Search

Intracellular vesicular transport through the cell's physically challenging, three dimensional (3D), actin cytoskeleton requires teams of Myosin Va (MyoVa) motors to navigate vastly different structures, such as the dense cortical actin network and bundled actin filaments in filopodia. To understand how MyoVa cargo transport is achieved through and along these cytoskeletal structures, we created 3 µm thick biomimetic actin networks in vitro, using actin organizing/binding proteins (fascin, Arp2/3, WASP, gelsolin) and imaged the networks using 3D STORM microscopy.

Myosin Va (MyoVa) motor teams collectively move intracellular cargo through a complex, three dimensional (3D), actin meshwork. To address how MyoVa motors navigate this physical challenge, we imaged fluorescently labeled 350nm synthetic lipid vesicles, bound with ∼10 MyoVa HMM motors, moving through 3D actin filament intersections strung between 3µm beads. Intersections were formed by filaments being ∼90 degrees to one another and separated by 30-250nm. As the vesicle surface contacted the intersecting actin filaments, two regions of potential motor engagement were created on the vesicle surface. Geometric constraints limited the engaged motor number at each contact point to no more than 3, as confirmed by laser trap stall force measurements. These points of vesicle-actin filament contact resulted in an effective tug of war between two motor teams that were physically separated on the vesicle surface. The vesicle's directional outcome at an intersection was visualized using high spatial and temporal 3D super resolution imaging. For vesicles that physically contacted both actin filaments at the intersection, 55% of vesicles continued straight through the intersection, 35% switched filaments, while only 10% terminated. These outcomes were successfully simulated by a model in which the directional outcome was dependent on the interfilament gap, the vesicle's azimuthal approach angle to the intersection, the stochastic attachment and detachment of motors within a team (i.e. a team's force generation), and the Brownian motion of the vesicle while attached to the actin filaments. All of these physical factors contributed to which motor ensemble won the tug of war at the intersection and thus the vesicle's directional outcome. This simplified biomimetic system and resultant data provide an experimental framework for understanding how MyoVa motor teams navigate their cargo through evermore complex 3D actin networks that exist within cells.

A cluster of basic residues on the surface of actin (K326, K328 and R147) are needed for F-actin to bind to acidic residues on tropomyosin in the blocked- and closed-states of the muscle thin filament (Li et al., 2011; Lehman et al., 2013; Orzechowski et al., 2013; Fischer et al., 2016). However, the binding of tropomyosin to actin is necessarily weak so that regulatory transitions across actin can proceed at low-energy cost. We propose that negatively charged actin residue D292 lying adjacent to residues K326, K328 and R147 is required to temper the binding of actin to tropomyosin allowing regulatory-switching.

Myosin II has two heads that are joined together by an alpha-helical coiled-coil rod, which can separate in the region adjacent to the head-rod junction (Trybus, K. M. 1994. J. Biol. Chem. 269:20819-20822). To test whether this flexibility at the head-rod junction is important for the mechanical performance of myosin, we used the optical trap to measure the unitary displacements of heavy meromyosin constructs in which a stable coiled-coil sequence derived from the leucine zipper was introduced into the myosin rod. The zipper was positioned either immediately after the heads (0-hep zip) or following 15 heptads of native sequence (15-hep zip). The unitary displacement (d) decreased from d = 9.7 +/- 0.6 nm for wild-type heavy meromyosin (WT HMM) to d = 0.1 +/- 0.3 nm for the 0-hep zip construct (mean +/- SE). Native values were restored in the 15-hep zip construct (d = 7.5 +/- 0.7 nm). We conclude that flexibility at the myosin head-rod junction, which is provided by an unstable coiled-coil region, is essential for optimal mechanical performance.

The double-headed myosin V molecular motor carries intracellular cargo processively along actin tracks in a hand-over-hand manner. To test this hypothesis at the molecular level, we observed single myosin V molecules that were differentially labeled with quantum dots having different emission spectra so that the position of each head could be identified with approximately 6-nm resolution in a total internal reflectance microscope. With this approach, the individual heads of a single myosin V molecule were observed taking 72-nm steps as they alternated positions on the actin filament during processive movement. In addition, the heads were separated by 36 nm during pauses in motion, suggesting attachment to actin along its helical repeat. The 36-nm interhead spacing, the 72-nm step size, and the observation that heads alternate between leading and trailing positions on actin are obvious predictions of the hand-over-hand model, thus confirming myosin V's mode of walking along an actin filament.

Kinesin and dynein are molecular motors that move in opposite directions on a microtubule. They often act on the same cargo, causing the cargo to frequently switch direction. Whether this back-and-forth motion results from a coordinating complex or from a tug-of-war between the two motors is currently unknown. We have applied single molecule fluorescence to determine that they are undergoing a synergistic tug-of-war. By synergistic, we mean that the combination of the two motors is able to bypass roadblocks along the microtubule. When a motor is driven by kinesin, it approaches a roadblock (either other microtubules or microtubule-associated-proteins), is forced to turn around, relying on dynein. After a few tries, the dynein appears to step sideways onto another protofilament, at which point, when the kinesin takes over, it is able to bypass the roadblock. We also tested the tug-of-war model inside of a cell by using an in vivo optical trap. The in vivo optical trap is able to measure the stall forces in a viscoelastic media, which is present inside of a cell. By comparing directional stall forces in vivo and in vitro, we found that when cargo is going in the positive microtubule direction, kinesin and dynein are pulling, with the dynein walking backwards. The net stall force equals the stall force of kinesin (≈ 7 pN) minus the stall forces of the number of dyneins (1.1 pN x ND, where ND, = 0 to 6). When moving in the negative microtubule direction, the stall force is just equal to a multiple of dynein's stall force (1.1 pN x ND), implying that kinesin has fallen off the microtubule.

