Paola Picotti obtained her PhD from the University of Padova, Italy. In 2007 she joined Ruedi Aebersold’s group at ETH Zuerich, where she worked as post-doctoral fellow. Her research focused on the development of targeted proteomic techniques, including selected reaction monitoring (SRM), and their application to the analysis of biological systems. She was also involved in the development of approaches and resources to promote the collection and dissemination of SRM assays and their extension to whole proteomes. In 2011 she started an independent research group at the Institute of Biochemistry of ETH Zuerich.

Selected Reaction Monitoring (SRM) is a targeted mass spectrometry technique which emerged in the field of proteomics as a complement to the untargeted shotgun methods. The main advantages of SRM become apparent when predetermined sets of proteins need to be measured across multiple samples in a consistent and accurate manner. The technology is however still in its infancy for protein analysis and several challenges need to be addressed to demonstrate the power of the technique in biological research and increase its widespread application in non-specialized laboratories.

To evaluate the applicability of SRM to studying biological networks, we applied it to the analysis of a metabolic network constituted by proteins covering a broad range of abundances and including a large number of families of isoenzymes, sharing high sequence overlap. Proteins in the network were quantified by SRM in yeast cells grown under a series of conditions inducing radically different metabolic setups and in a growth time-course of cells transiting through a series of metabolic phases. The quantitative dataset generated highlighted how yeast metabolism adapts to changing conditions of supply and demand of nutrients and suggested differential functionality for several isoenzymes. The application showed the potential of the technique to elucidate the dynamics of cellular networks through large number of perturbing conditions.

In order to expand the capabilities of the technique, we used an approach based on libraries of unpurified synthetic peptides to develop at high-throughput SRM assays for entire proteomes. We used the method to develop SRM assays for the ~6,000 proteins that constitute the proteome of S. cerevisiae and then expanded it to the generation of SRM assays for >90% of the human proteome. The synthetic peptide libraries were also used to generate gold-standard ion trap spectral libraries to be used for spectral matching of shotgun proteomic datasets in discovery-driven experiments. The described SRM assays and spectral libraries are currently being made publicly available via the SRMAtlas user interface, which supports their dissemination across different laboratoreis. The potential of such proteome maps, for targeted and discovery proteomics, as well as their limitations will be discussed.

Saccharomyces cerevisiae is a model organism for biological processes such as mitotic growth and meiotic development. Recent advances in genome biology have yielded data on yeast genomes, as well as the transcriptome, proteome and interactome. In the present work, we used a molecular systems biology approach by combining different high-throughput datasets, with the aim to gain new insight into growth and meiotic development processes. Duplicate total protein extracts from logarithmically growing diploid cells cultured in rich medium with glucose (YPD, fermentation) or acetate (YPA, respiration) were run onto a high resolution SDS-polyacrylamide gel, each lane was cut into 30 bands which were digested with Trypsin. Samples were then processed by shotgun LC-MS/MS analysis on a LTQ-Orbitrap XL instrument for protein identification. The resulting information was integrated with the output of our own whole-genome expression profiling experiments using tiling microarrays as well as DNA sequencing data and protein network data available via certified public repositories. We have determined the dynamic proteome of diploid yeast cells in two different metabolic states (fermentation of glucose and respiration of acetate). Our label-free method based on a simple protein extraction, and separation step identified essentially all (93%) of the known proteins present in vegetatively growing diploid SK1 budding yeast cells. The vast majority of the proteins were found in both replicates of each medium, underlining our method’s robustness. Small numbers of proteins were identified only in YPD or YPA, and several hundred were not detected in any of our experiments. The latter group contained proteins encoded by genes not expressed under the conditions we tested or by loci found to be deleted or mutated in SK1. Our data on the yeast SK1 proteome together with information on DNA and RNA yielded systems-level insight into vegetative growth under two different metabolic conditions and revealed its surprisingly distinct proteome as compared to the reference strain. This work paves the way for large-scale dynamic proteome profiling of complex processes such as meiotic development.

AICAR monophosphate (5-Aminoimidazole-4-carboxamide ribonucleoside-5’ monophosphate) is a metabolic intermediate of the de novo purine synthesis pathway, presenting highly promising metabolic and antiproliferative properties. One of its known activities is the induction of the AMP dependent kinase (AMPK), explaining many, but not all of its effects. The molecular mechanisms underlying the AMPK independent effects are poorly understood up to now. From yeast (S. cerevisiae) it is known that AICAR effects are multifactorial. This was demonstrated with genetic and metabolic analysis. To better tackle the complexity of the AICAR induced phenomena, we now applied another approach.

To systematically screen for cellular targets of AICAR, a proteomic approach was undertaken in yeast, using affinity chromatography and subsequent identification of the binding proteins by mass spectrometry. Like this 62 proteins (~1% of the proteome) were found in at least four of five experiments and defined as AICAR binding proteins. To test the specificity of the affinity chromatography, the same experiment was done for succinyl-AICAR (SAICAR). It occurs just one step before AICAR synthesis and is therefore chemically very close, differing only by an additional succinyl group. These experiments resulted in a list of 70 SAICAR binding proteins and strikingly only five are identical to the previously identified AICAR binding proteins. Interestingly the two different groups of binding proteins also belong to distinct functional groups of proteins. Our conclusion was that specificity of the proteomic approach used is acceptable and that a different action of the two molecules analyzed is expected. Detail studies of candidate proteins were thereafter undertaken in yeast, to further validate the approach.

We were furthermore interested in the identification of evolutionary conserved binding proteins, with the assumption in mind that this will allow us to reveal processes regulated by AICAR, conserved between species. We therefore proceeded with the identification of AICAR binding proteins in species from bacteria to man and have now a large overview over the most probable targets of this small molecule.

As a second approach to identify proteins affected by AICAR, a drug affinity responsive target sensitivity (DARTS) study is on the way now, applying mass spectrometric identification and label free quantification.

Bacillus cereus is a pathogenic agent causing human foodborne disease. Diarrheal syndrome may be the consequence of the production in the host’s small intestine of various extracellular factors, including enterotoxins. In human intestine, pathogens have to deal with a lack of carbohydrates, varying oxygen concentrations and oxydoreduction potential (ORP). While bacterial response to changes in oxygen level has been widely studied, adaptation to various POR has never been documented. We cultured B. cereus in a medium mimicking intestinal conditions under low and high-ORP anoxic conditions (pO2 = 0%) and in full oxic condition (pO2 = 100%). We compared (1) the early secreted and (2) the intracellular protein contents by means of a high-throughput label-free proteomic strategy. A total of 27 biological samples were in-depth analyzed following a system biology approach: record of high-throughput quantitative proteomic data, metabolism modeling, creation of mutant strains, high-throughput proteomics, and so on. After protein extraction and trypsinisation, nano-LC-MS/MS detection of the peptides resulted in a large body of MS/MS spectra acquired (above 1.5 millions). A total of 15,022 unique peptides were assigned to 1,373 proteins. We observed that most proteins identified in secretome are related to pathogenesis. Two-thirds have never been detected before. Changes in secreted protein levels depending on oxygen availability and ORP induce a modification of cytotoxicity towards Caco-2 human epithelial cells (Clair et al, 2010). Detailed study of intracellular proteome associated to metabolic data highlights that a global proteome remodeling occurs in low-ORP anaerobiosis compared to high-ORP anaerobiosis. Candidate proteins possibly involved in this remodeling were identified by analyzing proteomic data. Their major roles were then elucidated by (i) constructing strains lacking these proteins and (ii) analyzing the intracellular-proteome and secretome of the mutant strains (Clair et al, 2011).

References : Clair, Roussi, Armengaud, Duport (2010) Mol Cell Proteomics 9:1486-98. Clair, Armengaud, Duport (2011) submitted for publication

Proteasome is a multimeric protease complex, which degrades more than 80% of proteins in the cell. Due to its substrate wide range, proteasome is involved in the regulation of many important cellular processes (cell division, transcription, cell differentiation…). Proteasomes complexes consist of a 20S catalytic core particle comprising 14 to 17 different proteins assembled in four stacked ring-shape heptamers. This core complex can be associated to one or two regulatory particles of identical or different protein composition and can also recruit other so-called Proteasome Interacting Proteins (PIPs). Proteasome activity is tightly regulated by the nature of these associated proteins. Proteasome complexes are present in the cytoplasm (free, associated with cytoskeletal elements or bound to the endoplasmic reticulum) and in the nucleus but little is known about the distribution of the different associated regulators in each cell compartment.

To answer this question, we have developed an affinity purification-quantitative mass spectrometry strategy using a 20S proteasome core particle subunit as bait. This allows the enrichment of all combinations of 20S-regulator endogenous proteasome complexes in the cell.

To stabilize the labile association between the 20S core particle with its regulators and PIPs, we have performed in vivo cross-linking of protein-protein interactions using formaldehyde [1]. Formaldehyde was introduced very early in the proteomic workflow, so that real biological events could be frozen. Optimization of formaldehyde concentration was performed to maximize the specific activity of purified proteasomes, the yield of purification and the recovery of regulators and PIPs.

The fractionation protocol was also optimized to deal with the cross-linked cells and to obtain proteasome complexes specific of the cytoplasm, the microsome and the nucleus compartments.

A label-free quantitative proteomic workflow [2] was then performed to compare proteasome complexes diversity in each cellular compartment and in different leukemia cell lines exhibiting differences in proteasome activity.

Determining the composition in regulators of the proteasome, according to the cellular localization, will allow a better understanding of their functions.

1 Bousquet-Dubouch MP, Baudelet E, Guerin F, et al. Affinity purification strategy to capture human endogenous proteasome complexes diversity and to identify proteasome-interacting proteins. Mol Cell Proteomics 8, 1150-1164 (2009).

2 Mouton-Barbosa E, Roux-Dalvai F, Bouyssie D, et al. In-depth exploration of cerebrospinal fluid by combining peptide ligand library treatment and label-free protein quantification. Mol Cell Proteomics 9, 1006-1021 (2010).

Dr. Jennifer S. Brodbelt is a professor of chemistry at the University of Texas at Austin. She earned her B.S. degree in chemistry at the University of Virginia and her doctorate in chemistry at Purdue University under the supervision of Prof. Graham Cooks. After a post-doctoral position at the University of California at Santa Barbara, she began her academic career at the University of Texas in 1989. Her research interests focus on the development and application of photodissociation as an ion activation method for characterization of biological molecules, including peptides, proteins, nucleic acids, oligosaccharides, and lipids. Her group has also developed the use of mass spectrometry for evaluation of DNA/drug interactions.

The tremendous growth in the application of mass spectrometry for detection, quantification and characterization of biological molecules has spurred the exploration of new ion activation/dissociation methods. Although collision induced dissociation (CID) remains the gold standard for structural characterization of ions, it has several shortcomings (e.g. insufficient energy deposition, limited applicability for pinpointing post-translational modifications in peptides, etc.) that have stimulated the search for other activation methods. Photodissociation offers an especially promising alternative to traditional CID in ion traps, in addition to its success with FTICR and time-of-flight instruments.

There are several compelling advantages of using lasers to activate ions in a mass spectrometer. Photoactivation is a non-resonant process, meaning that both the selected precursor ions and primary fragment ions may be activated, leading to secondary dissociation and a richer array of fragment ions without requiring deliberate sequential stages of ion manipulation. The ability to vary energy deposition without resorting to multi-stage experiments (MSⁿ) is particularly important when analyzing macromolecules in which initial ion currents may be low and the total sample quantity is limited.

The reduction of ion losses is a critical feature when analyzing small ion populations while maintaining adequate detection sensitivity. Photoactivation can be utilized to analyze both negative and positive ions, so it offers the potential for broader characterization of the proteome. UV photodissociation using a laser offers fast, high energy deposition that yields good sequence coverage for peptide analysis.

L’obtention de données quantitatives à partir d’une expérience d’imagerie par spectrométrie de masse (MSI) reste à ce jour un challenge pour notre communauté scientifique. Cette quantification requière la mise au point d’une méthodologie particulière afin d’obtenir une meilleur reproductibilité et homogénéité des mesures (signal) au même titre qu’une courbe de calibration présentant un minimum de variabilité et une très bonne linéarité. Plusieurs méthodes ont été décrites dans la littérature mais aucune d’elles n’a d’applications universelles. Nous proposons ici deux approches à la fois différentes et complémentaires en vue de la quantification par imagerie MALDI-MS. Nous tenterons par ce fait de répondre aux principales difficultés à prendre en compte dans la démarche de quantification par MSI, à savoir, premièrement, la haute dépendance du signal détecté à la déposition/propriétés de la matrice MALDI ainsi qu’à sa capacité d’extraction ; deuxièmement, le rendement d’ionisation spécifique des molécules cibles ; et finalement, l’effet de suppression ionique du au tissue. Notre méthodologie a été appliquée dans le cadre de l’étude de la distribution de deux médicaments, le propranolol (Sigma-Aldrich) et le BDM31343, sur un modèle animal (souris) par imagerie MALDI-TOFMS (Autoflex Speed, Bruker). Les protocoles expérimentaux et les paramètres instrumentaux ont été adaptés à chacune des cibles. Les deux approches choisies pour cette étude consistent, pour la première, en l’utilisation d’un broyat d’un organe d’intérêt auquel est ajouté une gamme de concentration connue d’une molécule cible puis reconstitué par congélation et enfin analysé en même temps qu’un échantillon traité comportant le composé étudié. Cette première approche est particulièrement adaptée à l’étude de la distribution d’un composé cible dans un organe particulier, elle nécessite une plus moins longue préparation et un temps de retraitement des données minimum. La seconde approche, quant à elle, prend en compte l’effet de suppression ionique inhérent au tissu analysé par imagerie MALDI-MS en utilisant un facteur de normalisation, nommé TEC (Coefficient d’Extinction Tissulaire). Ce coefficient est calculé en comparant le signal d’une molécule cible sur et en dehors du tissu (support). Il permet de normaliser les données obtenues lors de l’analyse d’un échantillon traité par un médicament donné, la molécule cible précédemment citée. Cette approche a pour principaux atouts, un temps de préparation limité et d’être adapté à l’étude de plusieurs organes simultanément. Comme inconvénient, un traitement de donnés important est nécessaire. Afin de valider nos deux approches, les résultats issus de notre méthodologie ont été comparés à des études effectuées sur les mêmes échantillons par d’autres techniques analytiques (l’autoradiographie et la LC-MS/MS).

