Precise genome editing by homologous recombination.

Simple and efficient methods are presented for creating precise modifications of the zebrafish genome. Edited alleles are generated by homologous recombination between the host genome and double-stranded DNA (dsDNA) donor molecules, stimulated by the induction of double-strand breaks at targeted loci in the host genome. Because several kilobase-long tracts of sequence can be exchanged, multiple genome modifications can be generated simultaneously at a single locus. Methods are described for creating: (1) alleles with simple sequence changes or in-frame additions, (2) knockin/knockout alleles that express a reporter protein from an endogenous locus, and (3) conditional alleles in which exons are flanked by recombinogenic loxP sites. Significantly, our approach to genome editing allows the incorporation of a linked reporter gene into the donor sequences so that successfully edited alleles can be identified by virtue of expression of the reporter. Factors affecting the efficiency of genome editing are discussed, including the finding that dsDNA products of I-SceI meganuclease enzyme digestion are particularly effective as donor molecules for gene-editing events. Reagents and procedures are described for accomplishing efficient genome editing in the zebrafish.

Maternal and embryonic expression of zebrafish lef1.

Transcription factors of the TCF/LEF family interact with the Wnt signaling pathway to control transcription of downstream genes (Clevers, H., van de Wetering, M., 1997. TCF/LEF factor earn their wings. Trends Genet. 13, 485-489). We were interested in cloning family members which were expressed in zebrafish neural crest, because Wnt signaling modulates specification of neural crest fate (Dorsky, R.I., Moon, R.T., Raible, D.W., 1998. Control of neural crest cell fate by the Wnt signalling pathway. Nature 396, 370-373). We cloned a zebrafish homolog of lef1 and localized its chromosomal position by radiation hybrid mapping. lef1 is expressed in the neural crest as well as the tailbud and developing mesoderm, and is maternally expressed in zebrafish, unlike mouse and Xenopus homologs. In addition, we cloned two tcf3 genes and a homolog of tcf4, neither of which were strongly expressed in premigratory neural crest.

alyron, an insertional mutation affecting early neural crest development in zebrafish.

alyronz12 (aln) is a recessive lethal mutation that affects early stages of neural crest development in the zebrafish. alyron appears to be an insertional mutation as the mutation was generated following microinjection of plasmid DNA into one-cell embryos and the stably integrated transgenic sequences are closely linked to the mutation. The insertion site harbors multiple copies of the plasmid sequence that have experienced complex rearrangements. Host-insert junction fragments have been molecularly cloned and host sequences adjacent to the transgene have been used to map the mutation to the distal arm of linkage group 15. alyron function is required cell-autonomously in the neural crest lineage. alyron mutants have a severe but not complete deficit of premigratory neural crest as judged by reduced expression of several markers associated with early stages of neural crest development. Lack of premigratory neural crest is likely to account for the two most conspicuous characteristics of alyron mutants: the absence of body pigmentation and the inability to affect blood circulation. The neural crest phenotype of alyron mutants resembles that observed in mouse mutants that lack Pax-3 or both Wnt-1 and Wnt-3a function, and expression of the zebrafish homologues of these genes is greatly reduced in the dorsal neural keels of alyron mutants. In contrast, ventral neural keel identity appears unaffected. Given our findings that the mutation is unlinked to pax or wnt genes that have been described in the zebrafish, we propose that alyron is a novel gene function required for the specification and/or proliferative expansion of neural crest progenitors.

Delta-mediated specification of midline cell fates in zebrafish embryos.

BACKGROUND: Fate mapping studies have shown that progenitor cells of three vertebrate embryonic midline structures - the floorplate in the ventral neural tube, the notochord and the dorsal endoderm - occupy a common region prior to gastrulation. This common region of origin raises the possibility that interactions between midline progenitor cells are important for their specification prior to germ layer formation. RESULTS: One of four known zebrafish homologues of the Drosophila melanogaster cell-cell signaling gene Delta, deltaA (dlA), is expressed in the developing midline, where progenitor cells of the ectodermal floorplate, mesodermal notochord and dorsal endoderm lie close together before they occupy different germ layers. We used a reverse genetic strategy to isolate a missense mutation of dlA, dlAdx2, which coordinately disrupts the development of floorplate, notochord and dorsal endoderm. The dlAdx2 mutant embryos had reduced numbers of floorplate and hypochord cells; these cells lie above and beneath the notochord, respectively. In addition, mutant embryos had excess notochord cells. Expression of a dominant-negative form of Delta protein driven by mRNA microinjection produced a similar effect. In contrast, overexpression of dlA had the opposite effect: fewer trunk notochord cells and excess floorplate and hypochord cells. CONCLUSION: Our results indicate that Delta signaling is important for the specification of midline cells. The results are most consistent with the hypothesis that developmentally equivalent midline progenitor cells require Delta-mediated signaling prior to germ layer formation in order to be specified as floorplate, notochord or hypochord.