Pancreatic beta-cells secrete insulin in response to elevated blood glucose. After synthesis, insulin granules are transported initially by kinesin-I, then transferred to MyoVa within the actin cortex. Granule trafficking models propose that MyoVa tethers granules within the actin cortex, which acts as a secretion barrier. Upon glucose stimulation, actin cytoskeletal reorganization allows granule diffusion or transport by MyoVa to the plasma membrane. To determine how MyoVa and actin contribute to granule trafficking, we tracked eGFP-labeled granules in INS-1 cells (resolution: 23nm/50ms), analyzing their movement by mean square displacement. At rest, granules display 3 modes of motion: paused, diffusive, and directed. Glucose stimulation increases the percentage of granules undergoing MyoVa and/or kinesin-directed motion with little change in the dynamic movements of peripheral actin cytostructures. However, Jasplakinolide treatment, which changes actin's depolymerization rate, inhibits actin cytostructural dynamics and reduces the percentage of granules with directed movement. Therefore, a dynamic actin cytoskeleton is required for efficient granule transport. By introducing Qdot-labeled MyoVa into cells, we will track granules and the steps of associated MyoVa to distinguish between MyoVa's proposed roles as granule transporter or tether.

Myosin II is a two-headed motor protein with an elongated α-helical tail. The motor domain of each head interacts with actin to convert the chemical energy of ATP into movement. Myosin II activity in muscle and nonmuscle cells is switched off by intramolecular interaction between its two heads, which inhibits their activity. This has been shown by EM and image processing of myosin filaments and isolated myosin molecules. In switched-off single molecules, the myosin tail folds into three segments, with the interacting heads folded back on the tail. The interacting-heads motif is highly conserved, being found in vertebrate and invertebrate smooth and striated muscle and in nonmuscle cells. We are investigating its evolutionary origins by EM imaging of isolated myosin molecules in the off-state. In previous work we found that the motif was present as far back as Cnidaria (sea anemones), the earliest animals with muscles. Here we have studied additional animal and non-animal species. At high (0.5 M) salt, all the myosin IIs showed the typical appearance of an extended tail and non-interacting heads. At low salt (0.15 M), under relaxing conditions (MgATP), insect indirect flight and embryonic skeletal muscle myosins showed a folded tail and similar head-head interactions to other muscles. Three species of primitive, non-animal myosins gave differing results. Acanthamoeba (reported previously) and Schizosaccharomyces pombe showed extended tails and no head-head interactions. The tails of these two myosins were approximately 30-40% shorter than the animal myosins, possibly accounting for their inability to fold. In contrast, Dictyostelium, with a tail 10% longer than animal myosin, showed head-head interactions similar to animal myosin; however, the conformation of the folded tail was different. These results suggest that head-head interaction arose before the evolution of animals.

Myosin Va (myoVa) is a processive, actin-based motor involved in intracellular vesicular transport. Although capable of single myoVa transport in vitro, multiple myoVa motors transport intracellular vesicles, composed of phospholipid outer membranes. To determine how membrane fluidity affects the collective transport capacity of a myoVa ensemble, we synthesized liposomes with either fluid, DOPC or rigid, DPPC phospholipids. Varying number of myoVa motors were attached, and the liposome velocity on actin tracks observed by TIRF microscopy. Fluid DOPC liposomes with physiologically relevant myoVa surface densities (8 motors/200nm liposome) move at 498±228nm/s (n=282), faster (p

The fungal toxin cytochalasin D (CD) interferes with the normal dynamics of the actin cytoskeleton by binding to the barbed end of actin filaments. Despite its widespread use as a tool for studying actin-mediated processes, the exact location and nature of its binding to actin have not been previously determined. Here we describe two crystal structures of an expressed monomeric actin in complex with CD: one obtained by soaking preformed actin crystals with CD, and the other obtained by cocrystallization. The binding site for CD, in the hydrophobic cleft between actin subdomains 1 and 3, is the same in the two structures. Polar and hydrophobic contacts play equally important roles in CD binding, and six hydrogen bonds stabilize the actin-CD complex. Many unrelated actin-binding proteins and marine toxins target this cleft and the hydrophobic pocket at the front end of the cleft (viewing actin with subdomain 2 in the upper right corner). CD differs in that it binds to the back half of the cleft. The ability of CD to induce actin dimer formation and actin-catalyzed ATP hydrolysis may be related to its unique binding site and the necessity to fit its bulky macrocycle into this cleft. Contacts with residues lining this cleft appear to be crucial to capping and/or severing. The cocrystallized actin-CD structure also revealed changes in actin conformation. An approximately 6 degrees rotation of the smaller actin domain (subdomains 1 and 2) with respect to the larger domain (subdomains 3 and 4) results in small changes in crystal packing that allow the D-loop to adopt an extended loop structure instead of being disordered, as it is in most crystal structures of actin. We speculate that these changes represent a potential conformation that the actin monomer can adopt on the pathway to polymerization or in the filament.

Conventional kinesin-1 is a processive motor protein that moves unidirectionally on microtubules. We found that when full-length kinesin containing a HIS tag at its C-terminus is bound to an anti-HIS Quantum dot (Qdot), it shows diffusive movement on microtubules in the presence of either ATP or ADP. Diffusive behavior was first described for the depolymerizing kinesin-13, MCAK (Helenius et al., 2006). When bound to a carboxylated Qdot, the same kinesin construct moves processively in the presence of ATP, but does not interact with microtubules in ADP. Further investigation with a truncated construct lacking the last 75 amino acids (kinesin-ΔC) showed both unidirectional and diffusive movement on microtubules in solutions containing a mixture of ADP and ATP. The diffusion constant depends on the concentration of ADP/ATP. When tested in solutions containing only ADP, kinesin-ΔC shows purely diffusive movement. We interpret these data to imply that kinesin-1 diffuses on microtubules when it is in the inactive, folded conformation, and it moves processively when in its active, extended conformation. We speculate that in the folded state, kinesin with bound ADP retains a relatively high binding affinity for microtubules compared to extended kinesin, thus allowing it to diffuse.

The ATP hydrolysis rate and shortening velocity of muscle are load-dependent. At the molecular level, myosin generates force and motion by coupling ATP hydrolysis to lever arm rotation. When a laser trap was used to apply load to single heads of expressed smooth muscle myosin (S1), the ADP release kinetics accelerated with an assistive load and slowed with a resistive load; however, ATP binding was mostly unaffected. To investigate how load is communicated within the motor, a glycine located at the putative fulcrum of the lever arm was mutated to valine (G709V). In the absence of load, stopped-flow and laser trap studies showed that the mutation significantly slowed the rates of ADP release and ATP binding, accounting for the approximately 270-fold decrease in actin sliding velocity. The load dependence of the mutant's ADP release rate was the same as that of wild-type S1 (WT) despite the slower rate. In contrast, load accelerated ATP binding by approximately 20-fold, irrespective of loading direction. Imparting mechanical energy to the mutant motor partially reversed the slowed ATP binding by overcoming the elevated activation energy barrier. These results imply that conformational changes near the conserved G709 are critical for the transmission of mechanochemical information between myosin's active site and lever arm.

