A high-throughput compatible workflow for the biochemical identification and characterization of molecular glues

Abstract

Molecular glues are an emerging modality that induces or enhances an interaction between two proteins. Molecular glues can target proteins via proximity-induced degradation or sequestration and can, therefore, provide opportunities for therapeutic intervention to targets that cannot be modulated by traditional small-molecule approaches. Due to their modest molecular weight, molecular glues may not encounter the bioavailability issues associated with PROTACs. Characterization of molecular glues in hit finding and hit optimization settings can be challenging, as both the affinity of the glue for the target protein and the resulting improvement in affinity between the proteins of interest need to be assessed in parallel. Here, we propose and validate a workflow to derive both key parameters from a classic concentration response experiment. Furthermore, we provide a method for the rational determination of optimum biochemical assay conditions to identify and characterize molecular glues.

Keywords: drug discovery; drug screening; fluorescence resonance energy transfer (FRET); high-throughput screening (HTS); mathematical modeling; protein-protein interaction.

Figure 1

Possible binding schemes illustrating binary and ternary complex formation induced by PROTACSand molecular glues,L = Ligand (PROTAC or molecularglue), T = Target proteinandE = Effector protein.A, schematic illustration of the thermodynamic cycles of a PROTAC (L) induced ternary complex between a T and E. B, a general case illustrating all possible pathways and complexes that a molecular glue could elicit in a two-protein system. C, simplified version of (B) when a molecular glue binds to E but not to T. D, simplified version of (B) when a molecular glue binds to T but not to E.

Figure 2

Biochemical assays for the characterization of a molecular glue (modeled data).A, example modeled data from a Molecular Glue; Target protein matrix at fixed effector protein concentration (For the current case, we assume the molecular glue binds to the effector protein). Steady state data modeling was performed using KinTek (10), see experimental methods for details. B, same as (A) but showing data from a K_D shift assay format. The molecular glue and effector protein concentration is fixed, and the target protein is titrated. C, same as (A) but showing data from a concentration response assay format. Both target and effector proteins are fixed, and the molecular glue is titrated.

Figure 3

Characterization of the β-TrCP1:β-catenin molecular glue NRX-252262.A, Left: X-ray structure showing the binding pose of the NRX-103094 (an analogue of NRX-252262), bound to the β-catenin peptide:β-TrCP1 complex (6M91) (8). X-ray structure illustration was prepared using PyMOL (9). Orange = β-catenin peptide, Blue = β-TrCP1, Green = NRX-103094, Grey = SKP1. Right: The chemical structures of NRX-252262 and NRX-103094. B, determination of the basal and NRX-252262-induced K_D for the β-catenin:β-TrCP1 interaction. The FAM-labeled β-catenin peptide 1 was titrated in the presence or absence of a saturating concentration of NRX-252262. The basal K_D was too weak to fully define the binding isotherm, and therefore, the maximum signal value from NRX-252262 was used as a constraint during the basal K_D fit. C, concentration-response curves for NRX-252262 at different FAM-β-catenin peptide 1 concentrations: β-TrCP1 is at a fixed concentration while NRX-252262 is titrated. D, schematic illustration of the thermodynamic cycles of a molecular glue showing the dominant pathway at the bottom of a concentration-response curve (left), and the top of a concentration-response curve (right). The black arrow shows that the transition between these two states, the span of a concentration-response curve, is driven by the change from K_D¹ to αK_D¹. Molecular glue = L, Target protein = T, and Effector protein = E, Grayed out pathways and species are providing no or negligible contribution to the behaviour of the system under the stated conditions.

Figure 4

Modeled data showing the predicted relationship between glue induced K_D shifts (1/α) and the span of an EC₅₀ curve (S_n) for an EC₅₀ assay, which is set at different fractions of the basal K_D (fK_D). Results of modeling Equation 1. This plot shows how the normalized span (S_n) of an EC₅₀ curve changes for a wide range of glue-induced affinity shifts. fK_D refers to the fraction of basal K_D the EC₅₀ assay has been set at, e.g. fK_D of 1 = 1 × K_D, fK_D of 0.1 = 0.1 × K_D, fK_D of 0.01 = 0.01 × K_D.

Figure 5

Chemical structure of NRX-252262 and the seven analogues which were synthesized (Denoted as compounds 1–7).

