Revisiting the Role of the Mother Centriole in Centriole Biogenesis

See allHide authors and affiliations

Science  18 May 2007:
Vol. 316, Issue 5827, pp. 1046-1050
DOI: 10.1126/science.1142950


Centrioles duplicate once in each cell division cycle through so-called templated or canonical duplication. SAK, also called PLK4 (SAK/PLK4), a kinase implicated in tumor development, is an upstream regulator of canonical biogenesis necessary for centriole formation. We found that overexpression of SAK/PLK4 could induce amplification of centrioles in Drosophila embryos and their de novo formation in unfertilized eggs. Both processes required the activity of DSAS-6 and DSAS-4, two molecules required for canonical duplication. Thus, centriole biogenesis is a template-free self-assembly process triggered and regulated by molecules that ordinarily associate with the existing centriole. The mother centriole is not a bona fide template but a platform for a set of regulatory molecules that catalyzes and regulates daughter centriole assembly.

Centrioles are essential for the formation of cilia and flagella and for the organization of the centrosome (1). Normally, centrioles duplicate in coordination with the cell cycle. A new centriole, the daughter, arises orthogonally to each old one, the mother (1), in S phase. This led to the idea that the mother centriole templates the formation of the daughter (2, 3). However, daughter centrioles do not incorporate a substantial proportion of the mother (4), and centrioles can also form de novo when existing centrioles are naturally lost during development or are physically removed (57), questioning the idea of the mother centriole as a template.

SAK, also called PLK4 (SAK/PLK4), a kinase implicated in tumor development (8), is an upstream regulator of canonical centriole duplication and is necessary for centriole formation (9, 10). The Caenorhabditis elegans ZYG-1 kinase, a homolog of SAK, is part of a conserved module of proteins, which also includes SAS-6 and SAS-4, necessary for the normal centriole duplication cycle (1116). ZYG-1 is an upstream regulator in that process (17, 18), a role consistent with formation of multiple centrioles in cultured cells following overexpression of active SAK kinase (9, 10). The generation of multiple centrioles associated with high SAK expression also occurs physiologically in the olfactory mucosa (19). The Drosophila egg contains all the proteins necessary to make 213 centriole pairs (centrosomes) (20). Centrioles are naturally eliminated from the oocyte cytoplasm in the course of development and provided to the egg in the form of the basal body of the sperm (2023). Thus, we studied the consequences of overexpressing SAK in a cytoplasm that either contained centrioles (the embryo) or lacked them (the unfertilized egg).

Embryos overexpressing SAK did not develop (24) (fig. S1A and Fig. 1A) and were filled with free asters of microtubules not associated with spindles (Fig. 1, B and D, and fig. S1B). Those asters were focused around Drosophila pericentrin-like protein (D-PLP)–containing structures, a centriolar and pericentriolar material (PCM) marker (25) (fig. S1B). These centrosomes first appeared in 15- to 30-min-old embryos (Fig. 1B) and spread to fill the entire embryo after 2 to 3 hours (Fig. 1, C and D). The observed supernumerary centrosomes led to abnormal mitotic progression and impaired embryonic development, as observed previously upon microtubule depolymerization by colchicine treatment (26). To address the origins of those centrosomes, we examined the very early stages of embryonic development in embryos overexpressing SAK. Both the sperm aster around the incoming basal body and the first mitotic spindle were normal (Fig. 1E). However, at anaphase or telophase of the first mitosis, we observed more than two centrosomes at each pole (Fig. 1E), an indication of the onset of centrosome amplification. No other centrosomes were seen in the embryo at this stage. Moreover, we estimated that a minimum of 3700 centrosomes (equivalent to 12 duplication cycles) were present after 60 min in embryos overexpressing SAK. After 60 min, a wild-type embryo only showed 128 centrosomes. We observed duplicating centrioles in groups, suggesting they originated by duplication of a progenitor (Fig. 1, D and E, and fig. S1B). Thus, upon fertilization of eggs overexpressing SAK, the basal body of the sperm enters an environment that promotes accelerated canonical duplication, overriding any existing controls that would normally couple the centrosome and chromosome cycles.

Fig. 1.

