Screening system and method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction

Abstract

The present invention relates to a screening system for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction, the screening system comprising a first and second protein having affinity to each other, the system having a protein-protein association state and a protein-protein dissociation state, wherein switching between the both states is modulated depending on the presence, absence and / or concentration of the ligand, and results in a change of the screening system's spectroscopic properties. Wherein the system further comprises a condensate formed of condensate-forming molecules, wherein the condensate is configured to reversibly recruit the first and second protein independently, wherein the condensate facilitates the protein-protein association state, and wherein the first and / or second protein comprises a ligand-binding domain. The present invention further relates to a method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction and a condensate for use in such a method.

Description

[0001] Title: Screening system and method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction field

[0002] The present invention relates to a screening system for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction. The present invention further relates to a method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction and a condensate for use in such a method.

[0003] The identification and study of small molecule protein-protein interaction (PPI) stabilizers, also known as PPI stabilizing agents or molecular glues, is typically performed using biochemical assays such as fluorescence polarization or time- resolved fluorescence energy transfer (TR-FRET) in dilute solution or in cellular studies (using readouts such as BRET or functional assays). However, the characteristics and limitations of biochemical assay formats in solution are typically not well aligned with the screening needs for PPI stabilization (e.g. high protein concentrations, high compound concentrations, weak PPI affinities) and lack of properly reflecting the crowded environment within the cell, limiting biological relevance of the biochemical assay. Cellular studies are both time- and cost-intensive and the results are typically ambiguous due to complex read-out methods, potential off-target effects, and need for subsequent target identification.

[0004] Biological misregulation (loss or gain) of specific PPIs is often the cause of disease and hence therapeutic molecules that modulate PPIs, such as PPI stabilizing or inhibiting agents, are promising drug candidates. PPI modulating agents are small molecules that modulate the interactions between two or more proteins, either by direct binding at the protein-protein interface, or by allosteric mechanisms. Efficient screening to identify and study such PPI modulating agents is key for unlocking novel PPIs for drug discovery.

[0005] There is thus a need for providing a screening assay for studying the effect of PPI modulating agents on PPIs, wherein the screening assay overcomes the challenges identified above. Detailed description of the invention

[0006] In order to provide an efficient screening assay and to overcome the challenges of solution-based biochemical assays, the present invention provides hereto a screening system comprising a condensate that allows for a local high concentration of proteins that engage in PPIs within the condensate. By providing a condensate that allows for the uptake of proteins that engage in PPIs, the present invention herewith provides for an assay overcoming the challenges of solution-based biochemical assays, while at the same time mimicking the more molecular crowded environment of the cytoplasm and other cellular compartments, such as biomolecular condensates. The screening system of the present invention thus provides a unique environment to study ligands, using the stablizing effect of ligands on PPIs, such as small molecules, including PPI modulating agents, PPI stabilizing agents, drug candidates, and active pharmaceutical ingredients.

[0007] Condensates have been reported in literature. For example, the inventors described a coacervate-based artificial cell platform based on amylose-derived polyelectrolytes to reconstitute native interactions between client peptides and proteins and a 14-3-3 scaffold protein (see: van Veldhuisen, Thijs W., et al. “Enzymatic Regulation of Protein-Protein Interactions in Artificial Cells.” Advanced Materials 35.29 (2023): 2300947). The interactions of client peptides and the 14-3-3 scaffold protein were dependent on the client peptides’ phosphorylation state, which can be enzymatically regulated using phosphorylation and dephosphorylation processes. The scaffold protein 14-3-3 was efficiently immobilized in the coacervates, and it was shown that polypeptides with affinity motifs for 14-3-3 can be recruited to the coacervates based on their affinity for 14-3-3. Furthermore, a displacement experiment was used to demonstrate competitive binding between two peptides governed by their different affinity for 14-3-3. Besides peptides, proteins could be reversibly recruited by in situ phosphorylation using a kinase.

[0008] Other studies in literature reported the use of biomolecular condensates in the identification of condensate-modifying therapeutics (see, for example: Mitrea, Diana M., et al. “Modulating biomolecular condensates: a novel approach to drug discovery.” Nature Reviews Drug Discovery 21.11 (2022): 841-862). It is noted that the above-reported studies do focus on modulating the physical properties and / or the integrity of the condensate itself, by for example disassembly or reducing its size. The present invention, however, provides for a screening system comprising a condensate, wherein the condensate itself is used as a vehicle to study ligands using PPIs. In other words, the PPIs used in this system are modulated in the condensate without intentionally affecting the physical properties and integrity of the condensate itself. In other words, the proteins (also referred to as ‘protein constituents’) used in the screening system of the present invention do not affect the condensate constituents for forming the condensate. As such, the protein constituents, used in studying the ligand using PPIs, can be referred to as ‘protein cargo’ and wherein the modulation of the PPIs does not intentionally contribute to the structure, e.g., structural integrity, of the condensate. Given the condensate as used in the screening system of the present invention, the proteins can be therapeutically targeted without having a significant effect on physical properties of the condensate such as size, shape, viscosity, and the like. As such, the present invention does not relate to a screening system comprising biomolecular condensates wherein the protein or proteins part of the PPIs are an integral part of the structural constituents of the condensate, or in other words, wherein actual PPIs result in a physical change of the condensate.

[0009] In particular, in a first aspect, the present invention relates to a screening system for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction, wherein the screening system comprises: a first protein having at least one affinity motif; and a second protein having at least one affinity motif, said at least one affinity motif of the second protein having affinity towards said at least one affinity motif of the first protein and vice versa, wherein the screening system has a protein-protein association state in which the first and second proteins are associated and a protein-protein dissociation state in which the first and second proteins are not associated, and wherein switching between the protein-protein association state and the protein-protein dissociation state: is modulated depending on the presence, absence and / or concentration of the ligand; and results in a change of the screening system’s spectroscopic properties. Different to the screening assays provided in the prior art, the present invention differs in that the screening system of the present invention further comprises a condensate formed of condensate-forming molecules, said condensate is configured to reversibly recruit the first protein and the second protein independently, and further differs in that the condensate facilitates the formation of the protein-protein association state.

[0010] To provide a screening system that is configured to screen a ligand, the present invention provides for a ligand-dependent protein-protein interaction. To establish such a ligand-dependent protein-protein interaction, the screening system of the present invention further comprises a ligand-binding domain. The ligand-binding domain may be an allosteric ligand-binding domain.

[0011] Regarding the ligand-dependent protein-protein interaction it is noted that the first protein may comprise a ligand-binding domain. By providing a protein having a ligand-binding domain, the presence of a ligand will facilitate the PPI, allowing the screening system to switch to the protein-protein association state. In addition to the first protein comprising a ligand-binding domain, the second protein may also comprise a ligand-binding domain. Alternatively, the ligand-binding domain may be formed in the protein-protein association state. In such case, both proteins involved in the PPI do comprise complementary ligand-binding domains that form a ligand-binding domain in case both proteins bind to each other. One could understand that such alternative screening assay deviates from a screening assay wherein one of the proteins comprises a ligand-binding domain that is able to interact with a ligand on its own.

[0012] Given the screening system of the present invention, the present invention provides for an artificial dense and more cell-like environment that enables screening of the modulation of PPIs at significantly higher local concentrations than compared to solution-based biochemical assays. By providing a screening system comprising a condensate wherein the modulation of PPIs can be studied at significantly higher local concentrations, the screening system as provided by the present invention is able to study the frequently weak PPIs occurring in cells. As noted above, synthetic condensates such as the condensates used in the screening system of the present invention, have found applications in artificial cell research and research into the properties of native condensates. However, the use of a condensate in a screening system as an assay method for drug development has not been disclosed nor suggested in the art.

[0013] The term “screening system” as used herein may refer to an assay or method for screening ligands, such as PPI modulating agents. The term “screening” may refer, but is not limited to, the measuring of the effect of ligands, such as PPI modulating agents, on PPIs at a specific moment in time and / or may refer to monitoring the effect of PPI modulating agents on PPIs over time.

[0014] The term “condensate” as used herein may refer to biomolecular condensates, coacervates, and similar colloidal or organelle like structures. For example, a biomolecular condensate typically refers to an assembly of biomolecules into membrane-less organelles. The term “condensate” in a biomolecular condensate may refer to a dense ‘condensed’ phase rich in macromolecules, with higher macromolecular density than the surrounding environment. Biomolecular condensates are naturally occurring and can also be formed through diseased states or engineering of cells. A coacervate typically refers to a colloidal particle that is formed in aqueous medium by liquid-liquid phase separation. The term “coacervate” generally refers to a synthetic droplet, using purified macromolecules.

[0015] It is further noted that the condensate as used in the screening system of the present invention is adjustable in parameters such as charge ratio and type of multivalent polymer / protein building block. The composition of the condensate can be adjusted to facilitate a specific PPI. The condensate also allows for the adjustment of physical parameters such as ionic strength and pH, in contrast to the fixed environmental parameters in cellular studies.

[0016] The term “ligand” as used herein may refer to a small molecule, PPI modulating agent, PPI stabilizing agent, PPI destabilizing agent, drug candidate, active pharmaceutical ingredient, or the like. The ligand may also refer to a library of (candidate) PPI modulating agents. The ligand is able to interact with one or more of the proteins involved in the PPIs. The ligand is preferably selected from the group of compounds effecting PPIs. Ligands, also referred to as PPI modulating agents, may be referred to in the art as “molecular glues”. However, the term “molecular glues” excludes compounds having a PPI inhibiting effect. As such, the term “ligand” also includes compounds having an adverse or reducing effect on the PPIs. The term “spectroscopic detection of protein-protein interaction” as used herein refers to the detection of protein-protein interaction using spectroscopic assays. Such spectroscopic assays may include, but are not limited to, bioluminescence and fluorescence complementation assays, fluorescence anisotropy assays, fluorescence intensity, and fluorescence microscopy based assays (such as widefield or confocal microscopy).

