Myosins are a superfamily of molecular motors that convert the chemical energy of ATP hydrolysis into directed motion along the filamentous protein actin.
Myosins function by converting the chemical energy associated with ATP hydrolysis into force and directed motion. The goal of this review is to discuss the coupled chemical and mechanical cycles that allow this energy transduction to occur and to examine how this kinetic cycle is “tuned” by the various myosin classes to achieve distinct functions.
Myosin Structural Features
Although there are at least 35 distinct myosin classes implicated in a variety of cellular processes (Ordonitz and Kollmar 2007), all myosins share certain structural features, which are discussed in detail in “Myosin: Fundamental Properties and Structure.” A brief overview is provided here.
According to the lever arm model, the rigid neck domain acts as a lever that amplifies small movements within the motor domain. The force-generating swing of the neck domain is referred to as the powerstroke.
The motor domain consists primarily of α-helices surrounding a core seven-stranded β-sheet (Fig. 1b). The α-helices form a deep cleft (referred to as the 50-kDa cleft) in the motor extending from the nucleotide-binding site to the actin-binding interface. The motor domain is comprised of four subdomains (the N-terminal, upper 50-kDa (U50), lower 50-kDa (L50), and converter subdomains) that are connected by four linkers (the strut, switch II, relay, and SH1 helix) (Fig. 1c, d). Movements of the four domains are coupled as a result of these linkers, allowing for direct communication between the various parts of the motor domain.
The Common Myosin Kinetic Cycle
A myosin motor generates force and motion by coupling three cycles: (1) an ATPase cycle (consisting of ATP binding, ATP hydrolysis, and release of hydrolysis products), (2) an actin-binding cycle (consisting of binding and release of actin), and (3) a mechanical cycle (consisting of motions of the neck domain/lever arm). The resultant kinetic cycle is referred to as the actomyosin chemomechanical cycle. During their kinetic cycles, all studied myosin classes appear to transition through the same states, though not necessarily at the same rates. Detailed discussions of our current understanding of the actomyosin chemomechanical cycle can be found in (Sweeney and Houdusse 2010; Llinas et al. 2012). An overview is presented below.
States for Which Crystal Structures Have Been Solved
Starting from the left side of Fig. 2a, the motor begins in a state in which it is associated with actin and has no bound-nucleotide (AM), the so-called rigor state. The 50-kDa cleft is completely closed (both at the actin-binding interface and near the nucleotide-binding pocket) (Fig. 3, middle). In the absence of nucleotide, the β-sheet is highly twisted, positioning switch I away from the P-loop and close to switch II (Fig. 3, bottom).
Upon binding ATP, the post-rigor state (M⋅T) is formed where nucleotide-binding elements are repositioned to coordinate the nucleotide. In doing so, the cleft completely opens (Fig. 3, middle) causing the motor to dissociate from actin. In this state, the β-sheet partially untwists, positioning switch I near the nucleotide and relatively far from switch II (Fig. 3, bottom). The lever arm position changes only slightly between the rigor and post-rigor states.
Hydrolysis of ATP (M⋅D⋅P i ) is associated with partial closure of the 50-kDa cleft, involving only the inner part of the cleft near the nucleotide-binding pocket (Fig. 3, middle). This results in interactions between switch II and the γ-phosphate that trap the phosphate within the motor (Fig. 3, bottom). In addition, this conformation of switch II repositions the relay and SH1 helix such that the lever arm undergoes a conformational change that primes it for its next powerstroke, a motion referred to as the recovery stroke. The distortion of the β-sheet is fully untwisted in this state.
Phosphate Release and Generation of Force
Because crystal structures have not yet been solved for the remaining states, there is still debate as to how the chemomechanical cycle proceeds. It is likely that the motor binds weakly to actin (A*M⋅D⋅P i ) through an interaction between surface loops on the myosin (especially loop 2) and actin (Geeves and Conibear 1995; Furch et al. 1998). This is followed by three events: release of phosphate from the motor, strong rebinding of the motor to the actin, and the force-generating powerstroke. As a result of these transitions, the motor arrives at a state where it is strongly bound to actin and ADP, and its lever-arm has changed position (AM⋅D). While it is widely accepted that phosphate release and force generation are linked, the order of these three events has not been unequivocally determined.
Phosphate release from myosin precedes ADP, which means that the phosphate requires a “back door” from which it can exit the nucleotide-binding pocket since ADP is blocking its exit (Rayment et al. 1993). In addition, in the pre-powerstroke state, switch I and switch II prevent the phosphate from dissociating. Thus, formation of an escape route for phosphate requires a structural rearrangement of switch I or switch II. This would either require a rearrangement of switch I that allows phosphate to dissociate while maintaining coordination of the ADP, or a rearrangement of switch II that does not alter the lever arm position (see (Sweeney and Houdusse 2010) for a discussion of these two possible mechanisms).
Two mechanisms that have been proposed for how phosphate release is coupled to the powerstroke are presented in Fig. 2b. In the top model, the motor undergoes a transition to a state with moderate actin-affinity in which the back door is open (A**M⋅D⋅P i ). Phosphate is then able to dissociate from the motor (A**M⋅D), resulting in a rearrangement of switch II that leads to lever arm movement as well as strong-binding to actin (AM⋅D). Arguments supporting this model are presented in Sweeney and Houdusse 2010.
According to the second model (Fig. 2b, bottom), the weakly bound motor transitions to a strong actin-binding state (AM⋅D⋅P i ) and then undergoes its force-generating powerstroke (AM’⋅D⋅P i ). This is followed by phosphate release (AM⋅D). Arguments supporting this model are presented in Takagi et al. 2004. It is important to note, however, that Takagi et al. argue that force-generation must precede phosphate release. As they point out, if generation of force occurs when the motor goes from its weak to strong actin-binding state (A*M⋅D⋅P i to AM⋅D⋅P i ), perhaps through a rotation of the myosin head, then it is possible for phosphate release to occur either before or after the powerstroke associated with switch II motion. According to this alternate model (not shown in Fig. 2b), force generation would occur in two steps.
A Second Conformational Change is Associated with ADP Release
Before releasing ADP, the motor must undergo an additional lever arm motion that is associated with a transition to a state that binds ADP weakly (AM⋅D W ). Evidence supporting this conformational change is reviewed in Nyitrai and Geeves 2004. Though it is uncertain whether this conformational change can contribute to work production, it is thought that the force sensitivity of this step may be important to the force-dependent function of some myosins (see the discussion below).
At this point, the lever arm has returned to its post-powerstroke conformation. The β-sheet returns to its twisted state, again separating switch I and the P-loop, resulting in loss of coordination of the ADP. The motor releases ADP and returns to its rigor state (AM).
Thermal Ratchets and Powerstrokes
In the description of the chemomechanical cycle above, deterministic language has been primarily used to describe how myosin generates motion. The lever arm is envisioned as an elastic element, and the powerstroke consists of a relaxation of a strained conformation of the lever arm (similar to the contraction and relaxation of a stretched rubber band). In other words, the molecular motor is viewed as a “miniature version of a macroscopic device, employing springs, cogs, levers, and the like to effect motion” (Astumian and Derényi 1998).
However, there is an important difference between macroscopic and microscopic motors. Microscopic motors operate in an environment where viscosity dominates inertia and where random thermal fluctuations are significant. According to the powerstroke mechanism described above, thermal noise is viewed as something that must be overcome to create motion, and in the absence of this noise, the motor’s powerstroke could still occur.
An alternate mechanism envisions the motor as a thermal ratchet (or Brownian ratchet). A thermal ratchet harnesses (rather than fights) thermal fluctuations to generate its motion. According to such a model, the ATPase activity serves the purpose of rectifying this random motion. Thus, in the absence of thermal noise, the motor would be unable to generate motion.
Andrew Huxley presented the first thermal ratchet model of myosin (Huxley 1957). More recently, a possible thermal ratchet mechanism was described in which the motion of the lever arm is envisioned as a means of capturing Brownian motion rather than a means of directly generating motion (Houdusse and Sweeney 2001). Although this review will not go into the details of these models, discussions of the relationship between powerstroke and thermal ratchet models can be found in (Wang and Oster 2002; Howard 2009; Astumian 2010). It is also worth noting that the current model of myosin V motility, which is discussed in detail below, contains elements of both a powerstroke and thermal ratchet model.