Overexpression of SAK in Drosophila embryos leads to massive centrosome amplification. (A) Overexpression of upstream activation sequence (UAS)–SAK in embryos using the maternal driver V32-gal4. LC indicates loading control; WT, wild type. (B) Embryos overexpressing SAK become progressively full of free centrosomes nucleating asters. (C and D) The spreading of the centrosomes follows the wild-type spindle axial expansion pattern. Categories were as follows [according to the area occupied by the centrosomes within the embryo (24)]: 0 to 2%, 2 to 20%, 20 to 60%, and more than 60% area occupancy. α-tubulin is shown in green. An average of 60 embryos were counted in each category. Asterisks indicate polar bodies. Arrows indicate spindles. Scale bar indicates 50 μm. (E) Centrosome amplification in embryos is observed at the end of first mitosis. γ-tubulin is shown in red; α-tubulin, green; and DNA, blue. Scale bar, 10 μm. (Insets) γ-tubulin at 2×.

Uncoupling between centrosome and chromosome cycles occurs when embryos are arrested in S-phase–like conditions (27, 28). However, this did not seem to be so in this case, because proliferating cell nuclear antigen (PCNA) (29), which appears early in S phase, was not detected in DNA of SAK-overexpressing embryos (fig. S2).

We next asked whether SAK could promote centriolar assembly in the absence of centrioles. Centrioles were lost normally in oocytes overexpressing SAK (fig. S3). Yet observations of unfertilized eggs at varying developmental intervals revealed free centrosomes in eggs overexpressing SAK that had exited meiosis II (Fig. 2 and fig. S4) but never in wild-type eggs. Thus, in the absence of a basal body provided by the sperm, SAK can induce de novo formation of centrosomes. Whereas in embryos centrosomes appeared in a single cluster in the first mitotic spindle and spread throughout the cytoplasm (Fig. 1, D and E), in unfertilized eggs they appeared scattered at random positions, including at the anterior and posterior poles (Fig. 2D, arrows). The formation of the first centrioles started later in eggs than in embryos [at 30 min, 0 amplification in eggs versus 51% amplification in embryos; after 1 hour, the amounts were 18% versus 89%, respectively (Figs. 1B and 2B and fig. S5)], suggesting that centrioles take longer to be made in the absence of a template. However, once the first centrosomes had formed in eggs, their spreading in space and time was very similar to that seen in embryos (compare Fig. 1D and Fig. 2D), indicative of canonical biogenesis. Thus, once the first centrioles are formed de novo, they probably duplicate through the canonical pathway.

Fig. 2.

SAK induces de novo centrosome formation in Drosophila eggs. (A) Overexpression of UAS-SAK in eggs using the maternal driver V32-gal4. LC, loading control. (B) De novo centrosome formation starts after 30 min. (C and D) Denovo centrosomes appear randomly in space [arrows in (D)]. Categories were as follows: 0 to 2%, 2 to 60%, and more than 60% area occupancy. An average of 67 eggs was counted in each category. Scale bar, 50 μm. (E) Meiosis II occurs normally in V32-gal4/UAS-SAK eggs with no visible centrosomes (n = 110). i and ii indicate magnified fields. Scale bars, 50 μm (left) and 10 μm (inset) in each wild-type and overexpressing set. γ-tubulin, red; α-tubulin, green; and DNA, blue.

There is precedent for defects in de novo–formed centrioles (5, 30). We confirmed the presence of SAK (fig. S7) and two other molecules required for centriole duplication: DSAS-6 (fig. S6) (31) and DSAS-4 (16) (Fig. 3, A and B). We also detected PCM components, including γ-tubulin (Fig. 3A), centrosomin (CNN), and centrosomal protein 190 (CP190) (fig. S8, A and B). Moreover, electron microscopy showed that centrioles in both embryos and eggs overexpressing SAK were structurally normal (Fig. 3C). It also showed the presence of procentrioles next to the completed ones in both embryos and eggs (fig. S9 and Fig. 3C), a result suggesting that SAK-induced centrioles can duplicate.

Fig. 3.

