurements, with concentrations stated as those after mixing in the observation cell. All stopped-flow transients were either analyzed in software provided by High-Tech Scientific (Kinetic Studios) or GraphPad Prism. Concentrations are those after mixing unless stated otherwise.
      Rabbit skeletal actin was purified as previously described (

      Pardee, J. D., and Spudich, J. (1982) [18] Purification of muscle actin. in Methods in Enzymology, pp. 164–181, Structural and Contractile Proteins Part B: The Contractile Apparatus and the Cytoskeleton, Academic Press, 85, 164–181

      ) and labelled with pyrene as previously described (
      • Criddle A.H.
      • Geeves M.A.
      • Jeffries T.
      The use of actin labelled with N -(1-pyrenyl)iodoacetamide to study the interaction of actin with myosin subfragments and troponin/tropomyosin.
      ). Pyrene-labelled actin was excited at 365 nm using a Hg-Xe lamp, and emission of pyrene-labelled actin was detected after being passed through a KV399 cut-off filter. The binding of myosin S1 to actin quenches the fluorescence signal with dissociation of myosin S1 from actin leading to an increase in fluorescence. Using the scheme below (Scheme 1) it was possible to observe the interactions between actomyosin S1 and ATP or ADP.
      The binding of ATP to actomyosin (K1) is reversible followed by a rate-limiting isomerization of the complex leading the rapid dissociation of the myosin-ATP complex from actin (k+2). ADP and ATP both compete to bind to the nucleotide binding pocket of myosin S1 with ADP binding controlled by the dissociation constant KADP (=k+ADP/k-ADP).
      The dissociation reaction can be measured using pyrene-labelled actin in complex with S1. The S1 quenches the pyrene label on the actin. When ATP is rapidly mixed with the actomyosin complex, it leads to an increase in fluorescence. Using Scheme 1 and Equation 1 the constants K1k+2, k+2, and 1/K1 can be determined.
      kobs=K1k+2[ATP]1+K1[ATP]
      Equation 1


      Where K1k+2 is the second order rate constant for ATP binding to actomyosin S1, k+2 is the maximum rate of ATP induced dissociation, and 1/K1 is the ATP affinity for myosin S1.
      When ATP and ADP are in rapid competition for binding to the actomyosin complex, Equation 2 can be used to determine the ADP affinity (KADP).
      kobs=11+[ADP]KADP
      Equation 2


      The data were graphed in Graphpad Prism. Statistical significance was determined using an unpaired t-test.

      Optical Trapping

      Protein expression and purification

      Cardiac actin was purified from cryoground porcine ventricles as previously described (
      • Clippinger S.R.
      • Cloonan P.E.
      • Greenberg L.
      • Ernst M.
      • Stump W.T.
      • Greenberg M.J.
      Disrupted mechanobiology links the molecular and cellular phenotypes in familial dilated cardiomyopathy.
      ). SNAP-PDZ was expressed and purified as described for the motility experiments. Protein concentrations were determined spectroscopically.

      Optical trapping experiments

      Experiments were performed on a custom-built, microscope free dual-beam optical trap, described previously (
      • Blackwell T.
      • Stump W.T.
      • Clippinger S.R.
      • Greenberg M.J.
      Computational Tool for Ensemble Averaging of Single-Molecule Data.
      ). These experiments utilized the three-bead geometry, in which an actin filament is held between two optically trapped beads and brought close to a surface-bound bead that is sparsely coated with myosin (
      • Finer J.T.
      • Simmons R.M.
      • Spudich J.A.
      Single myosin molecule mechanics: piconewton forces and nanometre steps.
      ,

      Greenberg, M. J., Shuman, H., and Ostap, E. M. (2017) Measuring the Kinetic and Mechanical Properties of Non-processive Myosins Using Optical Tweezers. in Optical Tweezers: Methods and Protocols (Gennerich, A. ed), pp. 483–509, Methods in Molecular Biology, Springer, New York, NY, 10.1007/978-1-4939-6421-5_19

      ). All solutions were prepared in KMg25 buffer (60 mM MOPS pH 7.0, 25 mM KCl, 2 mM EGTA, 4 mM MgCl2, 1 mM DTT) unless otherwise specified. Before each experiment, myosin dead heads were removed by ultracentrifugation in the presence of 1 mM ATP and 2.17 μM phalloidin-stabilized actin (436,000 x g 30 min at 4 °C in an Optima MAX-TL ultracentrifuge equipped with a TLA 120.2 rotor (Beckman Coulter)). Rhodamine-phalloidin-stabilized actin filaments (2 μM) were prepared from cardiac actin and 15% biotinylated actin and rhodamine-phalloidin (3.75 μM). Streptavidin beads were blocked in 1 mg/mL BSA by three sequential cycles of suspension in 1 mg/mL BSA followed by centrifugation at 9,391 x g 3 min in a 5424R centrifuge (Eppendorf).
      Flow cells were coated sparsely with silica beads suspended in nitrocellulose in amyl acetate, as previously described (
      • Blackwell T.
      • Stump W.T.
      • Clippinger S.R.
      • Greenberg M.J.
      Computational Tool for Ensemble Averaging of Single-Molecule Data.
      ,

