Aurora A in cell division: kinase activity not required

(1)

Article

Reference

Aurora A in cell division: kinase activity not required

KRESS, Elsa, GOTTA, Monica

Abstract

Aurora A kinase is a key regulator of cell division, whose functions were attributed to its ability to phosphorylate diverse substrates. Aurora A is now shown to have a kinase-independent role in the regulation of chromatin-mediated microtubule assembly.

KRESS, Elsa, GOTTA, Monica. Aurora A in cell division: kinase activity not required. Nature Cell Biology , 2011, vol. 13, no. 6, p. 638-9

DOI : 10.1038/ncb2276 PMID : 21572422

Available at:

http://archive-ouverte.unige.ch/unige:25588

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

N E W S A N D V I E W S

Aurora A in cell division: kinase activity not required

Elsa Kress and Monica Gotta

Aurora A kinase is a key regulator of cell division, whose functions were attributed to its ability to phosphorylate diverse substrates.

Aurora A is now shown to have a kinase-independent role in the regulation of chromatin-mediated microtubule assembly.

Successful cell division depends on the precise spatio-temporal regulation of several events, including the formation of the mitotic spindle, which is necessary for chromosome seg- regation. Mitotic spindle formation requires separation of the duplicated centrosomes, centrosome maturation and, after nuclear envelope breakdown (NEBD), assembly of a bipolar spindle with microtubules emanat- ing from chromatin and centrosomes¹. The coordination of these events is often achieved by post-translational modifications such as phosphorylation, and thus protein kinases and phosphatases have important roles in cell cycle regulation. The serine/threonine kinase Aurora A is highly conserved from yeast to mammals and plays a key role in many aspects of cell division, including centrosome maturation and spindle assembly. The kinase activity of Aurora A depends on an activating phosphorylation of a threonine in its activation loop and the specificity of its functions is regulated by interaction with several cofactors, the best characterized of which is TPX2 (ref. 2).

In animal cells, microtubules are nucleated from two major subcellular locations during spindle assembly: chromatin and centrosomes.

Aurora A plays a central role in microtubule assembly at both sites, a function that was shown to depend on its kinase activity¹. On page 708 of this issue, Toya et al. reveal that Aurora A also has a kinase-independent role in the assembly of chromatin-induced microtubules in Caenorhabditis elegans³.

The authors used the single-cell C. elegans embryo as a model system. The oocyte lacks centrosomes, and thus the first division depends on the centrosome provided by the sperm. To form a proper bipolar spindle, this centrosome needs to duplicate and mature within a few minutes following fertilization⁴. AIR-1, the C. elegans Aurora A orthologue, has been previously shown to be important

for centrosome maturation. AIR-1 localizes in a doughnut shape at centrosomes and along the base of astral microtubules (Fig. 1a) and recruits the evolutionarily conserved microtubule nucleator γ-tubulin, as well as other proteins important for microtubule dynam- ics⁵. In addition, AIR-1 is essential for centrosome separation during spindle assembly after NEBD and it also regulates the timing of NEBD and mitotic entry^5–7 (Fig. 1a, b).

A previous study by Sugimoto and col- leagues showed that AIR-1 is required for the assembly of γ-tubulin-independent microtubules⁸. Whereas depletion of either γ-tubulin or AIR-1 resulted in only a reduc- tion of microtubule number and the assembly of monopolar spindles, depletion of both resulted in a failure to nucleate microtubules.

However, how AIR-1 contributes to microtubule assembly in a γ-tubulin-independent manner remained unclear.

In the current study the authors investigated the role of AIR-1 in microtubule assembly in greater detail. Using a combination of genet- ics and imaging they found that microtubule length and number are nearly constant during the cell cycle in AIR-1-depleted embryos, suggesting that γ-tubulin-dependent microtubules are present throughout the cell cycle. In contrast, microtubules increased in number and length during chromosome condensation in embryos depleted of γ-tubulin, indicating that the contribution of AIR-1 to microtubule assembly is temporally regulated. Chromo- somes can induce spindle assembly in meiosis or mitosis in many organisms and this process requires Aurora A¹, suggesting that the AIR-1-dependent/γ-tubulin-independent mechanism of spindle assembly could be chromatin-induced. Indeed, AIR-1 was also observed around condensed chromosomes in wild-type embryos in vivo (Fig. 1a, b), a localization not previously reported. Moreo- ver, when the one-cell embryo was genetically depleted of functional centrosomes¹, the chromatin-stimulated microtubules assembled in the absence of γ-tubulin, but not in the absence of AIR-1.

Surprisingly, in contrast to other systems¹, the kinase activity of AIR-1 was not required for chromatin-dependent microtubule assembly. Immunofluorescence microscopy experiments using phosphospecific antibodies indi- cated that kinase-active AIR-1 was restricted to the centre of centrosomes (Fig. 1a, b), and did not localize at chromatin in wild-type embryos or in embryos where microtubule assembly occurs only at chromatin. Although a kinase- inactive form of AIR-1 could not support centrosome maturation and separation after NEBD and was not sufficient for the viability of embryos depleted of endogenous AIR-1, it was able to promote microtubule assembly and growth from chromatin and was detected along these microtubules (Fig. 1b). Localization of the kinase-inactive form along microtubules was negatively regulated by endogenous AIR-1, suggesting that the balance between kinase- active and kinase-inactive AIR-1 is critical to regulate localization and function of AIR-1 in wild-type embryos.

How is the function of AIR-1 in chromatin- stimulated microtubule assembly regulated? In vertebrates, this process depends on the Aurora A activator TPX2 (ref. 1). The putative homo- logue of TPX2 in C. elegans is TPXL-1 (ref. 9).