Figure 6

Comparison between experimental and predicted S_n as a function of molecular glue-induced K_D shift.A, determination of glue-induced K_D values for β-TrCP1:β-catenin. The FAM-labeled β-catenin peptide was titrated in the absence or presence of 100 μM of each molecular glue. B, concentration-response curves for Compound 6 at different FAM-β-catenin peptide concentrations. β-TrCP1 is at a fixed concentration while Compound 6 is titrated. Colored double arrows indicate the Span of the EC₅₀ curves. C, plots showing the normalized span (S_n) of the concentration-response curves of panel B at different fK_Dversus the K_D fold shift values obtained from panel A. The solid lines represent the predicted relationship between span and K_D fold shift at each fK_D value according to Equation 1. D, plot illustrating the actual versus predicted span, demonstrating the performance of the model for the β-TrCP1:β-catenin glues. The red dotted line represents the line of identity.

Figure 7

Proposed workflow for molecular glue assay development and screening. Schematic showing a proposed workflow for the identification and characterization of molecular glues during a drug discovery campaign for cases where molecular glues bind to the effector protein. An iterative process of assay development and compound screening is required to ensure that assay conditions remain suitable as the molecular glue's affinity for the effector protein and the K_D shift they induce both improve during compound optimization cycles. A similar workflow can be used for molecular glues that bind to the target protein.

Figure 8

The relationship between signal, Span and K_D shift for an example molecular glue in a concentration-response and K_D shift experiment KinTek modeled data showing the relationship between Signal (Y) and Span (S^inf-S⁰) in a K_D shift (black) and concentration-response experiment (red) respectively, for an example molecular glue. Y = E:T + E:T:L, equivalent to the total amount of protein-protein complex.

NRX-252262: 4-((2,6-dichlorophenyl)thio)-3-(5,6-dimethoxyisoindoline-2-carbonyl)-6-(trifluoromethyl)pyridin-2(1H)-one: Synthesized following the procedures described by Simonetta et al. (8). White solid. 1H NMR (400 MHz, DMSO-d6) δ 3.71 (s, 3H), 3.76 (s, 3H), 4.43 to 4.7 (m, 2H), 4.77 (s, 2H), 6.31 (s, 1H), 7.00 (d, J = 17.2 Hz, 2H), 7.63 (dd, J = 8.6, 7.6 Hz, 1H), 7.76 (d, J = 8.0 Hz, 2H), 12.84 (s, 1H); MS (ESI) m/z = 544 [M-1].

4-Chloro-2-oxo-6-(trifluoromethyl)-1,2-dihydropyridine-3-carboxylic acid: Synthesized following the procedures described by Simonetta et al. (8). White solid. 1H NMR (500 MHz, DMSO-d6) δ 7.60 (s, 1H); MS (ESI) m/z = 240 [M-1].

4-chloro-3-(5,6-dimethoxyisoindoline-2-carbonyl)-6-(trifluoromethyl)pyridin-2(1H)-one: A solution of T4P in EtOAc (28.3 g, 39.33 mmol) was added dropwise to 4-chloro-2-oxo-6-(trifluoromethyl)-1,2-dihydropyridine-3-carboxylic acid (3.8 g, 15.73 mmol), 5,6-dimethoxyisoindoline HCl (4.07 g, 18.88 mmol) and N,N-Diisopropylethylamine (8.24 ml, 47.20 mmol) in DMF (20 ml) at 0°C. The resulting mixture was stirred at r.t. for 1 h 1M NaOH aq. (131 ml, 786.59 mmol) was added, and the resulting mixture was stirred at r.t. for 1 h. The reaction mixture was quenched with water (200 ml), extracted with EtOAc (3 × 100 ml), the organic layer was dried over Na₂SO₄, filtered, and evaporated to afford crude product. The crude product was purified by flash silica chromatography, elution gradient 0 to 25% EtOAc in petroleum ether. Pure fractions were evaporated to dryness to afford the title compound (5.80 g, 92%). White solid. 1H NMR (300 MHz, DMSO-d6) δ 3.70 (s, 3H), 3.76 (s, 3H), 4.36 to 4.56 (m, 2H), 4.76 (s, 2H), 6.89 (s, 1H), 7.02 (s, 1H), 7.68 (s, 1H), 13.10 (s, 1H); MS (ESI) m/z = 403 [M+1].

