Molecular consequences of the R453C hypertrophic cardiomyopathy mutation on human β-cardiac motor function

Ruth F. Sommesea,1, Jongmin Sunga,b,1, Suman Naga,1, Shirley Suttona, John C. Deaconc, Elizabeth Choea,d, Leslie A. Leinwandc, Kathleen Ruppela,e,2, and James A. Spudicha,2

Departments of aBiochemistry, ePediatrics (Cardiology), and dCancer Biology Program, Stanford University School of Medicine, Stanford, CA 94305; bDepartment of Applied Physics, Stanford University, Stanford, CA 94305; and cBioFrontiers Institute, Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, CO 80309

Contributed by James A. Spudich, May 19, 2013 (sent for review April 26, 2013) Cardiovascular disorders are the leading cause of morbidity and activity (8). Lowey et al. subsequently showed that the effects of mortality in the developed world, and hypertrophic cardiomyopa- the R403Q mutation in mouse cardiac myosin depended on the thy (HCM) is among the most frequently occurring inherited cardiac isoform into which the mutation was introduced. In α-cardiac disorders. HCM is caused by mutations in the encoding the myosin, they saw an increase in both ATPase activity and ve- fundamental force-generating machinery of the cardiac muscle, locity, whereas in the β-cardiac myosin, there was no significant including β-cardiac myosin. Here, we present a biomechanical anal- change in the velocity and actually a slight decrease in the ATPase ysis of the HCM-causing mutation, R453C, in the context of human activity (9). β-cardiac myosin. We found that this mutation causes a ∼30% de- As the R403Q data illustrate, the difficulty in interpreting the crease in the maximum ATPase of the human β-cardiac subfragment effects of these mutations has largely been due to the limited 1, the motor domain of myosin, and a similar percent decrease in the availability of homogeneous and fully active human motor pro- in vitro velocity. The major change in the R453C human β-cardiac teins and the inadequacy of commonly used animal models. Given subfragment 1 is a 50% increase in the intrinsic force of the motor that the HCM mutant phenotype results from single-residue compared with wild type, with no appreciable change in the stroke substitutions in β-MHC, observing the effects of such a mutation BIOCHEMISTRY size, as observed with a dual-beam optical trap. These results predict in a nonhuman background where there are many other that the overall force of the ensemble of myosin molecules in the residue differences from the human sequence is far from ideal. muscle should be higher in the R453C mutant compared with wild The majority of animal protein studies are derived from rodent type. Loaded in vitro motility assay confirms that the net force in the models that have a different cardiac MHC isoform composition ensemble is indeed increased. Overall, this study suggests that (predominantly α-MHC instead of β-MHC). Both α- and β-MHC the R453C mutation should result in a hypercontractile state in isoforms from rodents have >30 residue differences compared the muscle. with human β-cardiac myosin, and HCM-causing myosin muta- tions in mice have different biochemical effects depending on heart disease | optical trapping | single-molecule force measurements their backbone isoform (9, 10). Until recently, expression of hu- man cardiac muscle myosin has proven recalcitrant to common ypertrophic cardiomyopathy (HCM) is among the most recombinant protein expression methods. Additionally, cardiac Hcommon of inherited cardiovascular disorders, affecting ∼1in biopsies yield limited amounts of enzymatically active motor and 500 individuals (1). HCM is characterized by left ventricular hy- are heterogeneous in the proportion of the mutated form. This pertrophy, cardiomyocyte disarray, and myocardial fibrosis. HCM problem has now been addressed by a recently developed muscle patients typically have left ventricles with small cavities and pre- cell expression system that provides a source of active recombi- β served or even enhanced global systolic function, but impaired nant human -cardiac muscle myosin (11). β relaxation (2). Clinically, HCM is the most common cause of In this study, we have focused on the HCM-causing -cardiac fi β sudden cardiac death in young adults (3). The first genetic cause of myosin mutation R453C, located next to strand ve of the - fi pleated sheet of myosin ∼5 nm from the -binding region and familial HCM was identi ed in 1990 (4). Missense mutations in the ∼ Discussion major of the sarcomere have since been identified in HCM 3 nm from the nucleotide-binding site ( ). This muta- patients, and may account for up to 60% of all cases (2). Although tion is considered to be malignant and the small number of a number of these mutations have been studied using a variety of investigations into the functional biochemical and biophysical effects of this mutation have yielded contradictory results (12–14). approaches, there is no clear consensus as to the mechanism(s) by fi which these mutations give rise to the disease state. We have puri ed