s105 mutant embryos have a strong reduction in retinal cell number. (A) Wild-type (WT) and s105 mutant embryo at 3 dpf. The s105 mutant is characterized by small eyes. (B) Transverse vibratome sections through embryonic retinae counterstained with phalloidin to visualize plexiform layers (red) and propidium iodide to mark nuclei (green). In s105 mutants, the retina is smaller but some stratification of the retinal layers is present. (C) Transverse vibratome sections through embryonic retinae counterstained with DAPI to visualize nuclei (red), with the zn5 antibody to detect ganglion cells and the optic nerve (green), or with the zpr1 antibody to detect red/green double cones (green). GCL, ganglion cell layer; INL, inner nuclear layer; PRL, photoreceptor cell layer.

The zebrafish cap-g gene is mutated in s105. (A) Representation of the genetic map of the cap-g locus on linkage group 1 (LG1) and exon/intron structure of the cap-g transcript. Some of the markers utilized for cloning the mutation are indicated and the number of recombinants among 2982 meioses is indicated below each marker. (B) Schematic model of the associated condensin I complex. (C) Comparison of sequence data for wild-type and s105 mutant alleles. The s105 mutation generates a premature stop codon. (D) Schematic diagram of the Cap-G protein which contains several predicted HEAT domains. The s105 mutation generates a premature stop codon (red asterisk) that truncates more than half of the protein. (E) Comparison of cap-g expression with that of pcna by whole-mount in situ hybridization. Overlapping expression with pcna within the brain, the CMZ of the retina, which contains retinal stem cells, and within neuromasts of the lateral line organ (white asterisks) indicates that cap-g is required within proliferative cells.

Loss of CAP-G causes p53-mediated apoptosis within the retina. (A) Retinal expression of the proliferation marker pcna or of neurogenesis marker elval3 is not affected in cap-g^s105 mutants as detected by whole-mount in situ hybridizations at 3 dpf. (B) Transverse cryosections of embryonic retinae were stained against phosphorylated histone 3 which marks mitotic nuclei (red), TUNEL to detect apoptotic cells (green), and nuclei counterstained with DAPI (blue) at 24 hpf. Predominantly mitotic cells which divide at the ventricular side of the retina are apoptotic in cap-g^s105 mutants at 24 hpf. (C) Transverse vibratome sections of 4 dpf retinae stained against activated caspase 3 to detect apoptotic cells (red) and nuclei counterstained with propidium iodide (green). At this stage, proliferation is restricted to the CMZ. In cap-g^s105 mutants, cell death is restricted to the CMZ which indicates that proliferative cells are eliminated. (D) Transverse vibratome sections of embryonic retinae counterstained with phalloidin to visualize plexiform layers (red) and propidium iodide (green). cap-g^s105 mutants injected with MO^p53show a rescue of retinal development and display correct retinal layering. (E) Quantification of cell numbers within different retinal cell layers. Propidium iodide stained transverse retinal sections were used to determine average counts for wild-type (n = 9 section planes, 5 embryos), cap-g^s105 mutants (n = 11 section planes, 7 embryos) or cap-g^s105 mutant/p53 morphants (n = 9 section planes, 6 embryos). The average sum of cap-g^s105 mutant retinal cells is reduced by 65% compared with wild-type. In comparison, the average sum of cap-g^s105 mutant/p53 morphant retinal cells is reduced only by 43% compared with wild-type. Therefore, the severe reduction in retinal cell numbers is in part caused by p53-mediated apoptosis. Data represent average cell numbers per retina ± SD. T-test p-values for cell number differences in comparison to wild-type: ***, p < 0.001; and in comparison to cap-g^s105: #, p < 0.05; ### p < 0.001.

Defective sister chromatid separation and aberrant nuclear sizes and shapes upon loss of Cap-G. (A) Selected images from a timelapse recording of mitotic divisions at gastrula stages analyzed in Tg [H2A::GFP] transgenic embryos. cap-g morphant embryos display defective sister chromatid separation and abnormal chromatid morphology during anaphase. (B) Selected images from a timelapse analysis of mitotic divisions within the ventricular zone of the neural tube at 32 hpf. Progression from prometaphase, when condensed chromosomes are visible, to anaphase is significantly delayed in cap-g^s105 mutants. (C) Transverse vibratome sections through retinae counter-stained with propidium iodide reveal aberrant nuclear sizes and morphologies within cap-g^s105 mutants at 3 dpf. The strict retinal layering into inner INL and PRL is not recognizable. Red asterisk indicates decondensed nucleus with chromatid bridge/non-disjunction event. Similarly, the non-neural tail region of cap-g^s105 mutants contains cells with incomplete separation of chromatids (red asterisk). (D) In the wild-type retina, most nuclei in the photoreceptor layer have an elongated appearance. In cap-g^s105 mutant or cap-g^s105 mutant/p53 morphant retinae, most PRL nuclei fail to elongate. Data represent mean ± SD, n ≥ 115 for each retinal layer and genotype; **, p < 0.01; ***, p < 0.005. (E) Representative histograms from FACS analysis of propidium idodide stained nuclei suspensions. Whereas at 72 hpf most cells in the wild-type are diploid (2C; 1C = haploid genome equivalent), cap-g^s105 mutants harbor an increased fraction of cells with a genomic content >2C, especially in the tetraploid range (4C).

CAP-H and CAP-D2 are essential components of the condensin I complex. Morphant embryos show smaller heads and eyes. Transverse vibratome sections of embryonic retinae counterstained with phalloidin to visualize plexiform layers (red) and propidium iodide to mark nuclei (green). cap-h or cap-d2 morphant retinae are smaller in size. Similar to cap-g^s105 mutant retinae, nuclei are aberrantly shaped and sized which is indicative of polyploidy/genomic imbalances. INL, inner nuclear layer; PRL, photoreceptor cell layer.

Chromatid association of CAP-G during mitosis. Selected images from a timelapse recording of a CAP-G-mcherry fusion protein in a gastrula stage transgenic embryo. The fusion protein associates with chromatids after the breakdown of the nuclear envelope at the beginning of prometaphase (t = 4.0 min) where it remains until decondensation of chromosomes during telophase (t = 17.0 min). Asterisks indicate positions of segregating chromatids. The different mitotic stages are recognizable by transgenic H2A::GFP expression. M, metaphase; PM, prometaphase; T, telophase.

CAP-G-mcherry is mislocalized during gastrula stage mitoses upon loss of its interaction partners CAP-H and CAP-D2 and localizes to the cytoplasm. (A, B) The different mitotic stages are recognizable by transgenic H2A::GFP expression. M, metaphase; PM, prometaphase; T, telophase. (C, D) Occurrence of chromatid bridges in MO^cap-h+cap-d2 injected embryos. Red arrow indicates a chromatid bridge in between two anaphase nuclei. Red arrowhead marks aberrant genetic material associated with a decondensed nucleus.

Genes encoding condensin I complex proteins are expressed within highly proliferative tissues. Comparison of cap-g, cap-h and cap-d2 expression with that of pcna by whole-mount in situ hybridization. All genes display overlapping expression patterns throughout early development. Expression at the 512-cell stage indicates a strong maternal contribution. At 24 hpf, condensin I genes are most strongly expressed within brain, retina and spinal cord. Within the retina, expression of condensin I genes is within the CMZ which contains the retinal stem cells whereas expression is absent within postmitotic differentiated retinal cells.

Unillustrated author statements