Toya et al. showed that TPXL-1 is not required for AIR-1 localization on chromatin or for chromatin-stimulated microtubule assembly, consistent with previous data⁹ and with the fact that the AIR-1 kinase activity is not necessary for this process. However, TPXL-1 was needed for the localization of kinase-inactive AIR-1 along chromatin-stimulated microtubules and for their stabilization (Fig. 1b). Therefore, although TPXL-1 is not crucial for microtubule assembly at chromatin, it does have an important role in stabilizing microtubules.

These findings reveal a paradox: how can a putative activator of AIR-1 localize the inactive kinase? What prevents this pool of AIR-1 from being activated by TPXL-1? Although genetic and biochemical experiments suggested that, similarly to TPX2, TPXL-1 is an activator of AIR-1 in C. elegans⁹, Toya et al. show that the amount of active AIR-1 is not reduced on

Elsa Kress and Monica Gotta are at the CMU, Medical Faculty, University of Geneva, Geneva, Switzerland.

e-mail: monica.gotta@unige.ch DOI: 10.1038/ncb2276

638 NATURE CELL BIOLOGY VOLUME 13 | NUMBER 6 | JUNE 2011

(3)

N E W S A N D V I E W S

depletion of TPXL-1, suggesting that TPXL-1 might not be a key activator of AIR-1 in vivo.

Based on these data, other factors must be involved in activation of centrosomal AIR-1 (Fig. 1a). Several conserved potential activa- tors of Aurora A have been isolated in other systems and one of them emerges as a particularly good candidate. Recent data identify the conserved SPD2 protein as a centrosome- specific activator of Aurora A in mammalian cells¹⁰. In C. elegans, SPD-2 localizes at centrosomes and, like AIR-1, is required for centrosome maturation¹¹.

The present study convincingly shows that inactive AIR-1 has a role in chromatin-mediated microtubule assembly and stabilization in C. elegans (Fig. 1b) and also proposes the interesting possibility that the proper balance between the kinase-active and kinase-inactive states of AIR-1 is essential to allow timely assembly of a functional mitotic spindle. Future work will be needed to determine how these different

populations are spatio-temporally regulated and which cofactors are important to maintain their balance and coordinate their function.

Kinase-independent roles have been shown for kinases in other systems^12,13, and it will be important to investigate whether a kinase- independent Aurora A role also exists in other organisms. Interestingly, overexpression of both active and catalytically inactive Aurora A results in cell division defects in mammalian cells¹⁴. One possible interpretation of these data is that Aurora A also has a kinase-independent role in mammals. In view of the current findings it is worth revisiting the previous literature and directly investigating whether Aurora A could function independently of its kinase activity in vertebrates as well. This might be particularly important in light of the fact that Aurora A inhibitors are leading compounds for cancer therapy¹⁵. The potential conservation of the kinase-independent role of Aurora A would require a re-evaluation of the

mechanisms by which these inhibitors affect the growth of cancer cells.

COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.

1. Walczak, C. E. & Heald, R. Int. Rev. Cytol. 265, 111–

158 (2008).

2. Joukov, V. Cell Cycle 10, 895–903 (2011).

3. Toya, M. et al. Nat. Cell Biol. 13, 708–714 (2011).

4. Gonczy, P. & Rose, L. S. WormBook 1–20 (2005), ed.

The C. elegans Research Community, doi/10.1895/

wormbook.1.30.1, http://www.wormbook.org.

5. Hannak, E., Kirkham, M., Hyman, A. A. & Oegema, K.

J. Cell Biol. 155, 1109–1116 (2001).

6. Hachet, V., Canard, C. & Gonczy, P. Dev. Cell 12, 531–

541 (2007).

7. Portier, N. et al. Dev. Cell 12, 515–529 (2007).

8. Motegi, F., Velarde, N. V., Piano, F. & Sugimoto, A. Dev.

Cell 10, 509–520 (2006).

9. Ozlu, N. et al. Dev. Cell 9, 237–248 (2005).

10. Joukov, V. et al. Proc. Natl Acad. Sci. USA 107, 21022–

21027 (2010).

11. Muller-Reichert, T., Greenan, G., O’Toole, E. & Srayko, M. Cell Mol. Life Sci. 67, 2195–2213 (2010).

12. Zhao, X. et al. J. Cell Biol. 189, 955–965 (2010).

13. Abeliovich, H. et al. Mol. Biol. Cell 14, 477–490 (2003).

14. Meraldi, P., Honda, R. & Nigg, E. A. EMBO J. 21, 483–

492 (2002).

15. Dar, A. A. et al. Mol. Cancer Ther. 9, 268–278 (2010).

a TPXL-1 b

X (SPD-2?) AIR-1

P-AIR-1

AIR-1 P-AIR-1

AIR-1 P-AIR-1?

Localization

Centrosome separation

NEBD

Activation Microtubule

stabilization

Figure 1 Localization and function of kinase-active and kinase-inactive AIR-1. (a) Kinase-inactive AIR-1 (dark green) is localized in a doughnut shape at centrosomes and along the base of astral microtubules (red lines). This latter localization depends on TPXL-1. Activation of AIR-1 by phosphorylation of its activation loop is triggered by an unknown factor, possibly SPD-2. Kinase-active AIR-1 (light green) is localized to the centre of centrosomes, where it recruits γ-tubulin. NEBD also requires AIR-1 (see text) but whether active or inactive AIR-1 contributes to this process is not known. The brown circle represents the nucleus and the chromosomes are depicted in blue. (b) At metaphase, kinase-active AIR-1 maintains centrosome separation; the kinase-inactive form may trigger polymerization and stabilize chromatin-induced microtubules.

NATURE CELL BIOLOGY VOLUME 13 | NUMBER 6 | JUNE 2011 639