Surprise, Surprise: Hox Proteins Have Evolved

Image: Courtesy of William McGinnis LOTSALEGS: The six-limbed fruit fly, Drosophila melanogaster As school children are commonly taught, adult insects have six legs. No more, no less. But there's more to this story. The class Insecta is descended from multilimb ancestors, and most other living arthropods, including Insecta's closest living relative, the Crustacea, usually have at least five pairs of legs or leg-like appendages. How insects lost their limbs has interested those in the burgeoning field of "evo-devo," the study of developmental evolution. In particular, biologists have suspected that the insects' progression to six legs, and the evolution and modification of arthropod appendages in general, are driven by changes in Hox gene activity. Hox genes are clusters of regulatory genes that control development by turning other genes on and off. Biologists had attributed changes in the arthropod body plan to changes in Hox gene expression, not changes in the Hox genes themselves; what mattered was whether and where a particular Hox gene was expressed, not whether a new Hox gene had evolved.1,2 Biologists considered the Hox genes so conserved and their sequences so similar that their encoded proteins were practically interchangeable between species as disparate as flies and mice. But, recent discoveries are changing some researchers' minds: Hox proteins, it appears, have evolved. Moreover, these proteins were changed through the loss and gain of two domains--a finding that some researchers consider the most exciting aspect of the work. Researchers from the University of Wisconsin and the University of California, San Diego have shown how one Hox protein in particular, Ubx (Ultrabithorax), has changed so much over time that it functions differently in insects than it does in crustaceans and onychophorans, the arthropod's sister taxa from which the arthropods diverged 520 million years ago.3,4 Insects have evolved a Ubx protein that represses limb development, whereas crustaceans and onychophorans have each evolved slightly different Ubx proteins without the same limb-repression capability. In part, these differences in Ubx account for why insects have fewer limbs than their multilimb crustacean and onychophoran relatives. Together, the papers provide "perhaps the clearest example of a change in a regulatory protein being implicated so directly in developmental, or morphological, change during evolution," says Michalis Averof of the Institute of Molecular Biology and Biotechnology in Greece. "Until now, it was not clear whether such changes were involved." Biologists have generally considered protein evolution an unlikely mechanism of developmental change, says organismal biologist Nipam Patel, University of Chicago. "Most people haven't been focused on trying to find it." Because Hox regulatory proteins play so many roles and affect dozens or even hundreds of other genes, it had been generally believed that the penalties would be too great and too many biological processes would be disrupted if the proteins mutated. As Sean Carroll at the University of Wisconsin explains, scientists had this "tacit assumption that if I show you a Hox protein from Drosophila and the same protein from, for example, a frog, they are the same protein."3 The results were surprising. Says Averof: "I expected that if there were any changes in Hox proteins, they would be more subtle." Carroll also was taken aback. "Like a lot of science," he says, "you have a bias and you do some experiments that you think are going to confirm that bias and you get surprised." Another eyebrow-raiser was that a single protein would have such a profound morphological change in the animal body plan, says UC San Diego's William McGinnis, the second study's senior author.4 This evidence, he says, provides new fuel for the evolutionist argument against creationism. Creationists maintain that "there is no evidence for how you would accomplish a macroevolutionary change. This is [one] that we can actually explain at the molecular level." Image: Courtesy of Craig Brunetti, Jayne Selegue, Steve Paddock, and Sean Carroll, HHMI, University of Wisconsin-Madison GETTING THEIR AIR LEGS: Limb development in the Buckeye butterfly (Junonia sp.). The top image depicts a ventral view of the larval stage (caterpillar) showing the appendages. The bottom image shows the corresponding ventral view of an immunofluorescently labeled butterfly embryo using a confocal microscope. The Distal-less gene (in green) is expressed in the appendages' distal regions, Cubitus interruptus (in red) and engrailed (in blue) are expressed in the embryo's anterior and posterior compartments. THE STUDIES Both teams relied on a technique known as ectopic expression, which involves producing strains of transgenic fruit flies by replacing the Drosophila Ubx gene with either a crustacean or onychoporan Ubx gene. As Carroll explains, ectopic expression allows scientists to ask questions such as: What happens when a crustacean protein is expressed in a fly? If the proteins behave the same, then presumably they are the same. But if they behave differently, then "this is formal evidence that the protein itself has changed," says McGinnis. If the proteins were the same, then the crustacean Ubx protein should repress limb development in the same way that Drosophila