L’histologie permet de différencier simplement et rapidement sur un tissu biologique différentes structures (noyau, membrane) grâce à l’utilisation de colorants spécifiques. Par exemple, le Rouge Nil permet de colorer les lipides neutres. Néanmoins aucune information moléculaire détaillée n’est obtenue par cette technique. L’imagerie par spectrométrie de masse (ISM), de par la possibilité d’obtenir simultanément un grand nombre d’images, semble être une technique complémentaire à l’histologie. Jusqu’à maintenant, il était nécessaire d’utiliser des coupes adjacentes pour l’histologie et l’ISM afin d’établir des parallèles dans les résultats et d’en tirer des conclusions pertinentes. Pourtant, il est parfois difficile d’obtenir des coupes adjacentes adaptées pour les analyses. La possibilité de réaliser l’histologie et l’imagerie par SM sur la même coupe de tissu biologique a été l’objet de cette étude.

L’imagerie par spectrométrie de masse d’ions secondaires couplée à un analyseur à temps de vol (ToF-SIMS), qui ne nécessite aucun prétraitement de l’échantillon et qui n’est pas destructive, s’est révélée très adaptée pour cette étude. Des techniques de coloration de routine, telles que l’Hématoxyline-Eosine (HE) ou le Rouge Nil (NR) ont été appliquées sur des coupes de tissus biologiques avant ou après localisation de lipides par ToF-SIMS. Les coupes de 16 µm d’épaisseur ont été réalisées à l’aide d’un cryostat CM3050-S à -20°C, déposées sur des lames en verre puis séchées brièvement sous vide. Certaines de ces lames ont été analysées par ToF-SIMS puis colorées et analysées de nouveau par ToF-SIMS, d’autres ont été directement colorées puis analysées. Les images ioniques 500 µm x 500 µm (2 µm de résolution spatiale) ont été acquises avec un spectromètre de masse ToF-SIMS IV en modes d’ionisation positif et négatif.

•Dans le cas des coupes colorées puis analysées par ToF-SIMS, une délocalisation de certains ions tels que le cholestérol (m/z 369 ion fragment caractéristique en mode d’ionisation positif et m/z 385 en mode d’ionisation négatif) a été observée après coloration par HE et NR. Sur les coupes de cerveau de rat, les images ont montré que le cholestérol était toujours majoritaire au sein du corps calleux mais semble être anti-co-localisé après coloration. Au contraire, les images de ces mêmes ions après les lavages au tampon phosphate (une des étapes) ne montrent aucune délocalisation, et restent similaires aux images contrôles.

•De plus, les résultats préliminaires ont montré que l’analyse par ToF-SIMS n’empêche pas la coloration des coupes par la suite.

•Enfin, il a été possible de refaire une analyse de la surface par ToF-SIMS sur ces mêmes coupes pour évaluer la délocalisation et le rendement d’émission ionique secondaire.

En conclusion, il est possible d’obtenir les informations histologiques de coupes préalablement étudiées par spectrométrie de masse ToF-SIMS.

L’érythropoïétine humaine recombinante (rHuEPO) est une glycoprotéine de 30-34 kDa interdite d’utilisation par le Code des Courses. Cependant depuis quelques années, des cas de dopage avérés ont été observés chez le cheval aux USA et en Europe, mais aussi dernièrement chez le dromadaire. De nombreuses méthodes pour différencier l’EPO équine endogène de l’EPO humaine recombinante ont été développées aussi bien en LC-MS/MS qu’en western-blot (double-blotting, WADA). Toutefois, le temps de détection est relativement court (48h) et le screening réalisé par ELISA reste perfectible.

Afin d’améliorer ce temps de détection des rHuEPOs dans le plasma et l’urine équine, une colonne monolithe greffée avec des anticorps monoclonaux anti-EPO a été évaluée pour la préparation des échantillons. Cette nouvelle technologie, combinée avec l’affinité de l’EPO pour la lectine (MAIIA : Menbrane-Assisted Isoform ImmunoAssay) et analysée par un scanner haute résolution, a été utilisée pour le dépistage de l’EPO dans des échantillons plasmatiques et urinaires provenant de chevaux ayant reçu des traitements soit d’EPREX®, soit d’ARANESP®.

En parallèle, une méthode de purification basée exclusivement sur l’utilisation de la colonne monolithe anti-EPO et couplée à la spectrométrie de masse a été développée dans le but de confirmer la présence des rHuEPOs dans les différents milieux étudiés, via la détection de certains peptides caractéristiques. L’introduction de la mobilité ionique au travers du système FAIMS (high-Field Asymmetric Waveform Ion Mobility Spectrometry, ThermoFisherScientific) au couplage LC-MS/MS a permis d’augmenter la sélectivité des peptides et ainsi la sensibilité de notre méthode de détection.

Cette nouvelle combinaison analytique permet ainsi de confirmer la présence de rHuEPOs en un seul jour avec une limite de détection de l’ordre de 100 pg/mL. La méthodologie et les résultats obtenus seront présentés pour illustrer l’apport de ce nouveau développement pour le contrôle antidopage des chevaux.

Perdita Barran is a Senior Lecturer in Biophysical Chemistry at the University of Edinburgh. The Barran group has considerable experience in gas-phase ion chemistry, instrument development and the application of mass spectrometry to complex chemical and biological problems. Dr Barran has published over 50 papers with 595 citations to date, as well as being an invited/keynote speaker at 5 international conferences in the past two years. The Barran research group currently comprises of 7 graduate students, and 2 PDRA’s. Since Barran started as an independent researcher, 12 PhD students have successfully defended their PhD’s. Barran develops techniques to investigate changes in protein conformation that occur in response to chemical or physical intervention and relates them to their biological role. In 2009 in recognition of her achievements Barran was awarded the inaugural Joseph Black award by the RSC Analytical Division for her contribution to the developments of biological mass spectrometry.

The p53 protein is a transcription factor which plays a central role to tumour suppression. Loss of p53 function is correlated with the development of cancer in which the MDM2 protein binds to the N-terminal transcription activation domain shutting down function³. Mass spectrometry studies of WT p53 and mutants of the DNA binding domain following nano-electrospray from native conditions show a wide charge state range for charge states n= 8 - 18 for monomeric species of the general form [M+nH]ⁿ⁺. Mass spectrometry data from both instruments are qualitatively similar. The most abundant peaks are at n =9 and n=10 indicating the dominance of a compact structure. Multimeric species are observed for the WT and mutants, however the dimers shown in the WT are more intense. Generally, the collision cross-sections of the monomer are observed to increase with increasing charge in a stochastic fashion attributed to protein unfolding due principally to coulombic repulsion, which can be attributed to the ‘plasticity’ of these IDP’s . Multiple gas-phase conformers are resolved for a number of charge states, over a very wide charge state range. Subtle differences in unfolding transitions between the WT and mutated samples can be characterized. For instance the H115N mutant, altered at the boundary of loop I seems to be less structured – favoring the more extended conformations when compared with WT p53.,Conversely, the R249A/H115N mutant appears to be more compact. The thermally induced unfolding also shows interesting trends, and in particular reveals the resistance to unfolding of this important IDP. These results are interpreted in terms of the biological activity of these proteins and also in terms of the implications for the use of IM-MS to study this largely ignored, but critically important, class of proteins.

Selected Reaction Monitoring (SRM) carried out on triple quadrupole mass spectrometers coupled to liquid chromatography has been a reference method to develop quantitative analysis of small molecules in biological or environmental matrices for years and is currently emerging as a promising tool in clinical proteomics. However, sensitive assays in complex matrices are often hampered by the presence of co-eluted compounds that share redundant transitions with the target species. In the present work we document for the first time the potential interest of substituting classical gas-collision activation mode by laser photo-dissociation to track chromophore-derivatized target peptide absorbing at 532 nm diluted in a whole plasma trypsin hydrolysate. We anticipate that this technique coined photo-SRM, combined with chromophores endowed with reactivity towards generic chemical groups, might significantly improve the limit of quantification of classical SRM-based assays.

The present presentation focused on the development of a new technique dedicated to quantification, called photo-SRM, where non-discriminating collision-activated dissociation mode has been replaced by a more specific photo-dissociation process governed by the absorbing properties of a targeted compound. The first example photo-SRM experiments has been developed in-house on a conventional triple quadrupole mass spectrometer. Proof of principle of SRM-based monitoring by laser-induced dissociation instead of classical collision activated mode was carried out with the chromophore-derivatized-Oxytocin as a model molecule diluted in whole plasma trypsin hydrolysate. This preliminary investigation clearly demonstrates that both the selectivity and detection level may markedly be improved during SRM monitoring by replacing the classical mode of activation by gas collision by a photon excitation providing a judicious overlap between the absorbing properties of the target molecule and the excitation wavelength of the laser beam. The lack of cross contamination within the transition channel between the target molecule and co-eluted endogenous interferences results here in a 50-fold improvement of the limit of quantification. The question rises now about how photo-SRM might be further optimized and extended to molecules that do not exhibit natural or specific light absorption.

Novel aspect:

First demonstration of Photo-SRM tool for analytical quantification.

‘Matrix-free’ Laser Desorption/Ionization (LDI) mass spectrometry using specific inert surfaces to promote ion formation has been widely investigated the last decade.1 In addition to porous silicon through the original DIOS technique,2 different solid materials were tested as potent LDI-promoting agents, such as metals, carbon-based structures, porous particles, nanomaterials and more recently, ordered three-dimensional silicon architectures. Disposable ready-to-use Nano-Assisted Laser Desorption/Ionization (NALDITM) target was also developed for matrix-free LDI analysis of small molecules and peptides.3, 4

Following our recent development of a straightforward low cost LDI method based on the use of porous chromatography materials for the analysis of peptides5, we explored other inert silicon/silica-based materials. Taking into account that nanostructured silicon based materials were providing good detection sensitivity and that surface derivatization has a great influence on the ionization yield, 6 we choose to evaluate the efficiency of LDI-MS carried out on functionalized Silicon Nanowires for the analysis of a large variety of model peptides designed in our laboratory.

In this study, we focus on C18-Silicon Nanowires as matrix in LDI-MS for the analysis of small peptides mimicking proteolytic sequences (500-1700 Da). The detection of 30 peptides has been investigated on Nanowires prepared according to 3 different processes. We also realize a set of repeatability experiments on these peptides to assess the interspot and intraspot variations, representing about 800 MS analyses. Moreover, we characterize these peptides in MS/MS experiments and compare the fragmentation data with those obtained with sample conditioned in HCCA organic matrix. Finally, we carry out the analysis of various peptide mixtures in order to probe the potent spectral discrimination in LDI-MS and compare the results with the same experiment conducted with organic matrices in MALDI analyses. Through this set of experiments, we were able to assess the LDI performances in terms of sensitivity, repeatability and robustness of C18-Silicon Nanowires as ionizing surfaces within the framework of peptide analysis such as peptide mass fingerprinting in proteomics.

[1] Peteron DS. Mass Spectrom. Rev. 2007, 26: 19-34.

[2] Wei J, Buriak J, Siuzdak G. Nature 1999; 399: 243-246.

[3] Guénin E, Lecouvey M, Hardouin J. Rapid Commun. Mass Spectrom. 2009; 23, 1395-1400.

[4] Shenar N, Cantel S, Martinez J, Enjalbal C. Rapid Commun. Mass Spectrom. 2009; 23, 2371-2379.

[5] Shenar N, Martinez J, Enjalbal C. J.Am.Soc.Mass Spectrom. 2008; 19: 632-644.

[6] Piret, G. l.; Drobecq, H.; Coffinier, Y.; Melnyk, O.; Boukherroub, R. Langmuir 2009, 26 : 1354-1361.

Il est fondamental pour des études protéomiques de posséder des outils efficaces de fragmentation des peptides, afin d’accéder à leur séquence et à leurs éventuelles modifications post-traductionnelles (MPT). De nombreuses techniques cohabitent actuellement. La dissociation induite par collision couplée à une source electrospray permet d’obtenir de bons rendements de fragmentation mais n’autorise pas le positionnement de certaines MPTs. À l’inverse, le transfert électronique dissociatif (ETD) permet l’analyse de MPTs labiles mais les rendements de fragmentations sont très faibles. Le but de ce travail est de développer de nouveaux outils pour la fragmentation efficace de peptides modifiés ou non.

Le principe général du transfert électronique dissociatif à pression atmosphérique est basé sur l’utilisation de la photoionisation à pression atmosphérique (APPI). Le peptide en solution est tout d’abord transféré en phase gazeuse grâce à une source de type thermospray puis irradié par un faisceau intense de photons UV issu du synchrotron SOLEIL. Le rayonnement UV peut ioniser le solvant si l’énergie des photons est suffisante. L’avantage de ce couplage unique au monde est la possibilité de varier l’énergie des photons entre 5 et 20 eV, alors que les photons émis classiquement par une lampe à krypton ont des énergies fixes à 10,0 et 10,6 eV. Deux types d’expériences ont été menés : l’une consiste à faire varier la composition des solvants photoionisables, l’autre à additionner une molécule facilement photoionisable, appelée dopant.

