TL;DR
Researchers have developed a new coarse-grained model for disordered RNA, called CALVADOS-RNA, that is compatible with existing protein models. This model allows for more efficient and large-scale simulations of complex protein-RNA condensates, offering new insights into their formation and function in cellular processes.
What Problem Does This Paper Address?
Biomolecular condensates, formed by proteins and RNA, are crucial for various cellular activities. Understanding how these complex mixtures assemble and behave is essential for comprehending cellular organization and disease mechanisms. While coarse-grained models have been successful in simulating intrinsically disordered proteins (IDPs), a similar, compatible model for RNA has been lacking. This gap has hindered the ability to simulate and predict the behavior of mixed protein-RNA condensates, which are prevalent in cells. The authors aim to bridge this gap by developing a coarse-grained RNA model that can be seamlessly integrated with existing protein models, enabling the study of these complex systems.
Methods
The authors developed a two-bead-per-nucleotide coarse-grained (CG) model for disordered RNA, named CALVADOS-RNA. This model represents each nucleotide with two beads: one for the sugar-phosphate backbone and another for the nucleobase. The model was parameterized using a combined bottom-up and top-down approach.
- Bottom-up: Atomistic simulations and experimental data were used to determine local structural properties like bond lengths, angles, and stacking interactions.
- Top-down: Small-angle X-ray scattering (SAXS) data was employed to tune the "stickiness" parameters of the model, which govern attractive interactions, to reproduce experimental observations of RNA chain dimensions (radius of gyration) under varying ionic conditions.
The model incorporates bonded terms (bonds and angles), electrostatic interactions via the Debye-Hückel (DH) potential, and short-range attractive/repulsive interactions using the Ashbaugh-Hatch (AH) potential. For protein-RNA interactions, they utilized combining rules consistent with their existing CALVADOS protein model.
Datasets
The study utilized existing atomistic simulation data and experimental data for parameterization and validation:
- Atomistic simulation data: From previous studies on UCCAAUC sequences, polyA30, and polyU30.
- Experimental data:
- Small-angle X-ray scattering (SAXS) measurements for polyA30 and polyU30 to determine chain dimensions.
- Experimental measurements of the second virial coefficient for RNA-RNA interactions.
- Experimental data on the phase behavior of protein-RNA mixtures (FUSRGG3 and polyU40, MED1 and RNA) to validate the model's predictive capabilities.
No new datasets were explicitly introduced or benchmarked in this paper.
Key Insights and Ideas
- Bridging the gap: The development of a sequence-independent, two-bead CG model for RNA that is compatible with existing protein CG models is a significant step towards simulating complex biomolecular condensates.
- Hydropathy scale approach: Using a hydropathy scale for RNA, similar to that used for proteins, provides a unified framework for modeling mixed condensates.
- Balance of interactions: The model captures the interplay between electrostatic (backbone) and hydrophobic (base) interactions, which are crucial for RNA behavior and its interactions with proteins.
- Reentrant phase behavior: The model successfully reproduces reentrant phase behavior in protein-RNA mixtures, where phase separation occurs at intermediate RNA concentrations but not at very low or high concentrations. This highlights the model's ability to capture nuanced interaction dynamics.
- Selective partitioning: The model demonstrates the ability to predict how different proteins (transcription factors) partition into or are excluded from protein-RNA condensates, showing that this partitioning is dependent on the condensate's composition.
Key Results
- The CALVADOS-RNA model accurately reproduces local RNA geometry and intramolecular interactions based on atomistic simulations and experimental data.
- The model semi-quantitatively captures RNA-RNA interactions, as validated by comparing calculated and experimental virial coefficients.
- It accurately predicts protein-RNA interactions, including reentrant phase behavior and selective partitioning of transcription factors, by simulating mixtures of MED1 and FUSRGG3 with RNA.
- Simulations of MED1-RNA condensates showed that RNA concentration not only influences condensation but also alters the selectivity for IDRs of transcriptional regulators.
- The model successfully reproduced the de-mixing behavior of a prion-like protein (PLP) from RNA, consistent with experimental observations.
Limitations and Future Work
The authors acknowledge several limitations:
- Sequence independence: The current model is sequence-independent, meaning it doesn't capture specific base-base interactions that can drive RNA phase separation. Future work could involve incorporating sequence-dependent parameters.
- Structured RNA: The model is designed for disordered RNA and does not capture specific base-pairing or tertiary structures. Extending the model to include structured RNAs would require further development, possibly with harmonic constraints.
- Specific interactions: The model does not capture specific RNA-protein interactions, which are important for understanding the role of domain structures in phase separation.
- Cutoff distance effects: The study noted that the cutoff distance for the Debye-Hückel potential affects intermolecular interactions at low ionic strengths, suggesting that further investigation into optimizing cutoff distances based on ionic strength is needed.
- Sampling limitations: Simulating very low saturation concentrations (< 1 µM) was not feasible with standard molecular dynamics, suggesting a need for enhanced sampling methods for systems with strong phase separation.
Despite these limitations, the authors envision the CALVADOS-RNA model as a valuable tool for studying protein-RNA condensates and their biochemical functions, paving the way for more comprehensive simulations of cellular processes.
Summary updated