• Open Access

Neural control of gene recruitment in metazoans



Gene recruitment played a critical role in metazoan evolution. Yet, there is no consensus on whether it is an accidental event or a result of an inherent “gene recruiting” mechanism. The prevailing opinion among biologists is that gene recruitment results from random changes in genes or their regulatory regions, but the supporting evidence is poor and controversial. Herein, I present a mechanism in which gene recruitment is a neurally determined event, an adaptive response to changes in environmental conditions. In support of the hypothesis, I present evidence on the manipulative expression of genes in the central nervous system, as well as neurally determined examples of gene recruitment in transgenerational developmental plasticity and in evolution of metazoans Developmental Dynamics 240:1–8, 2011. © 2010 Wiley-Liss, Inc.


Gene recruitment has been frequent and crucial in evolution of metazoans but our knowledge on causes behind it is still poor. Its mechanism(s) remains one of the great enigmas of the modern biology. Understanding the nature of gene recruitment is essential to understanding the nature of evolutionary change in general and we cannot claim we understand the gene recruitment as long as we do not know the mechanism that brings it about.

For the purpose of this article, which deals with gene recruitment in metazoans alone, a narrow definition of gene recruitment is used, such that suggests integration of nonhousekeeping genes and their products in new gene regulatory networks (GRNs), developmental pathways, or signal cascades, which lead to new or changed, usually adaptive, phenotypic characters. Certainly, this, like all other definitions of gene recruitment, is a descriptive, noncausal definition.

The fact that the newly recruited gene product functionally is usually unchanged shows that changes in genes are not necessary for gene recruitment. If recruitment of genes in new GRNs requires no changes in genes or in the function of their products, then for the recruitment of a gene to occur all that is needed is the availability of gene products (and their receptors) at the time and place where the GRN is activated.

Since, obviously, a mechanism of gene recruitment was operational in the course of evolution, we have good reason to believe that such an evolutionarily advantageous mechanism would have been favored by natural selection, hence must be still be at work in extant organisms. This uniformitarianist principle suggests that the mechanism we are eagerly looking for must be somewhere among the extant genetic and epigenetic mechanisms.


One of the fundamental questions arising before the students of gene recruitment now is: What is the nature of the gene recruitment? Is it a spontaneous, randomly occurring event, a still unidentified “mutation,” or is it a deterministic event, resulting from activation of particular “recruiting” mechanism(s) in response to particular evolutionary pressures?

The prevailing opinion is that gene recruitment is a stochastic event based on mutations in the nucleotide sequence of genes, gene duplications or random changes in the regulatory sequences (True and Carroll, 2002). It has been argued that, because changes in genes and the function of proteins they code for are relatively frequent, such random changes may bring about gene recruitment. This assumption is refuted by the fact that recruitment of most genes in the course of evolution involved no changes in the function of proteins they code for. Similarly, gene duplications, such as those that occurred with Hox genes, are “not at all sufficient to account for the continuous diversification of lineages” as it is proven by the relatively rare occurrence in the key developmental regulatory gene families and the extremely low rate of gene duplications of once per gene per 100 million years. Hence, it is believed that the less investigated changes in regulatory sequences may represent the primary source of gene recruitment (Carroll, 2005).

If gene recruitment would be a contingent phenomenon, it is to be expected that in the course of evolution it would generate predominantly “neutral” recruitments such that produce no adaptive phenotypic results. But the evidence available irresistibly shows that gene recruitment is adaptive, i.e., it leads to new or changed phenotypic outcomes and the evidence on “neutral” recruitments is minimal, if existent. Even if the existence of neutral recruitment of genes will be demonstrated in the future, it would be almost impossible to prove that the “neutrality” of recruitment in extant organisms is authentic and not a result of loss of their function in the course of evolution.

