Researchers filmed the enzyme APE1 hunting DNA damage one molecule at a time. Its shapeless tail turned out to do the finding.
Every day, each cell in the human body loses somewhere between 10,000 and 20,000 individual letters from its DNA. The bases detach from the sugar backbone, leaving gaps called apurinic/apyrimidinic (AP) sites.
The opposite strand still carries the missing information. But the gap has to be found and patched before the cell tries to copy or read that stretch, and a human cell holds about 3 billion base pairs to search through.
A team of Korean researchers has now watched that search happen, one molecule at a time. The enzyme does not hunt at random. It slides.
The enzyme is APE1, short for apurinic/apyrimidinic endonuclease 1. It cuts the DNA backbone at an AP site so the rest of the base excision repair crew can rebuild the missing letter. Its chemistry has been described in detail for decades. How it locates a lesion in the first place had not been settled.
Gwangrog Lee of KAIST led the work with Ja Yil Lee of UNIST and Jejoong Yoo of Sungkyunkwan University. Ja Yil Lee also holds a position at the Center for Genomic Integrity at the Institute for Basic Science.
The team used three methods at once. Single-molecule FRET tracked the changing distance between a dye on the enzyme and a dye on the DNA.
A DNA curtain assay stretched out lambda phage DNA, 47,472 base pairs long, and filmed the enzyme traveling along it. All-atom simulations then modeled the same motion at the level of individual atoms.
The numbers were clear. APE1 binds to an AP site roughly 1,500 times more tightly than to undamaged DNA. On the stretched DNA, 86 percent of the enzyme molecules moved rather than remained stationary. And about 66 percent of the time, APE1 reached a damage site by sliding to it rather than colliding with it in solution.
The result that stands out involves a portion of the protein that lacks a fixed shape. APE1 has a structured nuclease domain at one end and an intrinsically disordered region, or IDR, at the other. Bacteria carry a related enzyme called ExoIII, and it has no such tail. Every eukaryotic version does.
So the team built two mutants. One had the disordered region deleted. The other had it duplicated. Removing it cut the probability of moving between damage sites by a factor of five. Doubling it changed little in the short-range test, which suggests a single copy already does most of the job.
Stripped of its tail, the enzyme could not bind undamaged DNA at all under the conditions tested. That matters more than it sounds.
Binding undamaged DNA is the entire point of sliding, since the enzyme has to hold onto the strand as it travels between lesions. The tail is the grip.
“This principle could provide a key clue for developing next-generation anticancer drugs,” said Professor Gwangrog Lee, who led the study at KAIST.
Magnesium has long been known as the ion APE1 needs to make its cut. The new work hands it a second job.
Movement between two damage sites was rare when magnesium was stripped out, at a probability of 0.14. Adding the ion raised that number, and the peak reached 0.1 millimolar, close to the free concentration inside a living cell.
A mutant with one magnesium-binding residue swapped out, E96A, showed no response at any concentration tested.
The proposed explanation is electrostatic. A cluster of negatively charged residues sits near the DNA-binding pocket, and DNA itself carries a negatively charged backbone. Magnesium neutralizes the cluster. Without it, the two repel each other, and the enzyme falls off.
The team also ran the APE1-DNA complex through AlphaFold3. The model laid the disordered tail out in an extended position, touching the DNA nowhere in particular. Both the simulations and the microscope showed something different.
The authors flag this result as a real limitation of current structure prediction tools when the question involves a shapeless, moving region rather than a folded one.
Their simulations also singled out one residue, arginine 177, sitting inside the structured domain. Contacts there held steady over long stretches of simulated time. The authors describe that as unexpected, since the residue had not been tied to this function before.
Work on APE1 as a cancer target has been running for years, on the logic that a tumor cell with a hobbled repair crew is easier to kill. It sits alongside a broader push to develop drugs targeting specific molecular weak spots.
Damaged DNA also accumulates with age, and separate research has linked DNA breaks to immune decline and tested ways to reverse aging in human cells.
One more result gives the field a concrete handle. The team ran a catalytically dead mutant, D210N, which finds AP sites normally but cannot cut them. It sat on damage roughly 2.5 times longer than the working enzyme, because it arrived and then had nothing to do.
Search and cut are separable steps. The enzyme uses its shapeless end to stay on the DNA and its folded end to do the chemistry, and those two functions can be pulled apart in the lab. For an enzyme this heavily studied, that is a new target.
The same molecule sits at the center of a much larger effort to read genetic information across the tree of life, from human cells to the genomes of every known eukaryotic species. Reading DNA and maintaining it turn out to be two very different problems.
The full study was published in the journal Nucleic Acids Research.
Donate spare computing power to Folding@home if you want to contribute directly to work like this. The project runs molecular dynamics simulations of proteins on volunteer machines and has been doing it since 2000.
Enroll in the All of Us Research Program if you live in the United States and want your own health data included in research. The program took a sharp budget cut in 2025 and now collects samples at a smaller scale.
Support the American Association for Cancer Research if funding is your preferred route. APE1 is part of the base excision repair pathway that several cancer drug programs build around.
Browse Zooniverse if you want to help with research without a lab, as volunteers there classify images and data for teams working across biology and medicine.
Use National DNA Day, observed on April 25, as a hook to share reliable information about genetics or to host an educational event. The date marks the 1953 publication of the double helix structure and the completion of the Human Genome Project in 2003.
—–
Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.
Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.
—–