Functional mapping of the 14-3-3 hub protein as a guide to design 14-3-3 molecular glues

Molecular glues represent an evolution in drug discovery, however, targeted stabilization of protein complexes remains challenging, owing to a paucity of drug design rules. The functional mapping of hotspots has been critical to protein–protein interaction (PPI) inhibitor research, however, the orthogonal approach to stabilize PPIs has not exploited this information. Utilizing the hub protein 14-3-3 as a case study we demonstrate that functional mapping of hotspots provides a triage map for 14-3-3 molecular glue development. Truncation and mutation studies allowed deconvoluting the energetic contributions of sidechain and backbone interactions of a 14-3-3-binding non-natural peptide. Three central 14-3-3 hotspots were identified and their thermodynamic characteristics profiled. In addition to the phospho-binding pocket; (i) Asn226, (ii) Lys122 and (iii) the hydrophobic patch formed by Leu218, Ile219 and Leu222 were critical for protein complex formation. Exploiting this hotspot information allowed a peptide-based molecular glue that elicits high cooperativity (α = 36) and selectively stabilizes the 14-3-3/ChREBP PPI to be uniquely developed.

Fluorescence anisotropy assay of 14-3-3 titration to fixed concentrations fluorescein-labelled peptide (10 nM) and schematic representation of phosphoserine containing peptide (orange) and the tested phosphoserine replacements being serine (blue), glutamic acid (yellow) and aspartic acid (green).  Figure S4. Sequences C-terminal truncated peptides 2e-i. Schematic representation and chemical structure of peptide 1 and the C-terminal truncations. Within these peptides we stepwise remove the C-terminal amino acids leading to peptides 2e-i.              Ac-EGRSAG pS IPGRRS-CONH2 6YOW 9 Table S8. Data collection and refinement statistics (molecular replacement) for 14-3-3c in complex with non-natural N-terminally truncated peptides 2c and 2d

Peptide synthesis
The peptides 1 and 2a-d were synthesized on 50 mol scale on a Rink amide MBHA resin (Novabiochem, 0.52 mmol/g loading) using an automated Intavis MultiPep RSi peptide synthesizer. Deprotection was performed twice per cycle with 20% v/v piperidine in DMF for 8 minutes. Fmocprotected amino acids were mixed in 5 equivalents of HBTU and 9 equivalents of DIPEA in DMF, and coupled to the resin. With the exception of phosphoserine, which was coupled once for 60 minutes, all amino acids were coupled to the resin twice for 30 minutes. Capping of unreacted amino acids was performed with a mixture of acetic anhydride/pyridine/DMF (1:1:3) for 5 minutes. Ala was incorporated as a spacer between the sequence and Lys(Alloc), to which a FITC fluorophore will be coupled for fluorescence anisotropy studies. The peptides 1 and 2e-i were synthesized manually on a 50 mol scale using Rink amide MBHA resin (Novabiochem, 0.52 mmol/g loading) via Fmoc based solid phase peptide synthesis. Deprotection was performed twice per cycle with 20% v/v piperidine in N,N-dimethylformamide (DMF) for five minutes. Fmoc-protected amino acids were mixed with 9 equivalents HBTU (0.38 M in DMF) and 15 equivalents of DIPEA, and coupled to the resin for 30 minutes (except pS, which is coupled for 60 minutes). Peptides 1 and 2e-i were labeled with fluorescein via a 5-(-fluorenylmehyloxycarbonyl-amino)-3oxapentanoic acid (O1-Pen) linker using 5 equivalents of FITC and 7.5 equivalents of DIPEA reacting overnight with continuous agitation. Removal of protecting groups and cleavage of the resin was performed by incubation in a mixture of trifluoroacetic acid (TFA), H2O, and triisopropylsilane (TIS) (96.5:2.5:1) for 2 hours with continuous agitation, followed by precipitation in an excess of ice-cold diethyl ether (Et2O). Peptides 1 and 2a-d were FITC labeled via a three-step approach. The free amine on the N-terminus was protected as follows: to a solution of di-tert-butyl decarbonate (10 eq, 400 mM) in DCM was added DIEA (10 eq), and solution was added to the resin. Coupling was allowed to proceed for 1 h. At this time, resin was washed 3x with DCM and coupling was repeated as described. Resin was washed 5x with DCM. Alloc removal was achieved as follows: resin was treated with a solution of tetrakis(triphenylphosphine)palladium(0) (0.5 eq, 20 mM) and phenylsilane (20 eq) in DCM, 2x 45 min. Resins were then washed 3x with DCM, then 3x with DMF. FITC was installed on the free amine on each C-terminal lysine by treating resin with fluorescein isothiocyanate isomer I (10 eq, 400 mM in 4:1 DMF:DCM) and DIEA (15 eq) for at least 1.5 h. Reactions were kept under aluminum foil for the duration of the coupling. Reaction mixtures were then drained and resins were washed 3x with DMF, 3x with DCM, and dried under reduced pressure. Removal of protecting groups and cleavage of the resin was performed by incubation in a mixture of trifluoroacetic acid (TFA), H2O, and triisopropylsilane (TIS) (96.5:2.5:1) for 2 hours with continuous agitation, followed by precipitation in excess of ice-cold diethyl ether (Et2O).
All peptides were purified using preparative HP-LC. This was performed using a Gemini S4 110A 150 x 21.20 mm column using ultrapure water with 0.1% formic acid (FA) and acetonitrile with 0.1% FA with various gradients. Correct mass and purity of peptides was identified using analytical liquidchromatography coupled with mass-spectrometry (LC-MS) was performed on a C4 Jupiter SuC4300A 150 x 2.0 mm column using ultrapure water with 0.1% formic acid (FA) and acetonitrile with 0.1% FA, in general with a gradient of 5% to 100% acetonitrile over 10 minutes, connected to a Thermo Fisher LCQ Fleet Ion Trap Mass Spectrometer. The purity of the samples was assessed using a UV detector at 254 nm. Measurements were performed directly after plate preparation, using a Tecan Infinite F500 plate reader at room temperature (lex: 485 ± 20 nm; lem: 535 ± 25 nm; mirror: Dichroic 510; flashes: 20; integration time: 50 ms; settle time: 0 ms; gain: optimal; and Z-position: calculated from well). Wells containing only FITC-peptide were used to set as G-factor at 35 mP. All data were analyzed using GraphPad Prism (7.00) for Windows and fitted using a four-parameter logistic model (4PL). Each measurement was performed in three independent experiments, average and standard deviations were calculated in Excel (see SI tables S1-S4).
G analysis: Based on the obtained binding affinities from the FA binding studies the fold change in binding affinity is determined by dividing the KD of one peptide with the KD from another peptide (fold change = KD peptide1 / KD peptide2 and Z-position: calculated from well). Wells containing only FITC-peptide were used to set as G-factor at 35 mP. All data were analyzed using GraphPad Prism (7.00) for Windows and fitted using a four-parameter logistic model (4PL). Data was obtained and averaged based on two independent experiments.

