The manuscript here provides a simple set of methods for analysing the secretion and diffusion of fluorescently tagged ligands in Xenopus. This provides a context for testing the ability of other proteins to modify ligand distribution and allowing experiments that may give insight into mechanisms regulating morphogen gradients.
This protocol describes a method to visualise ligands distributed across a field of cells. The ease of expressing exogenous proteins, together with the large size of their cells in early embryos, make Xenopus laevis a useful model for visualising GFP-tagged ligands. Synthetic mRNAs are efficiently translated after injection into early stage Xenopus embryos, and injections can be targeted to a single cell. When combined with a lineage tracer such as membrane tethered RFP, the injected cell (and its descendants) that are producing the overexpressed protein can easily be followed. This protocol describes a method for the production of fluorescently tagged Wnt and Shh ligands from injected mRNA. The methods involve the micro dissection of ectodermal explants (animal caps) and the analysis of ligand diffusion in multiple samples. By using confocal imaging, information about ligand secretion and diffusion over a field of cells can be obtained. Statistical analyses of confocal images provide quantitative data on the shape of ligand gradients. These methods may be useful to researchers who want to test the effects of factors that may regulate the shape of morphogen gradients.
During early embryonic development, cells are progressively committed to follow specific lineages of differentiation: this means a group of totipotent (or pluripotent) cells become gradually restricted to establishing populations of progenitor cells determined to give rise to one cell type. Cell-cell signalling is central to the regulation of lineage specification during embryonic development. Manipulation of these signals will be required to direct stem cells toward particular fates to support novel medical treatments.
A relatively small number of signalling pathways are reiterated during development, including pathways responding to the TGF superfamily (nodals and BMPs)1-2, FGFs3, Wnts4, and Hedgehogs5. These secreted proteins bind receptors present on the cell membrane to activate signal transduction thereby altering gene expression and/or cell behaviour. The tight regulation of cell signalling is essential for cell lineage specification and normal development. While the cross-talk among these pathways is important in determining cell fate, a single ligand can itself elicit distinct responses at different concentrations. Morphogen gradients were described over 100 years ago as a theory to explain how different cell types can derive from a field of cells6. Signalling molecules produced by one group of cells may diffuse over a certain range, decreasing in concentration with a greater distance from the source. Cells exposed to the signal will respond to the local concentration at their position in the field of cells, with cells at distinct positions responding differently to different levels of the signal. Evidence for the existence of morphogens comes from studies of the early Drosophila embryo7 and the wing disc8, as well as the vertebrate limb9 and neural tube10.
Methods are needed to investigate how morphogen gradients are established and to identify other molecules important in regulating these gradients. Elegant experiments using immunohistochemistry to visualise endogenous proteins in vivo in the context of different genetic backgrounds have been used to investigate morphogen gradients11-12. However, good antibodies and specific mutants are not always available, so we describe here a protocol using overexpression of fluorescent ligands in Xenopus, to provide an alternative, simple method to dissect how exogenous gene products can influence distribution of ligands across a field of cells. Xenopus laevis provides an excellent system to undertake these types of experiments as their embryos develop externally so they are accessible at the earliest stages. Their large size (1-1.5mm in diameter) simplifies microinjection and surgical manipulation and by blastula stages the cells are easy to image as they are still relatively big (about 20µm across). Overexpression studies in Xenopus are simple to do: mRNA injected into the early embryo can be targeted to particular cells and is efficiently translated.
The fluorescently tagged Wnt8a/Wnt11b-HA-eGFP constructs were generated using pCS2 Wnt8a-HA13, pCS2 Wnt11b-HA14 and eGFP. The HA peptide is important to include, not only to provide an additional molecular tag, but also because it is thought to act as a spacer separating the Wnt and eGFP proteins allowing both gene products to function. The construct used for the visualisation of Shh was previously used to generate a transgenic mouse expressing a Shh-eGFP fusion protein15; this was kindly provided by Andy McMahon. Importantly, the GFP tag for all the constructs is cloned 3’ to the signal sequence such that it is retained after processing. It is also essential to ensure that the final protein includes sequences required for modifications, such the addition of lipids as is the case for Shh and Wnt ligands.The cDNAs were subcloned into the pCS2+ expression vector which is optimised for the prodution of synthetic mRNA; it includes an SP6 promoter and polyadenylation signal (http://sitemaker.umich.edu/dlturner.vectors).