De novo– and canonical-formed centrosomes show centriolar and centrosomal markers and are structurally normal. (A and B) Centrosomes in both 0 to 1 hour embryos and eggs overexpressing SAK contain γ-tubulin, DSAS-6, DSAS-4, and SAK. (A) γ-tubulin, red; α-tubulin, green in left and red in middle and right; green fluorescent protein (GFP)–DSAS6, green; DSAS4-GFP, green; and DNA, blue. Scale bar, 10 μm. (B) SAK, green; D-PLP, red; and α-tubulin, blue. Scale bar, 10 μm. (C) De novo– and canonical-formed centrosomes are structurally normal by transmission electron microscopy. Bottom images are higher-magnification examples of centrioles in each condition. Asterisks indicate duplicating centrioles. Scale bars as indicated.

Our results show that SAK is sufficient to induce both canonical and de novo centriole biogenesis. If both rely on self-assembly of the structure, we would predict the use of the same regulatory molecules. We examined the dependency of SAK-promoted centriole biogenesis on DSAS-4 and DSAS-6. We took advantage of the fact that centrioles can be eliminated from Drosophila tissue culture cells (10). After depletion of SAK in four rounds of RNA interference (RNAi) over a period of 16 days, more than 80% of the cells lacked centrioles, presumably because the remainder are diluted in each division cycle (Fig. 4A) (10). Subsequent overexpression of SAK led to a clear increase in the number of cells with several centrosomes (from 4 to 48%) (Fig. 4, A and B). Depletion of DSAS-6 or DSAS-4 prevented SAK-induced centrosome biogenesis in cells with and without centrioles (Fig. 4, A to C, and fig. S10).

Fig. 4.

Canonical and de novo SAK-induced centrosome amplification is dependent on DSAS-6 and DSAS-4. (A and B) Overexpression of myc-SAK in S2 cells devoid of centrosomes leads to de novo centrosome formation that is dependent on DSAS-6 and DSAS-4. Cells were submitted to SAK RNAi for 16 days (four retransfections), leading to a population of cells where 80% showed no centrioles. Cells were then transfected with DNA (myc or mycSAK) and double-stranded RNA (GFP, DSAS4, or DSAS6) according to the scheme in (B) and stained for myc and D-PLP. (A) Myc and myc-SAK, green, and D-PLP, red. (B) The number of centrosomes per myc-positive cell was counted. The proportion of cells showing more than two centrosomes in each condition is shown as a ratio over the control (cells transfected with myc). (C) Overexpression of mycSAK in S2 cells leads to centrosome amplification that is dependent on DSAS-6 and DSAS-4. The same experiment as in (B) was performed, but this time in a population of cells not submitted previously to SAK RNAi. Scale bar, 10 μm. A minimum of 100 cells were counted for each condition in each of three independent experiments. Error bars indicate standard deviation.

Our results suggest that centriole biogenesis is a template-free self-assembly process that is locally triggered and regulated by molecules such as SAK, DSAS-6, and DSAS-4. What could be the role of the mother centriole? The presence of SAK at the centriole (Fig. 3B and fig. S7) (9, 10) and the fact that assembly is faster in the presence of centrioles (fig. S5) (5, 6) suggest that the mother centriole is not a bona fide template but a platform for regulatory molecules, hence catalyzing and regulating daughter centriole assembly. The establishment of that platform is probably less efficient in the absence of centrioles. The mother centriole could in principle establish a temporally and spatially regulated gradient of SAK activity, as demonstrated for RanGTP, a small guanosine triphosphatase involved in spindle assembly (32), perhaps counteracted in the cytoplasm by other molecules. Our data and that of other groups also point to a role for centrioles in regulating total centriole number, because their presence precludes de novo formation (Fig. 1D and fig. S11) (5, 6). This is true even in a large embryo (∼800 μm) containing very large amounts of SAK (Fig. 1D). Whether this indicates sequestering of active SAK or its substrates in existing centriolar structures or an active inhibitory effect of centrioles upon de novo assembly requires further study.

The regulation of SAK activity is essential in the control of centriole number (fig. S11) and may be a parameter that is regulated according to cellular needs, because multiciliated cells of the respiratory tract have high SAK levels (19). The activity of SAK may be inhibited in the acentriolar female meiosis, as de novo centrosome formation only occurs after meiosis exit in eggs overexpressing SAK (Fig. 2). Drosophila eggs and embryos should provide an ideal experimental system for further analyses of the control of centriole biogenesis and how it may go awry in cancer.

Supporting Online Material

Materials and Methods

Figs. S1 to S11


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article