Properties and molecular identity of NMDA receptors at synaptic and non-synaptic inputs in cerebellar molecular layer interneurons

Abstract

N-methyl-D-aspartate receptors (NMDARs) in cerebellar molecular layer interneurons (MLIs) are expressed and activated in unusual ways: at parallel fibre (PF) synapses they are only recruited by repetitive stimuli, suggesting an extrasynaptic location, whereas their activation by climbing fibre is purely mediated by spillover. NMDARs are thought to play an important role in plasticity at different levels of the cerebellar circuitry. Evaluation of the location, functional properties and physiological roles of NMDARs will be facilitated by knowledge of the NMDAR isoforms recruited. Here we show that MLI-NMDARs activated by both PF and climbing fibre inputs have similar kinetics and contain GluN2B but not GluN2A subunits. On the other hand, no evidence was found of functional NMDARs in the axons of MLIs. At the PF-Purkinje cell (PF-PC) synapse, the activation of GluN2A-containing NMDARs has been shown to be necessary for the induction of long-term depression (LTD). Our results therefore provide a clear distinction between the NMDARs located on MLIs and those involved in plasticity at PF-PC synapses.

Keywords: NMDA receptors; cerebellum; climbing fibres; interneuron; parallel fibres; plasticity.

Figure 1

Parallel Fibre and Climbing Fibre inputs activate GluN2B-containing NMDA receptors in Molecular Layer interneurons. (A,B) PF- or (C,D) CF-evoked NMDA currents in MLIs in control conditions (black) or in the presence of 300 nM free Zn²⁺ (red) or 300 nM Ro25-6981 (green). (A,C) Representative recordings (average of 30–50 sweeps). (B,D) Left panels: NMDA EPSC amplitude before (ctrl) and after (antag.) bath application of Zn²⁺ (red) or Ro25-6981 (green) for each cell. Right panels: Mean normalized NMDA-EPSC amplitude after Zn²⁺ or Ro25-6981.*p < 0.05. (E) Decay time constants of PF- or CF-evoked NMDA EPSCs in control conditions. (F,G) Decay time constants of CF- but not of PF-evoked NMDA EPSCs are slowed by Ro25-6981. (F) Scaled traces of representative recordings (average of 30–50 sweeps) in control conditions (black) and after 300 nM Ro25-6981 (green). For PF input, same cell as in (A). For CF input, different cell from (C). (G) PF (left) and CF (right) -evoked NMDA EPSC decay time constant before and after bath application of Ro25-6981 for each cell. *p < 0.05.

Figure 2

Molecular Layer interneurons express mostly GluN2B-containing NMDA receptors in their somatodendritic compartment. (A) Immunostaining showing GluN2B (red) colocalization with a neurobiotin-filled MLI somatodendritic compartment (green). Sum of 30 optical slices in a stack. Scale bar = 10 μm. (B,E) MLI somata immunostained for GluN2A (B) or GluN2B (E). Scale bars = 2 μm. (C,D) Examples of unstained PF synapses onto GluN2A-stained MLI soma extended from (B). Scale bars = 200 nm. (F) Examples of unstained PF synapses onto GluN2B-stained MLI soma extended from (E). Scale bar = 500 nm. (G) Quantification of the number of MLI somata faintly (gray) or strongly (black) stained for GluN2A or GluN2B in 17500 μm² areas.

Figure 3

Molecular Layer interneurons do not express functional axonal NMDA receptors. (A) Immunostaining showing the absence of GluN2B (red) colocalization with a neurobiotin-filled MLI axon (green). Sum of 56 optical slices in a stack. Scale bar = 10 μm. (B) Immunostaining for GluN2B showing GluN2B-stained putative inhibitory synapses (axon) on a GluN2B-stained MLI soma. Synapses are extended in the right panel. Scale bars: left panel 2 μm, right panel 0.5 μm. (C–F) Miniature IPSCs recorded in PCs (C,D) or in MLIs (E,F) in control conditions or in the presence of 15 μM (C,D) or 30 μM (E,F) NMDA. (C,E) Representative recordings. (D,F) Amplitude and frequency of mIPSCs for each cell before (ctrl), during (NMDA) or after (wash) bath-application of NMDA.