[0017] The first protein may include a scaffold protein. Preferably the first protein is a scaffold protein having over hundreds of interactions with different second and further proteins, also referred to in the art as “client proteins”. A suitable first protein acting as a scaffold protein is the 14-3-3 protein. However, other types of scaffold proteins may serve as a first protein used in the screening system of the present invention as well. It is noted that a screening system using a scaffold protein as the first protein is in particular suitable for the screening of ligands acting as PPI stabilizing agents, i.e. , ligands improving protein-protein interaction. Alternatively, a screening system using a scaffold protein as the first protein may also be suitable for the screening of inhibitors or other potentially interesting specific target candidates.

[0018] Alternatively, the first protein may be selected from the group consisting of nuclear receptors, including, for example PPARy. In case the first protein is selected from the group consisting of nuclear receptors, the second or further proteins are selected from the group of coregulators. By providing a screening system wherein the first protein is selected from the group consisting of nuclear receptors, the present invention provides for a screening system suitable for screening ligands that act in a different manner compared to ligands acting as PPI stabilizing agents, e.g., studying ligands having PPI destabilizing activity.

[0019] As already noted above, the screening system may comprise further proteins. Such further proteins have further affinity motifs, wherein said further affinity motifs have affinity towards the at least one affinity motif of the first protein.

[0020] The condensates of the present invention are formed by condensate-forming molecules. The condensate-forming molecules, also referred to as macromolecules, complex and de-mix from solution by liquid-liquid phase separation, yielding droplets with sizes matching cells and / or their compartments. By providing condensates formed of condensate-forming molecules, the condensates used in the screening system of the present invention mimic the crowdedness of the cellular environment by having a high concentration of condensate-forming molecules, e.g., (condensate-forming) macromolecules. Preferably, the condensate may comprise condensate-forming molecules having protein immobilizing motifs, wherein said protein immobilizing motifs having affinity or reactivity towards corresponding immobilizing motifs comprised on at least one of the proteins to immobilize the at least one of the proteins in the condensate. By providing condensate-forming molecules having protein immobilizing motifs, one or more of the proteins, e.g., a scaffold protein or nuclear receptor, may be stably incorporated (immobilized) in the condensate. In such configuration, the protein-protein association state is directed to and facilitated within the condensate as such. By immobilizing at least one of the proteins in the condensate, the proteinprotein interaction can primarily take place in the condensate itself.

[0021] An example of the condensate of the present invention may include the mixing of positively charged amylose, preferably modified with quaternary ammonium groups (Q-Am) and negatively charged amylose, preferably modified with carboxymethyl groups (Cm-Am) and amylose with an immobilizing motif, preferably, amylose modified with Ni2+complexed nitrilo triacetate (Ni-NTA-Am) to form a condensate, subsequently followed by the loading of His-tagged 14-3-3 protein and the addition of terpolymer in order to stabilize the condensate with a semipermeable membrane. A suitable terpolymer for stabilizing the condensate may include the synthetic block copolymer membrane poly(ethylene glycol)-co-poly(caprolactone-gradient-trimethylene carbonate)-co-poly(glutamic acid) (PEG-(PCLgTMC)-PGIu).

[0022] In a preferred embodiment, the condensate used in the screening system of the present invention is designed such that the condensate comprises a controlled uptake mechanism for facilitating the controlled uptake of proteins to be immobilized in the condensate.

[0023] In a preferred embodiment of the present invention the first protein is immobilized in the condensate. By immobilizing the first protein in the condensate, the second (or further) protein remains mobile and is allowed to freely move in and out the condensate. In such case the protein-protein association state may be determined on the localization of the second (or further) protein, i.e., the mobile protein.

[0024] In another preferred embodiment of the present invention, both the first protein and the second protein are immobilized in the condensate. In such case the PPI is highly dependent on the presence of a ligand, as both proteins are readily available in the condensate during the screening. Typically, the protein-protein association state may be determined based on spectroscopic changes, such as via a bioluminescence complementation assay.

[0025] In an embodiment of the present invention, the first protein is immobilized in the condensate and, optionally, the second protein is immobilized in the condensate, and wherein the screening system is configured to emit light in the protein-protein association state. In an alternative embodiment of the present invention, the first protein is immobilized in the condensate, and wherein the screening system is configured to emit light in the protein-protein dissociation state.

[0026] To provide a screening system that is configured to emit light, each of the first protein and the second protein may be labelled with a fluorescent dye. Preferably, the screening system is configured such that the fluorescent dye of the first protein interacts with the fluorescent dye of the second protein in the protein-protein association state to emit light.

[0027] In a second aspect of the invention, the present invention relates to a method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction, wherein the method comprises the steps of: a) providing the screening system of the present invention; b) providing a ligand; c) measuring the screening system’s spectroscopic properties in the absence of the ligand; d) contacting the ligand provided in step b) with the screening system provided in step a); e) measuring the screening system’s spectroscopic properties in the presence of the ligand; f) determining the change in the screening system’s spectroscopic properties based on the measurements provided in steps c) and e); and g) determining the quantitative and / or qualitative change in protein-protein interaction modulated by the ligand based on the change in the screening system’s spectroscopic properties determined in step f). In a third aspect of the invention, the present invention relates to the use of a condensate in a method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction according to the present invention.

[0028] Experimental data

[0029] Materials and instruments

[0030] 14-3-3o and 14-3-3y were kindly provided by Maxime van den Oetelaar, Carlo Verhoef, and Marloes Pennings. The c-Raf pS259 peptide was a kind gift from Dr Emira Visser. The ERRy peptide and the ERa ligand binding domain were provided by Dr Bente Somsen. SSBP4, CIP2A, and RIPK2 peptides were provided by Yannick Leurs. Fusicoccin analogues were provided by Dr Peter Cossar and Dr Sebastian Andrei. ERRy reactive stabilizers were provided by Dr Peter Cossar and Siebe van der Elzen. Full chemical identity of 14-3-3-binding peptides is given in Table 1. The PPARy coregulators are shown in Table 2. EP300 was synthesized in house for which Justin Houx is thanked. PRIPRAP and MED1 DRIP2 were purchased from Thermo Scientific. Unless otherwise noted, all other reagents were obtained from commercial suppliers and used without further purification.1H NMR spectra were collected on an A VANCE III HD (400 MHz) NMR spectrometer (Bruker). The1H NMR chemical shift values are reported in ppm relative to the residual solvent peak.

[0031] DNA molecular biology and cloning

[0032] Protein sequences and physicochemical properties are given in Table 3. All DNA was ordered through Integrated DNA Technologies (IDT). The constructs were codon-optimized using IDT’s built-in codon optimization tool for Escherichia coli (E. coli). The pET28a vector and gBIock dsDNA fragments were digested with the appropriate restriction enzymes (New England Biolabs). After ligation into the vector, the constructs were verified using Sanger sequencing (Azenta). Constructs were transformed into BL21 (DE3) E. Coli cells (Novagen). Expression of 14-3-3y-SmBiT, LgBiT-ERRy, and mCherryEA

[0033] For protein expression, 1 L or 0.5 L of 2xYT medium supplemented with 30 pg mL-1of kanamycin was used. After inoculation using an overnight culture grown at 37 °C, 250 rpm, the culture was grown to an optical density (OD600) of 0.6 at 37 °C, 140 rpm. Then, protein expression was induced by addition of isopropyl p-D-1- thiogalactopyranoside (IPTG) at a final concentration of 0.5 mM, with incubation overnight at 20 °C, 140 rpm. Cells were harvested by centrifugation at 4°C and 10OOOxg for 15 minutes. The cell pellets were resuspended in lysis buffer (50 mM T ris, 300 mM NaCI, 30 mM imidazole, pH 8.0 supplemented with 1 pL / 10 mL benzonase). Cells were lysed using an EmulsiFlexC3 High-Pressure homogenizer (Avestin) at 15,000 psi for three consecutive rounds. Cell debris and insoluble proteins were removed by centrifugation at 4 °C and 35000xg for 20 minutes. His-tagged proteins were purified from the soluble lysate using Ni-NTA affinity chromatography (His-Bind Resin, Novagen). The lysate was loaded onto the His-bind resin and washed twice with wash buffer (50 mM Tris, 300 mM NaCI, 60 mM imidazole, pH 8.0). The His- tagged proteins were eluted from the resin using elution buffer (50 mM Tris, 300 mM NaCI, 250 mM imidazole, pH 8.0). The eluted fractions were analyzed using SDS- PAGE (4-20% Mini-PROTEAN TGX Precast Protein Gel, Bio-Rad) and the purest fractions were pooled. The combined fractions were extensively dialyzed against condensate buffer (20 mM HEPES, 100 mM KCI, pH 7.5, freshly prepared with 100 pM tris(2-carboxyethyl)phosphine (TCEP) using membrane tubing with a molecular weight cut-off (MWCO) of 12-14 kDa (Fisher Scientific). Protein concentration was determined using an ND-1000 spectrophotometer (Thermo Scientific) at 280 nm with theoretical extinction coefficients as determined by the Expasy ProtParam tool as shown in Table 2. The purified proteins were aliquoted, flash-frozen in liquid N2, and stored at -80°C for single-use aliquots. The identity and purity of the protein samples were confirmed using liquid chromatography quadrupole time of flight mass spectrometry (LC-MS Q- ToF).

