Central and peripheral innervation patterns of defined axial motor units in larval zebrafish

Abstract

Spinal motor neurons and the peripheral muscle fibers they innervate form discrete motor units that execute movements of varying force and speed. Subsets of spinal motor neurons also exhibit axon collaterals that influence motor output centrally. Here, we have used in vivo imaging to anatomically characterize the central and peripheral innervation patterns of axial motor units in larval zebrafish. Using early born "primary" motor neurons and their division of epaxial and hypaxial muscle into four distinct quadrants as a reference, we define three distinct types of later born "secondary" motor units. The largest is "m-type" units, which innervate deeper fast-twitch muscle fibers via medial nerves. Next in size are "ms-type" secondaries, which innervate superficial fast-twitch and slow fibers via medial and septal nerves, followed by "s-type" units, which exclusively innervate superficial slow muscle fibers via septal nerves. All types of secondaries innervate up to four axial quadrants. Central axon collaterals are found in subsets of primaries based on soma position and predominantly in secondary fast-twitch units (m, ms) with increasing likelihood based on number of quadrants innervated. Collaterals are labeled by synaptophysin-tagged fluorescent proteins, but not PSD95, consistent with their output function. Also, PSD95 dendrite labeling reveals that larger motor units receive more excitatory synaptic input. Collaterals are largely restricted to the neuropil, however, perisomatic connections are observed between motor units. These observations suggest that recurrent interactions are dominated by motor neurons recruited during stronger movements and set the stage for functional investigations of recurrent motor circuitry in larval zebrafish.

Keywords: RRID: Addgene_74314; RRID: Addgene_74316; axial muscle; motor neurons; motor units; recurrent circuits; spinal cord; zebrafish.

Figure 1.

Visualizing the organization of axial muscle innervation in larval zebrafish. a, Side view of 5 day old Tg[mnx1:GFP] zebrafish illustrating spinal motor neurons and peripheral motor nerves in green/white. Rostral is to the left and dorsal is up. Dashed white line indicates region expanded in b. Note, some fluorescence is also observed in the brain. b, Contrast-inverted image of axial motor neurons and peripheral motor nerves between the 5^th and 20^th body segment (numbered) viewed from the side. Motor neurons form a longitudinal column near the top of the image. c, Schematic illustrating the medio-lateral distribution of fast-twitch muscle fibers (grey) and slow muscle fibers (red) in transverse view. Dorsal epaxial and ventral hypaxial musculature is separated by the horizontal myoseptum (black dashed horizonal line). Peripheral nerves originating from the spinal cord (sc) are distinguished by their medial (black) and septal (grey) projection patterns. Dashed circle indicates the notochord (nc). d, Unilateral transverse views taken from different body segments (numbered).

Figure 2.

Classifying secondary axial motor neurons based on medial and septal innervation patterns a, Contrast-inverted images of identified primary motor neurons in transverse view and their innervation of epaxial and hypaxial musculature. MiP, middle primary; dRoP, rostral primary, dorsal projection; vRoP, rostral primary, ventral projection; CaP, caudal primary; VaP, variable primary. The distinct types delineate four epaxial and hypaxial muscle groups (color coded and numbered I-IV on right). Dashed line indicating horizontal myoseptum (hm) is provided for reference here and in subsequent panels. Scale bar is the same for all panels, unless noted otherwise. b, Contrast-inverted images of secondary motor neurons in transverse view named based on their contribution to medial peripheral nerves and the epaxial/hypaxial muscle groups they innervate. Asterisks denote reconstructions instead of raw data in cases where one or more motor neurons were also in the field of view. Open red arrowheads illustrate medial nerve projections. c, Secondary motor neurons that contribute to both medial (open red arrowheads) and septal (filled red arrowheads) peripheral nerves. d, Secondary motor neurons that contribute to only septal (red arrowheads) nerves. e, Secondary motor neurons that terminate at the dorsal or ventral extremes of the axial musculature or near the horizontal myoseptum. f, Side on views of the nerve endings of motor neurons shown in panel e. Collapsed fluorescent images are superimposed on single optical sections of muscle visualized using differential interference contrast. g, Examples of secondaries from panels b-d superimposed on color coded primary motor neurons from panel a illustrate differences in the medio-lateral distribution of peripheral motor nerve endings between the different types. h, Quantification of the rostro-caudal distribution along the body of the different types of motor neuron.

Figure 3.

Comparing the morphological features of the different types of axial motor neurons a, p-type motor neurons (pI, pII, pIII, pIV) have been color coded to illustrate their tiling of the dorsal epaxial and ventral hypaxial musculature. Dashed line indicates the horizonal myoseptum (hm). b, Example of an m-type motor neuron that innervates all four axial muscle groups (mI-IV) superimposed on p-type motor neurons for reference. c, An ms-type motor neuron innervating all four muscle groups. Red arrowheads mark septal branches. d, An s-type motor neuron innervating three muscle groups. Note, the intermyotomal cleft runs caudally as you move to the surface of the muscle, so the collapsed image of the s-type peripheral axon is right-shifted relative to the primary reference image. e, Quantification of the soma diameter of the different types of motor neuron. Means sharing letters are not significantly different following one-way ANOVA and post-hoc Bonferroni tests. f, Quantification of length of the peripheral arbor of different types of motor neuron. Medians sharing letters are not significantly different following Kruskal-Wallis ANOVA and post-hoc Mann-Whitney U tests. g, Quantification of the dendritic arbor of different types of motor neuron. h, Quantification of soma diameter versus peripheral arbor length of secondary motor neurons. Correlation is reported for Spearman Rank test. i, Quantification of the soma diameter versus dendritic arbor length of secondary motor neurons. Correlation is reported for Spearman Rank test.

Figure 4.

Characterizing motor neurons with central axon collaterals a, Top: contrast-inverted image of spinal motor neurons in Tg[mnx1:GFP] fish viewed from the side. The ventral roots (vr) are indicated by red filled arrowheads. Bottom: differential interference contrast image of the same location in spinal cord shown above illustrates major landmarks for normalization (1, dorsal aspect of spinal cord, 0, ventral aspect of spinal cord, m, Mauthner axon). b, Contrast-inverted images of p-type motor neurons. Central axon collaterals are highlighted in red or with red open arrowheads. c, As in panel b. but for m-, ms-, and s-type motor neurons. d, Quantification of the dorso-ventral position of motor neuron somata normalized to the height of the spinal cord and their distance from the ventral root exit point. Blue indicates total dataset, while black indicates position of motor neurons with central axon collaterals, vr, ventral root; m, Mauthner axon. e, Quantification of normalized dorso-ventral soma positions of the different types of motor neuron. Means sharing letters are not significantly different following one-way ANOVA and post-hoc Bonferroni tests. f, Quantification of the central collateral length of different types of motor neuron. Means sharing letters are not significantly different following one-way ANOVA and post-hoc Bonferroni tests. g, Quantification of likelihood of central collateral among the different types of motor neuron. h, Quantification of the likelihood of central collaterals among p-type motor neurons, ordered from left to right by mean rostro-caudal soma position, pII = 42 ± 12 μm, pIII = 41 ± 21 μm, pI = 31 ± 10 μm, pIV = 8 ± 4 μm. Note that pIIIs contain VaPs = 13.03 ± 8 μm and vRoPs = 51.39 ± 12 μm. Since VaPs never exhibit central collaterals, this brings down the average. i, Quantification of central collateral presence in secondary motor neurons based on the number of axial muscle quadrants innervated.

Figure 5.

Characterizing the input/output features of the different types of motor neurons a, Schematic illustrating the strategy for identifying pre- and post-synaptic structures in motor neurons using fluorescently tagged synaptophysin (red, pre-synaptic), and PSD95 (orange, post-synaptic). b, Contrast-inverted images of motor neurons (grey) in transverse view from each type illustrating synaptophysin (Syp) expression (red). c, Contrast-inverted images of motor neurons from the side view illustrating synaptophysin (Syp) labeling on axon collaterals (left) and PSD95 labeling on dendrites and soma (right). Open black arrowheads demarcate central axon collaterals. Note, Syp and PSD95 labeling were not performed in the same motor neurons. Dorsal is up and rostral is to the left. d, Quantification of motor neuron peripheral synaptophysin puncta number versus peripheral axon length. Secondary motor neurons labeled in black and primary motor neurons in grey. Correlation is reported for Pearson’s test. e, Quantification of motor neuron central synaptophysin puncta number versus central collateral length. f, Quantification of PSD95 puncta number versus dendrite length. g, Quantification of PSD95 puncta number versus peripheral axon length. Correlation is reported for Pearson’s test.

