A model investigation of short-term synaptic plasticity tuned via Unc13 isoforms.

Short-term synaptic plasticity (STP) is a fundamental mechanism of neural computation supporting a variety of nervous system functions from sensory adaptation and gain control to working memory and decision making. At the presynaptic release site, an interplay between distinct (M)Unc13 protein isoforms is suggested to orchestrate depressing and facilitating components of STP. In this study, we introduce a modification of the well-established TsodyksMarkram Model (TMM) for STP. We constrain our model by in vivo intracellular recordings in the olfactory system of the fruit fly, where previous work suggested Unc13A to provide a phasic, depressing and Unc13B a tonic, facilitating release component. A combination of a facilitating and a depressing model component indeed allowed for accurate model fits. Differential knock-down experiments of the Unc13A and Unc13B gene variants provide biological model interpretation, linking the protein-specific molecular mechanisms to synaptic function and STP. Our mathematical formulation of protein-dependent STP can be readily and efficiently used to design biologically realistic spiking neural network models that feature different genetically defined synapse types.