4-((2,6-dichlorophenyl)thio)-2-oxo-6-(trifluoromethyl)-1,2-dihydropyridine-3-carboxylic acid: To a solution of 4-chloro-2-oxo-6-(trifluoromethyl)-1,2-dihydropyridine-3-carboxylic acid (480 mg, 1.99 mmol) and 2,6-dichlorobenzenethiol (534 mg, 2.98 mmol) in NMP (10 ml) was added K₂CO₃ (1099 mg, 7.95 mmol). The solution was stirred at 110 °C for 1 h. The reaction mixtue was diluted with EtOAc (10 ml) and washed with water (10 ml) and brine (10 ml). The aqueous layer was extracted with EtOAc (1 x 40 ml) and the combined organic layers were collected via a hydrophobic frit to afford crude product. The crude product was purified by reverse phase chromatography (Interchim C18-HP Flash column, 175 g), using decreasingly polar mixtures of water (containing by volume 1% NH₄OH (28–30% in H₂O)) and MeCN as eluents. Fractions containing the desired compound were evaporated to remove MeCN and acidified with 2M HCl. The aqueous solution was extracted with DCM (2 x 100 ml) before being concentrated in vacuo to afford the title compound (129 mg, 0.336 mmol, 16.9%). The aqueous layer was collected and filtered to afford a further crop of the title compound (353 mg, 0.918 mmol, 46.2%). Off white solid. 1H NMR (500 MHz, DMSO-d6) δ 3.18 (s, 3H), 6.03 (s, 1H), 7.65 (dd, J = 8.6, 7.6 Hz, 1H), 7.76 to 7.8 (m, 2H); MS (ESI) m/z = 382 [M-1].

Compound 1: 3-(5,6-dimethoxyisoindoline-2-carbonyl)-4-((2-methylbenzyl)thio)-6-(trifluoromethyl)pyridin-2(1H)-one: Into a 40 ml vial, 2-methylbenzylthiol (0.3 mmol, 42 mg) was added to 4-chloro-3-(5,6-dimethoxy-2,3-dihydro-1H-isoindole-2-carbonyl)-6-(trifluoromethyl)-1,2-dihydropyridin-2-one (0.10 mmol, 51 mg), NMP (1 ml), K₂CO₃ (0.3 mmol, 42 mg). The mixture was stirred at 150 °C for 3h. The crude product was purified by prep-HPLC with the following conditions: Xbridge prep C18 5vm 30∗150 mm and Welch Ultimate XB-C18:50∗250 mm,10vm. 5% - 90% MeCN in formic acid aq. (0.05%) or 5% - 90% MeCN in NH₄HCO₃ (10 mmol/L) to afford the title compound (7.8 mg, 14%). Off white solid. 1H NMR (600 MHz, DMSO-d6) δ 2.27 (s, 3H), 3.71 (s, 3H), 3.75 (s, 3H), 4.10 (d, J = 13.6 Hz, 1H), 4.37 (d, J = 13.6 Hz, 1H), 4.46 (d, J = 12.5 Hz, 1H), 4.49 (d, J = 12.2 Hz, 1H), 4.68 (s, 2H), 6.83 (s, 1H), 6.98 (s, 1H), 7.11 (dd, J = 7.6, 5.8 Hz, 2H), 7.15 to 7.21 (m, 1H), 7.27 to 7.31 (m, 1H), 7.56 (s, 1H), 12.39 (s, 1H). MS (ESI) m/z = 505 [M+1].

Compound 2: 4-((2,6-dichlorophenyl)thio)-3-(6-methylindoline-1-carbonyl)-6-(trifluoromethyl)pyridin-2(1H)-one: Into a 40 ml vial, 6-methylindoline (0.15 mmol, 20 mg) was added to 4-((2,6-dichlorophenyl)thio)-2-oxo-6-(trifluoromethyl)-1,2-dihydropyridine-3-carboxylic acid (0.10 mmol, 38 mg), HOBt (0.20 mmol, 27 mg), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (0.20 mmol, 31 mg), N,N-diisopropylethylamine (0.30 mmol, 39 mg), and DMF (2 ml). The mixture was stirred at room temperature for 16 h. The crude product was purified by Prep-HPLC with the following conditions: Xbridge prep C18 5 vm 30∗150 mm and Welch Ultimate XB-C18, 50∗250 mm,10 vm. 5% - 90% MeCN in formic acid aq. (0.05%) or 5%-90 % MeCN in NH₄HCO₃ (10 mmol/L) to afford the title compound (3.1 mg, 6%). Off white solid. 1H NMR (600 MHz, DMSO-d6) δ 2.34 (s, 3H), 3.12 (t, J = 8.6 Hz, 2H), 3.89 (q, J = 9.1 Hz, 1H), 3.98 (m, 1H), 6.30 (s, 1H), 6.94 (d, J = 7.6 Hz, 1H), 7.19 (d, J = 7.6 Hz, 1H), 7.63 (t, J = 8.1 Hz, 1H), 7.71 to 7.82 (m, 2H), 8.01 (s, 1H). MS (ESI) m/z = 499 [M+1].

