We have just seen a remarkable example of convergent evolution at the molecular level. Two membrane proteins independently evolved to carry out the same function. The two proteins have the same pattern of protein domains, but in reverse. Remarkable sequence similarity also exists, but again, seen only in reverse.
So how did this occur? We can think of this example of convergent evolution as the unfolding of the preadapted state. This original preadapted state channeled the blind watchmaker, much as a seeing eye dog can lead a blind man. And if it wasn’t for the lucky fact that the proteins are in reverse, an example of convergent evolution would be scored by everyone as an example of common descent.
So what was the preadapted state? Two key parts appear to be the original bacterial porin (which evolved to become Tom40) and the MTS-like sequences that exist on one out of every twenty bacterial proteins. Since the MTS sequence interacts with Tom40, the bacterial design was already poised to facilitate the endosymbiotic union. This interaction would set up a selection pressure that would be guided by the architecture and composition of the bacterial porin and the MTS. In essence, the combined demands of the porin and the MTS would function as bait to fish Tom20 out of the bacterial tool box of protein domains. Inside this toolbox was the TPR domain (a “bolt”) which seemed to fit the MTS nicely. From there, we simply attach a membrane spanning region, which could be picked up from many other bacterial proteins through duplication and recombination. In this case, natural selection did not stumble upon some solution, any solution, that just happened to work. No, no. It hit the same target – twice. And it did so roughly about the same time: after plants and animals/fungi split apart, but before animals and fungi split apart, and before green algae and plants split apart.
This design logic was laid out and explained in The Design Matrix. After surveying test tube experiments where scientists generated ATP-binding proteins from random peptides, I observed:
The whole experiment is an intelligent use of chance. First, you fish out the proteins that weakly express a function you are trying to find, which is easily accomplished by using the function itself as bait in the pool, cleansing away all the other sequences that do not meet your criteria. Once the candidates are isolated, you start over with them, but this time, the bait is more complex, as it is not merely the function, but also includes the sequence of the binder. The mutation steps that followed were built around a strategy that kept most of the identified sequence constant while tweaking on its periphery. The result was the isolation of a protein with improved function.
Life’s designer may have also made intelligent use of chance. Only in this case, the “bait” was not a simple molecule like ATP, nor a single complex of ATP and a protein. Instead, the bait could have been the entire cell, or set of heterogeneous cells. What the blind watchmaker could subsequently find was then constrained by the carefully chosen initial conditions. Just as the researchers, as artificial selectors, set up their in vitro selection experiment such that it was rigged to find ATP-binding proteins, so too may life’s initial conditions have been rigged by the design of the cell’s architecture and the choice of which components to employ. In such a case, this chosen state would then act as the surrogate for the artificial selector.
Tom20 represents a decent candidate for such design.