THE remarkable diversity of life on Earth stands as grand testimony to the creativity of evolution. Over the course of 500 million years, natural selection has fashioned wings for flight, fins for swimming, and legs for walking, and that’s just among the vertebrates. The capacity for evolutionary innovation – or, in buzzword form, “evolvability” – is built into the fabric of life.
Can evolvability itself evolve, with species becoming more or less evolvable over time? And if so, what factors affect evolvability and can they be studied in the natural world? Few questions are more fundamental to evolutionary theory, yet evolvability didn’t enter biological parlance until 1987, when arch-phrase-maker Richard Dawkins coined the term. In the intervening decades it has become a hot topic, though only recently has real-world evidence begun to put flesh on the bones of the theory.
Now there are studies aplenty to shed light on the factors that might constrain and enhance an organism’s capacity to evolve. They are also explaining crucial events in human evolution, such as the switch to walking on two legs and the emergence of our highly dexterous, tool-using hands.
Advertisement
One of the first hurdles had been to define exactly what “evolvability” means. The aim is to capture the capacity of a species or population to respond to natural selection. Since genetic variation is the raw material on which selection acts, the extent of this variation in a population provides a crude measure of evolvability.
When most researchers talk about evolvability, however, they mean something more subtle – not just how much genetic variation is present, but whether this variation translates into adaptive changes in the organism’s outward appearance and behaviour that could be shaped by natural selection. , a pioneer in the field who is based at Yale University, therefore defines evolvability as “the capacity to generate heritable phenotypic variation”. That is to say, variation in an organism’s body-type that can be passed from generation to generation.
The real question, of course, is what determines this capacity. Two factors are key. Perhaps the most fundamental of these is an organism’s “mutational robustness” – the capacity to develop normally despite the presence of genetic mutations. Since genes rarely act independently, a particular mutation could have a positive, negative or neutral effect on the organism, depending on the overall genetic background. Greater robustness could therefore be achieved by mechanisms that dampen the impact of mutations in any particular gene. In principle, that should increase an organism’s survival because it reduces the chance of potentially harmful changes to the organism’s body plan. But this buffering effect would also be the enemy of change, masking potentially beneficial variation and putting the brakes on the organism’s evolvability.
Or so it might seem. In fact, by neutralising the effects of otherwise harmful mutations, robustness preserves genetic variations that might otherwise be weeded out. That means organisms accumulate a wealth of hidden mutations within the population. Further genetic or environmental changes might then remove the buffering mechanisms and unmask the effects of these stored mutations, providing ready-made variation in the organism’s make-up. What might the mechanisms behind this robustness be?
“Environmental changes can unmask hidden mutations, providing ready-made variation when it’s needed”
The main players seem to be “heat-shock proteins”, according to studies led by at the Massachusetts Institute of Technology. HSPs ensure that other proteins always fold in the same stable three-dimensional shape, which is vital for their role within the cell. Under harsh conditions like scorching temperatures or high salinity, proteins can fold wrongly, preventing them from carrying out their function. It is here that the HSPs step in, acting as chaperones to guide the protein into the correct shape and allowing it to function properly even under troublesome circumstances.
Crucially, HSPs also ensure a protein folds into the same stable shape even in the face of genetic mutations that shuffle the protein’s sequence of amino acids. That permits hidden variation to build up over time without getting in the way of the protein’s everyday activities.
The structure and function of proteins govern all kinds of processes in the development of an organism. So when Lindquist’s team knocked out the HSPs in thale cress and fruit flies, the stored mutations suddenly became apparent in physical changes to the organisms, including new leaf shapes for the cress and changes to the shape of the flies’ eyeballs. The genes that code for HSPs don’t routinely stop functioning in natural populations but occasionally changes to the environment, like a radical change in diet, can overwhelm the HSP system and lead to similar effects, providing variation that can respond to the new evolutionary pressures precisely when it is needed most.
