Overview
Protein-protein interaction (PPI) networks are graph representations of the physical contacts between proteins within a cell. Each protein is a node, and each experimentally detected or computationally predicted interaction is an edge. These networks reveal the functional organization of the proteome, proteins that interact often participate in the same biological process, reside in the same cellular compartment, or form stable complexes. By studying network topology, researchers can identify highly connected hub proteins, functional modules, and pathways that are dysregulated in disease. PPI networks integrate data from multiple experimental techniques and are essential for systems biology.
Key Concepts
Experimental sources for PPI data include yeast two-hybrid screening, affinity purification followed by mass spectrometry (AP-MS), and co-immunoprecipitation. Each method captures different aspects of interactions, binary versus co-complex, stable versus transient. Network properties such as degree distribution, clustering coefficient, and betweenness centrality characterize the global architecture. Hub proteins that connect many partners are often essential for cell viability. Functional modules are densely connected subgraphs that correspond to protein complexes or signaling pathways. Diseases frequently arise from mutations that disrupt specific edges or nodes within these networks, a concept known as network medicine.
Practical Protocol
A practical AP-MS workflow begins with the expression of a tagged bait protein in the relevant cell line. The bait is commonly fused with a FLAG, GFP, or StrepII tag using stable transfection or lentiviral transduction. Cells expressing the tagged bait are lysed under gentle conditions - typically 0.5% NP-40 or digitonin buffer - to preserve protein complexes. The lysate is incubated with affinity resin, such as anti-FLAG agarose beads or StrepTactin Sepharose, for 2-4 hours at 4C. After extensive washing with lysis buffer to remove nonspecific binders, bound proteins are eluted using excess FLAG peptide or SDS elution buffer. The eluate is digested with trypsin and analyzed by LC-MS/MS. Raw data are searched against a protein database, and the resulting protein list is filtered to remove common contaminants using the CRAPome repository. Statistical filtering compares prey abundance in bait samples against controls using SAINT scoring, which integrates spectral counts and reproducibility across biological replicates. High-confidence interactors with SAINT probability above 0.95 are used to construct a PPI network in Cytoscape. For BioID experiments, the workflow is similar but uses a promiscuous biotin ligase fused to the bait, which biotinylates proximal proteins in living cells; biotinylated proteins are captured on streptavidin beads, enabling detection of transient and weak interactions. In an exemplary application, AP-MS of the human 26S proteasome identified 33 subunits and 10 assembly chaperones, revealing the biogenesis pathway of this essential complex. BioID has been used to map the nuclear envelope interactome, identifying over 300 proteins and discovering novel components of the nuclear pore complex.
Applications
PPI networks are used to predict protein function by guilt-by-association, an uncharacterized protein interacting with known DNA repair proteins is likely involved in DNA repair. In drug discovery, networks identify disease modules and prioritize therapeutic targets. The field builds on experimental methods such as the yeast two-hybrid system and antigen-antibody interactions for validation. Network analysis also contextualizes data from protein-protein interactions to generate mechanistic hypotheses about cellular regulation.
