Structure and function of myosin filaments

Structure and function of myosin filamentsRoger Craig and John L Woodhead

Myosin filaments interact with actin to generate muscle

contraction and many forms of cell motility. X-ray and electron

microscopy (EM) studies have revealed the general

organization of myosin molecules in relaxed filaments, but

technical difficulties have prevented a detailed description.

Recent studies using improved ultrastructural and image

analysis techniques are overcoming these problems. Three-

dimensional reconstructions using single-particle methods

have provided many new insights into the organization of the

myosin heads and tails. Docking of atomic structures into cryo-

EM density maps suggests how regulated myosin filaments are

‘switched off’, bringing about muscle relaxation. Additionally,

sequence analysis suggests probable interactions between

myosin tails in the backbone, whereas crystallographic and EM

studies are starting to reveal tail interactions directly in three

dimensions.

Addresses

Department of Cell Biology, University of Massachusetts Medical

School, 55 Lake Avenue North, Worcester, MA 01655, USA

Corresponding author: Craig, Roger ([email protected])

Current Opinion in Structural Biology 2006, 16:204–212

This review comes from a themed issue on

Macromolecular assemblages

Edited by Edward H Egelman and Andrew GW Leslie

Available online 24th March 2006

0959-440X/$ – see front matter

# 2006 Elsevier Ltd. All rights reserved.

DOI 10.1016/j.sbi.2006.03.006

IntroductionMyosin filaments (also called thick filaments) are key

components of muscle and non-muscle cells. In striated

muscle, they overlap with thin (actin-containing) fila-

ments in an orderly array, making a repeating pattern

of sarcomeres, the basic units of contraction [1]

(Figure 1a). In smooth and non-muscle cells, myosin

and actin filaments also form overlapping arrays, but their

organization is less ordered and more labile. Myosin

filaments play two key roles in muscle contraction and

cell motility. The myosin heads (or crossbridges), which

lie on the surface of the filaments, bring about contraction

by cyclic interaction with actin subunits in the thin

filaments. This ‘crossbridge cycle’ causes the thick and

thin filaments to slide past each other, producing move-

ment [2,3��]. Thick filaments in many types of muscle

and non-muscle cells also participate in regulating or


modulating contractile activity. Understanding how thick

filaments are assembled and carry out these functions

requires a detailed knowledge of their structure.

The structure of thick filaments has been elucidated over

the past 40 years by electron microscopy (EM), X-ray

diffraction and other techniques. Although this has pro-

vided many insights into the principles of myosin assem-

bly, critical molecular details have been missing and the

relationship of filament structure to function has been

unclear. Here, we review the key advances that have led

to our current understanding of thick filament structure

and its relation to function, with an emphasis on recent

progress. We focus on myosin organization in striated

muscle filaments in the relaxed state, in which the fila-

ment structure is stable and unperturbed by interaction

with actin. We also discuss myosin filaments in smooth

and non-muscle cells, and the role of myosin-binding

proteins in thick filament assembly and function.

Myosin molecules and thick filamentsThe major component of thick filaments is myosin II, a

member of the myosin superfamily of motor proteins that

produce motility by moving actin [4��]. Myosin II is an

elongated, two-headed molecule consisting of two iden-

tical heavy chains and two pairs of light chains [5]

(Figure 1d). The C-terminal half of each heavy chain

is a helical, whereas its N-terminal half folds into a

globular head region. The heads (also called subfragment

1 or S1) consist of a motor domain, which hydrolyzes ATP

and binds to actin, and a neck region (or light chain

domain), which contains a regulatory light chain and an

essential light chain, and functions as a lever responsible

for filament sliding in contraction [3��]. In many muscles,

the light chain domain regulates (or modulates) contrac-

tile activity, either by phosphorylation of the regulatory

light chains or by binding Ca2+ [6]. The a helices coil

around each other, forming a long coiled-coil tail that is

insoluble at physiological ionic strength.

Thick filaments are formed by the staggered association

of myosin tails running approximately parallel to the

filament axis. The distal (C-terminal) two-thirds of the

tail (light meromyosin [LMM]; Figure 1d) is responsible

for self-association and is tightly packed in the filament

backbone. The more soluble, proximal one-third of the

tail (subfragment 2 or S2) acts as a flexible link between

LMM and the heads, which lie on the filament surface,

where they can interact with actin. Non-myosin proteins,

usually in small amounts, may be incorporated into the

filament core, or serve structural or regulatory roles on the

filament surface.

www.sciencedirect.com

mailto:[email protected]

http://dx.doi.org/10.1016/j.sbi.2006.03.006

Myosin filament structure Craig and Woodhead 205

Figure 1

Myosin molecules and thick filaments. (a) Striated muscle sarcomere showing overlapping arrays of thick and thin filaments. Thick filaments are

connected at the M-line and thin filaments at the Z-line. The sarcomere shortens (bringing about contraction of muscle as a whole) by sliding of the thin

filament arrays towards each other, increasing their overlap with thick filaments. (b) Central portion of the thick filament model showing the helical array

(red dashed line) of pairs of myosin heads on the filament surface (each pair is represented by a yellow sphere). The green backbone contains myosin

tails. The model represents a filament with fourfold rotational symmetry. The axial distance between adjacent levels of heads is 14.5 nm and, in this

model, the helical repeat (43.5 nm) is three times the 14.5 nm spacing. (c) Schematic representation of the assembly of myosin molecules in a bipolar

thick filament. The head ends of the molecules point away from the filament center in each half. Tails have antiparallel and parallel overlaps in the

central bare zone (free of heads), but only parallel overlaps in the distal portions of the filament. (d) Myosin molecule. The tail (S2 and LMM) is a coiled-

coil formed from the C-terminal halves of each heavy chain (red and green). Heads (S1) comprise the motor domain (MD) and light chain domain (LCD),

which contains the essential light chain (ELC, blue) and regulatory light chain (RLC, yellow).