In neurons, microtubule motor driven transport is crucial for communication between processes and the cell body. Disruptions in transport are associated with a variety of neurodegenerative diseases. Previous studies implicate phosphorylation of serine 176 in kinesin-1 in the impaired axonal transport associated with Huntington's disease. In isolated squid axoplasm, introduction of pathogenic huntingtin protein activates the kinase, JNK3, which specifically phosphorylates kinesin at S176 (Morfini, Nature Neuroscience, 2009). The mechanism by which S176 modification leads to impaired transport is not very well understood. It is not known whether phosphorylation of kinesin alone is sufficient to cause impaired cargo transport. To investigate the effect of residue 176 on kinesin transport, we use optical trapping and single-molecule fluorescence imaging to study purified kinesin. We employ two constructs, S176A and S176D, truncated at residue 888 to remove the autoinhibition domain, resulting in constitutively active motors. There is no significant difference in the processivity, ATPase, or microtubule binding activity between the phosphomimetic S176D construct and the non-phosphorylatable S176A construct. However, we find that S176D has an attenuated stall force (5pN) compared to S176A (7 pN). Furthermore, polystyrene bead cargos coated with dynein and S176D are transported preferentially in the minus direction in comparison with cargos coated with equivalent concentrations of dynein and S176A. We also perform phosphorylation assays using JNK3 on both 888-truncated and full-length wild-type kinesin in which radiolabeling is used to quantify the percentage of protein phosphorylated in the assay. The pattern of stall force attenuation and directional bias observed for the S176D mutant is also observed in JNK3 phosphorylated samples in which 60-80% of the wild type protein has been phosphorylated. These results show that modification of serine 176 alone is sufficient to alter the behavior of kinesin.

Myosin Va is a processive, actin-based molecular motor that is critical for organelle transport. While transporting intracellular cargo, myosin Va faces significant physical barriers and directional challenges created by the complex actin cytoskeleton, a network of intersecting actin filaments and actin bundles. We have created an in vitro model system of fascin cross-linked unipolar actin bundles. While walking on an actin bundle, a single myosin Va motor switches filaments within the bundle with a high probability (24%). Although a single myosin Va is sufficient to transport cargo in vitro, intracellular cargo transport is driven by multiple motors. To understand the collective behavior of multiple motors, we have linked two myosin Va motors, with only one head of each motor labeled with either a red or green Qdot, via a third Qdot which acted as a cargo. If each motor walks on a different actin filament within the bundle, then the two motors may experience an off-axis load. The velocity and the run length of the 2-motor complex was reduced significantly from that of a single motor, suggesting that the motors interfere with each other's motion. Interestingly, the leading motor takes ∼10% back steps, indicating that it experiences a resistive load from its partner. Both the run length of the complex and the step lifetimes of the motors were correlated to the inter-motor distance, with the run length decreasing and the step lifetimes increasing with greater motor separation. Our data suggest that the two motors step independently when close together. However, when far apart, tension increases in their cargo-linkage, which results in inter-motor mechanical coupling. This study will provide insight into the mechanism of how multiple motors mechanically interact to transport cargo in vivo.

Full length MyoVa (FL-MyoVa) adopts a folded, inhibited conformation at low salt, stabilized by electrostatic head-tail interactions, and a fully extended, active conformation at high salt. TIRF microscopy was used to determine the effects of this equilibrium on processivity (i.e. velocity, run length, and stepping dynamics) of single Quantum dot (Qdot)-labeled FL-MyoVa motors. These values were compared to the constitutively active truncated HMM-MyoVa motor over a range of ionic strengths (25-200mM KCl). Surprisingly, at 25mM KCl 16% of actin-associated FL-MyoVa motors are processive, but with significantly slower velocities and shorter run lengths than HMM-MyoVa. The slower velocities result from FL-MyoVa transitioning between periods of “fast” (Vfast) and “slow” (Vslow

The activity of smooth and non-muscle myosin II is regulated by phosphorylation of the regulatory light chain (RLC) at ser19. The dephosphorylated state of full-length monomeric myosin is characterized by an asymmetric intramolecular head-head interaction that completely inhibits the ATPase activity, accompanied by a hairpin fold of the tail, which prevents filament assembly. The mechanism by which ser19 phosphorylation disrupts these head-head interactions is unknown. Computational modeling (Tama et al., 2005. J. Mol. Biol. 345, 837-854) suggested that formation of the inhibited state is characterized by both torsional and bending motions about the myosin heavy chain (HC) at a location between the RLC and the essential light chain (ELC). Based on this model, placement of the regulatory domain at this locus might alter relative motions between the ELC and the RLC and thereby disrupt the inhibited state. Here we derive an atomic model based on this hypothesis for the phosphorylated state of the smooth muscle myosin light chain domain (LCD). We use a homology model for the structure of the RLC and a largely α-helical structure for the regulatory domain. This model predicts a set of specific interactions between the regulatory domain and both the myosin HC and the ELC. Site directed mutagenesis combined with ATPase and motility assays was used to show that interactions between the highly basic N-terminus of the RLC with acidic residues on helix-A of the ELC are required for phosphorylation to activate smooth muscle myosin. These sites are well conserved in all the myosin II ELC sequences examined despite the lack of known interacting partners and when substituted, are usually substituted with aspartic acid. Supported by NIH grants AR47421, AR53975 and HL38113 and RR12255.

