Adrian Pasculescu

Systematic discovery of in vivo phosphorylation networks

Protein kinases control cellular decision processes by phosphorylating specific substrates. Thousands of in vivo phosphorylation sites have been identified, mostly by proteome-wide mapping. However, systematically matching these sites to specific kinases is presently infeasible, due to limited specificity of consensus motifs, and the influence of contextual factors, such as protein scaffolds, localization, and expression, on cellular substrate specificity. We have developed an approach (NetworKIN) that augments motif-based predictions with the network context of kinases and phosphoproteins. The latter provides 60%-80% of the computational capability to assign in vivo substrate specificity. NetworKIN pinpoints kinases responsible for specific phosphorylations and yields a 2.5-fold improvement in the accuracy with which phosphorylation networks can be constructed. Applying this approach to DNA damage signaling, we show that 53BP1 and Rad50 are phosphorylated by CDK1 and ATM, respectively. We describe a scalable strategy to evaluate predictions, which suggests that BCLAF1 is a GSK-3 substrate.

Linear Motif Atlas for Phosphorylation-Dependent Signaling

Created with both in vitro and in vivo data, NetPhorest is an atlas of consensus sequence motifs for 179 kinases and 104 phosphorylation-dependent binding domains and reveals new insight into phosphorylation-dependent signaling. An Atlas of Phosphorylation NetPhorest is a community resource that uses phylogenetic trees to organize data from both in vivo and in vitro experiments to derive sequence specificities for 179 kinases and 104 domains (SH2, PTB, BRCT, WW, and 14–3–3) that bind to phosphorylated sites. The resulting atlas of linear motifs revealed that oncogenic kinases tend to be less specific in the target sequences they phosphorylate than their non-oncogenic counterparts, that autophosphorylation sites tend to be more variable than other substrates of a given kinase, and that coupling interaction domains with kinase domains may allow phosphorylation site specificity to be low while still maintaining substrate specificity. Systematic and quantitative analysis of protein phosphorylation is revealing dynamic regulatory networks underlying cellular responses to environmental cues. However, matching these sites to the kinases that phosphorylate them and the phosphorylation-dependent binding domains that may subsequently bind to them remains a challenge. NetPhorest is an atlas of consensus sequence motifs that covers 179 kinases and 104 phosphorylation-dependent binding domains [Src homology 2 (SH2), phosphotyrosine binding (PTB), BRCA1 C-terminal (BRCT), WW, and 14–3–3]. The atlas reveals new aspects of signaling systems, including the observation that tyrosine kinases mutated in cancer have lower specificity than their non-oncogenic relatives. The resource is maintained by an automated pipeline, which uses phylogenetic trees to structure the currently available in vivo and in vitro data to derive probabilistic sequence models of linear motifs. The atlas is available as a community resource (http://netphorest.info).

NetworKIN: a resource for exploring cellular phosphorylation networks

Protein kinases control cellular responses by phosphorylating specific substrates. Recent proteome-wide mapping of protein phosphorylation sites by mass spectrometry has discovered thousands of in vivo sites. Systematically assigning all 518 human kinases to all these sites is a challenging problem. The NetworKIN database (http://networkin.info) integrates consensus substrate motifs with context modelling for improved prediction of cellular kinase–substrate relations. Based on the latest human phosphoproteome from the Phospho.ELM and PhosphoSite databases, the resource offers insight into phosphorylation-modulated interaction networks. Here, we describe how NetworKIN can be used for both global and targeted molecular studies. Via the web interface users can query the database of precomputed kinase–substrate relations or obtain predictions on novel phosphoproteins. The database currently contains a predicted phosphorylation network with 20 224 site-specific interactions involving 3978 phosphoproteins and 73 human kinases from 20 families.

Cell-Specific Information Processing in Segregating Populations of Eph Receptor Ephrin–Expressing Cells

