Computational Details

All the quantum mechanical calculations were performed using Gaussian 09 (Frisch et al., 2009). Systematic conformational searches were done for each compound in the gas phase using the MMFF force field, implemented in Spartan 14 (Spartan’14), and the results were validated using Macromodel (MacroModel, 2018) (MMFF force field, mixed torsional/low-mode sampling protocol) using an energy cutoff of 10 kcal/mol. The choice for the 10 kcal/mol of cutoff was set as a balance between reducing the overall CPU calculation time and minimizing the possibility of losing further contributing conformers. The numbers of unique conformations found within these boundaries were 92 for 1, 73 for 11epi-1, and 290 for 2. All conformers were kept for full geometry optimization at the RHF/3-21G level in gas phase. All structures within 5 kcal/mol from the corresponding global minima were reoptimized at the B3LYP/6-31G* level in gas phase. Frequency calculations were done at the same level to determine the nature of the stationary points found. The ECD calculations were carried out using the B3LYP/6-31G* optimized geometries. The excitation energies (nm) and rotatory strength (R) in dipole velocity (R _vel) of the first forty singlet excitations were calculated using TDDFT implemented in Gaussian 09 at the PBE0/def2-SVP and B3LYP/6-31G* levels from all significantly populated conformers, which were averaged using Boltzmann weighting. The Boltzmann amplitudes obtained by refining the Gibbs free energies of all compounds at the SMD/M06-2X/6-31G* level using methanol as solvent. The calculated rotatory strength were simulated into the ECD curve as the sum of Gaussians with 0.3 eV width at half-heights (σ), which were UV-corrected and scaled (Pescitelli and Bruhn., 2016).