Can I Please See Your License?

As we kick off the focus of this book—a deep dive into human genome replication—we’ll see that the process involves many different proteins operating in an extremely well-orchestrated and interconnected manner. Tight regulation of the process is necessary to ensure the genome is replicated once and only once and with extremely high accuracy.  

Although the copying of the human genome occurs during S-phase of the cell cycle, the events described in this post--that is, the attachment of the first initiator protein complex (called ORC) to a replication origin and the "licensing" of that origin--take place in early- and mid-G1. We'll soon see why this must be the case--why there must be temporal separation between origin licensing and DNA replication, itself.

A "licensed" replication origin is defined as one capable of initiating DNA replication in S-phase. Once an origin is licensed in G1, its ready to serve in S phase as the gathering point for the large array of replication-related proteins that will ultimately form two replisomes and two corresponding replication forks that will move in opposite directions away from the origin copying DNA.

The terms "replisome" and "replication fork" are often used to label the site of active DNA replication but they refer to slightly different things. "Replisome" refers to the totality of the protein machinery present at the site of DNA replication. "Replication fork" refers to the Y-shaped form the DNA takes at the site of replisome action. Think of the replication fork as the construction site and the replisome as the crew of machines and workers actively building there. Our focus here is the replisome and, specifically, how the replication protein machinery forms.

Licensing is the first step in replisome formation. It cannot be permitted in S phase because, if it could, an origin could be licensed in S phase multiple times which would cause the DNA near that origin to be replicated multiple times in S phase. That would create, in a word, a mess. Some genomic regions would be replicated once while others would be copied two, three, or more times.

Let me offer a human analogy for replication origin licensing. Let's say that the local county fair will be this June, and everyone in your very generous county is emailed one free, non-transferable ticket in May. Each ticket allows for only one entry to the fair. And no additional tickets can be purchased in June. Given that scenario, everyone in the county will be able to attend the fair once and only once in June.

Analogously, every replication origin will be licensed (receive a ticket) in G1 (May). No additional licensing (ticket purchasing) will be able to occur in S-phase (June). Thus, every replication origin (county resident) will be able to initiate replication (attend the fair) once and only once in S-phase (June)!

That's what G1 licensing accomplishes: it permits one and only one replication per segment of the genome in S phase. Let's take a look at how the cell physically licenses replication origins in G1.

ORC and Cdc6 arrive

The initial protein that lands on and designates a replication origin in G1 is human ORC (for Origin Recognition Complex). ORC (I'll drop "human" since we're not comparing to yeast in this post) is a protein complex made up of six different polypeptide subunits. It's shaped like a not quite fully closed hand that forms a not quite fully closed central channel. When ORC lands on a replication origin, the DNA lodges inside that central channel.

With ORC now in place, a protein called Cdc6 finds ORC and binds to it. Cdc6 attachessuch that it closes off ORC's central channel, trapping the DNA inside. In addition, Cdc6 binding to ORC alters the shape of ORC such that a molecular platform is formed. This platform will very soon serve as the landing spot for another protein complex critical for licensing called MCM2-7.

Cdc6 has an ATP binding site. And an ATP molecule must be attached to the Cdc6 for it to attach to ORC and close the gap. This ATP only needs to be attached. It doesn't have to be cleaved (with the associated release of energy). Soon, though, this Cdc6-attached ATP will have to be cleaved for other energy-requiring licensing steps to occur.

To summarize, we now have the ORC protein complex bound to a Cdc6 molecule with an ATP attached that effectively traps replication origin DNA inside ORC's central channel.

Cdt1 and MCM2-7 arrive

At the same time ORC and Cdc6 are doing their thing, elsewhere in the nucleus Cdt1--a so-called chaperone protein--attaches to another six-protein complex called MCM2-7. (The six subunits are named Mcm2-Mcm7... don't ask me what happened to Mcm1). Cdt1 binding to MCM2-7 stabilizes it in an "open-ring" conformation that's critical for its attachment to DNA. Cdt1 binding also modestly alters the shape in MCM2-7 that makes it compatible for initial loading onto the landing platform created when Cdc6 attached to ORC.

Once on the landing platform, the MCM2-7 now encircles the DNA. Doing this requires energy. This comes from cleavage of the ATP I previously mentioned that's lodged in Cdc6's ATP binding site. When that ATP is cleaved, the energy released affects the shape of Cdt1 which in turn affects the shape MCM2-7 in a way that facilitates DNA binding and encirclement. And this ATP cleavage has another effect: it causes the release of Cdt1 from MCM2-7. Cdt1 has served its role chaperoning MCM2-7 to the origin. It's no longer needed.

