Dynamics of Glutamatergic Drive Underlie Diverse Responses of Olfactory Bulb Outputs In Vivo

Glomerular glutamate signals across MT cell apical tufts show diverse inhalation-driven dynamics. A, Left, Mean fluorescence image of SF-iGluSnFR.A184V expression imaged in vivo. Right, Inhalation-triggered ΔF response maps for three odorants. B, ITA response traces for the glomeruli indicated in A. Traces are averages across 17 inhalations. Left, ITAs from one glomerulus (ROI 5) responsive to three odorants, illustrating odorant-specific ITA dynamics. Right, ITAs from different glomeruli responsive to each of the three odorants, illustrating glomerulus-specific dynamics. C, Cumulative fraction of %ΔF/F peak values across SF-iGluSnFR variants: black, A184S (high-affinity variant); gray, A184V (medium affinity variant); red, S72A (low-affinity variant). D, Cumulative fraction of decay rates across SF-iGluSnFR variants: Same colors as in C. E, Plot of ITA onset latency, peak latency, and duration (FWHM, full-width half-max) for all responsive glomerulus-odor pairs using SF-iGluSnFR.A184S. Notch box plots (left) of latency values (right). Notch: median, square: mean, box edges: 25th and 75th percentiles, n = 431 glomerulus-odor pairs, 5 mice. F, Cumulative fraction of median-subtracted, ITA onset latencies for SF-iGluSnFR.A184S signals, compared with those measured from GCaMP6f expressed in OSN axon terminals (OMP-Cre x Rosa-GCaMP6f mice), showing similar range of latencies for both measurements. G, Preferential iGluSnFR expression in sTCs versus mitral cells. Left, Confocal image of tissue section after Flex.AAV.SF-iGluSnFR injection into the superficial EPL of a CCK-IRES-Cre mouse. Note strongest expression in superficial EPL. Right, Similar image after Flex.AAV.SF-iGluSnFR injection into anterior piriform cortex in a Tbet-Cre mouse. H, Identical distributions of onset latencies, times-to-peak, and durations (FWHM) of ITA waveforms imaged from CCK⁺ sTC and pcMT cell populations. Range of onset latencies (10th–90th percentiles): 195–451 ms (CCK⁺) versus 192–447 ms (pcMTs); times to ITA peak: 373–938 ms (CCK⁺) versus 367–828 ms (pcMTs); FWHM: 313–1058 ms (CCK⁺) versus 326–1007 ms (pcMTs); n = 174 and 160 glomerulus-odor pairs, respectively, from 3 mice each. All statistical comparisons, p > 0.05, Kolmogorov–Smirnov test.

Glomerular glutamate signals show diverse temporal patterns across repeated inhalations of odorant. A, Mean fluorescence (left) and ΔF response maps for five odorants imaged with SF-iGluSnFR.A184S (right). Odor numbers on the left side of the odorant map indicate odorant identification code for listed odor name (for reference to B). Colored text indicates odorants whose responses are shown in panels B, C, F. B, Time series of SF-iGluSnFR.A184S signal from four glomeruli, showing continuous signal during three repetitions of 12 odorants, presented in random order. Shaded rectangle indicates time of odorant presentations (2-s duration). Signal is low-pass filtered at 0.5 Hz for display. C, Average time course of odorant-evoked responses in different glomeruli responsive to four of the 12 odorants from B. Insets show change in response pattern across the field of view from A over the duration of odorant presentation, expressed as Pearson’s r relative to the initial response in successive time bins of 387 ms (see text). Dotted line indicates r = 0. D, Waterfall plot showing time course of odorant-evoked SF-iGluSnFR.A184S signal in all significantly-responding glomerulus-odor pairs (one pair per row), normalized to the peak ΔF/F for each pair. Suppressive responses are low-pass filtered at 0.5 Hz, and excitatory responses are low pass filtered at 2 Hz. E, Proportion of excitatory (Exc.), suppressive (Sup) and biphasic (Exc. and Sup.) responses across the responsive population (right), as well as the entire population of glomeruli (left). F, Change in glutamate response patterns during odorant presentation, summarized over all presentations. Black plot shows mean correlation over time, averaged across all odorant presentations. Shaded area indicates SEM. Dot plot at right shows correlation coefficients at the final time bin for all presentations.

