Shaking the Myosin Family Tree Biochemical kinetics defines four types of myosin motor

Abstract

Although all myosin motors follow the same basic cross-bridge cycle, they display a large variety in the rates of transition between different states in the cycle, allowing each myosin to be finely tuned for a specific task. Traditionally, myosins have been classified by sequence analysis into a large number of sub-families (~35). Here we use a different method to classify the myosin family members which is based on biochemical and mechanical properties. The key properties that define the type of mechanical properties of the motor are duty ratio (defined as the fraction of the time myosin remains attached to actin during each cycle), thermodynamic coupling of actin and nucleotide binding to myosin and the degree of strain-sensitivity of the ADP release step. Based on these properties we propose to classify myosins into four different groups: (I) fast movers, (II) slow/efficient movers, (III) load-bearers and (IV) gates.

1. Introduction

Myosins comprise one of the three major families of molecular motors and together with kinesins and dyneins, are responsible for almost all large scale movement within and by eukaryotic cells. The most familiar of the myosins is myosin II, which is the major motor protein of muscle. However, the broader family of myosin motors is involved in a wide range of movement and transport processes (e.g. vesicle transport, phagocytosis, and cell division[1]. Yet other myosins are believed not to be involved in transport at all but to generate and sense mechanical forces in the cell [2]. We are now at the point where the study of the dazzling variety of behaviours and functions of different myosins is beginning to allow the underlying principles to emerge of how a prototypical myosin can be adapted for a myriad of different functions. Here we attempt to define some of these principles with the aim of developing an understanding of how myosin motors have been adapted for specific mechanical properties that allow them to fulfil their cellular role.

Most eukaryotic cells express more than a dozen myosin isoforms from the different family groups but the specific function of each myosin and its regulation within the cell is, in many cases, poorly defined. In the last ten years there has been huge progress in mapping the members of the myosin family (now ~ 35 subgroups[3]), exploring the cellular function of the different myosins, unravelling their regulation and defining the underlying molecular mechanism (see reviews in [4]). However the observation that a specific myosin (from one of the ~35 subgroups) is present at a specific location in a cell and at a specific time reveals relatively little about its detailed function. If we can define the type of mechanical activity a specific myosin is capable of, then it will help predict the cellular function of that myosin. The aim ought to be not to simply define these properties for each individual myosin but to understand the underlying principles that will allow the behaviour of the family as a whole to be defined.

We propose four types of myosin based on function: Fast movers (typified by vertebrate fast muscle myosin II), efficient movers & load bearers (vertebrate slow and smooth muscle myosins), strain sensors (mammalian myo1c, myo1b) and gated & processive motors (mammalian myo5a , myo5b). These four functional groups cut across the classification of myosin motors based on sequence analysis as different types of function can be found within a single structural class i.e. myosin classes I, II & V contain examples of more than one of the four putative functional groups (See Table 1). To understand how these different properties are produced we will examine the biochemical properties of myosins.

Table 1.

