Xenopus tropicalis Mutation Resource

Phenotyping Pipeline

In addition to stablish a mutant line for a requested gene by Tilling we will provide functional annotation for the genes we analyze. This will provide a jumping off point for genes which can be studied further, and will provide preliminary analysis relevant to mechanisms of organogenesis. Xenopus may in many cases provide a more advantageous system than the mouse for the study of mechanisms of development of specific organs. In particular future work will exploit experimental embryological tools and the relative accessibility of embryos at all developmental stages. In addition, due to the differences in genome structure and duplication, phenotypes may be manifest when a single gene is mutated in Xenopus, which may not be apparent if the cognate gene is duplicated or redundant in the mouse or zebrafish.

Schema of the Phenotyping Pipeline to describe the mutants identified by TILLING.

Schema of the Phenotyping Pipeline to describe the mutants identified by TILLING.


To this end we will perform three levels of phenotypic analysis on mutants with selected nonsense alleles.

  1. Firstly, we will observe mutant embryos at several defined stages of development through 5 days post fertilization (stage 45-48).
  2. Secondly, we will isolate RNA from mutants and wild-type siblings and profile and examine the expression of a set of about 2350 well characterized genes; these genes have a known tissue distribution of transcripts, or have been shown previously to change in expression during Xenopus or zebrafish development.
  3. Finally, we will stain embryos by in situ hybridizations with cocktails of probes designed to yield maximal information about cell and tissue specification.
Anatomical checklist for phenotype scoring.

Anatomical checklist for phenotype scoring.


1. Analysis by morphological deffects

For the matings of nonsense mutation carriers we will perform three levels of analysis. As a first analysis, the observation of living embryos using a dissecting microscope is simple and very informative. Using a checklist of organs and embryonic features we will score for the presence and status of each feature at defined stages (See Table). Various stages of development of wild-type Xenopus embryos are shown in Table 6 highlighting several tissues that will be scored. In addition to morphological defects live embryos will be assayed for behavioral phenotypes such as circling, touch response and motility. As can be seen in Figure 13, once the tadpole has cleared most of its yolk, it becomes quite transparent, and many organs can be distinguished in the dissecting microscope. In many cases, defects that are not easily observed at early stages will have obvious pleiotropic effects: for example, defects in the lymph heart, heart, pronephros or blood vessels can all lead to edema. Use of this specific checklist will ensure that subtle defects will not be missed. In cases of strong pleiotropic effects, such as those resulting in edema, in situ hybridization for structures such as the lymph heart (pax7, en1; Martin and Harland, in preparation) or pronephros (pax2, WT1; Xenbase) are used in a secondary screen at tadpole stages.

2. Analysis of gene expression by micro-arrays

To summarize, we will use 3 biological replicas with a dye swap for each, and 2 time points, yielding a total of 24 microarrays per gene analyzed. Image analysis will be carried out using the ScanArray (Perkin-Elmer) software that runs most of scanners at the Sanger Institute Microarray facility. The downstream analysis and quality control will be carried out using custom scripts, Bioconductor and GeneSpring (Silicon Genetics) which produces MIAME compatible data for submission to ArrayExpress and GEO.The data will be LOWESS normalized, then we will define regulated genes as: present in more than 75% of the hybridizations, show at least 1.5 fold change in expression, and are statistically significant within a time point (t-test p<0.05).

Our oligonucleotide probes have already been matched to the current X. tropicalis Ensembl build and GO terms of biological function are present in the GeneSpring genome definition files as are links to in situ image databases.

MA plots and principle component analysis will be used identify outlier slides (i.e. poor hybridizations etc.) (Dudoit et al., 2002). In addition, ~1/4 of all the elements on the arrays we have designed are control probes: spiked controls, 'housekeeping' gene controls, and reverse transcription controls (65mers at 500bp intervals 3'>5' along ornithine decarboxylase 1 and plakoglobin with antisense versions). A further ~10% of the genes are homologues of genes that do not change in between human cell lines or zebrafish development on Affymetrix arrays. Finally, all oligonucleotide probes are present in duplicate within a block to facilitate duplicate correlation and identification of pin/block effects.

3. Analysis by in situ hybridization:

From our previous experiments with expression cloning, and preliminary assays using mutagenesis, morphological examination detects defects with great sensitivity. However, this is often due to the pleiotropic effects of mutations. We do not yet know how many more subtle defects we may miss by using just morphological tests. Thus we will carry out a second level of analysis using in situ hybridization probes that have been selected to reveal specific tissues in a highly selective way. Cocktails of RNA probes can be used simultaneously on tadpoles, or probes that simultaneously reveal different structures can also be examined, as shown in the preliminary results. In situ hybridization is an important adjunct to the morphological screening, since alterations in several structures are difficult to see in whole embryos. Thus changes in hindbrain segmentation and neuronal identities have relied on scoring molecular markers. We will therefore routinely screen the mutants with a mix of probes that marks sub-regions of the nervous system, including the mediolateral mixture shown in the preliminary results, which marks the borders of the neural plate, intermediate interneurons, and the floor plate. This will be complemented by adding in sizzled, which stains the ventral midline. For rostrocaudal pattern, we will stain tailbud tadpoles. The mix will consist of stain for the forebrain with emx1 and nkx2.1, the isthmus with engrailed-2 and the hindbrain with krox20, the neural mixture will be rounded out with the spinal cord marker hoxb9. This mixture will also be spiked with hex for liver. In addition, emx1 will mark the kidney. Several of these markers are shown in X. tropicalis in (Wills et al., 2006). Although all mutant clutches will be stained with these mixtures, and examined for Mendelian ratios of defective embryos, there will be cases where we apply different probe mixtures depending on suggestions of phenotype (for example in edema we will distinguish between vascular, cardiac, lymphatic and pronephric defects), or in cases where there is a mouse mutant, or other literature on a gene, we will target the probe choice depending on possible expectations. In cases of interesting results, we will follow up with detailed time courses to document the onset of defective development, and examine the various fields of tissues in the vicinity of the defective tissue to determine whether there is likely to be a non-autonomous problem in development, for example, a failure of heart primordial to fuse in zebrafish was tracked to the failure to form the endoderm (Alexander and Stainier, 1999).

* quick link - http://q.sanger.ac.uk/uvswcvbh