Overexpression of a slit homologue impairs convergent extension of the mesoderm and causes cyclopia in embryonic zebrafish.

Slit is expressed in the midline of the central nervous system both in vertebrates and invertebrates. In Drosophila, it is the midline repellent acting as a ligand for the Roundabout (Robo) protein, the repulsive receptor which is expressed on the growth cones of the commissural neurons. We have isolated cDNA fragments of the zebrafish slit2 and slit3 homologues and found that both genes start to be expressed by the midgastrula stage well before the axonogenesis begins in the nervous system, both in the axial mesoderm, and slit2 in the anterior margin of the neural plate and slit3 in the polster at the anterior end of the prechordal mesoderm. Later, expression of slit2 mRNA is detected mainly in midline structures such as the floor plate cells and the hypochord, and in the anterior margins of the neural plates in the zebrafish embryo, while slit3 expression is observed in the anterior margin of the prechordal plate, the floorplate cells in the hindbrain, and the motor neurons both in the hindbrain and the spinal cord. To study the role of Slit in early embryos, we overexpressed Slit2 in the whole embryos either by injection of its mRNA into one-cell stage embryos or by heat-shock treatment of the transgenic embryos which carries the slit2 gene under control of the heat-shock promoter. Overexpression of Slit2 in such ways impaired the convergent extension movement of the mesoderm and the rostral migration of the cells in the dorsal diencephalon and resulted in cyclopia. Our results shed light on a novel aspect of Slit function as a regulatory factor of mesodermal cell movement during gastrulation.

Laser-induced gene expression in specific cells of transgenic zebrafish.

Over the past few years, a number of studies have described the generation of transgenic lines of zebrafish in which expression of reporters was driven by a variety of promoters. These lines opened up the real possibility that transgenics could be used to complement the genetic analysis of zebrafish development. Transgenic lines in which the expression of genes can be regulated both in space and time would be especially useful. Therefore, we have cloned the zebrafish promoter for the inducible hsp70 gene and made stable transgenic lines of zebrafish that express the reporter green fluorescent protein gene under the control of a hsp70 promoter. At normal temperatures, green fluorescent protein is not detectable in transgenic embryos with the exception of the lens, but is robustly expressed throughout the embryo following an increase in ambient temperature. Furthermore, we have taken advantage of the accessibility and optical clarity of the embryos to express green fluorescent protein in individual cells by focussing a sublethal laser microbeam onto them. The targeted cells appear to develop normally: cells migrate normally, neurons project axons that follow normal pathways, and progenitor cells divide and give rise to normal progeny cells. By generating other transgenic lines in which the hsp70 promoter regulates genes of interest, it should be possible to examine the in vivo activity of the gene products by laser-inducing specific cells to express them in zebrafish embryos. As a first test, we laser-induced single muscle cells to make zebrafish Sema3A1, a semaphorin that is repulsive for specific growth cones, in a hsp70-sema3A1 transgenic line of zebrafish and found that extension by the motor axons was retarded by the induced muscle.

Growth cones utilize both widespread and local directional cues in the zebrafish brain.

The distribution of cues that provide directional information for specific growth cones in the zebrafish brain was functionally assayed by transplanting epiphysial neurons to ectopic locations in the embryonic brain followed by determining the pathways taken by the donor axons. Epiphysial axons normally first extend ventrally from their position in the dorsal diencephalon and then turn and extend anteriorly in the ventral diencephalon. When transplanted to ectopic sites at other axial levels of the brain, where in principle the axons could extend in any direction, epiphysial axons consistently extended ventrally. Furthermore, following initial ventral extension ectopic epiphysial axons turned randomly in the anterior and posterior directions. These results suggest that the cues for ventral extension are widely distributed along the rostrocaudal axis of the zebrafish brain, but the cues for subsequent anterior extension are restricted to the site where the epiphysial axons normally turn longitudinally.

Molecular cloning, expression, and activity of zebrafish semaphorin Z1a.