      Greenberg, M. J., Shuman, H., and Ostap, E. M. (2017) Measuring the Kinetic and Mechanical Properties of Non-processive Myosins Using Optical Tweezers. in Optical Tweezers: Methods and Protocols (Gennerich, A. ed), pp. 483–509, Methods in Molecular Biology, Springer, New York, NY, 10.1007/978-1-4939-6421-5_19

      ). Each flow cell was loaded with 20 nM PDZ 5 min, blocked with 1 mg/mL BSA 5 min, and loaded with C-tagged myosin S1 (20-60 nM) 5 min. The surface was blocked by washing with 1 mg/mL BSA. The flow cell was then loaded with activation buffer (KMg25 with 1 mg/mL BSA, 1 μM ATP, 192 U/mL glucose oxidase, 48 μg/mL catalase, 1 mg/mL glucose, and ∼25 pM rhodamine-phalloidin-stabilized actin). This was followed by 4 μL of streptavidin beads suspended in 1 mg/mL BSA. The flow cell was then sealed with vacuum grease, and data were collected within 60 minutes of sealing.
      Rhodamine-phalloidin-stabilized actin filaments were attached to polystyrene beads using a biotin-streptavidin linkage, where actin contained 15% biotinylated actin and beads were coated with streptavidin, as previously described (
      • Blackwell T.
      • Stump W.T.
      • Clippinger S.R.
      • Greenberg M.J.
      Computational Tool for Ensemble Averaging of Single-Molecule Data.
      ,

      Greenberg, M. J., Shuman, H., and Ostap, E. M. (2017) Measuring the Kinetic and Mechanical Properties of Non-processive Myosins Using Optical Tweezers. in Optical Tweezers: Methods and Protocols (Gennerich, A. ed), pp. 483–509, Methods in Molecular Biology, Springer, New York, NY, 10.1007/978-1-4939-6421-5_19

      ). For each bead-actin-bead assembly, the trap stiffness was calculated from fitting of the power spectrum as previously described (

      Greenberg, M. J., Shuman, H., and Ostap, E. M. (2017) Measuring the Kinetic and Mechanical Properties of Non-processive Myosins Using Optical Tweezers. in Optical Tweezers: Methods and Protocols (Gennerich, A. ed), pp. 483–509, Methods in Molecular Biology, Springer, New York, NY, 10.1007/978-1-4939-6421-5_19

      ). Data were collected at 20 kHz and filtered to 10 kHz according to the Nyquist criterion.

      Analysis of single-molecule data

      All data from optical trapping experiments was analyzed using a custom-built MATLAB (MathWorks) program, SPASM (
      • Blackwell T.
      • Stump W.T.
      • Clippinger S.R.
      • Greenberg M.J.
      Computational Tool for Ensemble Averaging of Single-Molecule Data.
      ). Step size data are reported as mean ± standard error. Statistical testing of the step sizes was done using a 2-tailed Student’s t-test of individual binding interactions. Event durations were fit by single exponential functions using the MATLAB-based program MEMLET (
      • Woody M.S.
      • Lewis J.H.
      • Greenberg M.J.
      • Goldman Y.E.
      • Ostap E.M.
      MEMLET: An Easy-to-Use Tool for Data Fitting and Model Comparison Using Maximum-Likelihood Estimation.
      ) to determine the best-fit value for the detachment rate and the associated 95% confidence interval, which was determined by bootstrapping. Statistical testing of detachment rates was performed using a Mann-Whitney test of individual binding interactions.

      Single ATP turnover assay (SRX/DRX)

      Single ATP turnover experiments were performed by diluting myosin S1 to 0.8 μM with 1X assay buffer consisting of 30 mM KAc, 10 mM Tris pH 7.5, 4 mM MgCl2, 1 mM EDTA. Experiments were performed at 25 °C on a CLARIOstar Plate reader with two software-controlled injectors (BMG Labtech). Myosin S1 was added to one well of a 384 well black flat bottom plate (Corning, 3575) to achieve a 0.4 μM final concentration. 2'-(or-3')-O-(N-Methylanthraniloyl) Adenosine 5'-Triphosphate, Trisodium Salt (mant-ATP; ThermoFisher, M12417) was injected into the well at a final concentration of 0.8 μM followed by brief 3 s shake at 300 rpm. At 60 seconds, excess unlabeled ATP (final concentration of 4 mM) was injected and fluorescence was monitored every second for 940 seconds (365 nm excitation/450 nm emission). The fluorescence signal vs. time was plotted, normalized, and fit to a bi-exponential decay function using GraphPad Prism to obtain rates and amplitudes of the fast phase (DRX) and slow phase (SRX). Ambiguous fits were excluded across all data sets. Statistical significance was determined using a one-way ANOVA with Tukey’s multiple comparison test.

