Glutamate helps unmask the differences in driving forces for phase separation versus clustering of FET family proteins in sub-saturated solutions

Background

The internal cellular environment is organized through membrane-bound and membrane-less organelles (MLOs). MLOs are groupings of macromolecules that are not separated from the rest of the cell by any physical barrier [1]. The underlying interactions between these heterotypic macromolecules are the driving force for the formation and selectivity of MLOs [1]. Investigating how these interactions occur and are affected by the composition of the local cellular milieu helps us understand how the cell organizes itself functionally.

The self-associative properties of intrinsically disordered proteins (IDPs) are often central to the formation of MLOs [2]. This self-association is even seen in sub-saturated solutions, where visible de-mixing of the solution does not occur, and the interactions only result in transiently associating complexes of varying size [3]. In vitro, we often observe these interactions in super-saturated conditions, wherein molecules de-mix from the solution, akin to an oil-in-water emulsion, with droplets forming and fusing with one another. This phenomenon is known as “macro-phase separation,” but IDPs often only undergo this process at protein concentrations greater than what is seen endogenously. IDP heterogenous clustering is a key driver of MLO formation and macro-phase separation, and the molecular mediators of this clustering are often similar or identical to those that occur within MLOs and drive macro-phase separation [3].

In this preprint, the authors sought to understand the cellular milieu’s contribution to IDP self-association and to distinguish if these environments promote macro-phase separation or clustering interactions. To investigate these questions, a prototypical self-associative IDP, namely FUS, was assayed in solutions containing potassium and glutamate ions (KGlu), which are the most abundant and relevant intracellular ions [4]. The effect of glutamate was of particular interest, because it is a bulky molecule and carries a delocalized negative charge across two carboxylate oxygens, whereas a small ion, such as chloride, contains a strong centralized negative charge. This contributes to glutamate’s relatively weaker shielding effect on cations, causing a less significant reduction on intermolecular interactions, compared to chloride [5]. This work demonstrates the influential role of the cellular milieu on cellular organization and highlights glutamate as a modulator of intermolecular clustering interactions.

Key Findings

Macro-phase separation is minimally affected by glutamate compared to chloride

Using pelleting assays, the concentration at which FUS begins to form phase-separated condensates separate from the bulk solution (saturation concentration or C_sat) was determined in KCl and KGlu buffers. These results, coupled with microscopy, showed minor differences in macro-phase separation between KCl and KGlu, with similar C_sat values and condensates appearing at nearly the same concentration between the two buffers.

Clustering of FET proteins is increased in KGlu compared to KCl

With macro-phase separation relatively unaffected by the presence of glutamate, the authors sought to investigate the clustering of FUS molecules. For these questions, they employed a number of biophysical approaches. Nanoparticle tracking analysis (NTA) is a microscopy technique that records the Brownian motion of nanoparticle trackers and relates that movement to the size of the particle that the tracker is moving through [6]. In 100 mM KGlu, the authors measured a four-fold increase in the concentration of mesoscale clusters of FUS, compared to 100 mM KCl.

In addition to the NTA experiments, several microscopy techniques were used to confirm the formation of these transient complexes within the sub-saturated solution. These techniques all detected clustering of FUS molecules into complexes at FUS concentrations as low as 400 pM, far lower than endogenous FUS levels [7]. Finally, dynamic light scattering (DLS) was used to confirm the formation of higher-order complexes in proteins from the same family as FUS, namely, EWS and TAF15. These experiments suggest that clustering may be a relevant mode of self-association even when macro-phase separation is not readily occurring.

Tyrosine-arginine interactions between FUS prion-like domain and FUS RNA-binding domain are important mediators of clustering and phase separation

Having established the increased clustering propensity of FUS in KGlu solvents, the authors wanted to understand what interactions mediate this mode of self-association, and if they are similar interactions that drive macro-phase separation. FUS is known to undergo macro-phase separation through interactions between tyrosine residues within the N-terminal prion-like domain (PrLD) and arginine residues in the RNA-binding domain (RBD) [8].

As such, in both KCl and KGlu buffers, tyrosine-to-serine mutations in the PrLD significantly increased the C_­sat, demonstrating the lower propensity to undergo macro-phase separation. These mutations also made the clustering of FUS molecules undetectable in both solvents, suggesting that these tyrosine-serine interactions also mediate the self-association in sub-saturated solutions. This reduction in macro-phase separation and clustering was also seen when arginine residues in the RBD were mutated to glycine.

