Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2025 Sep 22;34(10):e70316. doi: 10.1002/pro.70316

Molecular basis of EWS interdomain self‐association and its role in condensate formation

Erich J Sohn 1,2, Kandarp A Sojitra 3, Leticia Rodrigues 1,2, Xiaoping Xu 1,2, Bess Frost 4, Jeetain Mittal 3,5,6,✉, David S Libich 1,2,✉
PMCID: PMC12451834  PMID: 40981365

Abstract

Ewing sarcoma, the second most common pediatric bone and soft tissue cancer, is caused by aberrant fusion of the RNA‐binding protein EWS (EWS) low‐complexity domain (EWSLCD) to the DNA‐binding domain of the transcription factor friend leukemia integration 1 (FLI1). The resulting fusion, EWS::FLI1, directly interacts with and engages in a dynamic interplay with EWS that drives tumorigenesis and regulates the function of both proteins. While EWSLCD is known to promote self‐association, the role of the RNA‐binding domains (RBDs) of EWS, which include arginine–glycine–glycine (RGG) repeat regions and a structured RNA‐recognition motif (RRM), remains less well understood. Here, we investigate the interplay between EWSLCD and RBDs using biomolecular condensation assays, microscopy, nuclear magnetic resonance (NMR) spectroscopy, and molecular simulations. Our studies reveal that RBDs differentially influence EWSLCD condensate formation and suggest that electrostatics and polypeptide‐chain length likely contribute to this interaction. NMR spectroscopy and molecular dynamics simulations further demonstrate that EWSLCD and the central RNA‐binding region, comprising the RRM and RGG2 domains, engage in transient, non‐specific interactions that are broadly distributed across both regions and involve diverse residue types. Specifically, tyrosine, polar residues, and proline within EWSLCD preferentially interact with arginine, glycine, and proline residues in the RBD. Atomistic simulations of EWS confirm that the full‐length protein exhibits a similar interaction profile with conserved chemical specificity, supporting a model in which a network of weak, distributed interdomain contacts underlies EWS self‐association. Together, these findings provide molecular insight into the mechanisms of EWS condensate formation and lay the groundwork for understanding how interdomain interactions regulate EWS and EWS::FLI1 function.

Keywords: biomolecular condensate, EWS, intrinsically disordered protein, molecular simulation, NMR

1. INTRODUCTION

Ewing sarcoma is the prototypical example of a family of related cancers caused by chromosomal translocations involving the FET family of RNA‐binding proteins, comprised of Fused in Sarcoma (FUS), RNA binding protein EWS (EWS), and TATA‐binding protein‐associated factor 2N (TAF15) (Delattre et al., 1992; Flucke et al., 2021; Turc‐Carel et al., 1983). The most common of these translocations fuses 264 residues from the low complexity domain (LCD) of EWS (EWSLCD) to the DNA‐binding domain of friend leukemia integration 1 (FLI1) (Fisher, 2014; Flucke et al., 2021; Maki et al., 2022). The resulting oncofusion, EWS::FLI1, is found in ~85% of Ewing sarcoma tumors, and its ability to form biomolecular condensates is required for oncogenic transformation (Bertolotti et al., 1998; Jaishankar et al., 1999; Johnson et al., 2017; Lessnick et al., 1995; Ng et al., 2007). Biomolecular condensation of EWS and EWS::FLI1 occurs via favorable interactions within and between EWSLCD domains (Ahmed et al., 2021; Johnson et al., 2024; Ng et al., 2007; Nosella et al., 2021; Spahn et al., 2003). These interactions promote EWS::FLI1 binding to gene enhancers (Chong et al., 2018; Johnson et al., 2017; Vasileva et al., 2024; Zuo et al., 2021) and recruit transcriptional co‐factors (Boulay et al., 2017; Kim et al., 2024; Kwon et al., 2013; Petermann et al., 1998), promoting subsequent oncogenic transactivation.

Wild type EWS and EWS::FLI1 are co‐expressed in affected individuals as only one of two EWS loci carries the aberrant fusion. Expression of the EWS::FLI1 fusion protein exerts a dominant negative repression over wild type EWS function, resulting in a phenotype that mimics EWS knockout (Boone et al., 2021; Embree et al., 2009; Gorthi et al., 2018; Jaishankar et al., 1999; Yang et al., 2000). Loss of endogenous EWS, as well as EWS::FLI1 expression, leads to poly(ADP‐ribose) polymerase 1 (PARP1) accumulation at sites of DNA damage and subsequent sensitivity to DNA damaging agents (Gorthi et al., 2018; Lee et al., 2020; Li et al., 2007). Furthermore, EWS has an antagonistic effect on the activity of EWS‐oncofusions, with homotypic EWSLCD‐EWSLCD interactions repressing transactivation (Chong et al., 2022), as well as heterotypic interactions between the RNA‐binding domains (RBDs) of EWS and the EWSLCD (Alex & Lee, 2005; Li & Lee, 2000). Understanding the molecular mechanisms driving EWS interdomain interactions is crucial for elucidating the interplay between EWS and EWS::FLI1.

In this work, we investigate interdomain interactions between the EWSLCD and RBDs and their ability to drive EWS condensate formation using turbidity and condensate partitioning assays, microscopy, nuclear magnetic resonance (NMR) titrations, and molecular simulations. We find a strikingly heterogeneous effect of the three arginine–glycine–glycine (RGG) repeat domains on EWSLCD condensate formation that mirrors their sequential and compositional heterogeneity. Extensive physiochemical characterization of condensate‐forming properties between EWSLCD and the central RNA‐binding region, which includes RGG2, reveals electrostatics (charge interactions) and EWSLCD polypeptide chain length as key drivers of the interdomain association. Using a fragment‐based NMR approach, we probed the atomistic details of the transient association between these two regions in the dilute phase. These experiments, along with atomistic simulations, uncover numerous non‐specific contacts across both polypeptide chains that drive the interdomain self‐association. In silico studies confirm that our fragment‐based approach effectively recapitulates the interactions that occur in full‐length EWS. Our analysis establishes the structural underpinnings of full‐length EWS condensate formation and offers a basis for exploring the molecular relationship between EWS::FLI1 and wild‐type EWS that drives Ewing sarcoma oncogenesis.