Dissecting Ephrin-Receptor Interaction Ephrins are transmembrane proteins that bind ephrin receptors on adjacent cells, leading to propagation of biochemical signals within both cells. Jørgensen et al. (p. 1502) devised a way to use differential isotopic labeling to distinguish cells engineered to express either the receptor or the ligand and to monitor bidirectional signaling events by mass spectrometry of the labeled peptides when the cells were mixed together. Signaling networks were constructed, and the information processing by the two interacting cell types was modeled. Changes in signaling within cells expressing just the ligand (ephrin) caused changes in the signal processing during the adjacent cells response to binding of the ephrin receptor. A proteomic strategy elucidates signaling networks between cells communicating through ephrin proteins and their receptors. Cells have self-organizing properties that control their behavior in complex tissues. Contact between cells expressing either B-type Eph receptors or their transmembrane ephrin ligands initiates bidirectional signals that regulate cell positioning. However, simultaneously investigating how information is processed in two interacting cell types remains a challenge. We implemented a proteomic strategy to systematically determine cell-specific signaling networks underlying EphB2- and ephrin-B1–controlled cell sorting. Quantitative mass spectrometric analysis of mixed populations of EphB2- and ephrin-B1–expressing cells that were labeled with different isotopes revealed cell-specific tyrosine phosphorylation events. Functional associations between these phosphotyrosine signaling networks and cell sorting were established with small interfering RNA screening. Data-driven network modeling revealed that signaling between mixed EphB2- and ephrin-B1–expressing cells is asymmetric and that the distinct cell types use different tyrosine kinases and targets to process signals induced by cell-cell contact. We provide systems- and cell-specific network models of contact-initiated signaling between two distinct cell types.

Comparative analysis reveals conserved protein phosphorylation networks implicated in multiple diseases.

Comparing the human phosphoproteome to that of flies, worms, and yeast reveals insight into evolution and disease. Phosphorylation Networks in Disease and Evolution Insights into the evolution of protein phosphorylation were revealed by combining the results from two computational analyses—a sequence-alignment approach and a kinase-substrate network alignment approach. The two approaches yielded different, but somewhat overlapping, sets of conserved phosphoproteins among humans and the model organisms. The first provided a set of genes encoding phosphoproteins that had positionally conserved phosphorylation sites, whereas the second included many functionally conserved phosphoproteins that lacked this positional conservation. Enrichment analysis of the genes identified through the kinase-substrate network approach suggested that genes encoding phosphorylated signaling hubs were enriched in disease-associated genes (defined by Online Mendelian Inheritance in Man), and both approaches showed that genes encoding conserved phosphoproteins were enriched in genes associated with cancer. The functional annotation of the two gene sets suggested that positional conservation is common in regions that are structurally constrained, such as those regulated by allosteric interactions, and that the kinase-substrate network method may aid in analyzing fast-evolving signaling processes, where functional conservation does not require positional conservation. The analysis also suggests that conserved regulatory networks may be involved in different diseases. Protein kinases enable cellular information processing. Although numerous human phosphorylation sites and their dynamics have been characterized, the evolutionary history and physiological importance of many signaling events remain unknown. Using target phosphoproteomes determined with a similar experimental and computational pipeline, we investigated the conservation of human phosphorylation events in distantly related model organisms (fly, worm, and yeast). With a sequence-alignment approach, we identified 479 phosphorylation events in 344 human proteins that appear to be positionally conserved over ~600 million years of evolution and hence are likely to be involved in fundamental cellular processes. This sequence-alignment analysis suggested that many phosphorylation sites evolve rapidly and therefore do not display strong evolutionary conservation in terms of sequence position in distantly related organisms. Thus, we devised a network-alignment approach to reconstruct conserved kinase-substrate networks, which identified 778 phosphorylation events in 698 human proteins. Both methods identified proteins tightly regulated by phosphorylation as well as signal integration hubs, and both types of phosphoproteins were enriched in proteins encoded by disease-associated genes. We analyzed the cellular functions and structural relationships for these conserved signaling events, noting the incomplete nature of current phosphoproteomes. Assessing phosphorylation conservation at both site and network levels proved useful for exploring both fast-evolving and ancient signaling events. We reveal that multiple complex diseases seem to converge within the conserved networks, suggesting that disease development might rely on common molecular networks.

Temporal regulation of EGF signalling networks by the scaffold protein Shc1

Cell-surface receptors frequently use scaffold proteins to recruit cytoplasmic targets, but the rationale for this is uncertain. Activated receptor tyrosine kinases, for example, engage scaffolds such as Shc1 that contain phosphotyrosine (pTyr)-binding (PTB) domains. Using quantitative mass spectrometry, here we show that mammalian Shc1 responds to epidermal growth factor (EGF) stimulation through multiple waves of distinct phosphorylation events and protein interactions. After stimulation, Shc1 rapidly binds a group of proteins that activate pro-mitogenic or survival pathways dependent on recruitment of the Grb2 adaptor to Shc1 pTyr sites. Akt-mediated feedback phosphorylation of Shc1 Ser 29 then recruits the Ptpn12 tyrosine phosphatase. This is followed by a sub-network of proteins involved in cytoskeletal reorganization, trafficking and signal termination that binds Shc1 with delayed kinetics, largely through the SgK269 pseudokinase/adaptor protein. Ptpn12 acts as a switch to convert Shc1 from pTyr/Grb2-based signalling to SgK269-mediated pathways that regulate cell invasion and morphogenesis. The Shc1 scaffold therefore directs the temporal flow of signalling information after EGF stimulation.

