Abstract
The phase separation of intrinsically disordered proteins is emerging as an important mechanism for cellular organization. However, efforts to connect protein sequences to the physical properties of condensates, i.e., the molecular grammar, are hampered by a lack of effective approaches for probing high-resolution structural details. Using a combination of multiscale simulations and fluorescence lifetime imaging microscopy experiments, we systematically explored a series of systems consisting of diblock elastin-like polypeptides (ELP). The simulations succeeded in reproducing the variation of condensate stability upon amino acid substitution and revealed different microenvi-ronments within a single condensate, which we verified with environmentally sensitive fluorophores. The interspersion of hydrophilic and hydrophobic residues and a lack of secondary structure formation result in a frustrated microphase separation, which explains both the strong correlation between ELP condensate stability and interfacial hydrophobicity scales, as well as the prevalence of protein-water hydrogen bonds. Our study uncovers new mechanisms for condensate stability and organization that may be broadly applicable.
eLife assessment
This important study investigates the structural organization of a series of diblock elastin-like polypeptide condensates. The methodology is highly compelling, as it combines multiscale simulations and fluorescence lifetime imaging microscopy experiments. The results increase our understanding of model biomolecular condensates. The manuscript would benefit from more details for the concepts and terminology introduced, as well as the analysis of the simulations.
Introduction
Biological condensates are found in both the cytosol1–4 and nucleus,5–10 playing essential roles in a variety of cellular processes11, 12 from stress response13 to genome organization.14, 15 Similar to membrane-bound organelles, they assemble a collection of molecules to raise the efficiency of sophisticated tasks. The lack of a membrane barrier allows fast material exchange between condensates and the cellular environment, rendering the molecular composition and stability of condensates more prone to regulations by external signals. 16–18
Intrinsically disordered proteins (IDPs) that promote multivalent, promiscuous interactions are key drivers of condensate formation. 19–21 Multiple mechanisms, including electrostatic, cation-π, π-π, hydrogen bonding, and hydrophobic interactions, contribute to the affinity among various chemical groups.22–24 Above a threshold concentration, as predicted by the Flory-Huggins theory,25 interactions among IDPs can drive liquid-liquid phase separation (LLPS) to produce a highly concentrated phase that nevertheless remains dynamic. The simplicity of theory, while insightful, may prove insufficient for a comprehensive understanding of biological condensates.
Much remains to be learned regarding the connection between amino acid sequences and protein phase behaviors, or the so-called “molecular grammar” of protein condensates.26–31 These systems often exhibit complex viscoelastic behaviors and substructures with a layered organization,32–35 defying the mean-field assumption in the Flory-Huggins theory. Several more advanced theories have been introduced to better account for long-lived structural features that might persist in the polymer network. The sticker and spacer model36, 37 has been adopted by Pappu and coworkers to account for strong and specific interactions that might form physical cross-links among protein molecules.38 In the meantime, the block copolymer theory may explain the microphase separation that can lead to layered structures.39, 40 Because of their inherent assumptions, different theories are more appropriate for some systems but not others. High-resolution structural characterizations of condensates could offer further insight into their organizational principles and the applicable theories.
Further decoding the condensate grammar could also benefit from studies of simpler systems. Natural proteins often utilize highly complex amino acid sequences, rendering the attribution of an individual residue’s contribution to collective phenomena challenging. On the other hand, elastin-like polypeptides (ELPs), which are composed of pentapeptide repeats of valine-proline-glycine-X-glycine, serve as excellent models for studying protein phase behaviors. ELPs undergo phase separation upon heating with a lower critical solution temperature (LCST).41–43 Importantly, the guest residue X can be modified to any amino acid except proline with modern engineering approaches, 44 enabling systematic characterization of the impact on condensate stability and organization upon introducing specific residues.
We combine multiscale simulations with fluorescence lifetime imaging microscopy (FLIM) to probe condensates formed by diblock ELPs and investigate the contribution of amino acid composition to condensate stability. The simulation approach allows the sampling of large-scale conformational rearrangements while providing atomistic resolution to quantify the solvation environment of individual residues. It succeeds in reproducing the stability of condensates, and reveals the formation of gyroid-like structures. Such structures deviate from weak micelles that have been proposed for diblock ELPs in solution, and result in heterogeneous microenvironments within the condensate, which we verify in-vitro with FLIM. In addition, we find that the condensate stability exhibits a striking correlation with a hydrophobicity scale derived from interfacial transfer free energy, supporting an interfacial microenvironment of the condensate interior. The chemical specificity of the microenvironment is dictated by the peptide sequence that prevents a complete microphase separation between hydrophobic and hydrophilic groups. As a result, all condensates remain highly solvated after phase separation, producing frustrated water molecules that maintain hydrogen bonding with the exposed peptide backbone which persists even in residues with hydrophobic side chains. Our study takes a significant stride at connecting amino acid sequences with the structural and physical properties of biological condensates.
Results
Multiscale simulations of diblock ELP condensates
Decoding the condensate grammar necessitates connecting protein sequences with collective physical properties. ELPs stand out because of their sequence simplicity and amenability for biological engineering, allowing systematic exploration and precise attribution of amino acid contributions. We focus on diblock ELPs of sequence (V-P-G-V-G)n-(V-P-G-X-G)n, which we abbreviate as VnXn (Fig. 1A). The guest residue in the first block is set as valine to promote phase separation, and we explore 20 systems in which every natural amino acid is substituted into the X position in the second block.
![](https://prod--epp.elifesciences.org/iiif/2/90750%2Fv1%2Fcontent%2F534967v1_fig1.tif/full/max/0/default.jpg)
Multiscale simulations enable thermodynamic and structural characterization of ELP condensates. (A) Illustration of the sequence for the simulated diblock ELPs that consist of five-amino acid repeats, where X is substituted with a guest amino acid. (B) Overview of the three-step multiscale simulation approach that gradually increases the model resolution. Simulations of V5L5 were used to produce the example configurations at each step. Peptides are shaded red-white, while water molecules, chlorine, and sodium, are colored in blue, green, and orange, respectively. Inserts of a G-V-G repeat and surrounding water molecules are shown to indicate the resolution of each model. (C) Correlation between the simulated surface tension (τ) of 20 ELP condensates and the transition temperatures (Tt) of related systems.41 ρ is the Pearson correlation coefficient between the two data sets, and the dashed line is the best fit between simulation and experimental data. Error bars represent the standard deviation of estimates from five independent time windows.