The second possibility of the cause of gene recruitment is that metazoans are endowed with capability for recruiting genes, for producing adaptive morphological, physiological, life history, and behavioral changes in response to the changing environment. It would essentially require an inherent potential of the organism to determine the spatial (where) and temporal (when) expression of its genes, suggesting manipulation of gene expression. On this possibility is based the hypothesis I present in this article. This hypothesis is inclusive of the role of genetic mechanisms and action of natural selection. Neurally determined recruitment of genes in evolution of metazoans is a process that takes place in time rather than an instantaneous “all-or-none” event and initially may affect individuals rather than whole populations (Cabej, 2008a). In all likelihood, initially the gene recruitments and the relevant new phenotypes may not have been as perfect as we see them presently. Over time, functionally favorable changes may accumulate in recruited genes and relevant GRNs, which may contribute to increased GRN efficiency and adaptive value of the corresponding phenotypic character. This suggests that the neurally determined gene recruitment also will be subjected to the action of natural selection in direction of increased survival rate of the favorably changed individuals and resulting increase of their proportion in the general population.

The crux of the problem with any hypothesis of directed recruitment of genes is that such hypotheses suggest a controversial epigenetic manipulation of spatiotemporal patterns of expression of genes. Hence, the first step in developing this hypothesis is to examine whether metazoans are capable of manipulating gene expression, i.e. achieve results that are different from those expected by classical modes of gene expression by using different, i.e. nongenetic, mechanisms of induction of gene expression.

Following the definition of “manipulation” as “shrewd or devious influence especially for one's own advantage” (The American Heritage Dictionary of the English Language, 2009), under manipulation of gene expression herein will be understood the ability of metazoans to express a gene/group of genes at the right time and cells in response to external/internal stimuli, which per se do not have access to genes and, even if they do, are incapable of inducing their expression.

Notwithstanding that any “manipulation” of gene expression may be intellectually irritating, there is compelling evidence that a mechanism of manipulative expression of nonhousekeeping genes for organism's “own advantage” is operational in extant metazoans from lower vertebrates to humans (Cabej, 2008b). And there is sufficient reason to believe that this mechanism of manipulative expression of genes has been operational in the course of metazoan evolution, at least since the Cambrian explosion, more than half a billion years ago.

Various external stimuli (visual, auditory, olfactory, tactile, etc., as well as changes in the temperature, photoperiod, moisture, social environment, etc.) and internal stimuli (variations in the level of hormones, growth factors, electrolytes, etc.) are known to trigger changes in the spatiotemporal patterns of expression of particular genes or groups of genes and resulting phenotypic changes in metazoans. These stimuli have no direct or natural causal relationship to the changes in gene expression, in the meaning that they per se, even in contact with relevant cells, are “unintelligible’ to genes, i.e., cannot induce their expression. The stimuli, which usually are received by sensory organs, become ‘intelligible” to particular genes only in the encoded form of electrical/chemical messages released by particular areas of the brain that result from their processing in specific neural circuits. The final output of the processing is a neuropeptide, neurotransmitter, or neuromodulator that triggers activation of a specific signal cascade, thus leading to a phenotypic result that normally adapts the organism to the change the stimulus represents.

Manipulation of gene expression is based on the processing of information received by the nervous system on various external and internal stimuli, which represents an indirect, “devious” route of expression of a gene, as opposed to the straightforward classical genetic model.

Let's briefly consider some paradigmatic examples of the manipulative expression of genes in the brain of vertebrates.


Male advertisement calls induce females of túngara frogs, Physalaemus pustulosus, to elevate estradiol levels (Lynch and Wilczynski, 2006) even though sound waves per se cannot induce expression of any gene. Females use an intricate way to achieve this result: they encode sound waves of the song in the auditory apparatus of the brain (Hoke et al., 2005) in the form of trains of electrical signals and send them for processing into a complex neurobiological maze with the final output leading to expression of the gene egr-1 and a modification of neuronal connections in the hypothalamus (Almli and Wilczynski, 2009) that by means of the hypothalamic-pituitary–ovarian axis leads to increased estradiol synthesis and secretion. The reason why only túngara frogs express the egr-1 gene in their brains, although other frog species have the same gene is because túngara frogs alone have evolved the capability to generate the chemical output necessary for expressing the egr-1 gene in relevant neurons in the brain by processing the species-specific male call in particular neural circuits.

Similarly, sound waves of male song in female zebra finches come in contact with cells of the outer ear but do not induce expression of any gene there. Only the output of the neural processing of the electrical signals into which the sound waves are converted in the female brain, results in expression of specific genes, Fos and ZENK (Bailey and Wade, 2003; Pinaud and Mello, 2007) as well as neurosteroids (estrogen; Remage-Healey et al., 2010) in neurons of particular brain nuclei even though they do not come in contact with the sound waves. Thus, in both the above cases, the processing of the song in the neural circuit creates a new, previously not existing, causal relationship between the song and specific genes, egr-1, Fos, Zenk, etc. All other birds have these same genes in their neurons but have not evolved an epigenetic mechanism for expression of these genes in response to male songs.