La première série d’expérience a permis de démontrer grâce à des balayage en énergie que pour des photons d’énergie supérieure au potentiel d’ionisation de l’éthanol, de l’isopropanol ou du butanol, un transfert d’électrons entre les molécules de solvant vers les ions multichargés de la substance P est observé conduisant à la formation en source d’ions fragments de type c, b et y, et à la disparition totale des ions multichargés. Cette réaction peut être comparée à celle obtenue par ETD mais avec un rendement bien plus important. Dans le cas du méthanol, cette réaction est par contre peu efficace et seul un transfert de proton entre le solvant et la substance P est observé.

La seconde série d’expériences a permis de démontrer que des fragments intenses de type c, b et y sont aussi obtenus lorsque les photons possèdent une énergie supérieure au seuil d’ionisation du dopant. L’analyse d’un peptide myristoylé a permis de confirmer ces phénomènes et la MPT a bien été localisée en N-ter grâce à une série presque complète d’ions de type c. Finalement l’analyse d’un peptide phosphorylé a conduit à la formation préférentielle d’une série complète d’ions de type b-H3PO4. .

La suite de ce travail consistera à développer un système de nébulisation peu chauffé afin de minimiser les fragmentations thermiques successives tout en contrôlant les processus ETD à pression atmosphérique.

La spectrométrie de masse après ionisation electrospray (ESI) est devenue une des techniques analytiques les plus utilisées et consiste à former des ions intacts en phase gazeuse par électronébulisation d’un analyte en solution. Cependant, les mécanismes mis en jeu lors de cette ionisation sont relativement peu connus, en particulier ceux du processus de désolvatation au sein de la plume entre l'émetteur electrospray et le capillaire d’entrée du spectromètre de masse. Des mesures spectroscopiques peuvent aider à étudier les propriétés chimiques et physiques des gouttelettes de la plume electrospray ainsi que les propriétés des ions au sein de ces gouttelettes.

Un montage expérimental permettant de réaliser des mesures de fluorescence au sein de la plume electrospray a été développé, tandis que des spectres de masse d’analytes sont enregistrés simultanément. Des images 2D de la plume electrospray ont été obtenues en mesurant la concentration (c.-à-d. le signal de fluorescence) d'une solution de colorant rhodamine 6G (Rh6G) avec une résolution spatiale d’environ 200µm. Afin de sonder les propriétés physiques et chimiques des gouttelettes au sein de la plume ESI, des colorants chromiques ont été utilisés. L'intensité de fluorescence ou la longueur d'onde de différents colorants chromiques est corrélée avec les propriétés chimiques et physiques, en utilisant les courbes d'étalonnage établies. Les spectres d'émission du Nile Red (NR) dans un mélange d'acétonitrile/eau ont été enregistrés afin de déterminer la composition de solvant dans la plume pour différents paramètres de spray. Les résultats indiquent clairement que la composition du solvant dans les gouttelettes n'est pas homogène dans la plume ESI, en raison de l'évaporation des solvants. La fluorescence induite par laser a également été employée pour profiler des changements de pH lors de l’évaporation des gouttelettes dans la plume ESI en mesurant des spectres d'émission du colorant pHchromic C.SNARF-1. Ce couplage de mesures de spectrométrie de masse et de spectroscopie optique permet d'étudier la relation entre l'état de charge observé d'anions de peptide et les changements de pH au sein du spray en fonction des paramètres de source tels que le débit et la température du gaz de gainage qui jouent sur le rendement d’évaporation de gouttelettes.

Dr. Overall is a Professor and Tier 1 Canada Research Chair in Metalloproteinase Proteomics and Systems Biology at the University of British Columbia, Centre for Blood Research. He completed his undergraduate, honours Science and Masters degrees at the University of Adelaide, South Australia; his Ph.D. in Biochemistry at the University of Toronto; and was a MRC Centennial Fellow in his post-doctoral work with Dr Michael Smith Nobel Laureate, Biotechnology Laboratory. He won the Institute of Musculoskeletal Health and Arthritis CIHR Award as 2002 CIHR Scientist of the Year, the University of British Columbia Killam Senior Researcher Award (Science) 2005, and was the Chair of the 2003 MMP and the 2010 Protease Gordon Research Conferences. On Sabbatical in 1997-1998 he was a Visiting Scientist at British Biotech Pharmaceuticals, Oxford, UK and in 2004/2008 he was a Visiting Scientist at the Expert Protease Platform, Novartis Pharma, Basel, Switzerland. He is currently an External Senior Fellow, Freiburg Institute for Advanced Studies - FRIAS, Albert-Ludwigs-Universität Freiburg. With over 8000 citations for his 173 papers and with an h factor of 52, he is a highly influential scientist in the field. He is the pioneer of Degradomics, with 6 Nature Review Papers on this, protease genomics, drug target validation, MMP therapeutics, and substrate discovery and has developed numerous approaches to decipher the roles of proteases in vivo by elucidating the protease and substrate degradomes through quantitative proteomics and N- and C-terminome analysis in cell based systems and in animal models.

Protein termini are truncated by proteolysis, but the extent to which this molds the proteome in vivo is unknown. In addition to constitutive proteolysis during protein synthesis and maturation, the processing of a mature protein often irreversibly changes its activity. Specific degradomics techniques are needed to rapidly identify and quantify the N- and C-terminomes in order to reveal the extent of post-translational modifications of protein termini and therefore the functional state of key molecules, the extent of proteolysis in a system, and to identify new protease substrates. We have devised new approaches to enrich for the N and C termini of proteins for high throughput terminome analyses of human tissues and cells for the Human Proteome project, in particular for chromosome 21 that is Canada’s focus for the HPP. Broad coverage N-terminome analysis necessitates a negative selection procedure as the variety of original mature protein N-terminal blocked peptides each present individual chemical hurdles for their enrichment by positive selection strategies.

We developed a combined N-terminomics and C-terminomics and protease substrate discovery degradomics platforms for the simultaneous quantitative analysis of the N-terminome and proteolysis on a proteome-wide scale called Terminal Amino Isotopic Labelling of Substrates (TAILS, Kleifeld et al Nature Biotech 28, 281-288; Prudova et al 2010 Mol Cell Proteomics; auf dem Keller et al 2010 Mol Cell Proteomics) and C-TAILS (Schilling et al Nature Methods 2010). By using novel polymers to deplete the internal tryptic peptides, TAILS suffers little from sample loss and low yields, so requiring only 100 microgram of sample and one MS/MS analysis per sample. By a three-day procedure with flexible labelling options, TAILS can be adapted to a variety of experimental situations including cell culture and complex biological sample analysis. Incorporating iTRAQ labelling iTRAQ-TAILS also provides wide coverage of all forms of naturally blocked N-terminal peptides and allows for their quantification through labelling of lysine side-chains in up to 8 samples. In addition to providing valuable proteome annotation this has several unique advantages. TAILS permits exploitation of the acetylated and other blocked mature protein N-terminal peptides as a statistical classifier that is then used to set isotope ratio cut offs that reveal protease activity.

We introduce a novel parameter evaluating ion intensity dependent quantification confidences of single peptide quantifications. Being a quantitative procedure, TAILS can analyse the substrate degradome of a broad specificity protease or one with no known specificity without manual data parsing, in the same experiment, and also do this in vivo. We have applied TAILS to a variety of proteases and compared protease knock out mice in models of arthritis, skin inflammation and models of breast cancer metastasis and pancreatic carcinoma in the RIP-Tag model. Typical analyses identify over 3000 N-terminal peptides from which we found that the removal of the N-terminal methionine is dependent upon the amino acid at position 2 with distinct preferences found for valine, glycine, alanine and serine. In one experiment, acetylation occurred on 731 original mature protein N-terminal peptides but at the initiator methionine in only 153 of these instances. In 578 cases, acetylation was at position 2 in the protein after removal of ¹Met, with alanine, serine and methionine being the preferred acetylated residues. Internal acetylation sites exhibit a distinct acetylation pattern that differs from the N-terminal acetylation. Finally N-terminal positional proteomics enables MS sample simplification with proteins identified in bronchoalvelar fluid and serum having abundances spanning a range greater than six orders of magnitude. This underscores the potential of TAILS to tackle the dynamic range analysis problem in complex proteomes and as a high throughput approach annotate the N and C termini in the HPP. 

The substrate repertoire of human tolloid-like proteinases (also known as BMP-1, mTLD, mTLL-1 and mTLL-2) has considerably expanded in the past 10 years to include several extracellular matrix components, growth factors and angiogenic factors. These extracellular metalloproteases are now thought to synchronize the biosynthesis of the extracellular matrix with the activation of several growth factors and appear as key players during embryogenesis and tissue repair. However, our current understanding of the physio-pathological roles of BMP-1/tolloid-like proteases is limited by the lack of high-throughput approaches to identify novel substrates and/or determine the relevant subset of substrates that are cleaved in specific biological contexts. Quantitative proteomics applied to cell-based assays and tissues is at the moment one of the most powerful techniques to address these questions.

Protease-generated fragments, released from the cell-surface or the extracellular matrix, are expected to be found in higher amounts in the supernatants of cells expressing active protease compared to cells expressing inactive protease and this difference can be monitored using iTRAQ™ labelling of tryptic peptides and quantification with LC-MS/MS. This technique has allowed the identification of several candidate substrates of BMP-1 which have then been studied by western blotting and classical biochemical techniques. However, secreted substrates that do not change cellular compartment after cleavage are missed by this technique and we are now switching to the more exhaustive TAILS technique (Terminal Amine Isotopic Labelling of Substrates) which relies on pre-labelling of whole proteins prior to fragmentation followed by a step of negative selection to enrich the sample in N-terminal peptides (1). Finally, iTRAQ™ labelling of whole proteins is also used in vitro to identify cleavage sites in complex extracellular substrates bearing glycosaminoglycan chains, as an off-gel alternative to Edman sequencing.

(1) Kleifeld, O.; Doucet, A.; auf dem, K. U.; Prudova, A.; Schilling, O.; Kainthan, R. K.; Starr, A. E.; Foster, L. J.; Kizhakkedathu, J. N.; Overall, C. M. Isotopic labeling of terminal amines in complex samples identifies protein N-termini and protease cleavage products. Nat. Biotechnol. 2010, 28 (3), 281-288.

Many bioinformatic tools developed in recent years for the quantification of MS data in label-free experiments use pattern-based methods and generation of LC-MS maps. Here, we applied the MFPaQ software, which uses an identity-based extraction approach, starting from peptide identification results to go back in the MS scans in order to retrieve peptide intensity values. In such label-free quantitative studies performed on complex protein samples, a compromise has to be found between reproducibility and high proteomic analytical coverage. The latter is often achieved through sample fractionation by 1D SDS-PAGE, which may induce experimental bias during the label-free comparison of samples processed and analyzed independently. We thus evaluated the efficiency of our label-free workflow with or without protein fractionation by 1D SDS-PAGE.

To this aim, a whole-cell lysate of primary human endothelial cells was analyzed by nanoLC-MS/MS, either in one analytical run, or after fractionation into 12 gel bands. Samples were run on an Orbitrap-Velos instrument with high sequencing speed in order to improve MS/MS sampling and analytical coverage. Technical gel replicates or LC-MS replicates were performed to evaluate variability and systematic errors due to upstream sample processing steps (electrophoretic migration, trypsin digestion, peptide extraction) or to final analytical steps (LC-MS measurement and bioinformatic extraction of peptide XICs by the software). Data normalization and integration procedures were used in MFPaQ to correct LC-MS variability and errors related to non-reproducible electrophoretic migration of proteins in the case of sample fractionation. Our results show that for one-shot analysis, label-free quantification can be achieved with MFPaQ on about 700 proteins with good accuracy (median CV of 5%, 99% proteins with CVs<48%). Sample fractionation largely improved the depth of proteomic coverage (about 3600 proteins quantified), and this was associated with a moderate decrease of quantitative measurement reproducibility (median CV of 7%, 99% proteins with CVs<62%).

The method was applied for the large-scale quantitative analysis of primary human endothelial cells stimulated by proinflammatory cytokines such as TNFa, INFß and IL1ß. It allowed us to identify and quantify about 5000 unique proteins, providing an in-depth analysis of the endothelial cell proteome and a detailed characterization of the proteomic variations associated with the inflammatory response.

Post translational modifications (PTM’s) are increasingly being revealed as key intermediates in pathogenesis pathways1-3. Understanding their role on the molecular level not only greatly increases our comprehension of disease mechanisms but provides an essential basis onto which human intervention strategies can eventually be built. Detection and localisation of PTM is not an easy task and mass spectrometry, which is a fast and sensitive analytical tool, is often implicated somewhere in the process.

In recent work we identified an important PTM on the type IV pili of Neisseria meningitidis. Pili are extracellular, filamentous organelles implicated in a variety of life processes including bacterial aggregation and host cell attachment. The pilus itself is a macro polymer built up of the repeating 17.5kDa protein unit pilE, which is arranged helically to create long and flexible fibres.

The glycerophosphate modification, present on serine 93 of this protein, is induced in vivo after several hours of host cell contact. We hypothesise that subsequent alteration of the pilus surface, once modified by the phosphoester, ultimately leads to the dissemination of the bacterium; a step that forcibly precedes invasive infection4.