We adopt a multiscale approach to balance accuracy and efficiency for simulating ELP condensates (Fig. 1B). The first stage of this strategy is to simulate condensate formation using a coarse-grained force field, MOFF,45–48 with one bead per amino acid and implicit solvation. These simulations are highly efficient for system relaxation and promoting large-scale conformational rearrangements that occur over slow timescales. We further converted these α-carbon based models to approximately four heavy atoms per bead resolution, and introduced explicit water and ions for simulations with the MARTINI force field. 49 The latest version of this force field (MARTINI3) provides a more balanced set of parameters for protein-protein interactions and has been applied to study biological condensates.50–53 These simulations further relax the protein configurations and the partition of water and counterions in condensed and dilute phases. Finally, we carried out explicit solvent all-atom simulations for 250 ns to produce an accurate characterization of the chemical environment of the condensates.
Utilizing the simulated condensate conformations, we computed various quantities to benchmark against experimental measurements. As a measure for condensate stability, we determined the surface tension, τ, using 100-µs-long MARTINI simulations performed with the NPNAT ensemble (Fig. 1C).54 Intuitively, condensates with higher surface tension are more stable. In addition, because ELPs undergo LLPS upon heating, more stable condensates show smaller values for the lower critical solution temperature (Tt).55 A negative correlation can be readily seen between computed surface tension and experimental Tt, supporting the simulation accuracy. We note that the experimental values were determined using different ELP sequences from those simulated here, and the relation between the surface tension and Tt is likely nonlinear, both contributing to the imperfect agreement.41 In another study,56 we adopted the same multiscale approach to determine the dielectric constant for several ELP condensates. The simulated values agree well with those determined using FLIM experiments.
Microphase separation of ELP condensates
Upon validating the accuracy of the simulation protocol for reproducing collective properties of ELP condensates, we examined their structural organization. Diblock copolymers are known to form a variety of morphologies.57–59 Hassouneh et al. argued that, in dilute solutions, ELP diblocks self-assemble into the so-called weak spherical micelles with dense cores and almost unstretched coronas.60 The authors attributed the unique features of ELP micelles to their low surface tension between the core and corona. At phase separation conditions, whether similar structural features are preserved in condensates remains unclear.
We found that most ELP condensates undergo a microphase separation. Representative configurations for four typical systems from MARTINI simulations are shown in the top panels of Fig. 2, with the V blocks in blue and X in red. Overall, the microphase separation manages to orient the more hydrophobic blocks towards the condensate interior as opposed to the interface. The relative density of guest blocks near the center of the condensate positively correlates with their hydrophobicity, as can be seen in both MARTINI (Extended Data Fig. 1) and all-atom (Extended Data Fig. 2) simulations.
![](https://prod--epp.elifesciences.org/iiif/2/90750%2Fv1%2Fcontent%2F534967v1_fig2.tif/full/max/0/default.jpg)
Internal organization of ELP condensates for (A) V5F5, (B) V5L5, (C) V5A5, and (D) V5G5. The uppermost panels present representative configurations from MARTINI simulations for each system. Periodic images along the x and y dimensions are shown for clarity, and the condensate-water interface is perpendicular to the z-axis. Only proteins are shown, with the X- and V-substituted halves of the peptides shown in red and blue, respectively. The central panels show contact maps between amino acids from different peptides. The blue and red bars indicate the V- and X-substituted half of the peptides. The lower panels plot the radial distribution functions, g(r), for amino acids only from the V-substituted half of the peptides (V5-V5), only from the X-substituted half of the peptides (X5-X5), and between the two halves (V5-X5). We limited the calculations to amino acid pairs from different peptides.
Microphase separation produced condensate morphologies deviating from homogeneous mixtures. In particular, the condensates appear to form gyroid-like structures (Extended Data Fig. 3A-D), in which the V and X blocks are intertwined. Such a morphology promotes interfacial contacts while maintaining substantial self-interactions as well. These qualitative observations are insensitive to the polymer length and system size in our simulations (Fig. S1 in the Supplementary Information). While longer diblock copolymers drive more prominent microphase separation, similar gyroid structures can also be observed. Notably, for more hydrophilic X blocks, the condensates begin to dissolve into a collection of micelle-like structures, consistent with the predictions by Hassouneh et al. in dilute solutions.60
![](https://prod--epp.elifesciences.org/iiif/2/90750%2Fv1%2Fcontent%2F534967v1_fig3.tif/full/max/0/default.jpg)
Experimental support of the microphase separation of ELP condensates. (A) Reversible ELP condensates formation via changing temperature. NHS ester fluorophores are attached at the amino-termini of V30X30 and X30V30 to detect the different microphysicochemical properties between the V-end and the X-end. (B) Structures of NHS-BODIPY and NHS-SBD, and FLIM images of V30A30, A30V30, V30G30 and G30V30 labeled with respective fluorophores. Scale bar: 5 µm. (C) Schematic diagram of optical tweezers stretching experiment. (D) Bright field images of V30A30 and A30V30 in the stretching experiment using the optical tweezer. Scale bar: 2 µm. (E) Normalized droplets fluorescence intensity changes while stretching (red curve: V30A30; blue curve: A30V30). Fifteen droplets
Quantitative characterization of the condensate interior supports gyroid-like organization as well. For example, significant contacts between V and X blocks can be readily seen in the inter-chain contact maps (Fig. 2, Extended Data Fig. 3E-X). We further computed the radial distribution functions for amino acids in various blocks (Fig. 2, Extended Data Fig. 4). For all systems, we found that the probability of finding cross-block contacts V-X is often comparable to the intra-block contacts between X-X or V-V, depending on the relative hydrophobicity of V and X. For example, the more hydrophobic F blocks exhibit strong self-clustering, much more prominent than V-V blocks that are comparable to V-F blocks. On the other hand, clustering among V-V blocks and V-A blocks is more substantial than among A-A blocks. These results are again robust with respect to system setups (Fig. S1 in the Supplementary Information).
