DNA study reveals basic principle of gene regulation

A study of where and how an enzyme cuts DNA may have inadvertently revealed a basic principle of gene regulation, say researchers in Boston Children's Hospital's Program in Cellular and Molecular Medicine (PCMM). The study, reported in the journal Cell, suggests that the cell can lock or "sandbox" genes and enzymes that act on them within loops of DNA and protein, confining their activity to minimize the risk of genetic disaster.

In our cells, DNA and its associated proteins—a combination called chromatin—are folded and wrapped in complex ways to form chromosomes. Researchers have long noted within a chromosome, the chromatin is organized into a series of loops. These loops can range in size from a few thousand to nearly 2.5 million base pairs, large enough to contain one or more complete genes.

The study team—led by co-first authors Jiazhi Hu, PhD, and Yu Zhang, PhD, and senior author Frederick Alt, PhD—believes these loops may form the backbone of a fundamental organizing principle for genomic processes.

"These loops are hardwired. The whole genome is organized into them," explained Alt, director of the PCMM and a professor of genetics at Harvard Medical School. "We think that they are there to restrict enzyme activity so that processes occur in a very organized fashion. We think this may be a basic biological principle that's going to show up in other processes like replication and transcription."

The enzyme the team studied, called RAG, cuts DNA at specific locations in antibody-producing B-cells, helping create the immune system's immense diversity of antibodies. Initially the researchers set out to see whether RAG would cut DNA more broadly across the genome. To do so, they used an assay called high-throughput genome-wide translocation sequencing (HTGTS, which Alt's lab has applied to study the precision of the gene editing systems like CRISPR/Cas9) to map all of the possible RAG targets in the genome.

The team was struck to find that while the answer is yes, RAG will cut DNA elsewhere in the genome, its activity normally is exclusively confined within chromatin loops. When they zoomed in, the researchers were surprised to find that RAG bound to the loops and physically travelled around them in one direction or another (depending on which way the enzyme was pointing when it bound to target sites) much like a train on a track, sampling the loop's DNA sequences looking for sites to cut. The enzyme stopped once it reached the junction point at the bottom of a loop (where the loop joins the rest of the genome).

Alt's team thinks that loops work in concert with the cell's DNA damage machinery to "sandbox" genes and the enzymes that act on them, containing them to reduce the risk that the cuts an enzyme makes to DNA will trigger damaging genetic events such as translocations (fusions of one chromosome to another, which can lead to cancer). Indeed, the study team showed that in cells lacking a DNA repair protein called ATM, RAG could escape a loop and make cuts elsewhere in the genome. These off-target cuts could generate potentially cancerous gene translocations.

"What this tells us is that an enzyme can bind in a loop, and its activity will be restricted to just within that loop," Alt said. "The loops take what could be very dangerous processes and lock them up."

He added that the team's basic discovery—that an enzyme that interacts with DNA can get into a loop and actively explore that loop in two directions—could teach us something about the basic principles of genome organization. The nature of the "engine" that moves RAG around loops, however, is not yet clear, and needs to be elucidated.

"Why have loops? Why did the genome evolve that way?" Alt asked. "We find that loops can restrict processes within them, and allow enzymes to sample sequences within them in directional ways. If RAG can do it, other enzymes like transcriptional or DNA polymerization factors should be able to do it too. With new technologies like HTGTS a lot more can be done to see how loops might organize such fundamental processes."