Many characteristics differentiate the chemistry of living systems from that of inorganic nature. Some of the best known include the homochirality of organic molecules within cells; maintenance of living systems away from thermodynamic equilibrium; and self-replication of key molecular components. During the last two decades, another differentiating characteristic – interactivity – has also become apparent.
Within a mixture of several thousand inorganic compounds one would not expect most of them to react or interact in any significant way with the others, but within a mixture of the several thousand organic molecules making up a living cell, all of the molecules interact in significant ways with many others. In fact, cells turn out to be extremely well-ordered chemical networks rather than randomised stew pots of protoplasmic compounds.
This recognition that living systems are composed of so-called ‘interactomes’ raises the fascinating problem of how such closed and ordered networks of interactions evolved from the random and rare interactions that characterise inorganic nature. This is a problem that I and my coworkers at Michigan State University, US, have been working on for many years.
The explanation for the evolution of interactivity seems to lie in molecular complementarity, which defines the stereospecific or shape-dependent non-covalent interaction between compounds. Such non-covalent binding can involve hydrogen, ionic, pi-pi overlap and charge transfer bonds, and van der Waals–London forces. Ephraim Katchalski of the Weizmann Institute in Israel demonstrated that the steric fit of the resulting molecular complexes can increase bonding energies by protecting the bonds from competition from solvents, seen for example in the phenomenon of water exclusion.
Well-known examples of molecular complementarity are the base pairing that permits DNA replication and transcription; enzyme–substrate and receptor–ligand specificity; receptor–second messenger protein interactions; the blood clotting cascade; and antibody–antigen recognition. These textbook examples illustrate the general assumption that molecular complementarity is limited to interactions involving macromolecules. Common wisdom has it that small molecules are too small to interact specifically or with enough affinity to make stable complexes. But if that is so, then from what precursors did each of these forms of macromolecular complementarity evolve?
My research has demonstrated that some small molecules can, in fact, bind to each other specifically and with sufficient affinity to be stable in solutions at ambient temperatures. For example, two molecules found in origins of life experiments, the amino acid glutamic acid and the dipeptide glycylglycine, bind to each other through four hydrogen bonds, stabilising each other against ultraviolet light (UV) degradation. Since the primordial Earth lacked an oxygen atmosphere, UV radiation levels would have been much higher than today, playing a key selective role in chemical evolution. The ability to survive high UV levels would have helped determine which chemicals contributed to the evolution of living systems and which were destroyed.
Similarly, ascorbic acid (vitamin C) binds to catecholamines, such as epinephrine (adrenalin), by forming a complex that acts as a pH buffer, retards oxidation of both molecules and protects the components against UV degradation. All of these characteristics could have been important during the origins of life and play roles in modern physiological systems: for example, ascorbate is a cofactor for the enzymes that produce catecholamines.
A similar complex between flavins, such as the B vitamin riboflavin, and indoles, like the amino acid tryptophan and the neurotransmitter serotonin, has been known since the 1930s. These complexes, like ascorbate–catecholamine complexes, also act as pH buffers and protect the constituents against oxidation, but, unlike ascorbate-catechol combinations, they are chemically reactive with each other. Evolution seems to have adapted this reactivity to the enzymes that metabolise indoleamines, which require flavin cofactors.
Such complexes are typical of a wide range of small molecule complementarities and illustrate the roles complementarity can play in evolution. On the one hand, complementarity can stabilise and protect molecules so that they are more likely to survive to participate in evolutionary processes. On the other, it can promote chemical reactivity, acting as a catalytic platform to encourage reactions that would otherwise be unlikely to occur. Both functions alter the distribution of compounds available in a chemical system by privileging those that form complexes.
My theory is that molecular complementarity was one of the key selective forces operating on evolution from the outset. Molecules were incorporated into evolutionary schema not as individual entities, but as sets of complementary modules. So what could evolution build with such complementary modules? The answer appears to be just about every form of macromolecular complementarity that we observe in living systems today.
Consider, as a case in point, the evolution of receptors and their ligands.
There are two possible evolutionary scenarios for receptor evolution. One is that receptors and ligands evolved separately and were selected at each stage of their evolution for properties independent of their ultimate joint function. Evolution simply assays every possible combination of receptor-like proteins with every possible ligand until something works. This is an incredibly wasteful way to evolve a system. Since a typical receptor is about 400 amino acids in length, evolution might have to generate an astronomical number of proteins (20400) to find one that functions as a receptor for any given ligand!
