News & Views | Published:

Structural biology

Actin in a twist

Nature volume 457, pages 389390 (22 January 2009) | Download Citation

How monomers of the cytoskeletal protein actin join to form the stable polymers crucial to muscle contraction and cellular motility has been a long-standing question. A state-of-the-art approach provides an answer.

The actin protein is abundant in all eukaryotic cells (those characterized by a membrane-bound nucleus), and is particularly prevalent in muscle, where it comprises about 20% of the total mass. Actin comes in two forms: monomeric, globular G-actin; and polymeric, filamentous F-actin. F-actin forms through polymerization of G-actin, a process that has a central role in cell motility. In non-muscle cells, for example, polymerization of F-actin induces cell movement by pushing structures such as membranes forwards1. Also, when one stands or walks, all of the tension causing muscle contraction is produced in muscle fibres by F-actin filaments interacting with the motor protein myosin2,3. Although the atomic structure of G-actin has been known for almost 20 years4, structural details of the F-actin monomer — which is similar, but not identical, to G-actin — have remained elusive. Reporting on page 441 of this issue, Oda et al.5 derive a high-resolution structure of F-actin from analysis of X-ray fibre-diffraction patterns to elucidate the transition from G- to F-actin.

The structure of G-actin has been determined independently more than 30 times. These data show that G-actin is a rather flat molecule built from two similar major domains (outer and inner) related by a pseudo-dyad, with a molecule of the nucleotide ATP, or its hydrolysed version ADP, bound between them (Fig. 1a). The terms outer and inner refer to the position of these domains in the F-actin structure. Electron-microscopy data6 have shown F-actin to be made of two chains that turn gradually round each other to form a right-handed 'long-pitch' helix. The inner domain is closer to the axis of this helix.

Figure 1: Monomer versus polymer.
Figure 1

a, G-actin monomers consist of two similar domains: outer (grey) and inner (blue). The names relate to the position of each domain in the F-actin helix — the inner domain is closer to the helix axis. The bound ADP molecule (magenta) is sandwiched between the inner and outer domains. (Data from ref. 9.) b, Oda et al.5 provide an atomic model of the F-actin helix. This helix repeats in six left-handed turns (measuring 35.7 nanometres). Each repeat contains 13 molecules so that the first molecule of each repeat (arrowed) is in an identical orientation. Because the rotation per molecule (167°) is close to 180°, the F-actin structure appears as two long-pitch helices slowly winding around each other. Each of the 14 G-actin molecules of the helix is shown in a different colour, apart from the grey–blue molecule that is shown in two colours for comparison with a (asterisk).

By contrast, solving the structure of F-actin has been challenging. Despite a plethora of modifications to the G-actin structure — including the use of various ATP analogues, drug binding, capping proteins or crosslinking to make the G-actin look like F-actin — all of the structures derived have been essentially the same as that of the G-actin monomer, and none has yielded the secret of the G- to F-actin transition.

The most detailed data on the structure of F-actin came from X-ray diffraction of arrays of this filamentous protein oriented in a liquid- crystalline gel7. Diffraction from an oriented gel gives a fibre diagram — a section through the diffraction pattern of a single fibrous molecule that has been averaged by spinning it around the fibre axis (cylindrical averaging). If the molecule has periodic repeats, as in F-actin, then the fibre diagram consists of a series of lines called layer lines. The best way to interpret such a diagram is to compute it from a starting model and then use a refinement process to modify the model and arrive at a better fit with experimental data. Unfortunately, cylindrical averaging causes loss of information. Furthermore, thermal movement of the molecules in the liquid-crystalline sample causes them to become disorientated, which leads to a smearing out of the layer lines. So success depends on preparing the sample with the best possible orientation — that is, with all molecular axes as parallel to the fibre axis as possible — and investigating it at the highest possible resolution.

Oda et al.5 used an intense magnetic field to improve the orientation of their sample, and a highly collimated, intense X-ray source to collect data. They started with a model made by placing the crystal structure of the G-actin molecule in the best orientation in the F-actin helix, rather like the original structure of F-actin that my colleagues and I proposed7. They then calculated the low-energy vibrational modes of the G-actin monomer and selected the combination of modes that best fitted the fibre diagram.

Using this improved structure, they repeated the procedure. When no further improvements could be made, the authors turned to simulated annealing. In this molecular dynamics procedure, a molecule is heated for a few picoseconds to agitate the atoms so as to sample a range of possible structures. The molecule is then slowly cooled, with the fit to the fibre diagram acting as a pseudo-force to steer the process to the correct structure8. So Oda et al. finally achieved a very good fit to the fibre diffraction pattern.

Despite the complexity of this procedure, the authors' F-actin structure5 (Fig. 1b) is convincing in its simplicity. The transition from G- to F-actin seems to involve a 20° rotation of the outer domain with respect to the inner domain about a rotation axis roughly at right angles to the helix axis (Fig. 2). In G-actin the two domains are related by a propeller-like twist. The 20° rotation reduces this twist and flattens the molecule. Apart from this rotation and a reorientation of a flexible loop at the top of the outer domain, no other substantial change seems to occur.

Figure 2: The secret of G- to F-actin transition.
Figure 2

a, The inner (blue) and outer (grey) domains of a single actin molecule as seen in Figure 1a. Traces of the local rotation axis and the helix axis are also shown. b, c, G-actin (b) and F-actin (c) as seen at approximately right angles compared with a, looking along the local rotation axis. Note that, in passing from G-actin to F-actin, the outer domain rotates by 20° with respect to the inner domain. Consequently, the F-actin structure is substantially flatter.

Actin polymerization, essential for cell motility, is driven by ATP hydrolysis. Whereas G-actin cannot hydrolyse ATP, F-actin can. Oda et al. find that one effect of the 20° rotation is to bring an evolutionarily conserved glutamine residue at position 137 — which is implicated in the ATP hydrolysis mechanism — closer to the β- and γ-phosphate groups of the ATP molecule. So the rotation may be the switch for turning on the ATP-hydrolysing activity of F-actin. In addition, the flattening of monomers within F-actin substantially alters the site on this protein to which the muscle protein myosin binds — an interaction essential for muscle contraction; this could explain why myosin binds with high affinity to F-actin but not at all to G-actin. Thus, the new structure will certainly become an essential ingredient in our understanding of cell motility and muscle contraction.

Finally, the flat F-actin is very much like a bacterial analogue of actin called MreB. So Oda and colleagues' structure also lends support to the idea that actin is a bridge between eukaryotes and prokaryotic organisms such as bacteria.


  1. 1.

    et al. Physiol. Rev. 83, 433–473 (2003).

  2. 2.

    Science 164, 1356–1366 (1969).

  3. 3.

    & Adv. Protein Chem. 71, 161–193 (2005).

  4. 4.

    et al. Nature 347, 37–44 (1990).

  5. 5.

    , , , & Nature 457, 441–445 (2009).| |

  6. 6.

    & J. Mol. Biol. 6, 46–60 (1963).

  7. 7.

    et al. Nature 347, 44–49 (1990).

  8. 8.

    & Acta Cryst. A49, 504–513 (1993).

  9. 9.

    , & Science 293, 708–711 (2001).

Download references

Author information


  1. Kenneth C. Holmes is at the Max Planck Institute for Medical Research, D69120 Heidelberg, Germany.

    • Kenneth C. Holmes


  1. Search for Kenneth C. Holmes in:

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing