Celsr3 is required for directionally unbiased acoustic startle escapes.

A) Schematic of a 5 dpf (days post-fertilization) larva (dorsal view) with a simplified diagram of the hindbrain circuit that drives the acoustic startle fast escape. +/- symbols indicate excitatory/inhibitory synapses. Dotted line indicates unknown connection. In response to a startling acoustic stimulus, Mauthner firing results in direct activation of contralateral motor neurons, leading to unilateral muscle contraction and a sharp turn (C1-bend). An iteration of this circuit model was previously used by the authors [87]. B) Representative right and left acoustic startle fast escapes in wild type 5 dpf larvae. Pre-Stimulus and C1-Bend: still images of when the acoustic stimulus was delivered (0 ms) and the highest angle of the Mauthner-dependent initial turn (C1-bend). Middle panel: 90 ms temporal projection of the response, including the C1-bend (begins arrow 1), counter bend (begins arrow 2), and subsequent swimming (begins arrow 3). Active Mauthner: predicted Mauthner that fired to drive right or left escape. C) Percentage of twenty acoustic stimuli that elicited any response (fast escape or response with >17 ms latency, with latency measured as first head movement following stimulus delivery). Each dot represents one fish. n = 126 siblings (mix of celsr3^+/+ and celsr3^-/+), 154 celsr3^-/- mutants. D) For larvae from (C), percentage of twenty acoustic stimuli that elicited a fast escape (latency <17 ms). 47% of mutant larvae responded within the sibling range (88–100%). E) For larvae from (D) that performed at least six fast escapes, initial C1-bend direction ratio (right versus left) for individual larva. Arrows indicate larvae that only turn one direction (“100% biased”; 24% of mutants). n = 126 siblings, 133 mutants.

Celsr3 is required for Mauthner development.

A,B) Tol-056 transgene expression of cytosolic GFP in Mauthner neurons in celsr3 sibling (A) and mutant (B) larvae. Mauthner axons are labeled with α-3A10 (neurofilament antibody) staining. All images and quantifications in figure are of 5 dpf larvae. Scale bar = 20μM, Z-projection. C) Mauthner cell number in celsr3 siblings (100% 2 M-cells; n = 51) and celsr3 mutants (12% 1 M-cell, 88% 2 M-cells; n = 67). D,E) Mauthner somas and axons in Tol-056 celsr3 sibling (D) and mutant (E). The Mauthner/spinal cord is pseudo-colored green for ease of visualization, while non-Mauthner/spinal cord GFP expression is magenta. Only z-frames with the Mauthner axon in focus are included. Arrows indicate where the Mauthner axons end and corresponding muscle segments, numbered anterior to posterior. Scale bar = 500μM, Z-projection. F,G) Magnified Mauthner axon ends in Tol-056 celsr3 sibling (F) and mutant (G). The top axon in (G) continues to segment 30. Scale bar = 50 μM, Z-projection. H) Mauthner axon lengths in celsr3 siblings and mutants measured by muscle segment containing the Mauthner axon end. Average sibling = segment 32 (n = 22 axons), mutant = segment 29 (n = 66 axons). I) Escape C1-bend ratio for celsr3 siblings and mutants that performed at least six fast escapes in response to 20 acoustic stimuli, normalized so all fish turning only one direction, left or right, appear at the same y position. n = 76 siblings, 6 mutants with one Mauthner (1M), 109 mutants with two Mauthners (2M). J) Escape frequency (Escape%) for individual Mauthners was calculated using total Escape% and turn bias ratio for individual fish; left C1-bend escape was assigned to right Mauthner, right C1-bend escape assigned to left Mauthner. n = 20 sibling Mauthners, 253 celsr3 mutant Mauthners. Points outlined in black indicate celsr3 mutant Mauthners from larvae with a single Mauthner (1M). Individual Mauthner escape% and axon length are not significantly correlated (simple linear regression, p = 0.63) in celsr3 mutants.

Celsr3 is required for spiral fiber axon guidance and glial localization to the axon cap.

