The evolution of novelty through subtle tinkering
The genealogies of European royalty are genuinely terrifying in their complexity. No matter which history book you open, someone is always marrying some cousin of the third degree. To give you an idea: there are at least 4,010 different lines of descent from William the Conqueror to Prince William. Go figure.

Protein families might rival European royal families in their complexity. But where genealogists can easily trace back names and histories of Prince William’s ancestors back to the 8th century AD, biologists have no such luck. They don’t know how ancestral proteins looked like or how they functioned, which makes it difficult to understand how current protein families evolved from ancestral proteins. Just imagine what it would be like to reconstruct William the Conquerer’s invasion of England by studying his descendants of today. Genealogists sure are a lucky bunch.

Statue of William the Conqueror in his birth place, Falaise.

But biologists have found ways to resurrect ancient proteins. The research group of Joe Thornton is specialized in these kind of reconstructions. The team has reconstructed the ancestor of the glucocorticoid receptor before. They could follow its evolutionary path as it changed its specificity to a different hormone.*

This time Thornton and his colleagues decided to take on the family of nuclear receptors (NRs). NRs are proteins that can regulate the activity of genes when they are bound by a specific hormone or molecule. Many NRs, such as the estrogen receptor, have internal cavities into which a hormone like estrogen fits perfectly. When the estrogen , other parts of the receptor change shape. This change allows the receptor to bind DNA and recruit proteins that influence the activity of nearby genes.

While many NRs operate in this way, others seem to be stuck in the ‘ON’ position forever. They will bind DNA whether they are bound by a molecule or not. Some of them even lack a binding cavity entirely.

A nuclear receptor (the blue/green spaghetti) binds to DNA (pink). The activity of the receptor is influenced by several molecules (red/white blobs).

Nuclear receptors are activated by a wide variety of chemicals and regulate many different processes. Since they all belong to the same family, it’s clear that novel functions evolved many times in this family. But how did this happen? Did evolution simply tack on new functions, like a creative painter adding scenery and characters to his painting? Or was evolution more subtle and did it slowly modify an existing set of proteins? The ancestral NR holds the key to these questions.

To reconstruct the ancestral NR, Thornton’s group analyzed a whole slew of biochemical, genomic and structural evidence. They first searched the genome of the sponge Amphimedon Queenslandica for NR genes. As one of the earliest branching animals, sponges are in a great spot to shed some light on the evolutionary history of the NR family.

In contrast to the dozens of NRs that are found in vertebrates, the team found only two members of the NR family in sponges. The new family tree of NRs shows that these two sponge proteins are orthologs of two NRs that were already present in the ancestor of all animals. As you can see in the phylogeny below, the NR family remained small in sponges. In other animal lineages the NRs were duplicated many times.

The family history of nuclear receptors. Gene duplications are marked in green, gene deletions in red. The ancestral NR was already duplicated in the ancestor of all animals. Figure from reference

The many NR sequences were too divergent to reconstruct the sequence and structure of the ancestral NR with high confidence. The team circumvented this problem by reconstructing the features of the ancestral receptor instead.

The reconstructed receptor was an activator that already required a partnering molecule for its activation. The biochemical activity and protein structure of the two sponge NRs support this reconstruction, since they both were bound and activated by fatty acids. It’s likely that fatty acids were the molecules that bound the ancestral NR.

What about the other receptors – the ones that don’t need any molecule for their activation?

The handful of exceptions—ligand-independent activators and pure repressors—are scattered across the tree and are in most cases nested deep within groups of liganded activators, indicating that these states are almost certainly derived.

This outcome implies that novel functions in the NR family have evolved through subtle tinkering with the basic bauplan of nuclear receptors. Small modifications in the size and shape of the binding cavity are enough to change the preference of a NR to a different chemical, for example. The NRs stuck in the ‘ON’ position evolved through simple mutations that stabilize the active form of the protein. While some of these can still have their baseline activity regulated by molecules, others have become completely independent by losing the binding cavity altogether.

