DNA Replication Machinery Sows Seeds of Own Demise

Platform Triggers Destruction of Replication Enabler

A dividing cell can be downright prissy about its DNA. It must replicate its genome once and only once so that each daughter cell receives a complete set of matching chromosomes, no more and no less.

Photo by Graham Ramsay Emily Arias and Johannes Walter discovered that DNA replication machinery can stop itself from copying the same piece of genome more than once before a cell divides.

To duplicate the several billion base pairs of a vertebrate genome, the cell divvies up the work, as if dispatching thousands of monks to transcribe a long document, everyone doing a couple of pages. A big question among biologists, said Johannes Walter, HMS associate professor of biological chemistry and molecular pharmacology, is How do cells make sure the monks don’t copy the same page twice? “This mechanism is arguably the most basic way that cells ensure the stability of their genomes,” he said.

Now graduate student Emily Arias, a Howard Hughes predoctoral fellow in Walter’s lab, has a new answer for that question. The DNA replication machinery contains the means of its own destruction, she and Walter report in the January Nature Cell Biology. “Just by the nature of beginning DNA synthesis,” she said, “the cell creates a way to prevent a second round. ”

In recent years, one protein has emerged as a pivotal player in both the start and finish of DNA replication. Without Cdt1, the cell cannot copy its genome. And as long as Cdt1 remains, the cell cannot stop copying the same piece of DNA over and over. To prepare for DNA replication, this essential protein loads the front blades on the fleet of replication machines that will move along the chromosomes like 50,000 snowplows, pushing histones and nucleosomes out of the way and separating the strands of DNA, baring them to the DNA polymerase scribes.

The cell starts the engines of the synthesis machinery by entering S phase. As the duplication of the genome is set in motion, Cdt1 disappears. The absence of Cdt1 undermines the ability of the replication machines to reassemble and roll down the same genetic block again. Only when mitosis is completed does Cdt1 reappear, ready to put new DNA copying machinery in place for the next round of cell division. So far, researchers have discovered two ways that cells get rid of Cdt1 and prevent redundant DNA copies. A cell can functionally neutralize Cdt1 by sequestering it in a complex with the inhibitor geminin (see Focus 1/12/2001), or it can send Cdt1 down the drain into the cellular garbage disposal.

The newly discovered mechanism involves the DNA replication machinery itself. Last year, in the January 2005 Genes & Development, Arias and Walter reported that Cdt1 destruction was coupled to DNA replication. When they prevented DNA replication, Cdt1 persisted unscathed. And during replication, Cdt1 appeared as a series of modified molecules on the chromatin, decorated by ubiquitin molecules, which label the protein as trash to be destroyed. The findings coupled Cdt1 destruction to DNA replication and revealed the replication fork as the smoking gun for Cdt1 ubiquitination.

Fingering the Trigger
Daunted by the prospect of trying to find Cdt1’s replication-protein partner by painstaking biochemical trial and error, the researchers reasoned through the clues. The molecular step required to trigger Cdt1 destruction coincided with the loading of proliferating cell nuclear antigen (PCNA), a replication factor known to bind to many different proteins. They looked up the sequence of Cdt1, and bingo. Arias and Walter found an eight–amino-acid sequence in Cdt1 that matched a motif known as the PIP box, which allows proteins to bind to PCNA. Then, they lined up Cdt1 homologues from all sequenced organisms. Except for yeast, all other multicellular organisms harbored a perfectly conserved PIP box in otherwise highly divergent regions of Cdt1.

Image courtesy of Emily Arias A copy cat’s one life. To copy its genome once, a dividing cell needs the protein Cdt1 to help assemble prereplication complexes at about 50,000 sites, dubbed origins. But once the replication machinery is in motion, the cell must get rid of Cdt1 to avoid multiple copies starting from the same origin. The doughnut-shaped replication protein PCNA not only tethers DNA polymerase so it does not fall off the DNA, researchers have discovered, but it also rids the cell of Cdt1 by routing it to the trash. Bound to PCNA, Cdt1 accepts the ubiquitin molecules that mark it for disposal.

“We got lucky,” Arias said. “We knew the answer as soon as we found the motif and how conserved it is. It took a year to do all the necessary experiments, but everything else was anticlimactic. Every protein that binds to PCNA has this signature motif.”

PCNA is most famous for its replication duties. The molecular doughnut encircles DNA like shower curtain rings on a rod, tethering the polymerase to the DNA. Besides polymerase, at least 30 other proteins involved in apoptosis, DNA repair, and cell cycle progression contain PCNA-binding PIP boxes.

In Walter’s lab, several experiments confirmed that PCNA is the molecular platform on which Cdt1 destruction is triggered. When Arias mutated the PIP box in Cdt1, for example, the protein hung around and induced multiple rounds of replication within the same S phase. Another set of experiments demonstrated that the destruction mechanism was sufficient by itself to prevent rereplication, eliminating Cdt1 even with a disabled geminin pathway.

The researchers also identified the E3 ubiquitin ligase containing DDB1 and probably Cul4 as the specific ubiquitination pathway, and they showed that the ligase also binds to replication forks via PCNA. Some experiments suggest that one Cdt1 after another binds to PCNA, accepts a ubiquitin from the ligase, and goes off to the proteasome to be destroyed, a process that continues until the Cdt1 is gone.

PCNA only binds with Cdt1 on the chromatin, despite constant PCNA levels throughout the cell cycle. This ensures that Cdt1 is destroyed only in S phase and not in G1, when Cdt1 is needed to set up the DNA-copying machines. “Our hypothesis is that when PCNA is on the DNA, there is a conformational change that makes Cdt1 binding more efficient. We’re testing that,” Arias said.

Not Just a Frog Thing
Most of the studies were conducted in a preparation of sperm chromatin added to egg extract, all from the African clawed toad, so Walter sought confirmation from longtime collaborator Anindya Dutta, formerly of Brigham and Women’s Hospital and now at the University of Virginia Health Sciences Center in Charlottesville, who studies similar processes in mammalian cells.

Last year in Dutta’s lab, HMS MD–PhD student David Takeda showed that human Cdt1 is degraded by at least two proteolysis pathways—one cyclin dependent and the other not. Walter’s results shaved an estimated year off a new, related project in Dutta’s lab. There, researchers confirmed that the same PCNA-dependent Cdt1 destruction was the cyclin-independent pathway they were seeking in human cells. Dutta’s lab first verified the mechanism in UV-damaged human cells, in which Cdt1 is also rapidly depleted, perhaps during the DNA repair process (which involves PCNA) and possibly to avoid replication of a damaged genome. Then the researchers repeated the experiments in the S phase. The findings were reported online in The Journal of Biological Chemistry on Jan. 9.

Another study this month shows the same mechanism occurs in response to DNA damage in fission yeast, perhaps using a PIP box with a different amino-acid sequence.

“You always have to worry it’s just a frog thing,” Arias said, “but in this case, it’s not.”

“The bigger picture tells us that PCNA has a role in protein degradation,” Dutta said. “Not only does it bring proteins together when it needs them, it helps get rid of proteins when the job is done by targeting the protein for degradation.” Dutta suspects other proteins involved in DNA and chromatin modifications may be disposed of in similar fashion.

Taking another perspective, Walter said that the PCNA mechanism represents an interesting way to destroy proteins at a specified place and time. He explained, “It’s unusual for proteolysis to take place on chromatin and be triggered by the docking of a protein involved in DNA metabolism.”

—Carol Cruzan Morton

top