Lingering connections
HSPs are not the be-all and end-all of robustness. Some proteins are intrinsically more robust than others, even without the help of HSPs. This too can affect an organism’s evolution. In 2006, for example, and colleagues at the California Institute of Technology in Pasadena showed that more robust proteins can pick up useful additional functions from new mutations without losing their basic structure and collapsing into a useless tangle (). And in 2008, and colleagues at Yale University showed that viruses bred to produce more robust proteins adapted to new evolutionary pressures, like higher temperatures, more quickly than less robust strains (). In other words, they were more evolvable.
Robustness is only half of the story of evolvability, however. The second key factor concerns a phenomenon known as integration – the way different body parts or traits appear to vary and evolve together. Integration between traits often results from their shared evolutionary history. Body parts like limbs, teeth, ribs and vertebrae that are repeated along the body axis, for example, arose through the direct duplication of certain genes way back in evolutionary history. The two copies will not be completely independent of one another because the expression of both will ultimately be governed by the same regulatory genes at a different point in the genome, meaning the two body parts would still tend to vary and evolve together.
Integration is also likely if the different parts are involved in the same function. The four fingers and thumb, for example, work together for grasping and manipulation. To preserve the optimal use of the hand, changes in one part – a longer finger, say – need to be complemented by corresponding changes in the other digits. Selection therefore favours a developmental system in which genetic changes affecting digit length produce a coordinated shift in all digits.
Like robustness, integration can be a double-edged sword. On one hand, it increases the capacity to generate coordinated, adaptive changes in body structures, which surely help to increase an individual’s chances of survival. On the other, it also constrains the possible evolutionary avenues an animal could take – capping its evolvability, in other words – since a potentially beneficial change to one trait could have a disastrous impact on the other traits to which it is linked.
Luckily for life on Earth, integration is not an all-or-nothing affair. It is becoming clear that the traits can be integrated to varying degrees (see “Evolvable dogs”). Sometimes existing integration can be uncoupled altogether, making each trait an independent “module” that is more evolvable.
Consider the evolution of wings in mammals. The fore and hind-limbs of mice and other rodents are very tightly integrated, so that changes in one limb pair, such as an increase in length, correlate almost perfectly with changes in the other pair, says , a biologist at Washington University in St. Louis, Missouri, who studied the link.
Bats, however, use their modified fore-limbs for flying and their hind-limbs for grasping – two very different tasks, which would hint that the limbs are quite loosely integrated. Sure enough, of the University of Calgary in Canada, and Nathan Young of the University of California, San Francisco, have recently found that the co-variation in the length of bones in the bat’s fore and hind-limbs tends to be much lower than in other mammals. The implication is that the ancestors of bats must have lost the genetic integration between fore and hind-limbs somewhere along the evolutionary path, opening the door for the evolution of wings ().
A similar story could explain various stages of primate evolution, too. Campbell Rolian at the University of Calgary, for example, has compared quadrupedal primate species such as macaques, in which the hands and feet share similar roles, with species like the great apes (humans, chimpanzees, gorillas and orangutans), in which the hands and feet perform independent functions.
As expected, Rolian found greater integration between the hands and feet of quadrupeds than in the great apes (). Another analysis of the entire fore and hind-limb, rather than the individual hands and feet, found a similar result, with roughly 40 per cent less co-variation in the ape limbs compared with the quadrupeds ().
The result was that “arms and legs could respond to natural selection with a greater degree of independence, thereby increasing their evolvability”, says Hallgrimsson. Ultimately, this enabled early humans to evolve longer legs adapted to walking and running while leaving arm length relatively unchanged. By the same logic, a shortening of the forearm, which would have facilitated tool use, was not constrained by corresponding changes on the lower leg that might have reduced their walking power.
