Several efforts have reported gels with over arrayed spots Gauss et al. While these studies are useful at pushing the technical boundaries of arraying, they do not represent a normal outcome. From a more realistic gel, perhaps, of the — proteins are abundant enough to be confidently assigned an identity using MS. If the sample analysed was derived from a whole cell or whole tissue there may be as many as 10, different gene products in the sample. The direct analysis of a complex protein lysate by MudPIT has the potential to identify a larger dynamic range and could enable the identification of thousands of proteins from a single sample McDonald et al.
The challenges faced by researchers attempting to obtain maximum proteomic depth from the sample are even further confounded by the dynamic range of proteins found with the cell Patterson, Plant samples isolated from green tissue contain large quantities of the chloroplast protein RuBisCO, which can obscure less abundant proteins both on gel arrays and when directly analysing complex lysates Komatsu et al. Furthermore, the majority of other proteins that are found in whole-cell or tissue lysates represent major metabolic enzymes and structural proteins, not the modulating and regulatory components such as transcription factors that are more often desired.
Thus much of the current proteomic analyses are surveying the abundant cellular components, while the majority of proteins within the cell are being precluded from the analysis. As a consequence when proteomic techniques are being used to interpret phenomena such as the effects of certain hormonal regimes, developmental processes or defence against a pathogen only a small fraction of abundant housekeeping proteins are really being investigated. Furthermore, even when these data on changes in particular protein abundance can be identified from the experimental analysis, they tend to refer to only individual members of complex biochemical pathways and thus do not provide a complete picture of the response occurring in whole pathways.
While disappointing, this is not really surprising given the fragmented nature of the display and analysis techniques. The use of subcellular experimental approaches can allow the analysis of proteins that are co-localized and are often functionally interactive. This also decreases the complexity of protein samples, allowing a greater depth of analysis of the expressed proteome. These approaches commonly employ cellular fractionation, centrifugationbased purification of organelle or cellular compartment and MS to identify peptides Millar, The chloroplast has been dissected through multiple studies that have explored its sub-organelle proteome including the thylakoid lumen Peltier et al.
Subsequently further studies have targeted complexes and used a direct lysate analysis approach Eubel et al. The proteome of nuclei has also been analysed Bae et al. The vacuole has been studied through analysis of the tonoplast membrane and the vacuole contents Carter et al. Only two small studies have been undertaken on the peroxisome isolated from cotyledons Fukao et al.
Several other studies have identified proteins from a variety of intracellular membrane systems including the Golgi and endoplasmic reticulum ER; Prime et al. The proteomes of the cell wall Chivasa et al. These studies provide a mechanism to penetrate deeper into cellular proteomes and can allow a more focused approach when undertaking comparative proteomic studies. The model plant Arabidopsis probably represents the source of the largest number of such sub-proteomic studies and these data have been recently compiled in an online SUBcellular database for Arabidopsis proteins called SUBA Heazlewood et al.
This database provides subcellular localizations for approximately non-redundant Arabidopsis proteins that have been identified through MS and illustrates the capabilities of subcellular fractionation as a means to address proteomic depth Table 1. Since no sequence data are obtained, unless using PSD , data matching is heavily reliant upon mass accuracy and sequence availability. Thus the real limitation to using PMF in proteomic studies is sequence availability Heazlewood and Millar, While several high-fidelity plant genomes have been made available in recent years e. Arabidopsis and rice , many other well-studied plant species lack sufficient genomic data for confident matching assignments to be readily undertaken using PMF.
Most of the interrogation programs such as Mascot Matrix Sciences or MSFit Protein Prospector will take these parameter shifts into account when scoring the matching process Perkins et al. Nonetheless, it is still in the hands of the researcher to interpret the final result, as nearly all data are bound to match some protein from large databases if parameters are broadened enough. The lack of a successful species-specific positive match using PMF can also lead researchers to attempt using cross-species matching.
Clearly genomic analysis has shown that there is a substantial degree of sequence conservation between many plant species. The problem faced for significant proteomic data matching is the requirement for identical masses of regions, specifically cleaved peptides under investigation, between proteins of divergent species.
The alternatives are to attempt to allow for error tolerances to account for amino acid substitutions. This will lead to much less confidence in the match and greatly increases the number of false and misleading identifications. Some attempts have been made to examine the opportunities available for cross-species matching with PMF data, but it is certainly not an optimal avenue for large-scale proteomic studies Liska and Shevchenko, The data format provides the mass of the analysed peptide as well as some fragmentation data which can either be used directly for interrogation with a database using available software e.
The use of either method will provide a far greater confidence in the resulting match, but just as with PMF matching, the final interpretation still requires some level of assessment. Since peptides can be concentrated and separated prior to analysis, far more data on each peptide can be obtained. This is not entirely accurate since this technique has obvious financial advantages, the availability and ease of instrument operation needs to be considered as does the ability to process hundreds or thousands of samples a week.
Such studies often exploit the availability of large-scale expressed sequence tags EST sequencing programs. One of the major issues in the area of quantitation is the problem of dynamic range, because protein abundances can cover five to six orders of magnitude in cell extracts. The ability to accurately resolve abundance across this dynamic range is still well outside the scope of current technologies.
Nonetheless, it is still possible to quantify a limited range with some accuracy. It is a comparatively sensitive staining method with a detection limit of around 5—10 ng of protein. The stain is relatively inexpensive, it is compatible with MS, it is simple to remove through washing and importantly it does not result in any modifications to the protein prior to MS Neuhoff et al.