2. RESULTS

2.1. Interdomain contacts occur throughout full‐length EWS

The EWS protein largely lacks secondary structure, with the entire SYGQP‐rich LCD (residues 1–280) and 67% of the RBD being intrinsically disordered (Figure S1a). The only folded domains are the RNA‐recognition motif (RRM) and zinc finger (ZnF), both of which contribute to nucleic acid binding (Figure S1a) (Lay et al., 2024; Selig et al., 2023). Within the central RNA‐binding region, EWSRRM‐RGG2, R 1, R 2, and heteronuclear nuclear Overhauser effect measurements reporting on ps‐ns time‐scale dynamics are consistent with the RRM adopting a folded structure, while the RGG2 remains dynamic and disordered under these conditions (Figure S1b). Intrinsic disorder within the C‐terminal RBD is due to the three RGG domains, of which RGG3 is the longest (90 residues), and RGG2 (66 residues) and RGG1 (59 residues) are somewhat similar in length (Figures 1a and S1a). RGG3 has a canonical RGG repeat structure, containing 12 RGG repeats, comprising 40% of the domain, with arginine and glycine accounting for 63% of all residues (Figure 1a). RGG1 and RGG2 are more divergent, containing only six and four RGG repeats, comprising 31% and 18% of each domain, respectively. Notably, RGG1 carries a net negative charge due to an abundance of acidic residues, particularly aspartate, while RGG2 is enriched (23%) in proline residues (Figure 1a). Sequence charge decoration, an order parameter that predicts charge patterning in intrinsically disordered regions (Sawle & Ghosh, 2015), predicts RGG1 has the most charge segregation (−3.046), with negative charges clustered at the C‐terminus.

FIGURE 1.

FIGURE 1

Open in a new tab

EWSLCD forms interdomain contacts with RNA‐binding domain (RBD) motifs. (a) RNA‐binding protein EWS (EWS) domain architecture, net charge of each domain, and amino acid distribution. The EWS::FLI1 breakpoint (264) is marked with a black triangle; red and blue lines are negatively and positively charged residues. (b) Average pairwise interchain van der Waals contact map shown as −ln(n contact,ij/n max), where n contact,ij is the average pairwise residue index interchain van der Waals contact and n max is the maximum values among all n contact,ij. (c) Summary of interdomain contact strength between EWS domain pairs represented as −ln(Σn contact,ij/(Σn contact,ij)max). (d) Fraction of total EWS interdomain contacts from co‐existence coarse grain (CG) simulations of EWS. Values represent the fraction of total interdomain contacts formed between specific domain combinations, calculated by summing all n contact,ij within each domain pair and normalizing by the total interdomain contacts across all domain combinations. FLI1, friend leukemia integration 1; LCD, low‐complexity domain; RGG, arginine–glycine–glycine; RRM, RNA‐recognition motif; ZnF, zinc finger, nuclear localization sequence.

To investigate how the sequence features of full‐length EWS govern the interdomain interactions within the condensed phase, we performed coarse‐grained co‐existence simulations, which revealed frequent contacts between the LCD and RBDs (Figure 1b–d). While the most frequent contacts occur between the LCD and RGG domains, we detect contacts with the folded RRM and ZnF as well as between the negatively charged region in RGG1 and other RGGs (Figure 1b–d). These interdomain interactions appear to enhance the condensate‐forming propensity of the protein, as full‐length EWS has a simulated saturation concentration (C sat) that is ~10 fold lower than EWSLCD (Figure S2a). Based on net charge, one might expect full‐length EWS to exhibit stronger electrostatic repulsion compared to EWSLCD (residues 1–264), potentially reducing its phase separation (Figure 1a). However, the opposite behavior is observed, suggesting that the multivalency inherent to the longer polypeptide chain of EWS as well as charge patterning may contribute to promoting phase separation (Figure S2a). Overall, these results clearly demonstrate the propensity for the LCD to form highly promiscuous intramolecular interactions throughout the entire polypeptide chain and especially with the three RGG domains.

2.2. RBDs of EWS impact EWSLCD condensate formation

To investigate the role of interdomain contacts in EWS condensate formation, we created two‐component systems combining EWSLCD with EWSRGG1‐RRM, EWSRRM‐RGG2, and EWSRGG3 (Figure S2b,c). In a single‐component system, both EWSLCD and EWS::FLI1 form condensates through diverse interaction modes, stimulated by dissolved ions (salting out), consistent with the effect of other LCDs (Bock et al., 2021; Johnson et al., 2024; Krainer et al., 2021; Martin et al., 2020; Murthy et al., 2019).

In the absence of NaCl, EWSLCD and EWSLCD + EWSRGG1‐RRM solutions have virtually no turbidity and lack visible condensates, consistent with previous reports for the EWSLCD (Nosella et al., 2021) under these conditions (Figure 2a,b). In the presence of 150 mM NaCl, the addition of EWSRGG1‐RRM diminished solution turbidity compared to the EWSLCD control, indicating reduced condensate formation, despite maintaining similar morphology as EWSLCD‐only condensates (Figure 2a,b). EWSRRM‐RGG2 and EWSRGG3 both substantially increased EWSLCD condensate formation in the absence of NaCl, although EWSRGG3 had a greater effect (Figure 2a). Upon the addition of 150 mM NaCl, EWSLCD + EWSRRM‐RGG2 and EWSLCD + EWSRGG3 condensate formation was diminished, with the EWSLCD + EWSRGG3 samples having a smaller change (Figure 2a). In the presence of NaCl, EWSLCD + EWSRRM‐RGG2 condensate morphology resembles that of EWSLCD‐only samples; however, in the absence of NaCl, they appear less round (Figure 2b). Condensates formed by EWSLCD + EWSRGG3 were a mixture of small and spherical, which remained suspended, and irregularly shaped, which appeared to settle quickly, spreading out heterogeneously across the slide, regardless of whether NaCl was present (Figure 2b).

FIGURE 2.

FIGURE 2

Open in a new tab

RNA‐binding protein EWS (EWS) RNA‐binding domain (RBD) motifs impact condensate formation propensity of EWSLCD. (a) Turbidity measurements, (b) bright field microscopy, scale bar = 20 μm, and (c) condensate partitioning of 50 μM EWSLCD only (black) and + 50 μM of EWSRGG1‐RRM (orange), EWSRRM‐RGG2 (green), or EWSRGG3 (purple). Lighter shade is with 150 mM NaCl. (d) Phase diagram for EWSLCD + EWSRRM‐RGG2 (filled circles = visible condensates). Turbidity measurements of EWSLCD + EWSRRM‐RGG2 across different (e) pH, scale bar = 10 μm, and (f) NaCl concentrations. (g) Fluorescence microscopy of select [NaCl] from (f), scale bar = 20 μm. P‐values: *P < 0.05, **P < 0.005. Error bars = standard deviation. LCD, low‐complexity domain; RGG, arginine–glycine–glycine; RRM, RNA‐recognition motif.

Partitioning assays were used to gain insight into the distinct effects of each RBD on EWSLCD condensate formation. In EWSLCD + EWSRGG1‐RRM samples, both domains remained completely soluble in the absence of NaCl, consistent with turbidity measurements (Figure 2c). Addition of NaCl had no effect on the solubility of EWSRGG1‐RRM, and only ~13% w/w of EWSLCD was driven into condensates (Figure 2c). Conversely, ~53% w/w EWSRRM‐RGG2 and ~51% w/w EWSRGG3 co‐partitioned into condensates with approximately equal amounts of EWSLCD (~60% and ~49% w/w, respectively) in the absence of NaCl (Figure 2c). Addition of 150 mM NaCl significantly reduced EWSRRM‐RGG2 (~6% w/w) and EWSLCD (~21% w/w) condensate co‐partitioning, while insignificantly changing the amounts of EWSRGG3 (~56% w/w) and EWSLCD (~53% w/w) that co‐partition. The unique impact of each RBD construct on EWSLCD condensates suggests a complex interplay of interdomain interactions that drive EWS self‐association that is regulated through distinct biochemical properties.

