Expression of Semaphorin7A in the zebrafish’s lateral line.

(A) A volumetric rendering of a PLL neuromast depicts the sensory axons (magenta) that branch from the lateral line nerve to arborize around the basolateral surface of the hair cell cluster (green). Additional cell types in the neuromast are not labeled. Larval age, 3 dpf. (B) A schematic drawing of the two variants of the Sema7A protein molecule depicts the full-length GPI-anchored form and the smaller, potentially secreted form. Both the molecules include a signal sequence (green) and a conserved sema domain (orange). Gel-based RT-PCR analysis indicates the presence of both the variants in developing larva. (C) A surface micrograph of a neuromast at 3 dpf depicts two pairs of mature hair cells (green dashed lines) and a pair of immature hair cells (grey dashed lines). Inset: among the three pairs of hair cell apices, the immature pair is indicated by arrowheads. Immunolabeling reveals that the Sema7A protein (orange) occurs consistently at the subapical region (arrowheads) and at the basolateral surface (arrows) of a hair cell. In this and in each of the subsequent neuromast images, anterior is to the left and dorsal to the top. (D) A plot quantifies developmental changes in the average Sema7A intensity in both rostrally and caudally polarized hair cells of neuromasts from 1.5 dpf to 4 dpf. The data stem from 18, 36, 39, and 40 hair cells in neuromasts of respectively 1.5 dpf, 2 dpf, 3 dpf, and 4 dpf larvae. (E) A plot quantifies the distribution of average Sema7A intensity along the hair cell’s apicobasal axis. The results stem from 52, 57, and 54 hair cells of neuromasts from 2 dpf, 3 dpf, and 4 dpf larvae. (F) An immunofluorescence image at the nuclear level of a 4 dpf neuromast shows the contact of the sensory arbors (magenta) with the basolateral surface of the hair cells (green). Immunolabeling for Sema7A (orange) reveals enrichment of the protein at the hair cell bases and sensory-axon interfaces (arrowheads). HC, hair cell; Scale bars, 5 µm; au, arbitrary unit; means ± SEMs; *** implies p < 0.001.

sema7a-/-mutants display aberrant sensory axon arborizations.

(A) In a micrograph of a 3 dpf sema7a-/- neuromast, Myosin VI (green) marks the hair cells (green dashed lines) in which the level of Sema7A (orange) is highly reduced, with sporadic localization in the subapical region (arrowheads) and none in the basolateral (arrow) region. A plot of normalized Sema7A intensity from 99 control and 100 sema7a-/- hair cells quantitates the effect. (B,C) Surface views of a control and a sema7a-/- neuromast at 4 dpf depict the interaction of the sensory arbors (magenta) with hair cell clusters (green). In the control, the arbors intimately contact the hair cells, with a few exceptions (arrowhead). In the sema7a-/-mutant, the arbors direct many aberrant projections (arrowheads) away from the hair cell cluster. The two immature hair cell pairs in the sema7a-/- neuromast are indicated by arrows. (D,E) Skeletonized networks portray the three-dimensional topology of the sensory arbors from the control and the sema7a-/- neuromasts depicted in panels B and C. The pseudocolored trajectories depict the increase in arbor contour length from each point of arborization, defined as the point at which the lateral line branch (magenta) contacts hair cell cluster. (F,G) Micrographs depict the skeletonized networks of the combined 4 dpf hair cell clusters, whose centers are located at (0,0). The X- and Y-coordinates represent the anteroposterior (AP) and the dorsoventral (DV) axes of the larva, respectively. Positive values of the X- and Y-coordinates represent the posterior and ventral directions, respectively. Combined skeletonized network traces from 27 control and 27 sema7a-/- mutant neuromasts are represented. H) The plot denotes the densities of the sensory arbors around the center of the combined hair cell clusters for 35 control (magenta) and 27 sema7a-/- (red) neuromasts at 4 dpf. The shaded area marks the region proximal to the boundary of the combined hair cell cluster. (I) The plot quantifies the degree of contact of the sensory arbors to their hair cell clusters in individual neuromasts from both control (magenta) and sema7a-/- mutants (red), each point represents a single neuromast. Thirty-three, 29, and 35 neuromasts were analyzed from 2 dpf, 3 dpf, and 4 dpf control larvae, respectively. Seventeen, 53, and 27 neuromasts were analyzed from 2 dpf, 3 dpf, and 4 dpf sema7a-/- mutant larvae, respectively. HC, hair cell; Scale bars, 5 µm; au, arbitrary unit; means ± SEMs; *** signifies p < 0.001; ** signifies p < 0.01.

Sema7A patterns the topology of the sensory arbor network .

