Cell Cycle

KIFC3 directs a centrosome cohesive force

Assembly of the mitotic spindle requires timely separation of the centrosomes. Their movement apart is driven by the plus-end-directed kinesin Eg5. A new study demonstrates that the kinesin KIFC3 provides an opposing microtubule-based cohesive force that modulates centrosome separation and ensures accurate chromosome segregation.

The mitotic spindle is established by microtubules that extend from its two poles1. In animal cells, the poles are defined by centrosomes, which consist of two barrel-shaped centrioles surrounded by microtubule-anchoring pericentriolar material (PCM)2. To ensure accurate chromosome segregation and successful mitosis, centrosome duplication is tightly regulated. The choreography of centrosome duplication involves the formation of new ‘daughter’ centrioles on the two extant ‘mother’ centrioles during S phase. A proteinaceous linker that contains the large coil-coiled proteins C-NAP1 and rootletin keeps the two resultant centrosomes in close proximity so that microtubules emanate from a single microtubule-organizing center3,4. At the onset of mitosis, this linker is removed by Never in mitosis A (NIMA)-related kinase 2A (NEK2A)5. The conserved motor protein Eg5 slides opposing microtubules apart and pushes the centrosomes in opposite directions, establishing a bipolar spindle structure6. The temporal regulation of centrosome separation is crucial for this process to work properly. In this edition of Nature Cell Biology, Hata, Schiebel and colleagues show that the kinesin KIFC3 opposes the separative force of Eg5 and directs the timely assembly of the mitotic spindle7.

The current study was presaged by findings demonstrating that intact microtubules keep interphase centrosomes together in human and rodent cell lines8, although the way in which this works was unclear. A more recent paper from the Schiebel group9 showed that ablation of centrosome cohesion by zinc finger nuclease (ZFN)-mediated disruption of CEP250 (which encodes C-NAP1) in hTERT-RPE1 cells led to increased centrosome separation, consistent with other work in which C-NAP1 was removed by short interfering RNA (siRNA) or genome editing3,10. Loss of microtubules from C-NAP1-deficient cells led to a dramatic increase in centrosome separation, indicating a microtubule-directed mechanism that ensured centrosome cohesion when the centriolar linker was removed9. Hypothesizing that this cohesion mechanism involves a microtubule minus-end-directed motor activity, Hata et al. used siRNA to deplete minus-end-directed kinesins and cytoplasmic dynein heavy chains from C-NAP1-deficient hTERT-RPE1 cells and monitored centrosome separation7. They also overexpressed each of the minus-end-directed kinesins. Knockdown of KIFC2 or KIFC3 exacerbated the centrosome separation phenotype in C-NAP1 nulls, and inducible KIFC3 overexpression suppressed this separation in a manner that required KIFC3’s motor activity and microtubules. Live-cell imaging revealed that the induced KIFC3 localized to centrosomes at the same time that the two centrosomes came back together. These findings indicate that KIFC3 generates a force between centrosomes that is sufficient for their close association even in the absence of the centrosome linker. Genome editing to disrupt KIFC3 bolstered this finding by demonstrating premature centrosome separation at the onset of mitosis, when the linker is removed7.

A previous study using siRNA in HeLa cells suggests that KIF25, another minus-end-directed kinesin, mediated centrosome cohesion, potentially in opposition to Eg5 (ref. 11). Hata et al. note that functional KIF25 is not expressed in either HeLa or hTERT-RPE 1 cells7. Treatment of C-NAP1-deficient hTERT-RPE1 cells with the KIF25 siRNA sequence (#2) used by Decarreau et al.11 caused an elevated level of centrosome separation, though this was not rescued by re-expression of siRNA-resistant KIF25. Furthermore, CRISPR-mediated editing of KIF25 did not cause centrosome separation during early mitosis. Together with the observation that KIF25 siRNA treatment caused reduced centrosomal KIFC3 localization, these findings suggest that the effects of KIF25 knockdown on centrosome proximity may be indirect and possibly occur through interfering with KIFC3 function7.

