Transcription of fgf8 is regulated by activating and repressive cis-elements at the midbrain-hindbrain boundary in zebrafish embryos

Dissection of the Fgf8 regulatory landscape by in vivo CRISPR-editing reveals extensive intra- and inter-enhancer redundancy

Developmental genes are often regulated by multiple elements with overlapping activity. Yet, in most cases, the relative function of those elements and their contribution to endogenous gene expression remain poorly characterized. An example of this phenomenon is that distinct sets of enhancers have been proposed to direct Fgf8 in the limb apical ectodermal ridge and the midbrain-hindbrain boundary. Using in vivo CRISPR/Cas9 genome engineering, we functionally dissect this complex regulatory ensemble and demonstrate two distinct regulatory logics. In the apical ectodermal ridge, the control of Fgf8 expression appears distributed between different enhancers. In contrast, we find that in the midbrain-hindbrain boundary, one of the three active enhancers is essential while the other two are dispensable. We further dissect the essential midbrain-hindbrain boundary enhancer to reveal that it is also composed by a mixture of essential and dispensable modules. Cross-species transgenic analysis of this enhancer suggests that its composition may have changed in the vertebrate lineage.

In vitro analysis of the transcriptional regulatory mechanism of zebrafish pou5f3

Zebrafish pou5f3 (previously named pou2), a close homologue of mouse Oct4, encodes a PouV-family transcription factor. pou5f3 has been implicated in diverse aspects of developmental regulation during embryogenesis. In the present study, we addressed the molecular function of Pou5f3 as a transcriptional regulator and the mechanism by which pou5f3 expression is transcriptionally regulated. We examined the influence of effector genes on the expression of the luciferase gene under the control of the upstream 2.1-kb regulatory DNA of pou5f3 (Luc-2.2) in HEK293T and P19 cells. We first confirmed that Pou5f3 functions as a transcriptional activator both in cultured cells and embryos, which confirmed autoregulation of pou5f3 in embryos. It was further shown that Luc-2.2 was activated synergistically by pou5f3 and sox3, which is similar to the co-operative activity of Oct4 and Sox2 in mice, although synergy between pou5f3 and sox2 was less obvious in this zebrafish system. The effects of pou5f3 deletion constructs on the regulation of Luc-2.2 expression revealed different roles for the three subregions of the N-terminal region in Pou5f3 in terms of its regulatory functions and co-operativity with Sox3. Electrophoretic mobility shift assays confirmed that Pou5f3 and Sox3 proteins specifically bind to adjacent sites in the 2.1-kb DNA and that there is an interaction between the two proteins. The synergy with sox3 was unique to pou5f3-the other POU factor genes examined did not show such synergy in Luc-2.2 regulation. Finally, functional interaction was observed between pou5f3 and sox3 in embryos in terms of the regulation of dorsoventral patterning and convergent extension movement. These findings together demonstrate co-operative functions of pou5f3 and sox3, which are frequently coexpressed in early embryos, in the regulation of early development.

Mammalian TBX1 preferentially binds and regulates downstream targets via a tandem T-site repeat

Haploinsufficiency or mutation of TBX1 is largely responsible for the etiology of physical malformations in individuals with velo-cardio-facial/DiGeorge syndrome (VCFS/DGS/22q11.2 deletion syndrome). TBX1 encodes a transcription factor protein that contains an evolutionarily conserved DNA binding domain termed the T-box that is shared with other family members. All T-box proteins, examined so far, bind to similar but not identical consensus DNA sequences, indicating that they have specific binding preferences. To identify the TBX1 specific consensus sequence, Systematic Evolution of Ligands by Exponential Enrichment (SELEX) was performed. In contrast to other TBX family members recognizing palindrome sequences, we found that TBX1 preferentially binds to a tandem repeat of 5′-AGGTGTGAAGGTGTGA-3′. We also identified a second consensus sequence comprised of a tandem repeat with a degenerated downstream site. We show that three known human disease-causing TBX1 missense mutations (F148Y, H194Q and G310S) do not alter nuclear localization, or disrupt binding to the tandem repeat consensus sequences, but they reduce transcriptional activity in cell culture reporter assays. To identify Tbx1-downstream genes, we performed an in silico genome wide analysis of potential cis-acting elements in DNA and found strong enrichment of genes required for developmental processes and transcriptional regulation. We found that TBX1 binds to 19 different loci in vitro, which may correspond to putative cis-acting binding sites. In situ hybridization coupled with luciferase gene reporter assays on three gene loci, Fgf8, Bmper, Otog-MyoD, show that these motifs are directly regulated by TBX1 in vitro. Collectively, the present studies establish new insights into molecular aspects of TBX1 binding to DNA. This work lays the groundwork for future in vivo studies, including chromatin immunoprecipitation followed by next generation sequencing (ChIP-Seq) to further elucidate the molecular pathogenesis of VCFS/DGS.