a Subfragment 1 (S1) construct of human β-cardiac myosin containing a truncated human MHC (residues One of the most commonly mutated genes is myosin heavy – chain 7 or MYH7, which encodes β-cardiac myosin heavy chain 1 808) and the human ventricular essential light chain (ELC) (β-MHC), the major cardiac muscle myosin isoform in the human (Fig. S1), followed by a short linker and a carboxy terminal eGFP heart. In general, HCM resulting from β-MHC mutations are characterized by early onset and severe left ventricular hyper- trophy (5). Although several mutations in β-MHC have been Author contributions: R.F.S., J.S., S.N., K.R., and J.A.S. designed research; R.F.S., J.S., S.N., fi S.S., E.C., and K.R. performed research; J.C.D., E.C., L.A.L., and K.R. contributed new studied over the past 15 y, there has been signi cant disagreement reagents/analytic tools; R.F.S., J.S., S.N., and J.A.S. analyzed data; and R.F.S., J.S., S.N., K.R., in the literature as to the effects of these mutations on the bio- and J.A.S. wrote the paper. chemical and biophysical properties of β-cardiac myosin (reviewed The authors declare no conflict of interest. in ref. 6). For example, the first cardiomyopathy-causing mutation Freely available online through the PNAS open access option. β fi in -cardiac myosin to be identi ed was R403Q. Initial studies on 1R.F.S., J.S., and S.N. contributed equally to this work. this mutant isolated from cardiac and soleus muscle biopsies 2To whom correspondence may be addressed. E-mail: [email protected] or kruppel@ reported a decrease in in vitro sliding velocity (7). Later studies stanford.edu. α using mouse -cardiac myosin containing the R403Q mutation This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. found an increase in velocity and also an increase in the ATPase 1073/pnas.1309493110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1309493110 PNAS Early Edition | 1of6 Downloaded by guest on September 26, 2021 moiety (SI Materials and Methods). We refer to this construct as saturating ATP concentration, or the strongly bound state time (ts) simply human β-cardiac S1 throughout. Here we present a de- (Fig. S2). For the R453C mutant and the WT, ts was ∼16 ms. These scription of the effects of this mutation on human β-cardiac S1 values are, however, only estimates given the difficulty of accu- function at both the single-molecule and ensemble levels. Sin- rately detecting and selecting short events in the trap, which are gle-molecule analysis revealed a significant increase in intrinsic challenging to distinguish from Brownian noise of the trapped force of the mutant motor by 50%, such that the net ensemble beads and instrument noise. Hence, a small change in the ts is force is higher than the wild-type (WT) motor, as confirmed by currently difficult to measure with our instrument. a loaded in vitro motility assay. Together, the data support a model whereby heart muscle containing this mutation would R453C Human β-Cardiac S1 Has a Lower Velocity than WT. We mea- be hypercontractile. sured the velocity of motor-driven actin filaments in methylcel- lulose as a function of actin filament length for both WT and Results R453C human β-cardiac S1. The average maximum velocity (vo) − R453C Human β-Cardiac S1 Has a Lower Maximum Actin-Activated for R453C at 23 °C for human β-cardiac S1 (610 ± 30 nm·s 1) −1 ATPase than WT. To measure the maximal rate of ATP turnover was significantly slower than that of WT (800 ± 40 nm·s )(P < (kcat) by the acto–myosin complex, we used an actin-activated 0.0001). The decrease in velocity for R453C compared with WT myosin ATPase assay at 23 °C. The kcat of R453C human was also observed at 30 °C (Fig. 2, Table S1). −1 β-cardiac S1 (5.0 ± 0.2 s ) was significantly lower (30%; P < The maximum velocity vo in the in vitro motility assay is related − 0.0002) than that of the WT motor (7.4 ± 0.4 s 1) (Fig. 1A, Table to the displacement generated by the myosin power stroke (i.e., S1). This difference was also seen with human β-cardiac S1 that the stroke size d) and the duration that myosin remains strongly was freshly recycled by binding to actin in the absence of ATP attached to actin (ts), such that vo= d/ts (15). Besides the estimates and then releasing active heads from the actin in 2 mM ATP (this from the trap measurements, the ts of S1 in saturating ATP process eliminates “dead heads” from the population). The Km concentration can be estimated by fitting the velocity as a function for WT was 38 ± 4 μM and for R453C was 28 ± 4 μM. of actin filament length as shown in Fig. 2 (15). Assuming d is ∼6 The inverse of the kcat is the amount of time it takes for the nm from the trap data, ts should be ∼10 ms for R453C and ∼7ms motor to complete one ATPase cycle, tc (Fig. 1B). Thus, the tc for for WT human β-cardiac S1 from the vo measurements at 23 °C. WT human β-cardiac S1 was 140 ± 10 ms and for R453C was 200 ± With this change in ts and the similar increase in tc, the fraction of 10 ms. These measures of tc point to a fundamental difference heads bound to