The work described here provides a simple protocol for comparing the secretion and diffusion of fluorescently tagged Wnt and Shh tagged ligands. By injecting defined amounts of synthetic mRNA, the protocol circumnavigates any problems associated with variable expression from different vectors using different promoters. These methods have recently been applied to investigate the effects of the heparan sulfate endosulfatase Sulf1 on Shh-eGFP and Wnt8a/Wnt11b-HA-eGFP secretion and diffusion in Xenopus16-17.
Ethics statement: Animal experiments were done under a UK Home office licence to MEP and the experiments carried out was approved by the University of York ethics committee in compliance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. (https://www.nc3rs.org.uk/arrive-guidelines).
1. Strategy to Generate Fluorescently Tagged Ligands
Below is an example protocol for subcloning Wnt8a-HA and eGFP into pCS2, in order to produce the in frame fusion construct Wnta-HA-eGFP:
2. mRNA Synthesis: Generating the Template
3. mRNA Synthesis: In Vitro Transcription
4. Generating Xenopus laevis Embryos
5. Microinjecting Xenopus Embryos
6. Excising Animal Cap Explants
7. Imaging
Note: Imaging was carried out using an inverted confocal microscope. Lambda mode was chosen for imaging as this allowed multiple fluorophores with overlapping signatures to be used, and removed the problem of potential sample movement between scans.
8. Image Analysis
Confocal analysis of animal cap explants expressing fluorescently tagged proteins provides an effective system for visualising ligand distribution under different experimental conditions. In one example, the distribution of GFP tagged Shh is shown (Figure 1). At the 2-cell stage, Xenopus embryos are injected into both cells with either with a control mRNA or with mRNA coding for Sulf1, an enzyme that modifies cell surface heparan sulfate and influences the Shh morphogen gradient16. These embryos are cultured until the 32-cell stage and a single cell is co-injected with mRNAs coding for Shh-GFP and memRFP (as a lineage tracer). This creates a clone of cells expressing Shh-GFP that is marked with the cell membrane tethered RFP. At the blastula stage, animal cap explants are taken for analysis by confocal microscopy. Figure 1 shows that under control conditions, Shh-GFP is secreted and diffuses outside of the region expressing the injected mRNAs. However, in the presence of Sulf1, Shh-GFP is more restricted in its distribution and, in this sample, is not detected outside the clone of cells producing it. The effects of Sulf1 on Shh-GFP distribution has been analysed more fully16.
When expressed in Xenopus embryos, fluorescently tagged Wnt ligands are secreted, accumulate on the cell membrane, and diffuse across the field of cells. In Figure 2, we characterise the quantitaive and qualitative properties of the two different fluorecently tagged Wnt ligands in cells injected and expressing the proteins. Figure 2A-H shows examples of how Wnt8a and Wnt11b-HA-eGFP accumulate on the membrane of animal cap cells. By comparing panels 2B and 2F it is clear that Wnt8a-HA-eGFP accumulates more efficiently on the cell membrane than Wnt11b-HA-eGFP. Quantitative information has been extracted from the confocal images using a combination of image and script analysis software (supplemental code file). Figure 2I shows the relative accumulation of Wnt8a-HA-eGFP on the cell membrane compared to Wnt11b-HA-eGFP. The data was obtained using a computer script that determines the total number of Wnt-HA-eGFP pixels co-localising with cell membrane pixels in each image. In another approach, the total number of Wnt-HA-eGFP puncta that co-localise with the cell membrane are counted using image analysis software (Figure 2J). Qualitative information about the size and shape of individual puncta can also be extracted from the data using image analysis software. In addition to fewer Wnt11b-HA-eGFP puncta co-localising with the cell membrane (Figure 2J) these puncta also have a smaller average size than Wnt8a-HA-eGFP puncta (Figure 2K). Moreover, Wnt11b-HA-eGFP puncta have a reduced circularity compared to Wnt8a-HA-eGFP puncta (Figure 2L). Circularity is a measure of how closely an object resembles a perfect circle, with 1 representing a perfect circle and 0.1 an elongated non-circular shape. This approach can be used to test any candidate regulator of Wnt ligand diffusion. To do this, control cells should be injected with a control mRNA coding for an inactive form of the candidate regulator or an irrelevant protein such as beta-galactosidase (lacZ); this provides a more valid control than simply not injecting the regulator. Data in these examples was analysed using the non-parametric test Mann-Whitney U18-19. A statistical analysis programme was used to perform the test.