An integrated holo-enhancer unit defines tissue and gene specificity of the Fgf8 regulatory landscape

Fgf8 encodes a key signaling factor, and its precise regulation is essential for embryo patterning. Here, we identified the regulatory modules that control Fgf8 expression during mammalian embryogenesis. These enhancers are interspersed with unrelated genes along a large region of 220 kb; yet they act on Fgf8 only. Intriguingly, this region also contains additional genuine enhancer activities that are not transformed into gene expression. Using genomic engineering strategies, we showed that these multiple and distinct regulatory modules act as a coherent unit and influence genes depending on their position rather than on their promoter sequence. These findings highlight how the structure of a locus regulates the autonomous intrinsic activities of the regulatory elements it contains and contributes to their tissue and target specificities. We discuss the implications of such regulatory systems regarding the evolution of gene expression and the impact of human genomic structural variations.

A SINE-derived element constitutes a unique modular enhancer for mammalian diencephalic Fgf8

Transposable elements, including short interspersed repetitive elements (SINEs), comprise nearly half the mammalian genome. Moreover, they are a major source of conserved non-coding elements (CNEs), which play important functional roles in regulating development-related genes, such as enhancing and silencing, serving for the diversification of morphological and physiological features among species. We previously reported a novel SINE family, AmnSINE1, as part of mammalian-specific CNEs. One AmnSINE1 locus, named AS071, showed an enhancer property in the developing mouse diencephalon. Indeed, AS071 appears to recapitulate the expression of diencephalic fibroblast growth factor 8 (Fgf8). Here we established three independent lines of AS071-transgenic mice and performed detailed expression profiling of AS071-enhanced lacZ in comparison with that of Fgf8 across embryonic stages. We demonstrate that AS071 is a distal enhancer that directs Fgf8 expression in the developing diencephalon. Furthermore, enhancer assays with constructs encoding partially deleted AS071 sequence revealed a unique modular organization in which AS071 contains at least three functionally distinct sub-elements that cooperatively direct the enhancer activity in three diencephalic domains, namely the dorsal midline and the lateral wall of the diencephalon, and the ventral midline of the hypothalamus. Interestingly, the AmnSINE1-derived sub-element was found to specify the enhancer activity to the ventral midline of the hypothalamus. To our knowledge, this is the first discovery of an enhancer element that could be separated into respective sub-elements that determine regional specificity and/or the core enhancing activity. These results potentiate our understanding of the evolution of retroposon-derived cis-regulatory elements as well as the basis for future studies of the molecular mechanism underlying the determination of domain-specificity of an enhancer.

Distribution and localization of fibroblast growth factor-8 in rat brain and nerve cells during neural stem/progenitor cell differentiation