[0034] Expression of PKA (catalytic subunit)

[0035] 2 L of TB auto-induction medium supplemented with kanamycin (30 pg mL-1) was inoculated using an overnight culture grown at 37 °C, 250 rpm. The culture was grown at 37 °C, 140 rpm, for 4 hours, after which protein expression was carried out overnight at 25 °C, 140 rpm. Cells were harvested by centrifugation at 4°C and 10000 xg for 15 minutes. The cell pellet was resuspended in lysis buffer (50 mM Tris, pH 8, 600 mM NaCI, 30 mM imidazole, 5% glycerol, pH 8.0 supplemented with 1 pL / 10 mL benzonase). Cells were lysed using an EmulsiFlexC3 High-Pressure homogenizer (Avestin) at 15,000 psi for three consecutive rounds. Cell debris was removed by centrifugation at 4 °C and 35000 xg for 20 minutes. PKA was purified from the soluble lysate using Ni-NTA affinity chromatography (His-Bind Resin, Novagen). The lysate was loaded onto the His-bind resin and washed twice with lysis buffer (50 mM Tris, pH 8, 600 mM NaCI, 30 mM imidazole, 5% glycerol, pH 8.0). The His-tagged proteins were eluted from the resin using elution buffer (50 mM Tris, pH 8, 600 mM NaCI, 250 mM imidazole, 5% glycerol, pH 8.0). Subsequently, the fractions were loaded on a preequilibrated 2 mL Strep-Tactin XT column (Iba Lifesciences). After 2 repeats of 5 column volumes of washing with wash buffer (100 mM Tris pH 8, 150 mM NaCI, 1 mM EDTA), the protein was eluted using freshly prepared Strep elution buffer (100 mM Tris pH 8, 150 mM NaCI, 1 mM EDTA, 50 mM biotin). The eluted fractions were analyzed using SDS-PAGE (4-20% Mini-PROTEAN TGX Precast Protein Gel, BioRad) and the purest fractions were pooled. The protein was extensively dialyzed against storage buffer (50 mM HEPES, 100 mM KCI, pH 7.5) using membrane tubing with a MWCO of 12-14 kDa (Fisher Scientific). Protein concentration was determined using an ND-1000 spectrophotometer (Thermo Scientific) at 280 nm with a theoretical extinction coefficient of 59270 M"1cm-1as determined by the Expasy ProtParam tool. The protein was aliquoted into single-use aliquots, flash-frozen in liquid N2, and stored at -80°C. The identity and purity of the protein sample were confirmed using LC-MS Q-ToF.

[0036] Expression of PPARy ligand binding domain

[0037] For the expression of PPARy LBD, 3x 2 L of TB medium supplemented with ampicillin (30 pg / mL) was used. After inoculation using overnight cultures grown at 37 °C, 250 rpm, the culture was grown to an optical density (GD600) of 0.8-1.0 at 37 °C. Then, protein expression was induced by addition of isopropyl p-D-1- thiogalactopyranoside (IPTG) at a final concentration of 0.1 mM, with incubation overnight at 18 °C, 140 rpm. Cells were harvested by centrifugation at 4°C and 15000xg for 10 minutes. The cell pellets were resuspended in lysis buffer (50 mM T ris, 300 mM NaCI, 5 mM MgCl2, pH 8.0 supplemented with 1 pL / 10 mL benzonase). Cells were lysed using an EmulsiFlexC3 High-Pressure homogenizer (Avestin) at 15,000 psi for two consecutive rounds. Cell debris and insoluble proteins were removed by centrifugation at 4 °C and 40000xg for 35 minutes. His-tagged proteins were purified from the soluble lysate using Ni-NTA affinity chromatography (Ni-NTA 5 mL Superflow Cartidge, Qiagen). The lysate was loaded onto the column and washed for six column volumes with lysis buffer. The His-tagged proteins were eluted from the column using elution buffer (50 mM Tris, 300 mM NaCI, 200 mM imidazole, pH 8.0). For further purification, size exclusion chromatography (SEC) was carried out on a Bio-Rad NGC system connected to a HiLoad 16 / 600 Superdex 200 pg column at a flow rate of 1 mL min-1, monitoring the absorbance at 280 nm. The column was equilibrated with SEC buffer (25 mM HEPES, 100 mM NaCI, 2 mM MgCI2, 1 mM TCEP, pH 7.4 and 10 % w / v glycerol). The purest fractions were combined, and the protein was concentrated using an Amicon Ultra Centrifugal Filter (10 kDa MWCO, Millipore). Protein concentration was determined using an ND-1000 spectrophotometer (Thermo Scientific) at 280 nm with theoretical extinction coefficients as determined by the Expasy ProtParam tool as shown in Table 2. The purified proteins were aliquoted, flash-frozen in liquid N2, and stored at -80°C for single-use aliquots. The identity and purity of the protein was confirmed using liquid chromatography quadrupole time of flight mass spectrometry (LC-MS Q-ToF).

[0038] Preparative phosphorylation of LgBiT-ERRy

[0039] LgBiT-ERRy was phosphorylated by TEV-protease treated PKA (which removes the His-tag from PKA). LgBiT-ERRy, TEV-treated PKA and ATP were added at the molar ratio 1 :0.02: 10 in phosphorylation buffer (20 mM HEPES, 100 mM NaCI, 20 mM MgCI2, pH 7.5) and incubated at 37°C for 50 min. After incubation, the solution was applied to a Ni-NTA gravity flow column. The flow-through containing TEV-cleaved PKA was discarded, and LgBiT-ERRy pS was eluted, combined, and exchanged into condensate buffer (20 mM HEPES, 100 mM KCI, pH 7.5) by extensive dialysis using membrane tubing of 12-14 kDa (Spectra / Por® 2 RC Dialysis Membrane, Fisher Scientific). Protein concentration was determined using an ND-1000 spectrophotometer (Thermo Scientific) at 280 nm as determined by the online ProtParam tool (ExPASy). Aliquots of pure protein were flash-frozen and stored at - 80°C. The degree of phosphorylation was analyzed using LC-MS Q-TOF.

[0040] LC-MS Q-ToF

[0041] Mass and purity of the proteins were determined using a high-resolution LC-MS Q-ToF system consisting of an ACQUITY LIPLC l-Class system (Waters) coupled to a Xevo G2 quadrupole time of flight. The protein was separated (0.3 mL min-1) on a column (Polaris C18A reverse phase column 2.0 x 100 mm, Agilent) using a 15-75% acetonitrile gradient in water supplemented with 0.1% v / v formic acid before analysis in positive mode in the mass spectrometer. The m / z spectra were deconvoluted using the MaxENTI algorithm in the Masslynx v4.1 software.

[0042] Labeling of proteins with fluorescent dyes

[0043] For aspecific fluorescent labeling of 14-3-3o, 14-3-3y, and PPARy, N- hydroxysuccinimide (NHS) ester activated chemistry was used. DyLight 405 or AF 647 NHS ester (Lumiprobe) were dissolved at 10 mg / mL in DMSO. Proteins were diluted at least tenfold into labeling buffer (0.1 M NaHCCh, pH 8.5) that was supplied with the dye directly before the protein was added. 1 ,5x to 3. Ox of molar excess of dye relative to the protein was used, and the mixture was incubated at 4 °C for 3 hours. Unreacted dye was removed twofold using PD Minitrap G-25 size exclusion column, equilibrated with condensate buffer (20 mM HEPES, 100 mM KCI, pH 7.5). The average labeling per protein was measured using the absorption of the dye at their maximum absorption wavelength and at 280 nm for the protein, using the manufacturer-provided extinction coefficients and A280 correction factors for the dyes.

[0044] Synthesis of amylose derivatives

[0045] Quaternized amylose (Q-Am), carboxymethylated amylose (Cm-Am), and nitrilotriacetic acid-modified amylose (NTA-Am) were synthesized using procedures based on those previously published (see: Mason, A. F., Buddingh, B. C., Williams, D. S. & van Hest, J. C. M. Hierarchical Self-Assembly of a Copolymer-Stabilized Coacervate Protocell. J. Am. Chem. Soc. 139, 17309-17312 (2017) and Altenburg, W. J. et al. Programmed spatial organization of biomacromolecules into discrete, coacervate-based protocells. Nat. Common. 11 , 6282 (2020)). For Q-Am, 12-16 kDa amylose (Carbosynth, 1.5 g) and NaOH (2.8 g) were dissolved in Milli-Q (14.25 mL) at 35 °C. After complete dissolution of the amylose, 3-chloro-2- hydroxypropyltrimethylammonium chloride solution (11.64 mL, 60 wt% in water) was added dropwise into the stirring reaction mixture, which was subsequently stirred overnight at 35 °C. Next, the mixture was neutralized with acetic acid and precipitated into cold ethanol (200 mL). The resulting precipitate was re-dissolved in Milli-Q water and dialyzed extensively against water using regenerated cellulose dialysis tubing (Spectrum Labs, USA) with a 3.5 kDa MWCO before lyophilization. This yielded Q-Am as colorless solids (5 g, ca. 80 % yield), with a degree of substitution of 0.8 as determined by1H NMR (D2O). For the LC-MS experiment to determine FC partitioning, a separate Q-Am batch was used due to insufficient material of the main Q-Am batch. The same procedure was used, and the degree of substitution was similar (0.7), as determined by1H NMR (D2O).