actin at steady-state (i.e., the duty ratio = ts/tc)is in the cross-bridge kinetics between R453C and WT human similar between R453C and WT. We note that this is the observed β-cardiac myosin. duty ratio at saturating ATP and saturating actin concentrations. The duty ratio is ∼5–10%, which is expected for β-cardiac myosin, R453C and WT Human β-Cardiac S1 Have a Similar Stroke Size and as it is a low duty ratio motor (16, 17). Strongly Bound State Time in the Optical Trap. We measured the stroke size (d) of WT and R453C human β-cardiac S1 at the single- R453C Human β-Cardiac S1 Has a Significantly Increased Intrinsic molecule level using a dual-beam laser trap assay at 23 °C. The Force Compared with WT. In addition to the mutant’s effect on magnitude of the displacement of the optically trapped bead at low motility, we investigated whether the force-generating capability trap forces is a measure of the stroke size (d) of the myosin pro- of the myosin molecule was affected, using a dual-beam optical duced during the power stroke (Fig. S2). There was no statistical trap. Here we determined the motion of the trapped beads with difference between the stroke sizes of R453C (6 ± 1 nm) and WT high spatial and temporal resolution under high load, achieved by (6 ± 1 nm) human β-cardiac S1. These values are consistent with a position feedback system. This feedback signal moves the trap a rotation of ∼70° of a truncated lever arm (consisting of only one position in response to the myosin stroke and can be used to light chain). quantify the intrinsic force (f) generated by a single myosin mol- It was also possible to get an estimate of the length of time that ecule. For the R453C mutant, the average intrinsic force pro- the myosin molecule is strongly bound to the actin filament in duced by a single myosin molecule was 2.1 ± 0.1 pN, which was

Fig. 1. Actin-activated ATPase. (A) Actin-activated ATPase activities for purified human WT (black) and R453C mutant (red) β-cardiac S1 at 23 °C. The data points shown are the average of multiple experiments (n = 4–6), from three protein preparations. The error bars represent SEM. The data were fittothe −1 Michaelis–Menten equation to obtain the kcat and Km values shown in Table S1. At 23 °C, WT has a kcat of 7.4 ± 0.4 s and Km of 38 ± 4 μM, whereas R453C has −1 a kcat of 5.0 ± 0.2 s and Km of 28 ± 4 μM. (B) The ATPase cycle illustrates the basic myosin states along with the presumed position of the lever arm. Strong actin-binding states of myosin are indicated in red, the weakly associated states of myosin in yellow, and actin in gray. The time spent for one cycle (tc) is equal to the sum of the times spent in the weakly (tw) and the strongly bound states (ts).

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1309493110 Sommese et al. Downloaded by guest on September 26, 2021 motor function under load using an ensemble of motors. Specif- ically, we used the in vitro motility assay in the presence of an external load to examine whether the R453C mutant’s ensemble force-generating ability was altered. We introduced a frictional load in the in vitro motility assay using α-, an actin-binding protein that leads to a molecular tug of war with myosin (18, 19). Typically, the amount of α-actinin needed to stop filament movement is reported as an indication of the myosin isometric force. As shown in Fig. 3B, the R453C mutant needed more α-actinin than WT to stop the same fraction of filaments from moving. Recognizing that the best model to describe such data is still under debate (18), we restrict ourselves to a qualitative de- scription of the difference between R453C and WT. Nevertheless, these ensemble force measurements show that R453C has a Fig. 2. Length-dependent actin filament velocity at 23 °C and 30 °C. The higher force-generating ability compared to WT and corroborates average length and average velocity distributions are shown for WT in black our single-molecule intrinsic force measurements. (A and C) and for R453C in red (B and D)at23°C(A and B) and 30 °C (C and D). For WT, average filament and velocity data were combined from ≥3 Discussion ≥ fi independent motor preparations and 3,000 lament tracks. For R453C, Although the clinical aspects of HCM have been studied for many average filament and velocity data were combined from ≥2 independent motor preparations and ≥2,000 filament tracks. The line in gray represents decades, there is no clear understanding of the underlying mo- fi v l lecular mechanisms leading to the disease. In the case of R453C, the t of the experimental velocity ( exp) and the actin length ( )tothe À Álk d ts the molecular data have not been consistent across studies. following equation: vexp = × 1 − 1 − (15). The d values are from our ts tc Palmer et al. (13) showed a decreased kcat when the mutation was

