Fat2 is an atypical cadherin with a large extracellular domain containing multiple cadherin repeats. Although members of the Fat family have been implicated in organ development, the role of Fat2 in neural development remains poorly understood. Here, we investigated the function of Fat2 in zebrafish, focusing on its involvement in the development of cerebellar granule cells. Granule cells in the rostromedial cerebellum project their axons, known as parallel fibers, to Purkinje cell dendrites. In contrast, granule cells in the caudolateral cerebellum project their axons to Purkinje cells and further extend them caudally to innervate crest cells, Purkinje-like neurons located in the dorsal hindbrain region called the medio-octaval lateralis. We found that in fat2 mutants, axons of caudolateral cerebellar granule cells projecting to crest cells exhibit abnormal branching and frequently overshoot their targets. These results indicate that Fat2 is required for proper axon formation in at least a subset of cerebellar granule cells.

During embryonic development, primitive and definitive waves of hematopoiesis take place to provide proper blood cells for each developmental stage, with the possible involvement of epigenetic factors. We previously found that lysine-specific demethylase 1 (LSD1/KDM1A) promotes primitive hematopoietic differentiation by shutting down the gene expression program of hemangioblasts in an Etv2/Etsrp-dependent manner. In the present study, we demonstrated that zebrafish LSD1 also plays important roles in definitive hematopoiesis in the development of hematopoietic stem and progenitor cells. A combination of genetic approaches and imaging analyses allowed us to show that LSD1 promotes the egress of hematopoietic stem and progenitor cells into the bloodstream during the endothelial-to-hematopoietic transition. Analysis of compound mutant lines with Etv2/Etsrp mutant zebrafish revealed that, unlike in primitive hematopoiesis, this function of LSD1 was independent of Etv2/Etsrp. The phenotype of LSD1 mutant zebrafish during the endothelial-to-hematopoietic transition was similar to that of previously reported compound knockout mice of Gfi1/Gfi1b, which forms a complex with LSD1 and represses endothelial genes. Moreover, co-knockdown of zebrafish Gfi1/Gfi1b genes inhibited the development of hematopoietic stem and progenitor cells. We therefore hypothesize that the shutdown of the Gfi1/Gfi1b-target genes during the endothelial-to-hematopoietic transition is one of the key evolutionarily conserved functions of LSD1 in definitive hematopoiesis.

LSD1/KDM1A is a widely conserved lysine-specific demethylase that removes methyl groups from methylated proteins, mainly histone H3. We previously isolated the zebrafish LSD1 gene and demonstrated that it is required for primitive hematopoiesis. Recently, a neuron-specific splicing variant of LSD1 was found in mammals and its specific functions and substrate specificities were reported. To our surprise, zebrafish LSD1 cDNA, which we previously analyzed, was corresponded to the neuron-specific variant in mammals. In this study, we investigated the structures and expression of LSD1 splicing variants in zebrafish and found all 4 types of LSD1 isoforms: LSD1, LSD1+2al, LSD1+8al and LSD1+2al8al. Interestingly, LSD1+8al/LSD1+2al8al, which correspond to mammalian neuron-specific variants, expressed ubiquitously in zebrafish. We also performed phenotypic rescue experiments of a zebrafish LSD1 mutant (kdm1a(it627)) using human and zebrafish LSD1 variants to identify which variant is involved in primitive hematopoiesis. Unexpectedly, the overexpression of all types of human and zebrafish variants was able to rescue the hematopoietic phenotypes in LSD1 mutants. Furthermore, enzymatic-deficient LSD1K661A (human) and K638A (zebrafish) were also able to rescue the mutant phenotypes. These results suggest that the LSD1 functions in zebrafish primitive hematopoiesis are free from any splicing-dependent regulation or demethylation reaction.