(A) A schematic drawing of the combined hair cell cluster (green) and associated sensory axons (magenta) shows diverse topological features of the arborization network, including nodes, loops, bare terminals, node degree, and bare-terminal curvature. The apex is defined as the center of the hair bundle cluster and the base is denoted by the node from which the arborizations extends. The relative organ height describes the position along the axis formed between the arborization initiation point and the apex of the hair cell cluster. (B) The correlation between the relative number of bare terminals and loops from 35 control neuromasts and 27 sema7a-/- neuromasts at 4 dpf shows a negative correlation in the control but a scattered distribution in the mutants. The same neuromasts were analyzed in the subsequent panels. R2, correlation coefficient. Dashed red line depicts the regression slope. (C) The relationship of contour lengths to bare terminals and loops reveals reduced looping in the mutants. (D) Quantification of the number of nodes per loop demonstrates a decrease in mutants. (E) Mutants in the same larvae manifest a broader distribution of loop lengths. (F) A schematic diagram of the extending bare terminals depicts regions of high and low curvature along the length of the arbor. The distributions of bare terminals between low-and high-curvature groups show an enrichment of long, highly curved arbors in the control and short, linear arbors in the mutants. Statistical analyses are displayed in Table 1.

Statistical comparison between control and sema7a-/- larvae for topological features.

One-sided Kolmogorov-Smirnov tests were used to assess the difference in node degree heights, nodes per loop, and loop lengths for control versus sema7a-/- neuromasts at 2 dpf, 3 dpf, and 4 dpf. In each of these cases, the difference in the distributions were not significantly different (p-values in red) between the control and the mutants at 2 dpf and 3 dpf but became significantly different (p-values in green) at 4 dpf. Chi-square tests were used to assess the difference in occurrences of bare terminals versus loops, as well as the occurrences of bare terminals with low versus high curvature. The distribution at 4 dpf shows significant difference between the control and the mutant larvae. The occurrence of significant differences at 4 dpf for diverse topological features suggest that Sema7A function is required to achieve an ordered network topology at a mature stage. Thirty-three, 29, and 35 neuromasts were analyzed from 2 dpf, 3 dpf, and 4 dpf control larvae, respectively, and 17, 53, and 27 neuromasts were analyzed from 2 dpf, 3 dpf, and 4 dpf sema7a-/- mutant larvae, respectively. * signifies p < 0.05.

Ectopic Sema7Asec diffusive cue provides neural guidance in vivo.

(A) A diagrammatic overview depicts the generation of a transgenic animal that expresses the Sema7Asec protein ectopically under the control of a thermally inducible promoter. Larvae with ectopic myotomal-integration near the network of sensory arbors (red line) were imaged to analyze arbor morphology. (B) A schematic drawing of a sensory arbor from a heat-shocked larva depicts an extended axonal projection (arrowhead) that reaches toward the myofibers expressing the ectopic Sema7Asec protein. Parameters that quantitate the accuracy of the extended axonal projections toward the ectopic Sema7Asec sources are denoted. (C) In a micrograph of an ectopically expressing Sema7Asec (orange) larva, the sensory arbor (magenta) extends two aberrant axonal projections. One elongates (cyan arrowhead) along the somite boundary to reach and contact an ectopically integrated muscle progenitor cell (white dashed line) and the other (red arrowhead) reenters the posterior lateral-line nerve while following a second ectopic source. The through-focus scans (i–iii) from the epidermis to the dermomyotome and the three-dimensional (3D) surface reconstruction (iv) reveal the intimate contact between the aberrant sensory arbor (arrowheads) and the muscle progenitor cell. (D) In a micrograph of an ectopically expressing Sema7Asec (orange) larva, the sensory arbor (magenta) extends a single aberrant axonal process (cyan arrowhead) to reach ectopically integrated myofibers (white dashed line). The through-focus scans (i–iii) from the epidermis to the dermomyotome and the 3D surface reconstruction (iv) reveal the proximal association of the aberrant sensory arbor (arrowheads) to the myofibers. Melanocytes (yellow arrowhead) along the horizontal myoseptum intermittently block the visibility of the lateral-line nerve. (E) An injected, but not heat-shocked, control larva does not express ectopic Sema7Asec and does not show aberrant projection from the sensory arbor. (F) A plot demonstrates the accuracy of 18 extended axonal arbors in finding ectopic Sema7Asec sources. Each circle represents a single ectopic integration event. The two pairs of numbers represent the minimal and maximal lengths of the projection path (black) and its corresponding source path (green). (G) A plot quantitates the distribution of projection-proximity length from 18 ectopic integration events. MPC, muscle progenitor cell; scale bars, 20 µm; means ± SEMs.

Sema7A deficiency impairs synaptic assembly.

(A,B) Maximal-intensity projections of micrographs from a control and a sema7a-/- neuromast depict the presynaptic aggregates marked by CTBP (cyan). The approximate hair cell basal region is outlined by dashed green circles. (C-F) Plots quantitate the numbers of presynaptic aggregates (C,D) and their areas (E,F), in each hair cell across developmental stages. One hundred and fifty-two, 317, and 325 hair cells were analyzed from 2 dpf, 3 dpf, and 4 dpf control larvae, respectively. One hundred and sixty, 216, and 177 hair cells were analyzed from 2 dpf, 3 dpf, and 4 dpf sema7a-/-mutant larvae, respectively. (I,J) Maximal-intensity projections from micrographs of a control and a sema7a-/- neuromast depict the postsynaptic aggregates marked by MAGUK (yellow). The approximate hair cell basal region is outlined by dashed green circles. (K-N) Plots quantitate the numbers of postsynaptic aggregates (K,L) and their areas (M,N), in each hair cell across developmental stages. One hundred and fifty, 218, and 167 hair cells were analyzed from 2 dpf, 3 dpf, and 4 dpf control larvae, respectively. Ninety-seven, 141, and 141 hair cells were analyzed from 2 dpf, 3 dpf, and 4 dpf sema7a-/- mutant larvae, respectively. Scale bars, 5 µm.

Expression of diverse neural guidance cues in developing neuromasts.

Analysis of single cell RNA-sequencing data (9) shows the expression of multiple neural guidance genes that are expressed across young hair cells (Y), mature hair cells (M), and support cells (SC). Dot sizes and colors signify the proportion of cells in each cluster that express a gene and the average strength of expression (log[(counts/10,000) +1]), respectively.

Expression of Semaphorin7A in the lateral line neuromast.

(A) Distinct sets of primers (arrows) identify the two variants of the Sema7A protein molecule. (B-D) Surface micrographs show hair cells marked with actin-GFP (green dashed lines) and the localization of Sema7A protein (orange) in the subapical (arrowheads) and basal regions (arrows) of the corresponding hair cells from neuromasts of developing control larvae. The inset in panel B depicts the apices of the immature hair cell pair. (E) Immunolabeling with only the secondary antibody fails to detect any Sema7A signal. (F) A schematic diagram of a single hair cell depicts the three distinct regions along the apicobasal axis of the cell. Sema7A intensity was measured along this apicobasal axis (red dashed line). (G). A surface micrograph depicts hair cells (green, left) whose subapical region (red dashed line) harbors the Golgi network (cyan, middle) where the Sema7A protein is enriched (orange, right). Scale bars, 5 µm.

sema7a-/- mutants exhibit aberrant sensory-axon arborization throughout development.

(A) Micrographs depict the overall morphologies and sema7a genomic profiles of 2 dpf control (top) and sema7a-/-larvae (bottom). (B) In a surface micrograph of a 3 dpf control neuromast, myosin VI (green) marks hair cells (green dashed lines). Sema7A (orange) protein is highly enriched in the subapical region (arrowheads) and basolateral (arrow) region of the hair cells. (C,D) Micrographs depict combined 4 dpf hair cell clusters, each of whose centers arise located at (0,0). The X- and Y-coordinates represent the anteroposterior (AP) and dorsoventral (DV) axes of the larva, respectively. Positive values of the X- and Y-coordinates represent posterior and ventral directions, respectively. Combined hair cell clusters from 27 control and 27 sema7a-/- mutant neuromasts are represented. (E,F) Micrographs depict the skeletonized networks of the combined 2 dpf hair cell clusters from 17 control and 17 sema7a-/- mutant neuromasts. (G,H) Micrographs depict the skeletonized networks of the combined 3 dpf hair cell clusters from 29 control and 29 sema7a-/- mutant neuromasts. (I) A plot quantitates the densities of the sensory arbors around the center of the combined hair cell clusters for 33 control (magenta) and 17 sema7a-/-(red) neuromasts at 2 dpf. The shaded area marks the region proximal to the boundary of the combined hair cell cluster. (J) A plot quantitates the densities of the sensory arbors around the center of the combined hair cell clusters for 29 control (magenta) and 53 sema7a-/- (red) neuromasts at 3 dpf. The shaded area marks the region proximal to the boundary of the combined hair cell cluster. Scale bar, 5 µm; au, arbitrary units.

Quantification of sensory arbor distributions and contacts with hair cell clusters.

(A) A schematic drawing of the combined hair cell cluster and surrounding region depicts concentric sections of equal area in distinct colors (cyan, green, yellow, and red). The arbor density is plotted as a histogram against the distance (black arrow) from the center (black circle) of the combined hair cell cluster. The histogram is represented as a density trace graph (magenta). (B) A schematic drawing of a posterior lateral-line neuromast depicts the posterior lateral-line (PLL) nerve (black), posterior lateral-line branch (black), the sensory arbors (magenta) that contact the hair cell cluster (green), and the sensory arbors that project away (blue) from the hair cell cluster.

Quantification of topological features across developmental stages and genotypes.

(A-C) The distribution of node degrees as a function of the height in hair cell clusters shows the difference in the network’s hierarchical organization between control and sema7a-/- larvae. Nodes of low degree are commonly observed between genotypes across developmental stages (A, B), but their distribution along the cluster height became significantly different at 4 dpf (C), with nodes in control larvae localizing higher along the axis than in the mutants. In this and the following two panels, 33, 29, and 35 neuromasts were analyzed from 2 dpf, 3 dpf, and 4 dpf control larvae, respectively and 17, 53, and 27 neuromasts were analyzed from 2 dpf, 3 dpf, and 4 dpf sema7a-/- mutant larvae, respectively (D, E) The correlations between the relative number of bare terminals and loops from control and sema7a-/- neuromasts reveal dispersive distributions at both 2 dpf (D) and 3 dpf (E) stages, suggesting that the networks are undergoing reorganization and yet to adopt an ordered topology. (F, G) In the distribution of arbors between bare terminals and loops with respect to their contour lengths, mutants’ arbors display significantly higher propensity to remain as bare terminals compared to the controls at both 2 dpf (F) and 3 dpf (G) stages. The samples were identical to those in the previous panel. Statistical analyses are displayed in Table 1.

Quantification of topological features between control and sema7a-/- larvae across developmental stages.

(A, B) Quantifications of the number of nodes per loop demonstrate a broader, but not significantly different, distribution in the mutants than the controls at both 2 dpf (A) and 3 dpf (B). In this and the following panels, 33 and 29 neuromasts were analyzed from 2 dpf and 3 dpf control larvae, respectively and 17 and 53 neuromasts were analyzed from 2 dpf and 3 dpf sema7a-/- mutant larvae, respectively. (C, D) Quantifications of loop lengths from the same larvae indicate a wider, but not significantly different, distribution in the mutants than the controls at both 2 dpf (C) and 3 dpf (D). (E, F) The distributions of bare terminals between low- and high-curvature groups from the same larvae show higher occurrences of long, curved arbors in the control and short, linear arbors in the mutants at both 2 dpf (E) and 3 dpf (F) stages. Statistical analyses are displayed in Table 1.

Ectopic Sema7Asec diffusive cues induce aberrant neural arborizations.

(A-A”) In a micrograph of an ectopically expressing Sema7Asec (orange) larva, the sensory arbors (magenta) extend multiple aberrant axonal projections (arrowheads) that reenter the lateral-line nerve while following an ectopic Sema7A source. Maximal-intensity projections of the sensory arbors and the ectopically expressing Sema7A cells (white dashed line) are shown in A’ and A”, respectively. (B-B”) Micrographs depict three individual incidents of ectopic sema7asec (orange) integrations that occurred distant from neuromasts. In each case, a set of myofibers expresses the ectopic Sema7Asec protein that induces neurite extensions (arrowheads) from the lateral-line nerve (magenta) toward itself. Scale bars, 20 µm.

sema7a-/- mutants display impaired synaptic assembly.

(A-F) The plots quantitate the average numbers and areas of presynaptic aggregates from control and sema7a-/- neuromasts. Significant decreases in the numbers and areas of presynaptic aggregates are observed across development. Three hundred and forty-one, 1229, and 1286 presynaptic aggregates were analyzed from 2 dpf, 3 dpf, and 4 dpf control larvae, respectively. Two hundred and thirty-five, 643, and 419 presynaptic aggregates were analyzed from 2 dpf, 3 dpf, and 4 dpf sema7a-/- mutant larvae, respectively. (G-L) The plots quantitate the average numbers and areas of postsynaptic aggregates from control and sema7a-/- neuromasts. Significant decreases in the numbers and areas of postsynaptic aggregates occurred during development. Three hundred and ninety-seven, 1243, and 1300 postsynaptic aggregates were analyzed from 2 dpf, 3 dpf, and 4 dpf control larvae, respectively. Three hundred and four, 651, and 451 postsynaptic aggregates were analyzed from 2 dpf, 3 dpf, and 4 dpf sema7a-/- mutant larvae, respectively. Means ± SEMs; *** implies p < 0.001.