Frog genetics: Xenopus tropicalis jumps into the future
Enrique Amaya,1* Martin F. Offield2 and Robert M. Grainger2
For over a century amphibian embryos have been one of the favoured systems for elucidating the mechanisms of early development1. The two main reasons for this are that the embryos develop externally and their relatively large size allows embryologists to perform microsurgery and manipulate the embryos experimentally in ways that are not as easy in other vertebrate embryos (e.g. mouse or zebrafish). Most of the early experiments were done on embryos from Rana (common frog), Triturus (European newt) and Ambystoma (American salamander). Since the 1950's, the favoured amphibian model system has gradually shifted to Xenopus laevis (South African clawed frog). However, neither X. laevis, nor other amphibian species, has provided the complementary powerful insights permitted by extensive genetic analyses. This may soon change though as investigators focus studies on a close relative of X. laevis, Xenopus (Silurana) tropicalis, that offers the possibility of such genetic manipulations.
Several aspects of X. laevis biology has made it an ideal laboratory animal but one which is not suited to genetic manipulation. Among its advantages are that: 1) X. laevis is primarily aquatic and can be maintained and bred easily in aquaria2; 2) unlike most other amphibians, X. laevis happily feeds on "dead" organic material; 3) X. laevis is very hardy and tolerate a wide range of living conditions; and most importantly 4) X. laevis can be induced to ovulate and mate anytime of the year following a simple injection of gonadotropic hormones. This discovery in the 1930's became the basis of a simple pregnancy test for humans3,4, and led to its worldwide distribution and use5. By the late 1950's and early 1960's more sensitive methods for detecting pregnancy were developed and X. laevis were no longer needed for this purpose. However by this time developmental biologists throughout the world had begun to exploit Xenopus embryos as a convenient model system. For the purpose of genetics, and some molecular studies, though, X. laevis is not the ideal system, in large part because it is effectively polyploid6-8.
Chromosome number and ploidy levels in Xenopus
Genome duplication has occurred several times in fish and amphibians9. The genus Xenopus provides a particularly striking example of this. The genus contains nearly twenty separate species which are all, except one, polyploid, ranging from tetraploid (4N) to dodecaploid (12N)7-8. Xenopus laevis was the first species within this group to be extensively studied and became the standard to compare other species within the genus10. Its chromosome number is 36, and based on karyotype morphology and the formation of bivalents during meiosis, it was classified as a diploid species. However subsequent work has shown that the ancestral lineage of X. laevis underwent a genome duplication event6. Therefore, X. laevis is more accurately considered a tetraploid species7-8. This fact is reinforced in virtually every paper that reports the isolation of a new gene since two paralogs are usually described. The implications of the polyploidy state of X. laevis extends beyond the nuisance of characterising two closely related paralogs for each gene. For one, it brings into the system a level of redundancy that makes genetic studies virtually impossible, though some studies do argue that the genome may not be functionally tetraploid at all loci11. But an additional consequence of this redundancy is the potential relaxation of selection for any particular paralog. In the most extreme case one might expect loss or complete silencing of one paralog. But in most cases, one might expect partial silencing in either all or a subset of the normal expression pattern of a gene. This could greatly complicate promoter analysis since important regulatory elements could be maintained in different paralogs. These are compelling reasons to explore an amphibian model system that is not polyploid, such as Xenopus (Silurana) tropicalis.
Xenopus (Silurana) tropicalis
Xenopus (Silurana) tropicalis is the only diploid species in the Xenopus genus. Based on karyotypic, morphological and molecular data, X. tropicalis forms a separate, but evolutionarily related, lineage from X. laevis6,7,13-15. A recent morphological study comparing X. tropicalis to other pipid frogs suggested that X. tropicalis might be more closely related to Hymenochirus than to X. laevis. Therefore the original genus of Silurana, coined by Gray in 1864 in his original description of the species12, was resurrected in order to make Xenopus monophyletic13. However a re-evaluation of the morphological data plus evidence based on molecular criteria have conclusively shown that X. tropicalis is monophyletic with the rest of the Xenopus genus7,14. For this reason, we continue to refer to the species as Xenopus tropicalis rather than Silurana tropicalis.
