How the zebrafish became a model organism

Walk into almost any developmental biology lab today and you will find racks of small aquaria holding an unassuming striped fish about three centimetres long. Danio rerio, the zebrafish, did not start out as a genetic powerhouse. It started out in pet shops. Its journey from home aquarium curiosity to one of the four or five most important model organisms in biology is a genuinely unusual story: it hinges on one geneticist's stubbornness, two rival mutagenesis screens on opposite sides of the Atlantic, and a genome sequencing effort that confirmed just how much this little fish has in common with us.

A phage geneticist looks for a vertebrate

George Streisinger was not a fish biologist by training. He spent the 1950s and 1960s working on bacteriophage genetics, first under Salvador Luria at the University of Illinois and later at Caltech, where he contributed to some of the foundational work on the genetic code and the structure of the T4 phage genome. By 1960 he had taken a post at the newly forming Institute of Molecular Biology at the University of Oregon in Eugene, which he helped establish.

Phage and bacterial genetics had answered many of the questions Streisinger and his contemporaries had set out to answer. What phage could not do was explain how a nervous system assembles itself, how cells know where to go and what to become inside a developing vertebrate. Streisinger wanted to bring the same kind of rigorous, mutation-based genetics that had worked so well for viruses and bacteria to a living vertebrate with a brain and a spinal cord. Colleagues later recalled that he described the zebrafish as "a phage with a backbone", a phrase that captures exactly what he was after: an organism simple and cheap enough to breed and screen like a microbe, but complex enough to have a real nervous system.

He was also, by his own account, a home aquarist, and had experience with fish through work with the ichthyologist Myron Gordon. Zebrafish, then usually called by their older name Brachydanio rerio, were already popular in the aquarium trade: hardy, small, and easy to breed. Streisinger and his colleague Charline Walker began working with them in the lab from the late 1960s, at a time when the idea of building an entirely new vertebrate genetic system from scratch, when mice already existed, struck many peers as an odd use of a career.

The scepticism was understandable. Mice had decades of genetics behind them by the 1970s, along with inbred strains and a growing toolkit. Zebrafish had none of that. What zebrafish did have was a set of practical advantages that, cumulatively, made them extraordinarily good for large-scale genetic work in a way mice never could be.

Why zebrafish, practically speaking

The case for zebrafish rests on a handful of very concrete biological facts, most of which come down to how the embryo develops.

• External, rapid development. Zebrafish eggs are fertilised outside the mother's body, so every stage of embryogenesis can be watched, manipulated, and photographed from the single-cell stage onwards, without surgery. By roughly 24 hours after fertilisation the embryo already has a defined body axis and rudimentary organs, including a beating heart.
• Optical transparency. The chorion and the embryo itself are essentially see-through in the early stages, which makes zebrafish exceptionally well suited to live imaging of cell division, migration, and tissue formation as they happen, rather than reconstructing them from fixed snapshots.
• High fecundity. A single breeding pair can produce clutches running into the hundreds of eggs, and can be bred repeatedly. For anyone trying to run a genetic screen, where the whole point is to examine large numbers of offspring for rare mutant phenotypes, that kind of output matters enormously.
• Low housing cost and small size. Adult zebrafish average under four centimetres in length, so many thousands can be kept in a modestly sized facility, at a fraction of the space and cost of a comparable mouse colony.
• Fast generation time. Zebrafish reach sexual maturity in around three months, which is slow compared with flies or worms but fast compared with mice, and matters a great deal when a genetic cross can take several generations to complete.

None of these facts, individually, were secrets in the 1970s. What Streisinger contributed was the genetic methodology needed to actually exploit them.

Cloning a vertebrate, fifteen years before Dolly

Streisinger's central technical problem was one familiar to anyone working with a diploid, sexually reproducing organism: recessive mutations are invisible in heterozygotes, and normal breeding to reveal them takes generations. He developed methods to bypass this using gynogenesis, activating eggs without allowing fertilisation by a normal sperm nucleus, combined with treatments to suppress the first meiotic or mitotic division. The result, published with Charline Walker and colleagues in Nature in 1981 under the title "Production of clones of homozygous diploid zebra fish (Brachydanio rerio)", was the first large-scale production of genetically uniform, homozygous clones of a vertebrate animal, some fifteen years before Dolly the sheep made animal cloning a household topic.

The paper mattered less as a curiosity than as a tool. Homozygous diploid clones meant that a mutation induced in one generation could be revealed and studied far more quickly than through conventional breeding, effectively giving zebrafish geneticists a shortcut around one of the slowest parts of classical vertebrate genetics. Combined with early chemical and radiation mutagenesis experiments in his lab, this work laid the genetic groundwork that later, much larger screens would build on directly.

Streisinger did not live to see the field's biggest expansion. He died in 1984, at 56, and Charles "Chuck" Kimmel, a colleague in the University of Oregon Institute of Neuroscience, took up leadership of zebrafish developmental genetics there. The Eugene lab's culture, generous with data, methodologically careful, oriented around the nervous system, carried on and trained many of the researchers who would go on to run zebrafish labs elsewhere.

