Phase separation and molecular ordering of the prion-like domain of the Arabidopsis thermosensory protein EARLY FLOWERING 3

Edited by Lucia C. Strader, Duke University, Durham, NC; received March 22, 2023; accepted June 6, 2023 by Editorial Board Member Joseph J. Kieber
July 3, 2023
120 (28) e2304714120

Significance

Phase separation of the protein EARLY FLOWERING 3 (ELF3) occurs as a function of temperature and is driven by its prion-like domain (PrLD). Here, we determine the dynamics and molecular basis for phase separation using complementary structural, biophysical, and microscopy techniques. We demonstrate that in the dilute phase, the PrLD of ELF3 forms a higher-order monodisperse oligomer, which is vital for phase separation. Changes in temperature trigger the formation of a condensed phase that shows increased molecular ordering. Subsequently, the liquid condensed state ages into a hydrogel, exhibiting molecular stacking based on small-angle X-ray scattering and atomic force and electron microscopy. These results resolve the structure of different physical states for a key plant thermosensor.

Abstract

Liquid–liquid phase separation (LLPS) is an important mechanism enabling the dynamic compartmentalization of macromolecules, including complex polymers such as proteins and nucleic acids, and occurs as a function of the physicochemical environment. In the model plant, Arabidopsis thaliana, LLPS by the protein EARLY FLOWERING3 (ELF3) occurs in a temperature-sensitive manner and controls thermoresponsive growth. ELF3 contains a largely unstructured prion-like domain (PrLD) that acts as a driver of LLPS in vivo and in vitro. The PrLD contains a poly-glutamine (polyQ) tract, whose length varies across natural Arabidopsis accessions. Here, we use a combination of biochemical, biophysical, and structural techniques to investigate the dilute and condensed phases of the ELF3 PrLD with varying polyQ lengths. We demonstrate that the dilute phase of the ELF3 PrLD forms a monodisperse higher-order oligomer that does not depend on the presence of the polyQ sequence. This species undergoes LLPS in a pH- and temperature-sensitive manner and the polyQ region of the protein tunes the initial stages of phase separation. The liquid phase rapidly undergoes aging and forms a hydrogel as shown by fluorescence and atomic force microscopies. Furthermore, we demonstrate that the hydrogel assumes a semiordered structure as determined by small-angle X-ray scattering, electron microscopy, and X-ray diffraction. These experiments demonstrate a rich structural landscape for a PrLD protein and provide a framework to describe the structural and biophysical properties of biomolecular condensates.
Compartmentalization into biomolecular condensates, or membraneless organelles, helps to regulate the biochemistry of the cell by dynamically concentrating and sequestering different components including proteins such as transcription factors, RNA-binding proteins and cofactors, and nucleic acids (19). Proteins with intrinsically disordered regions (IDRs) and low complexity prion-like domains (PrLD) often act as drivers of LLPS, separating into a highly concentrated protein-rich phase and a dilute phase under specific conditions (1012). While simple polymers have been successfully studied experimentally and modeled using theoretical methods such as course-grained (CG) simulation and atomistic models, quantifying and predicting the behavior of PrLD proteins as a function of physicochemical variables is challenging due to the complexity in the amino acid sequence of proteins (1319). The interactions that contribute to the metastable condensed phase include many transient, short-range interactions including pi–pi, cation–pi, dipole, electrostatic, and hydrophobic interactions, all of which may be present in a given polypeptide. These weak and low-specificity contacts are often present in disordered proteins and will occur intramolecularly in the dilute phase and both intra- and inter-molecularly in the condensed phase. The formation and dynamics of phase separation mediated by PrLD proteins are often highly sensitive to pH, ionic strength, and temperature and will vary as a function of the properties of the amino acids (i.e., polar, charged, hydrophobic, aromatic) in the PrLD sequence (2023). While this dynamic response of PrLD proteins is technically challenging to study, it may play a critical physiological function, allowing PrLD proteins to act as sensors of the in vivo cellular environment and to alter physiological responses accordingly. For example, the poly(A)-binding protein (Pab1) in yeast and the recently characterized circadian clock protein EARLY FLOWERING3 (ELF3) in Arabidopsis have been shown to act as direct in vivo temperature sensors and to alter developmental responses (24, 25). Thus, an understanding of the molecular basis of environmental sensing by PrLD proteins is a prerequisite to engineering these properties for the creation of tailored responses to stresses such as temperature changes in the cell.
ELF3 displays a well-characterized ability to undergo LLPS in vitro and in vivo and exists with natural sequence variation within the PrLD associated with specific phenotypes, suggesting a physiological role for LLPS (9, 26). ELF3 is a largely disordered protein with a C-terminal PrLD. The PrLD contains a poly-glutamine repeat (polyQ) that exhibits different lengths from 7 to 29 glutamines across 181 natural Arabidopsis accessions (26, 27). Previous experiments in Arabidopsis plants grown at 17 °C, 22 °C, and 27 °C demonstrated that the length of the polyQ has a mild but statistically significant effect on hypocotyl elongation, a commonly used measure of thermoresponsive growth (24). While ELF3 PrLD with seven glutamines (Q7), from the lab strain Columbia-0, has been shown to phase separate in vitro in a temperature-dependent manner, the effects of varying polyQ length on the dynamics of phase separation are not known (24). In order to better understand the biophysical basis of ELF3 condensation, we verified that ELF3 Q0, Q7, and Q20 could form puncta in vivo and performed biochemical, structural, and biomechanical studies of the corresponding PrLD regions, demonstrating that LLPS of ELF3 PrLD with varying polyQ lengths is sensitive to temperature and pH. We further show that the dilute phase, unlike a canonical intrinsically disordered protein, exists as a monodisperse higher-order oligomer and that upon liquid–liquid phase separation forms a new species with distinct microenvironments and different biomechanical properties. This species is able to further undergo aging into an ordered hydrogel. ELF3 PrLD exhibits a distinct structural landscape, with the dynamics and biomechanical properties of the condensed liquid and gel phases modulated by the length of the polyQ region.

Results

In Vivo Puncta Formation of ELF3 Q0, Q7, and Q20.

Based on our previous studies of the full-length and PrLD of ELF3, the protein undergoes LLPS in vivo and in vitro with the PrLD required and sufficient for phase separation (24). To verify formation and dynamics of condensate formation in vivo, the sequence of the ELF3 gene encoding polyQ lengths of 0 (deletion mutant), 7 (from accession Columbia-0), and 20 [from accession Sandåkra-2 (San-2)] glutamines tagged with mVenus was expressed under an inducible promoter in tobacco leaf epidermal cells. The formation of puncta was observed for all three constructs (Fig. 1). Fluorescence recovery after photobleaching (FRAP) experiments were performed to investigate the dynamics of the proteins in the puncta. The puncta for all constructs exhibited partial recovery after photobleaching, with fluorescence intensities after photobleaching ranging from about 50% for Q0 and Q7 (Fig. 1 A and B) to 30% for Q20 (Fig. 1C) of the prebleach intensity. This indicates that the puncta are composed of a mixture of mobile and immobile species, as has been observed for many other systems (28). The identification of the different species present in condensates, in particular the immobile species, is currently very limited. In order to more robustly characterize the dynamics and structure of the condensate, detailed in vitro studies of the PrLD region of the protein required for phase separation were performed.
Fig. 1.
Representative fluorescence recovery after photobleaching (FRAP) images of agroinfiltrated Nicotiana benthamiana leaf epidermal cells transiently expressing mVenus-ELF3. (AC) mVenus-ELF3 Q0 (A), Q7 (B), or Q20 (C), white boxes indicate photobleached areas. Panels at left are prior to bleaching, middle immediately after bleaching, and right after recovery. Mean recovery curves for ELF3 Q0, Q7, and Q20 are shown at far right and S.D. are shown. All curves are generated from 9 to 10 individual puncta measurements with each FRAP experiment performed on a different cell.

In Vitro Characterization of the ELF3 PrLD.

