Final Report Summary - PROSPECTS (Proteomics specification in time and space) Executive Summary: 'Proteomics' is the large-scale study of proteins, particularly their structures and functions. Since proteins are the functional units of the cells, proteomics holds the potential to answer some important questions that genomics left unsolved. The large European Union (EU) project PROSPECTS was a collaborative research effort in the proteomics field which brought together 10 leading research groups from around Europe, as well as Thermo Fisher Scientific, a mass spectrometry instrument manufacturer and chromatography company. The different groups aimed at improving the existing technology to develop insights into the cellular function of proteins and their aberration during diseases and thereby advancing the proteomics field to a 'second generation' state. With the innovative technology and methods developed in the PROSPECTS consortium, the proteomics community has now become able to deal with transient complexes and near complete descriptions of a proteome in time and space. The PROSPECTS consortium adopted a multidisciplinary approach to achieve its far reaching goal of quantitatively measuring the dynamics of a whole set of proteins in an experimental sample. In addition, the project focused on diseases related to folding stress which maximises its utility to the wider biomedical community. The new instrumentation that has been developed in close collaboration of the PROSPECTS partners is already commercially available and a big market success. Experimental workflows established with the improved equipment and finely tuned to the researchers' needs enable in-depth proteomics at unprecedented level and at considerably lower cost. PROSPECTS not only performed leading edge research in the field, but also invested considerable efforts into the dissemination of the results to the scientific community at various levels. Scientific results were published in over 150 peer-reviewed publications so far, many of these in the highest ranking journals in their field (e.g. Cell, Nature). Many of the breakthroughs were achieved through a close interdisciplinary collaboration between PROSPECTS partners, as is evident from 25 major joint publications. In addition, new software was developed and made freely available online. The PROSPECTS consortium also trained several PhD students and post-docs and organised many well-recognised workshops and seminars. The results of PROSPECTS have been made accessible to the public through a periodic newsletter and the project's website. To summarise, PROSPECTS demonstrated very close and fruitful interactions of industrial and academic partners leading to technology developments and biological insights of use not only to all consortium members, but also to the scientific community at large. At the end of its lifetime, the project has impressively fulfilled all expectations and the scientific advisors concluded that 'in summary, PROSPECTS is clearly one of the most successful efforts on the development and application of proteomics technologies.' Project Context and Objectives: In 2008, proteomics was a newly emerging field in biomedical research, which dealt with the large-scale identification and characterisation of large groups of proteins, or 'proteomes'. These can either be the components of a subcellular organelle or compartment, or even the entire protein complement of whole cells and tissues. Proteomics is essential in the functional annotation of the genome and in future attempts to build a quantitative, 'systems-based' description of cell biology. 'First generation' proteomics approaches, however, largely measured protein complexes and proteomes as homogeneous and static entities with little or no quantitative annotation. PROSPECTS was a proposal by world leaders in this young discipline to make a major advance, both by developing much more powerful instrumentation and by applying novel proteomics methods that would allow them to annotate quantitatively the human proteome with respect to protein localisation and dynamics. Complementary technologies, including mass spectrometry, cyro-electron microscopy and cell imaging were intended to be applied in innovative ways to capture transient protein complexes and the spatial and temporal dimensions of entire proteomes. The PROSPECTS consortium set out to develop these new proteomics technologies in a generic fashion to maximise their utility to the wider biomedical community. Consequently, they aimed at generating comprehensive data sets that would foster many downstream functional studies. PROSPECTS endeavoured devising approaches that would allow unique insights into the molecular basis of multiple forms of human disease, specifically neuro-degeneration and other diseases related to folding stress. The multidimensional data sets generated in PROSPECTS would be integrated using advanced data aggregation and machine learning and made available to the scientific community via annotated online public databases and used as a basis for a systems biological modelling of the human proteome, with spatial and temporal resolution within the cell. Project Results: Within the technology core of the PROSPECTS consortium the coordinating Mann group from the Max-Planck-Institute of Biochemistry (MPIB) Martinsried and its industrial partner Thermo Fisher Scientific, Bremen, the