DNA synthesis relies on a hoop-shaped complex, or "clamp" around DNA to keep it on track.
Processivity is a big issue for biological polymerases. An RNA polymerase needs to stay on its template until it reaches the end. Otherwise, the mRNA might be truncated and the resulting protein would be incomplete. Such proteins frequently have an activity opposite to that which the complete protein has, either because of a particular domain structure that puts key domains at the end, or just because it gums up the works instead of being a well-oiled cog in those works.
Likewise for DNA polymerases that replicate genomes. DNA polymerases that do small repairs may jump in for only a few nucleotides, but for earnest replication of entire genomes, you need a polymerase that chugs along reliably, for long distances. Evolution has come up with an elegant, if obvious, solution- mate the polymerase to another protein complex that firmly encircles the DNA like a hula hoop and doesn't let go. But this leads to other questions ... how does such a complex assemble where it is supposed to, and what happens to it later on?
Whatever the answers to those questions, this solution has been around for a very long time. The structures of the replication-associated sliding clamp from bacteria and from humans look virtually identical:
 |
| Human (left), and E. coli (right) sliding clamp protein complexes that facilitate DNA replication. Where does the DNA go? Obviously through the middle. These proteins do not bind to the DNA in any sequence specific way, but have positive charges arrayed around the inner ring, to gently stay in contact with negatively charged DNA. |
The clamp complexes assemble into extremely stable rings right after they are synthesized off the ribosomal assembly line. That means that they have to be pried apart again to get put on DNA. That is the job of "clamp loaders"- yet another protein complex that orchestrate the proper placement of these clamps. While the clamp proteins are pretty simple affairs- pure structures lacking any enzymatic activity- the clamp loaders are ATP-ases and quite dynamic.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3683903/
 |
| An extremely simplistic model of clamp loading, with the loader prying open the clamp to allow it to admit double stranded DNA. |
The job of a clamp loader is to recognize the correct location in DNA, to bind a free clamp complex, to pry open that complex, transiently allowing the double stranded DNA to enter, to detect when the DNA has successfully been loaded into the clamp, and lastly to detach again from the resulting complex. It is a big job, obviously. What is the correct location? It is a fork where single stranded DNA meets double stranded DNA- i.e. where DNA replication has been primed by a special replication origin selection, opening, and priming process. Clamp loaders are also rings, with five highly related subunits all around. Their ATPase activity (active on each subunit) allows them to twist, which is how they manage to twist the clamp as well, to open it up.
A recent paper extended structural knowledge about clamp loaders. These authors used the new cryo-electron microscopy methods to obtain high resolution structures of these complexes in a variety of states. These states were made available by virtue of using a slowly hydrolyzing form of ATP that gave them the whole spectrum, then frozen and photographed with electrons. They capture beautifully the sequence of events, where the loader first tentatively binds a clamp, and then opens it up wide enough to accept DNA. The DNA binding groove is not open until the clamp is bound, enforcing this forward sequence of clamp binding, then DNA binding (availability), then clamp opening. The authors do not, however, provide reason to think that this opening is dependent on DNA being already bound, so it is possible that this opening complex can be futile- opening and closing repeatedly before encountering the right replication fork. This is probably unlikely in practice, though, because the clamp loader complexes are quite rare and probably have other interactions that position them at replication forks before this process even begins. Additionally, the loader + clamp complex may remain open as long as necessary, until it encounters the right fork location.
 |
| An early form of the clamp loading complex is shown, when it first binds a clamp. The five subunits of the clamp loader are shown in color, while the clamp is shown at bottom in gray. In B, the contacts between the two complexes are shown in color, on the clamp, which is composed of proteins called PCNA. The different structures are of various stably inhibited forms, and the A-gate is where DNA is bound during the loading process. |
 |
| Later on, after the loader has expended some energy, it has more contacts apparent with the clamp (right) and has pried it open, wide enough to accept a strand of DNA. |
 |
| With DNA, the whole complex contracts a little again, and you can see the continuous flow of DNA (yellow) through the cleft of the loader complex down through the clamp complex. |
The last step is closure, when the clamp closes fully around the DNA, and the loader complex unbinds. This seems to be when ATP is actually hydrolyzed, implying that the opened complex is the ATP binding state, the closed (free) loader complex is the ADP bound state, and that successful DNA binding to is what stimulates ATP hydrolysis. Finally, another structure- a closeup of the DNA as bound, shows that the end of the primer strand, which sits right at the crux of the DNA fork, is specifically bound with its last base flipped around into the protein. This provides part of the mechanism of how the clamp loader feels its way to the right place at replication forks, after the briefly interacting polymerases that create such primers have fallen off. It is likely that this clamp loader complex engages in interaction with the final processive DNA polymerase to help it find the fork and clamp, but notably, the same face of the clamp interacts with both the loader and with that polymerase, so the handoff can not involve both binding at the same time at the same place.
 |
| A blowup of just the DNA within the fully bound complex above. Note how the top base of the yellow primer strand is flipped out from the rest of the helix, due to interactions with the loader protein complex. |