In this work a combination of top-down and bottom-up methods were used for structural characterisation of pilE and to localise the modifications of interest. However, an effective top-down approach for localisation of all PTM’s had at that time not been developed.

Here we present top-down results using both Orbitrap and FT-ICR mass spectrometers and a variety of fragmentation methods, ECD, ETD, CAD, HCAD, IRPMD and combinations thereof, for the localisation of the 5 modifications present on this protein. Unusually we also present comprehensive profiling of several of these fragmentation techniques performed on the pilE protein itself.

Aspects of the techniques, fragmentation mechanisms and applicability to other systems will be discussed along with the major implications of a “one shot” approach to understanding bacterial dissemination.

1S. Subramaniam et al., Science, 324, 1327-1330 (2009)

2A. Oueslati et al., Progress in Brain Research, 183, 115-145 (2010)

3J.K. Kimet al., Molecular & Cellular Proteomics, 2, 7, 453-62 (2003)

4J.Chamot-Rooke et al., Science, 331, 778-782 (2011)

The plasma membrane serotonin transporter (SERT) plays a critical role in the regulation of serotonergic transmission by enabling serotonin reuptake into the cells. This transporter is of major pharmacological and clinical interest, particularly as it represents one of the primary targets of several widely prescribed antidepressants. It is now well established that SERT does not function as an isolated protein. SERT functional activity is regulated by a combination of multiple mechanisms including both post-translational modifications and association with intracellular proteins. During the last decade, several SERT-interacting proteins have been identified, principally by means of yeast two-hybrid screens. We have recently used a proteomic approach that enabled us to characterize a reciprocal modulation of SERT and neuronal NO Synthase (nNOS) activity mediated by their physical interaction. To get further insight into SERT-associated protein complex, we used high-resolution mass spectrometry to identify novel proteins interacting with SERT C-terminus, or whole SERT protein expressed in two different cell culture models. This shotgun analysis of SERT interactome led us to identify several new partners of SERT. These include Calcineurin, a calcium-dependent serine/threonine phosphatase, ASCT2 (Alanine Serine Cysteine Transporter 2), a neutral amino acid transporter, VELI-3, a PDZ domain-containing protein which regulates plasma membrane distribution of several receptors and transporters, and a set of proteins related to the SNARE complex, possibly involved in SERT export to the plasma membrane. Moreover, we showed that both physical interaction of SERT with Calcineurin and Calcineurin phosphatase activity increase SERT plasma membrane expression and 5HT uptake via SERT. In addition, co-expression of VELI-3 with SERT likewise increases 5HT uptake, whereas co-expression of ASCT2 decreases SERT activity, possibly via modification of its glycosylation status. Collectively, these proteomic studies identify novel regulation mechanisms of SERT activity that might influence serotonergic transmission.

Christian Rolando is a senior scientist (Directeur de Recherche) at the CNRS (French National Center for Scientific Research) and works at the University of Lille. Christian Rolando is the head of the proteomics platform he founded in the early 2000. Christian Rolando has worked in mass spectrometry since his PhD, during which he managed the mass spectrometry facility at the Department of Chemistry of the Ecole Normale Supérieure in Paris. At this time the facility was equipped with two Varian MAT magnetic instruments (CH7 and 111) fitted with UV recorders… During the 90s Christian Rolando worked with the French company Nermag and participated in the development of the R3030 triple quadripole. He used this instrument to study organometallic ion-molecule reactions and he developed several analytical protocols combining organic reactions and mass spectrometry like the so-called thioacidolysis, which became the standard in lignin analysis. Since his arrival in Lille in the early 2000s, he developed a set of original analytical methodologies for proteomics, from sample preparation (new gels for electrophoresis, recipes based on pure chemicals avoiding the use of kits of unknown formulation) to data analysis (bioinformatics). He is also involved in several collaborations on different topics such as transfusion plasma safety, biomarkers of cigarette smoking or identification of proteins in cultural heritage artifacts. A part of Christian Rolando's research is also focused on the analytical applications of microfluidics and more recently on the development of two-dimensional FT-ICR MS. Christian Rolando is the co-author of 150 papers cited more than 2500 times and has directed 30 PhD theses. He was President of the Société Française de Spectrométrie de masse from 1994 to 1995 and he is now President of the Analytical Division of the French Chemical society and Deputy Director at the CNRS Institute of Chemistry in charge of Infrastructures.

The pulse sequence for 2D FT-ICR MS was first developed in 1988 [1] by MS and NMR teams respectively lead by Tino Gäumann and Geoffrey Bodenhausen [1] based on a key experiment demonstrating cyclotron excitation reversibility [2] and conveys detailed information for compounds in complex samples in one easily readable 2D mass spectrum without ion isolation. We revisited 2D FT-ICR MS [3] which, unlike 2D NMR, has no analytical applications yet. Gas-free fragmentation modes allow us to retain high resolution and sensitivity, and the advances in electronics and computer technology since 1988 enable broadband acquisition. The method, however, needs to be optimized in terms of scintillation noise, experiment time and necessary amounts of sample. In this study, we apply techniques adapted from NMR spectroscopy to optimize 2D FT-ICR MS.

Because in both dimensions the spectrum is obtained after Fourier transforming a signal that has been sampled at regular time intervals (in the vertical dimension, a delay, and in the horizontal dimension, a measurement date), the resolution of 2D mass spectra behaves in both dimensions like the resolution in one-dimensional FTMS: it increases with the number of data points and is inversely proportional to m/z ratio. In preliminary experiments, time transients were limited to 32k data points since the 32-bit-written software cannot handle files larger than 256 Mb. Using 2048 scans leads to hour-long experiments with a resolution of ~1000 in the MS and fragment ion MS/MS dimension but only a few hundred in the correlation parent ion dimension [3]. For example, we were able to distinguish the fragments of the ⁷⁹Br and ⁸¹Br isotopes and the ¹²C and ¹³C isotopes of singly charged ions of bromopride, a small brominated drug (molecular weight 344.25 g.mol^-1). The new 64-bit version of the NPK package developed by Marc-André Delsuc [4] and NMRTEC allows us to work on files larger than 16 Gb. An experiment on intact cytochrome C (molecular weight 12 kDalton) using 8192 scans of 512k points time transients shows line base separation for each multicharged ion bearing around 10 charges (ca 10 000 resolution) and quadrupole like (1 m/z unit) separation in the correlation dimension (ca 1000 resolution). Furthermore these experiments show that the resolution in both dimensions remains inversely proportional to m/z ratio, as expected.

All of our experiments were performed using nanoESI at concentrations of 1 pmol/µL, leading to sample consumption of less than 20 pmol per hour per experiment, but at the expense of signal intensity fluctuation which translates into scintillation noise in the 2D data. We adapted the NMR version of the Cadzow de-noising algorithm [5] for 2D FT-ICR mass spectra. This algorithm is based on the decomposition of the signal in singular values from which the signal at longer time may be predicted. The Cadzow noise reduction dramatically reduced fluctuations and brought out additional fragmentation peaks that were masked by the noise [6].

Results obtained using IRMPD and ECD fragmentation without any in-cell or quadrupole separation for simple compounds like individual peptides, the multicharged ion distribution of small proteins and more complex mixtures like tryptic protein digests will be presented. Finally future development of 2D FT-ICR will be briefly discussed at the sight of multidimensional NMR methodologies developed during the last twenty years. 

[1] P. Pfändler et al., J. Am. Chem. Soc. 110 (1988) 5625-5628

[2] A.G. Marshall et al., Chem. Phys. Letts. 105 (1984) 233-236

[3] M.A. van Agthoven et al., Int. J. Mass Spectrom., doi:10.1016/j.ijms.2010.10.034

[4] D. Tramesel et al., J. Magn. Res. 188 (2007) 56-67

[5] C. Brissac et al., J. Biomol. NMR 6 (1995) 361-365 

[6] M.A. van Agthoven et al., Rapid Commun. Mass Spectrom., 25 (2011) 1609–1616.

Le séquençage des peptides linéaires par spectrométrie de masse se fait conventionnellement en mode CID. Cette technique permet la dissociation des ions par apport d’énergie vibrationnelle impliquant des fragmentations induites par la charge. Dans le cas de l’ECD, la dissociation est causée par capture d’un électron lent menant à d’autres types de dissociations. Cette technique implique des mécanismes qui ne sont pas parfaitement connus notamment pour les peptides cycliques qui nécessitent deux fragmentations consécutives, ce qui rend l’attribution très complexe. Dans la littérature, le modèle de la cascade radicalaire est généralement considéré pour expliquer la formation de fragments à partir de peptide cycliques1.

Au cours d’un travail préliminaire effectué sur des peptides lasso, il est apparu que la formation d’ions b’•j en ECD est due a priori à des processus consécutifs induits par la charge [2]. Une première rupture liée à la capture d’un électron ouvre le cycle (clivage c/z) et une deuxième rupture induite par la charge conduit à la formation des ions de la série b’•j.

Au cours de ce travail la dissociation par capture d’électron (ECD) de peptides cycliques tels que la cyclosporine A, a été comparée aux dissociations induites par collisions (CID). Cette étude effectuée sur un ESI-FT-ICR (Fourier Transform Ion Cyclotron Resonance) a permis, grâce à la haute résolution en masse de cet appareil, d’analyser finement la composition des différents fragments produits par ECD. Nous avons montré que dans le cas de la cyclosporine, en ECD, il existe en plus des fragmentations induites par le radical (cascade radicalaire), une voie de fragmentation compétitive induite par la charge.

[1] Leymarie, N.; Costello, C. E.; O'Connor, P. B. Electron capture dissociation initiates a free radical reaction cascade. Journal of the American Chemical Society 2003, 125, 8949-8958.

[2] Zirah, S.; Afonso, C.; Linne, U.; Knappe, T. A.; Marahiel, M. A.; Rebuffat, S.; Tabet, J.-C. Topoisomer Differentiation of Molecular Knots by FTICR MS: Lessons from Class II Lasso Peptides. J. Am. Soc. Mass Spectrom. 2011, 22, 467-479.

As introduced by Roman Zubarev et al. in 1998, ECD has been described as leading mostly to c and z fragment types [1]. The observation of other fragment types (a, b, y, w) has also been described in the literature, but often as either secondary fragmentation of the major c/z ions, or as secondary or minor processes. Recent experimental work in our group on a pentapeptide series (AGXLK - X=A, D, E, S, W) of ions, as well as analysis of a database of 11 000 ECD spectra from doubly charged tryptic peptides [2] revealed that these pathways can represent the major ones for small (less than 9 residues) size peptides. Therefore, far from being an epiphenomenon, observed only for some very specific peptides, these fragmentation pathways can truly be observed in the ECD fragmentation of peptides present in bottom-up proteomics studies.

The database analysis not only revealed the extent of these atypical fragmentation behaviors, but also led to question assumptions on the fragmentation pathways leading to these ions. For instance, the w fragmentation pathway has usually [3] been ascribed to a secondary fragmentation of z ions, through a side-chain loss. Our database analysis shows that the w ions do not arise from the major z ions, therefore leading to the question of the mechanism accounting for their formation. No alternative mechanism is present in the literature to this date. A similar question arose for b ions formed by ECD, for which at least three pathways are proposed in the literature.

The structure of these fragments could provide hints on the mechanistic pathways followed to reach these products. In light of the questions raised above, infra-red action spectra of some w and b ECD fragments were recorded at the CLIO facility in Orsay. Briefly, ions of interest are produced from ECD fragmentation of selected precursors in the cell of an FT-ICR mass spectrometer and irradiated by a tunable infra-red free-electron laser beam. IRMPD fragmentation efficiency is recorded as a function of wavelength. Theoretical reference spectra for series of possible structures were produced through DFT geometry optimizations and frequency calculations and compared with the experimental spectra. For b ions, the IR action spectra of the corresponding CID fragments were also recorded and used for comparison.

This work led us to obtain the first IR spectra of w type ions, which seem to have the expected linear ketene structure. For b ions, some ions display a different experimental spectrum for the ECD fragments compared to the CID fragments, which would point out to a different formation mechanism.

[1] Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W., J Am Chem Soc 1998, 120, 3265.

[2] Van der Rest, G.; Hui, R.; Frison, G.; Chamot-Rooke, J., J. Am. Soc. Mass Spectrom. 2011, in press.

[3]Savitski, M. M.; Nielsen, M. L.; Zubarev, R. A., Anal. Chem. 2007, 79, 2296.

La fumée de cigarette est un contaminant majeur de l’air intérieur (indoor). De nombreux constituants de ces fumées, essentiellement gazeux, ont été identifiés. Certains présentent des propriétés toxiques, carcinogènes et/ou mutagènes. Outre ces espèces gazeuses, les fumées de cigarette contiennent une fraction particulaire importante. Les agences sanitaires et de santé sont actuellement préoccupées par l’examen de ces particules. Celles-ci possèdent non seulement des temps de séjour élevés dans l’organisme mais aussi une composition complexe et variée et sont le siège de l’adsorption d’un grand nombre de molécules issues de la combustion de la cigarette. A la différence de la pollution outdoor pour laquelle des normes existent en termes de contaminants particulaires (PM10 voire PM2.5), en atmosphère confinée aucune législation n’est encore proposée. Dans l’objectif d’accéder à la composition détaillée des particules de fumées de cigarettes, nous nous sommes intéressés plus particulièrement à la caractérisation et la comparaison des fumées inhalées par le fumeur (MSS), celles émanant de la cigarette entre deux bouffées (SSS), et les fumées expirées par le fumeur (EXS), ces deux dernières fractions étant celles auxquelles est exposé l’individu dans le cadre du tabagisme passif.