Neurons in the inferior olivary nuclei (IO neurons) send climbing fibers to Purkinje cells to elicit functions of the cerebellum. IO neurons and Purkinje cells are derived from neural progenitors expressing the proneural gene ptf1a. In this study, we found that the homeobox gene gsx2 was co-expressed with ptf1a in IO progenitors in zebrafish. Both gsx2 and ptf1a zebrafish mutants showed a strong reduction or loss of IO neurons. The expression of ptf1a was not affected in gsx2 mutants and vice versa. In IO progenitors, the ptf1a mutation increased apoptosis whereas the gsx2 mutation did not, suggesting that ptf1a and gsx2 are regulated independently of each other and have distinct roles. The fibroblast growth factors (Fgf) 3/8a and retinoic acid signals negatively and positively, respectively, regulated gsx2 expression and thereby the development of IO neurons. mafba and hox genes are at least partly involved in the Fgf- and retinoic acid-dependent regulation of IO neuronal development. Our results indicate that gsx2 mediates the rostro-caudal positional signals to specify the identity of IO neurons from ptf1a-expressing neural progenitors.

The cerebellum and the cerebellum-like structure in the mesencephalic tectum in zebrafish contain multiple cell types, including principal cells (i.e., Purkinje cells and type I neurons) and granule cells, that form neural circuits in which the principal cells receive and integrate inputs from granule cells and other neurons. It is largely unknown how these cells are positioned and how neural circuits form. While Reelin signaling is known to play an important role in cell positioning in the mammalian brain, its role in the formation of other vertebrate brains remains elusive. Here we found that zebrafish with mutations in Reelin or in the Reelin-signaling molecules Vldlr or Dab1a exhibited ectopic Purkinje cells, eurydendroid cells (projection neurons), and Bergmann glial cells in the cerebellum, and ectopic type I neurons in the tectum. The ectopic Purkinje cells and type I neurons received aberrant afferent fibers in these mutants. In wild-type zebrafish, reelin transcripts were detected in the internal granule cell layer, while Reelin protein was localized to the superficial layer of the cerebellum and the tectum. Laser ablation of the granule cell axons perturbed the localization of Reelin, and the mutation of both kif5aa and kif5ba, which encode major kinesin I components in the granule cells, disrupted the elongation of granule cell axons and the Reelin distribution. Our findings suggest that in zebrafish, (1) Reelin is transported from the granule cell soma to the superficial layer by axonal transport; (2) Reelin controls the migration of neurons and glial cells from the ventricular zone; and (3) Purkinje cells and type I neurons attract afferent axons during the formation of the cerebellum and the cerebellum-like structure.

BACKGROUND:Conventionally, comparison among amniotes - birds, mammals, and reptiles - has often been approached through analyses of mammals and, for comparison, birds. However, birds are morphologically and physiologically derived and, moreover, some parts of their genomes are recognized as difficult to sequence and/or assemble and are thus missing in genome assemblies. Therefore, sequencing the genomes of reptiles would aid comparative studies on amniotes by providing more comprehensive coverage to help understand the molecular mechanisms underpinning evolutionary changes.RESULTS:Herein, we present the whole genome sequences of the Madagascar ground gecko (Paroedura picta), a promising study system especially in developmental biology, and used it to identify changes in gene repertoire across amniotes. The genome-wide analysis of the Madagascar ground gecko allowed us to reconstruct a comprehensive set of gene phylogenies comprising 13,043 ortholog groups from diverse amniotes. Our study revealed 469 genes retained by some reptiles but absent from available genome-wide sequence data of both mammals and birds. Importantly, these genes, herein collectively designated as 'elusive' genes, exhibited high nucleotide substitution rates and uneven intra-genomic distribution. Furthermore, the genomic regions flanking these elusive genes exhibited distinct characteristics that tended to be associated with increased gene density, repeat element density, and GC content.CONCLUSION:This highly continuous and nearly complete genome assembly of the Madagascar ground gecko will facilitate the use of this species as an experimental animal in diverse fields of biology. Gene repertoire comparisons across amniotes further demonstrated that the fate of a duplicated gene can be affected by the intrinsic properties of its genomic location, which can persist for hundreds of millions of years.

