Determinants of sweetness in proteins: a topological approach

Sweet taste in mammals is accounted for by a single receptor that shares homology with a metabotropic glutamate receptor. Most sweeteners are small molecular weight molecules that interact with small cavities in the so-called Venus Flytrap domains of the sweet receptor. The mechanism of action of larger molecules such as sweet proteins is, however, more difficult to interpret. The first and still the only general mechanism proposed for the action of sweet proteins, the "wedge model," hypothesizes that proteins bind to an external binding site of the active conformation of the sweet receptor. Here, I have extended the concept that inspired the wedge model using a combination of structural analysis, bioinformatics tools, and a relatively large dataset of mutations of the two most extensively studied sweet proteins, monellin and brazzein. I show here that it is possible to single out, among the ensemble yielded by low-resolution docking, a unique complex that satisfies simple topological constraints. These models of the complexes of monellin and brazzein are fully consistent with experimental evidence, thus providing predicting power for further validation of the wedge model.