Après avoir validé en relation avec le laboratoire ASCAL les méthodes de prélèvements par le suivi de traceurs spécifiques, les différents types de particules sont analysées par désorption/ionisation laser couplée à la spectrométrie de masse FTICR. La haute résolution et la précision sur la mesure du rapport m/z permettent l’attribution de façon non ambigüe d’une formule brute aux 1000 à 2000 signaux observés dans la gamme m/z 150–500. L’emploi de représentations spécifiques (carte de Kendrick et diagramme de Van Krevelen) assiste la comparaison des différents types de fumées.

Les résultats montrent que la nature et la composition des MSS sont significativement différentes de celles des SSS pour lesquelles des composés à la fois moins oxygénés et moins saturés sont observés [1]. L’examen des EXS révèle quant à lui peu de différences de composition entre les deux individus soumis à l’étude mais une différence importante lorsqu’elles sont comparées aux MSS. De manière plus spécifique, il est constaté un enrichissement en composés aromatiques tels que les HAP et une diminution de la quantité des composés les plus polaires. Des échanges entre molécules adsorbées à la surface des particules et milieu physiologique seraient à même d’expliquer ce comportement.

[1] Schramm, S.; Carré, V.; Scheffler, J.-L.; Aubriet, F. Anal. Chem. 2011, 83, 133–142.

Several electrospray-mass spectrometry (ESI-MS)-based methods are available for determining the equilibrium association constant (Ka) between a protein and a small ligand, but current MS-based strategies are inadequate for measuring Ka of protein-protein interactions with accuracy. We expanded the application of ESI-MS-based titration to determine the strength of noncovalent interactions between proteins, forming a complex. Taking into account relative response factors (probability of being ionised, transmitted and detected), we determined Ka values of an equilibrium between dimers and tetramers at three different pH values (6.8, 3.4 and 8.4). We investigated the association of the lectin concanavalin A, whose dimer-tetramer equilibrium is affected by in-solution concentration and by pH. To calculate the constants of association (Kd) in solution, we also utilised isothermal titration calorimetry (ITC) for a comparison with MS-based titration method. At pH= 6.8 and pH= 8.4 the Ka values measured by MS and by ITC were in agreement. At pH= 3.4 we were able to measure the Ka only by MS, but not by ITC due to limited sensitivity of calorimetry.

Our investigation illustrates the great potential MS for calculating the binding strength of protein-protein interactions within noncovalent complexes. The main advantages of MS over ITC are its sensitivity (i.e. the required amount of sample is >100 times less than the one necessary for ITC) and the possibility to obtain precise information on composition of protein complexes, their stoichiometry, their subunit interactions and their assembly pathway. Moreover, our study shows the strong influence of response factors on determining accurate protein-protein association constants.

Since the 1980s Bruno Domon has applied novel techniques to characterize proteins and glycoproteins in academia and in the pharmaceutical/biotech industry. At MIT he pioneered, with Catherine E. Costello, the MS/MS fragmentation of glycoconjugates, and developed a systematic nomenclature. He worked in the industry (1988 – 2000), heading the mass spectrometry facility at Ciba in Basel and then at Biogen in Cambridge, USA. He went on to head the large-scale proteomics facility at Celera Genomics in Rockville, MD (2001-2004), then joined the group of Ruedi Aebersold at ETH Zurich (2004-2009), where he worked, as group leader, to develop alternate proteomics workflows and targeted approaches based on selected reaction monitoring, expanding this technology to encompass large-scale screening.

In 2010 Domon assumed leadership of the new Luxembourg Clinical Proteomics Center (LCP), and now heads the effort to develop high-throughput quantitative mass spectrometry methods to qualify and biomarkers in bodily fluids, with the goal of translating them into routine clinical assays. LCP is part of a wider effort in Luxembourg to push the envelope of translational biomedical research.

Protein biomarkers discovery of renal or urinary tract diseases in human urine has recently garnered tremendous interest. This fluid represents an ideal diagnostic analyte for bladder cancer, since the tumor is literally bathed in urine. The EU-FP7 funded DECanBio project aims at implementing a generic strategy for protein biomarker discovery and validation. Using “state of the art” MS instrumentation for quantitative proteins analysis, the primary objective is to provide a restricted number of urinary biomarkers; ultimately enabling the detection of bladder tumour recurrences. As a first step toward this goal, a label-free quantitative proteomics approach was used to compare urine protein profiles among a multi-site cohort consisting of 99 patients including various grades of bladder cancer, and age and sex matched controls.

A label-free quantitative proteomics approach based on the Accurate Mass and Time (AMT) tag strategy has been pursued. This strategy allows the analysis of the many samples (tens to hundreds) required at the candidate discovery phase in a reasonable time-frame (a few weeks) while providing a broad proteome coverage and a suitable dynamic range of peptides quantification. A multi-site cohort (France and Spain) was collected according to guidelines developed within the context of the DECanBio project. Proteins were extracted using organic solvent precipitation and digested in-solution as described earlier (Court et al. Proteomics, in press). Samples were analyzed in triplicate and in random order on a Ultimate 3000 nano-HPLC system (Dionex) coupled to an LTQ-Orbitrap XL mass spectrometer.

An AMT tag database was generated from >1200 LC-MS/MS analyses of bladder cancer incident, recurrent and control cases. This database constitutes an in-depth repository of all peptides previously detected in urine by our group and their corresponding proteins. It contains over 18,000 non-redundant peptide entries, representing more than 2,000 urinary proteins. In semi-quantitative LC-MS analyses, over 1,200 proteins were monitored and quantified over the discovery cohort. The peptides detected spanned a relative abundance range of more than 4 orders of magnitude. To extract lists of differentially abundant proteins between the cohort’s sub-populations, two different statistical analysis methods were applied: a modified version of the so-called Spectral Index (SpI) approach and a mixed effects ANOVA. The comparison of healthy versus cancer samples by the two methods resulted in a restricted list of proteins common to both approaches. These proteins are being reviewed using bioinformatics approaches to determine the final list of candidates, which will be evaluated on a different cohort during the subsequent phase of the DECanBio project using absolute quantification by LC-SRM.

N-terminal modifications play a major role for the fate of proteins in terms of activity, stability, or localisation [1, 2]. Such modifications remain poorly described and badly characterised in proteomic studies. Recent advances in the field allow now to specifically enrich and select N-terminal peptides in the course of proteome-wide MS analyses. These targeted approaches unravel as a result the extent and nature of the protein N-terminal modifications. Here, we aimed at studying such modifications in two projects using a simple and robust proteomic workflow targeted on N-terminal peptides. To challenge our early approach and identify putative bottlenecks, we compared the patterns of N-terminal modifications of the model plant Arabidopsis thaliana with those of human cell lines. Our data show strong convergence for some of the characterised modifications between the two organisms including N-terminal Methionine Excision (NME). NME is an essential and initial pathway of co-translational protein maturation_ENREF_5 [3]. Although proteins N-terminus acetylation show some similarities for substrates specificity in general, a large proportions of plant organelle-targeted proteins also feature post-translational N-a acetylation of the mature protein after removal of the transit peptide. This is a marked difference between plant and animal Kingdoms. As we identified that we also needed to recover unmodified protein N-terminus our enrichment technics was further with chemical acetylation with tri-deuterated acetate of free N-termini. This workflow works with either plant, animal, organellar or bacterial proteins. This strategy was successfully applied to investigate the influence of fumagillin as anticancer drug. Fumagillin is a potent inhibitor of METAP2, an enzyme in charge together with METAP1 of NME. After protein extraction and N-terminus peptides enrichments, we could assess the global impact of inhibiting METAP2 on NME in several cell lines. The data will be presented here and discussed.

References

1. Martinez et al. (2008) Proteomics 8, 2809-31. 

2. Meinnel & Giglione (2008) Proteomics 8, 626-49. 

3. Giglione et al. (2004) Cell. Mol. Life Sci. 61, 1455-74 

Keywords: proteolysis, N-terminal modifications, acetylation,

Background: Recently, proteomics delivered its’ first clinical application in routine with identification of microorganisms using MALDI-TOF. However this technology, based on protein fingerprints, provides only a probability of identification and is not well suited for antibiotic susceptibility testing (AST) or virulence detection. To overcome MALDI-TOF limitations, we propose to use a triple-quadrupole mass spectrometer working in Specific Reaction Monitoring (SRM) mode, coupled with a conventional bore chromatography.

Methods : Multiplexed SRM methods enabling an absolute identification of the microorganism as well as the determination of mechanisms of resistance, virulence and strain typing have been developed in our lab. After tryptic digestion either a colony or crude sample, is injected through a short chromatographic gradient (24 min) into a triple quadrupole mass spectrometer working in SRM mode.

Results: As an example, methicillin-resistant Staphylococcus aureus (MRSA) strains could be fully characterized in 24 minutes using the multiplexed detection of 36 peptides from 14 proteins, including Penicillin-Binding Protein 2a (PBP-2a) and Panton-Valentine Leucocidin (PVL) toxin. Candida albicans yeast pathogenicity could be characterized, as well, by quantification of Ergosterol metabolite for resistance prediction and Lipase-8 protein for virulence estimation.

Conclusions: Multiplexed and quantitative capabilities of SRM offer a unique possibility to develop targeted methods for characterization of microorganisms either after culture or directly on crude sample. The power of this state-of-the-art approach is the development speed, which one can adapt methods to capture the rapid microorganism evolution in biological mechanisms. This technology is expected to bring new proteomics applications into the clinic in the near future.

L’anthrax, maladie infectieuse aigüe, est lié à la bactérie Bacillus anthracis (Ba). Cette bactérie génère une structure résistante appelée « spore » qui constitue la forme de dissémination de la maladie. Le CDC (Centers for Disease Control) assigne Ba comme agent de classe A, le niveau de risque le plus élevé. L’utilisation des spores de Ba comme arme biologique a été soulignée lors de l’envoi de lettres contaminées aux Etats-Unis en 2001, qui causa 22 cas dont 5 décès.

Il est donc indispensable de détecter la présence de spores de Ba dans des échantillons environnementaux potentiellement contaminés. Une méthode efficace doit être à la fois sensible, spécifique, reproductible et rapide. La protéomique et la MS sont des approches pertinentes pour répondre à cette problématique.

La MS a ici été utilisée pour deux applications : la détection sensible des spores de Ba en mode ciblé de type MRM et la découverte de nouveaux biomarqueurs spécifiques. Nous avons tout d’abord développé un procédé analytique original de type immuno-LC/MS/MS adapté à la détection des spores de Ba. Il combine la spécificité et la sensibilité de deux techniques complémentaires : immunocapture et MS en mode ciblé. L’immunocapture est réalisée directement sur les spores intactes. Elle permet de les extraire spécifiquement de matrices environnementales complexes et entraine un gain en sensibilité. La seconde étape consiste à détecter en mode MRM la protéine SASP-B (Small Acid-soluble Spore protein B). Cette protéine est exprimée exclusivement dans les spores et en quantité abondante. Un isoforme particulier a été préalablement décrit pour les spores de Ba et permet ainsi de les discriminer des autres bactéries du groupe cereus phylogénétiquement très proches. La limite de détection (LOD) de notre approche est de 10^3 CFU/mL dans le lait et 7.10^3 spores détectées dans 10 mg de terre. Ces LOD sont au niveau de la dose infectieuse minimum (ID50 : 10^3-5.10^4) et proches de celles obtenues par PCR.

Cependant le suivi d’un seul biomarqueur peut présenter un biais en terme de spécificité. Nous avons ainsi cherché en parallèle à identifier de nouveaux marqueurs discriminant Ba des autres bactéries du groupe cereus, ceci afin de diversifier les protéines ciblées et ainsi augmenter encore la spécificité de détection. Dans ce but, des extraits de spores issus de 38 souches (10 Ba versus 28 B. cereus et B. thuringiensis) ont été analysés par MS haute résolution sur un LTQ-Orbitrap. Sept protéines présentant des mutations spécifiques de Ba ont été mises en évidence pour la première fois. Le suivi de ces marqueurs, en plus de la SASP-B, par MRM multiplex permet de détecter très spécifiquement les spores de Ba.

En conclusion, ces travaux présentent une utilisation originale de la MS pour la détection sensible et spécifique des spores de Ba. Ce procédé pourra être étendu à la détection d’autres bactéries de la menace.

My research interest concerns the functional interactions of plant cell organelles during development and differentiation, with an emphasis on chloroplast functions. We are using the tools of functional genomics, proteomics and systems biology to deal with the complexity of molecular responses and to unravel currently unknown dependencies in metabolic and regulatory networks. During our work we have established organellar protein inventories and we now use this information to analyze quantitative proteome and phosphoproteome dynamics under a variety of different conditions. Our research is currently focused around three major topics: i. Integration of plastids in cellular Ca²⁺ signaling, ii. Regulation and control of plastid protein import, and iii. Topology and dynamics of chloroplast phosphorylation networks.