Sub-cellular localization of mRNA is a widely used mechanism to ensure correct spatial and temporal expression of proteins within the cell. A paradigm for localizing mRNAs is the ASH1 transcript in budding yeast which is transported by Myo4p, a class V myosin motor. We previously showed that Myo4p is a single-headed motor that tightly binds the adapter protein She3p. She3p also interacts with the mRNA-binding protein She2p, which recruits two Myo4p/She3p motors. This double-headed motor complex is processive at low salt, but becomes destabilized at physiological ionic strength, suggesting that other factors are required for Myo4p motility. To understand how Myo4p functions in the cell, we increased the complexity of our system by adding labeled mRNA cargo to the motor complex, forming a fully reconstituted messenger ribonucleoprotein complex (mRNP). This mRNP shows robust processivity at physiological ionic strength, thus providing a checkpoint to ensure that only Myo4p motors that are integrated into an mRNP are motile. The ASH1 transcript contains four sequence elements called “zipcodes”, which bind She2p. To understand why ASH1 has multiple zipcodes, we reconstituted ASH1 mRNPs containing varying numbers of zipcodes. We find that transcripts with multiple zipcodes more effectively recruit a paired motor complex for transport. Metal-shadowed images of mRNPs show as many as 8 motors bound to native ASH1 transcripts, suggesting that these particles are likely optimized to move on the actin cables found in the cell. We find that mRNP motility on actin bundles shows dramatic enhancements in both run frequency and run length compared to single actin filaments. Thus, only by building complexity in vitro can one begin to fully understand how motors function in their cellular context.

Smooth muscle myosin (smM) and heavy meromyosin (smHMM) are activated by regulatory light chain (RLC) phosphorylation but the mechanism remains unclear. Dephosphorylated, inactive smHMM assumes a closed conformation with asymmetric intramolecular head-head interactions involving motor domains and the essential light chain (ELC) [Wendt et al., PNAS 98: 4361 (2001)]. The “free head” can bind to actin, but the actin-binding interface of the “blocked head” is involved in interactions with the free head. We report here a 3-D structure for phosphorylated, active smHMM obtained using electron crystallography of 2-D arrays, and an atomic model obtained by fitting using normal mode flexible fitting. Head-head interactions of phosphorylated smHMM resemble those found in the dephosphorylated state, but occur between separate molecules. The interface between heads of phosphorylated smHMM is less extensive and somewhat altered in orientation compared with that of dephosphorylated smHMM. The light chain binding domain of phosphorylated and several dephosphorylated myosin structures show systematic differences. However, the major difference appears to be the relationship between the motor domain and the ELC in a phosphorylated head compared to that of the “blocked head” of dephosphorylated smHMM. We hypothesize that RLC phosphorylation disrupts the inhibited conformation primarily by its effect on the “blocked head” rather than the “free head”. Singly phosphorylated smHMM is not compatible with the closed conformation if the “blocked head” is phosphorylated. The implications of this observation for myosin activation at low levels of phosphorylation in smooth muscle will be discussed. Supported by grants from the NIAMS, NHLBI and NSF-MCB.

Myosin Va is a motor protein responsible for vesicular transport inside eukaryotic cells. Its structure has been well-studied and is known to contain a flexible hinge region approximately half-way between the N-terminal motor domain and C-terminal globular tail. Previous studies have shown two distinct conformations for Myosin Va, one where the hinge remains extended and the protein is active, and one where the hinge allows the protein to fold, allowing auto-inhibition between the globular tail and motor regions. It is assumed, but never shown at the single-molecule level, that Myosin Va actively switches between these conformations at the timescales typical of their runs. Traditional two-dimensional fluorescence tracking techniques cannot adequately capture the conformation change. However, it is possible to track an out-of-focus particle in all three dimensions with a modified form of FIONA microscopy. The point spread function of the out-of-focus particle takes the shape of an Airy disc, with the width of the disc proportional to the distance by which the particle is out of focus. With this technique, the spatial resolution degrades quickly with drift, so we have further modified the microscope with an auto-focus system based on feedback of backscattered signal from a secondary IR laser. We present tracking data of quantum dots bound to full-length Myosin Va with < 5 nm spatial resolution in XY, < 30 nm in Z at a time resolution of 100 ms, which is sufficient to resolve conformation changes.

Full length myosin Va (FL-MyoVa) forms an inhibited, folded conformation at low salt, stabilized by interactions between the globular tails and the heads. High ionic strength disrupts this interaction, resulting in an extended, active processive motor. In vivo, it has been postulated that cargo binding disrupts the folded conformation and activates the motor. It is possible that splice variations in the tail (-B+D+F, melanocyte; +B-D-F, brain) could modify the ability of myosin Va to form the inhibited state. Two FL-MyoVa splice variants and an HMM-MyoVa, with biotin tags for Qdot labeling, were expressed in Sf9 cells. Sedimentation velocity experiments showed similar transitions from the folded-to-extended conformation for the two splice variants as a function of salt. TIRF microscopy was then used to observe processive runs on actin. The velocities of both FL-MyoVa splice variants were similar, and increased 270% (171-460nm/sec) with increasing KCl concentration (25-200mM). In contrast, the velocity of HMM-MyoVa increased by a more modest 50% (381-586nm/sec). The trajectories of the FL-MyoVa and HMM-MyoVa were also strikingly different. Both FL-MyoVa splice variants underwent processive runs that were interrupted by periods during which the motor dwelled at fixed points on the actin filament, presumably in the folded, inhibited state. At lower KCl concentration, FL-MyoVa dwelled approximately half of the total trajectory duration. Increasing ionic strength decreased duration of the dwells. HMM-MyoVa was fully active and maintained continuous processive movement at all KCl concentrations. The slower overall velocities for the FL-MyoVa splice variants, compared to HMM-MyoVa, results from inclusion of the dwell periods. We propose that during a processive run, a single FL-MyoVa can switch between an active and inhibited state without dissociating from actin, and that this phenomenon is independent of splice variations in the tail domain.

The coiled coil rod of smooth muscle myosin is important both for regulation of activity and optimal mechanical performance. Myosin with a phosphorylated light chain is active, while in the inhibited, dephosphorylated state the two heads form an asymmetric intramolecular interaction. The minimal myosin that can attain an “off” state has two heads and 15 heptads of coiled coil rod, a length approximately equal to that of the myosin head. This observation implies that there may be head-rod interactions in the inhibited state. Here we have determined the crystal structure of this region of the rod. Despite being a parallel, coiled coil dimer, the core arrangement is asymmetric. We propose that this asymmetry is wired into its sequence and crucial to its function. The core of the S2 segment is loosely packed in stretches and the two helical segments are locally off-register or staggered relative to one another. Staggered regions are centered on non-canonical core residues. This relative staggering causes three prominent bends in the coiled coil. Significant deviations from two-fold symmetry are observed in our structure, and to a lesser extent in equivalent crystal structures of S2 fragments from cardiac myosin. The larger variations in stagger and bend angles in the rods of smooth versus striated muscle myosins may explain in part why asymmetric head-head interactions are more prevalent in the thick filament regulated myosins.

