(218d) Enhancing Shotgun Proteomics Data Analysis with Bioinformatics Tools: Application to Cyanobacterial Proteomics

Shotgun proteomics, which uses multi-dimensional separations (usually via HPLC) of the complex mixture of peptides, has been shown to provide higher throughput and better results for hydrophobic and other proteins than is possible with two-dimensional electrophoresis (2DE). Although shotgun proteomics was originally based on orthogonal chromatography, recently the field has expanded by also employing some combinations of gel-based, liquid-phase IEF, and LC separation of proteins and peptides. Although the separation steps in shotgun proteomics have advantages over 2DE separation, there are also more challenges on the bioinformatics side, since peptides from the same protein are analyzed at separate times in the mass spectrometer. And in both approaches, there is a growing desire to obtain more information about the proteome. Thus, a goal of our work is to evaluate protein bioinformatics tools that enhance the data sets obtained by shotgun proteomics. To date, these tools include PSORTb for cellular location, the grand average hydropathy (GRAVY) index for hydrophobicity, LipoP for lipoproteins, and the exponentially modified protein abundance index (emAPI) for abundance.

We illustrate the use of these tools for mining shotgun proteomics data for the filamentous cyanobacterium Anabaena variabilis ATCC 29413, a model for N2 fixation. Cyanobacteria are photosynthetic prokaryotes notable for their ability to produce hydrogen and a variety of interesting secondary metabolites, with potential therapeutic application. As a result of the growing number of completed cyanobacterial genome projects, the development of post-genomics analysis for this important group of organisms has been accelerating. A. variabilis ATCC 29413 is a model cyanobacterium for N2 fixation. Using approaches for enhanced protein extraction, 646 proteins were identified by shotgun proteomics, which is more than double the previous results obtained using 2DE. The application of bioinformatics tools to the analysis of these data yielded additional important results, including confirmation that the improved extraction method and shotgun approach resulted in a significantly higher representation of basic and hydrophobic proteins. In addition, the use of the emPAI confirmed that the phycobilisome proteins are highly abundant. Additional insights, and reports on our efforts to improve data yield from mass spectra, will be presented.