Different functional proteomics tools are now available that enable the quantitative characterization of proteome dynamics and the mapping of posttranslational modifications. We report here examples how we used these tools to characterize the functional proteome of Arabidopsis and rice cell organelles, with a focus on plant-specific plastids. We analyzed the Arabidopsis proteome at genome-scale and provide quantitative information about organellar proteomes in different plant organs by “normalized spectral counting” (Baerenfaller et al., Science 320, 938-41; Baerenfaller et al., Integrative Biology 3, 225-237). For a functional characterization of plastid protein import, we analyzed the proteomes of protein import mutants and established protein N-termini to distinguish precursor from mature plastid proteins in WT, ppi1 and ppi2. These analyses revealed the accumulation of precursor proteins in the cytosol of protein import mutants. In order to assess the short-term regulation of the chloroplast proteome in response to environmental signals, we analyzed the chloroplast phosphoproteome and characterized its dynamics during a circadian cycle (Reiland et al., Plant Physiol. 150, 889-903). Phosphorylation motif utilization suggests that the phosphorylation network topology in chloroplasts is characterized by many-to-many relationships. To establish the kinase/substrate nodes in this network we performed comparative quantitative phosphoproteome profiling experiments with wildtype and kinase mutant plant material, starting out with the STN8 kinase. Differential protein phosphorylation was assessed by relative quantification with “extracted ion chromatograms” and the data were further validated by functional characterization of selected substrates (Reiland et al., Proc. Natl. Acad. Sci. USA, in press). We present here our data, comment on reliability and reproducibility and propose strategies to increase both at a reasonable cost.

Plants are exposed to various and changing environmental conditions. With respect to the availability of nutrients, a well-established and nearly general response is the adaptation of the expression level of those transporters ensuring the uptake of the corresponding nutrient. A large part of them is located at the plasma membrane and a general feature is that they belong to multigene families with high homology between members. However, the high level of homology may hamper precise identification of isoforms and prevent their quantification. For the same reason, immunological approaches are often not effective presently for a number of isoforms.

Beside shotgun strategies, however, dedicated approaches focusing on specific peptides and fragments are also emerging, but little used for sets of membranes proteins and never to date for plant membrane proteins from multigene families.

In this work, we set up a targeted Multiple Reaction Monitoring (MRM) proteomics approach, based on the use of proteotypic peptides and isotopically labeled analogs, for the simultaneous identification and quantification of transporter isoforms in Arabidopsis plasma membrane. Focus was given to four families of transporters, the proton pumps of the AHA (Autoinhibited H+ATPases) family, the water channels of the PIP (Plasma membrane Intrinsic Proteins) sub-family of aquaporins, the ammonium transporters of the AMT1 sub-family and members of the NRT1 and NRT2 groups of nitrate transporters.

In the strategy used, proteotypic peptides were selected firstly according to: their unicity in the whole Arabidopsis genome, their predicted potential ion-current response, their chromatographic behavior and their amino acid composition to avoid post-translational modification. Then, corresponding heavy forms of the 28 selected peptides were used for technical optimization (declustering potential, DP; collision energy, CE) using a stepwise procedure and LC-MS [1] and for quantification of the corresponding light version in biological samples.

This workflow was shown to allow the analysis of isoforms in the fourth families, including as well major membrane proteins as transporters of lower abundance. Similarly, it was used to investigate the effect of a salt stress on the expression of these latter 20 transporters in roots and identify the responding isoforms.

Such a resource (the peptides, their transitions and associated optimized detection conditions, as well as the synthetic peptides them-self) presents the advantage to be reusable and progressively expandable, thus opening the possibility of quantitative and isoform-specific investigation of the transportome in the model plant Arabidopsis.

[1] Lange, V., Picotti, P., Domon, B., Aebersold, R., Selected reaction monitoring for quantitative proteomics : a tutorial. Mol. Syst. Biol. 2008, 4, 222.

The serotonin 5-HT2A receptor is a primary target of psychedelic hallucinogens such as lysergic acid diethylamide (LSD), which reproduce some of the core symptoms of schizophrenia. An incompletely resolved paradox is that only some 5-HT2A receptor agonists exhibit hallucinogenic activity, whereas structurally related agonists with comparable affinity and activity do not.

Using quantitative phosphoproteomics combining SILAC, phosphopeptide enrichment by hydrophilic interaction chromatography /immobilized metal affinity chromatography and FT mass spectrometry, we compared the phosphoproteome in cells transiently expressing the 5-HT2A receptor under three conditions: no-stimulation; exposure to the hallucinogen DOI and exposure to the non-hallucinogenic agonist lisuride.

Among the 5,996 identified phosphopeptides, 454 sites were differentially phosphorylated upon exposure to DOI vs. lisuride. These include a serine phosphorylated upon exposure to DOI but not to lisuride and located in the i3 loop, a region important for 5HT2A desensitization. Mass spectrometry analysis of immunopurified receptor further confirmed such a differential phosphorylation. Correspondingly, exposure to hallucinogens induced a less pronounced receptor desensitization than exposure to non-hallucinogenic agonists.

In conclusion, this phosphoproteomic analysis reveals that 5-HT2A receptor stimulation by hallucinogenic vs. non hallucinogenic agonists induces contrasting phosphorylation patterns that may be related to their distinct behavioural responses. It also provides the first demonstration of differential phosphorylation of a G protein-coupled receptor upon stimulation by “biased” agonists.

The tyrosine kinase receptor (RTK) TrkA, which is activated by NGF (nerve growth factor) binding, plays a major role in differentiation and survival of both peripheral and central nervous system neurons. When PC12 cells, a model paradigm for these functions, are stimulated with NGF, they extend neurites and acquire characteristics similar to sympathetic neurons. After stimulation, TrkA undergoes tyrosine autophosphorylation creating docking sites that link several effectors and adaptors to the activation of downstream signaling pathways, which are characterized by a second, longer lasting, wave of protein phosphorylation targeted to serine/threonine residues. The Y490 and Y785 of the receptor play particularly important roles in these events and stimulate the activation of three main pathways: ERK1/2 via Ras, phosphatidyl inositol - 3- kinase (PI3K) and PLC?. However, a much broader array of phosphorylations is known to occur in both resting and stimulated cells. The aim of this study is to determine the profile of serine/threonine phosphorylation targets after stimulation of the TrkA receptor for 20’ and to characterize the precise involvement of Y490 and Y785 in these processes.

To avoid the stimulation of the second receptor of NGF (p75) and endogenous TrkA receptors, we used chimeric receptors composed of the ectodomain of the platelet-derived growth factor (PDGF) receptor and the transmembrane and endodomain of the TrkA receptor (denoted PTR), stably transfected into PC12 cells. The cells, grown in a medium containing heavy/light lysine/arginine (SILAC), were stimulated for 20 minutes with PDGF and lysed. The tryptic digest of proteins was enriched for phosphopeptides by a TiO2 column, further fractionated by strong cation exchange chromatography and analyzed by LC-MS/MS. The analysis of the phosphoproteome in cells expressing PTR showed quantitative changes in some previously known entities in signaling pathways, proving the specificity of the method, as well as many new modified entities. Interestingly, a comparison with a data set from HeLa (human) cells stimulated for the same time period by EGF (Olsen et al, 2006), showed a similar profile of protein kinase activation as judged by substrate specificity motifs. Analyses of cells expressing a PTR Y490F mutant and a PTR Y490/785F double mutant, stimulated under the same conditions, revealed three major classes of modifications: those dependent on Y490, those dependent on Y785 and those not dependent on either. This last group could be activated through the phosphorylation of one or more of the five other endodomain tyrosines of TrkA known to be phosphorylated but not previously identified as a docking/effector binding site(s). These activations are presumably not required for neurite proliferation but are likely involved in other functions stimulated by NGF.

This work was supported by NIH NCRR grant P41 RR001614.

Protein identification and characterization, has become one of the central activities in proteomics. The biological activity of many proteins is regulated by the extent and positions of posttranslational modifications (PTMs). More than 200 different PTMs have been described but only a minor fraction of those has been studied in detail. Indeed, most of these PTMs are lacking appropriate analytical methods that allow the sensitive, residue-resolved detection and localization of the modifications.

It appears that sulfation of tyrosine is a PTM occuring almost exclusively on secreted and trans-membrane spanning proteins. Evidence suggests up to 1% of all tyrosine residues of the total protein content in an organism can be sulfated.

Mass spectrometry has been developed as a key technology for protein modification analysis and many studies on sulfopeptides are MALDI-based. It has been reported that the sulfo-moiety from sulfopeptides is readily and quantitatively lost in positive ion mode. Alternatively, MALDI ion negative mode has proven to be useful for analysis of sulfopeptides as the lability of deprotonated anionic sulfomoieties is lower than protonated neutral ones.

Here we describe a peptide/small molecule non covalent interaction system allowing detection of mono- and polysulfated peptides in reflectron/linear positive mode by matrix-assisted desorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS). We will highlight strengths and limitations of this strategy and phospho-peptides detection approach will be discussed.

Prof. Yury O. Tsybin received his PhD degree in ion physics in 2004 from Uppsala University, Sweden performed under the supervision of Prof. Per Hakansson. For the next 2 years Tsybin was a postdoctoral research associate with Prof. Alan G. Marshall at the National High Magnetic Field Laboratory in the USA. Since 2006 Tsybin is a tenure-track assistant professor of physical and bioanalytical chemistry at the Ecole Polytechnique Fédérale de Lausanne (EPFL) in Switzerland where he heads the Biomolecular Mass Spectrometry Laboratory and serves as a Director of the MS Service Facility. Research interests of his group are around the high performance Fourier transform mass spectrometry method and technique development with subsequent applications in peptide and protein structure analysis. In 2011 Tsybin received the ERC Starting Grant to develop the super-resolution mass spectrometry for the applications in health and sustainability areas.

High magnetic field Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) provides the unbeatable analytical performance in terms of high resolution measurements. However, the high resolution measurements require a long experimental time, e.g., 1-20 seconds per single mass spectrum acquisition, and thus are incompatible with the current and near-future requirements for the high throughput and fast data acquisition, enforced by the short peptide and protein elution time from the chromatographic column. Here, we will discuss some of the possible solutions to this problem from both, hardware and signal processing development. The FT-ICR MS hardware development follows the implementation of ultra-high, e.g., 21 T, magnetic field environments together with higher acquisition speed and harmonized ICR ion traps. The recent progress in Orbitrap FTMS development has significantly reduced the gap between Orbitrap and ICR FTMS in terms of obtained resolving power required for the mainstream MS applications, 60-200k. Finally, the state-of-the-art time-of-flight (TOF) mass analyzers have demonstrated a substantial progress in the recent years and now offer 30-60k resolving powers achieved at comparable times to the high field ICR and Orbitrap FTMS and faster. On the other hand, incredible progress in the computational power and high frequency electronics opens the doors to the advanced signal processing development, which received a particular attention in the recent years. A number of groups, including ours, are in the process of tailoring the super-resolution signal processing methods, e.g., filter diagonalization method (FDM), to the needs of the ICR and Orbitrap MS. The goal is to replace the FT-based signal processing with the methods that would require 10-100 times shorter transient time-domain signals to yield a similar level of the resolving power.

Les grenouilles d’Amérique latine synthétisent et sécrètent au niveau de leur peau, des peptides biologiquement actifs. Dans cette étude, nous nous intéressons aux peptides issus de la peau de la rainette, Pachymedusa dacnicolor, afin d’isoler et caractériser de nouveaux peptides d’intérêt thérapeutique. Pour ce faire, nous travaillons sur des exsudats de peau. Les peptides contenus dans ces exsudats sont séparés par tamis moléculaire suivi d’HPLC. Les fractions obtenues sont testées pour leur activité biologique puis 4 fractions actives ont été analysées par spectrométrie de masse. Dans ces fractions, 6 peptides ont été identifiés et fragmentés par MALDI-TOF/TOF (séquençage de Novo). Trois d’entre eux avaient déjà été publiés (DMS-DA5, DMS-DA6 et DMS-DA8). Les trois autres peptides coexistent dans la quatrième fraction et possèdent une structure primaire N-terminale identique (résidus 1 à 29). Cette séquence commune, qui avait été déduite à partir de l'ADN de grenouille en 1998 est la suivante : ALWKTLLKKVGKVAGKAVLNAVTNMANQN-COOH. Les deux autres peptides de cette fraction possèdent ces 29 résidus +E et +EQ en C-terminal. Le peptide DMS-DA6 a été synthétisé sous forme amidée ou non en C-terminal. L’activité antimicrobienne des deux peptides a été testée sur deux souches de bactéries (Gram+ et -) et leur structure secondaire déterminée par dichroïsme circulaire.

Tous ces peptides de peau de grenouille contiennent généralement une grande quantité de résidus K, il a donc été nécessaire de recourir à une réaction d’acétylation afin de lever l’ambigüité Q/K qui existe du fait de la limitation dans la précision de mesure de masse de l’instrumentation MALDI-TOF (=20 ppm pour les peptides). L’acét! ylation des 4 fractions HPLC nous a permis de lever les ambiguïtés K/Q et nous a montré que cette réaction peut avoir lieu sur les fonctions amine primaires (N-terminal et résidus K) mais aussi, dans une moindre mesure, sur la fonction amine secondaire des résidus H. L’acétylation complète des peptides conduit à la disparition des sites protonables (si aucun résidu R dans la séquence). Ceci entraine la disparition de la molécule monoprotonée [M+H]+ au profit des espèces [M+Na]+ et [M+K]+. Dans le cas de la présence d’un résidu P dans la séquence de ces peptides, on observe une fragmentation source, préférentielle en N-terminal du résidu P. Ces peptides acétylés constituent de bons modèles pour aborder le phénomène d’ionisation en MALDI.

Glycosyl-Inositol-Phospho-Ceramides (GIPC) are the main sphingolipids of plant tissues, but the function of this complex family of compounds remains largely unknown. Therefore, the structural characterization of GIPC represents a critical step toward the understanding of their physiological function. In this work, GIPCs have been purified from plants cells (A. thaliana and N. tabacum cv. Bright Yellow 2, i.e. BY2), yielding multiple molecular species that have been analyzed by tandem mass spectrometry, using a combination of MALDI- and ESI-MS/MS. This method should open new avenues to decipher structure–function relationships between glycosphingolipids and membrane organisation.