An example of an internal stimulus. Estradiol, in the presence of its receptor, represses GnRH (gonadotropin-releasing hormone) but GnRH via FSH (follicle stimulating hormone) stimulates secretion of estradiol. The reason for the “paradox” is that higher levels of estradiol are perceived in the brain as a stimulus that is processed in a hypothalamic neural circuit with resulting removal of the normal inhibitory influence of noradrenaline and the opioid regulation of GABA (γ-aminobutyric acid) neurons (Dobson et al., 2003; Petersen et al., 2003; Smith and Jennes, 2001), leading to secretion of GnRH by specific hypothalamic neurons. Again, secretion of FSH, which is genetically impossible (estradiol blocks the respective neurohormonal cascade) occurs because of an epigenetic mechanism (inhibition of noradrenaline and opioid regulation of GABA neurons).

While the above, and numerous similar examples, represent no authentic cases of gene recruitment they demonstrate the ability of the central nervous system to manipulate expression of genes, i.e. to adaptively express specific genes in response to specific stimuli by nonclassical genetic mechanisms. This is made possible by the capability of the central nervous system to interpret and explicate to non-nervous cells the meaning of the stimulus in the language of gene. Thus, the central nervous system (CNS) is capable of establishing new, previously nonexisting, causal relationships between various stimuli and practically any gene.

Having illustrated the manipulation of gene expression through processing of information on various stimuli in the brain, now let's present some examples of developmental transgenerational plasticity, which clearly involve gene recruitment, i.e., integration of genes into new developmental pathways. Let's remember that transgenerational developmental plasticity is an epitome of the evolutionary change since, essentially, both phenomena consist of the appearance of inherited changes in phenotype. Being reproducible under natural and laboratory conditions, cases of transgenerational developmental plasticity may give us important clues about developmental mechanisms that enabled recruitment of genes in the evolutionary past that are impossible to be directly traced.


The water flea, Daphnia magna (Strauss, 1820), is a cyclic parthenogenetic crustacean that normally gives birth to female offspring only but when environmental conditions deteriorate (crowding, depletion of food resources) or presage deterioration (shortening of the photoperiod) (Stelzer, 2008), it adaptively produces a sexually reproducing generation (male + female individuals). At the molecular level the transition is triggered by neurally controlled secretion of methyl farnesoate, Daphnia's juvenoid hormone. The synthesis and secretion of methyl farnesoate by the mandibular organ (Liu and Laufer, 1998) in Daphnia is negatively controlled by neuropeptides MO-IH-1 (mandibular organ-inhibiting hormone-1) and MOIH-2 synthesized in, and secreted by, the secretory neurons of the X organ/sinus gland complex.

Transduction in the brain of the crustacean of the environmental signals into instructions (=information) for switching to sexual reproduction involves activation of the cholinergic system in the brain (Eads et al., 2008) and causes a pulse of methyl farnesoate, which binds its specific receptor during a critical period of oocyte maturation (Rider et al.2005) (Fig. 1).

Figure 1.

Illustration of the recruitment of genes in two different pathways leading to birth of asexual (parthenogenetic) and sexual generations in Daphnia magna. The point of divergence between the two pathways starts in the crustacean's brain with the failure of the X-organ/sinus gland to synthesize and secrete neuropeptides MO-IH1, MO-IH2, and CHH (left cascade) and recruitment of the methyl farnesoate in the case of sexual generation-producing Daphnias and recruitment of the above genes and prevention of the synthesis of methyl farnesoate in the case of production of the asexual generation.

What takes place at a molecular level is that, under normal conditions, Daphnia's brain “recruits” genes for neurohormones MO-IH-1 and MO-IH2 in the developmental pathway, thus determining the parthenogenetic reproduction, but when conditions in the environment deteriorate, the brain suppresses expression of these genes, leading to production and recruitment of methyl farnesoate and, consequently, to production of a sexually reproducing generation, with male and female individuals.