A) The axon cap (dotted red box in circuit diagram) is comprised of excitatory spiral fiber axons (blue) and feedback inhibitory axons (pink) synapsing with the Mauthner (green) axon initial segment (AIS), surrounded by astrocyte-like glia (dotted line). B-E) A wild type larva with j1229a transgene expression of cytosolic GFP in Mauthner and spiral fiber neurons (B,C), hcrt:Gal4, UAS:Kaede/hcrt::Kaede transgene expression in a subset of spiral fibers (B,D; protein mostly converted from green to red), and α-3A10 staining (B,E). White brackets in (B) indicate spiral fiber somas. Yellow brackets in (D) indicate spiral fiber axons at the axon cap. White arrows in (E) indicate spiral fiber axons at midline crossing. All images and quantifications in figure are of 5 dpf larvae. Scale bar = 20μM, Z-projection. F-J) Magnification of box in (B) showing both Mauthner axon caps with TO-PRO-3 labeling nuclei (F,G), α-S100B labeling glia (F,H), Kaede labeling spiral fiber axons (F,I), and α-3A10 labeling Mauthner axons (F,J). Yellow arrowheads in (J) indicate Mauthner AISs. Scale bar = 20μM, single Z-frame. K,L) celsr3 mutant larva with Tol-056:GFP (K) and hcrt::Kaede (K,L) transgene expression. White asterisks indicate spiral fiber soma (note: only a subset of spiral fiber neurons are labeled due to variegated transgene expression). Yellow brackets indicate normal spiral fiber axon targeting to Mauthner axon cap. Red arrow indicates inappropriate spiral fiber axon targeting. Scale bar = 20μM, Z-projection. M,N) Mauthner axon cap region of sibling (M) and celsr3 mutant (N) Tol-056:GFP larvae. Kv1.1 localizes to Mauthner axon caps (yellow brackets) that spiral fiber axons target and accumulates at ectopic positions (red arrows) where spiral fiber axons frequently misproject. Scale bar = 20μM, Z-projection (dorsal range included for posterior Mauthner axon cap region and ventral range included for anterior spiral fiber axon misprojections, for clarity). O) Schematic for quantifying spiral fiber misprojections, measured as accumulation of Kv1.1 in ectopic “axon caps,” as in (N) (red arrows), within 60μM anterior to Mauthner AIS and 60μM lateral to midline (boundaries of spiral fiber soma). Left brain hemisphere is normal, with 0 ectopic caps, while right brain hemisphere has 1 ectopic spiral fiber cluster. For quantification in (P), this example would be counted as 0 (for left) and 1 (for right). P) Quantification of “ectopic caps,” presumed to represent spiral fiber misprojections, assessed by counting numbers of large Kv1.1 accumulation (>5μM diameter) rostroventral to position of normal axon cap, as in (O). For each larva, the left and right hindbrain were counted independently (two “hemispheres” per larva). In addition to ectopic Kv1 accumulation, all sibling and celsr3 mutant hemispheres with a Mauthner present had Kv1.1 accumulation at the endogenous axon cap region surrounding the Mauthner AIS. celsr3 mutant brain hemispheres where Mauthner neurons were absent only had ectopic caps. n = 10 sibling hemispheres, 58 celsr3 mutant hemispheres with Mauthners, 9 celsr3 mutant hemispheres without Mauthners. Q) Schematic for quantifying anterior-posterior position of ectopic caps quantified in (P). 1 = positioned at Mauthner AIS, 2 = <20μM from correct position, 3 = posterior to spiral fiber soma, 4 = at posterior spiral fiber midline crossing, 5 = at anterior spiral fiber midline crossing. For example, ectopic cap in (O) would be counted as 4. When the Mauthner neuron was absent, additional axon tracts labeled by Kv1.1 were used to identify where correct axon cap positioning was expected. R) Quantification of anterior-posterior position of ectopic caps from (P). For brain hemispheres with two ectopic caps, each position was counted separately. n = 63 celsr3 mutant ectopic caps in hemispheres with Mauthners, 17 celsr3 mutant ectopic caps in hemispheres without Mauthners. S-X) Sibling (S-U) and celsr3 mutant (V-X) larvae with Glast:myrGFP-p2A-H2AmCherry (membrane GFP, nuclear mCherry) transgene labeling astrocytes, as well as Tol-056:GFP and hcrt::Kaede (partially converted from green to red) transgenes. Green channel (myrGFP, GFP, and Kaede) in S,V,T,W; red channel (H2AmCherry, Kaede) in S,V,U,X. White asterisks indicate spiral fiber soma (note: only a subset of spiral fiber neurons are labeled due to variegated transgene expression). Yellow brackets indicate endogenous Mauthner axon cap, red brackets indicate ectopic localization of glia, red arrow indicates misprojected spiral fiber axons, and yellow arrowheads indicate two example glia localized correctly to the endogenous axon cap. Scale bar = 20μM, Z-projection. Y) Quantification of percent of sibling or celsr3 mutant larvae with zero, one (left or right), or two (left and right) anterior ectopic glial clusters. All larvae for quantification had two Mauthners. n = 16 siblings, 15 celsr3 mutants. Z) For larvae in (Y), glia on one side of the hindbrain (left or right hemisphere) were categorized as correctly localized to only the endogenous axon cap at the Mauthner AIS, localized to both the endogenous axon cap and an anterior ectopic location, localized only to an ectopic location, or not observed/missing. n = 32 sibling hemispheres, 30 celsr3 mutant hemispheres.

Asymmetric spiral fiber input to Mauthner neurons in celsr3 mutants correlates with biased escapes.

A-D) Region of hindbrain with Mauthner and spiral fiber neurons from 5 dpf Tol-056:GFP larvae in (A) siblings with unbiased escapes (40–60% right/left C1-bend direction ratio; all categories >6 escapes over 20 stimuli), (B) celsr3 mutants with unbiased escapes, (C) celsr3 mutants that only made left C1-bends (right Mauthner predicted active), and (D) celsr3 mutants that only made right C1-bends (left Mauthner predicted active). Composite images are Z-projections with ventral range for anterior portion and dorsal range for posterior portion. Kv1.1 (left and middle panels) localizes to endogenous axon caps (brackets) and ectopic caps (red arrows in B-D) presumed to represent spiral fiber misprojections. α-Cx35 (left and right panels) labels spiral fiber synapses at endogenous (A-D) and ectopic (B-D) caps. White dashed brackets in (C,D) indicate the axon cap with reduced Kv1.1/Cx35 signal, which corresponds to the inactive Mauthner, based on escape bias. Scale bar = 20 μM.

celsr2 compensates for loss of celsr3 during Mauthner development.

A-F) j1229a:GFP transgene expression (A,B,D,E) labels Mauthner (A,B) and spiral fiber (asterisks; A,B,D,E) neurons in celsr2^-/- (A-C) and celsr3^-/-;celsr2^-/- (D-F) larvae. α-3A10 (A,C,D,F) labels spiral fiber axons at midline crossing (white arrowheads) and Mauthner axons (A,C). White dashed brackets (D-F) indicate positions where Mauthner neurons are absent. All images and quantifications in figure are of 5 dpf larvae. Scale bar = 20μM, Z-projection. G) Mauthner cell number in celsr3 sib;celsr2^-/- (100% 2 M-Cells; n = 44), celsr3^-/-;celsr2^+/+ (5% 1 M-cell, 95% 2 M-cells; n = 20), celsr3^-/-;celsr2^+/- (20% 0 M-cell, 39% 1 M-cell, 41% 2 M-cells; n = 41), and celsr3^-/-;celsr2^-/- (91% 0 M-cell, 9% 1 M-cell; n = 22). p-values calculated with Chi-square tests. H) Percentage of twenty acoustic stimuli that elicited any response for celsr3 sib;celsr2^-/- (n = 28) and celsr3^-/-;celsr2^-/- (n = 17) larvae. I) For larvae from (H), percentage of twenty acoustic stimuli that elicited a fast escape (latency <17 ms). Points outlined in black indicate celsr3^-/-;celsr2^-/- larvae with one Mauthner (5/17), while the remaining celsr3^-/-;celsr2^-/- larvae had zero Mauthners. J) For celsr3 sib;celsr2^-/- larvae from (I) that performed at least six escapes (100% of larvae), escape C1-bend ratio, normalized so all fish turning only one direction, left or right, appear at the same y position. K) Latency (time to first head movement) for individual fast escapes in celsr3^-/-;celsr2^-/- larvae with one (1M) or zero (0M) Mauthners. n = 19 responses 1M, 12 responses 0M.

FBMN migration is independent of celsr3.

A-D) isl1:GFP transgene expression of cytosolic GFP in 48 hours post-fertilization (hpf) facial branchiomotor neurons (FBMNs) in sibling (A), celsr2^-/- (B), celsr3^-/- (C), and celsr3^-/-;celsr2^-/- (D) embryos. α-3A10 labels Mauthner somas and axons (left panels). Approximate rhombomere boundaries (r4, r5, r6) are indicated with brackets (right panels). Asterisks in (C,D) denote presumptive dying Mauthners based on soma morphology. Scale bars = 20 μM, Z-projection. E) Percent of larvae with full (e.g. A,C), intermediate (e.g. D), and minimal (e.g. B) FBMN migration. n = 75 sib, 32 celsr3^-/-, 11 celsr2^-/-, 10 celsr3^-/-;celsr2^-/-. p-value calculated with Fisher’s exact test.

