Polo-like kinase 1 independently controls microtubule-nucleating capacity and size of the centrosome

Abstract

Centrosomes are composed of a centriolar core surrounded by a pericentriolar material (PCM) matrix that docks microtubule-nucleating γ-tubulin complexes. During mitotic entry, the PCM matrix increases in size and nucleating capacity in a process called centrosome maturation. Polo-like kinase 1 (PLK1) is recruited to centrosomes and phosphorylates PCM matrix proteins to drive their self-assembly, which leads to PCM expansion. Here, we show that in addition to controlling PCM expansion, PLK1 independently controls the generation of binding sites for γ-tubulin complexes on the PCM matrix. Selectively preventing the generation of PLK1-dependent γ-tubulin docking sites led to spindle defects and impaired chromosome segregation without affecting PCM expansion, highlighting the importance of phospho-regulated centrosomal γ-tubulin docking sites in spindle assembly. Inhibiting both γ-tubulin docking and PCM expansion by mutating substrate target sites recapitulated the effects of loss of centrosomal PLK1 on the ability of centrosomes to catalyze spindle assembly.

Auto-inhibition of Cnn binding to γ-TuRCs prevents ectopic microtubule nucleation and cell division defects

“…This idea is consistent with results from numerous mass spectrometry experiments showing that γ-TuRCs do not readily associate with CM1-domain proteins within the cytosol (Oegema et al, 1999;Choi et al, 2010;Hutchins et al, 2010;Teixidó-Travesa et al, 2012;Thawani et al, 2018;Liu et al, 2019;Wieczorek et al, 2019;Consolati et al, 2020). Binding of the human and C. elegans CM1 domain proteins, CDK5RAP2 and SPD-5, to γ-TuRCs involves phosphorylation (Hanafusa et al, 2015;Ohta et al, 2021), which can be a means to spatiotemporally control binding.…”

“…Our data is consistent with this possibility and, in our view, this is the most likely explanation. A similar mechanism has also been proposed in C. elegans (Ohta et al, 2021). Nevertheless, there are alternative possibilities, including that the CAI domain could recruit another protein that interferes with CM1 domain binding.…”

“…Phosphorylation appears to be an important mechanism for promoting binding between CM1 domain proteins and γ-TuRCs. This is true for human CDK5RAP2, C. elegans SPD-5, and S. cerevisiae SPC110, where the phosphorylation sites that promote binding have been identified either upstream or downstream of the CM1 domain (Hanafusa et al, 2015;Lin et al, 2014b;Ohta et al, 2021). We show that phospho-mimicking sites both upstream and downstream of the CM1 domain also promotes binding of Drosophila Cnn-C to γ-TuRCs.…”

“…We therefore considered other putative phosphorylation sites. Phosphorylation slightly downstream of the CM1 domain promotes binding to γ-TuRCs in humans and C. elegans CM1 domain proteins (Ohta et al, 2021;Hanafusa et al, 2015). While the sequence surrounding the CM1 domain is not conserved across diverse species (Figure S3B), we identified two serine residues (S173 and S186) downstream of Cnn's CM1 domain that were conserved in Drosophila species (Figure S3A).…”

“…How could CAI domain mediated auto-inhibition be relieved to allow efficient binding to γ-TuRCs at MTOCs? Studies in human cells, C. elegans, and S. cerevisiae have shown that binding of CM1 domain proteins to γ-TuSCs or γ-TuRCs is promoted by phosphorylation close to the CM1 domain (Hanafusa et al, 2015;Ohta et al, 2021;Lin et al, 2014b) (Figure S3B). Moreover, Cnn-C binds γ-TuRCs and is phosphorylated only at centrosomes (Zhang and Megraw, 2007;Conduit et al, 2014b;a), suggesting a possible link between binding and phosphorylation.…”