The dramatic inherited phenotypic changes of transition from the asexual (females alone) to sexual (female + male individuals) generation occur in whole populations and within a generation, unambiguously rejecting the possibility of involvement of genetic mechanisms (changes in genes or DNA in general) in the phenomenon. As shown in the figure, production of the sexual generation depends on an epigenetic mechanism (recruitment of MOIH genes) in the brain of the water flea.

The flesh fly, Sarcophaga bullata, overwinters through pupal diapause characterized by low physiological activity which confers clear advantages on the insect under cold weather. Since the fly has to produce its diapausing eggs earlier, when the weather is still warm, it uses a winter-foreshadowing cue, the shortening of day length that presages the approach of the cold season. It was observed that whether pupae go through diapause or not depends not on the photoperiod experienced by larvae themselves but on the photoperiod experienced by their mothers, what logically led to the conclusion that the character is proximately triggered by a maternal factor deposited in eggs. The nature of the maternal factor suppressing diapause in S. bullata, is not known but Denlinger et al. (1995) found a unique transcript in the ovary of SP (short photoperiod) females on the first day of adult life.

Extracts of brains of mothers that had experienced short photoperiod prevented appearance of diapause and further experiments showed that the active principle of the extracts was the neurotransmitter GABA. Based on this experimental fact, it was concluded that the neurotransmitter was the inducer of the predicted diapause-inducing factor. When applied to mother flies, which normally would diapause, the neurotransmitter drastically reduced the frequency of the appearance of diapause (Webb and Denlinger, 1998). Investigators believe that “The information transfer from mother's brain (the site of photoperiodic response) to her ovaries occurs sometime after pupariation but before the second day of adult life” (Henrich and Denlinger, 1982).

Thus, the change in the life history (diapause skipping) of the offspring involved no genetic changes; an epigenetic “recruitment” of the neurotransmitter GABA in a new developmental pathway, leading to the deposition in eggs of the diapause-suppressing maternal cytoplasmic factor is responsible for the change.

In response to shortening of photoperiod or dropping temperatures in the environment, females of many insects produce diapause eggs, which arrest embryonic development at the germ band stage. Induction of production of diapausing eggs in the maternal organism starts with reception of the external stimuli by exteroceptors, their conversion into trains of electrical signals and their processing in neural circuits in the brain. In the silkworm, Bombyx mori, the processing inactivates a GABA-ergic inhibitory mechanism, thus allowing secretion of a neuropeptide, the diapause hormone (DH), the product of the DH-PBAN (DH-pheromone biosynthesis activating neuropeptide) gene, by a pair of labial secretory neurons in the midline of the subesophageal ganglion (Sato et al., 1994; Ichikawa, 2003; Shiga and Numata, 2007). Axons of these neurons reach corpora cardiaca, from where they release DH in the hemolymph (Ichikawa et al., 1995).

What occurs in diapause-egg producing mothers is not any change in genetic information, but an epigenetic removal of the GABA-ergic inhibitory mechanism that enables “recruitment” of the DH, into the developmental pathway of production of diapausing eggs. The first difference observed between the DH-recruiting and DH-nonrecruiting insects is a difference in the firing activity patterns of respective labial neurons: labial neurons in DH-recruiting silkworms were active throughout pupal-adult development, while in DH-nonrecruiting silkworms these neurons were inactive until the last quarter of pupal-adult development (Ichikawa, 2003). This suggests that the epigenetic information encoded in the continued electrical spike trains of labial neurons of DH-recruiting silkworms is responsible for inactivation of the GABA-ergic inhibitory mechanism and for recruitment of DH in the signal cascade for producing diapausing eggs.

The above, and many similar examples, show that during the transgenerational developmental plasticity external stimuli, per se, cannot induce expression of any gene. Only after reception of these stimuli (changes in temperature, change in the day length, etc.) by sensory neurons and their processing in neural circuits, the insect brain releases chemical signals that induce manipulative expression and recruitment of genes in new signal cascades leading to new adaptive characters. The fact that all these cases of inherited developmental plasticity occur in the whole population and within one generation clearly shows that no specific changes, incremental or sudden, in genes or DNA, are related to them.

