Myosin subfragment-1 is sufficient to move actin filaments in vitro

The rotating crossbridge model for muscle contraction proposes that force is produced by a change in angle of the crossbridge between the overlapping thick and thin filaments. Myosin, the major component of the thick filament, is comprised of two heavy chains and two pairs of light chains. Together they form two globular heads, which give rise to the crossbridge in muscle, and a coiled-coil rod, which forms the shaft of the thick filament. The isolated head fragment, subfragment-1 (S1), contains the ATPase and actin-binding activities of myosin (Fig. 1). Although S1 seems to have the requisite enzymatic activity, direct evidence that S1 is sufficient to drive actin movement has been lacking. It has long been recognized that in vitro movement assays are an important approach for identifying the elements in muscle responsible for force generation. Hynes et al. showed that beads coated with heavy meromyosin (HMM), a soluble proteolytic fragment of myosin consisting of a part of the rod and the two heads, can move on Nitella actin filaments. Using the myosin-coated surface assay of Kron and Spudich, Harada et al. showed that single-headed myosin filaments bound to glass support movement of actin at nearly the same speed as intact myosin filaments. These studies show that the terminal portion of the rod and the two-headed nature of myosin are not required for movement. To restrict the region responsible for movement further, we have modified the myosin-coated surface assay by replacing the glass surface with a nitrocellulose film. Here we report that myosin filaments, soluble myosin, HMM or S1, when bound to a nitrocellulose film, support actin sliding movement (Fig. 2). That S1 is sufficient to cause sliding movement of actin filaments in vitro gives strong support to models of contraction that place the site of active movement in muscle within the myosin head.     

Location of the ATPase site of myosin determined by three-dimensional electron microscopy

Both ATP hydrolysis by myosin and the accompanying cyclic association-dissociation of actin and myosin are essential for muscle contraction. It is important for understanding the molecular mechanism of contraction to know the three-dimensional locations of the two major functional sites of myosin: the ATPase site and the actin-binding site. We have determined the position of the ATPase site of myosin using three-dimensional image reconstruction from electron micrographs and site-specific labelling with the avidin-biotin system. The ATPase site is about 5 nm from the tip of the myosin head and is about 4 nm away from the actin-binding site of myosin. This is the first report of the three-dimensional location of an enzyme active site by electron microscopy.     

Isolation and characterization of the G-actin-myosin head complex

The two main proteins involved in muscular contraction and cell motility, myosin and actin, possess the intrinsic property of being able to form filamentous structures. This property poses a serious impediment to the study of their structures and interactions, and a considerable effort has thus been made to isolate their functional domains. The globular part of myosin, subfragment-1 (S1), which possesses ATPase and actin-binding sites as well as supporting the movement of actin filaments during in vitro assays, has been isolated. But because S1 is efficient in inducing actin polymerization, as is myosin, it has not been possible to prepare and characterize a complex of S1 with monomeric actin (G-actin). We have now used chromatographically purified proteins to show that only the S1 isoenzyme carrying the A1 light-chain subunit promotes actin polymerization. The other isoenzyme, S1 (A2), carrying the A2 light-chain subunit, binds to actin, forming a tight complex of G-actin and S1 in a 1:1 ratio. This new functional difference between myosin isoforms directly implicates the A1 light-chain in myosin-induced actin polymerization. Additionally, this finding should lead to the purification of the stable G-actin-S1 complex needed to resolve the structure and to understand the molecular dynamics of the actin-myosin system.     

Sliding movement of single actin filaments on one-headed myosin filaments

The myosin molecule consists of two heads, each of which contains an enzymatic active site and an actin-binding site. The fundamental problem of whether the two heads function independently or cooperatively during muscle contraction has been studied by methods using an actomyosin thread, superprecipitation and chemical modification of muscle fibres. No clear conclusion has yet been reached. We have approached this question using an assay system in which sliding movements of fluorescently labelled single actin filaments along myosin filaments can be observed directly. Here, we report direct measurement of the sliding of single actin filaments along one-headed myosin filaments in which the density of heads was varied over a wide range. Our results show that cooperative interaction between the two heads of myosin is not essential for inducing the sliding movement of actin filaments.     

