Abstract 4016: Large scale, domain-based comparative protein interaction mapping in cancer cell signaling

Abstract

Abstract Protein interaction mapping has proven instrumental for the delineation and understanding of cancer signaling pathways, both in Homo sapiens and model organisms. We have recently published a drosophila protein interaction map centered on fly orthologs of human cancer-related and signaling proteins1. We now report the completion of the cognate human interaction map using the same high-throughput, domain-based yeast two-hybrid (Y2H) technology. 149 human homologs of the 102 drosophila proteins, including 29 human oncogenes and tumor suppressors2, were used as entry points to screen a highly complex, random-primed placenta cDNA library comprised of 10 million independent fragments in yeast. In order to strictly parallel the drosophila study, the same domains were selected as baits on human orthologs, resulting in 230 different screens. To ensure reproducible and exhaustive Y2H results, the library was screened at saturation using an optimized mating procedure that allowed for testing on average 50 million interactions per screen, corresponding to a 5-fold coverage of the library. As a consequence, multiple, independent fragments of the same interactant could be isolated, enabling the delineation of a minimal interacting domain and the computation of a confidence score. We identified 4,089 protein-protein interactions (PPI) connecting 2,412 human proteins. 1,212 PPI (30%) are of high confidence and connect 998 proteins. Interestingly, 54 additional human oncogenes or tumor suppressors were identified in the screening, and a sub-network of 1089 PPI linking at least one cancer protein was derived from the full map. The features of the drosophila and human parallel interaction maps were compared. Orthology relationships were combined with Y2H data and the conservation of interaction networks across evolution was examined. Experimentally determined interacting domains were also systematically confronted to known functional domains in both species, pointing to a significant enrichment of domains involved in protein binding. Finally, both experimental protein interaction maps were extended by including protein interaction data from the literature. We believe that this protein interaction map centered on human cancer signaling pathways will be an invaluable resource for researchers in the field. References : 1. Formstecher et al., 2005, Genome Research, 15:376 2. Futreal et al., 2004, Nature Review Cancer, 4(3):177 Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 4016.