To examine how KIFC3 contributes to centrosome cohesion, Hata et al.7 first examined the localization of KIFC3 within the centrosome and found it to be a component of the subdistal appendages (SDAs), structures that extend from the microtubule barrel wall of the mother centriole and facilitate microtubule binding2. Knockdown of SDA components ODF2, CEP128 or Ninein reduced centrosomal KIFC3 localization and increased centrosome separation in C-NAP1 knockouts7, as previously observed10. They then found that daughter centriole-specific impairment of PCM assembly by depletion of PCM components increased centrosome separation, which was not rescued by KIFC3 overexpression. These data show that microtubules associated with the SDAs and with the opposing daughter centriole PCM are needed for centrosome cohesion. Analogously to how the plus-end motor activity of Eg5 drives centrosomes apart6, Hata et al. suggest that these antiparallel microtubules can provide potential tracks for KIFC3 to direct movement of the centrosomes toward the minus ends and toward one another. Electron microscopy (EM) tomography analyses of microtubules in centriolar proximity provide support for this mechanism. The assembly of purified KIFC3 into a structure that was shown by negative-stain EM to have four globular domains linked by extended linker regions lends further credence to a bipolar tetrameric model for KIFC3, similar to Eg5 (ref. 7). Whether specific modifications of the relevant microtubules are needed to direct KIFC3’s association with them remains to be explored.

A significant insight of this work concerns the antagonistic balance between KIFC3 and Eg5 (Fig. 1). Hata et al. addressed this using inhibition of Eg5 and overexpression of KIFC3, generating key evidence for the opposing roles of the two activities, which determine centrosome spatial stability at the beginning of mitosis, upon NEK2-mediated removal of the linker7. After centrosome linker removal, mitotic-spindle assembly requires that the centrosomes move apart. Hata et al.7 show that centrosomal KIFC3 declines toward prometaphase, as Eg5 increases. Knockout of the NEK2 gene removed both the NEK2A and less well-characterized NEK2B isoforms and caused the retention of KIFC3 and reduced prophase centrosome separation. Conversely, removal of KIFC3 increased centrosome separation in NEK2 knockout cells. NEK2 thus drives both centrosome linker removal and the dissociation of KIFC3 from the centrosome. Even if NEK2 is absent, an Aurora-A-dependent activity will drive centrosome separation during prometaphase12. However, inhibition of Aurora A led to defective centrosome separation and a high level of chromosome mis-segregation in NEK2-deficient cells, a phenotype that was suppressed by KIFC3 knockdown7. These findings provide a temporal link between the loss of both the proteinaceous centrosome linker and the motor function of KIFC3 that will allow Eg5 to push centrosomes apart as mitosis proceeds. Interestingly, a similar process exists in yeast, in which antagonistic kinesin-14 and kinesin-5 activities direct the appropriate separation of spindle poles13. The study by Hata et al.7 thus provides evidence for an evolutionarily conserved mechanism that controls the spatial dynamics of the centrosomes during the early stages of mitotic spindle formation.

Fig. 1: Antagonistic activities of KIFC3 and Eg5 determine centrosome separation during the assembly of the mitotic spindle.

Cartoon illustrating the dynamics of centrosome separation in early mitosis. Centrosome cohesion is maintained in interphase through the C-NAP1 linker. NEK2 mediates the removal of this linker, and centrosome movement is determined by the balance between the cohesive activity of KIFC3 and Eg5-driven separation. NEK2-dependent loss of KIFC3 leads to the centrosomes moving further apart, allowing the establishment of the mitotic spindle.