optical trap measurements (6 nm), and the tc values are from our actin- in mouse α-cardiac myosin. Debold et al. (14) later found with activated ATPase measurements. The fitting parameters are k, which rep- mouse α-cardiac myosin no change in kcat or vo, but an increase in fi t resents the number of motor heads per unit length of actin lament, and s. the average ensemble force (Fensemble). These inconsistencies are The maximum unloaded in vitro velocity (v0) is equal to d/ts. likely due to both the isoform backbone and source of the purified cardiac myosin. Mouse α-cardiac myosin differs by 83 residues in BIOCHEMISTRY – β P < the head domain (human residues 1 808) from human -cardiac 50% higher ( 0.001) than the force generated by the WT myosin. When one is considering the effects of a single residue β ± human -cardiac S1 under the same conditions (1.4 0.1 pN) change on the function of human β-cardiac myosin, working with A (Fig. 3 , Fig. S3). that contain >80 residue differences is far from ideal. β fi By using a recently developed myoblast-based cardiac myosin The Ensemble Force of R453C Human -Cardiac S1 Is Signi cantly expression system (11) and focusing on purified human β-cardiac Higher than WT. In the context of the heart, individual cardiac myosin, we are now examining the effects of HCM and dilated myosin motors work as an ensemble to generate a contractile cardiomyopathy (DCM) mutations in the very protein in which force that propels blood throughout the circulatory system. The the diseased state is produced. Sata et al. (12) achieved some ensemble force should be the intrinsic force (f) multiplied by the expression of human β-cardiac myosin in SF9 cells, but with number of myosin heads interacting with the actin in a force- considerably lower ATPase activities than we report here for both producing state at any moment, which is determined by the duty WT and R453C β-cardiac myosin. Their large decreases in both ratio (ts/tc). That is, Fensemble = f · ts/tc · NT, where NT is the total ATPase and in vitro motility velocities for the R453C mutant number of heads in the overlap zone of the sarcomere. We compared with WT, as well as for other mutations they studied, therefore tested the effect of the R453C mutation on myosin likely reflect the difficulty of obtaining active recombinant human

Fig. 3. Intrinsic and ensemble force measurements of purified human WT and R453C mutant β-cardiac S1 at 23 °C. WT is shown in black and R453C in red. (A) Single-molecule intrinsic force measurements. Each data point is an independent protein preparation, which is averaged over two or more molecules. Each single molecule measurement contained more than 50 events on average. The error bars represent SEM. The horizontal bar is the average of all of the measurements. For the R453C, the average intrinsic force produced by a single S1 molecule is 2.1 ± 0.1 pN (n = 10), and for the WT S1, it is 1.4 ± 0.1 pN (n = 14). (B) Loaded in vitro motility measurements. For both WT and R453C, percent actin filaments moving at varying α-actinin concentrations were examined for at least four motor preparations. Error bars represent SEM (n = 2–5 surfaces) (Materials and Methods).

Sommese et al. PNAS Early Edition | 3of6 Downloaded by guest on September 26, 2021 Fig. 4. Human β-cardiac myosin S1 motor domain bound to ATP. Residues 1–808 of human β-cardiac myosin S1, in gray, bound by the ventricular ELC in blue. As shown in the right panel, the R453C residue (red) binds in the linker between the α-helix (yellow) and strand five (dark green) of the β-pleated sheet (green). This region in the core of the motor domain is also referred to the transducer (22). This figure is a combination of structure 2MYS and 4DB1.

β-cardiac myosin in heterologous expression systems, possibly due above, however, these results likely reflected the difficulty of to the lack of muscle-cell–specific chaperones (11). obtaining active motor from small amounts of biopsy material or The ∼30% decreases that we observed in both the kcat of the from heterologous expression systems. More recent studies of ATPase and the unloaded maximum velocity of movement of the muscle fibers, cardiac myofibrils, or cardiac myosins containing R453C mutant S1 compared with the WT S1 result in corre- various HCM mutations reveal a picture more consistent with sponding increases in both tc and ts (Fig. 1B). The duty ratio, ts/tc, these mutations resulting in enhanced contractility (reviewed in which is one determinant of the total ensemble force production ref. 6), in agreement with the results that we report here for R453C by the muscle, is therefore not changed much in the R453C β-cardiac myosin. It has been noted that patients with HCM often β-cardiac myosin (∼5% lower than WT). The major change in have enhanced contractile function, measured echocardiograph- the human β-cardiac S1 as a result of the R453C mutation is ically as an increased left ventricular ejection fraction (LVEF) (24, ∼ a 50% increase in intrinsic force, which should lead to hyper- 25). Recent echocardiographic studies of patients harboring contractile cardiac function. HCM-causing mutations in β-cardiac myosin who have no evi- It is important to point out that the duty ratio estimated in this dence of hypertrophy (so-called