Cell biology

Spinning actin to divide

When our cells divide, they are cut down the middle by a tightening belt of proteins. New work reveals that the protein filaments in this belt are made from scratch every time.

The ability of cells to multiply lies at the heart of many biological processes. In multicellular organisms such as ourselves, cell proliferation is essential for growth and development, and to replace cells spent by daily wear and tear. For single-celled species such as yeasts, proliferation is crucial because it is how these organisms reproduce. In order to proliferate, a cell has to duplicate its contents and then divide physically into two, distributing the duplicated contents evenly between the two new cells. This process of cell division — called cytokinesis — is carried out rather as if a thread encircling a boiled egg was gradually pulled tighter to constrict the egg and cut it across the middle. On page 82 of this issue, Pelham and Chang1 provide a clear picture of how the cell spins such a thread.

The thread encircling a dividing cell is a belt of proteins called the contractile ring (Fig. 1) and — unlike the thread around the boiled egg — is constructed within the cell, just beneath the cell membrane. The ring is composed of actin filaments and myosin proteins, well-known components of muscle. It is thought to constrict when oppositely orientated actin filaments slide over each other with the help of myosin, just as in contracting muscle.

Figure 1: Spinning out the contractile ring.

This ring, shown at the bottom, constricts to divide a proliferating cell into two. It is composed of antiparallel actin filaments, cross-linked by myosin molecules. Pelham and Chang1 have found that the filaments are formed from scratch at the site of division, and continuously assemble and disassemble. The process of filament formation is shown from the top. First, actin monomers assemble into trimers ('nucleation'); trimers act as a seed to which further monomers are added ('polymerization'). Pelham and Chang find that the Arp2/3 complex and the Cdc12 and Cdc3 proteins are needed for nucleation or polymerization. The filaments are then organized into the contractile ring.

Actin filaments are long helical polymers of globular actin monomers. The conventional view of how the contractile ring forms is that ready-made actin filaments in the cell periphery are recruited to the site of cytokinesis during or after segregation of the genetic material (mitosis), and are assembled into the ring by myosin or by actin-bundling proteins. But last year it was shown in dividing frog eggs that actin monomers are rapidly incorporated into actin filaments in the contractile ring as it is being constructed, implying that formation of new actin filaments is required2. So what does happen? Pelham and Chang's detailed analysis1 finds that the actin filaments are made from scratch and continuously assemble and disassemble. The authors also identify proteins that control this process.

All actin filaments form in two steps (Fig. 1). In the first — nucleation — three or four actin monomers assemble into a cluster, which acts as a seed. In the second step, called polymerization or elongation, monomers are successively added to the seeds, allowing rapid growth of the filament. Nucleation is the rate-limiting step, but is speeded up by a multiprotein complex named the Arp2/3 complex3.

Pelham and Chang started by analysing the involvement of de novo actin-filament formation in the construction of the contractile ring in the fission yeast Schizosaccharomyces pombe, which divides in a similar way to mammalian cells. They used an in vitro test of ring assembly, which involves permeabilizing the dividing yeast cells, and incubating them with fluorescently labelled actin monomers. If actin nucleation and polymerization occur, the fluorescent monomers will be incorporated into the ring. Indeed, the authors found that a fluorescent ring was produced in these cells, and that ring formation was sensitive to an inhibitor of actin polymerization. In contrast, a drug that caps the rapidly growing ends of existing actin filaments did not affect the rate of actin incorporation into the ring. So the results imply that actin filaments are formed from scratch and in situ in the contractile ring.

Pelham and Chang then looked at the contribution of the Arp2/3 complex to actin dynamics in the permeabilized-cell system. They found, first, that mutant cells lacking Arp3 did not show actin incorporation. Second, antibodies that bind to Arp3 interfere with incorporation of actin into the contractile ring. So the Arp2/3 complex is necessary for formation of contractile-ring actin filaments, presumably at the nucleation step.

The authors also reveal that another means of controlling filament formation is involved in ring construction in vivo in S. pombe. This regulatory mechanism consists of two proteins, Cdc12 and Cdc3. Cdc12 is a member of the formin family of proteins, which are found in organisms from yeasts to mammals and include Bni1 in budding yeast, Diaphanous in fruitflies and mDia in mammals4. Cdc3 is an actin-monomer-binding protein, also known as profilin. Formins related to Diaphanous bind profilin, and both are essential in cytokinesis4,5.

