Current status of the SWISS-2DPAGE database

Management and dissemination of MS proteomic data with PROTICdb: example of a quantitative comparison between methods of protein extraction

High throughput MS-based proteomic experiments generate large volumes of complex data and necessitate bioinformatics tools to facilitate their handling. Needs include means to archive data, to disseminate them to the scientific communities, and to organize and annotate them to facilitate their interpretation. We present here an evolution of PROTICdb, a database software that now handles MS data, including quantification. PROTICdb has been developed to be as independent as possible from tools used to produce the data. Biological samples and proteomics data are described using ontology terms. A Taverna workflow is embedded, thus permitting to automatically retrieve information related to identified proteins by querying external databases. Stored data can be displayed graphically and a "Query Builder" allows users to make sophisticated queries without knowledge on the underlying database structure. All resources can be accessed programmatically using a Java client API or RESTful web services, allowing the integration of PROTICdb in any portal. An example of application is presented, where proteins extracted from a maize leaf sample by four different methods were compared using a label-free shotgun method. Data are available at http://moulon.inra.fr/protic/public. PROTICdb thus provides means for data storage, enrichment, and dissemination of proteomics data.

Comparison of in-gel protein separation techniques commonly used for fractionation in mass spectrometry-based proteomic profiling

Fractionation of complex samples at the cellular, subcellular, protein, or peptide level is an indispensable strategy to improve the sensitivity in mass spectrometry-based proteomic profiling. This study revisits, evaluates, and compares the most common gel-based protein separation techniques i.e. 1D SDS-PAGE, 1D preparative SDS-PAGE, IEF-IPG, and 2D-PAGE in their performance as fractionation approaches in nano LC-ESI-MS/MS analysis of a mixture of protein standards and mitochondrial extracts isolated from rat liver. This work demonstrates that all the above techniques provide complementary protein identification results, but 1D SDS-PAGE and IEF-IPG had the highest number of identifications. The IEF-IPG technique resulted in the highest average number of detected peptides per protein. The 2D-PAGE was evaluated as a protein fractionation approach. This work shows that the recovery of proteins and resulting proteolytic digests is highly dependent on the total volume of the gel matrix. The performed comparison of the fractionation techniques demonstrates the potential of a combination of orthogonal 1D SDS-PAGE and IEF-IPG for the improved sensitivity of profiling without significant decrease in throughput.

Combined proteomic approach with SELDI-TOF-MS and peptide mass fingerprinting identified the rapid increase of monomeric transthyretin in rat cerebrospinal fluid after transient focal cerebral ischemia

Several proteins are known to be markedly expressed in the brain during cerebral ischemia, however the change in protein profiles within the cerebrospinal fluid (CSF) after an ischemic insult has not been fully elucidated. We studied the changes in the CSF proteome in rat transient middle cerebral artery occlusion model. Surface enhanced laser desorption/ionization time-of-flight (SELDI-TOF) mass spectrometry (MS) was used to detect the time-course changes in CSF protein patterns after transient focal brain ischemia. According to hierarchical cluster analysis by self-organising tree algorism (SOTA), the temporal pattern of protein peaks was divided into four groups: acute increase group, chronic increase group, gradual decrease group and unchanged group. In the acute increase group, the expression of a 13.6-kDa protein was markedly increased during the acute phase. The 13.6-kDa protein was identified as monomeric form of transthyretin using two-dimensional electrophoresis and peptide mass fingerprinting based on matrix-assisted laser desorption/ionization-time of flight mass spectrometry. The monomeric transthyretin may represent an ischemia-specific CSF marker to indicate the sequential changes according to ischemic insults of the brain.

Elevation of CRABP-I in the cerebrospinal fluid of patients with Moyamoya disease

BACKGROUND AND PURPOSE

The etiology of moyamoya disease (MMD) remains obscure. This study was undertaken to identify specific proteins associated with the pathogenesis of MMD.

METHODS

We studied cerebrospinal fluid (CSF) from 20 patients with angiographically confirmed MMD (4 boys and 16 girls; age range, 3 to 13 years; mean, 7.5 years) and 4 control patients with cerebral palsy who underwent selective dorsal rhizotomy (2 boys and 2 girls; age range, 5 to 10 years; mean, 7.3 years). CSF proteins were analyzed by 2-dimensional polyacrylamide gel electrophoresis, and protein identification was performed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. The presence of specific CSF protein in patients with MMD was confirmed by Western blotting. In addition, cerebral CSF was also tested in 7 patients who had other brain diseases but no MMD (2 boys and 5 girls; age range, 1 to 12 years; mean, 6.9 years).

RESULTS

We identified 1 polypeptide spot (Mr of 13 to 15 kDa and isoelectric point of 5 to 5.5) that was differentially expressed in the CSF samples of MMD patients (mean optical density intensity, 0.36+/-0.24; range, 0.05 to 0.92) and control spinal CSF samples (mean, 0.03+/-0.04; range, 0 to 0.08; P=0.002). This polypeptide was identified as cellular retinoic acid-binding protein (CRABP)-I. High levels of expression of CRABP-I in the CSF from 17 MMD children were confirmed by Western blotting.

CONCLUSIONS

The analysis of the CSF of MMD patients reveals high CRABP-I expression. The present study suggests that the elevation of CRABP-I in CSF may be a candidate for pathogenesis of MMD.

Whole Proteome pI Values Correlate with Subcellular Localizations of Proteins for Organisms within the Three Domains of Life

Isoelectric point (pI) values have long been a standard measure for distinguishing between proteins. This article analyzes distributions of pI values estimated computationally for all predicted ORFs in a selection of fully sequenced genomes. Histograms of pI values confirm the bimodality that has been observed previously for bacterial and archaeal genomes (Van Bogelen et al. 1999) and reveal a trimodality in eukaryotic genomes. A similar analysis on subsets of a nonredundant protein sequence database generated from the full database by selecting on subcellular localization shows that sequences annotated as corresponding to cytosolic and integral membrane proteins have pI distributions that appear to correspond with the two observed modes of bacteria and archaea. Furthermore, nuclear proteins have a broader distribution that may account for the third mode observed in eukaryotes. On the basis of this association between pI and subcellular localization, we conclude that the bimodal character of whole proteome pI values in bacteria and archaea and the trimodal character in eukaryotes are likely to be general properties of proteomes and are associated with the need for different pI values depending on subcellular localization. Our analyses also suggest that the proportions of proteomes consisting of membrane-associated proteins may be currently underestimated.

Cadmium-binding to transferrin in the plasma of the common carp Cyprinus carpio

In this study, it was investigated by autoradiography with radioactive cadmium after Western blotting of two-dimensional electrophoresis gels, to which proteins cadmium is mainly bound in plasma of common carp Cyprinus carpio. The obtained results demonstrate that in carp plasma, cadmium is primarily bound to two high molecular weight proteins. Relative small amounts are bound to a protein with M(r) approximately 60000. The other metal-binding protein, with M(r) approximately 70000 and pI approximately 6.7 was identified as transferrin. The conditional equilibrium constants for the binding of cadmium ions to the two metal-binding sites of this protein were calculated as logK(1)=5.40+/-0.12 and logK(2)=4.66+/-0.21, which are comparable to those of human transferrin under the same experimental conditions. Transport of cadmium in plasma of carp was found to be different from that of brown trout Salmo trutta and man, where cadmium is mainly bound to albumin and transferrin. The prominent binding of cadmium to transferrin can be explained by the absence or at least the very low concentrations in which albumin is present in carp plasma.

Bioinformatics in protein analysis

The chapter gives an overview of bioinformatic techniques of importance in protein analysis. These include database searches, sequence comparisons and structural predictions. Links to useful World Wide Web (WWW) pages are given in relation to each topic. Databases with biological information are reviewed with emphasis on databases for nucleotide sequences (EMBL, GenBank, DDBJ), genomes, amino acid sequences (Swissprot, PIR, TrEMBL, GenePept), and three-dimensional structures (PDB). Integrated user interfaces for databases (SRS and Entrez) are described. An introduction to databases of sequence patterns and protein families is also given (Prosite, Pfam, Blocks). Furthermore, the chapter describes the widespread methods for sequence comparisons, FASTA and BLAST, and the corresponding WWW services. The techniques involving multiple sequence alignments are also reviewed: alignment creation with the Clustal programs, phylogenetic tree calculation with the Clustal or Phylip packages and tree display using Drawtree, njplot or phylo_win. Finally, the chapter also treats the issue of structural prediction. Different methods for secondary structure predictions are described (Chou-Fasman, Garnier-Osguthorpe-Robson, Predator, PHD). Techniques for predicting membrane proteins, antigenic sites and postranslational modifications are also reviewed.