Force measurements by micromanipulation of a single actin filament by glass needles

Single actin filaments (approximately 7 nm in diameter) labelled with fluorescent phalloidin can be clearly seen by video-fluorescence microscopy. This technique has been used to observe motions of single filaments in solution and in several in vitro movement assays. In a further development of the technique, we report here a method to catch and manipulate a single actin filament (F-actin) by glass microneedles under conditions in which external force on the filament can be applied and measured. Using this method, we directly measured the tensile strength of a filament (the force necessary to break the bond between two actin monomers) and the force required for a filament to be moved by myosin or its proteolytic fragment bound to a glass surface in the presence of ATP. The first result shows that the tensile strength of the F-actin-phalloidin complex is comparable with the average force exerted on a single thin filament in muscle fibres during isometric contraction. This force is increased only slightly by tropomyosin. The second measurement shows that the myosin head (subfragment-1) can produce the same ATP-dependent force as intact myosin. The magnitude of this force is comparable with that produced by each head of myosin in muscle during isometric contraction.     

Mechanism of force generation by myosin heads in skeletal muscle

Muscles generate force and shortening in a cyclical interaction between the myosin head domains projecting from the myosin filaments and the adjacent actin filaments. Although many features of the dynamic performance of muscle are determined by the rates of attachment and detachment of myosin and actin, the primary event in force generation is thought to be a conformational change or 'working stroke' in the actin-bound myosin head. According to this hypothesis, the working stroke is much faster than attachment or detachment, but can be observed directly in the rapid force transients that follow step displacement of the filaments. Although many studies of the mechanism of muscle contraction have been based on this hypothesis, the alternative view-that the fast force transients are caused by fast components of attachment and detachment--has not been excluded definitively. Here we show that measurements of the axial motions of the myosin heads at ?ngstr?m resolution by a new X-ray interference technique rule out the rapid attachment/detachment hypothesis, and provide compelling support for the working stroke model of force generation.     

Alteration in crossbridge kinetics caused by mutations in actin

The generation of force during muscle contraction results from the interaction of myosin and actin. The kinetics of this force generation vary between different muscle types and within the same muscle type in different species. Most attention has focused on the role of myosin isoforms in determining these differences. The role of actin isoforms has received little attention, largely because of the lack of a suitable cell type in which the myosin isoform remains constant yet the actin isoforms vary. An alternative approach would be to examine the effect of actin mutations, however, most of these cause such gross disruption of muscle structure that mechanical measurements are impossible. We have now identified two actin mutations which, despite involving conserved amino acids, can assemble into virtually normal myofibrils. These amino-acid changes in actin significantly affect the kinetics of force generation by muscle fibres. One of the mutations is not in the putative myosin-binding site, demonstrating the importance of long-range effects of amino acids on actin function.     

A structural state of the myosin V motor without bound nucleotide

The myosin superfamily of molecular motors use ATP hydrolysis and actin-activated product release to produce directed movement and force. Although this is generally thought to involve movement of a mechanical lever arm attached to a motor core, the structural details of the rearrangement in myosin that drive the lever arm motion on actin attachment are unknown. Motivated by kinetic evidence that the processive unconventional myosin, myosin V, populates a unique state in the absence of nucleotide and actin, we obtained a 2.0 A structure of a myosin V fragment. Here we reveal a conformation of myosin without bound nucleotide. The nucleotide-binding site has adopted new conformations of the nucleotide-binding elements that reduce the affinity for the nucleotide. The major cleft in the molecule has closed, and the lever arm has assumed a position consistent with that in an actomyosin rigor complex. These changes have been accomplished by relative movements of the subdomains of the molecule, and reveal elements of the structural communication between the actin-binding interface and nucleotide-binding site of myosin that underlie the mechanism of chemo-mechanical transduction.     