      Molecular dynamics simulations

      We prepared homology models of S1 using the MODELLER software package. For the motor domain structural template, we selected a hybrid model of 5N6A with the converter domain of 5N69 because this part of 5N6A was poorly resolved. We followed this approach because both human and python MYH7b motor domains lacked experimental structures. The structural template for the light chain was taken from the light chain in 5N69. For homology modeling of MYH7b, we used the same motor domain and light chain sequences as the construct used in the in vitro experiments. For homology modeling of beta-cardiac myosin, we trimmed the motor domain sequence at the same location as Porter et al. (
      • Porter J.R.
      • Meller A.
      • Zimmerman M.I.
      • Greenberg M.J.
      • Bowman G.R.
      Conformational distributions of isolated myosin motor domains encode their mechanochemical properties.
      ) and used a human ventricular light chain sequence. Homology modeling is summarized in the summary of simulation systems table below.

      Summary of simulation systems

      Tabled 1
      ConstructMyosin Heavy Chain SequenceHeavy Chain Structural TemplateMyosin Light Chain SequenceLight Chain Structural Template
      β-MyHC S1hMYH75N6A/5N69hMYL35N69:H
      hMYH7b S1hMYH7b5N6A/5N69MYL1-3f5N69:H
      pMYH7b S1pMYH7b5N6A/5N69MYL1-3f5N69:H
      We prepared systems for simulation in GROMACS following the same procedure as Porter et al. (
      • Porter J.R.
      • Meller A.
      • Zimmerman M.I.
      • Greenberg M.J.
      • Bowman G.R.
      Conformational distributions of isolated myosin motor domains encode their mechanochemical properties.
      ). The AMBER03 force field and TIP3P water model were used. The protein structures were solvated in a dodecahedral box with a 1 nm pad of TIP3P water. Sodium and chloride ions were added to produce a neutral system at a concentration of 0.1 M NaCl. Each system was minimized using steepest descents until the maximum force on any atom decreased below 1,000 kJ/(mol x nm). The system was then equilibrated for 1 ns with heavy atoms placed under a position restraint at 300K maintained by the Bussi-Parinello thermostat.
      Production simulations were then performed on the [email protected] distributed computing platform using GROMACS 2020 (
      • Abraham M.J.
      • Murtola T.
      • Schulz R.
      • Páll S.
      • Smith J.C.
      • Hess B.
      • Lindahl E.
      GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers.
      ). All covalent bonds involving hydrogen were constrained using the LINCS algorithm with a LINCS order of 6 and a LINCS iter of 2 (
      • Hess B.
      • Bekker H.
      • Berendsen H.J.C.
      • Fraaije J.G.E.M.
      LINCS: A linear constraint solver for molecular simulations.
      ). Virtual sites were enabled to allow for a 4 fs timestep (
      • Feenstra K.A.
      • Hess B.
      • Berendsen H.J.C.
      Improving efficiency of large time-scale molecular dynamics simulations of hydrogen-rich systems.
      ). An aggregate total of 91.7, 78.5, and 79.0 microseconds of simulation data was obtained for β-MyHC S1, hMYH7b S1, and pMYH7b S1, respectively (see table of aggregate simulation times and modeling hyperparameters).

      Markov State Modeling

      Markov State Models were constructed by defining microstates using k-hybrid clustering with 3 rounds of refinement with k-medoids. A cluster radius of 8.0 2 was chosen and the Euclidean distance between residue sidechain solvent accessible surface area was used as a distance metric as previously described in Porter et al. (
      • Porter J.R.
      • Meller A.
      • Zimmerman M.I.
      • Greenberg M.J.
      • Bowman G.R.
      Conformational distributions of isolated myosin motor domains encode their mechanochemical properties.
      ) Markov State Models were fit separately for each isoform by adding a pseudocount of 1/n to each element of the transition counts matrix and row normalizing, as recommended by Zimmerman et al. (
      • Zimmerman M.I.
      • Porter J.R.
      • Sun X.
      • Silva R.R.
      • Bowman G.R.
      Choice of adaptive sampling strategy impacts state discovery, transition probabilities, and the apparent mechanism of conformational changes.
      ). Lag times were chosen by the implied timescales test (see table of aggregate simulation times and modeling hyperparameters and Figure S7). Important hyperparameters are listed in the table below. We find that our results are insensitive to the choice of these hyperparameters.

