Standard chemicals, anti-His antibody, and TRITC-phalloidin were purchased from Sigma; restriction enzymes, polymerases and DNA-modifying enzymes were purchased from MBI-Fermentas and Roche Applied Sciences. The 2’-(3’-)-O-(N′-Methylanthraniloyl) derivatives of ATP (mantATP) and ADP (mantADP) were purchased from Jena Bioscience.

Expression plasmids pDXA-MyoB-S332E-F387Y-2R and pDXA-MyoE-S336E-Y388F-2R, which encode the constitutively active motor domain constructs of Dd myosion-1B and Dd myosion-1E harboring mutations F387Y and Y388F, respectively, fused to an artificial lever arm (2R) and a C-terminal His₈-tag, were generated by PCR using mutagenesis primers 5′-GAT​TTC​AAA​ACC​ATA​AAT​ATC​TAA​AAT​ACC-3′ for introducing mutation F387Y in the motor domain of myosin-1B and 5′-CTC​AAA​ACC​AAA​GAT​ATC​AAG-3′ for introducing mutation Y388F in the motor domain of myosin-1E. E. coli strain XL1Blue (Stratagene, Heidelberg) was used for amplification of the plasmids. Myosin constructs were confirmed by sequencing. Wild-type and mutant constructs were produced in Dictyostelium discoideum and purified as described (Fujita-Becker et al., 2005; Dürrwang et al., 2006; Tsiavaliaris et al., 2008). Chicken skeletal actin was purified as described (Pardee and Aspudich, 1982). Pyrene-labeled actin was prepared from skeletal actin as described (Criddle et al., 1985).

Steady-state ATPase activity measurements were performed at 25°C in buffer containing 25 mM 2-(4-(2-Hydroxyethyl)-1-piperazinyl)-ethansulfonsäure (HEPES) pH = 7.3, 25 mM KCl, 5 mM MgCl₂, 1 mM Dithiothreitol and 1 mM ATP using the NADH-coupled assay (Diensthuber et al., 2015). Unless otherwise stated, transient kinetic measurements were performed at 20°C in experimental buffer containing 20 mM 3-(N-morpholino)propanesulfonic acid (MOPS) pH = 7.0, 100 mM KCl, 5 mM MgCl₂ and 1 mM DTT using a Hi-tech Scientific SF-61DX double-mixing stopped-flow system (TgK Scientific Limited, Bradford on Avon, U.K.). Data were analyzed according to the procedures and kinetic models described (Bagshaw et al., 1974; Millar and Geeves, 1983; Cremo and Geeves, 1998; Batra et al., 1999; Furch et al., 1999). Data for wild-type myosins were depicted from Dürrwang et al., 2006; Tsiavaliaris et al., 2008, respectively, and are listed in the tables as assigned, if not otherwise stated. In vitro motility assays were performed at 30 °C using an Olympus IX81 inverted fluorescence microscope as described (Taft et al., 2008). Penta∙His Antibody (Qiagen) was used for the specific attachment of wild-type and mutant myosin constructs on nitrocellulose coated coverslips. Actin filament tracking was performed using DiaTrack 3.05 software. Average actin sliding velocities were obtained from Gaussian fits to the velocity distributions using Origin 2022b software (OriginLab, Northampton, MA, USA). Landing assays were performed as described (Taft et al., 2008).

Mutation Y388F was introduced in the X-ray crystal structure of the Myo1E motor domain (pdb: 1LKX) using the Schrödinger Suite (Schrödinger Inc.), selecting the rotamers with the lowest sterical hindrance, and energy-minimization of the entire model was performed with MacroModel (Schrödinger Inc.), CNSsolve 1.2 (Brunger, 2007), and the OPLS3 force field. Molecular dynamics simulation of Myo1E^Y388F was carried out with Gromacs 4.0 (Hess et al., 2008) and OPLS (Optimized Potentials for Liquid Simulations) all-atom force field. The myosin was solvated with the TIP3P explicit water model and neutralized by addition of sodium counter ions as described (Preller et al., 2011). MD simulations were performed in a NpT ensemble (300 K, 1 bar) using Berendsen temperature coupling and Parrinello-Rahman pressure coupling. The particle-mesh Ewald method (Darden et al., 1993) was used for long-range electrostatic interactions. Short-range van der Waals and coulomb forces were treated with 12 Å cutoffs. A 2 fs time step was used during the production runs, and all bond lengths were constrained with the LINCS algorithm (Hess et al., 1997). The coordinates were optimized with the conjugate gradient algorithm to a final force of <10 kJ·mol^-1·nm^-1 after energy minimization with the steepest descent algorithm to a force of 1,000 kJ·mol^-1·nm^-1. The solvent molecules were equilibrated for 100 ps. A 4 ns equilibration of the entire system was performed to reach a plateau of the root mean square deviation of the backbone atoms. Production runs were performed for 100 ns? Myo1B homology models were generated with Modeller (Šali and Blundell, 1993) using the crystal structure of Myo1E motor domain (pdb: 1LKX) as template. Best models were selected according to the Modeller objective function and the discrete optimized protein energy score (DOPE). Images were generated with Pymol (DeLano scientific).

