The mechanism of interaction of sweet proteins with the T1R2-T1R3 sweet taste receptor has not yet been elucidated. Low molecular mass sweeteners and sweet proteins interact with the same receptor, the human T1R2-T1R3 receptor. The presence on the surface of the proteins of 'sweet fingers', i.e. protruding features with chemical groups similar to those of low molecular mass sweeteners that can probe the active site of the receptor, would be consistent with a single mechanism for the two classes of compounds. We have synthesized three cyclic peptides corresponding to the best potential 'sweet fingers' of brazzein, monellin and thaumatin, the sweet proteins whose structures are well characterized. NMR data show that all three peptides have a clear tendency, in aqueous solution, to assume hairpin conformations consistent with the conformation of the same sequences in the parent proteins. The peptide corresponding to the only possible loop of brazzein, c ], exists in solution in a well ordered hairpin conformation very similar to that of the same sequence in the parent protein. However, none of the peptides has a sweet taste. This finding strongly suggests that sweet proteins recognize a binding site different from the one that binds small molecular mass sweeteners. The data of the present work support an alternative mechanism of interaction, the 'wedge model', recently proposed for sweet proteinsKeywords: models; NMR; sweet proteins; sweeteners; taste receptors.Sweeteners are widely used by people affected by diseases linked to the consumption of carbohydrates, such as diabetes and obesity. Although sugar substitutes are now reasonably safe, there is a continuous search for new, safer ones. Their design requires a good understanding of the interaction of sweet molecules with their receptor. Ideally, one would like to determine the structure of sweet molecules inside the receptor but, although sweet taste receptors have been recently cloned and expressed [1,2], direct structural studies of membrane proteins are very difficult. In order to study the structureactivity relationship of sweet molecules it is possible to recur to indirect investigations. The great majority of sweet molecules are small molecular mass compounds [3] but a few sweet proteins have also been identified [4]. Low molecular mass sweet compounds and sweet proteins interact with the same receptor, the human T1R2-T1R3 receptor [5]; accordingly, it seems natural to look for a similar mechanism when trying to explain the taste of these proteins.Early models of the active site of the sweet receptor, proposed before the nature of the receptor was elucidated, were derived from the shape of conformationally rigid sweeteners, used as molecular molds [6][7][8][9][10][11][12]. Different models differ in details of the shape of the active site but there is a consensus on the presence, in the sweetener, of two groups involved in hydrogen bonding with the receptor, the so-called AH-B entity [6], and of an apolar group at a precise distance with respect to the AH...