ExPASy Home page | Site Map | Search ExPASy | Contact us | SWISS-2DPAGE |
Hosted by NCSC US | Mirror sites: | Canada | China | Korea | Switzerland | Taiwan |
Identification and characterization of all proteins expressed by a genome in biological samples represent major challenges in proteomics. Today commonly used high throughput approaches combine two-dimensional electrophoresis (2-DE) with peptide mass fingerprinting (PMF) analysis. Although automation is often possible, a number of limitations still adversely affects the rate of protein identification and annotation in 2-DE databases including the sequential excision process of pieces of gel containing protein; the enzymatic digestion step and the interpretation of mass spectra (reliability of identifications). Two approaches are currently underway in our laboratory to increase throughput of protein identification.
Manual protein identification from 2-DE gels is a time consuming technique that starts with spot excision and ends with the update of 2-DE databases. Nevertheless, this repetitive and exhaustive procedure could be easily mechanized using robots, thus increasing protein identification throughput. In our laboratory, automation of protein identification is currently in production using a spots excision robot from ARRM (Sydney, AUS), a liquid handling robot for protein digestion and peptide mixture sample loading from AB (Framingham, USA) as well as two powerful mass spectrometer (MALDI-TOF) also from AB. The major advantage of this robotized approach is the limitation of human interventions and thus an increase in reproducibility and a decrease of the number of potential errors. In our laboratory, this procedure currently gives rise to a throughput of 400 analyzed spots per person and per week. However, this approach has still a low to medium throughput of protein identification as well as a poor resolution. In order to circumvent these drawbacks we have investigated the concept of the molecular scanner described below.
The concept of the molecular scanner was to perform peptide mass fingerprinting analysis by scanning with a MS a PVDF membrane where digested proteins have been collected. To obtain this membrane, proteins previously separated by 2-DE or SDS-PAGE were digested during the electroblotting process from the gel to the PVDF membrane. The collecting PVDF membrane thus contained sets of digestion products of all proteins, each of them localized at discrete positions on the surface. An array of positions was defined on the membrane and was then scanned by a mass spectrometer (MALDI-TOF). A flexible and interactive tool was developed to automatically treat all MS data consecutively and to perform the various steps of the analysis, starting with peak detection and calibration. The positions on the sample plate were converted to apparent molecular weight (Mr) and pI values. The PMF data of all spectra, together with the calculated pI and Mr and other user defined parameters (mass tolerance, chemical modifications considered, species range, etc.) were automatically sent over Internet for protein identification to PeptIdent, a PMF identification tool developed in Geneva. The identification results of PeptIdent were represented as an annotated image. The program generated a first virtual 3-D image where the x and y coordinates related to pI and Mr values, respectively. The z values were represented in gray scale and reflected the signal intensity from the MS spectra related to the number of ions detected. The image file was stored in a graphical format that can be read by the Melanie 2-DE image analysis software package. The image also contained the identification results, which can be highlighted as labels in Melanie. The number of distinct attributes contained in the image reflects the number of dimensions the image virtually contains. These were pI, Mr, identification labels (SWISS-PROT or TrEMBL AC numbers, ID labels), peptide masses and MS intensities. Then for all potentially identified proteins, the annotations from PeptIdent (number of missed cleavages, annotated modifications, chemical modifications of cysteine and methionine residues, peptide sequences) were also available. From all the data contained in this multi-dimensional image the user can choose to filter and visualize only particular aspects. Proteins or peptides can be searched on the image by filtering part of the total information. Thus, a protein can be visualized by the positions where it has been identified. The z intensity can be a binary (black / white for present / absent, respectively) or a gray level. The darkness represents then either the number of peptides found to match the protein in the identification process using PeptIdent, or the sum of the MS intensities of the peptide masses matching the queried protein. Instead of searching for a protein, the user can specify and visualize a set of peptide masses. In this case, the image intensity scale can be defined from the number or the MS intensities of the masses detected out of the chosen list. This procedure permits to scan a 4 x 4cm 2-DE PVDF membrane in less than 12 hours with our 20-Hz laser MALDI-TOF-MS.
Protein identification is becoming a vital and necessary complement to the available fully sequenced genomes. To meet the challenge, newly developed techniques for high throughput protein identification using MALDI-MS and peptide mass fingerprint are needed. Two years ago, a parallel protein digestion (DPD) process was proposed. It provides after parallel digestion of all SDS-PAGE separated proteins a collecting PVDF membrane able to be scanned by MALDI. Acquired data are used to recreate a virtual multidimensional image. Protein electroblotting technique used during this DPD process is really different to the usual and standard electroblotting approach used at present in numerous laboratories. The goal of the current study was to obtain a quantitative evaluation of the effect of the electric field used with this technique in term of protein recovery on the collecting membrane. The effect of the pulsed electric field and the buffer composition were compared to a standard continuous transblotting process defined as a gold standard. Combination of the pulsed asymmetric electric field and CAPS buffer shown an average 65% increase of polypeptide recovery with the strongest effect for high Mr proteins. In conclusion, the present study highlighted a positive influence of the "shaking" effect of the asymmetric alternative voltage on gel protein extraction.
Genome sequencing projects produce large amounts of information that could be translated into potential protein sequences. Such amounts of material continuously increase protein database sizes. At present, 15 times more protein sequences are available in the SWISS-PROT and TrEMBL databases than 8 years ago in SWISS-PROT. One of the methods of choice for protein identification makes use of specific endoproteolytic cleavage followed by the MALDI-MS analysis of the digested product. Since 1993, when this technique was first demonstrated, the conditions required for a correct identification have changed dramatically. Whilst 4-5 peptides with an accuracy of 2-3 Da were sufficient for a correct identification in 1993, 10-13 peptides with less than 60 ppm mass error are now required for Human and E. coli proteins. This evolution is directly related to the continuous increase of protein database sizes, which causes an increase of the number of false positive matches in identification results. Utilisation of an information complement deduced from the primary protein sequence in the process of identification by peptide mass fingerprints can help to increase confidence in the identification results. In this article, we propose the exchange of labile hydrogen atoms with deuterium atoms. The exchange reaction with optimised techniques has shown an average 95% of hydrogen/deuterium exchange on tryptic peptides. This level of exchange was sufficient to single out one or more peptides from a list of potential candidate proteins due to the dependence of hydrogen/deuterium exchange on the peptide primary structure. This technique also has clear advantages in the identification of small proteins where direct protein identification is impaired by the limited number of endoproteolytic peptides. Then, primary sequence information obtained with this technique could help to identify proteins with high confidence without any expensive tandem mass spectrometer instruments.
A number of tools have been developed or adapted to handle automation of protein identification, and to help the characterization and more generally the interpretation of mass spectrometry related data. The tools described are either available on the ExPASy server or are developed for internal use.
ExPASy Home page | Site Map | Search ExPASy | Contact us | SWISS-2DPAGE |
Hosted by NCSC US | Mirror sites: | Canada | China | Korea | Switzerland | Taiwan |