While silver staining is generally more sensitive with detection limits around 0. These stains offer similar sensitivities to silver staining but provide a greater dynamic range and a far better level of compatibility with subsequent MS Lauber et al. There are still some advantages to using either Colloidal Coomassie or silver staining as scanning the gels for the comparative analysis requires a standard flatbed scanner, while fluorescent stains will require a charge-coupled device CCD camera or commercial LASER scanner for visualization Steinberg et al.
Furthermore, the fluorescent stains are relatively expensive which can be a factor if undertaking a large-scale gel-based comparative proteomic study Rose et al. Another problem with increasing levels of sensitivity offered by these newer stains is the inability to routinely identify the proteins they can stain at the low range. Although many mass spectrometers are marketed with promises of identifications at the femtomole or attomole level, this usually occurs under ideal conditions and is not a consistent outcome when generic settings are used in high-throughput MS protocols.
It should also be noted that these sensitivity levels are calculated on the isolated peptides supplied at the ionization source and do not take into account the efficiencies of peptide extraction from gels, sample handling, etc. For accurate comparative gel analysis, the scanned gels need to be analysed and some level of quantitation undertaken. There are a variety of commercially available software options as well as free packages to undertake these procedures Young et al. Several factors can affect the performance of these programs including: poor spot detection or the inability to discriminate between two valid spots, poor differentiation of proteins spots from background, and issues in matching spots between gels when undertaking a comparison building a replicate gel Rosengren et al.
Understanding every nuance of these programs is not a trivial exercise, with much time investment needed to allow the full exploitation of most of the features. Clearly one of the major problems when undertaking comparative proteomic studies using 2D-PAGE is the reproducibility between replicates and between samples. Inconsistencies and procedural variation between IEF focusing and arraying of samples can cause real problems with statistical analysis and validation.
More recently a fluorescent-labelling technique has been developed that attempts to overcome some of the technical discrepancies that invariably occur during 2D-PAGE. This method labels the two samples to be analysed with two different CyDye fluors and significantly can utilize the third fluor to label a loading control. The resultant gel is analysed using a fluorescent scanner with appropriate filters. The scanned images can be viewed independently to assess arraying of samples, but since the same protein in either of the samples being analysed will migrate to the same point in the gel after IEF and electrophoresis, the scanned images can be accurately overlayed, providing unsurpassed 2D-PAGE comparative proteomic capabilities Rose et al.
It is difficult to make too many inferences about abundances of the protein identified, since the process of peptide ionization is not similar for each of the peptides for a given protein. As a consequence techniques have been developed to provide relative quantitation of the same peptide between samples. This method uses thiol-reactive isotope-coded affinity tags ICAT to label cysteine residues in peptides from each sample.
The system employs two different ICAT tags to label the two samples for analysis. One tag contains eight hydrogen atoms light tag while the second contains eight deuterium atoms heavy tag , providing a distinct mass difference 8 amu. The samples are differentially labelled and mixed in equal amounts, digested and tagged peptides purified using a biotin moiety built into the tags.
The origins of the two peptides can be readily determined by the 8 amu mass difference between peaks. The relative intensity of each of the peptides can provide a comparative ratio to give relative quantitation of the protein identified Gygi et al. One of the problems that were soon identified with the ICAT approach is the requirement for the presence of cysteine in the protein for labelling and quantitation to occur. While several other labelling techniques have employed the use of stable isotopes such as 15N Oda et al. More recently an amine-specific isotope labelling system was developed that could provide tagging of all peptides in a sample.
The system employs four isobaric same mass tags that can be mixed to analyse four different samples simultaneously and is known as iTRAQ Ross et al. The system uses a similar logic to ICAT, where identical peptides will be analysed by the mass spectrometer simultaneously, and it provides strong quantitation capabilities without being affected by the vagaries of peptide ionization between samples. Relative or even absolute quantitation if a standard is labelled and included is obtained when peptides with the isobaric tags are fragmented, causing the release of a specific mass reporter for each tag Ross et al.
Since all peptides are labelled by this method, as opposed to the cysteine specificity of ICAT, the analysis process can be very complicated due to the increased abundance of ions in complex samples. Moreover, just like the analysis of complex lysates, there are limitations to the number of proteins that can be identified and thus quantified in these labelled samples. Although recent estimates have suggested that the number of functionally distinct proteins that could be produced by post-translational modifications may be 10— times this base number Rappsilber and Mann, A wide range of techniques can be used to characterize protein modifications using proteomic analysis techniques.
The 2D-PAGE array can be useful for the visualization of apparent post-translational modifications as many modifications alter the pI of the protein with only minor changes to the molecular mass Rabilloud, This uncontrolled modification results in a protein population with a variety of alterations and causes the appearance of a series of horizontal repates of the protein on the 2D-PAGE. Often these putative modifications are in fact non-biological and are a result of degraded urea in the sample buffer causing carbamylation Herbert et al. However, clearly biological protein modifications are also present in these analyses and these moieties do in fact result in pI shifts in the 2D-PAGE array Zhu et al.
The difficulty is in locating the proteins on the gel and characterizing the modified residues. Several approaches have been applied to this problem and by far the most successful has been studies of protein phosphorylation. While such approaches can provide a nice picture in the analysis process, increasingly mass spectrometric techniques rather than 2D-PAGE arrays are being utilized to confidently assign a plethora of biologically significant protein modifications including acetylation, phosphorylation, glycosylation and methylation Wilkins et al.