Two labs, one enormous screen

The single event most responsible for zebrafish's rise to prominence took place a decade after Streisinger's death, and it happened twice, in parallel, on two continents.

Through the early-to-mid 1990s, Christiane Nüsslein-Volhard's laboratory in Tübingen, Germany, and a parallel effort led by Wolfgang Driever, working with Mark Fishman's laboratory at Massachusetts General Hospital in Boston, independently ran large-scale chemical mutagenesis screens in zebrafish, using ethylnitrosourea (ENU) to induce mutations and then screening thousands of mutagenised genomes for visible developmental defects. Nüsslein-Volhard had already won a share of the 1995 Nobel Prize in Physiology or Medicine for her earlier screens identifying genes controlling body plan development in Drosophila; the zebrafish screens were, in effect, an attempt to ask the same kind of question in a vertebrate.

The scale of the undertaking was without precedent for a vertebrate model. The Tübingen screen alone scored close to 3,857 mutagenised genomes and identified over 4,000 mutants, of which more than 1,150 were characterised in detail and around 890 assigned to roughly 370 genes through complementation testing. Between the Tübingen and Boston efforts combined, researchers identified close to 1,200 mutants with developmental defects spanning almost every organ and tissue in the embryo and larva, from the heart and gut to the eye, ear, and nervous system.

The results were published together in a single, enormous special issue of the journal Development in December 1996: 37 papers spanning 481 pages, still the largest issue in the journal's history. It is difficult to overstate how much this single issue changed the field. Before 1996, zebrafish genetics was a promising but niche pursuit run out of a handful of labs. After it, hundreds of laboratories worldwide had access to a deep, shared catalogue of mutants affecting gastrulation, organogenesis, pigmentation, and behaviour, along with proof that forward genetic screens, the same unbiased "find the gene from the phenotype" logic that had worked in flies and worms, could be run at scale in a vertebrate.

The two screens were, in a sense, in friendly competition, but they were published together deliberately, alongside additional contributions from Frederick Bonhoeffer's laboratory, so that the field received the full picture at once rather than in fragments. Many of the genes identified in 1996 are still referred to today by the whimsical names given during the screens, a naming convention that has become one of the more recognisable quirks of zebrafish genetics.

Sequencing the genome

Mutagenesis screens tell you that a gene exists and roughly what it does. They do not, on their own, tell you what the gene is or how it compares with its counterpart in other species. That required a reference genome, and the effort to produce one for zebrafish was led by the Wellcome Trust Sanger Institute in the UK over more than a decade.

The finished, well-annotated reference genome sequence was published in Nature in 2013 by Kerstin Howe and a large international author list, under the title "The zebrafish reference genome sequence and its relationship to the human genome". Among its more striking findings, the paper reported that around 70% of human protein-coding genes have at least one obvious zebrafish counterpart, and that a substantial majority of genes known to be associated with human disease, cited in the paper as 84%, have a zebrafish orthologue. The zebrafish genome also turned out to have the largest protein-coding gene set of any vertebrate sequenced up to that point, with relatively few pseudogenes.

That combination, deep conservation with humans plus a well-annotated, high-quality assembly, turned the existing mutant collections from the 1996 screens, and the many screens that followed, into something far more powerful. A phenotype from a 1996 ENU mutant could now be mapped to an actual gene with reasonable confidence, and that gene could be checked directly against its human equivalent. The genome project did not create the zebrafish field, Streisinger and the 1996 screens had already done that, but it gave the field a common, comparable reference to work from, and it strengthened the case for zebrafish as a model directly relevant to human biology and disease rather than simply a convenient developmental system.

From aquarium curiosity to shared infrastructure

It is worth noting how much of this history depended on cooperation rather than competition. The 1981 cloning paper, the twin 1996 screens, and the 2013 genome project were all published as shared resources rather than held back for individual advantage, and all three were rapidly built upon by labs that had nothing to do with the original work. Community stock centres and mutant repositories that grew out of the 1996 screens made those thousands of mutant lines available to any lab that wanted them, an approach that has remained a hallmark of the zebrafish research community.

None of this was inevitable. Streisinger could easily have picked a different vertebrate, or given up when mouse geneticists asked why he was bothering with a pet-shop fish. The Tübingen and Boston groups could have kept their results to themselves rather than coordinating a joint publication. The Sanger Institute's genome project could have stalled, as sequencing projects for less biomedically prioritised organisms sometimes did. Instead, each step built cleanly on the last: a genetic method that made mutant discovery practical, a pair of screens that made mutants abundant, and a genome that made those mutants interpretable in human terms. Half a century after Streisinger first brought zebrafish into his Oregon lab, that combination is why the fish that started in a home aquarium now sits alongside the mouse and the fly as one of the standard tools of vertebrate genetics.