When the RBD arginine residues were mutated to lysine, macro-phase separation and clustering were both reduced in KCl buffers. However, interestingly, only macro-phase separation was reduced in KGlu buffers, while clustering was similar to WT. In sub-saturated solutions, clustering is mediated by cation-π interactions, with lysine still able to act as an effective cation to interact with the π system of the tyrosine aromatic rings. However, macro-phase separation is greatly influenced by solubility, which is increased when arginine residues are substituted for lysine. Arginine has a less favorable free energy of hydration compared to lysine, making it behave more hydrophobically and decreasing the solubility of the protein it is incorporated into [9].

Differences between glutamate and chloride in driving protein clustering may be due to the inability of glutamate to interact with backbone amides

Next, the authors turned to molecular dynamics simulations to understand how glutamate’s interaction with FUS differs from that of chloride. These experiments revealed lower interaction coefficients for chloride’s interaction with all amino acid residues, except arginine and lysine, compared to glutamate. These results suggest that the bulkier glutamate anion is able to interact with the side chain of positively charged residues, but is excluded from the peptide backbone, due to its larger size. Chloride, on the other hand, is small enough to associate with the backbone amide, providing a greater charge shielding effect compared to glutamate. This key difference in the interaction between glutamate and chloride with the peptide is likely a vital contributor to the increased clustering seen in KGlu buffers.

Why I chose this study

Understanding the atomic details underlying the interactions driving macromolecular self-association is fundamental to our understanding of how the cell is organized. This preprint drives forward the field of biomolecular and cellular organization by emphasizing the necessity of studying these phenomena in the context of the cellular milieu.

Additionally, the profound effect of solvent properties on the clustering of macromolecules is highlighted by the difference between cellularly abundant glutamate and chloride, which is commonly used for in vitro studies. This preprint identifies key interactions that occur, not just between macromolecules, but between the macromolecules and their cellular environment. This is often overlooked or not given its full due in the context of these biochemical systems.

Questions for the authors

1. Based on the increased self-associative clustering seen with glutamate, would you expect to see enhanced or higher affinity interactions between intrinsically disordered proteins, like FUS, and their protein binding partners? This could have implications for the “scaffolding” ability of these highly disordered proteins.
2. Other metabolites, which could be in relatively high concentration in particular regions of the cell, could have an outsized effect on the associative properties of molecules in that region. Alternatively, could the formation of these enzymatically-formed metabolites be having a compounding effect on the clustering, coupled to distinct cellular processes?

References

1. Wang, B. et al. Liquid-liquid phase separation in human health and diseases. Signal Transduct Target Ther 6, (2021).
2. Feric, M. & Misteli, T. Function moves biomolecular condensates in phase space. BioEssays 44, (2022).
3. Kar, M. et al. Phase-separating RNA-binding proteins form heterogeneous distributions of clusters in subsaturated solutions. Proc Natl Acad Sci U S A 119, (2022).
4. Milo, R., Phillips, R. & Orme, N. Cell Biology by the numbers. (Garland Science, 2015).
5. Kozlov, A. G. et al. How Glutamate Promotes Liquid-liquid Phase Separation and DNA Binding Cooperativity of E. coli SSB Protein. J Mol Biol 434, 167562 (2022).
6. Filipe, V., Hawe, A. & Jiskoot, W. Critical evaluation of Nanoparticle Tracking Analysis (NTA) by NanoSight for the measurement of nanoparticles and protein aggregates. Pharm Res 27, 796–810 (2010).
7. Patel, A. et al. A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease Mutation. Cell 162, 1066–1077 (2015).
8. Murthy, A. C. et al. Molecular interactions contributing to FUS SYGQ LC-RGG phase separation and co-partitioning with RNA polymerase II heptads. Nat Struct Mol Biol 28, 923–935 (2021).
9. Fossat, M. J., Zeng, X. & Pappu, R. V. Uncovering Differences in Hydration Free Energies and Structures for Model Compound Mimics of Charged Side Chains of Amino Acids. Journal of Physical Chemistry B 125, 4148–4161 (2021).

 

Tags: condensate, fet, fus, llps, membraneless organelle, microscopy, phase separation

Posted on: 3 October 2023 , updated on: 5 October 2023

doi: https://doi.org/10.1242/prelights.35670

Read preprint (No Ratings Yet)