Maternal rest is required for proper FBMN migration. (A) PCR genotyping of mating crosses to produce rest zygotic and maternal–zygotic (MZ) mutant embryos. The rest^sbu29 allele introduces a 7 bp deletion resulting in a smaller PCR product (red arrowhead). Homozygous mutant females (red box) produced eggs lacking maternal rest transcript. (B–F) Average intensity projection confocal images (dorsal views) of Tg(islet1:GFP); rest^sbu29 mutant embryos at 48 hpf, immunolabeled with EphA4a to allow visualization of r5 (red). Scale bar=20 µm. Decreasing levels of rest from B to F correlate with increasing disruption in FBMN migration. (B′–F′) Pie charts represent the average number of FBMNs counted in r4 (black), r5 (checkered), and r6 (white) as a percentage of total FBMNs in rest mutant embryos. The total number of FBMNs was not significantly different in embryos of different genotypes. (G) Statistical analysis of FBMN migration in rest^sbu29 mutants. The percentages of FBMNs that fully migrate to r6 (white) were compared using unpaired t test. Parentheses indicate number of embryos analyzed per group. Mean±SEM; ^*P<0.05; ^**P<0.01; ^***P<0.001; ^****P<0.0001; n.s.=not significant. (H and I) Tg(islet1:GFP) embryos co-injected with both splice- and translation-blocking morpholinos targeted against Rest recapitulate the MZrest^sbu29/sbu29 mutant phenotype at 48 hpf. Scale bar=20 µm. (J and K) MZrest^sbu29/sbu29 FBMNs fail to reach r6 at 72 hpf, indicating that the mutant phenotype is not merely a consequence of developmental delay.

Rest function is required within FBMNs for their normal migration. (A) Schematic of Rest protein domains, including eight zinc finger DNA binding domains, and corepressor binding domains at amino and carboxy termini. Dominant negative Rest4 (dnRest4) is a truncation of zebrafish Rest with ECDVLG (human sequence) added at the C-terminus (Mapp et al., 2011). REST-VP16 comprises human REST DNA-binding domain fused to VP16 activator with both corepressor binding domains removed (Immaneni et al., 2000). (B) Schematic of Tol2-mediated transgenesis to express mCherry-tagged constructs specifically in FBMNs. Coinjection of DNA constructs with transposase mRNA produces mosaic expression in Tg(islet1:GFP) FBMNs; migration was analyzed at 48 hpf, with r5 marked by EphA4a (red). Embryos classed according to number of mCherry+ FBMNs into five groups: zero (controls, C and H), 1–10, 11–20, 21–30, and >30; n= # embryos analyzed. (D–G) Increasing number of dnRest4 expressing FBMNs progressively disrupts migration. (I–L) Increasing number of REST-VP16 expressing FBMNs causes a greater disruption in migration. (C′–L′) Counts of Tg(islet1:GFP) FBMNs (green) in r4 (black), r5 (checkered) and r6 (white). A greater number of FBMNs remain in r4 and r5 as more FBMNs express each construct. (D′′–L′′) Counts of mCherry+ FBMNs (red/yellow) in each group of mCherry-expressing embryos. mCherry+ FBMNs are distributed in all three rhombomeres, suggesting FBMNs migrate as a collective group. Split channel images confirmed that all mCherry+ cells were also GFP+ as expected.

FBMNs lacking Rest repression send out aberrant neurites. (A and B) Average intensity projection confocal images (dorsal views) of Tg(islet1:GFP)-expressing FBMNs at 48 hpf. Scale bar=10 µm. (A) In wild type specimens protrusions across the midline and laterally (arrows) occur from FBMNs localized in r6. By contrast, in MZrest^sbu29/sbu29 mutant specimens protrusions laterally and across the midline frequently occur from FBMNs localized in r4 and r5 (B). (C–D′) Ectopic protrusions (arrows) in mCherry-REST-VP16 mosaic embryos are primarily sent out by FBMNs expressing the construct, as indicated by mCherry expression (D′). (E–G) Average number of neurites extended in r4 (black), r5 (checkered), and r6 (white) by FBMNs in rest mutant embryos (E) and those mosaically expressing dnRest4 (F) or REST-VP16 (G). (H–J) FBMNs expressing mosaic mCherry-REST-VP16 send out ectopic protrusions in r4 and r5 that are also Map2-positive, showing that dendritic branching occurs in inappropriate locations (arrows). (K and L) Lateral views of Tg(zCREST1:membRFP)-expressing FBMNs at 24 hpf. FBMNs in MZrest^sbu29/sbu29 embryos send out long protrusions during migration (arrow). Scale bar=10 µm.

Morphological analysis reveals precocious maturation of Rest-deficient FBMNs. (A–G) Analysis of FBMNs at 20 and 28 hpf in Tg(zCREST1:membRFP) wild type embryos (A and B) and MZrest^sbu29/sbu29 embryos (C and D). Average measurements from 5 embryos are shown for wild type and MZrest^sbu29/sbu29 (E–G) embryos at each stage. The wild type length-to-width ratio decreases from 20 to 28 hpf, as neurons become rounder; by contrast MZrest^sbu29/sbu29 FBMNs are round at both 20 and 28 hpf (E). Both number and length of protrusions increase over time in wild type and MZrest^sbu29/sbu29 embryos (F and G). MZrest^sbu29/sbu29 FBMNs have significantly more and longer protrusions than wild type FBMNs. (H–J) Volume measurements of FBMNs at 22 hpf. Whole cell and nuclear volume were measured for 10 “immature” neurons in r4 (yellow dots) versus 10 “mature” neurons in r6 (red dots) for 5 wild type (H) and 5 MZrest^sbu29/sbu29 (I) embryos. Wild type FBMN whole cell volume increases significantly as neurons migrate from r4 to r6; MZrest^sbu29/sbu29 FBMN whole cell volume does not alter as neurons migrate from r4 to r6 and is similar to that of wild type neurons in r6. Nuclear volume does not change significantly. Mean±SEM; ^*P<0.05; ^***P<0.001, ^****P<0.0001, n.s.=not significant.