	Duty Ratio	K’ADP (μM)	ADP release k’−AD (s−1)	Thermodynamic coupling K’AD/KD ( KDA/KA)	KCO	Myosin Type	Reference
Myosin-I
myo1a(acam)	0.05	93	-	-	>10	fast mover	[1, 2]
myo1d(dicty)		75	-	40	>10	fast mover	[3]
myo1b(acam)		53	84	-	>10	fast mover	[1]
myo1b(dicty)		47	233	118	>10	fast mover	[4]
myo1c(rat)		17.2/1.6a	20.6/1.9a	−/0.64a	≤0.1	strain sensor	[5]
myo1b(rat)		10/3a	8/3.5a	5	≤0.1	strain sensor	[6, 7]
myo1e(dicty)		12	30	2	nd	strain sensor	[8]
myo1e(hum)		3.9	93	8.9	≤0.1	strain sensor	[9]
myo1a(BBM)	0.05-0.1	2.6	8	-	0.6	strain sensor	[2]
Myosin-II
IFI(drosy)		409		55	>10	fast mover	[10]
myo2(dicty)		215	>100	15	>10	fast mover	[11]
fast sk (rab)	0.04	120	>500	30-60	>10	fast mover	[12, 13]
slow sk (bov)		9.6	94	4.8	5.3	slow/efficient mover	[14]
βcardiac (bov)		6.7	65	15-20	-	slow/efficient mover	[15]
soleus (rab)		11.3	58	21	0.3	slow/efficient mover	[16]
sm (chicken)		5	22	4.2	-	slow/efficient mover	[17]
nm2a(hum)	0.1	0.8	2	0.7	≤0.1	strain sensor	[18]
nm2b(hum)	0.3	0.15	0.35	0.2	≤0.1	strain sensor	[19]
Myosin-III
myo3a(hum)	0.9	11.1	6.9	0.7		strain sensor	[20]
myo3a(hum)	0.9	7-36	40	0.5	1.8	strain sensor	[21]
Myosin-V
myo5a(chick)	0.7	0.93	12-16	3.3		gate/processive	[22]
myo5b(hum)	0.8	3.7	12.2	-		gate/processive	[23]
myo5b(dicty)	0.2/0.8c	8	27.9/8c	5.1/1.6c		gate/processive	[24]
myo5c(hum)	0.1	0.3	15.7	1.5		strain sensor	[25]
myo5c(hum)	0.3	2.1	12.7	1.6		strain sensor	[26]
myo5(drosy)	0.1	32	150	1.9		strain sensor	[27]
Myosin-VI
myo6(pig)	0.9	8.8	5.5	1.5	≤0.1	gate/processive	[28]+ Robblee jbc2004
Myosin-VII
myo7a(drosy)	0.9	0.4	2.2	0.7	0.9	gate/processive	[29]+ Yang et al pnas2006
myo7b(drosy)	0.8	2-4	9.7	0.7-1.6	3.2	gate/processive	[30]
myo7b(mse)	0.8	0.37	7.0	1.8	0.22	gate/processive	[31]
Myosin-X
myo10(bov)	0.6-0.7	8.8	23	7.3		gate/processive	[32]
myo10(bov)	0.2	8.4	18	6.2			[33]
Myosin-XI
chara(c.coral)	<0.3	260	>2800	200	>10	fast mover	[34]
myo11(n.tab)	>0.8	42	95			gate/processive	[35]

^awith Ca²⁺/without Ca²⁺ ;

^bwithout actin/with actin ;

^clow [Mg²⁺] / high [Mg²⁺ ;

2. The Cross-bridge Cycle

The myosin family’s key biochemical characteristic is the ability to undergo a cyclical interaction with the actin filament, driven by ATP hydrolysis, to produce force and movement. All members of the family so far studied undergo the same ATP dependent biochemical cycle (reviewed by Geeves & Holmes [5] Sweeney & Houdusse [6]) but with large variations in the rate and equilibrium constants for the different events round the cycle. This leads to large changes in the life-time of the individual states in the cycle and the fraction of the total cycle time spent in each state. Since the mechanical properties of the states in the cycle vary, different types of motors can be generated by altering the life times of different states.

Myosins differ in the V_max of the ATPase and the velocity with which they can move along (or move) actin, they also differ in the duty ratio (the fraction of the ATPase cycle time spent attached to actin [7]). Here we emphasise how the myosins differ in the efficiency with which actin can displace ADP from myosin (thermodynamic and kinetic coupling), the extent to which the myosin can generate movement vs force; and the extent to which the cycle is dependent upon external mechanical load (strain sensitivity). In combination these characteristics define the type of mechanical activity of each myosin motor. Thus, the biochemical kinetics of a myosin can define the type of mechanical activity the motor undertakes and therefore help to predict the cellular function of the myosin. This will allow biochemical kinetic analysis to assign novel myosins to functional sub-groups.

We propose there are three primary signatures that in combination define the type of mechanical activity of the motor. The basic ideas have been outlined before ([7-8]; here we summarize and update the arguments and show how this leads to the definition of the four types of motor activity. These signatures are the thermodynamic and kinetic coupling of actin and ADP binding to myosin, the load sensitivity of the ADP release steps and the duty ratio. Each of these signatures is influenced by an isomerisation step that controls ADP release: the opening of the nucleotide pocket K_CO, (see step 5 in Fig 1). Fig 2 shows the relationship between these kinetic signatures and the 4 types of motor activity. We will first outline the basic cross-bridge cycle and then explore how the kinetic signatures can define the type of mechanical activity.

