In Vivo Imaging of Hedgehog Pathway Activation with a Nuclear Fluorescent Reporter

The Hedgehog (Hh) pathway is essential for embryonic development and tissue regeneration, and its dysregulation can lead to birth defects and tumorigenesis. Understanding how this signaling mechanism contributes to these processes would benefit from an ability to visualize Hedgehog pathway activity in live organisms, in real time, and with single-cell resolution. We report here the generation of transgenic zebrafish lines that express nuclear-localized mCherry fluorescent protein in a Gli transcription factor-dependent manner. As demonstrated by chemical and genetic perturbations, these lines faithfully report Hedgehog pathway state in individual cells and with high detection sensitivity. They will be valuable tools for studying dynamic Gli-dependent processes in vertebrates and for identifying new chemical and genetic regulators of the Hh pathway.


Introduction
Zebrafish have emerged as versatile models of vertebrate biology, due to their amenability to genetic and pharmacological manipulations, optical transparency during embryogenesis and larval development, and facile and economical husbandry [1][2][3][4]. They have been used extensively to investigate the molecular and cellular mechanisms that contribute to tissue patterning [5] and more recently have contributed to our understanding of tissue regeneration, tumorigenesis, metabolism, infectious disease, and behavior [6][7][8][9][10][11]. As the importance of teleost models in biomedical research continues to grow, transgenic lines that can provide real-time indicators of specific biological events will be increasingly valuable. Accordingly, zebrafish can be readily engineered to express fluorescent reporters in selected tissues or upon the activation of individual cellular pathways [4].
In vivo reporters of developmental pathways are particularly instrumental, given the pervasive role of these signaling mechanisms in vertebrate biology. Several zebrafish lines that allow the real-time observation of cellular responses to Wnt, Hedgehog (Hh), bone morphogenetic protein (BMP), and fibroblast growth factor (FGF) family members have been described [12][13][14][15][16][17][18][19]. Each of these zebrafish lines utilizes a fluorescent protein reporter driven by cisregulatory elements specific to the signaling pathway of interest. While these transgenic animals can reveal tissue-specific differences in cell signaling, the biological properties of their reporters can limit their utility. All but one of these lines utilizes fluorescent reporters without subcellular targeting motifs, and the cytoplasmic distribution of these proteins decreases their detection sensitivity. As a result, the reporters typically cannot be observed by fluorescence microscopy until several hours after their transcription commences, especially if destabilized versions are utilized to improve reporter dynamics. Second, although cell-wide dispersion of the fluorescence signal can provide useful, collateral information on cell morphology, it can also obscure differences in pathway activity among neighboring cells.
The ability of zebrafish with Gli-dependent enhanced green fluorescent protein (EGFP) [15], mCherry [16], or Kaede reporters [17] to emulate endogenous Hh pathway activity illustrates the utility and limitations of these transgenic lines. Zebrafish have four Hh isoforms, which have been classified as Sonic Hh (Shha and Shhb) and Indian Hh (Ihha and Ihhb) family members [20][21][22]. During the first 24 hours post fertilization (hpf), shha, shhb, and ihhb are expressed in distinct but overlapping domains, with transcripts initially detected within the dorsal mesoderm shortly after the onset of gastrulation (60% epiboly, 7 hpf). Hh ligand expression is restricted to the axial mesoderm as convergent extension proceeds, and these morphogens are produced by the developing notochord, medial floor plate, and ventral floor of the brain during somitogenesis. Zebrafish cells respond to Hh proteins through the Patched family of 12transmembrane receptors (Ptch1 and Ptch2) [23,24], leading to the activation of Smoothened (Smo) [25,26], a G protein-coupled receptor-like component, and Gli transcription factors (Gli1, Gli2a, Gli2b, and Gli3) [27][28][29][30]. The degree and duration of Hh pathway activation in these cells then regulates their differentiation. For example, somitic tissues form slow-twitch muscle fibers and muscle pioneer cells in response to moderate and high levels of notochord-derived Hh signals, respectively [31]. In contrast to these endogenous processes, the previously described EGFP and mCherry reporter lines do not exhibit Gli-dependent fluorescence until mid-somitogenesis (approximately 17 hpf) [15,16]. Nor can they clearly resolve the differences in Hh pathway activity that give rise to distinct muscle cell types during somite development. The photoconvertible Kaede line can be used to distinguish cell populations with temporally distinct Hh responses; however, the cytoplasmic reporter makes it challenging to differentiate cells with similar Hh signaling dynamics [17].
Targeting fluorescent reporter signals to the nucleus can help overcome these limitations, providing transgenic lines with improved signal-to-noise properties and single-cell resolution. In vivo reporters of Wnt signaling have been enhanced in this manner [14], and we therefore sought to establish new zebrafish lines that carry fluorescent, nuclear-localized reporters of Hh pathway state. Such genetically modified organisms could reveal how Hh signaling dynamically regulates formation of the brain, neural tube, somites, fins, and other tissues during zebrafish embryogenesis. Hh pathway activation also contributes to the regeneration of body parts such as a caudal fin [32,33], and its dysregulation can promote the onset and/or maintenance of certain cancers [34].
We have generated two Hh pathway reporter lines: one that expresses mCherry functionalized with a nuclear localization sequence (NLS) upon Gli activation and another that carries a destabilized form of the mCherry-NLS reporter. Gli-dependent mCherry fluorescence is clearly evident in these transgenic animals by the one-somite stage (11 hpf), and divergent responses to Hh signaling can be observed in these transgenic lines with single-cell resolution. Genetic and chemical perturbations further confirm that the fluorescent signals in these lines faithfully communicate Hh pathway state, and the reporter is functional in both embryos and adult fish. We anticipate that these transgenic lines will be valuable tools for studying Hh pathway-dependent development and physiology. They could also facilitate the discovery of new signaling components and small-molecule modulators, taking advantage of the zebrafish's amenability to genetic and chemical screens.

Ethics Statement
All experiments involving zebrafish were approved by the Institutional Animal Care and Use Committee at Stanford University (Protocol ID: 10511).
The 8xGliBS-IVS2-mCherry-NLS-polyA-Tol2 and 8xGliBS-IVS2-mCherry-NLS-Odc1-polyA-Tol2 plasmids were maxiprepped, treated with proteinase K (10 mg DNA, 1 mg/mL proteinase K, and 0.5% SDS in a 50-mL reaction at 55uC for 30 minutes), subjected to a phenol/chloroform extraction (aqueous volume increased to 150 mL with water, extracted twice with 150 mL 25:24:1 neutral-buffered phenol:chloroform:isoamyl alcohol (Invitrogen), and then twice again with pure chloroform), and then precipitated with an equal volume of isopropanol after adjusting the solution pH to 4.5 with 0.3 M sodium acetate. Zebrafish zygotes were then co-injected with this RNAse-free plasmid DNA and Tol2 transposase mRNA transcribed from pCS-TP [38] (kindly provided by Dr. Koichi Kawakami) using the Ambion SP6 mMessage mMachine kit (25 pg of each/embryo; 2-nL injection volume). The resulting adults were mated with wildtype AB fish to identify founders with germline transmission of the Hh pathway reporter, and those yielding F1 and F2 generations with robust, monoallelic reporter expression were used to establish transgenic colonies. Heterozygote Tg(8xGliBS:m-Cherry-NLS) and Tg(8xGliBS:mCherry-NLS-Odc1) lines in wildtype and mutant backgrounds were used in subsequent studies.

Tail amputations
Adult caudal fins were amputated by first anesthetizing the zebrafish with 0.2 mg/mL tricaine in fish-system water for approximately 5 minutes. The fish were next placed onto a clean paper towel, and their tails were snipped using spring scissors (Fine Science Tools, Cat. No. 91501-09). The fish were then allowed to recover in system water.

Transient genetic and molecular perturbations
Zebrafish shha mRNA was transcribed from a pSP64TS-shha plasmid [41] using the SP6 mMessage mMachine kit and then injected into zebrafish zygotes (150 pg/embryo). Cyclopamine (Infinity Pharmaceuticals) was dissolved in ethanol and applied to embryos at a final concentration of 100 mM, beginning at the 1cell or 10-somite stages.

Zebrafish imaging
Brightfield and fluorescence images were acquired using either a Leica DM4500B compound microscope equipped with a Retiga SRV camera or a Leica MZFLIII stereomicroscope with a Leica DFC480 camera. GFP and mCherry fluorescence was visualized using GFP and Texas Red filtersets, respectively. For live-imaging studies, zebrafish embryos were manually dechorionated, anesthetized in E3 medium containing 0.05% (w/v) tricaine mesylate, and then placed in agarose wells. Adult fish were similarly anesthetized but imaged on a dark surface.