These approaches utilize techniques like immobilized metal ion affinity chromatography IMAC to selectively recover and analyse phosphorylated peptides from complex lysates. Significantly such techniques can be used to enrich low-abundant modified proteins or peptides from complex lysates. The study of post-translational modifications in proteomics represents one of the major issues in this field. Currently studies analysing complex samples using MudPITtype analyses produce a significant number of unmatched ions or peptides that do not match any protein in the database Washburn et al.
Many such studies are undertaken on model systems where excellent genome sequence data are available. Whether many of these unmatched ions turn out to be peptides of known proteins that have been modified beyond recognition in some way remains to be seen. The ability to identify modified proteins and to conclusively map modifications to amino acid residues will provide a significant resource in the functional interpretation of the proteome, although the area is still in early development. This desire has developed from the realization that much of the data currently being produced is being lost.
Currently the data presented in the majority of publications is represented in binary format, with protein identifications simply listed within the manuscript. Consequently the primary data that were used to make these identifications are essentially inaccessible to the wider community. Moreover, much of the detail of the analysis that was undertaken in the experiment is lost.
There is a real fear that there are hidden gems in the primary data of many studies that will be simply discarded or just gather dust on computers around the world. With the development of better matching algorithms and the ongoing sequencing of more genomes, there is a possibility that much of these neglected data could be exploited by researchers in the future.
Attempts are currently underway to standardize the capture and storage of proteomic data forms. Currently most of the hardware and software designers are involved in producing a standard data format mzDATA , which can be readily examined and distributed for all major platforms Orchard et al.
There are also moves to produce a standardized management schema for proteomics, similar to the MIAME schema developed by the microarray community.
Book books library
Recently there have been a series of editorials from prominent proteomic journals which now minimally require the inclusion of the data file that was used for matching as well as a variety of other basic information on how the matching was undertaken during submission of manuscripts Beavis, ; Bradshaw, These decisions have come about because of the explosion in the use of proteomics by non-specialist researchers and the opinion that many applications of this technology were not stringent enough.
Much of these concerns stem from the problem that too many of the identifications that are currently being reported in the literature are likely to be false positives Carr et al. Such problems usually occur when data matching is pushed to its limit as in most circumstances mass spectral data will always match something in peptide databases. Such issues were not as serious when the majority of studies were carried out using 2D-PAGE which greatly limited the amount of data produced, and provided some empirical information about the protein being analysed molecular weight and pI.
The problem has become far more prevalent with the explosion in the use of direct analysis of complex samples where many thousands of spectra are produced and any physical relationship with the proteins being analysed is lost. With the number of large-scale proteomic analyses that are currently being undertaken, many researchers have found themselves moving away from the ubiquitous spreadsheet and into database platforms that allow them to more readily analyse the data flows produced Table 1.
These databases largely grow out of the desire for researchers to manage their own data in-house, but many flow into the public arena. The 2D-PAGE array was the primary analytical method used in proteomics for many years, and as such there are several 2D-PAGE-type databases that have attempted to provide reference maps of arrayed proteins from different species and different organs for the community at large.
The problems with 2D-PAGE databases are that they are largely limited to the cell lines and species that are chosen to be displayed. The ability to match 2D-PAGE arrays is problematic even when they are produced as replicates from a similar tissue source, so attempting to use online digital maps as accurate matching tools to a users own protein gels is exceedingly frustrating. Although there have been several attempts to use 2D-PAGE reference maps to identify proteins across plant species, to date these have not been entirely successful Mathesius et al. Furthermore, these databases are not equipped to deal with the newer proteomic techniques involving direct analysis of complex samples.
The proliferation of specific plant proteomic databases in recent years has probably reflected the movement of the community into more advanced proteomic studies. Much of this work has discarded the 2D-PAGE array and focused on specific sets of proteins within the cell. These subsets have invariably represented well characterized biochemical entities, such as the mitochondrion, the plastid and the nucleus Table 1.
These studies have not only defined distinct functional sets of proteins, but as many groups represent sub-fractions of the cellular milieu, they have also enabled proteomic depth of the cellular proteome. At this stage these databases generally provide the presence or absence of a protein from a given subset.
Proteomic databases will need to become far more dynamic in the future and will need to provide userannotation features as well as full integration with the genomic, transcriptomic and metabolomic datasets. Importantly they will need to allow tracking and archiving of experimental procedures and the ability to access the primary data.
Such features will allow the current databases to integrate with other areas providing advanced search and data-mining abilities that will be necessary in order to interpret the vast amount of information currently being produced. While at the present stage there is no organized data storage program for primary data produced through plant proteomics, there are moves in other organisms to establish some consensus in proteomic data handling that are likely to show the way to the plant proteomic community Taylor et al.
Table 1. Recently there have been a wide variety of freeware utilities to undertake many of the techniques that were previously only available from commercial products. There are some inherent limitations of these free utilities, namely many are very early releases of preliminary techniques that may not have been tested as widely as many of the larger commercial products. Nonetheless, many of these new tools provide far more flexible tools for analysis or manipulation of data. Currently this will involve the adoption of quantitation methods that employ the mass spectrometer directly Chelius et al.
