Zebrafish rest regulates developmental gene expression but not neurogenesis

Abstract

The transcriptional repressor Rest (Nrsf) recruits chromatin-modifying complexes to RE1 'silencer elements', which are associated with hundreds of neural genes. However, the requirement for Rest-mediated transcriptional regulation of embryonic development and cell fate is poorly understood. Conflicting views of the role of Rest in controlling cell fate have emerged from recent studies. To address these controversies, we examined the developmental requirement for Rest in zebrafish using zinc-finger nuclease-mediated gene targeting. We discovered that germ layer specification progresses normally in rest mutants despite derepression of target genes during embryogenesis. This analysis provides the first evidence that maternal rest is essential for repression of target genes during blastula stages. Surprisingly, neurogenesis proceeds largely normally in rest mutants, although abnormalities are observed within the nervous system, including defects in oligodendrocyte precursor cell development and a partial loss of facial branchiomotor neuron migration. Mutants progress normally through embryogenesis but many die as larvae (after 12 days). However, some homozygotes reach adulthood and are viable. We utilized an RE1/NRSE transgenic reporter system to dynamically monitor Rest activity. This analysis revealed that Rest is required to repress gene expression in mesodermal derivatives including muscle and notochord, as well as within the nervous system. Finally, we demonstrated that Rest is required for long-term repression of target genes in non-neural tissues in adult zebrafish. Our results point to a broad role for Rest in fine-tuning neural gene expression, rather than as a widespread regulator of neurogenesis or cell fate.

Fig. 1.

Zinc-finger nuclease-mediated targeting of zebrafish rest. (A) Sequence chromatograms from wild-type siblings (left) and rest mutants (right). The 7 bp deletion (Δ7) in SBU29 and 4 bp insertion (+4) in SBU34 are illustrated beneath. (B) Domain structure of wild-type Rest and truncated proteins produced by SBU29 and SBU34 mutations. (C) Animal pole views of 6-hpf wild-type and MZrest^sbu29/sbu29 embryos. Rest immunoreactivity, TOPRO-3 staining of nuclei, and merged channels are shown. Abundant cytoplasmic Rest protein is detected in wild-type embryos, but not in MZrest^sbu29/sbu29 mutants. Confocal images are single 1-μm stacks taken at 40× magnification.

Fig. 2.

Expression levels of RE1-containing genes during early development of rest mutants. qPCR analysis showing fold differences relative to MZrest^sbu29/sbu29 transcript levels (defined as 1). (A-C) gpr27, kcnh8 and snap25b are either not expressed or expressed at low levels in wild-type zebrafish embryos, but are expressed in MZrest^sbu29/sbu29 mutants. (D-F) bdnf, snap25a and grin1a show stage-specific alterations in expression in rest mutants. (G-L) Expression of pcad, cacng2, neurod, gfap, spop and bsx is not significantly altered in the rest mutant. Error bars indicate s.e. ND*, not detected.

Fig. 3.

Maternal rest suppresses target genes in mid-blastula embryos. qPCR analysis of (A) snap25b, (B) bdnf, (C) pcad and (D) gpr27 expression in mid-blastula zebrafish embryos. Comparison of transcript levels in wild-type, Zrest^sbu29/+ and Mrest^sbu29/+ embryos reveals derepression of target genes in Mrest but not Zrest embryos. Fold comparisons are relative to Mrest transcript levels (defined as 1). Maternal (M) and zygotic (Z) embryos were obtained by crossing female and male rest^sbu29/sbu29 mutants, respectively, to wild types. Error bars indicate s.e. ND*, not determined.

Fig. 4.

Histone deacetylase contribution to Rest-mediated inhibition of target genes is locus specific. qPCR analysis of RE1-containing gene expression at mid-blastula stage in wild-type and MZrest^sbu29/sbu29 zebrafish embryos treated with DMSO or the HDAC inhibitor TSA. The fold differences are relative to transcript levels in MZrest^sbu29/sbu29 DMSO (defined as 1). (A,B) TSA treatment enhances expression of snap25a and gpr27 in both MZrest and wild-type embryos compared with DMSO controls. (C) TSA treatment of wild-type embryos produces a much smaller increase in snap25b expression than does removing Rest function. This suggests that Rest-dependent histone methylation is central to repression of snap25b expression. (D-F) HDAC inhibition has little effect on pcad, spop and grin1a expression. Error bars indicate s.e.

Fig. 5.

Neurogenesis progresses normally in rest mutants. Whole-mount in situ hybridization for the pan-neural marker huC in stage-matched (A,B) 2-somite, (C,D) 8-somite, (E,F) 14-somite and (G,H) 24-hpf wild type or rest heterozygotes and MZrest^sbu29/sbu29 mutants. All views are dorsal, anterior to the left. MZrest mutants did not show ectopic or precocious huC expression. To control for subtle differences in developmental timing and genetic background, comparisons were undertaken in crosses of rest^sbu29/sbu29 females to rest^sbu29/+ males and between wild-type intercrosses and MZrest^sbu29/sbu29 intercrosses. Genotypes were determined by PCR.

Fig. 6.

Rest function is crucial for the regulation of oligodendrocyte precursor cells in the dorsal spinal cord and for facial branchiomotor migration. (A-D) Lateral views of live 50-hpf and 98-hpf Tg(olig2:GFP) (A,C) and Tg(olig2:GFP); MZrest^sbu29/sbu29 mutants (B,D). The number of migrating olig2⁺ oligodendrocyte precursors (arrows) is significantly reduced in rest mutants at 50 hpf (B) compared with controls (A). The number of migrating OPCs (arrows) in 98-hpf MZrest^sbu29/sbu29 mutants (D) is similar to controls (C). (E-G) Dorsal views of the hindbrain of 48-hpf Tg(islet1:GFP) transgenic zebrafish embryos. Representative embryos from a rest^sbu29/+, rest^sbu29/+; islet1:GPF intercross. (E) rest^+/+ embryo showing normal FBMN migration into r6-r7. (F) FBMNs in rest^sbu29/+ embryos have only subtle migration defects (arrows). (G) In rest^sbu29/sbu29 embryos, a significant proportion of FBMNs remain in r4-r5 (bracket).

Fig. 7.

RE1/NRSE reporter transgenic lines are ectopically expressed in rest mutants. (A-H′) Lateral (A-F) or dorsal (G,H) views of live RE1/NRSE transgenic reporter lines in wild-type (A,C,E,G) or rest^sbu29/sbu29 (B,D,F,H) backgrounds. Expression of each reporter was altered in the mutants. The boxed region in G and H is shown at higher magnification in G′ and H′. Arrows mark midline cells, which are not as numerous in the wild-type background. (I-L) The proportion of embryos in each phenotypic and genotypic class for each reporter. gmc606 and gmc641 larvae are 3 dpf; gmc607, 6 dpf; gmc632, 4 dpf.

Fig. 8.

Rest target genes are depressed in germ tissue of adult rest mutants. qPCR analysis of RE1-containing gene expression in wild-type and rest^sbu29/sbu29 mutant ovary and testes (three examples of each are shown). All fold differences are relative to the average of the transcript levels in the three mutants (defined as 1). (A-D) Levels of snap25b, kcnh8, gpr27 and pcad are enhanced in the rest mutant ovary. (E) Only kcnh8 expression was enhanced in rest mutant testes.