genotype positive/phenotype neg- study is the unloaded duty ratio. During contraction, though, ative or preclinical HCM patients) demonstrate that these individ- muscle is under load and load affects the kinetics of muscle me- uals also have increased LVEFs. This suggests that the enhanced chanics (Fenn Effect) (20). One recent study with smooth muscle contractility is due to the mutations themselves and not a result of myosin was able to examine the effect of load on the ts using an fi t the muscle hypertrophy (26). The mechanism by which this hyper- optical trap (21). One dif culty with measuring s for cardiac contractility could serve as a stimulus for hypertrophy is not known. myosin, though, is that the ts under saturating ATP concentrations fi fi This study demonstrates enhanced function of highly puri ed, is extremely short and dif cult to detect within the noise of the homogeneous human β-cardiac myosin containing an HCM- trap. An accurate measurement of the ts under varying load will be causing mutation. More studies on the biochemical and bio- key to understanding the effects of different β-cardiac human physical properties of a variety of HCM mutations studied in the myosin mutations on muscle function and power. human myosin context are needed to see whether this will prove to It is possible that the R453C mutation alters the “spring con- be more generally true. A future challenge will be to connect the stant” k of the motor, simplifying the intrinsic force as f = k · x, underlying biochemical and biophysical mechanisms to the sub- where x is the displacement due to the stroke. It is interesting that residue 453 lies very near strand five of the β-pleated sheet in the sequent patho-physiological changes observed in patients carrying core of the myosin head domain (Fig. 4). It is interesting to note HCM mutations. that there are no hypertrophic, dilated, or left ventricular non- Materials and Methods compaction cardiomyopathy-causing mutations that have been fi β Myosin Constructs and Protein Expression. The myosin proteins used in this identi ed in this strand of the -pleated sheet (there have been study were constructed and produced as previously described (11, 27), using numerous mutations found in the other six strands). It is possible the AdEasy Vector System (Qbiogene, Inc) with minor modifications. MHY7 fi that any alterations to this strand may have signi cant enough cDNA and 3 (MYL3) (ventricular ELC) were purchased consequences on motor function to be nonviable. This β-pleated from Open Biosystems (Thermo). A truncated version of MHY7 (residues sheet, also referred to as the transducer, undergoes a distortion or 1–808) was made, corresponding to a short S1. For construct details, see twist during the transition from the prestroke to poststroke states SI Materials and Methods. Replication-deficient recombinant adenoviruses (22). The strain in this sheet may well be part of the spring that were produced and amplified in HEK293 cells. Viral particles were purified establishes the intrinsic force-producing capability of the myosin from clarified cell lysates by sequential step and linear cesium chloride head. R453C lies at the bend between an α-helix and the 7-stranded gradients, and concentrated virus was stored in a glycerol buffer at –20 °C. sheet, and an increase in stiffness at the bend would contribute to Murine C2C12 myoblasts (ATCC) were cultured in GlutaMAX DMEM a larger spring constant. The overall spring constant must of course (Gibco) supplemented with 10% (vol/vol) FBS (Sigma) and 1% penicillin/ be related to the overall changes in the motor structure when streptomycin (Gibco). Forty-eight hours postdifferentiation in DMEM sup- plemented with 2% (vol/vol) horse serum (Sigma) and 1% penicillin/strep- transitioning between the prestroke and poststroke states. tomycin, cells were infected with 1 × 106–1 × 108 plaque-forming units of An early hypothesis regarding HCM pathophysiology was that both MHC and ELC viruses. Three to four days postinfection, myotubes HCM-causing mutations such as R453C resulted in decreased expressing recombinant MHC were collected by cell scraping and lysed. contractile activity, and hence the hypertrophy of the cardiac FLAG-tagged ELC and associated MHC was purified from clarified lysate with muscle was a compensatory response. This hypothesis appeared anti-FLAG resin. The human β-cardiac S1 was separated from endogenous to be corroborated by initial biochemical studies of HCM-caus- skeletal myosin and further purified by ion exchange chromatography on ing β-cardiac myosin mutant proteins (7, 12, 23). As mentioned a 1 mL HiTrap Q HP column (GE Healthcare) (Fig. S1).

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1309493110 Sommese et al. Downloaded by guest on September 26, 2021 Additional Protein Purification. Actin was prepared from fresh chicken breast discussion regarding analysis of velocities and motor cleanup, see SI Mate- skeletal muscle as previously described with slight modification (28). In the rials and Methods.

