News & Views | Published:

Building-blocks of embryogenesis

Nature Genetics volume 31, pages 125126 (2002) | Download Citation


Mutant screens have become a successful approach in vertebrates, but cloning the mutated genes can be an arduous task. Now an insertional mutagenesis method in the zebrafish bypasses this difficulty and is used to clone 75 mutated genes!

What does it take to make an embryo? The large-scale forward genetic screen has been a powerful method to identify important developmental genes in the fly and nematode, and they have now come to the forefront in vertebrates, in studies of both the zebrafish1,2 and mouse3,4,5,6. However, a major difficulty in these mutant screens is identifying which of the 30,000–60,000 genes in the genome is the one mutated.

The laboratory of Nancy Hopkins established a method to mutate and tag genes in the zebrafish that makes this arduous job of cloning the mutated gene relatively simple. Now they have gone on to use this method to carry out a large mutant screen. On page 135 of this issue, Golling, Hopkins and colleagues7 describe the molecular isolation of 75 mutated genes that are required for normal development of the zebrafish embryo—a stunning achievement for any organism, no less a vertebrate. This represents more mutated genes cloned than the entire zebrafish field has isolated over the past ten years! The spectrum of genes found indicates that developmental biologists have a long road ahead to understand embryonic development.

The viral mutagen

A large collection of embryonic zebrafish mutants have been generated using the chemical mutagen, ethylnitrosourea (ENU)1,2, which causes single base-pair changes in DNA. Isolation of the gene with that single base-pair change amid the 1.8 × 109 base pairs in the genome can be a challenging task. The majority of such mutated genes cloned in the zebrafish, about 50, have been identified through educated guesses8, based on studies of the mutant defects, previous knowledge of molecular pathways acting in the affected process, and co-localization of the chromosomal position of the mutation and the candidate gene. However, this approach is limited to those genes previously shown to act in the process of interest. To identify novel mutated genes, one must undertake the very difficult task of positional cloning. This has been accomplished for only a few genes in the zebrafish and typically takes years for each gene. The sequence of the zebrafish genome, expected in 2003, will make this approach easier, probably reducing the time to less than a year per gene.

To simplify the cloning process, the Hopkins group has developed a retrovirus as a mutagen in the zebrafish. The retrovirus, developed for human gene therapy, is a hybrid form of two viruses. It contains the genome of the Moloney murine leukemia virus, amenable to the stable transfer of exogenous genes into the mammalian genome, but uses the VSV-G envelope protein, which provides an outer coat for the virus and allows it to infect cells of a wide range of organisms, including those of zebrafish9. The Hopkins lab has optimized the generation of high titers of this virus and generated founder fish carrying an average of 25 independent insertions10. The mutagenicity of these insertions is about seven to nine-fold lower than that generated by ENU10,11 due to a low fraction of insertions disrupting genes. Thus, the insertional mutagenesis method initially requires substantially more work to isolate the same number of mutations compared to ENU mutagenesis, but the ease with which one can clone the mutated genes is extraordinary.

Golling et al.7 carried out a large-scale screen using this pseudotyped retrovirus as a mutagen, which then served as a molecular flagpost for efficiently cloning the mutated gene. The authors identified a similar spectrum of mutants as in the chemical mutagenesis screens: about one-third have specific developmental abnormalities and two-thirds have pleiotropic defects thought to be due to mutations in housekeeping genes. Only mutants with specific defects were kept from the chemical mutagenesis screens. With the ease of cloning the mutated genes—they could clone most mutated genes within two weeks!—the authors cloned all mutated genes from their screen, including those causing non-specific abnormalities. Thus, they have taken an unbiased approach at defining the genes required for vertebrate embryonic development.

What it takes: a glimpse

The authors clearly demonstrated the success of the insertional strategy. They broadly classify the mutant abnormalities and reveal the molecular nature of the cloned genes. In supplementary data, they provide brief descriptions of the mutant phenotypes, photographs of all mutants, as well as the gene structures, which will serve as an important resource for the zebrafish community, as well as others. The specific mutant phenotypes are similar to those of previous screens, ranging from defects in gastrulation to pigment formation and organ development.

The spectrum of genes suggests a bias in the previous screens toward receptors, ligands, and transcription factors. This is not totally unexpected, considering the majority were cloned by the candidate gene approach, the candidates being selected on the basis of previous studies. The hope in such a screen is to isolate novel genes not identifiable by the candidate approach. In this regard, the insertional screen is also a success: a large fraction of the genes, 20% of both the specific and non-specific–acting genes, are indeed novel. All the mutated genes have human orthologs and a few may be vertebrate-specific or animal-specific. Furthermore, the unbiased approach now demonstrates that most non-specific, pleiotropic mutant phenotypes are owing to mutations in genes encoding components of the general cellular machinery.

A good number of the mutants with specific defects are caused by mutant proteins relevant to cell biology or physiology. Some mutants may, in fact, have a primary defect in these processes, but display a specific defect owing to the tissue-restricted expression of the gene. Some may be hypomorphic alleles, which reduce, but do not eliminate gene expression. Alternatively, some of these gene products may have unknown functions or be acting in specific cell differentiation or other developmental processes; for example, to regulate the subcellular localization or modification of our favorite factors.

The retrovirus used by Golling et al. can be used as a gene-trap vector in the zebrafish, as recently reported10. It could also be used to generate a library of insertional alleles, similar to the Drosophila melanogaster P-element collection12, which might be used in screens for abnormalities that are not easily observable. And it could be used for reverse genetic approaches, through a sperm and DNA bank of insertions that would be screened by molecular methods for insertions in a gene of interest, similar to Tc1 insertions in Caenorhabditis elegans13.

The authors predict that, upon completion of their screen, they will have generated almost 500 insertional mutations. Cloning these genes will reveal a major fraction of the building blocks needed to make a vertebrate embryo. Whether all genes with essential functions in embryogenesis can be mutated by this retrovirus remains unclear; further use will reveal the extent, if any, of bias. Many of the specific mutants described here, although likely to exist in the chemical mutant collections, have not been previously studied. Thus, a major challenge lies ahead in deciphering the biological and molecular functions of these genes.


  1. 1.

    et al. Development 123, 1–36 (1996).

  2. 2.

    et al. Development 123, 37–46 (1996).

  3. 3.

    , & Proc. Natl Acad. Sci. USA 95, 485–490 (1998).

  4. 4.

    , & Methods 13, 379–395 (1997).

  5. 5.

    et al. Nature Genet. 25, 440–443 (2000).

  6. 6.

    et al. Nature Genet. 25, 444–447 (2000).

  7. 7.

    et al. Nature Genet. 31, 135–140 (2002).

  8. 8.

    & Genes Dev. 14, 755–762 (2000).

  9. 9.

    , , , & Proc. Natl Acad. Sci. USA 90, 8033–8037 (1993).

  10. 10.

    , , , & J. Virol. 76, 2192–2198 (2002).

  11. 11.

    et al. Genes Dev. 13, 2713–2724 (1999).

  12. 12.

    et al. Genetics 153, 135–177 (1999).

  13. 13.

    , , , & Proc. Natl Acad. Sci. USA 90, 7431–7435 (1993).

Download references

Author information


  1. Department of Cell and Developmental Biology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, USA.

    • Mary C. Mullins


  1. Search for Mary C. Mullins in:

About this article

Publication history



Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing