Functional and structural properties of strong inhibitory synapses in a binaural auditory circuit

Neuronal circuits are defined in their function and computational outcome by the synaptic interplay of Excitation and Inhibition (E/I). One step further in understanding the role of these opposite interactors in neuronal computations consists in determining their relative contribution, in terms of input strength and convergence. In the auditory brainstem, neurons in the Lateral Superior Olive (LSO) rely upon binaural E/I integration for computing sound-source localization. We made use of this circuit as a model for understanding the role of synaptic connectivity in information processing. We characterized the functional connectivity of the inhibitory and excitatory inputs onto a given LSO neuron. We found that LSO neurons receive few but large inhibitory inputs with a unitary conductance of 9 nS (under physiological intracellular Cl- concentration), although the unitary strength of inputs scattered over a 10-fold range. We estimated that unitary inhibitory postsynaptic currents (IPSCs) were mediated by a large number of synchronously released glycine vesicles (on average: 48 ± 39, range 9 - 115). The large quantal content suggests that these IPSCs are generated by a high number of functional release sites. Electron-microscopy based reconstructions revealed the structural specialization of the inhibitory axons which contacted the soma and proximal dendrites of a LSO neuron multiple times via large varicosities (diameter range 1 ¿ 11 µm), each containing on average three active zones. In contrast to the strong unitary IPSCs, excitatory inputs were numerous but had a small unitary conductance (~ 0.7 nS). A computational model of E/I integration showed that the strength and convergence of excitatory inputs determines how spontaneous firing is transmitted from upstream neurons, whereas the strength and convergence of inhibitory inputs determines the dynamic range of the tuning curve. We then aimed to study mechanisms underlying the developmental acquisition of strong inhibitory inputs on LSO neurons. Input-output curves were gradual at post-natal days (P) 5 and 6 but showed, in most cases, clear steps from P7 onwards, indicative of a reduction in the number of synaptic inputs and a growth in the strength of the individual inputs. In addition, we wished to investigate whether Synaptotagmin1 (Syt1) represents the Calcium sensor that guarantees the co-release of glutamate which has been known to occur at the immature inhibitory MNTB - LSO synapses. Genetic deletion of Syt1 caused asynchronous glutamate release only during the first 2 ¿ 3 post-natal days, suggesting that Syt1 might gradually become redundant with another Calcium sensor after ~ P5. In conclusion, this study highlights the functional and structural specializations of inhibitory synapses optimized for fulfilling the computational properties of binaural neurons devoted to sound localization. It also suggests that the basic synaptic connectivity at this connection is laid down in the first week of postnatal development.

RIM Determines Ca²+ Channel Density and Vesicle Docking at the Presynaptic Active Zone

Yunyun Han

Fast neurotransmitter release is essential for neuron-neuron communication and is initiated by the opening of voltage-gated Ca2+ channels close to docked vesicles at the presynaptic active zone. The high concentration of Ca2+ channels at the active zone is important to secure fast transmitter release. However, the mechanism that enriches Ca2+ channels at active zones is largely unknown, maybe because of the limited accessibility of most model synapses to direct measurements of Ca2+ signaling in the nerve terminal. RIM's (Rab3 interacting molecule) are scaffolding proteins in the active zone which interact with several presynaptic proteins. As revealed by previous studies in C. elegans and cultured neurons, RIM proteins affect synaptic transmission. However, their complex functions have not been clearly dissected, because direct access to the presynaptic nerve terminal has so far not been possible in synapses that are amenable to genetic manipulation. Here, we have established a Cre-lox based conditional KO approach at a presynaptically accessible CNS synapse, the calyx of Held. This has allowed us to use presynaptic recordings and Ca2+ uncaging, as well as electron microscopic analyses of synapses which have developed in vivo, to directly study the presynaptic functions of RIM proteins. We discovered three major functions of RIM proteins: First, RIM's hold Ca2+ channels in the presynaptic terminals to secure the high Ca2+ channel density at active zones. Removal of RIM proteins leads to a decrease of presynaptic Ca2+ current amplitude (by ∼ 50%) as well as to a decreased immunostaining signal using an antibody against a Ca2+ channel α-subunit (by ∼ 40%). Second, RIM proteins dock vesicles to the active zone and therefore control the size of readily releasable pool. In RIM1/2 cDKO synapses, the number of docked vesicles is reduced by ∼ 80% and the functional pool size is decreased by ∼ 75%. The good correlation of morphological and physiological data identified RIM as the first presynaptic protein in a CNS synapse with a clearly corresponding role in vesicle docking, and priming. Third, RIM proteins increase the release probability and speed up the release kinetics. This is carried out by two new functions: increasing the intrinsic Ca2+ sensitivity of release and tightly coupling docked vesicles to Ca2+ channels. Taken together, this study shows that RIM proteins function as the central organizers to coordinate active zone function: holding the Ca2+ channels in the presynaptic terminal, docking vesicles to the active zones, and regulating the release probability of any given readily-releasable vesicle. All the three functions together secure the fast vesicle release when the AP arrives in the presynaptic terminal.

EPFL2011

Structural basis and biophysical properties of synaptic plasticity studied by atomic force microscopy

Alexandre Yersin

In this thesis, we studied two systems important for synaptic plasticity, one presynaptic and another postsynaptic. The protein complex composed of VAMP 2, SNAP-25 and syntaxin 1 (SNARE complex) is essential for docking and fusion of neurotransmitter-filled synaptic vesicles with the presynaptic membrane. In order to better understand the fusion mechanisms, we reconstituted the synaptic SNARE complex in the imaging chamber of an atomic force microscope (AFM) and measured the interaction forces between its components. Each protein was tested against the two others, taken either individually or as binary complexes. This approach allowed us to determine specific interaction forces, dissociation kinetics of the SNAREs and led us to propose a sequence of formation for the complex. A theoretical model based on our measurements suggests that a minimum of four complexes is probably necessary for fusion to occur. We also showed that the regulatory protein nSec1 injected into the AFM chamber prevented the SNARE complex formation. Finally we measured the effect of tetanus toxin protease (TeTx) on the SNARE complex and its activity by on-line registration during TeTx injection. These experiments reveal that AFM is a valuable tool to investigate complicate interactions among a protein complex containing up to four components. Trafficking of glutamate receptors AMPA is closely related to postsynaptic induction of long-term potentiation. In a second part of this thesis, we used AFM to explore the distribution of AMPA receptors on-line at the surface of living cells in correlation with cellular viscoelastic properties. First, we showed that AFM permits to detect specifically AMPA receptors at the surface of living neurons and we proved that the specificity of detection is stable during time. Moreover, we were able to visualize variations in AMPA receptor density on surface in response to different stimulations. Finally, AFM also allowed us to study local viscoelastic properties of the cell and their modifications occurring during AMPA receptor trafficking. Our measurements reveal that AMPA receptors are situated at the cell surface on local stiffness maxima. We have noticed that an excitatory stimulation removes these stiffness maxima before producing internalization of receptors. Our results suggest also that AMPA receptors may be constituted of two subpopulations depending on the cell stiffness. A "hard" subpopulation may be preferentially internalized after an excitatory stimulation whereas a "soft" one seems to remain at the cell surface.

EPFL2004

RIM Determines Ca²+ Channel Density and Vesicle Docking at the Presynaptic Active Zone

Yunyun Han

EPFL2011

Structural basis and biophysical properties of synaptic plasticity studied by atomic force microscopy

Alexandre Yersin

EPFL2004