Evolution is biology’s powerful method of engineering. It works by generating many variants of DNA, RNA, and proteins inside cells and letting nature “select” the organism that performs best. Early farmers started taking advantage of evolution by interfering with natural selection and letting only the most productive livestock and crops mate.
In laboratories, researchers developed methods for directed evolution of proteins, especially enzymes and antibodies, that are used in household detergents, medicine, and industry. The problem with existing laboratory methods is that they apply a constant selection pressure, which means that they tend to produce proteins that are always strongly active, and biology does not work like that: signaling proteins, protein “switches”, and protein “logic gates” (proteins that combine multiple inputs to make a yes-or-no decision) change states over time.
For example, a protein may need to turn on briefly, then shut off, then turn on again. If directed evolution methods only select for one state, the other important states of a protein can lose function or stop switching properly, which can be biologically detrimental (e.g. kill a cell). And yet, directed evolution approaches have had a difficult time producing dynamic and multi‑state protein behaviors.
Light in the directed evolution darkness
Researchers led by Sahand Jamal Rahi at EPFL’s Laboratory of the Physics of Biological Systems have developed a method they call “optovolution,” which uses light to guide the evolution of proteins with dynamic, multi‑state, and computational functions—making yes-or-no decisions based on specific rules.
The work, which is published in Cell , brings directed evolution closer to how cells actually operate, where timing and switching matter as much as strength.
The team built their system in the budding yeast Saccharomyces cerevisiae , widely used to brew beer and a laboratory workhorse. They rewired the yeast’s cell cycle so that progression depended on the protein to be evolved, switching cleanly between off and on states.
The key was linking the protein’s output signal to a cell‑cycle regulator that is essential at one stage but toxic at another. If the protein of interest stayed on or off for too long, the yeast cell stalled or died. Only cells in which the protein oscillated correctly could keep dividing.
The researchers used light for precise external control. Employing optogenetics—a technique that switches genes on and off with light—the researchers could control the protein of interest so that it flipped states with timed light pulses. Each roughly 90‑minute cell cycle acted as a rapid pass‑fail test of whether the protein switched at the right moment. In this way, this optovolution method favors variants with better dynamics, without manual screening or repeated interventions.
New variants and new colors
Using optovolution, the team evolved several classes of proteins. First, they improved a widely used light‑controlled transcription factor. They obtained 19 new variants that were either more sensitive to light, less active in the dark, or responsive to green rather than only blue light. Until now, unlocking a response to any warmer color than blue light was widely considered extremely difficult to engineer, based on how these proteins absorb light.
The team also evolved a red‑light optogenetic system so that, in yeast, it no longer had to be supplemented with a chemical cofactor. Evolution discovered a mutation that disabled a normal yeast transport protein, and that unexpectedly allowed the system to use light‑sensitive molecules that are already present inside the cell, making it much easier to use in experiments.
Finally, the study showed that optovolution is not limited to light‑sensing proteins. The researchers evolved a transcription factor that behaves as a single-protein computer, which activated genes only when two different inputs were present at the same time – one light signal and one chemical signal.
Dynamic protein functions are at the heart of sensing, decision‑making, and control in biology, from how cells respond to stress to how they commit to divide. By making these behaviors continuously evolvable inside living cells, optovolution opens new possibilities for synthetic biology, biotechnology, and basic research.
The study could help scientists build smarter cellular circuits, develop optogenetic systems that can be independently controlled with different colors of light, and explore how complex protein behaviors emerge through evolution.
Other contributors
Reference
Vojislav Gligorovski, Marco Labagnara, Lorenzo Scutteri, Marius Blackholm, Andreas Möglich, Nahal Mansouri, Sahand Jamal Rahi. Light-directed evolution of dynamic, multi-state, and computational protein functionalities. Cell 06 March 2026. DOI: 10.1016/j.cell.2026.02.002
Cell
Light-directed evolution of dynamic, multi-state, and computational protein functionalities.
6-Mar-2026