      Aggregate simulation times and modeling hyperparameters

      Tabled 1
      ConstructAggregate Simulation Time (μs)Cluster Radius (2)Lag Time (ns)
      β-MyHC S191.28.07
      hMYH7b S178.58.07
      pMYH7b S178.98.07
      Total Simulation Time (μs)248.6

      DiffNets

      We employed DiffNets, a self-supervised deep learning framework to identify structural differences in the active site between the myosin isoform ensembles. We trained a DiffNet on the backbone and Cβ atoms of the active site. To select active site residues, we chose all residues within 7.5 angstroms of ADP, Mg, or Pi in the Pre-powerstroke crystal structure (PDB: 5N6A). We then converted the atomic coordinates for each isoform’s active site to DiffNets input following the normalization procedure described in Ward et al. (
      • Ward M.D.
      • Zimmerman M.I.
      • Meller A.
      • Chung M.
      • Swamidass S.J.
      • Bowman G.R.
      Deep learning the structural determinants of protein biochemical properties by comparing structural ensembles with DiffNets.
      ).
      We trained this network to learn a latent representation of the active site that could discriminate whether the active site was from either human or python MYH7b or β-MyHC. The initial labels were set to [0,1,1] for β-MyHC, human MYH7b, and python MYH7b respectively. These labels were iteratively refined as described in Ward et al. (
      • Ward M.D.
      • Zimmerman M.I.
      • Meller A.
      • Chung M.
      • Swamidass S.J.
      • Bowman G.R.
      Deep learning the structural determinants of protein biochemical properties by comparing structural ensembles with DiffNets.
      ). We selected expectation maximization bounds of [[0.1,0.5],[0.5,0.9],[0.5,0.9]] for β-MyHC, human MYH7b, and python MYH7b respectively. We used a total of 20 latent variables in the bottleneck layer. The DiffNet was trained for 10 epochs with 2 hidden layers, 10 epochs with 4 hidden layers (with the weights of the initial 2 layers frozen), and 25 ‘polish’ epochs where all weights were tunable. A learning rate of 0.0001 and a batch size of 32 were chosen. All training and analysis were performed using the open-source package available on GitHub at https://github.com/bowman-lab/diffnets.

      Statistical analysis

      All data except for optical trapping data and molecular dynamics simulations were graphed and analyzed in Graphpad Prism. Unless otherwise noted, mean ± standard deviation is represented and p < 0.05 was the threshold for significance. Sample size and legend for statistical significance are noted in each figure legend.

      DATA AVAILABILITY

      All data are included in the manuscript or supporting information. Raw data from this manuscript is available upon request from the corresponding author.

      SUPPORTING INFORMATION

      This article contains supporting information.

      CONFLICT OF INTEREST

      None declared.

      ACKNOWLEDGEMENTS

      We thank Dr. Angela Peter and Dr. Anastasia Karabina for supplying python tissue and preparing the protein gel for mass spectrometry. We thank Dr. Monika Dzieciatkowska and the University of Colorado School of Medicine Biological Mass Spectrometry Proteomics Core Facility for their mass spectrometry services. We thank the BioFrontiers Institute Advanced Light Microscopy Core for use of microscopy equipment and for imaging support.

