GABAergic innervation organizes synaptic and extrasynaptic GABAA receptor clustering in cultured hippocampal neurons

Abstract

We have studied the effects of GABAergic innervation on the clustering of GABA(A) receptors (GABA(A)Rs) in cultured hippocampal neurons. In the absence of GABAergic innervation, pyramidal cells form small (0.36 +/- 0.01 micrometer diameter) GABA(A)R clusters at their surface in the dendrites and soma. When receiving GABAergic innervation from glutamic acid decarboxylase-containing interneurons, pyramidal cells form large (1.62 +/- 0.08 micrometer breadth) GABA(A)R clusters at GABAergic synapses. This is accompanied by a disappearance of the small GABA(A)R clusters in the local area surrounding each GABAergic synapse. Although the large synaptic GABA(A)R clusters of any neuron contained all GABA(A)R subunits and isoforms expressed by that neuron, the small clusters not localized at GABAergic synapses showed significant heterogeneity in subunit and isoform composition. Another difference between large GABAergic and small non-GABAergic GABA(A)R clusters was that a significant proportion of the latter was juxtaposed to postsynaptic markers of glutamatergic synapses such as PSD-95 and AMPA receptor GluR1 subunit. The densities of both the glutamate receptor-associated and non-associated small GABA(A)R clusters were decreased in areas surrounding GABAergic synapses. However, no effect on the density or distribution of glutamate receptor clusters was observed. The results suggest that there are local signals generated at GABAergic synapses that induce both assembly of large synaptic GABA(A)R clusters at the synapse and disappearance of the small GABA(A)R clusters in the surrounding area. In the absence of GABAergic innervation, weaker GABA(A)R-clustering signals, generated at glutamatergic synapses, induce the formation of small postsynaptic GABA(A)R clusters that remain juxtaposed to glutamate receptors at glutamatergic synapses.

Fig. 1.

GABAergic innervation induces the formation of large GABA_AR clusters at GABAergic synapses. Smaller GABA_AR clusters are also present outside GABAergic synapses. Hippocampal neurons after 19–22 d in culture were immunolabeled with the rabbit anti-GABA_A receptor subunit γ₂ (A, D, G,J), mouse monoclonal anti-gephyrin (B, E), sheep anti-GAD (C,F, I, L), guinea pig anti-α₁ (H), or mouse monoclonal anti-β_2/3 (K).D–F show at high magnification the fields in A–C corresponding to theboxed area in A, respectively. Large clusters of GABA_A receptors (D,G, H, J, K,arrows) and gephyrin (E,arrows) colocalize with GAD-positive boutons (F, I, L,arrows), whereas small clusters of GABA_ARs (D, G, H,J, K, filled arrowheads) colocalize with small clusters of gephyrin (E,filled arrowheads) but not with GAD boutons. A number of small GABA_AR clusters not colocalizing with GAD contained only one of the two subunit classes (G,H, J, K, empty arrowheads). Scale bar (shown in A):A–C, 10 μm; (shown inD): D–L, 5 μm.

Fig. 2.

Various α subunit isoforms of the GABA_AR colocalize in all large GABAergic synaptic clusters, but not in all smaller non-GABAergic clusters. Hippocampal pyramidal neurons were triple labeled with the GABA_A subunit isoform-specific antibodies guinea pig anti-α₁(A, D), rabbit anti-α₂(B), or rabbit anti-α₃(E) in conjunction with sheep anti-GAD (C, F). All of the larger GAD-colocalizing GABA_AR clusters (A,B, D, E,arrows) and many smaller clusters (A,B, D, E, filled arrowheads) that do not colocalize GAD show the presence of the two α subunit isoforms. However, a significant population of the smaller clusters contained only one of the two α−subunit isoforms (A, B, D,E, empty arrowheads). Neurons were cultured for 19 d. Scale bar (shown in A): 5 μm.

Fig. 3.

Various β subunit isoforms of the GABA_AR colocalize in all large GABAergic synaptic clusters and most small non-GABAergic clusters. Cultured hippocampal neurons were triple labeled with the GABA_AR subunit isoform-specific antibodies rabbit anti-β₁(A), mouse monoclonal (62–3G1) anti-β_2/3 (B, E,H), rabbit anti-β₂(D), and rabbit anti-β₃(G) in conjunction with sheep anti-GAD (C, F). All of the larger GAD-colocalizing GABA_AR subunit isoform clusters (A, B, D,E, arrows) and many smaller clusters of β₁-IL and β_2/3 (A,B, filled arrowheads) or β₃and β_2/3 (G, H,arrows) that do not colocalize GAD show colocalization of the two β-subunit isoforms. β₂ expression was observed only within GAD-positive interneurons (F,filled arrowheads). Some small GABA_AR clusters contained only one of the two β-subunit isoforms examined (A, B, G,H, empty arrowheads). Neurons shown inA–F were cultured for 19 d; neuron in G and H was cultured for 28 d.A–C, G, andH show the processes of pyramidal neurons;D–F show the processes of an interneuron. Scale bar (shown in A): 5 μm.