Processive stepping of myosin V (myoV) on actin has been studied either by tracking the position of the tail, which follows the motion of the molecule as a whole, or by tracking the position of one or both heads. Here we combine these two approaches, and attach a quantum dot (Qdot) to one of the motor domains, and a bead to the tail. Using optical trapping and total internal reflection microscopy, the position of one head and the tail are simultaneously observed as myoV moves processively on an actin filament against increasing load. Our results show that the head (Qdot) moves continually with 72.9±10.3 nm step size, while the tail (bead) moves with a step size of 34.7±8.6 nm. For every two tail steps, the head moves only one step. One of the tail steps takes place concurrently with the head step. Back steps were occasionally observed. Analysis shows that before taking a back step, the head moves 68±11nm while the tail moves 31.9±9.7nm, which suggests that the leading head lands on the 11th actin subunit instead of its normal 13th actin subunit. Interestingly, during a backstep the tail moves −28.6±13.7nm, while the step size distribution for the head shows multiple peaks. This suggests that the head has multiple binding positions along the actin filaments, while the tail has a more defined conformation. Our observation supports a hand-over-hand model for processive movement of myoV, and reveals the cause of the back stepping behavior of myosin V under physiologically relevant loading forces (

Myosin Va (myoVa) and myosin VI (myoVI) are processive molecular motors that transport cargo in opposite directions on actin tracks. Since myoVa and myoVI may colocalize to the same cargo in vivo, these motors may undergo a tug of war. Therefore, we sought to characterize the stepping dynamics of single myoVa and myoVI motors in vitro as they mechanically interact when linked together by a Qdot cargo. Expressed myoVa-HMM with an N-terminal biotin tag were labeled with streptavidin-Qdots (565nm) while expressed dimerized myoVI-HMM were Qdot(655nm)-labeled on an exchanged calmodulin. The effective tug of war on actin filament tracks (25mM KCl, 2mM ATP, 22°C) was observed in TIRF with 6nm resolution, allowing individual steps to be detected. MyoVa won ∼80% of the time and regardless of which motor won, its stepping rate was reduced ∼50% below its unloaded value due to the resistive load of the opposing motor. Interestingly, as the winning motor stepped forwards (myoVa, 73nm; myoVI, 56nm) the opposing motor stepped backwards (myoVa, 68nm; myoVI, 65nm) at the same rate, although myoVI appeared to be dragged at times. Why does myoVa dominate when its stall force is similar to myoVI? Given the probability that both myoVa and myoVI take occasional backsteps and experience a 2-3-fold reduction in stepping rate when winning, we estimate based on optical trapping data (Altman et al., 2004; Kad et al., 2008) that myoVa exerts a 50% greater resistive load compared to myoVI, providing a potential advantage to myoVa. Differences in the length of the myoVa and myoVI constructs could lead to each motor experiencing different vectorial force components, the potential that this may influence the outcome of the tug of war is being investigated.

Long distance cargo transport in budding yeast is carried out not by kinesin, but along actin cables by two non-processive class V myosins, Myo2p and Myo4p. Overexpression of Smy1p, a kinesin-related protein, rescues the temperature sensitive myo2-66 mutant yeast strain, which is defective in Myo2p transport.1 The mechanism by which a kinesin family protein rescues actin-based transport is unknown, but does not require microtubules.2 To address this question, we expressed Smy1p and Myo2p in insect cells and characterized them in vitro. Smy1p does not move microtubules in an ensemble motility assay, and is not an active motor. Using total internal reflection fluorescence microscopy (TIRFM), we find that Smy1p does not bind strongly to microtubules, but diffuses along them in the presence or absence of ATP. Surprisingly, Smy1p also binds to and diffuses along actin-fascin bundles. This binding is ionic strength-dependent, indicating the interaction is electrostatic in nature. When a single Myo2p is attached to a quantum dot cargo, the complex does not move processively on actin bundles. However, when several Smy1p molecules are attached to the quantum dot in addition to a single Myo2p, the complex supports continuous, unidirectional movement. 46% of moving quantum dots run to the end of the actin bundle, with run lengths greater than 10 microns observed. We hypothesize that Smy1p acts as an electrostatic tether, keeping the quantum dot bound to actin after Myo2p undergoes its powerstroke. We propose that overexpression of Smy1p rescues the myo2-66 mutant by enhancing the binding of cargo to actin. A similar mechanism likely contributes to transport in wild-type cells when both Smy1p and Myo2p are present on the same cargo.1. SH Lillie and SS Brown (1992), Nature 356, 358-61.2. SH Lillie and SS Brown (1998), JCB 140, 873-83.

Myosin Va (myoVa) and myosin VI (myoVI) are processive molecular motors that transport cargo on actin tracks in opposite directions. We have shown that myoVa can effectively maneuver through an in vitro cytoskeletal model system composed of actin filament intersections and Arp2/3 branches (Ali et al. 2007). Here we challenge Quantum dot (Qdot)-labeled expressed myoVI with actin filament intersections and observed that myoVI maneuvers through intersections with the following statistics: 38% turned left or right with equal probability; 28% crossed over the intersecting actin filament; 34% terminated their run. The myoVI cross over probability is twice that of myoVa suggesting that the range of the myoVI leading head's diffusional search may be longer than myoVa. Similar to myoVa, myoVI has significant flexibility allowing it to turn at intersection angles up to 155°. When multiple myoVI were attached a Qdot, the turning probability increased to 53% whereas the cross over probability decreased to 15%. MyoVa and myoVI may be colocalized to the same cargo in vivo and to determine how these oppositely directed motors might interact during cargo transport, we attached both motors in a 1:1 ratio to a Qdot. We observed two types of movement associated with these myoVa/myoVI-labeled Qdots. A given Qdot would move in both the plus- or minus-end direction for periods of time at velocities appropriate for the specific motor, suggesting that myoVa and myoVI take turns transporting the Qdot. Other Qdots moved continuously but at velocities suggesting that both motors are simultaneous interacting with actin and undergoing an effective “tug of war.” These studies may help characterize how actin-based motors deliver their cargo through the complex actin network.