GIPC extracts were analyzed by MALDI–MS/MS in negative ion mode, using 2,6-dihydroxy-acetophenone (DHA) as a matrix (MALDI Q-ToF Premier, Waters). ESI-MS/MS analyses were performed with a 5500 QTRAP (AB Sciex) instrument.

Analyses performed by MALDI mass spectrometry provide a quick overview of the variety of GIPC structures found in a given type of plant cell. A comparison between several organisms is straightforward, using MS/MS to confirm the identifications. The ceramide part of identified GIPCs merely contains tri-hydroxylated long chain bases (t18:0 and t18:1) to which alpha-hydroxylated very long chain fatty acids are amidified, whereas the polar head part displays more diverse structures.

MALDI mass spectra of A.thaliana and BY-2 cell samples revealed the occurrence of several clusters of GIPCs that can be grouped in six series (A to F) which differ by their mass range. For A.thaliana and BY-2 cells, each series was composed by several GIPC species differing by 2, 14 and 16 Da, corresponding to different degrees of unsaturation, carbon number or hydroxylation of the fatty acid chain. Fatty acid chains spent from 20:0 to 27:0, and from h20 to h26 for hydroxylated species. Mass differences between series were accounted by the number of saccharide units, and were either 162 Da (hexose) or 132 Da (pentose). Thus, MALDI-MS revealed that these plant GIPC structures had an increasing number of saccharide units, from two (series A) to seven (series F). The major compounds belong to series A for both A.thaliana and BY-2 cells. This result is in agreement with the previous report of hexose(R1)-hexuronic acid-inositol-phosphoceramide as major species in A. thaliana and N. tabacum leaves [1], where R1 is a hydroxyl group and an amine or an acetylamine, respectively. Further ESI-MS/MS experiments were conducted to confirm the structure identifications. Fragmentations observed in MALDI-MS/MS (singly charged ions) and ESI-MS/MS (doubly charged ions) were complementary for in-depth structure characterization of individual GIPCs among A-F series.

Finally, up to 51 and 121 molecular species could be detected in A. thaliana and tobacco BY-2 cells, respectively.

[1] J.E. Markham et al. J. Biol. Chem. 2006, 281, 22684.

Acetochlor, Alachlor and Metolachlor are widely used herbicides frequently detected in river waters. Known as potential endocrine disruptors, their toxicity for human has been clearly established. The characterization of their ozonation and photolysis by-products has been the subjects of recent investigation since some of them could also been responsible for endocrine disruption. The structural elucidation of by-products is usually performed using LC-MS/MS and GC-MSn analysis for polar and little or not polar compounds, respectively. In GC-MS, the use of chemical ionization (CI) is expected to provide abundant MH+ ions allowing direct access to the molecular weight of the analyte while electron ionization (EI) provide abundant fragment ions but only trace amounts of M+. ions for the herbicides of interest. We have been recently puzzled by the strange behavior of Metolachlor which, unlike the other mentioned pesticides, provides a very abundant monochlorinated ion at m/z 295/297 under ammonia positive chemical ionization using an hybrid ion trap mass spectrometer in GC-MS coupling. This ion observed only at trace amounts when using methanol as reagent gas. It disappears when attempting to isolate it to perform MSn experiments. Curiously, this ion at m/z = M+12 is not observed for the herbicides Acetochlor and Alachlor which present very similar chemical structures. The chemical structures of the m/z 295 and m/z 297 ions and the explanation of the observed phenomenon were elucidated on the basis of experiments including isotopic labeling and modifications of the operating conditions of the ion trap mass spectrometer. This work allows to give new recommendations for an optimized use of hybrid source ion trap mass spectrometers.

Contrary to conventional polymerization techniques, plasma-polymerization at atmospheric pressure leads to thin, poorly soluble and highly cross-linked films, with huge potential for industrial applications. Their microstructure, affected by plasma process parameters and the nature of the precursors, is thus quite difficult to assess. To overcome these difficulties, a complete analytical strategy is proposed – combining NMR spectroscopy, organic synthesis and mass spectrometry – leading to a detailed molecular and structural characterization of some plasma-polymers.

The targeted samples are plasma-polymerized films deposited by means of Atmospheric Pressure plasma generated by Dielectric Barrier Discharge. Two siloxane precursors were plasma-polymerized, either cyclic (octamethylcyclotetrasiloxane, D4) or linear (hexamethyldisiloxane, HMDSO), to synthesize ppD4 and ppHMDSO deposits, respectively.

NMR spectroscopy offered a global vision of the deposits as well as an insight in structural features of their insoluble parts by the detection of well-known M, D, T or Q silicon functions for instance.

Based on the tandem mass spectrometric behavior of linear methyl-terminated PDMS, molecular and structural information was obtained for the electrosprayed ppD4 soluble part. Distributions of peaks spaced by 282 Da were detected and seen as [OSi(CH3)2]2(OSiCH3)2 cycles bound via either an oxygen atom (A, major species), a –O–CH2– group (B) or a linear –O–Si(CH3)2– backbone (C). The expected dimer of the distribution (A) as well as some potential oligomers of the (C) series were then synthesized by conventional chemistry and submitted to CID, so as to compare the MS/MS patterns of plasma and synthesized species and validate the structural elucidation.

The ESI-MS spectrum of ppHMDSO revealed at least eight oligomeric distributions, further noted Cn, with n=0-7 indicating the degree of cross-linkage. MS/MS spectra of the oligomers from a given Cn distribution showed n major product ion series. C0 oligomers exhibit MS/MS spectra quite similar to those obtained from linear CH3-PDMS with only one product ion series. In contrast, interpretation of CID data of more cross-linked oligomer adducts required alternative models, amongst commercial T-branched PDMS was shown to be promising for the C1 and C2 series. In addition, several new home-made cyclolinear PDMS would constitute promising models for the highest cross-linked distributions.

As a perspective, an aminolysis reaction allowed a entire dissolution of a homemade highly cross-linked silicon rubber, leading to hydroxy-terminated and cyclic siloxanes. Applied to plasma-polymers, it would be an elegant breakthrough for the complete structural characterization of plasma-polymers, allowing the “insoluble” part to be also mass-analyzed by ESI-MS.

Prof. Oliver Kohlbacher holds a Diplom in chemistry from Saarland University in Saarbrücken, where he also obtained a PhD in computer science. In 2000 he became a junior group leader in Saarbrücken working on protein-protein docking. After a postdoc at Celera Genomics (Rockville, MD, USA) he became full professor for bioinformatics at the University of Tübingen, where he is also a director of the center for bioinformatics and speaker of the department of computer science. He has been working on computational methods for the analysis of high-throughput data (proteomics, metabolomics, next-generation sequencing). He is one of the initiators of the OpenMS software package for computational mass spectrometry. Additional research interests are in the area of computer-aided drug design, computational immunology and systems biology.

Over the last decade HPLC-MS has become the workhorse in proteomics and metabolomics. Increasing sensitivity and resolution of the instruments give us unprecedented coverage of the proteomes and metabolomes under investigation. The flip side of this development is the amount and complexity of data produced. Similar to what we observe in next-generation sequencing, bioinformatics analysis of a high-throughput experiment is rapidly becoming the bottleneck.

In this talk, I will illustrate how wet-lab workflows have to be matched by appropriate data analysis workflows in order to enable more complex experiments. Based on our software package OpenMS/TOPP, I will illustrate how analysis workflows can be set up, adapted to specific experimental designs for different quantification strategies. The resulting analyses can then be run automatically. Automated workflows also enable more compute-intensive methods, as computer power scales much better than manual labor done by PhD students and postdocs. I will give a few examples, how these more complex analyses can result in a drastic increase in the coverage and accuracy of the experiment.

Contexte

Les progrès instrumentaux en spectrométrie de masse (MS) de ces 20 dernières années ont conduit au développement d’instruments générant des données MS/MS de plus en plus volumineuses (du fait d’une grande rapidité d’acquisition des spectres de fragmentation).

Par ailleurs, la soumission des résultats d’identification de protéines à partir de ces données MS/MS est de plus en plus réglementée par les journaux du domaine qui recommandent l’utilisation d’algorithmes transparents (open-source) et multiples si possible.

Dans ce contexte, afin de répondre au besoin croissant de puissance de calcul nécessaire à l’interprétation des données MS/MS, une suite logicielle bâtie sur des logiciels libres a été adaptée et améliorée sur une grille de calcul.

Méthodes et Résultats

La suite logicielle développée à l’IPHC permet :

- de créer, d’extraire, de concaténer, de formater des banques de séquences protéiques, notamment à partir des banques de séquences publiques accessibles telles que NCBInr, UniProtKB, UniProtKB/SwissProt, …

- de lancer des requêtes OMSSA (Open Mass Spectrometry Search Algorithm) pour l’identification de protéines à partir de données MS/MS sur la grille de calcul locale de l’IPHC (1024 cœurs de calculs, site TIER-2 de la grille LHC (Large Hadron Collider)) et mondiale.

- d’interpréter à haut débit des données MS/MS acquises sur des organismes non séquencés par séquençage de novo suivi de recherches d’homologies de séquences.

- d’extraire de manière automatisée les annotations fonctionnelles disponibles sur les protéines identifiées.

L’ensemble de ces outils est disponible sur le site https://msda.u-strasbg.fr.

Conclusion

L’adaptation de la suite logicielle sur une grille de calcul permet de répondre aux importants besoins de puissance de calcul non accessibles à ce jour dans les laboratoires de protéomique. En effet, selon le type de requête (nombre de spectres MS/MS, taille de la banque de séquences, recherche de modifications post-traductionnelles, séquençage de novo, …), un gain d’un facteur 100, voire 1000 est obtenu en routine grâce à la parallélisation des outils et à leur lancement sur une grille de calcul.

Many large scale phosphorylation studies ([1][2][3]) have been performed yielding thousands of identified phospho-petides and their positions of phosphorylated amino acids. These studies provide new insights into the paramount importance of phosphorylation as a regulator of biochemical processes. While the statistical false discovery rate (FDR) of the peptide identifications can be controlled by using decoy databases, the error in the assignment of the phosphorylation positions is rarely reported and usually only the position with the highest score is reported.

Recently, we introduced QuickMod, a tool to identify modified variants of MS/MS library spectra [4]. We showed that the spectra of unmodified peptides bear considerable similarity to their modified counterparts. Even though the peak intensities can be strongly altered, the fragmentation sites are quite conserved between modified and unmodified peptides. QuickMod implements a spectrum alignment and modification positioning algorithm which uses the information from the unmodified library spectra. It calculates and reports a score for all possible positions of the modifications on the peptide as well as a statistical significance for all these positions.

Using publicly available data from large scale phosphorylation studies we extract those phospho-peptides, which have a matching unmodified library spectrum of the same charge and fragmentation technique. We investigate the agreement of the phospho-site positions reported in these data and the results from our spectrum library search for those phospho-peptides. An initial study using a dataset of 1000 doubly and triply charged phosphor-peptide spectra showed that phospho-site assignments are often ambiguous when several serine and threonine residues are present in close proximity. Here, we investigate the accuracy of phospho-sites in more detail and on a larger dataset. We present some simple rules indicating which modification assignments can be trusted and which ones have to be taken with care.

[1] B. Bodenmiller et al., “PhosphoPep[mdash]a phosphoproteome resource for systems biology research in Drosophila Kc167 cells,” Mol Syst Biol, vol. 3, Oct. 2007.

[2] S. A. Beausoleil, J. Villen, S. A. Gerber, J. Rush, and S. P. Gygi, “A probability-based approach for high-throughput protein phosphorylation analysis and site localization,” Nat Biotech, vol. 24, no. 10, pp. 1285-1292, Oct. 2006.

[3] J. V. Olsen et al., “Global, In Vivo, and Site-Specific Phosphorylation Dynamics in Signaling Networks,” Cell, vol. 127, no. 3, pp. 635-648, Nov. 2006.

[4] E. Ahrne´, F. Nikitin, F. Lisacek, and M. Mu¨ller, “QuickMod: A Tool for Open Modification Spectrum Library Searches,” Journal of Proteome Research, http://dx.doi.org/10.1021/pr200152g

L'identification des protéines par les moteurs de recherche est une étape nécessaire aux analyses de protéomique \"bottom-up\" à haut débit. Elle est basée sur l'identification des peptides et sur leur assemblage le long des séquences protéiques. En général les moteurs de recherche se limitent au classement des protéines identifiées par ordre de score ou de probabilité. Lorsque plusieurs protéines contiennent des peptides identiques, un post-traitement est nécessaire pour lever les ambiguïtés. Classiquement il est basé sur l'élimination par regroupement des protéines identifiées uniquement par des peptides communs à d'autres protéines. Mais cela ne suffit pas pour gérer les cas d'intersections entre groupes de peptides communs.

Nous présentons ici un algorithme original basé sur un deuxième niveau de regroupement, qui réunit les protéines contenant au moins un peptide en commun. Il permet de détecter des protéines redondantes non éliminées par le premier regroupement, de connaître précisément le nombre de peptides communs et spécifiques à chaque protéine et d'estimer le nombre d'isoformes identifiées dans l'échantillon. Par ailleurs nous avons vérifié qu'à de rares exceptions près, ce niveau de classement regroupe les protéines par fonction. Nous présenterons un cas extrême d'utilisation avec une analyse de métaprotéomique bactérienne.

