Mapping potentiated synapses in the hippocampus during acquisition and consolidation of a contextual fear memory.

The acquisition of new information correlates with synaptic plasticity; in particular, potentiation of synaptic transmission plays a key role in the encoding of new memories (Rogerson et al., Nat Rev Neurosci 2014;15, 157–69). This leads to hypothesize that synaptic engrams, as opposed to cellular engrams, can represent the physical information storage unit of the brain. So far, research has focused on cellular engrams, mainly by detecting the activity-dependent expression of immediate early genes, such as c-fos (Kubik et al., Learn Mem 2007;14, 758–70; Ramirez et al., Front Behav Neurosci 2013;7, 226). Until recently, extending activity mapping to the synaptic level was hindered by the lack of a suitable tool. To this end, we designed the SynActive toolbox, and achieved tagging and labelling of potentiated synapses in vivo (Gobbo et al., Nat Comm 2017;8, 1629). Here, we applied SynActive to the cartography of synapse potentiation during the encoding of a contextual fear memory. A SynActive-controlled fluorescent reporter was delivered to the hippocampus CA1 via triple-electrode in utero electroporation; its expression was regulated by doxycycline. This allowed us to compare the distribution of potentiated synapses during the encoding (<90 min from the task) and the subsequent consolidation phases. Using a c-fos reporter, we simultaneously identified engram cells and engram synapses. Neurons activated during the encoding of a memory have a higher probability to be reactivated during the recall; however, the two sets do not overlap completely (Roy et al., Cell 2017;170, 1000–12; Ramirez et al., Science 2013;341, 387–91). Therefore, we also mapped c-fos activation the day after presenting the animal the conditioned context (recall) or an unrelated context, to understand if the number of potentiated synapses is predictive of the reactivation chance. Taking into account that consolidation occurs on a longer time scale than acquisition (van de Ven et al., Neuron 2016;92, 968–74) and based on our maps of potentiated synapses, we propose a two-stage model characterized by two epochs of synapse potentiation. Our findings provide new information on the establishment and maturation of a memory trace, and can be applied to understand the information flow during memory maturation.