ProHits: integrated software for mass spectrometry-based interaction proteomics

Affinity purification coupled with mass spectrometric identification (AP-MS) is now a method of choice for charting novel protein-protein interactions, and has been applied to a large number of both small scale and high-throughput studies1. However, general and intuitive computational tools for sample tracking, AP-MS data analysis, and annotation have not kept pace with rapid methodological and instrument improvements. To address this need, we developed the ProHits LIMS platform. ProHits is a complete open source software solution for MS-based interaction proteomics that manages the entire pipeline from raw MS data files to fully annotated protein-protein interaction datasets. ProHits was designed to provide an intuitive user interface from the biologists perspective, and can accommodate multiple instruments within a facility, multiple user groups, multiple laboratory locations, and any number of parallel projects. ProHits can manage all project scales, and supports common experimental pipelines, including those utilizing gel-based separation, gel-free analysis, and multi-dimensional protein or peptide separation. ProHits is a client-based HTML program written in PHP that runs a MySQL database on a dedicated server. The complete ProHits software solution consists of two main components: a Data Management module, and an Analyst module (Fig. 1a; see Supplementary Fig. 1 for data structure tables). These modules are supported by an Admin Office module, in which projects, instruments, user permissions and protein databases are managed (Supplementary Fig. 2). A simplified version of the software suite (“ProHits Lite”), consisting only of the Analyst module and Admin Office, is also available for users with pre-existing data management solutions or who receive pre-computed search results from analyses performed in a core MS facility (Supplementary Fig. 3). A step-by-step installation package, installation guide and user manual (see Supplementary Information) are available on the ProHits website (www.prohitsMS.com). Figure 1 Overview of ProHits. (a) Modular organisation of ProHits. The Data Management module backs up all raw mass spectrometry data from acquisition computers, and handles data conversion and database searches. The Analyst module organizes data by project, bait, ... In the Data Management module, raw data from all mass spectrometers in a facility or user group are copied to a single secure storage location in a scheduled manner. Data are organized in an instrument-specific manner, with folder and file organization mirroring the organization on the acquisition computer. ProHits also assigns unique identifiers to each folder and file. Log files and visual indicators of current connection status assist in monitoring the entire system. The Data Management module monitors the use of each instrument for reporting purposes (Supplementary Fig. 4–5). Raw MS files can be automatically converted to appropriate file formats using the open source ProteoWizard converters (http://proteowizard.sourceforge.net/). Converted files may be subjected to manual or automated database searches, followed by statistical analysis of the search results, according to any user-defined schedule; search engine parameters are also recorded to facilitate reporting and compliance with MIAPE guidelines2. Mascot3, X!Tandem4 and the TransProteomics Pipeline (TPP5) are fully integrated with ProHits via linked search engine servers (Supplementary Fig. 6–7). The Analyst module organizes data by project, bait, experiment and/or sample, for gel-based or gel-free approaches (Fig. 1a; for description of a gel-based project, see Supplementary Fig. 8). To create and analyze a gel-free affinity purification sample, the user specifies the bait gene name and species. ProHits automatically retrieves the amino acid sequence and other annotation from its associated database. Bait annotation may then be modified as necessary, for example to specify the presence of an epitope tag or mutation (Supplementary Fig. 9). A comprehensive annotation page tracks experimental details (Supplementary Fig. 10), including descriptions of the Sample, Affinity Purification protocol, Peptide Preparation methodology, and LC-MS/MS procedures. Controlled vocabulary lists for experimental descriptions can be added via drop-down menus to facilitate compliance with annotation guidelines such as MIAPE6 and MIMIx7, and to facilitate the organization and retrieval of data files. Free text notes for cross-referencing laboratory notebook pages, adding experimental details not captured in other sections, describing deviations from reference protocols and links to gel images or other file types may be added in the Experimental Detail page. Once an experiment is created, multiple samples may be linked to it, for example technical replicates of the same sample, or chromatographic fractions derived from the same preparation. All baits, experiments, samples and protocols are assigned unique identifiers. Once a sample is created, it is linked to both the relevant raw files and database search results. For multiple samples in HTP projects, automatic sample annotation may be established by using a standardized file naming system (Supplementary Fig. 11), or files may be manually linked. Alternatively, search results obtained outside of ProHits (with the X!Tandem or Mascot search engines) can be manually imported into the Analyst module (Supplementary Fig. 12). The ProHits Lite version enables uploading of external search results for users with an established MS data management system. In the Analyst module, mass spectrometry data can be explored in an intuitive manner, and results from individual samples, experiments or baits can be viewed and filtered (Supplementary Fig. 13–14). A user interface enables alignment of data from multiple baits or MS analyses using the Comparison tool. Data from individual MS runs, or derived from any user-defined sample group, are selected for visualization in a tabular format, for side-by-side comparisons (Fig. 1b; Supplementary Fig. 15–17). In the Comparison view, control groups and individual baits, experiments or samples are displayed by column. Proteins identified in each MS run or group of runs are displayed by row, and each cell corresponds to a putative protein hit, according to user-specified database search score cutoff. Cells display spectral count number, unique peptides, scores from search engines, and/or protein coverage information; a mouse-over function reveals all associated data for each cell in the table. For each protein displayed in the Comparison view, an associated Peptide link (Fig. 1b) may also be selected to reveal information such as sequence, location, spectral counts, and score, for each associated peptide. Importantly, all search results can be filtered. For example, ProHits allows for the removal of non-specific background proteins from the hit list, as defined by negative controls, search engine score thresholds, or contaminant lists. Links to the external NCBI and BioGRID8 databases are provided for each hit to facilitate data interpretation. Overlap with published interaction data housed in the BioGRID database8 can be displayed to allow immediate identification of new interaction partners. A flexible export function enables visualization in a graphical format with Cytoscape9, in which spectral counts, unique peptides, and search engine scores can be visualized as interaction edge attributes. The Analyst module also includes advanced search functions, bulk export functions for filtered or unfiltered data, and management of experimental protocols and background lists (e.g. Supplementary Fig. 18–20). Deposition of all mass spectrometry-associated data in public repositories is likely to become mandatory for publication of proteomics experiments2, 7, 10. Open access to raw files is essential for data reanalysis and cross-platform comparison; however, data submission to public repositories can be laborious due to strict formatting requirements. ProHits facilitates extraction of the necessary details in compliance with current standards, and generates Proteomic Standard Initiative (PSI) v2.5 compliant reports11, either in the MITAB format for BioGRID8 or in XML format for submission to IMEx consortium databases12, including IntAct13 (Supplementary Fig. 21). MS raw files associated with a given project can also be easily retrieved and grouped for submission to data repositories such as Tranche14. ProHits has developed to manage many large-scale in-house projects, including a systematic analysis of kinase and phosphatase interactions in yeast, consisting of 986 affinity purifications15. Smaller-scale projects from individual laboratories are readily handled in a similar manner. Examples of AP-MS data from both yeast and mammalian projects are provided in a demonstration version of ProHits at www.prohitsMS.com, and in Supplementary documents. The modular architecture of ProHits will accommodate additional new features, as dictated by future experimental and analytical needs. Although ProHits has been designed to handle protein interaction data, simple modifications of the open source code will enable straightforward adaptation to other proteomics workflows.