To further investigate the intriguing response of the EWSLCD + EWSRRM‐RGG2 system to NaCl, we created a phase diagram varying concentrations of EWSLCD and EWSRRM‐RGG2 (Figures 2d and S3a). In a single‐component system absent of NaCl, the C sat of EWSLCD is >150 μM (Johnson et al., 2024). Addition of equal molar ratios of EWSRRM‐RGG2 reduced the EWSLCD C sat to <25 μM. When EWSRRM‐RGG2 is present in molar excess, condensates form in solution at <10 μM EWSLCD (Figures 2d and S3a). Similarly, 50 μM EWSLCD was induced to form condensates by <5 μM EWSRRM‐RGG2 (Figures 2d and S3a). To investigate the role of multivalency, we reduced the EWSLCD length by fragmenting it into three overlapping regions: EWS1–120, EWS91–199, and EWS171–264 (Figure S2b,c) (Johnson et al., 2022). Surprisingly, none of the three fragments formed condensates with EWSRRM‐RGG2 when mixed at equimolar ratios up to 150 μM (Figure S3b), suggesting that multivalency and length are critical factors driving the association.

EWSLCD + EWSRRM‐RGG2 condensate formation was more complex than the initial turbidity and condensate partitioning assays revealed. While equimolar ratios (25 μM) of EWSLCD and EWSRRM‐RGG2 readily form condensates from pH 5 to 9, formation is inhibited at pH 10 and 11 (Figure 2e). In the absence of NaCl, robust turbidity is measured for the bi‐component EWSLCD + EWSRRM‐RGG2 system. Upon addition of up to 50 mM NaCl, turbidity is reduced (Figures 2f and S3c). At moderate NaCl concentrations (50–200 mM), condensate formation is almost completely inhibited, yet as NaCl increases further (>200 mM), turbidity again increases, indicating the reappearance of condensates (Figures 2f and S3c). Fluorescent microscopy images at selected NaCl concentrations reveal that EWSLCD and EWSRRM‐RGG2 colocalize into condensates in the absence of NaCl and that low levels of NaCl in solution greatly reduce the number and size of the condensates (Figure 2g). However, at higher NaCl concentrations, the condensates that reappear almost entirely exclude EWSRRM‐RGG2 and form smaller fusion‐defective condensates (Figure 2g), consistent with previous findings of single‐component EWSLCD condensates at high NaCl concentrations (Johnson et al., 2024). These two distinct modes of condensate formation reveal the complexity of EWS self‐association and the ability of EWS to drive molecular organization across a wide range of physiochemical conditions. The ability of soluble ions to prevent EWSLCD and EWSRRM‐RGG2 from co‐localizing into condensates suggests a potential role of electrostatic interactions in driving co‐condensation of these two domains.

2.3. Atomistic details of association between 15N‐EWSLCD and EWSRRM‐RGG2

We used NMR and molecular simulations to investigate the association and dynamics of the dilute phase interaction between EWSLCD and EWSRRM‐RGG2 as a surrogate for full‐length EWS. To abrogate pH changes and dilution effects we used a “reverse” titration protocol of the 15N‐EWSLCD fragments with EWSRRM‐RGG2 (Dcosta et al., 2024), and observed numerous small chemical shift perturbations (CSPs) dispersed across the sequence of each fragment (Figures 3 and S4a). Minor line broadening was also observed for residues across the three fragments, yet no obvious clustering indicative of a single binding site is evident (Figure S4b). EWS1–120 had the greatest proportion of both shifted and broadened peaks, suggesting that this region contains a higher density of interaction sites with EWSRRM‐RGG2 (Figures 3a and S4b). Both CSPs and line broadening were small in magnitude and non‐sequential, suggesting that EWSLCD non‐specifically interacts with EWSRRM‐RGG2 with minor contributions from residues throughout the polypeptide chain (Figure 3 and S4b). Consistent with these NMR observations, dilute phase atomistic simulations of EWSLCD fragments with EWSRRM‐RGG2 (Figure S5a–c) also revealed non‐specific contacts distributed across the lengths of the EWSLCD fragments (Figure 3). Among the three EWSLCD fragments, EWS1–120 appeared to form more contacts with EWSRRM‐RGG2 compared to the other fragments (Figure 3a). Notably, while residues within the 40–60 region of EWS1–120 exhibited higher contact frequencies, the interactions remained broadly distributed throughout the LCD, highlighting the overall non‐specific nature of these interactions (Figure 3a).

FIGURE 3.

FIGURE 3

Open in a new tab

EWSLCD interacts with EWSRRM‐RGG2 through contacts distributed across their chains. Chemical shift perturbations (CSPs) (top left panels), average per‐residue contacts (bottom left panels), and normalized average pairwise van der Waals contacts (right panels) for (a) 15N‐EWS1–120 (cyan), (b) 15N‐EWS91–199 (pink), and (c) 15N‐EWS171–264 (green) interacting with EWSRRM‐RGG2. CSPs are the maximum observed after addition of 3:1 molar ratio of EWSRRM‐RGG2 and color‐filled if CSPs >1 standard deviation (dashed black line). Asterisks indicate overlapped/ambiguous assignments (red) or prolines (black). Average per‐residue and pairwise van der Waals contacts were calculated from atomistic simulations of EWSRRM‐RGG2 with each EWSLCD fragment and are colored filled if contact values are > the mean for all three fragments (dashed black line). The average per‐residue contacts formed between EWSLCD fragments and EWSRRM‐RGG2 are shown along the left and bottom margins of each contact map. EWS, RNA‐binding protein EWS; RGG, arginine–glycine–glycine; RRM, RNA‐recognition motif.

2.4. Atomistic details of association between 15N‐EWSRRM‐RGG2 and EWSLCD

We next performed the inverse titration of 15N‐EWSRRM‐RGG2 with EWSLCD fragments. Since only 39% of non‐proline residues of RGG2 are assignable due to sequence degeneracy and spectral overlap, we stratified the unassignable peaks as glycine or arginine/other based on their 13C chemical shifts (Selig et al., 2023) (Figure 4). The unassigned peaks nearly all come from RGG2, as 98% of all non‐proline residues in the RRM are assigned, and thus this approach enabled tracking of RG (arginine‐glycine), RGG, and proline‐glycine‐glycine motifs even without sequence‐specific assignments. Titration of 15N‐EWSRRM‐RGG2 with each of the three EWSLCD fragments resulted in small CSPs and minimal line broadening for residues distributed in both the RRM and RGG2 regions (Figures 4 and S6). As with the 15N‐EWSLCD titrations, CSPs were small in magnitude and dispersed across the EWSRRM‐RGG2 polypeptide chain (Figure 4), supporting the conclusion that the interdomain association is largely non‐specific with minor contributions from numerous residue pairs. The increased chain flexibility provided by the sequence disorder found in RGG2 may permit the increased contact propensity with the intrinsically disordered EWSLCD fragments (Figure S1). Similarly, dilute phase atomistic simulations of EWSRRM‐RGG2 with EWSLCD fragments reveal mostly non‐specific contacts distributed across the length of RRM and RGG2 save for EWSRRM‐RGG2 residues 390–395, which interact with all three fragments and are part of a surface‐exposed loop within the RRM (Figure S7a,b). Moreover, the RGG2 domain displayed a higher number of interactions with EWSLCD fragments compared to the RRM (Figures 3 and S7a,b).