Fig. 4.

Large clusters of GABA_AR accumulate postsynaptically to active GABAergic presynaptic boutons and show exo-endocytosis of synaptic vesicles. Hippocampal cultures were triple labeled with guinea pig anti-γ₂ subunit (A), rabbit anti-VGAT (B), and sheep anti-GAD (C) antibodies. VGAT and GAD concentrate presynaptically in the interneuronal varicosities (A, B, arrows) contacting a pyramidal neuron and colocalize with large postsynaptic clusters of the γ₂ subunit-containing GABA_ARs (arrows) but do not colocalize with the small clusters of GABA_ARs (A–C,filled arrowheads). In live cell labeling conditions for the synaptic vesicle exo-endocytotic assay (D–I), active synapses were labeled with the Syt-N anti-synaptotagmin antibody at 37°C (E, arrows). However, when live cells were incubated at 4°C (to prevent exocytosis and endocytosis), no labeling of synaptotagmin with the Syt-N antibody was observed (H, arrows). After live cell labeling with the Syt-N antibodies, cells were fixed and incubated with guinea pig anti-γ₂ subunit and sheep anti-GAD to reveal the localization of large GABA_AR clusters (D,G, arrows) and GAD-containing presynaptic terminals (F, I, arrows). Note that the Syt-N labeling of active synapses occurs in GAD-containing terminals and colocalizes with large GABA_AR clusters (D, E, F,arrows). Neurons in A–Cwere cultured for 21 d, and neurons inD–I were cultured for 28 d. Scale bar (shown in A): 5 μm.

Fig. 5.

GABAergic innervation induces a reduction in the density of small clusters in dendritic areas adjacent to GABAergic synapses. A, Dendrite segments (25 μm long) from two dendrites from the same pyramidal neuron, one receiving GABAergic innervation (top panel) and another dendrite not receiving GABAergic innervation (bottom panel). GABA_AR clusters were visualized with a rabbit anti-γ₂ antibody. Note the presence of a large GABA_AR cluster (arrow in top panel) at the GABAergic synapse (identified by the colocalization of a GAD-containing bouton), and the lower density of small GABA_AR clusters in the adjacent area (noticeable up to 15 μm distance). B, The graph shows that the reduction of the density of small clusters is significant up to 15 μm from the synapse. Beyond 15 μm, the density of the clusters is similar to that of dendrites not receiving GABAergic innervation (control). Quantification of the average cluster density around the synapse was done in five 5 μm zones distal to the site of a GABAergic synapse and compared with non-GABAergic innervated dendritic areas of the same neuron in 22-d-old cultures. (***p < 0.005, **p < 0.01, *p < 0.05; n = 6 matched dendrite pairs).

Fig. 6.

A population of small GABA_AR clusters are associated with glutamatergic synaptic markers. GABAergic innervation reduces the local density of small GABA_AR clusters that associate with glutamatergic synapses within pyramidal neurons. Hippocampal neurons were labeled with rabbit anti-γ₂ (A), mouse monoclonal anti-PSD-95 (B), mouse monoclonal anti-β_2/3 (D), rabbit anti-GluR1 (E), guinea pig anti-α₁(G), and mouse monoclonal SV2 (H) in conjunction with GAD (C, F, I). The large GABA_AR clusters colocalized with GAD and SV2 in GABAergic synapses (A–I,arrows). Some small GABA_A subunit clusters that were not associated with GAD were associated with PSD-95 (A, B, filled arrowheads), GluR1 (D, E, filled arrowheads), and SV2 (G, H,filled arrowheads). Some small GABA_AR clusters were not associated with glutamatergic markers (A, B, G,H, empty arrowheads). A comparison between processes of the same neuron (the two lower horizontal dendrites in D–F) shows that the association between small GABA_AR clusters and GluR1 clusters in dendritic spines observed in the top dendrite that does not receive GABAergic innervation (D,E, filled arrowheads) is not seen in the GluR1-containing spines of the bottom dendrite that receives GABAergic innervation (D, E, empty arrowheads). Scale bar (shown in A): 5 μm.

Fig. 7.

Clusters of glutamatergic postsynaptic density protein PSD-95 can be apposed to presynaptic and postsynaptic GABAergic synaptic markers, but they remain segregated from the GABAergic markers. Hippocampal cultures were labeled with rabbit anti-γ₂ (A), mouse monoclonal anti-PSD-95 (B), and sheep anti-GAD (C). Large clusters of GABA_AR γ₂ subunit colocalize with GAD-positive presynaptic boutons (arrows). A single GABAergic synapse (A–C, arrows) has been magnified in the insets to show detail. Note colocalization of GAD terminal and the GABA_AR clusters and how the presynaptic GAD and postsynaptic GABA_AR clusters “avoid” the PSD-95 clusters (empty arrowheads) by taking a digitated shape, preserving the segregation of the two postsynaptic receptor clusters. Neurons were cultured for 21 d. Scale bar (shown in A): 5 μm; insetscale bar, 1 μm.