Structural remodeling of neural circuits via multisynaptic boutons

About This Grant

The overarching goal of this collaborative program is to understand how circuits in the mammalian brain are reorganized to encode new memories. We address this problem through detailed reconstruction of neuronal connectomes, single synapses, and glia at nanometer resolution using 3D Electron Microscopy (3D-EM). We combine 3D-EM with chemogenetic techniques for labeling cellular ensembles recruited for specific cognitive tasks and Artificial Intelligence (AI)-based computational tools for image segmentation. Using this interdisciplinary approach, we began to identify the morphological hallmarks of long-term associative memory in the mouse hippocampus. Our recent studies of the canonical CA3-CA1 pathway revealed principles by which pyramidal glutamatergic neurons (PNs) engaged during fear learning modify their local wiring diagrams, synaptic weights, and membrane organelles essential for energy metabolism and intracellular calcium buffering. Despite their broad physiological implications, these structural correlates of information storage share three features: (1) Their induction requires presynaptic activity elicited by sensory stimuli with negative valence; (2) Their manifestation transcends co-activated neurons; and (3) They involve multi-synaptic boutons (MSBs), atypical connections capable of simultaneously relaying neurotransmitter signals from one axonal terminal to several independent dendritic spines. Contrary to common dogma, we found that the initial cellular substrates of memory traces expand their connectivity via MSBs, thereby recruiting new neurons into the network while preserving the stable arrangements of individual synaptic sites on axons and dendrites. Taken together, these observations support the hypothesis that MSBs are pivotal for memory storage and that the structural plasticity of neural ensembles representing engrams does not adhere to traditional Hebbian rules. Our studies provide the first mechanistic explanation for representational drifts, a non-Hebbian phenomenon suggesting that population coding of a particular experience is not fixed over time. We will test out central hypotheses in the following specific aims: Aim 1. Investigate the spatiotemporal dynamics of non-Hebbian network remodeling via MSBs. We will determine if the synaptic architectures of an associative memory engram are reconfigured through MSBs globally or in a circuit-specific manner and will investigate the temporal dynamics of this process. Aim 2. Explore the physiological mechanisms of MSB morphogenesis. We will determine how the organization of MSBs reflects memory strength and will test if MSB morphogenesis is regulated by de novo protein synthesis, transcription, and synaptic activity. Aim 3. Define the composition of MSBs and their local microenvironment. We will comprehensively dissect the fine-scale architecture of MSBs and their postsynaptic partners. These analyses will involve reconstructions of active zones, PSDs, vesicles, other intracellular membrane organelles, astrocytes, and microglia.