The timing of DNA replication is thought to be linked to transcriptional activity — actively transcribed genes are replicated early, whereas transcriptionally silent genes replicate later. Nuclear localization has also been implicated in timing, with late-replicating genes clustering near the nuclear periphery. But what determines which genes replicate when? Although the exact mechanisms are not known, a paper in Current Biology now shows a clear link between replication timing and a group of proteins that are involved in transcriptional silencing.

Janet Leatherwood, Rolf Sternglanz and colleagues first tested whether a transcriptional silencer can regulate the time of replication initiation in budding yeast. To do this, they compared the replication of a yeast replication origin, ARS305, with that of the same sequence, but which also contained two copies of a transcriptional silencer called HMR-E that were integrated 225 bp downstream (ARS305:(HMR-E)2). As expected, ARS305 replication intermediates appeared 25 minutes after cells were released from arrest in G1. By contrast, no intermediates were detected from ARS305:(HMR-E)2. Using a yeast mutant strain that lacks the intra-S-phase checkpoint — which allows late-firing origins to be detected — the authors showed that ARS305:(HMR-E)2 can initiate replication, but that it fires about 30 minutes later than ARS305.

The HMR-E silencing system works by recruiting a complex of so-called Sir proteins through binding sites for the origin-recognition complex (ORC), Rap1 and Abf1. So, the authors next asked whether the effect of HMR-E on replication depends on the Sir complex. Mutation of SIR4 , SIR1 or the ORC-binding sites in (HMR-E)2 led to the early replication of ARS305:(HMR-E)2 — indeed, this sequence was replicated at the same time as ARS305.

It seems, then, that the Sir proteins are essential for resetting replication from early to late. But is simple targeting of a Sir protein to an early origin enough to make it fire late? To test this, the authors used a Gal4(1–147)–Sir4 hybrid protein (GBD–Sir4), which can silence transcription around GAL4-binding sites. They then inserted five such sites next to ARS305, to create ARS305:(G)5. As expected, GBD–Sir4 blocked the formation of replication intermediates at ARS305:(G)5. The authors then showed that GBD–Sir4 did not block replication initiation — rather, it reset the origin to fire late.

Leatherwood, Sternglanz and colleagues conclude that Sir proteins are a cause of late replication. But is the structure of the chromatin a more important determinant of replication timing than subnuclear localization? Yes it is, say the authors, as simply tethering ARS305:(G)5 to the nuclear periphery was not enough to confer late replication.