Tagging of potentiated synapses with a locally translated optogenetic reporter

Increasing evidence points to the importance of dendritic spines in the formation and allocation of memories, and alterations of spine number and physiology are associated with memory and cognitive disorders. Modifications of the activity of subsets of synapses are believed to be crucial for memory establishment. Indeed, treatments or conditions that affect synaptic potentiation almost inevitably lead to an impairment in the acquisition of memories. Thus, the potentiation of synaptic transmission is a likely correlate of the acquisition of associative memories. This suggests that, in the brain, memories are formed and stored as a persistent modification of the strength of a subset of synapses within a neural circuit. To test this hypothesis, it would be necessary to re-excite such a subset of synapses and observe a behavioural response coherent with the encoded memory. A precondition for this experimental test is the ability to detect potentiated synapses during the encoding phase. However, the development of a method to directly test this hypothesis is currently lacking. Here I developed a hybrid RNA/protein approach to selectively express proteins at potentiated synapses (SynActive). To open the possibility for recalling synaptic activity, SynActive can be used to express a light-sensitive membrane channel of the Channelrhodopsin2 family (SA-Ch). After the characterization of SA-Ch using primary neuronal culture SA-Ch was expressed in the hippocampus of living mice. By controlling the timing of its expression with doxycycline, I was able to tag and identify potentiated synapses during the exposure of mice to a novel context. This allowed the identification of a candidate synaptic engram encoding the representation of the explored context. Furthermore, SA-Ch was used to map the relative distribution of potentiated synapses within and across neurons. In both the hippocampal CA1 and dentate gyrus regions, I demonstrate the presence of clusters of potentiated synapses, whose dimension (i.e. number of synapses) is increased by the behavioural task. I also highlight differences between CA1 and DG in terms of distribution of potentiated synapses between different neurons. The results provide strong experimental evidence in support of proposed models for memory acquisition and activity flow in the hippocampal circuit. The SynActive approach can then be used to map potentiated synapses in the brain and will make it possible to re-activate neurons only at previously activated synapses, testing if the activity of the identified synaptic trace is sufficient to recall the memory of the encoded representation. Furthermore, it will provide an experimental way to expand the current neuron-tagging technologies in the investigation of memory processes to the synaptic level.