Myosin Va (myoVa) is an actin-based intracellular cargo transporter. In vitro experiments have established that a single myoVa moves processively along actin tracks, but less is known about how this motor operates within cells. Here we track the movement of a quantum dot (Qdot)-labeled myoVa HMM in COS-7 cells using total internal reflectance fluorescence microscopy. This labeling approach is unique in that it allows myoVa, instead of its cargo, to be tracked. Single-particle analysis showed short periods (≤0.5 s) of ATP-sensitive linear motion. The mean velocity of these trajectories was 604 nm/s and independent of the number of myoVa molecules attached to the Qdot. With high time (16.6 ms) and spatial (15 nm) resolution imaging, Qdot-labeled myoVa moved with sequential 75 nm steps per head, at a rate of 16 s−1, similarly to myoVa in vitro. Monte Carlo modeling suggests that the random nature of the trajectories represents processive myoVa motors undergoing a random walk through the dense and randomly oriented cortical actin network.

Class V myosins can be processive or non-processive, but both support cargo transport. Here we investigate the mechanism by which Myo4p, the single-headed non-processive class V myosin of Saccharomyces cerevisiae, can transport mRNA cargo from the mother to the bud tip. The adapter protein that couples the Myo4p/She3p motor complex to mRNA (She2p) is tetrameric, and thus can in principle recruit multiple motors. Total internal reflection fluorescence (TIRF) microscopy was used to show that one She2p tetramer recruits enough motors to support processive runs. Metal-shadowed images show two motors attached to a She2p tetramer. Deletion of a prominent α-helix that protrudes from the middle of She2p abolishes correct cellular localization of ASH1 mRNA, suggesting that it is a binding site for She3p. These results highlight that one strategy used by non-processive motors is to work together in small groups, which functionally allows them to support transport that is as robust as a single processive motor.

The budding yeast S. cerevisiae is an excellent model system to study cargo transport by myosin V. Cargo is transported from the mother cell to the growing bud exclusively by myosin V, and does not involve microtubule-based motors. We find that yeast myosin V (Myo2p) is non-processive in vitro,in agreement with previous results 1-3. This is surprising given that the cellular role of this motor is long-distance cargo transport. However, these experiments were performed using bare skeletal muscle actin filaments, which differ substantially from the native yeast actin track. Our goal is to reconstitute actin cables in vitro using yeast actin, yeast tropomyosin, and the actin bundling proteins fascin or fimbrin. Both isoforms of yeast tropomyosin stabilize yeast actin, resulting in much longer filaments. Preliminary data indicate that tropomyosin also enhances Myo2p function. TIRF microscopy was used to observe quantum dots transported by multiple Myo2p motors along the actin track. The presence of tropomyosin dramatically increased the run length and frequency of processive runs relative to bare actin filaments. We are currently testing if a single motor is capable of processive movement in the presence of tropomyosin. The effects of actin bundling on Myo2p function will also be assessed. Our results are consistent with the idea that the composition and structure of the actin track can greatly influence the properties of the motor.(1) Hodges et al., Curr Biol 19 (2009); (2) Dunn et al., JCB 178 (2007); (3) Reck-Peterson et al., JCB 153 (2001).

A prominent feature of most class V myosins is their ability to take multiple steps on actin without dissociating, known as processivity. Recent evidence showed that there are also non-processive class V myosins, both in humans and lower organisms. These motors need to work in ensembles to ensure continuous, unidirectional movement. The budding yeast Saccharomyces cerevisiae has two class V myosins, both of which are non-processive, but for different reasons. Myo2p is a dimeric motor with a low duty cycle, meaning that it spends a small portion of its cycle time strongly attached to actin. Myo4p has a high duty cycle motor, but is single-headed and thus cannot move processively as a single molecule. We propose that the association of Myo4p with its adapter protein She3p accounts for why it is single-headed. She3p is required for transport of all cargo of Myo4p (mRNA and cortical ER), and thus it has become a subunit of the motor. Myo2p, in contrast, moves many different cargoes (e.g. organelles and secretory vesicles), each with a unique adapter protein. Using a combination of in vitro and in vivo techniques, we probe how the features of each motor are uniquely suited for its particular cellular role. The involvement of other proteins that act as “processivity factors” will also be discussed.

Like their mammalian counterparts, class V myosins in S. cerevisiae (Myo2p and Myo4p) bind to various adapter proteins to target a particular cargo for transport. Myo4p uses the adapter proteins She3p and She2p in order to transport mRNA from the mother cell to the bud. She3p binds to the rod of Myo4p, and prevents it from dimerizing, thus forming a single-headed motor complex (Hodges et al., 2008; Bookwalter et al., 2009). Because the Myo4p/She3p complex is single-headed, the question arises as to whether enough motors can bind to a single She2p to enable continuous cargo transport. The She2p crystal structure suggested that She2p exists as a dimer (Niessing et al., 2004). In contrast, our sedimentation equilibrium measurements of She2p were consistent with formation of a tetramer in solution, in principle allowing for binding of four motor heads. We showed that Myo4p/She3p forms a complex with tetrameric She2p in the absence of mRNA, based on sedimentation velocity experiments and co-purification. Mutation of Ser 120 to Tyr converts She2p to a dimer. The ability of the motor complex to bind to this and other She2p mutants is being tested in order to map the binding interface. Total internal reflection fluorescence microscopy is being used to test whether the native She2p tetramer can bind enough single-headed motors to support continuous movement on actin. The ability of She2p mutants to support correct bud tip localization of ASH1 mRNA in living yeast cells will also be assessed. These studies will help elucidate how a non-processive single-headed motor can act as a cargo transporter.