      Supplementary data

      REFERENCES

        • Resnicow D.I.
        • Deacon J.C.
        • Warrick H.M.
        • Spudich J.A.
        • Leinwand L.A.
        Functional diversity among a family of human skeletal muscle myosin motors.
        Proceedings of the National Academy of Sciences. 2010; 107: 1053-1058
        • Deacon J.C.
        • Bloemink M.J.
        • Rezavandi H.
        • Geeves M.A.
        • Leinwand L.A.
        Identification of functional differences between recombinant human α and β cardiac myosin motors.
        Cell. Mol. Life Sci. 2012; 69: 2261-2277
        • Bloemink M.J.
        • Deacon J.C.
        • Resnicow D.I.
        • Leinwand L.A.
        • Geeves M.A.
        The Superfast Human Extraocular Myosin Is Kinetically Distinct from the Fast Skeletal IIa, IIb, and IId Isoforms.
        Journal of Biological Chemistry. 2013; 288: 27469-27479
        • Desjardins P.R.
        • Burkman J.M.
        • Shrager J.B.
        • Allmond L.A.
        • Stedman H.H.
        Evolutionary Implications of Three Novel Members of the Human Sarcomeric Myosin Heavy Chain Gene Family.
        Molecular Biology and Evolution. 2002; 19: 375-393
        • Bell M.L.
        • Buvoli M.
        • Leinwand L.A.
        Uncoupling of Expression of an Intronic MicroRNA and Its Myosin Host Gene by Exon Skipping.
        MCB. 2010; 30: 1937-1945
        • Peter A.K.
        • Rossi A.C.
        • Buvoli M.
        • Ozeroff C.D.
        • Crocini C.
        • Perry A.R.
        • Buvoli A.E.
        • Lee L.A.
        • Leinwand L.A.
        Expression of Normally Repressed Myosin Heavy Chain 7b in the Mammalian Heart Induces Dilated Cardiomyopathy.
        J Am Heart Assoc. 2019; 8e013318
        • Rossi A.C.
        • Mammucari C.
        • Argentini C.
        • Reggiani C.
        • Schiaffino S.
        Two novel/ancient myosins in mammalian skeletal muscles: MYH14/7b and MYH15 are expressed in extraocular muscles and muscle spindles: MYH14/7b and MYH15 expression in mammalian skeletal muscles.
        The Journal of Physiology. 2010; 588: 353-364
        • Mu L.
        • Wang J.
        • Su H.
        • Sanders I.
        Adult Human Upper Esophageal Sphincter Contains Specialized Muscle Fibers Expressing Unusual Myosin Heavy Chain Isoforms.
        J Histochem Cytochem. 2007; 55: 199-207
        • Haraksingh R.R.
        • Jahanbani F.
        • Rodriguez-Paris J.
        • Gelernter J.
        • Nadeau K.C.
        • Oghalai J.S.
        • Schrijver I.
        • Snyder M.P.
        Exome sequencing and genome-wide copy number variant mapping reveal novel associations with sensorineural hereditary hearing loss.
        BMC Genomics. 2014; 15: 1155
        • Rubio M.D.
        • Johnson R.
        • Miller C.A.
        • Huganir R.L.
        • Rumbaugh G.
        Regulation of Synapse Structure and Function by Distinct Myosin II Motors.
        Journal of Neuroscience. 2011; 31: 1448-1460
        • Machida S.
        • Noda S.
        • Takao A.
        • Nakazawa M.
        • Matsuoka R.
        Expression of slow skeletal myosin heavy chain 2 gene in Purkinje fiber cells in chick heart.
        Biology of the Cell. 2002; 94: 389-399
        • Lee L.A.
        • Broadwell L.J.
        • Buvoli M.
        • Leinwand L.A.
        Nonproductive Splicing Prevents Expression of MYH7b Protein in the Mammalian Heart.
        J Am Heart Assoc. 2021; 10e020965
        • Sommese R.F.
        • Sung J.
        • Nag S.
        • Sutton S.
        • Deacon J.C.
        • Choe E.
        • Leinwand L.A.
        • Ruppel K.
        • Spudich J.A.
        Molecular consequences of the R453C hypertrophic cardiomyopathy mutation on human -cardiac myosin motor function.
        Proceedings of the National Academy of Sciences. 2013; 110: 12607-12612
        • Spudich J.A.
        Hypertrophic and dilated cardiomyopathy: four decades of basic research on muscle lead to potential therapeutic approaches to these devastating genetic diseases.
        Biophys J. 2014; 106: 1236-1249
        • Nag S.
        • Sommese R.F.
        • Ujfalusi Z.
        • Combs A.
        • Langer S.
        • Sutton S.
        • Leinwand L.A.
        • Geeves M.A.
        • Ruppel K.M.
        • Spudich J.A.
        Contractility parameters of human β-cardiac myosin with the hypertrophic cardiomyopathy mutation R403Q show loss of motor function.
        Sci. Adv. 2015; 1e1500511
        • Finer J.T.
        • Simmons R.M.
        • Spudich J.A.
        Single myosin molecule mechanics: piconewton forces and nanometre steps.
        Nature. 1994; 368: 113-119
        • Blackwell T.
        • Stump W.T.
        • Clippinger S.R.
        • Greenberg M.J.
        Computational Tool for Ensemble Averaging of Single-Molecule Data.
        Biophysical Journal. 2021; 120: 10-20
        • Greenberg M.J.
        • Shuman H.
        • Ostap E.M.
        Inherent Force-Dependent Properties of β-Cardiac Myosin Contribute to the Force-Velocity Relationship of Cardiac Muscle.
        Biophysical Journal. 2014; 107: L41-L44
        • Woody M.S.
        • Greenberg M.J.
        • Barua B.
        • Winkelmann D.A.
        • Goldman Y.E.
        • Ostap E.M.
        Positive cardiac inotrope omecamtiv mecarbil activates muscle despite suppressing the myosin working stroke.
        Nat Commun. 2018; 9: 3838
        • Palmiter K.A.
        • Tyska M.J.
        • Haeberle J.R.
        • Alpert N.R.
        • Fananapazir L.
        • Warshaw D.M.
        R403Q and L908V mutant b-cardiac myosin from patients with familial hypertrophic cardiomyopathy exhibit enhanced mechanical performance at the single molecule level.
        Journal of Muscle Research and Cell Motility. 2000; 21: 609-620
        • Sung J.
        • Nag S.
        • Mortensen K.I.
        • Vestergaard C.L.
        • Sutton S.
        • Ruppel K.
        • Flyvbjerg H.
        • Spudich J.A.
        Harmonic force spectroscopy measures load-dependent kinetics of individual human β-cardiac myosin molecules.
        Nat Commun. 2015; 6: 7931
        • Capitanio M.
        • Canepari M.
        • Cacciafesta P.
        • Lombardi V.
        • Cicchi R.
        • Maffei M.
        • Pavone F.S.
        • Bottinelli R.
        Two independent mechanical events in the interaction cycle of skeletal muscle myosin with actin.
        Proceedings of the National Academy of Sciences. 2006; 103: 87-92
        • Greenberg M.J.
        • Lin T.
        • Goldman Y.E.
        • Shuman H.
