Signal transduction

Actin, cofilin and cognition

Patients with Williams syndrome, a multi-gene deletion syndrome, suffer from mild mental retardation and vascular disease. They also have defects in visuo-spatial cognition — a failure to integrate parts into a whole — that are linked to deletion of the gene that codes for LIM-kinase 1 (LIMK-1)1. Consistent with this cognitive defect, large amounts of LIMK-1 are expressed in neurons2, yet its molecular targets have remained elusive until now. On pages 805 and pages 809 of this issue, however, Arber et al.3 and Yang et al.4 report that LIMK-1 phosphorylates cofilin, an essential protein that is required for turnover of actin filaments.

During cell movements such as neuron outgrowth or leukocyte chemotaxis, actin filaments must be organized into a dense, dynamic meshwork. This forms at the leading edge of a cell, where actin polymerization drives forward movement, and it usually takes the form of thin sheets (lamellipodia) or spikes (filopodia). Cell movement is a dynamic process, and actin at the leading edge of the cell must be continuously depolymerized and then repolymerized to produce this movement5. Actin depolymerization limits the length of lamellipodia and enables the actin subunits to be recycled for further rounds of polymerization.

There is mounting evidence that the key enzyme required for actin depolymerization is cofilin. In vivo, cofilin has been shown to be essential for cytokinesis6, endocytosis7 and other cell processes that require rapid turnover of actin filaments8. In vitro, cofilin binds to both actin monomers and polymers, and promotes the disassembly of actin filaments. Cofilin is regulated by phosphorylation of the serine residue at position 3, which inhibits its actin-binding and depolymerization activities. Stimuli that induce the production of lamellipodia relieve this inhibition by causing the rapid dephosphorylation of cofilin9.

Arber et al.3 and Yang et al.4 now provide evidence that LIMK-1 phosphorylates (and therefore inactivates) cofilin. Both groups labelled mammalian cells with radioactive inorganic phosphate, and found that isolated LIMK-1 associates with only one phosphoprotein — cofilin. Moreover, LIMK-1, but not an inactive form of the enzyme, can phosphorylate recombinant cofilin. These findings account for the observations that overexpression of LIMK-1 leads to accumulation of excess actin filaments, whereas overexpression of a dominant-negative LIMK-1 (a mutant form that disrupts the wild-type activity) inhibits the accumulation of actin filaments.

For cells to move, signals from their peripheries must be relayed to proteins such as LIMK-1 and cofilin. What factors relay these signals? Arber et al. and Yang et al. show that the small GTPase Rac may be important in regulating the activity of LIMK-1. Rac regulates the actin reorganization that is required to form lamellipodia10, yet few of its protein targets have been identified. The authors3,4 found that Rac-dependent formation of lamellipodia is blocked by dominant-negative forms of LIMK-1, suggesting that LIMK-1 acts downstream in the Rac pathway. Moreover, dominant-negative Rac leads to decreased phosphorylation of cofilin, whereas activated Rac modestly increases phosphorylation. These results indicate that Rac activates LIMK-1 which, in turn, phosphorylates — and inactivates — cofilin (Fig. 1, overleaf).

Figure 1: Model for Rac-induced changes in actin dynamics to form a lamellipodium, based on the findings of Arber et al.3 and Yang et al.4.

When Rac is stimulated by growth factors, it activates LIM-kinase 1 (LIMK-1) which, in turn, phosphorylates cofilin. This may inhibit localized disassembly of actin filaments and help to release cofilin from the actin monomers. Rac-mediated stimulation of actin polymerization may occur through uncapping or increased nucleation of actin filaments. Both of these mechanisms may occur through production of phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P2) by phosphatidylinositol phosphate 5-kinase. The intermediates that activate phosphatidylinositol phosphate 5-kinase and LIMK-1, and the factor that stimulates dephosphorylation of cofilin, are unknown.

To induce the formation of lamellipodia, Rac must do more than simply inactivate cofilin — it must also induce the polymerization of actin. One possible mechanism involves Rac-induced increases in the levels of phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P2), which is thought to cause filament uncapping11. Removal of a capping protein from the end of an actin filament could then allow the polymerization of actin to resume. Actin also needs to be depolymerized for the formation of lamellipodia, so we might expect that the Rac pathway only transiently inactivates cofilin. Indeed, Yang et al.4 find no net change in cofilin phosphorylation when endogenous (not overexpressed) Rac is activated. Transient inactivation of cofilin at the leading edge could allow nearby actin filaments to grow. Additionally, cofilin phosphorylation may induce the release of recently depolymerized actin monomers. Continuous cycles of cofilin phosphorylation and dephosphorylation would allow both cofilin and actin to be recycled for further rounds of depolymerization and polymerization, respectively. Clearly, further work is needed to sort out the role of cofilin phosphorylation in actin dynamics, and to clarify the temporal and spatial regulation of actin depolymerization in cells.

Another exciting aspect of this work comes from studies on patients with Williams syndrome, who have only one copy of the LIMK-1 gene. Given that LIMK-1 regulates the turnover of actin filaments, why should people with Williams syndrome have defects in visuo-spatial cognition, as opposed to other processes that require actin dynamics? Perhaps neurons require high levels of LIMK-1 to finely regulate the turnover of actin filaments during axonal guidance. The PC12 neuronal cells containing dominant-negative LIMK-1 studied by Arber et al. may provide a clue as to what neurons with decreased levels of LIMK-1 look like. Although neuron outgrowth still occurred, the neurites contained dramatically fewer filopodia — finger-like projections that are thought to sense in which direction axons should grow. If this is true, a decrease in LIMK-1 could account for the abnormally clustered and aligned neurons seen in the brains of patients with Williams syndrome12.

To show that a decrease in LIMK-1 leads to abnormal neuronal wiring, we need to study people who lack a copy of the LIMK-1 gene (and not contiguous genes), as well as mouse knockout models. Future work will also need to address whether the visuo-spatial cognitive defect in people with Williams syndrome results exclusively from phosphorylation of cofilin by LIMK-1, or whether other substrates for LIMK-1 exist.


  1. 1

    Frangiskakis, J. M. et al. Cell 86, 59–69 (1996).

    Google Scholar 

  2. 2

    Okano, I. et al. J. Biol. Chem. 270, 31321–31330 (1995).

    Google Scholar 

  3. 3

    Arber, al. Nature 393, 805–809 (1998).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Yang, al. Nature 393, 809–812 (1998).

    ADS  CAS  Article  Google Scholar 

  5. 5

    Theriot, J. A. & Mitchison, T. J. Nature 352, 126–131 (1991).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Abe, H., Obinata, T., Minamide, L. S. & Bamburg, J. R. J. Cell Biol. 132, 871–885 (1996).

    Google Scholar 

  7. 7

    Lappalainen, P. & Drubin, D. G. Nature 388, 78–82 (1997).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Theriot, J. A. J. Cell Biol. 136, 1165–1168 (1997).

    Google Scholar 

  9. 9

    Moon, A. & Drubin, D. G. Mol. Biol. Cell 6, 1423–1431 (1995).

    Google Scholar 

  10. 10

    Ridley, A. J., Paterson, H. F., Johnston, C. L., Diekmann, D. & Hall, A. Cell 70, 401–410 (1992).

    Google Scholar 

  11. 11

    Hartwig, J. al. Cell 82, 643–653 (1995).

    Google Scholar 

  12. 12

    Galaburda, A. M., Wang, P. P., Bellugi, U. & Rossen, M. Neuroreport 5, 753–757 (1994).

    Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Rosenblatt, J., Mitchison, T. Actin, cofilin and cognition. Nature 393, 739–740 (1998).

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.