Striated muscle thick filamentsIn striated muscle, thick filaments are bipolar structures,

formed by antiparallel tail interactions at the filament

center (creating a ‘bare zone’ free of myosin heads) and

parallel interactions at either end (the head or ‘cross-

bridge’ region) [7] (Figure 1b,c). In most muscles, thick

filaments are connected to each other at their bare zones

via cytoskeletal bridges constituting the M-line


(Figure 1a) [8�]. These connections help organize the

filament lattice and maintain filament register.

Head organization

In the relaxed state, the myosin heads are helically or

quasi-helically ordered (reflecting the underlying organi-

zation of the tails), with an axial spacing of �14.5 nm

between adjacent levels (Figure 1b) [9,10]. This helical


206 Macromolecular assemblages

organization, first revealed by X-ray diffraction, was for

many years not observed by EM. The disordering of the

heads that occurred during filament preparation for nega-

tive staining was first overcome using invertebrate thick

filaments, which are relatively stable (Figure 2a) [11].

Fourier analysis and 3D helical reconstructions of nega-

tively stained invertebrate [12–15], and subsequently

vertebrate [16], filaments revealed information about

the number of helical tracks, as well as the conformation

and arrangement of the heads at moderate resolution

(Figure 2b).

Invertebrate thick filaments have become model systems

for study owing to their marked stability, which may

result from their ability to ‘switch off’ ATPase and

actin-binding activity (in contrast to vertebrates, in which

myosin ATPase is unregulated and the thin filaments

regulate actin-binding activity [6]). The number of par-

allel helical tracks of heads varies from four to seven in

invertebrates and their helical repeat varies from 30 to

48 nm, depending on both muscle type and species

[10,13,14,17]. Various interpretations of arthropod recon-

structions [13,14], in some cases based on docking with

the atomic structure of S1 [18], all suggested that the

myosin heads were splayed axially, with interactions

occurring between heads from different 14.5 nm levels.

However, the resolution was limited to �5 nm

(Figure 2b) and the accuracy of the atomic fitting was

therefore uncertain.

Figure 2

Head organization in striated muscle filaments. (a) Negatively stained thick

heads are apparent by viewing along the filament axis at a glancing angle.

resolution) based on negative-stain images. Data set from [13] surface rend

of tarantula thick filament (2.5 nm resolution) based on cryo-EM images [20

and the helical repeat is 43.5 nm. The ‘tilted-J’ (blue dots) represents a pair

with (b), which is at the same scale, is due to shrinkage with negative staini

resulting in loss of information and lower resolution in negative stain). (d) Re

myosin (PDB code 1i84) [21] at two 14.5 nm levels [20��]. In both the atomic

head of each pair (green) has its actin-binding site blocked by binding to th

light chain of the other (red). Interactions also occur between heads from di

has been modeled as an a-helical coiled-coil (the a-helices have the same

rod of density in the reconstruction that runs from the head–head junction d

blocked heads from both its own and the next 14.5 nm level. These multiple

interacting with actin, thus switching the filament off [20��,21,22].


A quite different organization is suggested by cryo-EM of

purified [19], frozen-hydrated tarantula filaments ana-

lyzed by single-particle reconstruction methods [20��].Cryo-EM preserves native structure without staining

artifacts, whereas single-particle reconstruction avoids

the problem of information overlap that can occur with

helical reconstruction methods [13]. With a resolution of

2.5 nm, the reconstruction clearly reveals the two heads of

a single myosin as a ‘tilted-J’ motif (Figure 2c). This motif

enables unambiguous fitting of the atomic model of an

‘off-state’ (dephosphorylated) myosin molecule from ver-

tebrate smooth muscle [21–23] (Figure 2d). The fit shows

that the heads of a single myosin are not splayed but lie

close together, interacting with each other and both

pointing towards the bare zone. These intramolecular

interactions between the heads of a single myosin, as

well as intermolecular interactions between the heads of

axially adjacent molecules, appear to block key functional

sites in the motor domain (Figure 2d), suggesting a simple

structural mechanism for switching off the filament in the

relaxed state [20��,21]. The reconstruction also reveals

how S2 connects the heads to the filament backbone and

shows possible interactions between S2 and the heads

[20��,23] (Figure 2d). The fit of a vertebrate smooth

muscle myosin molecule into an invertebrate striated

muscle filament is extraordinary, and suggests that the

intramolecular interacting-head structure is highly con-

served and likely to represent the structural basis of the

off-state in most regulated myosin filaments. An

filament from tarantula muscle (protein white). Helical tracks of

(b) Helical 3D reconstruction of tarantula thick filament (5 nm

ered using UCSF Chimera [63]. (c) Single-particle 3D reconstruction��]. The axial distance between adjacent levels of heads is 14.5 nm

of myosin heads. The larger diameter of this reconstruction compared

ng (the heads also appear to partially collapse onto the backbone,

construction in (c) fitted with the atomic model of smooth muscle

model (based on crystallographic studies) and the filament, one

e converter region (located in the motor domain) [3��] and essential

fferent levels. S2, whose position was uncertain in the atomic model,

colors as the corresponding heads from which they arise), filling a

own towards the filament backbone. S2 appears to interact with

interactions could prevent the heads from hydrolyzing ATP and



exception may be the highly specialized indirect flight

muscle of insects, for which modeling of the relaxed X-ray

diffraction pattern (to 6.5 nm resolution) suggests that

one myosin head projects out from the filament back-

bone, while the other wraps around it [24]. It will be

important to test this model against high-resolution cryo-

EM reconstructions of filaments.