![](https://prod--epp.elifesciences.org/iiif/2/90750%2Fv1%2Fcontent%2F534967v1_fig4.tif/full/max/0/default.jpg)
Interior of ELP condensates exhibits interfacial properties as a result of frustrated microphase separation. (A) Pearson correlation coefficient, ρ, between the stated measure of hydrophobicity and the condensate critical temperature, Tt, measured by Urry.41 To remove discrepencies in which end of the scale is hydrophobic and which end is hydrophilic, all hydrophobicity scales, including the Urry scale, are first normalized such that 1 corresponds to the most hydrophobic residue and 0 corresponds to the least hydrophobic residue. Hydrophobicity measures considered include experimental measures of water-solvent transfer free energies (water-1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine[POPC]-interface,65 water-octanol1,66 water-octanol2,67 water-ethanol,67, 68 water-cyclohexane,69 and water-N-cyclohexyl-2-pyrrolidone[CHP]70), atomic level analysis of moieties within amino acids (Kapcha),71 computational approaches to estimate hydrophobicity within disordered proteins (IDPs172 and IDPs273), bioinformatics techniques that approximate the burial propensity, or relative solvent accessibility (RSA) of amino acids (RSA174 and RSA275), and a method that mixed protein burial fraction with water-vapor transfer free energy (RSA/water-vapor76). The red box highlights the high correlation between Tt and water-octanol or water-POPC-interface transfer free energies. (B, C) Condensate organization from MAR-TINI simulations for V5F5 (B) and V5A5 (C). The top panels provide protein-only views for simulated condensates, with the guest residue X, glycine adjacent to the guest residue, and the remaining residues colored red, green, and blue, respectively. Condensate images are repeated periodically in the x/y plane. The bottom panel shows the overall radial distribution function from the guest amino acid to those amino acids native to the ELP sequence. (at size ∼ 4µm) were imaged and used for statistical analysis.
To experimentally test the microphase separation behavior uncovered in simulations, we studied the micro-physicochemical properties of the V-end and X-end of the peptides. We constructed diblock peptides with the combination of 30 pentameric repeats of V block and X (A or G) block, namely V30A30 and V30G30 (Experimental Sequences Section in the Supplementary Information). The amino-termini of V30A30 and V30G30 sequences were sub-sequently labeled with environmentally sensitive BODIPY or SBD fluorophores,61, 62 whose lifetime could be measured to quantify the viscosity or polarity of the V-end (Fig. 3A, left panel).63 We also constructed A30V30 and G30V30 diblock peptides, wherein the viscosity or polarity of the A-end or the G-end could be measured by fluorophores that are attached at the amino-terminus (Fig. 3A, right panel). Using FLIM, we found that the lifetime of BODIPY for the V-end (5.69 ns) was longer than that for the A-end (4.54 ns), suggesting that the V-end indeed has a higher microviscosity than the A-end (ηV = 2730.50 cp vs ηA= 1122.89 cp). Accordingly, the lifetime of SBD was longer for the V-end (10.30 ns) than the A-end (7.83 ns), indicating that the micropolarity of the V-end was lower than the A-end (ɛV = 9.64 vs ɛA = 16.00). These observations could be largely attributed to the greater extent of dehydration at the V-end due to its higher local peptide density. We further showed that the observed differences are not results of possible artifacts arising from any subtle distinctions between the two sequences V30A30 and A30V30 (Experimental Characterization of ELP Condensates Section in the Supplementary Information, Extended Data Fig. 5, and Extended Data Fig. 6). Similar results were observed using the V-G sequences. FLIM experiments revealed that the V-end was more viscous than the G-end (ηV = 3236.44 cp vs ηG= 2216.35 cp) and the V-end was less polar than the G-end (ɛV = 6.41 vs ɛG = 10.57). These experimental observations provided the first line of evidence to support the microphase separation, as suggested by the simulation results.
![](https://prod--epp.elifesciences.org/iiif/2/90750%2Fv1%2Fcontent%2F534967v1_fig5.tif/full/max/0/default.jpg)
Water hydrogen bonding environment of condensates from all-atom simulations. (A) Bar chart depicting the average number of protein-water (PW, in green) and protein-protein (PP, in red) hydrogen bond (HB) per residue for each condensate system. The x-axis presents the guest residue of the system and is ordered by the Urry scale. Error bars represent standard deviations of four independent estimates. (B) Water HB density for each of the 20 condensates, for three different residue selections. The metric is shown for all residues in the system (blue), guest residues (purple), and glycine residues adjacent to the guest residue in the sequence (orange). The x-axis presents the guest residue of the system and is ordered by the Urry scale. Error bars represent standard deviations of four independent estimates. (C) Correlation coefficients of the water HB density shown in part B with Urry Tt. (D) Schematics depicting two systems with low (i) and high (ii) water hydrogen bond density. The protein is depicted as blue and orange lines, and the water molecules are depicted as purple circles. Protein-protein and protein-water hydrogen bonds are drawn as red and green dashed lines respectively. (E) Visualizing intra-residue frustration and hydrogen bond density with illustrative snapshots for V5F5 (i) and V5G5 (ii) condensate. Side chain carbon atoms are shown in purple. Water molecules near the protein are explicitly depicted, with oxygen and hydrogen atoms colored in red and white. Water molecules hydrogen bonded to the protein are depicted as spheres (with hydrogen bonds shown in green lines).