Many cases of transgenerational developmental plasticity involve epigenetic changes in the activity of the central nervous system and start with hormonal signal cascades originating in the brain, demonstrating, among other things, that in metazoans “development comes under the control of the central nervous system” (Nijhout, 2003). However, in other cases of transgenerational developmental plasticity there is no visible involvement of the nervous system (West-Eberhard, 2003; Jablonka and Raz, 2009; Gilbert and Epel, 2009).

In view of the fact that transgenerational plasticity, as a phenomenon of the transmission of new characters to the offspring, is a special case of evolutionary change, the neural mechanisms of gene recruitment in transgenerational plasticity may be relevant to understanding the origin of evolutionary change.


The recruitment of a gene into a new developmental pathway requires the simultaneous restriction of the expression of the recruited gene only in the region where evolutionary change will take place. This is not an easy task for metazoans to accomplish because products of many genes (hormones, growth factors, secreted proteins, neuropeptides, etc.), by means of body fluids, circulate to practically every region and reach every cell of the body. And, most importantly, this spatial restriction has to evolve simultaneously with the integration of the gene in the new regulatory network, in order for the recruitment to succeed and not be eliminated by natural selection.

The spatial restriction of expression of recruited genes whose products circulate with body fluids is an unsolved problem from the point of view of the classical regulation of gene expression. An epigenetic model of determination of spatial patterns of expression of recruited genes is offered by the binary neural control of gene expression (Cabej, 2008c). It poses that, since mediators of the action of the circulating inducers of gene expression are specific receptors, it is the local innervation that by inducing expression of the above receptors determines whether, when, and where in the animal's body these genes will be recruited.

Here are some examples of the recruitment of genes that occurred in the course of evolution and indicate crucial involvement of the nervous system in the process.

Recruitment of Sex Pheromones in Salamanders

Salamanders of the plethodontid family use different ways of delivering their sex pheromone from the mental gland (located in their chin) to female nares. The pheromone consists mainly of SPF (sodefrin precursor-like factor) and PMF (plethodon modulatory factor), which were recruited as regulators of the male courtship behavior and female receptivity 50–100 million years ago (Palmer et al., 2007).

Pheromones are received by separate vomeronasal neurons, which convert them into electrical signals and transmit that information for processing in the olfactory centers, thus regulating female's mating behavior (Wirsig-Wiechmann et al., 2006). But males of the red-legged salamander, Plethodon shermani, approximately 27 million years ago recruited as pheromone another substance, PRF (plethodontid receptivity factor), to replace SPF (sodefrin precursor-like factor), although they have the gene for the latter. It is important to note that other modern plethodontids, such as Desmognathus ocoee (Nicholls) and Eurycea guttolineata, did not recruit PRF gene although they have it (Kiemnec-Tiburczy et al., 2009).

There is no direct evidence to answer the question: “What might have induced the evolutionary recruitment of the gene for PRF pheromone in P. shermani?”, but an examination of the mechanisms of expression of pheromones in the animal world may provide some relevant clues. Secretion of pheromones in all the cases studied so far depends on neural mechanisms. In invertebrates, e.g., secretion of pheromones is neurohumorally regulated by neuropeptides such as PBAN (Christensen et al., 1991; Abernathy et al., 1996; Teal et al., 1999; Groot et al., 2005; Nagalakshmi et al., 2007), whose pheromonotropic effect is mediated by a receptor that is produced in the form of three subtypes by alternative splicing (Kim et al., 2008).

Innervation of pheromone-producing organs is also sometime indispensable for secretion of pheromones (Teal et al., 1989, 1999; Christensen et al., 1991). Secretion of pheromones in invertebrates is also regulated by the nervous system in response to environmental cues such as photoperiod (Christensen et al., 1991) and even in response to social interactions and social stimuli (Krupp et al., 2008).

In vertebrates, especially in amphibians, including salamanders, pheromones are synthesized and secreted in response to visual stimuli, such as detection of predators in vicinity, i.e., by means of brain signals and neural pathways (Fraker et al., 2009) for there is no other way the visual stimuli can influence pheromone synthesis/secretion but through the brain. In the Japanese fire bellied newt, Cynops pyrrhogaster, the synthesis of the sex pheromone sodefrin is regulated by the pituitary prolactin (Rajchard, 2005; Kikuyama et al., 2005) and androgen hormones (Kikuyama et al., 2005), which are under direct and indirect control of specific hypothalamic neurons. In other amphibians secretion of sodefrin is centrally regulated by a neurohypophyseal hormone, arginine vasotocin (AVT) (Rajchard, 2005). In the peacock blenny, Salaria pavo, also pheromones seem to be regulated by the androgen 11-ketotestosterone (KT) (Serrano et al., 2008).

