Multi-Protein Complexes Studied by Mass Spectrometry

. Methods of biological mass spectrometry have improved at a dramatic pace ever since the discovery of electrospray ionization (ES) and Matrix assisted laser desorption ionization (MALDI). These methods now allow the relatively rapid identification of gel-separated proteins as well as the analysis of complex protein mixtures. Continuing efforts aim at increasing the sensitivity of mass spectrometric analysis which currently allows peptide sequencing at the low femtomole level, facilitating the sequencing of silver stained proteins and the detection of medium to low abundant proteins in complex mixtures. One of the goals of our groups has been to apply these methods towards the analysis of protein-protein interactions and the molecular elucidation of multiprotein complexes[1,2]. METHODS. Gel separated proteins were visualized, excised, and digested as previously described[3] and analyzed by nanoelectrospray sequencing as described[4]. For direct complex mixture analysis, peptides were enzymatically degraded and peptides were separated by HPLC and on-line infused into the mass spectrometer. Mass spectrometric analysis was on PE-Sciex QSTAR Pulsar. Protein databases, EST databases and the draft human genome sequence were searched with the MASCOTT program (Matrix Sciences) and the PepSea software suite (MDSProteomics). RESULTS. A large number of multiprotein complexes in mammalian cells have been analyzed by the mass spectrometric methods listed above. These complexes include the spliceosome, the centrosome, mitochondria, and many others. In depth analysis of the nucleolus by a layered approach involving MALDI mass spectrometry and directed sequencing with direct searching of the human draft genome yielded more than 260 proteins. Surprisingly, 30% of these proteins were novel or had no function attached to them (Andersen et al., in press). As in the case of the analysis of the human spliceosome, a number of the novel factors were tagged with the green or yellow fluorescent proteins to determine localization in vivo. Based on the results of this experiment, initial florescence and electron microscopy and the identity of the found proteins we estimate the purity of the preparation to be very high. Bioinformatic analysis of the found factors revealed no common targeting signal to the nucleolus. Subsequent analysis of the nucleolus by direct mass