![](https://prod--epp.elifesciences.org/iiif/2/90750%2Fv1%2Fcontent%2F534967v1_fig6.tif/full/max/0/default.jpg)
Comparison of protein conformation and solvation between ELP condensates and monomers. (A) Zoom in view of a single peptide chain (red) from atomistic simulations of V5A5 condensate (left) and monomer (right). Atoms within one nm of the peptide are shown, including water (cyan), chlorine (green), sodium (orange), and protein atoms from other chains (gray). (B) RSA of guest residues estimated from atomistic simulations of condensates (blue) and monomers (orange). For comparison, the corresponding values estimated using folded proteins are included as purple dots.75 Error bars represent the standard deviation of four independent time estimates. (C) Comparison of the radius of gyration (Rg) for peptides estimated from atomistic simulations of condensates (blue) and monomers (orange). The dashed line represents the expected Rg of an ideal chain, utilizing values that have been previously suggested for IDPs.81, 82 Error bars represent the standard deviation of four independent time estimates.
To further investigate these phenomena in real-time in motion, optical tweezers were used to stretch the droplet to observe the viscosity changes of different blocks. A single diblock condensate was captured by two optical traps (trap 1 and 2) at the same position, and it was stretched as trap 1 moved away from the original position (Fig. 3C-D). We used the peptides V30A30 and A30V30, with BODIPY labeled at their amino-terminus, to study the microviscosity changes where the fluorescence intensity decreases as the viscosity decrease. We found that the normalized fluorescence intensity of the V-end decreases greater than the A-end, indicating that the viscosity change of the V-end is larger (Fig. 3E). We anticipate that during stretching, condensates deform and molecules may deviate from the preferred chemical environment established with microphase separation. Such deviations impact the hydrophobic segments more, leading to more significant reduction in their interactions with surrounding molecules and correspondingly their viscosity.
Frustration produces condensate interior with interfacial properties
The strong dependence of molecular organization on amino acid hydrophobicity suggests that the solvation environment of individual residues might be a determining factor for condensate stability. To explore their relationship more quantitatively, we compared the critical temperature for ELP condensates measured by Urry41 to several hydrophobicity scales. Fig. 4A shows that water-octanol and water-POPC-interface transfer free energies best correlate with the critical temperature. Meanwhile, other hydrophobicity measures, even including those derived specifically for disordered proteins, do not match as well. We note that our conclusion here is not dependent on the limited number of hydrophobicity measures examined in this study, and can be seen in a previous clustering of 98 different hydrophobicity scales.64
The correlation analysis of the Urry critical temperature with various hydrophobicity scales supports a unique chemical environment emerging from the collective behaviors of condensates. To better understand the microenvironment surrounding guest residues in X blocks, we computed the radial distribution functions for them. As shown in Fig. 4B-C and Extended Data Fig. 7, the guest residues are immediately surrounded with glycine as the nearest neighbors. In the second shell, valine and proline residues appear. These profiles reveal the interfacial environment of the residues, which is consistent with a strong correlation of the Urry scale with water-POPC-interface transfer free energy. Therefore, ELP condensates do not simply feature a hydrophobic interior that separates from hydrophilic residues as seen for folded proteins, and hence the poor correlation shown in Fig. 4A with the burial propensity of amino acids (RSA1, RSA2).
The ELP peptide sequences partly dictate the interfacial environment. For example, the guest residues are juxtaposed between two glycine residues in the pentamer (Fig. 1A), creating the immediate shell of hydrophilic residues. However, the presence of valine in the pentamer prevents the movement of GXG motifs to the interface due to the clustering of hydrophobic residues that prefer the condensate interior. Therefore, while microphase separation occurs in ELP condensates, it is frustrated and cannot completely separate hydrophilic residues from hydrophobic ones due to constraints imposed by the amino acid sequence, creating unique microenvironments. Such a microenvironment arises from the collective behavior of many proteins and can deviate from individual chains, 77 explaining the less-than-ideal correlation between the Urry scale and the hydrophobicity scales optimized to reproduce the conformational ensemble of IDP monomers.
Solvation environment from atomistic simulations support interfacial properties
The interfacial environment uncovered in the previous section suggests the presence of significant solvation in the condensate interior. For condensates with lower surface tension, such as V5G5, the protein density near the center of the condensate is only around ∼ 30%. Even for the stable condensates with high surface tension, such as V5F5, the protein density near the center only reaches ∼ 70%. The relative density of protein, water, and ions in different systems generally follows the trend observed for the critical temperature (Extended Data Fig. 8 and Extended Data Fig. 9).
The large water mass fractions in the condensates prompt the question of the extent to which water is engaged in meeting the protein’s hydrogen bonding requirements. In particular, consistent with prior studies,42, 78, 79 ELPs remain largely disordered in the condensates without noticeable secondary structures (Fig. S2 in the Supplementary Information). Therefore, the exposed backbone lacks intrapeptide hydrogen bonding and must contact other peptides or water molecules. We determined the average number of hydrogen bonds per residue with ELP residues or water. As shown in Fig. 5A, over 75% of the hydrogen bonds are between protein and water, in all of the condensates. Prior work on a different ELP condensate42 similarly indicated that water-protein hydrogen bonds constitute most of the hydrogen bonds found in the system. Furthermore, comparison to a study of proteinprotein hydrogen bonding in X-ray structures reveals that there were over five times as many protein-protein hydrogen bonds per residue in these presumably quite ordered structures,80 compared to in the ELP condensates. This finding highlights the protein is less capable of meeting its own hydrogen bonding needs in ELP condensates, as compared to folded proteins. Therefore, high water concentration within the condensate is crucial to fulfilling the protein’s hydrogen bonding requirements.
Strikingly, given the importance of water molecules, the average number of hydrogen bonds per residue remained relatively constant across condensates, even when the amount of water inside the condensate varied by almost two folds. Therefore, we hypothesize that each water molecule in more hydrophobic condensates must engage more in hydrogen bonding than each water in hydrophilic condensates. To test this hypothesis, we determined the water hydrogen bond density, defined as the ratio of the number of protein-water hydrogen bonds divided by the total number of water molecules near the condensate (Fig. 5D and 5E). As shown in Fig. 5B and 5C, the water hydrogen bond density indeed negatively correlates with the condensate critical temperature, Tt, supporting our hypothesis that more hydrophobic condensates would have a higher hydrogen bond density.