Using optical imaging, turbidity assays (A440), and phase diagrams, we examined the effects of temperature, protein concentration, and pH on LLPS of ELF3 PrLD with polyQ lengths of 0, 7, and 20 glutamines (Fig. 2 and SI Appendix, Fig. S1A). All PrLD constructs underwent LLPS; however temperature, and pH had subtle but measurable effects on phase separation for the three constructs studied. Increasing temperature triggered LLPS for Q0, Q7, and Q20 constructs and exhibited reversibility as shown qualitatively (Fig. 2A) using temperature steps. Absorbance measurements at A440 were used to further examine this behavior, with Q0 and Q7 behaving similarly, exhibiting an LLPS transition temperature occurring at 31.2 ± 0.4 °C and 28.1 ± 0.3 °C, respectively (Fig. 2B). In comparison, the Q20 construct started phase separation at a lower temperature, as shown by the nonzero normalized absorbance between 15 °C and 20 °C and a more sloping bottom plateau as compared to Q0 and Q7. A440 gradually increased for Q20 as the temperature was raised with a Tm of 33.0 ± 1.5 °C (Fig. 2B). The Tm values were calculated based on the first temperature ramps, as denoted by the filled blue squares in Fig. 2B. The effects of temperature were largely reversible for Q0, Q7, and Q20 constructs tested, with the three PrLD proteins switching between the dilute and condensed phase as monitored by A440 (Fig. 2 A and B). It should be noted that, due to the experimental setup for the A440 measurements, where the highest temperature measured was 40 °C, the parameters extracted from the fitting, in particular for Q20, should be considered as estimations as no plateau was reached and above 40 °C the proteins began to irreversibly precipitate. The observed demixing of the ELF3 PrLD proteins with increasing temperature is indicative of hydrophobic and aromatic interactions driving phase separation due to the favorable enthalpic contributions to the free energy of the system under the temperature regime of interest and positive entropic effects due to counter ion and/or hydration water release, as has been observed for other LLPS systems (20, 2931). All constructs underwent phase separation under the conditions tested and expansion of the polyQ repeat from 0 to 20 glutamines had a relatively small effect on the transition temperature onset of phase separation.
Fig. 2.
ELF3 PrLD with varying polyQ lengths undergoes phase separation. (A) Reversible temperature induced phase separation for ELF PrLD Q0, Q7, and Q20 (4 mg/mL) visualized at 4 °C (Left), warmed to 22 °C (Center) and cooled to 4 °C (Right). (B) Turbidity assays (A440) of untagged ELF3 PrLD Q0, Q7, and Q20 as a function of temperature. Protein concentration was 15 μM (~0.4 mg/mL) for each construct. Temperature ramps are shown with red arrows indicating increasing or decreasing temperature. (C) Phase diagrams for untagged ELF3 PrLD Q0, Q7, and Q20 constructs as a function of protein concentration and pH. Open circles are dilute phase, dark blue indicates LLPS, and gray indicates precipitate/gel. LLPS conditions are shaded in light gray for clarity. Circles represent measurements at a specific pH (9.5, 9.0, 8.8, 8.6, 8.4, 8.2, 8.0, 7.8, 7.6, and 7.4) and a specific protein concentration (0.5, 1.0, 2.0, 4.0, 6.0 mg/mL). (D) Fluorescence recovery after photobleaching (FRAP) experiments, with samples at pH 7.8. Red arrows indicate regions that were photobleached and scale bars are shown. Recovery panels displayed are 3 min after photobleaching for all samples. Samples exhibited little fluorescence recovery due to formation of an immobile gel phase.
In addition to temperature variables, pH changes also affected phase separation, with the longer polyQ construct (Q20) forming spherical droplets over a wider range of protein concentrations and pH (pH 8.0 to 9.0) than the Q0 and Q7 constructs, which only underwent LLPS over a narrow pH range of 8.5 to 9.0 under the same ionic strength buffer conditions (Fig. 2C). Fusing a C-terminal GFP to the PrLD constructs resulted in the formation of spherical droplets with the same overall trend observed for the unfused proteins, with the Q20 construct undergoing phase separation at a pH range of 6.5 to 8.5 versus the Q7 and Q0 constructs that behaved in a similar manner, with phase separation occurring between pH 7 and 8.5 (SI Appendix, Fig. S1B). Extending the polyQ region results in a broader range of LLPS formation with respect to pH and temperature. This may be due to the relatively short polyQ tracts for Q0 and Q7, with longer polar glutamine stretches acting to prevent precipitation and keeping the protein in the condensed droplet phase over a wider range of pH and temperatures. Overall, polyQ variation had a relatively small effect on the initial stages of phase separation (i.e., onset of LLPS as a function of temperature and pH), but phase separation did not depend on the presence of polyglutamines.

Fluorescence Recovery after Photobleaching In Vitro.

While pH and temperature affected LLPS, we questioned whether there were inherent differences with respect to dynamics or stability of the condensed phase due to varying polyQ length. Fluorescence recovery after photobleaching (FRAP) experiments were performed on Q0, Q7, and Q20 PrLD constructs fused to GFP to investigate this possibility. All proteins formed droplets as the pH of the solution was reduced stepwise (0.2 pH units) from 9.4 to 8.4 and 40 min of equilibration after each pH change. These experiments revealed the formation of a low-mobility condensate with negligible fluorescence recovery over minutes (Fig. 2D). No clear differences were observed for the different samples, with all samples exhibiting formation of gel-like droplets. This suggests that while temperature change results in more reversible liquid–liquid phase separation, at least under fast temperature ramps, pH changes favored gel formation after initial condensation. FRAP experiments demonstrate that spherical condensates have different proportions of mobile and immobile species and this is highly dependent on preparation method and different variables including temperature and pH.

Biomechanical Measurements by AFM.

To further characterize the condensed phase and determine whether there were differences in the biomechanical properties of the samples, atomic force microscopy (AFM) and AFM-coupled confocal microscopy (32) experiments were performed (Fig. 3 and SI Appendix, Figs. S2 and S3). All samples were applied to a glass coverslip and imaged in liquid for AFM topography and force curve measurements. Lower pH (~7.6) and longer adhesion times resulted in the formation of the gel state, whereas higher pH (~8.2) and shorter adhesion times maintained the droplets in a more liquid-like state and droplets were considered a “highly viscous liquid” if they exhibited fusion with other droplets and at least 10% fluorescence recovery over 2 min. For these samples, in agreement with models describing nanoscale indentation of liquid interfaces (33, 34), the stiffness was calculated based on a linear fit to the force vs. distance curves with 9 to 13 measurements taken from the center of each droplet to avoid edge artifacts, which are present when the droplet size is comparable to the size of the AFM tip height (≈3.5 to 7 µm). Such a force versus indentation linear regime can directly be related to the liquid–liquid interfacial tension (34). The droplets for all samples exhibited variable mechanical properties, suggesting the presence of harder and softer regions within individual droplets, with measurements ranging from 3 to 9 mN/m (Table 1 and Fig. 3). This is likely due to liquid-hydrogel phases coexisting in the droplet and the more hydrogel present, the greater the deviation from linearity of the force as a function of distance. In contrast to the liquid condensed phase, the gel state exhibited no fluorescence recovery, no fusion, and a nonlinear force versus indentation regime that could be fit using a Hertz model (applicable to solid samples) and the Young’s modulus calculated (see comparison between indentation cycles performed on liquid-like and hydrogel droplets in SI Appendix, Fig. S2C, where deviation from linearity in the liquid phase curve can also be due to the presence of an out of contact Debye’s screening length region) (34). Interestingly, gel transition of the samples only partially affected overall droplet morphology since all droplets retained a quasispherical shape (Fig. 4); however, indentation cycles showed a nonlinear force vs. distance regime as opposed to the linear force vs. distance pattern observed for the highly viscous liquid droplet samples. In addition, the mechanical response changed for the hydrogel samples. For all hydrogel samples, the stiffness ranged between 5 and 100 kPa, with ELF3 PrLD Q0 showing a consistently lower Young’s modulus than the Q7 and Q20 constructs (Table 1). It should be noted that constructs in Fig. 4A exhibit higher rigidity than the values reported in Table 1 because they were imaged with a higher AFM loading rate (see Experimental Methods), required to decrease the acquisition time while providing sufficient pixels for high resolution, thus the apparent “stiffer” response at higher loading rate is due to viscoelastic behavior. Wide-field fluorescence and high-resolution AFM imaging of the samples further revealed a heterogeneity within the droplets, suggesting the formation of discrete microenvironments (Fig. 4). The microenvironments exhibited a stacked or layered structure, with varying step-like height profiles ranging from ~20 nm to 200 nm (Fig. 4D and SI Appendix, Fig. S3). These experiments represent the first examples, to our knowledge, of the use of AFM to determine the biomechanical properties of macromolecular condensates in the liquid phase, allowing us to probe the molecular organization of molecules within individual droplets. As sphericity is often one “characteristic” indicator of a liquid phase, these results reveal that dynamics and liquidity may be very low, with gelation occurring without loss of a spherical shape. In addition, within the same spherical droplet, multiple phases—liquid and gel—can coexist allowing the creation of distinct subenvironments within a condensate.
Fig. 3.
Coupled AFM and confocal fluorescence microscopy of ELF3 PrLD-GFP constructs. (AC) AFM topography mapping is shown at left, fluorescence of the same sample is shown at middle and stiffness measurements are shown at right for ELF3 Q0 (A), Q7 (B), and Q20 (C). Alignment between the AFM and confocal microscope was within 1 to 2 microns for the two microscopes. (Scale bar, 5 µm.) The color scales are shown below each panel. Q20 droplets were smaller and the scan size was adjusted accordingly to maintain lateral resolution.
Table 1.
Biomechanical properties of ELF3 PrLD. AFM measurements of condensate stiffness for liquid and gel phases*
 ELF3-GFP Q0ELF3-GFP Q7ELF3-GFP Q20
Liquid Phase (stiffness keff due to liquid interfacial tension)7.5 ± 1.4 mN/m4.6 ± 0.5 mN/m5.8 ± 1.9 mN/m
Gel Phase (stiffness- Elastic modulus in the linear elastic deformation range- Young’s modulus)7.4 ± 2.3 kPa102 ± 11 kPa59 ± 51 kPa
*All measurements are the average of n > 10.
Fig. 4.
Wide-field total internal reflection fluorescence microscopy and high-resolution AFM of ELF3-GFP PrLD gel droplets. (A) Young’s modulus measurements of the droplets showing inhomogeneous stiffness within individual droplets, (Scale bar, 1 µm.) (B) Fluorescence images of ELF3 PrLD Q0 (Left), Q7 (Middle), and Q20 (Right), (Scale bar, 1 µm.) Areas of interest are boxed and shown in close-up in C. (C) AFM topography of the droplets in A, (Scale bar, 1 µm.) (D) Height profile measurements across individual droplets (blue lines in C) showing a step pattern between flat layers.

SAXS of Dilute and Condensed Phases.

