Chirality imprinting and direct asymmetric reaction screening using a stereodynamic Brønsted/Lewis acid receptor

Molecular recognition, activation and dynamic self-assembly with Brønsted and Lewis acids play a central role across the chemical sciences including catalysis, crystal engineering, supramolecular architectures and drug design. Despite this general advance, the utilization of the corresponding binding motifs for fast and robust quantitative chemosensing of chiral compounds in a complicate matrix has remained challenging. Here we show that a stereodynamic probe carrying complementary boronic acid and urea units achieves this goal with hydroxy carboxylic acids. Synergistic dual-site binding and instantaneous chirality imprinting result in characteristic ultraviolet and CD readouts that allow instantaneous determination of the absolute configuration, enantiomeric excess and concentration of the target compound even in complex mixtures. The robustness and practicality of this strategy for high-throughput screening purposes is demonstrated. Comprehensive sensing of only 0.5 mg of a crude reaction mixture of an asymmetric reduction eliminates cumbersome work-up protocols and minimizes analysis time, labour and waste production.


Supplementary
). CD analysis was conducted immediately following the addition of the substrate using sample concentrations of 1.80 x 10 -4 M in ACN with a standard sensitivity of 100 mdeg, a data pitch of 0.5 nm, a bandwidth of 1 nm, a scanning speed of 500 nm s -1 and a response of 0.5 s using a quartz cuvette (1 cm path length). The data were baseline corrected and smoothed using a binomial equation (Supplementary Figures 2-8). Control experiments with 3-9 did not show any CD signal at the wavelength of interest.

Quantitative ee and concentration analysis 2.1. Ee determination using mandelic acid 3
The change in the CD as a function of sample ee was investigated using samples of the complexes derived from 1 and varying ee of 3. A stock solution of 1 (0.006 M in ACN) was prepared and 500 μL portions were placed in 4 mL vials. Into these vials, solutions of 3 (0.15 M in ACN) of varying enantiomeric composition (+100, +80, +60, +40, +20, 0, -20, -40, -60, -80, -100 %ee) were added. CD analysis was carried out as described above at 1.80 x 10 -4 M in ACN. The Cotton effect amplitudes at 300 and 330 nm were plotted against the enantiomeric excess of 3 ( Supplementary Figures 9-11). Five solutions of 1 were prepared and 3 was added at varying enantiomeric compositions. Using the regression equation obtained above and the measured CD intensity at 300 nm and 330 nm, the ee of these samples was determined (Supplementary Table 1).

Determination of the concentration of 3
The change in the UV signature of 1 upon addition of 3 was analyzed. A stock solution of 1 (0.006 M in ACN) was prepared and 500 μL portions were placed in 4 mL vials. A stock solution of 3 (0.15 mmol in 1 mL ACN) and Et3N (21 μL, 0.15 mmol) was also prepared. To the solutions of 1 was added 3 in varying amounts (0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 120, 140, 160, 180 and 200 mol%). UV spectra were collected with an average scanning time of 0.1 s, a data interval of 1nm, and a scan rate of 600 nm/min. The UV absorbance at 320 and 330 nm increased steadily upon addition of up to 1 equivalent of 3. When the concentration of 3 was in excess of 100 mol%, the UV absorbance remained mostly constant. Plotting and curve fitting of the UV absorbance at 320 and 330 nm against the molar ratio of [3]/[1] from 0 to 100 mol% gave linear equations ( Supplementary Figures 12-16). Five solutions of 3 at varying concentrations were prepared and analyzed as described above.
Using the regression equations obtained as described above and the UV absorbance at 320 and 330 nm, the concentration of these samples was determined (Supplementary Table 2).

NMR and CD analysis of the chemosensing mechanism
A solution of 1 (10 mg, 0.02 mmol), 3 (3.04 mg, 0.02 mmol) and Et3N (2.8 μL, 0.02 mmol) in ACN-d3 (0.5 mL) was subjected to 1 H and 11 B NMR analysis. The 1 H NMR spectra showed a strong downfield shift for the urea protons upon addition of 3 and Et3N. The 11 B NMR showed an upfield shift upon addition of 3 in both the presence and absence of Et3N. For all 11 B NMR spectra, a spectrum of pure ACN was subtracted to eliminate baseline noise. 11 B NMR analysis was also conducted using 1 with O-acetylmandelic acid 14 (3.95 mg, 0.02 mmol). No shift in the 11 B NMR was observed ( Supplementary Figures 20-22).
Four samples of sensor 1 and (R)-3 (1:1) were generated in anhydrous ACN. A stock solution of 1 was prepared by dissolving 11.8 mg (0.024 mmol) in 4 mL of anhydrous ACN. This stock solution was then separated into 0.5 mL (0.003 mmol) portions. To a stock solution of (R)-3 (22.8 mg, 0.15 mmol in 1 mL of anhydrous ACN) were added 21 L of Et3N (0.15 mmol). To each 0.5 mL sensor stock solution were then added 20 L (0.003 mmol) of the mandelic acid/Et3N solution. To these samples were added 0.0 (blue), 0.5 (red), 1.0 (green) and 10 (purple) molar equivalents of water and CD spectra were collected at a concentration of 1.8 x 10 -4 M (Supplementary Figure 23).