The cerebellum is derived from the dorsal part of the anterior-most hindbrain. The vertebrate cerebellum contains glutamatergic granule cells (GCs) and gamma-aminobutyric acid (GABA)ergic Purkinje cells (PCs). These cerebellar neurons are generated from neuronal progenitors or neural stem cells by mechanisms that are conserved among vertebrates. However, vertebrate cerebella are widely diverse with respect to their gross morphology and neural circuits. The cerebellum of cyclostomes, the basal vertebrates, has a negligible structure. Cartilaginous fishes have a cerebellum containing GCs, PCs, and deep cerebellar nuclei (DCNs), which include projection neurons. Ray-finned fish lack DCNs but have projection neurons termed eurydendroid cells (ECs) in the vicinity of the PCs. Among ray-finned fishes, the cerebellum of teleost zebrafish has a simple lobular structure, whereas that of weakly electric mormyrid fish is large and foliated. Amniotes, which include mammals, independently evolved a large, foliated cerebellum, which contains massive numbers of GCs and has functional connections with the dorsal telencephalon (neocortex). Recent studies of cyclostomes and cartilaginous fish suggest that the genetic program for cerebellum development was already encoded in the genome of ancestral vertebrates. In this review, we discuss how alterations of the genetic and cellular programs generated diversity of the cerebellum during evolution.

A spontaneous medaka ro mutant shows abnormal wobbling and rolling swimming behaviors. By positional cloning, we mapped the ro locus to a region containing the gene encoding Contactin1b (Cntn1b), which is an immunoglobulin (Ig)-superfamily domain-containing membrane-anchored protein. The ro mutant had a deletion in the cntn1b gene that introduced a premature stop codon. Furthermore, cntn1b mutants generated by the CRISPR/Cas9 system and trans-heterozygotes of the CRISPR mutant allele and ro had abnormal swimming behavior, indicating that the cntn1b gene was responsible for the ro-mutant phenotype. We also established zebrafish cntn1a and cntn1b mutants by transcription activator-like effector nucleases (TALENs). Zebrafish cntn1b but not cntn1a mutants showed abnormal swimming behaviors similar to those in the ro mutant, suggesting that Cntn1b plays a conserved role in the formation or function of the neural circuits that control swimming in teleosts. Although Cntn1-deficient mice have abnormal cerebellar neural circuitry, there was no apparent histological abnormality in the cerebellum of medaka or zebrafish cntn1b mutants. The medaka cntn1b mutants had defective optokinetic response (OKR) adaptation and abnormal rheotaxis (body positioning relative to water flow). Medaka and zebrafish cntn1b mutants are effective models for studying the neural circuits involved in motor learning and motor coordination.

The structure of the neural circuitry of the cerebellum, which functions in some types of motor learning and coordination, is generally conserved among vertebrates. However, some cerebellar features are species specific. It is not clear which genes are involved in forming these conserved and species‐specific structures and functions. This study uses zebrafish transgenic larvae expressing fluorescent proteins in granule cells, Purkinje cells, or other cerebellar neurons and glial cells to isolate each type of cerebellar cells by fluorescence‐activated cell sorting and to profile their gene expressions by RNA sequencing and in situ hybridization. We identify genes that are upregulated in granule cells or Purkinje cells, including many genes that are also expressed in mammalian cerebella. Comparison of the transcriptomes in granule cells and Purkinje cells in zebrafish larvae reveals that more developmental genes are expressed in granule cells, whereas more neuronal‐function genes are expressed in Purkinje cells. We show that some genes that are upregulated in granule cells or Purkinje cells are also expressed in the cerebellum‐like structures. Our data provide a platform for understanding the development and function of the cerebellar neural circuits in zebrafish and the evolution of cerebellar circuits in vertebrates. J. Comp. Neurol. 525:1558–1585, 2017. © 2016 Wiley Periodicals, Inc.