Myosin-Va (myoVa), one of the best characterized actin-based molecular motors, transports a variety of intracellular cargos. In order to bind a specific cargo, myoVa forms a tri-partite complex with a Rab effector protein (i.e. adapter) and a Rab GTPase protein (e.g. Rab27a) that is inserted in the granule membrane. MyoVa delivers insulin granules to the plasma membrane in pancreatic beta-cells. Interestingly, there are four known adapter proteins expressed in beta-cells, i.e. Granuphilin-A/B, Rabphilin and MyRIP, all of which bind myoVa. The role of these adapter proteins in cargo transport is poorly understood.Using TIRF microscopy, we measured the speed, run-length and stepping behavior of myoVa in presence of Qdot-labeled adapter proteins. At 25 mM KCl, the adapter proteins do not show appreciable activation of the inhibited myoVa motor. However, at physiological salt concentration, the adapter proteins significantly increase the run-length and the run-frequency of myoVa on actin filaments. Specifically, in the presence of Granuphilin A, the myoVa run-frequency increases ∼6-fold, with an ∼3.5-fold run-length enhancement as the motor steps (72nm) normally, but at half the speed. By labeling Granuphilin-A and MyRIP with a Qdot, we observed binding of these adapters directly to actin filaments, suggesting that they enhance the motor's run-length and slow speed by a tethering mechanism, similar to Melanophilin (Sckolnick, et al., 2013). In contrast, Granuphilin-B and Rabphilin have little binding affinity for actin. Nonetheless, they bind to and activate myoVa, because the full-length myoVa step size becomes regular like the constitutively active, truncated myoVa-HMM. A common feature of these adapter proteins is that they ensure that the motor remains active while attached to the cargo. However, only some adapter proteins have actin-tethering capacity, which may enhance the long-range vesicle transport. These functional differences may play synergistic roles in the cell.

Myosin Va is a processive molecular motor that transports intracellular cargo along actin filament tracks. In vivo, multiple myosin Va motors, attached to the same cargo, must interact but the mode of interaction is far from certain. We have shown that oppositely-directed myosins synchronized their stepping while engaged in a tug of war (Ali et al. 2011). Therefore, to understand the mechanical interactions between multiple motors of the same type, we have developed a simplified in vitro model in which two individual myosin Va were linked via a Qdot-cargo. To monitor each motor's stepping dynamics, one head of each motor was labeled with either a red or a green Qdot. For this two motor complex, velocity was reduced 1.3 - fold while run length increased 1.6 - fold. The leading motor must experience a resistive load from the trailing motor to account for the velocity reduction and why the leading motor has an 11% back step probability. When motors in the complex were close together (∼36 nm), their stepping appeared independent. However, when the distance between them grew larger (>72 nm), they began to synchronize their stepping as the tension in the linkage between them presumably rose. We relate the findings with a model of two coupled stochastic steppers in which the stochasticity of motor steps stretches the linkage, while the stiffness of the linkage limits the intermotor distance and synchronizes their stepping. Even in this simplified model system, mechanical interactions between two identical motors are complex but will help define the collective mechanics of larger motor ensembles.

There is a large gap between the conditions we use to study molecular motors in vitro, and the motor walking in a cellular environment. For simplicity, in vitro experiments on myosin V generally use single constitutively active truncated motor constructs that walk on individual bare actin filaments. Within the cell, however, actin generally has bound tropomyosin. Full-length motors are regulated by a folded-to-extended conformational transition, and adapter proteins that link the motor to cargo can affect this conformational transition. Adapter proteins, or the cargo itself, can also recruit multiple motors to improve the efficiency of cargo movement. The theme of this talk is that by re-creating a more native myosin complex and actin track, unexpected properties of the motor emerge that are likely to be important for cellular cargo transport. Several different class V myosins will be used to illustrate these principles.

The processive, hand-over-hand mechanism of myosin Va (myoVa) walking on actin has been intensively studied in vitro, but less is known about its behavior within cells. We previously showed that myoVa undergoes a random walk in COS-7 cells as it processively steps along actin tracks within the dense and randomly oriented cortical actin network (Nelson et al. BJ 97:509, 2009). Here we test how the processivity of myoV impacts on the observed cellular motion. A mutant construct with 3-fold shorter run lengths than wild-type myoVa (WT), and one with ∼1.5-fold longer run lengths, were introduced into cultured COS-7 cells by pinocytosis. The motion of Quantum dot (Qdot)-labeled single motors within the cultured cells was analyzed through high resolution TIRF microscopy and single particle tracking. Mean Squared Displacement (MSD) analysis of the motor:Qdot trajectories appear to be diffusive over short time scale (∼1s), and sub-diffusive over longer time scales (∼10s). Strikingly, the diffusion coefficients for the short time scales strictly correlate with the processivity of the motor, and range from 0.06μm2/s for the least processive motor, to 0.15μm2/s for the more processive variant. The non-processive and very slow myoVc, had the lowest diffusion coefficient of any of the constructs tested (0.019μm2/s). The observed diffusion coefficients and the sub-diffusive motion for longer time scales was successfully modeled through Monte Carlo simulations assuming that a processive myoVa motor will either cross over, turn or terminate at actin filament intersections within the randomly oriented actin meshwork. Once the motor terminates its run it undergoes restricted diffusion, being potentially confined within domains that are bounded by cytoskeletal or organellar structures. The motor-dependent cellular behavior supports the idea that the apparently wandering trajectories are random walks by active motors.

Myosin Va (myoVa) walks processively while carrying cargo towards the plus end of actin filaments. In cells, parallel actin filament bundles (e.g. stress fibers and filopodia) present a directional challenge to myoVa cargo transport. Therefore, we formed unipolar (fascin) and mixed polarity (alpha-actinin) actin bundles as tracks for expressed myoVa-HMM with a C-terminal biotin tag. In this assay, a single streptavidin-Qdot served as cargo for one or many (∼ 5) myoVa motors. Qdots transported by one or many myoVa molecules traveled in the same direction on unipolar bundles, while moving in either direction on mixed polarity bundles. Qdot speeds were the same regardless of bundle type or number of motors (400nm/s), and similar to that for one or many motors on a single actin filament (Nelson et al., 2009). However, run lengths for single motors were 2-3 times longer on bundles than previously observed on single actin filaments. This suggests that on parallel tracks the leading head has a greater number of actins within its reach, thus decreasing the probability of run termination. Interestingly, on mixed polarity bundles, we observed individual Qdots changing directions in the middle of a run, the frequency of which increases in the multiple motor case. It was not surprising that a Qdot with a single motor can switch directions on a mixed polarity bundle, given myoVa's inherent flexibility that allows it to turn up to 150o at actin filament intersections (Ali et al., 2007). These data also suggest that one or many myoVa molecules bound to a single cargo have the ability to jump tracks to neighboring actin filaments. With Qdot-labeling of the individual heads, high spatial resolution studies will confirm this on mixed polarity bundles, and determine whether the motors also wander on unipolar bundles.

