I Introduction
A central dogma of developmental biology today is that the fundamental genetic mechanisms that control the development of metazoans have been conserved evolutionarily, albeit frequently modified in their application. For example, invertebrates and vertebrates employ homologous signaling systems that act antagonistically to establish topologically equivalent, but spatially reversed, dorsal/ventral axes (De Robertis and Sasai, 1996). Based on mutant phenotype and protein structure, vertebrate ventralizing signals (e.g., BMP-2, BMP-4) are functionally homologous to Drosophila Decapentaplegic, which functions in dorsal determination in the fly, and the vertebrate dorsalizer Chordin is homologous to the Drosophila ventralizing signal, Short gastrulation. Nevertheless, some aspects of development are uniquely vertebrate. The neural crest, for example, is a group of migratory cells that arises in the embryo at the border between neural and nonneural ectoderm. These cells move to many regions of the embryo to form numerous tissues, including part of the cranial skeleton and the peripheral nervous system. The development of complex organ systems, such as the brain, heart, and kidneys, is another hallmark of vertebrates that is not easily studied in invertebrate genetic systems. For developmental analysis of vertebrates, the zebrafish, Danio rerio, has arguably emerged as the genetic system par excellence.
In December, 1996, the world of biological science witnessed the equivalent of Yogi Berra's ādĆ©jĆ vu all over again.ā That month's issue of the journal Development was devoted entirely to the description, in 37 articles, of approximately 2000 mutations that perturb development of the zebrafish (for highlights, see Currie (1996), Eisen (1996), Grunwald (1996), Holder and McMahon (1996)). This magnificent accomplishment, the result of two independent, large-scale mutagenic screens of the zebrafish genome and phenotypic analysis of embryonic development in the mutants obtained, approximates in a vertebrate the earlier saturation mutagenic screen in Drosophila (NĆ¼sslein-Volhard and Wieschaus, 1980). Indeed, two of the investigators leading the zebrafish screens, Christiane NĆ¼sslein-Volhard of the Max-Planck-Institut fĆ¼r Entwicklungsbiologie in TĆ¼bingen and Wolfgang Driever of the Massachusetts General Hospital (MGH) in Boston, were veterans of the Drosophila program. Working at the European Molecular Biology Laboratory in Heidelberg, āJanniā NĆ¼sslein-Volhard and her colleague Eric Wieschaus (corecipients with Edward Lewis of the 1995 Nobel Prize in Physiology or Medicine) conducted the now legendary Drosophila screen, and Driever, as a later member of the NĆ¼sslein-Volhard laboratory, analyzed many of the mutants to determine essential signaling pathways that control development of the fly's body plan. NĆ¼sslein-Volhard in TĆ¼bingen, and Driever and his colleague Mark Fishman at the MGH, subsequently applied the conceptual framework of the Drosophila screen to the fish. The community of developmental biologists owes these three individuals, and their many colleagues and collaborators, a tremendous debt of gratitude for this repeat performance.
II History of the Zebrafish System and Its Advantages and Disadvantages
These recent mutagenesis screens provided proof-of-principle that classical, forward genetics can be used to understand vertebrate development. The identification and study of mutations has been extraordinarily successful in providing an understanding of the early development of Drosophila and of the nematode worm, Caenorhabditis elegans. However, the same level of analysis of early developmental events in vertebrates has been more problematic. In the mouse, historically the species of choice for studies of vertebrate developmental genetics, much of embryogenesis is difficult to follow because it occurs within the mother's uterus. Beginning about 20 years ago at the University of Oregon, George Streisinger recognized the power of genetic analysis for understanding development and the advantages of a small tropical fish with external fertilization as a vertebrate for this approach. Streisinger selected the zebrafish, a freshwater fish commonly available in pet stores, because it has a relatively short generation time (2ā3 months), produces large clutches of embryos (100ā200 per mating), and provides easy access to all developmental stages. Zebrafish embryos are optically transparent throughout early development, which facilitates a host of embryological experiments and the rapid morphological screening of the live progeny of mutagenized fish for interesting mutations. Before his untimely death in 1984, Streisinger's group cloned the zebrafish (Streisinger et al., 1981) and developed techniques for mutagenesis (e.g., Grunwald and Streisinger, 1992a,b; Walker and Streisinger, 1983), genetic mapping (Streisinger et al., 1986), and clonal analysis of development by genetic mosaics (Streisinger et al., 1989). They also used F1 screens of mutagenized fish to isolate zygotic recessive lethal mutations with wonderfully curious embryonic phenotypes (Felsenfeld et al., 1991; Grunwald et al., 1988).
Streisinger's discoveries, as well as his enthusiasm and generosity, stimulated a number of other laboratories to begin using the zebrafish for developmental and genetic studies. Initially, all of these laboratories were also in Oregon. These groups have extended Streisinger's original studies by isolating and analyzing additional informative mutants (Halpern et al., 1993, 1995; Hatta et al., 1991; Kimmel, 1989; Kimmel et al., 1989) and have developed techniques for production of transgenic zebrafish (Stuart et al., 1988; Westerfield et al., 1993). Moreover, recent work has demonstrated the advantages of zebrafish for cellular studies of vertebrate embryonic development. The embryo is organized very simply (Kimmel et al., 1995) and has fewer cells than other vertebrate species under investigation (Kimmel and Westerfield, 1990). Its transparent cells are accessible for manipulative study. For example, cells can be injected with tracer dyes in intact, developing embryos to track emerging cell lineages (Kimmel and Warga, 1986) or axons growing to their targets (Eisen et al., 1986). Uniquely identified young cells can be ablated singly (Eisen et al., 1989) or transplanted individually to new positions (Eisen, 1991) to address positional influences on development at a level of precision that is unprecedented in any species. The combination of easy mutagenesis and powerful phenotypic screens of the earliest developmental stages eliminates, in principle, the biased detection of mutant phenotypes observed in the mouse, where scoring of mutants is generally restricted to neonatal and adult animals due to intrauterine development of the embryos. The more recent advent of tools for mapping mutations and candidate genes in the zebrafish genome has already begun to facilitate the isolation and functional analysis of genes required for normal development. Even small laboratories can conduct reasonably sized screens for new mutations, and the cost of a fish facility necessary to support such research is significantly lower than for the mouse.
Several disadvantages of the zebrafish system are also apparent. We presently lack in the zebrafish system methods to generate embryonic stem cells for gene āknock-outsā by homologous recombination. In the absence of such methods, we envision a cooperative and synergistic game of āping pongā between the zebrafish and mammalian research communities. Knock-out analysis of the mouse homologues of genes identified via study of zebrafish mutations should lead to a greater understanding of gene function in ver...