A paper in this week's Science pinpoints exactly that change. Bikard and coauthors report that, in the little flowering plant Arabidopsis thaliana (the plant world's answer to white lab mice and Drosophila fruit flies), it only takes one duplicated gene to finalize speciation [$-a]. It's a clear-cut case of a classic speciation scenario, Bateson-Dobzhansky-Muller incompatibility.
It all comes down to gene duplication, which I've discussed before in the context of the trouble it gives to genetic analysis. Making copies of an entire genome is an error-prone process, and sometimes a whole gene gets duplicated twice. If that extra copy is inherited, it means that the carrier has redundant coding for whatever the original gene does - so now one copy can mutate without affecting its carrier's fitness. Often this just results in loss of function for the mutating copy - sometimes it leads to new gene functions. In Arabidopsis, it's lead to reproductive incompatibility between two strains of the plant that took different evolutionary paths.
Photo by tico bassie.
Photo by tico bassie.
Bikard et al. noticed that, when they crossed two strains of Arabidopsis, the resulting seeds didn't include every possible combination of the parental strains' genes - and a few seeds grew short, not-quite-healthy looking roots when germinated. Some of the hybrid seeds just died in mid-development. With a lot more controlled crosses, the authors narrowed the candidate genes down to a pair that normally work together in synthesizing the essential amino acid histidine. Each of the two parental strains had working copies of the two genes - but when you crossed them, sometimes the seeds couldn't produce histidine, and so they snuffed it.
This looked like the above-mentioned (and awkwardly named) Bateson-Dobzhansky-Muller incompatibility [$-a], which is an old idea about how populations evolve reproductive incompatibilities to become separate species. Under B-D-M incompatibility, a new gene evolves in one population that doesn't work if it interacts with genes from the other. Imagine if Windows users didn't have to share documents with Mac users: as the two operating systems went through multiple redesigns and their respective versions of Microsoft Office(TM) were revised to keep up, it might no longer be possible to read a Mac-written Word document on a Windows machine.
Here, as Bikard et al. showed, one of the histidine-producing genes in Arabidopsis was accidentally duplicated - and one copy mutated into non-functionality. The catch is that, in the two partially incompatible strains, different copies went nonfunctional. So now, when the two lines are crossed, a small fraction of the seeds produced get nonfunctional copies of the duplicated gene. They die. And where once there were two strains of Arabidopsis thaliana, there's something a little more like two separate species, all because of what boils down to the flip of a coin.
D. Bikard, D. Patel, C. Le Mette, V. Giorgi, C. Camilleri, M.J. Bennett, O. Loudet (2009). Divergent evolution of duplicate genes leads to genetic incompatibilities within A. thaliana Science, 323 (5914), 623-6 DOI: 10.1126/science.1165917
K. Bomblies, D Weigel (2007). Arabidopsis — a model genus for speciation Current Op. Genet. & Dev., 17 (6), 500-4 DOI: 10.1016/j.gde.2007.09.006