The safety catch on DNA replication now has a structure, and a mutation that breaks it

A team at the MRC Laboratory of Medical Sciences and Imperial College London has published structural snapshots showing how the protein pair Sld3 and Sld7 verifies that the MCM2-7 helicase has been switched on, and then recruits Cdc45 to activate it ^[1][3]. The consequence is that a control step decades of genetics could infer but not see is now a specific set of protein surfaces, one of which the authors broke on purpose ^[4][13].

The setup is well established. Before a cell can copy its genome, it loads the DNA-copying motor, a six-subunit complex called MCM2-7, onto DNA, and deliberately keeps that motor inactive so replication does not start at the wrong time ^[2]. How the switch is thrown has not been fully understood despite decades of work ^[4].

Earlier work from other groups had shown that a flexible section of the Mcm4 subunit behaves like a safety catch, physically covering key surfaces on Mcm4 itself ^[5]. Speck's group reports that the same segment also covers surfaces on the neighbouring Mcm6 subunit ^[6], so the catch sits across two of the six subunits in the ring ^[7]. A phosphate tag added by the kinase DDK releases the catch and exposes those surfaces ^[8]. That is the structural account of how phosphorylation turns an inactive helicase into one that is ready to be activated ^[9].

What Sld3 does next is the part worth noting for anyone who cares about how biological checkpoints are actually implemented. According to the authors, Sld3, positioned by its partner Sld7, acts as a sensor: it binds the newly exposed regions of Mcm4 and Mcm6, and therefore only proceeds once release has occurred ^[10]. Sld3 then works as an adapter. It first anchors to Mcm2, senses that activation has happened, and repositions across the helicase to hand Cdc45 to a different site, the interface between Mcm2 and Mcm5 ^[11]. Cdc45 later becomes part of the active CMG helicase that unwinds the double helix ^[12].

The test is the useful bit. When the team changed the amino acids at the newly identified Sld3-Cdc45 contact, the machinery still bound the helicase but could no longer recruit Cdc45 ^[13]. That separates binding from delivery and makes the contact point load-bearing rather than incidental.

The study also caught an intermediate between the inactive helicase and the finished CMG motor, with Cdc45 in a partially connected state rather than its final position ^[14]. The next step, arrival of the GINS complex to stabilise Cdc45 and complete the machine, is proposed rather than observed here ^[14].

Two limits are stated plainly by the researchers. The work used yeast proteins, and while the core replication machinery is highly conserved, the suggestion that Treslin, the human counterpart of Sld3, recruits Cdc45 by a similar principle rests on structural evidence and has not been tested experimentally ^[15]. And there is no treatment attached: the authors frame the value as understanding how genomes are duplicated accurately, with errors linked to genome instability and to diseases in which replication is disrupted ^[16].

What to watch: whether the Treslin experiment gets done in a human system, and whether the mutated Sld3-Cdc45 interface holds up as a clean separation-of-function tool in cells rather than only in reconstituted structures ^[13][15].