[0046] In the case of Cm-Am, 12-16 kDa amylose (Carbosynth, 1.5 g) and NaOH (3.6 g) were dissolved in Milli-Q (15 mL) at 35 °C. After complete dissolution of the amylose, chloroacetic acid (2.7 g) was added dropwise into the stirring reaction mixture, which was subsequently stirred for 2 h at 35 °C. Next, the mixture was neutralized with acetic acid and precipitated into cold ethanol (200 mL). The resulting precipitate was redissolved in Milli-Q water and dialyzed extensively against water using regenerated cellulose dialysis tubing (Spectrum Labs, USA) with a 3.5 kDa MWCO before lyophilization. This yielded Cm-Am as colorless solids (5 g, ca. 80% yield), with a degree of substitution of 0.4 as determined by1H NMR (D2O).

[0047] Nitrilotriacetic acid-modified amylose (NTA-Am) was prepared via EDC / NHS activation of the Cm-Am carboxylic acid, followed by amide bond formation with an amine-functionalized NTA. First, Cm-Am (85 mg, 0.39 mmol eq.) was dissolved in 10 mM NaHPO4 buffer (10 mL) adjusted to pH 6 with 1 M HCI. To this was added / V- hydroxysuccinimide (67 mg, 0.58 mmol) and 1-ethyl-3-(3- dimethylaminopropyl)carbodiimide (222 mg, 1.16 mmol). The reaction mixture was then stirred for 2 h at room temperature. This mixture was subsequently concentrated using 3 kDa MWCO spin filters, diluted with 10 mM NaHPO4 buffer (adjusted to pH 8), and concentrated again to remove reagents and change the pH of the reaction medium for the next step. This centrifuge / dilution cycle was repeated a further two times. For the conjugation of the NTA group, / Va, / Vcr-bis(carboxymethyl)-L-lysine hydrate (152 mg, 0.58 mmol) was first dissolved in 10 mM NaHPC buffer (18 mL) with 5% DMSO and adjusted to pH 8. To this solution was added the NHS-activated Cm-Am, and the reaction mixture was left to stir at room temperature overnight. The reaction mixture was concentrated, dialyzed extensively against MilliQ water, and lyophilized to yield NTA-Am (160 mg, ca. 90%) as colorless solids with a degree of substitution of NTA groups of 0.09 as determined by1H NMR (D2O). In the1H NMR spectrum, signals corresponding to common EDC / NHS side reactions were observed; signals for p- alanine formation by NHS ring opening and signals for N-acylurea groups were found.

[0048] Synthesis of poly(ethylene glycol)-poly(caprolactone-gradient-trimethylene carbonate)-polyglutamic acid (PEG-b-PCLgPTMC-b-PGA) terpolymer

[0049] The terpolymer (PEG-b-PCLgPTMC-b-PGA) was synthesized as described by a previously published procedure (Mason, A. F., Buddingh, B. C., Williams, D. S. & van Hest, J. C. M. Hierarchical Self-Assembly of a Copolymer-Stabilized Coacervate Protocell. J. Am. Chem. Soc. 139, 17309-17312 (2017)). Step 1 : Preparation of poly(ethylene glycol)-poly(caprolactone-gradient-trimethylene carbonate) (PEG- PCLgTMC). The organocatalyzed ring-opening polymerization of e-caprolactone and trimethylene carbonate was performed, aiming for a composition of PEG44-PCLso-g- TMC50. Monomethoxy-PEG-OH macroinitiator (2 kDa, 0.5 mmol) was weighed into an oven-dried round bottom flask and dried at 140 °C under vacuum. After cooling down, distilled e-caprolactone (E-CL, 25 mmol) and recrystallized trimethylene carbonate (TMC, 25 mmol) were added under argon and dissolved in dry DCM (12 mL). The reaction was initiated by the addition of methanesulfonic acid (0.25 mmol ~ 125 pL). The reaction mixture was stirred at 25 °C in a water bath and reaction progress was regularly checked. After completion (4-6 h) the reaction was quenched with DI PEA (1.5 mL), and the polymer was precipitated into ice cold methanol and lyophilized. This yielded 6.7 g of a waxy solid (67% yield). GPC analysis (using a PL gel 5 pm mixed D column, with THF as the solvent and calibration using polystyrene standards) yielded a polydispersity of 1 .2.

[0050] Step 2: Chain-end modification with Boc-L-phenylalanine and deprotection. PEG-PCLgTMC (49 pmol) was dissolved in 5 mL acetonitrile. Then DMAP (25 pmol) and Boc-L-phenylalanine (0.25 mmol) were added, and the solution was cooled to 0 °C. After that, DCC (1.2 mmol) dissolved in € (1 mL) was added dropwise to the mixture. The reaction was stirred for 24 h at RT. After reaction completion the flask was placed in the freezer for 1 h and filtered through a plug of celite. The filtrate was concentrated and precipitated in cold Et20 (50 mL) to obtain the pure product. The resulting copolymer was then dissolved in DCM (5 mL), to which trifluoro acetic acid (5 mL) was added (on ice). The mixture was allowed to warm to RT and stirred for 2 h. After 2 h the solvent was evaporated and the copolymer was washed with NaHCCh, 1 M NaCI and brine. Then it was dried over MgSC , filtered, concentrated, and finally precipitated from ice cold Et20.1H NMR spectroscopy showed that the signal arising from the terminal TMC group had disappeared, due to addition of phenylalanine at the terminus, and aromatic protons were visible at around 7.2-7.3 ppm. GPC analysis before and after deprotection yielded polydispersities of 1.1 , indicating that TFA treatment did not facilitate copolymer hydrolysis.

[0051] Step 3: Polymerization and deprotection of N-carboxyanhydride y-benzyl L- glutamate (BLG-NCA). Phe-terminated copolymer (57 pmol) was weighed into a Schlenk flask and dissolved in dry DMF (3 mL). Then NCA-BLG (5.7 mmol) was added under argon and the reaction was left under a constant flow of N2 for 24 hrs. The product was precipitated into cold methanol and analyzed by1H NMR spectroscopy to confirm the overall composition and, in particular, the presence of benzylic and aromatic protons at 5.0-5.2 and 7.1-7.4 ppm, respectively. The resulting polymer (200 mg) was dissolved in THF (16 mL) and EtOH (2.5 mL) was added. The solution was degassed by bubbling N2 through the solution for 20 min. Pd / C (10 mg) was added, the flask was filled with H2 and the solution was left stirring overnight. After completion, the solution was filtered over celite. The filtrate was concentrated, precipitated into ice cold methanol, and lyophilized from dioxane. A colorless waxy solid was obtained.1H NMR spectroscopy was used to confirm successful deprotection of the PBLG units. The final composition of the terpolymer was PEG44P(CL63gTMC52)PGA?. The composition of the resulting copolymer was confirmed by1H NMR spectroscopy, comparing the protons of PEG (3.65 - 3.7 ppm), terminal methyl unit (singlet at 3.40 ppm) to PCL CH2 (multiplet at 2.40 - 2.25 ppm) and PTMC CH2 (multiplet at 2.2 - 1.8 PPm). Condensate preparation

[0052] Condensates were prepared based on a previously reported procedure (Mason, A. F., Buddingh, B. C., Williams, D. S. & van Hest, J. C. M. Hierarchical Self-Assembly of a Copolymer-Stabilized Coacervate Protocell. J. Am. Chem. Soc. 139, 17309- 17312 (2017)). Q-Am, Cm-Am, and NTA-Am were dissolved separately in condensate buffer (20 mM HEPES, 100 mM KCI, pH 7.5) at a concentration of 1 mg mL’1. First, buffer, BSA (final concentration 0.1 w / v%), and NTA-Am were added to 7.5 pM of NiCl2 (final concentration) in a 1 .5 mL tube, together with Cm-Am. The tube was placed in a MixMate shaker shaking at 1500 rpm (Eppendorf). Consecutively, Q-Am was added to induce coacervation in a 1.7:0.8:0.2 mass ratio of Q-Am:Cm-Am:NTA-Am, corresponding to a 2.5:0.8:0.2 charge ratio due to differing degrees of substitution, which was found to be the most stable. After 30 s, His-tagged protein cargo was added to the shaking solution. To achieve stabilized particles, 3.3 pL terpolymer (50 mg mL-1in methoxy-poly(ethylene glycol) 350 on condensate volume of 100 pL) was added after 6 min and the mixture was shaken for another 5-10 s. For microscopy, 25-100 pL of each sample was loaded on an 18 well microscopy slide with glass bottom (Ibidi).

[0053] LC-MS assay for partitioning of small molecules

[0054] Condensates were prepared using an adapted version of the normal protocol, since a highly concentrated and large condensate sample (1.1 mL) was needed for forming a condensate macrophase of sufficient scale (>10 pL). For this experiment, a different but similar batch of Q-Am was used because of the large amount of material needed (see synthetic procedure for Q-Am). Q-Am and Cm-Am were dissolved separately in condensate buffer (20 mM HEPES, 100 mM KCI, pH 7.5) at a concentration of 10 mg mL-1. NTA-Am was omitted since it was not available in a large quantity to perform the experiment. First, buffer, BSA (final concentration 0.1 w / v%), and Cm-Am were added to a 1.5 mL tube. The tube was placed in a MixMate shaker shaking at 1500 rpm (Eppendorf). Consecutively, Q-Am was added to induce coacervation in a 1.7: 1 mass ratio of Q-Am:Cm-Am, corresponding to a 2.5:1 charge ratio due to differing degrees of substitution. The solution was shaken for 6 minutes, after which it was removed from the shaker and 100 pM of FC was added. The sample was analyzed on a brightfield microscope to confirm the presence of condensates. The sample was incubated for 2 h to allow for equilibration of FC. Next, the sample was centrifuged to form a condensate macrophase (25 °C, 21300xg, 10 min). The supernatant was carefully removed from the condensate macrophase, consisting of ~15 pL. Next, a liquid-liquid extraction step was carried out to selectively extract FC without proteins and polymers. 10 pL of both the supernatant and the condensate macrophase were separately diluted with 90 pL of 1 M NaCI solution, to dissolve the condensate macrophase, in a 1.5 mL tube. Next, the aqueous solution was extracted with 250 pL of dichloromethane by means of vortexing for 30 s. The dichloromethane layer was carefully isolated and evaporated under a stream of argon. Next, the residue, containing FC, was dissolved in 100 pL of MilliQ / acetonitrile 95 / 5 v / v% supplemented with 0.1 % formic acid (FA) for LC-MS. LC-MS analysis was carried out with a system comprising a Phenomex kinetex® 2.6 pm EVO C18 50 x 2.1 mm column using a mixture of ultrapure water with 0.1% FA and acetonitrile with 0.1% FA. For this experiment, a custom method with isocratic elution at 35% acetonitrile was used. The system was connected to a Thermo Fisher LCQ Fleet Ion Trap Mass Spectrometer. 4 pL of the samples was injected, and the runs were performed for a total of three technical replicates. The data was processed as shown in Figure 15, after which the relative FC signal was used to determine the partitioning coefficient.