In contrast to invertebrates, all vertebrate thick filaments

have a similar structure, with the heads arranged in three

tracks and with a repeat of 43 nm. This was first estab-

lished in skeletal filaments [16] and has now also been

demonstrated in cardiac filaments [25�]. Suggestions that

vertebrate filaments are two stranded [26] are incorrect

[25�]. Unlike the helical arrays in invertebrates, the heads

in vertebrate muscle have only a quasi-helical arrange-

ment [9,16], their positions and possibly their conforma-

tions being perturbed, probably by the binding of myosin

to non-myosin proteins (see below). Using helical recon-

struction methods [27], it has been suggested that the

myosin heads are close to each other and project to high

radius, tilting away from the bare zone. The difference

between this proposed head orientation and that in tar-

antula muscle [20��] may result from the limited (�6 nm)

resolution of the reconstruction or from the absence of an

off-state in vertebrate filaments. Although helical recon-

struction provides only an average of the different head

positions and conformations, the perturbation is pre-

served in reconstructions using tilt views. In frog fila-

ments, it appears to involve differences in radial,

azimuthal and axial positions of the heads compared with

those of a strictly helical filament [16], but individual

heads are not resolved. Single-particle methods [28�]should distinguish different head conformations at dif-

ferent levels and improve the resolution, but may require

frozen-hydrated filaments for definitive conclusions to be

reached [20��].

Figure 3

Backbone structure of striated muscle thick filaments (transverse views, sho

crystal model. Each circle represents one myosin tail (based on [64]). (b) Su

three myosin tails. Based on [17]. (c) Projected density of a 43.5 nm repeat

Twelve subunits, 4 nm in diameter and of high density (red circles), represe

axis — as predicted by the subfilament model [17] for a filament with the ta

lower density material at higher radius represents heads.


The ordering of the heads that occurs in relaxing condi-

tions depends on the ‘closing’ of the switch 2 element

[3��] of the myosin active site, located in the motor

domain [29,30�]. This traps inorganic phosphate, inhibit-

ing its release from the head and minimizing ATP turn-

over — a key characteristic of relaxation. Relaxed

ordering of heads in regulated filaments produces

head–head interactions [20��] that contribute to the

switched-off state and may also underlie the efficient

switching on of contraction [30�].

Backbone structure

The backbone of the myosin filament consists of stag-

gered, closely packed myosin tails running approximately

parallel to the filament axis. The detailed packing of the

tails has not been determined because of their length

(�160 nm), narrow diameter (�2 nm) and tight packing.

Theoretical models have been proposed using equivalent

or quasi-equivalent interactions between myosin tails that

are similar between species, with only small changes

leading to the diverse filament symmetries found in

the animal kingdom [31]. In one class of model, the tails

assemble individually to form the backbone (molecular

crystal model; Figure 3a), whereas in another class they

form small groups (subfilaments) that make up the back-

bone (subfilament model [31]; Figure 3b). X-ray diffrac-

tion supports a 4 nm diameter subfilament model

(Figure 3b) as a common structural theme in crustacea,

with variations in subfilament number and tilt accounting

for filaments with different symmetries [17]. The recon-

struction of tarantula filaments provides direct support for

this model [20��] (Figures 2c and 3c), suggesting that it

might also be common throughout the arthropods. The

backbone organization in other invertebrates is not

known and remains speculative for vertebrate filaments.

The subfilament model could be modified slightly to

account for the quasi-helical symmetry in vertebrates

wn for a filament with fourfold rotational symmetry). (a) Molecular

bfilament model. Each subfilament is 4 nm in diameter and contains

of tarantula thick filament reconstruction (protein white) [20��].

nting the thick filament backbone, run parallel to the filament

rantula symmetry. Note that (c) is shown at a smaller scale than (a,b);



[17] and EM shows evidence of a substructure approxi-

mately parallel to the filament axis that is consistent with

subfilaments [32]. However, sequence considerations are

suggested to favor the molecular crystal model [33�]. Any

model of vertebrate filaments must account for their

fraying into three large subunits at low ionic strength [34].

Tail interactions

Insights into likely tail interactions in the filament back-

bone have come from sequence analysis of Caenorhabditiselegans myosin [35]. The tail shows a repeating structural

unit with a length of 28 residues, in which positively and

negatively charged regions are separated by 14 residues.

