Microarray analysis of the altered transcriptome due to suppression of endogenous pou5f3 in zebrafish embryos. (A–D) Venn diagrams showing the numbers of genes/probes (total number, 15,617) whose expression was downregulated (A, B) or upregulated (C, D) by en‐pou5f3 induction at 90% epiboly (left) or 3‐ss (right), respectively. (A, C) All genes whose expression was judged to have been significantly affected by the Wilcoxon signed‐rank test (upregulated, p < 0.0025; downregulated, p > 0.9975). (B, D) Genes whose expression was evidently affected by en‐pou5f3 induction by twice or more (signal log ratio, > 1 or < −1). E. Alteration of bHLH‐O gene expression due to en‐pou5f3 induction. Signal log ratios showing changes in expression levels, obtained by the microarray, are shown with a heat map for the two induction stages. The genes are arranged based on the signal log ratios at 90% epiboly. Green indicates downregulation, and magenta indicates upregulation. Blue letters indicate her‐related genes that function in a Notch‐independent manner, whereas the genes shown in red are Notch‐dependent.

Quantitative analysis of the expression of genes that were shown to markedly change in the expression level by the microarray analyses. Expression was reevaluated by qRT‐PCR to confirm the validity of the microarray data for the genes that showed marked alteration in the expression level (Group I–IV) and/or those involved in brain formation (Group II) (Table 1). Regarding several Group II genes and sebox that had shown strikingly altered expression in the microarray, we failed to reliably quantitate the expression for unknown reasons and not shown here. The experiments were repeated three times, resulting in essentially the same results. The ordinates represent the expression levels in the en‐pou5f3‐induced embryos (light gray bars) relative to those in wild‐type sibling embryos (dark gray bars) for respective genes. Error bars, standard errors of means. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Confirmation of the altered gene expression by WISH in en‐pou5f3‐induced embryos. The expression patterns of the genes that had been shown to markedly differ in expression levels by the microarray analysis (Table 1) were examined by WISH (A, Group I; B, Group II; C, Group III; D, Group IV). Embryos from crosses between en‐pou5f3^+/− fish and wild‐type fish were subjected to en‐pou5f3 induction at 90% epiboly (A, B) or 3‐ss (C, D), and after further development at 25°C for 1 h, the expression patterns of genes of interest were stained. After observations, respective embryos were genotyped. Dorsal views with anterior to the top. For each gene, the expression in wild‐type sibling (sib) and en‐pou5f3^+/− (en‐pou) embryos are shown on the top and bottom, respectively. The numbers of embryos showing the indicated expression patterns and total scored embryos are shown in the bottom right. di, diencephalon; dmb, diencephalon‐midbrain boundary; fb, forebrain; mb, midbrain; nc, neural crest; np, neural plate; pnc, proneural cluster. Scale bar, 200 μm.

Regulation of downstream genes by activated En‐Pou5f3 in the absence of protein synthesis. A. Structure of En‐Pou5f3‐ERT2 overexpressed in embryos by mRNA injection. B. Time schedule of En‐Pou5f3‐ERT2 activation by 4‐hydroxytamoxifen (4‐OHT) and inhibition of protein synthesis by cycloheximide (CHX). Embryos injected with en‐pou5f3‐ERT2 mRNA (150 pg/embryo) were treated for 2 h from 80% epiboly (8 hpf) with 4‐OHT, fixed at the bud stage, and examined for gene expression by WISH. To inhibit protein synthesis during activation, CHX was added 30 min before (7.5 hpf) addition of 4‐OHT (age matching). However, as development almost stopped as soon as CHX was added, CHX addition was delayed (100% epiboly) in some experiments (stage matching) so that the stage (bud stage) of staining was also at the bud stage. C–E. Expression of pax6b (C), her3 (D), and her5 (E) at the bud stage in embryos where En‐Pou5f3‐ERT2 was not activated (a, c) or activated (b, d) by 4‐OHT in the absence (a, b) or presence (c, d) of CHX. Dorsal views, with anterior to the top. Downregulation is marked with blue arrows. Numbers of embryos showing downregulation and total scored embryos are shown in the bottom right. fb, forebrain; mhb, midbrain‐hindbrain boundary; npp, neural progenitor pool; fb, forebrain. Scale bar, 200 μm.