L ρ = Z ∙ 1 − e − ρ ρ 0 n (1) d u t y   r a t i o = τ strong τ total ≈ 1 k − AD 1 k cat (2)

The ATPase activities of the wild-type myosin motor domain constructs Myo1B^wt and Myo1E^wt and the corresponding switch-2 mutants Myo1B^F387Y, Myo1E^Y388F, all carrying the S-to-E mutation at the TEDS site, which mimics the phosphorylation state and transforms the myosins into active, ATPase competent motors (Fujita-Becker et al., 2005; Dürrwang et al., 2006) were examined in the absence and presence of F-actin. Data for wild-type myosins were depicted from (Dürrwang et al., 2006; Tsiavaliaris et al., 2008) if not otherwise stated. Hyperbolic fit of the actin-dependence of the rate of ATP turnover according to Michaelis-Menten (Figure 1) yielded the basal ATPase activity (k _basal), the apparent equilibrium constant of half-maximal activation of maximum ATP turnover (K _app ), and actin-activated ATPase activity for saturating actin concentrations ( k _cat ). The steady-state parameters are summarized in Table 1. The mutants displayed reduced k _basal rates compared to the wild-types, while K _app remained largely unaffected. In the case of Myo1B^F387Y, the actin-activated ATPase activity increased almost linear, which allows only rough estimates of the apparent equilibrium constant (K _app ) and the maximum ATP turnover rate ( k _cat ). Myo1B^F387Y shows a 3-fold reduction in both, k _cat and the catalytic efficiency k _cat /K _app . Latter is a measure of how effectively actin activates the ATPase reaction as defined by the 2^nd order rate constant of actin binding in the presence of ATP. These parameters were unaffected in Myo1E^Y388F. Since we observed different impact of the mutations on the catalytic activity of the myosins, we extended the analysis to transient kinetic experiments and focused our investigations on those steps of the ATPase cycle that determine the occupation and lifetime of the strongly vs. weakly actin-bound states of the myosins.