FIGURE 4.

FIGURE 4

Open in a new tab

RNA‐recognition motif (RRM) and RGG2 form interdomain contacts with EWSLCD fragments. Chemical shift perturbations (CSPs) (top panels) and line broadening (bottom panels) of 15N‐EWSRRM‐RGG2 titrated with (3:1 molar ratio) (a) EWS1–120, (b) EWS91–199, or (c) EWS171–264. CSPs (line broadening) > (<) 1 standard deviation (dashed line) are plotted in cyan (EWS1–120), pink (EWS91–199), or green (EWS171–264). Asterisks mark unassigned/ambiguous/broadened residues (red) and proline (black). Unassigned glycine (G) or unclassified (R/X) resonances are plotted (right panels). EWS, RNA‐binding protein EWS; RGG, arginine–glycine–glycine.

Several EWSRRM‐RGG2 residues consistently shifted and/or broadened with the different EWSLCD fragments, suggesting that residue types and the local environment mediate interactions. Residues Y401, V442, and N452 were shifted in all three titrations, residues D359, N390, F430, and R455 were shifted in two of the three titrations, and residue N452 and two unassigned glycine residues were broadened in all three titrations (Figure S7c). Additionally, M451, N452, D493, and R494 peaks were all broadened by at least two of the three fragments, representing two pairs of neighboring residues that made consistent contact with EWSLCD (Figure S7c). CSPs and line broadening did not localize to a specific region of the EWSRRM‐RGG2 RRM domain (Figure S8a–d) and were dispersed between both the positively charged nucleic acid binding face and the opposite negatively charged face (Figure S8e). Analysis of simulation data reveals K391‐R392 and R446‐K447 consistently formed interactions with the three EWSLCD fragments (Figure S7a,b). Within RGG2, the sequence FPPRGPRGSR showed a higher level of interaction with all three EWSLCD fragments. These interactions remained transient, as reflected by fluctuations in contact probabilities (Figure S7b).

2.5. Numerous residue types contribute to the interaction between EWSLCD and EWSRRM‐RGG2

We stratified the observed CSPs and line broadening from EWSLCD and EWSRRM‐RGG2 titrations by amino acid side chain (Figure S9). For EWSLCD, shifts and line broadening occurred extensively across polar residues, as well as in acidic, aromatic, and hydrophobic residues, indicating that this interaction is driven by numerous non‐specific contacts (Figure S9a). Peak shifts and line broadening were detected in each type of residue within EWSRRM‐RGG2, with 36% of all residues shifted by at least one of the three EWSLCD fragments, >60% of assigned positively charged residues, and ~40% of assigned polar residues shifted (Figures S7c and S9b). Analysis of residue contributions from dilute phase simulations revealed that various residue positions in both EWSLCD fragments and EWSRRM‐RGG2 form transient contacts. In EWSLCD fragments, tyrosine, polar residues, proline, and aspartate contributed to interactions with EWSRRM‐RGG2 (Figure S9a). In EWSRRM‐RGG2, arginine, phenylalanine, polar residues, proline, and methionine appear to interact with the EWSLCD fragments, with arginine, glycine, and proline being prominent contributors (Figure S9b).

Together, these results suggest that the LCD or RBD lacks a single specific binding region that mediates their association. Instead, the interdomain association is stabilized by an accumulation of relatively weak, but highly abundant contacts between numerous residue pairs that are dispersed across the entire EWSLCD and both RRM and RGG2 (Figure S10). In the full EWSLCD, these interactions become numerous enough to be potent drivers of condensate formation, which are likely to be even more readily formed in full‐length EWS, where additional contacts are made between the LCD and the other RBDs.

2.6. Interdomain association between EWSLCD and EWSRRM‐RGG2 in full‐length EWS

To understand how interdomain interactions manifest in the context of full‐length EWS, we employed dilute‐phase atomistic simulations (Figure S5d). Over the course of 16 μs, the LCD of full‐length EWS interacted with parts of the RBD, preferring disordered segments over the folded domains, and the folded RBD elements predominantly interacted with their flanking regions (Figure 5a). Comparing the simulations of full‐length and EWSLCD fragments revealed that the overall pattern of interactions between EWSLCD and EWSRRM‐RGG2 is consistent (Figure 5b). An interesting deviation emerged in EWS between residues 150–200 (breakpoint between EWS99–199 and EWS171–264), which showed reduced interactions with EWSRRM‐RGG2 in the full‐length protein relative to the EWSLCD fragments (Figure 5b). This segment preferentially interacted with other RBDs, such as RGG1 and RGG3, suggesting that in the full‐length system, structural constraints restrict conformational flexibility, while the presence of additional RBDs provides alternative interaction partners, thereby limiting its ability to engage with RRM‐RGG2 (Figure 5b). In contrast, in isolated fragment simulations, this region was unconstrained and able to explore a broader conformational ensemble, allowing stronger interactions with EWSRRM‐RGG2 (Figure 5b). Indeed, inter‐residue distances between the EWSLCD fragments in their isolated state versus the full‐length protein revealed that EWSLCD fragments were more collapsed in the full‐length system (Figure S11a). Further, residue‐type analysis revealed consistent patterns of interactions between the full‐length and fragment simulations identifying tyrosine, polar residues, proline, and aspartate in EWSLCD, and arginine, proline, and glycine in EWSRRM‐RGG2 as the key residues contributing to the interaction network (Figures 5c and S11b). Representative snapshots from full‐length EWS simulations further support the transient and non‐specific nature of these interactions (Figure 5d). Additionally, analysis of total interdomain interactions formed by EWSLCD with different RBD domains confirms this distributed interaction pattern (Figure S12). Together, these results provide confidence that our biophysical studies of simplified systems capture the transient, non‐specific interactions that are relevant to self‐association and phase separation in the context of full‐length EWS.

FIGURE 5.