Importantly, integration comes in degrees. Even though the integration between the arms and legs has been significantly weakened over the years, some developmental connections linger that are strong enough to have changed the course of evolution in intriguing ways. For example, Rolian and Hallgrimsson, working with Daniel Lieberman of Harvard University, have recently discovered that evolutionary pressures on the feet might have prepared humankind’s highly dexterous hands for tool use and manual tasks. The team’s discovery came from a detailed comparison of the length of the bones making up each toe and corresponding finger (). They found that sufficient integration remain between the fingers and toes for them to have co-evolved to a certain degree.
Evolutionary pressures shaping the feet could therefore have changed the hands, or vice versa – but which way round? Using computer simulations to estimate the possible evolutionary pressures and corresponding changes to primate anatomy, the team suggests that natural selection acted primarily on the toes, enlarging the big toe and reducing the outer digits to stabilise the feet for walking.
“Evolutionary pressures on the feet might have prepared our hands for tool use and manual tasks”
The result, of course, was that as the big toe evolved, the thumb also grew accordingly. Completely by chance, that meant that the tips of the thumb and the fingers could now meet each other for the first time in evolutionary history, providing our ancestors with greater dexterity and precision gripping that became the key to the successful use of tools.
Evolvability through the ages
The take-home message is that animals are made up of a “nested hierarchy” of modules and integrated traits, says Hallgrimsson. So while the bones of the arms and hands show reduced integration in humans (and great apes generally) compared with quadrupedal monkeys, integration was nonetheless strong enough between human hands and feet to have profound evolutionary consequences. It is these specific patterns of integration and modularity, rather than either factor independently, that ultimately determines evolvability, he says.
Indeed, looking deeper into prehistory, it is easy to see how these factors could have played a crucial role throughout animal evolution. About 540 million years ago, the Cambrian explosion led to the emergence of the basic body plans of the 35 or so phyla of animals recognised today. Their common ancestor had not achieved a high level of integration or robustness, making it developmentally flexible and primed for evolutionary innovation. Evolution cashed in on that flexibility but soon pushed for greater developmental integration, more or less fixing the 35 body plans in the process.
That’s not to say evolvability has taken a nosedive since then. Although the developmental processes that produce the basic body plans of animals are too tightly integrated for fundamental change, a further drive towards greater modularity among parts of animals has increased their individual evolvability. It is this tinkering with bits and pieces of animal bodies, rather than radical re-invention of the body, that has fuelled the astonishing biological innovation among arthropods and vertebrates in particular.
Exactly what triggers the dissociation of integrated traits, prompting them to become increasingly modular, is still a burning issue. “This is the next big question,” says Wagner. It is possible that, in some cases, body parts become dissociated as a lucky accident that evolution then capitalises on. In the case of fore and hind-limb evolution, Wagner thinks it likely that selection for divergent functions in the arms and legs favoured any mechanism that could produce dissociation, but the issue is not settled. “Right now, we just don’t know why,” he says.
Getting a better handle on the process of dissociation will require a deeper understanding of the genetic mechanisms involved. Cheverud’s lab has already made steps in this direction by mapping genes that determine the integration among various characteristics in mice, though exactly how these genes do this remains unknown ().
These are early days in the empirical study of evolvability. Further progress will depend on researchers drawing together data on diverse aspects of biology, says , a biologist at the University of Manchester in the UK. A complete theory of integration, modularity and the developmental basis of evolvability will require connecting genetics and developmental biology with morphological studies in both experimental and natural settings. It will be a hard slog but, as Klingenberg says, “it’s where the fun starts”.
Evolvable dogs
The faces of man’s best friend come in an astonishing range – from the short, squashed face of a Pekingese to the slender snout of a collie. Studies by Abby Drake and Christian Klingenberg at the University of Manchester in the UK have shown that diversity in dog faces is comparable to the diversity apparent between all carnivore species.
But the remarkable variation in dogs has been achieved in just a few thousand years of selective breeding, and was possible thanks to a limited integration between the face and brain – a lack not normally found in other mammals. Intriguingly, this modularity is also found in wolves, coyotes and jackals, suggesting dog faces were always this evolvable – they simply needed the right evolutionary pressures to shape their snouts.
