Evolution doesn't make U-turns, according to a new study of proteins. The study shows that simply reversing selective pressure won't make a biomolecule revert to an earlier form. The finding confirms a much-debated biological law that, evolutionarily speaking, there's no going back.
Since the late 19th century, evolutionary biologists have debated whether evolution can go in reverse. If not, then evolution may depend on more than just natural selection. Multiple evolutionary paths could be possible through small chance events. It hasn't been easy to examine reversibility. Previous studies have focused on complex traits such as whale flippers, and scientists often lack sufficient information about ancestral traits or how present-day traits evolved.
So evolutionary biologist Joseph Thornton of the University of Oregon, Eugene, and his colleagues picked a more tractable subject: a single protein. His group has been studying the more than 450-million-year evolution of the glucocorticoid receptor (GR), a protein that binds to the stress hormone cortisol to control animals' response to it. Like all proteins, GR is made up of amino acids. By collecting the amino acid sequences of GR and related proteins from living animals, Thornton and his team previously constructed the GR evolutionary tree and resurrected sequences  of GR's ancestors.
This history reveals that GR has switched its hormone preference. Around the time cartilaginous fish such as sharks split off from bony fish, roughly 440 million years ago, the ancestral protein that the scientists call GR1 responded to both cortisol and the hormone aldosterone. But 40 million years later, when four-legged creatures started to appear, the descendent GR2 had become cortisol-specific.
During these 40 million years, 37 amino acids changed. Only two were necessary to alter the function: One put a kink in the protein's shape, making it unresponsive to both hormones, and another allowed the restructured molecule to interact with only cortisol. Thornton's team next wondered if they could make GR2 recognize both cortisol and aldosterone by reverting these amino acids, which they call group X, back to their GR1 state. The researchers report  today in Nature that this swap not only couldn't restore GR's original dual function but that it also killed the protein's ability to recognize any hormone.
So what blocked the way back? By comparing images of GR2 and a putative ancestral protein, the scientists fingered another five of the 37 GR1-to-GR2 mutations as the culprits. These changes probably occurred randomly after the X mutations and had no significant effect on the protein's function going forward. But in reverse, when the scientists tried to iron out the GR2 kink, these mutations caused protein parts to crash into one another. For GR2 to evolve back into GR1, these five mutations must be reversed first to avoid this molecular fender-bender. But because they have no effect on which hormone the protein recognizes, there would be no selective pressure to reverse these mutations. "They burn the bridge to return back to the ancestral function," Thornton says.
The study is "perhaps one of the most important papers in the last 10 to 15 years in evolutionary biology," says evolutionary biologist Gunter Wagner of Yale University. Not only does the study show how evolution can't go backward, Wagner says, but it also provides a detailed mechanism for why.