In vitro reconstitution of the entire chromatin replication pathway



  • Loading, locking and activation of the pre-replication complex reconstituted andvisualised in vitro

  • Naked DNA and chromatin replication entirely reconstituted in vitro

Eukaryotic genomes require multiple origins of replication to ensure that cells can divide rapidly and in a timely fashion; in the case of mammalian genomes, replication origins number in the tens of thousands. The problem of how complete and accurate eukaryotic replication is orchestrated has preoccupied John Diffley for decades. In the last five years, a series of papers from his lab has settled some longstanding questions, and opened the door to studying in exquisite detail how DNA replication occurs in the physiological context of the cell.

There are two distinct and mutually exclusive steps in the initiation of DNA replication. In the first step, which only occurs during G1 of the cell cycle, the pre-replicative complex (preRC) assembles on replication origins. The preRC contains an inactive form of the Mcm2-7 complex, a ring of six related ATPases that comprises the core motor domain of the replicative DNA helicase, the enzyme that separates and unwinds double stranded DNA into two single strands. In the second step, which takes place during S phase, the origins fire: the loaded DNA helicase is activated by a host of other protein components, and initiation factors also recruit DNA primase and DNA polymerase, the enzymes required for semi-conservative replication.

Over twenty years ago, Diffley decided that to properly understand DNA replication, he needed to study the entire process in vitro. This dauntingly long-term venture is what has recently come to fruition, yielding spectacular insights.

Prior to this quinquennium, Diffley’s lab had reconstituted the MCM helicase loading step from budding yeast extracts, purifying MCM and the other preRC complex components, ORC (the origin recognition complex), Cdc6 and Cdt1, using as bait synthetic origins linked to magnetic beads.

They showed that MCM loading by ORC, Cdc6 and Cdt1 is a two-step process, with ATP binding necessary for recruitment of the complex, and ATP hydrolysis by the MCM helicase required for its stabilisation on DNA. MCMs bind DNA as a face-to-face double hexamer, connected by their N-terminal rings, with the DNA running through the central channel (Abid Ali et al, 2017; in collaboration with Alessandro Costa’s group).

Recent work has elucidated how the MCM double hexamer is loaded onto DNA. First, two ORC binding sites are required, in a head-to-head orientation. In natural replication origins, one of these will be high affinity, and the other will be chosen from multiple low affinity ORC binding sites lying downstream. The process begins when one ORC binds specifically to the high affinity binding site, and recruits the first MCM hexamer, which in budding yeast is stably bound to the Cdt1 protein. Although different from the mammalian situation, the Cdt1 interaction has been very useful for elucidating the mechanism by which MCM is loaded. Using single-particle EM, the Diffley and Costa groups again collaborated to show that Cdt1 binds to the MCM ring and holds open a gate between Mcm2 and Mcm5. It is through this gate that the DNA is inserted into the central channel as ORC loads the MCM onto DNA (Frigola et al, 2017).

In the next step, ORC recruits MCM onto DNA via an interaction with Mcm3, and following hydrolysis of ATP by the MCM subunits, the Mcm2-5 gate closes, ORC, Cdt1 and Cdc6 are released, and the first MCM hexamer snaps shut around the DNA.

Loading of the second hexamer happens in a quasi-symmetrical fashion. Using a combination of biochemical methods (Coster and Diffley, 2017) and time-resolved cryo-EM (Miller et al, 2019; in collaboration with the Costa group), the Diffley and Costa labs showed that a second ORC uses a novel interface to bind the N-terminal face of the hexamer sitting at the first origin. This ORC, now bound to one of the lower affinity sites, loads a second hexamer by the same mechanism as the first. The two hexamers are then free to slide together; the lab is currently resolving exactly how this happens.

Image and quote

The MCM double-hexamer formation reaction visualised by cryo-EM

The MCM double-hexamer formation reaction visualised by cryo-EM

The size of the Crick means it’s big enough that I can almost always find someone with expertise I need for aspects of our work, but small enough that I know all of my colleagues well enough to freely contact them.

John Diffley

Once the double hexamer is assembled, it must be activated, giving rise to a new topological problem: how does an inactive helicase bound around double-stranded DNA transform into an active helicase working on single-stranded DNA? The double hexamer bound around the double-stranded DNA has to somehow melt the DNA inside its central channel, open up and extrude one strand, and then reclose around the correct single strand of DNA. Diffley’s approach to this problem has been to reconstitute the complete replication process in vitro, from loading and activating the MCM helicase to seeing leading and lagging strand synthesis, a tour de force involving the purification of 42 separate polypeptides that together form 16 separate replication factors (Yeeles et al, 2015; Yeeles et al, 2017).

Using this system, Diffley and his group have shown (Coster et al, 2014; Douglas et al, 2018) that the double hexamer hydrolyses ATP while it is being loaded onto DNA and remains bound to ADP. A set of firing factors are then recruited by two activating kinases, DDK and CDK, and this leads to release of ADP by the double hexamer.

Detailed dissection of the roles of DDK and CDK showed that DDK phosphorylates at least two of the MCM subunits, Mcm4 and Mcm6, which generates binding sites for two proteins called Sld3 and Sld7. Sld3 is important (Deegan et al, 2016) as it brings in Cdc45 protein, which is present in the active helicase. CDK phosphorylates Sld2 and Sld3, which can then bind Dbp11, bringing Sld2 and Sld3 together at the origin; phosphorylated Sld3 recruits other essential firing factors, and finally, Sld2 recruits GINS, required for the active helicase, and the leading strand DNA polymerase Polε.

Following these events, there is an important conformational change when ATP rebinds to the MCM subunits, triggering their separation into two single hexamers. ATP rebinding also leads to the stable association of Cdc45 and GINS, and finally, the first origin melts. What happens next is still a subject of enquiry, but involves Mcm10, which mediates displacement of one strand, giving rise to the active form of the helicase. Finally, the two active helicases generated from each MCM double hexamer move towards each other, somehow passing within the origin (Douglas et al 2018) as replication commences.

With in vitro reconstitution in hand, the DNA replication field now has a powerful way of beginning to dissect how the replicative machinery interacts with all of the other processes associated with DNA. Diffley is at the forefront of this work, collaborating with other Crick labs on among other things, sister chromatid cohesion and the mechanism of action of the repair-related replication checkpoint. Their current main focus, though, is chromatin assembly. For this last, they have developed an in vitro system which replicates chromatin at physiological rates, finding that an additional factor, the histone chaperone FACT, is essential for replication (Kurat et al, 2017). Intriguingly, examination of the replicated products reveals that the resident nucleosomes being displaced ahead of the replicative fork are being put back on labelled nascent DNA, providing an early glimpse of how epigenetic inheritance may work. The next target, as well as a deeper understanding of the processes currently under study, is the complete reconstitution of DNA replication in humans.

Model for helicase activation

Model for helicase activation