Network Biology Discovers Pathogens’ Contact Points in Host Protein-Protein Interaction Networks

Networks consist of systems’ components, referred to as ‘nodes’, and interactions between them, termed ‘edges’. The understanding of how networks function in the schema of an entire system is fueling the development of network based approaches. Such a framework is pertinent to evaluation of diverse networks including social media-fueled networks, telecommunication networks, banks transaction networks of businesses, information flow on the internet and biological networks. It is becoming increasingly evident that distinctive networks in nature and technology, regardless of their diversity and scale, display similar topological features. Understanding these topological properties and deciphering the network architecture can provide insights into the identification of new community structures, unknown signaling pathways, and novel relationships between host proteins.

Here, we applied a network science framework to identify host proteins targeted by pathogen effectors in two unrelated experimental protein-protein interaction networks. Towards this, we curated a comprehensive dataset of ~4,350 unique phenotypes in Arabidopsis. Unexpectedly, we demonstrated that hubs and bottlenecks are enriched in conditional phenotypes unlike centrality-lethality rule shown for yeast and other systems. We discovered that nodes located in close proximity of the network core are targeted by pathogen effectors and display enrichment in conditional, as well as immune-related phenotypes. Moreover, the average degree, betweenness and information centrality of these proteins is higher than the nodes in the periphery of the network. We also demonstrated plant immune system functions for these nodes. Finally, we expect that this dataset and network science framework will be useful for future studies including transcriptional regulatory networks.