In Figure 3, we investigate Wnt ligand diffusion. A single blastomere was injected at the 4-cell stage such that the animal cap explants represent a field of cells in which only some cells express either Wnt8a-HA-eGFP or Wnt11b-HA-eGFP. By including a cell lineage tracer, we can identify expressing versus non-expressing cells and this allows the range of Wnt8a-HA-eGFP and Wnt11b-HA-eGFP ligand diffusion away from the source cells to be analysed (Figure 3B and 3C). Due to the curvature of the animal cap explants, the maximum distance that could be reliably measured using this assay was 160µm, which was not enough to measure the absolute distance of ligand diffusion. However, by using a combination of confocal analysis, image analysis and scientific analysis and graphing software the overall shape of the Wnt-HA-eGFP morphogen gradient could be analysed (Figure 3D). This data can be used to investigate whether overexpressing another protein together with the tagged ligands in the same cells can affect Wnt secretion or diffusion. In another type of experiment (Figure 3E-3J) the effects of a potential regulator in receiving cells can be measured by overexpressing the candidate protein in clones of cells adjacent to cells expressing Wnt8a or Wnt11b-HA-eGFP. This allows any non-cell-autonomous effects of the regulator on Wnt-HA-eGFP ligand diffusion to be examined. Consistent with Figure 2, more Wnt8a-HA-eGFP can be detected diffusing away from cells than Wnt11b-HA-eGFP in both diffusion experiments.
The types of experiments described in this paper have been used to analyse the effect of Sulf1 on Shh and Wnt signalling16,17.
Figure 1. Modulation of Shh-GFP distribution can be detected by confocal analysis of Xenopus animal caps. (A) A cartoon depicting experimental assay where embryos are first injected with mRNA coding for Sulf1 or a control mRNA, the same embryos are later injected at the 32-cell stage with mRNA coding for Shh-GFP into a single cell. (B-C) Animal cap explants were taken at stage 8 and imaged after 4 hr. Images are shown at the edge of a Shh-GFP + memRFP expressing clone of cells. In control embryos (B), Shh-GFP is distributed away from the memRFP marked clone of signal producing cells. In embryos expressing Sulf1 (C), Shh-GFP is more restricted in its distribution, see16. memRFP is shown in magenta and Shh-GFP in green. Scale bars represent 20 µm. Please click here to view a larger version of this figure.
Figure 2. Quantitative and qualitative analysis of Wnt8a and Wnt11b-HA-eGFP puncta on the cell membrane. (A-H) Embryos were microinjected bi-laterally with mRNA encoding memRFP (500 pg) into the animal hemisphere at the two cell stage. In addition embryos were injected with mRNA encoding (A-D) Wnt8a-HA-eGFP (500 pg) or [E-H] Wnt11b-HA-eGFP (1ng). The embryos were also injected with mRNA coding for LacZ to provide a control for additional mRNA/protein when analysing the effects of any potential regulator. The white boxes in (C) and (G) mark the areas enlarged in panels (D) and (H) respectively. The total amount Wnt-HA-eGFP fluorescence co-localising with the cell membrane was calculated using image and script analysis software and then normalised (I). Qualitative information was extracted using image analysis software. Wnt8a and Wnt11b-HA-eGFP punctae were analysed for particle number (J), particle size (K) and particle circularity (L). Mann-Whitney U (**P<0.01), N=number of embryos. memRFP is shown in magenta, and Wnt8a/11b-HA-GFP is shown in green, scale bars represent 20µm. Please click here to view a larger version of this figure.