leading manufacturer of mass spectrometry instrumentation, have collaborated successfully in the improvement of instrumentation in mass spectrometry (MS) for years. In mid 2009, a novel ion trap-Orbitrap hybrid instrument, the 'Velos', was made commercially available. It incorporated several newly developed technologies which enhanced resolution and acquisition speed beyond all expectations. The Velos thus quickly became an essential tool in the experimental parts of PROSPECTS, contributing to the determination of in-depth proteomes in minute amounts of samples. Further additions to the Orbitrap family of mass spectrometry instruments were their successors, the Orbitrap ELITE and Q Exactive. The innovations target higher speed, resolving power and mass accuracy, sensitivity and m/z range. The new instruments became available already during the project's lifetime and contributed substantially to the success of the proteomics part of PROSPECTS. Used in combination with novel software for real-time analysis of the data, they make proteomic studies in the consortium and for proteomics research worldwide faster, cheaper and easier to use for a broad range of applications. A second breakthrough for MS technology was the optimisation of sample separation before application to MS instruments. The small company Proxeon, an expert on nano liquid chromatography, developed an approach to increase resolution by using nano ultra-high pressure liquid chromatography (UHPLC). The application of high pressure allows for utilisation of longer columns, which in turn led to a high increase in protein detection. A commercial version of the chromatography system, 'EASY-Spray', was launched in 2012. Using a novel workflow of a combination of EASY-Spray and mass spectrometric analysis on a Q Exactive instrument with a novel protocol for 'single-shot' analysis, high performance proteomics, extremely high sensitivity and coverage was achieved - the complete yeast proteome could be quantified in-depth by single shot analysis. Meanwhile this new technology has spread to proteomics labs worldwide and developed 'EASY-Spray' and long columns into a huge commercial success. Linking up cryo-electron tomography and MS of large complexes is a long term goal in structural cell biology, bearing the promise of valuable insights into the spatio-temporal organisation of proteomics networks in a close-to-living state. However, cryo-electron tomography is restricted to very thin samples, which has hindered its application to mammalian cells. The Baumeister group at the MPIB therefore explored and developed methods for the physical preparation of samples of suitable dimensions using the isolation or nano-machining (the machining and fabrication at the nanometre scale) of samples by optical and ion beam methods. They established efficient protocols to thin samples using a focused beam of heavy metal ions to slice away major parts of frozen cells grown on microscopy grids. This 'milling' of samples leaves a thin lamella amenable to electron microscopy, a first step to the investigation of mammalian cells at atomic resolution. Many biological functions require proteins to interact in larger complexes within cells. These assemblies of proteins are notoriously difficult to analyse because of their size and the fact that they are often unstable. The structural biology core of PROSPECTS therefore focused on the major challenge of devising methods to investigate such labile macromolecular protein complexes using multiple complimentary strategies to tackle this issue, which will be presented in the following paragraphs. The Baumeister group performed single particle analysis using cryo-electron microscopy, thus preserving even fragile complexes by rapid freezing of the biological sample. The major goal within PROSPECTS was the automation of data acquisition and interpretation to reduce the time required for structure determination significantly. They focused on the structural characterisation of the 26S proteasome, a complex responsible for protein degradation, which served as a blueprint for integrative structure determination. Using high-throughput data acquisition they have obtained more than 2,000,000 particles, which resulted in a approximately 7 Å resolution reconstruction of the 26S proteasome. With the help of site-specific cross-links (see below) they were able to generate an initial atomic model of the complete 26S proteasome, a milestone in the ubiquitin-proteasome field. The Aebersold group at the Eidgenössische Technische Hochschule (ETH) in Zürich develops cross-linking reagents that glue the components of macromolecular complexes together rather than quick freezing them. These linkers stabilise samples for structural analysis in an electron microscope. Also, the cross-linked complexes can be broken down in smaller peptides for an analysis by MS, to reveal the exact positions of cross-links between different parts of the complexes. This novel combination of MS and electron microscopy paved the way for investigating these large assemblies - zooming in from the whole complex to their individual building blocks. In this project, for the first time, the concept of cross-linking was extended to a protein interaction network - functionally related protein complexes that interact with each other on a higher level of organisation. In addition, new software (xQuest) was developed to interpret MS data and identify the cross-linking sites in a very efficient way. The specific technological advances were integrated into a complete and robust platform for the identification of cross-linking events. This approach was applied by different members of the consortium to a number of systems including 26S proteasome (see above) and TRiC, a so-called 'chaperonin' that helps to fold other proteins to their active form. This resulted in several publications in high impact journals demonstrating the efficient collaboration of different groups under the PROSPECTS framework. The Robinson lab at the University of Oxford also seeks to analyse large protein assemblies by MS, but in contrast to the analytical approaches above they use MS as a tool to directly analyse large macromolecular complexes and their interactions. The complex which was the main target for this study was the ribosome which is the protein factory of the cell. Ribosomes link amino acids together forming a nascent protein chain. During this process the ribosome interacts with a range of associated factors. The Robison lab succeeded in optimising the purification protocols in a way that enabled them to identify a range of associated factors, their intrinsic stoichiometries and networks. Based on these results they established new models of subunit arrangements and interaction maps for the ribosome. In addition, they developed a deuteration protocol, which enables them to follow folding and protection of the nascent amino acid chain as it is built by the ribosome. Extending the use of MS further, the Robinson lab is currently working to isolate the molecules of interest in the gas phase from a heterogeneous sample and deposit those on an electron microscopy (EM) grid for further analysis with EM. This original and ground-breaking method has already been successfully implemented with a first generation instrument in combination with optimised landing conditions that cause only a minimal distortion of the molecular complexes upon deposition. Currently, the developed method is being transferred to a new generation mass spectrometer (Q Exactive, which only became available during the end of the Prospect period) with higher ion throughput which will finally combine the efficient separation and landing setup with the high ion count necessary for an efficient cryo-EM pre-separation. The Serrano group at the Centre for Genomic Regulation (CRG) in Barcelona used another approach to investigate protein interactions by combining experiments with their expertise in modelling. The amount of proteins binding to each other depends on the chemical property of proteins, resulting in binding with different strength (binding affinity) and on the amount of protein present in a cell (protein abundance). One of the remaining challenges is to generate predictive mathematical computational models, where quantitative data (affinities and protein abundances) are used to capture the dynamic system behaviour (systems biology). The final aim is to generate computational models that predict cell function and protein interaction changes for both, any physiological cell type in the human body and cell types in tissues under disease conditions. In collaboration with the Hartl and Aebersold lab, the Serrano group have developed experimental and computational tools that enable: 1. the quantification of protein abundances in different cell types 2. the measurement of interaction strength between proteins 3. the computational prediction of interaction strength using three-dimensional structural information 4. distinguishing proteins that can bind to each other at the same time from those that can only bind in a consecutive way 5. investigating the interaction between proteins in their natural cell environment in an improved way (called FRET helper interactions) and 6. the computational simulation of larger protein interaction networks to calculate the amount of protein binding for each interaction and complex. Another important step in refining proteome analysis is the determination of the subcellular localisation of individual protein species. The localisation of proteins within the cell determines their potential for interaction and cell function, since interactions can occur only if partners reside in the same cell compartment in close vicinity. The large scale analysis of subcellular proteomes mostly relies on biochemical fractionation and advanced quantitative MS analysis (see above). Large efforts have been invested in establishing and evaluating reliable methods for cell fractionation, data analysis and their interpretation. Based on stable isotope labelling with amino acids (SILAC)-based MS, a variety of new generic approaches was established through a close collaborative effort of the PROSPECTS partners. Together, they addressed the relative distribution of the larger part of the proteome in the three major compartments of the cell - the membrane, cytosol and nucleus - in various settings to define a 'localisation index' for each protein. Around 12 600 proteins were mapped and quantified in HeLa cells and their distribution between the cell the membrane, cytosol and nucleus determined. The Andersen group at the Syddansk Universitet (SDU) in Odense developed an alternative approach for biochemical fractionation based on the analysis of the enrichment profiles