Identification of proteins from non-model organisms using mass spectrometry: application to a hibernating mammal

A major challenge in the life sciences is the extraction of detailed molecular information from plants and animals that are not among the handful of exhaustively studied "model organisms." As a consequence, certain species with novel phenotypes are often ignored due to the lack of searchable databases, tractable genetics, stock centers, and more recently, a sequenced genome. Characterization of phenotype at the molecular level commonly relies on the identification of differentially expressed proteins by combining database searching with tandem mass spectrometry (MS) of peptides derived from protein fragmentation. However, the identification of short peptides from nonmodel organisms can be hampered by the lack of sufficient amino acid sequence homology with proteins in existing databases; therefore, a database search strategy that encompasses both identity and homology can provide stronger evidence than a single search alone. The use of multiple algorithms for database searches may also increase the probability of correct protein identification since it is unlikely that each program would produce false negative or positive hits for the same peptides. In this study, four software packages, Mascot, Pro ID, Sequest, and Pro BLAST, were compared in their ability to identify proteins from the thirteen-lined ground squirrel (Spermophilus tridecemlineatus), a hibernating mammal that lacks a completely sequenced genome. Our results show similarities as well as the degree of variability among different software packages when the identical protein database is searched. In the process of this study, we identified the up-regulation of succinyl CoA-transferase (SCOT) in the heart of hibernators. SCOT is the rate-limiting enzyme in the catabolism of ketone bodies, an important alternative fuel source during hibernation.