|FENG, JIAN - JOHNS HOPKINS UNIVERSITY
|NAIMAN, DANIEL - JOHNS HOPKINS UNIVERSITY
Submitted to: American Society for Mass Spectrometry
Publication Type: Abstract Only
Publication Acceptance Date: 3/1/2006
Publication Date: 3/1/2006
Citation: Padliya, N.D., Garrett, W.M., Feng, J., Campbell, K., Naiman, D., Cooper, B. 2006. Towards the identification and detection of fungal plant pathogens via shotgun proteomics [abstract]. American Society for Mass Spectrometry.
Technical Abstract: Plant pathogenic fungi are responsible for billions of dollars of damage to economically important crops throughout the world. Thus, it is important to be able to detect fungal pathogens and diagnose disease to mitigate economic impacts. Many techniques are available for detection and diagnosis, including PCR and antibody-based methods. Unfortunately, these methods require development of pathogen-specific reagents prior to detection. We are investigating the use of mass spectrometry-based proteomics for detecting plant fungal pathogens since this method is reagent independent by comparison. Proteins were extracted from pure cultures of the fungal plant pathogens, Rhizoctonia solani, Fusarium graminearum, and Ustilago maydis. Digests of these protein mixtures were loaded on a self-packed C18/SCX fused-silica column. Each peptide mixture was eluted in a twelve-step process that included increasing concentrations of salt followed by an increasing gradient of mobile-phase at each step. Eluent was electrosprayed directly into the ESI source of the LCQ-Deca XP mass spectrometer. Automated peak recognition, dynamic exclusion, and MS/MS ion scanning of the three most intense parent ions was performed. MS/MS data was searched against the NCBInr database via Mascot. Mascot search results were parsed using Protein Panorama, our custom software that assembles a probability-based and parsimonious set of non-redundant proteins. We are interested in identifying MS/MS spectra of fungal plant pathogen peptides by searching against the NCBI non-redundant protein database (nr). The three fungal pathogens studied have varying degrees of representation in NCBI nr. Analysis of MS/MS data from U. maydis 521 via our Mascot and Protein Panorama pipeline led to the identification of 496 proteins at a 95% confidence level. It was found that 471/496 the proteins were comprised of peptide inferences from U. maydis. Analysis of MS/MS data from F. graminearum via our software pipeline led to the identification of 67 proteins at a 95% confidence level, of which 53 were comprised of peptide inferences from F. graminearum. On the other hand, analysis of MS/MS data from R. solani led to the identification of < 30 proteins. None of the inferences were derived from R. solani sequences; rather the inferences were made from sequences of other organisms such as Xylella fastidiosa, Neurospora crassa, and Caenorhabditis elegans. Since the genomes of U. maydis and F. graminearum are sequenced and well-represented in NCBInr, our data show that a high proportion of the peptides inferences were made from proteins for these fungal species. However, the genome information available for R. solani is very limited. Therefore, it is not surprising that this approach failed to identify any R. solani proteins and hence, the correct fungal plant pathogen itself. We have demonstrated that fungal plant pathogens that have a sequenced genome can be identified with a high-degree of confidence using spectral inferences. Since a majority of fungal plant pathogens are not sequenced, a mass spectrometry-based approach that relies on tandem mass spectra inferences will not be sufficient for disease diagnosis. It is possible, however, that an approach based on searching custom-made MS/MS libraries of fungal pathogen peptide biomarkers is much more viable.