Compound 3: 3-(5,6-dimethoxyisoindoline-2-carbonyl)-4-((4-ethylphenyl)thio)-6-(trifluoromethyl)pyridin-2(1H)-one: Into a 40 ml vial, 4-ethylbenzenethiol (0.3 mmol, 42 mg) was added to 4-chloro-3-(5,6-dimethoxy-2,3-dihydro-1H-isoindole-2-carbonyl)-6-(trifluoromethyl)-1,2-dihydropyridin-2-one (0.10 mmol, 51 mg), NMP (1 ml), K₂CO₃ (0.3 mmol, 42 mg). The mixture was stirred at 150 °C for 3 h. The crude product was purified by prep-HPLC with the following conditions: Xbridge prep C18 5vm 30∗150 mm and Welch Ultimate XB-C18:50∗250 mm,10vm. 5% - 90% MeCN in formic acid aq. (0.05%) or 5% - 90% MeCN in NH₄HCO₃ (10 mmol/L) to afford the title compound (7.2 mg, 14%). Off white solid. 1H NMR (600 MHz, DMSO-d6) δ 1.18 (t, J = 7.6 Hz, 3H), 2.65 (q, J = 7.6 Hz, 2H) 3.72 (s, 3H), 3.77 (s, 3H), 4.52 (d, J = 13.4 Hz, 1H), 4.57 (d, J = 13.7 Hz, 1H), 4.74 (s, 2H), 6.53 (s, 1H), 6.93 (s, 1H), 7.02 (s, 1H), 7.38 (d, J = 8.0 Hz, 2H), 7.54 (d, J = 7.9 Hz, 2H), 12.62 (s, 1H). MS (ESI) m/z = 505 [M+1].

Compound 4: 4-((3-chlorophenyl)thio)-3-(5,6-dimethoxyisoindoline-2-carbonyl)-6-(trifluoromethyl)pyridin-2(1H)-one: Into a 40 ml vial, 3-chlorophenylthiol (0.3 mmol, 43 mg) was added to 4-chloro-3-(5,6-dimethoxy-2,3-dihydro-1H-isoindole-2-carbonyl)-6-(trifluoromethyl)-1,2-dihydropyridin-2-one (0.10 mmol, 51 mg), NMP (1 ml), K₂CO₃ (0.3 mmol, 42 mg). The mixture was stirred at 150 °C for 3h. The crude product was purified by prep-HPLC with the following conditions: Xbridge prep C18 5 vm 30∗150 mm and Welch Ultimate XB-C18:50∗250 mm,10 vm. 5%–90% MeCN in formic acid aq. (0.05%) or 5%–90% MeCN in NH₄HCO₃ (10 mmol/L) to afford the title compound (5.4 mg, 10%). Off white solid. 1H NMR (600 MHz, DMSO-d6) δ 3.72 (s, 3H), 3.77 (s, 3H), 4.55 (d, J = 14.8 Hz, 1H), 4.59 (d, J = 13.1 Hz, 1H), 4.74 (s, 2H), 6.65 (s, 1H), 6.91 (s, 1H), 7.02 (s, 1H), 7.51 to 7.56 (m, 1H), 7.57 to 7.62 (m, 2H), 7.76 (s, 1H), 12.73 (s, 1H). MS (ESI) m/z = 511 [M+1].