Identification of two-dimensionally separated human cerebrospinal fluid proteins by N-terminal sequencing, matrix-assisted laser desorption/ionization--mass spectrometry, nanoliquid chromatography-electrospray ionization-time of flight-mass spectrometry, and tandem mass spectrometry

Optimal application of biological mass spectrometry (MS) in combination with two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) of human cerebrospinal fluid (CSF) can lead to the identification of new potential biological markers of neurological disorders. To this end, we analyzed a number of 2-D PAGE protein spots in a human CSF pool using spot co-localization, N-terminal sequencing, matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) and nanoliquid chromatography-electrospray ionization-time of flight-mass spectrometry (nanoLC-ESI-TOF-MS) with tandem MS switching. Our constructed CSF master contained 469 spots after image analysis and processing of 2-D gels. Upon visual inspection of our CSF master with the CSF pattern available on the ExPASy server, it was possible to locate and annotate 15 proteins. N-terminal sequence analysis and MALDI-MS peptide mass fingerprint analysis of both silver- and Coomassie Brilliant Blue (CBB) G-250-stained protein spots after in situ trypsin digest not only confirmed nine of the visually annotated spots but additionally resolved the identity of another 13 spots. Six of these proteins were not annotated on the 2-D ExPASy map: complement C3 alpha-chain (1321-1663), complement factor B, cystatin C, calgranulin A, hemoglobin beta-chain, and beta-2-microglobulin. It was clear that MALDI-MS identification from CBB G-250-stained, rather than from silver-stained, spots was more successful. In cases where no N-terminal sequence and/or no clear MALDI-MS result was available, nanoLC-ESI-TOF-MS and tandem MS automated switching was used to clarify and/or identify these protein spots by generating amino acid sequence tags. In addition, enrichment of the concentration of low-abundant proteins on 2-D PAGE was obtained by removal of albumin and immunoglobulins from the CSF pool using affinity chromatography. Subsequent analysis by 2-D PAGE of the fractionated CSF pool showed various new silver-stainable protein spots, of which four were identified by nanoLC-ESI-TOF-MS and tandem MS switching. No significant homology was found in either protein or DNA databases, indicating than these spots were unknown proteins.

An assessment of protein profiles from the marine oligotrophic ultramicrobacterium, Sphingomonas sp. strain RB2256

The protein expression profile of a novel marine oligotrophic ultramicrobacterium, Sphingomonas sp. strain RB2256, was investigated by two-dimensional polyacrylamide gel electrophoresis (2-D PAGE). Analytical reference maps were generated from mid-log phase batches and steady-state chemostat cultures with pH 4-8 immobilised pH gradients (IPGs) followed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis. The resolved proteins were detected by two different methods: radioactive labeling and silver staining. Protein profiles generated from analytical 2-D PAGE gels were compared and differential analysis was performed using Melanie II software. Both methods (radioactive labeling and silver staining) resulted in reproducible, high resolution gels (up to 1600 protein spots). This approach is proving to be a powerful tool for investigating the molecular basis of the unique physiology of this model oligotrophic microorganism.

A cerebrospinal fluid protein associated with moyamoya disease: report of three cases

OBJECTIVE

The pathogenesis of moyamoya disease is unknown. The purpose of this study was to detect proteins associated with the pathogenesis of moyamoya disease.

CLINICAL PRESENTATION

Cerebrospinal fluid (CSF) samples from three patients with moyamoya disease and four control patients who had cervical lesions but no intracranial lesion were studied.

INTERVENTION

CSF proteins separated by two-dimensional polyacrylamide gel electrophoresis were analyzed with the SWISS-2DPAGE and SWISS-PROT databases. In the CSF samples from all three patients with moyamoya disease, a polypeptide spot (Mr = 12,000, pI = 5.35) was observed. This spot was not evident in samples from the four control patients and has not been reported in the SWISS-2DPAGE and SWISS-PROT databases.

CONCLUSION

A CSF protein, which is possibly novel and associated with moyamoya disease, has been detected. The analysis of CSF by two-dimensional polyacrylamide gel electrophoresis may reveal a clue by which the molecular mechanism of moyamoya disease may be elucidated.