Fig 1. The ATP driven actin-myosin cross-bridge cycle.

The myosin cross-bridge is shown as consisting of 3 major parts. 1. The central core of the myosin head is shown in red (filled circle) with the lower 50k domain projecting up to make contact with one actin monomer (grey circles). The lower 50k domain is one half of the major cleft that splits the actin binding site. The P-loop that binds to the γPi of ATP is part of the central core. This red segment is drawn as a fixed reference point during the cycle. 2. The upper 50 K domain with SW-1 (switch 1) is shown in yellow. This is drawn as two parts; the upper jaw of the major cleft that splits the actin binding site and a yellow ring with SW-1 projecting into the center. 3. A blue ring represents the relay loop with SW-2 projecting to the center while the converter domain and the lever arm (shown with 2 light chains attached) is projected out from the ring. The nucleotide pocket is shown as lying on top of the central core (red) bounded by two circles one part of the upper 50 K domain (yellow) the other part of the relay/converter domain (blue). Each ring has a segment missing to represent the entrance to the nucleotide pocket.

A·M at the top left represents the rigor actin-myosin complex with the upper 50kDa domain cleft closed to allow both sides of the cleft to make contact with actin. The nucleotide pocket entrance is open with both SW1 & 2 open.

Step 1 ATP binds to myosin and myosin dissociates from actin. In detail ATP binds into the nucleotide pocket and the yellow segment rotates to close SW-1 onto the ATP in doing so the major cleft opens destroying the actin binding site leading to dissociation from actin.

Step 2 The recovery stroke and ATP-hydrolysis. The blue segment rotates to bring SW-2 into contact with ATP thus rotating the converter/light-chain binding domain to complete the recovery stroke or repriming of the motor while detached from actin. Only after both SW1 and SW2 are closed is ATP hydrolysed to form the stable M.ADP.Pi complex.

Step 3 Actin rebinding and the power stroke. After hydrolysis the lower 50kDa part of the cleft rebinds to actin, the cleft closes and both upper and lower 50kDa domains bind actin, this involves a rotation of the yellow segment which triggers both Pi release and the rotation of the blue segment generating a force – represented as a distortion of the converter domain. The exact order of cleft closure, Pi release and power stroke remains under debate. The location of the “elasticity” within the cross bridge is not defined.

Step 4 “Sliding”. Provided the force generated in step 3 is large enough the load is moved 5-10 nm by the cross bridge working-stroke. This is shown as a relative displacement of the end of the lever relative to the actin.

Step 5 Opening of the nucleotide pocket. Once the strain in the converter is dissipated by the sliding movement, the blue segment rotates further to open the nucleotide pocket and allows ADP to escape in step 6. An efficient motor requires that ADP release be limited until the cross-bridge has completed its movement or working stroke.

Step 4a/5a. The strain holding cross-bridge. If the force generated in step 3 is too small to move the load no sliding takes place. The further movement of the blue segment against the load is strongly inhibited (~5-100 fold for different myosins). The degree to which this rotation is inhibited in a load bearing myosin defines the load sensitivity of the myosin. This is a function of the size of the rotation, the length of the lever arm and the stiffness of the “spring” in the structure.

Figure 2. Four groups of myosin function: fast movers, slow/efficient movers, strain sensors and gated/processive movers.

For each group the correlation with the three kinetic signatures (duty ratio, K_AD/K_D and load-dependence) is indicated, together with the isomerisation constant K_CO. K_CO is the equilibrium constant for the step that opens the nucleotide binding pocket before ADP can be released. Fast movers, such as fast muscle myosin-II, have a low duty ratio and load-dependence and a high thermodynamic coupling ratio (K_AD/K_D) and K_CO value. Gated myosins, such as myosin 5a, have a high duty ratio and load-dependence but low values for the thermodynamic coupling and K_CO. The diagram is drawn to emphasise the graded nature of the change in the parameters across the series. There is not a step change in the values and therefore the boundary between each type of motor function is not exact.