Generation of zebrafish with nuclear Hh pathway reporters
To visualize in vivo Hh signaling with high detection sensitivity and single-cell resolution, we sought to generate transgenic zebrafish carrying Gli-dependent, nuclear-localized fluorescent reporters. Hh pathway-driven expression of exogenous genes in cultured cells or live organisms has been previously achieved using a minimal d-crystallin promoter and eight tandem Gli binding sites derived from the murine Fox2A floor plate enhancer [39]. Therefore we prepared reporter constructs that coupled these regulatory elements with sequences encoding either mCherry-NLS alone or the nuclear-localized fluorescent protein tagged with an ornithine decarboxylase-derived destabilizing peptide [42] (mCherry-NLS-Odc1) ( Figure 1A). We reasoned that the fast maturation kinetics [40] and discrete subcellular localization of mCherry-NLS reporter would allow the rapid detection of Hh pathway activation in individual cells; addition of the destabilizing domain would enhance reporter turnover and help reveal temporal changes in Gli function. Both constructs were individ- ually cloned into a vector with flanking Tol2 transposase recognition elements, and the resulting plasmids were injected into zebrafish zygotes with Tol2 mRNA.
Approximately 10-20% of adult zebrafish subjected to this transgenic procedure exhibited germline integration of the Glidependent reporter, and these F0 founders (both male and female) typically gave rise to F1 carriers with a transmission efficiency of 5%. For each transgenic line, three founders were chosen for further analysis, and we identified those yielding progeny with robust, Hh pathway-dependent reporter expression that segregated as a single insertion allele in the F2 generation. Based on these criteria, we maintained one line each for the Tg(8xGliBS:m-Cherry-NLS) and Tg(8xGliBS:mCherry-NLS-Odc1) transgenes.
Both reporter lines exhibited prominent mCherry fluorescence in the somites, ventral brain and neural tube, and other Hhresponsive tissues ( Figure 1B-C). We observed some differences in reporter activity within the brain of larval-stage animals (84 hpf), likely due to the presence or absence of the Odc1-derived destabilizing element. Although cerebellar and telencephalic tissues exhibited reporter expression in the two transgenic lines, mCherry fluorescence was undetectable in the hindbrain and optic tectum of Tg(8xGliBS:mCherry-NLS-Odc1) larvae. These observations suggest that Hh morphogen-dependent patterning of these neural domains is largely complete by 84 hpf, where its actions within the cerebellum and telecephalon are ongoing. Differences between the two reporter lines could also be discerned at earlier stages upon addition of the Smo antagonist cyclopamine [43]. While mCherry fluorescence persisted in 30-hpf Tg(8xGliBS:m-Cherry-NLS) zebrafish previously exposed to cyclopamine for several hours, similarly treated Tg(8xGliBS:mCherry-NLS-Odc1) embryos lost mCherry signals in the ventroanterior brain, ventral spinal cord and posterior somites ( Figure 2). Thus, the Odc1derived destabilization domain increases Hh reporter turnover.

Nuclear Hh pathway reporters convey cellular activity with improved sensitivity
Due to the enhanced temporal resolution afforded by the mCherry-NLS-Odc1 reporter, we focused on this transgenic line for further analysis of embryonic Hh pathway activity. Nuclear mCherry fluorescence could be first detected by the onset of somitogenesis (11 hpf), several hours earlier than in previous Hh reporter lines [15,16]. Mesodermal adaxial cells and overlying neural plate cells in the midline exhibited reporter expression, which progressively increased during somite formation and differentiation ( Figure 3A-D and Movie S1). Upon the completion of somitogenesis at 24 hpf, varying levels of mCherry fluorescence could be observed within myotome. Mononucleate fibers near the horizontal myoseptum exhibited the highest nuclear signals, and  more distally positioned mononucleate fibers had mid-level reporter expression ( Figure 3E). Multinucleated cells within the somitic mass did not express detectable levels of nuclear mCherry. These differences coincide with the locations of muscle pioneer cells and superficial slow-twitch muscle fibers, which respectively achieve high and intermediate thresholds of Hh pathway activity in response to notochord-derived morphogens [31]. Indeed, the lateral migration of mCherry-positive slow-twitch muscle precursors within each somite can be visualized through time-lapse videomicroscopy of Tg(8xGliBS:mCherry-NLS-Odc1) embryos (Movie S2). In contrast, the multinucleate fast-twitch muscle fibers that comprise the myotome bulk are not known to be Hh ligandresponsive [31]. Thus, physiologically important variations in Hh signaling can be discerned with single-cell resolution in the Tg(8xGliBS:mCherry-NLS-Odc1) line.