Positive Selection of Tyrosine Loss in Metazoan Evolution

Cataloging Kinase Targets Protein phosphorylation is a central mechanism in the control of many biological processes (see the Perspective by Collins). It remains a challenge to determine the complete range of substrates and phosphorylation sites altered by a kinase like cyclin-dependent kinase 1 (Cdk1), which controls cell division in yeast. Holt et al. (p. 1682) engineered a strain of yeast to express a modified Cdk1 molecule that could be inhibited by a specific small-molecule inhibitor. The range of Cdk1-dependent phosphorylation was assessed by quantitative mass spectrometry, which revealed many previously uncharacterized substrates for Cdk1. In addition to phosphorylation on serine and threonine residues, which appears to be evolutionarily ancient, tyrosine phosphorylation occurs primarily in multicellular organisms. Tan et al. (p. 1686, published online 9 July) compared the overall presence of tyrosine residues in human proteins (which are frequently phosphorylated) and in yeast proteins (which are not). Loss of tyrosine residues has occurred during evolution, presumably to reduce adventitious tyrosine phosphorylation. Evolution of tyrosine phosphorylation as a signaling mechanism may have coincided with loss of tyrosine residues to avoid noise. John Nash showed that within a complex system, individuals are best off if they make the best decision that they can, taking into account the decisions of the other individuals. Here, we investigate whether similar principles influence the evolution of signaling networks in multicellular animals. Specifically, by analyzing a set of metazoan species we observed a striking negative correlation of genomically encoded tyrosine content with biological complexity (as measured by the number of cell types in each organism). We discuss how this observed tyrosine loss correlates with the expansion of tyrosine kinases in the evolution of the metazoan lineage and how it may relate to the optimization of signaling systems in multicellular animals. We propose that this phenomenon illustrates genome-wide adaptive evolution to accommodate beneficial genetic perturbation.

