These are my notes from lecture 10 in Harvard’s BCMP 200: Molecular Biology course, delivered by Johannes Walter on September 26, 2014.
When two replication forks meet each other head-on, there will be two forks of dsDNA meeting at a plectonemic supercoil of the original dsDNA duplex. DNA topoisomerase I and II change the linking number Lk in units of 1 and 2 respectively, and either can help to resolve this plectoneme. However it is currently believed that the dsDNA arms beyond the fork are actually rotated to relieve the topological stress. Once the stress is relieved, somehow the last bit of DNA in between the forks must be replicated (apparently we don’t know how this is done). Once this occurs, the two daughter duplexes will be catenated (bent around each other once) and another topoisomerase will be required to separate them from each other. One school of thought holds that the two replication forks slow down as they get closer to one another, and that the daughter duplexes disentange before the forks complete replication. Another is that the two forks collide in a steric clash which then has to be resolved.
Xenopus egg extracts are often used for studying DNA replication, because they contain huge stockpiles of the proteins needed to perform replication. The eggs are originally arrested in metaphase II of meiosis, but when they are crushed and the contents extracted, they are not really in any particular phase, and the components present can be used to perform replication. You can add geminin to license the origins, and then add nuclear extract containing CDK activity to promote initiation of replication. When you add plasmids to these extracts, ORC will bind the plasmid in a sequence-independent fashion, depositing MCM2-7 in a sequence-independent fashion. This sequence independence is crucial for being able to replicate an arbitrary DNA sequence (even non-eukaryotic sequences) in Xenopus extracts. Once replication of a plasmid completes, you end up with monomeric daughter plasmids - i.e. the two circles are not catenated.
You can add 32P-labeled ATP to radiolabel the nascent daughter plasmids, and then run them on a gel to characterize them. You can also use immunodepletion (pulling down proteins with a high-affinity antibody conjugated to a sepharose bead) to remove any protein(s) from the extracts to see what happens to replication in their absence. For instance you can show that immunodepletion of ORC prevents replication, and then you can add recombinant ORC back in to rescue replication (thus ruling out the possibility that you prevented replication by virtue of depleting some other protein that co-IPed with ORC).