Technologies for Proteome-Wide Discovery of Extracellular Host-Pathogen Interactions

Martinez-Martin, Nadia

As an addendum to the vast amount of knowledge acquired using MS approaches and some of the additional methodologies discussed in this review, bioinformatics offers an in silico systems biology approach that reveals a global perspective on host-pathogen interactions. Advances in computation have been fundamental to dissect the complex datasets generated in many genome-wide MSbased studies and have enabled the reconstruction of largescale host-pathogens PPI networks, providing fundamental insights into viral disease and hence host biology [15-18, 34-36, 117] . Although the computational tools available for analysis of large datasets, in some cases developed in association with some of the high throughput screens mentioned in this review, certainly deserve a focused chapter, a couple of observations are specially notable. Commonly observed in these studies is that the intracellular viral effectors preferentially target host proteins that act as hubs (proteins with many interacting partners) or bottlenecks (proteins central to many pathways in the network) [15, 16, 36] . For example, Dyer and colleagues built a network of host-pathogen PPI by integrating published information from 190 pathogens [36] . Supporting previous findings, this analysis indicated that pathogen-encoded proteins preferentially interfere with host molecules that control critical cellular processes, such as cell death or nuclear transportation, possibly as a strategy to maximize control of the host machinery given limited genomic resources. Interestingly, this study highlighted a small set of extracellular host proteins recurrently targeted by several of the viral and bacterial pathogens analyzed, including cell surface receptors such as VEGFR2/KDR and collagen, possibly indicating previously unrecognized roles in the immune response against pathogens. Although informative, the analysis performed by Dyer and collaborators was skewed towards viruses, with a prominent enrichment in HIV strains [36] . More extensive analyses encompassing other human viruses and bacterial pathogens may reveal general strategies of immunomodulation and potential human targets suitable to therapeutic intervention. Interestingly, increasing evidence suggests that virus-host interactions are governed by principles distinct to those that dictate within-host interactions [20, 28, 85, 87, 118] . Notably, detailed analyses carried out by the Xia group highlighted significant differences between virushost and within-host (also called endogenous) interactions, such as the tendency of viral proteins to compete with host proteins for binding to a given receptor in the absence of sequence similarity with the host counterpart or the observation that viral molecules have evolved multiple short linear motifs capable of mediating a number of diverse interactions [20, 118] , features that are consistent with the multifunctional capabilities of some pathogen-encoded proteins [20, 22, 28, [77] [78] [79] [80] 118] . Altogether, bioinformatics analysis of virushost interactions suggest that virus-mediated targeting of host proteins is characterized by signatures of pleiotropy, economy, and convergent evolution, conclusions that are supported by emerging experimental data. Followed by thorough biological experimentation such computationalbased systems biology approaches will provide a unique tool to help decipher basic global principles of pathogenhost interaction and may reveal novel ePPIs amenable to therapeut