of proteins in centrifugation gradients, the so-called 'protein correlation profiling'. Proteins are identified as constituents of cell compartments or organelles due to co-purification with relevant marker proteins for that organelle. Sophisticated software is required to single out genuine organelle proteins from co-purifying contaminants. In a first study protein correlation profiling helped identify a proteome of 230 centrosomal proteins whose centrosomal localisation was independently confirmed by immuno-localisation in collaboration with the experts from the Protein Atlas (Royal Institute of Technology (KTH), Stockholm, see below). Another study identified proteins required for protein degradation mediated by autophagosomes. Over the past years, the Uhlén group at the KTH has been working towards their aim of generating antibodies for all human proteins known as the Human Protein Atlas. They played a crucial role in PROSPECTS by providing antibodies from their Protein Atlas to their collaborators. During the course of PROSPECTS, over 16 000 antibodies generated by the Human Protein Atlas project have been used for characterising localisation of the corresponding 13 000 human proteins. For instance to characterise the human centrosomal proteome, identify autophagosome-associated proteins, characterise mouse-tissues as well as characterise signaling pathways, protein isoforms and turnover. In addition a novel method for absolute quantification of protein copy numbers using mass-spectrometry was developed in close collaboration with PROSPECTS partners using Human Protein Atlas project antigens as spike-in reagents. Also, a streamlined approach for systematic localisation of proteins on a subcellular level using high-resolution microscopy has been established, including validation of results using gene silencing. Furthermore the level of details in the annotations has been refined and the number of subcellular structures identified doubled. Generated data on protein expression profiles and subcellular localisation are made publicly available to the research community via the Protein Atlas database (www.proteinatlas.org). The latest version of the atlas, v11.0 comprises expression profiles for 75 % of all human genes. In close collaboration with the Mann, Uhlén and Andersen group, the Lamond lab at the University of Dundee have successfully developed new methodologies and an efficient workflow for the quantitative measurement of protein distributions between cellular compartments using SILAC MS. A reliable procedure for fractionating cells into distinct subcellular compartments was established, including cytoplasmic, nuclear and nucleolar compartments. In addition, they developed methods for fractionation using detergent solubility, which provide separation of cytoplasmic, membrane, soluble nuclear, chromatin and cytoskeletal compartments. This methodology enables spatial proteomics analysis on smaller numbers of cells and provides information on the behavior of distinct protein pools in separate subcellular compartments, which was not observed previously by analysing only whole cell extracts. The specificity of fractionation has been confirmed by the orthogonal technique of western blotting, using antibodies recognising proteins known to localise to these compartments (i.e. tubulin, syntaxin 6, histone H3 and fibrillarin). Using these techniques they have analysed in detail three human cell lines (HeLa, U2OS and HCT116 cells) under various experimental conditions, including deoxyribonucleic acid (DNA) damage, ultraviolet (UV) stress, protein-degradation inhibition. They have also developed a software environment for the management of these quantitative MS-based proteomics data, called PepTracker (http://www.peptracker.com/) which includes a set of tools for the convenient analysis and graphical representation of the protein distributions. They have developed within PepTracker an online database for sharing the data with the community, called 'The Encyclopedia of Proteome Dynamics'. In addition, the Lamond lab focused on the redistribution of proteins between cellular compartments in response to stress and genetic manipulations. They developed and refined the required methodology and workflow for the quantitative measurement of dynamic protein localisation during these cellular perturbations using SILAC-based quantitative MS. They focused on the analysis of protein localisation by sub-cellular fractionation, with a biological focus on analysis of the nucleolus as a general strategy. This allowed them to perform quantitative MS based proteomics that could be directly compared with quantitative fluorescence microscopy experiments analysing the same endogenous proteins. They validated the approach and methodology and compared the analysis of green-fluorescent-protein (GFP)-tagged proteins with the analysis of the equivalent endogenous protein using specific antibodies from the Human Protein Atlas group (see above). These data have demonstrated the accuracy of the spatial proteomics datasets generated thus far. Importantly, pulse-labelling strategies combined with SILAC MS were employed to characterise the turnover rates and half-lives of proteins encoded by more than 5 000 human genes. In these studies, cells were exposed to heavy isotope-substituted amino acids (typically 13C-labeled arginine and lysine) for relatively