Expression of c-ret in the zebrafish embryo: potential roles in motoneuronal development.

We have isolated and characterized the zebrafish ortholog of c-ret, a gene essential for renal organogenesis and enteric nervous system development in mammals. During zebrafish embryogenesis c-ret transcripts are expressed in a number of tissues including spinal motoneurons, pronephric ducts, cranial ganglia, pharyngeal arches, and the enteric nervous system. We have examined in detail the expression of c-ret during the development of identified spinal primary motoneurons. c-ret expression is regulated in a cell type-specific manner among the three primary motoneurons. c-ret is expressed at its highest levels in caudal primary (CaP) motoneurons and transcripts can be detected shortly before the expression of the CaP-specific gene, islet2. We suggest that c-ret may play a role in specifying CaP cell identity. c-ret is expressed at low levels in the other primary motoneurons and also in a subset of secondary motoneurons, suggesting that it may also play a broader role in motoneuronal survival or maintenance.

tbx6, a Brachyury-related gene expressed by ventral mesendodermal precursors in the zebrafish embryo.

Classical embryology experiments have indicated the existence of dorsal-type and ventral-type mesoderms that arise as a consequence of mesoderm induction during vertebrate development. Here we report that the zebrafish tbx6 gene, a member of the Brachyury-related T-box family of genes, is exclusively expressed by ventral mesendoderm. Three observations link the expression of tbx6 to ventral mesoderm specification. First, the gene is initially expressed at the onset of gastrulation within a ventrolateral subpopulation of cells that express the pan-mesodermal gene, no tail (Brachyury). Second, the mesoderm-inducing factors activin and bFGF activate tbx6 expression in animal caps. Third, dorsalization of the mesendodermal precursor population following exposure of embryos to lithium ions causes down-regulation of tbx6 transcription. tbx6 is expressed transiently in the involuting derivatives of the ventral mesendoderm, which give rise to nonaxial mesodermal tissues; its expression is extinguished as tissue differentiation progresses. Transcription of tbx6 commences about an hour after initiation of expression of the pan-mesendodermal gene no tail and the organizer gene goosecoid. The dependence of tbx6 expression on no tail activity was examined in no tail mutant embryos. The activation of tbx6 transcription in ventral mesoderm does not depend on no tail gene activity. However, no tail appears to contribute to the maintenance of normal levels of tbx6 transcription and may be required for tbx6 transcription in the developing tail.

The zebrafish gene cloche acts upstream of a flk-1 homologue to regulate endothelial cell differentiation.

The zebrafish cloche mutation affects both the endothelial and hematopoietic lineages at a very early stage (Stainier, D. Y. R., Weinstein, B. M., Detrich, H. W., Zon, L. I. and Fishman, M. C. (1995). Development 121, 3141-3150). The most striking vascular phenotype is the absence of endocardial cells from the heart. Microscopic examination of mutant embryos reveals the presence of endothelial-like cells in the lower trunk and tail regions while head vessels appear to be missing, indicating a molecular diversification of the endothelial lineage. Cell transplantation experiments show that cloche acts cell-autonomously within the endothelial lineage. To analyze further the role of cloche in regulating endothelial cell differentiation, we have examined the expression of flk-1 and tie, two receptor tyrosine kinase genes expressed early and sequentially in the endothelial lineage. In wild-type fish, flk-1-positive cells are found throughout the embryo and differentiate to form the nascent vasculature. In cloche mutants, flk-1-positive cells are found only in the lower trunk and tail regions, and this expression is delayed as compared to wild-type. Unlike the flk-1-positive cells in wild-type embryos, those in cloche mutants do not go on to express tie, suggesting that their differentiation is halted at an early stage. We also find that the cloche mutation is not linked to flk-1. These data indicate that cloche affects the differentiation of all endothelial cells and that it acts at a very early stage, either by directly regulating flk-1 expression or by controlling the differentiation of cells that normally develop to express flk-1. cloche mutants also have a blood deficit and their hematopoietic tissues show no expression of the hematopoietic transcription factor genes GATA-1 or GATA-2 at early stages. Because the appearance of distinct levels of flk-1 expression is delayed in cloche mutants, we examined GATA-1 expression at late embryonic stages and found some blood cell differentiation that appears to be limited to the region lined by the flk-1-expressing cells. The spatial restriction of blood in the ventroposterior-most region of cloche mutant embryos may be indicative of a ventral source of signal(s) controlling hematopoietic differentiation. In addition, the restricted colocalization of blood and endothelium in cloche mutants suggests that important interactions occur between these two lineages during normal development.