The present study explored the distribution and localization of fibroblast growth factor-8 and its potential receptor, fibroblast growth factor receptor-3, in adult rat brain in vivo and in nerve cells during differentiation of neural stem/progenitor cells in vitro. Immunohistochemistry was used to examine the distribution of fibroblast growth factor-8 in adult rat brain in vivo. Localization of fibroblast growth factor-8 and fibroblast growth factor receptor-3 in cells during neural stem/progenitor cell differentiation in vitro was detected by immunofluorescence. Flow cytometry and immunofluorescence were used to evaluate the effect of an anti-fibroblast growth factor-8 antibody on neural stem/progenitor cell differentiation and expansion in vitro. Results from this study confirmed that fibroblast growth factor-8 was mainly distributed in adult midbrain, namely the substantia nigra, compact part, dorsal tier, substantia nigra and reticular part, but was not detected in the forebrain comprising the caudate putamen and striatum. Unusual results were obtained in retrosplenial locations of adult rat brain. We found that fibroblast growth factor-8 and fibroblast growth factor receptor-3 were distributed on the cell membrane and in the cytoplasm of nerve cells using immunohistochemistry and immunofluorescence analyses. We considered that the distribution of fibroblast growth factor-8 and fibroblast growth factor receptor-3 in neural cells corresponded to the characteristics of fibroblast growth factor-8, a secretory factor. Addition of an anti-fibroblast growth factor-8 antibody to cultures significantly affected the rate of expansion and differentiation of neural stem/progenitor cells. In contrast, addition of recombinant fibroblast growth factor-8 to differentiation medium promoted neural stem/progenitor cell differentiation and increased the final yields of dopaminergic neurons and total neurons. Our study may help delineate the important roles of fibroblast growth factor-8 in brain activities and neural stem/progenitor cell differentiation.

Regulation of XFGF8 gene expression through SRY (sex-determining region Y)-box 2 in developing Xenopus embryos

Fibroblast growth factors (FGFs) function as mitogens and morphogens during vertebrate development. In the present study, to characterise the regulatory mechanism of FGF8 gene expression in developing Xenopus embryos the upstream region of the Xenopus FGF8 (XFGF8) gene was isolated. The upstream region of the XFGF8 gene contains two putative binding sites for the SRY (sex-determining region Y)-box 2 (SOX2) transcription factor. A reporter assay with serially deleted constructs revealed that the putative SOX2-binding motif may be a critical cis-element for XFGF8 gene activation in developing Xenopus embryos. Furthermore, Xenopus SOX2 (XSOX2) physically interacted with the SOX2-binding motif within the upstream region of the XFGF8 gene in vitro and in vivo. Depletion of endogenous XSOX2 resulted in loss of XFGF8 gene expression in midbrain–hindbrain junction, auditory placode, lens placode and forebrain in developing Xenopus embryos. Collectively, our results suggest that XSOX2 directly upregulates XFGF8 gene expression in the early embryonic development of Xenopus.

A lineage specific enhancer drives Otx2 expression in teleost organizer tissues

In mouse Otx2 plays essential roles in anterior-posterior axis formation and head development in anterior visceral endoderm and anterior mesendoderm. The Otx2 expression in these sites is regulated by VE and CM enhancers at the 5' proximal to the translation start site, and we proposed that these enhancers would have been established in ancestral sarcoptergians after divergence from actinopterigians for the use of Otx2 as the head organizer gene (Kurokawa et al., 2010). This would make doubtful an earlier proposal of ours that a 1.1 kb fragment located at +14.4 to +15.5 kb 3' (3'En) of fugu Otx2a gene harbors enhancers phylogenetically and functionally homologous to mouse VE and CM enhancers (Kimura-Yoshida et al., 2007). In the present study, we demonstrate that fugu Otx2a is not expressed in the dorsal margin of blastoderm, shield and early anterior mesendoderm, and that the fugu Otx2a 3'En do not exhibit activities at these sites of fugu embryos. We conclude that the fugu Otx2a 3'En does not harbor an organizer enhancer, but encodes an enhancer for the expression in later anterior mesendodermal tissues. Instead, in fugu embryos Otx2b is expressed in the dorsal margin of blastoderm at blastula stage and shield at 50% epiboly, and this expression is directed by an enhancer, 5'En, located at -1000 to -800 bp, which is uniquely conserved among teleost Otx2b orthologues.

FGF receptor gene expression and its regulation by FGF signaling during early zebrafish development

The expression of all four fgfr genes was extensively examined throughout early embryogenesis of the zebrafish (Danio rerio). fgfr1 alone was expressed maternally throughout the blastoderm, and then zygotically in the anterior neural plate and presomitic mesoderm. fgfr4 expression was first detected in late blastulae and was gradually restricted to the brain. fgfr2 and fgfr3 expression were initiated in early and late gastrulae, respectively; fgfr2 was expressed in the anterior neural plate and somitic mesoderm, whereas fgfr3 was activated in the axial mesoderm and then in the midbrain and somitic mesoderm. During somitogenesis, each of these fgfr genes was expressed in a characteristic manner in the brain. Using an FGF signal inhibitor, dominant-negative FGF receptors and fgf8.1/fgf8a mutants, we found that fgfr expression is directly or indirectly regulated by FGF signaling during epiboly and at the end of somitogenesis, revealing the presence of an autoregulatory mechanism.