The basic cross-bridge cycle is shown in Fig 1 and is the simplest version of the cycle that is suited to our discussion. Every myosin studied to date (that has an active ATPase), goes through the same cycle and the figure legend outlines the major events in each step of the cycle. In many cases the steps shown consist of two or more underlying molecular events e.g. Step-1 involves ATP binding, an induced change in conformation of the myosin (switch 1 closure and cleft opening) followed by rapid actin dissociation. The simpler version is used here and the reader is referred to more specialist reviews for details [5-6].

Step 5 in the cycle is of particular interest here as it has not been characterised for all myosins. It represents an isomerisation of the A.M.D complex that is required before ADP can be released from the nucleotide pocket. The equilibrium between these two A.M.D states is well characterised for some slow/efficient movers, strain sensors and gates/processive myosins (see references in Table 1). We propose that this isomerisation is a universal step in all myosins, as shown in Fig 1. The step involves a change in structure that is opposed by any load on the head. In the absence of load the equilibrium between the two states varies for different myosin types.

3. Myosin Functional Signatures

3.1 Thermodynamic & kinetic coupling between ADP and actin binding

The requirement for rapid release of ADP from fast motors tends to make actin binding very efficient at displacing ADP from fast myosins i.e. the ratio of ADP affinity for actomyosin (K_AD) to the ADP affinity for myosin (K_D) is high (K_AD/K_D > 20)[8]. For processive myosins and strain-sensing myosins, this ratio tends towards a value of 1 and slow load-bearing myosins have a K_AD/K_D ratio around 5. The kinetic coupling k_−AD/k_−D (ratio of rate constants of ADP release from actomyosin and myosin) tends to have higher values but the same relative order >100 for fast myosins, ≤ 10 for processive motors [8]

For myosins other than the fast motor group, ADP release is observed as a two-step process as shown in Fig 1 and in more detail in Scheme 1 below

Scheme 1

In this model the actin-myosin complex must first isomerise to allow ADP to be released and the apparent affinity of ADP for A.M and thermodynamic coupling is controlled by the isomerisation step K_CO. As discussed by Nyitrai & Geeves the apparent ADP affinity K’_ADP, for this two step system is defined as:

$${\mathrm{K}}_{\mathrm{ADP}}^{’}={\mathrm{K}}_{\mathrm{CO}}{\mathrm{K}}_{\mathrm{ADP}}∕(1+{\mathrm{K}}_{\mathrm{CO}})$$

Thus if K_CO is >>1 then K’_ADP = K_ADP and if K_CO is <<1 the K’_ADP = K_COK_ADP.

The actual ADP release equilibrium constant K_ADP is weak in all cases (~100 μM). Thus the value of K_CO is the driver of the ADP affinity and k_+CO is rate-limiting for ADP release. A large value of k_+CO results in very fast ADP release, a large value of K_CO and strong coupling of actin and ADP binding.

Fast myosins are predominantly in the open pocket form and therefore have a large value of K_CO (≥ 20, equilibrium towards the right) and a significant free energy change associated with the step. For slow/efficient movers, strain sensors and gated/processive myosins, a lower value of K_CO (< 10) means that the acto-myosin-ADP complex is now in both the closed and open pocket form and the small value of the equilibrium constant means there is little free energy change associated with the step. It can therefore be influenced by other factors. In fact, a classic gated myosin would be predominantly in the closed form (K_CO < 1) and require an additional signal to allow ADP to be released. Thus, the thermodynamic coupling is controlled by the value of K_CO. In fast movers, actin binding can rapidly open the pocket and displace ADP. For the other groups ADP release is tuned to the mechanical activity (see strain dependence below). In kinetic terms the important factor is k_+CO , which controls how fast actin induces displacement of ADP.