Semaphorins/collapsins are a large family of secreted and cell surface molecules that are thought to guide growth cones to their targets. Although some members are clearly repulsive to specific growth cones in vitro, the in vivo role of many of these molecules in vertebrate embryos is still unclear. As a first step towards clarifying the in vivo role of semaphorins/collapsins, we analyzed semaZ1a in the simple and well-characterized zebrafish embryo. SemaZ1a is a secreted molecule that is highly homologous to Sema III/D/collapsin-1, and it can collapse chick dorsal root ganglion growth cones in vitro. It is expressed in highly specific patterns within the developing embryo, which suggests that it influences outgrowth by a variety of growth cones including those of the posterior lateral line ganglion. Consistent with this hypothesis, the peripherally extending growth cones of posterior lateral line neurons retract and partially collapse during normal outgrowth.

The zebrafish detour gene is essential for cranial but not spinal motor neuron induction.

The zebrafish detour (dtr) mutation generates a novel neuronal phenotype. In dtr mutants, most cranial motor neurons, especially the branchiomotor, are missing. However, spinal motor neurons are generated normally. The loss of cranial motor neurons is not due to aberrant hindbrain patterning, failure of neurogenesis, increased cell death or absence of hh expression. Furthermore, activation of the Hh pathway, which normally induces branchiomotor neurons, fails to induce motor neurons in the dtr hindbrain. Despite this, not all Hh-mediated regulation of hindbrain development is abolished since the regulation of a neural gene by Hh is intact in the dtr hindbrain. Finally, dtr can function cell autonomously to induce branchiomotor neurons. These results suggest that detour encodes a component of the Hh signaling pathway that is essential for the induction of motor neurons in the hindbrain but not in the spinal cord and that dtr function is required for the induction of only a subset of Hh-mediated events in the hindbrain.

Molecular cloning and developmental expression of a zebrafish axonal glycoprotein similar to TAG-1.

TAG-1 is a mammalian cell adhesion molecule of the immunoglobulin superfamily that is expressed transiently by a subset of neurons and serves as a fertile substrate for neurite outgrowth in vitro (Furley, A.H., Morton, S.B., Manalo, D., Karagogeos, S., Dodd, H., Jessell, T.M., 1990 The axonal glycoprotein TAG-1 is an immunoglobulin superfamily member with neurite outgrowth promoting activity. Cell 61, 157-170). In order to examine the in vivo function of this molecule, we have cloned a zebrafish tag1-like cDNA and analyzed its expression patterns. tag1 Is expressed transiently by specific subsets of neurons when they are projecting their axons or when they are migrating. The specific and dynamic pattern of expression of zebrafish tag1 is consistent with its proposed role in axon guidance and cell migration.

Analysis of a Zebrafish semaphorin reveals potential functions in vivo.

The semaphorin/collapsin gene family is a large and diverse family encoding both secreted and transmembrane proteins, some of which are thought to act as repulsive axon guidance molecules. However, the function of most semaphorins is still unknown. We have cloned and characterized several semaphorins in the zebrafish in order to assess their in vivo function. Zebrafish semaZ2 is expressed in a dynamic and restricted pattern during the period of axon outgrowth that indicates potential roles in the guidance of several axon pathways. Analysis of mutant zebrafish with reduced semaZ2 expression reveals axon pathfinding errors that implicate SemaZ2 in normal guidance.

Molecular cloning and expression of two novel zebrafish semaphorins.

The large, conserved semaphorin/collapsin gene family encodes putative axon guidance molecules. We describe the cloning and expression of two n ovel zebrafish semaphorins that represent an increase in the size and diversity of the family. These semaphorins are expressed in unique and dynamic patterns during development.

Role of sonic hedgehog in branchiomotor neuron induction in zebrafish.

The role of zebrafish hedgehog genes in branchiomotor neuron development was analyzed by examining mutations that affect the expression of the hedgehog genes and by overexpressing these genes in embryos. In cyclops mutants, reduction in sonic hedgehog (shh) expression, and elimination of tiggy-winkle hedgehog (twhh) expression, correlated with reductions in branchiomotor neuron populations. Furthermore, branchiomotor neurons were restored in cyclops mutants when shh or twhh was overexpressed. These results suggest that Shh and/or Twhh play an important role in the induction of branchiomotor neurons in vivo. In sonic-you (syu) mutants, where Shh activity was reduced or eliminated due to mutations in shh, branchiomotor neurons were reduced in number in a rhombomere-specific fashion, but never eliminated. Similarly, spinal motor neurons were reduced, but not eliminated, in syu mutants. These results demonstrate that Shh is not solely responsible for inducing branchiomotor and spinal motor neurons, and suggest that Shh and Twhh may function as partially redundant signals for motor neuron induction in zebrafish.