Melanophilin (Mlph) is a cargo adapter protein that links melanosomes via Rab27a to myosin Va (myoVa) for transport along actin. In the absence of cargo, full length myoVa is in equilibrium between a folded, inactive and an extended, active conformation. This equilibrium causes a deviation from a hand-over-hand stepping pattern due to altered gating between the heads (Armstrong et al. 2012). Cargo binding has been suggested to activate myoVa for transport. Here we used single molecule TIRF assays at near physiological ionic strength (150mM KCl) to determine the effect of Mlph on myoVa processivity. In the absence of Mlph, Qdot labeled full-length myoVa moved at a median velocity of 443nm/s, and showed the altered stepping pattern previously seen at lower ionic strength. Addition of Mlph recruited 14-times more motors to move processively, consistent with a simple model of cargo activation. The myoVa-Mlph complex also showed increased run lengths, with many motors traveling to the ends of the actin filament. In the presence of Mlph, myoVa moved much more slowly (median velocity=75nm/s). The speed distribution was asymmetrical and similar to speeds of melanosome movement observed in vivo. In the presence of Mlph, myoVa showed normal gating between the heads, and hand-over-hand steps on actin typical of a fully-active motor. Based on mutagenesis of Mlph the enhanced processivity depended on a positively charged cluster of amino acids in the actin binding site of Mlph. This suggests that Mlph acts as an electrostatic tether to limit myoVa dissociation from actin, a property likely to favor the transfer of melanosomes to adjacent keratinocytes in vivo. More generally, our results suggest that adapter proteins which link motors to cargo can affect motor properties in ways favorable for their biological role.

The number of processive motors attached to a cellular cargo influences its transport behavior. Altering either motor number or the ratio of different classes of motors can therefore be a mechanism to regulate intracellular transport. Jamison et al. recently showed that two kinesin-1 motors coupled by a DNA scaffold have transport properties that are often dominated by one of the motors. Here we perform a similar experiment with myosin Va (myoVa), which has a larger step size (∼36nm) and walks on a smaller track than kinesin. A heterodimeric myoVa was labeled on only one head with either a red or green quantum dot (Qdot). Two myoVa molecules were then linked to an ∼50 nm long double-stranded DNA scaffold. Only complexes with one red and one green Qdot were analyzed. Our results show that the complex has increased run length (∼1μm) compared to a single myoVa (∼0.6μm). Average run lengths are, however, smaller than those predicted for two myosins assuming motor stepping, binding, and detachment is unaffected by intermotor interactions. Furthermore, the motor complex moved with reduced velocity (0.19 μm/s versus 0.27 μm/s for the single motor case). A histogram of the distances between the labeled heads of the two motors contains multiple peaks at ∼85, 130 and 165 nm, indicating the system is flexible. The distance between motors changes in time and the stepping pattern of the two motors are variable, suggesting asynchronous motor stepping. After the first motor binds to actin, the second motor binds at ∼10 s-1. Our findings suggest that the walking behavior of two myoVa molecules is altered when they are coupled mechanically, but perhaps in a different way than multiple kinesins. Our technique constitutes a unique tool to understand collective motor behavior.

Full-length myosin Va (myoVa) is auto-inhibited via a motor domain-globular tail interaction, unlike the truncated constitutively active myoVa-HMM. One potential mechanism to activate the full-length motor is cargo binding to the tail, which would compete with the head-tail interaction and trigger the molecule to extend and be activated for transport.In the absence of cargo, it was recently shown that full-length myoVa has two modes of interaction with actin in the presence of MgATP (Armstrong et al.). Most motors bind to actin but do not move, while the remainder show processive motion, but with a variable stepping pattern and altered gating. Here we investigate how binding of melanophilin (Mlph), which links the melanocyte-specific isoform of myoVa to the Rab27a(GTP)-melanosome complex, affects the properties of myoVa at the single-molecule level.In the absence of Mlph at 150mM KCl, a subset of Quantum dot labeled full-length myoVa moved at a median velocity of 566nm/s with the variable stepping pattern previously described, suggesting altered gating under these conditions. Addition of Mlph recruited 7-times more motors to move processively, consistent with a simple model of cargo activation. The myoVa-Mlph complex also showed increased run lengths, with many traveling to the ends of the actin filament. In the presence of Mlph, myoVa moved much more slowly (median velocity=76nm/s), leading to longer travel times on actin. When Mlph was bound to the motor the step sizes were normally distributed around 60 ± 14nm (SD) steps. Therefore, while myoVa moves more slowly along actin in the presence of the cargo adapter protein Mlph, it covers a greater distance with a more uniform and efficient stepping pattern. This slower processive movement could potentially facilitate binding of the Rab27a(GTP)-melanosome complex.

A feature of most class V myosins is their ability to move processively on actin. The budding yeast Saccharomyces cerevisiae has a non-processive class V myosin, Myo4p, which is a single-headed but high duty cycle motor. Its cellular role is to asymmetrically transport more than 20 different mRNAs, a widely used strategy to polarize a protein within the cell. The most studied mRNA is ASH1, which is moved by Myo4p to the bud tip to repress mating type switching in the daughter cell. Here we determine the features of Myo4p that are necessary for correct localization of ASH1 mRNA to the daughter cell. This process requires the adapter protein She3p, and the mRNA binding protein She2p, which binds ASH1 at specific localization elements called zip codes. Based on a series of chimeric constructs, we showed that the rod region of Myo4p, but not the globular tail, is essential for correct localization of ASH1 mRNA. The rod thus contains the primary binding site for She3p, consistent with our earlier in vitro studies (Hodges et al., 2008). To test if mRNA localization is more efficient when two motors are coupled together, we compared transport by a constitutive dimer of Myo4p/She3p with a constitutive monomer. Correct ASH1 mRNA localization was achieved equally well with both constructs. This may reflect the fact that many mRNAs and thus many motors are part of the translocation complex. Our results show that the most important feature for correct localization is the retention of coupling between all the members of the complex (Myo4p- She3p-She2p-ASH1 mRNA), which is aided by She3p being a tightly bound subunit of Myo4p.