Noncoding conserved sequences found in the her3 upstream DNA. A. Comparison of the expression patterns in the neural plate between pou5f3 and her3. The expression of her3, stained red, was compared by two‐color WISH with that of pou5f3, stained in blue, at the bud stage. Dorsal views of flat‐mount preparations with anterior to the left. Broken lines mark the midbrain‐hindbrain boundary (mhb). Co‐expression of pou5f3 and her3 is marked with arrowheads. r1/2–4, rhombomeres 1/2–4. Scale bar, 100 μm. B. Comparison of the nucleotide sequences of the upstream 5.0‐kb DNA plus the coding regions between zebrafish her3 and spotted gar her3 using rVISTA. Prediction of the binding sequences for transcription factors (Oct, Sox, Nanog) were performed in parallel. Noncoding conserved regions (NCR‐1, NCR‐2) are marked with blue frames. Red vertical bars shown above indicate the zebrafish sequences to which transcription factors were predicted to bind, whereas green bars indicate the binding sites shared by the two species. C. Consensus transcription factor binding sites identified in the conserved sequences. Binding sites in the NCRs were searched by Match‐1.0 Public (vertebrates, cut‐off to minimize the sum of both error rates), and the sites for Oct (red) and Sox (green) are shown. ‘+’ and ‘–’ represent orientations of the binding sites. D. Structures of the constructs used in reporter assays. The genomic organization of her3 is shown at the top. The two NCRs are shown with blue ovals. Below are shown the structures of the EGFP and luciferase reporter constructs, in which the upstream 4.0‐kb DNA of her3 was ligated to the reporter genes. The EGFP reporter construct (Her3[−4.0]‐EGFP) was used for live imaging in embryos and cultured cells. The luciferase reporter Her3[−4.0]‐Luc and its deletion constructs, Her3[−4.0]dNCR1‐Luc, Her3[−4.0]dNCR2‐Luc, and Her3[−4.0]dNCRs‐Luc, were used in in vitro reporter assays.

Expression of Her3[−4.0]‐EGFP in developing embryos. A. EGFP fluorescence in 24‐hpf embryos injected with pHer3[−4.0]‐EGFP DNA. Left column shows bright‐field images, the middle row shows fluorescence images, and the right column shows merged views. Dorsal views with anterior to the left. Expression rates and embryo numbers examined are shown at the bottom right. Scale bar, 200 μm. B. Endogenous expression of her3 in the brain was examined by WISH at 24 hpf. Dorsal views of the anterior brain (a) and hindbrain (b) are shown with anterior to the left. C. Histogram showing fluorescence patterns in embryos injected with the reporter gene. The ordinate shows the percentages of embryos showing the expression patterns shown along the abscissa (n = 30). +, expression in the anterior tegmentum (atg), midbrain‐hindbrain boundary (mhb), or hindbrain (hb); ++, expression in atg and mhb or mhb and hb; +++, expression in atg, mhb, and hb. ±, nonspecific expression. mb, midbrain; ey, eye (optic vesicle). Two examples of +++ embryos and one example of ++ embryos are shown, whereas no embryos were scored as +. Weak ectopic expression observed when fluorescence was intense is marked with white arrowheads.

Transcriptional regulation of her3 in P19 cells undergoing neural differentiation. A. Time schedule for inducing neuronal differentiation by exposure to retinoic acid (RA). Horizontal arrows indicate the durations after RA addition at 0 h. B, C. Expression of the backbone luciferase plasmid (pGL4, left) and pHer3[−4.0]‐Luc (right) after exposure of P19 cells to RA for the duration shown on the abscissa. The expression levels of firefly luciferase were standardized by Renilla luciferase expression as an internal control. Blue and orange bars indicate the reporter expression in the absence or presence of RA, respectively, relative to the expression of pGL4 at 18 h in the absence of RA. The experiments were repeated three times, resulting in essentially the same results. Error bars, standard deviations of means. *, p < 0.05.

Transcriptional regulation of her3 by Pou5f3 and SoxB1 in cultured cells. A, B. Effects of Pou5f3 and Sox3 (A) or Pou5f3 and Sox2 (B) on the expression of pHer3[−4. 0]‐Luc in HEK293T (left) or P19 (right) cells. The expression levels of the reporter in the presence of the effectors indicated on the abscissa are shown relative to the control level when egfp was used as an effector. C. Effects of Pou5f3 and Sox3 on the expression of luciferase constructs lacking NCR1, NCR2, or both. The expression levels of the reporter genes in HEK293T cells in the presence of the expression plasmids for egfp (blue), pou5f3 (magenta), sox3 (gray), and pou5f3 plus sox3 (green) are shown relative to the luciferase expression for pGL4 in the presence of egfp. The expression levels of firefly luciferase were standardized by Renilla luciferase expression as an internal control. The experiments were repeated three times, resulting in essentially the same results. Error bars, standard deviations of means. *, p < 0.05; **, p < 0.01; ***, p < 0.001.