The water hydrogen bond density also highlights the frustration in an interfacial environment. For example, when we limit the analysis to water molecules near the guest residues, X, the hydrogen bond density now becomes positively correlated with Tt. Therefore, locally, protein-water interactions are strongly influenced by the chemistry of side chains, and water molecules have fewer hydrogen bonding opportunities with more hydrophobic residues. However, the opposite trend is observed for the hydrogen bond density near glycine residues that immediately surround the guest residues. Therefore, the exposed backbones retain water molecules even when hydrophobic side chains do not provide hydrogen bonding opportunities, and the inseparability between the two further contributes to the observed interfacial property of condensates.
As a complementary measure of the solvation environment of individual residues, we computed the relative solvent accessibility (RSA) for the guest amino acids using atomistic simulations (Fig. 6A-B). The RSA measures the solvent-accessible surface area normalized by the maximum possible solvent exposure. RSA values in ELP condensates are much lower than those computed from all-atom simulations of ELP monomers, consistent with the increase of polymer density upon phase separation. However, they are almost always higher than the values estimated for folded proteins due to many water molecules roaming inside the condensates.75 Therefore, while ELP condensates provide a mechanism to shield hydrophobic amino acids from solvent, they are less effective in doing so than folded proteins. While the amino acid RSA values increase upon phase separation, unlike protein folding, the increase does not result from polymer collapse. We observe a significant expansion for almost all the studied peptides upon condensation, supporting the notion that water is a poor solvent for ELP (Fig. 6C). The expansion is most evident for more hydrophobic ELPs, the radii of gyration of which are now comparable to that of an ideal Gaussian chain. Notably, the sizes of the two separate peptide blocks are similar in most systems, supporting an interfacial environment that can accommodate both hydrophobic and hydrophilic residues (Extended Data Fig. 10).
CONCLUSIONS and DISCUSSION
We carried out multiscale simulations to elucidate the connection between protein sequences and emergent properties of ELP condensates. The multiscale approach overcomes challenges in sampling slow conformational rearrangements to produce equilibrium atomistic condensate configurations. The simulations successfully reproduced condensate stability variation upon amino acid substitution. We found that diblock ELPs undergo microphase separation, with the more hydrophobic blocks localizing towards the condensate interior. The more hydrophobic blocks increase the local protein density, resulting in higher viscosity and lower dielectric constants, as determined in FLIM experiments. Unlike the weak micelles formed in solution, the condensates exhibit gyroid-like morphology that promotes contact between blocks. In contrast to the interior of folded proteins, ELP condensates are significantly solvated with water occupying over 30% of mass density.
Our study uncovered a remarkably simple grammar for determining the stability of ELP condensates. A single quantity specific to the guest residue suffices to predict the emergent properties of the condensate, as evidenced by the strong correlation of the critical temperatures with the transfer free energies. The lack of higher-order effects is striking, supporting a mean field approximation to account for the contribution of individual residues in the system. However, the hydrophobicity scale is itself sensitive to the local microenvironment that is sequence specific. Such a microenvironment arises from the collective behavior of many proteins and can deviate from that of individual chains. Future work on systems with double amino acid substitutions could elucidate the generality of the mean field interpretation and the additivity of individual contributions.
While the transferability of the extracted rule for condensate stability to other systems remains to be shown, there are reasons to be hopeful. For instance, we anticipate that the frustrated microphase separation that produces the interfacial properties is a general principle for condensate organization. For most IDPs, the hydrophobic and hydrophilic residues are often interspersed throughout the chain to prevent the collapse of individual proteins83, 84 and, correspondingly, the complete phase separation as in the lamellar phase. In addition, the hydrophilic backbones are expected to be exposed due to a lack of secondary structure formation, contributing to the interfacial property in the presence of hydrophobic side chains. As such, we anticipate that future research will discover the broad presence of the frustrated microphase separation motif observed here.
Methods
Molecular dynamics simulation details
To better understand the properties of biological condensates, we performed simulations on 40 diblock ELPs with the sequence (V-P-G-V-G)n-(V-P-G-X-G)n. We set n = 5 to balance computational efficiency and polymer topological effects. As shown in Fig. S1 in the Supplementary Information, the results presented in the main text are qualitatively robust with respect to the system size and peptide length.
The multiscale simulations of ELP condensates started with an implicit solvent model, MOFF,45 and a similar homopolymer model, to generate initial configurations for each system. Next, we converted the α−carbon only configurations into those consistent with the MARTINI force field50 for coarse-grained, explicit solvent simulations. These simulations measured the coexistence between the condensate and solvent, lasted for 100 µs, and were conducted in the NPNAT ensemble. Finally, we converted the MARTINI configurations into atomistic structures for simulations with CHARMM36m85 force field, with protein-water interactions shifted to better characterize IDPs. Atomistic simulations were conducted in the NPNAT ensemble and lasted 250 ns. All simulations were conducted using the software GROMACS86 and are described in detail in the Condensate simulation details section of the Supplementary Information.
In addition to the condensates, we performed atomistic simulations of ELP monomers. These simulations began from RoseTTAFold predictions of the peptide structure 87, 88 and lasted for 1 µs in the NPT ensemble. Full details are available in the Monomer simulation details section of the Supplementary Information.
Analysis of condensate simulations
Surface tension
In explicit solvent simulations, two coexisting phases are present, in which proteins form a dense slab in the x-y plane, and the solvent creates a dilute phase in the z-dimension.89, 90 We controlled the pressure normal to the protein-solvent interface by performing simulations in the NPNAT ensemble.54 The surface tension (τ) was calculated from these simulations as
where Lz is the length of the Z-dimension and N is the number of protein droplets. PXX, PYY, and PZZ are the components of the pressure tensor along the three axes.