But is slow and steady the only way evolution beats? It’s clear that many new gene families (such as the nuclear receptors) evolved in the animal ancestor. Could it be that evolution operates in two gears, shifting from bursts of invention to periods of stable molecular tinkering, in a molecular variant of punctuated equilibrium?

Thornton thinks it’s likely that purifying selection will lead to retention of molecular innovations (‘molecular stasis’, if you will) once they have become incorporated into biological systems. However, he warns that punctuated stasis has not been proven to play a role in molecular evolution so far. To prove that evolution can shift gears, it would need to be shown that there were times in which new gene families were generated in unusual numbers and that physiological and developmental systems underwent notable reorganizations.

Thornton’s work has previously been misinterpreted by intelligent design enthusiasts as showing that evolution can only be responsible for tiny changes in existing systems. Thornton is not afraid that this will happen again, “Our experiments are driven by the desire to understand the nature of evolutionary processes and the historical causes of protein structure and function. But I enjoy showing how our work provides additional evidence against the claims of ID proponents.”, he said.

I can only agree with him: the work by Thornton’s group does not deflate the role of evolution in generating complexity and novelty. On the contrary, it is a testament to the power of evolution. Small modifications to the core functionality of nuclear receptors were enough to generate an incredible diversity of functionality. While they don’t reveal their story easily, the tales of past tinkering are alive and well in the descendants of the ancient receptors.

Bridgham JT, Eick GN, Larroux C, Deshpande K, Harms MJ, Gauthier ME, Ortlund EA, Degnan BM, & Thornton JW (2010). Protein evolution by molecular tinkering: diversification of the nuclear receptor superfamily from a ligand-dependent ancestor. PLoS biology, 8 (10) PMID: 20957188

Sources for pictures 1 and 2.

* For an excellent writeup of this work, see Carl Zimmer’s piece in the NY times.

You might also like:

    Unpopular Genes
    Vampire bats care little for sweet blood
    Resurrecting ancient apples and proteins

3 comments to The evolution of novelty through subtle tinkering

  • [...] The evolution of novelty through subtle tinkering [...]

  • Genealogies of proteins seem to be even worse than genealogies of bacteria! That was really well explained though.

    I had the idea in my head that one of the reasons punctuated evolution was observed was because steady random mutations would remain mostly ‘silent’ until one of those changes lead to a difference (i.e a frame shift or a previously off region being turned on) that suddenly lead to all the other mutations becoming suddenly more relevant leading to a (relatively) large shift in phenotype. I’m not sure how correct that idea is though.

    Gotta love the English monarchy though; they’re the German branch of a Dutch connection to the Scottish ancestors of a Welshman who could trace his way back via illegitimate children to a French bastard. :p

    • At least the proteins in this family follow a nice pattern of duplication and diversification. Things get really messy when fusion proteins and repeated gains/losses get involved. But again, it are the bacteria that make protein evolution this complicated!

      The mechanism you describe is similar to the mechanism of epistasis that Thornton’s group has investigated before and which I wrote on here (on a different study). So called silent mutations will build up in every protein. While they don’t change the function of the protein, they still constrain the future directions this protein’s evolution can take. They also make it almost impossible to re-evolve a previous function, since the silent mutations have become the foundation on which the new function has been built.

      When I think of a punctuated mechanism in protein evolution, I tend to think of processes like domain shuffling. In domain shuffling, functional protein domains get combined in new ways, as if evolution is playing with lego bricks. Some of the NR domains have clear analogs in non-metazoans for example, although they’ve not been combined in a single protein. A clear picture on the importance of steady versus inventive molecular evolution hasn’t emerged yet, but it’s definitely interested to see where things are headed :).

      And Dutch royalty is the same I guess! Our ‘founding father’ already came from the Elzas, between France and Germany. Things got more mixed up and complicated from there ;).

You must be logged in to post a comment.

Subscribe without commenting