An empirical calibration of a Karplus relationship between the JHH-vicinal coupling constant and the corresponding dihedral angle is achieved for the H-N-CO-H" sequence in amides and peptides. A first approach is made by referring to several heterocyclic rigid molecules such as 2-pyrrolidinone, 3-isoquinuclidone, and other ring structures. The results so obtained are subsequently confirmed by use of open-chain peptide model compounds previously studied, as regard to their conformation in solution, by infrared spectroscopy. The quite satisfactory agreement between these two sets of experimental data leads us to conclude that the measured J coupling constant is not much influenced by the cis or trans conformation of the amide linkage. In the same way, it seems that the "strain effect" (k, the variation of J resulting from a distortion of the bond angles) for the H-N-C*-H* sequence is likely to be smaller than the A J correction estimated through the formulas commonly used in the case of the H-C-C-H ethane-like sequence.In most cases, the extensive application of nmr spectroscopy for elucidating the conformations of peptide molecules mainly relies on the measurement of the vicinal H-N-Ca-Ha coupling constant. From the theoretical work of Karplus,2a it is usually assumed that this coupling constant J is related to the dihedral angle 8 by ( Figure 1).The relationship between 0 and the rotational angle @ commonly used, according to the IUPAC-IUB recommendationsZb in peptide conformational studies, isThe nature and the hybridization state of the atoms involved in the coupling greatly influence the values of the coefficients A, B, and C; it is therefore necessary to parametrize, through an empirical calibration for each series of homologous compounds, the most appropriate Karplus relation.In the case of amides and peptides, such an attempt has been undertaken simultaneously by Bystrov, et by Ramachandran, et al.,5 and by o u r s e l v e~.~,~ The calibration by Bystrov, et ~l . ,~ relies mainly on measurements performed with two standard compounds, i.e., N-methylformamide and N-acetyldialanyl methyl ester.Moreover, it takes account of an erroneous J coupling constant of 8.2 Hz previously reported on the basis of the nmr spectrum of 6-phenyldihydroura~il.~ This value has subsequently been rectified and corrected to 2.6 H z .~ For the same purpose, Ramachandran, et ~l . ,~ make use of eight flexible compounds. Because of the numerous conformational states available for such molecules, the resulting unequivocal plotting of J us. 0 obviously needs an averaged estimate based on some theoretical analysis, the reliability of which remains questionable.10 Furthermore, up to this date, theory is mainly concerned with molecules "in uacuo," free from any intermolecular constraints. This condition is not fulfilled in the experiments reported by these authors, the measurements being generally performed with samples taken in the pure liquid state or dissolved in dimethyl sulfoxide, a solvent which is well known to interact strongly wit...