A multiple-linear-regression analysis (MLRA) has been carried to study intermolecular hydrogen bonding and as model out using the Kamlet-Abboud-Taft (KAT) solvatochromic pacompounds for the peptide bond (1-3). In particular, Kamrameters in order to elucidate and quantify the solvent effects let et al. (4) used the solvatochromic comparison method on the 17 O chemical shifts of N-methylformamide (NMF), N,Nto unravel and rationalize solvent effects on the 15 N chemical dimethylformamide (DMF), N-methylacetamide (NMA), and shifts of amides. In recent years, 17 O chemical shifts d( 17 O) N,N-dimethylacetamide (DMA). The chemical shifts of the four have received attention as a structural probe in amides and molecules show the same dependence (in ppm) on the solvent peptides (5-15) since these shifts are extremely sensitive, polarity-polarizability, i.e., 022p*. The influence of the solvent in general, to both solvation and substituent effects (5-17). hydrogen-bond-donor (HBD) acidities is slightly larger for the A simple model for separating the different contributions acetamides NMA and DMA, i.e., 048a, than for the formamides to the 17 O nuclear shieldings due to hydrogen-bonding hydra-NMF and DMF, i.e., 042a. The influence of the solvent hydrogenbond-acceptor (HBA) basicities is negligible for the nonprotic mol-tion at various sites in amide molecules was suggested by ecules DMF and DMA but significant for the protic molecules Burgar, Amour, Fiat et al. (BAF) some years ago (5, 9). NMF and NMA, i.e., 09b. The effect of substituting the N-H Recently, an improved solvation model of amides and pephydrogen by a methyl group amounts to 05.9 ppm in NMF and 5.4 tides based on 17 O chemical shifts was presented by Gerothappm in NMA. The effect of substituting the O|C{H hydrogen nassis and Vakka (13, 15). These authors demonstrated that amounts to 5.5 ppm in NMF and 16.8 ppm in DMF. The model both long-range dipole-dipole interactions and specific hyof specific hydration sites of amides by I. P. Gerothanassis and drogen bonds due to solvation of molecules by H 2 O at the C. Vakka [J. Org. Chem. 59, 2341 (1994)] is settled in a more amide oxygen induce significant and specific modifications quantitative basis and the model by M. I. Burgar, T. E. St. Amour, of the 17 O shielding which are larger than originally considand D. Fiat [ J. Phys. Chem. 85, 502 (1981)] is critically evaluated. ered (5, 9). On the contrary, solvation of the N-H amide 17 O hydration shifts have been calculated for formamide (FOR) by hydrogen induces a small modification of this shielding. the ab initio LORG method at the 6-31G* level. For a formamide surrounded by the four in-plane molecules of water in the first In this paper, the model by Gerothanassis and Vakka hydration shell, the calculated 17 O shift change due to the four (13,15) is reduced to a more quantitative basis by applicahydrogen bonds, 083.2 ppm, is smaller than the empirical hydration of the Kamlet-Abboud-Taft (KAT) parameters using tion shift, 0100 ppm. The 17 O shift change from each out-of-plane linear solvation-shift relationships (18-21). The particular water molecule hydrogen-bonded to the amide oxygen is 018.0 equations for each amide are embodied into a general equappm. These LORG results support the conclusion that no more tion for the four molecules which allow us to evaluate critithan four water molecules are hydrogen-bonded to the amide oxycally the BAF model (5, 9) as well as to determine reliable gen in formamide. ᭧ 1997 Academic Press values for the substituent effects. Ab initio calculations of 17 O hydration shifts have been carried out with the aim to gain information about the structure of aqueous formamide.