Regulation of netrin-1a expression by hedgehog proteins.

Netrins, a family of growth cone guidance molecules, are expressed both in the ventral neural tube and in subsets of mesodermal cells. In an effort to better understand the regulation of netrins, we examined the expression of netrin-1a in mutant cyclops, no tail, and floating head zebrafish embryos, in which axial midline structures are perturbed. Netrin-1a expression requires signals present in notochord and floor plate cells. In the myotome, but not the neural tube, netrin-1a expression requires sonic hedgehog. In embryos lacking sonic hedgehog, the sonic-you locus, netrin-1a expression is reduced or absent in the myotomes but present in the neural tube. Embryos lacking sonic hedgehog express tiggy-winkle hedgehog in the floor plate, suggesting that, in the neural tube, tiggy-winkle hedgehog can compensate for the lack of sonic hedgehog in inducing netrin-1a expression. Ectopic expression of sonic hedgehog, tiggy-winkle hedgehog, or echidna hedgehog induces ectopic netrin-1a expression in the neural tube, and ectopic expression of sonic hedgehog or tiggy-winkle hedgehog, but not echidna hedgehog, induces ectopic netrin-1a expression in somites. These data demonstrate that in vertebrates netrin expression is regulated by Hedgehog signaling.

Zebrafish semaphorin Z1a collapses specific growth cones and alters their pathway in vivo.

The semaphorin/collapsin gene family encodes secreted and transmembrane proteins several of which can repulse growth cones. Although the in vitro activity of Semaphorin III/D/Collapsin 1 is clear, recent analyses of two different strains of semaphorin III/D/collapsin 1 knockout mice have generated conflicting findings. In order to clarify the in vivo action of this molecule, we analyzed sema Z1a, a zebrafish homolog of semaphorin III/D/collapsin 1. The expression pattern of sema Z1a suggested that it delimited the pathway of the growth cones of a specific set of sensory neurons, the posterior ganglion of the lateral line, in zebrafish. To examine the in vivo action of this molecule, we analyzed (1) the pathways followed by lateral line growth cones in mutants in which the expression of sema Z1a is altered in an interesting way, (2) response of lateral line growth cones to exogenous Sema Z1a in living embryos, and (3) the pathway followed by lateral line growth cones when Sema Z1a is misexpressed by cells along their normal route. The results suggest that a repulsive action of Sema Z1a helps guide the growth cones of the lateral line along their normal pathway.

Development of branchiomotor neurons in zebrafish.

The mechanisms underlying neuronal specification and axonogenesis in the vertebrate hindbrain are poorly understood. To address these questions, we have employed anatomical methods and mutational analysis to characterize the branchiomotor neurons in the zebrafish embryo. The zebrafish branchiomotor system is similar to those in the chick and mouse, except for the location of the nVII and nIX branchiomotor neurons. Developmental analyses of genes expressed by branchiomotor neurons suggest that the different location of the nVII neurons in the zebrafish may result from cell migration. To gain insight into the mechanisms underlying the organization and axonogenesis of these neurons, we examined the development of the branchiomotor pathways in neuronal mutants. The valentino b337 mutation blocks the formation of rhombomeres 5 and 6, and severely affects the development of the nVII and nIX motor nuclei. The cyclops b16 mutation deletes ventral midline cells in the neural tube, and leads to a severe disruption of most branchiomotor nuclei and axon pathways. These results demonstrate that rhombomere-specific cues and ventral midline cells play important roles in the development of the branchiomotor pathways.

Axon tracts correlate with netrin-1a expression in the zebrafish embryo.

