Computational and Molecular Properties of Starburst Amacrine Cell Synapses Differ With Postsynaptic Cell Type

Joseph Pottackal,Jonathan B. Demb,Joshua H. Singer

doi:10.3389/fncel.2021.660773

Joseph Pottackal, Jonathan B. Demb + Show 1 more

Open Access

https://doi.org/10.3389/fncel.2021.660773

Copy DOI

Abstract

A presynaptic neuron can increase its computational capacity by transmitting functionally distinct signals to each of its postsynaptic cell types. To determine whether such computational specialization occurs over fine spatial scales within a neurite arbor, we investigated computation at output synapses of the starburst amacrine cell (SAC), a critical component of the classical direction-selective (DS) circuit in the retina. The SAC is a non-spiking interneuron that co-releases GABA and acetylcholine and forms closely spaced (<5 μm) inhibitory synapses onto two postsynaptic cell types: DS ganglion cells (DSGCs) and neighboring SACs. During dynamic optogenetic stimulation of SACs in mouse retina, whole-cell recordings of inhibitory postsynaptic currents revealed that GABAergic synapses onto DSGCs exhibit stronger low-pass filtering than those onto neighboring SACs. Computational analyses suggest that this filtering difference can be explained primarily by presynaptic properties, rather than those of the postsynaptic cells per se. Consistent with functionally diverse SAC presynapses, blockade of N-type voltage-gated calcium channels abolished GABAergic currents in SACs but only moderately reduced GABAergic and cholinergic currents in DSGCs. These results jointly demonstrate how specialization of synaptic outputs could enhance parallel processing in a compact interneuron over fine spatial scales. Moreover, the distinct transmission kinetics of GABAergic SAC synapses are poised to support the functional diversity of inhibition within DS circuitry.

Highlights

Within a neural circuit, divergence permits the activity of one presynaptic cell to influence multiple postsynaptic cell types in parallel
To evaluate the possibility that GABAergic starburst amacrine cell (SAC) synapses differ systematically according to postsynaptic cell type, we combined optogenetic stimulation of SACs with linear systems analysis of inhibitory postsynaptic currents (IPSCs) evoked in SACs and DS ganglion cells (DSGCs) (Figure 1B)
Evoked IPSCs recorded in either a SAC or a DSGC were quantified using LN cascade analysis, which generates a computational model consisting of two components: (1) a linear filter, which captures kinetic properties of the modeled synapse; and (2) a static nonlinearity, which captures timeindependent properties of the synapse, including rectification and saturation (Figure 2A; Pottackal et al, 2020)

Summary

Introduction

Divergence permits the activity of one presynaptic cell to influence multiple postsynaptic cell types in parallel. To investigate synaptic mechanisms for divergent output signals, we leveraged the well-defined connectivity within the direction-selective (DS) circuit of the mature mouse retina This circuit depends critically on the starburst amacrine cell (SAC), an axon-less, non-spiking interneuron that provides GABAergic inhibition to both neighboring SACs and DS ganglion cells (DSGCs) (Fried et al, 2002; Lee and Zhou, 2006; Kostadinov and Sanes, 2015; Ding et al, 2016), as well as cholinergic excitation to DSGCs but not SACs (Zheng et al, 2004; Lee et al, 2010). Whether GABAergic synapses from SACs onto distinct postsynaptic cell types differ in their computational properties and, if so, whether these differences arise pre- or postsynaptically

Methods

Results

Conclusion