A mutation in zebrafish affecting a localized cellular function required for normal ear development.

Zebrafish holds great potential as a model system for studying inner ear development because genetic techniques are highly efficient and inner ear development is a conspicuous and manipulable feature of zebrafish embryogenesis. Here we describe analysis of a semilethal dominant mutation, termed monolith (mnl), that specifically perturbs formation of the anterior (utricular) otolith in the developing ear. Other than the utricular otolith deficiency, all structures in the ear appear morphologically normal in mutant embryos, including posterior otoliths and all sensory epithelia. Expression patterns of several ear marker genes (msxC, msxD, and dlx3) also appear normal in the mutant. To identify the cell type(s) affected by the mnl mutation, chimeras were generated by transplanting dye-labeled +/+ cells into unlabeled mnl/mnl host embryos. Roughly half of such chimeras formed utricular otoliths normally, indicating that the transplanted wild-type cells rescued their mutant hosts. Detailed analysis of +/+ cell fates revealed that virtually all chimeras in which +/+ cells formed support cells in the utricular sensory epithelium were rescued. In contrast, wild-type cells forming other cell types (such as hair cells) or colonizing other regions of the host were not sufficient to facilitate rescue. These data indicate that support cells are required for normal otolith formation, providing the first experimentally established role for support cells in vertebrate sensory epithelia. The data also provide the first clear indication that otolith formation is controlled independently in different regions of the ear by localized cellular functions.

Developmental regulation of zebrafish MyoD in wild-type, no tail and spadetail embryos.

We describe the isolation of the zebrafish MyoD gene and its expression in wild-type embryos and in two mutants with altered somite development, no tail (ntl) and spadetail (spt). In the wild-type embryo, MyoD expression first occurs in an early phase, extending from mid-gastrula to just prior to somite formation, in which cells directly adjacent to the axial mesoderm express the gene. In subsequent phases, during the anterior-to-posterior wave of somite formation and maturation, expression occurs within particular regions of each somite. In spt embryos, which lack normal paraxial mesoderm due to incorrect cell migration, early MyoD expression is not observed and transcripts are instead first detected in small groups of trunk cells that will develop into aberrant myotomal-like structures. In ntl embryos, which lack notochords and tails, the early phase of MyoD expression is also absent. However, the later phase of expression within the developing somites appears to occur at the normal time in the ntl mutants, indicating that the presomitogenesis and somitogenesis phases of MyoD expression can be uncoupled. In addition, we demonstrate that the entire paraxial mesoderm of wild-type embryos has the potential to express MyoD when Sonic hedgehog is expressed ubiquitously in the embryo, and that this potential is lost in some of the cells of the paraxial mesoderm lineage in no tail and spadetail embryos. We also show that MyoD expression precedes myogenin expression and follows or is coincident with expression of snaill in some regions that express this gene.

Efficient induction of point mutations allowing recovery of specific locus mutations in zebrafish.

A technique is described that greatly increases the efficiency of recovering specific locus point mutations in zebrafish (Danio rerio). Founder individuals that were mosaic for point mutations were produced by mutagenizing postmeiotic gametes with the alkylating agent N-ethyl-N-nitrosourea. Under optimal conditions, each founder carried an average of 10 mutations affecting genes required for embryogenesis. Moreover, approximately 2% of these founders transmitted new mutations at any prespecified pigmentation locus. Analyses of new pigmentation mutations confirmed that most were likely to be point mutations. Thus, mutagenesis of postmeiotic gametes with N-ethyl-N-nitrosourea yielded frequencies of point mutations at specific loci that were 10- to 15-fold higher than previously achieved in zebrafish. Our procedure should, therefore, greatly facilitate recovery of multiple mutant alleles at any locus of interest.