Figure 3. Measuring the range of diffusion of fluorescently tagged Wnt ligands. (A) Diagram depicting the assay used to measure Wnt8a and Wnt11b-HA-eGFP secretion and diffusion away from expressing cells. (B-C) mRNA encoding (B) memCerulean (600 pg), LacZ (4ng) and Wnt8a-HA-eGFP (2ng) or (C) memCerulean (600 pg), LacZ (4ng) and Wnt11b-HA-eGFP (2 ng) was injected into the animal hemisphere of one blastomere at the four cell stage. (D) The range of diffusion of Wnt8a-HA-eGFP and Wnt11b-HA-eGFP was measured through a control background. (E) Diagram depicting the assay used to measure Wnt8a and Wnt11b-HA-eGFP diffusion through a background expressing LacZ, see method for details. (F-G) mRNA encoding memCerulean (600 pg) and (F and H) Wnt8a-HA-eGFP (2 ng) or (G and I) Wnt11b-HA-eGFP was injected into the animal hemisphere of one blastomere at the four cell stage. An adjacent blastomere was injected with mRNA encoding memRFP (600 pg) and LacZ (4 ng) see Key for details. (J) The range of Wnt8a-HA-eGFP and Wnt11b-HA-eGFP diffusion was measured through a background expressing LacZ. The data was quantified and plotted using confocal analysis, image analysis and scientific analysis and graphing software. memCerulean (blue (C-D) and Yellow (F and H)), Wnt-HA-eGFP (green), memRFP (magenta), scale bars represent 20 µm. Please click here to download a larger version of this file.
Table 1. Primers used to subclone Wnt8a/Wnt11b-HA-eGFP and Shh-GFP into pCS2.
Reaction | Components | |
Example CS2+ digest | 1.5 μg of pCS2+ | |
2 μl of Restriction enzyme 1 | ||
2 μl of Restriction enzyme 2 | ||
5 μl of Restriction enzyme buffer (10X) | ||
Made up to 50 μl with molecular grade water | ||
Example mRNA synthesis | 2 μl linearised template | |
2 μl 10x Megascript trx mix | ||
2 μl 50mM ATP | ||
2 μl 50mM CTP | ||
2 μl 50mM UTP | ||
2 μl 5mM GTP | ||
2.5 μl 40mM Cap Analog (m7G(5') | ||
2 μl SP6 enzyme mix | ||
3.5 μl Molecular grade water | ||
Example PCR reaction | 0.5 μl High fidelity DNA polymerase (2,000 units/ml) | |
2 ng Template DNA | ||
2.5 μl of forward and reverse primers (10μM) | ||
0.5 μl of dNTPs | ||
5 μl of DNA ploymerase buffer (10X) | ||
Made up to 50 μl with molecular grade water | ||
Example PCR conditions | Intial denaturation 2 min at 98 °C | |
15 sec 98 °C | ||
15 sec 65 °C | 30 cycles | |
40 sec 72 °C | ||
Final extension 10 min at 72 °C | ||
Example PCR product digest | 28 μl of PCR product | |
2 μl of Restriction enzyme 1 | ||
2 μl of Restriction enzyme 2 | ||
5 μl of Restriction enzyme buffer (10X) | ||
13 μl Molecular grade water | ||
Example T4 ligation | 1 μl Cut CS2+ | |
3 μl Cut PCR product 1 | ||
3 μl Cut PCR product 2 | ||
1 μl of T4 DNA ligase | ||
1 μl T4 DNA ligase buffer (10X) | ||
1 μl Molecular grade water | ||
Example template digest | 5 ug Plasmid DNA | |
10 μl Restriction enzyme buffer (10X) | ||
3 μl Not1 (except Shh-GFP, Kpn1 is used) | ||
Made up to 100 μl with molecular grade water |
Table 2. Example reaction conditions used for subcloning and producing synthetic mRNA for Wnt8a-HA-eGFP.
