We can gain insights into evolution by studying the sequence in which new features are acquired. But studying loss of features has its benefits, too. When a certain trait is lost multiple times in distinct groups of organisms, powerful statistical approaches can identify its genomic underpinnings. A study by Campagna et al. in Evolution1 sheds light on the genetic changes associated with a loss of flight in birds. They compare the whole genomes of 59 individual steamer ducks (of the genus Tachyeres) to examine loss of flight as it is evolving.
Steamer ducks occupy coastal habitats and lakes in southern Chile, southern Argentina and the Falkland Islands2. They show a distinctive escape behaviour called steaming — rapid, synchronized paddling of their wings and feet across water that mimics the action of their namesake, paddle-steaming boats (Fig. 1). Of the four recognized species, three (T. brachypterus, T. pteneres and T. leucocephalus) are characterized by their inability to fly2. Some heavier, male ducks of the usually flighted species, T. patachonicus, are also unable to fly, because their wing loading (the ratio of body weight to wing surface area) is higher than that of their lighter counterparts.
All steamer ducks also walk proficiently on land, and dive to feed and to escape predators. Unlike puffins and penguins, which use wing movements in foraging and feeding, they do not steam to acquire food. However, they do use their wings when diving underwater, and the flight muscles in flightless species are only slightly proportionally smaller relative to body mass than in steamer ducks that can fly2.
It has been debated whether the flightless species of steamer duck each independently lost the ability to fly or are all descended from a single flightless branch of ducks2. Resolving this debate would provide insight into some of the environmental or ecological factors that might promote flight loss.
Steamer ducks are an evolutionarily young group — estimated to be only about 2 million years old. Through their genome comparison, Campagna et al. show that the evolution of flightlessness in the two continental species, T. pteneres and T. leucocephalus, occurred early in the clade’s history and within a relatively short time frame. By contrast, T. patachonicus and the coastal T. brachypterus are more closely related (they diverged only recently), and indeed might still interbreed. Overall, the authors’ genome comparison suggests that flightlessness might have evolved independently on as many as three occasions, although there are alternative interpretations.
Campagna et al. also identified the parts of the genome that contain the highest number of differences in DNA sequence between flighted and flightless individuals, by mining the genomes for single nucleotide polymorphisms (SNPs): substitutions of single nucleotides at specific points in the DNA sequence. The authors correlated measurements of wing bones and bone proportions in the sequenced birds with the genome data, so that they could distinguish wing-shape-related genetic differences between individuals from those that were not relevant to wing shape or that had occurred by chance. Notably, some T. patachonicus and T. brachypterus ducks exhibited a mixture of both flight- and flightlessness-related versions of the genetic sequences linked to wing length. Thus, the evolution of flight loss seems to be caught in the act in steamer ducks.
Most of the SNPs that Campagna et al. found to be associated with differences in limb measurements occurred in or near a gene called DYRK1A. Thus, the authors suggest that changes in DYRK1A expression and function might contribute to the reduction in limb length relative to body weight that is observed in flightless individuals. They also note that mice that carry more copies of DYRK1A than normal show limb-skeleton differences3. Moreover, increases in the number of copies of DYRK1A have been implicated in certain symptoms of Down syndrome in humans, including differences in body size and the length of long bones, particularly those in the forelimbs4. Although Campagna et al. were unable to examine the number of copies of DYRK1A in Tachyeres, future work could examine the effects of observed genetic differences in bird development experimentally.
Flightless species are highly diverse, and flight loss has evolved in very different contexts. It has occurred after the acquisition or elaboration of an aquatic mode of locomotion, such as diving or steaming, and in largely terrestrial contexts in which there are few predators. As an example of the latter scenario, rails, which are relatives of cranes, have lost the ability to fly on nearly every oceanic island on which they have landed (and sometimes repeatedly on the same island5).
Regardless of the different contexts that might promote the loss of flight, in all cases of flight loss a reduction in the length of the wings relative to the rest of the body results in the wing loading becoming too high to allow flight. However, other changes in the wing musculature, skin and feathers, as well as the sensory systems and the rest of the skeleton, vary considerably among different flightless species, and it not always clear whether these changes are related to loss of flight or to other factors. For example, it is worth noting that the genetic and wing-shape changes associated with flight loss in steamer ducks are proposed to have occurred at the same time that these birds acquired steaming behaviour. Wings are typically relatively short in birds that use them to move through water. Thus, whether the genetic changes that affect wing shape are associated with the acquisition of steaming, or with the loss of flight, is difficult to determine.
The past few years have seen other substantial developments in research into the genetics of flight loss6,7. One study6 identified differences between the genomes of three flying species of cormorant and their flightless relative, Phalacrocorax harrisi. Many of these variations were in or around genes involved in the function of cell protrusions called cilia, which mediate cell signals required for skeletal development. However, the flight muscles and associated parts of the sternal bones of P. harrisi are much smaller than those of its flighted relatives (differences not observed between the flightless and flighted steamer ducks2).
Another study7 investigated a different basis for flight loss in ratites — a group of birds that includes the cassowary, ostrich and kiwi, and in which flight was lost multiple times in the deep past. Differences between flighted and flightless species were identified in regions of DNA that regulate the expression of genes involved in laying down the structure of the forelimb (but were distinct from the changes seen in the steamer ducks). Changes in the expression of several of these genes during development result in short forelimbs7.
The diverse mechanisms underlying flightlessness that have been identified in these genomic studies are not necessarily incompatible with each other. Indeed, an emerging perspective is that the genetic mechanisms that lead to changes in wing shape and length might be as diverse as the ecological contexts in which flight loss has occurred. Perhaps this is not surprising. Studies of digit reduction in mammals have shown similarly diverse mechanisms8,9, and different genetic mechanisms underlie adaptations to high altitude in closely related hummingbird species10. More work with museum collections11, and in developmental biology and anatomy, is needed to advance our understanding of the genetic changes that underpin traits such as flightlessness.
Nature 572, 182-184 (2019)