Structure of the actin-myosin interface

The topography of the rigor complex between F-actin and myosin heads (S1) has been investigated by carbodiimide zero-length cross-linking. The results demonstrate for the first time that the 95,000-molecular weight (95K) heavy chain of the myosin head enters into van der Waals contact with two neighbouring actin monomers; one is bound to the 50K domain and the other to the 20K domain of the myosin chain. The covalent F-actin-S1 complex can be isolated; it shows a vastly elevated Mg2+-ATPase. Each pair of actin subunits in the thin filament seems to act as a functional unit for specific binding of a myosin head and stimulation of its Mg2+-ATPase activity.     

A gelsolin-like Ca2+-dependent actin-binding domain in villin

Villin is an actin-binding protein of relative molecular mass (Mr) 95,000 found in the core bundle of microfilaments in brush border microvilli from intestine. In physiological calcium concentrations (less than 1 microM), villin crosslinks actin filaments into bundles. However, in free calcium concentrations (greater than 1 microM), villin severs actin filaments into short pieces. To understand how villin can sever and bundle actin filaments, we are studying the molecular basis of villin-actin binding interactions by identifying important actin-binding domains in villin. Here, we report the purification and preliminary characterization of a 44,000-Mr fragment of villin which contains a calcium-dependent actin-severing activity. In addition, the partial amino-acid sequence from the amino terminus of this fragment reveals homology with a 16-residue region near the amino terminus of gelsolin, an actin-severing protein found in many cells and sera. The sequence homology suggests a common structural basis for the calcium-regulated actin-severing properties shared by villin and gelsolin.     

Sliding distance between actin and myosin filaments per ATP molecule hydrolysed in skinned muscle fibres

Muscle contraction is generally thought to be driven by tilting of the 19-nm-long myosin head, part of the thick filament, while attached to actin, part of the thin filament. This motion would produce about 12 nm of filament sliding. Recent estimates of the sliding distance per ATP molecule hydrolysed by actomyosin in vitro vary widely from 8 nm to greater than or equal to 200 nm. The latter value is incompatible with a power stroke incorporating a single tilting motion of the head. We have measured the isotonic sliding distance per ATP molecule hydrolysed during the interaction between myosin and actin in skinned muscle fibres. We directly estimated the proportion of simultaneously attached actomyosin complexes and their ATP use. We report here that at low loads the interaction distance is at least 40 nm. This distance corresponds to the length of the power stroke plus the filament sliding while actomyosin crossbridges bear negative drag forces. If the power stroke is 12 nm, then our results indicate the drag distance to be at least 28 nm. Our results could also be explained by multiple power strokes per ATP molecule hydrolysed.     

Atomic model of a myosin filament in the relaxed state

Contraction of muscle involves the cyclic interaction of myosin heads on the thick filaments with actin subunits in the thin filaments. Muscles relax when this interaction is blocked by molecular switches on either or both filaments. Insight into the relaxed (switched OFF) structure of myosin has come from electron microscopic studies of smooth muscle myosin molecules, which are regulated by phosphorylation. These studies suggest that the OFF state is achieved by an asymmetric, intramolecular interaction between the actin-binding region of one head and the converter region of the other, switching both heads off. Although this is a plausible model for relaxation based on isolated myosin molecules, it does not reveal whether this structure is present in native myosin filaments. Here we analyse the structure of a phosphorylation-regulated striated muscle thick filament using cryo-electron microscopy. Three-dimensional reconstruction and atomic fitting studies suggest that the 'interacting-head' structure is also present in the filament, and that it may underlie the relaxed state of thick filaments in both smooth and myosin-regulated striated muscles over a wide range of species.     

Low Ca2+ impedes cross-bridge detachment in chemically skinned Taenia coli

Muscle force is generated by cycling cross-bridges between actin and myosin filaments. In smooth muscle, cyclic attachment and detachment of cross-bridges is thought to be induced by a Ca2+- and calmodulin-dependent myosin light chain kinase which phosphorylates myosin. The relaxation that occurs after Ca2+ removal is usually ascribed to dephosphorylation of myosin by a phosphatase as non-phosphorylated myosin is unable to form force-generating criss-bridges. Recently, Dillon et al. claimed, however, that dephosphorylation of attached cross-bridges may impede cross-bridge detachment, thus forming so-called 'latch bridges'. Here we present evidence that after a Ca2+- and calmodulin-induced contraction of chemically skinned guinea pig Taenia coli, the rapid removal of Ca2+ impedes the detachment of the myosin cross-bridges from the actin filament; force can then be maintained without energy consumption. The extremely slowly detaching cross-bridges which maintain the force after Ca2+ removal may indeed correspond to the 'latch bridges' mentioned above.     