Clustering
Since the condensate can diffuse along the z-axis, i.e., the direction perpendicular to the interface, we aligned the simulated configurations using the position of the largest cluster of peptide molecules. To find the largest cluster, we first computed a center of mass contact matrix between peptide molecules with a distance cutoff of 4 nm. Then, using a depth first search algorithm,91 we identified the size and location of the largest protein cluster. The system was then shifted such that the z coordinate of the center of mass of the largest cluster is at zero. We implemented the alignment using the software MDAnalysis.92, 93
Contact map
Protein contact maps were calculated using a cutoff distance of 0.6 nm.
Specifically, two residues were defined as in contact if any pair of their coarse-grained beads were within 0.6 nm.
Radial distribution function
Radial distribution functions were calculated using the InterRDF function of MDAnalysis between 0.00001 nm and 2 nm. For analyzing the microphase separation of each condensate, we divided each protein into the V-substituted blocks and X-substituted blocks. This radial distribution function excluded intra-chain contacts to better capture how inter-chain contacts determine condensate organization. To analyze the origin of frustration within diblock phase separation, we computed the radial distribution function of each guest amino acid to the amino acids native to the sequence. Intra-chain contacts were included to fully capture the local environment experienced by the guest amino acid.
Relative solvent accessibility
The relative solvent accessibility (RSA) was measured based on the solvent accessible surface area (SASA) of guest amino acids determined from all atom simulations. We first calculated the SASA for each residue using GROMACS tools sasa, with a probe radius of 0.14 nm.94 We then averaged across all repeats of guest amino acids, and normalized the SASA by the maximum possible value for a given amino acid to compute RSA75 (see Table S1).
Radius of gyration
The estimated Rg from atomistic simulations were compared to the following analytical expression81
where N is the length of the polymer, b = 0.38nm is the bond length, and lp = 0.40nm is the persistence length.
Secondary structure
Secondary structure for protein condensates was calculated from atomistic simulations using the GROMACS do dssp command.95, 96 Figure S2 displays the ordered fraction of secondary structures, which includes α-helices, β-sheets, β-bridges, and turns.
Hydrogen bonding
Protein-protein and protein-water hydrogen bonds were determined using the hydrogenbonds package within MDAnalysis.92, 93, 97 A hydrogen bond was defined as having a donor heavy atom within 3.30 A of an acceptor heavy atom and a donor heavy atom-hydrogen atom-acceptor heavy atom angle of 135.0◦ or larger. The nearby water count for each analysis selection was determined by counting the water molecules whose oxygen atoms were within 4.0 A of the heavy atoms of this selection. Water hydrogen bond density for each residue selection for each frame was computed as the ratio of protein-water hydrogen bonds to the number of nearby water molecules (Fig. S3 in the Supplementary Information).
Experimental methods
Fluorescence lifetime imaging microscopy (FLIM)
The proteins of interest were labeled with NHS ester fluorophore. Droplets were formed with the final concentration of 70 µM ELP in 2 M NaCl or NH4SO4 for V-A and V-G diblocks, respectively. A drop of droplets containing solution was placed on a 0.17 mm coverslip with a 500 µm spacer. Images were acquired by Leica Falcon Fluorescence Microscope equipped with Wil pulse laser and 63X/0.12 oil-immersion objective. The BODIPY was excited at 488 nm and the SBD was excited at 448 nm. The fluorescence lifetime fitting and image analysis were performed in LAS X and Image J.
Optical tweezer experiments
The protocol describes the preparation and manipulation of a sample containing a BODIPY-labeled ELP protein using an optical tweezer, specifically the C-Trap from LULICKS. The goal is to form a single large condensate from small droplets and then stretch it using the optical tweezers while monitoring its fluorescence intensity. The first step is to mix the BODIPY-labeled ELP protein with NaCl solution to obtain a final concentration of 20 µM protein in 2 M NaCl. The sample is then loaded into the flow channel of the optical tweezer. The dual trap mode of the C-Trap is then used to move the two traps to the same position. The sample is flushed until a single large condensate of about 4 µm is caught by the dual trap via the fusion of small droplets. The traps are then moved along with the droplet to another flow channel containing 2 M NaCl buffer. The channel and pressure are turned off to avoid any additional force on the droplet. The next step is to stretch the droplet by moving trap 1 0.2 µm each time. Fluorescent images of each stretch are taken using a 488 nm excitation laser. The mean fluorescence intensity of the whole droplet is then quantified using Image J.
Acknowledgements
This work was supported by the National Institutes of Health (Grant R35GM133580, B.Z.), the Research Center for Industries of the Future (RCIF) at Westlake University (X.Z), PEW Biomedical Scholars Program 00033066 (X.Z.), and the National Science Foundation under CHE-1654415 (A.P.W.). A.L. acknowledges support from the National Science Foundation Graduate Research Fellowship Program (grant number 1745302). D.S. acknowledges support from the Hertz Foundation and the National Science Foundation Graduate Research Fellowship Program (grant number 2141064).