First, we studied nucleotide interactions according to Figure 2 applying the kinetic models as described (Geeves et al., 1984; Taylor, 1991; Franz et al., 2021). Nucleotide binding in the absence of actin was mainly unaffected by the mutations (Table 2). However, both mutants displayed accelerated ADP release rates (k _-D, Figure 3) and two to three-fold weaker affinities for ADP than the corresponding wild-types (K_D; Table 2). The Y-to-F mutation in Myo1E^Y388F accelerated both rates of the two-step ADP release (Dürrwang et al., 2006) by up to four-fold, whereas mutant and wild-type Myo1B displayed single-step ADP release kinetics that differed by approx. two-fold. We interpret the accelerated ADP release as the major contributor of the reduced ADP affinity. In the presence of actin, equilibrium and rate constants of the interaction with ATP, defined by K ₁ and k ₊₂ (Figure 2), were differently affected in the mutants (Table 2). Myo1B^F387Y displayed five-fold slower ATP-binding (K ₁ k ₊₂ ) and three-fold lower ATP affinity of the A·M state. Myo1E^Y388F showed wild-type like ATP-binding behavior and a higher ATP-affinity of the A·M state than the wild-type. The tyrosine appears to disturb high affinity ATP binding by affecting the isomerization of the A·M·T state to the A-M*·T state. Pronounced changes were also observed for the ADP affinity of the A·M state (K _AD ) determined from the inhibition of the ATP-induced dissociation of the actomyosin complex with increasing ADP concentrations (Figure 4). Here, it is important to highlight the monophasic (Figures 4A, D) vs. biphasic actomyosin dissociation kinetics (Figures 4B, C) between wild-type and mutants. Monophasic dissociation kinetics are typical for myosins with a low affinity for ADP (Tsiavaliaris et al., 2002). A biphasic actomyosin dissociation reaction, in which the amplitude of the fast phase decreases and the amplitude of the slow phase increases with excess ADP, is indicative for a highly favorable A·M·D state, where ADP and ATP compete for the same binding site (Batra et al., 1999; Dürrwang et al., 2006). For both single and biphasic dissociation reactions, the ADP dependence of the rate (Figures 4A, D) and amplitude (Figures 4B, C) could be described by hyperbolic functions yielding K_AD values listed in Table 2. Myo1B^F387Y displayed a five-fold higher affinity for ADP than Myo1B^wt, whereas the ADP affinity of the actin-bound state of Myo1E^Y388F (A·M·D) was more than 6-fold decreased compared to wild-type. Thus, the F-to-Y substitution in Myo1B^F387Y appears to strengthen ADP binding and concomitantly prolong the strongly actin-bound states of the motor, while the Y-to-F substitution in Myo1E^Y388F induces the opposite effects. This altered ADP affinity is also reflected in the ADP release kinetics ( k _-AD , Figure 5), which were almost 2-fold decelerated for Myo1B^F387Y and more than six-fold accelerated for Myo1E^Y388F compared to the wild-types (Table 2). The mutations also displayed an inverse effect of the actin affinity of the ADP-bound states. Myo1B^F387Y displayed a more than 30-fold higher actin affinity in complex with ADP (K _DA ) compared to the wild-type, whereas Myo1E^Y388F showed a more than six-fold reduced actin affinity (Table 3). These prominent changes suggest an important role of the tyrosine for the mechano-chemical coupling mechanism of the class-1 myosins, which we investigated further.

It is well known that k _-AD is the major determining parameter of the duty ratio (Mikhailenko et al., 2008; Taft et al., 2008; Diensthuber et al., 2015). In case that k _-AD is of similar order as k _cat , the ADP release can limit the overall ATPase, which is a discernible characteristic of processive motors and motors made for tension, such as Mmmyosin-5a, Ddmyosin-5b, Ddmyosin-1E, Hsmyosin-7a, or Ntmyosin-11 (Dürrwang et al., 2006; Watanabe et al., 2006; Sakamoto et al., 2008; Taft et al., 2008; Diensthuber et al., 2015). Contrary, the rate-limiting step of low duty ratio myosins is commonly the actin-accelerated P_i-release, which precedes the fast dissociation of Mg²⁺ADP (Takagi et al., 2004). Previously, we and others have reported the existence of an equilibrium between magnesium free (A·M·D) and magnesium bound actomyosin-ADP (A·M·Mg²⁺D) states in high-duty ratio myosins (De La Cruz et al., 2000; Hannemann et al., 2005; Rosenfeld et al., 2005; Taft et al., 2008; Jacobs et al., 2011) through which product dissociation can occur either sequentially, where Mg²⁺ is released prior to ADP (Rosenfeld et al., 2005) or together with ADP (Chizhov et al., 2013). The preference for either path depends on free Mg²⁺-ions, which shift the equilibrium towards the simultaneous release of Mg²⁺ and ADP (Chizhov et al., 2013). Therefore, the fraction of time the myosin remains strongly attached to actin can be affected by free Mg²⁺-ions. For high duty ratio myosins we have shown that elevated, physiologically relevant free Mg²⁺ concentrations can inhibit the ADP release to such an extent that Mg²⁺ADP dissociation from the actin-bound states becomes the rate-limiting step, strongly influencing the motile properties and the duty ratio of these motors (Fujita-Becker et al., 2005; Dürrwang et al., 2006; Taft et al., 2008; Nagy et al., 2010; Diensthuber et al., 2015; Amrute-Nayak et al., 2019). The low duty ratio myosin Myo1B^wt (Tsiavaliaris et al., 2008) or skeletal class-2 myosin isoforms do not display such Mg²⁺-dependence of the ADP-release kinetics and motor activity (Taft et al., 2008). To examine whether the myosin isoform-specific sensitivity towards free Mg²⁺-ions was affected by the mutation, we measured the rate of ADP release from the actin-bound states using the fluorescent analogue mantADP. By displacing the bound mantADP with excess ADP in the presence of increasing concentrations of free Mg²⁺-ions at constant ionic strength (Figure 5), we obtained the rates of ADP release from single exponential fits of the fluorescence decays (Figures 5A, C). Interestingly, Myo1B^F387Y displayed ADP release rates (k _-AD) that were dependent on free Mg²⁺, contrary to the wild-type, which did not show such a behavior (Figure 5B). The rates declined hyperbolically from initially 400 ± 25 s^-1 (at 0.05 mM free Mg²⁺) to 120 ± 25 s^-1 (at 10 mM free Mg²⁺) with an apparent inhibition constant K_i ^Mg of 0.45 ± 0.3 mM. The Myo1E constructs displayed exactly the opposite behavior: Myo1E^wt showed a Mg²⁺-dependence of k _-AD as previously published (Dürrwang et al., 2006), whereas Myo1E^Y388F displayed ADP release rates that were independent of free Mg²⁺-ions (Figure 5D) yielding values comparable to those of Myo1B^wt (Figure 5B). All other experimentally determined kinetic parameters listed in the tables showed no or only minor changes. We therefore omitted graphical representation of the data. In summary, the tyrosine in switch-2 of class-1 myosins appears to play a determining role for Mg²⁺ADP release, which could also affect the motile properties and duty ratio of the myosins.