Several features makes X. tropicalis an ideal vertebrate model system for developmental biology. A comparison of an adult X. tropicalis and X. laevis (Fig. 1) indicates a highly similar overall morphology, though X. tropicalis is considerably smaller. It shares all of the advantages discussed previously for X. laevis. However, the haploid genome size per nucleus of X. tropicalis (1.78 pg or 1.7 X 109 bps) is nearly half of X. laevis (3.2 pg or 3.1 X 109 bps)15. Similarly the chromosome complement of X. tropicalis is 20 chromosome pairs, instead of 36 for X. laevis16. Among amphibians, X. tropicalis has one of the smallest genomes. For example the genome size of salamanders range from 17-96 pg per haploid nucleus (10-50 times larger than the genome of X. tropicalis!)17 Among non-amphibian model systems, X. tropicalis compares favourably, being nearly half the genome size of mouse (3 X 109 bps per haploid nucleus) and about the same as zebrafish (1.8 X 109 bps).
In addition to a small genome size and diploid state, X. tropicalis offers additional advantages as a system. It has a relatively short generation time of 4 to 6 months, and this may be truncated further under optimal lab conditions. For comparison, the generation time of X. laevis is 1-2 years. In fact, X. tropicalis has perhaps the shortest generation time of any amphibian17. X. tropicalis is also significantly smaller than X. laevis (see Fig. 1), allowing more animals to be housed in equivalent space. Despite their small size, X. tropicalis females lay 1000-3000 eggs per ovulation. The embryos are smaller than those of X. laevis (0.7-0.8 mm versus 1.0-1.3 mm, respectively), but are sufficiently large to undertake the subtle manipulations performed in X. laevis and show less variability in development from batch to batch. In addition, because of the close relationship of X. tropicalis and X. laevis, many sequences between the two species are highly conserved, as shown by successful in situ hybridization of X. tropicalis embryos with a number of X. laevis probes (N. Hirsch and R. Grainger, unpublished). All of these features make X. tropicalis an exceptional vertebrate model system where classical embryological approaches can be combined with modern molecular approaches. However the most exciting aspect of X. tropicalis is its potential of becoming a powerful genetic system.
At present loss of function experiments in the frog are mainly limited to missexpression of dominant negative mutations, which interfere with the function of endogenous genes18,19. However these experiments are often plagued by the fact that dominant negative mutations are rarely specific to one gene product. Therefore there is a well justified need to develop strategies for producing null mutations of endogenous genes in the frog. Within the Xenopus genus, such strategies must be developed in X. tropicalis, the only diploid species.
Transgenesis in Xenopus tropicalis
The final impetus for developing X. tropicalis as a vertebrate genetic model system is the recent advent of a method for generating transgenic frog embryos20. Briefly, the procedure involves integrating DNA into isolated sperm nuclei in vitro, followed by transplantation of these manipulated nuclei into unfertilised eggs, thus generating transgenic embryos. The method is very efficient, resulting in the production of hundreds of transgenic embryos in a few hours. One of the great advantages of this system over transgenesis in mice or zebrafish is that the transgene is integrated into the male genome prior to fertilisation, therefore the resulting embryos are not chimeric and breeding of animals is not required. Indeed this method of transgenesis has mainly been used in X. laevis, where the 1-2 year generation time makes multi-generation experiments unthinkable. However, such experiments are not only conceivable, but practical in X. tropicalis.
The transgenesis technique, originally developed for X. laevis, has been modified for X. tropicalis (M. Offield and R. Grainger, unpublished). The ability to generate transgenic X. tropicalis opens a powerful array of possibilities. For example, transgenic lines could, in principle, be established to label any structure. After mutagenesis, these lines could be used to easily screen for absence or abnormal development of the structure. For example a transgenic line has been made containing the g1-crystallin promoter driving the expression of the green fluorescent protein (GFP) in the lens (see Fig. 2). A mutant screen in this transgenic line would greatly facilitate the identification of mutations affecting lens development. This transgenic line can also be used in explant, recombinant or transplantation experiments to easily assay for lens induction.
