All cells need to sense and respond to their environment, to know when to activate genes, build proteins, and carry out their basic functions. One of the most well-studied cellular responses is how they react during times of stress, such as when the temperature gets too high or there aren't enough nutrients around to sustain activity. When this happens, cells gather strands of RNA and proteins into stress granules, dense clumps of material generally known as biomolecular condensates.
For decades, scientists have puzzled over why cells form these condensates, but new research from the University of Chicago shows that cells use condensates to save RNA messages they're working on when stress occurs so that they can prioritize new stress-relevant messages.
"This is how cells change their mind," said D. Allan Drummond, PhD , Associate Professor of Biochemistry and Molecular Biology at UChicago and a senior author of the study. "These little condensates are present all the time, and not just during stress. They're omnipresent in cells, and this seems to be a way that all eukaryotic cells couple transcriptional responses and translational responses." The work was published on November 24, 2025, in Molecular Cell .
'Cells tend to know exactly what they're doing'
Despite the years of study, the exact function of stress granules isn't well understood. They seem to be a way for cells to save their work in progress and hold it for safekeeping until conditions become favorable again. One popular hypothesis is that when stress hits, genes switch off. They stop transcribing new strands of messenger RNA (mRNA), which quit translating into proteins. Any of these leftover mRNAs and associated proteins pile up into stress granules as a side-effect of the genetic translation shutdown.
Drummond says he has always been skeptical that this was just an accident. "I just have this impression from long experience that cells tend to know exactly what they're doing, Having these stress granules just as a biophysical side effect never scanned for me," he said.
Stress granules are large enough that you can see them through a microscope. Until now, many scientists have taken this as evidence of when and how they form. Stress out a cell, and the granules appear; relieve the stress, and they go away. But Drummond and his team wanted to see if there was more to it than what they could see, in both stressed and non-stressed cells. So, they opened cells from budding yeast, a classic model organism for studying cellular biology, and spun the contents in a centrifuge. This separated the heavier materials, like condensates, from smaller molecules like free-floating mRNA and proteins.
In the past when scientists studied the contents of stress granules, they found that mostly long strands of mRNA were clumped together, because they were "stickier" or became more easily tangled with other molecules. But when Drummond's team analyzed their cells using their new methods, they saw that almost all mRNAs were pulled into the condensates, regardless of length—with one exception. Newly synthesized mRNAs that were made after the stressful conditions began were excluded.
"This means that instead of picking and choosing, what the cell is doing when stress happens is it takes all pre-existing messenger RNAs and shoves them into condensates," Drummond said. It didn't matter what genetic sequence was encoded on the mRNAs, or how long they were—it was all about timing.
In a separate paper also published in Molecular Cell , a group of Drummond's colleagues from ETH Zurich in Switzerland also saw that mRNAs produced before stress caused by glucose withdrawal were also lumped into condensates, while those generated later escaped.
Drummond believes this happens because the ends of each RNA strand, where the process of translating proteins normally takes place, are freed up when translation stops. Now these mRNAs can be pulled into condensates by their ends. The researchers aren't sure how the new, post-stress mRNAs escape this fate, but it could be because their ends are specially protected when these mRNAs are newly made, allowing them to avoid condensation and stay busy translating proteins.
Tiny condensates for normal biological processes
Drummond's team conducted further biochemical tests and found that in addition to the larger, conventional stress granules, the cells formed many smaller condensates. These "translation initiation inhibited condensates," or TIICs ("ticks"), form whenever there is a pause in translation. This happens all the time in cells for many reasons and normal biological processes, not just stress.
The researchers did a series of tests to block the process of protein translation and confirmed that these TIICs do indeed form, whether the cell was stressed or not. They even tested a drug that prevents stress granule formation, and the TIICs still formed.
"Whether these condensates form doesn't depend on stress in particular," Drummond said. "In fact, in all cells, all the time, anytime you get a transient pause in translation, things get pulled into one of these little nanocondensates." They could be storing mRNAs and proteins for later use, or if conditions change and they're no longer needed, they degrade.
For Drummond, this is an answer to a question that his lab has been working on for more than 10 years. "This is the best thing my lab has ever done," he said. "We can see all the old phenomena in a new way: before, the cells were thought to be suffering from side-effects; now, we see that the cell knows exactly what it's doing. It's using these condensates in an adaptive and programmed way, not just as a stress response but to accomplish its ends by rearranging cellular matter in time and space."
The study, " Transcriptome-wide mRNP condensation precedes stress granule formation and excludes new mRNAs ," was supported by the National Institutes of Health, the European Union's Horizon 2020 Marie Skłodowska-Curie Grant, the Wellcome Trust, and the Helen Hay Whitney Foundation.
Additional authors include Hendrik Glauninger, Caitlin J. Wong Hickernell, Karen M. Velez, Sneha Paul, Jingyi Fei, and Tobin R. Sosnick from UChicago; Jared A.M. Bard from Texas A&M University; Edo M. Airoldi from Temple University, Weihan Li and Robert H. Singer from Albert Einstein College of Medicine; and Edward W. J. Wallace from the University of Edinburgh. Glauninger, Bard, and Wong Hickernell are co-first authors.