Strong charge attraction is predicted between tails stag-

gered by odd multiples of 14 residues. Attraction is

especially strong with staggers of 98 or 294 residues,

corresponding to displacements of 14.4 and 43.2 nm

between adjacent tails — two of the fundamental peri-

odicities seen in myosin filaments. Analysis of the

sequences of vertebrate skeletal, smooth and non-muscle

myosin tails reveals the presence of long-range oscilla-

tions of charge and the importance of skip residues

(single-residue insertions in the a-helical sequence that

cause local instability of the coiled-coil) in determining

the dominant stagger interactions [33�]. Longer odd-

number multiples of 14.4 nm staggers are considered to

stabilize, and even-number multiples to destabilize, thick

filament structure.

Non-myosin components

Striated muscle thick filaments contain non-myosin com-

ponents that play structural and regulatory roles. Most

invertebrate filaments contain a core of paramyosin

(a coiled-coil protein about 124 nm in length, similar to

the myosin tail) in molar ratios up to 50% of myosin [36].

This may serve to strengthen these filaments against the

high loads they must bear due to their greater length (and

their larger number of actin filament interactions) com-

pared with vertebrate muscle. In Drosophila flight muscle,

paramyosin must be phosphorylated for optimal power

output [37]. Most invertebrates also contain small ver-

sions of titin-like proteins (see below), which contribute

Figure 4

Non-myosin proteins in vertebrate striated muscle thick filaments. Schemat

filament (green backbone, myosin heads omitted) and myosin-associated p

overlap with molecules from the opposite half of the sarcomere), along the

filament and the Z-line), where they have an elastic domain, and ending at t

Myosin-binding proteins (MyBP, yellow), primarily MyBP-C, occupy 11 sites


to the mechanical properties of these muscles [38,39].

Substantial molar amounts of additional proteins may also

be present, particularly in insect muscles (myofilin [40],

flightin [40,41] and myosin rod protein [42]), where they

play a role in filament assembly and stability, and in

modulation of contraction. The 3D structural organiza-

tion is not known for any of these non-myosin

components.

Thick filaments from vertebrate skeletal and cardiac

muscle are associated both with the giant elastic protein

titin and with myosin-binding proteins. Titin (also called

connectin) is a giant polypeptide (�3 MDa) that extends

from the center of the thick filament through the I-band

(where the thick filaments are absent) to the Z-line

[39,43] (Figure 4). It functions as a template for the

assembly of the sarcomere during development and as

a molecular spring responsible for key aspects of the

mechanical behavior of muscle. In the region associated

with the thick filament, titin consists of immunoglobulin-

like (Ig) and fibronectin-like (Fn) domains, arranged in

super-repeat patterns with a periodicity of �43 nm.

These repeats seem to dictate the positioning of the

myosin-binding proteins (Figure 4) and the organization

of myosin. There appear to be six molecules of titin

associated with each half of a thick filament, probably

running parallel to the filament axis on the filament

surface [39], but their 3D organization remains uncertain

[32,44].

The myosin-binding proteins of vertebrate thick fila-

ments (primarily myosin-binding protein C [MyBP-C])

bind at 11 sites, 43 nm apart (every third level of heads),

in each half of the filament [43,45�]. They are an impor-

tant target of mutations in familial hypertrophic cardio-

myopathy [45�]. Like titin, these myosin-binding proteins

(probably three molecules at each axial level) are made

largely from Ig and Fn modules. MyBP-C presumably

contributes to thick filament images and reconstructions,

but distinguishing it from the dominant organization of

myosin molecules remains difficult [27] and its 3D orga-

nization therefore remains unknown. The C-terminal end

ic diagram of a sarcomere (thin filaments omitted) showing a thick

roteins. Titin molecules (blue) run from the M-line (where they

thick filament, then through the I-band (the zone between the thick

he Z-line, where they would overlap with titin from the next sarcomere.

43 nm apart in the proximal portions of each half thick filament.



of MyBP-C may wrap around the filament backbone,

forming a collar [45�], or run axially along the filament

surface [46], while its N-terminal end may project from

the filament, interacting with S2 [45�,47] and possibly

extending as far as the neighboring actin filament [46].

MyBP-C may play a role in sarcomere assembly [45�] and,

in cardiac muscle, it appears to modulate contractility by

interacting with S2 in a phosphorylation-dependent way

[45�,47].

Smooth muscle and non-muscle thickfilamentsMyosin filaments are also present in smooth muscle and

non-muscle cells, where they pull on actin to produce

filament sliding, as in striated muscle. In vertebrates,

smooth muscle and non-muscle filaments are more labile

than those of striated muscle. Under in vitro relaxing

conditions, dephosphorylated filaments disassemble into

monomers with a folded structure and low ATPase activ-

ity. Upon phosphorylation, the molecules unfold and

assemble into filaments that have high ATPase activity

and are capable of interacting with actin [48�]. Phosphor-

ylation-dependent regulation of ATPase activity and

filament assembly provides a simple device for the sto-

rage of inactive molecules in non-muscle cells, and for

their transient assembly into filaments when and where

they are needed for motility. In the case of smooth

muscle, the lability of dephosphorylated filaments caused

uncertainty as to their very existence in the relaxed state

[48�]. It now appears that, although filaments are present

in relaxed smooth muscle, their number or length may, in

some cases, increase upon activation by assembly from a

pool of the folded monomers [48�,49�].

