Optical analysis of synaptic plasticity in rat hippocampus

Abstract

Long-term potentiation (LTP) in the CA1 region of the hippocampus is dependent on NMDA receptor activation. Downstream of NMDA receptor signaling, the activation of α-calcium/calmodulin-dependent protein kinase II (αCaMKII) is both necessary and sufficient for the induction of this form of LTP. It has been shown that αCaMKII accumulates in spines after glutamate application or ‘chemical LTP’. This postsynaptic accumulation of αCaMKII could be a key step for the induction of LTP, because it localizes the activated kinase close to the substrates of synaptic potentiation. It is not clear, however, what the threshold, time course of αCaMKII translocation are, and whether it is specific to the stimulated synapses only. To address these three questions, I combined optical stimulation techniques (Channelrhodopsin-2 stimulation and two-photon glutamate uncaging) with optical measurements of calcium transients and αCaMKII concentration. This ‘all-optical’ approach made it possible to investigate synapse-specific changes during the induction of LTP. I could show that coincident activity of pre- and postsynaptic cells was needed to trigger the translocation of αCaMKII. Functional potentiation could be measured immediately after stimulation, whereas αCaMKII accumulation reached its peak ~10 min later. This points to an additional structural role of αCaMKII at the postsynaptic density. Both αCaMKII fractions, the cytoplasmic fraction and postsynaptic bound αCaMKII, increased after optical LTP induction. These changes were restricted to stimulated spines. In spines that showed a persistent volume increase, the amount of bound αCaMKII was increased by a factor of two after 30-40 minutes. A second very interesting finding was the close correlation between spine volume changes and LTP, in terms of the time course, induction threshold and specificity. The optical LTP protocol led to a lasting volume increase only in the stimulated spines, but not in directly neighboring spines on the same dendrite. Spine volume reached its maximum immediately after stimulation. Since my all-optical approach relied heavily on the use of a newly identified light-gated cation channel (Channelrhodopsin-2, ChR2), I finally also characterized light activation of ChR2 in hippocampal pyramidal cells in detail. Neuronal activity could be controlled by blue light with millisecond precision. No direct activation of ChR2 was observed by two-photon imaging lasers, making it possible to combine the ChR2 stimulation technique with two-photon imaging. This led to a third important finding: the release probability of ChR2-expressing axonal terminals was increased if the action potential was induced by light. As a result, pairing of light stimulation with postsynaptic depolarization induced reliable long-term potentiation at CA1 synapses. In summary, the new all-optical approach that combines ChR2 stimulation, two-photon glutamate uncaging, and optical measurements of calcium transients and protein concentration, provides a new avenue for investigating plasticity at the level of single synapses. The induction of LTP in single synapses revealed that accumulation of αCaMKII is input specific thus validating Hebb’s postulate on a micrometer scale.