Evolving an enzyme–substrate pair or any other macromolecularly complementary system would involve sorting through the same huge numbers of permutations.
An alternative scenario is that receptors and ligands co-evolved and were selected at each stage of evolution for their joint functionality. Donard Dwyer, a neuroscientist at Louisiana State University, US, was the first to propose such a co-evolutionary scenario. He noticed that some receptors for peptide ligands have sequences in the receptor binding site that mimic the ligand. Moreover, these peptide ligands self-aggregate. That is to say, they are self-complementary. This pair of observations led Dwyer to propose that peptide receptors evolved from self-complementary peptides. First, the gene encoding the peptide is duplicated; a common evolutionary occurrence. One member of the gene pair is conserved and continues to produce the peptide ligand. The other becomes the basis for the receptor.
Dwyer’s insight led me to realise that receptors have three basic components, each of which can be evolved as separate modules. One is the extracellular receptor sequence that Dwyer describes, which is selected for its self-protective properties. The second is a region that is selected for its ability to span and stabilise a lipid membrane. The third is an intracellular sequence selected for its ability to act as part of catalytic system. By linking the genes encoding these three components, a prototype receptor is formed. This prototype receptor will be selected for a novel, emergent property never seen in inorganic nature: it converts an action on one side of a membrane – binding to the receptor – into a different action on the other side of the membrane: a chemical reaction. Such unidirectional transduction of information is another unique aspect of living systems.
The scenario just described is somewhat simplified, because a random evolutionary process would mix and match all the different elements of such a protoreceptor in every way possible, as well as mix these elements with modular elements from other systems. Most of the results would not yield functional proteins and the systems would not survive. But selecting out a handful of peptides that could act as receptor sequences (perhaps 103 out of 2010 possibilities, assuming a peptide 10 amino acids in length), a handful of transmembrane spanners (another 103 out of 2010 possibilities), and a handful of chemical reactors (another 103 out of 2010 possibilities) and then generating all their possible permutations (103 x 103 x 103 = 109) by means of the mixing and matching of all the possible complementary modules is far more efficient than exploring the 20400 possible receptor sequences that a purely random approach to receptor evolution would entail.
Interactivity was present from the start of chemical evolution. In consequence, there is a ‘molecular palaeontology’ that reveals the complementary modularity within evolution that is just as revealing of our origins as are geological palaeontology and the evolutionary history revealed in our genes.
Again, an example is illustrative. Besides binding to catecholamines, we discovered that ascorbate also binds to short peptides. Some of these peptides are similar to glutathione, a tripeptide found in all cells that mops up reactive oxygen species and recycles oxidised ascorbate back into ascorbate. Glutathione and ascorbate are important components of the cellular antioxidant system. The ascorbate-binding peptides that we have identified are found not inside cells, but on their surfaces. One set is found in the cores of ascorbate transport proteins and probably serve as ‘handles’ by which ascorbate is moved inside cells. The other set is found on catecholamine receptors where they enhance binding and activity. In other words, evolution appears to have adapted direct ascorbate–catecholamine binding into receptor-mediated ascorbate-catecholamine co-binding.
Natural chemical selection
We have also discovered that catecholamines bind to opiate drugs, such as morphine, and to opioid peptides, such as enkephalins and endorphins, the body’s natural opiates. The nature of molecular complementarity is such that since opiates, opioids and ascorbate all bind to catecholamines, opiates and opioids also bind to the same peptides to which ascorbate binds. Thus, we have discovered that opiates and opioids, like ascorbate, enhance binding and activity of catecholamines for their receptors.
Moreover, these complementarities in binding and in function are carried over to the opiate and opioid receptors. Opiate and opioid receptors turn out to have catecholamine binding sites that enhance opiate and opioid binding and activity. So once again, evolution appears to have adapted catecholamine–opioid binding into receptor-mediated opiate-catecholamine co-binding.
In short, when we look at modern physiological control systems, their molecular palaeontology reveals a basic set of small molecule modules used and re-used in multiple permutations. Evolution has used natural chemical selection criteria to winnow out unstable and unreactive compounds to leave those that can bind to each other to form stable complexes or to carry out novel chemical reactions.
By limiting itself to using and re-using such complementary modules, evolution has not only drastically limited the permutations it must explore to generate functional systems, but also ensured that those systems will be composed, from the very outset, of massively interconnected and interactive subunits; hence the origins of one of the most characteristic properties of living systems, their interactomes.
Robert Root-Bernstein is professor of physiology at Michigan State University, East Lansing, Michigan, US.