Solution | Components |
Cysteine-HCL | 0.1X NAM |
2.5% L-cysteine hydrochloride monohydrate (pH7.8) | |
NAM salts | 110 mM NaCL |
2 mM KCl | |
1 mM CA(NO3)2 | |
0.1 mM EDTA | |
NAM/2 | 0.5X NAM salts |
5 mM HEPES pH7.4 | |
0.25 mM Bicarbonate | |
25 ug/ml Gentamycin | |
NAM/3 + Ficoll | 0.33X NAM salts |
5 mM HEPES pH7.4 | |
0.25 mM Bicarbonate | |
2 5ug/ml Gentamycin | |
5 % Ficoll | |
NAM/10 | 0.1X NAM salts |
5 mM HEPES pH7.4 | |
25 ug/ml Gentamycin |
Table 3. Solutions used during the production and microinjection of Xenopus laevis embryos.
An essential part of this protocol is generating biologically active ligands that are normally processed, secreted, and able to elicit a response in the receiving cell, despite having a fluorescent moiety attached. It is critical to establish that the fluorescently-tagged gene product is biologically active using an appropriate assay. For Shh-GFP, the ability to activate the expression of ptc1 was confirmed16. Wnt8a has a remarkably potent ability to induce a secondary axis when expressed in a single ventral blastomere20-21, and Wnt8a-HA-eGFP was shown retain this biological activity. Wnt11b inhibits activin treated ectodermal explants from undergoing convergent extension21-22, and Wnt11b-HA-eGFP also retains its biological activity17. The ability of the tagged ligands to elicit responses equivalent to the untagged proteins suggests that the conclusions based on experiments using the fluorescent ligands will be relevant to the normal protein. The addition of the HA epitope between the ligand and the fluorescent protein provided a spacer region that allows the Wnt constructs to be both active and fluorescent.
The major limitation of this type of analysis is that it does not directly inform on endogenous morphogen gradients. For instance, the diffusion of Shh across the field of cells in animal cap may not reflect the way endogenous Shh moves through the columnar neural epithelium in the embryo. However, the simplicity of this protocol allows experiments where the effect of exogenous regulators on the distribution of ligands can be tested. The results from these approaches can form the basis of hypotheses to be tested using more demanding in vivo analyses. For instance, the ability of over-expressed Sulf1 to restrict the distribution of Shh-GFP using the methods described in this paper pointed to the potential for Sulf1 in modulating the Shh morphogen gradient. This was directly tested and validated using antisense morpholino knock-down of Sulf1 and immunocytochemistry to visualise endogenous Shh in the neural tube16. A similar method to ours has been used to investigate specific mechanisms that could regulate ligand diffusion. Smith and colleagues demonstrated that a GFP-tagged nodal (Xnr2) used simple diffusion, and not cell extensions (cytonemes) or transcytosis (using vesicles) to form a gradient23.
Further elaborations of these methods are possible where other proteins that block particular pathways can be expressed in targeted cells to investigate whether a response to the signalling molecule is needed as an intermediary to relay signals from one cell to another. For instance, expressing a dominant negative activin receptor between the source of activin and the responding cells showed that simple ligand diffusion could elicit a response several cell diameters away from the source even with non-responsive cells in between24. This conclusion was supported by work in zebrafish where wild type cells can respond to a source of nodal, despite being surrounded by non-responsive mutant cells lacking an essential co-receptor for nodal25. Other studies have used zebrafish to investigate diffusion of fluorescently tagged ligands28,29. Some studies has observed that epitope tagging the C-terminus of Wnt proteins can impact on signalling activity, possibly because of the highly conserved cysteines that contribute to protein conformation. Our previous work has shown that including a spacer (such as the HA tag) results in tagged Wnt ligands that are biologically active17. This is likely because the spacer provides flexibility needed for ligand –receptor interaction, such as in the thumb-forefinger model of Wnt-Frz binding30.