In order to further investigate the transition from the dilute to condensed phase, condensed phase dynamics, and the potential structuration and ordering of microenvironments within the condensate, we performed a series of small-angle X-ray scattering (SAXS) experiments (SI Appendix, Table S1 and Fig. 5). Scattering curves were determined for each sample (ELF3 PrLD Q0, Q7 and Q20) in the dilute phase using an online HPLC purification step in order to remove any aggregates (Fig. 5A). All proteins eluted as symmetrical peaks after the void volume of the column. In the dilute phase with monodisperse samples, there should be little or no contribution to the scattering from other species such as aggregates or the condensed phase, and the scattering curve will then reflect the form factor, P(q), providing information on the size and shape of the individual molecules in solution. Based on polymer theory, the conformation the molecules adopt in the dilute phase can be predictive of their phase behavior, providing a technically simple way to gain insight into the condensed phase (35). Calculating the radius of gyration (Rg) based on fitting the linear portion of the Guinier region yields values between 72 and 76 Å for the three samples (ELF3 PrLD Q0, Q7, and Q20) and the maximum dimension, Dmax, values of 272, 273, and 294 Å, respectively. Surprisingly, based on these measurements, the estimated molecular weights, even in the dilute phase, of the molecular species correspond to a multimeric ~28 to 30-mer assembly (SI Appendix, Table S1). This higher-order oligomeric state was further confirmed by size-exclusion chromatography multiangle laser light scattering (SI Appendix, Fig. S7). As expected, the largest Rg and Dmax values were calculated for ELF3 PrLD Q20 due to its increased length from expansion of the polyQ region (Table 2). Furthermore, these dilute phase measurements demonstrate that for all PrLDs, the species in solution is relatively globular, as shown in the Kratky plot (Fig. 5B). The compact spherical structure of a monodisperse higher-order oligomer, in contrast to an extended conformation often observed for intrinsically disordered proteins, coupled with the lack of predicted secondary or tertiary structure, suggests that the presence of weak multivalent interactions may be sufficient to exclude solvent and that hydrophobic amino acid sequences lacking predicted secondary structure, partially organize the protein molecules, likely a prerequisite for phase separation by ELF3 (36). As liquid–liquid phase separation requires multivalent interactions, the dilute phase organization may recapitulate these types of interactions intramolecularly or as smaller oligomeric units, with proteins predicted to phase separate potentially exhibiting a more compact structure versus an extended conformation from nonphase separating disordered proteins (36).
Fig. 5.
ELF3 PrLD in the dilute and condensed phase. (A) SAXS scattering curve for ELF3 PrLD Q0 (green), Q7 (pink), and Q20 (dark blue) in the dilute phase measured with online HPLC purification. Inset shows a schematic representation of the dilute and condensed phases as a function of temperature. (B) Corresponding Kratky plots for the curves shown in A The Gaussian shape denotes a globular species in solution. (C) SAXS scattering curves for ELF3 PrLD Q0 (green), Q7 (pink), and Q20 (dark blue) samples in the condensed phase after aging. The structure factor peak due to long-range ordering of molecules within the condensed phase is indicated by an arrow. (D) Corresponding Kratky plots for condensed phase scattering curves are shown and colour coded as above. The shape of the Kratky plot indicates a more extended and less globular species with the structure factor peak showing more prominently. SAXS scattering curves in A and C show a vertical offset by a factor of 10 for ELF3 PrLD Q7 and a factor of 100 for Q0 to aid in viewing.
Table 2.
Radius of gyration for dilute and condensed ELF3 PrLDs at room temperature
 RgQ0 (Å)RgQ7 (Å)RgQ20 (Å)
Dilute Phase72.473.475.6
Condensed Phase118128134
LLPS of the samples was triggered by a temperature increase and resulted in changes to the scattering curves (Fig. 5C and SI Appendix, Fig. S5). In the liquid–liquid separated phase, the SAXS measurements demonstrate an increase in the Rg and a more extended conformation based on the Kratky plot calculations (Table 2 and Fig. 5D). Changes in Rg in the condensed state have been previously observed for proteins undergoing LLPS and ascribed to nucleation events (36). The LLPS samples are more heterogeneous than the online HPLC purified samples in the dilute phase, as reflected in the upturn of the scattering curve as I(q) approaches 0 (Fig. 5C), but exhibit a clear trend in Rg, with an increase in the Rg in the condensed phase, suggesting possible elongation of the polypeptide chains or an increase in the overall size. In addition to these changes in the Rg, a structure factor peak, S(q), formed in the low q region of the scattering curves that increased in height with increasing temperature (Fig. 5C and SI Appendix, Fig. S5). This S(q) (structure factor) peak is indicative of loose lamellae within the condensed phase and a calculated spacing (dhkl) of approximately 155, 163 and 167 Å for ELF3 PrLD Q0, Q7, and Q20, respectively, similar to the calculated layer spacing observed in AFM experiments. It should be noted that the layered arrangement is distinct from the characteristic fibril formation observed for amyloids, (37) but a less compact lamellar phase characteristically observed for lipid membranes and polymer lamellae (38, 39). To further investigate the temperature effects to the condensed phase, temperature ramps ELF3 PrLD Q7 and Q20 exhibited an increase in the structure factor peak as the temperature was increased, whereas ELF3 PrLD Q0 exhibited little change, suggesting it had passed through the liquid phase and had undergone aging to the more stable hydrogel, losing the capacity to easily transition between the condensed and dilute phases (SI Appendix, Fig. S5). These results demonstrate that in solution and prior to condensation, the ELF3 PrLD is a homogenous globular oligomeric ~30-mer species. An increase in temperature triggers condensation and a long-range ordering of a portion of the ELF3 PrLD oligomers, which adopt a stacked arrangement within the condensate. Some elongation of the globular species in the condensed phase occurs based on an increase in the radius of gyration, but oligomerization is likely maintained as the stacking unit. As the amount of stacking increases over time, the reversibility between the dilute and condensed phase decreases due to the more extensive intermolecular interactions.

Structural Characterization of the Gel Phase.

The ordered species in the condensate were further characterized using stable gel samples, obtained by aging of the condensed liquid phase at pH 7.7 to 8 for ~1 h at 4 °C for Q0, Q7, and Q20. These samples no longer exhibited reversibility between dilute and condensed phases with temperature or pH change based on incubation of the samples at lower temperatures and dilution with high pH buffer and coalesced into large nonspherical amorphous species. Gel samples were characterized using transmission electron microscopy (TEM) and X-ray diffraction. Negative stain TEM confirmed regions with a stacked structure in the gel phase for all polyQ constructs, with a stack spacing estimated at ~40 to 50 Å (Fig. 6A). This stack spacing is smaller than observed for AFM and SAXS experiments and may be due to the longer aging of the sample and/or dehydration effects of sample preparation; however, a highly ordered organization of molecules is observed. The TEM experiments exhibit inhomogeneity in the samples with certain regions exhibiting well-ordered molecular stacks and other regions showing no clear organization.
Fig. 6.
ELF3 PrLD exhibits stacked molecular organization in the condensed phase. (A) Negative stain TEM images of ELF3 PrLD Q0, Q7, and Q20 showing a layered structure in the gel condensed phase. Layered structure can be clearly observed in magnified images of the boxed regions. (B) X-ray diffraction image for ELF3 PrLD Q20 showing a diffuse powder ring at 4 to 5 Å. A close-up is shown with the diffraction ring indicated by the black arrow.
X-ray diffraction experiments for the ELF3 PrLD Q20 construct exhibited a weak powder diffraction ring at 4 to 5 Å (Fig. 6B), similar to observations for hydrogel-forming FUS constructs, denoting a long-range molecular ordering with a powder diffraction ring arising from the distance between sheets of polypeptide chains (40). The ELF3 PrLD Q0 and Q7 constructs showed very diffuse scattering in the 4 to 5 Å range (SI Appendix, Fig. S6), due to either diffraction measurements from an amorphous region exhibiting less molecular stacking or indicating an overall less well-ordered hydrogel for the ELF3 PrLD Q0 and Q7 constructs. Electron microscopy and X-ray diffraction experiments provide complementary structural information of the molecular and atomic ordering in the gel phase. Stack spacing between layers of molecules is observed in the negative stain electron micrographs. X-ray powder diffraction rings indicate long-range ordering due to ELF3 PrLD molecules aligning and interacting within each layer.

