How reliable could economic Hartree–Fock computations be in studying large, folded peptides? A comparative HF and DFT case study on N- and C-protected aspartic acid

In this study, potential energy hypersurfaces have been generated and analyzed for each of the nine possible backbone (BB) conformations for both the endo and exo forms of N-acetyl-l-aspartic acid N′-methylamide. Ab initio calculations were carried out at RHF/3-21G, RHF/6-31G(d), and B3LYP/6-31G(d) levels for all backbone conformations. The relative energies, as well as stabilization energies exerted by the sidechain (SC) on the backbone, were calculated for all stable conformers. All sidechain–sidechain (HO⋯OC), backbone–backbone (N–H⋯OC), and sidechain–backbone (N–H⋯OC; N–H⋯OH) hydrogen bond interactions were analyzed. The appearance of the traditionally absent αL and εL conformers may be recognized as special geometric orientation which the aspartyl residue manifests during peptide folding or ligand docking in a receptor that contains aspartic acids in its ligand recognition sites. At all three levels of theory, there exists a trend between the hydrogen bond distance and ring size. In addition, strikingly high correlations between the torsional angles (R2=0.9937 for RHF/6-31G(d) versus RHF/3-21G; R2=0.9967 for B3LYP/6-31G(d) versus RHF/6-31G(d); R2=0.9914 for B3LYP/6-31G(d) versus RHF/3-21G) and between the ΔE values in kcal/mol (R2=0.9424 for RHF/6-31G(d) versus RHF/3-21G; R2=0.9108 for B3LYP/6-31G(d) versus RHF/6-31G(d); R2=0.9434 B3LYP/6-31G(d) versus RHF/3-21G) found at the different ab initio levels suggest that calculations carried out at the lower levels (i.e. at RHF/3-21G) are still significant.