Netrins are secreted molecules that can attract or repel growth cones from a variety of organisms. In order to clarify the extent and scope of the effects of netrins for guiding growth cones, we have analyzed netrin-1a within the relatively simple and well-characterized nervous system of zebrafish embryos. netrin-1a is expressed in dynamic patterns that suggest that it guides the growth cones of a wide variety of neurons. The spatiotemporal relationship of netrin-1a expression and extending growth cones further suggests that netrins may act to delineate specific pathways and stimulate axonal outgrowth in addition to attracting and repelling growth cones. Furthermore, aberrant outgrowth by commissural growth cones in the spinal cords of floating head mutants, in which netrin-1a expression is altered, is consistent with an in vivo, chemoattractive action of netrin-1a. These data suggest that netrins act on many growth cones and influence their behavior in a variety of ways.

The notochord and floor plate guide growth cones in the zebrafish spinal cord.

Elimination of the floor plate, a row of distinctive cells at the ventral midline of the spinal cord, dramatically increased the frequency of errors made by specific growth cones in the zebrafish embryo. This demonstrated that the floor plate participated in guiding specific growth cones at the ventral midline of the spinal cord. However, since a significant proportion of these growth cones followed their normal pathway despite the absence of the floor plate, we hypothesized that a second source of cues must exist at or near the ventral midline. We tested whether the notochord, which is located just ventral to the spinal cord, was an additional source of pathfinding cues by eliminating it prior to axonogenesis. Laser ablation of the notochord in wildtype embryos increased errors by spinal growth cones. Likewise, spinal growth cones made errors in no tail mutants that are missing the notochord but not the floor plate. These results demonstrate that the notochord also participates in guiding growth cones at the ventral midline. Furthermore, removal of both the floor plate and notochord resulted in a higher error rate than removal of either one alone. These results suggest that the notochord and floor plate normally act in concert to guide spinal growth cones at the ventral midline.

Development of the zebrafish nervous system: genetic analysis and manipulation.

The accessibility and simplicity of the zebrafish embryo have led to fruitful examinations of how vertebrate embryos develop, at the cellular level. Recently, several groups have initiated large-scale mutagenesis in zebrafish and begun to generate transgenic zebrafish. The goals of these endeavors are to identify developmentally important genes and to delineate their in vivo function. If successful, the two approaches should significantly enhance our understanding of how genes control development in a vertebrate embryo.

The molecular cloning and characterization of potential chick DM-GRASP homologs in zebrafish and mouse.

A full-length zebrafish cDNA clone and a partial mouse cDNA clone similar to chick DM-GRASP were isolated and analyzed. The nucleotide sequence of the full-length zebrafish clone shares 54% identity, and predicts 39% amino acid identity, with chick DM-GRASP. The partial mouse clone shares 76% nucleotide identity, and predicts 76% amino acid identity, with chick DM-GRASP. The predicted proteins encoded by both of these clones exhibit conserved structural domains that are characteristic of the chick protein. These features may identify them as a distinct subfamily within the immunoglobulin superfamily of cell adhesion molecules. Expression of the zebrafish DM-GRASP protein is similar to chick DM-GRASP and is principally restricted to a small subset of developing sensory and motor neurons during axonogenesis. Zebrafish DM-GRASP expression was temporally regulated and limited to specific axon domains. This regional expression correlated with fasciculated axon domains. These results suggest that the zebrafish and mouse cDNA clones represent the respective fish and mammalian homologs of chick DM-GRASP. The highly selective expression of zebrafish DM-GRASP suggests that it is involved in the selective fasciculation and guidance of axons along their normal pathways.

Axonal outgrowth within the abnormal scaffold of brain tracts in a zebrafish mutant.

