Voltage imaging identifies spinal circuits that modulate locomotor adaptation in zebrafish

Abstract

Motor systems must continuously adapt their output to maintain a desired trajectory. While the spinal circuits underlying rhythmic locomotion are well described, little is known about how the network modulates its output strength. A major challenge has been the difficulty of recording from spinal neurons during behavior. Here, we use voltage imaging to map the membrane potential of large populations of glutamatergic neurons throughout the spinal cord of the larval zebrafish during fictive swimming in a virtual environment. We characterized a previously undescribed subpopulation of tonic-spiking ventral V3 neurons whose spike rate correlated with swimming strength and bout length. Optogenetic activation of V3 neurons led to stronger swimming and longer bouts but did not affect tail beat frequency. Genetic ablation of V3 neurons led to reduced locomotor adaptation. The power of voltage imaging allowed us to identify V3 neurons as a critical driver of locomotor adaptation in zebrafish.

Keywords: V3 neurons; locomotor adaptation; spinal motor circuits; voltage imaging; zebrafish.

Figure 1.. Voltage imaging in spinal neurons during fictive swimming.

(A) Schematic of the light-sheet microscope, ventral nerve root recording, and closed loop visual feedback. (B) zArchon1-GFP expression in the Tg(vGlut2a:Gal4; UAS:zArchon1-GFP) transgenic line. Top: Two-photon image of GFP expression marker; Bottom: light-sheet image of zArchon1 fluorescence in the same focal plane. Scale bar 10 μm. (C) Fluorescence traces showing n = 13 simultaneously recorded neurons, divided into oscillating (red) and non-oscillating (purple) activity patterns. Bottom: processed VNR signal and derived swim strength. (D) Left: VNR signal and fluorescence of two simultaneously recorded oscillating cells. Right: VNR cycle-triggered averages. Scale bar 10 μm. (E) Relationship between cycle-triggered amplitude and dorso-ventral position for n = 344 neurons, 9 fish. The pie charts show the distribution of activity types in the dorsal and ventral subpopulations. (F) Phase map of simultaneously recorded oscillating neurons. Cell bodies are colored according to their phase relative to the VNR signal. Scale bar 50 μm. (G) Transverse view showing cell body positions of oscillating neurons, color coded by phase relative to VNR. Non-oscillating cells in gray, dotted lines indicate position of the lateral neuropil region and the central canal. (H) Relationship of average phase and cell body position along the rostro-caudal axis. The two populations indicate cell bodies on the left and the right side of the spinal cord. The slope indicates the phase delay along the tail. Slope = −0.51 ± 0.06 π radians/mm, r² = 0.96. See also Fig. S1–3.

Figure 2.. Voltage imaging characterization of V3 neurons.

(A) Top: Example V3 recording and spike rates of all V3 neurons aligned to start and end of swim bouts. Spike rate for each neuron was averaged over 5 to 25 bouts. Neurons sorted by bout activity index I (Methods). Middle: grand average spike rates for all neurons and subthreshold fluorescence with spike computationally removed. Bottom: Corresponding mean VNR signals aligned to start and end of swim bouts. Shaded area denotes s.e.m. (n = 189 bouts). (B) Left: spike raster plot of a single V3 neuron for 80 swim cycles. Spikes were uniformly distributed throughout VNR phase. Right: VNR-triggered grand average spike rate of V3 neurons showing no phase-dependent modulation in spike rate. (C) Rapid inhibition of two V3 neurons on the left (red) and right (blue) sides of the spinal cord during a swim bout. (D) Top: Mean fluorescence transient during mid-bout inhibitory events, aligned relative to preceding VNR peak and color coded according to the cell body position on the left-right axis (n = 16 cells, 5–18 events per cell). Middle: Corresponding mean VNR signal. Bottom: histogram of threshold crossings showing two-peaked distribution separated by half a VNR cycle. See also Fig. S4–5.

Figure 3.. Action potential propagation and V3 morphology.