The study of post-translational modifications and how these changes affect biological processes is still a largely unexplored area in plant proteomics. However, these techniques only exist in the context of biological questions and it is here that we need to maintain our understanding of issues of interest in plant biology in order to do interesting experiments and link our findings to those of our colleagues working with very different techniques.
Much of our current work in plant proteomics tends to be binary identifications, and for this to link with other large data analysis technologies, adequate quantitation will be required. Such information will provide the ability to more closely link protein function with global protein abundances, overlaying them into regulatory and metabolic networks within the cell and ultimately play a part in explaining how plants grow, develop and react to their environment. References Alexandersson, E. Plant Cell Physiol. Altschul, S. Nucleic Acid.
Andon, N. Proteomics, 2, — Bae, M. Plant J. Beavis, R. Proteome Res. Bodnar, W. Mass Spectrom. Bordini, E. Rapid Commun. Bradshaw, R. Calikowski, T. Carpentier, S. Proteomics, 5, — Carter, C. Plant Cell, 16, — Chelius, D. Chen, H. Chevallet, M. Electrophoresis, 19, — Chivasa, S. Electrophoresis, 23, — Eubel, H. Supercomplexes and a unique composition of complex II. Plant Physiol.
Plant Method. Ferro, M. Fido, R. Humana Press Inc. Friso, G. Froehlich, J. Fukao, Y. Garwood, K. BMC Genom. Gatlin, C. Gauss, C. I Brain proteins: separation by two-dimensional electrophoresis and identification by mass spectrometry and genetic variation.
International Plant Proteomics Organization (INPPO)
Electrophoresis, 20, — Giavalisco, P. Electrophoresis, 24, — Electrophoresis, 9, — Granier, F. Gygi, S. Han, D. Haslam, R. Havlis, J. Heazlewood, J. Plant Biol. Acta, , — FEBS Lett. Heinemeyer, J. Phytochemistry, 65, — Herbert, B. Electrophoresis, 22, — Hurkman, W. James, P. Protein Sci. Kleffmann, T. Natl Acad. USA, 99, — Komatsu, S. Kruft, V. Lauber, W. Link, A. Liska, A. Proteomics, 3, 19— Liu, H. Lopez, M. Electrophoresis, 21, — Mann, M. Mathesius, U. Proteomics, 1, — McDonald, W. Millar, A. Plant Physiol, , — Mithoefer, A. Plant Pathol. Neubauer, G. Neuhoff, V.
Newton, R. Oda, Y. USA, 96, — Orchard, S. Proteomics, 4, — Patterson, S. Electrophoresis, 17, — Peck, S. Plant Cell, 13, — Peltier, J. Plant Cell, 14, — Pendle, A. Cell, 16, — Perdew, G. Perkins, D. Pierpoint, W. Methods in Molecular Biology, Vol. Prime, T. Prince, J. Quackenbush, J. Rabilloud, T.
Proteomics, 2, 3— Electrophoresis, 18, — Rappsilber, J. Rose, J. Rosengren, A. Proteomics, 3, — Ross, P. Santoni, V. Biochimie, 81, — Schubert, M. Shevchenko, A. USA, 93, — Shimaoka, T. Skopp, R. Steinberg, T. Szponarski, W. Taylor, C. Wang, X. Plant Cell, 17, — Washburn, M. Whitelegge, J. Wilkins, M. Wolters, D. Wu, C. Yates III, J. Young, N. Bioinformatics, 20, — Zhu, K. This page intentionally left blank 2 Proteomic analysis of post-translational modifications by mass spectrometry Albrecht Gruhler and Ole N.
Jensen 2. Mass spectrometry MS based analytical strategies are already providing detailed insights into the molecular functions of PTMs, including phosphorylation, glycosylation and acylation. Emerging quantitative proteomic methods will soon enable detailed spatial and temporal studies of protein modifications, thereby revealing dynamic aspects of cellular regulatory networks in plants. We present an overview of recent analytical strategies for the systematic determination of PTMs by MS.
The pace of development is propelled by integration of novel computational techniques, ever more advanced and sensitive analytical methods and optimized genetic and biochemical approaches to study complex biological systems. In early proteomics efforts, the main tools were two-dimensional 2D electrophoresis for protein separation and image analysis of protein spot intensities for protein quantitation.
The emergence of soft ionization MS techniques i. The combination of multidimensional capillary chromatography and tandem MS soon emerged as an increasingly robust approach in proteomics Washburn et al. PTMs of proteins present a formidable analytical challenge. First of all, more than different types of PTMs have been characterized Krishna and Wold, and new ones are regularly reported see e. Delta-Mass list at www. PTMs come in many sizes and with a wide range of physicochemical properties Table 2. PTMs generate a large diversity of gene products due to Table 2. There are a multitude of relevant references for the different PTMs and we apologize to all the authors we could not cite due to the limited space.
Since PTMs alter the molecular weight of proteins, the mapping, identification and characterization of individual modifications are often achieved by MS analysis of intact proteins, when feasible, and then by MS analysis of the proteolytically derived peptides Figure 2. In the last few years, significant progress in proteomics analysis of proteins and PTMs in plants has been made Huber and Hardin, ; Laugesen et al. In the following sections, we will describe approaches to the detection and characterization of PTM proteins by MS.
The discussion will cover not only plants, but also include examples from other species, where appropriate. Several recent reviews are recommended Jensen, , ; Kirkpatrick et al. This method, as effective as it is for the analysis of many low-complexity samples, poses serious challenges to the characterization of proteins and PTMs on a system-wide basis. One of the main reasons is the large number of distinct peptides generated from the protein content of any cell.