Composition of the peptide fraction in human blood plasma: database of circulating human peptides

A database was established from human hemofiltrate (HF) that consisted of a mass database and a sequence database, with the aim of analyzing the composition of the peptide fraction in human blood. To establish a mass database, all 480 fractions of a peptide bank generated from HF were analyzed by MALDI-TOF mass spectrometry. Using this method, over 20000 molecular masses representing native, circulating peptides were detected. Estimation of repeatedly detected masses suggests that approximately 5000 different peptides were recorded. More than 95% of the detected masses are smaller than 15000, indicating that HF predominantly contains peptides. The sequence database contains over 340 entries from 75 different protein and peptide precursors. 55% of the entries are fragments from plasma proteins (fibrinogen A 13%, albumin 10%, beta2-microglobulin 8.5%, cystatin C 7%, and fibrinogen B 6%). Seven percent of the entries represent peptide hormones, growth factors and cytokines. Thirty-three percent belong to protein families such as complement factors, enzymes, enzyme inhibitors and transport proteins. Five percent represent novel peptides of which some show homology to known peptide and protein families. The coexistence of processed peptide fragments, biologically active peptides and peptide precursors suggests that HF reflects the peptide composition of plasma. Interestingly, protein modules such as EGF domains (meprin Aalpha-fragments), somatomedin-B domains (vitronectin fragments), thyroglobulin domains (insulin like growth factor-binding proteins), and Kazal-type inhibitor domains were identified. Alignment of sequenced fragments to their precursor proteins and the analysis of their cleavage sites revealed that there are different processing pathways of plasma proteins in vivo.

A computer-assisted two-dimensional gel electrophoresis approach for studying the variations in protein expression related to an induced functional repression of NFkappaB in lymphoblastoid cell lines

Strategies are needed for conclusive interpretation of two-dimensional gel electrophoresis (2-D PAGE) maps in order to identify pertinent differences in protein expression during regulation of the transcription of discrete sets of genes. The model used in this study was a human lymphoblastoid cell line in which a functional repression of the transcription factors NFkappaB was obtained by induction of overexpression of IkappaBalpha, a physiological inhibitor of NFkappaB. The analytical methodology used relies on the comparison of two sets of 2-D PAGE maps for detecting differences in protein expression between samples overexpressing or not overexpressing IkappaBalpha. The analysis was based on a combination of an automatic computerized analysis, constituting an actual aid for deciding, and of an interactive visual validation, corresponding to the interpretation of computer propositions. This strategy is proposed as a rapid way to detect potential variations in protein expression applicable to any biological model. In this study, correspondence analysis data made it possible to discrimate between the samples overexpressing or not overexpressing IkappaBalpha, and pointed out some of the potential meaningful spots characterizing the samples in which NFkappaB was active. Then, after visual validation of the computer data, 53 polypeptides were considered to be different in the two classes of gels. Five polypeptides were specifically found in both samples overexpressing IkappaBalpha. The overexpression of IkappaB also induced a lower expression of 11 polypeptides. Finally, 15 polypeptides were only expressed in samples in which IkappaBalpha was not overexpressed and, consequently, in which NFkappaB factors were active. Thus, these polypeptides are candidates for further analysis as putative target gene products of NFkappaB.

Identification of yeast proteins from two-dimensional gels: working out spot cross-contamination

With the complete sequence of the yeast genome now available, efforts by many laboratories are underway to identify each of the spots on two-dimensional (2-D) gels corresponding to the most abundant yeast proteins. The high mass accuracy now attainable using matrix assisted laser desorption/ionization (MALDI)-mass spectrometry equipped with delayed extraction simplifies the process of identification, such that many spots can be unambiguously identified in a short period of time merely by using peptide mass fingerprinting and generally available database matching programs. Although it is not always possible to match spots between gels run by different laboratories, proteins generally yield the same abundant proteolytic fragments when tryptic digestions are performed. Databases containing these signature peptides not only simplify the task of reidentifying proteins from different gels, but also make it possible to identify small amounts of cross-contaminating proteins from different spots, as well as common extraneous contaminants such as human keratins. In this paper, we present data on the identification of > 20 previously unreported yeast proteins from 2-D gels. Some novel proteins were identified from randomly analyzed spots. Focusing on 14 spots in a narrow-pH-range gel, we demonstrate how organizing peak-table data and peptide match-list data into databases enables the identification of a larger percentage of the peaks.