MIT finds genes reshape DNA folding

DNA is not just a storage medium. It is also a physical object — twisted, stretched, and crowded inside the nucleus — and that physical state can change what nearby genes do. That matters because most gene-control stories focus on chemistry: proteins bind here, a tag gets added there, a gene turns on or off. But cells also have mechanics, and those mechanics have been harder to pin down. MIT researchers now say they’ve directly shown one of those mechanical effects in human cells. When one gene is transcribed, the act of reading it rewinds and unwinds nearby DNA in ways that can either help or hinder neighboring genes. The team published the work in *Science* on April 30, 2026, with Christopher Johnstone as lead author and Katie Galloway as senior author. ### What is the actual new idea? The core claim is simple. A gene turning on does not stay local. As RNA polymerase moves along DNA to copy a gene into RNA, it changes the twisting of the DNA around it. DNA upstream gets looser. DNA downstream gets tighter. That shift changes how easy it is for polymerase to latch onto neighboring genes and transcribe them too. So neighboring genes are not just sharing regulatory proteins or enhancers. They are also sharing mechanical stress. ### Why would upstream and downstream behave differently? Because transcription twists DNA unevenly. Think of a phone cord or a rope getting torque pushed through it — one side relaxes while the other side bunches up. The analogy is not perfect, but it captures the point: motion at one spot sends strain along the line. In DNA, that strain shows up as supercoiling. Looser DNA upstream is easier to open and read. Tighter DNA downstream is harder to access. That gives the effect a direction. It is not just “neighboring genes influence each other.” It is “where the neighbor sits matters.” ### What did the MIT team actually test? They focused on what they call gene syntax — the order and orientation of genes on a DNA strand. The group had modeled this idea back in 2022. In the new paper, they tested it experimentally in human cells using engineered reporter circuits and chromatin-structure measurements. They looked at arrangements like tandem, divergent, and convergent genes, then measured how switching one gene on changed the behavior of another. The result was that syntax-specific coupling showed up clearly. Upstream genes tended to be activated. Downstream genes tended to be suppressed. ### Why is DNA folding in the headline? Because the twist does not stay as a tiny local kink. The study ties transcription-driven supercoiling to larger three-dimensional DNA structures. The team used Region Capture Micro-C to measure induction-dependent formation of supercoiled plectonemes — basically braided, writhed DNA conformations — and linked those structures to changes in neighboring gene expression. So this is not only about “gene A nudges gene B.” It is about transcription reshaping local chromatin architecture. ### Why does this matter for synthetic biology? Most synthetic gene circuits are designed mainly around biochemical parts — promoters, repressors, enhancers, feedback loops. This work says the layout itself is also a design parameter. Put the same genes in a different order, or flip their orientation, and the circuit can behave differently even if the parts list stays the same. That is a big deal for anyone trying to build reliable cell therapies, biosensors, or programmable cell states. ### Does this change how we think about natural genomes too? Yes — at least as an inference from the paper. If gene order can mechanically couple expression, then genome organization may encode regulatory logic more directly than people assumed. Not just which genes exist, but where they sit relative to each other, could shape coordinated activity. The supplementary material even points to analyses across multiple eukaryotic genomes, suggesting this is not just a weird lab artifact. ### What is the bottom line? The bottom line is that genes are not isolated lines of code. They behave more like parts mounted on the same flexible machine. MIT’s study adds a physical layer to gene regulation — one where transcription changes DNA folding, and DNA folding feeds back on transcription. That makes biology messier, but also more programmable.