Structure-function analysis on the level of individual synapses

Excitatory synapses in the mammalian brain are made on small protrusions of the postsynaptic cell called dendritic spines. Dendritic spines are highly variable in their morphology and in their microanatomy (e.g. presence of subsynaptic organelles). It is unclear whether and how variability in spine morphological and anatomical properties translates into differences in synaptic function. Using two photon imaging, we analyzed how spine properties can affect synaptic signals and the potential for synaptic plasticity at single identified spine synapses. We show that synaptic signals can be tightly regulated on the level of individual synapses and that differences in spine morphology and microanatomy regulate synaptic function. We also provide evidence for the existence of functionally distinct populations of synapses in regard to their potential for synaptic plasticity. The present thesis is subdivided into three main sections. The first section is dedicated to the analysis of the function of specialized subsynaptic organelles in regulating synaptic plasticity. In the second section we studied the impact of spine morphology on synaptic signals and in the third section we examined whether critical proteins can be tagged to individual synapses in response to plasticity inducing stimuli.
In pyramidal cells, only a subset of dendritic spines contains endoplasmic reticulum (ER). Spine ER often forms a ‘spine apparatus’, a specialized organelle with unknown function. It is unclear whether these specialized subsynaptic structures can affect the function of the synapse on the spine head. The possible involvement of spine ER in shaping spine calcium transients, a key trigger for synaptic plasticity, raises the possibility that spine ER could modulate the potential of a given synapse to undergo activity dependent modifications. Using a genetic approach to label the ER in living neurons, we find that the ER preferentially localizes to spines containing strong synapses. We demonstrate that spine ER represents a specialized calcium signaling machinery required for the induction of metabotropic glutamate receptor dependent long term depression at individual synapses. We demonstrate that different subsets of synapses exist in regard to their potential to undergo specific forms of plasticity. Spine ER represents the anatomical correlate for a mechanism by which strong synapses can be retuned in an activity dependent manner.
Dendritic spines are separated from their parent dendrite by a thin spine neck. The spine neck slows down diffusion of molecules from the spine head to the parent dendrite, allowing spine-specific action of second messengers and activated enzymes. The resistance of the spine neck is crucial in determining whether spines can also be considered electrical compartments. Only a high enough spine neck resistance leads to electrical compartmentalization and activation of voltage gated channels in the spine in response to synaptic stimulation. We show that spine neck resistance can change in an activity dependent manner. Using single spine calcium imaging as a reporter of NMDA receptor activation and spine head depolarization, we show that spines can indeed act as electrical compartments. Using pharmacological experiments and modeling, we demonstrate that different voltage dependent channels cooperatively participate in shaping spine head depolarization and spine calcium transients. We also show that in vivo the spine neck resistance is higher compared to the situation in acutely sliced brain tissue, demonstrating that in the living animal a higher fraction of spines can be considered electrical compartments compared to the in vitro situation. We provide strong evidence that the spine neck can profoundly affect synaptic calcium signals. Biochemical and electrical compartmentalization is dynamically regulated in an activity dependent way.
Spine calcium signals can activate key signaling cascades responsible for the induction of synaptic plasticity. Long term potentiation (LTP) has been shown to require the activity of CaMKII, a serine/ threonine kinase. A chemical protocol leading to LTP has been shown to induce translocation of CaMKII to dendritic spines. It is however unclear whether this molecule acts at single synapses or whether it can spread and modulate neighboring synapses in response to more physiological protocols. Using a new optical approach to induce LTP at single visualized synapses, we show that LTP induction is accompanied by a long-lasting increase of CaMKII at the stimulated synapse. This increase was specific to the stimulated spine and did not spread to neighboring spines. We provide evidence that CaMKII acts locally, on the micrometer scale, to regulate plasticity. We show that the concentration of proteins involved in regulating synaptic plasticity can be tightly regulated at the level of single synapses.