To investigate this, we generated motor domain constructs with artificial lever arms and performed in vitro motility assays as previously described (Taft et al., 2008). The experiments were performed under constant ionic strength and revealed an inhibitory effect of free Mg²⁺ on the actin sliding velocity of Myo1E^wt and Myo1B^F387Y but not on that of Myo1B^wt and Myo1E^Y388F (Figures 6A, C). We note that for Myo1E^wt, the rate constant for ADP release ( k _-AD ) at excess free Mg²⁺-concentrations is similar to k _cat and thus the rate-limiting parameter of the ATPase. For the mutant Myo1E^Y388F, the ADP release is not a rate limiting factor of the ATPase cycle, since it proceeds by more than one order of magnitude faster than the steady-state ATP turnover ( k _cat ). This hints for a low duty ratio. This kinetic data reveal that the tyrosine is an important residue in switch-2 that controls Mg²⁺ADP release and potentially the weak-to-strong actin binding interactions. The thermodynamic coupling constants (K _AD /K_D; K _DA /K _A ) and the kinetic coupling constant ( k- _AD / k _-D) of nucleotide and actin binding are valuable parameters related to weak-to-strong actin binding transitions (Bloemink and Geeves, 2011). They report how effectively actin can displace ADP, providing predictions of the duty ratio of a myosin. High duty ratio myosins display low coupling constants that tend to approximate unity or acquire values below 1, since the displacement of ADP from myosin by actin is slowed down, often rate limiting the ATPase cycle, which contributes to prolonged population of the strongly actin bound states. Myo1E^wt displays typical coupling constants (K _AD /K _D = 1.7 ± 0.3; K _DA /K _A = 2.0 ± 0.6; k- _AD / k _-D1 = 4.4 ± 0.7; k- _AD / k _-D2 = 15.8 ± 3.1; Table 2, 3) of a high duty ratio motor, whereas the coupling constants of Myo1B^wt (K _AD /K _D = 118 ± 32; K _DA /K _A = 126 ± 36; k- _AD / k _-D = 333 ± 29; Table 2, 3) resemble those of fast motors with a low duty ratio characterized by an effective actin-stimulated ADP release (Bloemink and Geeves, 2011). Notably, the mutations almost reversed the coupling parameters. The Y-to-F substitution in Myo1E^Y388F led to an increase of the coupling constants by two- to seven-fold (K _AD /K _D = 3.1 ± 0.7; K _DA /K _A = 15.3 ± 5.0; k- _AD / k _-D1 = 6.8 ± 0.8; k- _AD / k _-D2 = 34.8 ± 4.9; Table 2, 3), which indicates a decrease of the duty ratio. The F-to-Y substitution in Myo1B^F387Y caused a decrease of all coupling constants (K _AD /K _D = 9 ± 2; K _DA /K _A = 8 ± 3; k- _AD / k _-D = 115 ± 11; Table 2, 3), indicating that the duty ratio is also affected. Using Eq. 2 (Watanabe et al., 2006; Ito et al., 2007), we calculated the duty ratios (Table 2) revealing for the mutant Myo1E^Y388F indeed a lower duty ratio (0.07 ± 0.01) compared to Myo1E^wt (0.55 ± 0.13). For Myo1B^wt we obtained a low duty ratio (0.01 ± 0.01), while the calculation of the duty ratio of Myo1B^F387Y predicts a value > 0.01.