Homology of a yeast actin-binding protein to signal transduction proteins and myosin-I

In yeast, the cortical actin cytoskeleton seems to specify sites of growth of the cell surface. Because the actin-binding protein ABP1p is associated with the cortical cytoskeleton of Saccharomyces cerevisiae, it might be involved in the spatial organization of cell surface growth. ABP1p is localized to the cortical cytoskeleton and its overproduction causes assembly of the cortical actin cytoskeleton at inappropriate sites on the cell surface, resulting in delocalized surface growth. We have now cloned and sequenced the gene encoding ABP1p. ABP1p is a novel protein with a 50 amino-acid C-terminal domain that is very similar to the SH3 domain in the non-catalytic region of nonreceptor tyrosine kinases (including those encoded by the proto-oncogenes c-src and c-abl), in phospholipase C gamma and in alpha-spectrin. We also identified an SH3-related motif in the actin-binding tail domain of myosin-I. The identification of SH3 domains in a family of otherwise unrelated proteins that associate with the membrane cytoskeleton indicates that this domain might serve to bring together signal transduction proteins and their targets or regulators, or both, in the membrane cytoskeleton.     

Site-directed mutagenesis of the regulatory light-chain Ca2+/Mg2+ binding site and its role in hybrid myosins

The regulatory light chains, small polypeptides located on the myosin head, regulate the interaction of myosin with actin in response to either Ca2+ or phosphorylation. The demonstration that the regulatory light chains on scallop myosin can be replaced by light chains from other myosins has allowed us to compare the functional capabilities of different light chains, but has not enabled us to probe the role of features, such as the Ca2+/Mg2+ binding site, that are common to all of them. Here, we describe the use of site-directed mutagenesis to study the function of that site. We synthesized the chicken skeletal myosin light chain in Escherichia coli and constructed mutants with substitutions within the Ca2+/Mg2+ binding site. When the aspartate residues at the first and sixth Ca2+ coordination positions are replaced by uncharged alanines, the light chains have a reduced Ca2+ binding capacity but still bind to scallop myosin with high affinity. Unlike the wild-type skeletal light chain which inhibits myosin interaction with actin, the mutants activate it. Thus, an intact Ca2+/Mg2+ binding site in the N-terminal region of the light chain is essential for regulating the interaction of myosin with actin.     

The motor protein myosin-I produces its working stroke in two steps

Many types of cellular motility, including muscle contraction, are driven by the cyclical interaction of the motor protein myosin with actin filaments, coupled to the breakdown of ATP. It is thought that myosin binds to actin and then produces force and movement as it 'tilts' or 'rocks' into one or more subsequent, stable conformations. Here we use an optical-tweezers transducer to measure the mechanical transitions made by a single myosin head while it is attached to actin. We find that two members of the myosin-I family, rat liver myosin-I of relative molecular mass 130,000 (M(r) 130K) and chick intestinal brush-border myosin-I, produce movement in two distinct steps. The initial movement (of roughly 6 nanometres) is produced within 10 milliseconds of actomyosin binding, and the second step (of roughly 5.5 nanometres) occurs after a variable time delay. The duration of the period following the second step is also variable and depends on the concentration of ATP. At the highest time resolution possible (about 1 millisecond), we cannot detect this second step when studying the single-headed subfragment-1 of fast skeletal muscle myosin II. The slower kinetics of myosin-I have allowed us to observe the separate mechanical states that contribute to its working stroke.     

