PROJECT SUMMARY Genetic mutations that confer autism risk often occur in genes that are expressed at the glutamate synapse. The protein products of these genes form a highly interconnected protein interaction network (PIN), and represent attractive therapeutic targets since they are expressed throughout the lifespan and can be acutely targeted with small molecule drugs. However, the dynamic, network-scale behavior of this PIN in normal or disease states is poorly understood. Here, we apply a novel PIN-mapping technology, quantitative multiplex co-immunoprecipitation, to explore the input-output relationships of an autism-linked PIN at the glutamate synapse as it responds to physiological inputs. Our target system is a 20-member PIN, consisting of glutamate receptors, scaffolds, and signal transduction molecules; mutations in the genes encoding all target proteins have been genetically linked to autism. We first show that, in wild-type animals, our target PIN changes its pattern of co-associations in a stereotyped manner in response to acute stimulation with KCl or glutamate, using cultured neurons or acute slices. We then model the input-output relationships of the PIN system, and demonstrate that the PIN produces specific, recognizable signatures in response to stimulation through the mGluR or NMDA receptors. In the context of physiological glutamate stimulation, the PIN integrates the two inputs to produce a coordinated cellular response- potentiation or de-potentiation. Based on these and other preliminary observations and published data, we propose that mutations that contribute to autism risk disrupt information flow through this PIN, such that the balance between LTP-like potentiation and LTD-like depotentiaion is altered, ultimately leading to an organism-level imbalance between excitation and inhibition. We will test this hypothesis by modeling the PIN response to mGluR or NMDA stimulation in three distinct, well-characterized animal models of autism- the Fragile X knockout, Shank3 knockout, and Ube3a overexpressing models. We will characterize the input-output relationships for mGluR or NMDA stimulation, and mathematically model their integration using a vector transformation model in principal component space. We will define specific mechanisms by which autism-linked mutations disrupt either input-output relationships, or disrupt signal integration in the context of physiological stimulation. In addition, we will treat two of our animal models (Shank3 and Fragile X) with drugs that have been previously demonstrated to rescue autism- like behaviors. We will model the response of the PIN to the drug with or without concurrent stimulation to define a PIN signature associated with behavioral rescue. In summary we propose to (1) define normal information flow through a PIN consisting of the protein products of autism-liked genes; (2) define how information flow is disrupted in mouse models of autism, with the goal of understanding the system...