The next step for advancing our studies is to include a visual readout for the activation of the signalling pathway. For instance, expressing GFP-tagged Dvl26 in cells distant from the source of ligand will allow the range across which Wnt11b has ability to signal to be measured. The range of Wnt8a signalling activity could be shown using antibody staining to measure nuclear localisation of b-catenin in cells27 at a distance from the source. These types of experiments are possible using the techniques described in this paper. By employing a variety of different fluorescent proteins or fluorophores, an additional level of information would be provided where one not only measures the distribution of a ligand but the also the output of the signalling pathways, and in this way determining the threshold range of a morphogen.
Xenopus laevis is a useful model for visualising GFP-tagged ligands and further details on the methods for procuring embryos, mRNA synthesis, microinjection and dissection are available31.
The authors have nothing to disclose.
This work was funded by a BBSRC grant to MEP (BB/H010297/1), a BBSRC quota studentship to SAR, and an MRC studentship to SWF.
Agarose | Melford | MB 1200 | |
Ammonium acetate | Ambion | From Megascript SP6 Kit AM 1330 | |
Bicarbonate | VWR International | RC-091 | |
Calcium nitrate | Sigma | C13961 | |
Cap analog (m7G(5')) | Applied Biosystems | AM 8050 | |
Chloroform | Sigma | C 2432 | |
L-Cysteine hydrochloride monohydrate | Sigma | C7880 | |
dNTPs | Invitrogen | 18427-013 | |
Ethylenediaminetetraacetic acid (EDTA) | Sigma | O3690 | |
Ethanol | VWR International | 20821.33 | |
Ficoll 400 | Sigma | F 4375 | |
Fiji image J software | N/A | N/A | Free download http://fiji.sc/Fiji |
Gentamycin | Melford | G 0124 | |
Glacial acetic acid | Fisher Scientific | A/0400/PB17 | |
Glass cover slips, No.1.5 | Scientific Laboratory Supplies | 22X22-SGJ3015. 22X50-SGJ3030 | |
Glass needle puller | Narishige | Narishige PC -10 | |
Glass pull needles | Drummond Scientific | 3-000-203-G/X | |
Human chronic gonadotropin (HCG) | Intervet | ||
Isopropanol | Fisher Scientific | P/7500/PB17 | |
Lithium chloride (LiCl) | Sigma | L-7026 | |
LSM710 and Zen software (2008-2010) | Carl Zeiss | ||
Matlab software | Mathworks | http://uk.mathworks.com/ | |
Molecular grade water | Fisher Scientific | BP 2819-10 | |
Nail varnish | Boots | Bar code 3600530 373048 | |
Spectrophotometer | Lab.tech International | ND-1000 / ND8000 | |
Petri dish (55mm) | VWR International | 391-0865 | |
Phenol-chloroform | Sigma | P3803 | |
Photoshop software | Adobe | N/A | http://www.photoshop.com/products |
High fidelity DNA polymerase and buffers | Biolabs | M0530S Buffer – M0531S | |
Potassium chloride (KCl) | Fisher Scientific | P/4280/53 | |
PVC insulation tape | Onecall | SH5006MPK | |
Gel extraction kit | Qiagen | S28704 | |
Restriction enzymes buffers | Roche | SuRE/CUT Buffer Set 11082 035 001 | |
RNAse-free DNAse | Promega | ME10A | |
Steel back single edge blades | Personna | 66-0403-0000 | |
Sodium chloride (NaCl) | Fisher Scientific | 27810.364 | |
SP6 transcription kit | Ambion | AM1330 | |
Glass slides | Thermo Fisher | SHE 2505 | |
Tris base | Invitrogen | 15504-020 | |
Tungsten needles | N/A | N/A | homemade |
Zen lite software | Carl Zeiss | N/A | Free download http://www.zeiss.co.uk/microscopy/en_gb/downloads/zen.html |