[0055] Fluorescence anisotropy (FA) assays

[0056] 14-3-3 or PPARy were titrated in a 2-fold dilution series to 10 of 100 nM of fluorescently labeled peptides. For 14-3-3, condensate buffer was used (20 mM HEPES, 100 mM KCI, pH 7.5) supplemented with 0.1 % (v / v) of Tween 20 and 1 mg / mL of bovine serum albumin (BSA) to prevent aspecific hydrophobic interactions. For PPARy the same buffer was used with 100 mM of NaCI instead of 100 mM of KCI. Dilution series were prepared in low volume, non-binding polystyrene 384 well plates (Corning 4514 Black Round Bottom 384-well plates). Measurements were performed directly after plate preparation using a Tecan Spark plate reader at room temperature. The following settings were used for the FITC peptides: excitation 485 ± 20 nm; emission: 535 ± 25 nm; mirror: Dichroic 510; number of flashes: 30; integration time: 40 ps; settle time: 1 ms; gain: optimal; and Z-position: calculated from well. For RBITC- labeled HSPB6, the following settings were used: excitation 535 ± 25 nm; emission: 590 ± 20 nm; mirror: Dichroic 560; number of flashes: 30; integration time: 40 ps; settle time: 1 ms; gain: optimal; and Z-position: calculated from well. Wells containing only the labelled peptide were used to set as G-factor at 35 mAU. All data were analyzed using GraphPad Prism (version 10.0.3) and fitted using a four-parameter logistic model (4PL) to determine binding affinities (KD). All results are based on triplicates, with the mean and standard error determined by GraphPad.

[0057] Brightfield microscopy

[0058] Brightfield images were acquired using a Zeiss Axio Observer D1 microscope coupled with an AxioCamMR3 camera, at an objective with 20* magnification.

[0059] Confocal laser scanning microscopy

[0060] Confocal laser scanning microscopy (Leica TCS SP8) was used for analysis of condensates with fluorescent cargo. The system was equipped with a 405 nm laser, a 488 nm laser (used for FITC and mCherryEA), 552 nm laser (used for TAMRA, SNARF-4F, and mCherryEA), and 638 nm laser (used for AF 647) and a hybrid detector (HyD). For the 488 nm laser channel, emission was collected between 498 and 550 nm. For the 552 laser channel, emission was collected between 562 and 630 nm. Finally, for the 638 laser channel, emission was collected between 648 and 710 nm. Laser power and detector gain were optimized for each different construct and concentration to use the maximum number of gray values of the detector. For single timepoint measurements, an HC PL APO CS2 63* water immersion objective with a numerical aperture (NA) of 1.20 was used. Images were acquired with a resolution of 1024 x 1024 pixels and a pixel dwell time of 1.2 ps. For kinetic measurements, an HC PL APO CS2 20x dry objective with an NA of 0.75 was used. Images were acquired with a resolution of 1024 x 1024 or 512 x 512 and a scanning speed of 100 Hz at specified timepoints on certain positions using the Mark and Find tool. The pinhole was set to 1 Airy Unit for the wavelength of maximum emission for each fluorophore.

[0061] Image processing and analysis

[0062] All images were processed and analyzed with Fiji (Imaged). For micrographs of the 14-3-3 and PPARy channels, the brightness was digitally adjusted equally for enhanced visibility. The channels of the client were not adjusted and uncropped images in the SI were not adjusted either. For quantification of the internal fluorescence intensity, a threshold was applied to images in the 14-3-3 channel, converting it into a binary image. Next, the images were dilated using a maximum filter of radius 1 pixel, to make particle outlines more pronounced and particle recognition more reliable. Next, a watershed function was applied to separate adjacent condensates into individual regions of interest (ROIs). Using the particle analysis tool with appropriate cutoff values to select condensates as ROIs, fluorescence intensity was quantified. The ROIs recognized in the 14-3-3 channel were also redirected to the client peptide or protein channels, which is especially important for experiments over time since the 14-3-3 channel is relatively constant in fluorescence. Alternatively, for images without 14-3-3, Cy5-labeled Cm-Am was used to determine the condensate outlines. ROIs were visually checked to make sure that only condensates were selected. Next, recognized ROIs were filled and measured using the particle analysis tool, redirected to the peptide channel. The intensity was determined for each selected particle.

[0063] Crystallography figures

[0064] Figures of crystal structures were generated using Chimera 1.17.1.

[0065] Table 1. Overview of 14-3-3-binding peptides used in this work. Ahx = aminohexanoic acid, 01 pen = (2-aminoethoxy)acetic acid, p = phosphate. FC-A = Fusicoccin A.

[0066] Table 2. Overview of PPARy-binding peptides used in this work. PA = p-alanine.

[0067] Table 3. Amino acid sequences and physicochemical parameters of the proteins used in this work. Strep-tags are shown in blue, His-tags are shown in red, TEV protease sites are shown in orange, and 14-3-3-binding domains are shown in green. Text in bold shows thrombin cleavage sites. Underlined residues indicate residues that can be phosphorylated. Physicochemical parameters were calculated using the online ProtParam tool (ExPASy). The extinction coefficient (E) is given with all Cys residues reduced.a)Mass without phosphorylated residues. Results

[0068] Clients of 14-3-3 are recruited to synthetic condensates upon treatment with a molecular glue

[0069] The condensates consisted of amylose-derived polyelectrolytes, with the positively charged quaternized amylose (Q-Am, in excess) and negatively charged carboxymethylated amylose (Cm-Am). His-tagged 14-3-3 was recruited to the condensates by interactions with the nitrilo triacetic acid modified amylose complexed with Ni2+(Ni-NTA-Am). Stabilization of the condensates against fusion was achieved by a semipermeable membrane that is formed by a triblock copolymer (terpolymer).

[0070] In a first experiment, several fluorescein isothiocyanate (FITC)-labeled peptides, derived from full length 14-3-3-binding proteins, were selected because of their sensitivity to PPI stabilization by the natural product Fusicoccin A (FC, here referring to 3’-deacetylated FC). Cancerous inhibitor of protein phosphatase 2A (CIP2A), single-stranded DNA-binding protein 4 (SSBP4), and receptor-interacting Ser / Thr-protein kinase 2 (RIPK2) are moderate affinity 14-3-3 binders, that bind with their phosphorylated C-terminus to 14-3-3 (Figures 9-11). The 14-3-3o isoform was loaded in the condensates yielding a local 14-3-3 concentration of 16 ± 1 pM (Figure 12). The loading of the condensates with 14-3-3 induced some recruitment of the client peptides as compared to condensates without 14-3-3 (Figure 1). All three FC- dependent 14-3-3 client peptides showed further enhanced recruitment to the condensates upon treatment with FC, as analyzed by confocal microscopy (Figure 1). Satisfyingly, c-Raf pS259, a control peptide that binds to 14-3-3 but is not stabilized by FC, did not show a significant condensate recruitment in response to the molecular glue (Figures 1G,H and 13).

[0071] The SSBP4 peptide was the strongest binding FC-responsive client, and it also showed the smallest enhancement in recruitment upon treatment with the molecular glue (1.2-fold, Figure 1A). With a tenfold reduction in the concentrations of 14-3-3o and SSBP4, the interaction was more sensitive to FC with a 1.9-fold change in SSBP4 recruitment, highlighting the system’s capability for the stabilization of weak PPIs (Figure 14). These results show that PPI stabilization of a set of 14-3-3 clients with molecular glues indeed works within the environment of synthetic condensates, with specificity of the molecular glues towards the PPIs at hand. Molecular-glue induced client recruitment is dose-dependent

[0072] Next, the dependence of the FC concentration was analyzed on the recruitment of CIP2A to the condensates (Figure 1 i). There was a significant effect on client recruitment starting from an FC concentration of 1 pM (10 eq. relative to 14-3-3 and CIP2A), with recruitment further enhanced at increasing FC concentrations. Saturation of the recruitment at lower FC concentrations might have indicated strong partitioning of the molecular glue to the condensates. The partitioning of such drug molecules to condensates has implications for their local efficacy, and hence a liquid chromatography coupled to mass spectrometry (LC-MS) assay was developed to quantify the partitioning of FC (Figure 15). With this assay, the partitioning coefficient of FC to the condensate phase was determined to be 0.96 ± 0.02, which means that its concentration inside the condensates is similar to its concentration in bulk solution.