Spiral fiber axon guidance is partially independent from Mauthner development A) Quantification of percent of twenty acoustic stimuli that resulted in a fast escape (latency <17 ms). All images and quantifications in figure are of 5 dpf larvae. n = 55 celsr3 siblings, 51 celsr3 sib;fzd3a^-/-, 56 celsr3^-/-, 92 celsr3^-/-;fzd3a^-/+, 45 celsr3^-/-;fzd3a^-/-. B) Mauthner cell number in larvae from (A). n = 53 celsr3 sib;fzd3a^-/- (100% 2 M-cells), 55 celsr3^-/- (15% 1 M-cell, 85% 2 M-cells), 86 celsr3^-/-;fzd3a^+/- (7% 1 M-cell, 93% 2 M-cells), 45 celsr3^-/-;fzd3a^-/- (36% 0 M-cells, 33% 1 M-cell, 31% 2 M-cells). C) Escape C1-bend ratio for larvae from (A) with two Mauthners (B) that performed at least six fast escapes in response to 20 acoustic stimuli, normalized so all fish turning only one direction, left or right, appear at the same y position. n = 53 celsr3 siblings, 48 celsr3 sib;fzd3a^-/-, 41 celsr3^-/-, 71 celsr3^-/-;fzd3a^-/+, 10 celsr3^-/-;fzd3a^-/-. D) Mauthner cell number in control (no hoxb1a:Gal4, UAS:fzd3aΔC/hoxb1a::fzd3aΔC transgenes) and hoxb1a::fzd3aΔC larvae. n = 28 control (100% 2 M-cells), 52 hoxb1a::fzd3aΔC (12% 0 M-cells, 27% 1 M-cell, 62% 2 M-cells). E) Quantification of percent of twenty acoustic stimuli that resulted in a fast escape (latency <17 ms). n = 30 control, 65 hoxb1a::fzd3aΔC. F) Escape C1-bend ratio for larvae from (E) with two Mauthners that performed at least six fast escapes in response to 20 acoustic stimuli, normalized so all fish turning only one direction, left or right, appear at the same y position. n = 23 control, 27 hoxb1a::fzd3aΔC. G,H) j1229a:GFP labels Mauthner cells and spiral fiber neurons in hoxb1a::fzd3aΔC larvae while α-S100B labels glia. Yellow brackets indicate correctly positioned GFP+ spiral fiber axons at the Mauthner AIS, while white dashed brackets indicate displaced spiral fiber axon caps, due to either an absent Mauthner (G) or mispositioned Mauthner AIS (H). Yellow arrowheads in (H) denote aberrant path of left Mauthner axon, which continues posteriorly out of the Z-range displayed. Scale bars = 20 μM, Z-projection (ventral anterior range and dorsal posterior range for GFP, for clarity; division marked by dotted line). I) Quantification of “ectopic caps,” as in Fig 3O and 3P. In brain hemispheres with no Mauthners, any large Kv1.1 accumulations were considered “ectopic” since no endogenous Mauthner AIS was present. n = 16 control hemispheres, 60 hoxb1a::fzd3aΔC hemispheres with Mauthners, 32 hemispheres without Mauthners. J) Quantification of anterior-posterior position of ectopic caps from (K), as in Fig 3Q and 3R. n = 32 hoxb1a::fzd3aΔC ectopic caps in hemispheres without Mauthners.

The PCP pathway directs A-P axon guidance of Mauthner cells.

A-B) Wild type hindbrain with α-3A10 labeling Mauthner somas and axons. Mauthner axons grow posteriorly, towards the bottom of the image. All images and quantification in figure are from 30 hpf embryos. B) Two representative examples of hoxb1a::fzd3aΔC-GFP transgenic hindbrains with Mauthner somas and axons labeled with α-3A10. Yellow arrowheads indicate axons from left Mauthners and red arrows indicate axons from right Mauthners. In top panels, the left axon grows correctly posteriorly while the right axon grows incorrectly towards the anterior, and in bottom panels, both axons grow incorrectly towards the anterior. Scale bars = 20 μM, Z-projection. C) Quantification of Mauthner axon phenotypes in hoxb1a::fzd3aΔC positive (n = 18) and celsr3^-/-;celsr2^-/- (n = 19) embryos. D,E) Sibling (D) and celsr3^-/-;celsr2^-/- (E) embryos with hoxb1a in situ (magenta) labeling hindbrain rhombomere 4 and nefma in situ (green) labeling reticulospinal neurons, including Mauthner cells (arrowheads). Scale bars = 20 μM, Z-projection. F,G) Mauthner soma and axons are labeled with α-3A10 in siblings (F) and celsr3^-/-;celsr2^-/- (G) embryos. Yellow arrowheads indicate axon from left Mauthner growing correctly towards the spinal cord but displaced laterally. Red arrowheads indicate axon from right Mauthner growing incorrectly towards the anterior. Dotted lines indicate midline and encircle developing ears, for reference. Scale bars = 20 μM, Z-projection.