The ample evidence on the neural control and regulation of the synthesis and secretion of pheromones in extant animals strongly suggests that recruitment of the PRF gene in P. shermani in the course of evolution involved the nervous system or even has been neurally determined. If, as empirical observations have shown, recruitment of pheromone genes depends on neural mechanisms and structures in extant salamanders, there is reason to believe that these same neural mechanisms and structures may have been responsible for recruitment of these pheromones in the course of their evolution.

Pedomorphosis is the reaching of sexual maturity and reproduction during larval stage in originally metamorphosing species. This is the case with Ambystoma tigrinum species complex, which have lost the ability to metamorphose, thus evolving into facultative or obligatory pedomorphic species. Metamorphosis in salamanders is proximally induced by a surge in the level of the hormone thyroxine that results from neural activation of the hypothalamus–pituitary–thyroid cascade, which is timed by “hypothalamic maturation comprising neurons of several regulatory centers and culminating at the time of the secretory surge” (Rosenkilde and Ussing, 1996) and by “release of catecholamines at their synaptic contacts with cell bodies and dendrites of TRH (thyrotropin-releasing hormone) -containing neurons in the hypothalamus (Strand, 1999), triggering a complex signal cascade that, in a simplified form, looks as follows:

Catecholamines in the nonhypothalamic brain → hypothalamic TRH → pituitary TSH (thyroid-stimulating hormone) → TH (thyroid hormone) by the thyroid gland → TR (thyroid hormone receptor) and formation of complexes TH/TRremodeling of chromatinprimary gene expression response → secondary gene expression response.

Pedomorphic salamanders have all the genes involved in the metamorphic developmental pathway, but they fail to recruit one or more of the elements of the above neurally determined cascade that metamorphic salamanders do.

In favor of the manipulative neural determination of the evolution of pedomorphosis also speaks the fact that these salamanders may be experimentally induced to revert to metamorphic development by stressful conditions (e.g., captivity stress) that cause general disturbance in the central nervous system (Rosenkilde and Ussing, 1996).

Pax6 Gene

Recruitment of genes has been frequent in the evolution of lens crystallin. The eye anlage, i.e., the optic vesicle, is a forebrain neuroepithelial evagination. Its neurons in contact with the head ectoderm induce formation of the lens placode. Developing into optic cup with an inner layer of retinal neurons and innervated by the optic nerve, it engulfs the developing lens. Among other genes, Pax6, a highly conserved transcription factor, is of crucial importance in eye development, especially in lens development.

Recruitment of this gene in neurons of the optic vesicle is indispensable for the development of the lens, even when Pax6 expresses in the lens placode (Cantor-Soler and Adler, 2006; Adler and Cantor-Soler, 2007). Pax6 binding site is found in the regulatory regions of most vertebrate crystallin genes and appears to be involved in the recruitment of multiple lens crystallin genes. The neural control of recruitment of Pax6 in the regulatory network of lens crystallin during the early individual development strongly suggests that the nervous system must have also been responsible for the initial recruitment of lens crystallin in metazoan evolution.

The recruitment of the Pax6 gene during the individual development is an obvious manifestation of its evolutionary recruitment that occurred somewhere in the course of species' phylogeny. The developmental recruitment of a gene is the only way evolutionary gene recruitment can be perpetuated and not be lost over successive generations, just as the recruitment of a gene in the process of development is the only way its evolutionary recruitment can take place. The evolutionary and developmental recruitment of genes represents two sides of the same evo-devo coin.

The loss of eyes in the cave dwelling forms of the Mexican tetra (Astyanax mexicanus) also illustrates the role of the central nervous system as the ultimate source of the recruitment of the gene Pax6 in the lens GRN.