The role of specific axonal tracts for the guidance of growth cones was investigated by examining axonal outgrowth within the abnormal brain tracts of zebrafish cyclops mutants. Normally, the earliest differentiating neurons in the zebrafish brain establish a simple scaffold of axonal tracts. Later-developing axons follow cell-specific pathways within this axonal scaffold. In cyclops embryos, this scaffold is perturbed due to the deletion of some ventromedial neurons that establish parts of the axonal scaffold and the development of an abnormal crease in the brain. In these mutant embryos, the growth cones projected by the neurons of the nucleus of the posterior commissure (nuc PC) are deprived of the two tracts of axons that they sequentially follow to first extend ventrally, then posteriorly. These growth cones respond to the abnormal scaffold in several interesting ways. First, nuc PC growth cones initially always extend ventrally as in wild-type embryos. This suggests that for the first portion of their pathway the axons they normally follow are not required for proper navigation. Second, approximately half of the nuc PC growth cones follow aberrant longitudinal pathways after the first portion of their pathway. This suggests that for the longitudinal portion of the pathway, specific growth cone/axon interactions are important for guiding growth cones. Third, although approximately half of the nuc PC growth cones follow aberrant longitudinal pathways, the rest follow normal pathways despite the absence of the axons that they normally follow. This suggests that cues independent of these axons may be capable of guiding nuc PC growth cones as well. These results suggest that different guidance cues or combinations of cues guide specific growth cones along different portions of their pathway.

Pathway selection by growth cones in the zebrafish central nervous system.

The accessibility and simplicity of the zebrafish embryo have led to a detailed characterization of pathfinding by identifiable growth cones in both the spinal cord and brain. These growth cones follow precise, cell-specific pathways to their targets. Analyses of pathfinding in mutant embryos and wild-type embryos following laser ablation or transplantation of specific cells demonstrate that growth cones accomplish this by interacting with specific cellular cues in their environment, many of which are likely to be redundant. These experiments suggest that a combination of separate pathway and directionality cues are required for pathfinding by growth cones. Growth cones distinguish different pathways by sensing specific pathway cues and know what direction they should extend on a pathway by sensing widely distributed directionality cues.

Axonal trajectories and distribution of GABAergic spinal neurons in wildtype and mutant zebrafish lacking floor plate cells.

The role of the midline floor plate cells in the neuronal differentiation of the spinal cord was examined by comparing putative GABAergic neurons in wildtype zebrafish embryos with those in cyc-1 mutant embryos. The mutation produces a pleiotropic recessive lethal phenotype and is severe in rostral brain regions, but its direct effect in the caudal hindbrain and the spinal cord is apparently restricted to the depletion of the midline floor plate cells. In wildtype embryos, an antibody against the neurotransmitter GABA labeled the cell bodies, axons, and growth cones of three classes of previously identified neurons; dorsal longitudinal neurons (DoLA), commissural secondary ascending neurons (CoSA), and ventral longitudinal neurons (VeLD). A novel ventral cell type, Kolmer-Agduhr (KA) neurons, was also labeled. In the cyc-1 mutant, abnormalities were observed in some, but not all, of the GABAreactive CoSA, VeLD, and KA axons, while the axonal trajectories of DoLA neurons were not affected. Furthermore, the number of KA cells was reduced in the mutant while the numbers of the other GABAreactive cells were unperturbed. These observations corroborate our earlier hypothesis that the floor plate cells are one of several guidance cues that direct axonal outgrowth near the ventral midline of the spinal cord. They also suggest that the floor plate cells may play a role in the cellular differentiation of the spinal cord of zebrafish embryos.

The paired domain-containing nuclear factor pax[b] is expressed in specific commissural interneurons in zebrafish embryos.

The zebrafish paired box (Pax) genes are expressed in the early neural tube and are thought to be transcription factors that regulate the differentiation of cells in the central nervous system (CNS). The protein product of one of these Pax genes, pax[b], is detectable as a nuclear antigen in all the regions of the embryo that transcribe the gene including the posterior midbrain, the nephritic primordium, the Wolffian duct, the optic stalk, and, in specific neurons, in the hindbrain and spinal cord. The timing and pattern of axonal outgrowth by the early pax[b]-positive neurons suggest that they are the commissural secondary ascending (CoSA) interneurons in the spinal cord; the primary commissural interneurons (MiD2c and MiD3c) in hindbrain rhombomeres mi2 and mi3; and a previously unclassified set of commissural interneurons that we termed the commissural caudalrhombomere ascending (CoCaA) interneurons in the caudal hindbrain. In contrast, the Mauthner interneurons do not express pax[b] early in development. Shortly after the appearance of the first pax[b]-positive interneurons, additional nuclei adjacent to the first pax[b]-positive interneurons become pax[b] positive. This pattern of expression suggests that the pax[b] protein may be involved in determining the identity of specific commissural interneurons.