FIGURE 5

Open in a new tab

RNA‐recognition motif (RRM) and RGG2 form interdomain contacts with the low‐complexity domain (LCD) in full‐length RNA‐binding protein EWS (EWS). (a) Normalized average pairwise van der Waals contacts from single chain atomistic simulations of full‐length EWS. (b) Average per‐residue contacts between EWSLCD and EWSRRM‐RGG2. Contacts formed by EWSLCD regions, comparing full‐length EWS versus individual LCD fragments in fragment simulations (top) and contacts formed by EWSRRM‐RGG2, comparing interactions with LCD in full‐length EWS versus with individual LCD fragments in fragment simulations (bottom). EWSLCD,FL (extracted from atomistic simulations of single chain full‐length EWS, black), EWS1–120 (cyan), EWS91–199 (pink), or EWS171–264 (green). (c) Average pairwise interdomain residue type contacts between EWSLCD and EWSRRM‐RGG2 extracted from single chain atomistic simulations of full‐length EWS; backbone–backbone (red), backbone–sidechain (blue), and sidechain–sidechain (green), first residue is from EWSLCD, second residue is from EWSRRM‐RGG2. (d) Examples of EWSLCD‐EWSRRM‐RGG2 ensemble images (EWSLCD, blue; RRM, yellow; RGG2, red) from atomistic simulations of full‐length EWS. RGG, arginine–glycine–glycine.

3. DISCUSSION

Many of the functions of EWS and EWS::FLI1 are thought to be mediated through condensate formation that is primarily driven by inter‐ and intra‐molecular interactions between multiple residue types within the EWSLCD (Chong et al., 2018; Johnson et al., 2024; Kato et al., 2012; Kwon et al., 2013; Nosella et al., 2021; Sundara Rajan et al., 2024). However, studies have also demonstrated that interdomain interactions between the LCD and RBD of FUS (and other RNA‐binding proteins) are important for condensate formation (Martin et al., 2021; Murthy et al., 2021; Wake et al., 2025). Here we uncover similar, transient, non‐specific interactions between the EWS LCD and RBDs and discover that charge and sequence patterning within the RGG motifs differentially influence EWS phase separation. Our findings are consistent with the growing body of work pointing to both the patterning of interaction sites and the quantity of contacts, rather than specific location, as key drivers of biomolecular condensation (Galagedera et al., 2023; Lin et al., 2017; Martin et al., 2020; Rekhi et al., 2023; Rekhi et al., 2024).

Recent studies on the miscibility of intrinsically disordered proteins (IDPs) highlighted that client partitioning is governed by the balance between homotypic and heterotypic interactions between the client and scaffold proteins (Rana et al., 2024; Welles et al., 2024). Moreover, studies have highlighted that charge segregation enhances compaction and condensate formation (Das & Pappu, 2013; Sawle & Ghosh, 2015; Schuster et al., 2020). Therefore, our observations of reduced co‐partitioning of EWSRGG1‐RRM compared to EWSRRM‐RGG2 and EWSRGG3 with EWSLCD may arise from sequence differences between the RGG motifs. In addition to having weaker heterotypic interactions with EWSLCD, the charge patterning in RGG1 may increase its homotypic strength (keeping it below the threshold for phase separation), allowing it to compete with heterotypic interactions, while the heterotypic interactions of EWSRRM‐RGG2 and EWSRGG3 with EWSLCD appear sufficient to support co‐partitioning under similar conditions. Co‐partitioning can also be aided by homotypic interactions within the client domains, provided that heterotypic interactions are sufficiently strong (Welles et al., 2024). Notably, upon the addition of salt, the partitioning of both EWSRRM‐RGG2 and EWSRGG3 into condensates is reduced, likely due to weakened electrostatic interactions that disrupt both homotypic and heterotypic contacts. However, condensate partitioning assays and imaging tell a more complex story, wherein salt addition only significantly affects the condensates with EWSRRM‐RGG2. Specifically, EWSRRM‐RGG2 shows reduced partitioning as NaCl concentration increases, while EWSLCD droplet partitioning initially decreases but then rapidly re‐enters the condensed phase in a salt‐dependent manner. This reentrant behavior could result from electrostatic screening effects at low salt concentrations followed by salting‐out effects at higher concentrations (Krainer et al., 2021; Lin et al., 2025). While EWSLCD condensates remain spherical over time, the addition of EWSRRM‐RGG2 or EWSRGG3 causes non‐spherical droplet morphologies, an effect more pronounced with EWSRGG3. This greater surface‐wetting effect may be due to the greater positive charge carried by EWSRGG3. However, in our previous work, the surface‐wetting effect was not detected with relatively inflexible poly‐L‐lysine peptides (Selig et al., 2025), suggesting that chain flexibility likely also plays an important role in how the proteins in a condensate interact with surfaces.

Kamagata et al. recently reported a size‐dependent condensate exclusion effect in FUS droplets, where the partitioning of folded proteins is restricted while disordered domains are preferentially recruited (Kamagata et al., 2022). The retention of the folded RRM on EWSRRM‐RGG2 may thus limit its incorporation into condensates relative to the fully disordered EWSRGG3. Additionally, the relative lengths and net‐charge differences between the EWS RGG domains affect their condensate partitioning and influence condensate dynamics. For example, the distinct two‐phase regime governed by ionic strength in the EWSLCD + EWSRRM‐RGG2 condensates highlights the sophisticated regulatory factors governing condensate formation and the highly complex interplay between the different domains of EWS. We observe a notable enrichment of contacts involving tyrosine, polar residues, proline, and aspartate in EWSLCD with arginine, glycine, polar residues, and proline in EWSRRM‐RGG2, suggesting a chemically selective interaction profile involving π–π interactions, hydrogen bonding, and electrostatics. Simulations identify a region within RGG2 (FPPRGPRGSR) in which arginines occur at regular intervals, with proline positioned nearby. Given that proline can engage in CH‐π (Zondlo, 2013) and CH‐O (Daniecki et al., 2022) interactions and impart local conformational rigidity (Hazra et al., 2023; Mateos et al., 2020; Perez et al., 2014), this arrangement may support favorable local contacts and interaction dynamics. Our atomistic simulations of full‐length EWS reveal that while interaction patterns change in the context of the full‐length protein, with the LCD gaining more opportunities to interact with multiple RBDs, the overall chemical interaction specificity remains consistent with that observed in the fragment simulations.

4. CONCLUSION

Our study shows EWS RBD‐derived domains interact non‐specifically with its LCD and that these interactions lack well‐defined or structured binding sites and instead engage in weak, transient interactions involving multiple residue type pairs. As in vitro condensates are defined by a complex network of multiple biomolecular partners, we expect that inclusion of nucleic acids, particularly RNA, and other relevant partners such as EWS::FLI1 will alter EWS dynamics in vivo. Indeed, the presence of the FLI1DBD has a pronounced effect on EWSLCD condensates (Selig et al., 2025), effects which are likely to be altered in the presence of the EWS RBD. These findings provide the basis for understanding how EWS interdomain self‐association occurs and influences condensate formation and emphasize the need for further investigation into the biological consequences of disrupting these interactions through mutagenesis, post‐translational modification, and modulation with other biomolecules, particularly nucleic acids.