Odorant concentration systematically impacts glutamate signaling onto MT cells. A, Waterfall plots (left) and traces (right) showing time course from select glomeruli showing original iGluSnFR signals in response to three concentrations of odorant for three different odorants (1×, 3×, 10× indicate relative concentration values; see text). All data are from the same experiment and field of view. Only glomeruli showing significant responses at all three concentrations are shown. Inhalation frequency, 2 Hz; odorant duration, 4 s. Signals are unfiltered. Glomerular identity is color coded with lines pointing to the traces. White dotted line indicates odor onset. T₁, T₂, T_max (10×, methyl valerate panel) indicates time points used for T₂–T₁/T_max measurements in B. B, Change in iGluSnFR response amplitude from beginning to end of odorant presentation (T₂–T₁/T_max; see text), summarized for all analyzed glomerulus-odorant pairs (n = 189 glomerulus-odorant pairs from 4 mice). Notch: median, square: mean, box edges: 25th and 75th percentiles. Note high variability and slight but significant trend toward adaptation at all concentrations [one sample t test comparing T₂–T₁/T_max; to zero, 1×: −0.19 ± 0.35 (mean ± SD), p = 2 × 10⁻¹², 3×: −0.14 ± 0.40, p = 5 × 10⁻⁶, 10×: −0.26 ± 0.44, p = 8 × 10⁻¹⁴, n = 189 glomerulus-odorant pairs]. Lower bars indicate paired t tests between concentrations; ***p < 0.001; n.s., not significant. C, Coherence of iGluSnFR signal to the 2-Hz inhalation frequency (see text), measured for the same glomerulus-odorant pairs in B, showing high variability across glomeruli and a decrease in coherence at higher concentrations. D, Coherence values as a function of concentration, showing a significant effect of odorant on concentration-coherence functions. Dot: median, error bars: quartiles.

Diversity of glutamate dynamics imaged from MT cells in the awake mouse. A, Top, Mean SF-iGluSnFR.A184V fluorescence. Bottom, ΔF response map (ethyl butyrate) from the dorsal OB of an awake, head-fixed mouse. Right, Glutamate signals from three glomeruli (shown in A) during a single presentation of odorant (ethyl butyrate). Top trace shows respiration/sniffing as measured with an external flow sensor, with inhalation oriented downward. Lower left, snippet of signals from each glomerulus, illustrating temporal lag between different glomeruli and consistency of dynamics with each sniff. Signals are unfiltered. Lower right shows ITA waveforms for each glomerulus, generated across 137 sniffs over six presentations of ethyl butyrate. Vertical line indicates time of inhalation onset. B, Left, Range of ITA onset (black) and peak (red) latencies, relative to median values, compiled across 108 glomerulus-odorant pairs (14 odorants, seven fields of view). Notch: median, box edges: 25th and 75th percentiles, square: mean values. Right, Spread of latency values seen across multiple responsive glomeruli imaged in the same field of view for the same odorant (defined as the difference between maximum (lat_max) and minimum (lat_min) latencies), for 17 unique odorant presentations. C, SF-iGluSnFR.A184V signals imaged from four glomeruli from the same FOV showing responses to each of two odorants, showing distinct temporal response patterns across glomeruli and odorants. Traces are unfiltered and from a single presentation. Top trace shows respiration/sniffing for each trial; vertical dotted lines indicate inhalation peak for reference. D, SF-iGluSnFR.A184V signals imaged from three glomeruli, taken from a separate field of view in the same mouse, showing distinct response patterns to a third odorant (3-octen-2-one). Note weak suppressive response in one glomerulus (top trace). E, Increasing odorant concentration elicits changes in glutamate signal dynamics in awake mice. Traces show averaged responses (eight presentations) from four glomeruli in the same field of view to a low (1:1000 dilution) versus high (1:10) concentration of ethyl tiglate. Note that the response in the most sensitive glomerulus shows increased adaptation at the higher concentration, while other glomeruli show facilitating responses. F, Changes in glutamate response patterns over the course of odorant presentation, calculated as described for Figure 3, in four awake mice, 33 odorant-FOVs (two mice, SF-iGluSnFR.A184V; two mice, SF-iGluSnFR.A184S). Plots of individual odorant responses are shown in gray, thick black line shows mean, shaded region is quartiles (25th and 75th percentiles). Note high variability in degree of decorrelation over time.