Dysregulation of the centrosome separation mechanism affects the mitotic spindle and cell division, as well as centrosome–Golgi association and cell migration7,9,11. Mazo et al. also disrupted the SDAs by knockdown of specific components required for their function in C-NAP1-ablated hTERT-RPE1 cells and found that centrosomes were further separated than those in the C-NAP1 mutants10, consistent with the findings by Hata et al.7. In addition, they observed that ablation of both C-NAP1 and the SDAs led to a relocalization of the primary cilium, a sensory structure formed by the extension of the mother centriole, to the cell surface, changing cellular responses to fluid flow and cellular signalling10. Although few data exist on the pathological effects of centrosome cohesion defects, a truncating mutation in CEP250 that causes premature centrosome separation with otherwise normal centrioles gives rise to Caprine-like Generalized Hypoplasia Syndrome (SHGC) in Montbéliarde cattle14. The symptoms and cell division anomalies in SHGC are similar to those seen in conditions in which mutations affect the PCM, such as Seckel syndrome14, although it is unknown whether centrosome cohesion is relevant in these disorders. Furthermore, disruption of the linker component rootletin potentiates chromosome mis-segregation in colorectal cancer15. In summary, the current study describes a previously unappreciated mechanism that determines how centrosomes regulate mitotic spindle assembly, with implications for cell signalling processes, cell movement and migration during development. The extent to which centrosome cohesion is established by the canonical centrosomal linker varies between HeLa, hTERT-RPE1 and U2OS cells9, indicating differences in the mechanics of centrosome cohesion between cell lines and highlighting the need for further understanding of the dynamic regulation of centrosome cohesion.


  1. 1.

    Prosser, S. L. & Pelletier, L. Nat. Rev. Mol. Cell Biol. 18, 187–201 (2017).

    CAS  Article  Google Scholar 

  2. 2.

    Conduit, P. T., Wainman, A. & Raff, J. W. Nat. Rev. Mol. Cell Biol. 16, 611–624 (2015).

    CAS  Article  Google Scholar 

  3. 3.

    Bahe, S., Stierhof, Y. D., Wilkinson, C. J., Leiss, F. & Nigg, E. A. J. Cell Biol. 171, 27–33 (2005).

    CAS  Article  Google Scholar 

  4. 4.

    Mayor, T., Stierhof, Y. D., Tanaka, K., Fry, A. M. & Nigg, E. A. J. Cell Biol. 151, 837–846 (2000).

    CAS  Article  Google Scholar 

  5. 5.

    Faragher, A. J. & Fry, A. M. Mol. Biol. Cell 14, 2876–2889 (2003).

    CAS  Article  Google Scholar 

  6. 6.

    Kapitein, L. C. et al. Nature 435, 114–118 (2005).

    CAS  Article  Google Scholar 

  7. 7.

    Hata, S. et al. Nat. Cell Biol. https://doi.org/10.1038/s41556-019-0382-6 (2019).

    Article  Google Scholar 

  8. 8.

    Jean, C., Tollon, Y., Raynaud-Messina, B. & Wright, M. Eur. J. Cell Biol. 78, 549–560 (1999).

    CAS  Article  Google Scholar 

  9. 9.

    Panic, M., Hata, S., Neuner, A. & Schiebel, E. PLoS Genet. 11, e1005243 (2015).

    Article  Google Scholar 

  10. 10.

    Mazo, G., Soplop, N., Wang, W. J., Uryu, K. & Tsou, M. F. Dev. Cell 39, 424–437 (2016).

    CAS  Article  Google Scholar 

  11. 11.

    Decarreau, J. et al. Nat. Cell Biol. 19, 384–390 (2017).

    CAS  Article  Google Scholar 

  12. 12.

    Mardin, B. R. et al. Nat. Cell Biol. 12, 1166–1176 (2010).

    CAS  Article  Google Scholar 

  13. 13.

    Saunders, W., Lengyel, V. & Hoyt, M. A. Mol. Biol. Cell 8, 1025–1033 (1997).

    CAS  Article  Google Scholar 

  14. 14.

    Floriot, S. et al. Nat. Commun. 6, 6894 (2015).

    CAS  Article  Google Scholar 

  15. 15.

    Remo, A. et al. Mol. Cancer Res. 16, 1385–1395 (2018).

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Ciaran G. Morrison.

Ethics declarations

Competing interests

The author declares no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Morrison, C.G. KIFC3 directs a centrosome cohesive force. Nat Cell Biol 21, 1057–1059 (2019). https://doi.org/10.1038/s41556-019-0385-3

Download citation


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

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