Stereodivergent Chirality Transfer by Noncovalent Control of Disulfide Bonds

Controlling dynamic stereochemistry is an important challenge, as it is not only inherent to protein structure and function but often governs supramolecular systems and self-assembly. Typically, disulfide bonds exhibit stereodivergent behavior in proteins; however, how chiral information is transmitted to disulfide bonds remains unclear. Here, we report that hydrogen bonds are essential in the control of disulfide chirality and enable stereodivergent chirality transfer. The formation of S–S···H–N hydrogen bonds in solution can drive conformational adaption to allow intramolecular chirality transfer, while the formation of C=O···H–N hydrogen bonds results in supramolecular chirality transfer to form antiparallel helically self-assembled solid-state architectures. The dependence on the structural information encoded in the homochiral amino acid building blocks reveals the remarkable dynamic stereochemical space accessible through noncovalent chirality transmission.


DFT calculations
Computational analysis was employed to optimize the lower energy conformer structures of the ground state minima of AA-L-Ala, MAA-L-Ala and MAA-L-t-Leu. Due to the dynamic nature of the molecules involved, all the structures were pre-screened using the CREST driver in the xTB software [1] using the semiempirical GFN2-xTB level. The conformers thus obtained were then reoptimized with DFT at ωB97X-D/def2-TZVP level. All DFT optimizations were conducted with the Gaussian 16, Rev B.01 software package. [2] All minima were confirmed to be such due to the absence of imaginary frequencies and the Gibbs free energies obtained after calculating the Hessian were used to sort the conformers. The electronic circular dichroism was calculated at the TD-ωB97X-D/def2-TZVP level over the first 30 singlet transitions. All xyz coordinates of the most stable conformers considered in this work are provided as separate additional file.

Single crystal preparation
The single crystals were prepared by slow evaporation of the solvent at low temperature (0 ~ 4 o C). The sample powders were dissolved in diethyl ether to obtain homogeneous yellow solution. Then 1 vol. equivalent amount of heptane was added and mixed. The resulting homogeneous solution was filtrated two times by cotton filter and then transferred into glass vials, which were sealed by para film with a few small holes. Then the vials were placed into a ventilated fridge to evaporate the solvents slowly. The crystals can be collected after 2 ~ 3 days and then stored in dark for measurement.

X-ray single crystal analysis
A single crystal sample was mounted on top of a cryoloop and transferred into the cold nitrogen stream (100 K) of a Bruker-AXS D8 Venture diffractometer. Data collection and reduction was performed using the Bruker software suite APEX3. [3] The final unit cell was obtained from the xyz centroids of 9824 reflections after integration. A multiscan absorption correction was applied, based on the intensities of symmetry-related reflections measured at different angular settings (SADABS). The structures were solved by direct methods using SHELXT. [4] and refinement of the structure was performed using SHLELXL. [5] The hydrogen atoms were generated by geometrical considerations, constrained to idealized geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms.

AA-L-Ala:
Asparagusic acid (AA) (1.06 g; 10 mmol) was dissolved in CH2Cl2 (50 mL) in flask, forming a homogeneous bright yellow solution. The solution was cooled in ice bath and then EDCI (1.5 eq) and HOBT (1.5 eq) were added and completely dissolved under continuous stirring. After stirring for 15 min, L-Alanine methyl ester hydrochloride salt (2.06 g; 15 mmol) was deprotonated by DIPEA (2.6 mL; 15 mmol) in CH2Cl2, which was dropwise added into the mixture solution. Then the reaction solution was stirred at room temperature overnight. After confirming the full conversion of reactants by TLC tracing, the mixture solution was washed by 1 M HCl (aq) (30 mL × 3), H2O (30 mL × 1), saturated NaHCO3 (30 mL × 3), and brine (30 mL × 1). The organic phase was dried by anhydrous Na2SO4, and then purified by flash chromatography (SiO2, CH2Cl2/methanol = 200 : 0.25 to 200 : 1). The products were collected by evaporating the solvents under vacuum, affording colorless needle crystals (1.32 g; yield = 56%), which can be stored in dark for several months. 1

