Proteogenomic analysis of Mycobacterium tuberculosis by high resolution mass spectrometry

The genome sequencing of H37Rv strain of Mycobacterium tuberculosis was completed in 1998 followed by the whole genome sequencing of a clinical isolate, CDC1551 in 2002. Since then, the genomic sequences of a number of other strains have become available making it one of the better studied pathogenic bacterial species at the genomic level. However, annotation of its genome remains challenging because of high GC content and dissimilarity to other model prokaryotes. To this end, we carried out an in-depth proteogenomic analysis of the M. tuberculosis H37Rv strain using Fourier transform mass spectrometry with high resolution at both MS and tandem MS levels. In all, we identified 3176 proteins from Mycobacterium tuberculosis representing ~80% of its total predicted gene count. In addition to protein database search, we carried out a genome database search, which led to identification of ~250 novel peptides. Based on these novel genome search-specific peptides, we discovered 41 novel protein coding genes in the H37Rv genome. Using peptide evidence and alternative gene prediction tools, we also corrected 79 gene models. Finally, mass spectrometric data from N terminus-derived peptides confirmed 727 existing annotations for translational start sites while correcting those for 33 proteins. We report creation of a high confidence set of protein coding regions in Mycobacterium tuberculosis genome obtained by high resolution tandem mass-spectrometry at both precursor and fragment detection steps for the first time. This proteogenomic approach should be generally applicable to other organisms whose genomes have already been sequenced for obtaining a more accurate catalogue of protein-coding genes.     

Integration with the human genome of peptide sequences obtained by high-throughput mass spectrometry

A crucial aim upon the completion of the human genome is the verification and functional annotation of all predicted genes and their protein products. Here we describe the mapping of peptides derived from accurate interpretations of protein tandem mass spectrometry (MS) data to eukaryotic genomes and the generation of an expandable resource for integration of data from many diverse proteomics experiments. Furthermore, we demonstrate that peptide identifications obtained from high-throughput proteomics can be integrated on a large scale with the human genome. This resource could serve as an expandable repository for MS-derived proteome information.     

Enhanced homology searching through genome reading frame predetermination

MOTIVATION: Many bioinformatic approaches exist for finding novel genes within genomic sequence data. Traditionally, homology search-based methods are often the first approach employed in determining whether a novel gene exists that is similar to a known gene. Unfortunately, distantly related genes or motifs often are difficult to find using single query-based homology search algorithms against large sequence datasets such as the human genome. Therefore, the motivation behind this work was to develop an approach to enhance the sensitivity of traditional single query-based homology algorithms against genomic data without losing search selectivity. RESULTS: We demonstrate that by searching against a genome fragmented into all possible reading frames, the sensitivity of homology-based searches is enhanced without degrading its selectivity. Using the ETS-domain, bromodomain and acetyl-CoA acetyltransferase gene as queries, we were able to demonstrate that direct protein-protein searches using BLAST2P or FASTA3 against a human genome segmented among all possible reading frames and translated was substantially more sensitive than traditional protein-DNA searches against a raw genomic sequence using an application such as TBLAST2N. Receiver operating characteristic analysis was employed to demonstrate that the algorithms remained selective, while comparisons of the algorithms showed that the protein-protein searches were more sensitive in identifying hits. Therefore, through the overprediction of reading frames by this method and the increased sensitivity of protein-protein based homology search algorithms, a genome can be deeply mined, potentially finding hits overlooked by protein-DNA searches against raw genomic data.     

Integrated approach for manual evaluation of peptides identified by searching protein sequence databases with tandem mass spectra

Quantitative proteomics relies on accurate protein identification, which often is carried out by automated searching of a sequence database with tandem mass spectra of peptides. When these spectra contain limited information, automated searches may lead to incorrect peptide identifications. It is therefore necessary to validate the identifications by careful manual inspection of the mass spectra. Not only is this task time-consuming, but the reliability of the validation varies with the experience of the analyst. Here, we report a systematic approach to evaluating peptide identifications made by automated search algorithms. The method is based on the principle that the candidate peptide sequence should adequately explain the observed fragment ions. Also, the mass errors of neighboring fragments should be similar. To evaluate our method, we studied tandem mass spectra obtained from tryptic digests of E. coli and HeLa cells. Candidate peptides were identified with the automated search engine Mascot and subjected to the manual validation method. The method found correct peptide identifications that were given low Mascot scores (e.g., 20-25) and incorrect peptide identifications that were given high Mascot scores (e.g., 40-50). The method comprehensively detected false results from searches designed to produce incorrect identifications. Comparison of the tandem mass spectra of synthetic candidate peptides to the spectra obtained from the complex peptide mixtures confirmed the accuracy of the evaluation method. Thus, the evaluation approach described here could help boost the accuracy of protein identification, increase number of peptides identified, and provide a step toward developing a more accurate next-generation algorithm for protein identification.     

Proteomic profiling and protein identification by MALDI-TOF mass spectrometry in unsequenced parasitic nematodes

Lack of genomic sequence data and the relatively high cost of tandem mass spectrometry have hampered proteomic investigations into helminths, such as resolving the mechanism underpinning globally reported anthelmintic resistance. Whilst detailed mechanisms of resistance remain unknown for the majority of drug-parasite interactions, gene mutations and changes in gene and protein expression are proposed key aspects of resistance. Comparative proteomic analysis of drug-resistant and -susceptible nematodes may reveal protein profiles reflecting drug-related phenotypes. Using the gastro-intestinal nematode, Haemonchus contortus as case study, we report the application of freely available expressed sequence tag (EST) datasets to support proteomic studies in unsequenced nematodes. EST datasets were translated to theoretical protein sequences to generate a searchable database. In conjunction with matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF-MS), Peptide Mass Fingerprint (PMF) searching of databases enabled a cost-effective protein identification strategy. The effectiveness of this approach was verified in comparison with MS/MS de novo sequencing with searching of the same EST protein database and subsequent searches of the NCBInr protein database using the Basic Local Alignment Search Tool (BLAST) to provide protein annotation. Of 100 proteins from 2-DE gel spots, 62 were identified by MALDI-TOF-MS and PMF searching of the EST database. Twenty randomly selected spots were analysed by electrospray MS/MS and MASCOT Ion Searches of the same database. The resulting sequences were subjected to BLAST searches of the NCBI protein database to provide annotation of the proteins and confirm concordance in protein identity from both approaches. Further confirmation of protein identifications from the MS/MS data were obtained by de novo sequencing of peptides, followed by FASTS algorithm searches of the EST putative protein database. This study demonstrates the cost-effective use of available EST databases and inexpensive, accessible MALDI-TOF MS in conjunction with PMF for reliable protein identification in unsequenced organisms.     

Validating divergent ORF annotation of the Mycobacterium leprae genome through a full translation data set and peptide identification by tandem mass spectrometry

Mycobacterium leprae has undergone extensive degenerative evolution, with a large number of pseudogenes. It is also the organism with the greatest divergence between gene annotations from independent institutes. Therefore, M. leprae is a good model to verify the currently predicted coding sequence regions between different annotations, to identify new ones and to investigate the expression of pseudogenes. We submitted a total extract of the bacteria isolated from Armadillo to Gel‐LC‐MS/MS using a linear quadrupole ion trap‐Orbitrap mass spectrometer. Spectra were analyzed using the Leproma (1614 genes and 1133 pseudogenes) and TIGR (5446 genes) databases and a database containing the full genome translation. We identified a total of 1046 proteins, including five proteins encoded by previously predicted pseudogenes, which upon closer inspection appeared to be proper genes. Only 11 of the additional annotations by TIGR were verified. We also identified six tryptic peptides from five proteins from regions not considered to be coding sequences, in addition to peptides from two unannotated gene candidates that overlap with other genes. Our data show that the Leproma annotation of M. leprae is quite accurate, and there were no peptide observations corresponding to true pseudogenes, except for a new gene candidate, overlapping with an essential enolase on the complementary strand.     

Mass spectrometry at the interface of proteomics and genomics

With the onset of modern DNA sequencing technologies, genomics is experiencing a revolution in terms of quantity and quality of sequencing data. Rapidly growing numbers of sequenced genomes and metagenomes present a tremendous challenge for bioinformatics tools that predict protein-coding regions. Experimental evidence of expressed genomic regions, both at the RNA and protein level, is becoming invaluable for genome annotation and training of gene prediction algorithms. Evidence of gene expression at the protein level using mass spectrometry-based proteomics is increasingly used in refinement of raw genome sequencing data. In a typical "proteogenomics" experiment, the whole proteome of an organism is extracted, digested into peptides and measured by a mass spectrometer. The peptide fragmentation spectra are identified by searching against a six-frame translation of the raw genomic assembly, thus enabling the identification of hitherto unpredicted protein-coding genomic regions. Application of mass spectrometry to genome annotation presents a range of challenges to the standard workflows in proteomics, especially in terms of proteome coverage and database search strategies. Here we provide an overview of the field and argue that the latest mass spectrometry technologies that enable high mass accuracy at high acquisition rates will prove to be especially well suited for proteogenomics applications.     