short periods (typically 1-8 hours), rather than incubated for 6 days or more to ensure complete substitution as in standard SILAC protocols. Sampling of protein populations in specific purified organelles or substructures at one hour time intervals following the heavy isotope amino acid pulse revealed the rate at which newly translated proteins replaced the pre-existing proteins in each respective subcellular location. Another topic of interest for the Lamond group was the mitotic cell cycle. It is a fundamental and highly regulated biological process that is important for all organisms. Regulatory pathways that control cell cycle progression provide critical links between signal transduction of extracellular and intracellular states and the timing of major cellular events, such as DNA replication and cell division. The Lamond lab combined centrifugal elutriation with the analysis workflows developed previously to follow protein expression and localisation across the cell cycle, including within the nucleolar compartment. The PROSPECTS collaborators' original approach using antibodies and other proteomics methods (e.g. SILAC, MS, fluorescence microscopy) in a complementary way was unprecedented in the proteomics field thus far. It constitutes an important advantage over previous research, because it enabled the collaborators to cross-validate their data using more than just one method which in turn yielded more robust results. In addition, thanks to antibody-based methods, different MS protocols which usually deal with protein fragments in vitro could be tested on real tissue. Shedding light on molecular basis for neurodegenerative diseases was another important focus of the PROSPECTS consortium which was mainly addressed by the Hartl group at the MPIB. Together with the Mann group, they investigated the influence of folding stress and neurodegenerative disease proteins on the spatial and temporal proteome. The chaperone system ensures that proteins get folded correctly into their native three dimensional structures and also maintain these structures under stress conditions. Insufficiently folded proteins tend to form toxic aggregates and a wide range of diseases such as Alzheimer's disease, Huntington's disease Amylotrophic Lateral Syndrome and Parkinson's disease is associated with such aggregates. Firstly, they studied the proteome-wide effects of an inhibitor of the important chaperone Hsp90. Inhibition of Hsp90 induced a stress response with activation of chaperones and translational pathways. A wide range of kinases was found to be down-regulated thus confirming that Hsp90 is involved in the folding of kinases. Since protein phosphorylation is one of the key regulatory events in tumor growth, this result explains the broad antitumor activities of Hsp90 inhibitors. Secondly, they focused on the mechanisms of toxicity exerted by amyloid-like proteins in human cells. Most proteins were found to associate with aggregates directly after their synthesis so preventing them from functioning properly. Aggregation thus targets a metastable proteome, thereby causing multifactorial toxicity. Thirdly, the Hartl group investigated the proteome changes during aging in stressed human cells and worms. The ability of old cells to maintain the proteostatic balance under stress conditions was found to be compromised by increased levels of aggregates and decreased levels of ribosomal proteins. Also in worms, a decrease in ribosomal subunit proteins during the aging process appeared to contribute to aging, whereas an increase in proteasome core subunits reflects an attempt to degrade misfolded proteins. Importantly, the lifespan of worms could be extended by RNAi mediated down-regulation of aggregation-prone proteins. In summary, important pathways and mechanisms in protein folding diseases and aging were revealed. The Linial group at Hebrew University was set to establish PROSPECTS' warehouse for experimental data. Due to the rapidly growing amounts of experimental data obtained with high throughput technologies, data warehousing has become increasingly important, but also increasingly complex. The Linial group further developed their proprietary ProtoDB database for this task and combined it with a set of web tools for protein analysis. It now covers almost 20 million sequences and their automatic 'annotations' - a description of protein characteristics like size, abundance, expression levels and patterns. An important aspect for PROSPECTS partners was the incorporation of quantitative data such as the number of peptides, number and types of phosphorylation sites, the subcellular localisation and the nature of other common modifications. Experimental quantitative data on expression level, retention time, turnover, localisation index and more, can be added. The platform has been tested with numerous data sets produced by PROSPECTS and has thereby contributed to gaining insights into biological processes. In summary, the ProtoDB and its associated Pandora visualisation tool serve as a protein centric view of the current biological knowledge. Potential Impact: The array of PROSPECTS-related improvements reaches out beyond the expert community. The novel instrumentation developed during PROSPECTS lifetime is already commercially available and a big market success. Experimental workflows adapted to or developed with the improved equipment enable