        • Ostap E.M.
        Myosin IC generates power over a range of loads via a new tension-sensing mechanism.
        Proc. Natl. Acad. Sci. U.S.A. 2012; https://doi.org/10.1073/pnas.1207811109
        • Laakso J.M.
        • Lewis J.H.
        • Shuman H.
        • Ostap E.M.
        Myosin I Can Act As a Molecular Force Sensor.
        Science. 2008; 321: 133-136
        • Veigel C.
        • Coluccio L.M.
        • Jontes J.D.
        • Sparrow J.C.
        • Milligan R.A.
        • Molloy J.E.
        The motor protein myosin-I produces its working stroke in two steps.
        Nature. 1999; 398: 530-533
        • Veigel C.
        • Wang F.
        • Bartoo M.L.
        • Sellers J.R.
        • Molloy J.E.
        The gated gait of the processive molecular motor, myosin V.
        Nat Cell Biol. 2002; 4: 59-65
        • Chen C.
        • Greenberg M.J.
        • Laakso J.M.
        • Ostap E.M.
        • Goldman Y.E.
        • Shuman H.
        Kinetic Schemes for Post-Synchronized Single Molecule Dynamics.
        Biophysical Journal. 2012; 102: L23-L25
        • Anderson R.L.
        • Trivedi D.V.
        • Sarkar S.S.
        • Henze M.
        • Ma W.
        • Gong H.
        • Rogers C.S.
        • Gorham J.M.
        • Wong F.L.
        • Morck M.M.
        • Seidman J.G.
        • Ruppel K.M.
        • Irving T.C.
        • Cooke R.
        • Green E.M.
        • Spudich J.A.
        Deciphering the super relaxed state of human β-cardiac myosin and the mode of action of mavacamten from myosin molecules to muscle fibers.
        Proc Natl Acad Sci USA. 2018; 115: E8143-E8152
        • Nag S.
        • Trivedi D.V.
        To lie or not to lie: Super-relaxing with myosins.
        eLife. 2021; 10e63703
        • Rohde J.A.
        • Roopnarine O.
        • Thomas D.D.
        • Muretta J.M.
        Mavacamten stabilizes an autoinhibited state of two-headed cardiac myosin.
        Proc Natl Acad Sci USA. 2018; 115: E7486-E7494
        • Stewart M.A.
        • Franks-Skiba K.
        • Chen S.
        • Cooke R.
        Myosin ATP turnover rate is a mechanism involved in thermogenesis in resting skeletal muscle fibers.
        Proceedings of the National Academy of Sciences. 2010; 107: 430-435
        • Gollapudi S.K.
        • Yu M.
        • Gan Q.-F.
        • Nag S.
        Synthetic thick filaments: A new avenue for better understanding the myosin super-relaxed state in healthy, diseased, and mavacamten-treated cardiac systems.
        Journal of Biological Chemistry. 2021; 296100114
        • Planelles-Herrero V.J.
        • Hartman J.J.
        • Robert-Paganin J.
        • Malik F.I.
        • Houdusse A.
        Mechanistic and structural basis for activation of cardiac myosin force production by omecamtiv mecarbil.
        Nat Commun. 2017; 8: 190
        • Porter J.R.
        • Meller A.
        • Zimmerman M.I.
        • Greenberg M.J.
        • Bowman G.R.
        Conformational distributions of isolated myosin motor domains encode their mechanochemical properties.
        eLife. 2020; 9e55132
        • Heissler S.M.
        • Arora A.S.
        • Billington N.
        • Sellers J.R.
        • Chinthalapudi K.
        Cryo-EM structure of the autoinhibited state of myosin-2.
        Science Advances. 2021; 7eabk3273
        • Ward M.D.
        • Zimmerman M.I.
        • Meller A.
        • Chung M.
        • Swamidass S.J.
        • Bowman G.R.
        Deep learning the structural determinants of protein biochemical properties by comparing structural ensembles with DiffNets.
        Nat Commun. 2021; 12: 3023
        • Pande V.S.
        • Beauchamp K.
        • Bowman G.R.
        Everything you wanted to know about Markov State Models but were afraid to ask.
        Methods. 2010; 52: 99-105
        • Llinas P.
        • Isabet T.
        • Song L.
        • Ropars V.
        • Zong B.
        • Benisty H.
        • Sirigu S.
        • Morris C.
        • Kikuti C.
        • Safer D.
        • Sweeney H.L.
        • Houdusse A.
        How Actin Initiates the Motor Activity of Myosin.
        Developmental Cell. 2015; 33: 401-412
        • Ikeda D.
        • Ono Y.
        • Snell P.
        • Edwards Y.J.K.
        • Elgar G.
        • Watabe S.
        Divergent evolution of the myosin heavy chain gene family in fish and tetrapods: evidence from comparative genomic analysis.
        Physiological Genomics. 2007; 32: 1-15
        • McGuigan K.
        Evolution of Sarcomeric Myosin Heavy Chain Genes: Evidence from Fish.
        Molecular Biology and Evolution. 2004; 21: 1042-1056
        • Carlyle B.C.
        • Kitchen R.R.
        • Kanyo J.E.
        • Voss E.Z.
        • Pletikos M.
        • Sousa A.M.M.
        • Lam T.T.
        • Gerstein M.B.
        • Sestan N.
        • Nairn A.C.
        A multiregional proteomic survey of the postnatal human brain.
        Nat Neurosci. 2017; 20: 1787-1795
        • Guo Z.
        • Shao C.
        • Zhang Y.
        • Qiu W.
        • Li W.
        • Zhu W.
        • Yang Q.
        • Huang Y.
        • Pan L.
        • Dong Y.
        • Sun H.
        • Xiao X.
        • Sun W.
        • Ma C.
        • Zhang L.
        A Global Multiregional Proteomic Map of the Human Cerebral Cortex.
        Genomics, Proteomics & Bioinformatics. 2021; https://doi.org/10.1016/j.gpb.2021.08.008
        • Xu J.
        • Patassini S.
        • Rustogi N.
        • Riba-Garcia I.
        • Hale B.D.
        • Phillips A.M.
        • Waldvogel H.
        • Haines R.
        • Bradbury P.
        • Stevens A.
        • Faull R.L.M.
        • Dowsey A.W.
        • Cooper G.J.S.
        • Unwin R.D.
        Regional protein expression in human Alzheimer’s brain correlates with disease severity.
        Commun Biol. 2019; 2: 1-15
        • Hooijman P.
        • Stewart M.A.
        • Cooke R.
        A New State of Cardiac Myosin with Very Slow ATP Turnover: A Potential Cardioprotective Mechanism in the Heart.
        Biophysical Journal. 2011; 100: 1969-1976
        • McNamara J.W.
        • Li A.
        • dos Remedios C.G.
        • Cooke R.
        The role of super-relaxed myosin in skeletal and cardiac muscle.
        Biophys Rev. 2015; 7: 5-14
      1. Rasicci, D. V., Tiwari, P., Desetty, R., Sadler, F. W., Sivaramakrishnan, S., Craig, R., and Yengo, C. M. (2022) Dilated cardiomyopathy mutation E525K in human beta-cardiac myosin stabilizes the interacting heads motif and super-relaxed state of myosin. 10.1101/2022.02.18.480995

