Bulk sequencing identifies genes that label early and late-born subsets of extraocular motor neurons.

(A) (Left) Schematic of the spatial pattern of extraocular motor nuclei nIII and nIV in the larval zebrafish. Motor neuron sub-populations labeled by isl1 (IR/MR/SO/SR) are color-coded by birthdate, with populations born before 33 hpf in magenta (“Early-Born”) and those born after 33 hpf in green (“Late-Born”). (Right) Dorsal-view confocal image of nIII/nIV motor neurons in a 50 hpf Tg(isl1:Kaede) larva photoconverted at 33 hpf. Only dorsal nIII neurons express photoconverted Kaede (magenta), ventral nIII and nIV neurons express only unconverted Kaede (green). (B) Volcano plot of gene expression fold change vs. adjusted p-value for 28,807 sequenced genes in Early-Born and Late-Born RNA libraries. Genes were significantly differentially-expressed when fold change > 2, and adjusted p-value < 0.01. 695 genes were differentially expressed, with 411 higher-expressed in early-born neurons (magenta) and 284 higher in late-born neurons (green). (C) PANTHER19.0 testing for over-representation of gene ontology terms for molecular function among the 695 differentially expressed genes. (D) Fluorescent in-situ hybridization examples for 2 candidate genes (ccdc85a and sim1a) predicted to be enriched in early-born OMNs. Probes targeting candidate RNA (red channel) were hybridized in Tg(isl1:Kaede) fish photoconverted at 33 hpf. Early-born neurons in dorsal nIII overlap with probe expression. Merge image shows the location of neurons expressing unconverted Kaede (white channel), used to draw the boundaries of nIII and nIV (dashed lines). Arrows indicate expression in dorsal nIII. Scale bar 20 µm. (E) Fluorescent in-situ hybridization examples for 2 candidate genes (alcama and sema3aa) predicted to be enriched in late-born OMNs. Probes targeting candidate RNA (red channel) were hybridized in Tg(isl1:Kaede) fish photoconverted at 33 hpf. Probe expression for late-born candidates is seen in ventral nIII and nIV, and does not overlap with photoconverted early-born neurons. Merge image shows the location of neurons expressing unconverted Kaede (white channel), used to draw the boundaries of nIII and nIV (dashed lines). Arrows indicate expression in ventral nIII or nIV. Scale bar 20 µm. (F) Venn diagram summarizing the spatial expression pattern of 25 candidate gene probes for fluorescent in-situ hybridization. Diagram circles indicate the subpopulations of OMNs where strongest probe expression was observed. Gene name color (magenta or green) indicates the population in which candidates were predicted to be enriched based on bulk RNA sequencing data.

Single cell sequencing identifies genes expressed in sub-populations of nIII/nIV motor neurons.

(A) Dissection schematic of Tg(isl1:GFP) ocular motor neurons (top) using targeted suction by micropipette. Example image of the same Tg(isl1:GFP) fish before and after targeted dissection of motor neurons in nIII and nIV show that cells of interest (dashed circle) are removed completely, while other nearby GFP+ cell populations (nV) remain intact. (B) Uniform Manifold Approximation and Projection (UMAP) representation of gene expression in 84 isl1:GFP+ cells dissected and sequenced at 48 hpf. Expression of canonical extraocular motor neuron gene markers show that analyzed cells have high expression of expected transcription factors (TFs) and cholinergic neuronal markers, consistent with a population of extraocular motor neurons. (C) UMAP representation of 84 isl1:GFP+ cells split into 3 cell clusters by gene expression profiles, representing genetic sub-populations of sequenced extraocular motor neurons. (D) Heat map of gene expression across all 82 cells split by UMAP cluster identify (rows) for the top 20 marker genes for each cluster (columns). Arrows below gene names indicate the marker genes that were highly expressed in late-born (green) or early-born (magenta) ocular motor neurons in previous bulk sequencing experiments. (E) UMAP representation of gene expression in 84 isl1:GFP+ cells for select gene markers of cluster 0, 1, and 2. (F) Venn diagram summarizing the spatial expression pattern of 25 cluster marker gene probes for fluorescent in-situ hybridization. Circles indicate the subpopulations of OMNs where strongest probe expression was observed. Gene name color indicates the UMAP cluster for which that gene was a top marker. (G) Fluorescent in-situ hybridization examples for candidate genes (nav2a and otx2a) marking cluster 0, (H) candidate genes (inab and mafba) marking cluster 1, (I) and candidate genes (nrp1a and sim1a) marking cluster 2 in single-cell sequencing experiments. Arrows indicate expression in ventral nIII (G), nIV (H), or dorsal nIII (I).

Both transiently- and persistently-expressed select candidate genes remain subtype-specific at multiple developmental stages.

(A) Fluorescent in situ hybridization for sim1a and nrp1a, candidate gene markers of dorsal nIII neurons, (B) for nav2a and tacc2, candidate gene markers of ventral nIII neurons, (C) and for inab and mafba, candidate gene markers of nIV neurons, at 33, 50, and 80 hpf in ventral and dorsal nIII and nIV. Arrows indicate expression in dorsal nIII (A), ventral nIII (B) or nIV (C).

Subtype-specific gene mutations cause vertical eye rotation deficits.

(A) Schematic of body tilt stimulus in eye rotation behavioral assay. Gray dotted box shows the trapezoidal velocity profile and angle of the tilt. Scale bars 2 seconds. (B) Example eye (solid) and platform (dotted) position during one stimulus step. Scale bar 0.5 seconds. (C) Left: eye position of a 7 dpf wildtype larva during each 15° nose-up tilt step of an experiment. Right: eye velocity of the same fish during nose-up tilts. Dashed gray line at the peak tilt velocity (35°/s). Scale bar 0.5 seconds. (D) Top: Average eye velocity ± SEM of wildtype siblings (WT, gray), phox2a heterozygotes (Het, blue), and phox2a homozygous mutants (Mut, pink) (n= 7, 9, and 6 fish) in response to pitch tilts in the nose-up (left) and nose-down (right) direction. Scale bar 0.5 seconds. Bottom: Gain (peak eye velocity / peak platform velocity) for each fish during nose-up (left) and nose-down (right) tilts. Wildtype versus mutant nose-up gains, PTukey’s = 1.4 × 10-4; heterozygote versus mutant nose-up gains, PTukey’s = 0.0017; wildtype versus mutant nose-down gains, PTukey’s = 8.1 × 10-5; heterozygote versus mutant nose-down gains, PTukey’s = 0.025. Color indicates genotype as above. Gray dashed line at gain = 1 (E) Top: Average eye velocity ± SEM of wildtype siblings (gray), sim1a heterozygotes (blue), and sim1a homozygous mutants (pink) (n= 6, 9, and 8 fish). Scale bar 0.5 seconds. Bottom: Gain for each fish. Wildtype versus mutant nose-up gains, PTukey’s = 0.063; heterozygote versus mutant nose-up gains, PTukey’s = 0.016; wildtype versus mutant nose-down gains, PTukey’s = 2.2 × 10-4; heterozygote versus mutant nose-down gains, PTukey’s = 0.0013. (F) Top: Average eye velocity ± SEM of wildtype siblings (gray), nav2a heterozygotes (blue), and nav2a homozygous mutants (pink) (n= 7, 10, and 9 fish). Scale bar 0.5 seconds. Bottom: Gain for each fish. Wildtype versus mutant nose-up gains, PTukey’s = 0.63; heterozygote versus mut nose-up gains, PTukey’s = 0.99; wildtype versus mutant nose-down gains, PTukey’s = 0.73; heterozygote versus mutant nose-down gains, PTukey’s = 0.96. (G) Top: Average eye velocity ± SEM of wildtype siblings (gray), onecut1 heterozygotes (blue), and onecut1 homozygous mutants (pink) (n= 9, 13, and 7 fish). Scale bar 0.5 seconds. Bottom: Gain for each fish. Wildtype versus mutant nose-up gains, PTukey’s = 0.73; heterozygote versus mutant nose-up gains, PTukey’s = 0.75; wildtype versus mutant nose-down gains, PTukey’s = 0.88; heterozygote versus mut nose-down gains, PTukey’s = 0.81.

phox2a and sim1a phenotypes likely reflect impairments to motor neurons.