Reannotation of Shewanella oneidensis genome

As more and more complete bacterial genome sequences become available, the genome annotation of previously sequenced genomes may become quickly outdated. This is primarily due to the discovery and functional characterization of new genes. We have reannotated the recently published genome of Shewanella oneidensis with the following results: 51 new genes have been identified, and functional annotation has been added to the 97 genes, including 15 new and 82 existing ones with previously unassigned function. The identification of new genes was achieved by predicting the protein coding regions using the HMM-based program GeneMark.hmm. Subsequent comparison of the predicted gene products to the non-redundant protein database using BLAST and the COG (Clusters of Orthologous Groups) database using COGNITOR provided for the functional annotation.     

Concerted action of the new Genomic Peptide Finder and AUGUSTUS allows for automated proteogenomic annotation of the Chlamydomonas reinhardtii genome

The use and development of post-genomic tools naturally depends on large-scale genome sequencing projects. The usefulness of post-genomic applications is dependent on the accuracy of genome annotations, for which the correct identification of intron-exon borders in complex genomes of eukaryotic organisms is often an error-prone task. Although automated algorithms for predicting intron-exon structures are available, supporting exon evidence is necessary to achieve comprehensive genome annotation. Besides cDNA and EST support, peptides identified via MS/MS can be used as extrinsic evidence in a proteogenomic approach. We describe an improved version of the Genomic Peptide Finder (GPF), which aligns de novo predicted amino acid sequences to the genomic DNA sequence of an organism while correcting for peptide sequencing errors and accounting for the possibility of splicing. We have coupled GPF and the gene finding program AUGUSTUS in a way that provides automatic structural annotations of the Chlamydomonas reinhardtii genome, using highly unbiased GPF evidence. A comparison of the AUGUSTUS gene set incorporating GPF evidence to the standard JGI FM4 (Filtered Models 4) gene set reveals 932 GPF peptides that are not contained in the Filtered Models 4 gene set. Furthermore, the GPF evidence improved the AUGUSTUS gene models by altering 65 gene models and adding three previously unidentified genes.     

Proteomic Detection of Non-Annotated Protein-Coding Genes in Pseudomonas fluorescens Pf0-1

Genome sequences are annotated by computational prediction of coding sequences, followed by similarity searches such as BLAST, which provide a layer of possible functional information. While the existence of processes such as alternative splicing complicates matters for eukaryote genomes, the view of bacterial genomes as a linear series of closely spaced genes leads to the assumption that computational annotations that predict such arrangements completely describe the coding capacity of bacterial genomes. We undertook a proteomic study to identify proteins expressed by Pseudomonas fluorescens Pf0-1 from genes that were not predicted during the genome annotation. Mapping peptides to the Pf0-1 genome sequence identified sixteen non-annotated protein-coding regions, of which nine were antisense to predicted genes, six were intergenic, and one read in the same direction as an annotated gene but in a different frame. The expression of all but one of the newly discovered genes was verified by RT-PCR. Few clues as to the function of the new genes were gleaned from informatic analyses, but potential orthologs in other Pseudomonas genomes were identified for eight of the new genes. The 16 newly identified genes improve the quality of the Pf0-1 genome annotation, and the detection of antisense protein-coding genes indicates the under-appreciated complexity of bacterial genome organization.     

Robust method for proteome analysis by MS/MS using an entire translated genome: demonstration on the ciliome of Tetrahymena thermophila

To improve the utility of increasingly large numbers of available unannotated and initially poorly annotated genomic sequences for proteome analysis, we demonstrate that effective protein identification can be made on a large and unannotated genome. The strategy developed is to translate the unannotated genome sequence into amino acid sequence encoding putative proteins in all six reading frames, to identify peptides by tandem mass spectrometry (MS/MS), to localize them on the genome sequence, and to preliminarily annotate the protein via a similarity search by BLAST. These tasks have been optimized and automated. Optimization to obtain multiple peptide matches in effect extends the searchable region and results in more robust protein identification. The viability of this strategy is demonstrated with the identification of 223 cilia proteins in the unicellular eukaryotic model organism Tetrahymena thermophila, whose initial genomic sequence draft was released in November 2003. To the best of our knowledge, this is the first demonstration of large-scale protein identification based on such a large, unannotated genome. Of the 223 cilia proteins, 84 have no similarity to proteins in NCBI's nonredundant (nr) database. This methodology allows identifying the locations of the genes encoding these novel proteins, which is a necessary first step to downstream functional genomic experimentation.     