[0073] Screening of a panel of molecular glues yields differentiated client recruitment

[0074] After demonstrating the FC-responsive recruitment of several 14-3-3 clients, a specific PPI against several molecular glues with varying stabilizing efficacy was screened (Figure 2). A panel of semisynthetic FC analogues that vary in stabilization factor (SF), defined as the fold difference between the binary KD and the EC50 in the presence of the compound, was selected. The 14-3-3o / CIP2A interaction was screened against 6 of these FC analogues, varying in stabilization factors (Figures 2A and 2B). At a concentration of 100 pM, the molecular glues FC-THF, FC, FC-31 , and FC-NAc yielded significant enhancement in CIP2A recruitment (Figures 2C and 2D). Convincingly, these molecular glues are also among the best performing ones in the FA assay (Figure 2B). FC-J shows differentiated performance in the FA assay and the condensates and this may be due to compound-specific partitioning effects.

[0075] Competitive binding of multiple clients and a molecular glue drives affinity-based displacement

[0076] A non-FC-responsive client peptide, derived from heat shock protein beta-6 (HSPB6), was selected to analyze competitive binding between two 14-3-3 clients under the influence of molecular glues. The 14-3-3o / HSPB6 interaction is of moderate affinity (KD: 0.99 ± 0.03 pM) and is not stabilized by FC (Figure 16). The rhodamine B isothiocyanate (RBITC)-labeled HSPB6 peptide was added to 14-3-3o-loaded condensates together with the FITC-labeled CIP2A peptide, with both peptides added at a 1.5-fold molar excess relative to 14-3-3o. Interestingly, the HSPB6 peptide was displaced to a great degree by CIP2A even in the absence of stabilizer (Figure 17), although the 14-3-3o / CIP2A interaction is weaker in bulk solution (KD > 10 pM, Figure 9). This differentiated behavior within the condensates might be caused by the peptide charge, since HSPB6 is charge-neutral at neutral pH, whereas CIP2A has a net charge of -3 that favors partitioning to the positively charged condensates (Table 1).

[0077] In the presence of FC, CIP2A binds 14-3-3o with 40-fold greater EC50 than HSPB6 (Figure 9). In the condensates however, the addition of FC or the strong molecular glue FC-NAc did not produce further displacement of HSPB6 in the samples containing both peptides (Figure 17).

[0078] Molecular glues can also destabilize PPIs with off-target clients in case of overlap between their binding modes. HSPB6 occupies the FC binding pocket of 14- 3-3o, and its responsivity was tested towards FC and FC-NAc in the condensates. FC- NAc has a higher intrinsic affinity for apo 14-3-3o (KD": 0.01 mM) than FC (KD": 0.3 mM), as confirmed by a compound titration (Figure 16). Indeed, Figures 2E and 2F show that only FC-NAc yielded significant release of HSPB6 from the condensates, with the molecular glue acting as PPI inhibitor for this client. Hence, the condensate system is also suitable for screening of off-target and inhibitory effects on PPIs.

[0079] Plate-reader based screening of molecular glues is facilitated by a split luciferase assay

[0080] Efficient assays for the screening of molecular glues benefit the screening of compound libraries, which are typically carried out in plate reader format. To this end, a plate reader assay format for molecular glues in the condensates was developed for the 14-3-3y / ERRy interaction (Figure 3A). Here, the large part of split NanoLuc, LgBiT, was fused to the ERRy domain, and the small part, SmBiT, to 14-3-3y. Both fusion proteins were anchored in the condensates by His-tag interactions with Ni-NTA-Am (Figure 3B). Upon increased proximity of the luciferase subunits by PPI stabilization, an enhanced bioluminescence signal was expected. The molecular glues for this interaction have covalent mechanisms of action, forming an imine bond with Lys122 of 14-3-3y. In addition, the disulfide provides reactivity towards a Cys residue of ERRy, thereby crosslinking the ternary PPI complex. A panel of 5 molecular glues was selected featuring double or single reactive handles and SFioo pM values of 1 to 37, as shown in Figure 3C. 14-3-3y and ERRy were loaded in the condensates at 250 pM, and the samples were analyzed after overnight incubation in the presence of 100 pM of various molecular glues (Figure 3D). The phosphorylated LgBiT-ERRy yielded an enhanced bioluminescence signal compared to its unphosphorylated form due to its affinity for 14-3-3y. The reactive molecular glues R1 , R2, and R3 showed an additional, significantly enhanced, bioluminescence. Satisfyingly, the control reactive molecular glue R5, featuring only an aldehyde, produced no change in bioluminescence. Furthermore, the non-phosphorylated LgBiT- ERRy (no pS) was unresponsive to the reactive molecular glue R1.

[0081] The condensates have adjustable environments that empower screening under specific conditions

[0082] Various cellular compartments such as biomolecular condensates have characteristic local physicochemical environments, with unique pH values as an example. Here, the adjustability of the condensates towards specific environments was demonstrated using the 14-3-3y / ERRy covalent molecular glues.

[0083] The FITC-labeled ERRy peptide was added to condensates with 14-3-3y, and an enhanced recruitment was observed at higher pH values in the presence of R1 (Figures 4A and 4B). Interestingly, the recruitment of ERRy was lower at pH 8.5 compared to pH 8.0. Possibly, elevated pH values drove off-target reactivity towards other Lys residues in the condensates, as this effect was not observed in bulk solution. The binary 14-3-3y / ERRy interaction was mostly pH-insensitive in condensates (Figure 4C). Using the ratiometric pH sensor protein mCherryEA inside the condensates, it was shown that the local pH values inside the condensates closely matched the pH in bulk solution across this pH range (Figure 18).

[0084] In the assays with the ERRy reactive stabilizers, tris(2-carboxyethyl)phosphine (TCEP) was used as a catalyst for disulfide exchange. At higher concentrations of TCEP, disulfide formation is halted completely and the PPI stabilization of ERRy in the condensates is prevented (Figures 4D and 4E), which is in agreement with the stabilization of the same interaction in bulk solution. 14-3-3 full length clients

[0085] Although peptide derivatives of clients can be used as mimics for their binding properties, full length proteins are more therapeutically interesting to screen as they possess a more complex structure with different interfaces that might influence the performance of molecular glues. Hence, it was proposed to use the folded domain of a native protein, the estrogen receptor a ligand binding domain (Era LBD). Here, significant recruitment of the Era LBD to the condensate interface upon treatment with 3’-deac FC-A was shown (Figure 5). Without wishing to be bound by theory, it is believed that Era LBD has patches that are unfavourable to uptake in the condensate, and that this is the cause of the protein sticking to or aggregating on the surface. It was also shown that a negative control FC-NAg, does not yield enhanced recruitment of Era LBD to the condensate (Figure 5F). This data shows that the system could be used for studies into clients with folded domains or full length clients.

[0086] Stabilization of nuclear receptor PPIs show the broad applicability of the condensate system

[0087] Next, a different type of PPI, not involving 14-3-3, was studied to investigate the general applicability of the condensate system for screening small molecule modulators of PPIs. The nuclear receptor PPARy regulates fatty acid storage and glucose metabolism and is involved in diseases such as cancer and diabetes. Natural and synthetic PPARy ligands find clinical use, and a range of partial as well as full agonists have been reported. Ligand binding to PPARy induces conformational changes, which then leads to changes in coregulator binding affinities (Figure 6). Due to its therapeutic relevance and allosteric mode of PPI stabilization, PPARy was selected as an ideal target for the validation of the condensate system towards a broader range of druggable PPIs.

[0088] PPARy PPI stabilizers

[0089] The ligand binding domain of PPARy was loaded in the coacervates in the same manner as 14-3-3, that is by His tag-Ni-NTA-amylose interactions. Small molecule ligands can bind in its ligand binding pocket, which stabilizes interactions with coactivating polypeptides in an allosteric manner (Figure 6A). These small molecule ligands bind with intrinsic affinity to the binding pocket. It was shown that the coregulator peptide PRIPRAP can be recruited by different commercially available ligands (Figure 6B-6E), and that the recruitment depends on the intrinsic affinity of the ligand (Figure 6F). Hereby, it was shown that the condensate platform is not only suited to molecular glues at the PPI interface, but also other PPI stabilizing agents that act in a different manner, in this case by allosteric effects with intrinsic affinity for a defined binding pocket. It was also shown that PPIs of different classes of proteins, in this case nuclear receptors, can be studied in the condensates and that the system is more broadly applicable.

[0090] In more detail, His-tagged PPARy (400 nM) was loaded in the condensates by means of interactions with the Ni-NTA-amylose (Figure 6B), yielding a local PPARy concentration of 2.0 ± 0.4 pM (Figure 19). A panel of commercially available PPAR ligands was selected that have varying intrinsic affinity for PPARy and cooperativity for coregulator binding, shown in Figure 6C. Next, this panel was screened in the PPARy-loaded condensates at a ligand concentration of 10 pM, monitoring the recruitment of the coregulator peptide PRIPRAP by confocal microscopy (Figures 6D- 6E). The intrinsic affinity of the ligand, KD", was found to correlate with the coregulator recruitment to the condensates (Figure 6F). The ligands MRL24 (partial agonist) and tesaglitazar (full agonist) yielded significant recruitment of PRIPRAP relative to the DMSO control. The dose-dependency of Tesaglitazar was also tested for the recruitment of PRIPRAP (Figure 6G). Here, a significant effect was found in the condensates only at 10 pM compound, which indicates that tesaglitazar did not partition effectively.