The yeast genome as an example encodes approximately genes, giving rise to more than , tryptic peptides in an in silico digestion of the corresponding gene products with trypsin. Considering PTMs will increase this number several fold. In addition, some PTMs are substoichiometric and transient, meaning that a particular PTM attachment site is not modified on all proteins present in a cell at a given time. Other PTMs are heterogeneous, as is the case for glycosylation, where differential modification can give rise to several glycan structures with different masses occupying the same site of a protein.
These characteristics of PTMs often result in low amounts of modified peptides that are available for analysis by MS and pose great demands on the sensitivity of mass spectrometers. Some of the PTMs e. Therefore, the sensitivity of many proteomic approaches depends on the reduction of sample complexity, often achieved by fractionation or the enrichment of post-translationally modified proteins and peptides. Figure 2. The combination of protein and peptide separation techniques with advanced MS provide the means to achieve high specificity, selectivity and sensitivity in proteomics experiments.
Basic steps of a proteomics experiment marked in bold are the cell lysis, protein extraction, peptide generation and analysis by MS. Chromatographic fractionation and affinity based enrichment techniques are very useful for the analysis of post-translationally modified proteins and peptides. They can be employed during different stages of sample preparation. The most direct proteomic approach to analyse a cell or tissue sample by MS comprises only a few steps Figure 2.
These include homogenization of the tissue, the disruption of the cells, the extraction of proteins often aided by detergents or denaturing buffer systems and the generation of peptides by proteolytic cleavage of proteins. For an LC-MS analysis, the peptide mixture is further fractionated by reversed phase chromatography prior to injection into the mass spectrometer.
In addition to the reversed phase separation of peptides, additional orthogonal chromatography steps often serve to further segregate the peptide mixture into different fractions that can be analysed separately by LC-MS. It has been widely applied to the large-scale identification of proteins and PTMs Issaq et al. Protein amounts are commonly determined by densitometry of Coomassie and silver stained gels or by Western blotting. More recently, fluorescent dyes have been developed that allow more accurate quantitation over a larger linear range Patton, Fractionation and enrichment strategies can be employed at all levels of the sample preparation and can often be combined with each other.
One way to reduce the number of proteins is the isolation of individual subcellular compartments such as cell organelles, membranes or protein complexes Brunet et al. However, the most efficient way to systematically investigate PTMs is the affinity enrichment of either modified proteins or peptides.
Methods have been developed that are based on different techniques such as affinity chromatography, PTM specific antibodies or PTM-directed chemical modification of proteins or peptides. A number of these applications are described below. Many PTMs are regulated by the cell and in order to investigate their dynamics, quantitative proteomic studies are desirable.
Advances in MS based technologies within the last few years have tremendously enlarged the repertoire of methods that can be used for quantitation of proteins and PTMs. Most of them rely on stable isotope labelling of a control sample, which is mixed and analysed together with the sample of interest Julka and Regnier, Thereby relative amounts of the unlabelled and labelled peptide pair — which have similar physical properties — can be determined from the same spectrum. Introduction of the isotopic label can occur at different stages of the proteomic experiment, dependent on the type of labelling method used.
Cell cultures can be labelled in vivo either by growing them within the presence of an 15N-containing nitrogen source Oda et al. Alternatively, stable isotopes can be introduced by chemical modifications of proteins and peptides Leitner and Lindner, There exists a broad variety of reagents that have affinity for different reactive groups, for example, thiols, free amines or carboxylic groups. Commercially available reagents include isotope coded affinity tags ICAT reagents that covalently bind to cysteines Gygi et al.
Alternatively, MS experiments of different samples can be compared to each other by aligning and matching identical peptides and comparing their relative intensities Peters et al. This requires standardized conditions for sample preparation, liquid chromatography and mass spectrometric analysis, in order to achieve good reproducibility between individual samples.
The combination of high peptide mass accuracy accurate mass tags and highly reproducible chromatographic separation technologies may eventually provide the means for routine, comparative analysis of complex peptide samples Pasa-Tolic et al. So far, this has not been and maybe never will be achieved, since even the most comprehensive studies of cells or subproteomes only identify a fraction of the total number of proteins present. In higher organisms, the percentage of identified proteins drops, due to the larger complexity of their genomes, not even taking into account that many genes give rise to splice variants, which are not yet completely annotated, and that most proteins bear PTMs.
This lack in completeness is due partly to the widely varying physical properties of distinct proteins and PTMs and the vast differences in protein abundance within cells and tissues. In addition, current proteomics methods are limited, both with regards to sample preparation and to the sensitivity of the MS instrumentation. Some PTMs e. For these reasons, identification of modified proteins and peptides and characterization of the PTMs are not comprehensive in complex samples, but even with these limitations, the present technologies provide powerful tools to study PTMs, which allow for the detection of hundreds of modified peptides in a single experiment.
In the following paragraphs, we will give examples for proteomics approaches that have been used to analyse classes of PTMs. There will be references to plant studies; however, not all of the methods have been applied to plants on a systemwide basis yet. Cell differentiation, propagation and responses to environmental clues are governed by cell signalling pathways that involve protein phosphorylation. For this reason, the study of protein phosphorylation has long been the focus of many investigations and numerous methods have been employed for the detection and modification of phosphoproteins and phosphorylation sites.