References
- (1.)Germline P Granules Are Liquid Droplets That Localize by Controlled Dissolution/CondensationScience 324:1729–1732
- (2.)Regulation of biomolecular condensates by interfacial protein clustersScience 373:1218–1224
- (3.)A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease MutationCell 162:1066–1077
- (4.)A Membraneless Organelle Associated with the Endoplasmic Reticulum Enables 3’UTR-Mediated Protein-Protein InteractionsCell 175:1492–1506
- (5.)Coexisting Liquid Phases Underlie Nucleolar SubcompartmentsCell 165:1686–1697
- (6.)Liquid droplet formation by HP1α suggests a role for phase separation in heterochromatinNature 547:236–240
- (7.)Phase separation drives heterochromatin domain formationNature 547:241–245
- (8.)Coactivator condensation at super-enhancers links phase separation and gene controlScience 361
- (9.)Single-stranded nucleic acid binding and coacervation by linker histone H1Nat. Struct. Mol. Biol 29:463–471
- (10.)On the stability and layered organization of protein-DNA condensatesBiophys. J 121:1727–1737
- (11.)Biomolecular condensates: Organizers of cellular biochemistryNat. Rev. Mol. Cell Biol 18:285–298
- (12.)Intrinsically disordered proteins in overcrowded milieu: Membrane-less organelles, phase separation, and intrinsic disorderCurr. Opin. Struct. Biol 44:18–30
- (13.)Stress-Triggered Phase Separation Is an Adaptive, Evolutionarily Tuned ResponseCell 168:1028–1040
- (14.)Multiscale Modeling of Genome Organization with Maximum Entropy OptimizationJ. Chem. Phys 155
- (15.)Biomolecular Condensates in the NucleusTrends Biochem. Sci :961–977
- (16.)A phase separation model predicts key features of transcriptional controlCell 169:13–23
- (17.)Composition-dependent thermodynamics of intracellular phase separationNature 581:209–214
- (18.)Partitioning of cancer therapeutics in nuclear condensatesScience 368:1386–1392
- (19.)Liquid-Liquid Phase Separation in BiologyAnnu. Rev. Cell Dev. Biol 30:39–58
- (20.)Protein Phase Separation: A New Phase in Cell BiologyTrends Cell Biol 28:420–435
- (21.)Conformational Dynamics of Intrinsically Disordered Proteins Regulate Biomolecular Condensate ChemistryChem. Rev 122:6719–6748
- (22.)Biomolecular Phase Separation : From Molecular Driving Forces to Macroscopic PropertiesAnnu. Rev. Phys. Chem 71:1–23
- (23.)Comparative roles of charge, π, and hydrophobic interactions in sequence-dependent phase separation of intrinsically disordered proteinsProc. Natl. Acad. Sci. U.S.A 117:28795–28805
- (24.)Molecular interactions contributing to FUS SYGQ LC-RGG phase separation and co-partitioning with RNA polymerase II heptadsNat. Struct. Mol. Biol 28:923–935
- (25.)Thermodynamics of high polymer solutionsJ. Chem. Phys 10
- (26.)Biomolecular Condensates: Sequence Determinants of Phase Separation, Microstructural Organization, Enzymatic Activity, and Material PropertiesJ. Phys. Chem. B 125:3441–3451
- (27.)Physical Principles Underlying the Complex Biology of Intracellular Phase TransitionsAnnu. Rev. Biophys 49:107–133
- (28.)A Molecular Grammar Governing the Driving Forces for Phase Separation of Prion-like RNA Binding ProteinsCell 174:688–699
- (29.)Elucidating binding mechanisms and dynamics of intrinsically disordered protein complexes using NMR spectroscopyCurr. Opin. Struct. Biol 54:10–18
- (30.)The disordered P granule protein LAF-1 drives phase separation into droplets with tunable viscosity and dynamicsProc. Natl. Acad. Sci. U.S.A 112:7189–7194
- (31.)Learning the chemical grammar of biomolecular condensatesNat. Chem. Biol
- (32.)Molecular Determinants for the Layering and Coarsening of Biological CondensatesAggregate
- (33.)A conceptual framework for understanding phase separation and addressing open questions and challengesMol. Cell 82:2201–2214
- (34.)Phase-separating RNA-binding proteins form heterogeneous distributions of clusters in subsaturated solutionsProc. Natl. Acad. Sci. U.S.A 119
- (35.)Single fluorogen imaging reveals spatial inhomogeneities within biomolecular condensatesbioRxiv
- (36.)Theory of Thermoreversible GelationMacromolecules 22:1988–1994
- (37.)Thermoreversible Gelation in Solutions of Associative Polymers. 1. StaticsMacromolecules 31:1373–1385
- (38.)Intrinsically disordered linkers determine the interplay between phase separation and gelation in multivalent proteinsElife 6:1–31
- (39.)Theory of Microphase Separation in Block CopolymersMacromolecules 13:1602–1617
- (40.)Block Copolymer Thermodynamics: Theory and ExperimentAnnu. Rev. Phys. Chem 41:525–557
- (41.)Physical chemistry of biological free energy transduction as demonstrated by elastic protein-based polymersJ. Phys. Chem. B 101:11007–11028
- (42.)The liquid structure of elastineLife 6
- (43.)Thermal Compaction of Disordered and Elastin-like Polypeptides: A Temperature-Dependent, Sequence-Specific Coarse-Grained Simulation ModelBiomacromolecules 21:3523–3538
- (44.)Recursive directional ligation by plasmid reconstruction allows rapid and seamless cloning of oligomeric genesBiomacromolecules 11:944–952
- (45.)Consistent Force Field Captures Homologue-Resolved HP1 Phase SeparationJ. Chem. Theory Comput 17:3134–3144
- (46.)Improving Coarse-Grained Protein Force Fields with Small-Angle X-ray Scattering DataJ. Phys. Chem. B 123:1026–1034
- (47.)Maximum Entropy Optimized Force Field for Intrinsically Disordered ProteinsJ. Chem. Theory Comput 16:773–781
- (48.)Unifying coarse-grained force fields for folded and disordered proteinsCurr. Opin. Struct. Biol 72:63–70
- (49.)The MARTINI force field: Coarse grained model for biomolecular simulationsJ. Phys. Chem. B 111:7812–7824
- (50.)Martini 3: a general purpose force field for coarse-grained molecular dynamicsNat. Methods 18:382–388
- (51.)Improving Martini 3 for Disordered and Multidomain ProteinsJ. Chem. Theory Comput 18:2033–2041