3.2 The strain sensitivity of the ADP release step

Nyitrai and Geeves [8] argued that the load dependence of ADP release is central in defining the type of mechanical activity a myosin generates. All myosins require a load dependent event during the attached part of the cycle to prevent detachment of the cross-bridge before the working stroke (step 4 & 5 in Fig 1) is complete. This ensures efficient coupling of the energy from ATP hydrolysis with the work done by the motor. This is most easily visualised as a mechanism to keep ADP in the nucleotide pocket until the working stroke is complete. As depicted in Fig 1, step 5 is required to allow ADP to escape from the nucleotide pocket and involves an additional swing of the lever arm in the same direction of the working stroke. Thus if the working stoke is incomplete, the lever arm will be carrying a load and the additional swing will be inhibited.

There is considerable evidence that ADP release induces a change in the position of the lever arm of actin-myosin which could generate a movement similar to and in the same direction as that of the power-stroke. [9-11] [12] The thermodynamics of ADP release coupled to the conformational change in step 5 makes it unlikely in most cases to be an additional power- or working-stroke but the thermodynamics does predict a strain-sensing mechanism [2, 13]. Myosins that have been identified with a well-developed ADP associated stroke (myosin 1a/b/c, MV) are ones that operate as tension sensors rather than rapid motors. A load-dependent 2^nd mechanical step in single molecule optical trap studies was consistent with this view. [14-17] S Studies using improved methods in both the optical trap [18], and in X-ray diffraction [19] Iwamoto have broadened the range of myosins which show this ADP-linked mechanical event, consistent with this being a universal process. The load dependence of ADP release of a given motor will depend upon 3 major factors: the equilibrium constant of the isomerisation step K_CO (step 5, defining the free energy change associated with the isomerisation), the length of the lever arm and the stiffness of the “elastic element” bearing the load. Nyitrai & Geeves discussed in more detail the relationship between the lever arm, the elasticity and K_CO.

3.3 The duty ratio

The duty ratio defines the fraction of time a myosin motor spends attached to actin during each ATP hydrolysis cycle. Fast myosins have a low duty ratio (0.05-0.1) and work in groups to produce effective movement. In contrast processive myosins have a large duty ratio (>0.7) allowing a single two-headed myosin to walk along an actin filament and dragging a cargo without detaching from the actin. Other myosin motor types are intermediate between the fast and processive myosins and some motors can adjust their duty ratio in response to the load or to other cellular signals, such as changes in calcium [20], magnesium[21-22], as well as the well know regulation signals like phosphorylation-state or actin concentration. The duty ratio is controlled by the net rate constants defining the detachment of myosin from actin and reattachment of myosin to actin. Detachment of myosin from actin is controlled by ATP binding to the actomyosin cross-bridge (Figure 1, step 1) or by ADP release (Figure 1 step 5/6). For fast myosins, ADP-release was found to be fast and too fast to limit the shortening velocity (k_−AD > 1000 s⁻¹) whereas ATP-binding and subsequent dissociation from actin was of the right magnitude to control the shortening velocity [23-24]. For slow and efficient moving myosins the maximum shortening velocity is limited by ADP release or the isomerisation controlling ADP release.[24-25] These slower load-bearing myosins, in contrast to fast myosins, have a longer-lived A.M.ADP state which can bear load but slower ADP release slows motility. The equilibrium between the two AM.ADP states defined by step 5 in Fig 1 is shifted more towards the closed state for the strain-sensors and processive myosins thereby increasing the life-times of the AM.ADP state and thus enhancing the load-bearing properties and increasing the duty ratio.

The lifetime of the actin-detached part of the cycle is limited by the repriming step which occurs during the detached part of the cycle and is coupled to the ATP hydrolysis step (see Figure 1, step 2). For fast myosins the ATP hydrolysis step is > 50 s-¹ whereas the slower load-bearing myosins maintain a low duty ratio by reducing the rate constant of the hydrolysis step in-line with the ADP release rate constant. [24-25] Processive motors and sensors in contrast have a very fast hydrolysis step compared to their ADP-release rates, in order to reduce the detached part of the cycle to a minimum.[26] , Henn, Yang Taft)

4. Assignment of myosin to motor functional groups

Based on their kinetic signatures, we can now begin to assign myosins that have been studied to date, to one of the four types of myosin function groups, as shown in Table 1. These assignments are provisional in that not all of the myosins have each of the signatures defined. The Table does not include load sensitivity as this has been measured for only a small number of myosins, (non-muscle MII, MV, MVI, M1c [15-17, 27]). The values used to define for the boundaries between groups are to some extend arbitrary. For example the duty ratio has a small value (<0.5) for both fast movers and the slow/efficient force holders and they may not be distinguished on this category. Gates/processive motors to have large duty ratio >0.7 whereas for strain sensor the value may be more ambiguous. They may well cycle with a low duty ratio until they reach a stall force. K_CO values are large for fast motors (≥10) and are rarely defined. We have therefore included the K’_AD and k’_−AD values and these are expected to be >50 μM and >100 s-1 respectively.