Compound 5: 4-((2,6-dichlorobenzyl)thio)-3-(5,6-dimethoxyisoindoline-2-carbonyl)-6-(trifluoromethyl)pyridin-2(1H)-one: Into a 40 ml vial, 2,6-dichlorobenzyl thiol (0.3 mmol, 58 mg) was added to 4-chloro-3-(5,6-dimethoxy-2,3-dihydro-1H-isoindole-2-carbonyl)-6-(trifluoromethyl)-1,2-dihydropyridin-2-one (0.10 mmol, 51 mg), NMP (1 ml), K₂CO₃ (0.3 mmol, 42 mg). The mixture was stirred at 150 °C for 3 h. The crude product was purified by prep-HPLC with the following conditions: Xbridge prep C18 5 vm 30∗150 mm and Welch Ultimate XB-C18:50∗250 mm,10 vm. 5% - 90% MeCN in formic acid aq. (0.05%) or 5% - 90% MeCN in NH₄HCO₃ (10 mmol/L) to afford the title compound (5.5 mg, 9%). Off white solid. 1H NMR (600 MHz, DMSO-d6) δ 3.70 (s, 3H), 3.75 (s, 3H), 4.22 (d, J = 13.7 Hz, 1H), 4.45 (d, J = 13.6 Hz, 1H), 4.56 to 4.72 (m, 4H), 6.87 (s, 1H), 6.97 (s, 1H), 7.34 (dd, J = 8.6, 7.7 Hz, 1H), 7.47 (d, J = 8.1 Hz, 2H), 7.64 (s, 1H), 12.53 (s, 1H). MS (ESI) m/z = 559 [M+1].

Compound 6: 3-(5,6-dimethoxyisoindoline-2-carbonyl)-4-((2,4-dimethylphenyl)thio)-6-(trifluoromethyl)pyridin-2(1H)-one: Into a 40 ml vial, 2,4-dimethyl-thiophenol (0.3 mmol, 42 mg) was added to 4-chloro-3-(5,6-dimethoxy-2,3-dihydro-1H-isoindole-2-carbonyl)-6-(trifluoromethyl)-1,2-dihydropyridin-2-one (0.10 mmol, 51 mg), NMP (1 ml), K₂CO₃ (0.3 mmol, 42 mg). The mixture was stirred at 150 °C for 3 h. The crude product was purified by prep-HPLC with the following conditions: Xbridge prep C18 5 vm 30∗150 mm and Welch Ultimate XB-C18:50∗250 mm,10 vm. 5% - 90% MeCN in formic acid aq. (0.05%) or 5% - 90% MeCN in NH₄HCO₃ (10 mmol/L) to afford the title compound (8.9 mg, 16%). Off white solid. 1H NMR (600 MHz, DMSO-d6) δ 2.28 (s, 3H), 2.32 (s, 3H), 3.71 (s, 3H), 3.76 (s, 3H), 4.49 (d, J = 14.6 Hz, 1H), 4.58 (d, J = 13.7 Hz, 1H), 4.75 (s, 2H), 6.34 (s, 1H), 6.94 (s, 1H), 7.02 (s, 1H), 7.16 (d, J = 7.9 Hz, 1H), 7.30 (s, 1H), 7.48 (d, J = 7.8 Hz, 1H), 12.58 (s, 1H). MS (ESI) m/z = 505 [M+1].

Compound 7: N-(4-(1H-pyrazol-4-yl)phenyl)-4-((2,6-dichlorophenyl)thio)-2-oxo-6-(trifluoromethyl)-1,2-dihydropyridine-3-carboxamide: Into a 40 ml vial, 4-(1H-pyrazol-4-yl)aniline (0.15 mmol, 24 mg) was added to 4-((2,6-dichlorophenyl)thio)-2-oxo-6-(trifluoromethyl)-1,2-dihydropyridine-3-carboxylic acid (0.10 mmol, 38 mg), HOBt (0.20 mmol, 27 mg), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (0.20 mmol, 31 mg), N,N-diisopropylethylamine (0.30 mmol, 39 mg), and DMF (2 ml). The mixture was stirred at room temperature for 16 h. The crude product was purified by Prep-HPLC with the following conditions: Xbridge prep C18 5 vm 30∗150 mm and Welch Ultimate XB-C18, 50∗250 mm,10 vm. 5% to 90% MeCN in formic acid aq. (0.05%) or 5% - 90% MeCN in NH₄HCO₃ (10 mmol/L) to afford the title compound (4.2 mg, 8%). Off white solid. 1H NMR (600 MHz, DMSO-d6) δ 5.89 (s, 1H), 7.57 to 7.62 (m, 3H), 7.67 to 7.7 (m, 2H), 7.73 (d, J = 8.1 Hz, 2H), 8.02 (s, 2H). MS (ESI) m/z = 525 [M+1].