I have a more fundamental question about validation of the biological significance of protein-protein interactions. My colleague recently found an interesting protein-protein interaction confirmed by diverse pull down assays and he speculated that interaction between these proteins, letís call them A and B confers an interesting phenotype in vivo (the phenotype is triggered by overexpression of protein A). By analysing the aa of protein A at the N-terminus that interact with protein B he generated a N-terminally truncated protein A that cannot bind to protein B as confirmed by GST pulldown. He then made transgenic mice overexpressing the truncated protein A and surprisingly the mice lost the phenotype. For him this result was proof that the binding interaction between protein A and B is responsible for the phenotype.
In the meantime I generated transgenic mice expressing protein A with a single point mutation within the binding region that abolishes interaction with protein B as well. The block of interaction between protein A and B was confirmed by co-IP using tissue from my transgenic mice so there is no doubt the interaction is abolished in vivo. Unexpectedly, my transgenics retained the phenotype and seem to disprove his theory. Obviously there is now a controversy about how both results should be interpreted and possibly reconciled. Although I am not really an expert in protein biology in my opinion the interaction between protein A and B is insignificant for the phenotype. It is plausible that my colleague changed the conformation of protein A by the applied truncation in a way that substantially altered its function. I would appreciate any other thoughts on this awkward topic.
Thank you very much in advance.