Phosphoproteins are then usually separated by gel electrophoresis and detected by autoradiography. Recently, a fluorescent dye called Pro-Q diamond was introduced that specifically stains phosphoproteins in 1- or 2D gels and has the potential to replace the more hazardous 32P labelling Martin et al.
Alternatively, phosphorylation site specific antibodies can be employed for the detection and characterization of protein phosphorylation, either by use of Western blotting or in immunoprecipitation experiments. A number of different methods can be found in the literature, which have been implemented on different types of mass spectrometers Mann et al. They centre on the detection of different phosphoamino acid specific fragments generated during the analysis either by post-source decay or during intentional peptide fragmentation in both positive and negative ion modes.
However, the intrinsic chemical properties of the phosphate groups, namely their low pKa value and their relatively labile phosphoester bonds, present challenges to their analysis by MS. Hydrophilic phosphopeptides might not bind to the reversed phase material commonly used for sample purification or LC-MS Larsen et al. In addition, their ionization is believed to be less efficient than that of unphosphorylated peptides, leading to their suppression in complex sample mixtures.
And finally, during peptide sequencing, the loss of phosphoric acid from phospho-serine and -threonine is favoured Figure 2. Combined with the fact, phosphorylated proteins often are of low abundance, many efforts have been made to specifically enrich phosphopeptides prior to their analysis by MS. Starting from a whole cell lysate, more than phosphopeptides containing close to phosphorylation sites could be identified in a single mass spectrometric experiment on an ion trap mass spectrometer.
IMAC binds preferentially phosphopeptides, but has also affinity towards acidic peptides, which decreases the specificity of purification of phosphopeptides due to concomitant binding of unphosphorylated Plasma membrane fractions from Arabidopsis thaliana were treated with trypsin and the generated peptide mixture was fractionated by strong anion exchange chromatography SAX. Fragment ion signals corresponding to phosphoserine and to dehydro-alanine Dha after neutral loss of H3PO4, respectively, were observed.
Therefore, methylation of carboxyl groups was performed prior to enrichment of phosphopeptides on the IMAC column. Methyl estrification of peptides reduces unspecific binding to IMAC beads, however additional steps of sample handling, including desalting and lyophilization are necessary, which require relatively large sample amounts and might entail sample loss, partial methylation of peptides and side reaction such as deamidation of glutamines and aspartames.
Methyl esterification may also be the reason for the large number of doubly or multiply phosphorylated peptides observed in this study, since in the absence of free carboxylic groups peptides with multiple phosphates are expected to bind stronger to the Fe—NTA complexes than singly phosphorylated peptides. Plasma membrane fractions were prepared from Arabidopsis suspension cells by phase partitioning, washed with Na2CO3 in the presence of the detergent Brij that facilitates the formation of inside-out vesicles, where the cytoplasmic moieties of membrane proteins are accessible on the outside of the vesicles.
These vesicles were treated with trypsin and released peptides were fractionated by strong anion exchange SAX chromatography.
Bioinformatic analysis of this large-scale dataset revealed a series of novel features of plant plasma membrane phosphoproteins, including receptor-like kinases Nuhse et al. A similar phosphoproteomic approach has been used to quantitatively investigate the activation of an MAP kinase pathway in yeast in response to alpha-factor, the mating hormone of yeast Gruhler et al. Auxotrophic yeast cells were labelled by growing them in the presence of [13C]-arginine and [13C]-lysine.
Equal amounts of isotope encoded cells and pheromone treated cells were then mixed and analysed together. It has been shown previously that at pH 2. Therefore, the first fractions of an SCX gradient are enriched for phosphopeptides Beausoleil et al. In this study, it has been demonstrated for the first time, that quantitative proteomics is capable to characterize a prototypic signalling pathway. Recently, TiO2 has been described as a new solid support for the purification of phosphopeptides, which exhibits less unwanted binding of acidic peptides than IMAC, but still retains high affinity for phosphorylated peptides Pinkse et al.
Pinkse and co-workers demonstrated the potential of TiO2 columns for the affinity purification of phosphopeptides with on-line multidimensional LC-MS by coupling a TiO2 column to reversed phase pre- and analytical columns. From a tryptic digest of autophosphorylated cGMP protein kinase, phosphopeptides were selectively retained on the TiO2 column and thereby separated from other peptides, which were analysed by LC-MS. Larsen and co-workers modified the method for off-line analysis of phosphopeptides with TiO2 micro-columns.
The low pH value of this buffer decreases the unspecific binding of acidic peptides to TiO2 as compared with the acetic acid based binding buffer used by Pinkse et al.
Download Product Flyer
This improvement in selectivity at lower pH has subsequently also been reported for the enrichment of phosphopeptides with IMAC Kokubu et al. Likewise, the addition of DHB improves markedly the specificity of phosphopeptide purification. The authors explain this phenomenon by the existence of two different types of binding sites on the surface of TiO2 beads, which are preferentially occupied by either phosphate groups or acidic amino acids via their carboxyl groups. DHB is thought to compete with acidic peptides for binding sites leaving the phosphate binding unaffected.