Dual-color imaging reveals high correspondence between presynaptic and postsynaptic signals in MT cells of the same glomerulus. A, Post hoc confocal image showing coexpression of jRGECO1a (magenta) and SF-iGluSnFR.A184S (A184S, green) in MT cells of a Tbet-Cre mouse. B, Dual-color two-photon imaging of SF.iGluSnFR.A184S and jRGECO1a signals from MT cells. Left, Odorant-evoked ΔF response maps for SF-iGluSnFR and jRGECO1a signals (magenta imaged simultaneously). Note jRGECO1a signal in dendrites of several MT cells exiting the central glomerulus, with SF-iGluSnFR signal confined to the glomerular neuropil. Right, Traces showing time course of the fluorescence signal in each channel (1-Hz inhalation). Top traces are from glomerular neuropil; lower traces are from dendrites outside of the glomerulus. C, High correspondence in calcium signals imaged from different MT cell subcompartments. Left, Images show mean fluorescence and ΔF odorant response map (ethyl tiglate) for GCaMP6f signals imaged in a CCK-IRES-Cre: Rosa-GCaMP6f cross. Right, Overlaid traces showing time course of GCaMP6f signal from the neuropile of two glomeruli (blue, orange traces) and, for each, the soma of a tufted cell innervating each glomerulus. For the glomerulus on the right, the signal from the primary dendrite of a second tufted cell innervating the same glomerulus is also shown. Near-synchronous, inhalation-driven transients are seen in all compartments, with a slightly slower rise and slower decay in the somata. Latency differences between the two glomeruli are also present in their respective cells’ somata. D, Pseudocolor odorant-evoked ITA response maps across the green (A184S) and red (jRGECO1a) channels. Arrows indicate ROIs with traces plotted in D. E, ITA traces taken from different glomeruli activated by the odorant in C, with different onset latencies, times to peak and durations in different glomeruli. Left traces, SF-iGluSnFR.A184S. Right traces, jRGECO1a; traces from the same glomerulus are shown with the same color in each set. Each ITA trace is scaled to the same maximum. Vertical lines indicate peak time for the signal in each glomerulus. The relative order of peak times is the same for both signals. F, Traces showing SF-iGluSnFR.A184V and jRGECO1a signals imaged simultaneously from a glomerulus, with high correspondence between spontaneously-occurring transients in the absence of odorant stimulation. Vertical lines mark transients seen in both signals; downward arrows mark transients seen in the green (A184S) but not red (jRGECO1a) channels; upward arrow marks transient seen in the red but not the green channel. G, Dual-color imaging of SF.iGluSnFR.A184S and jRGECO1a signals from MT cells in the awake mouse. Left, Composite dual-color ΔF response map showing SF-iGluSnFR (green) and jRGECO1a (magenta) signals evoked by cyclohexylamine. Right, Traces showing fluorescence signal taken from the glomerular neuropile (top) and soma (bottom) of an innervating TC. Traces are average of eight presentations. H, Trial-averaged ΔF response maps and traces for SF-iGluSnFR.A184S and jRGECO1a signals imaged from a single presentation of odorant in an awake mouse. Top trace shows respiration measured via external flow sensor. The SF-iGluSnFR signal clearly follows each inhalation in only one of the three glomeruli shown, while the jRGECO1a signal does not follow inhalations in any glomeruli.