A Supramolecular Stabilizer of the 14‐3‐3ζ/ERα Protein‐Protein Interaction with a Synergistic Mode of Action

Abstract We report on a stabilizer of the interaction between 14‐3‐3ζ and the Estrogen Receptor alpha (ERα). ERα is a driver in the majority of breast cancers and 14‐3‐3 proteins are negative regulators of this nuclear receptor, making the stabilization of this protein‐protein interaction (PPI) an interesting strategy. The stabilizer (1) consists of three symmetric peptidic arms containing an arginine mimetic, previously described as the GCP motif. 1 stabilizes the 14‐3‐3ζ/ERα interaction synergistically with the natural product Fusicoccin‐A and was thus hypothesized to bind to a different site. This is supported by computational analysis of 1 binding to the binary complex of 14‐3‐3 and an ERα‐derived phosphopeptide. Furthermore, 1 shows selectivity towards 14‐3‐3ζ/ERα interaction over other 14‐3‐3 client‐derived phosphomotifs. These data provide a solid support of a new binding mode for a supramolecular 14‐3‐3ζ/ERα PPI stabilizer.

. a) Direct binding as observed from titrations of 14-3-3 to various client-derived fluorescein-labelled phosphopeptides (ER, TASK3, C-Raf, Tau and Cdc25B). Selected protein concentrations for compound titration experiments are indicated for each client. b) Titration data for 1 to the phosphopeptides and 14-3-3.

Computational studies
The molecular models for the complexes 1:14-3-3 were prepared in three general steps. First, probe 1 was docked into the binding sites of the modulator protein using AutoDock4.0 [3] in the presence of the 5-mer peptide 591-FPATpV-595 of the human Estrogen receptor Peptide alpha (ER), which contains the phosphorylated Thr594 (pT), and fusicoccin (FC-A). The Cartesian coordinates of the trimeric ER/14-3-3/FC-A complex were taken from its crystal structure at a resolution of 2.1 A (PDB entry. 4JDD). [4] Complex 1 was divided by arms and docked in an incremental manner, where each of the arms was added gradually to the system. The grid boxes where centered at pT binding site and at the central pore with a size of 60x60x60 Å. 13 torsional degree of freedom were included for 1 and a genetic algorithm (GA) was used for the docking, including 150 number of individuals per population and 20 runs. The GA was performed several times. After visual inspection, the best solution was selected for refinement using force field MD simulation.
The optimized geometries of FC-A, GCP and Tp (named as TPO) and their force-field parameters were computed following the standard method (HF/6-31G**//HF/3-21G) using the RESP ESP charge Derive Server Development. [5] All the atoms were described as AMBER atom type and the point charges were described as RESP charges. After fitting to the atoms the electrostatic potential was computed using the program antechamber of AmberTools18. [6] The leaprcff14SB force field was used in all the MD simulations. The MD simulations were run on GPUs using the pmemd.cuda_SPFP module implemented in Amber18.
Both ER/14-3-3/FC-A/1 and ER/14-3-31 complexes were embedded in a truncated octahedral box of ca. 24000 TIP3P water molecules [7] that extended 12 Å away from any solute atom and 21 Na + ions were added to ensure charge neutrality. The system was relaxed by energy minimization in three consecutive steps (3 x 5000 cycles), in which after the first 1000 cycles the minimization method was switched from steepest descendent to conjugate graduate. The resulting system was heated from 100 to 300 K during 200 ps with a time step of 0.1 fs and with the position of all the solute atoms restrained with a harmonic constant of 200 kcal mol -1 Å -2 . The Langevin thermostat (friction coefficients 1.0 ps -1 ) was employed for the temperature regulation and the simulation was run with fixed volume (NVT ensemble). The harmonic restraints were gradually reduced in six steps from 100 to 5 kcal mol -1 Å -2 . Then, the density of the system was equilibrated for 20 ps using a time step of 0.2 fs by fixing the pressure, using the Langevin thermostat with isotropic pressure scaling (NPT ensemble), and allowing the volume of the box to change. 5 independent MD simulations of 100 ns were submitted for each of the systems (5 x 100 ns per system) at 300 K with a time step of 2 fs. To generate different initial conditions, the velocities of the system were randomly assigned during each of the heating steps. A harmonic restraint of 5 kcal mol -1 Å -2 was imposed on the alpha carbon atoms of the protein and ER in order to keep the secondary structure. The cut-off distance for the nonbonded interactions was 10 Å and periodic boundary conditions were used. Electrostatic interactions were treated by using the smooth particle mesh Ewald (PME) method [8] with a grid spacing of 1 Å. The SHAKE algorithm [9] was applied to all bonds involving hydrogen atoms. The trajectories were analyzed with the module cpptraj of AmberTools18 and the binding energies ∆Gbind of the 5-mer peptide 591-FPATpV-595 of ER/14-3-3FC-A/1 computed with MM-ISMSA. [ Figure S3. Evolution of distances d1-d4 (Å) along 5 independent 100ns-restrained MD simulations of each of the ER/14-3-3/FC-A/1 (labeled as FC-A) and ER/14-3-3/1 (no FC-A) complexes in solution. In each case, both binding sites in 14-3-3where monitored. The alpha carbon atoms of the protein where restrained with a harmonic force constant of 5 kcal mol -1 Å -2 .