Differences Between the Myosin Classes
Although all myosins undergo the same chemomechanical cycle, different myosins must be able to perform disparate cellular functions. Variability in the structure and kinetics of the various classes allows different myosins to use the common chemomechanical cycle to perform different tasks. Some of the ways in which different myosins vary and how this variability impacts their function are highlighted below.
Structural Diversity Among the Myosin Classes
Conserved Domains within the Tail
Because the tail domain is the region of the motor that varies the most among the myosin classes, many of the differences in motor function can be attributed to differences in this region. Tails typically contain a number of conserved domains that allow for protein-protein and protein-lipid interactions, as well as additional functions such as kinase activity. Tail domain diversity, and its relation to myosin function, is reviewed in Krendel and Mooseker 2005.
For a myosin to move cargo along an actin filament, it must undergo numerous steps along the actin without dissociating, a capability known as processivity. Because it is not expected that a single motor can processively move along actin (with the notable exception of myosin IX (Inoue et al. 2002; Post et al. 2002; Liao et al. 2010)), the tail domains of many myosins contain self-assembly domains (typically coiled coil forming domains) that bring together two myosin motors. Dimeric motors have been observed to move processively along an actin filament in a hand-over-hand fashion (e.g., see Churchmann et al. 2005). Dimerization may also be regulated for some myosins. For example, myosin VI is normally monomeric, and yet it forms a dimer when bound to its cargo (Llinas et al. 2012). Thus, the motor is activated to move processively only when it is needed.
Lever Arm Length
According to the lever arm model, the length of the neck domain of a myosin is proportional to its step size and to the rate at which it moves along actin, a prediction that has been verified by a variety of experiments (reviewed in Sun and Goldman 2011). Lever arm length plays an important role in the processive stepping of the two-headed processive motor myosin V. Because of the relatively long neck domain of this motor, which binds six light chains, the two motor domains are able to span the 36 nm pseudo-repeat of an actin filament. Thus, the motor can translocate along a single face of actin without rotating around the filament (Sun and Goldman 2011).
Direction of Motion
Actin is a polarized filament, and most characterized myosins move along the filament toward the so-called (+) end of the filament. Myosin VI is the only motor known to move toward the (−) end, which allows it to achieve unique functions from other myosins. How reverse directionality is achieved is described in Llinas et al. 2012; some of the details are discussed here. Interestingly, much of the motor domain of myosin VI is similar to other (+) end directed myosins. Reverse directionality primarily arises through an insert (insert-2) that is located between the converter and the lever arm. Insert-2 forms a novel calmodulin-binding motif, and the bound light chain interacts with the converter, stabilizing a new position of the lever arm. In addition, in its pre-powerstroke state, the converter adopts a unique fold that allows the lever arm to rotate by 180° during its powerstroke (instead of the ∼70° rotation seen with myosin II). Thus, myosin VI still exhibits the same chemomechanical cycle as other myosins and conforms to the lever arm model, though it exhibits a unique swinging of its lever arm due to these structural modifications.
Kinetic Tuning Among the Myosin Classes
The rates of transition between the states can vary greatly for different myosins, thus affecting the lifetimes of the various states. In this way, myosins are “kinetically tuned” to achieve their disparate cellular functions. A detailed discussion of how kinetic tuning dictates myosin function can be found in Bloemink and Geeves 2011. Some of the important properties of a myosin’s chemomechanical cycle that are tuned in this way are described below:
Average Cycle Time (tc)
Myosins vary in how fast they pass through their chemomechanical cycles, which affects the rate at which an individual motor can move (or move along) actin. If the average time for a motor to complete its cycle is t c , then the average speed that a single motor can move along actin is d/t c , where d is the step size of the motor.
