Number and relative abundance of synaptic vesicles in functionally distinct priming states determine synaptic strength and short-term plasticity.

Heterogeneity in synaptic strength and short-term plasticity (STP) was characterized in post-hearing rat calyx of Held synapses at near-physiological external [Ca] under control conditions and after experimentally induced synaptic potentiation. Kinetic modelling was combined with non-negative tensor factorization (NTF) to separate changes in synaptic vesicle (SV) priming kinetics from those in SV fusion probability (p ). Heterogeneous synaptic strength and STP under control conditions can be fully accounted for by assuming a uniform p among calyx synapses yet profound synapse-to-synapse variation in the resting equilibrium of SVs in functionally distinct priming states. Although synaptic potentiation induced by either elevated resting [Ca], elevated external [Ca] or stimulation of the diacylglycerol (DAG) signalling pathway leads to seemingly similar changes, that is, stronger synapses with less facilitation and more pronounced depression, the underlying mechanisms are different. Specifically, synaptic potentiation induced by the DAG mimetic and Munc13/PKC activator phorbol 12,13-dibutyrate (PDBu) only moderately enhances p but strongly increases the abundance of fusion-competent maturely primed SVs, demonstrating that the dynamic equilibrium of differentially primed SVs critically determines synaptic strength and STP. Activation of the DAG pathway not only stimulates priming at resting [Ca] but further promotes SV pool replenishment at elevated [Ca] following pool-depleting stimulus trains. A two-step priming and fusion scheme which recapitulates the sequential build-up of the molecular SV fusion machinery is capable of reproducing experimentally induced changes in synaptic strength and STP in numerical simulations with a small number of plausible model parameter changes. KEY POINTS: A relatively simple two-step synaptic vesicle (SV) priming and fusion scheme is capable of reproducing experimentally induced changes in synaptic strength and short-term plasticity with a small number of plausible parameter changes. The combination of non-negative tensor factorization (NTF)-decomposition analysis and state modelling allows one to separate experimentally induced changes in SV priming kinetics from those in SV fusion probability. A relatively low sensitivity of the SV priming equilibrium to changes in resting [Ca] suggests that the amplitude of the 'effective' action potential (AP)-induced Ca transient is quite large, likely representing contributions of global and local Ca signals. Enhanced synaptic strength and stronger depression after stimulation of the diacylglycerol (DAG) signalling pathway is primarily caused by enhanced SV priming, leading to increased abundance of maturely primed SVs at rest with comparably small changes in SV fusion probability. Application of DAG mimetics enhances the Ca-dependent acceleration of SV priming causing a faster recovery of synaptic strength after pool-depleting stimuli.