Cytotoxicity assay
The relative in vitro cytotoxicities of compound 1 against 293T normal cells and HeLa cancer cells were assessed by using the Alamar Blue assay. Briefly, the cells were seeded in 96-well plates at a density of 1 × 10 5 cells per well in 100 L of DMEM containing 10% fetal bovine serum, supplemented with 50 UꞏmL −1 penicillin and 50 UꞏmL −1 streptomycin, and cultured in 5% CO2 at 37 °C for 24 h. Then, compound 1 with different concentrations was added into each well, and the two types of cells were further incubated for 24 h.
Subsequently, 10 L of Alamar Blue solution was added into each well and incubated for another 3 h in 5 % CO2 at 37° C. After that, the medium in each well was transferred to another black 96-well plate, and the fluorescence was measured at 590 nm using a multimode reader (GloMax-Multi+Detection System, Promega). Untreated cells in medium were used as the blank control. All experiments were carried out with three replicates. The cytotoxicity was expressed as the percentage of the cell viability relative to the blank control.

293T Normal Cells
HeLa Cancer Cells Figure S5. Cytotoxicity of 1 in Hela and 293T cell cultures after 24h incubation.

Chemistry
a. General Remarks Solvents were dried and distilled before use. Millipore water was obtained with a Micropure from TKA. All reactions were carried out in oven dried glassware. Microwave assisted SPPS was carried out with a CEM Discover. Analytical TLC was carried out on SiO2 aluminum foils ALUGRAM SIL G/UV254 from Macherey Nagel. The analytical "High Performance Liquid Chromatography" (HPLC) was done with Dionex HPLC apparatus: P680 pump, ASI-100 automated sample injector, UVD-340U UV detector, UltiMate 3000 Column Compartment. Commercially available HPLC grade solvents were used as eluents and solvent mixtures are reported in volume percent. Lyophilization was carried out with an Alpha 1-4 2D plus freeze drying apparatus from Christ. Reversed phase column chromatography was done with an Armen Instrument Spot Flash Liquid Chromatography MPLC apparatus with RediSep C-18 Reversed-Phase columns. 1 H-and 13 C-NMR spectra were recorded on a Brucker DRX 500 MHz spectrometer at ambient temperature. The chemical shifts are reported in parts per million (ppm) relative to the deuterated solvent DMSO-d6. The following abbreviations are used for peak multiplicities: s, singlet; d, doublet; m, multiplet; br, broad. High resolution ESI mass spectra were recorded with a Bruker BioTOF III spectrometer. Determination of pH values was carried out with a pH-Meter766 Calimatic from Knick.

b. Synthesis and characterization
Compounds 1, 2 and 5 were synthesized as described in the literature. [11]