Cooperativity analysis:
To determine the cooperativity parameters from the 2D-titration of 14-3-3/ChREBP/2d (as described above) we have used the general framework for straightforward model construction of multi-component thermodynamic equilibrium systems as described by Geertjens et al. (2021). 10 This general platform generates a model to describe multi-component equilibrium systems when given a system description.
In our case we gave the following system description: R + P = RP; KD I RP + C = RPC; KD II /  R + C = RC; KD II RC + P = RPC; KD I /  with R = 14-3-3, P = non-natural peptide (labelled), and C = ChREBP. The framework determined based on this system description the equilibrium equations needed to determine KD II and . The data from 2Dtitrations was provided to the model including the KD I at 1.96 mM, P_tot = 10 nM, and the variable concentrations of 14-3-3 and ChREBP at each datapoint. From this the model determined the KD II and  (with 95% confidence interval) and the error-landscape of the determine parameters.

Isothermal titration calorimetry (ITC)
Final dialysis fluid from protein expression was frozen as 2 mL aliquots to serve as ITC buffer. Protein and peptide were dissolved and diluted in this buffer to reported concentration. DMSO was added if reported and matched in cell and syringe. Samples were degassed for 10 min prior to measurement at 450 mmHg. The reference cell was filled with 300 L degassed MilliQ water and sample cell with 300 L of peptide or protein mixture. Syringe was loaded with at least 200 L protein or peptide sample. Measurements were performed on an Affinity ITC LV (TA instruments), with injection size set to 2 L, stirring speed of 150 rpm and temperature at 25 °C. The data was processed and analyzed in NanoAnalyze v3.11. The baseline was manually inspected and corrected, after which a blank constant model was fitted to correct for the heat of injection. Subsequently an independent model was fitted, which the Nanoanalyse uses to report the thermodynamic binding properties reported in this paper. Crystallography 14-3-3C/2c: Peptide 2c was soaked in preformed crystals of 14-3-3c (truncated after T231 to reduce flexibility) with cJun peptide 11 , which grew in 28% (v/v) PEG400, 5% glycerol, 0.2 M CaCl2, 0.1 M HEPES pH 7.5 within two weeks. The soaked crystal was fished after 14 days of incubation and flashfrozen in liquid nitrogen. Diffraction data was in-house collected at 100 K. X-ray diffraction data were