LINKING THE PROTEINS—ELUCIDATION OF PROTEOME-SCALE NETWORKS USING MASS SPECTROMETRY Delphine Pflieger, 1 Florence Gonnet, 1 Sergio de la Fuente van Bentem, 2 Heribert Hirt, 3,4 and Alberto de la Fuente 5 * 1 Laboratoire Analyse et Mode ´lisation pour la Biologie et l’Environnement, Universite ´ d’Evry Val d’Essonne, CNRS UMR 8587, Evry, France 2 Syngenta Seeds B.V., P.O. Box 2, 1600 AA Enkhuizen, the Netherlands 3 URGV Plant Genomics, INRA/CNRS/University of Evry, 2 rue Gaston Cremieux, F-91057, France 4 MFPL, University of Vienna, Dr. Bohrgasse 9, A-1030 Vienna, Austria 5 CRS4 Bioinformatica, c/o Parco Tecnologico POLARIS, Edificio 1, Loc. Piscina Manna, 09010 Pula, Italy Received 31 March 2009; received (revised) 5 October 2009; accepted 5 October 2009 Published online 24 May 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/mas.20278 Proteomes are intricate. Typically, thousands of proteins interact through physical association and post-translational modifications (PTMs) to give rise to the emergent functions of cells. Understanding these functions requires one to study proteomes as ‘‘systems’’ rather than collections of individual protein molecules. The abstraction of the interacting proteome to ‘‘protein networks’’ has recently gained much attention, as networks are effective representations, that lose specific molecular details, but provide the ability to see the proteome as a whole. Mostly two aspects of the proteome have been represented by network models: proteome-wide physical protein–protein-binding interactions organized into Protein Interaction Networks (PINs), and proteome-wide PTM rela- tions organized into Protein Signaling Networks (PSNs). Mass spectrometry (MS) techniques have been shown to be essential to reveal both of these aspects on a proteome-wide scale. Techniques such as affinity purification followed by MS have been used to elucidate protein–protein interactions, and MS- based quantitative phosphoproteomics is critical to understand the structure and dynamics of signaling through the proteome. We here review the current state-of-the-art MS-based analytical pipelines for the purpose to characterize proteome-scale networks. # 2010 Wiley Periodicals, Inc., Mass Spec Rev 30: 268–297, 2011 Keywords: mass spectrometry; systems biology; network biology; proteomics; phosphorylation I. INTRODUCTION To understand the functioning of living cells requires one to study them as systems rather than as a collection of individual molecules. The study of systems that consist of thousands of interacting molecular species is daunting, and simplifying abstractions are necessary. The abstraction of intracellular processes into ‘‘networks’’ is particularly fruitful (Oltvai & Barabasi, 2002; Barabasi & Oltvai, 2004). Networks provide a clear representation of complicated relationships among large numbers of elements, and are used in scientific disciplines as diverse as sociology, epidemiology, molecular biology, and physics. This network approach applied to complex bio- molecular systems has led to insights into how such systems could have evolved (Wagner, 2001, 2003), and has shed light on the interplay between structure and function (Oltvai & Barabasi, 2002; Barabasi & Oltvai, 2004; Milo et al., 2004; Alon, 2007; Pieroni et al., 2008). The main goal is to relate the structure, or ‘‘topology,’’ and dynamics of networks to their biological functions. Investigating the global topological organization of networks of proteins will provide insights into functional organization of proteomes. Future advances will enable us to understand diseases in terms of complex networks (Goh et al., 2007; Kann, 2007). Novel experimental strategies that rely on mass spectrometry (MS) have proven to be of crucial importance to gather the relevant information to establish proteome-wide network models. A. Network Terminology Complex Network Analysis is a quantitative framework to investigate large complex networks with techniques from graph theory, statistical physics, dynamical systems, and other fields. Application of this framework to the field of proteomics requires formulating the proteome as networks of proteins. In such networks, ‘‘nodes’’represent the proteins as system components. Nodes are usually visually represented as small circles. The ‘‘edges’’ in such networks represent certain relationships between the nodes, sometimes called ‘‘connections,’’ or ‘‘links.’’ Depending on the nature of the interaction, the edges might be directed, distinguishing between a source and a target, or undirected. In the present context a directed edge could correspond to a signal transfer between proteins: the source node transfers information to the target node, by means of a PTM, such as phosphorylation. Directed edges are often depicted as arrows that start at the source node and end at the target node. An undirected edge could, for instance, represent a physical-binding interaction between two proteins: in this case, there is no direction of information flow. Undirected edges are simply visualized as lines between two nodes. A network with only directed edges is called a ‘‘directed network,’’ one with only undirected edges an ‘‘undirected network,’’ and networks with both types of edges are called ‘‘mixed’’ networks. An edge can Mass Spectrometry Reviews, 2011, 30, 268– 297 # 2010 by Wiley Periodicals, Inc. ———— *Correspondence to: Alberto de la Fuente, CRS4 Bioinformatica, c/o Parco Tecnologico POLARIS, Edificio 1, Loc. Piscina Manna, 09010 Pula, Italy. E-mail: alf@crs4.it