The homeobox genes vox and vent are redundant repressors of dorsal fates in zebrafish.

Ventralizing transcriptional repressors in the Vox/Vent family have been proposed to be important regulators of dorsoventral patterning in the early embryo. While the zebrafish genes vox (vega1) and vent (vega2) both have ventralizing activity in overexpression assays, loss-of-function studies are needed to determine whether these genes have distinct or redundant functions in dorsoventral patterning and to provide critical tests of the proposed regulatory interactions among vox, vent and other genes that act to establish the dorsoventral axis. We show that vox and vent are redundant repressors of dorsal fates in zebrafish. Mutants that lack vox function have little or no dorsoventral patterning defect, and inactivation of either vox or vent by injection of antisense morpholino oligonucleotides has little or no effect on the embryo. In contrast, embryos that lack both vox and vent function have a dorsalized phenotype. Expression of dorsal mesodermal genes, including chordin, goosecoid and bozozok, is strongly expanded in embryos that lack vox and vent function, indicating that the redundant action of vox and vent is required to restrict dorsal genes to their appropriate territories. Our genetic analysis indicates that the dorsalizing transcription factor Bozozok promotes dorsal fates indirectly, by antagonizing the expression of vox and vent. In turn, vox and vent repress chordin expression, restricting its function as an antagonist of ventral fates to the dorsal side of the embryo. Our results support a model in which BMP signaling induces the expression of ventral genes, while vox and vent act redundantly to prevent the expression of chordin, goosecoid and other dorsal genes in the lateral and ventral mesendoderm.

Patterning the early zebrafish by the opposing actions of bozozok and vox/vent.

Fish and frog embryos are patterned along the dorsal-ventral axis during the gastrula stage by opposing gradients of Bmps and Bmp inhibitory proteins. Three transcriptional repressors with partially overlapping expression domains have been proposed to be important mediators of Bmp function in Xenopus. We find that two related factors are expressed in the early zebrafish embryo. Although these factors are considerably divergent from the related Xenopus genes, they are expressed in domains similar to those of their Xenopus relatives throughout embryogenesis. Both of the zebrafish genes, which we have named vox and vent, are potent ventralizing factors in both zebrafish and Xenopus embryos. Using mutants in the Bmp pathway, we find that there are Bmp-dependent and Bmp-independent domains of vox expression, whereas vent is mostly dependent upon Bmp signaling. We show that ectopic vox or vent negatively regulates expression of the early dorsal gene bozozok (boz) and that ectopic boz eliminates vox and vent expression. Moreover, the normal exclusion of vox and vent from the organizer region is lost in boz mutant embryos. Our results show that boz and vox/vent are mutually antagonistic and indicate that the early establishment of the size of the organizer domain is dependent on an interplay between these early expressed transcriptional repressors.

Regulation of dorsal gene expression in Xenopus by the ventralizing homeodomain gene Vox.

Patterning in the vertebrate embryo is controlled by an interplay between signals from the dorsal organizer and the ventrally expressed BMPs. Here we examine the function of Vox, a homeodomain-containing gene that is activated by the ventralizing signal BMP-4. Inhibition of BMP signaling using a dominant negative BMP receptor (DeltaBMPR) leads to the ectopic activation of dorsal genes in the ventral marginal zone, and this activation is prevented by co-injection of Vox. chordin is the most strongly activated of those genes that are up-regulated by DeltaBMPR and is the gene most strongly inhibited by Vox expression. We demonstrate that Vox acts as a transcriptional repressor, showing that the activity of native Vox is mimicked by a Vox-repressor fusion (VoxEnR) and that a Vox-activator fusion (VoxG4A) acts as an antimorph, causing the formation of a partial secondary axis when expressed on the ventral side of the embryo. Although Vox can ectopically activate BMP-4 expression in whole embryos, we see no activation of BMP-4 by VoxG4A, demonstrating that this activation is indirect. Using a hormone-inducible version of VoxG4A, we find that a critical time window for Vox function is during the late blastula period. Using this construct, we demonstrate that only a subset of dorsal genes is directly repressed by Vox, revealing that there are different modes of regulation for organizer genes. Since the major direct target for Vox repression is chordin, we propose that Vox acts in establishing a BMP-4 morphogen gradient by restricting the expression domain of chordin.

Spatial regulation of floating head expression in the developing notochord.