        • Lee L.A.
        • Karabina A.
        • Broadwell L.J.
        • Leinwand L.A.
        The ancient sarcomeric myosins found in specialized muscles.
        Skeletal Muscle. 2019; 9: 7
        • Porter J.D.
        • Baker R.S.
        • Ragusa R.J.
        • Brueckner J.K.
        Extraocular muscles: Basic and clinical aspects of structure and function.
        Survey of Ophthalmology. 1995; 39: 451-484
        • Schachat F.
        • Briggs M.M.
        Phylogenetic implications of the superfast myosin in extraocular muscles.
        Journal of Experimental Biology. 2002; 205: 2189-2201
        • Billington N.
        • Beach J.R.
        • Heissler S.M.
        • Remmert K.
        • Guzik-Lendrum S.
        • Nagy A.
        • Takagi Y.
        • Shao L.
        • Li D.
        • Yang Y.
        • Zhang Y.
        • Barzik M.
        • Betzig E.
        • Hammer J.A.
        • Sellers J.R.
        Myosin 18A Coassembles with Nonmuscle Myosin 2 to Form Mixed Bipolar Filaments.
        Current Biology. 2015; 25: 942-948
        • Andersen J.B.
        • Rourke B.C.
        • Caiozzo V.J.
        • Bennett A.F.
        • Hicks J.W.
        Postprandial cardiac hypertrophy in pythons.
        Nature. 2005; 434: 37-38
        • Riquelme C.A.
        • Magida J.A.
        • Harrison B.C.
        • Wall C.E.
        • Marr T.G.
        • Secor S.M.
        • Leinwand L.A.
        Fatty Acids Identified in the Burmese Python Promote Beneficial Cardiac Growth.
        Science. 2011; 334: 528-531
        • Secor S.M.
        • Diamond J.
        A vertebrate model of extreme physiological regulation.
        Nature. 1998; 395: 659-662
        • Keller A.
        • Nesvizhskii A.I.
        • Kolker E.
        • Aebersold R.
        Empirical Statistical Model To Estimate the Accuracy of Peptide Identifications Made by MS/MS and Database Search.
        Anal. Chem. 2002; 74: 5383-5392
        • Nesvizhskii A.I.
        • Keller A.
        • Kolker E.
        • Aebersold R.
        A Statistical Model for Identifying Proteins by Tandem Mass Spectrometry.
        Anal. Chem. 2003; 75: 4646-4658
      2. Toyoshima, Y. Y., Kron, S. J., McNally, E. M., Niebling, K. R., Toyoshima, C., and Spudich, J. A. (1987) Myosin subfragment-l is sufficient to move actin filaments in vitro

