Fast-spiking interneuron detonation drives high-fidelity inhibition in the olfactory bulb

Inhibitory circuits in the mammalian olfactory bulb (OB) dynamically reformat olfactory information as it propagates from peripheral receptors to downstream cortex. To gain mechanistic insight into how specific OB interneuron types support this sensory processing, we examine unitary synaptic interactions between excitatory mitral and tufted cells (MTCs), the OB projection cells, and a conserved population of anaxonic external plexiform layer interneurons (EPL-INs) using pair and quartet whole-cell recordings in acute mouse brain slices. Physiological, morphological, neurochemical, and synaptic analyses divide EPL-INs into distinct subtypes and reveal that parvalbumin-expressing fast-spiking EPL-INs (FSIs) perisomatically innervate MTCs with release-competent dendrites and synaptically detonate to mediate fast, short-latency recurrent and lateral inhibition. Sparse MTC synchronization supralinearly increases this high-fidelity inhibition, while sensory afferent activation combined with single-cell silencing reveals that individual FSIs account for a substantial fraction of total network-driven MTC lateral inhibition. OB output is thus powerfully shaped by detonation-driven high-fidelity perisomatic inhibition.


Figure S2 .
Figure S2.EPL-IN subtypes exhibit stark differences in most intrinsic biophysical properties A-E: Diverse responses used to calculate intrinsic biophysical properties for the example FSI from Figure 1A, including: spontaneous activity at resting membrane potential (A), mean response to negative step current injection, with single-exponential fit (dashed black line) (B), mean spike waveform evoked by 1-ms suprathreshold current injection (C), firing rate-current relationship (D), and interspike interval (ISI) coefficient of variation evoked by positive step current injection (E).Spontaneous spike in A truncated to better visualize synaptic activity.Inset in A: magnification of spontaneous synaptic events.F-T: Same as A-E for the example FSI and RSIs from Figure 1C,E,G.Insets in A,F,K,P are identically scaled.

Figure S3 .
Figure S3.PV expression distinguishes FSIs from RSIs Intracellular NB and post hoc PV staining with 50-μm magnified inset of somata (left), inverted NB (middle), and step current-evoked spiking (right) of a panel of EPL-INs.Spiking responses are color-coded to reflect FSI vs. RSI physiology, as in Figure 1.

Figure S4 .
Figure S4.Neither FSIs nor RSIs are dopaminergic Intracellular NB and post hoc TH staining with 50-μm magnified inset of somata (left), inverted NB (middle), and step current-evoked spiking (right) of a panel of EPL-INs.Spiking responses are color-coded to reflect FSI vs.RSI physiology, as in Figure1.An example TH + SAC exhibiting tonic spontaneous firing is additionally included as positive control for TH staining.

Figure S5 .
Figure S5.VIP expression poorly distinguishes FSIs and RSIs Intracellular NB and post hoc VIP staining with 50-μm magnified inset of somata (left; single optical confocal planes), inverted NB (middle; maximum-intensity confocal projection), and step current-evoked spiking (right) of a panel of EPL-INs.Spiking responses are color-coded to reflect FSI vs. RSI physiology, as in Figure 1.

Figure S6 .
Figure S6.Unitary MTC-to-RSI excitation in a solitary example was distinctly weaker than FSI excitation A,B: Step current-evoked spiking response (A) and unitary synaptic interactions (B) for the solitary MTC-RSI pair exhibiting significant unitary MTC-to-RSI excitation (morphology not recovered).Postsynaptic RSI voltage shown on same scale as Figure 1O,P for comparison to postsynaptic FSI responses.Inset: magnification of mean postsynaptic RSI voltage.C: The MTC-to-RSI uEPSP amplitude was markedly weaker than FSI uEPSPs (n=69).

Figure S8 .
Figure S8.Unitary MTC-FSI synaptic pharmacology A: Recording from an example MTC-FSI pair (morphology not recovered) showing MTC-to-FSI uEPSP amplitudes before and after combined bath application of glutamatergic antagonists NBQX (10 μM) and AP5 (50 μM) and subsequent application of GABAAR antagonist Gabazine (10 μM).B: Postsynaptic FSI voltages from the pair in A. Traces in each subplot correspond to the bracketed trials in A. C: Unitary MTC-to-FSI excitation was blocked by combined application of NBQX and AP5 and partially recovered upon wash-out in 4 MTC-FSI pairs

Figure S9 .
Figure S9.MTC-FSI connectivity is restricted to infraglomerular layers A: Example MTC-FSI pair with MTC apical dendrite truncated prior to entering glomerular layer (open arrowhead).B,C: FSI fast-spiking response to step current injection (B) and unitary synaptic connectivity with MTC (C).D-F: Same as A-C for a second example MTC-FSI pair.

Figure S12 .
Figure S12.Syt2 clusters do not selectively target MTC axon initial segments Single confocal optical plane (left) and maximum-intensity projection (right; ~50 μm depth) of Syt2 and axon initial segment component TRIM-46 in the OB, revealing an absence of clear Chandelier-like innervation of MTCs.

Figure S13 .
Figure S13.Comparison of spontaneously alternating FSI detonation vs. uEPSP trials reveals highfidelity recurrent MTC inhibition A,B: Example MTC-FSI pair (A) in which unitary MTC release triggers FSI detonation on some trials (light green) and uEPSPs on other trials (dark green) (B).C: Subtraction of mean MTC currents across FSI detonation vs. uEPSP trials from B isolates IPSC waveforms time-locked to FSI detonation.D-F: Same as A-C for an example MTC-FSI pair recorded in current-clamp, revealing isolation of an IPSP waveform time-locked to FSI detonation.