In the first day or so of life, the fruit fly Drosophila melanogaster grows from a glistening oval egg into a wriggly maggot. As it does so, regular stripes of gene activity appear that eventually translate into the segments of the maggot’s body.
These regimented patterns caught the eye of physicist-turned-biologist Hans Meinhardt, who began putting forward mathematical models to explain them, based in part on the kind of reaction–diffusion ideas proposed by Alan Turing in 1952. Then in 1978, a year before Newman and Frisch’s paper on Turing patterning in the limb, theoretical biologist Stuart Kauffman laid out how he thought the same system was in action in the developing fly.
Kauffman suggested that the neat lines in the Drosophila embryo might be created by Turing’s reaction–diffusion method, with interactions of activators and inhibitors generating stripes of different molecules across the growing body. These, he suggested, are then refined and interpreted to create segments with individual characteristics – bearing wings or legs, for example – depending on where they are.
Unlike the situation with the developing ‘fingers’ in the chicken limb, where Turing’s ideas were quickly rejected, these models were taken seriously by developmental biologists working with real flies in the lab. Leader among these was German embryologist Christiane (Janni) Nüsslein-Volhard, whose dogged search for genes that affect Drosophila development was key to laying open the genetic recipe book that makes a fruit fly.
Michael Akam, an expert in insect development at the University of Cambridge, recalls the situation at the time. “There were a lot of different models. Janni talked a lot to Hans Meinhardt when they were doing the original screens for genes in Drosophila and she expected to find Turing-type mechanisms.” But it turned out not to be so.
At first glance the beautiful repeating stripes in the fly seemed consistent with what would be expected from a self-organising, Turing-like pattern. And mistakes, or mutations, in a single gene seemed to change the pattern across the whole organism – another key hallmark. But on closer inspection, it became apparent that this was not the case. Nüsslein-Volhard and the rest of the fly brigade started to find mutations that seemed to have specific effects on just one stripe, which didn’t fit with the idea of a self-striping system.
Piece by piece the picture slotted together. In the fruit fly, it turns out that the stripes are inelegantly generated by a complex interplay of underlying gradients of morphogens. Turing’s ghost had led them astray. Akam explains, “I remember Janni saying, ‘We were misled for years by Meinhardt’s models and they turned out to be completely unhelpful.’”
By the late 1980s it looked like the idea that Turing mechanisms could explain the creation of any part of a living thing had been finally laid to rest. The stripy fruit fly maggots that looked for all the world like they were patterned by a Turing mechanism turned out not to be so. This supported the prevailing dogma that the digits in the limb were patterned by underlying gradients of morphogen chemicals, rather than a self-organising Turing system.
But with new evidence coming to light showing that Turing patterns underlie a number of biological phenomena, including limb patterning, it’s time for a rethink.
At the Centre for Genomic Regulation in Barcelona, Spain, James Sharpe suspects that the famous stripes in a fruit fly embryo might have more to do with Turing patterning than the rest of the developmental biology community are prepared to believe. When I ask him about this he chuckles guiltily. “Personally, I think that even in the fly things need to be re-evaluated, but I can’t really say it publicly. The Drosophila biologists absolutely killed that idea and won’t contemplate anything other than fairly pure positional information.”
As a Drosophila biologist himself, Michael Akam is a little more circumspect. “I think we were unfairly negative about modelling, and what’s happened in the last few years has been a vindication of some of the modellers’ views, actually. We now understand quite a lot more about the fly, and some of the things that were highlighted as being most peculiar and unexpected about it we would now see as characteristics that evolved at some point in insects but were not originally part of their segmentation process.”
But could there be Turing mechanisms at work? Akam also thinks there probably are ancestral repetitive processes going on to make insect segments, which have subsequently been overwritten by the complex gene patterns laid on top by thousands of years of evolution – filling in the painting by numbers on top of a pre-existing pattern, rather than actually generating a pattern. “Even if they’re not Turing patterns, then they’re some kind of physical processes. It must be of some comfort to the modellers.”