Encoding Physical Properties of the Cytosol

Biomolecular condensates are membrane-free compartments that spatially and temporally organize biochemistry in cells. Nucleic acids are frequence drivers and architectural elements in condensates however the molecular grammar that dictates the composition and physical states of condensates remains mysterious. To examine how RNA sequence and structure impact the form and function of condensates, we created an evolutionary algorithm that designs shuffled mRNA sequences which maximize or minimize predicted free energies of folding while preserving mass, known protein binding sites, encoded amino acid sequence, and nucleotide composition. RNAs with minimized free energies of folding contain many stable duplexes, while RNAs with maximized energies have unstable structures and long single-stranded regions. Using mass photometry, we show that RNAs with stable structures are mostly monomeric, while RNAs with unstable structures can multimerize. When mixed with protein, these differential RNA-RNA interactions affect dense phase compositions. Specifically, unstable RNAs form condensates with high RNA and low protein concentrations, while the converse is true for RNAs with stable structures. Using microbead rheology, we observe that unstable RNAs form condensates with long viscoelastic relaxation times that increase with condensate age, indicating elastic properties at many timescales. By contrast, stable RNAs form less elastic condensates with shorter relaxation times that increase to a lesser degree with time. We hypothesize that such aging is controlled by the exchange of intra- for inter-molecular base pairing among RNAs whose folding free energies determine the timescale of such exchanges. Thus, RNA structure can have multi-scale consequences for biomolecular condensates, with impacts on condensate composition and material properties.