Hormone Evolution

Hormones cannot bind to DNA on their own but must first be bound to a receptor. How this linkage evolved is not clear. Did every mutation in a receptor have to weather the blows of natural selection until a comparable mutation arose in its ligand? This seems unlikely, but complex organisms depend extensively on such coevolution, the mechanisms of which in general are not well understood. This is another element of the "problems" of specificity, cross-reactivity, and coevolution that face systems that have multiple components evolving by modularity and duplication. However, some aspects of hormones and their receptors may help elucidate this.

A number of "orphan" nuclear receptors have been identified (by their shared amino acid sequence similarity with receptors for which ligands are known) for which a ligand has not yet been found and may not exist. Indeed, some receptors (i.e., TFs encoded by genes in a receptor gene family and having the prototypical receptor structure) seem to be able to function as TFs independently of ligand. Two possible evolutionary scenarios have been proposed: either the ancestral nuclear receptors were orphan and bound to DNA as homodimers, some of them later gaining ligand-binding capability independently (Laudet 1997), or the ancestors were liganded and some of the receptors independently lost their ligand-binding capability (Escriva et al. 2000). This supposes that the ligands—hormones—existed before ligand-receptor binding became part of the working cellular repertoire, and in fact, as we have seen, plants make steroid hormones, as do fungi, so the ancestral molecule was clearly ancient.

The affinity with which a ligand binds to its receptor varies, and indeed, some hormones bind to more than one receptor and some receptors bind more than one ligand. This kind of cross-reactivity might in principle "confuse" the organism, and there is certainly variation in nature that could reflect this (e.g., deviations from stereotypy in morphological or behavioral sexual dimorphism among individuals), but probably also provides a measure of protection via redundancy.

Finally, all of the nuclear receptors that have been identified in cnidarians are unli-ganded, suggesting that this is the ancestral state (Grasso et al. 2001). The most parsimonious explanation of the evolution of hormones and receptors is perhaps that an ancestral receptor used in some other way evolved characteristics that also would bind a ligand and, after a gene duplication event and subsequent mutation, each receptor acquired the capability of binding to different ligands. This could have involved a period of cross-reactivity (which if too harmful would have been removed by selection), followed by divergence of function. We are largely handwaving here, however, because this does not explain why receptors in different subfamilies now bind ligands with no structural homology (that is, they are not themselves related to each other), how photoreceptors came to respond to light rather than chemical ligands, whether protohormones were TFs that alone could bind DNA, and why they may have lost that ability, among other provocative questions.

Perhaps the tightly paired ligands and receptors we observe and catalog today are those for which appropriate cross-reaction or coevolution happen to have occurred, whereas other signals evolved independently of ligand (e.g., some of the orphan receptors, which in fact are the majority of nuclear receptors) or ligands evolved independently of receptors (nitric oxide is an example) or have disappeared without them. The exchangeability of regulatory mechanisms means that in principle any one of many possible receptor-ligand pairs could carry out a given regulatory function and what we see today may just be those that happened to have been used. At the same time, similar systems and the use of gene family members in different species for the same job, or within species for related jobs, clearly suggests that coevolution has been important.

In this context it is interesting to ask why a truly unliganded receptor would maintain a viable ligand binding domain (that is, why would mutation not have erased the organized nature of that domain, given no selective pressure maintaining its ligand binding structure?). This could, of course, simply reflect limits on our current knowledge; ligands continue to be found for "orphan" receptors (Gustafsson 1999), thus causing the receptor to be termed "adopted" (Chawla et al. 2001). It is now recognized that some adopted receptors play a role in lipid sensing, forming heterodimers with retinoid X receptors, with low-affinity dietary lipids as ligands (Chawla, Repa et al. 2001). Also, as research goes, a gene may be long studied in the context in which it was originally found, and only later discovered to have entirely different functions. Such discoveries often tip off distant homologies among the functions, but this need not be the case.

However, orphan receptors commonly bind DNA as monomers rather than the heterodimers that steroids bind as, suggesting that this class of receptor acts in a way that is different from other receptors, whether by being truly unliganded or something else not yet understood. If any of these receptors truly have no necessary ligand, the fact that they still share the structure of liganded receptors may represent their occasional ability to cross-react in ways that have enough importance to be supported by selection, such as some undocumented specialization. On close inspection, history suggests that genes in the same family may not be as completely redundant as some experiments initially suggest.


Steroid hormones seem to have been highly conserved until the divergence of jawed from jawless vertebrates some 450mya. Figure 10-1 shows the structure of a number of steroid hormones. The structural diversity is produced by variation in the synthetic pathways and hence the genes that encode the synthesis of these molecules. Genes for nuclear hormone receptors form a large gene superfamily.


Peptide hormones are typically very short strings of amino acids. Comparative sequencing suggests that after the divergence of jawed vertebrates peptide hormone genes underwent periods of sustained and rapid bursts of change, although there seems to be significant variation in the basal evolutionary rates of these genes (Liu et al. 2001; Wallis 2000). As has been suggested by studies of other genes in the vertebrate lineage, genes for protein hormones show some evidence of two whole-genome or at least large chromosomal duplication events early in the evolution of vertebrates; for example, when there is a single gene for a given hormone in Drosophila, four paralogous genes are found in vertebrates, creating large gene families for many hormones (Ohno 2001), and some teleost (bony) fish are tetraploid. Nonprimate mammals and the bush baby, for example, have a single-copy growth hormone gene, whereas humans and the rhesus monkey have five (Liu, Makova et al. 2001).

Coevolution at Higher Levels

Genes related to life history stages, such as different types of metamorphosis in various animal groups, must experience various forms of coevolution with other

A. Adrenal Steroid Hormones

B. Gonadal Steroid Hormones

Was this article helpful?

0 0

Post a comment