Back

RNA structure directs RNA partitioning and is actively disrupted inside stress granules to enable cellular recovery

Zhu, B.; Lim, J. J. A.; Zhang, Y.; Tan, W. T.; Zhang, J.; Xiang, J. S. N.; Huber, R.; Khong, A.; Wan, Y.

2025-06-05 molecular biology Community evaluation
10.1101/2025.06.05.658090 bioRxiv
Show abstract

RNA structures play important roles in liquid-liquid phase separation. However, how it is regulated during stress response and stress granule formation is still under studied. Here, we performed in vivo RNA structure probing before and after sodium arsenite treatment, and in stress granules. While RNAs generally become more double-stranded upon stress, they maintain their single-strandedness inside stress granules. We showed that RNA single-strandedness enables increased inclusion inside stress granules and that stress granule-enriched RNAs form fewer intra- and intermolecular RNA-RNA interactions. Additionally, several RNA binding proteins including SRSF1 are enriched in differential structure regions. eCLIP analysis revealed that SRSF1 binds to single-stranded regions along RNAs, and increased SRSF1 binding enabled better inclusion of RNAs in stress granules, whereas depletion of SRSF1 decreased stress granule formation under mild oxidative stress. We also observed the active unwinding of RNAs inside stress granules regulated by helicases, including DDX3X, and showed that inhibition of DDX3X results in slower dissolution of stress granules during recovery. Our study reveals the existence of multiple mechanisms to maintain RNA single-strandedness inside stress granules and to allow reversibility of stress granule formation, highlighting the importance of regulating RNA structure to enable cellular plasticity and stress response.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.