The majority of phosphorylation occurs on serine and threonine residues, whereas tyrosine phosphorylation only accounts for 0. Nevertheless, tyrosine phosphorylation is essential in many cell signalling pathways in mammalian cells and deregulation often contributes to oncogenesis Blume-Jensen and Hunter, In particular, receptor tyrosine kinases e. Tyrosine phosphorylation in plant cells is at present largely uncharacterized, with the exception of MAP kinases that have been shown to participate in various stress responses Zhang and Klessig, Since phosphotyrosine-containing peptides comprise a minor fraction of all phosphopeptides, generally only few have been detected in large-scale phosphoproteomics studies that have employed enrichment strategies, which did not discriminate between different types of phosphorylation.
In contrast, efficient enrichment of phosphotyrosine-containing proteins can be achieved with the help of antibodies with high affinity for phosphotyrosine residues. Two recent elegant studies combined immunoprecipitation of tyrosine-phosphorylated proteins with in vivo stable isotope labelling to follow the kinetics and regulation of tyrosine phosphorylation Blagoev et al. In HeLa cells, the analysis of five different time points of EGF stimulation could follow the up- and downregulation of tyrosine phosphorylation of the EGF receptor and many of its known interacting proteins Blagoev et al.
Likewise, analysis of EGF and platelet derived growth factor PDGF signalling in mesenchymal stem cells identified not only proteins activated by both growth factors, but also differences between the two pathways. Since only EGF, but not PDGF, triggers the formation of bone-forming cells, these results may help to determine the factors necessary for stem cell differentiation Kratchmarova et al. However, purification of phosphotyrosine proteins and analysis by MS often only identifies the protein, but not the site of phosphorylation.
Therefore, IMAC has been used to isolate phosphopeptides from proteins immunopurified with phosphotyrosine antibodies Salek et al. With this strategy, a large number of phosphotyrosine peptides could be identified by LC-MS in a single analysis. They identified 78 phosphorylation sites from 58 proteins, 52 of which had been previously associated with EGF signalling.
Phosphorylation of the EGF receptor on tyrosines and was strongly increased 5 min after stimulation of the cells followed by slow dephosphorylation. Cluster analysis with self-organizing maps allowed the identification of proteins with similar phosphorylation profiles. Recently, phosphotyrosine specific antibodies have also been applied to the affinity purification of phosphotyrosine peptides Rush et al. Investigating tyrosine phosphorylation in three different cancer cell lines, the authors prepared protein extracts that were enzymatically digested and subsequently partitioned into three fractions by reversed phase solid phase extraction.
Peptides from each fraction were incubated with immobilized phosphotyrosine antibody. Captured phosphopeptides were eluted with 0. Employment of a panel of four different proteases, trypsin, chymotrypsin, endoproteinase GluC and elastase, increased phosphopeptide yield by a factor of more than two and increased the confidence of a positive identification, because in many cases several distinct peptides were detected for a specific phosphorylation site.
With this approach, close to phosphotyrosine peptides were identified from three cell lines, making it the largest study of tyrosine phosphorylation to date. In plants, protein glycosylation plays a role in metabolism and cell remodelling mediated by, for example, plasma membrane proteins with N-linked glycan structures or GPI anchors, and cytosolic proteins with O-linked glycans.
Oligosaccharides are very diverse with regards to their structure and composition and their complete characterization by MS is, therefore, not trivial, usually requiring purification of the glycoprotein. Determination of the glycan structure can be achieved by chromatographic separation of the different molecular species, followed by a combination of successive treatment with exoglycosidases and MS or by direct fragmentation MS analysis Zaia, In many cases however, the sites of carbohydrate attachment are of interest. Isolation of glycoproteins or glycopeptides and the subsequent removal of the oligosaccharide chain allow the identification of glycosylation sites from complex samples by MS.
Different strategies to enrich glycoproteins or glycopeptides have been described in the literature. These include the use of lectins, which bind more or less specifically certain classes of oligosaccharide chains, hydrophilic or normal phase chromatography and chemical modification with subsequent attachment of affinity tags. In the following, we will give examples for these methods. Lectins are defined as carbohydrate binding proteins that do not modify the oligosaccharides, a property which propagated their widespread use in the biochemical study of glycoproteins.
Lectins are found in many organisms, including plants and mammals, where they have important roles in cell—cell interactions, innate immunity, angiogenesis and modulation of tumour progression. They exhibit diverse affinities towards different carbohydrate structures see Figure During the enzymatic cleavage, the Asn residue to which the glycan is attached is converted to aspartic acid. If the reaction occurs in the presence of HO, aspartic acid incorporates two 18O atoms shifting the molecular weight of the peptide by 4 Da.
This mass shift facilitates the identification of formerly N-glycosylated peptides by MS and their discrimination from nonenzymatic deamidation of asparagines that can occur either in vivo or during sample preparation. Kaji and co-workers used Concanavalin A to capture glycoproteins from C. Cis-diol-groups in oligosaccharides are first oxidized with peroxide to form free aldehyde groups, and then incubated with immobilized hydrazide. Thereby, glycopeptides are specifically and covalently linked to the hydrazide resin and can be released by enzymatic deglycosylation with PNGase F.
Application of this method to a membrane fraction of a prostate cancer epithelial cell line led to the identification of glycopeptides from more than 60 unique proteins Zhang et al. Most carbohydrate chains are hydrophilic due to the presence of polar groups such as hydroxyl, carboxyl, aminoacyl or sulphate groups. This property can be exploited for their enrichment by hydrophilic interaction liquid chromatography HILIC. Applying a mixture of tryptic peptides from several glycoproteins to an HILIC column made it possible to selectively enrich glycopeptides on the column and to subsequently analyse them by MS.