5. MATERIALS AND METHODS

Protein constructs were expressed in Escherichia coli then purified via immobilized metal ion and size‐exclusion chromatography. EWSLCD and EWSRRM‐RGG2 were fluorescently labeled via transpeptidase. Turbidity and condensate partitioning assays were performed to assess the effect of different regions of the EWS RBD on EWSLCD condensate formation. 1H,15N‐HSQC NMR titrations were used to analyze the interaction between fragments of the EWSLCD and EWSRRM‐RGG2. Coexistence coarse grain (CG) simulations of full‐length EWS and EWSLCD were used to evaluate relative phase separation propensities, while atomistic simulations in the dilute phase were employed to examine interactions between EWSLCD fragments and EWSRRM‐RGG2. Additionally, atomistic simulations of full‐length EWS in the dilute phase were conducted to investigate interdomain interactions in the context of the full‐length protein. Detailed descriptions of the materials and methods used in this study are included in Supporting Information S1.

AUTHOR CONTRIBUTIONS

Erich J. Sohn: Conceptualization; investigation; formal analysis; visualization; writing – original draft; writing – review and editing. Kandarp A. Sojitra: Conceptualization; investigation; software; formal analysis; visualization; writing – original draft; writing – review and editing. Leticia Rodrigues: Investigation; writing – review and editing. Xiaoping Xu: Investigation; writing – review and editing. Bess Frost: Resources; supervision; writing – review and editing. Jeetain Mittal: Conceptualization; formal analysis; funding acquisition; resources; software; supervision; writing – review and editing; project administration. David S. Libich: Conceptualization; formal analysis; funding acquisition; resources; supervision; writing – review and editing; project administration.

CONFLICT OF INTEREST STATEMENT

The authors declare no competing interests.

Supporting information

Data S1. Supporting Information.

PRO-34-e70316-s001.pdf (14.2MB, pdf)

ACKNOWLEDGMENTS

This study was funded in part by the Alex's Lemonade Stand Foundation 1335092, the Welch Foundation A‐2113‐20220331 (Jeetain Mittal). A portion of this work was performed at Brown University with support from the Department of Molecular Biology, Cell Biology & Biochemistry. We gratefully acknowledge the computational resources provided by the Texas A&M High Performance Research Computing (HPRC). This work is based upon research conducted in the Structural Biology Core Facilities, a part of the Institutional Research Cores at the University of Texas Health Science Center at San Antonio, supported by the National Institutes of Health (R01GM140127, R35GM153338, F31CA295030, P30CA054174).

Sohn EJ, Sojitra KA, Rodrigues L, Xu X, Frost B, Mittal J, et al. Molecular basis of EWS interdomain self‐association and its role in condensate formation. Protein Science. 2025;34(10):e70316. 10.1002/pro.70316

Erich J. Sohn and Kandarp A. Sojitra contributed equally.

Review Editor: Jean Baum

Contributor Information

Jeetain Mittal, Email: jeetain@tamu.edu.

David S. Libich, Email: libich@uthscsa.edu.

DATA AVAILABILITY STATEMENT

NMRPipe processing scripts are available upon reasonable request; expression plasmids were deposited at Addgene: EWSLCD (180467), EWS1–120 (180464), EWS99–199 (180465), EWS171–264 (180466), EWSRGG1‐RRM (238369), EWSRRM‐RGG2 (188046), and EWSRGG3 (238370).

