Differential contributions of pre- and postsynaptic components in tuning high-frequency short-term synaptic plasticity

Activity-dependent plasticity of synaptic transmission is a key mechanism underlying learning and memory. During high-frequency short-term synaptic plasticity (HF-STP), the amplitude of synaptic responses changes upon presynaptic stimulation on a timescale of milliseconds. HF-STP is important for information processing in the brain, serving particularly for temporal integration. However, the precise functions of HF-STP, and its impact on information processing, remain unknown. It is widely accepted that HF-STP is regulated primarily by presynaptic mechanisms. However, postsynaptic mechanisms have been shown to also regulate HF-STP, although their role here remains to be fully understood. Previously, we demonstrated that AMPA receptors (AMPARs), the main excitatory receptors in the brain, are trafficked between the extrasynaptic and synaptic compartments through surface diffusion that complements endo- and exocytosis. Here, I studied the functional role of AMPAR surface diffusion in HF-STP in integrated slice tissue models with intact synaptic connectivity. I use the AP-GluA2 knock-in (KI) mouse model, we developed, where GluA2 subunits of AMPARs are tagged with a 15 amino acid biotinylation acceptor peptide (AP-tag) and can be specifically biotinylated when co-expressed with an endoplasmic reticulum resident biotin ligase (BirAER), and immobilized on the cell surface with a biotin-binding protein NeutrAvidin. Using this toolset, I show that immobilization of endogenous AMPARs modulates HF-STP by increasing synaptic depression in the Schaffer collateral-CA1 synapse of organotypic hippocampal slices. This effect is reversed when AMPAR desensitization blockers are applied, suggesting that the modulation of HF-STP is achieved by preventing the replacement of desensitized AMPARs in the synapse. Moreover, when imaging glutamate release with the iGluSnFr sensor we find no change in presynaptic glutamate release upon AMPAR immobilization. Altogether this strongly suggests a postsynaptic contribution of AMPAR mobility in regulating HF-STP. Surprisingly, AMPAR cross-link has no effect on HF-STP in SC-CA1 synapses in ex vivo brain slices. Next, I demonstrate that AMPAR immobilization strongly increases synaptic depression in the L4-L2/3 synapse in the primary somatosensory cortex (S1) of ex vivo brain slices, with no consecutive change in glutamate release measured with iGluSnFR. We find that the rate of AMPAR surface diffusion is higher in the L2/3 than CA1 synapses. This points to a synapse-specific effect of AMPAR mobility on shaping HF-STP that depends on the mutual contribution of presynaptic glutamate release, AMPAR desensitization and surface diffusion. I demonstrate that GSG1L, an AMPAR auxiliary protein expressed highly in the cortex but not CA1, can regulate AMPAR surface diffusion and HF-STP. Next, we show that AMPAR mobility tunes synaptic calcium integration and network activity. Finally, we demonstrate that AMPAR immobilization downstream of paCaMKII activation and LTP induction results in increased synaptic depression in intact tissue. In the last part of the project, I show that AMPAR immobilization can also produce an opposing effect and promote synaptic facilitation, an effect found in juvenile SC-CA1 synapse as well as the lateral perforant path to granule cell synapse in the adult dentate gyrus. This finding further emphasizes the synapse and development specific effects of AMPAR immobilization on HF-STP. Altogether, I determined the respective contributions of presynaptic transmitter release, postsynaptic AMPAR biophysics and mobility in HF-STP and identified physiological processes which act upon AMPAR kinetics and mobility to regulate HF-STP in the brain. Moreover, AMPAR cross-link is a promising tool to achieve cell-specific blockade of HF-STP that may allow the transition from modeling-based evidence of HF-STP roles in brain function to experimental evidence.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.