The zebrafish homeobox gene floating head (flh) is essential for notochord development and is one of the earliest genes to be expressed in notochord precursors. To understand how flh is regulated during notochord development, we compared the wild-type flh expression pattern to that in embryos mutant for flh and no tail (ntl), the zebrafish homologue of Brachyury. In the early gastrula, the pattern of flh expression is not affected in either flh or ntl mutants, implying that the initial establishment of a gastrula notochord domain is independent of the function of these genes. However, flh RNA is expressed at lower levels in flh mutants suggesting that flh positively regulates its own expression. During gastrulation, flh mutants show an abrupt loss of flh expression in cells which have involuted and entered the hypoblast, while the rest of the expression pattern appears normal, thus flh+ function is specifically required to maintain flh expression in hypoblast cells. The anterior-most part of the notochord rudiment differentially maintains flh expression in both wild types and flh mutant embryos, suggesting that there is unique regulation of flh in this region of the developing notochord. In ntl mutants the spatial pattern of flh expression is altered as early as the late gastrula stage, becoming broad and diffuse. We hypothesize that ntl+ is required for the proper convergence movements of flh-expressing cells.

Specification of cell fates at the dorsal margin of the zebrafish gastrula.

Using fate mapping techniques, we have analyzed development of cells of the dorsal marginal region in wild-type and mutant zebrafish. We define a domain in the early gastrula that is located just at the margin and centered on the dorsal midline, in which most cells generate clones that develop exclusively as notochord. The borders of the notochord domain are sharp at the level of single cells, and coincide almost exactly with the border of the expression domain of the homeobox gene floating head (flh; zebrafish homologue of Xnot), a gene essential for notochord development. In flh mutants, cells in the notochord domain generate clones of muscle cells. In contrast, notochord domain cells form mesenchyme in embryos mutant for no tail (ntl; zebrafish homologue of Brachyury). A minority of cells in the notochord domain in wild-type embryos develop as unrestricted mesoderm, invariably located in the tail, suggesting that early gastrula expression of flh does not restrict cellular potential to the notochord fate. The unrestricted tail mesodermal fate is also expressed by the forerunner cells, a cluster of cells located outside the blastoderm, adjacent to the notochord domain. We show that cells can leave the dorsal blastoderm to join the forerunners, suggesting that relocation between fate map domains might respecify notochord domain cells to the tail mesodermal fate. An intermediate fate of the forerunners is to form the epithelial lining of Kupffer's vesicle, a transient structure of the teleost tailbud. The forerunners appear to generate the entire structure of Kupffer's vesicle, which also develops in most flh mutants. Although forerunner cells are present in ntl mutants, Kupffer's vesicle never appears, which is correlated with the later severe disruption of tail development.

A homeobox gene essential for zebrafish notochord development.

The notochord is a midline mesodermal structure with an essential patterning function in all vertebrate embryos. Zebrafish floating head (flh) mutants lack a notochord, but develop with prechordal plate and other mesodermal derivatives, indicating that flh functions specifically in notochord development. We show that floating head is the zebrafish homologue of Xnot, a homeobox gene expressed in the amphibian organizer and notochord. We propose that flh regulates notochord precursor cell fate.

The gene encoding rat insulinlike growth factor-binding protein 1 is rapidly and highly induced in regenerating liver.

The liver is an epithelioid organ that can regenerate following partial hepatectomy. Although it is composed mainly of hepatocytes, it has a complex, multicellular architecture, implying that intercellular communications must exist during regeneration. As in other mitogen-stimulated cells, immediate-early growth response genes induced in the absence of prior protein synthesis are likely to play an important regulatory role in the regenerative process. Through differential screening of regenerating liver cDNA libraries, we found that one of the most highly expressed immediate-early genes in liver regeneration encodes the rat homolog of the low-molecular-weight insulinlike growth factor (IGF)-binding protein (IGFBP-1). This protein has been implicated in enhancing the mitogenic effect of IGF on tissues. IGFBP-1 gene induction is transcriptionally mediated and specific to regenerating liver, as the gene is not expressed in mitogen-stimulated fibroblasts. IGFBP-1 expression has been shown to increase under low-insulin conditions such as diabetes, and the complex regulation of expression is indicated by our finding that insulin treatment of H35 rat hepatoma cells, which induces proliferation, also causes a rapid decrease in transcription and expression of the IGFBP-1 gene. Of note, IGFBP-1 mRNA is abundant in fetal rat liver, implying that it participates in normal liver growth and development. Although regenerating liver cells continue to produce IGF-I, we did not detect IGF-I receptor mRNA during the first 24 h after hepatectomy. However, some IGFBPs may act to enhance the activity of IGF-I independently of IGF-I receptors. Thus, IGF-1 and IGFBPs may interact with hepatocytes or nonparenchymal liver cells, through either IGF-I or novel receptors. In this way, IGFBP-I and IGF-I could act in a paracrine and/or autocrine fashion in maintaining normal liver architecture during regeneration.