Granule cells (GCs) are the major glutamatergic neurons in the cerebellum, and GC axon formation is an initial step in establishing functional cerebellar circuits. In the zebrafish cerebellum, GCs can be classified into rostromedial and caudolateral groups, according to the locations of their somata in the corresponding cerebellar lobes. The axons of the GCs in the caudolateral lobes terminate on crest cells in the dorsal hindbrain, as well as forming en passant synapses with Purkinje cells in the cerebellum. In the zebrafish mutant shiomaneki, the caudolateral GCs extend aberrant axons. Positional cloning revealed that the shiomaneki (sio) gene locus encodes Col4a6, a subunit of type IV collagen, which, in a complex with Col4a5, is a basement membrane (BM) component. Both col4a5 and col4a6 mutants displayed similar abnormalities in the axogenesis of GCs and retinal ganglion cells (RGCs). Although type IV collagen is reported to control axon targeting by regulating the concentration gradient of an axonal guidance molecule Slit, Slit overexpression did not affect the GC axons. The structure of the BM surrounding the tectum and dorsal hindbrain was disorganized in the col4a5 and col4a6 mutants. Moreover, the abnormal axogenesis of the caudolateral GCs and the RGCs was coupled with aberrant BM structures in the type IV collagen mutants. The regrowth of GC axons after experimental ablation revealed that the original and newly formed axons displayed similar branching and extension abnormalities in the col4a6 mutants. These results collectively suggest that type IV collagen controls GC axon formation by regulating the integrity of the BM, which provides axons with the correct path to their targets.

Significance Epigenetic regulation of cell fate determination is one of the hottest topics today. In this study, we isolated and characterized a zebrafish mutant in histone Lys-specific demethylase 1 (LSD1/KDM1A), and found that LSD1 plays a role in the initiation of hematopoietic differentiation in the hemangioblast, a bipotent cell that can give rise to hematopoietic or endothelial progenitors. In addition, we identified the major function of LSD1 in hemangioblasts to be the down-regulation of Ets variant 2 (Etv2), a critical regulator of hemangioblast development. Our results suggest that the LSD1-dependent shutdown of the Etv2 gene is the significant event required for hemangioblasts to initiate hematopoietic differentiation.

The hemangioblast is a progenitor cell with the capacity to give rise to both hematopoietic and endothelial progenitors. Currently, the regulatory mechanisms underlying hemangioblast formation are being elucidated, whereas those controllers for the selection of hematopoietic or endothelial fates still remain a mystery. To answer these questions, we screened for zebrafish mutants that have defects in the hemangioblast expression of Gata1, which is never expressed in endothelial progenitors. One of the isolated mutants, it627, showed not only down-regulation of hematopoietic genes but also up-regulation of endothelial genes. We identified the gene responsible for the it627 mutant as the zebrafish homolog of Lys-specific demethylase 1 (LSD1/KDM1A). Surprisingly, the hematopoietic defects in lsd1(it627) embryos were rescued by the gene knockdown of the Ets variant 2 gene (etv2), an essential regulator for vasculogenesis. Our results suggest that the LSD1-dependent shutdown of Etv2 gene expression may be a significant event required for hemangioblasts to initiate hematopoietic differentiation.

The development of vertebrate neurons requires a change in membrane phosphatidylcholine (PC) metabolism. Although PC hydrolysis is essential for enhanced axonal outgrowth mediated by phospholipase D (PLD), less is known about the determinants of PC metabolism on dendritic arborization. We show that protein arginine methyltransferase 8 (PRMT8) acts as a phospholipase that directly hydrolyzes PC, generating choline and phosphatidic acid. We found that PRMT8 knockout mice (prmt8(-/-)) displayed abnormal motor behaviors, including hindlimb clasping and hyperactivity. Moreover, prmt8(-/-) mice and TALEN-induced zebrafish prmt8 mutants and morphants showed abnormal phenotypes, including the development of dendritic trees in Purkinje cells and altered cerebellar structure. Choline and acetylcholine levels were significantly decreased, whereas PC levels were increased, in the cerebellum of prmt8(-/-) mice. Our findings suggest that PRMT8 acts both as an arginine methyltransferase and as a PC-hydrolyzing PLD that is essential for proper neurological functions.