To quickly review where we are, ORC is still bound to Cdt6 and to the replication origin. Next to the ORC-Cdc6 complex, one MCM2-7 protein complex has been loaded onto the DNA. But now the cell needs a second MCM2-7 to be loaded.

Let's do it again!

Now we repeat the process. A second Cdt1 attaches to a second MCM2-7 complex in the nucleoplasm and chaperones it to the ORC-Cdc6 platform. Once there, the second MCM2-7 is loaded onto the DNA such that the heads (the fronts) of the two MCM2-7s will face each other. In this "head-to-head" configuration the two MCM2-7s are referred to as the MCM-double hexamer, or MCM-DH.

Now we have the two MCM2-7s loaded, but the cell needs to clean up a few things. This requires energy so a second ATP attached to Cdc6 will have to be cleaved. This second ATP cleavage releases the second Cdt1. Following the departure of the second Cdt1, the ATP cleavage also causes Cdc6 to be released from ORC.

The result of all of this activity? Two MCM2-7 complexes attached to the replication origin DNA in a head-to-head configuration (MCM-DH) with ORC still present.

Preventing re-licensing

The activities I just described happen in early- and mid-G1 phase to 30,000-50,000 human replication origins. These origins are now all capable of initiating replication in S phase.

But the question I posed at the outset of this post concerned how the cell ensures that replication origins fire only once during S phase. So far, I’ve only described how origins are licensed, not how they are later prevented from being licensed. Let’s broach that now.

We've discussed the cell cycle and have emphasizing how a family of master regulator proteins called cyclins determine the phase of the cell cycle at a given point in time. Cyclins do this by activating proteins called CDKs (cyclin-dependent kinases) which, once activated, use ATP hydrolysis to phosphorylate (add a small chemical phosphate group to) different proteins in the nucleus. CDK phosphorylation of proteins effects the appropriate gene expression pattern for that phase and also turns specific proteins on or off.

Thus, in a nutshell, re-licensing in S phase is prevented because the specific cyclins active during late G1 and S phase (cyclin E and later cyclin A) turn on a specific new CDK (CDK2). CDK2 in turn phosphorylates and deactivates all the proteins required for licensing: Cdc6, Cdt1 and ORC. With these three proteins deactivated, replication origins can no longer be licensed in S phase. That’s how licensing is restricted to G1.

For those interested, let’s look at the details. In early- and mid-G1, cyclin D is active. Cyclin D activates CDK4 and CDK6. These two CDKs perform their jobs as master regulators. They define gene expression in early- and mid-G1. But neither is capable of phosphorylating (and thus deactivating) Cdc6, Cdt1 or ORC. So, conditions are such in early- and mid-G1 that Cdc6, Cdt1, and ORC are all in their un-phosphorylated, active forms and are busy licensing replication origins.

Everything changes in late G1 and S phases. In late G1, cyclin D levels decrease, and cyclin E levels increase. Cyclin E activates its partner CDK2. Because cyclin D is no longer present, CDK4 and CDK6 are also inactive. Soon after, in S phase, cyclin E levels decline and cyclin A is expressed, which activates CDK1 and CDK2. CDK2 will play the leading role in our story. It is active in both late G1 and S phases.

As mentioned, activated CDK2 will target Cdc6, Cdt1 and ORC for phosphorylation. What is the effect of these phosphorylations? I'll go one by one.

Cdc6: CDK2 phosphorylation of Cdc6 in late G1 and in S phase marks the protein for export out of the nucleus and for degradation.

Cdt1: Phosphorylation of Cdt1 flags it for the attachment of another protein degradation marker called ubiquitin. Also in S phase, another non-kinase protein called geminin starts to be expressed. It, too, binds to Cdt1, rendering it inactive. So there are two redundant mechanisms for Cdt1 inactivation.

ORC: CDK2 phosphorylates ORC in late G1, weakening its replication origin licensing ability. Then, in S phase, ORC is phosphorylated again by CDK2. The second phosphorylation dissociates ORC from the replication origin and leads to its destruction.

Thus, in late G1 and S phases, CDK2 phosphorylates all three of the proteins required for origin licensing. Phosphorylation renders all of them inactive, either by export out of the nucleus, by degradation, or by attaching to some other protein that impedes function. With these proteins out of the picture, no replication origin licensing can occur in S phase. Tricky!