Pharmaceutical polymorph control in a drug-mimetic supramolecular gel

A supramolecular gel designed to chemically mimic the structure of a pharmaceutical compound controls the polymorphic outcome of the crystallization of the substrate.

2-Nitrophenylisocyanate (2.5 g, 15.2 mmol) was added to a solution of 4,4-methylenebis(2,6 diethylaniline) (2.36 g, 7.6 mmol) in dry chloroform (400 mL) at room temperature. The resulting mixture was stirred under nitrogen and a gelatinous precipitate start appearing after 5 minutes. The mixture was then heated under reflux for 18 hours and cooled to room temperature, filtered, washed with chloroform (4 x 50 mL) and dried. The yellow precipitate was then triturated with 200 ml chloroform for 24h and then filtered, washed with chloroform (100 ml) and dried. The solid was ground to give 2 as a fine pale yellow powder (3.9 g, 6.1 mmol, 80 %  (

Gel Preparation
Compounds 1, 2, 8, 9 and 10 were tested in a range of solvents for evidence of gel formation. The compound (0.01 g) was heated in 1 mL of solvent (1 % w/v) in a sealed vial until fully dissolved and then cooled to room temperature. After 24 h, gel formation was characterised by a simple vial inversion test; if the solvent was fully immobilised it was considered to have gelled. Compound 8 and 9 failed to gel any of the solvent tested but compound 10 was able to form gel in nitromethane and nitrobenzene.

Gel Characterisation
T gel Characterisation of Compound 2 T gel was measured by the drop ball method using a custom-made glass ball (0.25 g). In a typical experiment, the gelator (0.01 g) was heated in 1 mL of toluene (1 % w/v) in a sealed vial until fully dissolved. The solution was cooled to room temperature and after 12 hours the vial was opened and the glass ball was placed on the gel surface, sealed and gradually heated in an oil bath. The temperature at which the ball drops into the bottom of the vial was recorded as the gel dissociation temperature (T gel ) and was found to be 92 ⁰C at 1 wt %. The minimum gel concentration (MGC) of 2 in toluene was found to be 0.0075g/mL

Rheology of Compound 2
Rheology experiments were performed using a TA Instruments Advanced Rheometer 2000. A concentric cylinder couette geometry (25 mm rough plate) with a gap of 2500 μm and 2mL of sample was used in each case. 0.02 g/mL of gelator in toluene (2 mL) was prepared in a glass vial, sealed and carefully heated until the gelator had fully dissolved. The hot gelator solution was transferred into the glass cylindrical mould (diameter 30 mm) on the rough plate, cooled to 20°C and equilibrated for 30 minutes. Oscillatory stress sweep measurement was performed over a range of 0.01-100 Pa with a constant frequency value of 1 Hz. In addition frequency sweep measurements were performed at 1% w/v.

Structures of non-ROY-specific gelators used in control crystallizations
The control gelators 3 -7 used in the initial screen are shown below. Compounds 3, 4 and 6 have been reported previously as described in the main text. Compounds 5 and 7 were prepared as detailed below.
Crystal form observed at time interval Gelator ROY mg/mL toluene 24 hours 48 hours 72 hours 96 hours 2 weeks >1 month 100 Typical IR spectra for ON (blue), R (red) and Y (green) polymorphs of ROY grown from gels.   Figure S3. DrySyn Multi-reaction station used for controlled crystallization experiments.

Figure S4
Crystallization of ROY in gels of 2. In a typical experiment, 1 mL toulene was added to gelator (10 mg) and ROY (100 mg) in a vial, sealed and heated to 140 ⁰C in a DrySyn Multi-reaction station until all solids had completely dissolved. The vials were removed from the reaction station and allowed to cool to room temperature.

Figure S5
Vertically offset XRPD data for gelator 2. The bulk solid was prepared from chloroform and then purified by washing the compound with copious quantities of chloroform (see experimental).

Conformational searches
A search for conformers of 2 was performed using a low-mode conformational search (I. Kolossváry, W. C. Guida, J. Am. Chem. Soc. 1996, 118, 5011-5019.) method, as implemented in MacroModel (Schrodinger LLC, New York, NY, MacroModel, V9.0, 2011.). This is a mode-following algorithm -a starting molecular geometry is perturbed along one or a combination its calculated normal modes before re-minimising. The OPLS-AA (W. L. Jorgensen and J. Tirado-Rives, J. Am. Chem. Soc., 1988, 110, 1657-1666 force field was used in these searches. Minimum and maximum move distances of 3 and 6 Å were applied and 32,000 search steps were performed. A gradient of < 0.05 kJ mol -1 Å -1 was set as a criterion for convergence of geometry optimisations. All conformations within a 50 kJ mol -1 window of the global minimum were saved from the initial search, both to keep all conformers that might be relevant to crystal packing and to allow for significant inaccuracies of the force field. Duplicate molecular geometries were identified and removed first using an all-atom RMS deviation of atomic positions (within Macromodel), with a 0.05 Å tolerance, followed by clustering based on selected dihedral angles (performed using in-house software), with tolerances of 5 RMS and 10 maximum dihedral angle difference (in degrees) to identify duplicate conformers.
Specific searches for anti-anti conformer were performed, starting with both urea groups in the anti-anti conformation, again performing 32,000 search steps, but allowing higher energy conformers to be saved.