page 999). Chromosomes do so by capturing spindle microtubules from both poles.
Spindle assembly and DNA replication occur simultaneously in budding yeast. So if DNA synthesis stalls—from a lack of nucleotides, for instance—yeast cells must prevent untimely spindle elongation until replication resumes. This S-phase checkpoint is controlled by the kinase Rad53, which both maintains replication forks during stalls and prevents spindle elongation. The new results suggest these processes may be linked mechanistically.
By maintaining fork integrity as nucleotides begin to run out, Rad53 may ensure that at least some centromeres are replicated (budding yeast centromeres are copied from early-firing origins) even if the rest of the genome is not. This would allow a few chromosome to achieve bipolar attachment and generate traction forces to resist spindle elongation. As would be expected based on this model, chromosomes engineered with two centromeres bypassed the need for Rad53 during replication stalls. So did cells containing mini chromosomes with origins immediately adjacent to their centromeres (so that forks have only a short distance to travel).
Mutations that interfere with centromere–spindle attachments caused spindle elongation during replication stalls, similar to that seen in rad53 mutants. Mutation of a kinase that promotes bipolar attachment, called Ipl1, had similar effects.
The S-phase checkpoint also down-regulates proteins required for anaphase spindle extension (Krishnan et al. Mol. Cell. 2004. 16:687–700). Thus, bipolar chromosome–spindle attachments and the regulation of spindle dynamics may combine to block untimely spindle extension during S-phase arrest.