Vertebrate smooth muscle filaments have a different

arrangement of myosin molecules from that of striated

filaments. In this non-helical, ‘side-polar’ structure, the

orientation of the myosin molecules reverses on opposite

sides of the filament, rather than at opposite ends [50,51]

(Figure 5). The tail interactions responsible for this

different mode of assembly may be explained by subtle

differences in the tail sequence compared with striated

Figure 5

Schematic diagram of a side-polar thick filament. In the side-polar mode of

interactions occur along the entire length of the filament (c.f. bipolar filamen

have opposite polarity. Functionally, the filament is still bipolar and able to p

double arrows indicate how bipolar myosin dimers could easily associate w

disassembly processes thought to occur in some smooth muscles in differe


muscle myosin [33�]. The side-polar structure would still

pull thin filaments of opposite polarity towards each

other, and could help to account for the extreme short-

ening of which smooth muscle is capable [50]. In vivo,

smooth muscle filaments appear to have four molecules at

each 14.4 nm level [52], but the lability and non-helical

structure of the filaments has so far prevented 3D recon-

struction. Two-dimensional crystals of vertebrate smooth

muscle myosin show that the heads of individual mole-

cules interact with each other in the dephosphorylated

state, switching each other off [21,22]. This is the same

interaction that has now been demonstrated in off-state

striated muscle filaments from tarantula [20��](Figure 2d). The presence of this interhead interaction

across phyla and muscle types implies that it will also be

present in native smooth muscle myosin filaments.

Smooth muscle filaments contain, in addition to myosin,

smitin (a titin analog, thought to play a role in smooth

muscle mechanics) and two other proteins (telokin and a

38 kDa protein), which may be involved in stabilizing

filaments in relaxed muscle [48�]. The 3D organization of

these components is unknown.

Invertebrate smooth muscles, such as the catch muscles

of molluscs, contain large bipolar thick filaments with a

large core of paramyosin. In these filaments, paramyosin

has a paracrystalline organization, with the myosin form-

ing a thin surface layer of loosely attached molecules,

possibly organized in a helical fashion [53]. The thick

paramyosin core may strengthen these large filaments

against the high tensions that they must bear. These

filaments also contain small amounts of additional pro-

teins, including catchin and twitchin, which may play a

role in the catch mechanism of invertebrate smooth

muscle [53].

Assembly of thick filamentsThe correct assembly of myosin molecules into thick

filaments in muscle and non-muscle cells has several

key requirements, which have been well demonstrated

by genetic and other analyses [54]. These requirements

assembly, found in smooth muscle thick filaments, antiparallel tail

ts), so that opposite sides, rather than opposite ends, of the filament

ull actin filaments in opposite directions from opposite ends. The

ith and dissociate from the filament ends, consistent with assembly-

nt physiological states.



Figure 6

Myosin tail domains involved in thick filament assembly. Schematic representation of the myosin tail sequence showing the locations of various

regions thought to be involved in the correct assembly of thick filaments (composite diagram based on different myosins; not all sites are involved

in all myosins). These include a negative charge cluster, N1 (red), and two positive charge clusters, P1 (blue) and P2 (purple) [58��]; an assembly

competence domain or ACD (yellow) [55,58��]; the non-helical tailpiece or NHT (green) [56,57,58��,60,61�]; and four skip residues (dark blue) [33�].

Smooth muscle and non-muscle myosins lack the indicated skip residue.

vary by system, but may include: the correct level of

myosin expression; the presence of critical domains of

myosin; expression and proper assembly of non-myosin

proteins (e.g. paramyosin, titin, MyBP-C); regulation of

assembly by phosphorylation of the heavy or light chains;

and the correct functioning of chaperone proteins [54].

Numerous studies point to the special significance of the

C-terminal region of the myosin tail in the assembly of

thick filaments. Near their C terminus, both paramyosin

and myosin tails have a short segment of coiled-coil with

different charge and hydrophobicity properties compared

with the rest of the helix, and the molecule usually

terminates in a non-helical tailpiece (NHT) (Figures

1d and 6). The specialized C-terminal region of the

coiled-coil is found in vertebrate and invertebrate sarco-

meric, smooth and non-muscle myosins. Its different

domains have been suggested to play an essential role

in the initiation and modulation of filament assembly in

these systems [55–57,58��]. Interaction between critical

charge patches of the C-terminal coiled-coil may account

for the nucleation of assembly by antiparallel interaction,

followed by elongation by parallel addition of molecules

[58��]. In Dictyostelium, a large-scale (196-residue) charge

repeat in the C-terminal half of the tail (in addition to the

28-residue repeat) is important for forming thick fila-

ments [59]. As in other non-muscle systems, the assembly

process is delicately balanced, with disassembly being

achieved, in this case, by phosphorylation of three sites in

the C-terminal region. In contrast to the C-terminal

coiled-coil, the NHT does not generally appear to be

required for filament assembly [56,57,58��,60]. It may

instead be important for modulating filament stability

and structure in vertebrate smooth muscle [60], or for

maintaining proper sarcomere structure in striated muscle

[57].

Acanthamoeba may potentially be used to obtain a detailed

understanding of thick filament assembly, including, in

particular, the significance of the C-terminal region [61�].


Acanthamoeba myosin assembles into octameric filaments

in vitro and in vivo. Truncation studies have defined key

regions of the tail essential for the sequential steps of

assembly. Crystallographic studies of such defined tail

fragments may provide the first insights at high resolution

into the 3D interactions involved in the assembly process

[62��]. Acanthamoeba may be a model for the early stages

of assembly of other species of filament.