Contribution of early cells to the fate map of the zebrafish gastrula.

Previously, a tissue-specific fate map was compiled for the gastrula stage of the zebrafish embryo, indicating that development subsequent to this stage follows a reproducible pattern. Here it is shown that each early zebrafish blastomere normally contributes to a subset of the gastrula and thus gives rise to a limited array of tissues. However, the final contribution that any early blastomere makes to the fate map in the gastrula cannot be predicted because of variability in both the position of the future dorsoventral axis with respect to the early cleavage blastomeres and the scattering of daughter cells as the gastrula is formed. Therefore, early cell divisions of the zebrafish embryo cannot reproducibly segregate determinants of tissue fates.

Something's fishy here--rethinking cell movements and cell fate in the zebrafish embryo.

A recent paper has challenged the prevailing view that zebrafish blastomeres undergo extensive cell mixing and are unrestricted in their cell fates during early development. Here, we offer a model that resolves apparent contradictions between studies that have examined the origins of cell fate in the zebrafish embryo. The model reconsiders cell movements during epiboly, and how cell mixing, or the lack of it, can affect the predictability of cell fate.

The DVR-1 (Vg1) transcript of zebrafish is maternally supplied and distributed throughout the embryo.

It is not known how region- or tissue-specific differences are generated in the zebrafish embryo. To look at the potential role of maternal transcripts in generating cell diversity, we have isolated and characterized the zebrafish homologue of Xenopus DVR-1 (Vg1), a maternally supplied RNA that encodes a member of the transforming growth factor-beta superfamily. The zebrafish DVR-1 RNA is maternally supplied and its protein product shares a high degree of sequence identity with Xenopus DVR-1. These conserved features indicate that DVR-1 is likely to have an essential function in early embryogenesis. However, unlike the frog transcript, which is restricted to vegetal cells, DVR-1 RNA is distributed equally among all zebrafish blastomeres. We suggest that the ubiquitous distribution of DVR-1 RNA reflects a significant aspect of the developmental strategy of the zebrafish in which each blastomere retains an equivalent developmental potential throughout the cleavage period.

Lithium perturbation and goosecoid expression identify a dorsal specification pathway in the pregastrula zebrafish.

The zebrafish dorsoventral axis can first be distinguished at gastrulation, upon formation of the embryonic shield, the site of the organizer. We have asked whether the shield is specified before gastrulation. First, we show that brief exposure of premidblastula embryos to lithium, which is known to shut down the phospho-inositol signaling pathway, produces excessive shield formation and extreme hyper-dorsal development. Second, we show that the zebrafish goosecoid homeobox gene is activated at or just after the midblastula in a localized domain of cells that subsequently populate the most anterior region of the incipient shield and axial hypoblast, goosecoid expression is elevated and radialized by early lithium treatment, suggesting that goosecoid plays a role in establishing the organizer and shield. Our results demonstrate that the zebrafish dorsal axis is signaled by a pathway initiated in the cleavage-stage embryo. Furthermore, they provide novel insights into anterior morphogenesis.

Induction of recessive lethal and specific locus mutations in the zebrafish with ethyl nitrosourea.

Recessive lethal mutations and mutations at the gol-1 locus were induced in the zebrafish by exposure of mature sperm to the alkylating agent ethyl nitrosourea (ENU). Embryonic lethal phenotypes were recognized among the parthenogenetic progeny of mutagenized animals or among the progeny of daughters of mutagenized animals. Novel specific locus mutations were identified by the failure of mutagenized chromosomes to complement pre-existing mutant alleles at the gol-1 locus. Each mutagenized individual harboured approximately 10 embryonic lethal mutations in its germ line and about 1 in 500 mutagenized animals harboured a new mutation at the gol-1 locus. Three lines of evidence indicate that the majority of mutations that were recovered following treatment of mature sperm with ENU were probably point mutations. First, the soma and germ lines of mutagenized animals were mosaic, as expected following simple alkylation of sperm DNA. Second, mutations induced by ENU at the gol-1 locus affected pigmentation but not viability, unlike the majority of mutations induced at this locus with gamma-irradiation. Third, the ratio of specific locus:recessive lethal mutations induced by ENU was approximately 50-fold lower than the ratio observed following mutagenesis with gamma-rays. Comparison of the incidence with which embryonic recessive lethal mutations were induced with the incidence with which specific locus mutations arose indicates that there are greater than 5000 genes essential to the development and viability of the zebrafish embryo.