Expressed peptide tags: an additional layer of data for genome annotation

While genome sequencing is becoming ever more routine, genome annotation remains a challenging process. Identification of the coding sequences within the genomic milieu presents a tremendous challenge, especially for eukaryotes with their complex gene architectures. Here, we present a method to assist the annotation process through the use of proteomic data and bioinformatics. Mass spectra of digested protein preparations of the organism of interest were acquired and searched against a protein database created by a six-frame translation of the genome. The identified peptides were mapped back to the genome, compared to the current annotation, and then categorized as supporting or extending the current genome annotation. We named the classified peptides Expressed Peptide Tags (EPTs). The well-annotated bacterium Rhodopseudomonas palustris was used as a control for the method and showed a high degree of correlation between EPT mapping and the current annotation, with 86% of the EPTs confirming existing gene calls and less than 1% of the EPTs expanding on the current annotation. The eukaryotic plant pathogens Phytophthora ramorum and Phytophthora sojae, whose genomes have been recently sequenced and are much less well-annotated, were also subjected to this method. A series of algorithmic steps were taken to increase the confidence of EPT identification for these organisms, including generation of smaller subdatabases to be searched against, and definition of EPT criteria that accommodates the more complex eukaryotic gene architecture. As expected, the analysis of the Phytophthora species showed less correlation between EPT mapping and their current annotation. While approximately 76% of Phytophthora EPTs supported the current annotation, a portion of them (7.7% and 12.9% for P. ramorum and P. sojae, respectively) suggested modification to current gene calls or identified novel genes that were missed by the current genome annotation of these organisms.     

Mass spectrometry-based prokaryote gene annotation

MS combined with database searching has become the preferred method for identifying proteins present in cell or tissue samples. The technique enables us to execute large-scale proteome analyses of species whose genomes have already been sequenced. Searching mass spectrometric data against protein databases composed of annotated genes has been widely conducted. However, there are some issues with this technique; wrong annotations in protein databases cause deterioration in the accuracy of protein identification, and only proteins that have already been annotated can be identified. We propose a new framework that can detect correct ORFs by integrating an MS/MS proteomic data mapping and a knowledge-based system regarding the translation initiation sites. This technique can provide correction of predicted coding sequences, together with the possibility of identifying novel genes. We have developed a computational system; it should first conduct the probabilistic peptide-matching against all possible translational frames using MS/MS data, then search for discriminative DNA patterns around the detected peptides, and lastly integrate the facts using empirical knowledge stored in knowledge bases to obtain correct ORFs. We used photosynthetic bacteria Synechocystis sp. PCC6803 as a sample prokaryote, resulting in the finding of 14 N-terminus annotation errors and several new candidate genes.     

DAtA: database of Arabidopsis thaliana annotation

The Database of Arabidopsis thaliana Annotation (D At A) was created to enable easy access to and analysis of all the Arabidopsis genome project annotation. The database was constructed using the completed A.thaliana genomic sequence data currently in GenBank. An automated annotation process was used to predict coding sequences for GenBank records that do not include annotation. D At A also contains protein motifs and protein similarities derived from searches of the proteins in D At A with motif databases and the non-redundant protein database. The database is routinely updated to include new GenBank submissions for Arabidopsis genomic sequences and new Blast and protein motif search results. A web interface to D At A allows coding sequences to be searched by name, comment, blast similarity or motif field. In addition, browse options present lists of either all the protein names or identified motifs present in the sequenced A.thaliana genome. The database can be accessed at http://baggage. stanford.edu/group/arabprotein/     

Strategic proteome analysis of Candida magnoliae with an unsequenced genome

Erythritol is a noncariogenic, low calorie sweetener. It is safe for people with diabetes and obese people. Candida magnoliae is an industrially important organism because of its ability to produce erythritol as a major product. The genome of C. magnoliae has not been sequenced yet, limiting the available proteome database. Therefore, systematic approaches were employed to construct the proteome map of C. magnoliae. Proteomic analysis with systematic approaches is based on two-dimensional electrophoresis, matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS), tandem mass spectrometry (MS/MS) and database interrogation. First, 24 spots were analyzed using peptide mass fingerprinting along with MALDI-TOF MS with high mass accuracy. Only four spots were reliably identified as carbonyl reductase and its isoforms. The reason for low sequence coverage seemed to be that these identification strategies were based on the presence of the protein database obtained from the publicly accessible genome database and the availability of cross-species protein identification. MS/MS (MS/MS ion search and de novo sequencing) in combination with similarity searches allowed successful identification of 39 spots. Several proteins including transaldolase identified by MS/MS ion searches were further confirmed by partial sequences from the expressed sequence tag database. In this study, 51 protein spots were analyzed and then potentially identified. The identified proteins were involved in glycolysis, stress response, other essential metabolisms and cell structures.