ConclusionsMyosin filaments comprise a diverse array of macromo-

lecular assemblies. Although their basic structure has

been understood for more than 40 years, progress towards

an atomic-level description has been slow. Recent

innovative studies have finally revealed details of the

organization of the heads and tails in tarantula filaments

at near-atomic level [20��], suggesting a structural model

for myosin filament regulation. Sequence analysis and

crystallographic studies of other systems are providing

new insights into the structural basis of thick filament

assembly. Important unresolved issues remain. Cryo-EM

and single-particle reconstruction of additional thick fila-

ment species should reveal whether subfilaments and an

interacting head structure, both seen in tarantula, are

common themes of thick filament structure. Improve-

ments in resolution may directly resolve individual myo-

sin tails and non-myosin proteins, and also provide more

detail on head interactions. Interpretation of structure

should be aided by further sequencing of the relevant

proteins from these species. Understanding vertebrate

smooth muscle thick filament structure will require tomo-

graphic or single-particle reconstruction, as these fila-

ments lack helical symmetry. Further progress in

understanding assembly can be expected from higher

resolution EM and X-ray studies.

AcknowledgementsOur thick filament studies are supported by grant AR34711 from theNational Institutes of Health. Rendering and atomic fitting ofreconstructions was carried out using UCSF Chimera [63].



References and recommended readingPapers of particular interest, published within the annual period ofreview, have been highlighted as:

� of special interest

�� of outstanding interest

1. Craig R, Padron R: Molecular structure of the sarcomere. InMyology, edn 3. Edited by Engel AG, Franzini-Armstrong C.NewYork, NY: McGraw-Hill; 2004:129-166.

2. Huxley HE: The mechanism of muscular contraction. Science1969, 164:1356-1366.

3.��

Geeves MA, Holmes KC: The molecular mechanism of musclecontraction. Adv Protein Chem 2005, 71:161-193.

This detailed review of the atomic basis of the crossbridge mechanism ofmuscle contraction relates crystallographic structures of S1 to biochem-ical and kinetic data.

4.��

Krendel M, Mooseker MS: Myosins: tails (and heads) offunctional diversity. Physiology (Bethesda) 2005, 20:239-251.

The authors summarize current knowledge of myosin superfamily struc-ture and function, placing myosin II in context.

5. Sweeney HL, Houdusse A: Mammalian muscle myosin. InMyology, edn 3. Edited by Engel AG, Franzini-Armstrong C.NewYork, NY: McGraw-Hill; 2004:167-186.

6. Perry SV: Activation of the contractile mechanism by calcium.In Myology, edn 3. Edited by Engel AG, Franzini-Armstrong C.NewYork, NY: McGraw-Hill; 2004:281-306.

7. Huxley HE: Electron microscope studies on the structure ofnatural and synthetic protein filaments from striated muscle.J Mol Biol 1963, 7:281-308.

8.�

Agarkova I, Perriard JC: The M-band: an elastic web thatcrosslinks thick filaments in the center of the sarcomere.Trends Cell Biol 2005, 15:477-485.

A review of the structural components and function of the striated muscleM-line.

9. Huxley HE, Brown W: The low-angle X-ray diagram ofvertebrate striated muscle and its behaviour duringcontraction and rigor. J Mol Biol 1967, 30:383-434.

10. Wray JS, Vibert PJ, Cohen C: Diversity of cross-bridgeconfigurations in invertebrate muscles. Nature 1975,257:561-564.

11. Kensler RW, Levine RJ: An electron microscopic and opticaldiffraction analysis of the structure of Limulus telson musclethick filaments. J Cell Biol 1982, 92:443-451.

12. Vibert P, Craig R: Electron microscopy and image analysis ofmyosin filaments from scallop striated muscle. J Mol Biol 1983,165:303-320.

13. Crowther RA, Padron R, Craig R: Arrangement of the headsof myosin in relaxed thick filaments from tarantula muscle.J Mol Biol 1985, 184:429-439.

14. Stewart M, Kensler RW, Levine RJ: Three-dimensionalreconstruction of thick filaments from Limulus and scorpionmuscle. J Cell Biol 1985, 101:402-411.

15. Morris EP, Squire JM, Fuller GW: The 4-stranded helicalarrangement of myosin heads on insect (Lethocerus) flightmuscle thick filaments. J Struct Biol 1991, 107:237-249.

16. Stewart M, Kensler RW: Arrangement of myosin heads inrelaxed thick filaments from frog skeletal muscle. J Mol Biol1986, 192:831-851.

17. Wray JS: Structure of the backbone in myosin filaments ofmuscle. Nature 1979, 277:37-40.

18. Offer G, Knight PJ, Burgess SA, Alamo L, Padron R: A new modelfor the surface arrangement of myosin molecules in tarantulathick filaments. J Mol Biol 2000, 298:239-260.

19. Hidalgo C, Padron R, Horowitz R, Zhao FQ, Craig R: Purificationof native myosin filaments from muscle. Biophys J 2001,81:2817-2826.


20.��

Woodhead JL, Zhao FQ, Craig R, Egelman EH, Alamo L, Padron R:Atomic model of a myosin filament in the relaxed state.Nature 2005, 436:1195-1199.

The first reconstruction of a thick filament to unambiguously reveal theorganization of myosin heads and the probable arrangement of the tails insubfilaments. The head organization directly suggests how these fila-ments are switched off in the relaxed state.

