Sometime between 360 and 390 million years ago, a group of fishes made the move to life on land. Along the way, their fins gradually transformed into weight-supporting limbs with distinct elbow and wrist joints. Fins became legs. Swimmers became walkers.
Some aspects of this evolutionary transition are now clear. We know how the bones of those pioneering fish evolved into those that I’m now using to type, thanks to a number of beautiful fossils. We’re even starting to uncover the genetic changes that were involved in this makeover.
But there are still mysteries. How, for example, did the owners of the first legs control their limbs?
Hold your arm out and bend it at the elbow. Several things are happening: your biceps are contracting, but your triceps are also relaxing. This happens because of modules in your spine—groups of neurons that coordinate a large number of movements that go together. “A module might move your entire arm rather than just one pair of muscles, or coordinate the muscles in your arm and trunk,” explains Martha Bagnall from Northwestern University.
These preset programmes simplify the job of controlling our limbs. They’re like the .exe files that that can launch complicated computer programmes upon a quick double-click.
Fish don’t seem to have any obvious counterparts for these modules. Instead, they have much simpler neural circuits that separately control the left and right halves of their bodies. These circuits fire alternately, so that muscles on one side of the fish contract while those on the other relax. This makes their bodies undulate from side to side, propelling them through the water.
How do you get from this simple left-right set-up to the complicated modules that control your own limbs?  “We don’t really know,” says Bagnall. “There are reviews, but they are more or less hand-waving.”
Together with colleague David McLean, she has found an important clue, and one that contradicts the textbook view of fish swimming.
She started by recording the signals that arrive at the motor neurons in a zebrafish’s muscles, and tell them to fire.  If the textbook view of fish swimming is correct, you would expect these signals to arrive at the same time on either side of a swimming animal, so their muscles can contract as a synchronised team.
That’s not what happens. Instead, Bagnall found tiny differences in the timing of the signals reaching the top and bottom half of the fish’s muscles—just a few milliseconds, but differences nonetheless.
For example, the motor neurons on the fish’s top-right all get synchronised signals, as do those on its bottom-right… but these quadrants in sync with each other. What’s more, Bagnall and McLean found that they get signals from different sets of neurons. There are four distinct circuits, rather than two.
In normal swimming, the top and bottom circuits are mostly indistinguishable, but they come into their own when the fish needs to roll. If the zebrafish is lying on its right side, it sends stronger signals into the motor neurons controlling its bottom-left and top-right quadrants. The resulting muscular contractions help the fish to right itself.
“We think this is a good example of modular organisation in the fish spinal cord,” says Bagnall. “They’re controlling these four quadrants more or less separately, and that simplifies the job of a command that produces rolling and self-righting.”
She suggests that these simple modules could have been templates for the more complex ones that control our limbs. The top circuits could have evolved to control extensor muscles like our triceps, while the bottom ones evolved to control flexors like our biceps. “It’s definitely just a hypothesis at this stage,” says Bagnall. “But since there hasn’t been any plausible hypothesis before now, we’re pretty excited about it.”
“It’s a good story, and it’s beautifully laid out,” says Ole Kiehn from the Karolinska Institute, who studies how neurons produce movements. The team needs to work out several details, like how the separate top and bottom modules are coordinated in the fish, but their discovery is interesting in itself and their hypothesis is sound.
Martyn Goulding from the Salk Institute, who studies spinal circuits, is more circumspect. “Given the prominent role that limb locomotion played in the transition of vertebrates to a terrestrial environment, the issue is an extremely interesting one,” he says. “However, I am not sure how the modular organization that they see with respect to axial muscle control easily translates into a system for the control of flexors and extensors.”
For example, he notes that limb muscles come from a different group of embryonic cells to trunk and spine muscles. There’s a superficial parallel between the top and bottom halves of the trunk muscles, and the top (extensor) and bottom (flexor) muscles in the fin and limb, but these may not translate exactly.
Bagnall agrees that it’ll take a lot more work to check her hypothesis. She’s now trying to check if the neurons controlling the zebrafish’s top and bottom halves rely on different sets of genes. If they do, she’ll check the flexors and extensors of limbed animals to see if they share the same genetic signatures.