Average Strongly Bound Time (ts)
When motors work in ensembles, the rate at which they can move actin is determined not by t c , but by the average time that the motor remains strongly bound to the actin during a single chemomechanical cycle, t s . It is thus not surprising that the fastest known myosin, Chara corallina myosin XI, exhibits an extremely small value of t s (Ito et al. 2007). As another example, t s of smooth muscle myosin II is much larger than for skeletal myosin forms, resulting in a slower rate of contraction and increased force-generation for smooth muscle (Guilford et al. 1997).
A motor’s duty ratio is defined as the fraction of time that the myosin spends strongly bound to actin during a single cycle (t s /t c ). In general, a low duty ratio motor spends most of its cycle occupying the low actin-affinity states, and a high duty ratio motor spends most of its cycle occupying the high actin-affinity states. For a population of motors, the duty ratio is also the steady-state fraction of actin-bound molecules.
For skeletal muscle myosin II, which works in large ensembles within a sarcomere, the slowest step in its chemomechanical cycle is actin-activated phosphate release. Thus, the motor has a low-duty ratio, which minimizes the drag resulting from multiple motors remaining bound to a filament during contraction. For the two-headed myosin V, on the other hand, the slowest step is ADP release from actomyosin. Thus, the motor has a high duty ratio, enabling the motor’s processivity since it is unlikely that both of the motor domains will detach from actin at the same time.
Force-Dependent Kinetic Tuning of Myosins
Force applied to a molecular motor has the effect of specifically perturbing mechanical transitions within the chemomechanical cycle. Thus, force affects the kinetic rates of those transitions involving conformational changes, potentially altering t s , t c , and the duty ratio. A striking example of force-dependent kinetics is seen with a mammalian myosin 1b. A backward force on this motor slows the apparent rate of ADP release. This is thought to arise because the applied force affects the conformational change between the strong ADP-binding (AM·D) and weak ADP-binding states (AM·D W ). As a result of this force-sensitivity, the duty ratio of this motor changes from <0.2 in the absence of force to >0.9 in the presence of 2 pN of backwards force (Laakso et al. 2008). This allows the motor to move its cargo in the absence of force. In the presence of force, however, the motor’s kinetic cycle is more appropriate for maintaining tension and anchoring its bound cargo to actin.
Another interesting example of force-dependent kinetic tuning comes from the processive motion of myosin V. As discussed earlier, both the dimerization domain and the high duty ratio of myosin V are key to the motor’s processivity. However, the average number of consecutive steps observed for a myosin V moving along actin is much greater than would be expected given its duty ratio (Sellers and Veigel 2006). This discrepancy has led to a model of kinetic “gating,” in which the two motor domains coordinate their chemomechanical cycles to ensure that the trailing head of a walking motor passes through its ATPase cycle before the lead head releases from the filament.
In Fig. 4b, a model for the myosin V kinetic pathway is presented that includes a gating mechanism. This model is presented and discussed in Dunn and Spudich 2007. The motor starts in a state with both heads bound to actin and ADP. The trailing head is in a post-powerstroke conformation. The lead head, however, is prevented from undergoing its powerstroke due to intramolecular forces, and is thus shown in a strained, curved state. As a result of this backward force, the chemomechanical cycle of the lead head is slowed, preventing it from releasing from actin before the trailing head.
The rear head then releases its ADP and binds ATP, causing it to dissociate from actin. This removes the intramolecular strain, allowing the lead head to undergo its powerstroke, which moves the motor’s center of mass forward by ∼20 nm. The unbound motor then hydrolyzes ATP and undergoes a diffusional search until it finds its next actin-binding site, moving the center of mass an additional ∼15 nm. It then binds to actin, releases phosphate, and undergoes its powerstroke, returning to the initial strained state after having completed a single ∼35-nm step.
Alternative models have been proposed for the myosin V stepping mechanism. In particular, the state of the gated lead head of the motor has not been unequivocally determined. For example, a model has been proposed in which the lead head is kinetically stalled in a pre-powerstroke conformation with bound ADP and phosphate (De La Cruz and Ostap 2004).
To generate motion and force, myosins must coordinate chemical and mechanical events. Although different myosin classes may perform vastly different cellular functions, all appear to pass through the same states during their chemomechanical cycles. Variability among the different classes results from structural and kinetic differences that allow these motors to utilize the same kinetic cycle to perform distinct roles.