Combinatorial proteomic analysis of intercellular signaling applied to the CD28 T-cell costimulatory receptor

Significance Intracellular signaling during complex cell–cell interactions, such as between immune cells, provides essential cues leading to cell responses. Global characterization of these signaling events is critical for systematically exploring and understanding how they eventually control cell fate. However, proteome-wide characterization of intercellular signaling under physiologically relevant conditions involving multiple interacting receptors during cell–cell interactions remains challenging. We developed an integrated proteomic strategy for quantitatively profiling intercellular-signaling events mediated by protein phosphorylation and protein–protein interaction. We applied this approach to determine the influence of a single receptor-ligand pair during T-cell stimulation by blocking the interaction of the CD28 costimulatory receptor with its ligand. This approach is generally applicable to other transmembrane receptors involved in signaling during complex cell interactions. Systematic characterization of intercellular signaling approximating the physiological conditions of stimulation that involve direct cell–cell contact is challenging. We describe a proteomic strategy to analyze physiological signaling mediated by the T-cell costimulatory receptor CD28. We identified signaling pathways activated by CD28 during direct cell–cell contact by global analysis of protein phosphorylation. To define immediate CD28 targets, we used phosphorylated forms of the CD28 cytoplasmic region to obtain the CD28 interactome. The interaction profiles of selected CD28-interacting proteins were further characterized in vivo for amplifying the CD28 interactome. The combination of the global phosphorylation and interactome analyses revealed broad regulation of CD28 and its interactome by phosphorylation. Among the cellular phosphoproteins influenced by CD28 signaling, CapZ-interacting protein (CapZIP), a regulator of the actin cytoskeleton, was implicated by functional studies. The combinatorial approach applied herein is widely applicable for characterizing signaling networks associated with membrane receptors with short cytoplasmic tails.

The Adaptor Protein p66Shc Inhibits mTOR-Dependent Anabolic Metabolism

The adaptor protein p66Shc unexpectedly inhibits glucose metabolism, which could be a useful therapeutic target for diabetes or cancer. Negatively Adapting Metabolism Deficiency of the adaptor protein p66Shc, which inhibits signaling through receptor tyrosine kinases, improves the glucose tolerance of mice that are a genetic model of obesity. Soliman et al. found that in both transformed and nontransformed cells, p66Shc silencing enhanced glucose metabolism, increased the abundance of intermediates in various biosynthetic pathways, and increased cell size. The change in metabolism required the mTOR pathway, which couples energy, nutrient, and growth factor (including insulin) signals to processes that mediate cellular growth and proliferation. Thus, p66Shc suppresses mTOR signaling, and manipulating the abundance of p66Shc or interfering with p66Shc function could be used to treat diseases characterized by disruptions in this pathway, including diabetes and cancer. Adaptor proteins link surface receptors to intracellular signaling pathways and potentially control the way cells respond to nutrient availability. Mice deficient in p66Shc, the most recently evolved isoform of the Shc1 adaptor proteins and a mediator of receptor tyrosine kinase signaling, display resistance to diabetes and obesity. Using quantitative mass spectrometry, we found that p66Shc inhibited glucose metabolism. Depletion of p66Shc enhanced glycolysis and increased the allocation of glucose-derived carbon into anabolic metabolism, characteristics of a metabolic shift called the Warburg effect. This change in metabolism was mediated by the mammalian target of rapamycin (mTOR) because inhibition of mTOR with rapamycin reversed the glycolytic phenotype caused by p66Shc deficiency. Thus, unlike the other isoforms of Shc1, p66Shc appears to antagonize insulin and mTOR signaling, which limits glucose uptake and metabolism. Our results identify a critical inhibitory role for p66Shc in anabolic metabolism.