The immediate-early growth response in regenerating liver and insulin-stimulated H-35 cells: comparison with serum-stimulated 3T3 cells and identification of 41 novel immediate-early genes.

Liver regeneration provides a unique system for analysis of mitogenesis in intact, fully developed animals. Cellular immediate-early genes likely play an important role in cell cycle regulation and have been extensively studied in mitogen-stimulated fibroblasts lymphocytes but not in liver. We have begun to characterize the immediate-early growth response genes of mitogen-stimulated liver cells, specifically, regenerating liver and insulin-stimulated Reuber H-35 hepatoma cells, and to address differences in growth response between different cell types. Through subtraction and differential screening of cDNA libraries from regenerating liver and insulin-treated H-35 cells, we have extensively characterized 341 differentially expressed clones and identified 52 immediate-early genes. These genes have been partially sequenced and subjected to Northern (RNA) blot analysis, and 41 appear to be novel. Surprisingly, two-thirds of these genes are also expressed in BALB/c 3T3 cells, but only 10 were identified in previous studies of 3T3 cells, and of these, 6 include well-known genes like jun and fos, and only 4 are novel. Approximately one-third of the immediate-early genes identified in mitogen-stimulated liver cells or serum-stimulated NIH 3T3 cells are expressed in a tissue-specific fashion, indicating that cell type-specific regulation of the proliferative response occurs during the immediate-early period. Our findings indicate that the immediate-early response is unusually complex for the first step in a regulatory cascade, suggesting that multiple pathways must be activated. The abundance of immediate-early genes and the highly varied pattern of their expression in different cell types suggest that the tissue specificity of the proliferative response arises from the particular set of these genes expressed in a given tissue.

Immediate-early gene expression differs between regenerating liver, insulin-stimulated H-35 cells, and mitogen-stimulated Balb/c 3T3 cells. Liver-specific induction patterns of gene 33, phosphoenolpyruvate carboxykinase, and the jun, fos, and egr families.

Immediate-early genes, whose expression increases independent of de novo protein synthesis during the transition from quiescence to proliferation, are postulated to play important regulatory roles in the growth response. The complement of immediate-early genes expressed must depend on the milieu of preexisting transcription factors in the quiescent cell as well as the type of mitogenic stimulation and, thus, may differ between cell types. We have begun characterizing the immediate-early response in regenerating liver and insulin-stimulated Reuber H-35 hepatoma cells in comparison with previously published results from mitogen-stimulated Balb/c 3T3 fibroblasts. The proliferating H-35 and regenerating liver cells maintain their similarity to quiescent liver as demonstrated by their continued production of the liver-specific albumin, CCAAT/enhancer binding protein, and phosphoenolpyruvate carboxykinase messenger RNAs (mRNA). Surprisingly, the phosphoenolpyruvate carboxykinase gene, which undergoes down-regulation in insulin-treated H-35 cells, was cloned by differential screening of a subtraction-enriched regenerating liver cDNA library and is an immediate-early gene in regenerating liver. H-35 cells treated with either insulin or phorbol 12-myristate 13-acetate express elevated levels of the jun genes, and phorbol 12-myristate 13-acetate pretreatment fails to abolish the insulin response, indicating that it does not depend on protein kinase C. jun family gene expression in regenerating liver differs from that in mitogen-treated fibroblasts in that the time course of expression of c-jun and junB is prolonged, and junD mRNA levels distinctly increase. Additionally, although c-fos and egr-1 mRNAs are expressed at elevated levels in stimulated liver cells, fos-B, fra-1, and egr-2 are not, which suggests that factors in addition to the serum response factor participate in the regulation of immediate-early gene induction. Interestingly, gene 33, which was cloned from a regenerating liver cDNA library by differential screening and lacks a recognizable serum response element, functions as an immediate-early gene in regenerating liver and in mitogen-treated H-35 and Balb/c 3T3 cells. These results suggest that gene 33 participates in the transition from quiescence to proliferation in many mitogen-treated cells in addition to its previously reported involvement in hormone responses. Overall, the results presented here suggest that the immediate-early response varies considerably between regenerating liver and mitogen-stimulated fibroblasts and could involve multiple, preexisting, tissue-specific, transcription-activating proteins.