Dendritic spines as chemical and electrical compartments : a two-photon imaging study in the hippocampus of the rat

Most excitatory synapses are located on small dendritic protrusions called spines. So far, the function of dendritic spines is not fully understood. Imaging experiments have shown that spines can compartmentalize second messengers such as Ca2+. In addition to chemical compartmentalization, spines could play a role in shaping excitatory postsynaptic potentials (EPSPs) by activating voltage-dependent conductances in the spine head, thus serve as electrical amplifier. The electrical resistance of the spine neck is essential for influencing synaptic potentials. We measured the diffusion coupling between spine heads and parent dendrites of CA1 pyramidal cells using fluorescence recovery after photobleaching (FRAP) to estimate the resistance of the spine neck. Our data indicate that the diffusional coupling between spine and parent dendrite is highly plastic. Postsynaptic depolarization led to dramatic reduction in the diffusional coupling between spine head and parent dendrite, indicating a proportional rise in the electrical resistance. But is the ohmic resistance sufficent to electrically isolate the synapse? We used two-photon Ca2+ imaging combined with modeling to address this question. We found two different classes of synapses in the CA1 region of the hippocampus. One class producrd clearly detectable Ca2+ transients in current clamp (functional spines), whereas an other class showed hardly any Ca2+ influx under current clamp conditions (‘Ca2+-silent spines’). Interestingly, this group of Ca2+-silent spines showed Ca2+ responses following a brief burst of presynaptic action potentials that were much larger than in functional spines, indicating differences in both presynaptic release properties and postsynaptic receptor densities. The Ca2+ transients in the funciotnal spines were inhibited by blocking either NMDA receptors, AMPA receptors, or R-type Ca2+ channels. We concluded that Ca2+ transients were dependent on the joint activation of these channels, which all contributed to spine depolarization in a synergistic fashion. To estimate the depolarization in individual functional spines, we used the voltage-dependence of the NMDA receptors. Two-photon imaging allowed us to measure NMDA receptormediated Ca2+ currents in individual spines. The voltage-dependence of synaptic NMDA receptors was steeper than previously thought. Using the Ca2+ imaging data
combined with modeling, we predicted that EPSPs reach amplitudes of ~55 mV in functional spines, approaching the synaptic reversal potential. Functional spines are electrically isolated from the parent dendrite by a high resistance neck and amplify synaptic currents through the activation of high-voltage activated Ca2+ channels.
The spine neck resistance appears to have a strong effect on the Ca2+ transients needed to induce synaptic plasticity. Our biophysical model predicted that only spines with high spine neck resistance experience supralinear Ca2+ transients after paring presynaptic activity with a postsynaptic action potential. Furthermore, spines with a high spine neck resistance are more sensitive to the precise timing the presynaptic and the postsynaptic action potentials then in spine with low neck resistance.
By integrating data from diffusion measurements, calcium imaging and pharmacology into a single quantitative model, we have gained new insights into the complex interaction between chemical and electrical signaling at individual synapses. The newly discovered spine neck plasticity might be an important mechanism to set the threshold for the induction of functional synaptic plasticity. To test this hypothesis in the future, new methods are currently being developed to quantify plasticity on the single-synapse level.