Retinoic acid-dependent establishment of positional information in the hindbrain was conserved during vertebrate evolution

Zebrafish hoxb1b is expressed during epiboly in the posterior neural plate, with its anterior boundary at the prospective r4 region providing a positional cue for hindbrain formation. A similar function and expression is known for Hoxa1 in mice, suggesting a shared regulatory mechanism for hindbrain patterning in vertebrate embryos. To understand the evolution of the regulatory mechanisms of key genes in patterning of the central nervous system, we examined how hoxb1b transcription is regulated in zebrafish embryos and compared the regulatory mechanisms between mammals and teleosts that have undergone an additional genome duplication. By promoter analysis, we found that the expression of the reporter gene recapitulated hoxb1b expression when driven in transgenic embryos by a combination of the upstream 8.0-kb DNA and downstream 4.6-kb DNA. Furthermore, reporter expression expanded anteriorly when transgenic embryos were exposed to retinoic acid (RA) or LiCl, or injected with fgf3/8 mRNA, implicating the flanking DNA examined here in the responsiveness of hoxb1b to posteriorizing signals. We further identified at least two functional RA responsive elements in the downstream DNA that were shown to be major regulators of early hoxb1b expression during gastrulation, while the upstream DNA, which harbors repetitive sequences with apparent similarity to the autoregulatory sequence of mouse Hoxb1, contributed only to later hoxb1b expression, during somitogenesis. Possible implications in vertebrate evolution are discussed based on these findings.

Cis-regulation and chromosomal rearrangement of the fgf8 locus after the teleost/tetrapod split

The complex expression pattern of fibroblast growth factor 8 (Fgf8) and the cellular responses dependent on concentration of its mRNA in vertebrates suggest that Fgf8 should be tightly controlled at the transcriptional level. We found zebrafish conserved noncoding elements (CNEs) with pan-vertebrate as well as fish-specific orthologous sequences from across 200 kb of the zebrafish fgf8a genomic regulatory block to direct reporter expression in patterns consistent with the expression pattern of fgf8a. These included elements from inside the introns of the skin-specific slc2a15a and the ubiquitously expressed fbxw4 bystander genes. The fgf8a/fbxw4 gene pair, which has remained joined throughout three whole genome duplications in chordate evolution, is inverted in teleost genomes, but CNEs across both evolutionary breakpoints showed specific activity. While some CNEs directed highly reproducible expression patterns, others were subject to variation but showed, in a subset of transgenes, expression in the apical ectodermal ridge, the anterior boundaries of somites and the midbrain-hindbrain boundary, specific Fgf8 signaling domains, suggesting that their activity may be context specific. A human element with tetrapod-specific orthologous sequences directed reporter expression to the vasculature, possibly corresponding to a tetrapod innovation. We conclude that fgf8a transcriptional regulation employs pan-vertebrate and teleost-specific enhancers dispersed over three genes in the zebrafish genome.

Autoregulatory loop and retinoic acid repression regulate pou2/pou5f1 gene expression in the zebrafish embryonic brain

Zebrafish pou5f1, also known as pou2, encodes a POU-family transcription factor that is transiently expressed in the prospective midbrain and anterior hindbrain during gastrulation, governing brain development. In the present study, we found that the main regulatory elements reside in the proximal upstream DNA sequence from -2.2 to -0.12 kb (the -2.2/-0.1 region). The electrophoretic gel mobility shift assay (EMSA) revealed four functional octamer sequences that can associate with zebrafish Pou2/Pou5f1. The expression of mutated reporter constructs, as well as EMSA, suggested that these four octamer sequences operate in a cooperative manner to drive expression in the mid/hindbrain. We also identified a retinoic acid (RA)-responsive element in this proximal region, which was required to repress transcription in the posterior part of the embryo. These data provide a scheme wherein pou2/pou5f1 expression in zebrafish embryos is regulated by both an autoregulatory loop and repression by RA emanating from the posterior mesoderm.