(A) Schematic of vestibulo-ocular reflex circuitry. (B) Fluorescent in situ hybridization for phox2a in the stato-acoustic ganglion (1, left), tangential nucleus (2, center), and oculomotor and trochlear nuclei (3, right) at 5 dpf. Scale bars 25 µm. Data from 20. (C) Total number of isl1+ cells in nIII/nIV of phox2a wildtype control siblings (gray) and homozygous mutants (pink). Data from 20. (D) Number of isl1+ cells in dorsal nIII, ventral nIII, and nIV of sim1a wildtype control siblings (gray) and homozygous mutants (pink). (E) Number of isl1+ cells in dorsal nIII, ventral nIII, and nIV of nav2a wildtype control siblings (gray) and homozygous mutants (pink). (F) Number of isl1+ cells in dorsal nIII, ventral nIII, and nIV of onecut1 wildtype control siblings (gray) and homozygous mutants (pink). (G) Lateral view of fluorescent in situ hybridization for sim1a in the stato-acoustic ganglion (SAG, 1) and tangential nucleus (Tan, 2) at 33 (left), 50 (center), and 80 hpf (right). Scale bars 25 µm.

extraocular motor neuron birthdate varies with dorsoventral position

(A) Percent of Tg(isl1:Kaede) neurons that were born prior to the time of photoconversion in dorsal nIII (blue), ventral nIII (red) and nIV (yellow). Lines are averages across all fish, ribbons represent ± 1 standard deviation or range. N = 4 fish (18 hpf), 3 fish (22 hpf), 2 fish (26 hpf), 4 fish (30 hpf), 5 fish (34 hpf), 3 fish (42 hpf), 2 fish (46 hpf), 4 fish (50 hpf), and 4 fish (54 hpf) (B)Dorsoventral position of photoconverted Tg(isl1:Kaede) neurons in nIII (left) and nIV (right) colored by photoconversion timepoint (18-54 hpf).

Fluorescent in situ hybridization expression for candidate genes identified in bulk and single-cell RNA sequencing of extraocular motor neurons in nIII and nIV.

Maximum intensity projections of RNA probe expression in extraocular motor neurons split by dorsal (dorsal nIII and nIV) and ventral (ventral nIII) location. Black dashed outlines correspond to the location of extraocular motor neuron populations labeled by Tg(isl1:Kaede) (Kaede expression not shown). Scale bar 20 µm. (A) RNA probe expression for genes identified by bulk sequencing of late- (top, green) and early-born (bottom, pink) cells. Red arrows indicate expression in populations in which candidates were predicted to be enriched based on sequencing data. Red Ts indicate expression in other populations. (B) RNA probe expression for genes identified by both bulk and single-cell sequencing. Gene names are colored by late- (green) and early-born (pink) groups and organized by cluster (top: cluster 0, gray; center: cluster 1, orange; bottom: cluster 2, blue). Red arrows indicate expression in populations in which candidates were predicted to be enriched based on sequencing data. Red Ts indicate expression in other populations. (C) RNA probe expression for genes identified by single-cell sequencing. Genes are organized by cluster (top: cluster 0, gray; bottom: cluster 1, orange). Red arrows indicate expression in populations in which candidates were predicted to be enriched based on sequencing data. Red Ts indicate expression in other populations.

CRISPR/Cas9 mediated mutagenesis of select candidate genes

(A) Top: Schematic representing mutation created in nav2a. Red star indicates location of guides against nav2a DNA. Filled boxes show exons, open boxes show UTRs, and horizontal lines show introns. Bottom: Left shows DNA sequence in wildtype and nav2ad8 alleles. Red dashed lines indicate deleted sequence. STOP box indicates predicted premature stop codon due to deletion. Right shows predicted protein sequence. (B) Top: Schematic representing mutation created in sim1a. Bottom: DNA sequence and predicted protein in wildtype and sim1ad17 alleles. (C) Top: Schematic representing mutation created in onecut1. Bottom: DNA sequence and predicted protein in wildtype and onecut1d14 alleles.