      3. Pardee, J. D., and Spudich, J. (1982) [18] Purification of muscle actin. in Methods in Enzymology, pp. 164–181, Structural and Contractile Proteins Part B: The Contractile Apparatus and the Cytoskeleton, Academic Press, 85, 164–181

        • Aksel T.
        • Choe Yu E.
        • Sutton S.
        • Ruppel K.M.
        • Spudich J.A.
        Ensemble force changes that result from human cardiac myosin mutations and a small-molecule effector.
        Cell Rep. 2015; 11: 910-920
      4. Frangi, A. F., Niessen, W. J., Vincken, K. L., and Viergever, M. A. (1998) Multiscale vessel enhancement filtering. in Medical Image Computing and Computer-Assisted Intervention - MICCAI’98 (Wells, W. M., Colchester, A., and Delp, S. eds), pp. 130–137, Lecture Notes in Computer Science, Springer Berlin Heidelberg, Berlin, Heidelberg, 1496, 130–137

        • Jaqaman K.
        • Loerke D.
        • Mettlen M.
        • Kuwata H.
        • Grinstein S.
        • Schmid S.L.
        • Danuser G.
        Robust single particle tracking in live cell time-lapse sequences.
        Nat Methods. 2008; 5: 695-702
      5. Tay, J. W., and Cameron, J. C. (2020) CyAn: A MATLAB toolbox for image and data analysis of cyanobacteria. 10.1101/2020.07.28.225219

        • Criddle A.H.
        • Geeves M.A.
        • Jeffries T.
        The use of actin labelled with N -(1-pyrenyl)iodoacetamide to study the interaction of actin with myosin subfragments and troponin/tropomyosin.
        Biochemical Journal. 1985; 232: 343-349
        • Clippinger S.R.
        • Cloonan P.E.
        • Greenberg L.
        • Ernst M.
        • Stump W.T.
        • Greenberg M.J.
        Disrupted mechanobiology links the molecular and cellular phenotypes in familial dilated cardiomyopathy.
        Proc Natl Acad Sci USA. 2019; 116: 17831-17840
      6. Greenberg, M. J., Shuman, H., and Ostap, E. M. (2017) Measuring the Kinetic and Mechanical Properties of Non-processive Myosins Using Optical Tweezers. in Optical Tweezers: Methods and Protocols (Gennerich, A. ed), pp. 483–509, Methods in Molecular Biology, Springer, New York, NY, 10.1007/978-1-4939-6421-5_19

        • Woody M.S.
        • Lewis J.H.
        • Greenberg M.J.
        • Goldman Y.E.
        • Ostap E.M.
        MEMLET: An Easy-to-Use Tool for Data Fitting and Model Comparison Using Maximum-Likelihood Estimation.
        Biophysical Journal. 2016; 111: 273-282
        • Abraham M.J.
        • Murtola T.
        • Schulz R.
        • Páll S.
        • Smith J.C.
        • Hess B.
        • Lindahl E.
        GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers.
        SoftwareX. 2015; (1–2): 19-25
        • Hess B.
        • Bekker H.
        • Berendsen H.J.C.
        • Fraaije J.G.E.M.
        LINCS: A linear constraint solver for molecular simulations.
        Journal of Computational Chemistry. 1997; 18: 1463-1472
        • Feenstra K.A.
        • Hess B.
        • Berendsen H.J.C.
        Improving efficiency of large time-scale molecular dynamics simulations of hydrogen-rich systems.
        Journal of Computational Chemistry. 1999; 20: 786-798
        • Zimmerman M.I.
        • Porter J.R.
        • Sun X.
        • Silva R.R.
        • Bowman G.R.
        Choice of adaptive sampling strategy impacts state discovery, transition probabilities, and the apparent mechanism of conformational changes.
        J Chem Theory Comput. 2018; 14: 5459-5475