REFERENCES

  1. Ahmed NS, Harrell LM, Wieland DR, Lay MA, Thompson VF, Schwartz JC. Fusion protein EWS‐FLI1 is incorporated into a protein granule in cells. RNA. 2021;27(8):920–932. 10.1261/rna.078827.121 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alex D, Lee KA. RGG‐boxes of the EWS oncoprotein repress a range of transcriptional activation domains. Nucleic Acids Res. 2005;33(4):1323–1331. 10.1093/nar/gki270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bertolotti A, Melot T, Acker J, Vigneron M, Delattre O, Tora L. EWS, but not EWS‐FLI‐1, is associated with both TFIID and RNA polymerase II: interactions between two members of the TET family, EWS and hTAFII68, and subunits of TFIID and RNA polymerase II complexes. Mol Cell Biol. 1998;18(3):1489–1497. 10.1128/MCB.18.3.1489 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bock AS, Murthy AC, Tang WS, Jovic N, Shewmaker F, Mittal J, et al. N‐terminal acetylation modestly enhances phase separation and reduces aggregation of the low‐complexity domain of RNA‐binding protein fused in sarcoma. Protein Sci. 2021;30(7):1337–1349. 10.1002/pro.4029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Boone MA, Taslim C, Crow JC, Selich‐Anderson J, Byrum AK, Showpnil IA, et al. The FLI portion of EWS/FLI contributes a transcriptional regulatory function that is distinct and separable from its DNA‐binding function in Ewing sarcoma. Oncogene. 2021;40(29):4759–4769. 10.1038/s41388-021-01876-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Boulay G, Sandoval GJ, Riggi N, Iyer S, Buisson R, Naigles B, et al. Cancer‐specific retargeting of BAF complexes by a prion‐like domain. Cell. 2017;171(1):163–178.e19. 10.1016/j.cell.2017.07.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chong S, Dugast‐Darzacq C, Liu Z, Dong P, Dailey GM, Cattoglio C, et al. Imaging dynamic and selective low‐complexity domain interactions that control gene transcription. Science. 2018;361(6400):eaar2555. 10.1126/science.aar2555 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chong S, Graham TGW, Dugast‐Darzacq C, Dailey GM, Darzacq X, Tjian R. Tuning levels of low‐complexity domain interactions to modulate endogenous oncogenic transcription. Mol Cell. 2022;82(11):2084–2097.e5. 10.1016/j.molcel.2022.04.007 [DOI] [PubMed] [Google Scholar]
  9. Daniecki NJ, Bhatt MR, Yap GPA, Zondlo NJ. Proline C‐H bonds as loci for proline assembly via C‐H/O interactions. Chembiochem. 2022;23(24):e202200409. 10.1002/cbic.202200409 [DOI] [PubMed] [Google Scholar]
  10. Das RK, Pappu RV. Conformations of intrinsically disordered proteins are influenced by linear sequence distributions of oppositely charged residues. Proc Natl Acad Sci U S A. 2013;110(33):13392–13397. 10.1073/pnas.1304749110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dcosta N, Black M, Huang R. A simulation‐guided “swapping” protocol for NMR titrations to study protein–protein interactions. Can J Chem. 2024;102(11):734–744. 10.1139/cjc-2024-0031 [DOI] [Google Scholar]
  12. Delattre O, Zucman J, Plougastel B, Desmaze C, Melot T, Peter M, et al. Gene fusion with an ETS DNA‐binding domain caused by chromosome translocation in human tumours. Nature. 1992;359(6391):162–165. 10.1038/359162a0 [DOI] [PubMed] [Google Scholar]
  13. Embree LJ, Azuma M, Hickstein DD. Ewing sarcoma fusion protein EWSR1/FLI1 interacts with EWSR1 leading to mitotic defects in zebrafish embryos and human cell lines. Cancer Res. 2009;69(10):4363–4371. 10.1158/0008-5472.CAN-08-3229 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fisher C. The diversity of soft tissue tumours with EWSR1 gene rearrangements: a review. Histopathology. 2014;64(1):134–150. 10.1111/his.12269 [DOI] [PubMed] [Google Scholar]
  15. Flucke U, van Noesel MM, Siozopoulou V, Creytens D, Tops BBJ, van Gorp JM, et al. EWSR1‐the Most common rearranged gene in soft tissue lesions, which also occurs in different bone lesions: an updated review. Diagnostics. 2021;11(6):1093. 10.3390/diagnostics11061093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Galagedera SKK, Dao TP, Enos SE, Chaudhuri A, Schmit JD, Castaneda CA. Polyubiquitin ligand‐induced phase transitions are optimized by spacing between ubiquitin units. Proc Natl Acad Sci U S A. 2023;120(42):e2306638120. 10.1073/pnas.2306638120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gorthi A, Romero JC, Loranc E, Cao L, Lawrence LA, Goodale E, et al. EWS‐FLI1 increases transcription to cause R‐loops and block BRCA1 repair in Ewing sarcoma. Nature. 2018;555(7696):387–391. 10.1038/nature25748 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hazra MK, Gilron Y, Levy Y. Not only expansion: proline content and density also induce disordered protein conformation compaction. J Mol Biol. 2023;435(17):168196. 10.1016/j.jmb.2023.168196 [DOI] [PubMed] [Google Scholar]
  19. Jaishankar S, Zhang J, Roussel MF, Baker SJ. Transforming activity of EWS/FLI is not strictly dependent upon DNA‐binding activity. Oncogene. 1999;18:5592–5597. [DOI] [PubMed] [Google Scholar]
  20. Johnson KM, Mahler NR, Saund RS, Theisen ER, Taslim C, Callender NW, et al. Role for the EWS domain of EWS/FLI in binding GGAA‐microsatellites required for Ewing sarcoma anchorage independent growth. Proc Natl Acad Sci U S A. 2017;114(37):9870–9875. 10.1073/pnas.1701872114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Johnson CN, Sojitra KA, Sohn EJ, Moreno‐Romero AK, Baudin A, Xu X, et al. Insights into molecular diversity within the FUS/EWS/TAF15 protein family: unraveling phase separation of the N‐terminal low‐complexity domain from RNA‐binding protein EWS. J Am Chem Soc. 2024;146(12):8071–8085. 10.1021/jacs.3c12034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Johnson CN, Xu X, Holloway SP, Libich DS. The (1)H, (15)N and (13)C resonance assignments of the low‐complexity domain from the oncogenic fusion protein EWS‐FLI1. Biomol NMR Assign. 2022;16(1):67–73. 10.1007/s12104-021-10061-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kamagata K, Iwaki N, Kanbayashi S, Banerjee T, Chiba R, Gaudon V, et al. Structure‐dependent recruitment and diffusion of guest proteins in liquid droplets of FUS. Sci Rep. 2022;12(1):7101. 10.1038/s41598-022-11177-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kato M, Han TW, Xie S, Shi K, Du X, Wu LC, et al. Cell‐free formation of RNA granules: low complexity sequence domains form dynamic fibers within hydrogels. Cell. 2012;149(4):753–767. 10.1016/j.cell.2012.04.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kim YR, Joo J, Lee HJ, Kim C, Park JC, Yu YS, et al. Prion‐like domain mediated phase separation of ARID1A promotes oncogenic potential of Ewing's sarcoma. Nat Commun. 2024;15(1):6569. 10.1038/s41467-024-51050-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Krainer G, Welsh TJ, Joseph JA, Espinosa JR, Wittmann S, de Csillery E, et al. Reentrant liquid condensate phase of proteins is stabilized by hydrophobic and non‐ionic interactions. Nat Commun. 2021;12(1):1085. 10.1038/s41467-021-21181-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kwon I, Kato M, Xiang S, Wu L, Theodoropoulos P, Mirzaei H, et al. Phosphorylation‐regulated binding of RNA polymerase II to fibrous polymers of low‐complexity domains. Cell. 2013;155(5):1049–1060. 10.1016/j.cell.2013.10.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lay MA, Thompson VF, Adelakun AD, Schwartz JC. Ewing sarcoma related protein 1 recognizes R‐loops by binding DNA forks. Biopolymers. 2024;115(3):e23576. 10.1002/bip.23576 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lee SG, Kim N, Kim SM, Park IB, Kim H, Kim S, et al. Ewing sarcoma protein promotes dissociation of poly(ADP‐ribose) polymerase 1 from chromatin. EMBO Rep. 2020;21(11):e48676. 10.15252/embr.201948676 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lessnick SL, Braun BS, Denny CT, May WA. Multiple domains mediate transformation by the Ewing's sarcoma EWS/FLI‐1 fusion gene. Oncogene. 1995;10(3):423–431. [PubMed] [Google Scholar]
  31. Li KK, Lee KA. Transcriptional activation by the Ewing's sarcoma (EWS) oncogene can be cis‐repressed by the EWS RNA‐binding domain. J Biol Chem. 2000;275(30):23053–23058. 10.1074/jbc.M002961200 [DOI] [PubMed] [Google Scholar]