[0091] The Tesaglitazar-dependent recruitment of coregulators was validated with two additional coregulators that bind with high affinity (MED1 DRIP-2, KD 647 ± 62 nM) or low affinity (EP300, KD >10 pM), as shown in Figure 20. For these additional coregulators, a ligand-mediated recruitment was observed to the condensates as well, which indicates that a wide range of coregulators can be studied in this system.

[0092] After demonstrating the efficacy of molecular glues in the condensate system, it was demonstrated that the system is also applicable to screen more complex interactions of competing ligands. GW9662 is a covalent PPARy antagonist that blocks the effect of agonists by covalent anchoring in the agonist binding pocket. Condensates were prepared with PPARy, and the coregulator PRIPRAP was added in the presence of DMSO (control), GW9662 (10 pM), tesaglitazar (10 pM), or both compounds (Figure 7). GW9662 did not influence PRIPRAP recruitment. Convincingly, the antagonist abolished the effect of tesaglitazar, and the PRIPRAP recruitment was found to be similar to the DMSO control. The condensate system is suitable for the optimization of competitively binding ligands, as ligands as well as the interacting proteins can be supplied in adjustable stoichiometries.

[0093] Brief description of the figures

[0094] Figures 1A-H show the confocal micrographs of 14-3-3o-loaded condensates supplied with client peptides (100 nM) in the absence (DMSO control) or presence of 100 pM of FC, including quantification. Scale bar: 25 pm. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N > 49 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison, with significance regarded as P < 0.05. The line shows the mean, and the symbols show the individual measurements. Figure 11 shows confocal micrographs of 14-3-3o-loaded condensates supplied with the CIP2A peptide (100 nM) in the absence (DMSO control) or presence of various concentrations of FC, incubated overnight. Scale bar: 25 pm. Figure 1J shows quantification of micrographs of 14-3-3o-loaded condensates supplied with the CIP2A peptide (100 nM) in the absence (DMSO control) or presence of various concentrations of FC. Statistical analysis was performed by one-way ANOVA with Dunnet’s test with correction for multiple comparisons, compared to the DMSO control. N > 55 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison, with significance regarded as P < 0.05. The line shows the mean, and the symbols show the individual measurements.

[0095] Figure 2A shows structures of selected FC analogues. The differences are highlighted in red. Figure 2B shows protein titration of 14-3-3o to the CIP2A peptide as measured by FA assay in the presence of DMSO (control) or 100 pM of various FC analogues. Symbols represent the mean of a triplicate measurement, with the error bars partly obscured by the symbols. Lines show fits by a 4-parameter logistic model. The resulting KD obtained from the fit is shown as mean with the standard error. For the DMSO and FC-Nag samples, no reliable KD could be fitted due to the absence of an upper plateau. For the sample with FC-Nac, no reliable KD could be fitted due to the absence of a lower plateau, surpassing the assay limit at 10 nM peptide concentration. Figure 2C shows confocal micrographs of 14-3-3o-loaded condensates supplied with CIP2A (100 nM) in the absence (DMSO control) or presence of 1 pM of FC analogues. Samples were incubated overnight (scale bar: 25 pm). Figure 2D shows quantification of micrographs of 14-3-3o-loaded condensates and CIP2A (100 nM) in the absence (DMSO control) or presence of 1 pM of FC analogues. Statistical analysis was performed by one-way ANOVA with Dunnett’s test with correction for multiple comparisons to the DMSO control, with N > 89 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison, with significance regarded as P < 0.05. The line shows the mean, and the symbols show the individual measurements. Figure 2E shows confocal micrographs of 14-3-3o- loaded condensates supplied with HSPB6 (150 nM) in the absence (DMSO control) or presence of FC analogues. Samples were incubated overnight (scale bar: 25 pm). Figure 2F shows quantification of micrographs of 14-3-3o-loaded condensates and HSPB6 (100 nM) in the absence (DMSO control) or presence of FC analogues. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N > 52 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison, with significance regarded as P < 0.05. The line shows the mean, and the symbols show the individual measurements.

[0096] Figure 3A shows a schematic overview of split luciferase assay for the 14-3- 3y / ERRy interaction, stabilized by reactive molecular glues. Figure 3B shows an assay setup of the split luciferase assay inside condensates. Figure 3C shows a panel of reactive molecular glues for the 14-3-3y / ERRy interaction, with stabilization factor at 100 pM compound concentration (SFIOOMM). Figure 3D shows a split luciferase assay for the dynamic covalent stabilization of the 14-3-3y / ERRy interaction inside condensates. Assay conditions: 250 pM of LgBiT-ERRy, phosphorylated or lacking the phosphorylation site (no pS), 250 pM of 14-3-3y-SmBiT, 100 pM of reactive compounds R1-R5, 50 pM of TCEP, and 1 v / v% of DMSO. The samples were incubated overnight, after which the substrate furimazine was added (1 : 1000 dilution) and the samples were measured immediately. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N = 3 independent condensate samples. P values are shown above the comparison. The bar plot shows the mean ± standard deviation, and the symbols show the individual measurements.

[0097] Figure 4A shows confocal micrographs of 14-3-3y-loaded condensates supplied with FITC-ERRy (100 nM) in the presence of 100 pM of stabilizer R1 at various pH values and in the presence of 50 pM of TCEP. Samples were incubated overnight (scale bar: 25 pm). Figures 4B and 4C shows quantification of micrographs of 14-3-3y-loaded condensates supplied with FITC-ERRy (100 nM) in the presence of 100 pM of stabilizer R1 (Figure 4B) or DMSO (Figure 4C) at various pH values. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N > 47 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison, with significance regarded as P < 0.05. The line shows the mean, and the symbols show the individual measurements. Figure 4D shows confocal micrographs of 14-3-3y- loaded condensates supplied with FITC-ERRy (100 nM) in the presence of 100 pM of stabilizer R1 at various TCEP concentrations and pH 7.5. Samples were incubated overnight (scale bar: 25 pm). Figure 4E shows quantification of micrographs of 14-3- 3y-loaded condensates supplied with FITC-ERRy (100 nM) in the presence of 100 pM of stabilizer at various TCEP concentrations. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N > 37 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison, with significance regarded as P < 0.05. The line shows the mean, and the symbols show the individual measurements.

[0098] Figure 5A shows the Era LBD which is a folded domain that forms a dimer, and of which its disordered F-domain can bind to 14-3-3 in a phosphorylation-dependent manner. This interaction can be modulated by 3’-deac FC-A. Figures 5B and 5C show the confocal micrographs showing that Era LBD is recruited to the condensates in a 14-3-3-dependent manner and a phosphorylation-dependent manner (scale bar: 25 pm). Figure 5D shows the quantification of the data in Figures 5B and 5C, respectively. Statistical differences were analysed by Tukey’s test with multiple comparisons, with N = 3 independent experiments. Where solid lines represent the mean, ns: no statistical difference, ****: p < 0.0001. Figure 5E shows the confocal micrographs demonstrating the 3’-deac FC-A-dependent recruitment of Era LBD to the condensates (scale bar: 25 pm). Figure 5F shows the quantification of micrographs in Figure 5E. Statistical differences were analysed by Tukey’s test with multiple comparisons, with N = 3 independent experiments. Where solid lines represent the mean, ns: no statistical difference, *: p < 0.05.

[0099] Figure 6A shows a schematic overview of ligand and coregulator binding equilibria. Figure 6B shows the crystal structure of the ligand binding domain of PPARy with bound coregulator peptide PRIPRAP and ligand rosiglitazone, PDB: 6ONJ. Figure 6C shows a schematic overview of ligand-dependent recruitment of coregulators in condensates loaded with PPARy. Figure 6D shows an overview of commercial PPARy ligands. Figure 6E shows confocal micrographs of ligand-dependent recruitment of PRIPRAP coregulator (100 nM). Ligands were added at a concentration of 10 pM. Samples were incubated overnight. Scale bar: 25 pm. Figures 6F and 6G show quantification of PRIPRAP from micrographs taken at the conditions of panel E. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N > 46 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison, with significance regarded as P < 0.05. The line shows the mean, and the symbols show the individual measurements. Panel 6G shows the data plotted against the intrinsic affinity of the ligands (KDII).

[0100] Figure 7 shows that the inverse agonist GW9662 negates the effect of the agonist ligand Tesaglitazar. Figure 7A shows a schematic overview of coregulator recruitment into PPARy-loaded condensates upon treatment with agonist and / or antagonist. Figure 7B shows confocal micrographs of condensates loaded with PPARy (400 nM) and in the presence of coregulator PRIPRAP (100 nM), treated with Tesaglitazar (agonist, 10 pM) and / or GW9662 (antagonist, 10 pM). Samples were incubated overnight. Scale bar: 25 pm. Figure 7C shows quantification of confocal micrographs showing PPARy-, agonist- and / or antagonist-dependent recruitment of the coregulator PRIPRAP. Statistical differences were analysed by Tukey’s test with correction for multiple comparisons, with N > 132 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparisons, with significance regarded as P < 0.05. Solid lines represent the mean, and the symbols represent individual measurements.

[0101] Figure 8 provides an overview of examples of assay designs highlighted in the current technology. The first design includes a mobile polypeptide which is localized to the condensate upon treatment with a PPI stabilizing agent (top). The addition of the PPI stabilizing agents leads to binding of the mobile polypeptide to the immobilized polypeptide to which it has enhanced affinity in the presence of the PPI stabilizing agent. The second assay setup includes condensates with two or more interacting polypeptides immobilized in the condensate (bottom). The interaction of these polypeptides is enhanced by the addition of a PPI stabilizing agent.

[0102] Additional information

[0103] Figure 9. Protein titration of 14-3-3o to the CIP2A peptide (10 nM), as measured by fluorescence anisotropy assay in the presence of DMSO (control) or 100 pM of FC. Symbols represent the mean of a technical triplicate, with the error bars partly obscured by the symbols. Lines show fits by a 4-parameter logistic model. The resulting KD or EC50 obtained from the fit is shown as mean with the standard error. For the DMSO sample, no reliable KD could be fitted due to the absence of an upper plateau.

[0104] Figure 10. Protein titration of 14-3-3o to the SSBP4 peptide (10 nM), as measured by fluorescence anisotropy assay in the presence of DMSO (control) or 100 pM of FC. Symbols represent the mean of a technical triplicate, with the error bars partly obscured by the symbols. Lines show fits by a 4-parameter logistic model. The resulting KD or EC50 obtained from the fit is shown as mean with the standard error.

[0105] Figure 11. Protein titration of 14-3-3o to the RIP2K peptide (10 nM), as measured by fluorescence anisotropy assay in the presence of DMSO (control) or 100 pM of FC. Symbols represent the mean of a technical triplicate, with the error bars partly obscured by the symbols. Lines show fits by a 4-parameter logistic model. The resulting KD or EC50 obtained from the fit is shown as mean with the standard error. For the DMSO sample, no reliable KD could be fitted due to the absence of an upper plateau.

[0106] Figure 12. Quantification of the concentration of 14-3-3o inside the condensates. A) Calibration curve of AF 647 in solution measured using the same settings as 14-3-3. The formula was obtained by linear regression. B) Quantification of 14-3-3o (AF 647-labeled) across three independent condensate samples loaded with 100 nM of 14-3-3o. In the calculation, the mean degree of labeling of 14-3-3o with AF 647 was used, which was determined to be 1 .69. The dashed lines show the mean, and the dotted lines show the quartiles. N > 124 condensates.

[0107] Figure 13. Protein titration of 14-3-3o to the non-FC-responsive c-Raf pS259 peptide (10 nM), as measured by fluorescence anisotropy assay in the presence of DMSO (control) or 100 pM of FC. Symbols represent the mean of a technical triplicate, with the error bars partly obscured by the symbols. Lines show fits by a 4-parameter logistic model. For both titrations, no reliable KD or EC50 could be fitted due to the absence of an upper plateau.

[0108] Figure 14. A) Confocal micrographs of 14-3-3o-loaded condensates (10 nM of 14-3-3o) supplied with SSBP4 (10 nM) in the absence (DMSO control) or presence of 100 pM of FC. Scale bar: 25 pm. b) Quantification of micrographs of samples used in panel a. Statistical analysis was performed by unpaired two-tailed t-test, with N > 114 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison. The line shows the mean, and the symbols show the individual measurements.

[0109] Figure 15. LC-MS assay for the determination of the partitioning coefficient of FC. A) Schematic overview of sample preparation and measurement of the partitioning coefficient. Condensates are prepared from Q-Am and Cm-Am at 10x the usual amylose concentration, yielding unstabilized condensates. The terpolymer was omitted because it is challenging to purify it for the LC-MS method, and the samples were pelleted into a condensate macrophase anyway. Ni-NTA-amylose was omitted due to the large scale of the experiment and insufficient material. A representative microscopy image taken directly after preparation is shown as inset, scale bar 100 pm. The sample was equilibrated and centrifuged to yield a condensate macrophase (photo of the sample shown), after which the relative concentration of FC in the supernatant and the condensate macrophase was determined by LC-MS. B) Chromatograms of the supernatant and condensate macrophase, after extraction, showing the FC peak. C) Mass spectra derived from the chromatograms shown in panel b. d) Determination of the partitioning coefficient by relative peak areas of the LC-MS chromatograms. N=3 technical replicates based on the same sample.

[0110] Figure 16. A) Protein titration of 14-3-3o to the RBITC-labeled HSPB6 peptide (100 nM), as measured by fluorescence anisotropy assay. Symbols represent the mean of a technical triplicate, with the error bars partly obscured by the symbols. Lines show fits by a 4-parameter logistic model. The resulting KD obtained from the fit, shown as mean with the standard error. B) Compound titration of FC or FC-Nac to RBITC- labeled HSPB6 (100 nM) in the presence of 14-3-3o (1 pM). The highest concentration of compound is 200 pM in both cases. Symbols represent the mean of a triplicate measurement, with the error bars partly obscured by the symbols.

[0111] Figure 17. A) Confocal micrographs of 14-3-3o-loaded condensates supplied with CIP2A and HSPB6 (150 nM each) in the absence (DMSO control) or presence of FC analogues. Scale bar: 25 pm. b, c) Quantification of b) CIP2A and c) HSPB6 in samples at the conditions of panel a. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N > 53 condensates across 2 imaging positions in 3 independent samples. P values are shown above the comparison. The line shows the mean, and the symbols show the individual measurements.

[0112] Figure 18. A) Schematic overview of mCherryEA structure and uptake into synthetic condensates. The structure shown is that of the non-mutated mCherry, PDB: 2H5Q. b) Calibration of mCherryEA in bulk solution (50 pM) using confocal microscopy and calculation of the ratio of excitation at 488 nm or 552 nm, with collection of emission between 562-700 nm in both settings. C) Confocal micrographs of condensates prepared at different pH (bulk pH values shown), with 500 nM of mCherryEA. Excitation was performed at 488 nm or 552 nm, with collection of emission between 562-700 nm in both settings. D) Plot of the determined local pH values in the condensates relative to the buffer pH (bulk), as determined from the samples in panel c and the calibration curve.

[0113] Figure 19. Measurement of the local concentration of PPARy in the condensates. A) Calibration curve of AF 647 in condensate buffer by confocal microscopy, using the same settings used for the condensate samples. The formula was obtained by linear regression. B) Confocal micrograph of condensates loaded with AF 647-labeled PPARy (400 nM bulk concentration). C) Quantification of micrographs of independent condensate samples containing 400 nM of AF 647-labeled PPARy, revealing the distribution of local concentrations of PPARy in the condensates. In the calculation, the degree of dye labeling of the protein was used, which was determined to be 0.38. N > 130 condensates analyzed for each sample. Figure 20. A) Confocal micrographs of ligand-dependent uptake of MED1 DRIP-2 coregulator. Scale bar: 25 pm. b) Quantification of MED1 DRIP-2 from micrographs. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N = 3 independent condensate samples. P values are shown above the comparison. The lines show the means, and the symbols show the individual measurements. C) Confocal micrographs of liganddependent uptake of EP300 coregulator. Scale bar: 25 pm. d) Quantification of EP300 from micrographs. Statistical analysis was performed by one-way ANOVA with Tukey’s test with correction for multiple comparisons, with N = 3 independent condensate samples. P values are shown above the comparison. The lines show the means, and the symbols show the individual measurements.

Claims

CLAIMS1. Screening system for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction, the screening system comprising: a first protein having at least one affinity motif; and a second protein having at least one affinity motif, said at least one affinity motif of the second protein having affinity towards said at least one affinity motif of the first protein and vice versa, wherein the screening system has a protein-protein association state in which the first and second proteins are associated and a protein-protein dissociation state in which the first and second proteins are not associated, wherein switching between the protein-protein association state and the protein-protein dissociation state: is modulated depending on the presence, absence and / or concentration of the ligand; and results in a change of the screening system’s spectroscopic properties, characterised in that the screening system further comprises a condensate formed of condensate-forming molecules, said condensate is configured to reversibly recruit the first protein and the second protein independently, in that the condensate facilitates the formation of the protein-protein association state, and in that a ligand-binding domain is provided by: the first protein and / or the second protein; or the first and second protein in the protein-protein association state.

2. Screening system according to claim 1 , wherein the screening system comprises further proteins having further affinity motifs, said further affinity motifs having affinity towards the at least one affinity motif of the first protein.

3. Screening system according to claim 1 or 2, wherein the condensate comprises condensate-forming molecules having protein immobilizing motifs, said protein immobilizing motifs having affinity or reactivity towards corresponding immobilizing motifs comprised on at least one of the proteins to immobilize the at least one of the proteins in the condensate.

4. Screening system according to any of the preceding claims, wherein: the first protein is immobilized in the condensate; or the first protein and the second protein are immobilized in the condensate.

5. Screening system according to any of the preceding claims, wherein the first protein is immobilized in the condensate and, optionally, the second protein is immobilized in the condensate, and wherein the screening system is configured to emit light in the protein-protein association state.

6. Screening system according to any of claims 1-4, wherein the first protein is immobilized in the condensate, and wherein the screening system is configured to emit light in the protein-protein dissociation state.

7. Screening system according to any of the preceding claims, wherein the ligandbinding domain is an allosteric ligand-binding domain.

8. Screening system according to any of the preceding claims, wherein each of the first protein and the second protein is labelled with a fluorescent dye.

9. Screening system according to claim 8, wherein the fluorescent dye of the first protein interacts with the fluorescent dye of the second protein in the protein-protein association state to emit light.

10. Method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction, the method comprising the steps of: a) providing the screening system according to any of the preceding claims; b) providing a ligand; c) measuring the screening system’s spectroscopic properties in the absence of the ligand; d) contacting the ligand provided in step b) with the screening system provided in step a); e) measuring the screening system’s spectroscopic properties in the presence of the ligand; f) determining the change in the screening system’s spectroscopic properties based on the measurements provided in steps c) and e); andg) determining the quantitative and / or qualitative change in protein-protein interaction modulated by the ligand based on the change in the screening system’s spectroscopic properties determined in step f).

11. Use of a condensate in a method for screening a ligand using protein-protein interaction and spectroscopic detection of said protein-protein interaction according to claim 10.

Images