The transgenesis technique in X. tropicalis also has the potential of being used as an insertional mutagen. The preferable method for insertional mutagenesis will a gene trap approach21-22. For this, transgenic X. tropicalis embryos carrying multiple integrations of a splice acceptor - GFP construct will be generated. Since the construct lacks a promoter, the marker gene can only be transcribed and translated if it integrates into an exon or intron of an actively transcribed endogenous gene, thereby simultaneously disrupting the gene and acting as a locus- specific marker of the gene. This would greatly facilitate the identification and cloning of the disrupted genes, a task which is not necessarily straightforward following chemical mutagenesis. In practice hundreds of transgenic embryos, each carrying multiple integration sites, could be generated per day and screened for GFP expression. A huge advantage of using this approach in frog embryos is that they develop externally, therefore GFP expression can be assayed in living embryos at any stage. Most embryos will not express the marker gene. These will be discarded and only the few that express will be nurtured to maturity, thus great reducing the number of embryos that must be carried to the next generation. Preliminary experiments in X. laevis strongly suggest that using a gene trap approach will be productive (O. Bronchain and E. Amaya, unpublished).
The powerful manipulations that one can perform on amphibian embryos have been used to reveal important principles about development for over a century. As we approach the next century many laboratories are beginning to work with Xenopus tropicalis . In this system it will be possible to overlay this rich embryological history with the power of genetic manipulations, creating an armamentarium of approaches as we look toward revealing a new generation of concepts about vertebrate embryonic development.
1 Hamburger, V. (1988) The Heritage of Experimental Embryology: Hans Spemann and the Organizer. Oxford University Press, Oxford.
2 Wu, M. and Gerhart, J. (1991) in Xenopus laevis: Practical Uses in Cell and Molecular Biology (Kay, B.K. and Peng, H.B., eds), Meth. Cell Biol. 36, 3-18.
3 Bellerby, C.W. (1934) Nature 133, 494-495.
4 Shapiro, H.A. and Zwarenstein, H. (1934) Nature 133, 762.
5 Weisman, A.I. and Coates C.W. (1944)The South African Frog (Xenopus laevis) in Pregnancy Diagnosis: A Research Bulletin. New York Biologic Research Foundation, New York.
6 Bisbee, C.A., Baker, M.A., Wilson, A.C. Hadiji-Azimi, I. and Fischberg M. (1977) Science 195, 785-787.
7 Cannatella, D.C. and de Sa, R.O. (1993) Syst. Biol. 42, 476-507.
8 Graf, J.-D. and Kobel, H.R. (1991) in Xenopus laevis: Practical Uses in Cell and Molecular Biology (Kay, B.K. and Peng, H.B., eds), Meth. Cell Biol. 36, 19-34.
9 Ohno, S. (1970) Evolution by Gene Duplication. Springer-Verlag, Berlin.
10 Tymowska, J. and Kobel H.R. (1972) Cytogenetics 11, 270-278.
11 Krotoski, D.M., Reinschmidt, D.C. and Tompkins, R. (1985) J. Exp. Zool. 233, 443-449.
12 Gray, J.E. (1864) Ann. Mag. Nat Hist. 14, 315-316.
13 Cannatella, D.C. and Trueb, L. (1988) Zool. J. Linn. Soc. 94, 1-38.
14 de Sa, R.O. and Hillis, D.M. (1990) Mol. Biol. Evol. 7, 365-376.
15 Tymowska, J. (1973) Cytogenet. Cell Genet. 12, 297-304.
16 Thiebaud, C.H. and Fischberg, M. (1977) Chromosoma 59, 253-257.
17 Duellman, W.E. and Trueb, L. (1986) Biology of Amphibians. McGraw-Hill Book Co., New York.
18 Herskowitz, I. Nature 329, 219-222 (1987).
19 Amaya, E., Musci, T.J. & Kirschner, M.W. Cell 66, 257-270 (1991).
20 Kroll, K.K. and Amaya, E. (1996) Development 122, 3173-3183.
21. Gossler, A., Joyner, A.L., Rossant, J. & Skarnes, W.C. Science 244, 463-465 (1989).
22. Skarnes, W.C., Auerbach, B.A. & Joyner, A.L. Genes and Dev. 6, 903-918 (1992).
Figure 1. Comparison of an adult female X. laevis (left) versus an adult femaleX. tropicalis (right).
Figure 2. Transgenic X. tropicalis frog expressing green fluorescent protein (GFP) specifically in the lens of the eye. The transgene in this frog contains the X. laevis g1-crystallin promoter controlling the expression of GFP. Inset shows a magnified view of one of the eyes.