Electron cryo-microscopy shows how strong binding of myosin to actin releases nucleotide

Muscle contraction involves the cyclic interaction of the myosin cross-bridges with the actin filament, which is coupled to steps in the hydrolysis of ATP. While bound to actin each cross-bridge undergoes a conformational change, often referred to as the "power stroke", which moves the actin filament past the myosin filaments; this is associated with the release of the products of ATP hydrolysis and a stronger binding of myosin to actin. The association of a new ATP molecule weakens the binding again, and the attached cross-bridge rapidly dissociates from actin. The nucleotide is then hydrolysed, the conformational change reverses, and the myosin cross-bridge reattaches to actin. X-ray crystallography has determined the structural basis of the power stroke, but it is still not clear why the binding of actin weakens that of the nucleotide and vice versa. Here we describe, by fitting atomic models of actin and the myosin cross-bridge into high-resolution electron cryo-microscopy three-dimensional reconstructions, the molecular basis of this linkage. The closing of the actin-binding cleft when actin binds is structurally coupled to the opening of the nucleotide-binding pocket.     

Plasma and cytoplasmic gelsolins are encoded by a single gene and contain a duplicated actin-binding domain

Gelsolin is representative of a class of actin-modulating proteins found in lower eukaryotes to mammals, which sever actin filaments. Gelsolin found in the cytoplasm of cells is functionally similar to a mammalian plasma protein of similar size, originally called ADF or brevin. Human plasma and rabbit macrophage gelsolins differ by the presence of a 25-amino-acid residue extension on plasma gelsolin which appears to account for the difference in relative molecular mass (Mr) between the proteins as assessed by SDS-polyacrylamide gel electrophoresis (PAGE), 93,000 (93K) and 90K, respectively. Here we report the isolation of full-length human plasma gelsolin complementary DNA clones from a HepG2 library. The inferred amino-acid sequence reveals the presence of a signal peptide, a long tandem repeat that matches the actin-binding domains of gelsolin, a tetrapeptide present in actin and extended regions of identical sequence with rabbit macrophage gelsolin. Southern blot analysis indicates that a single gene in the haploid genome encodes both protein forms.     

Molecular engineering of a backwards-moving myosin motor

All members of the diverse myosin superfamily have a highly conserved globular motor domain that contains the actin- and nucleotide-binding sites and produces force and movement. The light-chain-binding domain connects the motor domain to a variety of functionally specialized tail domains and amplifies small structural changes in the motor domain through rotation of a lever arm. Myosins move on polarized actin filaments either forwards to the barbed (+) or backwards to the pointed (-) end. Here, we describe the engineering of an artificial backwards-moving myosin from three pre-existing molecular building blocks. These blocks are: a forward-moving class I myosin motor domain, a directional inverter formed by a four-helix bundle segment of human guanylate-binding protein-1 and an artificial lever arm formed by two alpha-actinin repeats. Our results prove that reverse-direction movement of myosins can be achieved simply by rotating the direction of the lever arm 180 degrees.     

Visualization of an unstable coiled coil from the scallop myosin rod

Alpha-helical coiled coils in muscle exemplify simplicity and economy of protein design: small variations in sequence lead to remarkable diversity in cellular functions. Myosin II is the key protein in muscle contraction, and the molecule's two-chain alpha-helical coiled-coil rod region--towards the carboxy terminus of the heavy chain--has unusual structural and dynamic features. The amino-terminal subfragment-2 (S2) domains of the rods can swing out from the thick filament backbone at a hinge in the coiled coil, allowing the two myosin 'heads' and their motor domains to interact with actin and generate tension. Most of the S2 rod appears to be a flexible coiled coil, but studies suggest that the structure at the N-terminal region is unstable, and unwinding or bending of the alpha-helices near the head-rod junction seems necessary for many of myosin's functional properties. Here we show the physical basis of a particularly weak coiled-coil segment by determining the 2.5-A-resolution crystal structure of a leucine-zipper-stabilized fragment of the scallop striated-muscle myosin rod adjacent to the head-rod junction. The N-terminal 14 residues are poorly ordered; the rest of the S2 segment forms a flexible coiled coil with poorly packed core residues. The unusual absence of interhelical salt bridges here exposes apolar core atoms to solvent.