Molecular surgery: playing with network edges

ResearchBlogging.orgProtein X interacts with protein Y, what are the phenotypic consequences? And what is the impact of the X-Y partnership in the whole protein-protein interaction network? To address this question, scientists often remove specific network nodes by eliminating (knock-out) or downregulating (knock-down) the gene encoding one protein product (i.e. X). This is a poor strategy, because usually X interacts not only with Y, but also with P, Q, R, S, T, U, V, W and Z. Thus, X-KO strategy is too much invasive and prone to promptly activate compensatory mechanisms in the newtork redundancy program, making the phenotypic interpretation a pain in the ASS. To better understand networks, one should be less invasive and knock-down edges other than nodes.
This new 'molecular edge surgery' can be done now, thanks to a reporter gene: the yeast Cytosine Deaminase (yCD) and to Po Hien Ear and Stephen Michnick and their Nature Methods paper. The yCD reporter can be succesfully splitted and used in classic protein-fragment complementation assays (PCA). The measure of reporter activity in this case is life or death, as yCD allows both positive and negative selection in yeast. So, if protein X interacts both with Y and Z, the molecular edge surgery is done performing first the selection with a library of mutant alleles X* screened against Z (selected for death: disruption of X-Z interaction), the same mutants are then screened against Y (selected for growth: retaining X-Y interaction).
At the end of the streamlined selection one should obtain 'edgetic' mutants preserving the X-Y edge but eliminating the X-Z edge of the protein-protein interaction network. The yCD-PCA assay should provide more granular insights into the relationship between loss of a specific edge on a network and the Phenotype and eventually trace those consequences back to the protein structure. This seems also a beautiful technique for prototyping drugs able to antagonize one 'bad' protein-protein interaction, while maintaining the 'good' one.

---/ citation /--- --- ---
Ear, P., & Michnick, S. (2009). A general life-death selection strategy for dissecting protein functions Nature Methods, 6 (11), 813-816 DOI: 10.1038/NMETH.1389