Discussion

Proteins that drive LLPS are under active study not only due to their key biological role in cellular compartmentalization but also for diverse applications in biomaterials and drug delivery (20, 4145). Characterization of the dilute and condensed phases and the effects of exogenous variables including temperature and pH is experimentally challenging due to the complexity of the protein polymer and the intra- and inter-molecular interactions at play during the nucleation and condensation process. Here, we demonstrate the utility of complementary biophysical and structural techniques to determine the dynamics and structure of different phases of a PrLD protein as a function of pH and temperature. Phase diagrams, A440 assays, and SAXS measurements demonstrate that the dilute phase is a higher-order oligomer and formation of the condensed phase is triggered by pH changes and/or an increase in temperature. While phase separation of the ELF3 PrLD does not require the first polyQ region, variation in polyQ length, which occurs in natural Arabidopsis accessions, tunes the onset and range of LLPS as a function of temperature and pH. Different ELF3 proteins from accessions studied to date act as thermosensors in vivo, with phase separation triggered by an increase in temperature (9). This behavior, as demonstrated here, is recapitulated in vitro. The pH sensitivity observed in vitro may also relate to the thermosensory activity of the protein, although this is speculative. As the temperature of a solution increases, the pH will decrease and there is evidence that this occurs intracellularly, albeit with proton pumps acting to buffer these pH changes (46, 47). A decrease in pH favors LLPS of ELF3 PrLD, for both the untagged and GFP fused PrLD and this may be true for ELF3 in complex with different protein partners, whose activity will vary due to transition from a mobile dilute phase, to a concentrated liquid phase and eventually to an “aged” low-mobility phase.
During the aging process, an ordered gel phase quickly forms in vitro, with AFM, SAXS, EM, and X-ray diffraction experiments of the gel phase all consistent with a stacked/layered structure in the biomolecular condensate and extensive ordering. Hydrogel formation likely occurs in vivo, as puncta do not exhibit full fluorescence recovery, particularly for the ELF3 PrLD Q20. While aging of liquid condensates into low-mobility gel phases in vitro and in vivo is common, its functional significance is unclear as are its effects on cargo proteins (48). It should be noted that ELF3 acts as a scaffold, interacting with many different protein partners including PHYTOCHROME INTERACTING FACTORS (PIFs), phytochromes, the COP1-SPA complex and the circadian clock proteins, TIMING OF CAB EXPRESSION, ELF4, GIGANTEA and LUX (49). These interactions change the localization of ELF3 to different condensates (i.e., ELF3 nuclear bodies, photobodies) and it is likely that liquid versus gel formation will be context dependent and tuned by the presence of different protein and/or nucleic acid constituents and their relative concentrations within the condensate. Indeed, photobodies, which contain ELF3, are highly dynamic condensates (50). The biological roles of these different phases dilute, highly dynamic liquids and low-mobility gels are not well understood and the formation of a gel matrix by one constituent does not preclude the presence of mobile macromolecules within the condensate. Thus, the physiological function of these different phases is a central and open question requiring further investigation. Testing the in vivo role of dilute, liquid phase, and gel condensates requires generating mutants which decouple oligomerization and phase separation from other protein functions such as complex formation and localization. Based on these studies, however, mutations in the ELF3 PrLD targeting regions putatively important for oligomerization, LLPS, and molecular ordering will allow us to address this question. The detailed structural and biophysical studies of the surprisingly higher-order oligomeric state in the dilute phase and the demixing properties of the protein combined with quantitative measurements provide a foundation for altering these properties and, in the longer term, determining the physiological function of different dilute and condensed phases.
The varying conformations and dynamics of the ELF3 PrLD polypeptide–from a surprisingly large and monodisperse oligomeric species in the dilute phase to a solvent-excluded liquid condensed phase and finally to a highly ordered hydrogel–demonstrate how the intrinsically disordered polypeptide is able to access different regions of structure space as a function of the physicochemical environment. The low-affinity and transient interactions of the ELF3 PrLD, consisting of both intra- and inter-molecular interactions, may facilitate the priming of the molecules for phase separation and condensation via the formation of a large oligomeric species in the dilute phase (51, 52). While the ELF3 PrLD has little predicted secondary structure, it is still able to form a defined and homogeneous ~28 to 30 oligomeric assembly in solution. Based on the SAXS studies, this species is compact and globular, likely serving as an intermediate for phase separation which would occur via the fusion of oligomeric spheres. Whether this state is specific to the ELF3 PrLD or more common to PrLD phase separating proteins is unknown and will require further study.
Examination of the primary amino acid sequence of ELF3 PrLD demonstrates a number of proline clusters and aromatic residues that may act as “stickers” for LLPS and/or oligomerization, although the specific amino acids driving this remain to be determined. In silico predictions suggests the PrLD overall has a high propensity for phase separation, with no obvious short motifs or specific amino acids, such as polyQ repeats or aromatic residues directing LLPS (5356). However, in other systems that undergo LLPS and fibril or hydrogel formation, various motifs important for steric zippers, such as those in amyloid fibrils, and kinked beta sheets have been identified. Studies of LLPS in the protein FUS, for example, reveal the presence of specific peptide motifs that act as drivers for phase separation and hydrogel formation. These low-complexity, aromatic-rich, kinked segments (LARKs) have been shown to self-assemble into stacked structures and protofilaments, with short LARK hexapeptides able to form kinked beta sheets that are less stable than beta amyloid fibrils but still exhibit long-range ordering (40, 57). The ELF3 PrLD does not possess canonical LARK peptide sequences, and increasing temperature generally disrupts LARK-type peptide interactions, in contrast to increasing temperature triggering LLPS for ELF3 PrLD, suggesting that other mechanisms are critical for the observed self-organizing behavior of the protein that likely depend on specific hydrophobic interactions, which we are currently investigating (20).
While focused on a specific PrLD, these studies provide a framework for the in vitro characterization of condensate-forming biomolecules using complementary biophysical and structural techniques. The in vivo implications and prevalence of long-range molecule ordering of PrLD-containing proteins such as ELF3 are still to be fully explored. This may be a much more general phenomenon for other condensate-forming species. For example, the formation of discrete heterogeneous assemblies, or pleiomorphic ensembles, within the same droplet has been observed in vivo for scaffold proteins and signaling complexes, such as the PDGF receptor complex and the Wnt signalosome, allowing for the cooperative assembly of different components while retaining some conformational flexibility of the components (58, 59). Stacked structures in condensates have been observed for other IDR-containing proteins such as FLOE1, a plant water-sensory protein important for seed germination. Substitution of tyrosine and phenylalanine for serine residues in the aspartate-serine-rich (DS) domain of FLOE1 resulted in gel formation of the mutant in vivo with ordered structures observed by TEM and 3-D tomography (60). Phase separation and gelation of FLOE1 were driven by aromatic residues in the glutamine/proline/serine-rich (QPS) domain, in particular tyrosine residues.
Taken together, these experiments have allowed a detailed in vitro characterization of ELF3 PrLD condensates using complementary biophysical and structural techniques that have not been previously applied to biological molecules in dilute and condensed phases. The molecular ordering observed in vitro for ELF3 PrLD may be a much more general phenomenon than previously appreciated and may help explain the variability in fluorescence recovery after photobleaching observed for many membraneless organelles (5, 51, 53). An important and remaining challenge is to determine the physiological role of condensed species with different mobilities and whether long-range order and molecular stacking have a physiological role in ELF3 function. This will require further study including the reengineering of the PrLD of ELF3 to alter LLPS and an examination of the effects in vivo at the cellular and phenotypic level.

Materials and Methods

Protein Expression and Purification of ELF3PrLD and ELF3PrLD-GFP.

ELF3 PrLD (Q7, residues 388-625, AT2G25930, Arabidopsis thaliana ecotype Columbia), Q0, Q20, ELF3 PrLD-GFP Q7, Q0, and Q20 were cloned into the expression vector pESPRIT2 using the Aat II and Not I sites as previously described (9). To generate the GFP constructs, the stop codon for ELF3 PrLD Q0, Q7, and Q20 was removed by site-directed mutagenesis using the QuikChange protocol (Agilent) and a GFP-tag was added to the C-terminus as previously described (45). All proteins were overexpressed in Escherichia coli BL21-CodonPlus-RIL cells (Agilent). Proteins used for phase diagrams, FRAP, AFM, SAXS, EM, and X-ray diffraction experiments were expressed and purified as follows. Briefly, cells were grown in LB media supplemented with 50 μg mL−1 kanamycin and 35 μg mL−1 chloramphenicol at 37 °C and 120 rpm. At OD600nm = 0.8, the temperature was lowered to 18 °C and protein expression was induced with 1 mM isopropyl-b-D-1-thiogalactopyranoside (IPTG). After 16 h, the cells were harvested by centrifugation at 5000 × g and 4 °C for 15 min. For all constructs, the cells were resuspended in 100 mM Bis-Tris propane pH 9.4, 300 mM NaCl, 20 mM imidazole, 1 mM TCEP (tris(2-carboxyethyl)phosphine (TCEP), and EDTA-free protease inhibitors (ThermoFisher). Cells were lysed by sonication, and the cell debris was pelleted at 50000 × g and 4 °C for 30 min. The supernatant was applied to a 1 mL Ni-NTA column preequilibrated in resuspension buffer (lysis buffer without protease inhibitors), washed with resuspension buffer, and high salt buffer (100 mM Bis-Tris propane pH 9.4, 1 M NaCl, 20 mM imidazole, 1 mM TCEP). The proteins were eluted in 100 mM Bis-Tris-propane pH 9.4, 300 mM NaCl, 300 mM imidazole, and 1 mM TCEP. Protein purity was determined by SDS-PAGE, and the fractions of interest were pooled and dialyzed for ~2 h at 4 °C in 50 mM Bis-Tris-propane pH 9.4, 500 mM NaCl, and 1 mM TCEP. The final protein concentration was 4 to 8 mg/mL for all samples. Bis-Tris-propane was used due to its wide buffering pH range. For the turbidity assays the Bis-Tris-propane was replaced with CAPS (N-cyclohexyl-3-aminopropanesulfonic acid) buffer, as the proteins were stable and soluble at this pH at protein concentrations. Purification was performed as for above, with the Bis-tris propane buffer replaced by CAPS pH 9.7. Final dialysis was performed against 50 mM CAPS pH 9.7, 200 mM NaCl, 1 mM TCEP and proteins diluted to ~0.4 mg/mL for turbidity assays.

Transient Expression in Nicotiana benthamiana for FRAP.

The coding sequence of ELF3Q0, ELF3Q7, and ELF3Q20 were cloned into the GreenGate cloning system by digestion and ligation. For mVenus-tagged constructs, GreenGate reactions with a β-estradiol inducible promoter were performed as previously described (61). The Agrobacterium tumefaciens strain harboring the pSOUP helper plasmid was transformed with the above-described plasmids. mVenus-tagged fusion proteins were coexpressed with the p19 silencing repressor in epidermal leaf cells of N. benthamiana. Induction with β-estradiol was carried out as described (62).

In Vivo FRAP Measurements and Analysis of ELF3 Condensates.

FRAP measurements were carried out with a confocal laser scanning microscope (inverted LSM880, ZEISS) equipped with a 40× water objective (C-Apochromat, NA 1.2, ZEISS). mVenus was excited with an argon laser at 514 nm. For FRAP measurements, single condensates in or around the nucleus were chosen. Time series of 65 frames were acquired to observe fluorescence recovery. Acquisition times were approximately 300 ms per frame. Fluorescence recovery was analyzed individually for each bleached condensate by drawing narrow ROIs (63). For normalization, calculation of mobile fraction and half-life (τ½) a plugin written at the Stowers Institute for medical research (https://research.stowers.org/imagejplugins/index.html) was used in Fiji. Then, 9 to 10 individual puncta were photobleached from different cells, and used in each analysis. FRAP curves were generated from all data and mean curves with SD are shown. Experiments were performed at room temperature.

Light Scattering Turbidity Assays.

The light scattering assay was performed in a Cary 100 UV–vis spectrometer (Agilent Technologies UK Ltd., Stockport, UK) as previously reported(24). Briefly, the absorbance at 440 nm was monitored for samples containing buffer alone [50 mM N-cyclohexyl-3-aminopropanesulfonic acid (CAPS) pH 9.7, 200 mM NaCl, 1 mM TCEP], ELF3 PrLD Q0, Q7 or Q20 (15 µM, ~ 0.4 mg/mL) in quartz cuvettes (path length 10 mm) with increasing temperature (10 to 40 °C; 1 °C min−1), and the spectra were normalized with respect to buffer alone. A transition temperature (Tm) was determined by fitting the spectrum with a four-parameter sigmoidal equation using GraphPad Prism 9.4.0 and the sigmoidal 4PL equation, which is defined as y = A + (B-A)/(1+(Tm/x))C, where y is the normalized turbidity at 440 nm, x is the temperature in °C, A and B are bottom and top plateaus with the same units as y, C is the slope factor or Hill slope, and Tm is the midpoint transition temperature. Reported values are from the curves shown in Fig. 2A where the filled blue squares for the first temperature increase from 10 to 40 °C. Tm values were 31.2 ± 0.4 °C (ELF3 PrLD Q0), 28.1 ± 0.3 °C (ELF3 PrLD Q7), and 33.0 ± 1.5 °C (ELF3 PrLD Q20). Due to the experimental setup, where the highest temperature measured was 40 °C, the parameters extracted from the fitting, in particular for ELF3 PrLD Q20 should be considered as estimations as no plateau was reached, and above 40 °C, the proteins began to irreversibly precipitate. To monitor reversibility, the turbidity was monitored for an increasing temperature ramp (10 to 40 °C; 1 °C min−1) followed by decreasing the temperature (40 to 10 °C; 1 °C min−1) and this cycle was repeated twice in total for ELF3 PrLD Q0, Q7, and Q20.

