Random Mutations Turn RNA Droplets Into Moving Protocells

Scientists in Germany have taken one of the boldest ideas in biology — building a living cell from scratch — and made it work with little more than RNA and random typos. A research team led by molecular biologist Kerstin Göpfrich at Heidelberg University has shown that simple RNA Droplets, given a few random genetic mutations, transform into cell-like structures with hollow internal cavities, the ability to move, and even a way to store cargo. A mutation rate of around three percent was all it took to produce these complex changes, according to a press release from Heidelberg University. The findings were published in Nature Chemical Engineering on 7 October 2026, and the full study is available via the published paper.

That might sound like science fiction, but it lands squarely in one of the oldest questions in science: how did life emerge from non-living matter? For synthetic biologists, a minimal cell must at least keep hold of its genetic material while still being able to change — change is the raw material of evolution. This experiment suggests that a blob of RNA can already tick both boxes, without a single piece of cell membrane involved.

Droplets that keep their own blueprints

The system is built around an RNA nanostar, a single RNA strand that folds as it is transcribed into four arms with sticky ends, called kissing loops. When enough of these nanostars float in solution, they clump together into liquid-like droplets through a process known as liquid-liquid phase separation — think of oil drops forming in water. Crucially, the DNA template that encodes each nanostar is not washed away when the droplets form. As the RNA Droplets assemble, the nascent RNA drags its own blueprint inside with it, creating a physical link between the genetic instructions and the droplet itself.

“That establishes a basic prerequisite for a system that can evolve,” says William Verstraeten, a doctoral candidate in Göpfrich's research group. This matters because evolution needs heritable information tied to something that can be selected. Without that link, a droplet is just chemistry; with it, the door to Darwinian selection opens a crack.

The droplets are also built differently from most lab-made condensates studied before. Older protocell models typically assemble through loose electrostatic attraction between oppositely charged molecules — a generic glue. Here, the nanostars hold together through sequence-specific base pairing, which means the droplet's physical properties are literally written into its genetic code. Change the sequence, and you change the droplet. That programmability is what let the team test what mutations do to a proto-cell in the first place.

Mutations did the opposite of what was expected

Deliberately scrambling the DNA template was supposed to break the system. The researchers expected mutated droplets to dissolve or fall apart. Instead, the mutated droplets built something new: large internal cavities known as vacuoles, which remained stable for days or even weeks. “Instead, they formed large cavities known as vacuoles, which remained stable for days or even weeks,” says Göpfrich. “They take up small molecules and can store this cargo inside them.”

The threshold was surprisingly low. In a second round of experiments using the standard tools of directed evolution, the team found that roughly seven altered bases per nanostar template — a mutation rate of only about three percent — was enough to trigger vacuole formation. The sequencing work that pinned down the responsible mutations was carried out by the group of Michael Platten at the German Cancer Research Center, using a specialized long-read sequencing technology. That allowed the team to reverse-engineer a defined “model mutant” and reproduce the hollow-droplet effect on demand.

The physics behind the cavities comes down to a traffic jam. As transcription continues, fresh RNA keeps pouring into the already viscous droplet interior faster than it can diffuse, creating local instabilities that open up an internal cavity — in the words of physicist Christoph Weber of the University of Augsburg, whose theoretical modeling mapped the whole process. The experiments confirmed two conditions for vacuole formation: the droplets had to exceed a radius of more than twenty micrometers, and contain over two percent modified material.

Then came the surprise twist. The cavities shift the droplet's center of mass as they form and grow, producing tiny jerky movements on the micrometer scale. Tracking experiments showed that mutated droplets traveled significantly farther than their unmutated counterparts, collided with one another more frequently, and fused into larger droplets as a result. The vacuoles also function as selective storage compartments: small molecules like a 10-nanometer dextran marker accumulated inside them, while larger 50-nanometer lipid vesicles were excluded from the dense RNA matrix.

Why this is a big deal — and why it is not life yet

The story here sits inside a decades-long timeline of trying to recreate life's origins. In the 1920s, Alexander Oparin proposed that coacervate droplets could have been the first compartments on Earth. By 2018, researchers had shown that RNA stays chemically active inside such membraneless droplets. In 2024, teams at the University of Chicago showed how rainwater could help wrap those droplets in mesh-like walls. This study goes one step further: it shows that the droplets can evolve new functions on their own, driven purely by mutation — the engine that powers all of biology.

That is the “why it matters” part for the world beyond the lab. Because the droplets are fully programmable, the researchers see a path to materials that evolve toward a function, rather than being designed for it. Within Heidelberg University's SynthImmune Cluster of Excellence, the team wants to use mutation and selection to grow synthetic immune effectors — artificial cells that could one day develop specific immune functions on their own. Further down the road, the same platform could lead to smart drug-delivery systems that evolve to respond to specific cellular environments.

The honest counterpoint: these droplets are not alive. They cannot copy their DNA blueprints, pass genes to offspring, or metabolize energy — the hallmarks of even the simplest bacterium. What they satisfy, strictly speaking, are the prerequisites for Darwinian evolution: retaining genetic material, mutation-driven variation, and new selectable traits. Getting from there to open-ended evolution — where the system invents genuinely new functions generation after generation — is the next mountain. But the gap between a chemical mixture and something that behaves, moves, and stores cargo like a cell just got a lot thinner.