MAA-R-Butyl-ester:
MAA (1.64 g; 10 mmol) was dissolved in CH2Cl2 (100 mL) in flask, forming a homogeneous bright yellow solution. The solution was cooled in ice bath and then EDCI (2.33 g; 15 mmol) and DMAP (122 mg; 1 mmol) were added and completely dissolved under continuous stirring. After stirring for 5 min, the (S)-(+)-2-butanol (1.0 g; 14 mmol) were dissolved in CH2Cl2, which was dropwise added into the mixture solution. Then the reaction solution was stirred at room temperature overnight. After confirming the full conversion of reactants by TLC tracing, the mixture solution was washed by H2O (50 mL × 3), saturated NaHCO3 (50 mL × 2), and brine (50 mL × 1). The organic phase was dried by anhydrous Na2SO4, and then purified by flash chromatography (SiO2, pentane/CH2Cl2 = 10 : 1 to 5 : 1). The products were collected by evaporating the solvents under vacuum, affording yellow oil (0.68 g; yield = 32%), which can be stored in diluted solutions in dark.

MAA-N-Me-L-Ala:
MAA (1.64 g; 10 mmol) was dissolved in CH2Cl2 (100 mL) in flask, forming a homogeneous bright yellow solution. The solution was cooled in ice bath and then EDCI (2.33 g; 15 mmol) and DMAP (122 mg; 1 mmol) were added and completely dissolved under continuous stirring. After stirring for 5 min, methyl (2S)-2-(methylamino)propanoate (1.0 g; 8.5 mmol) were dissolved in CH2Cl2, which was dropwise added into the mixture solution. Then the reaction solution was stirred at room temperature for four days. The conversion of the reaction is very slow possibly due to the steric hinderance. After that the mixture solution was washed by diluted HCl aqueous solution (0.5 M; 50 mL × 2), saturated NaHCO3 (50 mL × 2), and brine (50 mL × 1). The organic phase was dried by anhydrous Na2SO4, and then purified by flash chromatography (SiO2, CH2Cl2/CH3OH = 400 : 1 to 200 : 1). The products were collected by evaporating the solvents under vacuum, affording yellow oil (~15 mg; yield = ~1%). The freshly prepared compounds are immediately used for spectroscopic characterization.
Yellow oil 1 H NMR (CDCl3, 300 MHz, 298 K, ppm): δ = 4.93 (t, J = 7.2 Hz, 1H), 3.70 (s, 3H), 3.62 (dd, J1 = 12.0 Hz, J2 = 11.5 Hz, 3H), 3.05 (d, J = 11.7 Hz, 2H), 1.48 (s, 3H), 1.41 (d, J = 7.5 Hz, 2H). 13    Polarity-dependency of the CD spectra (A) and UV-Vis spectra (B) of MAA-L-Ala. Optical path = 10 mm. Acetonitrile solution exhibited very low ellipticity in CD spectra, and meanwhile slightly red-shifted absorption band in UV-Vis spectra, indicating the higher conformational freedom of the disulfide five-membered ring in acetonitrile, which enables more metastable and more planar conformers with red-shifted absorption bands.         Figure S13 with the intensities scaled using the Boltzmann averaged population.       IR spectra of the amide vibration band (νNH) of MAA-L-Ala in 10 mM CDCl3 solutions at 293K and the van't Hoff fitting plot according to VT-NMR spectra. The peaks were separated into two major peaks (the H-bonded and the unbonded), whose integration area was further used to calculate the equilibrium constant at 293K (K293K). To evaluate the thermodynamic parameters, van't Hoff plots were fit by combining the IR data and temperature-varied 1 H NMR spectra. The detailed fitting method has been described in Method section. Red band indicates 95% confidence intervals.

MAA-L-Leu
Van't Hoff plot: ΔH = -1.62 ± 0.06 kcal/mol ΔS = -24.14 ± 0.91 J mol -1 K -1       Fig. S62B, it can be estimated that the species of P helicity (CD band I at 364 nm) is the predominant species; (B) Simulated CD spectra of the different conformations of MAA-L-t-Leu as shown in Figure S61; (C) Simulated CD spectra of the different conformations of MAA-Lt-Leu as shown in Figure S61, with the intensities scaled using the Boltzmann averaged population.    Figure S63. B. Simulated CD spectra of the different conformations of AA-L-Ala as shown in Figure S63, with the intensities scaled using the Boltzmann averaged population.