  32. Li H, Watford W, Li C, Parmelee A, Bryant MA, Deng C, et al. Ewing sarcoma gene EWS is essential for meiosis and B lymphocyte development. J Clin Invest. 2007;117(5):1314–1323. 10.1172/JCI31222 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lin Y, Currie SL, Rosen MK. Intrinsically disordered sequences enable modulation of protein phase separation through distributed tyrosine motifs. J Biol Chem. 2017;292(46):19110–19120. 10.1074/jbc.M117.800466 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lin Y‐H, Kim TH, Das S, Pal T, Wessén J, Rangadurai AK, et al. Electrostatics of salt‐dependent reentrant phase behaviors highlights diverse roles of ATP in biomolecular condensates. eLife. 2025;13:RP100284. 10.7554/Elife100284.3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Maki RG, Grohar PJ, Antonescu CR. Ewing sarcoma and related FET family translocation‐associated round cell tumors: a century of clinical and scientific progress. Genes Chromosomes Cancer. 2022;61(8):509–517. 10.1002/gcc.23050 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Martin EW, Holehouse AS, Peran I, Farag M, Incicco JJ, Bremer A, et al. Valence and patterning of aromatic residues determine the phase behavior of prion‐like domains. Science. 2020;367(6478):694–699. 10.1126/science.aaw8653 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Martin EW, Thomasen FE, Milkovic NM, Cuneo MJ, Grace CR, Nourse A, et al. Interplay of folded domains and the disordered low‐complexity domain in mediating hnRNPA1 phase separation. Nucleic Acids Res. 2021;49(5):2931–2945. 10.1093/nar/gkab063 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Mateos B, Conrad‐Billroth C, Schiavina M, Beier A, Kontaxis G, Konrat R, et al. The ambivalent role of proline residues in an intrinsically disordered protein: from disorder promoters to compaction facilitators. J Mol Biol. 2020;432(9):3093–3111. 10.1016/j.jmb.2019.11.015 [DOI] [PubMed] [Google Scholar]
  39. Murthy AC, Dignon GL, Kan Y, Zerze GH, Parekh SH, Mittal J, et al. Molecular interactions underlying liquid‐liquid phase separation of the FUS low‐complexity domain. Nat Struct Mol Biol. 2019;26(7):637–648. 10.1038/s41594-019-0250-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Murthy AC, Tang WS, Jovic N, Janke AM, Seo DH, Perdikari TM, et al. Molecular interactions contributing to FUS SYGQ LC‐RGG phase separation and co‐partitioning with RNA polymerase II heptads. Nat Struct Mol Biol. 2021;28(11):923–935. 10.1038/s41594-021-00677-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Ng KP, Potikyan G, Savene RO, Denny CT, Uversky VN, Lee KA. Multiple aromatic side chains within a disordered structure are critical for transcription and transforming activity of EWS family oncoproteins. Proc Natl Acad Sci U S A. 2007;104(2):479–484. 10.1073/pnas.0607007104 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Nosella ML, Tereshchenko M, Pritisanac I, Chong PA, Toretsky JA, Lee HO, et al. O‐linked‐N‐acetylglucosaminylation of the RNA‐binding protein EWS N‐terminal low complexity region reduces phase separation and enhances condensate dynamics. J Am Chem Soc. 2021;143(30):11520–11534. 10.1021/jacs.1c04194 [DOI] [PubMed] [Google Scholar]
  43. Perez RB, Tischer A, Auton M, Whitten ST. Alanine and proline content modulate global sensitivity to discrete perturbations in disordered proteins. Proteins. 2014;82(12):3373–3384. 10.1002/prot.24692 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Petermann R, Mossier BM, Aryee DN, Khazak V, Golemis EA, Kovar H. Oncogenic EWS‐Fli1 interacts with hsRPB7, a subunit of human RNA polymerase II. Oncogene. 1998;17(5):603–610. 10.1038/sj.onc.1201964 [DOI] [PubMed] [Google Scholar]
  45. Rana U, Xu K, Narayanan A, Walls MT, Panagiotopoulos AZ, Avalos JL, et al. Asymmetric oligomerization state and sequence patterning can tune multiphase condensate miscibility. Nat Chem. 2024;16(7):1073–1082. 10.1038/s41557-024-01456-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Rekhi S, Garcia CG, Barai M, Rizuan A, Schuster BS, Kiick KL, et al. Expanding the molecular language of protein liquid‐liquid phase separation. Nat Chem. 2024;16(7):1113–1124. 10.1038/s41557-024-01489-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Rekhi S, Sundaravadivelu Devarajan D, Howard MP, Kim YC, Nikoubashman A, Mittal J. Role of strong localized vs weak distributed interactions in disordered protein phase separation. J Phys Chem B. 2023;127(17):3829–3838. 10.1021/acs.jpcb.3c00830 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Sawle L, Ghosh K. A theoretical method to compute sequence dependent configurational properties in charged polymers and proteins. J Chem Phys. 2015;143(8):085101. 10.1063/1.4929391 [DOI] [PubMed] [Google Scholar]
  49. Schuster BS, Dignon GL, Tang WS, Kelley FM, Ranganath AK, Jahnke CN, et al. Identifying sequence perturbations to an intrinsically disordered protein that determine its phase‐separation behavior. Proc Natl Acad Sci U S A. 2020;117(21):11421–11431. 10.1073/pnas.2000223117 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Selig EE, Bhura R, White MR, Akula S, Hoffman RD, Tovar CN, et al. Biochemical and biophysical characterization of the nucleic acid binding properties of the RNA/DNA binding protein EWS. Biopolymers. 2023;114(5):e23536. 10.1002/bip.23536 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Selig EE, Sohn EJ, Stoja A, Moreno‐Romero AK, Akula S, Xu X, et al. Phase separation of the oncogenic fusion protein EWS::FLI1 is modulated by its DNA‐binding domain. Proc Natl Acad Sci U S A. 2025;122(20):e2221823122. 10.1073/pnas.2221823122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Spahn L, Siligan C, Bachmaier R, Schmid JA, Aryee DN, Kovar H. Homotypic and heterotypic interactions of EWS, FLI1 and their oncogenic fusion protein. Oncogene. 2003;22(44):6819–6829. 10.1038/sj.onc.1206810 [DOI] [PubMed] [Google Scholar]
  53. Sundara Rajan S, Ebegboni VJ, Pichling P, Ludwig KR, Jones TL, Chari R, et al. Endogenous EWSR1 exists in two visual modalities that reflect its associations with nucleic acids and concentration at sites of active transcription. Mol Cell Biol. 2024;44(3):103–122. 10.1080/10985549.2024.2315425 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Turc‐Carel C, Philip I, Berger MP, Philip T, Lenoir G. Chromosomal translocation (11; 22) in cell lines of Ewing's sarcoma. CR Seances Acad Sci III. 1983;296(23):1101–1103. [PubMed] [Google Scholar]
  55. Vasileva E, Arata C, Luo Y, Burgos R, Crump JG, Amatruda JF. Origin of Ewing sarcoma by embryonic reprogramming of neural crest to mesoderm. bioRxiv. 2024;2024.10.27.620438‐2024.10.27.38. 10.1101/2024.10.27.620438 [DOI] [Google Scholar]
  56. Wake N, Weng SL, Zheng T, Wang SH, Kirilenko V, Mittal J, et al. Expanding the molecular grammar of polar residues and arginine in FUS phase separation. Nat Chem Biol. 2025;21:1–1088. 10.1038/s41589-024-01828-6 [DOI] [PubMed] [Google Scholar]
  57. Welles RM, Sojitra KA, Garabedian MV, Xia B, Wang W, Guan M, et al. Determinants that enable disordered protein assembly into discrete condensed phases. Nat Chem. 2024;16(7):1062–1072. 10.1038/s41557-023-01423-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Yang L, Chansky HA, Hickstein DD. EWS.Fli‐1 fusion protein interacts with hyperphosphorylated RNA polymerase II and interferes with serine‐arginine protein‐mediated RNA splicing. J Biol Chem. 2000;275(48):37612–37618. 10.1074/jbc.M005739200 [DOI] [PubMed] [Google Scholar]
  59. Zondlo NJ. Aromatic‐proline interactions: electronically tunable CH/pi interactions. Acc Chem Res. 2013;46(4):1039–1049. 10.1021/ar300087y [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zuo L, Zhang G, Massett M, Cheng J, Guo Z, Wang L, et al. Loci‐specific phase separation of FET fusion oncoproteins promotes gene transcription. Nat Commun. 2021;12(1):1491. 10.1038/s41467-021-21690-7 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1. Supporting Information.

PRO-34-e70316-s001.pdf (14.2MB, pdf)

Data Availability Statement

NMRPipe processing scripts are available upon reasonable request; expression plasmids were deposited at Addgene: EWSLCD (180467), EWS1–120 (180464), EWS99–199 (180465), EWS171–264 (180466), EWSRGG1‐RRM (238369), EWSRRM‐RGG2 (188046), and EWSRGG3 (238370).

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES