The figure shows a one-cell embryo, stained for ZYG-1 (green) and β-tubulin (red). ZYG-1 localizes to the poles. © (2001) Elsevier Science.

Centrosomes — the main microtubule organizing centres of the cell — make sure that, upon division, each cell receives an equal number of chromosomes. To this end, centrosomes duplicate once per cell cycle, separate, and then nucleate the two end points of the mitotic spindle. The cell is thought to use two types of regulator to ensure that this crucial duplication takes place: first, cell-cycle factors, which regulate when centrosome duplication takes place; and second, intrinsic factors, which ensure only one round of duplication occurs per cell cycle. Now, reporting in Cell, O'Connell and colleagues show that ZYG-1, a protein kinase that localizes to centrosomes, fufils the second criterion — it mediates duplication but is not required for cell-cycle progression. They also highlight a division of labour between paternal and maternal ZYG-1 — a regulatory mechanism that ensures coordination between the gametes during fertilization.

The zyg-1 gene was previously identified in Caenorhabditis elegans as being essential for cell division. In its absence, cells fail to form a bipolar spindle but instead produce a monopolar spindle. This could occur either because the centrosomes (which consist of a pair of centrioles) have not reproduced properly or because they cannot separate. Looking more closely, O'Connell and colleagues were surprised to find that the monopolar spindles contained a single unpaired centriole — a finding that was not expected in either model. This led them to conclude that ZYG-1 is required for centriole formation.

During fertilization, a pair of centrioles is usually provided by the sperm, indicating that the defect might be due to a paternal effect. Using genetics, the authors asked what happens when only the paternal copy of zyg-1 is lost. They found that the first cell division was blocked owing to the absence of a bipolar spindle. But, surprisingly, during the following divisions, a normal spindle formed and cell division proceeded. Extrapolating from this, O'Connell and colleagues proposed that maternal zyg-1 takes over after the first round of division — a prediction that they then confirmed by temperature-shift experiments.

Still intrigued by the presence of a single unpaired centriole in zyg-1 mutant zygotes, the authors next asked when paternal zyg-1 activity is required, before or after fertilization. They found that paternal ZYG-1 is required before fertilization, during male meiosis, for the synthesis of a second centriole.

Cloning of zyg-1 revealed it to be a unique protein kinase that does not show homology to any specific protein kinase family. The authors also could not tell whether it is a serine or a tyrosine kinase. So is the kinase domain functional? To test this, they compared wild-type and kinase-dead versions of ZYG-1 in vitro. Although they could not find other substrates, they showed that ZYG-1 undergoes auto-phosphorylation. Finally, molecular complementation experiments indicated that the kinase activity is required for rescue of the centrosome duplication defect.

Does the expression pattern of zyg-1 support a role in centrosome duplication? To address this question, O'Connell and colleagues raised an antibody against ZYG-1 and showed that it associates almost exclusively at the centrosomes at the time of duplication (see figure).

As ZYG-1 is required throughout development, the authors propose that it forms part of the basal centrosome duplication machinery, but what it interacts with remains to be seen. Interestingly, they note that the function of ZYG-1 resembles that of Mps1 during assembly of the spindle pole body (yeast's equivalent of the centrosome), indicating that, despite the differences between the division apparatus in mammals and yeast, the basic molecular mechanisms for nucleating microtubules during division might be conserved.