Induction of mutations in the zebrafish with ultraviolet light.

Recessive lethal germline and specific locus somatic mutations were induced efficiently in the zebrafish by exposure of mature sperm to UV light. Mutagenesis of sperm yielded mosaic individuals: clones bearing novel mutations represented approximately 12-25% of the haploid germ cells and 25-50% of the somatic tissue. Simple methods are described for the reliable identification and propagation of newly arising developmental mutations in zebrafish.

Clonal origins of cells in the pigmented retina of the zebrafish eye.

Mosaic analysis has been used to study the clonal basis of the development of the pigmented retina of the zebrafish, Brachydanio rerio. Zebrafish embryos heterozygous for a recessive mutation at the gol-1 locus were exposed to gamma-irradiation at various developmental stages to create mosaic individuals consisting of wild-type pigmented cells and a clone of pigmentless (golden) cells in the eye. The contribution of individual embryonic cells to the pigmented retina was measured and the total number of cells in the embryo that contributed descendants to this tissue was determined. Until the 32-cell stage, almost every blastomere has some descendants that participate in the formation of the pigmented retina of the zebrafish. During subsequent cell divisions, up to the several thousand-cell stage, the number of ancestral cells is constant: approximately 40 cells are present that will give rise to progeny in the pigmented retina. Analysis of the size of clones in the pigmented retina indicates that the cells of this tissue do not arise through a rigid series of cell divisions originating in the early embryo. The findings that each cleavage stage cell contributes to the pigmented retina and yet the contribution of such cells is highly variable are consistent with the interpretation that clonal descendants of different blastomeres normally intermix extensively prior to formation of the pigmented retina.

A neural degeneration mutation that spares primary neurons in the zebrafish.

We describe an embryonic lethal mutation in the zebrafish Brachydanio rerio that specifically affects the viability of most cells in the embryonic central nervous system (CNS). The mutation ned-1 (b39rl) was induced with gamma-irradiation and segregates as a single recessive allele closely linked to its centromere. It produces massive cell death in the CNS but a small set of specific neurons, including Rohon-Beard sensory neurons, large hindbrain interneurons, and primary motoneurons, survive embryogenesis and are functional. Synaptic connections between embryonic motoneurons and muscle cells appear physiologically normal, and the normally observed spontaneous flexions are present. Correlated with the presence of sensory neurons and interneurons, mutant embryos display reflexive movements in response to mechanical stimulation. Together, the surviving neurons, called primary neurons, form a class of cells that are prominent in size and arise early during development. Thus, this mutation may define a function that is differentially required by developmentally distinguishable sets of cells in the embryonic CNS.

P80: a tumor-related protein found in many lymphomas of mice.

Examination of syngeneic tumor regressor sera prepared by immunization of mice with several different lymphomas revealed a common pattern of reactivity to proteins expressed in these tumors. Antibodies present in these sera immunoprecipitate a triplet of proteins of 115,000 mol wt (p115), 80,000 mol wt (p80), and 32,000 mol wt (p32) from many but not all T cell lymphomas of mice. P80, the predominant molecular species immunoprecipitated with these sera, is a nonglycosylated, phosphoprotein that does not appear to be expressed at the cell surface. Comparison of the tryptic peptides of p32 and p80 indicated that the peptides found in p32 are a subset of those found in p80. Comparison of the tryptic peptides of p80 with those of the p120 gag-fusion protein of Abelson murine leukemia virus demonstrated that p80 and p120 did not share tryptic peptides. Comparison of the partial proteolytic products generated by treatment of p80 molecules from different tumors with V8 protease did not reveal heterogeneity in p80 among tumors of different strains of mice. Direct labeling and competition blocking experiments with lysates from normal cells failed to provide evidence of p80 synthesis in normal thymus, spleen, or bone marrow. Thus, p80 is a biochemically identified tumor-related antigen of mouse lymphomas.