Ee and concentration analysis of mandelic acid 3 obtained by reduction of phenylglyoxylic acid 12 with (+)-DIP-Cl
Solutions of phenylglyoxylic acid, 12, (10 mg, 0.06 mmol), an amine additive (0.06 mmol) and (+)-DIP-Cl (21.4 mg, 0.067 mmol) in 0.5 mL of anhydrous solvent were stirred in 4 mL vials under air for 12 hours at room temperature ( Supplementary Figures 24 and 25). The reaction was quenched with 1M NaOH (100 μL, 0.1 mmol) and H2O2 (10 μL, 30% in H2O, 0.3 mmol) and stirred for 30 minutes. 1M HCl was then added (150 μL) and the solvent was removed in vacuo. From the crude reaction mixture, 1 mg of the white solid was removed for UV and CD analysis. For traditional analysis (gravimetry and chiral HPLC) the remaining portion of the material was purified by flash chromatography on silica gel (EtOAc) to give 3 as a white solid. Each column consumed ~100 mL of solvent and required ~16 minutes, including column packing, collection, and solvent removal. For chiral HPLC analysis, 3 was converted to the methyl ester 13 by refluxing in 5 mL of anhydrous methanol for three hours in the presence of p-TSA (0.1 molar equivalent). The ee of the methyl mandelate 13 was determined by HPLC on a Chiralcel OD column using hexane:i-PrOH (80:20 v/v) as mobile phase at 1 mL/min, t1, (R) = 5.6 min, t2, (S) = 9.2 min and required ~15 mL of solvent and ~12 minutes per sample (Supplementary Figures 26-31).
The crude solid (0.5 mg) was added to a 0.5 mL solution of the sensor (1.48 mg, 0.003 mmol) in ACN. Triethylamine (4.2 μL, 0.03 mmol) was then added. CD analysis was conducted as described above. If a CD signal was observed, subsequent UV analysis was conducted as described above. The molar ratio  of [3]/[1] was calculated using equations 1 and 2 derived from the calibration curves generated from the UV absorbances at 320 and 330 nm. The average value  was then used to determine the enantiomeric excess using equations 4 and 5, derived from the calibration curves at 300 and 330 nm. The previously determined UV and CD calibration curves were used for all analyses and new calibrations were not required. The results are listed in Supplementary Tables 3-5 CD sensing of the reaction mixtures of experiments 6-9 and 11-16 did not show a measurable ee and these runs were not further analyzed. The results of all other runs were analyzed by UV/CD sensing and traditional gravimetric analysis of isolated product and chiral HPLC. The direct chiroptical chemosensing of the yield, ee, and sense of asymmetric induction from 0.5 mg of the crude reaction mixture of 3 obtained by reduction of phenylglyoxylic acid with (+)-DIP-Cl required significantly less time and solvent than gravimetry and chiral HPLC. Each optical measurement required less than 30 seconds and solvent consumption was reduced to less than 3 mL (for sample dilution) for each sample (a single sample was used for both CD and UV analysis). Importantly, the chirality sensing of the crude product mixtures eliminates elaborate purification steps as well as product derivatization for ee analysis. By contrast, the traditional analysis required purification via flash chromatography followed by gravimetric analysis and chiral HPLC of the methyl ester of 3. This led to significantly higher solvent consumption (120 mL) and analysis time (3.5 hours) per sample (Supplementary Table 6).

Crystallography
A single crystal of compound 2 was obtained by slow evaporation of a concentrated ACN solution (Supplementary Figure 35). Crystallographic analysis was performed at 100 K using a Siemens platform diffractometer with graphite monochromated Mo-Kα radiation (λ = 0.71073 Å). Data were integrated and corrected using the Apex 2 program. The structure was solved by direct methods and refined with full-matrix least-square analysis using SHELX-97-2 software. Non-hydrogen atoms were refined with anisotropic displacement parameters. The asymmetric unit contains two molecules of 2. Crystal structure data: Formula C17H10F6N2O, M = 372.27, crystal dimensions 0.21 x 0.20 x 0.14 mm, triclinic, space group P-1, a = 9.0933(4) Å, b = 13.1316 (5)