The eyeless form of A. mexicanus evolved rapidly, at least twice, within the last 10,000 years, in cave-dwelling populations of this species (Dowling et al., 2002; Jeffery, 2005). Leading investigators in the field have shown that no genetic changes are involved in the evolutionary loss of eyes in the cavefish, i.e., that not only Pax6 and crystallin genes are retained but “eye gene cascades are completely operational in cavefish embryos” (Jeffery, 2005). They also think that not only does the Modern Synthesis fail to account for the loss of eyes in the Mexican cavefish but other hypotheses (neutral mutation hypothesis and energy conservancy) similarly fail (Jeffery, 2005).

Experimental evidence shows that the loss of the lens and the vestigialization of eyes in the cavefish is controlled by signals emanating from outside the eye itself (Jeffery, 2005) and more precisely by neural signals as proven, among others, by the fact that lens formation occurs only in the presence of the retina and/or optic vesicle (Goss, 1969; Furuta and Hogan, 1998; Reza and Yasuda 2004a,b) that activates the lens GRN. Unilateral ablation of the prospective retinal region of the neural plate prevents formation of lens in the operated side of the head ectoderm (Kamachi et al., 1998).

What takes place in blind cavefish embryos is that the Pax6 expression domain in the neural plate/neural tube is restricted and reduced in size as a result of the expansion of the Shh expression domain. This restriction of the domain of the recruitment of Pax6 and the accompanying expansion of Shh expression in the neural plate/neural tube is the initiating element in the causal chain that leads to the inhibition of the recruitment of crystallin genes and development of lens in the cavefish form of A. mexicanus. That this change in Pax6 and Shh expression in the neutral plate/neural tube starts the causal chain of the loss of eyes is corroborated by experimental manipulation of these domains that mimics loss of eyes in embryos of the eyed surface form of the fish (Yamamoto et al., 2004).

Gene Recruitment by Neural Crest Cells

In vertebrates gene recruitment frequently has been determined by neural crest cells migrating from the neural tube to the sites where the new trait develops. For example, in experiments with the transplantation of the cranial neural tube between duck and quail neural crests, neural crest cells of the donor recruit genes that determine development of donor-specific cranial feather features in the recipient species (quail features in ducks and duck features in quails; Eames and Schneider, 2005, 2008).

These experiments have demonstrated that neural crest cells are provided with epigenetic patterning information before leaving the neural tube of the donor (Schneider and Helms, 2003; Trainor et al., 2002) and “the proper program of events governing the migration of crest may need first to be established in the hindbrain to allow migratory crest cells to interpret and respond to environmental signals set up through a series of tissue interactions.” (Trainor et al., 2002).

Studies on Darwin's finches have shown that the differences in the size and the shape of their beaks are related to the size of the regions where these neural crest cells recruit BMP4 (bone morphogenetic protein-4, which is responsible for the beak shape), BMP2 and BMP7 (for the size of the beak), as well as calmodulin (Abzhanov et al., 2006).

Birds lost their teeth approximately 60–80 million years ago. However, it is not the loss of genes but rather “changes in the relative position of a lateral signaling center over competent odontogenic mesenchyme” (Harris et al., 2006) that seem to be the cause of the loss. It is experimentally demonstrated that although birds have lost dentition their embryos may be induced to form tooth structures simply by homotopically and homochronically transplanting mouse neural tube of mesencephalon and rhombencephalon origin, but not from grafts of prosencephalon origin (Mitsiadis et al., 2003). Thus, particular regions of the neural tube provide the epigenetic information necessary for recruiting the odontogenic genes and the odontogenic GRN.

We might never be able to directly demonstrate how gene recruitment took place in any particular case in the course of metazoan evolution, but the representative evidence presented in this article represents weighty premises for a general inference on the nature and origin of the gene recruitment.

When we remember:

  • 1That the central nervous system is responsible for gene recruitments that bring about inherited phenotypic changes in cases of transgenerational developmental plasticity;
  • 2That the CNS is uniquely involved in the regulation of developmental pathways for changed/new characters in the examples presented in this article and numerous others;
  • 3That the developmental pathways leading to the appearance of changed/new characters start with chemical signals from the central nervous system;
  • 4That the failure of the CNS to produce the signal that triggers the relevant developmental pathway prevents the development of the new/changed character, although all the genes of the pathway are present and functional.

When all the above facts are taken into consideration, one cannot help but think that the central nervous system has been crucially involved in gene recruitment throughout metazoan evolution.