- (52.)Simulation of FUS Protein Condensates with an Adapted Coarse-Grained ModelJ. Chem. Theory Comput 17:525–537
- (53.)Coacervate formation studied by explicit solvent coarse-grain molecular dynamics with the MartiniChem. Sci
- (54.)Computer simulation of liquid/liquid interfacesI. Theory and application to octane/water. J. Chem. Phys 103:10252–10266
- (55.)Polymer Physics
- (56.)Micropolarity governs the structural organization of biomolecular condensatesBioRxiv :1–20
- (57.)Block copolymers-designer soft materialsPhys. Today 52:32–38
- (58.)Self-assembly of block copolymersChem. Soc. Rev 41:5969–5985
- (59.)Diblock copolymer micelles in a dilute solutionMacromolecules 38:5330–5351
- (60.)Elastin-like Polypeptide Diblock Copolymers Self-Assemble into Weak MicellesMacromolecules 48:4183–4195
- (61.)AgHalo: A Facile Fluorogenic Sensor to Detect Drug-Induced Proteome StressAngew. Chemie – Int. Ed. 56:8672–8676
- (62.)A dual-functional BODIPY-based molecular rotor probe reveals different viscosity of protein aggregates in live cellsAggregate :2–8
- (63.)An Optical Technique for Mapping Microviscosity Dynamics in Cellular OrganellesACS Nano 12:4398–4407
- (64.)50 years of amino acid hydrophobicity scales: Revisiting the capacity for peptide classificationBiol. Res 49:1–19
- (65.)Experimentally determined hydrophobicity scale for proteins at membrane interfacesNat. Struct. Biol 3:842–848
- (66.)Solvation energies of amino acid side chains and backbone in a family of host – Guest pentapeptidesBiochemistry 35:5109–5124
- (67.)Physicochemical Basis of Amino Acid Hydrophobicity Scales: Evaluation of Four New Scales of Amino Acid Hydrophobicity Coefficients Derived from RP-HPLC of PeptidesAnal. Chem 67:1210–1219
- (68.)The characterization of amino acid sequences in proteins by statistical methodsJ. Theor. Biol 21:170–201
- (69.)Comparing the Polarities of the Amino Acids: Side-Chain Distribution Coefficients between the Vapor Phase, Cyclohexane, 1-Octanol, and Neutral Aqueous SolutionBiochemistry :1664–1670
- (70.)A simple experimental model for hydrophobic interactions in proteinsJ. Biol. Chem 259:2910–2912
- (71.)A simple atomic-level hydrophobicity scale reveals protein interfacial structureJ. Mol. Biol 426:484–498
- (72.)Accurate model of liquid – liquid phase behavior of intrinsically disordered proteins from optimization of single-chain propertiesProc. Natl. Acad. Sci. U.S.A
- (73.)A Data-driven Hydrophobicity Scale for Predicting Liquid-Liquid Phase Separation of ProteinsJ. Phys. Chem. B 125:4046–4056
- (74.)Hydrophobicity of Amino Acid Residues in Globular ProteinsScience 229
- (75.)Maximum allowed solvent accessibilites of residues in proteinsPLoS ONE 8
- (76.)A simple method for displaying the hydropathic character of a proteinJ. Mol. Biol 157:105–132
- (77.)Phase behaviour of disordered proteins underlying low density and high permeability of liquid organellesNat. Chem 9
- (78.)Direct observation of structure and dynamics during phase separation of an elastomeric proteinProc. Natl. Acad. Sci. U.S.A 114:E4408–E4415
- (79.)Insights into Structure and Aggregation Behavior of Elastin-like Polypeptide Coacervates: All-Atom Molecular Dynamics SimulationsJ. Phys. Chem. B 125:8627–8635
- (80.)Hydrogen bonding in globular proteinsJ. Mol. Biol 226:1143–1159
- (81.)Polymer scaling laws of unfolded and intrinsically disordered proteins quantified with single-molecule spectroscopyProc. Natl. Acad. Sci. U.S.A 109:16155–16160
- (82.)Relation between single-molecule properties and phase behavior of intrinsically disordered proteinsProc. Natl. Acad. Sci. U.S.A 201804177
- (83.)Valence and patterning of aromatic residues determine the phase behavior of prion-like domainsScience 367:694–699
- (84.)Hydropathy Patterning Complements Charge Patterning to Describe Conformational Preferences of Disordered ProteinsJ. Phys. Chem. Lett 11:3408–3415
- (85.)CHARMM36m: an improved force field for folded and intrinsically disordered proteinsNat. Methods 14:71–73
- (86.)GROMACS: A message-passing parallel molecular dynamics implementationComput. Phys. Commun 91:43–56
- (87.)Accurate prediction of protein structures and interactions using a three-track neural networkScience 373:871–876
- (88.)Rosetta in CASP4: Progress in ab initio protein structure predictionProteins 45:119–126
- (89.)Sequence determinants of protein phase behavior from a coarse-grained modelPLoS Comput. Biol 14:1–23
- (90.)Triple-point coexistence properties of the lennard-jones systemChem. Phys. Lett 51:155–159
- (91.)Analyzing and Driving Cluster Formation in Atomistic SimulationsJ. Chem. Theory and Comput 13:1317–1327
- (92.)MDAnalysis: A Python Package for the Rapid Analysis of Molecular Dynamics SimulationsProceedings of the 15th Python in Science Conference
- (93.)MDAnalysis: a toolkit for the analysis of molecular dynamics simulationsJ. Comput. Chem 32:2319–2327
- (94.)The Double Cubic Lattice Method: Efficient Approaches to Numerical Integration of Surface Area and Volume and to Dot Surface Contouring of Molecular AssembliesJ. Comput. Chem 16:273–284
- (95.)A series of PDB related databases for everyday needsNucleic Acids Res 43:D364–D368
- (96.)Dictionary of Protein Secondary Structure: Pattern Recognition of Hydrogen-Bonded and GeometricalBiopolymers 22:2577–2637
- (97.)On the interaction of hyaluronic acid with synovial fluid lipid membranesPhys. Chem. Chem. Phys 21:9845–9857
Article and author information
Version history
- Preprint posted:
- Sent for peer review:
- Reviewed Preprint version 1:
- Reviewed Preprint version 2:
- Reviewed Preprint version 3:
Copyright
© 2023, Latham et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
- views
- 629
- downloads
- 40
- citations
- 0
Views, downloads and citations are aggregated across all versions of this paper published by eLife.