Although the duty ratios calculated from the kinetic data correlate well with the coupling parameters, the values appear to be underestimated, particularly for the Myo1B constructs. This is due to the high uncertainty in determining k _cat . To experimentally validate the effect of the mutations on the duty ratio, we performed landing assays at two boundary conditions (0.25 mM and 5 mM free Mg²⁺ concentrations) to account for the Mg²⁺-effect observed in the kinetics and motility assays (Figure 5; Figures 6A, C). The number of motile landing events decreased with decreasing motor densities (Figures 6B, D). The dependency of the landing rates on motor density was best fit by Eq. 1 as described (Hancock and Howard, 1998). From the reciprocal of n, the duty ratio could be calculated (Sumiyoshi et al., 2007), revealing that the duty ratio of Myo1E^wt increased from 0.38 ± 0.11 at 0.25 mM free Mg²⁺ to 0.91 ± 0.07 at 5 mM free Mg²⁺, whereas the duty ratio of Myo1E^Y388F dropped to 0.14 ± 0.02 at 0.25 mM free Mg²⁺ and 0.2 ± 0.04 at 5 mM free Mg²⁺, respectively. This demonstrates the loss of the Mg²⁺-sensitivity of the motor (Table 4) consistent with Mg²⁺-insensitive ADP-release kinetics (Figure 5) and Mg²⁺-insensitive motor activity (Figures 6A, C). Myo1B^wt was significantly more difficult to handle in the landing assays. At densities below 5000 motor molecules µm^-2, we observed almost no landing events, which speaks for a low duty ratio of the motor. Therefore, there are no data available for Myo1B^wt. Myo1B^F387Y displayed landing events that were dependent on both, free Mg²⁺-ions and motor density, yielding duty ratios of 0.14 ± 0.02 at 0.25 mM Mg²⁺ and 0.5 ± 0.1 at 5 mM free Mg²⁺. The estimation of the duty ratio for Myo1B^F387Y should be taken with care, since the number of moving actin filaments at low motor densities became increasingly less. However, the data clearly demonstrate that the introduction of the tyrosine in Myo1B significantly increases the duty ratio of the myosin.