(A) (i) Spatial footprint of a V3 neuron with one neurite marked in red. (ii) Spike-triggered average action potential waveform sampled at discrete 1 ms camera frames (staircase) and interpolated smooth waveforms, from the proximal (brown) and distal (green) ends of the neurite indicated in (i). (iii) Action potential propagation along the neurite marked in (i). Color indicates the normalized fluorescence. Black line indicates the sub-Nyquist interpolated peak timing. (B) Action potential propagation along two neurons. Color denotes time delay, scale bar 20 μm. Action potential initiation sites (black stars) are consistent with action potential initiation at the axon initial segment. (C) Linear fits to the action potential propagation of seven neurons (gray lines) and the average fit (black line). Same colored dots are from the same neuron, each dot represents the binned value over 18 μm or neurite length. The mean conduction velocity was 0.19 ± 0.07 m/s (mean ± s.e.m.). (D) Schematic of activity-based segmentation. Voltage traces measured at the soma were used to calculate an average spike triggered movie and spike waveform. Pixelwise correlation of the spike waveform and the movie revealed the neuron morphology (Methods). Scale bar 10 μm. (E) Top: Average zArchon1 fluorescence from a voltage imaging movie (25,000 frames). Middle: map of spike-triggered average fluorescence amplitude for each neuron in the recording. Bottom: Composite image with a different color for each functionally identified cell. Scale bar 20 μm. (F) Tracing of a photoconverted V3 neuron in the Tg(vGlut2a:Gal4; UAS:Kaede) line. Red line shows extent of axon. Inset shows morphology near the soma. Scale bar 10 μm. (G) Top: morphology of photoconverted V3 neurons. Cell bodies were aligned relative to each other along the rostral-caudal axis and relative to the ventral margin of the spinal cord along the dorso-ventral axis. Bottom: magnified region. Positions were shifted along rostral-caudal axis for better visualization. (i) V3 neurons with bifurcating axons. (ii) V3 neurons with descending axons. See also Fig. S5.

Figure 4.. V3 activity correlates with increased swim strength.

(A) Eight simultaneously recorded V3 neurons during swims of different strengths. All neurons increased firing during stronger swim bouts. (B) V3 activity correlated with swim strength and bout duration but not cycle frequency. Each dot represents the average of all bouts in each category from one field of view. Red line denotes average ± s.e.m. V3 activity was defined as >50% of V3 neurons in the field of view spiking at least once during the bout. N = 9 field of views from 7 fish. Significance threshold α = 0.017 after Bonferroni correction, paired t-test.

Figure 5.. Optogenetic activation of V3 neurons increases swim strength in spinalized larvae.

(A) Lateral view of a spinal cord segment of Tg(vGlut2a:loxP-DsRed-loxP-ChR-GFP; nkx2.2:Cre) larva expressing ChR-GFP in V3 neurons. Scale bar 20 μm. (B) Schematic of the experimental setup. (C) Spontaneous and optogenetically evoked tail oscillations in the presence of 100 μM NMDA. Blue indicates ChR stimulation. (D) Average bout probability at increasing concentrations of NMDA during ChR stimulation (blue). N = 10 fish, averaged over 20 repetitions. (E) Optogenetic stimulation increased bout duration and swim strength, but not beat frequency. Experiments performed at 100 μM NMDA. For statistical model, see Supplementary Table 1.

Figure 6.. V3 neurons are necessary for swim speed adaptation.

(A) lateral view of a spinal cord of Tg(vGlut2aGFP; sim1a:loxP-DsRed-loxP-DTA) without (left panel) or with (right panel) Tg(hox4a/9a:Cre). Cre expression led to the loss of most V3 neurons as seen from the absence of DsRed-expressing cells in the ventral spinal cord. Scale bar 20 μm. (B) Schematic of the behavioral testing setup. Free-swimming larvae were presented with a moving grating from below and recorded from above. (C) V3 ablated larvae showed reduced modulation of bout speed in response to changes in OMR grating speed, but no difference in tail beat frequency or bout duration. Each dot represents the median value for one fish, the line the average value across fish, error bars are s.e.m. N = 28 ablated and 40 control fish. See also Fig. S6. For statistical model, see Supplementary Table 2.