This method was then applied to the characterization of N-glycosylation sites from human plasma. This one residual carbohydrate group serves as a marker for N-glycosylation that provides unambiguous evidence for the position of the oligosaccharide attachment site but at the same time allows the identification of the peptide sequence by MS fragmentation. The analysis of the deglycosylated peptides from the HILIC fractions resulted in the identification of 37 N-glycosylated peptides.
Strikingly, no glycopeptides were found in the flow-through fractions from the HILIC column, demonstrating the high efficiency of this method to isolate glycopeptides. The previously described approaches aim at the identification of glycosylation sites rather than the characterization of the carbohydrates. The latter task is challenging, because of the heterogeneity of glycosylation, where often many structurally and composition-wise different oligosaccharides occupy one glycosylation site, thereby leading to a spreading of the signal into several glycopeptide peaks during analysis of by MS.
In addition to this reduction in signal intensity, oligosaccharides may interfere with peptide sequencing by MS complication, the concomitant determination of glycosylation sites. Therefore, these analyses generally require high amounts of purified protein. Larsen and co-workers recently introduced a strategy that allowed them to characterize glycan structure and glycosylation sites from gel separated proteins in the low picomole range Larsen et al. The remaining part was further digested with protease K, an unspecific proteinase that further cleaves the tryptic peptides.
However, the bulky glycan group sterically hinders complete cleavage of glycopeptides, rendering a few amino acids around the glycosylation site protected. Since the protein has been identified, these small peptides are sufficient to pinpoint the glycosylation site. The sample is then further purified by successive desalting on micro-columns with reversed phase and graphite, respectively. Elution from the graphite column allows the analysis of the glycopeptides by MS. Using this strategy, the authors identified 13 different glycans from two N-glycosylation sites of ovalbumin and characterized the N-glycosylation of two contaminating proteins of commercial ovalbumin.
The basic structure of a GPI group comprises a phosphoethanolamine, a core tetrasaccharide consisting of three mannose and one glucosamine and the phosphatidylinositol group with the diacylglycerol that provides the membrane anchorage. There exist a variety of modifications to this core structure including alterations of the fatty acid groups and addition of carbohydrates or ethanolamine Menon, Modification of proteins with GPI takes place in the lumen of the endoplasmic reticulum ER , where a GPI transamidase cleaves at the carboxyterminus of proteins and catalyses an amide bond formation between the phosphoethanolamine of the GPI anchor and the new C-terminal amino acid.
GPI-AP can be selectively released from the membrane by the action of the GPI specific phospholipase C that cleaves between the phosphoinositol and the diacylglycerol. The lipid moiety of the GPI anchor is firmly inserted into one leaflet of the membrane and confers the characteristics of an integral membrane protein to the GPI-AP, such as resistance to alkaline extraction by Na2CO3.
Due to their hydrophobicity, GPI-APs and other membrane proteins will accumulate in the detergent phase during a partitioning experiment with the detergent Triton X Treatment of this fraction with phospholipase C will cleave the diacylglycerol and release the GPI-APs from the membranes, so that they will be extracted into the soluble phase in a subsequent partitioning experiment. This combination of Triton X partitioning and phospholipase C treatment has been used by several groups to isolate and identify GPI-APs in proteomic studies Fivaz et al.
Unexpectedly, two of these proteins contained hydrophilic or charged residues in their hydrophobic termini. This information was used to modify the search algorithm, which led to the prediction of nine novel candidate GPI-APs from Arabidopsis. Phospholipase C specific proteins were excised from the gel and identified by automated LC-MS analysis. Bioinformatic analysis of these proteins by three different search algorithms categorized all of these proteins as putative GPI-APs by at least one of the programs.
The proteomic approach applied in these two cases proved very efficient for isolating and identifying a large number of GPI-APs. However, proof of modification of the detected GPI-proteins was acquired indirectly by phospholipase C sensitivity and prediction of a GPI attachment sequence motif. For none of these proteins was the cleavage site or the structure of the GPI moiety determined by MS, although this is feasible when microgram amounts of GPI-protein are available Omaetxebarria et al.
Another study focused on short arabinogalactan AG peptides from Arabidopsis, a subgroup of the AG family of proteins, which are plasma membrane anchored or part of the extracellular matrix in plants and have been implicated functions in plant growth, embryogenesis and apoptosis. The mature part of the proteins is rich in hydroxyproline, alanine, serine and threonine, and contains large AG oligosaccharide chains attached to hydroxyproline and additional shorter glycans attached either to hydroxyproline or serine. Precipitated AGPs were chemically deglycosylated with anhydrous hydrogen fluoride, which cleaves between the phosphoethanolamine and the carbohydrate moiety of the GPI anchor.
We are also pleased to announce that both Polish country representatives of INPPO found their place in the society board. Further information will soon be available on the Society's website ptp. Weckwerth W Green systems biology - From single genomes, proteomes and metabolomes to ecosystems research and biotechnology. J Proteomics This article explains the foundations of Green Systems Biology and presents an iterative workflow to investigate the genotype-phenotype-map in plant ecology, evolution and biotechnology.
The socio-economic relevance of these research fields is further discussed. Thank you for your cooperation and support. Contact Abhijit Sarkar at contactinppo gmail. This special issue reports on much of the progress and reflects the current situation of plant proteomics at least in Europe. Et cetera.
Why study plants?