Les recherches systématiques de modifications post-traductionnelles conduisent souvent à une incertitude de position quand plusieurs résidus susceptibles d'être modifiés sont proches dans la séquence. En général, les moteurs de recherche proposent les positions les plus probables en analysant séparément chaque spectre, ce qui conduit parfois à des résultats contradictoires pour un même site suivant le spectre analysé.

Nous présentons ici un algorithme innovant qui regroupe l'ensemble des spectres MS/MS identifiant une modification sur des résidus voisins, de manière à tenir compte de l'ensemble des spectres simultanément pour lever l'incertitude de position. Ceci permet de dénombrer correctement les sites modifiés et les protéines qui les contiennent. Nous présenterons un exemple d'utilisation sur une analyse du phosphoprotéome du maïs.

Ces deux algorithmes ont été implémentés dans le logiciel X!Tandem Pipeline utilisant le moteur de recherche X!tandem. Il est disponible gratuitement à http://pappso.inra.fr/bioinfo/xtandempipeline/ et distribué sous licence GPL.

Dr. Pascal Kintz has a degree in Pharmacy (1985), a Diplôme d'Etudes Approfondies in Molecular Pharmacology and a PhD in Toxicology (1989) of the Université Louis Pasteur in Strasbourg. He was Associate Director of the Institute of Legal Medicine of Strasbourg and Associate Professor of Legal Medicine until the end of 2004. Then he was Head of the Scientific Affairs at ChemTox Laboratory, a private structure in Strasbourg, France (2005-2010). Currently, he is consultant in Toxicology, President of his own company, X-Pertise Consulting.

His main topics of interest include: alternative specimens with a special focus on hair and oral fluid, pharmacology of drugs of abuse, postmortem toxicology and doping control.

He is active in several national and international scientific societies, such as Société Française de Toxicologie Analytique, SFTA (President 1997-2003), The International Association of Forensic Toxicologists, TIAFT (President 2005-2008) and the Society of Hair Testing (Founding Member in 1995, President since 2008).

He received the TIAFT Award for Excellence in 2001

He is an Expert for Justice for the Court of Appeal of Colmar since 1992, appointed by the Court of Cassation since 2007, for Pharmacology / Toxicology and blood alcohol determination and an Expert certified by the Gesellschaft für Toxicologische und Forensische Chemie (Germany) and Eurotox.

Dr Kintz has published more than 300 papers in peer-reviewed journals. He is associate editor of Journal of Analytical Toxicology and regular reviewer for Journal of Chromatography, Forensic Science International, Clinical Chemistry, Journal of Pharmaceutical Sciences, and Annales de Toxicologie Analytique.

The use of a drug to modify a person’s behaviour for criminal gain is not a recent phenomenon. However, the recent increase in reports of drug-facilitated crimes (sexual assault, robbery) has caused alarm in the general public. Drugs involved can be pharmaceuticals, such as benzodiazepines (flunitrazepam, lorazepam ...), hypnotics (zopiclone, zolpidem), sedatives (scopolamine, neuroleptics, some anti-H1) or anaesthetics (GHB, ketamine), drugs of abuse, such as cannabis, ecstasy or LSD, or more often ethanol.

To perform successful toxicological examinations, the analyst must follow some important rules : 1. obtain as soon as possible the corresponding biological specimens (blood, urine and hair), 2. use sophisticated analytical techniques (LC/MS, HS/GC/MS, tandem mass spectrometry); and 3. take care on the interpretation of the findings.

Even after the publication of these guidelines for the clinicians, in most cases specimens are collected at best 24 hours after the crime has occurred.

Drugs used to facilitate sexual assaults can be difficult to detect (active products at low dosages, chemical instability), possess amnesic properties and can be rapidly cleared from the body (short half-life). Prohibiting immunoassays and using only hyphenated techniques, substances can be found in blood for 6 hours to 2 days and in urine for 12 hours to 5 days. In these situations, blood or even urine can be of poor interest. This is the reason why this laboratory developed an original approach based on hair testing.

Hair was suggested as a valuable specimen in situations where, as a result of a delay in reporting the crime, natural processes have eliminated the drug from typical biological specimens. While there is a lot of papers focused on the identification of drugs in hair following chronic drug use, those dealing with a single dose are very scarce.

This laboratory recommends to wait for 3-4 weeks after the offense and then collect 4 strands of about 100 hair. One strand will be used to test for drugs of abuse (mostly for cannabis, but also for ecstasy related compounds and cocaine that are sometimes observed), one for GHB and the other one for a screening of 30 various sedatives. The last strand can be used for a potential counter-analysis. After decontamination, hair is then segmented as follows : 0 to 2 cm (segment corresponding to the period of crime), 2 to 4 and 4 to 6 cm (which should be drug-free). For GHB, segments are of 3 mm (n=8).

Conventional GC/MS can be used to test for drugs of abuse, but given the expected concentrations to measure in low weight segments (in order to avoid the shave the victim), GHB and sedatives are tested by GC-MS/MS and UPLC-MS/MS, respectively.

The experience of the authors will be documented in cases involving GHB, zolpidem, bromazepam, alprazolam, scopolamine, alimemazine, diphenhydramine ...

Hair analysis may be a useful adjunct to conventional drug testing in sexual assault. It should not be considered as an alternative to blood and urine analyses, but as a complement. This approach may find useful applications, but appears very expensive, given the number of analyses to achieve with sophisticated equipment.

La RMN et la spectrométrie de masse sont deux outils complémentaires en métabolomique. Alors que la RMN est désormais largement reconnue dans le domaine de la métabolomique de par sa robustesse, la spectrométrie de masse est quant à elle en pleine expansion, en particulier depuis l’arrivée des colonnes chromatographiques de granulométrie faible et l’arrivée de la spectrométrie de masse à très haute résolution de type Orbitrap. Dans ce travail, nous nous sommes intéressés à développer une méthode associant en parallèle la RMN et le couplage UHPLC-HRMS, qui a été appliquée à une étude des effets du bisphénol A à faibles doses au niveau cellulaire (cellules HepaRG). Nous nous sommes particulièrement intéressés à l’optimisation des paramètres de retraitement des données de spectrométrie de masse par le pack XCMS.

Des cellules HepaRG en culture ont été exposées ou non à des faibles doses de bisphénol A (DMSO versus BPA 10e-6, 10e-9 et 10e-12mol/L). Les extraits cellulaires ont ensuite été analysés tout d’abord par RMN du proton (600MHz, Bruker) puis par couplage UHPLC-FTMS, constitué d’un système RSLC3000 (Dionex), d’un gradient de phase mobiles composées de H2O/méthanol/acide acétique 95/5/0,1 et méthanol/acide acétique 100/0,1 éluées à 0,3 mL/min, et d’une colonne chromatographique Hypersil Gold (100x2,1 mm, 1,9 µm) (Thermo Scientific) à 30°C. La détection a été réalisée avec un spectromètre de masse LTQ-Orbitrap XL (Thermo Scientific) en mode FTMS avec une source d’ionisation Electrospray fonctionnant en mode positif ou négatif.

Tout comme les autres logiciels de traitement de données LC-MS de métabolomique, le pack XCMS utilise deux grandes fonctions : la détection/intégration de pics chromatographiques au sein des signaux bruits, et le regroupement/alignement des variables (couple m/z et tR) entre les échantillons. Ces deux fonctions sont basées sur différents algorithmes présents dans le pack XCMS, mais qui nécessitent le réglage de paramètres à la fois nombreux et peu explicites. Une interface développée au laboratoire (MetMS) rend le paramétrage de XCMS plus aisé depuis l’importation des fichiers, jusqu’à la création de la liste des pics détectés, par l’utilisation de fenêtres de commande. De plus, cette interface permet de visualiser les spectres de masse et les chromatogrammes à chaque étape du traitement XCMS. Ainsi, nous avons optimisé chaque paramètre XCMS à nos conditions d’analyses, par l’étude de la définition de ces paramètres et leur influence sur la détection de pics.

Les données ainsi générées par RMN et UHPLC-HRMS ont été traitées par des méthodes statistiques multivariées. La régression PLS-DA a permis une bonne séparation des différents groupes. L’ensemble du processus mis en place depuis le traitement des échantillons jusqu’au traitement des données produites sera présenté en s’appuyant sur l’application traitée.

Les caractères « globaux » et « non ciblés » d’une étude métabolomique en font une approche de choix dans de nombreux domaines. Appliquée à la sécurité chimique des aliments, la métabolomique laisse entrevoir de nombreuses perspectives comme, par exemple, l’ambition de découvrir de nouveaux biomarqueurs révélateurs d’une contamination de la (des) denrée(s). Dans ce contexte, et en raison de l’universalité de la détection et des niveaux de sensibilité atteints, la spectrométrie de masse s’est naturellement imposée comme l’une des technologies de choix pour la prise d’empreintes métabolomiques.

Dans le cadre de ses recherches appliquées à la détection des pratiques frauduleuses en élevage, le LABERCA a développé depuis plus de 5 ans une stratégie complète de prise d’empreinte métabolomique (LC-ESI-Orbitrap(HRMS) // XCMS // SIMCA P+) afin de révéler de potentiels candidats biomarqueurs d’un traitement anabolisant. Appliquée à différentes familles de composés (Hormone de croissance, stéroïdes, ?-agonistes), l’approche à dores et déjà fait ses preuves [1-8] puisque de robustes modèles statistiques permettant de discriminer des groupes d’animaux (i.e : contrôles & traités) ont été établis et l’identification structurale, étape critique et délicate, a pu être réalisée pour certains candidats biomarqueurs. L’objectif est à présent de développer un outil de dépistage basé sur le suivi ciblé des biomarqueurs identifiés, par le développement de méthodes quantitatives par spectrométrie de masse (UPLC-MS/MS). Il s’agit en particulier de travailler à l’optimisation de la détection de ces composés, par l’évaluation de techniques chromatographies alternatives (HILIC) ou encore à l’aide d’une quantification plus fine (dilution isotopique). Enfin un modèle combinant les concentrations de chacun des candidats biomarqueurs permettra de valider l’approche et la pertinence de leur suivi pour mettre en évidence une suspicion de pratique anabolisante en élevage.

[1] Courant F., Pinel G. et al., Analyst, 2009, 134:1637-1646.

[2] Kieken F., Pinel G. et al., Anal Bioanal Chem, 2009, 394:2119-2128.

[3] Anizan S., Bichon E. et al., J Chrom A, 2010, 1217:6652-6660.

[4] Pinel G., Mooney M. et al., TrAC, 2010, 29:1269-1280.

[5] Kieken F., Pinel G. et al., Metabolomics, 2011, 7:84-93.

[6] Antignac J.P., Courant F. et al., TrAC, 2011, 30:292-301.

[7] Pinel G., Weigel S. et al., Anal. Chim. Acta, 2011, in press.

[8] Anizan S., Di Nardo D. et al., Anal. Chim. Acta, 2011, in press.

L'amplitude et l'évolution du dopage demande le développement de techniques de détection (screening) plus rapides et plus fiables. Une des principales difficultés des laboratoires de contrôle est l'émergence continue des substances et des méthodes dopantes. Les nouveaux composés présentent les mêmes propriétés dopantes, mais leur structure est légèrement modifiée et, par conséquent, ils ne peuvent pas être détectés par les moyens classiques.

Afin de relever ce défi, nous avons choisi une approche métabonomique. Contrairement aux méthodes utilisées à l'heure actuelle, cette approche n'est pas axée sur la détection des substances illicites, mais sur leurs effets sur le métabolisme de l'athlète. Plus précisément, cette étude vise à mettre en évidence des différences d’ordre métabolique entre différents groupes d’individus tels que : des athlètes non-dopés, des athlètes dopés, des volontaires, etc. La stratégie choisie comprend cinq étapes, à savoir : le choix des échantillons, l’obtention des empreintes moléculaires, le traitement statistique des données, la sélection de biomarqueurs potentiels et l’identification de ces derniers. L’urine est une matrice extrêmement complexe contenant un nombre impressionnant de métabolites avec des propriétés chimiques et des concentrations très variables. De ce fait, après une étude de faisabilité réalisée sur une plateforme LC-Q-TOF, l’étude a été orientée vers la spectrométrie de masse de très haute résolution FT-ICR.

Cette étude a été réalisée sur une cohorte de 50 échantillons d’urine dont 30 échantillons ont été testés positifs (glucocorticoïdes et béta-2-Agonistes) par l’AFLD; et 20 forment le groupe des contrôles (athlètes non dopés et volontaires). Les échantillons ont été analysés dans une même séquence sur un spectromètre de masse FT-ICR en mode positif et négatif. Les empreintes moléculaires ainsi obtenues ont été ensuite traitées par analyse multivariée (APC et OPLS-DA) dans le but de classifier les échantillons et mettre en évidence des profils discriminants.

La précision de mesure (< 1 ppm) ainsi que la sensibilité de la technique utilisée ont conduit à l’obtention d’empreintes moléculaires riches en information et très qualitatives. Ainsi, plusieurs composés ayant des profils potentiellement discriminants ont été sélectionnés. Ensuite, en utilisant un générateur de formules brutes et la plateforme MassTrix qui regroupe plusieurs bases de données une vision globale de la classification des échantillons a été obtenue.

La comparaison des signatures urinaires a montré que la métabonomique pourrait être un outil complémentaire pour obtenir des informations et des profils riches à partir de l’urine. De plus, l’étape de préparation d’échantillon minimale (dilution) et le temps d’analyse très court font d’elle une technique adaptée à l’analyse des grandes cohortes d’échantillons, notamment dans des études cliniques.