The cerebellum is involved in some forms of motor coordination and motor learning. Here we isolated transgenic (Tg) zebrafish lines that express a modified version of Gal4-VP16 (GFF) in the cerebellar neural circuits: granule, Purkinje, or eurydendroid cells, Bergmann glia, or the neurons in the inferior olive nuclei (IO) which send climbing fibers to Purkinje cells, with the transposon Tol2 system. By combining GFF lines with Tg lines carrying a reporter gene located downstream of Gal4 binding sequences (upstream activating sequence: UAS), we investigated the anatomy and developmental processes of the cerebellar neural circuitry. Combining an IO-specific Gal4 line with a UAS reporter line expressing the photoconvertible fluorescent protein Kaede demonstrated the contralateral projections of climbing fibers. Combining a granule cell-specific Gal4 line with a UAS reporter line expressing wheat germ agglutinin (WGA) confirmed direct and/or indirect connections of granule cells with Purkinje cells, eurydendroid cells, and IO neurons in zebrafish. Time-lapse analysis of a granule cell-specific Gal4 line revealed initial random movements and ventral migration of granule cell nuclei. Transgenesis of a reporter gene with another transposon Tol1 system visualized neuronal structure at a single cell resolution. Our findings indicate the usefulness of these zebrafish Gal4 Tg lines for studying the development and function of cerebellar neural circuits.

The primitive blood circulation requires intravascular plasma flow [1]. However, it remains unclear whether the onset of earliest blood circulation is dependent solely on establishment of a functional circulatory organ or whether it also requires active processes inherent in blood cells. In this study, we present novel mechanisms for the onset of blood circulation by monitoring fluorescently labeled blood precursors and blood vessels in zebrafish. The earliest blood circulation occurs synchronously. This synchrony is achieved by the retention of erythroid precursors on the lumen of the vasculature after their invasion from the subaortic region, and then by simultaneous release of these precursors into the flow. Morphological and biochemical analyses suggest that the onset of blood circulation accompanies disruption of blood cell-to-vessel adhesion and requires metalloprotease-dependent processes. ADAM8, a member of the a disintegrin and metalloprotease (ADAM) family, mediates the onset of blood circulation. In ADAM8-depleted embryos, erythroid cells fail to detach from the vascular lumen and stagnate. Expression of a protease-defective ADAM8 in erythroid cells causes dominant-negative effects on blood circulation, suggesting cell-autonomous roles of ADAM8. Based on these findings, we propose that the first erythroid cells require both flow-dependent passive and proteolysis-dependent active processes to enter the circulation.

Phenotypic rescue experiments have been commonly used in zebrafish since it is convenient to study the causality of mutant phenotypes just by injecting mRNA into embryos. However, this strategy is only effective for phenotypes at early embryonic stages due to mRNA instability. For later developmental stages, DNA constructs are used to express exogenous genes, while it is usually ineffective owing to the problem of mosaicism. This study attempted to solve the problem by using Tol2-mediated transgenesis. As a model case, we used vlad tepes (vlt), a zebrafish gata1 mutant, whose phenotypes have never been able to be rescued at later stages by transient rescue experiments. Blood cell-specific transgenic expression of gata1 was driven by its own promoter/enhancer elements. The co-injection of a Tol2-donor plasmid containing gata1 cDNA and transposase mRNA efficiently rescued the bloodless phenotypes of vlt even in day 12 larvae when definitive erythropoiesis took place with primitive erythropoiesis. This Tol2-mediated rescue is therefore considered to be a quick and easy method for analyzing the mutant phenotypes in zebrafish.

Pi class GSTs (glutathione S-transferases) are a member of the vertebrate GST family of proteins that catalyse the conjugation of GSH to electrophilic compounds. The expression of Pi class GST genes can be induced by exposure to electrophiles. We demonstrated previously that the transcription factor Nrf 2 (NF-E2 p45-related factor 2) mediates this induction, not only in mammals, but also in fish. In the present study, we have isolated the genomic region of zebrafish containing the genes gstp1 and gstp2. The regulatory regions of zebrafish gstp1 and gstp2 have been examined by GFP (green fluorescent protein)-reporter gene analyses using microinjection into zebrafish embryos. Deletion and point-mutation analyses of the gstp1 promoter showed that an ARE (antioxidant-responsive element)-like sequence is located 50 bp upstream of the transcription initiation site which is essential for Nrf 2 transactivation. Using EMSA (electrophoretic mobility-shift assay) analysis we showed that zebrafish Nrf 2-MafK heterodimer specifically bound to this sequence. All the vertebrate Pi class GST genes harbour a similar ARE-like sequence in their promoter regions. We propose that this sequence is a conserved target site for Nrf 2 in the Pi class GST genes.