in-depth proteomics at unprecedented level and at considerably lower cost. The close collaboration between a leading research group and the leading manufacturer in the field pays off in this respect with novel instrumentation finely tuned to the demands of the researchers. The ease of use of PROSPECTS' novel technologies and the steadily decreasing amounts of material required for analysis now enable a more general use of proteomics in the biological and biomedical communities. PROSPECTS not only performed leading edge research in the field, but also invested considerable efforts into the dissemination of the results to the scientific community at various levels. Scientific results were published in over 150 peer-reviewed publications so far, many of these in the highest ranking journals in their field (e.g. Cell, Nature). Many of the breakthroughs were achieved through a close interdisciplinary collaboration between PROSPECTS partners, as is evident from 25 major joint publications. One highlight here was the publication of a full 'Special issue' in the journal Molecular and Cellular Proteomics (March 2012), where PROSPECTS partners jointly presented 16 original research publications describing a selection of project results and their view on the future of proteomics. Various software applications and data bases were established or further developed within PROSPECTS, such as packages for data acquisition and analysis for MS, cryo-electron tomography and Single Particle Analysis and for the modelling of protein folding and protein interaction. Similarly, the data sets obtained through PROSPECTS research are compiled in dedicated data bases. All of these resources are freely available to the public online. A compiled annotated list of tools is available on the PROSPECTS website. Extensive and continuous training is essential to maintain the high standards in a research field that heavily relies on highly specialised technology. PROSPECTS dedicated large efforts to this aspect - during its lifetime the project trained over 30 PhD students and 70 post-doctoral researchers. These next generation proteomics experts not only learned how to use state of the art technology, but also had the opportunity to build up their own professional network with the leading researchers in the field during project meetings, lab visits and joint experiments. Many of these young scientists already moved on to new labs, building up their own groups, or joined companies where they now apply and spread the expertise and new knowledge they acquired through PROSPECTS. Reaching out far beyond the project, open workshops and courses were held for the scientific community. The most prominent of these is the series of MaxQuant Summer Schools founded by the Mann group during the PROSPECTS lifetime and held annually at the Max Planck Institute in Martinsried. The School started in 2009 with around 50 participants and received extraordinary attention in the Proteomics field since. Originally planned as a single event, they have meanwhile developed into a successful and growing serial. It last took place in 2012 at the MPI of Biochemistry in Martinsried. The School was heavily oversubscribed and a new record number of participants (2011: 105, 2012:150) was admitted. These Summer Schools have attained high reputation in the proteomics field which was reflected in a further increase in applications over the last year. To cope better with the very high - and still growing - interest in these courses the organising Mann department has decided to continue the programme beyond the PROSPECTS lifetime. As an additional means to accommodate a larger audience and to further spread information on MaxQuant, the ten lectures held in the 2012 School have been videotaped and uploaded to Youtube into a dedicated MaxQuant channel. Within the last three months, 4 950 views were recorded. Currently, the first 'post-PROSPECTS' Summer School 2013 is in preparation for which only 2/3 of applicants can be admitted due to heavy over-subscription. To communicate with the general public, the ongoing work and the results of the PROSPECTS partners were regularly issued in the Project Newsletter and on the PROSPECTS website. In addition, the PROSPECTS coordinator is partner in the CommHERE EU-Project ('Communication of European Health Research') which created a network of major research institutions to coordinate communication actions aimed at the media and the general public. One activity of CommHERE was the establishment of the dedicated website http://www.horizonhealth.eu/ presenting EU-projects in health research. The site was officially launched in Brussels on 21 March 2013 and features PROSPECTS (see http://horizonhealth.eu/project/proteomics-specification-time-and-space/89). In summary, PROSPECTS was a highly efficient and very successful collaborative project. An important key to its success was not only the excellence of its members, but also their ability to recognise opportunities for fruitful collaborations in which different groups, scientific and industrial, contributed their skill sets to address important biological and biomedical questions. Therefore, PROSPECTS sets an example how successful and innovative a Europe-wide collaboration between leading scientists and their industrial partners can be. List of Websites: http://www.prospects-fp7.eu/ prospect@biochem.mpg.de
