DNA Polymerase Works in Short Bursts Rather than One Long Stretch

Biology lectures teach students that when a cell’s replication machinery comes together, DNA polymerase takes off down the double-helix like a car on a highway, continuously replicating the strand. If an error is made, the same enzyme stops, reverses, and corrects the mistake, then returns to its ceaseless journey to the end of the line.

However, research published in the last 10 to 15 years increasingly challenged that model, suggesting that DNA replication and proofreading involves multiple polymerases.^1,2 Now, in a publication in Nature Communications, a team at Vrije University Amsterdam provided additional evidence that DNA polymerase does not replicate DNA as continuously as once believed.³

“As opposed to those proteins sitting on the DNA very stably bound, they come and go all the time,” said Antoine van Oijen, a molecular biophysicist at the University of Sydney who was not involved in the work. “It’s almost like you’re changing tires while you’re driving.” Unraveling DNA polymerase’s activity can help scientists better understand DNA replication and repair and explore how these processes fail and lead to diseases, such as cancer.

“This finding [happened] by chance,” said Longfu Xu, a biophysicist and postdoctoral researcher in Gijs Wuite’s lab at Vrije University Amsterdam. The team initially set out to study another replication machinery protein, single-stranded DNA-binding protein, and its interaction with DNA polymerase. However, they first needed to establish how the two proteins interacted with DNA independently. While exploring DNA polymerase activity, the team observed something unexpected: The proteins rapidly hopped on and off the nucleic acid.

To determine where DNA polymerase was and what it was doing, Xu and his colleagues combined two methods: confocal microscopy and optical tweezers. The team stretched an 8,000 kilobase strand of DNA between two laser-locked beads. The tethered DNA strand was comprised of a double-stranded (dsDNA) segment that became single-stranded (ssDNA). Using lasers, they applied different amounts of force to mimic the tension DNA normally experiences during replication or proofreading to experimentally promote these enzymatic functions. Then, by adding a fluorescent tag to DNA polymerase, they tracked the enzyme’s progress along and binding dynamics to DNA.

“We can apply tensions on DNA, and we can also visualize the polymerase movements on DNA, but these two data sets are independent. We wanted to correlate them,” Xu said. He explained, though, that synchronizing these two data sets and mapping the protein’s path along the strand posed challenges. However, by tracking fluorescent protein binding at the dsDNA-ssDNA junction the team could overlay these two pieces of information to reveal DNA polymerase’s behavior.

The researchers observed that, on average, a single DNA polymerase molecule remained bound to the nucleic acid at the junction for slightly more than one second—far from the continuous binding that most textbooks describe. Further contrasting from the dogma, during this time, a single enzyme only performed either extension or proofreading, occasionally also pausing on the DNA; rather than backing up to fix an error, the enzyme detached from the nucleic acid to let another bind.

“The idea of having a motor that you put in reverse sounds very appealing to us, but it's much more efficient to throw the motor out,” Wuite explained. Unlike cars, a cell has multiple DNA polymerase motors, so an enzyme that is already in the configuration needed to bind the DNA and correct the error can take over. This exchange takes less energy than the same protein changing conformation to fulfill a different function.

However, DNA polymerase’s activity appeared seamless and uniform, so the team considered that a process existed to help one enzyme pick up where another left off, acting like a memory. They analyzed one extension event and observed that polymerases unbound and rebound multiple times, but each time, they resumed the same function.

To study this further, the team assessed the activity state—enzymatic or paused—before, during, and after a fluorescent polymerase bound the DNA over the course of several experiments. They found that the most common pattern was for the activity to be the same at all three observational points, whether the enzymatic period was during exonuclease repair or DNA extension.

“This experiment is really the nail in the coffin of this model where everything is sitting stably on the DNA,” van Oijen said. He added that structural studies will be important for adding additional context to these mechanisms.

“The real dream is, of course, to see all those different components at the same time, working at the junction where you have single-stranded and double-stranded DNA meet,” Wuite said. Xu and his colleagues have started to work on these experiments.

The researchers are also applying this dual approach to other questions, like studying chromosome segregation. “What you read in biology books about the organization of chromosomes is largely fantasy,” Wuite said. “With our tools, we actually can take some steps forward, and then actually, by understanding how they're really organized, we understand some of the basic things that go right or go wrong.”

Disclosure of Conflicts of Interest: Gijs Wuite is a cofounder and shareholder of LUMICKS, a biological research instrument company, and holds patents related to the methods and technologies described in this story.

References