Purification and proteomic profiling of PSD-95 interactors from in vivo potentiated synapses

The acquisition of new memories is accompanied by long-lasting modifications in the strength of information transmission between neurons i.e., synaptic plasticity. The structural substrate for synaptic plasticity is the reorganization of the protein content of the synapses. This includes changes in the subunit composition of pre-existing complexes, synthesis or accumulation/relocalization of new proteins, and the formation of new interactions, which can be induced or stabilized by posttranslational modifications (Herring & Nicoll, Ann Review Physiol 2016;78, 351–65). The postsynaptic density (PSD) is a hub of these processes; indeed, PSD-95 interacts with many structural proteins (Shank3, PSD-93) and effectors (AMPARs, NMDARs, CaMKII) (Okabe, Mol Cell Neurosci 2007;34, 503–18). While the set of PSD-95 interactors in the forebrain has been defined by means of proteomic analysis (Fernández et al., Mol Syst Biol 2009;5, 269), to date it has not been possible to describe how the interactome of PSD-95 changes in response to synapse potentiation. To fill this gap, we exploited the SynActive toolbox, which we recently developed to achieve specific expression of proteins of interests at potentiated synapses (Gobbo et al., Nat Comm 2017;8, 1629), to selectively purify the interactors of PSD-95 from potentiated synapses in vivo. A SynActive-controlled, FLAG-tagged version of PSD-95 was delivered via AAV to the hippocampus of mice, which were subsequently challenged with contextual fear conditioning. PSDs were then isolated by affinity purification and their protein content analysed by mass spectrometry. As a reference set, we analysed the interactomics of FLAG-tagged PSD-95 constitutively expressed in home caged animals. Our proteomics data, validated by Western Blot, are the first report of an unbiased comparison of the protein content of the PSD between unstimulated and potentiated synapses. Our data provide a deeper understanding of the mechanisms acting in concert at the synapse during learning. This approach can be applied also to mouse models for neurodegenerative pathologies, like Alzheimer’s disease, to look for structural alterations at precocious stages of the disease, thus helping in the search for early therapeutic interventions.