The mammalian sense of taste involves the analysis of complex mixtures of analytes present in food and beverages. The mechanism of taste involves a metabolomic fingerprint of the solution, and in many cases the sense of taste can distinguish subtly different mixtures, as well as differentiate minor differences in chemical structures. Most tastants, and odorants as well, are identified through a composite of responses from non-specific interactions. The pattern created by the simultaneous response of a series of differential receptors is specific for a particular set of stimuli. For example, wine tasters are able to distinguish very subtle taste characteristics as a key to the success of their careers. Our plan is to use wine as a test-bed solution to optimize the art of differential sensing techniques. The reason for our choice of wine is its complexity and the very unique chemical structures. Further, wine is an excellent choice because human test panels are available for comparison to our artificial approach. We propose to target several classes of analytes present in wine: carboxylic acids, sugars, and tannins (and analogs). These classes present interesting challenges for molecular recognition and fingerprinting. We will use "in-hand" receptors for carboxylates and sugars. Alternatively, the receptors for the tannins will be derived from combinatorial chemistry. Further, we plan to create specific receptors for those components of wine known to have health advantages: resveratrol and quercetin. To accomplish our goals, a collaboration has been established between Dr. Eric Anslyn at the University of Texas at Austin and Dr. Hildegarde Heymann at the University of California Davis. Dr. Anslyn will create the complex organic receptors, the signaling protocols, and apply the appropriate pattern recognition protocols. Dr. Heymann will use more standard approaches to metabolomic profiling, and will oversee trained human test panels. The data collected at U.T. and U.C.D. will be used to answer several questions as described herein, a few of which are: 1) How many receptors, and what structures, are needed to accomplish the fingerprinting of our analyte classes? 2) What correlations will we find between fingerprints of the analyte classes created at U.T. and at U.C.D., 3) Will specific chemicals dominate the fingerprints and test panel responses at U.C.D., 4) Will specific chemicals found important to the U.C.D. fingerprints be evident in the U.T. fingerprints, even if we did not train on those specific chemicals, 5) Which chemical fingerprints found at either U.C.D. or U.T. will best reflect the sensory panel input, and 6) Can fingerprints found at U.T. be used in a predictive manner for sensory panel response. The broad impact of this proposal resides in exploring a general approach that can have impact on the fields of medical, environmental, defense, and food diagnostics. We feel that the power of an array of synthetic receptors, when coupled with pattern recognition protocols, cannot be surpassed for array sensor applications. In practice, synthetic receptors suffer interference from similar analytes due to their simplicity. Therefore, synthetic receptors are naturally cross-reactive, the exact attribute that is desired in an array setting. We predict that this attribute of synthetic receptors will allow chemists to use them to analyze solutions for which the components are not exactly known. Furthermore, the use of synthetic combinatorial chemistry in the creation of unnatural receptors naturally compliments this requirement of cross-reactivity. The "big-picture" goal of this proposal is to teach this lesson to supramolecular and analytical chemists, while the choice of wine to prove the techniques leads to a general societal interest and dissemination of the results in non-traditional academic media.