Figure 6.

Assessing the topographic organization of the input/output of the different types of motor neurons a, Left: A confocal z-stack in transverse view illustrating the medio-lateral (ML) and dorso-ventral (DV) distribution of motor neuron somata and neuropil within the spinal cord in a Tg[mnx1-3×125bp:Gal4-VP16;UAS:pTagRFP] fish. Right: Schematic demonstrating the normalization according to anatomical landmarks, which include the spinal cord (dashed lines) and the Mauthner axon (m). b, Depiction of axon trajectory and proximity to the Mauthner axon (circles) in transverse view of the different types of motor neuron prior to exiting the spinal cord. D, dorsal; L, lateral. c, Left: Quantification of the maximum normalized dendrite dorso-ventral (DV) height versus the normalized soma DV position. Correlation is reported for Spearman Rank test. Right: Quantification of the maximum normalized central collateral DV height versus the normalized soma DV position. Correlation is reported for Pearson’s test. d, Contour density plot of the normalized medio-lateral and dorso-ventral distribution of the dendrites (black) and central collaterals (grey) of motor neurons with central collaterals (n = 148). The innermost contour represents the highest filament density. e, Contour density plot of the normalized medio-lateral and dorso-ventral distribution of PSD95 puncta of the different types of motor neuron (p-type, n = 19; m-type, n = 10; ms-type, n = 11; s-type, n = 6). The innermost contour represents the highest puncta density. f, Contour density plot of the normalized medio-lateral and dorso-ventral distribution of synaptophysin (Syp) puncta of the different types of motor neuron (p-type, n =15; m-type, n = 12; ms-type, n = 5). g, Contour density plot of the normalized medio-lateral and dorso-ventral distribution of synaptophysin and PSD95 puncta from all the types of motor neuron (n = 46 motor neurons with PSD95 labeling and n = 32 motor neurons with synaptophysin labeling).

Figure 7.

Identifying putative motor neuron-motor neuron contacts a, Single optical sections viewed from the side of putative perisomatic contacts (open red arrowheads) between individual motor neurons of different types labeled by injections and all axial motor neurons in the Tg[mnx1-3×125bp:Gal4-VP16;UAS:pTagRFP] line. Dorsal is up and rostral is left. b, Quantification of dorso-ventral location of p- (n = 5), m- (n = 7), and ms-type (n = 5) motor neurons and their putative motor neuron targets. c, Collapsed fluorescent image of an mII-III and msII-III secondary motor neuron in muscle segments 16 and 17, respectively, superimposed on a differential interference contrast (DIC) image to illustrate segment boundaries. d, Top: Contrasted-inverted image of motor neurons from panel c in grey and synaptophysin (Syp) puncta labeling in red. Blue arrowhead marks central collateral from rostral m-type neuron. Bottom: Single optical sections of neuron labeled with asterisk above, depicting putative perisomatic connectivity from the mII-III to the msII-III neuron (open black arrowheads).

Figure 8.

Summary of the main observations. a, Schematic depicting the major divisions of the axial muscle based on primary motor neuron innervation. sc, spinal cord; nc, notochord. b, Schematic depicting the proposed muscle innervation territories m-type units, which innervate deep fast-twitch muscle. Here and in subsequent panels, territories are superimposed on primary quadrants. Solid lines represent observed units, while dashed lines indicate units that would be predicted to occur but were not found here. c, Schematic of ms-type units, which innervate more superficial fast-twitch and slow fibers. d, Schematic of s-type units, which innervate superficial slow fibers. Relatively immature s-types (s’) terminate in the extremities noted by red circles. e, Schematic depicting the proposed recurrent circuitry involving motor neurons defined by peripheral innervation patterns. Fast-twitch units (m-type and ms-type) are more likely to have collaterals and be interconnected. We also propose fast units would be more likely to innervate the equivalent of Renshaw cells in fish, namely V1 interneurons (see Discussion for details), which have more potent inhibitory inputs to slow units.