Interestingly, the Arp2/3 complex and the formins have been shown to induce the formation of distinct actin structures: patches and cables, respectively, in budding yeast6,7; and lamellipodia (sheet-like cellular extensions) and stress fibres, respectively, in mammalian cells3,8. More intriguingly, the formin Bni1 can nucleate actin polymerization independently of the Arp2/3 complex in vitro, and the presence of profilin markedly accelerates subsequent polymerization9,10. It seems, then, that two mechanisms for actin nucleation and polymerization — one involving the Arp2/3 complex, and the other requiring formins and profilins — are recruited to, and work at, the site of cell division.

Finally, Pelham and Chang studied the stability of the contractile ring, and found that it is highly dynamic, with its components exchanging every minute. To show this, they used the technique of 'fluorescence recovery after photobleaching'. They tagged the protein tropomyosin or a subunit of myosin — both of which are actin-binding components of the contractile ring — with green fluorescent protein. When they bleached the fluorescence with laser light, they found that it recovered with a half-time of less than 30 seconds, indicating that the ring is broken down and reconstructed with new fluorescent components within this time period. Similar dynamics of myosin in the contractile ring have been reported previously11, but the rate of exchange determined by Pelham and Chang is much faster. This requirement for a continuous supply of filament-binding proteins implies that not only actin polymerization, but also actin depolymerization, occurs in the ring, and explains the previously puzzling requirement for the actin-depolymerizing factor ADF/cofilin12.

So Pelham and Chang1 have shown that the contractile ring undergoes continuous remodelling to carry out its functions. Given the previous observations of dividing frog eggs2, it seems that this mechanism is not limited to fission yeast, but applies more generally. What are the implications of these results? One of the authors' main findings is the need for the formin Cdc12 in ring construction. Although not explicitly shown for Cdc12, Diaphanous-related formins are targets of the Rho protein4,5, which works as a molecular switch in various cell processes and has been suggested to regulate cytokinesis13. So this new study once again highlights the role of Rho in cell division. Studies of the functions and dynamics of Rho in cytokinesis might resolve several issues, such as how the plane of division is determined, and how the timing of cleavage is regulated.


  1. 1

    Pelham, R. J. & Chang, F. Nature 419, 82–86 (2002).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Noguchi, T. & Mabuchi, I. J. Cell Sci. 114, 401–412 (2001).

    CAS  PubMed  Google Scholar 

  3. 3

    Higgs, H. N. & Pollard, T. D. Annu. Rev. Biochem. 70, 649–676 (2001).

    CAS  Article  Google Scholar 

  4. 4

    Wasserman, S. Trends Cell Biol. 8, 111–115 (1998).

    CAS  Article  Google Scholar 

  5. 5

    Watanabe, N. et al. EMBO J. 16, 3044–3056 (1997).

    CAS  Article  Google Scholar 

  6. 6

    Evangelista, M. et al. Nature Cell Biol. 4, 32–41 (2002).

    CAS  Article  Google Scholar 

  7. 7

    Sagot, I., Klee, S. K. & Pellman, D. Nature Cell Biol. 4, 42–50 (2002).

    CAS  Article  Google Scholar 

  8. 8

    Watanabe, N. et al. Nature Cell Biol. 1, 136–143 (1999).

    CAS  Article  Google Scholar 

  9. 9

    Pruyne, D. et al. Science 297, 612–615 (2002).

    ADS  CAS  Article  Google Scholar 

  10. 10

    Sagot, I. et al. Nature Cell Biol. 4, 626–631 (2002); advance online publication, 22 July 2002 (doi:10.1038/ncb834).

    CAS  Article  Google Scholar 

  11. 11

    Wong, K. C. Y. et al. Curr. Biol. 12, 724–729 (2002).

    CAS  Article  Google Scholar 

  12. 12

    Abe, H. et al. J. Cell Biol. 132, 871–885 (1996).

    CAS  Article  Google Scholar 

  13. 13

    Mabuchi, I. et al. Zygote 1, 325–331 (1993).

    CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Narumiya, S., Mabuchi, I. Spinning actin to divide. Nature 419, 27–28 (2002). https://doi.org/10.1038/419027a

Download citation

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.


Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing