a Shank3 protein diagrams of human SHANK3 and duplicated zebrafish Shank3a and Shank3b show where in the zebrafish proteins four independent CRISPR-Cas 9 indel alleles introduce frameshift mutations. Protein interaction domains indicated in human SHANK3 are more highly conserved in zebrafish Shank3a than Shank3b (SPN = Shank/ProSAP N-terminal, ANK = ankyrin repeats, SH = SRC Homology 3, PDZ = post-synaptic density protein/disc large/zonula occludens-1, PRR = proline-rich region that includes interaction domains with H = Homer, C = cortactin, A = actin binding protein 1, and SAM = sterile alpha motif). Each shank3abΔN (purple) and shank3abΔC (orange) mutant model has similar mutations in Shank3a and Shank3b paralogs: shank3abΔN mutations are in ankyrin repeat regions and shank3abΔC mutations are in the proline-rich region. b Coronal cryosections from 6dpf larvae were stained with antibodies against the glutamatergic post-synaptic scaffolding proteins PSD-95 and Shank3, with a representative sample shown for each genotype. Synapses in wild-type cerebellum (Ce) stain for both PSD-95 (green arrowheads) and Shank3 (magenta arrowheads) puncta, some of which colocalize (white arrowheads), compared to shank3ab mutants that stain for PSD-95, but not Shank3. TeO= Optic Tectum; Scale bars represent 10 µm. c Visual motor responses (VMR) are shown as line graphs of median distance traveled in 30 s ±SE to four cycles of lights-on to lights-off transitions. White and black boxes below the x-axis indicate alternating lights-on and lights-off, respectively. Exact sample sizes of biologically independent samples are indicated below the Paired dot plots and apply to plots in (c–e). d Paired dot plots compare median swimming distances per larva of the four light transitions in the 30 s before and after the lights-on to lights-off transition. Within genotype comparisons were conducted using Dunn-Bonferroni p-value corrected t-tests (e) Box plots compare distance traveled during the first 30 s of dark between WT and shank3ab mutant models. Boxes denote the median, 1st and 3rd quartile, while whiskers represent the minimum and maximum values. Groups were statistically compared using a Kruskal–Wallis ANOVA, and when p < 0.05, were followed by a Dunn’s multiple comparison test. P-value asterisks represent; p < 0.05 - *, p < 0.01 - **, p < 0.001 - ***, p < 0.0001-****. Source data for plots are provided in Supplementary Data 2.

Brain-wide activity maps were generated by using phosphorylated-ERK (pERK) antibody staining as a proxy for neuronal activity. a Individual larval stacks were registered for use with the Z-brain atlas and MAP-mapping MATLAB scripts (Randlett et al. ¹², Engert lab) Individual pERK stacks left were then divided by total-ERK (tERK; middle), providing normalized pERK/tERK signal (right). b Median p/tERK values were then calculated for every voxel within the brain for each genotype and light condition (Exact sample sizes of biologically independent samples for each condition and genotype (n = lights-on/lights-off); wild-type (n = 16/19), shank3abΔN (n = 19/21) and shank3abΔC (n = 16/15) are indicated on group median images in (b) and also apply to (c) and (d). c Mann–Whitney U z-scores were calculated, comparing lights-off and lights-on, with magenta indicating increased activity during the transition to lights-off (e.g. Medulla Oblongata, MO) and green indicating increased activity during the transition to lights-on (e.g. Optic Tectum, TeO). Regions within the brain that are black did not reach the P < 10⁻⁵ cut-off. d In comparison to wild-type, shank3abΔ−/− mutant models respond to the lights-off condition (magenta) with activation of their pineal (P), but fail to show activation in the MO and spinal cord (sc). a–c All images are 20 µm dorsal z-projections. d Whole brain z- and x-projections. Scale bars = 50 µm.

a A cartoon shows how cells from wild type donor embryos marked by a ubiquitously expressed dTomato fluorescent protein (ubi:zebrabow) are transplanted into the presumptive hindbrain of shank3ab−/− mutant recipient embryos at mid-gastrulation stages. b Chimeric embryos at 1 day post-fertilization (dpf), with donor cells expressing the fluorescent protein (false-colored in cyan) in recipient shank3abΔN−/− or shank3abΔC−/− embryos. Chimeric six-day-old larvae (shank3ab−/−:Zb-T) were imaged to determine the fate of the transplanted cells. c Confocal images of chimeric larvae at 6 dpf following behavioral screening, demonstrating transplanted cells in rescued larvae populate the dorsal/rostral brainstem nuclei. Individual representative larvae are numbered 1-3, with the three averaged in the right most stack. d VMR line graphs, median +/− SE, (d–f) and (e) paired dot plots show lights-off behavioral phenotypes are rescued in both shank3abΔ−/− mutant models with wild-type-derived brain stems (shank3abΔ−/−:Zb-T). Exact sample sizes of biologically independent samples for each genotype and chimera are indicated below each plot and also apply to d and f. Within shank3 model comparisons were conducted using Dunn–Bonferroni p-value corrected t-tests. f Box plots displaying median swimming distances for individuals following the first 30 s following lights-off. Individual values are medians representing all four lights-off transitions for individual larvae. Boxes denote the median, 1st and 3rd quartile, while whiskers represent the minimum and maximum values. Groups were statistically compared using Kruskal-Wallis one-way ANOVA, and when p < 0.05, were followed by Dunn’s multiple comparisons. P-value asterisks represent; p < 0.05 - *, p < 0.01 - **, p < 0.001 - ***, p < 0.0001-****. Scale bars = 100 µm (b); 50 µm (c). Source data for plots are provided in Supplementary Data 2.

a z-stack image from CobraZ reference and 26 segment atlas. Regions outlined in black denote the locus coeruleus (LC) and medulla oblongata (MO). Scale bar = 100 µm. b Box plots comparing relative volume of LC and MO brain segments in wild-type, shank3abΔn and shank3abΔc mutants, and shank3ab mutants with wild-type transplanted brain stems (shank3abΔZb-T). Exact sample sizes of biologically independent samples for each condition and genotype (n = lights-on/lights-off); wild-type (n = 16/19), shank3abΔN (n = 19/21) and shank3abΔC (n = 16/15). Boxes denote the median, 1st and 3rd quartile, while whiskers represent the minimum and maximum values. Brain segment sizes were analyzed using a non-parametric Kruskal–Wallis one-way ANOVA and followed by a Dunn’s corrected multiple values comparison. P-value asterisks represent; p < 0.05 - *, p < 0.01 - **, p < 0.001 - ***, p < 0.0001-****. Source data for plots are provided in Supplementary Data 2.