21. Wendt T, Taylor D, Trybus KM, Taylor K: Three-dimensionalimage reconstruction of dephosphorylated smooth muscleheavy meromyosin reveals asymmetry in the interactionbetween myosin heads and placement of subfragment 2.Proc Natl Acad Sci USA 2001, 98:4361-4366.

22. Liu J, Wendt T, Taylor D, Taylor K: Refined model of the 10Sconformation of smooth muscle myosin by cryo-electronmicroscopy 3D image reconstruction. J Mol Biol 2003,329:963-972.

23. Tama F, Feig M, Liu J, Brooks CL III, Taylor KA: The requirementfor mechanical coupling between head and S2 domains insmooth muscle myosin ATPase regulation and its implicationsfor dimeric motor function. J Mol Biol 2005, 345:837-854.

24. Al Khayat HA, Hudson L, Reedy MK, Irving TC, Squire JM: Myosinhead configuration in relaxed insect flight muscle: X-raymodeled resting cross-bridges in a pre-powerstroke state arepoised for actin binding. Biophys J 2003, 85:1063-1079.

25.�

Kensler RW: The mammalian cardiac muscle thick filament:crossbridge arrangement. J Struct Biol 2005, 149:303-312.

An elegant study of rabbit cardiac thick filaments using negative stainingand Fourier analysis.

26. Skubiszak L, Kowalczyk L: The vertebrate skeletal muscle thickfilaments are not three-stranded. Reinterpretation of someexperimental data. Acta Biochim Pol 2002, 49:841-853.

27. Eakins F, Al Khayat HA, Kensler RW, Morris EP, Squire JM: 3Dstructure of fish muscle myosin filaments. J Struct Biol 2002,137:154-163.

28.�

Al Khayat HA, Morris EP, Squire JM: Single particle analysis: anew approach to solving the 3D structure of myosin filaments.J Muscle Res Cell Motil 2004, 25:635-644.

This useful review outlines the use of single-particle methods to recon-struct filamentous particles, including those lacking strict helical sym-metry, such as vertebrate thick filaments.

29. Xu S, Offer G, Gu J, White HD, Yu LC: Temperature and liganddependence of conformation and helical order in myosinfilaments. Biochemistry 2003, 42:390-401.

30.�

Zoghbi ME, Woodhead JL, Craig R, Padron R: Helical order intarantula thick filaments requires the ‘‘closed’’ conformationof the myosin head. J Mol Biol 2004, 342:1223-1236.

A negative-stain EM study of tarantula thick filaments in the presence ofnucleotide analogs simulating different states of ATP. Quantification ofhead ordering by Fourier transformation of images shows that only analogsthat produce the switch 2 closed state cause the heads to be ordered.

31. Chew MW, Squire JM: Packing of alpha-helical coiled-coilmyosin rods in vertebrate muscle thick filaments. J Struct Biol1995, 115:233-249.

32. Kensler RW: The mammalian cardiac muscle thick filament:backbone contributions to meridional reflections. J Struct Biol2005, 149:313-324.

33.�

Straussman R, Squire JM, Ben Ya’acov A, Ravid S: Skip residuesand charge interactions in myosin II coiled-coils: implicationsfor molecular packing. J Mol Biol 2005, 353:613-628.

This detailed analysis of the sequence and charge patterns of the tails ofhuman striated, smooth and non-muscle myosin molecules suggestslikely tail interactions in thick filaments.

34. Maw MC, Rowe AJ: Fraying of A-filaments into threesubfilaments. Nature 1980, 286:412-414.

35. McLachlan AD, Karn J: Periodic features in the amino acidsequence of nematode myosin rod. J Mol Biol 1983,164:605-626.

36. Levine RJ, Elfvin M, Dewey MM, Walcott B: Paramyosin ininvertebrate muscles. II. Content in relation to structure andfunction. J Cell Biol 1976, 71:273-279.



37. Liu H, Miller MS, Swank DM, Kronert WA, Maughan DW,Bernstein SI: Paramyosin phosphorylation site disruptionaffects indirect flight muscle stiffness and power generation inDrosophila melanogaster. Proc Natl Acad Sci USA 2005,102:10522-10527.

38. Bullard B, Linke WA, Leonard K: Varieties of elastic protein ininvertebrate muscles. J Muscle Res Cell Motil 2002, 23:435-447.

39. Tskhovrebova L, Trinick J: Titin: properties and familyrelationships. Nat Rev Mol Cell Biol 2003, 4:679-689.

40. Qiu F, Brendel S, Cunha PM, Astola N, Song B, Furlong EE,Leonard KR, Bullard B: Myofilin, a protein in the thick filamentsof insect muscle. J Cell Sci 2005, 118:1527-1536.

41. Ayer G, Vigoreaux JO: Flightin is a myosin rod binding protein.Cell Biochem Biophys 2003, 38:41-54.

42. Polyak E, Standiford DM, Yakopson V, Emerson CP Jr,Franzini-Armstrong C: Contribution of myosin rod protein to thestructural organization of adult and embryonic muscles inDrosophila. J Mol Biol 2003, 331:1077-1091.

43. Clark KA, McElhinny AS, Beckerle MC, Gregorio CC: Striatedmuscle cytoarchitecture: an intricate web of form andfunction. Annu Rev Cell Dev Biol 2002, 18:637-706.

44. Cantino ME, Chew MW, Luther PK, Morris E, Squire JM: Structureand nucleotide-dependent changes of thick filaments inrelaxed and rigor plaice fin muscle. J Struct Biol 2002,137:164-175.