Nuclear Hh pathway reporters respond to chemical and genetic perturbations
We next sought to confirm the fidelity of the Tg(8xGliBS:m-Cherry-NLS-Odc1) reporter by assessing its responsiveness to chemically and genetically induced changes in Hh pathway state. We first treated Tg(8xGliBS:mCherry-NLS-Odc1) zygotes with cyclopamine. Consistent with the resulting Hh loss-of-function phenotypes, including ventral body curvature and U-shaped somites [25,26], mCherry fluorescence was undetectable in these fish ( Figure 3F-G). We similarly observed that overexpression of Shha through mRNA injection caused strong mCherry upregulation throughout the embryo ( Figure 3H).
As a third method to validate reporter fidelity, we analyzed Tg(8xGliBS:mCherry-NLS-Odc1) reporter output in mutant backgrounds with varying degrees of Hh pathway dysregulation, including the loss-of-function mutants shha t4 [35], smo hi1640 [26], and gli1 ts269 [36] (Figure 3I-L). We also assessed reporter activity in dzip1 tm79a mutants [36] (Figure 3M), which exhibit decreased Hh target gene expression in the ventral brain and neural tube but increased pathway activation in somitic tissues [44,45]. Consistent with the morphological and molecular phenotypes associated with these mutant alleles, we observed tissue-specific changes in reporter expression. Embryos lacking Smo function resembled cyclopamine-treated fish and were essentially devoid of detectable mCherry fluorescence. In comparison, shha t4 and gli1 ts269 mutants lacked mCherry signals within the myotome but maintained a reduced reporter activity in anterior tissues, and dzip1 tm79a mutants upregulated and downregulated Hh reporter expression within the somites and ventral brain, respectively. Taken together, our observations demonstrate that the Glidependent nuclear reporters are sensitive, specific indicators of Hh pathway state.

Nuclear Hh pathway reporters function in adult zebrafish
We concluded our studies by investigating the functionality of our Hh pathway reporters in adult zebrafish. Caudal fin regeneration is a convenient model of Hh pathway-dependent physiology; previous studies have established clear domains of Hh ligand production and response during this process [32], and fin structures are particularly amenable to live imaging. After fin amputation, epithelial cells rapidly cover the wound and a blastema of de-differentiated cells forms within each bony ray segment [46,47]. The blastema then proliferates and gives rise to daughter cells that reconstitute the fin in an epimorphic manner. During this regenerative process, shha is initially expressed around 30 hours post amputation (hpa) in a subset of basal epidermal cells at distal end of each ray [32]. Shha-responsive ptch2-positive cells can subsequently be detected at 40 hpa in the epidermis, first overlapping with the distal shha-positive domain and then extending proximally [32]. These signaling events regulate dermal bone patterning as the ray segments reform from blastema-derived cells.
To simultaneously visualize shha-producing cells and their ptch2-expressing responders, we crossed Tg(-2.4shha:GFP-ABC) [37] and Tg(8xGliBS:mCherry-NLS) zebrafish to obtain progeny carrying both fluorescent reporters. Shha-expressing cells in these transgenic organisms are labeled with GFP, and the corresponding cellular responses can be assessed by mCherry fluorescence. We cut the caudal fins of adult Tg(-2.4shha:GFP-ABC;8xGliBS:m-Cherry-NLS) zebrafish and monitored changes in GFP and mCherry reporter expression at 3 days post amputation. As expected, amputation caused a dramatic increase in GFP expression at the distal end of each fin ray ( Figure 4A-B), and mCherry-expressing cells populated interray regions immediately proximal to the GFP-positive domains ( Figure 4C).

Conclusion
Taken together, our results demonstrate the utility of nuclearlocalized mCherry reporters of Hh pathway activity, particularly when used in conjunction with optically transparent model organisms such as the zebrafish. The Tg(8xGliBS:mCherry-NLS) and Tg(8xGliBS:mCherry-NLS-Odc1) fish described here have certain advantages over previous Hh pathway reporter lines, including: (1) higher detection sensitivity that enables real-time observation of Hh signaling by the onset of somitogenesis; and (2) greater cellular resolution that allows pathway activity differences between neighboring cells to be discerned. We anticipate that these transgenic organisms will be valuable tools for studying the Hh pathway-dependent processes that contribute to embryonic development, tissue homeostasis, and tumorigenesis. They could also facilitate the identification of new genetic and chemical modulators of Hh signal transduction through high-content, image-based screens.

Supporting Information
Movie S1 Real-time imaging of Hh signaling during early somitogenesis. Nuclear mCherry expression in a Tg(8xGliBS:mCherry-NLS-Odc1) zebrafish between 11 and 15 hpf. The midline region between the future sixth and ninth somites was imaged at a rate of 1 frame/minute, and the movie is shown at rate of 30 frames/second. Embryo orientation: dorsal view and anterior up. Field of view: 180 mm6180 mm.

(MOV)
Movie S2 Real-time imaging of slow-twitch muscle precursor migration. Nuclear mCherry expression in a Tg(8xGliBS:mCherry-NLS-Odc1) zebrafish between 13 and 17 hpf. The midline region between the second and seventh somites was imaged at a rate of 1 frame/minute, and the movie is shown at rate of 30 frames/second. Embryo orientation: dorsal view and anterior up. Field of view: 260 mm6260 mm. (MOV)