Finally, we used the available X-ray structure of Myo1E (pdb: 1LKX) (Kollmar et al., 2002) and generated energy minimized homology models of Myo1E^Y388F, which we applied to molecular dynamics simulations to obtain insights into potential conformational changes induced by the mutation (Figure 7). The superimposed structures in complex with ADP-VO₄ show Y388 in switch-2 of wild-type Myo1E^wt to form a 2.9 Å hydrogen bond with the β7-sheet residue L183 of the transducer (Figure 7A). This conformation appears to stabilize the Mg²⁺-ion in the binding pocket of the wild-type myosin. The structural model of Myo1E^Y388F predicts the loss of this hydrogen bond (Figure 7B). As a consequence switch-2 is slightly shifted, which enables residue D386 to strengthen the hydrogen bond with P-loop residue T108. This in turn weakens the T108-mediated coordination of the Mg²⁺-ion. This conformational shift of switch-2 appears to additionally affect the Mg²⁺-coordination mediated via switch-1 residue S158. Consequently, these small conformational rearrangements of switch-2 appear to disrupt the coordination sphere of Mg²⁺, which is apparently associated with an accelerated Mg²⁺-ADP release and the inability of the Myo1E^Y388F to sense free Mg²⁺-ions. For Myo1B we generated energy-minimized homology models (Figure 7B) to show that a hydrogen bond between Y387 in the mutant and L183 can be formed like in Myo1E^wt, which is absent in the wild-type (Figure 7B, blue residues). This tyrosine-mediated hydrogen bond may be critical for stabilizing Mg²⁺ and ADP, which could explain the Mg²⁺-dependent suppression of the ADP-release from acto·Myo1B^F387Y and the prolonged strongly actin-bound states responsible for the higher duty ratio. These structural considerations are supported by X-ray and cryoEM structures of Myo5, revealing that the tyrosine in switch-2 is important for stabilizing the rigor conformation in high duty ratio motors via the interaction with leucine in the ß7 sheet (Coureux et al., 2004; Pospich et al., 2021). Mechanistically, the effect of free Mg²⁺ can be interpreted as previously described for Myo5b (Rosenfeld et al., 2005; Chizhov et al., 2013). The dissociation of the Mg²⁺ can occur prior or in complex with ADP. Thus, increased concentrations of free Mg²⁺-ions can slow down the release of the nucleotide. Our structural models attribute this to the altered switch-2 conformation induced by the mutations affecting the ADP affinity of the actomyosin complex. The tyrosine stabilizes the Mg²⁺·ADP state, while the phenylalanine causes accelerated release rates. However, the F-to-Y mutation in MyoB^F387Y did not alter the ADP dissociation from the acto-MyoB^F387Y complex to rate-limit the overall ATPase cycle time, as in the case of MyoBF387Y. Obviously, other structural elements including switch-1, loop-1, W-helix, and loop-2, all of which have been related to product release (Bobkov et al., 1996; Sweeney et al., 1998; Murphy and Spudich, 1999; Clark et al., 2005; Bloemink et al., 2014; Franz et al., 2021), contribute additionally in fine-tuning the ADP release kinetics. With respect to switch-2, our data suggest the tyrosine in class-1 myosins suppresses the actin-induced acceleration of ADP release by strengthening the coordination of Mg²⁺ and ADP through the stabilization of the conformation of the surrounding active site elements.

The presence of the tyrosine in the switch-2 consensus sequence of myosins is not strictly indicative for a high duty ratio. For example, mammalian class-I myosins contain a tyrosine and display under unloaded conditions characteristics of a low duty ratio motor (Greenberg and Ostap, 2013), whereas under tension (Greenberg et al., 2012; Pyrpassopoulos et al., 2016) and/or Ca²⁺-dependent binding of calmodulin to the neck region (Lewis et al., 2012), they exhibit altered rates and equilibrium constants of the transitions in the ATPase cycle, which can increase the duty ratio (Laakso et al., 2008). Interestingly, the class-VI myosin members are high duty ratio motors, although they possess an alanine at the Y388-equivalent position (De la Cruz et al., 2001). Furthermore, the exchange of the tyrosine in switch-2 of mammalian myosin-5a by an alanine resulted in an increased processive behaviour of the motor and strengthened the free Mg²⁺-dependent suppression of ADP release from actomyosin, however, at reduced speed (Nagy et al., 2010). These findings suggest that natural variations of structural elements besides switch-2, including those mediating actin interactions, such as loop-2, loop-4, the CM-loop, activation loop, helix-loop-helix (Goodson et al., 1999; Onishi et al., 2006; Diensthuber et al., 2015) have a dominant role in the regulation of the duty ratio, since they couple actin-interactions to product release as exemplary shown for class-V and class-I (Yengo and Sweeney, 2004; Lieto-Trivedi et al., 2007).

Here, we provide kinetic and molecular insights into the product release mechanisms of two functionally distinct class-1 myosins and reveal that switch-2 dictates the duty ratio of the motors and their ability to perform rapid movement, processive motion, or gating function. The Mg²⁺-sensitivity of the class-1 myosins appears to be an important feature in cellular processes that require a switching between fast contractility and slow tension bearing, such as during endocytosis, where the myosins act as motorized cross-linkers between the membrane and actin cytoskeleton systems to provide the contractile forces to accomplish actin and membrane remodelling (Manenschijn et al., 2019). Additionally, class-1 myosins can act as tension-sensitive tethers or even transporters (McIntosh and Ostap, 2016). Thus, it would be interesting to analyse how switch-2 in conjunction with other structural elements implicated in regulating Mg²⁺ADP release define the duty ratio of individual myosin-1 isoforms to perform distinct and multiple types of molecular functions.