Ubx does. Drosophila was the testbed for a couple of reasons, according to Carroll: Most Ubx ectopic expression tools were developed specifically for use in Drosophila; no such technique exists for use in crustaceans. Using ectopic expression to compare insect and onychophoran Ubx, Carroll and coauthor Ron Galant, also from Wisconsin, identified the QA domain of the Ubx protein that has limb repression capabilities and is present in insects but absent in onychophorans. The occurrence of the QA domain in insects at least partially explains why they have fewer limbs than onychophorans. Using the same technique to compare insect (Drosophila) and crustacean Ubx, the McGinnis team identified a second repression domain that appears in both Drosophila and crustaceans. However, crustaceans also have an additional inhibitory domain that stops the activity of this second repression domain. McGinnis and his collaborators took their study one step further by actually mapping the inhibitory domain and identifying the handful of amino acid changes that account for its functional difference between crustaceans and insects. They surveyed several species of insects and arthropods and found a perfect correlation between the presence or absence of a series of serine and/or threonine phosphorylation sites and limb repression. "None of the insects had these specific amino acid residues," explains McGinnis, "but all of the multilimb crustaceans did." Thus, it appears that insects can re-press limb development not only by gaining a unique QA repression domain but also by losing a serine/ threonine inhibitory do-main. For some scientists, the fact that a Hox protein can be broken down into these different domains--each apparently with its own evolutionary trajectory--is the most exciting aspect of these two studies. "We're learning that proteins are in a sense modular," says Patel. "These studies really nicely illustrate the concept. They show multiple parts of the protein all contributing to phenotypic change." Still, the new evidence does not mean that changes in Ubx are the sole agents that suppress arthropod limbs. "There is no doubt that both protein evolution and changes in gene expression play a role in phenotypic change," says Patel. Averof agrees: "Morphological changes are still predominantly associated with changes in expression patterns. What this work tells us is that protein changes could also be playing a role. Evolution just takes advantage of whatever variation it has dealt to it." Photo: Courtesy of William McGuinnis William McGinnis (standing) and colleague Matt Ronshaugen BUT WHY SIX LEGS? The QA domain has been static and stable in insects for 300 million years, says Carroll, so the selective pressure to maintain it has presumably been very high. Put another way, this domain and the six-leggedness associated with it have been, for whatever reason, evolutionarily advantageous. But what's so great about six legs? The only evolutionary aspect to answer this question is that these data reveal "that changes in a protein can contribute to phenotypic evolution," Carroll warns, not why these changes occurred. Still, biologists can speculate. Perhaps the insect's six-legged body plan is better suited for a terrestrial lifestyle, Carroll says. McGinnis agrees. Insects, which he says are "the most successful landborne arthropods," probably do not need that many limbs on land. Resistance to air is much less than to water, so maybe having more than six legs would be an impediment. Terrestrial crustaceans, such as pillbugs, are notoriously slow movers. Nor do the new findings explain how Ubx has evolved. "I fully believe that these changes came about in ordinary microevolutionary fashion," says Carroll, that is, gradually and over a long period of time. However, based on these studies, he says, one can infer only what changed, not how it changed. But in a recently published study,5 Carroll and colleagues present evidence suggesting that Hox genes evolve gradually in multiple, intermediate steps, as opposed to sudden large ones. McGinnis agrees that Ubx evolution likely occurred over millions of years, one amino acid at a time. He and his research team are "trying to test this by looking at Ubx in insects with partial limbs on their abdomens." No matter why or how these changes in Ubx happened, there is no question that they did. Says Averof: "The demonstration ... will perhaps make people more optimistic of finding such changes in the future." Leslie Pray (lpray@nasw.org) is a freelance writer in Leverett, Mass. References 1. M. Levine, "How insects lose their limbs," Nature, 415:848-9, 2002. 2. A. Pavlopoulos, M. Averof, "Developmental evolution: Hox proteins ring the changes," Current Biology, 12:R291-3, April 16, 2002. 3. R. Galant, S.B. Carroll, "Evolution of a transcriptional repression domain in an insect Hox protein," Nature, 415:910-3, 2002. 4. M. Ronshaugen et al., "Hox protein mutation and macroevolution of the insect body plan," Nature, 415:914-7, 2002. 5. R. Galant et al., "Hox repression of a target gene: extradenticle-independent, additive action through multiple monomer binding sites," Development, 129:3115-26, July 2002.

Surprise, Surprise: Hox Proteins Have Evolved

Interested in reading more?

Become a Member of

Image: Courtesy of William McGinnis
LOTSALEGS: The six-limbed fruit fly, Drosophila melanogaster