polymerization step, actin paracrystals were formed by adding MgCl2 to a final concentration of 50 mM instead of KCl, and only 0.2 mM ATP was Loaded in Vitro Motility. Loaded in vitro motility assays were performed using used at this step. The solution was then slowly stirred at room temperature α-actinin as a load (34). After anti-GFP antibody was incubated with the for 1 h to ensure complete formation of paracrystals. Actin was stored with surface, excess antibody was washed out with AB, followed by the desired 0.02% sodium azide. Actin was cycled from G- to F-actin freshly for each α-actinin concentration. The surface was then blocked with BSA and motor assay and only used for up to 5 d before being cycled again. flowed in as described above. For each condition, multiple 30-s movies were Full-length human (29) was expressed and purified with mod- taken in each flow cell and movies were then analyzed to determine the ifications to previous methods as discussed in SI Materials and Methods (30). fraction of filaments moving during those 30 s. The concentration of motor For the loaded in vitro motility assays, α-actinin was purified from chicken and anti-GFP antibody was held constant for each α-actinin concentration. gizzards as previously described with a few minor changes (31). After the Measuring the fraction of filaments stuck proved to be the most consistent diethylaminoethanol (DEAE)-Sepharose and S-Sepharose steps described, the way of measuring the effect of α-actinin on filament movement, as has been protein was bound to a hydroxyapatite column and eluted with a 0–250 mM previously observed (34). potassium phosphate gradient. Binding to actin filaments was confirmed by sedimentation analysis. Optical Trap. Experiments were performed in a manner similar to those de- scribed for the unloaded in vitro motility assay. Typical myosin concentrations Actin-Activated ATPase Assay. For the ATPase, gelsolin was added to actin at used were ∼200 pM to ensure binding events from a single myosin molecule. a ratio of 1:1,000. Gelsolin at this concentration acted to decrease the viscosity ABBSA containing ATP, TMR–phalloidin labeled biotin–actin filaments, of the actin, and thereby decrease pipetting error, without interfering with streptavidin-coated polystyrene beads, 1 mM phalloidin, and the oxygen- the ATPase accuracy (Fig. S4). Actin-activated ATPase assays were then per- scavenging and ATP regeneration systems described above was flowed formed as previously described using a colorimetric readout (32). Briefly, through the chamber. The chamber was sealed with vacuum grease, and myosin was diluted to a final concentration of 0.03–0.1 μM (with 3–5 times as each such slide was used up to 1 h. much for the basal myosin ATPase control in the absence of actin to amplify Detailed experimental setup, experimental procedure, and the data the signal) with 2 mM ATP (CalBiochem) and actin at concentrations ranging analysis are described in SI Materials and Methods. Briefly, each beam μ fi from 0 to 100 M. The nal buffer conditions were 10 mM Imidazole, pH 7.5, trapped a streptavidin-coated polystyrene bead (diameter of 1 μm) that was 5 mM KCl, 3 mM MgCl2, and 1 mM DTT. The reaction was performed at attached to the ends of a fluorescently labeled biotin–actin filament. The fi either 23 °C or 30 °C with shaking in a Thermo Scienti c Multiskan GO, and bead–actin–bead system, called actin dumbbell, was tightly stretched and fi four to ve time points were taken for each concentration. As the cardiac S1 was brought into close proximity to the surface of a silica bead pedestal activity was linear over the time period of the assay, an ATP-regenerating (diameter of 1.5 μm). Typical trap stiffness for the experiments was 0.04 pN/ −1 < BIOCHEMISTRY system was not necessary. Basal activity ( 0.2 s ) was subtracted to get nm (weak trap) and 0.2 pN/nm (strong trap), estimated from the equi- fi – actin-activated ATPase activity. Data were t to a Michaelis Menten equa- partition theorem or the power spectral density analysis (35). ADP release k tion to determine the maximal activity ( cat) and the associated actin con- rate measurements were performed at 2 mM ATP. Multiple binding events K stant for myosin ( m) using OriginLab. Statistical analysis to determine from a single molecule (Fig. S2A) were selected and binned into a histogram fi ’ t P signi cance was performed using a standard Welch s test. All of our (Fig. S2B). An exponential distribution with a maximum likelihood estima- < values were 0.01. tion was used to extract the mean strongly bound state time. Stroke size measurements were collected at low ATP concentration, typically 0.5 μM, to Unloaded in Vitro Motility. The basic method followed our previously de- ensure long binding times. Gaussian distribution with a least-square fitting scribed motility assay (33) with the following modifications. Coverslips were was used to extract the mean stroke size from the individual molecules (Fig. pretreated with nitrocellulose (1% in amyl acetate, Ernst Fullam) diluted to S2C). Force measurements were carried out under similar conditions as 0.2% in amyl acetate (Sigma). Before blocking of the surface with 1 mg/mL stroke size measurements. We used a force feedback control of the trap BSA, anti-GFP antibody (Millipore) was flowed in to coat the coverslip such beam to measure the isometric force generated by the motor (Fig. S3). that β-cardiac myosin constructs could be attached to the coverslip by way of their C-terminal eGFP tag. The assay buffer (AB) used was 25 mM Imidazole ACKNOWLEDGMENTS. The authors thank Prof. Henrik Flyvberg and Kim (pH 7.5), 25 mM KCl, 4 mM MgCl2, 1 mM EGTA, 1 mM DTT. Both anti-GFP Mortensen for suggestions and help in the analysis of the data; Paige antibody and myosin concentrations were varied to determine the surface Shaklee for her work with the initial protein expression system; Chao Liu for density that supported maximal velocity. The final motility solution of AB helping develop the motility threshold algorithm; Ivan Rayment for pro- with 1 mg/mL BSA (ABBSA) contained methylcellulose at a concentration of viding the crystal structures; Craig Buckley for assisting with the optical trap 0.5% (33), tetramethylrhodamine (TMR)–phalloidin labeled actin, 2 mM ATP, setup; and Prof. Alexander Dunn and Prof. Zev Bryant for discussions on the optical trap experiments. This work was funded by National Institutes of an oxygen-scavenging system (0.3–0.4% glucose, 0.25 μg/mL glucose oxi- μ Health (NIH) Grant GM33289 (to J.A.S.), NIH Grant GM29090 (to L.A.L.), dase, 0.45 g/mL catalase), and an ATP regeneration system [1 mM phos- Stanford Interdisciplinary Graduate Fellowship (to R.F.S.), Stanford Dean’s phocreatine (PCR), 0.1 mg/mL creatine phosphokinase (CPK)]. Movies were Postdoctoral Fellowship (to S.N.), Stanford Bio-X Fellowship (to J.S.), and obtained at 23 °C and 30 °C at a frame rate of 1 Hz using a Nikon Ti-E NIH Clinical and Translational Science Award Grant KL2 RR025743 and Ad- inverted microscope with Andor iXon+EMCCD camera model DU885. For vanced Residency Training at Stanford Fellowship (to E.C.).

1. Maron BJ, et al. (1995) Prevalence of hypertrophic cardiomyopathy in a general 9. Lowey S, et al. (2008) Functional effects of the hypertrophic cardiomyopathy R403Q population of young adults. Echocardiographic analysis of 4111 subjects in the CAR- mutation are different in an alpha- or beta-myosin heavy chain backbone. J Biol Chem DIA Study. Coronary Artery Risk Development in (Young) Adults. Circulation 92(4): 283(29):20579–20589. 785–789. 10. Lowey S, Bretton V, Gulick J, Robbins J, Trybus KM (2013) Transgenic mouse alpha-and 2. Ramaraj R (2008) Hypertrophic cardiomyopathy: Etiology, diagnosis, and treatment. beta-cardiac myosins containing the R403Q mutation show isoform dependent tran- Cardiol Rev – 16(4):172 180. sient kinetic differences. J Biol Chem 288(21):14780–14787, 10.1074/jbc.M113.450668. 3. Harvey PA, Leinwand LA (2011) The cell biology of disease: Cellular mechanisms of 11. Resnicow DI, Deacon JC, Warrick HM, Spudich JA, Leinwand LA (2010) Functional J Cell Biol – cardiomyopathy. 194(3):355 365. diversity among a family of human skeletal muscle myosin motors. Proc Natl Acad Sci 4. Geisterfer-Lowrance AA, et al. (1990) A molecular basis for familial hypertrophic USA 107(3):1053–1058. cardiomyopathy: A beta cardiac myosin heavy chain missense mutation. Cell 12. Sata M, Ikebe M (1996) Functional analysis of the mutations in the human cardiac 62(5):999–1006. beta-myosin that are responsible for familial hypertrophic cardiomyopathy. Implica- 5. Frey N, Luedde M, Katus HA (2012) Mechanisms of disease: Hypertrophic cardiomy- tion for the clinical outcome. J Clin Invest 98(12):2866–2873. opathy. Nat Rev Cardiol 9(2):91–100. 13. Palmer BM, et al. (2004) Differential cross-bridge kinetics of FHC myosin mutations 6. Moore JR, Leinwand L, Warshaw DM (2012) Understanding cardiomyopathy pheno- Am J Physiol Heart Circ types based on the functional impact of mutations in the myosin motor. Circ Res R403Q and R453C in heterozygous mouse myocardium. Physiol – 111(3):375–385. 287(1):H91 H99. 7. Cuda G, Fananapazir L, Epstein ND, Sellers JR (1997) The in vitro motility activity of 14. Debold EP, et al. (2007) Hypertrophic and mutations differ- beta-cardiac myosin depends on the nature of the beta-myosin heavy chain gene entially affect the molecular force generation of mouse alpha-cardiac myosin in the mutation in hypertrophic cardiomyopathy. J Muscle Res Cell Motil 18(3):275–283. laser trap assay. Am J Physiol Heart Circ Physiol 293(1):H284–H291. 8. Tyska MJ, et al. (2000) Single-molecule mechanics of R403Q cardiac myosin isolated 15. Uyeda TQP, Kron SJ, Spudich JA (1990) Myosin step size. Estimation from slow sliding from the mouse model of familial hypertrophic cardiomyopathy. Circ Res 86(7): movement of actin over low densities of heavy meromyosin. J Mol Biol 214(3): 737–744. 699–710.

Sommese et al. PNAS Early Edition | 5of6 Downloaded by guest on September 26, 2021 16. De La Cruz EM, Ostap EM (2004) Relating biochemistry and function in the myosin 25. Olivotto I, et al. (2010) Spectrum and clinical significance of systolic function and superfamily. Curr Opin Cell Biol 16(1):61–67. myocardial fibrosis assessed by cardiovascular magnetic resonance in hypertrophic 17. De La Cruz EM, Wells AL, Sweeney HL, Ostap EM (2000) Actin and light chain isoform cardiomyopathy. Am J Cardiol 106(2):261–267. dependence of myosin V kinetics. Biochemistry 39(46):14196–14202. 26. Ho CY, et al. (2002) Assessment of diastolic function with Doppler tissue imaging to predict 18. Greenberg MJ, Moore JR (2010) The molecular basis of frictional loads in the in vitro genotype in preclinical hypertrophic cardiomyopathy. Circulation 105(25):2992–2997. motility assay with applications to the study of the loaded mechanochemistry of 27. Deacon JC, Bloemink MJ, Rezavandi H, Geeves MA, Leinwand LA (2012) Identification α β molecular motors. (Hoboken) 67(5):273–285. of functional differences between recombinant human and cardiac myosin mo- Cell Mol Life Sci – 19. Janson LW, Sellers JR, Taylor DL (1992) Actin-binding proteins regulate the work tors. 69(13):2261 2277. fi Methods Enzymol performed by myosin II motors on single actin filaments. Cell Motil Cytoskeleton 28. Pardee JD, Spudich JA (1982) Puri cation of muscle actin. 85(Pt B): 164–181. 22(4):274–280. 29. Way M, Gooch J, Pope B, Weeds AG (1989) Expression of human plasma gelsolin in 20. Fenn WO (1923) A quantitative comparison between the energy liberated and the Escherichia coli and dissection of actin binding sites by segmental deletion muta- work performed by the isolation sartorius muscle of the frog. J Physiol 58(2/3): genesis. J Cell Biol 109(2):593–605. 0175–0203. 30. Le Clainche C, Carlier MF (2004) Actin-based motility assay. Curr Protoc Cell Biol 21. Veigel C, Molloy JE, Schmitz S, Kendrick-Jones J (2003) Load-dependent kinetics of Chapter 12:Unit 12.7. force production by smooth muscle myosin measured with optical tweezers. Nat Cell 31. Feramisco JR, Burridge K (1980) A rapid purification of alpha-actinin, filamin, and Biol – 5(11):980 986. a 130,000-dalton protein from smooth muscle. J Biol Chem 255(3):1194–1199. 22. Coureux PD, Sweeney HL, Houdusse A (2004) Three myosin V structures delineate 32. Trybus KM (2000) Biochemical studies of myosin. Methods 22(4):327–335. EMBO J – essential features of chemo-mechanical transduction. 23(23):4527 4537. 33. Kron SJ, Toyoshima YY, Uyeda TQ, Spudich JA (1991) Assays for actin sliding move- 23. Cuda G, Fananapazir L, Zhu WS, Sellers JR, Epstein ND (1993) Skeletal muscle ex- ment over myosin-coated surfaces. Methods Enzymol 196:399–416. pression and abnormal function of beta-myosin in hypertrophic cardiomyopathy. 34. Bing W, Knott A, Marston SB (2000) A simple method for measuring the relative force J Clin Invest 91(6):2861–2865. exerted by myosin on actin filaments in the in vitro motility assay: Evidence that tropo- 24. Wynne JBE (1992) The cardiomyopathies and myocarditis: Toxic, chemical and phys- myosin and increase force in single thin filaments. Biochem J 350(Pt 3):693–699. ical damage to the heart. Braunwald’s Heart Disease: A Textbook of Cardiovascular 35. Berg-Sorensen K, Flyvbjerg H (2004) Power spectrum analysis for optical tweezers. Rev Medicine (Saunders, Philadelphia), pp 1394–1450. Sci Instrum 75(3):594–612.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1309493110 Sommese et al. Downloaded by guest on September 26, 2021