For several myosins the duty ratio and the thermodynamic coupling measured may not be an absolute value but dependent upon calcium or magnesium concentration, on the local actin concentration or on the load. We have quoted values wherever possible at the conditions close to the unloaded resting cellular conditions i.e. low calcium, modest free magnesium and 0.1 −1 mM actin. Where possible we have indicated the effect of calcium or magnesium.

Table 1 shows that class-I myosins contain both slow/efficient movers and strain sensors. The slow movers in class-I have weaker ADP-affinity and larger thermodynamic coupling ratios compared to the strain sensors, but all have a low duty ratio. The myosin-I class has been divided into 2 subclasses [7] or 4 subclasses[28-30]and this 2 subclass division agrees with our assignment of myosin class I into fast movers (subclass-1) and strain sensors (subclass-2).

The myosin-II class shows three different myosin types: fast movers, slow/efficient movers and strain sensors. The fast movers (fast muscle myosin-2 mammalian myosin isoforms 2a/2b/2x, Drosophila IFM myosin and myosin-2 from Dictyostelium) have very weak ADP affinity (K_AD > 100 μM and K_CO >> 1), fast ADP-release which is measured as a single phase (k_−AD and k_−CO if present > 100 s⁻¹), a large thermodynamic coupling constant (K_AD/K_D > 10) and a low duty ratio (0.04). The isomerisation step K_CO is too fast to measure for the fast movers and ADP release is measured as a single phase. For the slow/efficient movers, such as slow skeletal or cardiac isoform 1, and smooth muscle myosin, the ADP-release step is biphasic and K_CO can be measured. The non-muscle myosin IIA/IIB can be assigned to the strain-sensors since they have very tight ADP affinity, a very small thermodynamic coupling (K_AD/K_D <1) and low but increased duty ratio (0.1-0.3). The class-V myosin group has both strain sensors (myosin-5c) and processive myosins (myosin5a/5b) and the main difference is the large duty ratio found for myo5a/b (>0.7) compared to myo5c (0.1-0.3). Class XI myosins contain two myosin types which are very different, based on their kinetic signatures with myosin-XI from alga Chara corallina (Chara) being a fast mover and myosin-XI from Nicotiana tabacum (n.tab) a processive mover. Both myosin-XI members are extremely fast, however Chara myosin has a low duty ratio and very weak ADP affinity whereas myosin-XI from n.tab is a processive myosin with a high duty ratio.

The definition, in kinetic and thermodynamic terms, of the behaviour of different groups of the myosin family may allow us to correlate the mechanical function with specific sequences of the myosins. Currently the protein structure or specifically the sequence motifs that generate the different functional properties of myosins are not well defined. For some isforms Loop 1 at the entrance to the nucleotide pocket, or Loop 2 at the actin binding site have been shown to be important effectors of myosin properties but this is by no means a universal mechanism even within a single myosin group. Some myosin IIs ( eg vertebrate smooth muscle , scallop catch and stated muscle) change isoform properties by alternately splicing loop 1, others (eg Drosophila) produce numerous alternately spliced isoforms but not change Loop1 at all.

As discussed above, the motor activities of the myosins do not correlate with the simple assignment to family groups. Myosin groups I, II & V each have members with both high and low duty ratios, high and low thermodynamic coupling between nucleotide and actin binding, and fast moving vs motors that move very slowly if at all and may operate as strain sensors within the cytoskeleton. If the family grouping indicates the evolutionary history of the myosin family then this observation implies that mechanical activities such as duty ratio and thermodynamic coupling have evolved separately within the family groups. I.e there are both divergent and convergent elements within the family history.