Rest targets are inappropriately expressed in rest and pk1b mutants. In situ hybridization analysis of Rest target gene expression in Tg(islet1:GFP) embryos at 24 hpf in wild type, MZrest^sbu29/sbu29, and pk1b^fh122/fh122 specimens. Rest targets selected for analysis are adam23a (A–C′), snap25b (D–F′), srrm4 (G–I′), and syt4 (J–L′). We also analyzed the non-Rest target and FBMN maturation marker cmet (M–O2). All five genes are expressed in fully migrated FBMNs at 48 hpf. Upper panels show co-expression with islet1:GFP to label FBMNs, with gene expression alone shown below. At this stage (24 hpf), the selected Rest targets, as well as the maturation marker cmet, are not expressed in wild type FBMNs, but are expressed in both rest and pk1b mutant FBMNs (arrows).

Srrm4 knockdown rescues FBMN migration in rest mutants. (A) Model of Rest/Srrm4 regulation of FBMN maturation and the effect on FBMN migration. (B–E) Average projection dorsal views of 48 hpf FBMNs (green) in Tg(islet1:GFP) embryos immunolabeled with EphA4a to mark r5 (red). FBMN migration is unaffected in Srrm4 morphants (C) as compared to uninjected controls (B). MZrest^sbu29/sbu29 mutant FBMN migration (D) is partially rescued with knockdown of Srrm4 (E). (F) Quantitative counts of FBMN migration in r4 (black), r5 (checkered), and r6 (white) as % of all FBMNs. A significantly increased number of FBMNs migrate to r6 in MZrest^sbu29/sbu29 mutants injected with Srrm4 MO as compared to uninjected mutants. Mean±SEM; ^****P<0.0001. (G–J) Analysis of FBMNs at 22 and 30 hpf in Tg(zCREST1:membRFP);MZrest^sbu29/sbu29 embryos (G and H), and Tg(zCREST1:membRFP);MZrest^sbu29/sbu29 embryos injected with Srrm4 MO (I and J). Average morphological measurements from 5 embryos are shown at each stage (K–N). Mutants injected with Srrm4 MO have FBMNs that are more elongated than those injected with Control MO, and remain so at 30 hpf (K). They also send out fewer and shorter protrusions than MZrest^sbu29/sbu29 at both early and late stages (L–M). Mutants injected with Srrm4 MO have a smaller cytoplasmic volume compared to mutant controls at 22 hpf, which swells significantly at the later stage compared to the unchanged volume in MZrest^sbu29/sbu29 mutants (N).

Nuclear-localized Rest rescues FBMN migration in Rest-deficient but not Pk1b-deficient embryos. (A) Schematic of Rest protein domains with the added NLS sequence PKKKRKV at the C-terminus (Rest-NLS). This construct was tagged with mCherry and cloned into both pCS2+ and Tol2 vectors. (B–D) Expression of mCherry-tagged Rest-NLS (injected mRNA) in blastula stage embryos (5 hpf), compared with expression of Rest protein expression using immunohistochemistry. Rest-NLS is correctly localized to the nucleus while endogenous Rest is expressed in the cytoplasm and nucleus (Kok et al., 2012). (E–H) Expression of mCherry-Rest-NLS specifically in FBMNs using Tol2 transgenesis (as described in Fig. 2) in Tg(islet1:GFP);MZrest^sbu29/sbu29 mutant embryos. Embryos were classed into four groups based on the number of mCherry+ FBMNs present: zero (controls, E), 1–10, 11–20, and 21–30. (F–H) Increasing the number of FBMNs expressing Rest-NLS rescues FBMN migration to r6. (E′–H′) Counts of Tg(islet1:GFP) FBMNs (green) in r4 (black), r5 (checkered) and r6 (white). (F′′–H′′) Counts of mCherry+ FBMNs (red/yellow) in each group of mCherry-expressing embryos. mCherry+ FBMNs are more likely to reside in r6, suggesting their ability to “pull along” mutant FBMNs lacking the construct to complete migration. Split channel images confirmed that mCherry+ cells were also GFP+ as expected. (I–K) Rest-NLS expression within Pk1b mutant FBMNs using Tol2 transgenesis does not rescue migration out of r4.

Cell transplantation reveals that Rest function is required specifically in FBMNs for their complete migration. (A) Schematic of transplantation approach. Blastula stage donor-derived MZrest^sbu29/sbu29 mutant cells were transplanted into the hindbrain primordia of gastrula stage wild type hosts. To enable targeting of the hindbrain primordium, each host embryo was positioned with the dorsal shield at a 45° angle as indicated. (B–E) Average intensity projection confocal images (dorsal views, anterior to the left) of 48 hpf transplanted and immunolabeled embryos. Donor-derived FBMNs and EphA4a staining (r3 and r5) in red, host FBMNs in green, and all transplanted cells in blue (Alexa Fluor 647-labeled). (B′–E′) Red channel only, showing position of donor-derived FBMNs. Arrows indicate FBMNs that did not migrate to r6; arrowheads indicate long projections extended by MZrest^sbu29/sbu29 donor-derived cells. (F) Percent of total donor FBMNs from wild type or MZrest^sbu29/sbu29 embryos located in r4, r5, or r6 after transplantation into wild type hosts.