Visualization of liquid-liquid phase transitions using a tiny G-quadruplex binding protein

Abstract

Liquid-liquid phase transitions govern a wide range of protein-protein and protein-RNA interactions. Although the importance of multivalency and protein disorder in driving these transitions is clear, there is limited knowledge concerning the structural basis of phase transitions or the conformational changes that accompany this process. In this work, we found that a small human protein, SERF2, is important for the formation of stress granules. We determined the solution NMR structure ensemble of SERF2. We show that SERF2 specifically interacts with non-canonical tetrahelical RNA structures called G-quadruplexes, structures linked to stress granule formation. The biophysical amenability of both SERF2 and RNA G4 quadruplexes have allowed us to characterize the multivalent protein-RNA interactions involved in liquid-liquid phase transitions, the role that protein disorder plays in these transitions, identify the specific contacts involved, and describe how these interactions impact the structural dynamics of the components enabling a detailed understanding of the structural transitions involved in early stages of ribonucleoprotein condensate formation.

Fig. 1|. High-throughput screening for SERF2 binding specificity.

a Fluorescence polarization assay used to measure the binding affinity of SERF2 to 6-FAM labeled poly ribo-polynucleotides and rG4 quadruplex forming sequences, as indicated. Data are presented as mean values ± standard deviations obtained from three independent replicates. b A schematic representation of RNA Bind-n-Seq experiments using RNA pools. A randomized DNA oligonucleotide pool was transcribed to RNA and folded in a buffer containing either KCl which favors rG4 formation or LiCl which disfavors rG4 folding. An additional RNA pool was made by replacing guanines (G) with 7-deaza (7dG) to eliminate rG4 quadruplex folding. These pools were mixed with GST-SBP-SERF2, and bound RNA was isolated and sequenced with ~2 × 10⁶ reads. c RNA Bind-n-Seq analysis of the 6-mers enrichment in KCl versus RNA that was made with 7dG in a sample mixture containing 50 nM SERF2. The guanine-rich 6mers in the top 5 kmers are labeled. d Enrichment of rG4 quadruplex patterns in different conditions were classified by G4 quadruplex strength in buffer containing 50 nM SERF2. Sequences containing 3 or more guanines in the G-tetrad are referred to as strong rG4s, while sequences with ≥8 guanines but lacking a defined G4 forming motif as defined by QGRS mapper program are referred to as non rG4s. Data are presented as mean values ± standard deviations for each condition obtained from two different RNA library preparations. e FOREST analysis: average binding intensities of SERF2 to a library containing 1800 folded human pre-miRNAs (orange curve) and to 10 defined rG4 structures (grey curve). The p-value shown was determined by the two-tailed Brunner-Munzel test. f The fluorescence polarization plot that shows binding affinity measurements between SERF2 and three different 6-FAM labeled rG4s. The binding assays in a and f were done with varied protein concentrations mixed with 20 nM rG4 quadruplex or polynucleotides at room temperature in 20 mM NaPi (pH 7.4), 100 mM KCl. The standard deviations shown were calculated from three independent replicates, and data are presented as mean values.

Fig. 2|. RNA interaction drives liquid-liquid phase transition in SERF2.

a Phase regime illustrating phase transitions for SERF2 as a function of total RNA concentration using RNA extracted from HeLa cells (left panel). Fluorescence imaging shows gel-like structures present in solutions containing 50 μM SERF2 mixed with 200 ng of total HeLa cell RNA containing 10% (w/v) PEG8000, incubated for 30 minutes at room temperature (right panel). b Dynamics and recovery of Cy5-labeled SERF2 in total HeLa cell RNA droplets, obtained by FRAP analysis, suggest that the mesh-like condensates observed in a, right panel are dynamic and reversible. Standard deviations were calculated by analyzing 6 isolated droplets subjected to FRAP, and data are presented as mean values. c Fluorescence images show 50 μM SERF2, dissolved in 20 mM NaPi (pH 7.4),100 mM KCl, readily undergoes phase transition (right) when mixed with equimolar concentrations of three different rG4s, TERRA23, (G4C2)4, and (UG4U)6. Similar results were obtained in three independent repeated experiments. The sample mixture contains 1/200^th Cy-5 labeled SERF2 (purple) and 6-FAM rG4 (green), as indicated in the figure inset. Phase regimes illustrating phase transitions for SERF2 and TERRA23 rG4 in 10% (w/v) PEG8000, at varying protein and RNA concentrations, are shown on the right panel. d Phase diagram showing SERF2 and TERRA23 rG4 sample mixtures undergoing phase separation, in varying salt and PEG8000 concentrations. e Two-component FRAP analysis was done to measure the recovery rates of SERF2 (purple) and TERRA23 rG4 (green) in SERF2-rG4 droplets. Data are presented as mean values, and standard errors were calculated by analyzing 8 isolated droplets subjected to FRAP. The pre-bleached, after-bleached (0 s), and recovered droplets (300 s) are shown above the FRAP plot. The FRAP data were fitted in GraphPad Prism, using a non-linear regression, one-phase association model, to obtain the recovery halftime (t1/2) reported in the text.

Fig. 3 |. High-resolution NMR structure of SERF2 reveals that its disordered and dynamic N-terminal domain binds TERRA rG4.

a 20 best NMR ensemble model structures of SERF2. The average converged helical structure spanning residues 33–47 in SERF2 is shown as the helix in orange. b 1H-15N NMR assignment (red spectrum) of 100 μM human SERF2 mixed with 20 μM (yellow) and 50 μM (purple) TERRA23 rG4 at 4 °C. Spectral zooms on the right illustrate chemical shift changes in several N- and C-terminal residues at 2:1 protein: rG4 ratio. c The 1H-15N chemical shift perturbations (left y-axis) were calculated from (b) and plotted as color-shaded peaks for each assigned residue in SERF2 at increasing TERRA23 rG4 concentrations. The yellow and purple colors in the graph correlate to their corresponding spectrum shown in (b), unassigned peaks are denoted with an asterisk. The heteroNOE values of SERF2 in the absence of TERRA23 rG4 were plotted to highlight the dynamic regions in SERF2. d 15N relaxation rates R2/R1 demonstrating that a significant change in dynamics occurs for 200 μM SERF2 on its own (pink dots) and in the presence of 100 μM interacting TERRA23 rG4 (grey squares), as a function of residue number. Error bars represent the standard error from per-residue exponential fitting of peak intensities in NMRFAM-Sparky. e-f 2D analysis plots derived from analytical ultracentrifugation experiments for 4.7 μM TERRA12 rG4 without (e) or mixed with a 2-molar excess SERF2 (f). The partial concentration shown in color on the right y-axis represents the abundance of individual species in the sample solution. g-h Cartoon shows the top (g) and side-view (h) of SERF2 and TERRA rG4 complex. Note the quadrupole-like and planar interactions shown respectively, as ellipses in (g) and vertical slabs in (h). Residues generating the quadrupole-like interactions and distorting the TERRA rG4 structure (PDB ID: 2M18) are labeled, and hydrogen bonds are indicated with dashed lines. (i) EMSA gel-shift assay of 5 μM TERRA23 rG4 mixed with an equimolar amount of wild-type SERF2 and different lysine to alanine SERF2 mutants. Similar results were obtained in two independent repeated experiments.

Fig. 4|. All-atom MD simulation approach to study the structure of SERF2-TERRA rG4 phase separated condensate.

a A cubic all-atom MD simulation box encapsulating randomly distributed 30 molecules of SERF2 (orange), 30 molecules of the TERRA10 rG4 (PDB ID: 2M18, blue), molecules of PEG are shown in pink, Cl⁻ in green, and K⁺ in grey. b-c Surface representation of the structure of SERF2-TERRA rG4 ring-shaped droplet-like structure found in the condensed phase (b), and lower-ordered oligomers (c), that were obtained at time 0.5 μs in the MD simulation. The enlarged all-atom cartoon structures of the condensed droplet-phase and the lower-ordered 1:2 SERF2:TERRA rG4 dilute-phase oligomers are shown on the top. The three distinct contact sites in SERF2 in the lower-ordered complex structure (c, top) is shown and the TERRA rG4 interacting SERF2 residues are labeled. The high-resolution images in (b, top) show a representative interaction network that involves the three critical binding sites in SERF2 located in the disordered N-terminus. The SERF2 interacting residues are represented with an uppercase letter and TERRA rG4 nucleotide in a smaller-case letter (e.g. R11-U14 denotes Arg3 and uracil 14 in SERF2 and TERRA rG4, respectively). (d) Representative all-atom MD snapshot at 0.5 μs reveals the emergence of a droplet-like condensate featuring a ring-shaped assembly of SERF2 (orange) and TERRA rG4 (blue) complexes, stabilized in the presence of PEG crowding agents (magenta). (e) Quantitative analysis of PEG localization reveals time-dependent enrichment of PEG molecules within 6 Å of either SERF2 or TERRA rG4 over the course of the 0.5 μs simulation. (f) Time-resolved quantification of Hoogsteen hydrogen bonds within the TERRA rG4 structure confirms high structural integrity of the quadruplex core during the entire 0.5 μs trajectory. Color bar on right indicates the number of protein molecules simultaneously in contact with the rG4, demonstrating that even under high protein association, the G4 core remains structurally preserved. (g) Mean squared displacement or MSD analysis reveals diffusion of SERF2 in the condensates. The 3D diffusion constant (D) extracted from the linear regime of MSD (t) using equation $D=\frac{\text{MSD}\left(\text{t}\right)}{6\text{t}}$ is calculated to be 1.07 μm²/s.

Fig. 5|. Single-molecule imaging and tracking of SERF2 interactions with TERRA rG4s.

a Single-molecule fluorescence microscopy shows a single droplet of 50 μM SERF2 (spiked with picomolar Cy5 labeled SERF2) mixed with equimolar TERRA12 rG4s (left panel). Scale bar is 3.2 μm. Similar results were obtained in three independent repeated experiments. The corresponding single-particle tracking image of SERF2 molecules within the single droplet generated using ImageJ is shown on the right. B MSD plot of SERF2 (average MSD ≈ 0.95±0.23 μm²) obtained from a single particle tracking fluorescence microscopy experiment. The shaded area represents the standard error across tracked particles. Data are presented as mean values ± standard error. c Distribution of diffusion coefficients (D) extracted from tracking trajectories of individual SERF2-Cy5 molecules inside droplets shows a skewed and heterogeneous mobility profile centered around D ≈ 0.4–0.6 μm²/s, consistent with prior reports of protein diffusion in crowded, phase-separated compartments^,. d FRET distance histogram between donor and acceptor was determined based on the FRET efficiency in SERF2 (T2C and A51C) and TERRA12 rG4 phase-separated droplets containing picomolar Cy3-Cy5 labeled SERF2 in dilute-phase (left) and condensed-phase (right) samples separated by centrifugation (see methods).

Fig. 6|. SERF2 colocalizes with stress granules upon various stress conditions.

a Immunofluorescence images show that endogenous SERF2 is predominantly distributed in the nucleolus of fixed U2OS cells, as evidenced by staining with the nucleolar marker fibrillarin. b Colocalization analysis of fibrillin and SERF2 obtained from (a). Thirteen foci from nine cells in three biological replicates were subjected to Pearson’s correlation coefficient analysis. c SERF2 forms cytoplasmic foci and colocalizes with the core stress granule marker protein G3BP1 in different stress conditions. d This plot shows the quantification of stress granules retrieved from (c) under various stress conditions containing both SERF2 and G3BP1. At least twenty-six foci from three biological replicates were subjected to Pearson’s correlation coefficient analysis. e Fixed U2OS immunofluorescence cell images showing the DAPI stained nucleus (blue) and oxidative (sodium arsenite) stress-induced granules containing G4s(red), FUS (purple) and SERF2 (green) as detected by the BG4 antibody and the FUS and SERF2 proteins. f Plot showing SERF2 colocalization with FUS and G4sretrieved from (e) as measured by the Pearson’s coefficient. Twenty foci from three biological replicates were analyzed. The scale bars in (a, c, and e) indicate 10 μm.

Fig. 7|. SERF2 regulates stress granule formation and dynamics.

a Immunofluorescence of SERF2 and G3BP1 in fixed U2OS cells treated with 0.5 mM sodium arsenite for 1 hour. b Plot shows the percentage of stress granule positive cells after sodium arsenite treatment calculated from images shown in (a). Error bars were calculated from four independent experiments, and data are presented as mean values. c Live-cell imaging of EGFP-FUS HeLa Kyoto cells treated either with a control RNA (siCTRL) or an RNA targeting SERF (siSERF2), treated with different stressors (0.5 mM Sodium arsenite, 0.4 M Sorbitol, or 10 μM MG132) for 1 hour. Scale bars in (a and c) are 10 μm. d Plot shows percentage of stress granule positive cells under different stress treatments calculated from images shown in (c). At least 50 cells from three biological replicates were analyzed, and data are presented as mean values. **** shown in (b and d) indicates P < 0.0001. e-f These graphs show FRAP recovery curves in EGFP-FUS HeLa Kyoto live cells with siCTRL or siSERF2 conditions treated with 0.4 M sorbitol and (e) 0.5 mM sodium arsenite (f). Standard deviations were calculated by analyzing four different foci in three replicates subjected to FRAP, and data are presented as mean values.

Fig. 8|. SERF2 facilitates the phase transition of the stress granule core protein G3BP1 and enhances G3BP1 fluidity.

a DIC and fluorescence images showing co-phase separation of SERF2 (purple) and G3BP1 (unlabeled) with 12.5 ng/μL HeLa total RNA. Similar results were obtained in three independent repeated experiments. b DIC and fluorescence images show SERF2 facilitates G3BP1-RNA condensation in samples containing the indicated SERF2 and G3BP1 concentrations and 12.5 ng/μL HeLa total RNA. Similar results were obtained in three independent repeated experiments. Scale bars in (a and b) are 10 μm. c Phase diagram showing G3BP1 phase transition in a non-crowding condition and a crowding condition that contained 2.5% (w/v) PEG8000 in a 20 mM NaPi, pH 7.4, 100 mM KCl buffer. G3BP1 phase transition was measured in the absence or presence of SERF2 or rG4 or mixtures of SERF2 and rG4s as indicated. d Time-series images of 50 μM G3BP1 condensates after photobleaching in the absence or presence of SERF2, total RNA, and TERRA23 rG4s prepared in 20 mM NaPi, 100 mM KCl (pH 7.4) containing 5% (w/v) PEG8000. e FRAP recovery plots of G3BP1-Cy5 (left) and SERF2-AF488 (right) obtained from (d), the graph colors correspond to the plus and minus signs of the sample mixture shown in (d). Recovery halftime is calculated by non-linear regression fit and one-phase association using GraphPad prism. Scale bar is 5 μm. Standard deviations were calculated by analyzing 8 isolated droplets subjected to FRAP, and data are presented as mean values.