Phase Diagrams.

To generate phase diagrams for ELF3 PrLD Q0, Q7, and Q20 and ELF3 PrLD Q0-, Q7-, Q20-GFP, the pH of the dialysis buffer (50 mM Bis-Tris propane pH 9.4, 500 mM NaCl, and 1 mM TCEP) was gradually decreased by 0.2 pH units, and at each pH, the solution was allowed to equilibrate for at least 40 min. The solution was monitored for increased turbidity due to liquid droplet formation visually and by optical microscopic examination using the Olympus CKX41 bright-field microscope with a Photometrics CoolSNAP cf2 camera at 20× magnification. All solutions were at 4 °C and kept in the cold room except for brief visualization under a microscope at room temperature on cooled glass slides (4 °C). For each phase diagram, measurements were performed at pH 9.5, 9.0, 8.8, 8.6, 8.4, 8.2, 8.0, 7.8, 7.6, 7.4, and 5 different protein concentrations (0.5, 1.0, 2.0, 4.0, 6.0 mg/mL).

Temperature-Induced LLPS Images.

Samples of ELF3 PrLD Q0, Q7, and Q20 at ~ 4 mg/mL in 50 mM Bis-Tris propane, pH 9.2, 500 mM NaCl, and 1 mM TCEP (Q0 and Q7) or 50 mM Bis-Tris propane, pH 8.8, 500 mM NaCl, and 1 mM TCEP (Q20), 500 mM NaCl, 1 mM TCEP were visualized at 4 °C, 22 °C, and 4 °C using an Olympus CKX41 bright-field microscope with a Photometrics CoolSNAP cf2 camera at 20× magnification. 50 μL samples were cooled to ~ 4 °C and 5 μL solution was applied to a cooled glass slide and quickly imaged. The samples were then heated to 22 °C in a heating block for ~5 min to induce phase separation. In addition, 5 μL solution was applied to a room-temperature glass slide and imaged. The tubes were then cooled to ~4 °C for 15 min and imaged as described.

Confocal Imaging and Fluorescence Microscopy In Vitro.

For droplet visualization and photobleaching experiments of ELF3 PrLD Q0-, Q7-, and Q20-GFP proteins, liquid droplet formation was induced by dialysis as described above with GFP labeled ELF3 PrLD protein concentration of ~4 mg/mL and starting dialysis conditions of 50 mM Bis-Tris-propane pH 9.4, 500 mM NaCl with the pH gradually reduced in units of 0.2 until the solution became turbid (~pH 8.2 to 7.8). All dialysis steps were performed in a cold room at 4 °C. Once turbidity was observed, a 10 μL drop of solution was applied to a glass slide, and all subsequent measurements were performed at room temperature. The drop was covered with a cover slip and mounted and visualized using an objective-based total internal reflection fluorescence (TIRF) microscopy instrument composed of a Nikon Eclipse Ti, an azimuthal iLas2 TIRF illuminator (Roper Scientific), a 60× numerical aperture 1.49 TIRF objective lens followed by a 1.5× magnification lens, and an Evolve 512 camera (Photometrics). For photobleaching experiments, droplets were allowed to adhere to the coverslip prior to photobleaching to minimize droplet movement during the experiment. Acquisition times were approximately 1 s per image for 30 s and then every 10 s for 270 s. Droplet size was ~2 to 5 µm with a bleaching area of ~ 1 µm. Time-lapse images were acquired at 530 nm.

Correlative AFM-Fluorescence Experiments.

AFM-confocal fluorescence microscopy.

AFM coupled to confocal fluorescence microscope was developed in-house, allowing AFM imaging and fluorescence recovery after photobleaching experiments on the same droplets (SI Appendix, Fig. S3A) (32). AFM images were acquired using a Nanowizard 4 (JPK Instruments, Bruker) mounted on a Zeiss inverted optical microscope and equipped with a Tip Assisted Optics (TAO) module and a Vortis-SPM control unit. The AFM cantilever optical beam deflection system makes use of an infrared low-coherence light source (emission centered at 980 nm). A custom-made confocal microscope was coupled to the AFM using a supercontinuum laser (Rock-PP, Leukos) as laser source at 20 MHz equipped with an oil immersion objective with a 1.4 numerical aperture (Plan-Apochromat 100×, Zeiss). Fluorescence was collected after a pinhole of 100 μm diameter size (P100D, Thorlabs) by an avalanche photodetector (SPCM-AQR-15, PerkinElmer) connected to an SPC-150 (Becker & Hickl) TCSPC card. An ET800sp short-pass filter (Chroma) was used to filter out the light source of the AFM optical beam deflection system. The excitation laser power was measured after the objective at the sample level with a S170C microscope slide power sensor and a PM100 energy meter (both purchased from Thorlabs) and was set in all the experiments to 1 μW. Confocal images and FRAP were acquired using a 488/10 nm excitation filter and a 525/39 nm emission filter. Simultaneous AFM/confocal images were collected with the AFM tip and confocal spot positions fixed and coaligned while the sample was scanned using the TAO module. Full alignment was obtained using at first white field illumination. Then, the fine-tuned was achieved measuring the increase of tip luminescence when aligned with the confocal spot. Finally, while imaging the sample, any mismatch between topography and confocal image was adjusted by correcting the tip position.

AFM-wide-field fluorescence microscopy.

AFM-coupled wide-field fluorescence microscope was developed in-house, allowing AFM imaging and epifluorescence or Total Internal Reflection Fluorescence (TIRF) imaging (SI Appendix, Fig. S3B) (64, 65). and based on a Nanowizard 4 (JPK Instruments, Bruker) mounted on a Zeiss inverted optical microscope. The custom-made epifluorescence/TIRF microscope was coupled to the AFM using a LX 488-50 OBIS laser source (Coherent). We used an oil immersion objective with a 1.4 numerical aperture (Plan-Apochromat 100×, Zeiss). Fluorescence was collected with an EmCCD iXon Ultra897 (Andor) camera. The setup makes use of a 1.5× telescope to obtain a final imaging magnification of 150-fold, corresponding to a camera pixel size of 81 nm. An ET800sp short-pass filter (Chroma) was used in the emission optical path to filter out the light source of the AFM optical beam deflection system. The excitation laser wavelength was centered at 488nm and the power was measured before the objective with a PM100 energy meter (purchased from Thorlabs) and was optimized in all the experiments in the range of 1–5 μW. We used an acousto-optic tuneable filter (AOTFnc-400.650-TN, AA opto-electronics) to modulate the laser intensity and record fluorescence images using an ET525/50 nm (Chroma) as emission filter.
In both correlative AFM–TIRF/confocal, AFM images were acquired in QI mode with a scan size ranging from 10 μm × 10 μm to 50 μm × 50 μm and with 256 × 256 or 128/128 lines/pixels. Quantitative imaging mode (QI) was used to generate a force curve for each recorded pixel. Force curves were employed to evaluate droplets mechanical response. Typical force–distance curves were recorded with a loading rate of ≈10 μm/s, with large indentation cycle lengths ranging from 5 µm to 10 µm and maximal peak force of 1 to 1.5 nN. All parameters were optimized to provide image acquisition stability, considering sample low rigidity and having the cantilever immersed in a crowded environment with micrometric large droplets diffusing. Curves were then employed to extract all mechanical parameters for both liquid and gel phase constructs reported in Table 1, as well for the AFM images in Figs. 3 and 4. Gel phase constructs reported in Fig. 4, less dynamic on the coverslip and in solution, exhibited more stable force vs distance curves and could be imaged using a tip speed of 200 μm/s, ensuring a faster acquisition time. Because of the micrometric height of the droplet comparable with the size of the AFM tip, only the curves acquired on the top of the droplets were considered, in order to avoid edge artifacts, nonnegligible when the pyramidal tip facets are in contact with the droplets and present in the AFM images reported in Figs. 3 and 4C.
From the force curves on droplets in liquid phase, the stiffness keff was calculated based on a linear fit to the force vs indentation δ (Eq. 1) (33, 34).
F=keffδ,
[1]
Indentation cycles performed on droplets in gel phase were treated using a Hertz contact model, leading to the evaluation of the Young’s modulus E (Eq. 2).
F=43E1-υ2tanθδ2,
[2]
where ν is the Poisson’s ratio of 0.5, conventionally used for soft incompressible isotropic materials which are elastically deformed, and θ is the pyramidal tip angle. A comparison between indentation cycle performed on LLPS liquid phase (red) or gel condensates (blue) is shown in SI Appendix, Fig. S4 where Eqs. 1 and 2 have been used to fit the experimental data. The Young’s modulus images reported in Fig. 4A exhibit a higher E compared to the values reported in table, which is likely due to the higher AFM tip speed (loading rate) employed for the acquisition of the force vs distance curves, suggesting a viscoelastic behavior for all constructs in the gel phase.
Biolever mini (AC40TS, Olympus) and MSNL (Bruker) AFM cantilevers were purchased from Nano Bruker. Biolevers mini have a resonance frequency of ≈30 kHz in liquid, nominal stiffness of 0.1 N/m, and a 7-μm high tetrahedral tip. MSNL are V-shaped cantilevers with small tip radius (≈2 nm) suited for high-resolution imaging, and nominal stiffness ranging from 0.01 to 0.6 N/m, depending on the cantilever. Images in Figs. 3 and 4 in the main manuscript were generated using AC40TS and MSNL AFM cantilevers, respectively. In all AFM experiments, the inverse optical lever sensitivity and lever stiffness of the cantilevers were calibrated using the “contact-free” method of the JPK AFM instruments, making use of a combination of a Sader and thermal methods (66, 67). For droplets in gel phase, we used the MSNL-E for Q0 and Q7 constructs and the MSNL-F for the Q20 ones, with nominal spring constant of 0.1 N/m and 0.6 N/m, respectively, assuming θ=22.5 to fit experimental data with Eq. 2. The fundamental resonance was used with a correction of 0.817 at room temperature and in liquid environment. Circular glass coverslips (25 mm diameter, 165 μm thick) were purchased from Marienfeld. They were cleaned by a first cycle of sonication in 1 M KOH for 15 min, rinsed with deionized water 20 times, and finally subjected to a second sonication cycle in deionized water for 15 min. For measurements in the liquid phase, GFP-labeled ELF3 PrLD Q0, Q7, and Q20 samples in 50 mM Bis-Tris propane, pH 9.4, 500 mM NaCl, and 1 mM TCEP at 4 mg/mL were dialyzed stepwise against decreasing pH of 0.2 each 45 min until pH 8.2. For measurements in the gel phase, GFP-labeled ELF3 PrLD Q0, Q7, and Q20 samples in 50 mM Bis-Tris propane, pH 9.4, 500 mM NaCl, and 1 mM TCEP at 4 mg/mL were dialyzed stepwise for 45 min until pH 9.0, followed by pH 8.0 and finally pH 7.6 for 15 min. The fast pH changes led to the formation of spherical droplets with gel-like properties. The final protein concentration used for all measurements was ~2 mg/mL as calculated using A280 measurements on a NanoDrop (ThermoFisher) and adjusted by dilution with the respective final dialysis buffer. A 10 µL sample of protein was applied to a clean glass cover slip and mounted on the AFM. After incubation for 2 min to allow droplets to adhere to the surface, 300 µL of buffer A was added and finally imaged by correlative AFM-TIRF fluorescence (Fig. 4 in main manuscript). Total sample preparation time for AFM measurements was ~ 15 to 20 min. including sample slide preparation, cantilever tuning, and image acquisition time (11 min for a 256 × 256 pixel image). We have noted that droplets become more gel-like over time as they are allowed to adhere to the surface of the microscope slide.
From correlative confocal-AFM experiments, the LLPS liquid-like droplet contact angle can be evaluated. For a spherical droplet, it is given by ref. 68
θ=2tan-1(h+ra).
[3]
θ can be evaluated using the morphological information obtained from both confocal (parameter a) and AFM images (parameter h+r, maximal height of the droplet) as shown in SI Appendix, Fig. S5. Shadow vertically asymmetry in SI Appendix, Fig. S5C can be partially due to the tetrahedral geometry of the probe as well as to the 10° cantilever inclination with respect to the glass coverslip (imposed by the AFM cantilever holder). Horizontal asymmetry can be partially due to the excitation beam not fully perpendicular to the glass coverslip and to a horizontal inclination of the cantilever. Contact angles for droplets were found in the range between 120° and 150° suggesting a low wettability associated to quasispherical droplet shapes as shown in Fig. 3 in main manuscript.

Transmission Electron Microscopy Experiments.

Samples of gel phase Q0, Q7, and Q20 ELF3 PrLD were prepared by stepwise dialysis of proteins at 4 mg/mL in 50 mM Bis-Tris propane, pH 9.4, 500 mM NaCl, and 1 mM TCEP. The pH was decreased in units of 0.2, with each dialysis step performed for 40 min at 4 °C until pH 7.8. The sample became cloudy and then formed amorphous gel-like deposits, which were stored for at least 1 h at 4 °C before image collection. The viscous gel was resuspended by fast pipetting to obtain a suspension of small gel pieces. Around 4 µL of sample was applied to the mica-carbon interface of a mica sheet with a layer of evaporated carbon film. The carbon film was floated off the mica sheet in ~600 µL of 2% (wt/vol) sodium silicotungstate. The sample was then transferred onto a 400-mesh copper TEM grid and air-dried. Images were taken on a Themo Fisher Technai F20 microscope operating at 200 kV. Images were acquired using a Thermo Fisher Ceta camera.

SAXS Measurements.

SAXS experiments were performed at the European Synchrotron Radiation Facility (ESRF) on the BioSAXS beamline BM29 (69, 70) for dilute phase, HPLC experiments and on beamline B21 at Diamond Light Source (SM26313-1, SM26314-2) for temperature ramp experiments (71). For dilute phase experiments, an online HPLC system (Shimadzu) was attached directly to the sample inlet valve of the BM29 sample changer. Protein samples at 4 mg/ml (ELF3 PrLD Q0) or ~8 mg/mL (ELF3 PrLD Q7, Q20) were loaded into vials and automatically injected onto the column (Superose 6 3.2/300 Increase GE Healthcare) via an integrated syringe system. Buffers were degassed online, and a flow rate of 0.05 mL/min at room temperature was used for all sample runs. For measurements of the dilute phase for ELF3 PrLD Q0, Q7, and Q20, 50 mM Bis-Tris propane pH 9.4, 1 M NaCl, and 1 mM TCEP was used for all samples. Prior to each, run the column was equilibrated with at least three column volumes of buffer and the baseline was monitored. All data from the run were collected using a sample to detector (Pilatus 2M Dectris) distance of 2.81 m corresponding to a q range of 0.008 to 0.45 Å−1. Due to column separation of the sample, some dilution effects will occur prior to measurement. Then, 1,800 frames (2 s/frame) were collected. Initial data processing was performed automatically using the Dahu pipeline, generating radially integrated, calibrated, and normalized 1-D profiles for each frame (72, 73). All frames were then further processed using Scatter IV (74); briefly, the frames were dropped into the software and 50 to 100 frames were selected for the background buffer. Using the heat plot, a selection of similar frames (cyan) were selected and merged. Thirty frames corresponding to the highest protein concentration were merged and used for all further data processing and model fitting.
For the temperature ramp experiment for ELF3 PrLD Q7, samples were placed into the BioSAXS robot (Arinax, France) with both the sample changer and the vacuum cell both set at 4 °C. Then, 50 µL of sample and matching buffer containing 50 mM CAPS, pH 9.7, 300 mM NaCl, 1mM TCEP were individually loaded and 10 frames of 1 s were taken. Temperature was raised in both the sample changer and the vacuum cell and incubated for 5 min before measurement. Temperature series of 4 °C, 12 to 15 °C, and 22 to 24 °C were taken. All frames were compared to the initial frame and matching frames were merged for buffer and samples. Scatter IV10 was then used to subtract buffer from sample and to generate Rg and plots.

X-ray Diffraction Measurements.

X-ray diffraction experiments for Q0, Q7, and Q20 ELF3PrLD were performed on beamline ID23-2 at the ESRF. Briefly, samples were purified as described above and dialyzed directly against 50 mM BTP, pH 7.8, 250 mM NaCl, and 1 mM TCEP at a concentration of ~ 4 mg/mL for 3 h, after which time the protein formed an amorphous gel. The sample was spun down to collect the gel. The gel was directly mounted in a mesh litholoop (Molecular Dimensions). Diffraction was measured at room temperature. Diffuse powder rings were observed at ~4 to 5 Å, consistent with van der Waal’s distance between interacting molecules.

Multiangle Laser Light Scattering Experiments.

First, 50 µl of ELF3 PrD Q0, Q7, or Q20 at a concentration of ~4 to 7mg/mL were loaded onto an S200 Increase size-exclusion column (Superdex 200 Increase10/300 GL, GE Healthcare) at a flow rate of 0.5 mL min−1. Then, the column was preequilibrated with 50 mM Bis Tris propane at pH 9.4, 1 M NaCl, 1 mM TCEP and connected to a Hitachi Elite LaChrom UV detector and LAChrome Pump L-2130, a multiangle laser light-scattering detector (DAWN HELEOS II, Wyatt Technology Corporation) and a refractive-index detector (Optilab T-rEX, Wyatt Technology Corporation). The data were processed with the ASTRA 6.1.7.17 software (Wyatt Technology Corporation).

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.

Acknowledgments

We thank the B21 local contact Katsuaki Inoue from Diamond Light Source for his help during the SAXS temperature ramp experiments experiment. We thank Dr. Guy Schoehn for his support with EM experiments. We would like to thank the Partnership for Soft Condensed Matter (PSCM) at the ESRF for providing the lab space and equipment. This project received support from the ANR (ANR-19-CE20-0021 and ANR-21-CE11-0037) and GRAL, a program from the Chemistry and Biology Health Graduate School of the University Grenoble Alpes (ANR-17-EURE-0003). GRAL support for the μLife imaging facility was provided. This work benefited from access to the MX-Grenoble, an Instruct-ERIC centre within the Grenoble Partnership for Structural Biology (PSB). The X-ray diffraction and SAXS dilute and condensed phase experiments were performed on beamline ID23-2 and BM29 at the European Synchrotron Radiation Facility (ESRF), Grenoble, France. Financial support was provided by Instruct-ERIC (PID 13317). This work used the platforms of the Grenoble Instruct-ERIC center (ISBG; UAR 3518 CNRS-CEA-UGA-EMBL) within the Grenoble Partnership for Structural Biology (PSB), supported by FRISBI (ANR-10-INBS-0005-02). The electron microscope facility is supported by the Auvergne-Rhône-Alpes Region, the Fondation Recherche Medicale (FRM), the fonds FEDER and the GIS-Infrastructures en Biologie Sante et Agronomie (IBiSA). IBS acknowledges integration into the Interdisciplinary Research Institute of Grenoble (IRIG, CEA). P.-E.M. and L.C. acknowledge the support from CNRS Momentum program (2017) and from the Plan Cancer Equipment 2016. The CBS is a member of the France-BioImaging (FBI), national infrastructure supported by the French National Research Agency (ANR-10-INBS-04-01) and of the GIS IBISA (Infrastructures en Biologie Santé et Agronomie). J.R.K is supported by an MRC Career Development Award (MR/W01632X/1).

Author contributions

S.H., Y.S., L.C., M.D.T., and C.Z. designed research; S.H., J.R.K., V.I.S., A.D., W.L.L., N.L., A.P., M.H.N., L.C., and M.D.T. performed research; S.H., P.-E.M., M.B., P.A.W., Y.S., L.C., M.D.T., and C.Z. analyzed data; and S.H., L.C., M.D.T., and C.Z. wrote the paper.

Competing interests

The authors declare no competing interest.

Supporting Information

Appendix 01 (PDF)

References

1
C. P. Brangwynne et al., Germline P granules are liquid droplets that localize by controlled dissolution/condensation. Science 324, 1729–1732 (2009).
2
M. Feric et al., Coexisting liquid phases underlie nucleolar subcompartments. Cell 165, 1686–1697 (2016).
3
J. A. Riback et al., Composition-dependent thermodynamics of intracellular phase separation. Nature 581, 209–214 (2020).
4
H. Zhang et al., RNA controls PolyQ protein phase transitions. Mol. Cell 60, 220–230 (2015).
5
Y. Shin et al., Liquid nuclear condensates mechanically sense and restructure the genome. Cell 175, 1481–1491.e13 (2018).
6
T. M. Franzmann et al., Phase separation of a yeast prion protein promotes cellular fitness. Science 359, eaao5654 (2018).
7
J. Guillén-Boixet et al., RNA-induced conformational switching and clustering of G3BP drive stress granule assembly by condensation. Cell 181, 346–361.e17 (2020).
8
C. Iserman et al., Condensation of Ded1p promotes a translational switch from housekeeping to stress protein production. Cell 181, 818–831.e19 (2020).
9
J.-H. Jung et al., A prion-like domain in ELF3 functions as a thermosensor in Arabidopsis. Nature 585, 256–260 (2020).
10
S. Alberti et al., A user’s guide for phase separation assays with purified proteins. J. Mol. Biol. 430, 4806–4820 (2018).
11
D. S. W. Protter et al., Intrinsically disordered regions can contribute promiscuous interactions to RNP granule assembly. Cell Rep. 22, 1401–1412 (2018).
12
E. W. Martin, A. S. Holehouse, Intrinsically disordered protein regions and phase separation: Sequence determinants of assembly or lack thereof. Emerg. Top. Life Sci. 4, 307–329 (2020).
13
M. Paloni, R. Bailly, L. Ciandrini, A. Barducci, Unraveling molecular interactions in liquid-liquid phase separation of disordered proteins by atomistic simulations. J. Phys. Chem. B 124, 9009–9016 (2020).
14
S. Rauscher, R. Pomès, The liquid structure of elastin. Elife 6, e26526 (2017).
15
A. C. Murthy et al., Molecular interactions underlying liquid-liquid phase separation of the FUS low-complexity domain. Nat. Struct. Mol. Biol. 26, 637–648 (2019).
16
B. S. Schuster et al., Identifying sequence perturbations to an intrinsically disordered protein that determine its phase-separation behavior. Proc. Natl. Acad. Sci. U.S.A. 117, 11421–11431 (2020).
17
G. L. Dignon, W. Zheng, Y. C. Kim, R. B. Best, J. Mittal, Sequence determinants of protein phase behavior from a coarse-grained model. PLoS Comput. Biol. 14, e1005941 (2018).
18
S. Das, A. N. Amin, Y.-H. Lin, H. S. Chan, Coarse-grained residue-based models of disordered protein condensates: utility and limitations of simple charge pattern parameters. Phys. Chem. Chem. Phys. 20, 28558–28574 (2018).
19
A. Statt, H. Casademunt, C. P. Brangwynne, A. Z. Panagiotopoulos, Model for disordered proteins with strongly sequence-dependent liquid phase behavior. J. Chem. Phys. 152, 075101 (2020).
20
G. L. Dignon, W. Zheng, Y. C. Kim, J. Mittal, Temperature-controlled liquid-liquid phase separation of disordered proteins. ACS Cent. Sci. 5, 821–830 (2019).
21
S. Elbaum-Garfinkle et al., The disordered P granule protein LAF-1 drives phase separation into droplets with tunable viscosity and dynamics. Proc. Natl. Acad. Sci. U.S.A. 112, 7189–7194 (2015).
22
K. A. Burke, A. M. Janke, C. L. Rhine, N. L. Fawzi, Residue-by-residue view of in vitro FUS granules that bind the C-terminal domain of RNA polymerase II. Mol. Cell 60, 231–241 (2015).
23
S. Kroschwald et al., Different material states of Pub1 condensates define distinct modes of stress adaptation and recovery. Cell Rep. 23, 3327–3339 (2018).
24
J. Jh et al., A prion-like domain in ELF3 functions as a thermosensor in Arabidopsis. Nature 585, 256–260 (2020).
25
J. A. Riback et al., Stress-triggered phase separation is an adaptive. Evolutionarily tuned response. Cell 168, 1028–1040.e19 (2017).
26
S. F. Undurraga et al., Background-dependent effects of polyglutamine variation in the Arabidopsis thaliana gene ELF3. Proc. Natl. Acad. Sci. U.S.A. 109, 19363–19367 (2012).
27
T. Tajima, A. O. Oda, M. Nakagawa, H. Kamada, T. Mizoguchi, Natural variation of polyglutamine repeats of a circadian clock gene ELF 3 in Arabidopsis. Plant Biotechnol. 24, 237–240 (2007).
28
D. T. McSwiggen, M. Mir, X. Darzacq, R. Tjian, Evaluating phase separation in live cells: Diagnosis, caveats, and functional consequences. Genes Dev. 33, 1619–1634 (2019).
29
S. Park et al., Dehydration entropy drives liquid-liquid phase separation by molecular crowding. Commun. Chem. 3, 1–12 (2020).
30
Y. Lin, Narrow equilibrium window for complex coacervation of tau and RNA under cellular conditions. ELife 8, e42571 (2019).
31
F. G. Quiroz, A. Chilkoti, Sequence heuristics to encode phase behaviour in intrinsically disordered protein polymers. Nat. Mater 14, 1164–1171 (2015).
32
T. F. D. Fernandes, O. Saavedra-Villanueva, E. Margeat, P.-E. Milhiet, L. Costa, Synchronous, crosstalk-free correlative AFM and confocal microscopies/spectroscopies. Sci. Rep. 10, 7098 (2020).
33
L. Costa, G. Li-Destri, D. Pontoni, O. Konovalov, N. H. Thomson, Liquid-liquid interfacial imaging using atomic force microscopy. Adv. Mater. Interfaces 4, 1700203 (2017).
34
D. Y. C. Chan, R. R. Dagastine, L. R. White, Forces between a rigid probe particle and a liquid interface. J. Colloid Interface Sci. 236, 141–154 (2001).
35
E. W. Martin, J. B. Hopkins, T. Mittag, “Chapter Seven - Small-angle X-ray scattering experiments of monodisperse intrinsically disordered protein samples close to the solubility limit” in Methods in Enzymology, Liquid-Liquid Phase Coexistence and Membraneless Organelles, C. D. Keating, Ed. (Academic Press, 2021), pp. 185–222.
36
E. W. Martin et al., A multi-step nucleation process determines the kinetics of prion-like domain phase separation. Nat. Commun. 12, 4513 (2021).
37
T. Shirahama, A. S. Cohen, Structure of amyloid fibrils after negative staining and high-resolution electron microscopy. Nature 206, 737–738 (1965).
38
A. Meister, A. Blume, Cryo)Transmission electron microscopy of phospholipid model membranes interacting with amphiphilic and polyphilic molecules. Polymers (Basel) 9, 521 (2017).
39
M. J. Hope, M. B. Bally, G. Webb, P. R. Cullis, Production of large unilamellar vesicles by a rapid extrusion procedure: characterization of size distribution, trapped volume and ability to maintain a membrane potential. Biochim. Biophys. Acta 812, 55–65 (1985).
40
M. P. Hughes et al., Atomic structures of low-complexity protein segments reveal kinked β sheets that assemble networks. Science 359, 698–701 (2018).
41
C. M. Bellingham et al., Recombinant human elastin polypeptides self-assemble into biomaterials with elastin-like properties. Biopolymers 70, 445–455 (2003).
42
R. E. Lyons et al., Comparisons of recombinant resilin-like proteins: repetitive domains are sufficient to confer resilin-like properties. Biomacromolecules 10, 3009–3014 (2009).
43
C. Garcia Garcia, K. L. Kiick, Methods for producing microstructured hydrogels for targeted applications in biology. Acta Biomater 84, 34–48 (2019).
44
B. S. Schuster et al., Controllable protein phase separation and modular recruitment to form responsive membraneless organelles. Nat. Commun. 9, 2985 (2018).
45
Y. Shin et al., Spatiotemporal control of intracellular phase transitions using light-activated optodroplets. Cell 168, 159–171.e14 (2017).
46
P. Aducci, R. Federico, G. Carpinelli, F. Podo, Temperature dependence of intracellular pH in higher plant cells: A 31 P nuclear magnetic resonance study on maize root tips. Planta 156, 579–582 (1982).
47
A. Pecherina et al., Whole-plant measure of temperature-induced changes in the cytosolic pH of potato plants using genetically encoded fluorescent sensor Pt-GFP. Agriculture 11, 1131 (2021).
48
C. L. Cuevas-Velazquez, J. R. Dinneny, Organization out of disorder: Liquid–liquid phase separation in plants. Curr. Opin. Plant Biol. 45, 68–74 (2018).
49
H. Huang et al., Identification of evening complex associated proteins in arabidopsis by affinity purification and mass spectrometry. Mol. Cell Proteomics 15, 201–217 (2016).
50
J. Rausenberger et al., An integrative model for phytochrome B mediated photomorphogenesis: From protein dynamics to physiology. PLoS One 5, e10721 (2010).
51
P. Tompa, M. Fuxreiter, Fuzzy complexes: Polymorphism and structural disorder in protein-protein interactions. Trends Biochem. Sci. 33, 2–8 (2008).
52
H. Wu, M. Fuxreiter, A structure and dynamics continuum of higher-order assemblies: Amyloids signalosomes and granules. Cell 165, 1055–1066 (2016).
53
M. Vendruscolo, M. Fuxreiter, Sequence determinants of the aggregation of proteins within condensates generated by liquid-liquid phase separation. J. Mol. Biol. 434, 167201 (2022).
54
M. Hardenberg, A. Horvath, V. Ambrus, M. Fuxreiter, M. Vendruscolo, Widespread occurrence of the droplet state of proteins in the human proteome. Proc. Natl. Acad. Sci. U.S.A. 117, 33254–33262 (2020).
55
A. Hatos, S. C. E. Tosatto, M. Vendruscolo, M. Fuxreiter, FuzDrop on AlphaFold: Visualizing the sequence-dependent propensity of liquid–liquid phase separation and aggregation of proteins. Nucleic Acids Res. 50, W337–W344 (2022).
56
X. Chu et al., Prediction of liquid–liquid phase separating proteins using machine learning. BMC Bioinformatics 23, 72 (2022).
57
M. P. Hughes, L. Goldschmidt, D. S. Eisenberg, Prevalence and species distribution of the low-complexity, amyloid-like, reversible, kinked segment structural motif in amyloid-like fibrils. J. Biol. Chem. 297, 101194 (2021).
58
E. Korkmazhan, P. Tompa, A. R. Dunn, The role of ordered cooperative assembly in biomolecular condensates. Nat. Rev. Mol. Cell Biol. 22, 647–648 (2021).
59
B. J. Mayer, M. L. Blinov, L. M. Loew, Molecular machines or pleiomorphic ensembles: Signaling complexes revisited. J. Biol. 8, 81 (2009).
60
Y. Dorone et al., A prion-like protein regulator of seed germination undergoes hydration-dependent phase separation. Cell 184, 4284–4298.e27 (2021).
61
A. Lampropoulos et al., GreenGate - A novel, versatile, and efficient cloning system for plant transgenesis. PLOS One 8, e83043 (2013).
62
A. Bleckmann, S. Weidtkamp-Peters, C. A. M. Seidel, R. Simon, Stem cell signaling in arabidopsis requires CRN to localize CLV2 to the plasma membrane. Plant Physiol. 152, 166–176 (2010).
63
J. Schindelin et al., Fiji: An open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
64
S. Dahmane et al., Nanoscale organization of tetraspanins during HIV-1 budding by correlative dSTORM/AFM. Nanoscale 11, 6036–6044 (2019).
65
A. Vial et al., Correlative AFM and fluorescence imaging demonstrate nanoscale membrane remodeling and ring-like and tubular structure formation by septins. Nanoscale 13, 12484–12493 (2021).
66
J. E. Sader et al., A virtual instrument to standardise the calibration of atomic force microscope cantilevers. Rev. Sci. Instrum. 87, 093711 (2016).
67
R. Proksch, T. E. Schäffer, J. P. Cleveland, R. C. Callahan, M. B. Viani, Finite optical spot size and position corrections in thermal spring constant calibration. Nanotechnology 15, 1344–1350 (2004).
68
D. Nadargi, S. Latthe, H. Hirashima, A. Rao, Studies on rheological properties of methyltriethoxysilane (MTES) based flexible superhydrophobic silica aerogels. Microporous Mesoporous Mater. 117, 617–626 (2009).
69
P. Pernot et al., Upgraded ESRF BM29 beamline for SAXS on macromolecules in solution. J. Synchrotron Rad. 20, 660–664 (2013).
70
M. D. Tully et al., BioSAXS at european synchrotron radiation facility – extremely brilliant source: BM29 with an upgraded source, detector, robot, sample environment, data collection and analysis software. J. Synchrotron Rad. 30, 258–266 (2023).
71
N. P. Cowieson et al., Beamline B21: High-throughput small-angle X-ray scattering at diamond light source. J. Synchrotron Rad. 27, 1438–1446 (2020).
72
M. F. Incardona et al., EDNA: A framework for plugin-based applications applied to X-ray experiment online data analysis. J. Synchrotron Rad. 16, 872–879 (2009).
73
J. Kieffer et al., New data analysis for BioSAXS at the ESRF. J. Synchrotron Rad. 29, 1318–1328 (2022).
74
M. D. Tully, N. Tarbouriech, R. P. Rambo, S. Hutin, Analysis of SEC-SAXS data via EFA deconvolution and scatter. J. Vis. Exp. 28, e61578 (2021).

Information & Authors

Information

Published in

The cover image for PNAS Vol.120; No.28
Proceedings of the National Academy of Sciences
Vol. 120 | No. 28
July 11, 2023
PubMed: 37399408

Classifications

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.

Submission history

Received: March 22, 2023
Accepted: June 6, 2023
Published online: July 3, 2023
Published in issue: July 11, 2023

Keywords

  1. phase separation
  2. thermosensing
  3. EARLY FLOWERING 3
  4. small angle X-ray scattering
  5. atomic force microscopy

Acknowledgments

We thank the B21 local contact Katsuaki Inoue from Diamond Light Source for his help during the SAXS temperature ramp experiments experiment. We thank Dr. Guy Schoehn for his support with EM experiments. We would like to thank the Partnership for Soft Condensed Matter (PSCM) at the ESRF for providing the lab space and equipment. This project received support from the ANR (ANR-19-CE20-0021 and ANR-21-CE11-0037) and GRAL, a program from the Chemistry and Biology Health Graduate School of the University Grenoble Alpes (ANR-17-EURE-0003). GRAL support for the μLife imaging facility was provided. This work benefited from access to the MX-Grenoble, an Instruct-ERIC centre within the Grenoble Partnership for Structural Biology (PSB). The X-ray diffraction and SAXS dilute and condensed phase experiments were performed on beamline ID23-2 and BM29 at the European Synchrotron Radiation Facility (ESRF), Grenoble, France. Financial support was provided by Instruct-ERIC (PID 13317). This work used the platforms of the Grenoble Instruct-ERIC center (ISBG; UAR 3518 CNRS-CEA-UGA-EMBL) within the Grenoble Partnership for Structural Biology (PSB), supported by FRISBI (ANR-10-INBS-0005-02). The electron microscope facility is supported by the Auvergne-Rhône-Alpes Region, the Fondation Recherche Medicale (FRM), the fonds FEDER and the GIS-Infrastructures en Biologie Sante et Agronomie (IBiSA). IBS acknowledges integration into the Interdisciplinary Research Institute of Grenoble (IRIG, CEA). P.-E.M. and L.C. acknowledge the support from CNRS Momentum program (2017) and from the Plan Cancer Equipment 2016. The CBS is a member of the France-BioImaging (FBI), national infrastructure supported by the French National Research Agency (ANR-10-INBS-04-01) and of the GIS IBISA (Infrastructures en Biologie Santé et Agronomie). J.R.K is supported by an MRC Career Development Award (MR/W01632X/1).
Author contributions
S.H., Y.S., L.C., M.D.T., and C.Z. designed research; S.H., J.R.K., V.I.S., A.D., W.L.L., N.L., A.P., M.H.N., L.C., and M.D.T. performed research; S.H., P.-E.M., M.B., P.A.W., Y.S., L.C., M.D.T., and C.Z. analyzed data; and S.H., L.C., M.D.T., and C.Z. wrote the paper.
Competing interests
The authors declare no competing interest.

Notes

This article is a PNAS Direct Submission. L.C.S. is a guest editor invited by the Editorial Board.

Authors

Affiliations

Laboratoire de Physiologie Cellulaire et Végétale, University Grenoble Alpes, Centre national de la recherche scientifique, Commissariat à l'énergie atomique et aux énergies alternatives, Institut national de recherche pour l’agriculture, l’alimentation et l’environnement, Institut de recherche interdisciplinaire de Grenoble, Grenoble 38054, France
Janet R. Kumita
Department of Pharmacology, University of Cambridge, Cambridge CB2 1PD, United Kingdom
Institute for Developmental Genetics, Heinrich-Heine University, Düsseldorf D-40225, Germany
Institute for Developmental Genetics, Heinrich-Heine University, Düsseldorf D-40225, Germany
University Grenoble Alpes, Commissariat à l'énergie atomique et aux énergies alternatives, Centre national de la recherche scientifique, Institut de Biologie Structurale, Institut de recherche interdisciplinaire de Grenoble, Grenoble 38000, France
Nessim Louafi
Centre de Biologie Structurale, University Montpellier, Centre national de la recherche scientifique, Institut national de la santé et de la recherche médicale, Montpellier 34090, France
European Synchrotron Radiation Facility, Structural Biology Group, Grenoble 38000, France
Centre de Biologie Structurale, University Montpellier, Centre national de la recherche scientifique, Institut national de la santé et de la recherche médicale, Montpellier 34090, France
Martin Blackledge
University Grenoble Alpes, Commissariat à l'énergie atomique et aux énergies alternatives, Centre national de la recherche scientifique, Institut de Biologie Structurale, Institut de recherche interdisciplinaire de Grenoble, Grenoble 38000, France
Max H. Nanao
European Synchrotron Radiation Facility, Structural Biology Group, Grenoble 38000, France
Leibniz-Institut für Gemüse- und Zierpflanzenbau, 14979 Grossbeeren, Germany
Institute of Biochemistry and Biology, University of Potsdam, 14476 Potsdam, Germany
Yvonne Stahl
Institute for Developmental Genetics, Heinrich-Heine University, Düsseldorf D-40225, Germany
Cluster of Excellence on Plant Sciences, Heinrich-Heine University, Düsseldorf D-40225, Germany
Centre de Biologie Structurale, University Montpellier, Centre national de la recherche scientifique, Institut national de la santé et de la recherche médicale, Montpellier 34090, France
European Synchrotron Radiation Facility, Structural Biology Group, Grenoble 38000, France
Laboratoire de Physiologie Cellulaire et Végétale, University Grenoble Alpes, Centre national de la recherche scientifique, Commissariat à l'énergie atomique et aux énergies alternatives, Institut national de recherche pour l’agriculture, l’alimentation et l’environnement, Institut de recherche interdisciplinaire de Grenoble, Grenoble 38054, France

Notes

1
To whom correspondence may be addressed. Email: [email protected], [email protected], or [email protected].

Metrics & Citations

Metrics

Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.


Citation statements




Altmetrics

Citations

Export the article citation data by selecting a format from the list below and clicking Export.

Cited by

    Loading...

    View Options

    View options

    PDF format

    Download this article as a PDF file

    DOWNLOAD PDF

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to access the full text.

    Single Article Purchase

    Phase separation and molecular ordering of the prion-like domain of the Arabidopsis thermosensory protein EARLY FLOWERING 3
    Proceedings of the National Academy of Sciences
    • Vol. 120
    • No. 28

    Figures

    Tables

    Media

    Share

    Share

    Share article link

    Share on social media