45.�

Flashman E, Redwood C, Moolman-Smook J, Watkins H: Cardiacmyosin binding protein C: its role in physiology and disease.Circ Res 2004, 94:1279-1289.

A review of the structure and function of MyBP-C, comparing differentmodels of its arrangement on the thick filament, and discussing its role inphysiology and disease.

46. Squire JM, Luther PK, Knupp C: Structural evidence for theinteraction of C-protein (MyBP-C) with actin and sequenceidentification of a possible actin-binding domain. J Mol Biol2003, 331:713-724.

47. Harris SP, Rostkova E, Gautel M, Moss RL: Binding of myosinbinding protein-C to myosin subfragment S2 affectscontractility independent of a tether mechanism. Circ Res2004, 95:930-936.

48.�

Seow CY: Myosin filament assembly in an ever-changingmyofilament lattice of smooth muscle. Am J Physiol Cell Physiol2005, 289:C1363-C1368.

A concise review of vertebrate smooth muscle thick filament composi-tion, structure, assembly and stability.

49.�

Smolensky AV, Ragozzino J, Gilbert SH, Seow CY, Ford LE:Length-dependent filament formation assessed frombirefringence increases during activation of porcine trachealmuscle. J Physiol 2005, 563:517-527.

Birefringence studies of tracheal smooth muscle are consistent with anincrease in thick filament number or length upon activation. This in vivostudy supports and discusses earlier EM and birefringence evidence infavor of changes in filament density in different physiological andmechanical states of smooth muscle.

50. Craig R, Megerman J: Assembly of smooth muscle myosin intoside-polar filaments. J Cell Biol 1977, 75:990-996.

51. Xu JQ, Harder BA, Uman P, Craig R: Myosin filament structure invertebrate smooth muscle. J Cell Biol 1996, 134:53-66.


52. Tonino P, Simon M, Craig R: Mass determination of nativesmooth muscle myosin filaments by scanning transmissionelectron microscopy. J Mol Biol 2002, 318:999-1007.

53. Chantler PD: Scallop adductor muscles: structure andfunction. In Scallops: Biology, Ecology and Aquaculture. Editedby Parsons J, Shumway SE. Elsevier; 2005:231-320.

54. Landsverk ML, Epstein HF: Genetic analysis of myosin IIassembly and organization in model organisms. Cell Mol LifeSci 2005, 62:2270-2282.

55. Cohen C, Parry DA: A conserved C-terminal assembly region inparamyosin and myosin rods. J Struct Biol 1998, 122:180-187.

56. Ikebe M, Komatsu S, Woodhead JL, Mabuchi K, Ikebe R, Saito J,Craig R, Higashihara M: The tip of the coiled-coil rod determinesthe filament formation of smooth muscle and nonmusclemyosin. J Biol Chem 2001, 276:30293-30300.

57. Hoppe PE, Andrews RC, Parikh PD: Differential requirementfor the nonhelical tailpiece and the C terminus of the myosinrod in Caenorhabditis elegans muscle. Mol Biol Cell 2003,14:1677-1690.

58.��

Nakasawa T, Takahashi M, Matsuzawa F, Aikawa S,Togashi Y, Saitoh T, Yamagishi A, Yazawa M: Critical regionsfor assembly of vertebrate nonmuscle myosin II. Biochemistry2005, 44:174-183.

By studying the solubility of truncated rod fragments of non-musclemyosin, two domains in the C-terminal region of the coiled-coil areidentified as critical for thick filament assembly. Interactions betweenpositive and negative charge clusters in these regions are proposed toinitiate filament assembly.

59. Hostetter D, Rice S, Dean S, Altman D, McMahon PM, Sutton S,Tripathy A, Spudich JA: Dictyostelium myosin bipolar thickfilament formation: importance of charge and specificdomains of the myosin rod. PLoS Biol 2004, 2:e356.

60. Rovner AS, Fagnant PM, Lowey S, Trybus KM: The carboxyl-terminal isoforms of smooth muscle myosin heavy chaindetermine thick filament assembly properties. J Cell Biol 2002,156:113-123.

61.�

Turbedsky K, Pollard TD: Assembly of Acanthamoeba myosin-IIminifilaments. Definition of C-terminal residues required toform coiled-coils, dimers, and octamers. J Mol Biol 2005,345:351-361.

Constructs of Acanthamoeba myosin II tail are used to define the taildomains required for assembly into the octameric minifilaments that mayoccur in vivo. This model system may provide clues to the initial stages ofthe assembly of thick filaments in other systems.

62.��

Turbedsky K, Pollard TD, Yeager M: Assembly of Acanthamoebamyosin-II minifilaments. Model of anti-parallel dimers basedon EM and X-ray diffraction of 2D and 3D crystals. J Mol Biol2005, 345:363-373.

An elegant 3D EM and X-ray crystallographic study of the interactions inthe antiparallel dimer of coiled-coil tails that guide the first step of myosinassembly in Acanthamoeba thick filaments.

63. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM,Meng EC, Ferrin TE: UCSF Chimera–a visualization system forexploratory research and analysis. J Comput Chem 2004,25:1605-1612.

64. Squire JM: The Structural Basis of Muscular Contraction.New York, NY: Plenum Press; 1981.


Structure and function of myosin filaments

Documents

Transcript of Structure and function of myosin filaments